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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2e34b8bd-8243-4c1f-a281-e5464e3fda7c | authorship-detection-of-sms-messages-using | 1403.1314 | null | http://arxiv.org/abs/1403.1314v1 | http://arxiv.org/pdf/1403.1314v1.pdf | Authorship detection of SMS messages using unigrams | SMS messaging is a popular media of communication. Because of its popularity
and privacy, it could be used for many illegal purposes. Additionally, since
they are part of the day to day life, SMSes can be used as evidence for many
legal disputes. Since a cellular phone might be accessible to people close to
the owner, it is important to establish the fact that the sender of the message
is indeed the owner of the phone. For this purpose, the straight forward
solutions seem to be the use of popular stylometric methods. However, in
comparison with the data used for stylometry in the literature, SMSes have
unusual characteristics making it hard or impossible to apply these methods in
a conventional way. Our target is to come up with a method of authorship
detection of SMS messages that could still give a usable accuracy. We argue
that, considering the methods of author attribution, the best method that could
be applied to SMS messages is an n-gram method. To prove our point, we checked
two different methods of distribution comparison with varying number of
training and testing data. We specifically try to compare how well our
algorithms work under less amount of testing data and large number of candidate
authors (which we believe to be the real world scenario) against controlled
tests with less number of authors and selected SMSes with large number of
words. To counter the lack of information in an SMS message, we propose the
method of stacking together few SMSes. | ['P. Herath', 'R. G. Ragel', 'U. Senanayake'] | 2014-03-06 | null | null | null | null | ['author-attribution'] | ['natural-language-processing'] | [ 5.11132367e-03 1.20645680e-01 -1.29650896e-02 -4.07587029e-02
-1.69269979e-01 -6.49474084e-01 1.06957078e+00 4.14880484e-01
-5.79797685e-01 1.12243676e+00 4.79713343e-02 -6.66015744e-01
-1.28143542e-02 -8.21991384e-01 -2.95487732e-01 -5.21535814e-01
2.52491504e-01 8.13517928e-01 3.98021668e-01 -3.73652309e-01
8.40622842e-01 7.30246782e-01 -1.47175694e+00 8.14424977e-02
8.09498966e-01 5.97088993e-01 -2.07306203e-02 6.52568340e-01
-3.84908974e-01 5.42531431e-01 -1.15389204e+00 -7.77586401e-01
1.04798034e-01 -4.91546482e-01 -8.63300920e-01 1.34411782e-01
2.14531749e-01 -1.39600635e-01 7.38356411e-02 8.99892271e-01
4.38584983e-01 -8.54423195e-02 7.84030259e-01 -1.16539598e+00
-7.27878958e-02 5.44626296e-01 -5.56005299e-01 2.94177264e-01
7.57920682e-01 -1.86319441e-01 6.66802406e-01 -3.12840044e-01
7.16098666e-01 9.99788404e-01 4.54040945e-01 1.83711678e-01
-9.98398006e-01 -6.45693600e-01 -2.93978393e-01 -7.20154215e-03
-1.19054353e+00 -4.01751548e-01 5.98670125e-01 -4.52808887e-01
1.19353466e-01 5.16593218e-01 5.11279643e-01 1.26535630e+00
-2.68813483e-02 3.37872535e-01 1.70302606e+00 -6.45620286e-01
1.07905947e-01 7.97956288e-01 6.07737228e-02 3.42348069e-01
6.31122291e-01 -3.93802911e-01 -3.25383961e-01 -5.31623781e-01
5.20589411e-01 -2.96054006e-01 -2.17884824e-01 1.51335627e-01
-1.10697699e+00 8.35144043e-01 -2.84909248e-01 8.79038393e-01
-2.98198700e-01 -2.92238176e-01 4.65548486e-01 3.48572075e-01
3.58942419e-01 6.37570441e-01 -2.44160593e-01 -6.45599782e-01
-1.16485870e+00 2.05037698e-01 1.10244226e+00 4.30407733e-01
2.71251559e-01 -3.72376829e-01 2.97893554e-01 7.73665071e-01
1.40754521e-01 3.60143960e-01 7.11936831e-01 -4.24609095e-01
3.91018718e-01 7.57360399e-01 2.48833686e-01 -1.35803366e+00
-8.48875418e-02 -4.47444677e-01 -6.44730508e-01 3.64377871e-02
1.07521713e+00 -3.07997465e-02 -2.52863050e-01 1.28872395e+00
1.61732078e-01 1.14062419e-02 -1.84770644e-01 6.13310039e-01
4.57764208e-01 4.47244525e-01 -3.73586774e-01 -4.12566155e-01
1.41401887e+00 -1.30196497e-01 -6.55182123e-01 2.64292002e-01
6.45866156e-01 -1.25299835e+00 9.35871720e-01 5.69726944e-01
-5.31093061e-01 -1.80627599e-01 -7.11343706e-01 4.58095968e-01
-6.51869178e-01 1.73931316e-01 3.53283554e-01 1.28735602e+00
-6.48814321e-01 7.84949243e-01 -3.06219399e-01 -7.77064979e-01
8.46829265e-02 4.16329533e-01 -4.29219961e-01 3.90196741e-01
-1.11891568e+00 9.97371793e-01 2.49316841e-01 -7.97616318e-02
7.84112737e-02 1.58709109e-01 -2.41445407e-01 -2.12813154e-01
1.57576665e-01 -1.67132720e-01 9.38702643e-01 -1.07353640e+00
-1.40201437e+00 1.08759165e+00 -2.17413649e-01 -3.53150427e-01
9.12600398e-01 1.32136613e-01 -4.52169180e-01 -2.15580370e-02
1.17098711e-01 1.08472660e-01 7.81751275e-01 -1.27933562e+00
-5.59091508e-01 -1.61162764e-01 -1.27988579e-02 -3.98433983e-01
-3.78444672e-01 2.66900063e-01 -8.84811860e-03 -5.12564719e-01
3.80351990e-02 -1.01674163e+00 3.18599164e-01 -7.11687863e-01
-4.52092290e-01 -1.53602302e-01 7.95019388e-01 -9.88470972e-01
1.39146960e+00 -1.73778200e+00 -3.28108400e-01 6.22913361e-01
9.74578336e-02 4.04796362e-01 4.50202852e-01 8.16682637e-01
1.74174502e-01 4.19385642e-01 2.64007803e-02 -3.40897709e-01
3.58659811e-02 1.70603301e-02 -3.41079921e-01 6.39961183e-01
-1.75691038e-01 5.03736854e-01 -7.07178056e-01 -7.75246143e-01
6.53435439e-02 1.77604705e-01 1.49327153e-02 -1.54320952e-02
2.60297090e-01 5.65596879e-01 -3.58707488e-01 5.75489104e-01
6.10628009e-01 1.73334479e-02 1.90394923e-01 2.56731123e-01
-9.08954144e-02 3.88700515e-01 -1.53681731e+00 7.34710276e-01
-3.44035476e-01 9.51104224e-01 -1.91699326e-01 -7.98575044e-01
1.08938920e+00 3.10439438e-01 2.01846763e-01 -5.82811952e-01
4.88313377e-01 7.61437416e-01 2.35210598e-01 -3.97366792e-01
7.56483316e-01 -9.97098163e-02 -1.83862541e-02 5.29797435e-01
-4.29276854e-01 4.84443493e-02 5.10606945e-01 1.24557294e-01
8.94042253e-01 -1.90598696e-01 4.51922894e-01 -1.35962889e-01
7.20713258e-01 -3.46189588e-01 1.20798327e-01 5.38120270e-01
-5.47547229e-02 4.13020432e-01 8.64492238e-01 -1.90235376e-01
-1.15257347e+00 -5.66716313e-01 -2.26973101e-01 5.73258698e-01
-1.48248896e-01 -1.94423348e-01 -8.02507043e-01 -6.80631340e-01
-1.53511941e-01 5.34172475e-01 -3.00953090e-01 4.97372299e-01
-4.31053162e-01 -8.28846097e-01 7.70871222e-01 -2.69553542e-01
4.14147139e-01 -8.57485473e-01 -6.97713554e-01 1.19058482e-01
-2.16903389e-01 -1.00440753e+00 -5.07421121e-02 -1.19163297e-01
-8.06940138e-01 -1.37988949e+00 -9.18500006e-01 -3.27852130e-01
6.34487748e-01 7.65743256e-02 8.00973952e-01 4.91560131e-01
1.84870780e-01 -3.21099795e-02 -6.04677737e-01 -2.75490582e-01
-9.82048035e-01 3.55896235e-01 1.07321367e-01 1.67974591e-01
3.64800781e-01 -6.11434937e-01 -2.96063513e-01 4.68343288e-01
-9.02996302e-01 -3.60596895e-01 3.96318972e-01 6.58332765e-01
-3.86477947e-01 1.37531549e-01 5.23728192e-01 -1.08574498e+00
9.68215704e-01 -4.54830855e-01 -4.80874687e-01 4.33982275e-02
-4.61947441e-01 -6.23159967e-02 7.59281337e-01 -4.76719499e-01
-4.62330610e-01 -3.72922361e-01 -2.63731062e-01 3.71620059e-01
-4.52806115e-01 2.09796533e-01 1.74484274e-03 -2.00220764e-01
5.61726630e-01 1.67891160e-01 1.98895693e-01 -4.60779965e-01
-1.92243457e-01 1.37242615e+00 -2.45133918e-02 -4.41241711e-01
7.99417317e-01 2.39478081e-01 1.28315389e-01 -1.20821190e+00
-1.32048368e-01 -6.58039391e-01 -5.08830667e-01 -2.36974463e-01
2.68955350e-01 -2.70357698e-01 -1.18014598e+00 4.91177231e-01
-1.41650248e+00 1.75164551e-01 2.12490484e-01 4.14121062e-01
-1.94659665e-01 8.65397513e-01 -3.76195222e-01 -1.30476809e+00
1.69637024e-01 -1.12578487e+00 8.03632438e-01 1.08836554e-02
-9.08781052e-01 -1.00221324e+00 -1.34534642e-01 5.57659626e-01
3.82527471e-01 3.12068880e-01 7.40774155e-01 -1.04613149e+00
-1.21816307e-01 -7.15142787e-01 -7.43690208e-02 2.81374902e-01
3.34506601e-01 2.43969485e-01 -8.54082584e-01 -3.35510187e-02
3.63620780e-02 1.17546424e-01 4.30797428e-01 -4.40282188e-02
9.45778251e-01 -3.29109222e-01 -1.79745287e-01 -9.98146236e-02
1.23235548e+00 1.00153446e-01 9.65761721e-01 6.72343791e-01
3.04223239e-01 9.45448577e-01 3.87592226e-01 4.24268246e-01
1.64096147e-01 8.57929111e-01 2.31089413e-01 2.13771015e-01
2.33890846e-01 -2.11656287e-01 1.50417969e-01 8.37175548e-01
-3.32465410e-01 -5.28247893e-01 -8.39782476e-01 3.26575547e-01
-1.58624637e+00 -1.12781942e+00 -7.36693621e-01 2.64053178e+00
4.84427065e-01 4.85112160e-01 6.65540934e-01 8.12242985e-01
9.02643204e-01 -1.21362187e-01 3.27672362e-01 -6.63933396e-01
-1.93935782e-01 -3.86262941e-03 7.40877450e-01 4.83420968e-01
-7.48104692e-01 3.82905573e-01 5.77281666e+00 9.63526487e-01
-1.23521388e+00 1.15277417e-01 7.71556497e-01 2.38085940e-01
-1.06707364e-01 1.47908077e-01 -8.23890626e-01 1.24151945e+00
1.00535154e+00 2.58839935e-01 7.61792287e-02 2.39979357e-01
4.86492902e-01 -7.61760414e-01 -7.88990378e-01 1.06119609e+00
1.03385940e-01 -1.07521188e+00 -1.12274915e-01 5.32739937e-01
5.53394198e-01 -5.75045109e-01 -2.71007299e-01 -1.59739524e-01
-2.94202417e-01 -1.15710044e+00 5.65811872e-01 5.47582209e-01
3.71442527e-01 -5.31130791e-01 1.28717852e+00 7.29075909e-01
-6.21840119e-01 1.25689790e-01 -4.77590263e-02 -3.70248228e-01
2.27717489e-01 6.25696719e-01 -9.62011695e-01 4.14991856e-01
8.02700743e-02 6.90711895e-04 -6.67669833e-01 1.23937690e+00
-8.60001445e-02 6.83707535e-01 -5.05664766e-01 -5.70965827e-01
1.44501448e-01 -2.73662150e-01 5.59335947e-01 1.19834197e+00
6.20312929e-01 -4.18504655e-01 -3.59078020e-01 3.64396751e-01
6.09777309e-02 5.45535922e-01 -8.20102870e-01 -2.47118875e-01
5.02035558e-01 1.03053975e+00 -1.20503235e+00 -1.55971572e-01
-1.61632419e-01 1.00345755e+00 -4.61671166e-02 -4.89740968e-02
-5.26089787e-01 -2.76111364e-01 6.34552762e-02 8.16704392e-01
-1.17939912e-01 -2.66693711e-01 -3.43998820e-01 -9.37368810e-01
9.92161930e-02 -1.20923066e+00 4.61899154e-02 -2.47998863e-01
-1.12604010e+00 4.04074103e-01 -6.57349676e-02 -1.29777730e+00
-3.14851969e-01 -6.50290251e-01 -6.12464488e-01 1.01262784e+00
-1.03748775e+00 -8.84277523e-01 4.16421592e-02 1.50384975e-03
2.10955575e-01 -5.04162312e-01 6.68252110e-01 3.71307909e-01
-1.52862072e-01 4.89088625e-01 2.46851534e-01 1.10441715e-01
7.94302881e-01 -1.13033366e+00 2.11644024e-02 7.15166569e-01
2.81516165e-01 8.57160747e-01 1.20501220e+00 -9.25053060e-01
-9.12048101e-01 -2.01085299e-01 1.44992375e+00 -5.99030495e-01
7.93484330e-01 -3.46991360e-01 -7.62822926e-01 1.10323846e-01
2.03576371e-01 -5.79191983e-01 5.60549557e-01 1.73676461e-01
1.78326383e-01 1.47819892e-01 -1.17309070e+00 3.86794508e-01
4.41037714e-01 -3.88038188e-01 -6.58014178e-01 2.94900447e-01
-2.49860018e-01 -2.00040713e-01 -6.50611818e-01 -1.41414955e-01
7.13938653e-01 -1.18311036e+00 4.12366867e-01 -1.51442319e-01
4.05592799e-01 -2.38382056e-01 7.29917139e-02 -1.09166396e+00
4.28852528e-01 -8.02592754e-01 2.45983586e-01 1.67977941e+00
4.25619394e-01 -9.88473296e-01 9.33439255e-01 3.75191033e-01
5.81036985e-01 -7.15819240e-01 -1.08633161e+00 -7.75638402e-01
-1.29002988e-01 -1.93118513e-01 5.56500256e-01 1.01216424e+00
7.42823631e-02 1.03210233e-01 -5.39382339e-01 -2.55906701e-01
2.89067656e-01 -2.63981313e-01 1.07761323e+00 -1.57708490e+00
-2.98361808e-01 -6.61503613e-01 -6.92488551e-01 -6.67521715e-01
7.02249855e-02 -4.40845102e-01 -3.65338922e-01 -1.28986645e+00
7.62724876e-02 -6.97777271e-01 2.91736513e-01 -1.64409578e-01
6.79415017e-02 3.62006307e-01 1.65809467e-01 3.96009654e-01
2.53380612e-02 -4.06214036e-02 1.02253973e+00 1.33832052e-01
-4.57565784e-02 4.41890985e-01 -6.21550083e-01 6.25881016e-01
8.15154612e-01 -4.90378231e-01 -2.56111473e-02 3.44969422e-01
5.90054929e-01 -1.75678641e-01 2.63340592e-01 -7.60553360e-01
1.83338895e-01 -1.30858824e-01 1.88846469e-01 -4.14077640e-01
2.01681718e-01 -9.03640985e-01 1.49265110e-01 4.59947854e-01
5.64109311e-02 1.87046796e-01 -2.37560824e-01 2.65454561e-01
-2.52746642e-01 -8.70607138e-01 5.04120529e-01 -2.21622154e-01
-1.82120740e-01 -3.53488445e-01 -6.10965133e-01 -3.00770253e-01
9.16656256e-01 -7.53917634e-01 -4.01168883e-01 -6.14266634e-01
-2.10509032e-01 -1.69420302e-01 9.42799866e-01 8.28297436e-02
1.59493044e-01 -9.44503307e-01 -5.22491872e-01 -2.06988901e-01
-6.17253445e-02 -6.98136091e-01 -3.74328405e-01 1.18391788e+00
-8.54807258e-01 3.81799847e-01 -1.58024594e-01 -2.37993568e-01
-1.50738800e+00 1.79917231e-01 2.64088698e-02 -1.73532173e-01
-2.14042604e-01 2.25163639e-01 -6.43672347e-01 -1.81233853e-01
9.15990621e-02 5.21854125e-02 -4.69379693e-01 5.54408550e-01
3.87735784e-01 6.35537326e-01 2.45349318e-01 -1.01554668e+00
-4.04261023e-01 3.26736450e-01 1.44441485e-01 -4.05402035e-01
9.60862219e-01 -2.20381454e-01 -3.28745633e-01 8.30449402e-01
1.02456594e+00 8.82285595e-01 -3.38262439e-01 3.49172771e-01
3.70207638e-01 -7.51762569e-01 -4.30017442e-01 -4.17304575e-01
-4.14368153e-01 7.37635255e-01 4.11227882e-01 9.48823512e-01
6.50083423e-01 -2.58782148e-01 7.18447506e-01 1.64152637e-01
4.67328548e-01 -1.34336674e+00 -2.56816715e-01 3.36225301e-01
4.38544780e-01 -1.46600127e+00 1.37858763e-01 -3.91457856e-01
-4.67958868e-01 1.38735294e+00 -2.05550678e-02 7.38578662e-02
3.02291602e-01 1.09896749e-01 4.75388430e-02 -4.45281640e-02
-1.51159018e-01 -2.06572950e-01 1.54369041e-01 4.06569213e-01
8.00700486e-01 1.21036902e-01 -1.20323384e+00 1.34521931e-01
-5.58631063e-01 -2.94854306e-02 9.63235795e-01 7.39648163e-01
-5.30290127e-01 -1.67998338e+00 -8.22727442e-01 7.94658601e-01
-8.76594484e-01 2.13016078e-01 -8.42391253e-01 9.99972165e-01
3.00090522e-01 1.20250058e+00 -2.09527194e-01 -2.94994652e-01
4.09793854e-02 2.52054453e-01 3.68468702e-01 -3.74035805e-01
-8.13590169e-01 -2.75203109e-01 5.48624814e-01 1.28951207e-01
-5.99877656e-01 -7.38218725e-01 -6.78630710e-01 -8.79214644e-01
-5.05363226e-01 5.76751471e-01 1.01065755e+00 1.30460131e+00
-1.76665977e-01 3.89930676e-03 6.11832201e-01 -4.22570288e-01
-3.92476380e-01 -1.17772329e+00 -7.17426598e-01 7.00870037e-01
-4.47084270e-02 -6.82060659e-01 -5.54988444e-01 -1.31233722e-01] | [9.48448371887207, 10.609443664550781] |
6de18dd4-a15c-4544-b9c6-33455650056d | layered-neural-atlases-for-consistent-video | 2109.11418 | null | https://arxiv.org/abs/2109.11418v1 | https://arxiv.org/pdf/2109.11418v1.pdf | Layered Neural Atlases for Consistent Video Editing | We present a method that decomposes, or "unwraps", an input video into a set of layered 2D atlases, each providing a unified representation of the appearance of an object (or background) over the video. For each pixel in the video, our method estimates its corresponding 2D coordinate in each of the atlases, giving us a consistent parameterization of the video, along with an associated alpha (opacity) value. Importantly, we design our atlases to be interpretable and semantic, which facilitates easy and intuitive editing in the atlas domain, with minimal manual work required. Edits applied to a single 2D atlas (or input video frame) are automatically and consistently mapped back to the original video frames, while preserving occlusions, deformation, and other complex scene effects such as shadows and reflections. Our method employs a coordinate-based Multilayer Perceptron (MLP) representation for mappings, atlases, and alphas, which are jointly optimized on a per-video basis, using a combination of video reconstruction and regularization losses. By operating purely in 2D, our method does not require any prior 3D knowledge about scene geometry or camera poses, and can handle complex dynamic real world videos. We demonstrate various video editing applications, including texture mapping, video style transfer, image-to-video texture transfer, and segmentation/labeling propagation, all automatically produced by editing a single 2D atlas image. | ['Tali Dekel', 'Oliver Wang', 'Dolev Ofri', 'Yoni Kasten'] | 2021-09-23 | null | null | null | null | ['video-style-transfer', 'video-reconstruction'] | ['computer-vision', 'computer-vision'] | [ 4.30353850e-01 8.51702988e-02 4.91926484e-02 -2.81142980e-01
-5.28520226e-01 -5.98363996e-01 3.50049496e-01 -9.60815847e-02
-1.73176408e-01 4.16768163e-01 5.03451983e-03 9.07731801e-02
3.70884389e-01 -5.85938036e-01 -1.14978409e+00 -7.09503651e-01
3.83951589e-02 3.96622121e-01 3.88392836e-01 1.93214953e-01
1.03119230e-02 7.50090837e-01 -1.38782144e+00 2.72245616e-01
4.39309686e-01 1.07362247e+00 3.92105043e-01 8.88045490e-01
-3.51768509e-02 8.67141008e-01 -4.07912672e-01 -3.60334486e-01
3.95839453e-01 -1.07101649e-01 -5.75258672e-01 6.19896948e-01
8.13457310e-01 -5.35266638e-01 -6.27681911e-01 1.11690581e+00
-2.25034039e-02 2.42469847e-01 5.92539608e-01 -1.06055069e+00
-5.38985968e-01 -1.48846069e-02 -8.32531989e-01 -1.88040867e-01
4.75496978e-01 -5.08334190e-02 4.26750630e-01 -8.99257123e-01
8.83983314e-01 1.17439973e+00 7.44509935e-01 4.30849046e-01
-1.54588616e+00 -1.85495391e-01 3.19971085e-01 -2.03992441e-01
-1.40959215e+00 -3.93951297e-01 9.11228955e-01 -7.29988337e-01
6.91232622e-01 3.57511759e-01 9.35094774e-01 8.43743205e-01
4.06182945e-01 6.59819365e-01 8.59291613e-01 -4.79375750e-01
1.49568006e-01 2.30773374e-01 -3.92859697e-01 1.14439249e+00
-3.74044180e-01 -1.16336182e-01 -4.37385648e-01 -1.08519495e-01
1.45156693e+00 -7.99293965e-02 -6.13036752e-01 -7.16730535e-01
-1.32225227e+00 4.22632247e-01 -1.22630903e-02 -3.06255758e-01
-2.88037598e-01 3.14632535e-01 2.06946954e-01 1.30512238e-01
7.55713463e-01 -3.85653935e-02 -3.32872182e-01 1.94594357e-02
-8.11851799e-01 1.11647427e-01 5.25129497e-01 1.04420304e+00
9.21740592e-01 1.51403040e-01 1.35754630e-01 8.01362157e-01
3.04282546e-01 6.55327439e-01 9.81908366e-02 -1.61465538e+00
2.67990738e-01 2.33169824e-01 1.61286995e-01 -1.19004178e+00
-1.47853360e-01 7.00506419e-02 -8.12550426e-01 6.76053226e-01
2.51042902e-01 5.37252538e-02 -8.50692213e-01 1.84215140e+00
6.08609915e-01 3.62652600e-01 -3.15540671e-01 8.50973845e-01
5.05531430e-01 7.99116910e-01 -9.52000767e-02 -2.10609540e-01
1.16036284e+00 -8.98721218e-01 -5.50893784e-01 -1.45074055e-01
1.10711277e-01 -9.15909946e-01 8.96405697e-01 3.59656662e-01
-1.45564246e+00 -5.15708923e-01 -9.74796951e-01 -3.52457404e-01
-7.15027153e-02 -4.30367775e-02 2.92236477e-01 3.03238273e-01
-1.19954038e+00 5.97586930e-01 -9.82280016e-01 -1.65321976e-01
1.49970159e-01 2.89412916e-01 -5.96317708e-01 1.83488965e-01
-7.80775666e-01 8.96354437e-01 3.00459322e-02 4.08694558e-02
-7.39871681e-01 -8.13955367e-01 -1.04160380e+00 -1.43009245e-01
3.21047723e-01 -8.44159484e-01 1.14832222e+00 -1.25962818e+00
-1.74661088e+00 1.25804174e+00 -3.58095348e-01 6.83326786e-03
5.35625279e-01 -1.82613254e-01 -1.32375970e-01 3.98866147e-01
-8.41947198e-02 9.20286894e-01 1.27715135e+00 -1.49878764e+00
-4.26209122e-01 -1.92918897e-01 2.58449972e-01 5.84237278e-01
2.68382598e-02 4.59321775e-02 -1.21268094e+00 -9.56866443e-01
1.21646322e-01 -8.42267334e-01 -1.36705056e-01 7.21146226e-01
-3.38470340e-01 5.61058760e-01 9.37114656e-01 -1.02191603e+00
9.47675645e-01 -2.33894324e+00 7.26362288e-01 2.58399844e-01
3.00079703e-01 -3.03198904e-01 -2.46230140e-02 -1.32937074e-01
-4.13380526e-02 -2.75336355e-01 -6.09225512e-01 -5.83002627e-01
-2.78126508e-01 3.00081849e-01 -2.66779184e-01 6.21364236e-01
2.07842775e-02 5.03898382e-01 -8.27884078e-01 -6.45944357e-01
6.55397236e-01 8.41579556e-01 -7.29791701e-01 3.71208459e-01
-4.08671200e-01 8.03494573e-01 -2.40520149e-01 5.68562448e-01
5.98732471e-01 -1.38840735e-01 9.29622203e-02 -4.71322894e-01
-6.23408332e-02 -2.63344586e-01 -1.17079425e+00 1.80960608e+00
-6.07655168e-01 8.14047158e-01 5.62160134e-01 -6.44844532e-01
5.81906378e-01 3.00735712e-01 8.32551539e-01 -4.34634686e-01
-8.51157960e-03 -7.89441988e-02 -8.13535690e-01 -5.18969536e-01
4.24369305e-01 5.24008349e-02 1.52816683e-01 6.12317264e-01
2.24122666e-02 -4.16317493e-01 -1.25781864e-01 1.30869403e-01
7.97258317e-01 5.58579564e-01 -2.86308490e-02 -2.78514177e-02
2.36320198e-01 -2.73223877e-01 3.94775361e-01 4.17344183e-01
4.68898304e-02 9.97261584e-01 4.55614895e-01 -4.90206510e-01
-1.34270370e+00 -1.41250002e+00 -1.75318673e-01 7.83247232e-01
3.75079960e-01 -2.16738373e-01 -1.14816189e+00 -4.64806110e-01
-1.03330754e-01 3.76389891e-01 -5.26780903e-01 1.21059522e-01
-7.63665497e-01 -4.22425300e-01 1.39300808e-01 3.80542248e-01
2.52598584e-01 -7.94272482e-01 -5.87758124e-01 2.38694295e-01
-2.88311660e-01 -1.28141987e+00 -1.07869923e+00 -1.44035444e-02
-9.14108872e-01 -1.12367642e+00 -6.59457922e-01 -8.77841592e-01
1.11459315e+00 2.96633124e-01 1.06804347e+00 -3.66925672e-02
-2.89293379e-01 8.02171767e-01 9.49428976e-02 2.21242100e-01
-6.01300120e-01 -6.25379384e-01 1.85334355e-01 4.56004649e-01
-3.03840518e-01 -7.17948020e-01 -4.03340191e-01 3.55823338e-01
-1.08306384e+00 5.13584673e-01 7.22326413e-02 6.20311677e-01
9.88719046e-01 -8.93329084e-02 -3.18486750e-01 -8.18985522e-01
6.46838844e-02 -1.34226888e-01 -9.09517407e-01 3.01631898e-01
8.57514516e-02 -7.96850026e-02 5.87063730e-01 -3.93508106e-01
-1.21026611e+00 3.20665002e-01 8.56815502e-02 -1.07014465e+00
-1.28320649e-01 1.34932786e-01 -2.86489874e-01 -4.34424222e-01
3.31682742e-01 1.04021072e-01 2.59205908e-01 -4.84866410e-01
6.16889596e-01 2.73936361e-01 9.60610867e-01 -7.24066734e-01
7.31084824e-01 7.58953452e-01 -5.07468805e-02 -9.99101460e-01
-7.54559577e-01 -5.50056808e-03 -9.96974170e-01 -5.02921283e-01
1.12009907e+00 -7.75268316e-01 -6.20254397e-01 6.14524722e-01
-1.37156200e+00 -5.83215415e-01 -2.60114998e-01 5.04362643e-01
-7.96886921e-01 5.56703210e-01 -8.05046976e-01 -3.44422281e-01
8.66141096e-02 -1.44595051e+00 1.14376974e+00 3.20736505e-02
-3.09921741e-01 -1.20770907e+00 -1.06757306e-01 2.59174347e-01
1.36007965e-01 3.81831557e-01 1.20794308e+00 3.55204046e-01
-9.61597085e-01 1.74897686e-02 -1.02022350e-01 5.51735163e-01
2.07979828e-01 3.64025205e-01 -1.01535237e+00 -1.73595205e-01
1.08323537e-01 -1.28144259e-03 4.54801023e-01 6.58663094e-01
1.51708496e+00 -4.42220241e-01 -1.15817688e-01 1.09496582e+00
1.13659048e+00 2.65686482e-01 4.25961852e-01 3.72606553e-02
1.05769038e+00 5.53530812e-01 1.61978021e-01 2.96065927e-01
2.11079717e-01 1.06508470e+00 4.00365800e-01 -3.60148907e-01
-2.85036266e-01 -1.67305663e-01 5.61609983e-01 8.71155739e-01
-1.24495186e-01 -1.68904278e-03 -3.16628635e-01 -2.31494140e-02
-1.65202248e+00 -8.22784126e-01 1.61416054e-01 2.48414612e+00
8.54797602e-01 -5.10623679e-02 -1.38300940e-01 -3.58192950e-01
8.90309811e-01 2.46221349e-01 -5.71015120e-01 -2.82332093e-01
-1.17602557e-01 1.57738272e-02 5.66710830e-01 9.84101295e-01
-1.30610037e+00 7.58910716e-01 7.05562687e+00 6.23329699e-01
-1.28135455e+00 2.49099322e-02 7.01608062e-01 -2.69755453e-01
-4.86348420e-01 -9.44319516e-02 -1.78960159e-01 3.88921291e-01
3.87842059e-01 1.85607091e-01 8.24504972e-01 6.18225157e-01
2.22085282e-01 -1.34643927e-01 -1.29470873e+00 1.03725255e+00
1.87416509e-01 -1.61348498e+00 1.78031102e-01 -1.64442688e-01
7.19918847e-01 -1.13458239e-01 5.07509373e-02 -3.83876264e-01
9.57204178e-02 -7.47604847e-01 1.26493812e+00 7.76376128e-01
1.18750834e+00 -4.16676730e-01 7.02804625e-02 -1.04436465e-01
-1.02114904e+00 3.52280676e-01 -4.02360857e-02 3.65235627e-01
4.29723501e-01 4.52134520e-01 -2.55282849e-01 8.87630060e-02
7.69415021e-01 7.12935150e-01 -1.55811340e-01 7.25492239e-01
1.20436354e-02 1.66029647e-01 -4.68907595e-01 6.16272688e-01
-7.66576752e-02 -5.90240896e-01 8.33557248e-01 1.07250738e+00
2.01071545e-01 2.53157943e-01 3.93016160e-01 8.81835580e-01
-2.06747606e-01 -1.71739250e-01 -5.19497216e-01 3.17628384e-01
4.54640567e-01 1.11586034e+00 -8.51934373e-01 -4.21720237e-01
-5.59971392e-01 1.26968920e+00 9.81238410e-02 7.62381613e-01
-8.94109845e-01 -1.52109474e-01 7.85137057e-01 3.83939296e-01
1.79222032e-01 -2.49962017e-01 -2.87975311e-01 -1.35396206e+00
2.96102971e-01 -6.68699384e-01 6.04213029e-02 -1.17541480e+00
-1.01383233e+00 5.99238575e-01 1.76320836e-01 -1.41547692e+00
-9.56120640e-02 -7.26281703e-01 -4.50819284e-01 6.84413075e-01
-1.00962675e+00 -8.75267446e-01 -2.78317362e-01 6.77932799e-01
5.15740573e-01 -4.41991538e-02 7.30109870e-01 5.06430447e-01
-4.95634347e-01 3.06258112e-01 1.72381923e-01 3.39085795e-02
6.51995122e-01 -1.10202372e+00 2.93469548e-01 6.60542905e-01
1.37699381e-01 3.06612700e-01 6.20622754e-01 -5.01844883e-01
-1.44319725e+00 -1.13065290e+00 3.53709280e-01 -6.75051928e-01
4.23553348e-01 -3.78207535e-01 -9.70292687e-01 9.95768130e-01
-3.90139967e-02 1.22708693e-01 2.49278635e-01 -2.89003760e-01
-2.03523591e-01 -1.89367101e-01 -1.12964511e+00 7.64394045e-01
8.95542920e-01 -7.58257806e-01 -2.08368421e-01 3.50430816e-01
7.69731939e-01 -1.08476198e+00 -8.56163740e-01 2.08742879e-02
6.27150178e-01 -9.96549308e-01 1.23598719e+00 -3.21829706e-01
3.54197443e-01 -5.52902937e-01 -2.37383589e-01 -1.14828122e+00
-2.29200333e-01 -8.01539361e-01 -2.63857275e-01 8.15093935e-01
6.54047057e-02 -2.09749892e-01 6.54104650e-01 8.63020420e-01
-3.75740081e-01 -7.24344850e-01 -9.92371440e-01 -2.49436066e-01
-4.86602724e-01 -5.62334061e-01 3.73847306e-01 9.81474459e-01
-4.62891787e-01 -1.05812863e-01 -5.52995622e-01 4.46689934e-01
8.83428037e-01 -5.30075654e-02 7.24226952e-01 -7.73164093e-01
-5.30319870e-01 -4.71759319e-01 -4.35667843e-01 -1.24034131e+00
1.69933647e-01 -7.59692311e-01 -8.87802709e-03 -1.08544481e+00
7.95281306e-02 -4.00878280e-01 1.20896026e-01 3.80717546e-01
2.88787156e-01 5.20750940e-01 2.07790777e-01 3.35335761e-01
-2.33642414e-01 4.45553601e-01 1.33320022e+00 -1.63910627e-01
-2.04258829e-01 -1.39745221e-01 -2.22778335e-01 1.25932729e+00
2.45496631e-01 -3.06500584e-01 -3.11025739e-01 -9.37827289e-01
-1.14953764e-01 4.46240455e-01 5.73590696e-01 -7.37401664e-01
-2.51854695e-02 -3.31318885e-01 5.91089785e-01 -2.55395919e-01
7.39198506e-01 -8.45271647e-01 5.95088959e-01 -7.27677271e-02
-2.60270655e-01 1.00999609e-01 2.15420753e-01 5.82717180e-01
-1.10902995e-01 -9.38677639e-02 9.12563920e-01 -1.72005460e-01
-5.79223394e-01 5.10077596e-01 -2.93118507e-01 -1.72853902e-01
1.02435887e+00 -5.46482921e-01 2.44249091e-01 -4.80856866e-01
-1.04748058e+00 3.45197320e-02 9.60401714e-01 3.29276800e-01
7.02225983e-01 -1.37931609e+00 -4.07907099e-01 6.48476243e-01
-1.75591797e-01 3.00540537e-01 5.24949193e-01 6.83945835e-01
-1.05679059e+00 -1.17180750e-01 -2.50249922e-01 -8.35350037e-01
-1.12773728e+00 3.58863622e-01 6.61044955e-01 1.74221233e-01
-1.11317003e+00 7.86530793e-01 7.33069539e-01 -3.65734965e-01
5.24503708e-01 -3.64208430e-01 2.72065997e-01 -3.01769167e-01
5.49760580e-01 9.91620570e-02 -9.75876898e-02 -7.15150177e-01
-1.53602436e-01 1.05277431e+00 1.43965513e-01 -3.49356413e-01
1.03411591e+00 -4.07749832e-01 -3.08910131e-01 4.50275689e-01
1.30613625e+00 2.20523298e-01 -2.03111887e+00 -2.39888132e-01
-5.38251638e-01 -7.27327585e-01 1.68368489e-01 -4.34315771e-01
-1.30600035e+00 7.14691341e-01 2.39433289e-01 2.48310007e-02
1.17799914e+00 -2.14837641e-02 7.63127506e-01 1.42289162e-01
3.08016658e-01 -9.92481947e-01 -5.67652322e-02 3.47251832e-01
9.53734279e-01 -8.81201804e-01 -3.52429086e-03 -3.46477866e-01
-6.16039455e-01 1.20841563e+00 4.57780570e-01 6.53527156e-02
7.29155302e-01 4.67305779e-01 2.29102045e-01 -4.21137884e-02
-4.88855124e-01 5.21465540e-01 2.44232729e-01 6.23471558e-01
1.42838329e-01 -4.04929034e-02 5.69444418e-01 -9.44746211e-02
7.84681961e-02 -1.93879977e-01 5.49223244e-01 7.50207841e-01
-2.86977679e-01 -7.91874647e-01 -6.26911640e-01 2.60237277e-01
-1.69752955e-01 5.41783199e-02 -1.53845930e-02 5.91945231e-01
9.42213461e-02 2.33494937e-01 5.56153595e-01 -2.11116880e-01
1.80890262e-01 -1.02012053e-01 9.59829271e-01 -4.63842422e-01
-3.16481702e-02 4.09539610e-01 -9.87379327e-02 -7.99176693e-01
-4.96058762e-01 -5.64469278e-01 -1.24704027e+00 -4.03466374e-01
1.10975914e-01 -2.16422603e-01 5.84858835e-01 7.52985895e-01
2.38349408e-01 2.72342801e-01 7.01444447e-01 -1.47171474e+00
1.00973934e-01 -3.90848875e-01 -6.96283937e-01 4.35226113e-01
5.09598494e-01 -7.63072610e-01 -2.38726467e-01 7.91422307e-01] | [9.06655502319336, -2.789391040802002] |
3f3fbf74-5997-45d5-abff-85a0fc12c02a | fine-grained-hard-negative-mining | 2301.01079 | null | https://arxiv.org/abs/2301.01079v1 | https://arxiv.org/pdf/2301.01079v1.pdf | Fine-Grained Hard Negative Mining: Generalizing Mitosis Detection with a Fifth of the MIDOG 2022 Dataset | Making histopathology image classifiers robust to a wide range of real-world variability is a challenging task. Here, we describe a candidate deep learning solution for the Mitosis Domain Generalization Challenge 2022 (MIDOG) to address the problem of generalization for mitosis detection in images of hematoxylin-eosin-stained histology slides under high variability (scanner, tissue type and species variability). Our approach consists in training a rotation-invariant deep learning model using aggressive data augmentation with a training set enriched with hard negative examples and automatically selected negative examples from the unlabeled part of the challenge dataset. To optimize the performance of our models, we investigated a hard negative mining regime search procedure that lead us to train our best model using a subset of image patches representing 19.6% of our training partition of the challenge dataset. Our candidate model ensemble achieved a F1-score of .697 on the final test set after automated evaluation on the challenge platform, achieving the third best overall score in the MIDOG 2022 Challenge. | ['Viktor H. Koelzer', 'Maxime W. Lafarge'] | 2023-01-03 | null | null | null | null | ['mitosis-detection'] | ['medical'] | [ 3.98268878e-01 4.70456988e-01 -1.70868173e-01 -3.89859766e-01
-1.24333549e+00 -6.04824603e-01 5.57618201e-01 2.95930535e-01
-8.43198836e-01 9.25654292e-01 -1.99467257e-01 -3.21618021e-01
-1.19811587e-01 -4.98145640e-01 -4.26957160e-01 -1.05511594e+00
-4.55581620e-02 1.05189204e+00 1.70901358e-01 -1.54444957e-02
4.26280871e-02 6.78996801e-01 -1.13420403e+00 7.00601101e-01
5.78193367e-01 1.04592133e+00 3.41135338e-02 9.44443643e-01
4.27417278e-01 4.94910032e-01 -7.99123168e-01 -1.42917544e-01
2.88071245e-01 -2.19749138e-01 -9.97898579e-01 9.68947709e-02
5.96564889e-01 1.20743960e-01 -1.13050163e-01 9.08352315e-01
1.01824367e+00 -3.97821277e-01 8.91985238e-01 -1.19553995e+00
-5.58096915e-02 6.38742566e-01 -5.74447095e-01 7.16865838e-01
-4.74720329e-01 4.74175125e-01 5.46646178e-01 -6.25518978e-01
1.26374137e+00 3.50442827e-01 8.51836860e-01 8.94996464e-01
-1.46695149e+00 -6.62292480e-01 -2.08653912e-01 2.16754735e-01
-1.64682984e+00 -2.06716642e-01 2.58532166e-01 -6.10756636e-01
9.53549027e-01 3.75310242e-01 6.57041907e-01 1.14620245e+00
5.19271791e-01 5.19315779e-01 1.19398344e+00 -1.75150678e-01
4.10729408e-01 4.59106714e-01 1.77875251e-01 4.26563025e-01
3.00652802e-01 4.74528298e-02 -3.90280485e-01 -1.02832280e-01
3.88072491e-01 -5.03353655e-01 -3.65111768e-01 -3.81754518e-01
-1.11685252e+00 5.82403064e-01 5.08086264e-01 3.43980193e-01
-1.51313469e-01 -1.57358050e-01 5.08102953e-01 4.24011797e-01
4.10622358e-01 7.00948715e-01 -7.22582638e-01 3.56375426e-01
-8.39287221e-01 2.34904692e-01 6.45243287e-01 4.76230264e-01
4.47411031e-01 -3.61151963e-01 -3.15410912e-01 6.68233275e-01
-1.39815331e-01 1.70006931e-01 1.02276003e+00 -4.90740716e-01
6.59940019e-02 7.65828013e-01 -4.29628223e-01 -2.85172880e-01
-1.13513207e+00 -1.22190571e+00 -1.18808246e+00 1.63110688e-01
6.61485851e-01 -2.30210081e-01 -1.47057283e+00 1.71596074e+00
4.02185559e-01 -1.08471110e-01 1.07556067e-01 5.50900638e-01
9.68353748e-01 1.71951782e-02 3.11896294e-01 -1.66152731e-01
1.19336033e+00 -6.08677924e-01 -1.71464622e-01 -1.45775184e-01
1.26675868e+00 -3.05806905e-01 7.50537515e-01 4.69734818e-01
-5.27434528e-01 -2.31540725e-01 -1.19156075e+00 9.84760746e-02
-6.36497855e-01 1.81979239e-01 3.14883024e-01 4.38247561e-01
-1.34284484e+00 3.11004072e-01 -7.55049050e-01 -4.80092257e-01
7.02340126e-01 9.67196882e-01 -7.34974921e-01 -6.23991340e-02
-7.57791340e-01 7.86806047e-01 5.98917484e-01 -8.64518136e-02
-1.02596176e+00 -1.20573485e+00 -4.77440417e-01 -8.56940076e-02
2.80321063e-03 -7.86056578e-01 1.15482867e+00 -1.01241052e+00
-6.86610401e-01 1.62240827e+00 2.96016842e-01 -8.26683760e-01
9.12426949e-01 5.41622400e-01 -1.41255453e-01 1.95335746e-01
1.41036838e-01 1.07465982e+00 5.02838314e-01 -1.08178723e+00
-7.10852563e-01 -5.61457396e-01 -5.80675423e-01 -1.84492752e-01
-3.74783248e-01 -6.11248136e-01 -1.61918521e-01 -5.13155818e-01
-2.63023609e-03 -1.24587226e+00 -4.55595702e-01 -3.07537854e-01
-3.40890020e-01 -1.61353871e-01 5.08878708e-01 -6.27060950e-01
8.99826109e-01 -2.13717151e+00 -8.71967524e-02 5.06309032e-01
3.99992913e-01 1.26969829e-01 -3.64877045e-01 -2.37246394e-01
-5.31446934e-01 3.97380143e-01 -4.62496206e-02 -1.79433659e-01
-2.23022193e-01 -1.76746726e-01 1.20475441e-01 7.69998968e-01
4.48845953e-01 8.40892017e-01 -7.74938047e-01 -5.71284592e-01
-2.63398945e-01 2.48988315e-01 -4.44396406e-01 -3.06889392e-03
-2.91611314e-01 3.51996928e-01 8.76084939e-02 8.86223137e-01
7.21456110e-01 -5.92167735e-01 3.60217720e-01 -2.02853739e-01
5.80434084e-01 -3.19458097e-01 -7.79491544e-01 1.32224739e+00
-1.50140179e-02 7.32337832e-01 -3.99710611e-02 -5.67869365e-01
8.39305580e-01 1.23391494e-01 7.83056200e-01 -7.93954134e-01
1.55859351e-01 4.50032234e-01 4.95685220e-01 -5.67763746e-01
1.17680728e-01 -3.62358719e-01 -9.28791538e-02 3.16498846e-01
5.21080613e-01 -9.71266031e-02 2.31907308e-01 -2.51481514e-02
1.69753599e+00 -4.20452267e-01 3.97217989e-01 -5.66228986e-01
5.94821811e-01 2.41276309e-01 6.50550425e-01 6.69346809e-01
-7.96345592e-01 8.79344106e-01 6.22588575e-01 -8.49671006e-01
-1.05797493e+00 -8.18902612e-01 -6.21237814e-01 7.74223685e-01
-3.97497922e-01 1.03017315e-01 -6.18530035e-01 -1.14692545e+00
-1.07491873e-01 2.61511385e-01 -1.31838727e+00 -4.29075301e-01
-4.86891240e-01 -1.56052542e+00 6.45051360e-01 4.61643428e-01
3.42455536e-01 -9.64786232e-01 -4.58907574e-01 -6.16025254e-02
-2.13893689e-03 -1.04893911e+00 4.90485737e-03 8.97266030e-01
-5.86601019e-01 -1.43390906e+00 -7.07970858e-01 -8.86839032e-01
9.03827608e-01 -3.32602024e-01 1.36945283e+00 1.45640582e-01
-8.88637841e-01 2.97352728e-02 -1.13316223e-01 -6.39539421e-01
-7.74450541e-01 6.41458929e-01 -2.27445170e-01 -1.09275587e-01
6.65009081e-01 -1.09413885e-01 -6.29904687e-01 2.17004657e-01
-9.74591255e-01 -1.39413580e-01 8.67648959e-01 1.23194587e+00
1.00498462e+00 -1.19961359e-01 8.91701877e-01 -9.90528822e-01
2.31987312e-01 -4.75441962e-01 -3.99809092e-01 2.55224764e-01
-4.56985325e-01 -2.01284975e-01 4.63887542e-01 -6.02111161e-01
-5.26152551e-01 3.35146427e-01 2.55375057e-02 -1.30538329e-01
-1.43482685e-01 5.38892269e-01 -2.10260339e-02 -2.88633287e-01
1.26307726e+00 7.31151551e-02 1.54766157e-01 1.46001071e-01
-2.75202394e-01 4.75782245e-01 7.38192916e-01 -2.03329921e-01
5.22073269e-01 5.35363078e-01 4.30200636e-01 -4.48826402e-01
-7.77737439e-01 -5.27993023e-01 -6.33137286e-01 -1.96948946e-01
6.79003537e-01 -8.91463697e-01 -3.32345575e-01 5.86127400e-01
-5.87669432e-01 -8.07118654e-01 -3.79384905e-01 3.20039868e-01
-4.43734944e-01 -1.12509750e-01 -4.01982069e-01 -2.60587633e-01
-6.29986465e-01 -1.04947436e+00 1.13424003e+00 2.64029920e-01
-5.32443762e-01 -1.03454018e+00 3.50452244e-01 4.75978613e-01
3.39597434e-01 6.74213529e-01 1.09844291e+00 -1.49979937e+00
-2.44211897e-01 -5.39551914e-01 -9.97059792e-03 2.06048042e-01
-7.19814375e-02 1.71883088e-02 -1.28760028e+00 -6.74541712e-01
-2.13124990e-01 -5.55674672e-01 1.19365633e+00 3.16361398e-01
1.44254649e+00 1.80667356e-01 -7.80775547e-01 1.02127409e+00
1.50156868e+00 -1.28473729e-01 4.19336796e-01 8.64845872e-01
1.50505155e-01 5.61743855e-01 5.26394904e-01 2.75124311e-01
4.96159121e-02 2.98874736e-01 6.10718012e-01 -4.21199352e-01
-2.00799584e-01 2.37114072e-01 -3.65211517e-01 -1.48087949e-01
3.38596910e-01 -2.43045256e-01 -1.61662006e+00 8.52943540e-01
-1.33567309e+00 -6.55794680e-01 1.36081547e-01 2.06475902e+00
8.12791228e-01 2.52782375e-01 6.12356476e-02 2.12134331e-01
5.87896049e-01 -4.90230918e-01 -6.86341703e-01 6.55807778e-02
-4.81830150e-01 2.60806561e-01 7.57646739e-01 1.50268108e-01
-1.51768947e+00 5.06475270e-01 6.36747837e+00 8.56912673e-01
-1.45434427e+00 -1.79096550e-01 1.42543769e+00 -3.70960474e-01
1.04575299e-01 -6.46095991e-01 -8.96208644e-01 2.32998773e-01
1.12207460e+00 -2.78990865e-01 -2.20361426e-01 5.97137511e-01
-1.15541026e-01 -2.50224490e-02 -1.33942115e+00 7.36143768e-01
1.76246524e-01 -1.65099669e+00 -4.35738862e-02 5.14834046e-01
1.09700668e+00 3.48393977e-01 5.38888752e-01 5.77692986e-01
1.89561024e-01 -1.44750512e+00 -4.25680876e-02 2.47138292e-01
9.88366902e-01 -6.56731844e-01 1.47704864e+00 3.63020271e-01
-4.14620161e-01 -2.63552129e-01 -8.08505435e-03 6.45040095e-01
-5.21509886e-01 5.49187720e-01 -1.81542969e+00 2.66406178e-01
6.35574758e-01 2.99831599e-01 -1.28753197e+00 1.28473794e+00
5.57155907e-01 3.59501511e-01 -4.67506528e-01 4.42961007e-02
5.39529808e-02 7.31520474e-01 1.75021991e-01 1.43043101e+00
2.11112931e-01 -2.81561106e-01 -4.58768159e-02 3.52614641e-01
-3.49572390e-01 -2.79051829e-02 -3.48904759e-01 1.62997723e-01
3.31427842e-01 1.46800363e+00 -1.15113127e+00 -2.21100196e-01
3.24970812e-01 3.15329015e-01 3.78183454e-01 1.29225060e-01
-6.08458698e-01 2.65759919e-02 1.55516103e-01 2.81957209e-01
1.88218087e-01 6.04705632e-01 -5.39442837e-01 -7.63746083e-01
-2.37540320e-01 -1.23876691e+00 8.51128280e-01 -2.73709267e-01
-1.38253427e+00 7.82100558e-01 -3.07280064e-01 -1.28960288e+00
-3.17245096e-01 -7.74038315e-01 -4.01485682e-01 6.71136975e-01
-1.36208880e+00 -1.33735061e+00 -5.51669657e-01 3.43460381e-01
1.35030925e-01 -5.48726737e-01 9.77765441e-01 9.52906460e-02
-3.84659052e-01 1.16235185e+00 1.19699232e-01 1.05939955e-01
9.71435249e-01 -1.44575381e+00 2.40691751e-02 4.38639075e-01
-3.40072781e-01 2.69301385e-01 6.07904851e-01 -3.58937413e-01
-8.65958750e-01 -1.47005355e+00 7.64278114e-01 -4.98284459e-01
6.35354757e-01 -2.21086726e-01 -8.39368761e-01 6.54197335e-01
-2.05715433e-01 3.74732077e-01 1.40791178e+00 -2.08485737e-01
-2.17713252e-01 -1.85997337e-01 -1.53983331e+00 3.46890539e-01
4.94153231e-01 -3.05836320e-01 3.50993127e-02 5.71093857e-01
1.93647683e-01 -6.08118474e-01 -9.49145496e-01 8.91885400e-01
3.52855116e-01 -5.49645126e-01 5.81394136e-01 -7.46726334e-01
3.48214298e-01 -1.80122599e-01 -4.56554480e-02 -1.28148353e+00
-2.84741670e-01 -2.17465982e-01 1.69773221e-01 7.46710539e-01
6.95325255e-01 -2.04558715e-01 1.40773416e+00 3.82243156e-01
-1.34995371e-01 -1.16148913e+00 -1.35999548e+00 -4.46803331e-01
5.68893135e-01 1.05989903e-01 2.98526287e-01 8.51034522e-01
-1.29055142e-01 2.32642099e-01 3.16113293e-01 1.59806862e-01
4.96248484e-01 -1.35252491e-01 8.16816866e-01 -1.28471315e+00
-3.12783748e-01 -6.22242689e-01 -1.00256729e+00 -1.95714235e-02
2.15558931e-01 -1.17820299e+00 -1.14582419e-01 -1.19100237e+00
6.27774477e-01 -3.35890204e-01 -5.77156126e-01 4.45077598e-01
-1.89594939e-01 6.85428083e-01 -1.00027330e-01 1.11594774e-01
-6.40860379e-01 -1.46029636e-01 1.05390036e+00 -7.32861876e-01
-4.36303131e-02 1.95557922e-02 -8.77690017e-01 4.06790763e-01
8.11455190e-01 -4.81010079e-01 -2.59370774e-01 -7.33945370e-02
5.76026291e-02 -2.07383275e-01 4.05364752e-01 -1.24875486e+00
1.87513724e-01 -2.01180592e-01 9.97059882e-01 -7.36672282e-01
1.53359443e-01 -5.71377456e-01 2.78208107e-01 8.11578929e-01
-5.29345572e-01 -1.94169119e-01 4.63245720e-01 4.54839498e-01
-5.89855909e-02 -2.32757572e-02 1.09094381e+00 -5.39069735e-02
-5.12887120e-01 4.42745656e-01 -2.70903021e-01 2.75106337e-02
1.31401086e+00 -2.68798321e-01 -9.10159647e-01 1.40989140e-01
-1.28353250e+00 4.10873264e-01 6.12857044e-01 9.15460810e-02
2.46627659e-01 -9.73118901e-01 -1.02290285e+00 1.66216820e-01
4.87229556e-01 2.92781264e-01 5.07211924e-01 7.86863685e-01
-7.59695530e-01 3.72345597e-01 -4.10429537e-01 -9.87149119e-01
-1.41575646e+00 4.98636365e-01 1.11151981e+00 -9.80198562e-01
-1.76602349e-01 1.34842551e+00 2.42361560e-01 -4.97384548e-01
2.44049221e-01 -1.33298323e-01 -2.13802516e-01 -9.28976759e-03
5.37426412e-01 6.28081784e-02 6.02484465e-01 -3.84912640e-01
-5.49664736e-01 1.19213127e-01 -4.33543652e-01 1.53450236e-01
1.32586765e+00 5.04889369e-01 1.59820262e-02 6.92434385e-02
1.52791715e+00 -4.10144776e-01 -8.89559329e-01 1.07006475e-01
1.51735291e-01 1.50533035e-01 -6.53653145e-02 -1.26111758e+00
-1.11207378e+00 5.83100796e-01 1.21269345e+00 2.44683400e-02
1.17297232e+00 1.37129724e-02 9.35732946e-02 5.68169534e-01
4.23248522e-02 -1.07310653e+00 1.03162967e-01 3.22883427e-01
7.49492705e-01 -1.59694719e+00 1.46221757e-01 -7.78341666e-02
-3.01822305e-01 1.17720997e+00 1.07847750e+00 1.89383850e-02
5.38589656e-01 5.11193573e-01 4.71160501e-01 -3.06947947e-01
-1.33627999e+00 8.18808377e-02 2.49686558e-02 1.03397191e+00
3.60831261e-01 1.05773862e-02 -2.13547021e-01 7.81669915e-01
-4.55193132e-01 1.76473454e-01 6.60509109e-01 7.40238369e-01
-3.13863844e-01 -7.10136235e-01 -1.40219271e-01 6.76820159e-01
-5.73230445e-01 2.36917868e-01 -7.19449580e-01 1.03477359e+00
2.29559049e-01 3.74415934e-01 1.80677727e-01 -5.66817641e-01
1.31789982e-01 7.09879845e-02 3.51214677e-01 -5.85023344e-01
-9.27011549e-01 -1.26793966e-01 -4.39924113e-02 -8.95583481e-02
-9.16931182e-02 -4.84896779e-01 -1.03949738e+00 -4.46476080e-02
-7.11598992e-02 -1.00224040e-01 4.82661247e-01 6.78519249e-01
2.43662432e-01 7.19752848e-01 5.08804262e-01 -5.77030063e-01
-9.27805305e-01 -1.11010516e+00 -5.28183281e-01 3.95841181e-01
6.19903207e-01 -3.33078891e-01 -5.61852932e-01 2.77698755e-01] | [15.126038551330566, -3.035402774810791] |
0d752f9e-1828-4535-bb0c-e13f373db566 | coach2vec-autoencoding-the-playing-style-of | 2106.15444 | null | https://arxiv.org/abs/2106.15444v1 | https://arxiv.org/pdf/2106.15444v1.pdf | Coach2vec: autoencoding the playing style of soccer coaches | Capturing the playing style of professional soccer coaches is a complex, and yet barely explored, task in sports analytics. Nowadays, the availability of digital data describing every relevant spatio-temporal aspect of soccer matches, allows for capturing and analyzing the playing style of players, teams, and coaches in an automatic way. In this paper, we present coach2vec, a workflow to capture the playing style of professional coaches using match event streams and artificial intelligence. Coach2vec extracts ball possessions from each match, clusters them based on their similarity, and reconstructs the typical ball possessions of coaches. Then, it uses an autoencoder, a type of artificial neural network, to obtain a concise representation (encoding) of the playing style of each coach. Our experiments, conducted on soccer-logs describing the last four seasons of the Italian first division, reveal interesting similarities between prominent coaches, paving the road to the simulation of playing styles and the quantitative comparison of professional coaches. | ['Luca Pappalardo', 'Paolo Cintia'] | 2021-06-29 | null | null | null | null | ['sports-analytics'] | ['computer-vision'] | [-4.32047993e-01 -5.76357126e-01 -5.11529446e-02 -6.27563819e-02
-3.27794552e-01 -6.10911548e-01 4.05666292e-01 7.27368534e-01
-7.13029623e-01 1.15146898e-01 5.61806619e-01 4.46712285e-01
-4.16950822e-01 -1.13533258e+00 -6.78363979e-01 -5.34050882e-01
-8.47932100e-02 7.50170887e-01 2.80679375e-01 -7.77892530e-01
1.07614040e-01 6.19599879e-01 -2.25848651e+00 7.72696614e-01
2.23976448e-02 7.82727242e-01 1.03608929e-01 9.34835255e-01
-6.07831217e-03 1.15619314e+00 -7.79891431e-01 -6.68527067e-01
3.29721630e-01 -4.66626287e-01 -5.21910369e-01 4.33387347e-02
2.11441711e-01 2.80343026e-01 -8.14227700e-01 5.74994445e-01
1.95804060e-01 5.97343206e-01 4.54376966e-01 -9.64892030e-01
1.77689403e-01 9.27834570e-01 -7.53220618e-02 5.03458679e-01
6.08494341e-01 1.92227080e-01 1.09217143e+00 -2.43613526e-01
7.49211848e-01 8.01212311e-01 9.08975303e-01 -1.86709732e-01
-9.19462144e-01 -6.15666389e-01 -2.50981331e-01 8.95468891e-01
-1.44084549e+00 1.25849843e-01 6.39267921e-01 -8.78066301e-01
4.85933572e-01 2.30858520e-01 1.75399530e+00 1.16189098e+00
1.40398979e-01 1.00251603e+00 9.71430659e-01 -2.10893050e-01
2.26008803e-01 -6.15414716e-02 4.05784756e-01 1.81589842e-01
-8.03411752e-02 3.37196350e-01 -1.05423713e+00 2.18799829e-01
4.96719629e-01 3.73444259e-01 1.12020612e-01 -1.64217994e-01
-1.19129252e+00 8.42408597e-01 2.28449702e-01 4.22572285e-01
-8.33357394e-01 2.06681713e-01 7.96636164e-01 2.19714478e-01
1.52613863e-01 4.08457816e-01 -3.82601433e-02 -8.82829607e-01
-1.06330299e+00 1.02104032e+00 9.13022697e-01 5.11608720e-01
6.65966034e-01 -6.06423244e-02 -2.10651055e-01 5.69634497e-01
-2.48835683e-01 1.43416405e-01 5.79664886e-01 -7.08589137e-01
2.03111812e-01 8.07277977e-01 -1.23808697e-01 -1.31529200e+00
-5.27171195e-01 -4.89141613e-01 -3.70340437e-01 4.95748669e-02
7.47453034e-01 4.69621159e-02 -1.51474521e-01 1.07592392e+00
3.04690808e-01 2.71543354e-01 -1.30760059e-01 1.18471587e+00
7.71935225e-01 5.30907571e-01 -9.78175849e-02 3.39417756e-01
1.90563083e+00 -5.71093798e-01 -5.97499311e-01 -7.00867623e-02
1.08272113e-01 -5.33753574e-01 8.58539701e-01 8.95460546e-01
-1.18687522e+00 -1.07593155e+00 -9.75562155e-01 2.24223703e-01
-3.96448255e-01 1.42646730e-01 9.39050913e-01 5.22608638e-01
-1.29269943e-01 9.68375087e-01 -1.04345930e+00 -1.65118143e-01
1.70581669e-01 -8.13239887e-02 -6.09382272e-01 1.78439632e-01
-1.10693991e+00 6.77138388e-01 7.52345681e-01 -3.96262743e-02
-9.18020487e-01 -1.01328528e+00 -1.04523396e+00 2.66025126e-01
6.10401511e-01 -2.46787295e-01 9.78876233e-01 -8.77470553e-01
-1.34973764e+00 8.31734598e-01 4.11868542e-01 -8.97360742e-01
4.16195273e-01 -4.46178794e-01 -6.23139560e-01 -3.45399268e-02
1.12881549e-01 -1.40189365e-01 3.36036056e-01 -7.54165709e-01
-1.07266569e+00 -5.08944988e-01 -1.38997376e-01 3.71908657e-02
-2.53174722e-01 6.12141863e-02 -7.40034342e-01 -5.77705204e-01
1.00587189e-01 -7.58647144e-01 -2.45378718e-01 -7.17295647e-01
-1.16191328e-01 -2.98022777e-01 -1.38055146e-01 -5.57266355e-01
1.43719804e+00 -2.45620942e+00 4.03872699e-01 4.05429482e-01
3.40275973e-01 -5.04791439e-02 3.03968489e-01 8.15816760e-01
7.64603838e-02 -7.95678496e-01 1.68733194e-01 -3.43667641e-02
2.37518921e-01 4.11357969e-01 -3.90065610e-02 5.44027746e-01
-2.49861091e-01 6.67698205e-01 -9.02809918e-01 -3.48292470e-01
4.21519488e-01 2.34404713e-01 -5.46734214e-01 3.26346934e-01
1.01141343e-02 9.01026353e-02 -8.30852762e-02 3.15312743e-01
2.49856621e-01 4.51890886e-01 3.23072523e-01 -2.79342294e-01
-5.12412846e-01 1.49110248e-02 -1.70808482e+00 1.87256002e+00
-4.09477443e-01 8.31343055e-01 -2.09766418e-01 -1.25350022e+00
1.17562819e+00 1.04484903e-02 9.34951901e-01 -6.65580988e-01
3.40601146e-01 -1.59141868e-01 -1.54091939e-01 -7.72550106e-01
1.07581758e+00 -2.23878875e-01 -7.32572496e-01 4.81945217e-01
5.48184693e-01 2.53770314e-02 8.58325005e-01 -1.13241903e-01
9.98412192e-01 1.44770741e-01 2.31070518e-01 1.38137102e-01
1.83065847e-01 3.49824607e-01 6.51066780e-01 7.52778590e-01
-3.85852135e-03 6.48200452e-01 5.61560929e-01 -1.22356975e+00
-9.74578202e-01 -1.06343925e+00 4.19371635e-01 1.28732109e+00
2.22365204e-02 -8.20661843e-01 -7.11325705e-01 3.78535651e-02
2.68326461e-01 2.54951596e-01 -9.99099970e-01 -3.79918575e-01
-7.92809367e-01 -5.78446746e-01 8.29914868e-01 6.25907063e-01
6.89284578e-02 -1.17267120e+00 -1.13149321e+00 6.30813181e-01
-4.32063311e-01 -1.09523618e+00 -9.56503823e-02 2.84235716e-01
-3.37393343e-01 -1.31448531e+00 -4.21203166e-01 -4.40649897e-01
-3.47175568e-01 -1.90929592e-01 1.59404969e+00 -2.59154499e-01
-6.48687661e-01 9.36793834e-02 -5.72990239e-01 -6.03903413e-01
-3.17888230e-01 5.27449958e-02 2.64949590e-01 5.14209867e-01
7.68130660e-01 -6.28732920e-01 -3.72460932e-01 2.81749278e-01
-7.75921822e-01 -2.49272957e-01 4.31316286e-01 4.79691982e-01
8.67443800e-01 2.59284407e-01 -3.48473579e-01 -6.18254185e-01
4.53967631e-01 -6.95970476e-01 -3.09007227e-01 -1.38768077e-01
-6.77347034e-02 -3.23252797e-01 6.08241856e-01 -4.89104420e-01
-5.23109376e-01 1.52935863e-01 -4.03468966e-01 -5.47017932e-01
-7.40950406e-01 4.79688734e-01 2.86082476e-02 6.04930818e-01
9.54778433e-01 5.41115940e-01 -2.57252306e-01 -9.68208790e-01
2.91478723e-01 4.22825128e-01 1.25157714e+00 -6.59937918e-01
5.44329047e-01 7.01992154e-01 -2.90486127e-01 -9.07348931e-01
-7.55577624e-01 -1.05607200e+00 -7.59216189e-01 -7.25363970e-01
1.04799914e+00 -8.76545906e-01 -1.42553067e+00 4.91722226e-01
-7.62904823e-01 -6.57248721e-02 -1.07844830e+00 1.02502918e+00
-6.69280708e-01 8.83611441e-02 -6.14770591e-01 -5.84541857e-01
1.08343892e-01 -7.49756575e-01 6.29440844e-01 4.06151295e-01
-4.83825237e-01 -5.22919238e-01 7.94681251e-01 7.43503869e-01
-2.57554233e-01 5.48864186e-01 2.83050358e-01 -8.15769136e-01
4.33353670e-02 -5.46078920e-01 3.56502354e-01 9.92414653e-02
-3.74259472e-01 -7.89348781e-02 -6.40938520e-01 1.12137452e-01
-3.81902218e-01 4.90052104e-02 7.65002728e-01 4.73695576e-01
1.08834350e+00 6.52453825e-02 1.35569662e-01 8.24714720e-01
1.21467566e+00 -2.05366239e-01 4.93155509e-01 4.71453667e-01
6.15816653e-01 6.29854083e-01 5.73895454e-01 9.04137790e-01
5.01362562e-01 9.42315876e-01 2.99674183e-01 1.33008793e-01
5.42932302e-02 -6.34432614e-01 3.40800673e-01 8.75062227e-01
-8.93868029e-01 1.29498795e-01 -7.55754054e-01 8.39172959e-01
-1.64370716e+00 -1.48003197e+00 -4.40706164e-01 2.02089477e+00
4.32986528e-01 1.58977732e-01 1.00621700e+00 8.14303577e-01
5.06467104e-01 1.96076617e-01 7.34952688e-02 -6.69911683e-01
-6.07484654e-02 6.94958806e-01 6.49257660e-01 -9.91647989e-02
-9.97162759e-01 9.15900469e-01 5.59107494e+00 1.01587272e+00
-9.06116307e-01 1.43086821e-01 -1.97472543e-01 -5.24471462e-01
2.13817835e-01 -1.55503348e-01 -6.48814201e-01 5.81301093e-01
1.00928247e+00 -8.01451132e-02 3.81749094e-01 7.66827285e-01
1.70071408e-01 -3.02083548e-02 -9.54159439e-01 1.03095305e+00
2.58175075e-01 -1.57474542e+00 -3.97429377e-01 -1.28490075e-01
2.01748937e-01 -6.22856915e-02 -2.32106373e-01 6.07707381e-01
5.16867638e-01 -9.11923528e-01 1.26376855e+00 9.26233530e-01
1.69941649e-01 -1.22425461e+00 9.17993069e-01 2.91793704e-01
-1.24044347e+00 -4.68941957e-01 -1.54806808e-01 -6.45571053e-01
3.02128345e-01 3.40308309e-01 -2.95163393e-01 6.88605249e-01
1.33534980e+00 1.01251960e+00 -4.55112040e-01 9.43597198e-01
7.80399740e-02 7.59298563e-01 -2.30784595e-01 -1.27515614e-01
3.19788694e-01 -2.95439094e-01 5.11157513e-01 1.33815134e+00
2.13313118e-01 5.01761027e-02 3.36316586e-01 8.29080701e-01
2.98285842e-01 3.66306692e-01 -4.44017053e-01 -1.50383890e-01
1.19187854e-01 1.31173408e+00 -6.64184272e-01 -3.99708062e-01
6.80267019e-03 6.77227795e-01 8.54463130e-02 -2.14309067e-01
-6.72383726e-01 -3.72179568e-01 1.07006347e+00 4.57221210e-01
3.05406481e-01 -2.77395546e-01 -9.76029709e-02 -9.65419114e-01
-1.04096122e-01 -1.07682228e+00 7.54334390e-01 -4.18036848e-01
-9.53889847e-01 5.53183496e-01 -6.56280592e-02 -1.52830327e+00
-3.45217764e-01 -3.37242156e-01 -6.11585081e-01 4.41069990e-01
-5.95251679e-01 -8.45555127e-01 -4.96905774e-01 7.28560507e-01
6.61451876e-01 -4.79743630e-01 5.59083402e-01 4.62620527e-01
-5.24152696e-01 3.50875914e-01 -2.21545640e-02 5.06029248e-01
2.43772253e-01 -1.26231086e+00 2.91208029e-01 4.73457009e-01
1.04277313e+00 1.12215519e-01 9.47337866e-01 -3.38084489e-01
-1.40000737e+00 -5.81985295e-01 7.72634029e-01 -3.51397604e-01
8.11668277e-01 -4.19450849e-01 -8.07867944e-01 4.21686143e-01
-1.34029567e-01 -1.57975122e-01 1.11443985e+00 4.72985685e-01
-8.36560875e-02 -4.33368593e-01 -3.76920640e-01 3.97222400e-01
7.86500692e-01 -4.76223320e-01 -8.36346090e-01 8.85351077e-02
-5.17367721e-02 -5.10025918e-01 -1.38987815e+00 -1.55499235e-01
8.86758268e-01 -1.32216668e+00 1.28779113e+00 -8.07988405e-01
6.85718179e-01 -8.78707021e-02 -4.20078263e-03 -1.45829749e+00
-2.34901056e-01 -2.50956148e-01 1.37464494e-01 8.03066373e-01
-3.70058030e-01 3.52089763e-01 9.53739047e-01 -1.14481777e-01
-3.11005712e-01 -2.98906535e-01 -8.37508619e-01 -5.78578770e-01
-7.56173730e-02 -1.15287054e+00 9.10716772e-01 6.82680309e-01
1.00157201e-01 -2.22131789e-01 -7.07211971e-01 -7.17541054e-02
5.02040863e-01 2.68778175e-01 1.13941050e+00 -1.43225038e+00
-9.19408023e-01 -6.69883609e-01 -1.26377571e+00 -5.74149430e-01
3.84479798e-02 -9.31566775e-01 -1.48642495e-01 -9.00120854e-01
-1.04184911e-01 -8.68132934e-02 -3.84881794e-01 -3.05331834e-02
2.10462049e-01 6.46144569e-01 4.48754460e-01 3.44833098e-02
-6.59322679e-01 1.07504860e-01 8.63184094e-01 -1.61071002e-01
-2.88555950e-01 4.05601323e-01 -2.70460457e-01 7.97924459e-01
5.64975917e-01 -4.80752021e-01 3.92595589e-01 -2.39935532e-01
5.64215302e-01 3.12543303e-01 5.04197657e-01 -1.58275175e+00
3.14005733e-01 -2.92462826e-01 2.22017825e-01 -7.65474260e-01
4.23686594e-01 -6.39789820e-01 4.26777005e-01 6.57484293e-01
-3.36788148e-01 4.17644754e-02 1.50982812e-01 5.69180667e-01
-7.39612699e-01 -1.66531771e-01 3.28199834e-01 -2.64773130e-01
-1.04890716e+00 3.10201883e-01 -1.06310225e+00 6.09470457e-02
9.22555208e-01 -3.63884747e-01 6.33827388e-01 -1.71278223e-01
-1.16651428e+00 -1.73354015e-01 9.40330699e-02 5.79125762e-01
3.25826496e-01 -1.32563245e+00 -1.07716572e+00 4.75798488e-01
4.57817495e-01 -4.44749832e-01 9.02638018e-01 6.15455747e-01
-8.91282976e-01 -2.41839573e-01 -8.36488307e-01 -6.09265983e-01
-1.22739530e+00 5.18465400e-01 2.56734699e-01 -3.65016252e-01
-1.02758372e+00 6.56846344e-01 -4.51237321e-01 -2.60538161e-01
-5.02045490e-02 -1.99147761e-01 -8.35822046e-01 6.33430302e-01
6.67852581e-01 8.81095350e-01 2.89843202e-01 -8.70898068e-01
-1.70369759e-01 4.36145574e-01 3.44765276e-01 -2.84835268e-02
1.57071853e+00 3.52463424e-01 2.07148969e-01 8.56710851e-01
7.88899720e-01 1.10059537e-01 -1.03624296e+00 -3.20848554e-01
1.40623987e-01 -5.16661823e-01 -1.53077647e-01 -2.10757226e-01
-1.37765133e+00 9.76838768e-01 5.41151226e-01 2.21651271e-01
8.47806156e-01 -2.39535477e-02 1.08379316e+00 1.94346234e-01
3.31142545e-01 -1.41842413e+00 -1.38477221e-01 5.60286641e-01
3.69326174e-01 -8.10192883e-01 -2.26174325e-01 1.75464064e-01
-9.74210501e-01 1.06072700e+00 1.99648008e-01 -4.52061594e-01
6.02877557e-01 4.21896756e-01 2.27880761e-01 -5.08612275e-01
-5.28517306e-01 -4.85850155e-01 4.15326387e-01 4.35236841e-01
1.86288089e-01 5.42745590e-01 -3.87648761e-01 1.40870452e+00
-9.62320089e-01 5.75956255e-02 5.45644879e-01 5.87363958e-01
-2.83235937e-01 -9.62561548e-01 -6.10673845e-01 2.03664362e-01
-5.29668748e-01 4.13903564e-01 -9.24095511e-02 7.88161337e-01
9.25737560e-01 7.41771460e-01 6.59720063e-01 -8.73578668e-01
1.01474094e+00 -2.82361954e-02 4.42080945e-01 -5.10221899e-01
-1.57590914e+00 -2.98569322e-01 -1.86992154e-01 -9.18732643e-01
-9.60006267e-02 -8.56428504e-01 -7.60606110e-01 -7.42427826e-01
2.61079609e-01 5.08999825e-01 7.25041330e-01 8.40905666e-01
-1.12531513e-01 7.42654681e-01 6.09917462e-01 -1.02254987e+00
3.86822149e-02 -8.96447420e-01 -1.24414623e+00 1.09045410e+00
-2.08966136e-01 -8.12631428e-01 1.19696133e-01 1.58265814e-01] | [7.2382330894470215, 0.26829901337623596] |
b9d4ac20-69b6-4da7-b670-3e1619e2f460 | scaling-strategies-for-on-device-low | 2303.03005 | null | https://arxiv.org/abs/2303.03005v1 | https://arxiv.org/pdf/2303.03005v1.pdf | Scaling strategies for on-device low-complexity source separation with Conv-Tasnet | Recently, several very effective neural approaches for single-channel speech separation have been presented in the literature. However, due to the size and complexity of these models, their use on low-resource devices, e.g. for hearing aids, and earphones, is still a challenge and established solutions are not available yet. Although approaches based on either pruning or compressing neural models have been proposed, the design of a model architecture suitable for a certain application domain often requires heuristic procedures not easily portable to different low-resource platforms. Given the modular nature of the well-known Conv-Tasnet speech separation architecture, in this paper we consider three parameters that directly control the overall size of the model, namely: the number of residual blocks, the number of repetitions of the separation blocks and the number of channels in the depth-wise convolutions, and experimentally evaluate how they affect the speech separation performance. In particular, experiments carried out on the Libri2Mix show that the number of dilated 1D-Conv blocks is the most critical parameter and that the usage of extra-dilation in the residual blocks allows reducing the performance drop. | ['Alessio Brutti', 'Daniele Falavigna', 'Francesco Paissan', 'Mohamed Nabih Ali'] | 2023-03-06 | null | null | null | null | ['speech-separation'] | ['speech'] | [ 1.24013811e-01 -2.02765599e-01 2.18302056e-01 -1.73122257e-01
-1.09280184e-01 -1.65916592e-01 1.89722478e-01 1.95155680e-01
-8.68193924e-01 4.35637414e-01 -1.05443478e-01 -5.77676892e-01
-3.21998000e-01 -3.57521027e-01 -3.26372266e-01 -8.11719120e-01
1.61152184e-02 3.32378410e-02 4.10931826e-01 -9.23246816e-02
1.92787405e-02 7.16686606e-01 -1.81934559e+00 6.16813861e-02
9.47283506e-01 1.12046337e+00 5.27842581e-01 6.09114468e-01
-2.42969900e-01 3.12398821e-01 -7.85282493e-01 -1.39311180e-01
2.75036365e-01 -4.14541721e-01 -2.90862441e-01 5.19053489e-02
1.76539242e-01 -2.11424306e-01 -2.54089564e-01 1.00412023e+00
8.40087175e-01 1.45168081e-01 5.38390815e-01 -7.01876760e-01
1.28292486e-01 7.50884056e-01 -2.99937576e-01 5.55393517e-01
-2.88608372e-01 -1.42242938e-01 6.70196116e-01 -6.91675663e-01
1.29176095e-01 1.03130531e+00 4.38201159e-01 4.40240562e-01
-1.24687064e+00 -6.98412240e-01 2.38261431e-01 2.33036220e-01
-1.31671202e+00 -9.04178679e-01 6.27558887e-01 -1.94516689e-01
1.17519069e+00 2.56447792e-01 6.25727773e-01 7.32079327e-01
-2.21002892e-01 4.67870563e-01 1.00476980e+00 -7.70627677e-01
4.90576595e-01 3.62476408e-01 2.32249469e-01 2.61361331e-01
4.84527260e-01 -2.82603383e-01 -5.20938277e-01 1.22660562e-01
7.10340500e-01 -2.70313352e-01 -4.43647712e-01 -2.37847805e-01
-4.56564695e-01 5.33195436e-01 3.46341878e-01 9.05047834e-01
-2.25744277e-01 -2.65658535e-02 3.76082957e-01 2.29034856e-01
3.86465579e-01 4.62379366e-01 -3.74392271e-01 -1.14843756e-01
-1.34541035e+00 1.51139982e-02 6.96140885e-01 6.70117080e-01
4.07057106e-01 2.64279485e-01 2.29842246e-01 1.02683532e+00
1.24383144e-01 1.14436857e-01 7.05637276e-01 -3.20366949e-01
6.30194783e-01 5.37092745e-01 -2.62569577e-01 -8.79827559e-01
-4.12559688e-01 -7.13101685e-01 -8.64709318e-01 8.32914934e-02
5.93112588e-01 -4.03732240e-01 -1.00311899e+00 1.38906682e+00
-4.15376425e-02 5.66917695e-02 1.42894477e-01 8.97489429e-01
7.10557163e-01 4.98741210e-01 -4.82216068e-02 -2.05627427e-01
1.32492828e+00 -1.05140805e+00 -7.77568042e-01 -6.13174915e-01
3.23038608e-01 -9.19530272e-01 6.64519012e-01 4.25672114e-01
-1.35005915e+00 -6.20277047e-01 -1.32382917e+00 2.89936624e-02
-4.72399861e-01 5.80995858e-01 3.72661769e-01 8.45204473e-01
-9.40317571e-01 5.70012987e-01 -8.98104250e-01 -1.77644476e-01
1.57214373e-01 7.44361222e-01 -2.27115452e-01 6.82061836e-02
-9.19923782e-01 8.21059883e-01 4.80282903e-01 6.57535613e-01
-5.06024778e-01 -2.84501791e-01 -4.01787877e-01 7.87641644e-01
2.86691397e-01 -3.42928916e-01 8.06729019e-01 -9.85400081e-01
-1.53023589e+00 3.92793417e-01 -8.47526416e-02 -6.74308658e-01
3.39013696e-01 -2.75490910e-01 -4.15226400e-01 1.65266767e-01
-6.57165945e-01 6.41636908e-01 1.21674597e+00 -9.25295413e-01
-6.36271417e-01 -3.12598079e-01 2.36666620e-01 2.09963083e-01
-8.77285302e-01 1.64337739e-01 -5.40225506e-01 -6.18939757e-01
2.76121229e-01 -8.29467833e-01 -3.77902061e-01 -3.29314113e-01
-1.64450914e-01 1.72759354e-01 5.72605133e-01 -7.42557704e-01
1.45194435e+00 -2.64640951e+00 2.80439287e-01 2.43721783e-01
-2.09242683e-02 8.04516792e-01 -9.47135966e-03 1.61639750e-01
-7.80461878e-02 -5.58611676e-02 -1.91934928e-01 -6.12750232e-01
-2.53543675e-01 2.17430834e-02 -5.57463951e-02 2.47914046e-01
-1.02261193e-01 1.30275980e-01 -2.72630692e-01 -1.59717143e-01
3.51753473e-01 8.19358110e-01 -6.53749287e-01 2.04941496e-01
1.84073597e-01 1.11936189e-01 -1.14479974e-01 5.35335131e-02
8.28413367e-01 1.72075152e-01 2.56241173e-01 -9.00300145e-02
-3.22668344e-01 6.89624131e-01 -1.35736859e+00 1.39996374e+00
-6.79281890e-01 7.98723936e-01 5.93578577e-01 -9.29819524e-01
9.77208436e-01 5.74667096e-01 9.67280567e-03 -7.00483441e-01
4.82062221e-01 7.05724299e-01 4.78964329e-01 -1.09597147e-01
4.55518633e-01 -2.94830855e-02 4.59191054e-01 -4.60682586e-02
-1.90664344e-02 2.95685674e-03 4.26929086e-01 -2.68311322e-01
8.52218270e-01 -4.11532640e-01 4.37568128e-02 -3.82202238e-01
6.70109272e-01 -4.87608254e-01 3.52475345e-01 2.95406044e-01
7.52895996e-02 6.24887049e-01 4.45504099e-01 -1.53399810e-01
-7.75423825e-01 -4.94482011e-01 -1.69410706e-01 7.46056855e-01
-5.86058907e-02 -3.89557391e-01 -1.19814754e+00 3.60227376e-03
-3.96787614e-01 4.31127191e-01 1.81992576e-02 -3.48544307e-02
-7.86213338e-01 -7.75481641e-01 5.94726086e-01 3.91312510e-01
4.92017835e-01 -9.24481213e-01 -1.09045696e+00 4.28979337e-01
2.41325095e-01 -1.33885121e+00 -1.49064824e-01 7.95260668e-01
-1.12688315e+00 -7.39791930e-01 -8.30117643e-01 -9.76220846e-01
7.58843958e-01 3.83156061e-01 6.01466894e-01 1.80642903e-01
-1.84795111e-01 -2.48197958e-01 -2.73720682e-01 -3.44405591e-01
-1.44680545e-01 4.78166789e-01 5.52828005e-03 8.27140659e-02
1.98666185e-01 -6.49923205e-01 -4.84484285e-01 2.94534564e-01
-9.89028871e-01 -4.80055399e-02 6.48034275e-01 6.48841083e-01
2.97449529e-01 5.25240183e-01 4.41004008e-01 -4.72845048e-01
7.98246026e-01 1.80100590e-01 -7.31700063e-01 3.93598117e-02
-2.71335095e-01 5.63193373e-02 8.18544924e-01 -5.21072447e-01
-8.86565745e-01 -1.72448173e-01 -3.86267781e-01 -4.07706589e-01
-2.65077055e-01 3.53456527e-01 -3.68019402e-01 -7.94012193e-03
1.92890912e-01 -3.07830740e-02 -1.00001819e-01 -7.27133214e-01
-1.01209566e-01 8.49982858e-01 5.01863919e-02 -2.09822208e-02
3.79863173e-01 9.98943001e-02 -2.56712288e-01 -1.28889132e+00
-2.71660179e-01 -3.89812022e-01 -4.39289927e-01 1.03385672e-01
8.16508770e-01 -8.64861488e-01 -4.95836675e-01 6.32252276e-01
-1.15924132e+00 -2.52994925e-01 -3.36725800e-03 6.67669058e-01
-9.49022844e-02 1.46906465e-01 -3.81802708e-01 -7.98822105e-01
-4.62709188e-01 -1.65421236e+00 3.56622517e-01 3.96932065e-01
-2.33770832e-01 -7.44364083e-01 -5.36133170e-01 2.80801743e-01
8.53692234e-01 -4.48601216e-01 1.20110524e+00 -5.85609376e-01
-4.55049813e-01 -6.88006282e-02 -5.53849377e-02 8.23161125e-01
9.39129815e-02 -1.59819841e-01 -1.17559612e+00 -2.74761051e-01
3.44145060e-01 1.90579772e-01 9.11348224e-01 6.45942450e-01
1.03861928e+00 -1.29828781e-01 -2.51565069e-01 5.27477562e-01
1.12462866e+00 4.09533054e-01 6.14404142e-01 2.21699312e-01
3.77847284e-01 6.60509348e-01 8.50908160e-02 1.43921420e-01
-2.41519570e-01 8.39305997e-01 3.02749038e-01 -1.84212774e-01
-3.46551389e-01 2.58342296e-01 1.77558064e-01 1.17252994e+00
-1.10084161e-01 -2.87232757e-01 -7.08342016e-01 5.28627396e-01
-1.38792241e+00 -5.87063432e-01 1.29649311e-01 2.42098284e+00
5.27381659e-01 4.83173728e-01 -7.09067360e-02 7.43149757e-01
6.60673022e-01 2.11155459e-01 -1.91537231e-01 -7.00686276e-01
-1.67092010e-01 5.06687641e-01 5.24295568e-01 4.50048029e-01
-7.79001653e-01 6.94548845e-01 5.70955420e+00 9.24720824e-01
-1.31667078e+00 -2.84595112e-03 4.64482367e-01 -6.38561487e-01
1.05046835e-02 -1.74605712e-01 -9.57341671e-01 4.77179646e-01
1.16740572e+00 2.94376463e-01 6.15670085e-01 6.55134320e-01
9.79997069e-02 -2.46344373e-01 -9.38401222e-01 1.14876819e+00
-6.28480017e-02 -8.85482252e-01 -1.91558465e-01 7.71616623e-02
3.31977844e-01 -1.20718665e-01 7.10339621e-02 -1.06657716e-02
-5.68450391e-01 -8.40606153e-01 9.59591687e-01 -1.06580392e-01
3.46580803e-01 -9.58116710e-01 7.77141929e-01 2.32980028e-01
-1.00667775e+00 -3.54439884e-01 -4.59787220e-01 1.25642624e-02
4.78064828e-02 7.49549568e-01 -6.06025994e-01 1.60341501e-01
7.35649228e-01 3.87909561e-02 -4.87859905e-01 1.22332418e+00
-2.47817189e-01 5.50174236e-01 -5.53035557e-01 -9.73454937e-02
3.37217808e-01 -8.44248310e-02 3.85026038e-01 1.19041181e+00
3.81133229e-01 -1.01596154e-01 -4.61737722e-01 4.49573457e-01
1.06186360e-01 7.47755170e-02 -6.29644245e-02 -7.80195817e-02
4.88663584e-01 1.07617640e+00 -9.05064523e-01 9.11630988e-02
-4.13608313e-01 5.87534606e-01 2.64489293e-01 4.44490284e-01
-6.45552695e-01 -5.15124321e-01 9.14122045e-01 4.05045688e-01
7.98931658e-01 -4.65667129e-01 -4.19292420e-01 -7.09721923e-01
3.05164099e-01 -1.04443240e+00 -1.23375833e-01 -3.27261984e-01
-4.49866563e-01 1.11941564e+00 -8.47642720e-02 -8.68837118e-01
-1.16317250e-01 -8.11271966e-01 -3.84676367e-01 9.47410703e-01
-1.48086619e+00 -4.59251583e-01 -4.53104526e-02 3.32208246e-01
6.12028241e-01 -3.32050800e-01 7.37874329e-01 7.28562415e-01
-8.81869435e-01 7.12281466e-01 1.15660489e-01 -7.84727186e-02
3.77492815e-01 -6.21972859e-01 3.45737815e-01 9.49470818e-01
1.78751975e-01 8.70886266e-01 7.70153165e-01 -1.27555914e-02
-1.09579325e+00 -2.15936095e-01 7.00833201e-01 5.96599400e-01
3.11544180e-01 -7.02710509e-01 -9.77208197e-01 7.23148212e-02
2.45154291e-01 -1.55993655e-01 5.99923551e-01 6.81331158e-02
1.76302972e-03 -3.98980498e-01 -7.49409795e-01 7.16858745e-01
7.59153783e-01 -3.34816486e-01 -2.70301223e-01 -3.69587630e-01
3.60897362e-01 -3.92483950e-01 -3.89942646e-01 1.81232378e-01
4.16919380e-01 -1.25180686e+00 8.40906084e-01 -8.29915926e-02
7.16905817e-02 -2.70028561e-01 7.55089894e-02 -1.30296826e+00
-6.15959764e-02 -3.94632906e-01 -3.57056372e-02 1.09071970e+00
6.15730703e-01 -5.86107016e-01 6.83667183e-01 6.10638142e-01
-2.07550138e-01 -8.57133567e-01 -1.27268934e+00 -6.98117435e-01
-2.47218996e-01 -2.29342923e-01 5.47988832e-01 2.22170025e-01
-2.50247717e-01 3.57722014e-01 -1.62745535e-01 1.66500315e-01
1.04562994e-02 -2.06613913e-01 4.71235275e-01 -1.16644061e+00
-4.87099141e-01 -6.72387779e-01 -3.90741527e-01 -1.15484440e+00
-5.89386374e-02 -3.83959621e-01 5.79925105e-02 -1.33663380e+00
-3.47633988e-01 -7.53316104e-01 -1.84694782e-01 2.99195737e-01
-8.63718763e-02 -1.31104782e-01 3.66021514e-01 1.55783091e-02
7.31821358e-02 3.60908240e-01 7.49512374e-01 -9.65665430e-02
-5.21852791e-01 8.03113058e-02 -4.20060098e-01 8.61186981e-01
7.46355712e-01 -4.40727323e-01 -5.26298344e-01 -8.63243461e-01
-2.91741081e-02 -2.00583935e-01 1.20751057e-02 -1.54070640e+00
5.17767906e-01 3.48668903e-01 2.80870199e-02 -3.86682540e-01
7.47037470e-01 -1.20208097e+00 1.29813507e-01 3.91366959e-01
-1.03232093e-01 4.25451249e-03 6.25042319e-01 1.46349430e-01
-5.51921010e-01 -8.19597840e-01 8.00000370e-01 1.06618598e-01
-4.28364813e-01 -2.36525580e-01 -4.09871787e-01 -2.57660151e-01
7.24937320e-01 -4.30427462e-01 7.04747289e-02 -1.30535543e-01
-6.38921380e-01 -2.82011300e-01 1.72274247e-01 3.91101003e-01
4.00910914e-01 -6.56661630e-01 -3.55124325e-01 5.82790911e-01
-4.81240183e-01 1.11421891e-01 3.79372686e-01 7.92311370e-01
-5.35684288e-01 5.90693831e-01 -2.82439739e-01 -3.23735029e-01
-1.56642878e+00 3.56348574e-01 4.47531044e-01 -1.58528432e-01
-5.10075152e-01 8.88102949e-01 6.00635782e-02 2.39185587e-01
7.73570120e-01 -5.84150672e-01 -2.66720891e-01 2.39394039e-01
6.76065445e-01 5.66167593e-01 5.84969938e-01 -4.53191876e-01
-3.49607021e-01 4.47896153e-01 -2.28314579e-01 -4.06770445e-02
1.41154325e+00 -5.93767613e-02 -1.41986236e-01 1.77701741e-01
8.74179244e-01 -5.79907522e-02 -1.05138862e+00 5.13754226e-02
5.23271598e-03 -2.35406369e-01 3.88567895e-01 -2.95533180e-01
-1.29502058e+00 1.22627831e+00 6.47766531e-01 3.32859874e-01
1.49315870e+00 -4.09647346e-01 6.21559381e-01 2.24799141e-01
2.36697897e-01 -1.09280419e+00 -2.39486679e-01 5.02153277e-01
6.18494451e-01 -5.33698082e-01 -9.69419777e-02 -6.35006785e-01
-2.18082353e-01 1.06110585e+00 4.68039364e-01 1.43754750e-01
5.69726765e-01 6.79411709e-01 5.71059063e-02 1.92433193e-01
-3.92208159e-01 -2.05724001e-01 1.08646072e-01 2.04333395e-01
6.04234576e-01 -6.57153353e-02 -5.56223571e-01 5.50371170e-01
-2.31533006e-01 -1.65554911e-01 3.20623875e-01 6.66642785e-01
-4.92481321e-01 -1.17135704e+00 -4.76875156e-01 2.79159129e-01
-5.44879615e-01 -2.07154512e-01 -1.46431506e-01 4.80108380e-01
3.04371357e-01 1.16814315e+00 2.94090092e-01 -9.50353146e-02
4.45534170e-01 7.06058517e-02 4.87765878e-01 -5.31042635e-01
-9.07075703e-01 3.03441852e-01 2.57159881e-02 6.22427836e-02
-4.46921706e-01 -3.90643418e-01 -1.07635415e+00 -1.89200401e-01
-5.78460753e-01 1.58323273e-01 1.01102066e+00 8.78490031e-01
3.84003997e-01 8.08647633e-01 6.13646321e-02 -8.72390032e-01
-4.26519781e-01 -1.15569961e+00 -6.27457798e-01 -5.97666018e-03
2.79397845e-01 -7.13008821e-01 -4.98307139e-01 -2.20568478e-01] | [14.84547233581543, 5.855870246887207] |
4a3d868d-1f86-4492-b896-8a47335ddc57 | chrono-at-semeval-2018-task-6-a-system-for | null | null | https://aclanthology.org/S18-1012 | https://aclanthology.org/S18-1012.pdf | Chrono at SemEval-2018 Task 6: A System for Normalizing Temporal Expressions | Temporal information extraction is a challenging task. Here we describe Chrono, a hybrid rule-based and machine learning system that identifies temporal expressions in text and normalizes them into the SCATE schema. After minor parsing logic adjustments, Chrono has emerged as the top performing system for SemEval 2018 Task 6: Parsing Time Normalizations. | ['Luke Maffey', 'Nicholas Morgan', 'Bridget McInnes', 'Amy Olex'] | 2018-06-01 | null | null | null | semeval-2018-6 | ['timex-normalization', 'temporal-information-extraction'] | ['natural-language-processing', 'natural-language-processing'] | [ 7.41819069e-02 5.33334985e-02 -9.34374928e-01 -5.55625141e-01
-4.16683197e-01 -8.56423318e-01 9.62895274e-01 6.42577767e-01
-8.64833951e-01 6.17873847e-01 4.68040794e-01 -3.28662872e-01
-3.00375730e-01 -4.19551581e-01 -2.34534740e-01 -1.91540286e-01
-3.08992535e-01 6.39684498e-01 2.17830583e-01 -3.44342381e-01
-7.03529269e-02 4.70339000e-01 -1.07262230e+00 3.00364226e-01
2.94132739e-01 8.57282162e-01 -1.09689045e+00 6.68060064e-01
-2.09241778e-01 1.23936105e+00 -4.91839528e-01 -5.64938664e-01
1.02322444e-01 -9.04627666e-02 -1.06908560e+00 -9.20706391e-01
2.57278651e-01 1.12784594e-01 -9.05571282e-01 5.74140668e-01
2.08171710e-01 3.77787888e-01 5.36980927e-01 -1.60984325e+00
-2.28786901e-01 1.53709269e+00 -3.62807244e-01 8.66492927e-01
3.31365108e-01 -1.54563680e-01 9.22626853e-01 -2.25565240e-01
1.07369614e+00 1.24581373e+00 8.18779469e-01 7.19318211e-01
-1.35839558e+00 -6.11714900e-01 1.47039399e-01 4.29614723e-01
-1.14959824e+00 -4.41002071e-01 5.40605187e-01 -3.54686797e-01
1.75331914e+00 2.74022013e-01 5.68598449e-01 1.61174738e+00
5.10056794e-01 6.90370917e-01 9.61321831e-01 -5.11611164e-01
2.24074379e-01 -8.46839845e-01 8.05583000e-01 2.51821697e-01
1.25229883e-03 6.49896218e-04 -9.97753620e-01 8.52415413e-02
-4.64910418e-02 -5.99205256e-01 3.70198339e-01 5.24177969e-01
-1.30691457e+00 3.76490444e-01 -1.07183494e-01 3.83579910e-01
-3.08526576e-01 3.80929857e-01 1.25930059e+00 2.63890684e-01
6.50695205e-01 5.37507772e-01 -8.83752525e-01 -8.22700620e-01
-7.80245662e-01 5.75207710e-01 5.65837145e-01 8.77790809e-01
-1.58906817e-01 -2.12918103e-01 -5.23206294e-01 4.76523131e-01
-2.04412580e-01 3.80546115e-02 4.65134889e-01 -9.78122652e-01
6.17688835e-01 5.94399035e-01 5.82583472e-02 -4.23747838e-01
-9.57884848e-01 1.03587341e-02 -5.04374862e-01 -3.23018074e-01
6.50618196e-01 -3.63321006e-01 -1.06374943e+00 2.26590133e+00
2.45445088e-01 1.49878785e-02 1.38704166e-01 1.91431820e-01
8.16626608e-01 6.26698017e-01 8.12919140e-01 -7.41504192e-01
1.35095906e+00 -7.26401687e-01 -1.49915469e+00 -3.35787386e-01
1.00017643e+00 -5.82869053e-01 6.34644687e-01 3.11494768e-01
-9.32645261e-01 1.82565510e-01 -1.19552147e+00 -4.56689149e-01
-8.54573905e-01 -2.29206696e-01 1.30178547e+00 2.38510326e-01
-7.09958673e-01 8.83938551e-01 -1.43374920e+00 -8.55421185e-01
2.38966376e-01 3.09965700e-01 -4.77938443e-01 5.69004655e-01
-1.60352254e+00 1.39110339e+00 9.69369769e-01 -1.83531642e-02
-7.87782744e-02 -1.01598406e+00 -9.87360835e-01 -4.24856216e-01
5.44130981e-01 -5.39458871e-01 1.88008273e+00 -2.24551093e-02
-1.29740667e+00 1.07782626e+00 -4.35266465e-01 -1.07115781e+00
5.27844071e-01 -2.39369825e-01 -1.13991928e+00 -3.12872052e-01
4.28329647e-01 4.29915816e-01 1.73009545e-01 2.88244992e-01
-8.38142037e-01 -1.71268746e-01 -1.07226551e-01 -2.84310818e-01
-1.97870165e-01 8.06186318e-01 -6.19687319e-01 -9.57843840e-01
1.97786808e-01 -8.31143558e-01 -1.41674876e-01 -5.78738809e-01
-2.42736101e-01 -9.81881857e-01 5.04084051e-01 -7.11185217e-01
2.28425407e+00 -1.73013377e+00 -4.90004718e-02 7.76139125e-02
-1.79535784e-02 -1.84811279e-01 1.54688418e-01 4.12285507e-01
-7.50447512e-01 3.12333643e-01 -1.65802255e-01 -3.23188215e-01
1.60264909e-01 3.50164175e-01 -4.94403064e-01 3.83372545e-01
4.33527939e-02 1.01502395e+00 -1.13367903e+00 -7.38586664e-01
1.88273519e-01 -1.51451781e-01 -2.44546056e-01 -2.07713917e-01
-3.83174896e-01 3.02425832e-01 7.47591108e-02 5.62486410e-01
2.75227934e-01 3.41051102e-01 3.39543223e-01 -2.79374123e-01
-3.77495080e-01 9.93496656e-01 -8.63506615e-01 1.92094219e+00
-2.83272028e-01 7.98827708e-01 -8.15261722e-01 -3.21676493e-01
4.36719865e-01 4.66280937e-01 1.11830604e+00 -1.16511917e+00
4.62389886e-01 -6.51620254e-02 -1.24860987e-01 -5.09832203e-01
4.09387410e-01 -3.76640335e-02 -8.05520773e-01 5.77008247e-01
2.95051783e-01 8.48948583e-02 1.07683837e+00 3.56252253e-01
1.48746383e+00 5.66832125e-01 1.00725996e+00 -2.73005009e-01
2.22453356e-01 2.39193156e-01 1.13171434e+00 2.72797406e-01
-5.24068654e-01 1.37072746e-02 7.48861432e-01 -9.42528784e-01
-1.04031920e+00 -1.00607169e+00 -1.58119887e-01 1.30434203e+00
-4.59140867e-01 -1.33712876e+00 -5.83394170e-01 -7.02321053e-01
-1.81720793e-01 1.07359934e+00 -1.14995205e+00 -1.78517133e-01
-1.10360968e+00 -1.05810261e+00 1.28836894e+00 9.84818697e-01
-1.48284614e-01 -9.61222887e-01 -5.37507117e-01 4.37873989e-01
-6.46291435e-01 -1.50334215e+00 -5.99809706e-01 8.83401394e-01
-6.86224282e-01 -1.22597742e+00 4.33107018e-01 -2.73402005e-01
2.13604961e-02 -6.03878498e-01 1.10705769e+00 -7.27814078e-01
-2.09772646e-01 -7.31629655e-02 -3.61205846e-01 -8.44162643e-01
-3.22228849e-01 3.92954409e-01 4.18118626e-01 -3.59456778e-01
8.87639940e-01 -8.85338247e-01 1.00424271e-02 1.38389831e-02
-6.07086539e-01 9.18735266e-02 -3.43011677e-01 5.30598201e-02
6.06916428e-01 1.18601164e-02 1.72027454e-01 -9.12043512e-01
5.51423788e-01 -3.02142203e-01 -7.20793784e-01 3.84010941e-01
-8.88863742e-01 4.06246185e-01 4.32060421e-01 -5.29062331e-01
-1.10179150e+00 1.04926772e-01 -2.17348970e-02 5.33585027e-02
8.48257393e-02 5.88606775e-01 -7.42378905e-02 5.88381648e-01
9.64704514e-01 -4.92663354e-01 -8.06940198e-01 -3.30438912e-01
6.78396523e-01 1.47724170e-02 1.25573242e+00 -8.14877689e-01
5.44472277e-01 4.66985136e-01 2.51403898e-01 -1.97129458e-01
-1.27944160e+00 -3.95187199e-01 -1.17134440e+00 -3.14091235e-01
1.12409019e+00 -5.85865438e-01 -7.53885806e-01 4.33923632e-01
-1.34513164e+00 -5.74585199e-01 -5.40274560e-01 4.08855677e-01
-4.97877657e-01 -1.68060273e-01 -7.39078879e-01 -5.23103654e-01
-3.71890455e-01 -2.35094264e-01 4.96598661e-01 1.56367868e-01
-1.28475368e+00 -8.67652357e-01 4.96062011e-01 5.47350049e-02
-2.94721443e-02 8.99772286e-01 8.09131742e-01 -7.98226178e-01
3.96545857e-01 -3.00315678e-01 2.23685399e-01 -3.42600226e-01
1.63961619e-01 7.34850705e-01 -6.87886536e-01 4.43534791e-01
-4.11006808e-01 -4.10574973e-02 6.56199574e-01 3.68413448e-01
9.52634275e-01 -2.01033115e-01 -4.31400388e-01 8.19216847e-01
9.49353874e-01 5.23943722e-01 2.71726042e-01 8.08355451e-01
5.01173496e-01 5.65488875e-01 9.62099850e-01 4.99700695e-01
5.21197021e-01 7.52244532e-01 5.43029532e-02 4.67432201e-01
-4.87332419e-02 -3.61741781e-01 4.07700211e-01 8.52483332e-01
-1.32410079e-01 -2.19465196e-01 -1.41596878e+00 6.66192651e-01
-2.22021937e+00 -1.05541992e+00 -6.88982964e-01 1.58921468e+00
9.78381515e-01 6.36186182e-01 2.17304528e-01 2.98347592e-01
4.74424452e-01 4.01340842e-01 -3.17068070e-01 -7.74085104e-01
-5.65783679e-01 5.50217330e-01 9.36982095e-01 1.81143478e-01
-1.78340197e+00 1.39940941e+00 7.35646296e+00 5.65909803e-01
-9.76420999e-01 2.48399481e-01 6.55416399e-02 -3.80398542e-01
1.90210566e-01 1.70254037e-01 -6.59366012e-01 1.88378513e-01
1.67085564e+00 -7.30619848e-01 5.01690865e-01 3.93338352e-01
6.04686439e-01 -3.79372127e-02 -1.67604530e+00 1.11278903e+00
-3.61705661e-01 -1.19839811e+00 -2.65617788e-01 -5.69051981e-01
4.57732379e-01 2.30655447e-01 -4.52033818e-01 4.95105684e-01
5.98756075e-01 -9.19380546e-01 1.41288376e+00 5.26524842e-01
4.66263860e-01 -5.24260700e-01 5.74548721e-01 -8.60712081e-02
-1.49222517e+00 5.26344450e-03 5.76349974e-01 -2.08867714e-01
6.77675009e-01 4.14866358e-01 -4.71476227e-01 6.35294199e-01
9.59238708e-01 1.10056579e+00 -8.47168803e-01 5.92837155e-01
-5.82264423e-01 8.45315814e-01 -6.22403920e-01 3.23666781e-01
1.97302662e-02 1.03829846e-01 7.01528549e-01 1.55812907e+00
-1.72454730e-01 3.98539603e-01 -3.51879895e-01 2.77606964e-01
-2.62082875e-01 9.50847045e-02 -4.09602672e-01 -4.56772447e-01
5.63033640e-01 9.69017446e-01 -1.00829887e+00 -4.36148256e-01
-1.38397992e-01 6.43217325e-01 3.80651146e-01 1.33167654e-02
-1.63414884e+00 -2.50628293e-01 5.75039387e-01 -1.14970066e-01
-3.15519392e-01 -5.21545827e-01 -8.20163548e-01 -1.32441390e+00
3.68162058e-02 -5.42762518e-01 1.39848375e+00 -8.11000884e-01
-8.94280791e-01 3.47460002e-01 6.22218251e-01 -1.00819683e+00
-2.91960984e-01 -4.75956351e-01 -6.16537511e-01 3.06728303e-01
-7.61260629e-01 -1.12683928e+00 7.00142011e-02 5.54252863e-01
3.87567431e-01 2.47873008e-01 6.02284133e-01 6.30177855e-01
-1.01172364e+00 6.71624839e-01 -5.10674655e-01 2.34000117e-01
1.10825622e+00 -1.39568090e+00 1.01130700e+00 1.34480357e+00
-8.56659412e-02 7.94362187e-01 1.11898363e+00 -8.27345610e-01
-1.10436296e+00 -1.50145042e+00 1.74319506e+00 -9.00535822e-01
1.35968614e+00 -3.73797804e-01 -8.05096805e-01 1.30758476e+00
3.76684159e-01 -2.25809380e-01 5.08697867e-01 3.21632981e-01
-9.25176322e-01 -1.90747350e-01 -6.81002975e-01 8.49439979e-01
1.50281692e+00 -6.85217321e-01 -1.11976671e+00 2.72030085e-01
8.50232363e-01 -8.08709562e-01 -1.20026720e+00 7.33893216e-01
4.89927530e-01 -1.10567108e-01 5.39164245e-01 -1.20565343e+00
3.11924368e-01 -4.15793210e-01 -1.72091406e-02 -6.64302051e-01
-2.35967278e-01 -1.34690809e+00 -5.42099357e-01 1.71366346e+00
7.01859534e-01 -4.59134996e-01 4.77757663e-01 9.03111160e-01
-1.32227242e-01 -2.14650303e-01 -1.11259902e+00 -9.46352959e-01
-4.74259369e-02 -1.25084734e+00 6.08525038e-01 1.35029113e+00
5.49469113e-01 4.51139569e-01 -2.50915200e-01 -2.19546005e-01
3.58910561e-01 -3.82078409e-01 3.73609811e-01 -1.11319435e+00
4.58904445e-01 -7.40978062e-01 -4.47770208e-02 -1.40281260e-01
4.67976451e-01 -1.10691285e+00 -1.09425671e-01 -1.31317294e+00
-1.11058161e-01 1.43589020e-01 -2.68420249e-01 1.24058294e+00
2.83828080e-01 9.66591910e-02 -1.36696324e-01 -1.46952257e-01
-8.13440502e-01 2.36493126e-01 5.44552863e-01 -3.54262322e-01
-4.73624289e-01 -1.14363074e-01 -5.04097342e-01 7.58877158e-01
8.40091109e-01 -8.66481543e-01 -1.38530821e-01 -3.37062836e-01
8.80805254e-01 -9.06026885e-02 -1.93593219e-01 -6.12774253e-01
6.13804698e-01 -5.98158956e-01 -9.03847292e-02 -9.96517241e-01
-2.80066550e-01 -5.36307454e-01 4.04800296e-01 3.32046300e-01
-4.43521112e-01 7.37627566e-01 7.84218550e-01 1.34750038e-01
-2.03703076e-01 3.51273239e-01 8.75674188e-01 3.66995931e-01
-8.93027067e-01 2.98206508e-01 -6.22925520e-01 3.06839168e-01
1.10619736e+00 1.40745059e-01 -7.36352444e-01 2.10809499e-01
-1.10695422e+00 6.42905295e-01 -1.27825886e-01 9.07061040e-01
-1.91995218e-01 -1.42120385e+00 -4.14294004e-01 -6.00740969e-01
2.88229465e-01 -2.03375414e-01 6.36773482e-02 1.26929510e+00
-5.14421761e-01 7.86496818e-01 -1.29256532e-01 -2.99732208e-01
-1.31795835e+00 6.80583298e-01 2.95479685e-01 -4.79238421e-01
-7.81839550e-01 4.57221419e-01 -7.02190757e-01 -3.28275055e-01
4.10865009e-01 -9.17727828e-01 -3.29065144e-01 5.90233743e-01
3.66429180e-01 6.14772320e-01 5.35329640e-01 -4.92817760e-01
-8.96347702e-01 2.39754021e-01 -2.28585992e-02 -5.61561525e-01
1.56488025e+00 6.98473006e-02 -8.06733608e-01 7.38780081e-01
8.91033769e-01 -1.60687819e-01 -5.60875177e-01 -1.35442480e-01
9.50576365e-01 3.54944408e-01 -1.78902343e-01 -9.37544644e-01
-3.27531964e-01 2.19731912e-01 9.87362787e-02 -1.73152715e-01
1.34990466e+00 3.09971869e-02 7.64081180e-01 2.40716293e-01
4.09183018e-02 -1.56617904e+00 -6.06530547e-01 1.21817660e+00
6.77486479e-01 -8.05424869e-01 1.72437370e-01 -3.40440303e-01
-2.31233701e-01 1.28895450e+00 7.86957443e-01 3.42532635e-01
7.18593180e-01 6.09362066e-01 1.43359363e-01 -1.60703748e-01
-1.22806096e+00 -2.30073825e-01 3.34691375e-01 2.08483815e-01
5.70649743e-01 2.13244230e-01 -9.21243966e-01 1.07367146e+00
-7.20161319e-01 -3.97235900e-02 4.63929772e-01 9.39097762e-01
4.88130480e-01 -1.33945167e+00 -1.39042422e-01 2.23143488e-01
-7.92147577e-01 -1.30035236e-01 -5.12990534e-01 1.14268970e+00
2.66349494e-01 8.99827719e-01 1.44462302e-01 -4.86634403e-01
7.10124016e-01 4.35886890e-01 5.99595249e-01 -3.79976511e-01
-8.37805688e-01 -1.62668288e-01 5.94407439e-01 -1.13214064e+00
-6.74371123e-01 -1.13019216e+00 -1.71249604e+00 -3.16971838e-01
4.44012851e-01 4.06379066e-02 4.29303259e-01 1.22604191e+00
-3.98618467e-02 9.44126308e-01 5.09258620e-02 -1.65333569e-01
2.21745968e-01 -8.68673444e-01 1.58006564e-01 5.02813041e-01
6.21965639e-02 -5.84733784e-01 -2.82071948e-01 3.46465379e-01] | [9.063789367675781, 9.216462135314941] |
435697de-8575-422e-8411-14d301ec503f | cycle4completion-unpaired-point-cloud | 2103.07838 | null | https://arxiv.org/abs/2103.07838v2 | https://arxiv.org/pdf/2103.07838v2.pdf | Cycle4Completion: Unpaired Point Cloud Completion using Cycle Transformation with Missing Region Coding | In this paper, we present a novel unpaired point cloud completion network, named Cycle4Completion, to infer the complete geometries from a partial 3D object. Previous unpaired completion methods merely focus on the learning of geometric correspondence from incomplete shapes to complete shapes, and ignore the learning in the reverse direction, which makes them suffer from low completion accuracy due to the limited 3D shape understanding ability. To address this problem, we propose two simultaneous cycle transformations between the latent spaces of complete shapes and incomplete ones. The insight of cycle transformation is to promote networks to understand 3D shapes by learning to generate complete or incomplete shapes from their complementary ones. Specifically, the first cycle transforms shapes from incomplete domain to complete domain, and then projects them back to the incomplete domain. This process learns the geometric characteristic of complete shapes, and maintains the shape consistency between the complete prediction and the incomplete input. Similarly, the inverse cycle transformation starts from complete domain to incomplete domain, and goes back to complete domain to learn the characteristic of incomplete shapes. We provide a comprehensive evaluation in experiments, which shows that our model with the learned bidirectional geometry correspondence outperforms state-of-the-art unpaired completion methods. | ['Yu-Shen Liu', 'Wen Zheng', 'Pengfei Wan', 'Yan-Pei Cao', 'Zhizhong Han', 'Xin Wen'] | 2021-03-14 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Wen_Cycle4Completion_Unpaired_Point_Cloud_Completion_Using_Cycle_Transformation_With_Missing_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Wen_Cycle4Completion_Unpaired_Point_Cloud_Completion_Using_Cycle_Transformation_With_Missing_CVPR_2021_paper.pdf | cvpr-2021-1 | ['point-cloud-completion'] | ['computer-vision'] | [ 2.38636918e-02 9.14321095e-02 3.47295329e-02 -4.15730119e-01
-5.16870558e-01 -8.77108991e-01 5.68857610e-01 -4.75913912e-01
2.88919598e-01 4.16681588e-01 1.03602737e-01 -7.70990700e-02
-2.78802812e-02 -1.03734410e+00 -9.00826752e-01 -6.41273439e-01
4.59525377e-01 1.06331825e+00 -2.02713460e-02 -8.45422745e-02
9.41056684e-02 5.89664876e-01 -1.15853798e+00 2.36510068e-01
8.51808012e-01 7.68240869e-01 3.66052985e-01 2.20962629e-01
-4.83631253e-01 1.21433578e-01 9.98120010e-02 -1.28147721e-01
4.93514955e-01 -5.47666140e-02 -5.57497323e-01 8.54481012e-02
3.57802540e-01 -7.31241524e-01 -5.03960848e-01 8.41081023e-01
7.58733228e-02 -2.37542227e-01 9.02330756e-01 -1.43612719e+00
-1.23403442e+00 1.60477042e-01 -4.95141923e-01 -7.29198754e-01
2.84581691e-01 4.52263430e-02 8.20617676e-01 -1.42590332e+00
8.18207681e-01 1.39826322e+00 6.53457403e-01 5.49186587e-01
-1.20765412e+00 -8.43389809e-01 2.54277527e-01 -3.46208155e-01
-1.31817484e+00 -1.76654354e-01 1.32439208e+00 -4.83631015e-01
5.90363085e-01 -3.86030190e-02 7.05112517e-01 8.25633705e-01
-2.16041952e-01 7.41824567e-01 9.13274050e-01 -3.84577177e-02
1.11982979e-01 -2.26808101e-01 -2.17433468e-01 6.01572871e-01
2.80394614e-01 3.31202269e-01 -8.79906639e-02 -5.55198677e-02
1.15796161e+00 4.05583441e-01 -1.97406471e-01 -8.62315536e-01
-1.28970766e+00 5.42619348e-01 6.56275094e-01 1.13464646e-01
-3.17131400e-01 -3.74757312e-02 -1.62203744e-01 1.71596110e-01
5.22051275e-01 1.61832482e-01 -5.53645849e-01 1.92007512e-01
-6.51768684e-01 5.10663629e-01 7.68393815e-01 1.54617274e+00
1.27448928e+00 -1.25798821e-01 9.11711529e-02 5.19032836e-01
5.78253746e-01 7.35855043e-01 -1.56216413e-01 -1.13080966e+00
8.92162144e-01 1.01229727e+00 1.08975604e-01 -1.04074407e+00
-1.13890871e-01 -5.95755816e-01 -1.02401865e+00 2.79637754e-01
3.32097769e-01 1.66562349e-01 -8.87476146e-01 1.71520782e+00
3.64410758e-01 2.15805173e-01 6.99163079e-02 9.38858330e-01
9.87784147e-01 7.48151362e-01 -4.31329578e-01 2.44582042e-01
9.77167547e-01 -8.32581580e-01 -1.65560216e-01 -2.87901796e-02
4.79699671e-01 -8.25238764e-01 1.04088259e+00 7.90095702e-02
-1.09003770e+00 -6.03893101e-01 -9.52711582e-01 -4.44426835e-01
-1.15509927e-01 3.90206665e-01 3.65135074e-01 4.53874096e-02
-9.08109963e-01 7.12964058e-01 -8.16913784e-01 4.51575927e-02
6.16150141e-01 1.83625340e-01 -5.41124284e-01 -5.04100144e-01
-5.46330929e-01 4.39389318e-01 1.68002397e-01 5.06171286e-02
-9.27487493e-01 -1.16288650e+00 -9.25362825e-01 3.26132216e-02
1.76026195e-01 -1.03063166e+00 1.01980996e+00 -5.23392558e-01
-1.20879853e+00 8.77967298e-01 -2.10336626e-01 1.92417130e-01
6.72439456e-01 2.32954323e-02 1.03586890e-01 -1.12610534e-01
1.54766351e-01 1.03733766e+00 8.60510707e-01 -1.70831406e+00
-1.96405843e-01 -6.76284373e-01 1.14953786e-01 3.20301890e-01
-3.21394056e-02 -8.45167220e-01 -6.71798289e-01 -6.84329927e-01
9.03163970e-01 -8.26677442e-01 1.51453093e-02 5.96005261e-01
-3.73510122e-01 -2.50643373e-01 1.27472544e+00 -6.75955176e-01
4.70323443e-01 -2.15303707e+00 3.86613220e-01 2.45209470e-01
4.10145074e-01 -2.33899355e-01 -4.89749134e-01 4.05548036e-01
-2.40490034e-01 1.10406265e-01 -6.29054487e-01 -8.72618198e-01
1.74237452e-02 4.55275387e-01 -6.89264953e-01 2.37695396e-01
3.29007685e-01 1.00774252e+00 -1.00347543e+00 -1.66988969e-01
1.11099698e-01 6.57382965e-01 -6.68814301e-01 3.80021960e-01
-4.28352058e-01 9.63172495e-01 -6.83589101e-01 6.48264885e-01
1.28833783e+00 -1.84498399e-01 -1.73742592e-01 -4.33587641e-01
-1.60830736e-01 2.10082188e-01 -1.00008571e+00 2.17583752e+00
-4.40924466e-01 1.53953940e-01 1.61823109e-01 -8.48180473e-01
1.56214356e+00 2.93421149e-01 4.84515816e-01 -5.28618515e-01
-1.92759976e-01 3.48489106e-01 -4.57687676e-01 -2.23053515e-01
2.22929940e-01 -3.77569288e-01 2.44916335e-01 6.83245242e-01
-2.17547834e-01 -8.76967967e-01 -3.36627275e-01 1.28749043e-01
5.54227710e-01 8.63288403e-01 -3.35425168e-01 7.21024573e-02
5.68450689e-01 -1.22326069e-01 7.73072481e-01 1.41122952e-01
1.76096484e-01 1.09714687e+00 5.28255641e-01 -7.31576800e-01
-1.45823729e+00 -1.64518273e+00 2.39799753e-01 2.10340604e-01
4.59280640e-01 -1.66072533e-01 -4.71019685e-01 -6.99614942e-01
1.56809092e-01 5.58797359e-01 -6.09474957e-01 -1.40135318e-01
-6.58358574e-01 -8.64851847e-02 2.45583713e-01 6.35655940e-01
7.47122645e-01 -9.13432002e-01 1.42653897e-01 -4.09689285e-02
-2.74015009e-01 -9.77138996e-01 -6.86885118e-01 -4.75849509e-01
-1.31116271e+00 -1.04927003e+00 -7.72634208e-01 -1.10000217e+00
1.19011879e+00 4.32798624e-01 9.61812556e-01 4.31183845e-01
1.84222028e-01 1.99497998e-01 -1.64841115e-01 -7.87853003e-02
-4.39787775e-01 3.63012180e-02 -8.55738297e-02 4.09487225e-02
8.48254859e-02 -1.24774563e+00 -5.87798774e-01 3.86847705e-01
-9.48950291e-01 6.02455854e-01 7.18962967e-01 6.76595569e-01
9.06290591e-01 -3.17024380e-01 3.99057090e-01 -7.32034504e-01
1.69704929e-01 -3.49891245e-01 -6.10301316e-01 2.15392381e-01
-3.44600439e-01 2.65910894e-01 7.38188922e-01 -1.80863604e-01
-1.08094943e+00 4.64798033e-01 -2.25379527e-01 -9.44690406e-01
-3.06535333e-01 1.77273765e-01 -5.39175630e-01 1.06643304e-01
2.91846097e-01 4.84272152e-01 2.33712912e-01 -9.23058331e-01
4.83218640e-01 2.11910293e-01 6.99645877e-01 -9.35604036e-01
1.31698465e+00 8.17652881e-01 1.36894882e-01 -5.01410365e-01
-6.62101507e-01 -1.69437170e-01 -1.01179075e+00 8.44046548e-02
7.27916420e-01 -9.02670026e-01 -4.07627702e-01 6.16530836e-01
-1.51376522e+00 -2.22518399e-01 -3.78012061e-01 2.44577959e-01
-6.99451864e-01 6.07028067e-01 -4.36779648e-01 -2.42524028e-01
-2.13887095e-01 -1.13282025e+00 1.42592525e+00 -4.11987863e-02
1.48437008e-01 -8.33262682e-01 1.11485206e-01 4.93178554e-02
4.21697237e-02 3.78395736e-01 1.28705120e+00 -9.39547643e-02
-9.52673852e-01 -1.20023422e-01 -4.39421773e-01 2.99843818e-01
3.71050149e-01 -5.79779297e-02 -7.87874401e-01 -3.22156787e-01
9.47740227e-02 -1.34495765e-01 5.90000272e-01 1.73232734e-01
1.18666470e+00 -5.16351759e-02 -3.55834693e-01 1.06516993e+00
1.24766386e+00 -1.18063539e-01 5.95164776e-01 -1.83219060e-01
9.16820347e-01 6.84481621e-01 3.44734669e-01 2.55323678e-01
6.56025410e-01 4.18818504e-01 7.87037373e-01 -1.50917191e-02
-2.94111609e-01 -1.01868701e+00 -2.10956186e-02 1.00467849e+00
-1.28277645e-01 2.25929409e-01 -8.15264344e-01 3.68921965e-01
-1.48876381e+00 -6.25903368e-01 -1.55573279e-01 2.19367290e+00
4.78588313e-01 -3.25189978e-02 -2.42994249e-01 -8.76664221e-02
7.88554311e-01 1.15511194e-01 -7.01023400e-01 1.62781477e-01
-3.14977728e-02 -8.77476707e-02 -1.86951682e-02 6.02963746e-01
-6.50026619e-01 7.86701024e-01 5.40838099e+00 6.16577566e-01
-8.76646340e-01 -1.19324036e-01 4.91563499e-01 3.40114921e-01
-1.08580303e+00 5.61639071e-01 -4.62218344e-01 2.30707929e-01
-1.57469898e-01 1.10516669e-02 6.23969972e-01 7.74632215e-01
-3.53739709e-02 5.89393139e-01 -1.43970537e+00 1.07951641e+00
-9.17954072e-02 -1.48418033e+00 5.33400238e-01 1.30160823e-01
9.63984132e-01 -2.21072733e-01 2.11863220e-02 2.63786942e-01
1.66509882e-01 -8.56261790e-01 8.30579698e-01 7.72206426e-01
1.08753085e+00 -5.17529011e-01 2.96816617e-01 8.26666594e-01
-1.57577491e+00 2.80702502e-01 -6.18067384e-01 -2.08171055e-01
1.77701056e-01 5.30899584e-01 -6.86715841e-01 7.97466099e-01
3.81233841e-01 1.08195472e+00 -2.80137718e-01 7.75950789e-01
-4.93128091e-01 1.03769526e-01 -3.92455846e-01 3.28694552e-01
7.98212215e-02 -7.98642039e-01 6.18288994e-01 4.11462784e-01
7.26220548e-01 2.49861300e-01 2.82944620e-01 1.69770908e+00
-3.27014327e-01 -2.55276889e-01 -9.94204521e-01 1.66063800e-01
7.07987249e-01 1.20418561e+00 -5.95258653e-01 -2.38106266e-01
-3.47821087e-01 8.47467184e-01 4.15004849e-01 5.55667400e-01
-4.93205220e-01 -7.88281336e-02 4.80164409e-01 5.35334311e-02
1.86151952e-01 -6.47515595e-01 -8.01957965e-01 -1.08097208e+00
3.37398797e-01 -1.68179557e-01 -5.33849094e-03 -1.11872780e+00
-1.39279687e+00 4.35169667e-01 2.65490692e-02 -1.68975973e+00
1.73645630e-01 -4.58869606e-01 -1.08725107e+00 1.05281007e+00
-1.50331509e+00 -1.57908869e+00 -5.28774977e-01 5.84020555e-01
3.63991618e-01 4.56542242e-03 6.79331958e-01 1.49968296e-01
2.78904065e-02 3.72768462e-01 2.55049206e-03 2.17022657e-01
4.53774750e-01 -1.00155962e+00 7.35012650e-01 4.43088233e-01
-5.11455126e-02 5.26055932e-01 1.87988356e-01 -8.43339205e-01
-1.52232850e+00 -1.11009884e+00 8.23426366e-01 -4.55103785e-01
-8.04794282e-02 -4.63903069e-01 -1.04263592e+00 6.86237931e-01
-3.74321818e-01 1.05273314e-01 1.69684753e-01 -2.03651503e-01
-6.84237242e-01 -9.15418044e-02 -1.05491507e+00 6.99240863e-01
1.38995481e+00 -5.61075866e-01 -8.76657248e-01 2.55174078e-02
1.08256900e+00 -5.46779931e-01 -7.17758179e-01 5.09438932e-01
6.91718400e-01 -7.94520855e-01 1.03586662e+00 -5.68617761e-01
8.24956298e-01 -5.11064351e-01 -2.24084049e-01 -1.43750286e+00
-3.45535785e-01 -3.62214297e-01 4.92850691e-02 1.10389578e+00
2.24272192e-01 -6.06177509e-01 1.16218257e+00 4.95530397e-01
-6.46855712e-01 -8.33463788e-01 -8.38767350e-01 -7.30090261e-01
6.40899539e-01 -4.98694718e-01 1.21376526e+00 9.82680261e-01
-6.35535657e-01 1.26873985e-01 -1.16332233e-01 3.09538662e-01
6.58863366e-01 9.14567709e-01 9.56809461e-01 -1.31368458e+00
-1.15144186e-01 -1.93783388e-01 -6.85236529e-02 -1.65067661e+00
1.75119847e-01 -1.21708083e+00 -1.85869619e-01 -1.63062489e+00
1.92242444e-01 -9.28405523e-01 3.11793208e-01 5.01915574e-01
1.45618588e-01 2.65613258e-01 2.70953894e-01 4.43343371e-01
5.50076440e-02 1.13528490e+00 2.10286999e+00 -3.19367439e-01
-2.73482680e-01 8.25182870e-02 -6.71955347e-01 8.22382510e-01
4.83890474e-01 -4.15882945e-01 -5.33118665e-01 -8.72673988e-01
3.63047540e-01 2.03501016e-01 5.48185527e-01 -6.10999823e-01
2.31973618e-01 -2.20144272e-01 4.89837170e-01 -1.18105066e+00
5.19705951e-01 -1.01839185e+00 2.90249586e-01 3.40302676e-01
5.43303695e-03 3.99649255e-02 -7.52415136e-02 5.65633476e-01
-9.37688425e-02 -9.51967090e-02 4.17794287e-01 -2.71614730e-01
-3.62837881e-01 9.92149413e-01 4.94313836e-01 8.74033943e-02
7.95358181e-01 -5.36077261e-01 -1.19046003e-01 -3.41223001e-01
-9.56034422e-01 3.34079206e-01 9.12128687e-01 6.12305999e-01
1.06796324e+00 -1.93573749e+00 -9.15392220e-01 7.51480818e-01
5.30305728e-02 8.82663012e-01 3.78490478e-01 4.85531151e-01
-6.03086054e-01 1.40889525e-01 -2.98359960e-01 -9.65067208e-01
-7.82924354e-01 4.97395098e-01 3.26576501e-01 -6.47063106e-02
-8.13465178e-01 4.87931997e-01 7.50098288e-01 -1.22412193e+00
8.39893147e-02 -4.07725245e-01 1.95832908e-01 -4.69856054e-01
1.19129114e-01 2.57472396e-01 -4.57113085e-04 -5.74213684e-01
-1.19490735e-01 1.16356909e+00 1.21939316e-01 3.23256187e-04
1.55259919e+00 7.36185312e-02 -3.09504241e-01 1.85165301e-01
1.51820993e+00 -1.37260020e-01 -1.64818585e+00 -3.68431449e-01
-5.01479387e-01 -7.41590321e-01 -4.36479598e-01 -3.25666338e-01
-1.17932630e+00 1.15265453e+00 1.66938394e-01 -3.69048476e-01
8.94348145e-01 1.47150204e-01 8.42573464e-01 4.05278474e-01
4.73578036e-01 -5.90848565e-01 2.28116021e-01 7.74594128e-01
1.33961952e+00 -8.77948463e-01 -6.65123165e-02 -7.67174006e-01
-2.74030924e-01 1.26659238e+00 7.45780170e-01 -4.70504403e-01
8.97314310e-01 -6.56788126e-02 -3.14742774e-01 -3.76834750e-01
-5.48920453e-01 3.32593381e-01 4.20095056e-01 9.12126839e-01
2.44887769e-02 1.35302886e-01 1.53191686e-01 5.70644140e-01
-3.02651495e-01 -1.01965964e-01 1.55395895e-01 6.03369117e-01
-7.11813122e-02 -1.43048835e+00 -4.85793054e-01 3.43507022e-01
3.93305272e-01 1.62685499e-01 -5.31866133e-01 7.11716235e-01
3.66182566e-01 4.90832299e-01 3.00738424e-01 -3.92496854e-01
5.32213330e-01 -5.34992218e-02 4.74414885e-01 -7.43289649e-01
1.11173242e-01 1.36757176e-02 -4.19567198e-01 -3.30655307e-01
-1.68809891e-01 -6.16914034e-01 -1.42998910e+00 -2.79573858e-01
-4.59538773e-03 -1.61249712e-02 4.96589601e-01 1.00606585e+00
5.06026566e-01 1.77599669e-01 8.37029457e-01 -1.23852801e+00
-6.84321702e-01 -6.72088504e-01 -5.08741438e-01 7.43727207e-01
2.93954432e-01 -7.87828743e-01 -2.75898069e-01 -5.83569221e-02] | [8.461918830871582, -3.5997531414031982] |
67cc264a-440e-42e5-bcb8-4324940b1af3 | adaptive-recursive-neural-network-for-target | null | null | https://aclanthology.org/P14-2009 | https://aclanthology.org/P14-2009.pdf | Adaptive Recursive Neural Network for Target-dependent Twitter Sentiment Classification | null | ['Li Dong', 'Furu Wei', 'Chuanqi Tan', 'Ming Zhou', 'Ke Xu', 'Duyu Tang'] | 2014-06-01 | null | null | null | acl-2014-6 | ['twitter-sentiment-analysis'] | ['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] | [-7.253398418426514, 3.7227776050567627] |
4a00d5bf-a83e-4382-b6c2-91c10b3b71f6 | multi-agent-path-finding-via-tree-lstm | 2210.12933 | null | https://arxiv.org/abs/2210.12933v2 | https://arxiv.org/pdf/2210.12933v2.pdf | Multi-Agent Path Finding via Tree LSTM | In recent years, Multi-Agent Path Finding (MAPF) has attracted attention from the fields of both Operations Research (OR) and Reinforcement Learning (RL). However, in the 2021 Flatland3 Challenge, a competition on MAPF, the best RL method scored only 27.9, far less than the best OR method. This paper proposes a new RL solution to Flatland3 Challenge, which scores 125.3, several times higher than the best RL solution before. We creatively apply a novel network architecture, TreeLSTM, to MAPF in our solution. Together with several other RL techniques, including reward shaping, multiple-phase training, and centralized control, our solution is comparable to the top 2-3 OR methods. | ['Xiaolong Zhu', 'Jiaxin Chen', 'Qimai Li', 'Kunjie Zhang', 'Yuhao Jiang'] | 2022-10-24 | null | null | null | null | ['multi-agent-path-finding'] | ['playing-games'] | [-7.12330863e-02 4.84026849e-01 -7.06574261e-01 2.25418091e-01
-6.79943860e-01 -6.04519665e-01 6.14630342e-01 1.18961237e-01
-8.64942253e-01 1.42806876e+00 4.22934592e-01 -5.74113965e-01
-6.18026257e-01 -7.39738584e-01 -7.00458050e-01 -4.30254072e-01
-6.56842291e-01 7.78486431e-01 2.91204005e-01 -7.25920737e-01
4.20413733e-01 3.82229865e-01 -1.08823359e+00 -6.64362013e-02
6.47734880e-01 7.65483379e-01 2.36557454e-01 7.72286654e-01
-3.01467683e-02 1.25663722e+00 -7.90231407e-01 1.65047213e-01
4.13482577e-01 -1.87349647e-01 -1.07209301e+00 -6.93860114e-01
-2.01503873e-01 -4.94155079e-01 -5.11921525e-01 3.28844160e-01
6.77430630e-01 4.53226238e-01 1.59046382e-01 -1.64408457e+00
-3.28927517e-01 1.10251951e+00 -5.82440138e-01 1.76798627e-01
4.31873620e-01 1.96339965e-01 1.01332414e+00 -5.09955548e-02
8.35338593e-01 1.31403863e+00 4.69390512e-01 6.82474315e-01
-8.30080211e-01 -4.93963897e-01 5.36598325e-01 2.40731597e-01
-8.45198989e-01 -4.02896762e-01 2.30865583e-01 2.45091990e-01
1.52747202e+00 -1.30347773e-01 9.42495644e-01 1.04811716e+00
6.04070365e-01 1.03668272e+00 1.15535331e+00 -3.49252254e-01
3.96255195e-01 -2.43006006e-01 -5.87679565e-01 7.11390793e-01
-7.60143399e-02 7.02396214e-01 -5.12666762e-01 -2.18491003e-01
9.76125896e-01 -4.51160431e-01 -2.95032449e-02 -3.23486239e-01
-1.38963223e+00 1.00096536e+00 6.90432727e-01 2.38471180e-02
-8.49181116e-01 9.77269828e-01 2.26221010e-01 4.32950586e-01
6.33706301e-02 9.22274768e-01 -7.53087878e-01 -5.22027493e-01
-3.74437720e-01 6.82625055e-01 9.14417267e-01 7.45079219e-01
4.53735501e-01 3.47947985e-01 -1.99001759e-01 4.96900469e-01
4.56008762e-01 3.78116697e-01 1.19105600e-01 -1.77271628e+00
4.24879372e-01 2.63178110e-01 5.38641751e-01 -7.13609397e-01
-8.70130539e-01 -6.23391032e-01 -4.45330858e-01 5.95852077e-01
2.90374041e-01 -5.95924497e-01 -5.55720925e-01 1.73484504e+00
-1.05749004e-01 1.90861002e-02 2.62692004e-01 6.30604923e-01
5.36592245e-01 9.08870816e-01 -1.05908215e-01 -9.28485617e-02
5.86253405e-01 -1.42519987e+00 -5.33505380e-01 -4.62349832e-01
5.68607748e-01 -2.89377153e-01 8.25173140e-01 6.46984518e-01
-1.38129628e+00 9.46587995e-02 -1.04510593e+00 6.11608684e-01
-3.83028716e-01 -5.81535220e-01 9.56092060e-01 4.05582637e-01
-1.62167072e+00 7.56451130e-01 -5.94289005e-01 -5.10129511e-01
2.95323670e-01 5.56786597e-01 -2.83636414e-02 -2.20385402e-01
-1.43866587e+00 1.30498135e+00 4.07078743e-01 -1.53080776e-01
-1.25285041e+00 -5.18695354e-01 -4.92711514e-01 9.85030159e-02
1.03133488e+00 -7.77097881e-01 1.58449554e+00 -2.03775123e-01
-2.08954740e+00 1.07089337e-02 3.90787899e-01 -8.85040104e-01
3.83249521e-01 -3.08968157e-01 -3.01617265e-01 -8.17898810e-02
2.64426284e-02 1.17676389e+00 4.11347926e-01 -1.16403592e+00
-9.70946372e-01 3.56868446e-01 5.13822675e-01 5.02433121e-01
2.33103782e-01 -1.19753107e-01 2.07084641e-01 -1.62644163e-01
-4.23324585e-01 -8.54626298e-01 -9.48118627e-01 -4.43706065e-01
1.73589606e-02 -4.71199989e-01 3.38699341e-01 -2.27783173e-01
1.11739230e+00 -1.41679156e+00 7.53422454e-02 2.70928472e-01
9.57145467e-02 -1.70827538e-01 -7.00026393e-01 8.73967350e-01
3.33373398e-01 2.12774754e-01 2.20803916e-01 1.32965133e-01
3.92505497e-01 5.01660168e-01 -1.04799606e-01 2.28771716e-01
2.25349575e-01 1.05036604e+00 -1.33673418e+00 -5.37925474e-02
2.02626940e-02 -1.31011412e-01 -7.17833281e-01 -1.75338894e-01
-6.15513742e-01 3.70085210e-01 -4.19387043e-01 6.73568308e-01
2.19713017e-01 -8.89694914e-02 3.50914091e-01 6.60614967e-01
-4.29218799e-01 3.61658663e-01 -9.74967360e-01 2.11439347e+00
-5.32374740e-01 5.61652124e-01 -3.11456285e-02 -6.06603205e-01
8.99968684e-01 1.78699642e-01 1.01764143e+00 -1.28628898e+00
-1.97916001e-01 4.76379335e-01 2.47171879e-01 -2.17898011e-01
8.68556201e-01 2.26446271e-01 -2.83917665e-01 6.34979725e-01
-2.60051012e-01 -5.20566218e-02 4.14860159e-01 2.60255843e-01
1.79778552e+00 5.82284212e-01 1.39908373e-01 -2.71870524e-01
2.13486299e-01 3.37350994e-01 6.15063667e-01 1.16614187e+00
-6.60826921e-01 -2.31862903e-01 6.95606411e-01 -6.28635287e-01
-7.60551274e-01 -1.02211607e+00 6.64029062e-01 1.23135459e+00
2.28132397e-01 -5.01701832e-01 -2.60577917e-01 -8.34883511e-01
2.49125361e-01 7.18826115e-01 -3.92405123e-01 -5.93434498e-02
-1.04254293e+00 -5.31783879e-01 7.06300318e-01 3.21383834e-01
8.00504506e-01 -1.54289567e+00 -8.57816398e-01 7.97564745e-01
-2.23759502e-01 -8.86993408e-01 -2.50822455e-01 4.43789482e-01
-5.48601925e-01 -9.57164049e-01 -7.12400258e-01 -5.60238361e-01
1.62498169e-02 3.08197886e-01 1.30282474e+00 -3.99454460e-02
1.43387079e-01 3.16073358e-01 -5.39982557e-01 -3.09478223e-01
-4.75014091e-01 3.11057985e-01 7.95211196e-02 -9.37654674e-01
-1.28772303e-01 -3.08731914e-01 -7.79628336e-01 2.80211508e-01
-2.28749231e-01 -2.62511671e-01 9.70760524e-01 7.61096060e-01
1.16898209e-01 2.81068653e-01 1.20556188e+00 -3.68532270e-01
1.07161951e+00 -4.47294235e-01 -6.37673318e-01 2.13822469e-01
-1.04500318e+00 8.96556601e-02 5.32001793e-01 -1.08782992e-01
-8.14951479e-01 -2.56004125e-01 4.58648913e-02 -2.09453218e-02
6.28158450e-02 6.83918893e-01 6.45304859e-01 -4.18730825e-01
6.79684222e-01 -9.54914689e-02 2.19077766e-01 6.87442124e-02
4.71229613e-01 7.88097009e-02 1.75426587e-01 -6.12046123e-01
5.91885388e-01 3.92616950e-02 1.58359796e-01 -1.48403928e-01
-3.10319006e-01 -1.59013748e-01 5.93247972e-02 -4.78629261e-01
5.51280797e-01 -7.40125656e-01 -1.16389549e+00 1.87896386e-01
-1.01681280e+00 -1.07814074e+00 -3.93582344e-01 4.45675999e-01
-1.04376805e+00 -1.87437102e-01 -6.96697414e-01 -8.45548928e-01
-2.35468462e-01 -9.56309557e-01 3.77445132e-01 3.10892999e-01
-2.59508528e-02 -8.70823026e-01 5.07794738e-01 1.36905938e-01
1.09414148e+00 3.66183579e-01 7.95379817e-01 -2.77669996e-01
-7.45662510e-01 1.05421491e-01 -1.66894987e-01 -1.97699636e-01
-1.21925227e-01 -2.00810283e-01 -3.45627815e-01 -5.23291111e-01
-6.18194103e-01 -6.32944822e-01 7.26144373e-01 7.52379715e-01
4.39153224e-01 -3.52926224e-01 -3.75984341e-01 1.42608033e-02
1.40100515e+00 7.97111809e-01 7.19662130e-01 1.41310763e+00
3.93932797e-02 5.12076616e-01 9.93548870e-01 6.95219755e-01
8.58781338e-01 5.44389129e-01 8.19585919e-01 2.09287740e-02
1.55720219e-01 -4.11092162e-01 8.45490396e-01 4.79258239e-01
-1.86528206e-01 -6.16249025e-01 -1.18025208e+00 2.22518638e-01
-2.24337673e+00 -9.94202554e-01 2.22133487e-01 1.86826289e+00
2.94720411e-01 4.82048631e-01 4.83309180e-01 -2.05075711e-01
3.85449260e-01 2.54291385e-01 -8.29624355e-01 -7.72943974e-01
-1.16102241e-01 2.87515700e-01 1.05746436e+00 5.19845486e-01
-8.46069157e-01 1.21909106e+00 7.90666056e+00 6.83276296e-01
-6.44888103e-01 -3.40506099e-02 2.25830674e-01 -2.14064956e-01
-1.81634545e-01 8.01634043e-02 -4.35308963e-01 -7.01321587e-02
1.27384841e+00 -4.82413530e-01 1.13929975e+00 5.39852858e-01
5.61846554e-01 -4.21304405e-01 -6.19478405e-01 4.69630301e-01
-2.69666135e-01 -1.47869802e+00 -3.66854012e-01 2.10948139e-01
8.73075187e-01 4.89445329e-01 -3.78869027e-02 1.01373208e+00
1.31604505e+00 -1.31224382e+00 4.69839990e-01 3.98176253e-01
2.21727178e-01 -1.11326253e+00 6.38636112e-01 4.10470456e-01
-1.15022349e+00 -6.43277943e-01 -2.27970377e-01 -3.64663213e-01
3.71034384e-01 -9.70852599e-02 -1.12785697e+00 7.49016702e-01
7.45445728e-01 7.29257822e-01 -8.72711688e-02 1.33109128e+00
-3.09508204e-01 2.33279750e-01 -2.77924687e-01 -4.96006072e-01
9.50196326e-01 4.30943482e-02 6.26729369e-01 6.02257073e-01
6.05426766e-02 -2.44993314e-01 4.22562540e-01 5.57086229e-01
8.15470889e-03 -3.18856835e-01 -6.28208280e-01 -1.44849405e-01
5.73029280e-01 1.20755100e+00 -4.91398096e-01 7.51522556e-02
-3.78844552e-02 4.22784239e-01 3.17825913e-01 4.21197444e-01
-9.29331720e-01 -3.15344512e-01 5.17913222e-01 -2.99413651e-01
-7.87372068e-02 -4.05682713e-01 2.17147946e-01 -3.86321783e-01
-3.10924023e-01 -8.78206074e-01 3.05621237e-01 -6.70016825e-01
-9.00780499e-01 5.41837096e-01 5.63292019e-02 -1.08683014e+00
-6.13404572e-01 -3.67803544e-01 -5.25750935e-01 4.38800007e-01
-2.20023060e+00 -7.01525509e-01 4.74108867e-02 3.20050567e-01
4.57119703e-01 -4.61300343e-01 8.19132984e-01 3.27182710e-01
-4.62939203e-01 2.33955935e-01 2.77113199e-01 -5.19988000e-01
5.63992977e-01 -1.29672968e+00 6.43004239e-01 3.34078312e-01
-4.72538948e-01 4.73806933e-02 4.23634887e-01 -5.27563810e-01
-1.62164330e+00 -7.57937491e-01 4.10106719e-01 1.06234299e-02
9.02071714e-01 -4.78380658e-02 -1.01543449e-01 5.89972019e-01
6.97369754e-01 -6.32185102e-01 1.59219548e-01 9.13357064e-02
3.41375798e-01 3.49187776e-02 -1.15081108e+00 7.51500010e-01
1.06166804e+00 1.46831319e-01 4.44097631e-02 3.18446070e-01
1.10789955e+00 -4.35025960e-01 -8.06396604e-01 3.71893793e-01
4.77225959e-01 -8.69296014e-01 8.48778248e-01 -6.17927730e-01
2.26174384e-01 -1.73275724e-01 -2.46442899e-01 -1.92436683e+00
-6.16053224e-01 -1.14520621e+00 -1.44211411e-01 5.38799703e-01
5.50432205e-01 -9.42402303e-01 1.26897073e+00 1.46867394e-01
-3.79607588e-01 -1.10777867e+00 -1.02781558e+00 -1.26271379e+00
3.28616291e-01 -7.00521320e-02 7.18256593e-01 6.31439984e-01
2.09566802e-02 2.11638302e-01 -5.68830729e-01 -2.56647974e-01
4.66395020e-01 -2.79524505e-01 6.64580584e-01 -1.06020451e+00
-3.43737304e-01 -8.82435441e-01 1.66923016e-01 -1.18811238e+00
9.42694768e-02 -7.20094860e-01 2.60978460e-01 -2.25001359e+00
-2.85385489e-01 -8.64083767e-01 -5.05894244e-01 7.77771473e-01
4.65459108e-01 -4.09438074e-01 5.19751728e-01 7.69264325e-02
-7.72777796e-01 8.53366852e-01 1.65380323e+00 -1.29037768e-01
-4.51177299e-01 -3.24129090e-02 -7.53350675e-01 4.34801042e-01
1.33246148e+00 -4.95995760e-01 -4.26646799e-01 -4.16043669e-01
6.59859776e-01 7.67343104e-01 1.91616312e-01 -9.84298527e-01
4.66449440e-01 -7.01566815e-01 -6.12273850e-02 -6.45933092e-01
3.39832246e-01 -6.79481030e-01 -1.47485197e-01 1.03676283e+00
-3.99043590e-01 5.82670033e-01 3.62728268e-01 5.36152482e-01
9.90012884e-02 -1.22326024e-01 2.51100093e-01 -4.89896923e-01
-1.06504905e+00 2.42465302e-01 -8.20360899e-01 1.25841260e-01
1.12835658e+00 -5.85896662e-03 -1.13731670e+00 -7.98871219e-01
-4.64366466e-01 1.07952046e+00 -6.95607513e-02 4.80900824e-01
9.29394245e-01 -1.20735645e+00 -8.96548450e-01 -3.75453234e-01
-4.82473135e-01 -8.03565010e-02 -6.19624667e-02 8.47891033e-01
-6.05646610e-01 6.03468895e-01 -5.84112585e-01 3.25425714e-02
-2.84317434e-01 1.95177108e-01 5.45102537e-01 -9.45512295e-01
-5.94926596e-01 5.20882070e-01 -6.09703302e-01 -8.93975735e-01
5.10107994e-01 1.46824494e-01 -2.22710967e-01 -2.16524571e-01
1.67208821e-01 7.73836672e-01 -3.82465094e-01 3.34143192e-01
-4.45536554e-01 1.81439519e-01 -1.01534106e-01 -5.72057545e-01
1.46720064e+00 -3.16261128e-02 1.37269318e-01 2.11743400e-01
4.78500813e-01 -4.31484610e-01 -1.20555794e+00 1.20283939e-01
3.54229510e-01 -2.50016540e-01 7.70380199e-02 -1.53111172e+00
-9.50150311e-01 1.43434256e-01 3.85007143e-01 2.78808206e-01
8.43191564e-01 -3.99059057e-01 9.13351536e-01 9.33896124e-01
1.00748944e+00 -1.32656157e+00 4.64086741e-01 9.91992235e-01
9.37562764e-01 -1.03042924e+00 -6.65666908e-02 2.79807329e-01
-7.16065347e-01 1.04575288e+00 8.88239026e-01 -4.61214453e-01
2.69616038e-01 3.33000749e-01 -4.12664488e-02 -8.16184059e-02
-1.41091037e+00 -2.40460441e-01 -5.65978289e-01 8.87637377e-01
-2.47727800e-02 1.86257828e-02 -9.51770619e-02 -1.88288931e-02
-1.18677706e-01 2.77048349e-01 8.86724770e-01 1.31299245e+00
-8.45252991e-01 -1.33961344e+00 -6.61131740e-02 4.10438627e-01
-1.53654218e-01 -1.67316012e-02 -3.07954371e-01 1.13736796e+00
-4.03968215e-01 1.23466718e+00 -6.39000833e-02 -3.93074751e-01
5.35351217e-01 -5.88437796e-01 4.45129484e-01 -1.12383418e-01
-9.53487396e-01 -7.92589113e-02 6.94981575e-01 -1.13235629e+00
-4.26818997e-01 -4.14811850e-01 -1.70391929e+00 -8.15333545e-01
-3.99081297e-02 2.29065418e-01 6.72583103e-01 5.59701562e-01
3.68467659e-01 8.94607544e-01 8.25341046e-01 -9.42838967e-01
-4.79821146e-01 -5.42430937e-01 -5.44317842e-01 -6.48144782e-01
2.44985551e-01 -7.87864387e-01 -9.82911363e-02 -1.24307859e+00] | [3.776365041732788, 1.6314611434936523] |
fecbe4f7-dd6b-40c0-819f-75d23c6d9f97 | robust-reinforcement-learning-objectives-for | 2305.18820 | null | https://arxiv.org/abs/2305.18820v1 | https://arxiv.org/pdf/2305.18820v1.pdf | Robust Reinforcement Learning Objectives for Sequential Recommender Systems | Attention-based sequential recommendation methods have demonstrated promising results by accurately capturing users' dynamic interests from historical interactions. In addition to generating superior user representations, recent studies have begun integrating reinforcement learning (RL) into these models. Framing sequential recommendation as an RL problem with reward signals, unlocks developing recommender systems (RS) that consider a vital aspect-incorporating direct user feedback in the form of rewards to deliver a more personalized experience. Nonetheless, employing RL algorithms presents challenges, including off-policy training, expansive combinatorial action spaces, and the scarcity of datasets with sufficient reward signals. Contemporary approaches have attempted to combine RL and sequential modeling, incorporating contrastive-based objectives and negative sampling strategies for training the RL component. In this study, we further emphasize the efficacy of contrastive-based objectives paired with augmentation to address datasets with extended horizons. Additionally, we recognize the potential instability issues that may arise during the application of negative sampling. These challenges primarily stem from the data imbalance prevalent in real-world datasets, which is a common issue in offline RL contexts. While our established baselines attempt to mitigate this through various techniques, instability remains an issue. Therefore, we introduce an enhanced methodology aimed at providing a more effective solution to these challenges. | ['Lili Meng', 'Dave Evans', 'Tristan Sylvain', 'Melissa Mozifian'] | 2023-05-30 | null | null | null | null | ['sequential-recommendation', 'offline-rl'] | ['miscellaneous', 'playing-games'] | [ 3.29218626e-01 -1.32112622e-01 -5.76855183e-01 -2.69762605e-01
-7.34304965e-01 -3.91314507e-01 5.35191357e-01 -1.37013923e-02
-3.29360783e-01 8.82100284e-01 4.82161969e-01 -2.88249046e-01
-4.18474078e-01 -5.49171746e-01 -5.00330269e-01 -4.46805000e-01
-3.62628371e-01 1.56449258e-01 -1.91914573e-01 -4.90716904e-01
5.36132157e-01 2.48257652e-01 -1.87453258e+00 1.24546640e-01
1.19905412e+00 9.73083615e-01 2.11376160e-01 5.42416215e-01
-1.33975118e-01 1.17792022e+00 -6.77898288e-01 -2.47628018e-01
3.25622171e-01 -6.18988633e-01 -4.17051524e-01 -6.40308335e-02
1.80846512e-01 -6.27893984e-01 -1.77694321e-01 4.23320115e-01
6.87708318e-01 9.07583773e-01 4.17170256e-01 -1.40733516e+00
-8.71358633e-01 6.03767872e-01 -4.55591887e-01 3.20984483e-01
5.61088145e-01 3.11026245e-01 1.46738088e+00 -5.10730982e-01
2.97056854e-01 1.13395250e+00 5.39463162e-01 7.64475346e-01
-1.23456919e+00 -5.45342982e-01 7.66604364e-01 2.46453986e-01
-6.82742894e-01 -4.55668360e-01 8.17192018e-01 -2.15212166e-01
1.06946814e+00 4.86740619e-01 7.63752043e-01 1.37425625e+00
-2.53661036e-01 1.23335016e+00 7.73609638e-01 -2.75865674e-01
2.98221707e-01 1.83066711e-01 1.10213466e-01 -1.75787558e-04
5.34047075e-02 4.51944351e-01 -4.05040503e-01 -1.74217120e-01
6.04593098e-01 2.56798923e-01 -1.99060768e-01 -2.85128742e-01
-8.36031497e-01 1.00925994e+00 3.28541547e-01 7.50353262e-02
-7.48069108e-01 -6.79927692e-03 2.56990075e-01 2.52963930e-01
4.37247097e-01 1.08865607e+00 -4.66190785e-01 -4.76488054e-01
-8.98591161e-01 5.90970457e-01 6.94699049e-01 7.09415197e-01
3.10561866e-01 2.71885216e-01 -5.58124125e-01 9.22720492e-01
2.76293129e-01 1.29988268e-01 6.95117712e-01 -1.06572795e+00
4.78215814e-01 5.08767903e-01 4.61586744e-01 -9.31548834e-01
-2.78177619e-01 -7.82456934e-01 -4.14787441e-01 -1.05953617e-02
3.39969337e-01 -3.20842057e-01 -4.95759040e-01 1.97115672e+00
2.89724410e-01 3.49069476e-01 1.17507778e-01 1.00428605e+00
3.97913098e-01 4.64667827e-01 2.72436827e-01 -4.07223433e-01
8.33072782e-01 -1.03447604e+00 -8.13822687e-01 -2.99604148e-01
4.45010483e-01 -4.81457978e-01 1.34325087e+00 5.03581464e-01
-1.14117205e+00 -3.65654796e-01 -1.00828075e+00 2.35719770e-01
-4.69511822e-02 4.85151857e-02 8.05552959e-01 6.33387864e-01
-8.86661410e-01 8.38268518e-01 -5.63589394e-01 -2.35494792e-01
3.48459154e-01 4.84526604e-01 4.36483324e-01 9.44171399e-02
-1.37534404e+00 8.46084595e-01 -2.36994937e-01 2.05112830e-01
-4.48448539e-01 -8.13771904e-01 -5.74478388e-01 2.07805350e-01
7.85671890e-01 -3.93258274e-01 1.42425263e+00 -1.35018134e+00
-1.79172957e+00 -1.74326062e-01 1.89859077e-01 -5.37260115e-01
6.17259383e-01 -4.55905557e-01 -5.05651832e-01 -2.82112241e-01
-2.58690953e-01 3.01694125e-01 7.31175721e-01 -1.09537876e+00
-6.85956776e-01 -1.33821398e-01 4.48092282e-01 5.85306227e-01
-5.13812423e-01 -2.11141646e-01 -1.04855090e-01 -7.50589371e-01
-4.75371480e-01 -8.93675268e-01 -6.16005421e-01 -5.17493725e-01
2.09373280e-01 -2.78822243e-01 5.08041501e-01 -4.90880847e-01
1.64551950e+00 -2.00066471e+00 2.54520550e-02 1.21919066e-01
1.64348334e-02 4.00032163e-01 -4.17260677e-01 5.22172511e-01
2.16790646e-01 6.47673234e-02 2.68274844e-01 -3.74790698e-01
5.01205102e-02 6.79246113e-02 -4.97529477e-01 3.58631089e-02
2.97229677e-01 8.40798557e-01 -1.10452652e+00 2.99101062e-02
7.53701404e-02 3.18744898e-01 -1.10623682e+00 4.52973992e-01
-5.15208185e-01 4.59237278e-01 -4.54911232e-01 6.39261901e-01
3.16569984e-01 -2.31795415e-01 4.30499822e-01 2.07567409e-01
-1.46757469e-01 5.17034233e-01 -1.33113694e+00 1.11001050e+00
-4.64289576e-01 2.24430919e-01 -1.70602620e-01 -1.02859020e+00
8.66097331e-01 1.57932937e-01 7.88731992e-01 -1.08401251e+00
-2.37682592e-02 -1.06446557e-01 9.67798233e-02 -5.02103627e-01
1.17926586e+00 5.98586425e-02 3.12794864e-01 6.91885531e-01
-2.16599867e-01 3.34626585e-01 1.66323394e-01 1.28571346e-01
1.00701594e+00 5.15898168e-01 8.39698687e-02 1.64411202e-01
2.22987026e-01 -2.41071090e-01 7.51166463e-01 9.52092052e-01
-3.18473816e-01 4.51169550e-01 3.89101386e-01 -1.15118824e-01
-6.80973351e-01 -4.94605958e-01 2.51933515e-01 1.54770100e+00
-5.64574376e-02 -3.41622621e-01 -5.77145033e-02 -6.22484028e-01
1.91470295e-01 9.40807879e-01 -5.57460427e-01 -3.65583122e-01
-4.89011079e-01 -1.00973129e+00 1.41941264e-01 5.98903894e-01
-6.82945549e-02 -1.26180387e+00 -8.25585902e-01 6.06023848e-01
-2.87198216e-01 -6.87236965e-01 -6.39335513e-01 1.24179423e-01
-9.19698298e-01 -9.40426171e-01 -7.30949938e-01 -2.89629757e-01
4.77401018e-01 5.62756062e-01 1.23123884e+00 7.59388953e-02
-9.18292552e-02 6.88346982e-01 -5.38995922e-01 -3.62045854e-01
-9.17613879e-02 1.30026266e-01 1.08245611e-01 -6.01390116e-02
1.50798976e-01 -5.61878443e-01 -8.73705804e-01 3.50150645e-01
-7.84561574e-01 -2.31665239e-01 4.49743927e-01 9.53895211e-01
1.19282901e-01 -3.52285266e-01 1.35612762e+00 -9.74482894e-01
1.13200355e+00 -9.37554896e-01 -4.43911582e-01 2.49964029e-01
-1.08454418e+00 9.07963235e-03 7.98831165e-01 -8.60287786e-01
-1.20275629e+00 -4.61164355e-01 -8.53951201e-02 -3.05026263e-01
2.17856079e-01 5.62379956e-01 2.57738799e-01 3.55103344e-01
7.54953325e-01 1.66068189e-02 2.34360546e-01 -2.11657643e-01
3.63297015e-01 6.45467699e-01 -2.60007530e-02 -7.34259427e-01
1.94514140e-01 -5.11900485e-02 -4.60233182e-01 -3.96660894e-01
-8.97265911e-01 -4.71447796e-01 1.42243385e-01 -4.09561813e-01
3.39553729e-02 -8.24162781e-01 -8.58767629e-01 9.05490071e-02
-4.42349553e-01 -5.38150966e-01 -5.08125007e-01 4.99452472e-01
-6.07989728e-01 2.26186126e-01 -4.57023680e-01 -1.37212539e+00
-4.78195995e-01 -1.07678449e+00 3.46251875e-01 5.20550907e-01
-3.87367040e-01 -8.18830848e-01 1.17534652e-01 4.56598461e-01
8.40278506e-01 -4.79232892e-02 5.85749447e-01 -7.08808959e-01
-4.16020930e-01 -2.16645617e-02 7.99142793e-02 2.55373031e-01
1.34200588e-01 6.87531754e-02 -9.00475442e-01 -4.48644102e-01
-2.51223058e-01 -6.32578671e-01 3.96274716e-01 3.23881835e-01
1.01732218e+00 -3.82860959e-01 1.61939472e-01 1.88755974e-01
1.18553662e+00 4.48394865e-01 3.58428299e-01 4.93025452e-01
2.56139010e-01 6.28550231e-01 1.04784155e+00 8.98235023e-01
6.75477803e-01 6.72043204e-01 4.46775556e-01 4.13465723e-02
1.36478946e-01 -2.69474566e-01 4.67208922e-01 4.07777190e-01
-2.52469629e-01 -3.40662181e-01 -3.87316495e-01 3.18526924e-01
-2.15977478e+00 -1.21043742e+00 3.50830257e-01 2.48792076e+00
8.01155269e-01 6.17626645e-02 5.74938893e-01 1.00555457e-01
3.70775521e-01 1.27118260e-01 -1.03136432e+00 -4.18063879e-01
-1.42728696e-02 -3.35607417e-02 2.10162744e-01 1.83019131e-01
-6.76820338e-01 6.07865214e-01 6.43862343e+00 3.03543091e-01
-1.31735098e+00 -3.61052901e-01 6.73558652e-01 -6.57970667e-01
-5.51097572e-01 -2.26561099e-01 -6.26394749e-01 4.33851331e-01
1.02503991e+00 -1.32316843e-01 9.00281906e-01 7.31927454e-01
4.89278167e-01 -9.70278606e-02 -1.02645743e+00 6.26740694e-01
2.68741697e-02 -1.01411307e+00 -3.25609148e-01 1.76037923e-01
9.29722369e-01 -4.08303663e-02 3.03004175e-01 1.03267586e+00
5.29558659e-01 -8.03510368e-01 6.36159301e-01 6.62450850e-01
3.99703622e-01 -7.53383756e-01 5.12495875e-01 3.05542916e-01
-7.68796861e-01 -5.84264100e-01 -1.82697847e-01 -3.89647156e-01
-2.36276202e-02 2.50161856e-01 -7.52531648e-01 5.17449677e-01
4.53631967e-01 8.59170794e-01 -3.54260206e-01 1.05225420e+00
2.06473380e-01 6.26238406e-01 -9.77900475e-02 -3.07967812e-01
2.12473407e-01 -3.95708948e-01 2.77888387e-01 7.45345950e-01
3.36178690e-01 1.40833244e-01 1.84382841e-01 5.67885816e-01
1.38028068e-02 3.91174823e-01 -4.43797112e-01 -4.08531755e-01
4.98668164e-01 1.17817712e+00 -3.91397417e-01 -1.46525413e-01
-6.10768616e-01 5.62693179e-01 6.02002323e-01 4.54220146e-01
-8.52621973e-01 8.15601051e-02 7.04686582e-01 6.59421226e-03
2.55593508e-01 -1.33625090e-01 -2.02647761e-01 -1.11733449e+00
-1.51705950e-01 -1.31042349e+00 4.80983377e-01 -3.31584185e-01
-1.31930339e+00 3.31071675e-01 -2.92451352e-01 -1.40398526e+00
-5.25655687e-01 7.24476874e-02 -5.38172841e-01 6.07926965e-01
-1.68301904e+00 -7.09433675e-01 -7.58434385e-02 3.21383446e-01
7.31995881e-01 -5.80931269e-02 6.87714636e-01 5.12005091e-01
-7.93444991e-01 8.20502818e-01 3.43971372e-01 -6.84457064e-01
8.98966730e-01 -1.20882297e+00 1.85357016e-02 5.55497706e-01
-1.95999727e-01 7.24555612e-01 7.43903041e-01 -4.00723845e-01
-1.61136794e+00 -8.10478926e-01 4.10983860e-01 -4.06056195e-01
4.45018709e-01 -1.54651597e-01 -8.52373779e-01 5.66521347e-01
-9.02564973e-02 -2.66620636e-01 9.22258019e-01 5.65636158e-01
1.38287157e-01 -8.70339572e-02 -1.08730853e+00 8.93271148e-01
1.06551719e+00 -1.50344715e-01 -3.20479840e-01 -8.86039957e-02
6.35437250e-01 -4.15469140e-01 -8.01736414e-01 2.24628851e-01
8.31052601e-01 -7.97429740e-01 9.60704684e-01 -7.30647266e-01
4.39499497e-01 7.76015893e-02 7.77106434e-02 -1.47256815e+00
-3.09838146e-01 -7.56253004e-01 -4.67008084e-01 1.36262310e+00
3.44888270e-01 -5.84383309e-01 8.37710500e-01 1.11114645e+00
1.07531779e-01 -9.54284191e-01 -3.20027024e-01 -4.74783987e-01
-2.20026046e-01 -4.61678296e-01 6.67061388e-01 9.34262156e-01
2.46468022e-01 2.31877849e-01 -1.02556336e+00 -2.13445589e-01
2.10316703e-01 1.97025970e-01 7.97774374e-01 -1.13217938e+00
-4.98910338e-01 -5.63678324e-01 3.78462702e-01 -1.19832671e+00
-2.65976209e-02 -4.50812131e-01 1.27978712e-01 -1.27382457e+00
-1.31071985e-01 -8.09088647e-01 -7.12775469e-01 2.55510509e-01
-4.50516790e-01 -1.29005298e-01 3.97461206e-01 1.36016801e-01
-8.10266078e-01 9.12125468e-01 1.29877150e+00 2.34542623e-01
-7.29865015e-01 4.78247434e-01 -1.22868288e+00 2.20997125e-01
8.89063358e-01 -2.08951339e-01 -7.92932570e-01 -4.72396277e-02
4.15763170e-01 2.95037806e-01 -1.26438543e-01 -7.09576249e-01
1.73646003e-01 -5.34532249e-01 1.73275769e-01 -4.84548807e-01
2.87943691e-01 -7.13265598e-01 -1.06556028e-01 1.81889683e-01
-7.37074912e-01 2.62006730e-01 2.27913007e-01 7.48224318e-01
5.60716614e-02 -7.33484626e-02 3.47719699e-01 -4.87500355e-02
-5.69586933e-01 1.87555417e-01 -4.08908874e-01 1.11725867e-01
8.71207535e-01 -1.65811941e-01 -1.20843001e-01 -7.44287193e-01
-5.70846379e-01 7.30582595e-01 1.70671716e-01 7.84070969e-01
4.00394112e-01 -1.09022081e+00 -3.44701886e-01 1.60624206e-01
-7.01057464e-02 -3.51590872e-01 4.49591607e-01 8.66800368e-01
2.43643090e-01 2.35366181e-01 -2.35295251e-01 1.64879337e-02
-9.55214322e-01 4.64364707e-01 2.10642844e-01 -5.18436015e-01
-3.60020548e-01 4.66242045e-01 -2.31871113e-01 -4.67848241e-01
6.48518622e-01 -2.33008161e-01 -6.32296145e-01 5.08715332e-01
4.69933569e-01 7.26524413e-01 8.98104720e-03 1.37723321e-02
-3.71979810e-02 -2.04428524e-01 -2.00107276e-01 -4.52347286e-02
1.47701657e+00 -3.40275347e-01 5.13904095e-01 4.03452992e-01
5.67066312e-01 4.95104939e-02 -1.58508503e+00 -3.92700315e-01
1.15797244e-01 -4.76889163e-01 1.55193821e-01 -1.12791646e+00
-9.38627183e-01 4.26403224e-01 5.82650363e-01 3.95389855e-01
1.13453126e+00 -5.56952655e-01 7.76295185e-01 2.23431528e-01
1.73721105e-01 -1.39990163e+00 2.97990650e-01 4.61794406e-01
8.09817135e-01 -1.50874352e+00 3.36736701e-02 2.31355578e-01
-1.02283323e+00 9.04887855e-01 8.70713174e-01 -7.60810748e-02
3.60846668e-01 -9.03331563e-02 6.96574152e-02 2.13384628e-01
-1.22663832e+00 -3.02002341e-01 2.06820548e-01 3.41940790e-01
7.55963743e-01 1.97673850e-02 -5.79378188e-01 8.92278016e-01
1.18767612e-01 3.16455066e-01 6.01455688e-01 1.15305579e+00
-6.54635429e-02 -1.24957585e+00 -2.17871994e-01 9.43868101e-01
-5.17029166e-01 -1.07123747e-01 1.01426309e-02 4.98485655e-01
-2.81039029e-01 1.09732699e+00 1.95972156e-02 -4.74497676e-01
5.20291984e-01 -1.26788437e-01 2.13430524e-01 -5.40987134e-01
-9.33980465e-01 1.57853603e-01 2.03371629e-01 -5.75216889e-01
-3.22813898e-01 -9.40152466e-01 -1.02344167e+00 -1.53587252e-01
-4.80068564e-01 2.00091213e-01 4.95356023e-01 9.11056399e-01
6.23070598e-01 8.15562189e-01 8.81164074e-01 -9.71375644e-01
-1.19344425e+00 -9.14578497e-01 -4.55474436e-01 3.86143774e-01
4.05680686e-01 -7.73390412e-01 -4.19571042e-01 -5.97407639e-01] | [4.2249836921691895, 2.351856231689453] |
28cb714f-1687-4428-b984-333d2a4f870f | hallucinating-very-low-resolution-and | 1811.04645 | null | http://arxiv.org/abs/1811.04645v4 | http://arxiv.org/pdf/1811.04645v4.pdf | Hallucinating very low-resolution and obscured face images | Most of the face hallucination methods are designed for complete inputs. They
will not work well if the inputs are very tiny or contaminated by large
occlusion. Inspired by this fact, we propose an obscured face hallucination
network(OFHNet). The OFHNet consists of four parts: an inpainting network, an
upsampling network, a discriminative network, and a fixed facial landmark
detection network. The inpainting network restores the low-resolution(LR)
obscured face images. The following upsampling network is to upsample the
output of inpainting network. In order to ensure the generated
high-resolution(HR) face images more photo-realistic, we utilize the
discriminative network and the facial landmark detection network to better the
result of upsampling network. In addition, we present a semantic structure
loss, which makes the generated HR face images more pleasing. Extensive
experiments show that our framework can restore the appealing HR face images
from 1/4 missing area LR face images with a challenging scaling factor of 8x. | ['Ting Sun', 'Song Ding', 'Lianping Yang', 'Bin Shao', 'Xiangde Zhang'] | 2018-11-12 | null | null | null | null | ['face-hallucination'] | ['computer-vision'] | [ 1.83009028e-01 4.04356688e-01 1.43546626e-01 -3.86098623e-01
-4.59068775e-01 8.13093036e-02 2.14016527e-01 -1.07883656e+00
1.17168754e-01 7.77887583e-01 5.48607528e-01 3.46176863e-01
3.12308580e-01 -8.71272445e-01 -8.35961580e-01 -6.64418876e-01
3.67953449e-01 -7.80948773e-02 -1.47595823e-01 -2.04537600e-01
-1.94532722e-01 7.32909560e-01 -1.59140313e+00 4.53238547e-01
7.54111648e-01 1.03335822e+00 3.04156721e-01 1.22004887e-02
8.65150616e-02 7.45648265e-01 -5.64403057e-01 -8.13066289e-02
4.85302895e-01 -6.24189794e-01 -3.18104953e-01 4.16555822e-01
3.54929715e-01 -9.68561828e-01 -7.81218588e-01 1.16055024e+00
6.59226954e-01 1.17129870e-01 3.15869600e-01 -1.21051180e+00
-1.05442667e+00 4.19301242e-01 -1.03476310e+00 -1.92691639e-01
4.97854561e-01 2.36379076e-02 1.08580090e-01 -1.36515749e+00
6.00713730e-01 2.04460859e+00 5.99128246e-01 7.73658931e-01
-1.07033682e+00 -9.57728088e-01 -1.57617137e-01 -4.79802229e-02
-1.61432517e+00 -9.27520454e-01 1.24042082e+00 1.48942158e-01
1.90314993e-01 1.55879498e-01 3.57081473e-01 1.08996856e+00
-6.17064647e-02 4.94008183e-01 1.07499528e+00 -2.60128289e-01
-1.08507752e-01 1.82859406e-01 -5.47681630e-01 6.15505278e-01
1.66498557e-01 2.16578946e-01 -3.17543715e-01 -1.13502577e-01
1.40371144e+00 6.04700208e-01 -6.91961229e-01 -1.87935326e-02
-6.01173639e-01 6.26049876e-01 6.14951909e-01 2.31475793e-02
-3.97194624e-01 -7.25743175e-02 2.17940956e-02 3.82562459e-01
6.39719605e-01 1.31964609e-01 3.11867166e-02 6.48878932e-01
-1.10349381e+00 2.65023522e-02 3.06569308e-01 9.75759447e-01
9.40535247e-01 3.55725676e-01 -1.35329917e-01 1.29691625e+00
3.96053404e-01 4.93292212e-01 3.05481911e-01 -1.25575197e+00
3.34964186e-01 3.85416627e-01 7.88247585e-02 -1.03230095e+00
-4.77983207e-02 -2.23530635e-01 -1.16302967e+00 3.82830590e-01
-7.22275674e-02 -1.20697752e-01 -8.62250328e-01 1.83203816e+00
3.32160830e-01 1.63937956e-01 1.78172868e-02 1.22074437e+00
8.77786040e-01 8.43704939e-01 -3.21228355e-01 -5.55894256e-01
1.34127355e+00 -1.15359581e+00 -1.10070646e+00 -1.36346653e-01
-2.78940111e-01 -1.02740157e+00 9.60519433e-01 2.78887063e-01
-1.36166131e+00 -1.00887609e+00 -1.04769409e+00 -4.72216219e-01
1.65883183e-01 3.25771809e-01 2.49599352e-01 2.22996786e-01
-9.85647023e-01 5.41384399e-01 -2.35225558e-01 -1.12806395e-01
7.85902143e-01 1.93208963e-01 -5.73518753e-01 -6.37644410e-01
-1.12130284e+00 6.00486159e-01 1.32136717e-02 3.84156138e-01
-1.18623006e+00 -6.13719225e-01 -1.01349235e+00 8.05805251e-02
2.08248258e-01 -5.92775583e-01 8.71969163e-01 -1.27574790e+00
-1.42647517e+00 6.29582345e-01 -2.04658836e-01 6.02713563e-02
5.73612213e-01 -1.44805655e-01 -6.09242022e-01 4.33791369e-01
1.01658411e-01 7.48144865e-01 1.42208648e+00 -1.62863016e+00
-1.10588722e-01 -6.26622200e-01 -3.54427189e-01 2.81497866e-01
-2.20149636e-01 -2.25458536e-02 -4.22333717e-01 -8.99515152e-01
1.63044482e-01 -4.15476352e-01 4.56492305e-02 5.17471790e-01
-4.84537870e-01 3.98627758e-01 1.52868652e+00 -1.04557860e+00
7.50629008e-01 -2.49650049e+00 -7.17258081e-02 -1.04794830e-01
3.59130412e-01 2.41202354e-01 -5.12367189e-01 2.56976932e-01
-3.93376201e-01 -1.89266652e-01 -9.67619270e-02 -6.81884408e-01
-4.40250635e-01 2.95329809e-01 -7.41614997e-01 5.43610215e-01
3.46527129e-01 6.65565073e-01 -5.92950106e-01 -4.92624193e-01
1.80304393e-01 1.13845646e+00 -4.24839586e-01 5.85080981e-01
-2.61626244e-02 4.75163758e-01 -3.84989202e-01 7.72163153e-01
1.47455013e+00 5.28872609e-02 -9.18789729e-02 -3.92184854e-01
2.93484679e-03 -3.02304000e-01 -1.17607784e+00 1.64046764e+00
-2.08137959e-01 2.38028154e-01 5.66451073e-01 -3.61309409e-01
1.30647540e+00 3.69266421e-01 2.75184661e-01 -6.47164702e-01
6.77286685e-02 1.01339139e-01 -5.14216304e-01 -6.07458711e-01
9.43064466e-02 -4.20647144e-01 6.51854038e-01 2.60378212e-01
-9.42043886e-02 7.81577379e-02 -4.99246180e-01 2.17480268e-02
6.27917171e-01 1.62645712e-01 -5.77351823e-02 1.68749510e-04
6.60185456e-01 -7.52151132e-01 8.56684446e-01 1.79517330e-04
9.11590364e-03 1.23737717e+00 4.81901109e-01 -6.57937407e-01
-1.22757924e+00 -9.07952189e-01 -1.55469403e-01 5.43130875e-01
3.29499245e-01 -5.83365969e-02 -1.01850724e+00 -2.94102788e-01
-1.52913943e-01 2.91069001e-01 -6.48789227e-01 -4.13576245e-01
-5.71489334e-01 -3.56656253e-01 5.14212549e-01 4.12718922e-01
1.11153388e+00 -1.26819634e+00 -2.93276925e-02 -2.33041309e-02
-2.76811689e-01 -1.01102924e+00 -9.09152806e-01 -5.03721833e-01
-7.35541701e-01 -9.37784135e-01 -1.05257010e+00 -1.03967929e+00
1.09598589e+00 6.76310301e-01 6.45555496e-01 2.15210468e-01
-1.32410631e-01 -3.28459829e-01 -1.26892179e-01 -7.60217756e-02
-2.04434484e-01 -5.09352505e-01 1.31304905e-01 4.50711340e-01
6.20307364e-02 -9.49845076e-01 -7.37893462e-01 4.13629115e-01
-1.18828845e+00 3.12374562e-01 7.25608289e-01 9.10598814e-01
5.26577592e-01 2.03959629e-01 5.85899472e-01 -6.42608762e-01
4.51705277e-01 -3.89131099e-01 -3.34104270e-01 -5.94001561e-02
-3.49293500e-01 -1.85466066e-01 1.05088413e+00 -6.25298679e-01
-1.41665268e+00 9.55014396e-03 -2.23656386e-01 -1.04312956e+00
-3.01651619e-02 -3.30665916e-01 -7.92528212e-01 -2.41018593e-01
4.98688042e-01 2.95558125e-01 4.36869264e-01 -7.87508130e-01
2.22077832e-01 7.27972686e-01 6.74503982e-01 -3.23720336e-01
1.00053430e+00 7.24206805e-01 -6.59700856e-02 -7.69365907e-01
-8.37580383e-01 2.30633914e-01 -2.02953681e-01 -1.15750298e-01
6.94283724e-01 -1.17327464e+00 -4.85152990e-01 3.82231355e-01
-1.27112842e+00 5.13014868e-02 -4.08574015e-01 2.96713084e-01
-4.39904660e-01 3.91757220e-01 -8.55050623e-01 -8.31542373e-01
-3.11049461e-01 -1.12119293e+00 1.28932226e+00 6.61350906e-01
4.20260519e-01 -2.72451580e-01 -3.43530118e-01 3.94372433e-01
5.51716387e-01 3.39932829e-01 6.90778315e-01 2.30288982e-01
-5.88074625e-01 2.12331727e-01 -5.65457761e-01 6.12199366e-01
3.44060779e-01 -2.33506352e-01 -1.27121937e+00 -4.71399426e-01
4.24190223e-01 -3.07338566e-01 8.43818545e-01 2.46870175e-01
1.39786065e+00 -7.35750198e-01 -1.67573377e-01 1.00717783e+00
1.35246909e+00 1.01467147e-01 1.21540296e+00 -2.22540259e-01
7.58904099e-01 7.24449337e-01 5.37142336e-01 3.17801178e-01
6.77652806e-02 5.48640549e-01 5.56856334e-01 -6.38812482e-01
-5.86873651e-01 -7.58101940e-01 4.72237825e-01 3.98589730e-01
-9.27508771e-02 8.21937919e-02 -9.87251848e-02 2.41231292e-01
-1.59397674e+00 -1.12610078e+00 3.33416045e-01 2.03525472e+00
8.07888389e-01 -3.65570128e-01 -1.09875903e-01 6.59973696e-02
9.68147337e-01 3.11432332e-01 -6.82338178e-01 -8.53795260e-02
-3.30782205e-01 6.78986013e-02 -1.23970844e-01 5.63700020e-01
-4.62477177e-01 8.71638417e-01 5.54463625e+00 8.62087131e-01
-1.07714343e+00 1.71114951e-01 1.03255355e+00 -2.71065794e-02
-4.70587045e-01 -3.36082876e-02 -6.05472922e-01 5.96669853e-01
1.93821207e-01 2.56347954e-01 7.79112339e-01 9.48600352e-01
4.06565517e-01 2.47501209e-01 -6.25662863e-01 1.29179001e+00
4.74939853e-01 -1.04696560e+00 3.67756188e-01 1.26462281e-01
6.04025543e-01 -5.42166650e-01 2.42963240e-01 4.09582220e-02
-2.38994867e-01 -1.36964250e+00 4.81122285e-01 7.37484038e-01
1.26211214e+00 -1.13408399e+00 4.32038069e-01 2.04598054e-01
-1.05942833e+00 -2.11401135e-01 -9.10844326e-01 1.78674042e-01
1.48500308e-01 8.07786286e-01 -2.73932934e-01 4.54385757e-01
7.48606443e-01 5.56522369e-01 -2.70315260e-01 5.52855372e-01
-3.03004116e-01 1.88476574e-02 -7.71726146e-02 8.73240590e-01
-1.92233801e-01 -3.76320273e-01 4.69299674e-01 6.45961821e-01
3.54426444e-01 4.02288854e-01 -4.89600142e-03 1.23477662e+00
-6.37140989e-01 -1.61589548e-01 -8.09728563e-01 3.34443718e-01
6.13238990e-01 1.47379220e+00 -1.71863154e-01 -1.44836590e-01
-2.31838346e-01 1.30071998e+00 1.51081368e-01 6.03901803e-01
-7.81187713e-01 -4.84658480e-01 6.87301934e-01 5.25919378e-01
9.37711895e-02 6.75397813e-02 1.50497645e-01 -1.19072866e+00
2.58839637e-01 -9.76736724e-01 -4.96987812e-02 -1.20812488e+00
-1.21107733e+00 9.25434411e-01 -3.42527419e-01 -1.16652143e+00
1.60181493e-01 -1.26498163e-01 -6.42930150e-01 1.14149714e+00
-1.78909802e+00 -1.24013412e+00 -7.92624176e-01 8.01838815e-01
6.04578018e-01 -1.29975289e-01 5.21838069e-01 5.44595778e-01
-6.76908374e-01 5.27084291e-01 -2.19594672e-01 1.01450264e-01
9.73292410e-01 -4.93367970e-01 -3.48971114e-02 8.02552700e-01
-4.91873294e-01 7.17814803e-01 5.02348542e-01 -6.46314442e-01
-1.36384273e+00 -1.50997591e+00 4.51769859e-01 1.27922058e-01
-3.30244660e-01 -3.11024457e-01 -1.10240376e+00 6.28915012e-01
4.37138490e-02 4.42955196e-01 1.59618214e-01 -6.55231774e-01
-6.05880201e-01 -5.08690357e-01 -1.74701011e+00 5.89347243e-01
1.13062763e+00 -5.14681339e-01 -4.63132381e-01 3.18222225e-01
9.02682126e-01 -2.63464570e-01 -7.63629019e-01 3.49994361e-01
4.91814047e-01 -1.08239925e+00 1.10666776e+00 -8.50906819e-02
7.67673194e-01 -6.32321596e-01 -1.28560245e-01 -1.10793793e+00
-3.22805405e-01 -8.30033600e-01 -1.15006961e-01 1.36379802e+00
-1.32177278e-01 -6.99006736e-01 5.07800758e-01 2.21604317e-01
-5.79815954e-02 -7.29113877e-01 -6.83340490e-01 -5.71685135e-01
-2.93983668e-01 3.64931375e-01 1.07498181e+00 1.01728117e+00
-3.68247658e-01 4.26066697e-01 -8.83626044e-01 2.86685228e-01
9.97853339e-01 1.10276826e-01 6.48601234e-01 -1.04320526e+00
-7.16252113e-03 1.61958084e-01 -3.34540084e-02 -1.10364151e+00
3.83844078e-02 -4.13957924e-01 -1.55028433e-01 -1.13470411e+00
2.87934601e-01 -1.23325616e-01 1.40598893e-01 4.97890085e-01
2.35743057e-02 5.67094266e-01 -4.17357832e-02 5.02005458e-01
-4.10639420e-02 1.03965569e+00 1.86164665e+00 2.27828398e-01
-1.30982280e-01 -3.88802767e-01 -1.00772774e+00 7.18367100e-01
6.37485981e-01 -2.67356038e-01 -5.87899685e-01 -5.09736419e-01
-4.10092056e-01 4.70841140e-01 3.89874607e-01 -9.45703566e-01
2.81413198e-02 -7.59684294e-02 1.16356778e+00 -3.78734469e-01
7.00534284e-01 -8.65849435e-01 4.44682777e-01 2.55547911e-01
-1.56668782e-01 -1.38719827e-01 -2.91802157e-02 3.74269545e-01
-3.62201035e-01 2.43803129e-01 1.41589820e+00 -2.19268605e-01
-1.30928114e-01 6.61741793e-01 3.21346849e-01 -2.63498604e-01
9.83894825e-01 -3.14084321e-01 -5.02295911e-01 -5.72639406e-01
-3.63355219e-01 7.92873874e-02 7.06656635e-01 5.76272011e-01
1.20136833e+00 -1.56701255e+00 -7.90982187e-01 8.65985692e-01
-3.72816622e-01 2.24076048e-01 4.84705567e-01 5.11189640e-01
-5.80349684e-01 -4.07613404e-02 -6.63078785e-01 -1.93187930e-02
-1.05519700e+00 8.46288323e-01 3.89138907e-01 2.97307730e-01
-9.50380862e-01 5.06646335e-01 5.45982599e-01 -2.96652794e-01
2.26559311e-01 2.51799554e-01 -2.47433379e-01 -2.28447631e-01
1.14766026e+00 3.26508254e-01 -3.82244349e-01 -8.15236270e-01
-9.76415798e-02 6.39397979e-01 -2.14024931e-01 -1.08415976e-01
1.52490580e+00 -3.36812288e-01 -4.60663646e-01 -1.65546507e-01
1.27948058e+00 1.08904906e-01 -1.50168073e+00 -2.31714830e-01
-9.31434810e-01 -8.42176855e-01 -2.48680618e-02 -3.41039658e-01
-1.46596348e+00 8.08160841e-01 5.14800370e-01 -1.36834830e-01
1.64169049e+00 -3.04547936e-01 1.02190173e+00 -1.31070316e-01
3.00687492e-01 -8.29687297e-01 2.68938094e-01 1.00280091e-01
1.38692939e+00 -8.97840619e-01 5.73980398e-02 -7.07396746e-01
-4.62777823e-01 1.16733873e+00 9.25342739e-01 -3.53741765e-01
4.91589457e-01 2.95328408e-01 -7.82912225e-02 -9.23620686e-02
-4.99080628e-01 1.80279046e-01 -4.54239696e-02 6.38891876e-01
-1.09211439e-02 -3.68726403e-01 -8.32773075e-02 6.24672174e-01
-8.46335515e-02 1.31597564e-01 5.61467826e-01 3.93391758e-01
-5.08551955e-01 -7.94404507e-01 -7.60391176e-01 1.34446397e-01
-3.42688799e-01 4.21891883e-02 -1.56690449e-01 5.83306849e-01
3.70949447e-01 9.83050585e-01 -6.47428408e-02 -3.55986178e-01
3.14122647e-01 -2.26469398e-01 3.51323754e-01 -5.39186537e-01
-4.12820205e-02 2.67179489e-01 -3.47593755e-01 -8.31827044e-01
-2.03439705e-02 -8.61101784e-03 -1.16023886e+00 -5.31359613e-01
-2.18022689e-01 2.90051997e-02 3.00956845e-01 4.69672173e-01
4.31422710e-01 3.86693865e-01 1.09858525e+00 -9.14321423e-01
-4.23870951e-01 -1.03731120e+00 -9.38248456e-01 2.93746710e-01
6.97263598e-01 -4.34106261e-01 -4.69395459e-01 4.01198044e-02] | [12.812640190124512, -0.04418937489390373] |
bd5d88eb-e0e0-481f-8707-e0ebe879191e | style-agnostic-3d-reconstruction-via | 2110.10784 | null | https://arxiv.org/abs/2110.10784v1 | https://arxiv.org/pdf/2110.10784v1.pdf | Style Agnostic 3D Reconstruction via Adversarial Style Transfer | Reconstructing the 3D geometry of an object from an image is a major challenge in computer vision. Recently introduced differentiable renderers can be leveraged to learn the 3D geometry of objects from 2D images, but those approaches require additional supervision to enable the renderer to produce an output that can be compared to the input image. This can be scene information or constraints such as object silhouettes, uniform backgrounds, material, texture, and lighting. In this paper, we propose an approach that enables a differentiable rendering-based learning of 3D objects from images with backgrounds without the need for silhouette supervision. Instead of trying to render an image close to the input, we propose an adversarial style-transfer and domain adaptation pipeline that allows to translate the input image domain to the rendered image domain. This allows us to directly compare between a translated image and the differentiable rendering of a 3D object reconstruction in order to train the 3D object reconstruction network. We show that the approach learns 3D geometry from images with backgrounds and provides a better performance than constrained methods for single-view 3D object reconstruction on this task. | ['Hilde Kuehne', 'Oliver Deussen', 'Bastian Goldluecke', 'Felix Petersen'] | 2021-10-20 | null | null | null | null | ['3d-object-reconstruction', 'object-reconstruction'] | ['computer-vision', 'computer-vision'] | [ 6.14218295e-01 2.77657151e-01 4.26082939e-01 -4.81017977e-01
-6.72580779e-01 -8.40515137e-01 6.11447096e-01 -4.19568300e-01
-2.91567773e-01 2.33746573e-01 -3.66584510e-01 -3.27894241e-01
5.26426554e-01 -8.33119154e-01 -1.34487450e+00 -5.11857331e-01
4.43569988e-01 7.70129502e-01 3.88503551e-01 -6.68101236e-02
1.89214960e-01 9.45387959e-01 -1.48457658e+00 1.92369163e-01
4.67796952e-01 9.18660283e-01 3.23592007e-01 1.04273582e+00
-2.96701610e-01 3.42612326e-01 -5.47281802e-01 -1.69007584e-01
7.85936117e-01 -3.92966717e-01 -6.13471508e-01 6.31305218e-01
9.75863695e-01 -7.89115548e-01 -9.42535922e-02 9.81782794e-01
2.29606837e-01 -8.56038462e-03 9.00012314e-01 -9.45247591e-01
-5.31309664e-01 -2.94962019e-01 -5.13346910e-01 -3.02777737e-01
6.35629892e-01 3.94028500e-02 4.31338757e-01 -8.16521645e-01
9.86571968e-01 1.48822534e+00 4.23422992e-01 6.27281547e-01
-1.52728152e+00 -2.68213272e-01 3.57050478e-01 -3.84470344e-01
-9.19500351e-01 -2.46596202e-01 1.09859717e+00 -6.70747399e-01
6.20728612e-01 2.24366471e-01 7.07128227e-01 9.92327332e-01
-1.08381681e-01 4.56970006e-01 1.32986057e+00 -6.18241072e-01
3.28562468e-01 3.05267543e-01 -4.35339123e-01 7.55007505e-01
-2.01624900e-01 2.20094621e-01 -1.89329520e-01 2.72016022e-02
1.42729235e+00 1.58752441e-01 -4.52680022e-01 -8.49397779e-01
-1.14286339e+00 5.00483215e-01 5.69135785e-01 -3.38401705e-01
-2.70583451e-01 1.05285704e-01 -9.00279731e-03 2.66474187e-01
6.57245398e-01 3.85634661e-01 -4.17784810e-01 2.26535693e-01
-6.24054551e-01 2.76320547e-01 8.41492712e-01 1.06079435e+00
9.95623827e-01 -9.11067799e-03 1.50152460e-01 1.74206167e-01
1.84411332e-01 6.90895677e-01 -2.29569018e-01 -1.27301145e+00
4.01251584e-01 5.55158198e-01 3.65692705e-01 -6.10343218e-01
9.72792506e-02 6.51115477e-02 -4.16473925e-01 1.05665743e+00
6.26673758e-01 1.79674610e-01 -1.06411874e+00 1.35639179e+00
7.16921151e-01 1.95946187e-01 4.94923349e-03 1.12655306e+00
6.53344691e-01 5.52281022e-01 -5.58767796e-01 3.05116326e-01
1.01447392e+00 -6.53578579e-01 -1.19749092e-01 -2.69887447e-01
4.23749499e-02 -7.10040390e-01 1.27477312e+00 2.16516510e-01
-1.24253392e+00 -5.05418122e-01 -9.58365500e-01 -4.66262996e-01
-2.39273474e-01 -2.52388358e-01 3.80545370e-02 3.65230918e-01
-8.90344441e-01 7.69926965e-01 -1.03045022e+00 -2.01944530e-01
4.29102629e-01 2.50078827e-01 -5.55790305e-01 -1.94634572e-01
-4.83700752e-01 9.13144469e-01 3.29400837e-01 -2.37663433e-01
-1.07346678e+00 -8.80061746e-01 -9.62321460e-01 -2.42609963e-01
3.75702143e-01 -8.10707808e-01 1.00617671e+00 -1.24303913e+00
-1.73836970e+00 1.42357326e+00 -7.45690688e-02 -1.77173670e-02
9.22663093e-01 -5.90512939e-02 4.43991244e-01 3.72993112e-01
-1.70185313e-01 4.06979680e-01 1.31566465e+00 -1.59993219e+00
-1.94910496e-01 -6.36862576e-01 4.46518689e-01 3.92388999e-01
3.85176212e-01 -2.01315477e-01 -5.40233552e-01 -3.35146785e-01
2.30686739e-01 -8.17494154e-01 -1.05151013e-01 8.54380071e-01
-4.93977159e-01 3.11179578e-01 1.09065318e+00 -7.44004428e-01
-2.25716472e-01 -2.07985783e+00 3.92161757e-01 1.23094752e-01
2.80958470e-02 1.66807666e-01 -2.96523839e-01 -2.85112355e-02
1.56352967e-02 -9.31069721e-03 -2.68168658e-01 -7.52389491e-01
-1.62523806e-01 3.94191533e-01 -3.18874985e-01 6.43794417e-01
6.46436572e-01 9.44597840e-01 -8.29795659e-01 -1.29004315e-01
4.94454056e-01 9.52985525e-01 -6.56615555e-01 8.17683160e-01
-7.11662531e-01 1.28521240e+00 -4.01141226e-01 2.41865650e-01
9.77245271e-01 -1.07208438e-01 -1.65305868e-01 -1.12449817e-01
3.81301418e-02 2.19363108e-01 -1.03107214e+00 1.93508089e+00
-6.05308414e-01 5.01475155e-01 3.66964459e-01 -1.06241691e+00
1.01658094e+00 2.12454185e-01 1.49968356e-01 -4.66961503e-01
1.32512242e-01 -6.35227840e-03 -5.37563503e-01 -3.86058956e-01
4.73208465e-02 -4.23229724e-01 2.24084526e-01 5.11458993e-01
-6.76210225e-02 -1.12367642e+00 -4.67530131e-01 -3.73112857e-02
6.98077679e-01 8.74261320e-01 -1.76158950e-01 1.48439288e-01
3.65698934e-01 -1.66661918e-01 2.38611326e-01 2.89338022e-01
4.69297141e-01 1.08938456e+00 5.07991672e-01 -5.56774437e-01
-1.53454483e+00 -1.50919807e+00 -4.90670092e-02 5.74555457e-01
2.90426046e-01 2.74296433e-01 -7.13178039e-01 -7.93581843e-01
4.81398255e-02 6.12410605e-01 -6.16572380e-01 9.12041962e-02
-7.83451855e-01 2.68299192e-01 1.79442987e-01 4.15985256e-01
3.93307060e-01 -7.03003645e-01 -9.96215105e-01 -1.34554937e-01
-2.16290224e-02 -1.36996353e+00 -5.79788029e-01 1.83467790e-01
-9.37656641e-01 -1.12597108e+00 -8.00234497e-01 -7.77292907e-01
1.19228280e+00 1.56943396e-01 1.26902187e+00 -2.03787889e-02
-3.33416104e-01 5.77178419e-01 -6.74953908e-02 -4.36429590e-01
-7.75669158e-01 -4.49501574e-01 -1.09911703e-01 1.28387079e-01
-2.42376789e-01 -8.22835922e-01 -5.32799363e-01 3.79547179e-01
-1.02918386e+00 4.39588577e-01 1.65148333e-01 6.51858330e-01
9.45775092e-01 -2.45901078e-01 -6.24323227e-02 -1.02107513e+00
-7.23250359e-02 9.08899829e-02 -1.07070935e+00 1.71710730e-01
-4.08920944e-02 1.87871829e-01 7.50868738e-01 -8.45868170e-01
-1.04020607e+00 4.71738577e-01 -1.24065384e-01 -9.52744603e-01
-4.24979597e-01 -1.27770409e-01 -4.27141368e-01 -2.22863048e-01
7.57328629e-01 1.77523598e-01 2.95970798e-01 -7.14462698e-01
5.48708260e-01 3.39719057e-01 6.93531454e-01 -7.44174361e-01
1.17037034e+00 9.39298093e-01 1.86071336e-01 -5.53027034e-01
-8.53092968e-01 -1.35359541e-01 -1.18440330e+00 -1.10182904e-01
1.03731465e+00 -8.16220701e-01 -5.08803189e-01 2.09635735e-01
-1.39884377e+00 -7.50504017e-01 -4.32874709e-01 2.50938445e-01
-1.03145158e+00 1.17133908e-01 -3.19010854e-01 -5.15866339e-01
-3.05071343e-02 -1.25537050e+00 1.52086222e+00 -2.25964785e-02
1.53358355e-01 -9.89104748e-01 -3.43788236e-01 3.43347996e-01
2.83976167e-01 8.53498280e-01 1.02954674e+00 -5.45876883e-02
-9.49492455e-01 -2.19715104e-01 -2.29120851e-01 6.70240641e-01
2.71325767e-01 -1.31909549e-01 -1.23066866e+00 -2.31966600e-01
2.73477465e-01 -4.65159357e-01 3.47799152e-01 1.48193881e-01
1.07956409e+00 -3.32558513e-01 -1.25762567e-01 1.09510911e+00
1.48720443e+00 3.35942134e-02 5.53999126e-01 6.90347282e-03
9.39000189e-01 6.59242332e-01 3.93099546e-01 7.20620900e-02
1.58709258e-01 7.76263893e-01 7.86547124e-01 -3.74521673e-01
-3.42728883e-01 -5.35506129e-01 2.04951987e-01 1.24039285e-01
-1.28604233e-01 1.00101404e-01 -6.80047095e-01 7.52070919e-02
-1.27592957e+00 -6.11554503e-01 2.77366266e-02 2.53081822e+00
7.33974516e-01 2.14265566e-02 -9.82587859e-02 -6.35147989e-02
4.00256991e-01 -2.10204169e-01 -1.05999291e+00 -3.53159577e-01
1.25810027e-01 1.93268403e-01 3.43387038e-01 8.00721288e-01
-7.37731934e-01 8.20945024e-01 5.90802622e+00 1.79844454e-01
-1.43620372e+00 -1.04323752e-01 4.50823873e-01 1.06934840e-02
-5.41679859e-01 4.98794354e-02 -4.28419143e-01 1.26136884e-01
2.73016006e-01 5.72436787e-02 7.12537050e-01 6.16651773e-01
-2.96080988e-02 3.78140770e-02 -1.66919136e+00 1.07595742e+00
2.61800528e-01 -1.14292097e+00 1.71447411e-01 1.47300422e-01
7.65565991e-01 -8.16329718e-02 1.60868138e-01 -6.41723201e-02
2.68224955e-01 -1.00407898e+00 9.36733305e-01 5.30605555e-01
1.04024589e+00 -3.67106527e-01 2.17616379e-01 7.82472670e-01
-7.14643419e-01 4.80511814e-01 -3.61066788e-01 -9.89394858e-02
3.23650599e-01 3.70015293e-01 -1.04311621e+00 4.48447317e-01
6.25338435e-01 4.70227033e-01 -2.46130213e-01 6.11506820e-01
-2.95724541e-01 1.45483166e-01 -4.83464450e-01 4.52296495e-01
-1.93150397e-02 -5.14254749e-01 6.99263990e-01 6.14005327e-01
2.03670949e-01 8.67612734e-02 2.61980623e-01 1.46351171e+00
-8.71851891e-02 -1.76664516e-01 -1.03863204e+00 1.99634999e-01
-1.10003568e-01 9.40613747e-01 -6.51903808e-01 -2.10763991e-01
-3.70735198e-01 1.41609359e+00 4.98414338e-01 6.06545866e-01
-6.13393128e-01 -2.17805454e-03 4.96336281e-01 4.59357202e-01
4.78708148e-01 -4.83368337e-01 -9.72604901e-02 -1.30730855e+00
2.36896604e-01 -6.25937581e-01 -2.75259167e-01 -1.23720157e+00
-1.08160651e+00 6.15712166e-01 5.00187464e-02 -1.31940341e+00
-1.39678702e-01 -8.72741997e-01 -5.93727231e-01 1.31903136e+00
-1.65604842e+00 -1.09341908e+00 -4.72217381e-01 6.83261812e-01
4.17195320e-01 1.73038140e-01 7.73022294e-01 -8.07951465e-02
2.30386332e-01 1.59479484e-01 -1.44791871e-01 1.82759404e-01
6.57075047e-01 -1.40397739e+00 5.71420252e-01 5.71958363e-01
1.36826932e-01 2.38471821e-01 5.31025946e-01 -3.51823956e-01
-1.59053552e+00 -1.08832419e+00 1.70581967e-01 -9.90272105e-01
2.48190258e-02 -7.45420456e-01 -9.27087128e-01 8.00025463e-01
-5.04661426e-02 4.22219157e-01 8.65133181e-02 -4.00250584e-01
-6.44948125e-01 4.85863164e-02 -1.34298265e+00 6.18044794e-01
1.00718570e+00 -6.91366196e-01 -4.40075099e-01 3.01526815e-01
7.35614479e-01 -1.08678675e+00 -7.62402773e-01 1.00191645e-01
4.96738255e-01 -9.88120854e-01 1.17654014e+00 -6.84303463e-01
4.89614874e-01 -5.56499183e-01 -2.30131701e-01 -1.40732729e+00
3.31304103e-01 -4.83894318e-01 1.26593828e-01 8.26847255e-01
9.11151618e-02 -6.35285139e-01 7.84423292e-01 9.64348316e-01
-1.77240372e-03 -5.06144106e-01 -7.64901698e-01 -7.78661311e-01
3.38093966e-01 -3.08283001e-01 5.32702565e-01 6.66502833e-01
-7.58536458e-01 2.00713947e-01 -6.90457970e-02 5.00704288e-01
7.99221098e-01 5.66885531e-01 1.27152193e+00 -1.22870505e+00
-4.56450999e-01 -8.56114998e-02 -5.53063452e-01 -1.45681453e+00
4.03009325e-01 -1.10495734e+00 7.82673433e-02 -1.42678952e+00
-3.06717660e-02 -6.25069618e-01 4.61779803e-01 2.34851182e-01
3.46979983e-02 3.82185280e-01 3.47610146e-01 1.16672337e-01
-6.11709319e-02 3.96183163e-01 1.82855237e+00 -1.31282106e-01
-2.32299268e-01 2.00448394e-01 -2.55217820e-01 8.43433857e-01
5.44361949e-01 -4.68558520e-01 -4.06644881e-01 -8.21989000e-01
-1.53871803e-02 2.04548463e-01 8.67759049e-01 -4.37153339e-01
-1.52702600e-01 -8.05657431e-02 7.10420370e-01 -4.60152924e-01
6.91661000e-01 -1.13469255e+00 2.16306910e-01 1.08613916e-01
-3.85031044e-01 -1.23816535e-01 1.99456260e-01 5.37805915e-01
2.19801351e-01 -2.32570201e-01 8.74132037e-01 -3.97488505e-01
-1.81299314e-01 4.83577311e-01 1.86175436e-01 1.39539137e-01
8.13392282e-01 -4.92696553e-01 2.16879144e-01 -3.57087851e-01
-8.18498611e-01 -6.82343394e-02 1.23196101e+00 2.96138793e-01
8.96333396e-01 -1.18056417e+00 -7.34408975e-01 6.71624541e-01
-8.78346786e-02 8.52387607e-01 -1.48612544e-01 3.32969159e-01
-8.70703280e-01 -1.25313967e-01 -2.45862052e-01 -1.01878417e+00
-1.24815762e+00 7.99721360e-01 7.36347198e-01 1.59875363e-01
-1.14133000e+00 5.73629439e-01 6.06202245e-01 -9.49026525e-01
3.09540331e-01 -5.20109355e-01 6.36944294e-01 -6.29248857e-01
4.48637486e-01 -4.27459441e-02 -1.73235219e-03 -4.81982380e-01
-8.99602100e-03 9.90993321e-01 1.43385351e-01 -3.52565467e-01
1.26936305e+00 -6.58959746e-02 4.35359254e-02 5.27819574e-01
1.48795998e+00 7.42666097e-03 -2.08425617e+00 -4.38922435e-01
-5.05892694e-01 -8.96144867e-01 -1.23861404e-02 -6.14214599e-01
-1.03938949e+00 1.13382983e+00 4.39617097e-01 -6.77426830e-02
8.93918753e-01 2.15582222e-01 4.70460802e-01 2.96928883e-01
5.51253259e-01 -5.56584120e-01 3.92536283e-01 4.18971330e-01
1.17648315e+00 -1.30933583e+00 -1.20037831e-01 -5.08079588e-01
-4.46126819e-01 1.23813903e+00 3.67079645e-01 -4.59100872e-01
7.26763964e-01 4.20085758e-01 3.45931053e-01 -8.19242373e-02
-2.42094904e-01 -6.75051287e-02 5.00447989e-01 9.91992235e-01
1.10010719e-02 -2.06134245e-01 6.96076632e-01 -2.04052463e-01
-1.03788733e-01 -2.33493760e-01 3.44508559e-01 7.03762233e-01
2.18621492e-02 -1.15300179e+00 -5.41419923e-01 6.37814179e-02
-2.77693212e-01 1.91244751e-01 -4.08798277e-01 5.85570991e-01
8.71975124e-02 4.32165712e-01 2.79184461e-01 1.76103488e-01
5.79304338e-01 4.44572698e-03 1.05522418e+00 -9.15918946e-01
-1.27645135e-01 1.08849734e-01 -3.47228646e-01 -6.55972004e-01
-4.74499464e-01 -4.80999321e-01 -1.19373691e+00 7.29186013e-02
-1.51329949e-01 -3.21592838e-01 9.57830071e-01 7.40441144e-01
1.64832652e-01 2.75256246e-01 7.76871979e-01 -1.65637517e+00
-3.95587325e-01 -4.37559724e-01 -5.33019543e-01 7.46466815e-01
7.28779137e-01 -5.71459651e-01 -4.44403768e-01 4.81657535e-01] | [8.754859924316406, -3.0932705402374268] |
37f3cec9-cba8-40a2-bed6-019204802623 | causality-crossing-and-analyticity-in | 2102.02433 | null | https://arxiv.org/abs/2102.02433v1 | https://arxiv.org/pdf/2102.02433v1.pdf | Causality, Crossing and Analyticity in Conformal Field Theories | Analyticity and crossing properties of four point function are investigated in conformal field theories in the frameworks of Wightman axioms. A Hermitian scalar conformal field, satisfying the Wightman axioms, is considered. The crucial role of microcausality in deriving analyticity domains is discussed and domains of analyticity are presented. A pair of permuted Wightman functions are envisaged. The crossing property is derived by appealing to the technique of analytic completion for the pair of permuted Wightman functions. The operator product expansion of a pair of scalar fields is studied and analyticity property of the matrix elements of composite fields, appearing in the operator product expansion, is investigated. An integral representation is presented for the commutator of composite fields where microcausality is a key ingredient. Three fundamental theorems of axiomatic local field theories; namely, PCT theorem, the theorem proving equivalence between PCT theorem and weak local commutativity and the edge-of-the-wedge theorem are invoked to derive a conformal bootstrap equation rigorously. | ['Jnanadeva Maharana'] | 2021-02-04 | null | null | null | null | ['automated-theorem-proving', 'automated-theorem-proving'] | ['miscellaneous', 'reasoning'] | [ 1.39022887e-01 4.49632496e-01 3.47074509e-01 -2.91171908e-01
-4.55035940e-02 -5.72824717e-01 1.08900690e+00 -4.73106056e-01
-1.42412439e-01 6.38439834e-01 1.19122267e-01 -4.67754304e-01
-1.01344597e+00 -7.66508341e-01 -2.89208502e-01 -1.10563076e+00
-3.97921681e-01 2.03222215e-01 1.56403184e-01 -9.53036845e-01
3.06689650e-01 3.84510309e-01 -1.25384545e+00 -8.67466703e-02
9.46158886e-01 8.28338325e-01 -1.21252917e-01 8.15052032e-01
-2.61461101e-02 1.11104620e+00 1.41594380e-01 -2.90480047e-01
7.17884183e-01 -6.97114646e-01 -1.24620759e+00 -2.32090175e-01
3.42670679e-01 1.00376882e-01 -8.65217224e-02 1.27656627e+00
-1.27459377e-01 1.77461997e-01 1.06260645e+00 -1.44963729e+00
-1.13251460e+00 3.01473618e-01 -2.51589984e-01 1.85752094e-01
9.56731215e-02 -8.86094809e-01 1.23318565e+00 -1.11235297e+00
4.79939729e-01 1.12498558e+00 5.97294629e-01 3.57483029e-01
-1.42115057e+00 -1.14510275e-01 -7.11520553e-01 2.90115178e-01
-1.47110462e+00 -1.65045738e-01 5.52779257e-01 -6.58207715e-01
3.49357724e-01 6.49244905e-01 4.98322874e-01 1.83385834e-01
1.09425306e+00 -2.63457894e-01 1.03872943e+00 -1.03768027e+00
2.24374980e-01 2.44728819e-01 3.41510713e-01 8.07110846e-01
2.22858101e-01 1.12317495e-01 -2.07358465e-01 -4.24562454e-01
9.64483261e-01 -1.50540084e-01 2.04629395e-02 -8.64109159e-01
-1.31336093e+00 1.05972028e+00 -2.29123272e-02 7.81088293e-01
-1.83191881e-01 -7.67455250e-02 4.91882637e-02 3.15479308e-01
1.61147356e-01 2.20227197e-01 3.94301210e-03 3.83001953e-01
-1.52118132e-01 4.29665387e-01 7.02520132e-01 1.33764172e+00
6.23892367e-01 -3.37099165e-01 2.69400626e-01 1.52654782e-01
2.07899123e-01 8.18803430e-01 -1.23348065e-01 -1.44175708e+00
-4.00892571e-02 2.30414867e-01 1.60189345e-01 -7.57773221e-01
-5.35068989e-01 -1.05887741e-01 -8.24027956e-01 3.17054421e-01
4.26571697e-01 2.77111620e-01 2.11933866e-01 1.68063962e+00
3.24772090e-01 -3.57195169e-01 2.58580267e-01 5.09948552e-01
9.63516533e-02 3.68117988e-01 -1.22758135e-01 -4.50473756e-01
1.28791606e+00 -1.65831640e-01 -8.31743538e-01 7.71171391e-01
8.93425643e-01 -7.27496743e-01 5.58906674e-01 1.83664307e-01
-1.44163001e+00 -1.70553476e-01 -9.87057865e-01 -1.60369873e-01
-1.82306498e-01 -4.04149383e-01 6.91799283e-01 2.12390915e-01
-1.20473349e+00 9.12199318e-01 -2.95442224e-01 -5.22272468e-01
-2.82660991e-01 2.72509187e-01 -7.80979276e-01 5.01024008e-01
-1.18963087e+00 1.31861353e+00 -9.46169198e-02 -1.04582414e-01
-1.09589624e-03 -9.78350461e-01 -5.05301535e-01 1.77092925e-02
-4.81535137e-01 -5.69582880e-01 1.32798314e+00 -1.08956134e+00
-9.03621852e-01 6.28498912e-01 1.84847996e-01 -2.63030678e-01
2.23398700e-01 2.54379570e-01 -8.30730736e-01 4.89394218e-01
5.35342813e-01 -1.97627265e-02 5.35104513e-01 -6.63445830e-01
-5.93995929e-01 -2.26398334e-01 2.73721993e-01 -3.61953638e-02
-1.17902763e-01 9.15764496e-02 1.07256353e+00 -1.92558259e-01
4.09200996e-01 -6.66478276e-01 3.45272809e-01 -1.64078489e-01
2.80810893e-01 -4.43478286e-01 7.33003914e-01 -3.01682502e-01
8.90435696e-01 -2.31987023e+00 1.60816833e-01 4.98544753e-01
4.37082082e-01 -7.04212964e-01 8.27117562e-02 8.30194533e-01
-8.90785098e-01 -1.07238002e-01 -2.27388993e-01 6.51556611e-01
4.25171047e-01 -8.70453864e-02 -4.50103313e-01 1.23207688e+00
2.95695543e-01 4.29360420e-01 -9.26230550e-01 -6.72608733e-01
1.01098381e-01 5.17546594e-01 -7.10635185e-01 -4.39458668e-01
1.61410615e-01 3.25799793e-01 -4.58399147e-01 -1.03820905e-01
1.16672087e+00 -2.14455687e-02 1.33771688e-01 -3.69379491e-01
-7.85435498e-01 1.52115151e-01 -1.04326940e+00 1.01257682e+00
-1.38895944e-01 2.21495390e-01 1.69176742e-01 -1.12798035e+00
6.66231632e-01 4.59520727e-01 7.04713404e-01 -1.17741573e+00
1.73991472e-01 2.45481521e-01 3.11008573e-01 -3.37385029e-01
3.12273920e-01 -8.23333740e-01 -2.59754598e-01 4.00527358e-01
1.45236671e-01 4.81556505e-02 3.72136205e-01 6.60965621e-01
1.16329694e+00 2.64207304e-01 4.50486422e-01 -1.77981436e+00
1.04015458e+00 -1.68366551e-01 2.54604638e-01 1.93809435e-01
-4.58489247e-02 1.87326550e-01 9.54528153e-01 -5.13192058e-01
-1.54094315e+00 -1.53624570e+00 -9.99679029e-01 1.03835642e+00
3.41689616e-01 -4.81419921e-01 -8.14393878e-01 2.07847562e-02
-2.58981228e-01 1.04685116e+00 -1.11440027e+00 -2.17092574e-01
-3.57410520e-01 -4.10360903e-01 1.15927510e-01 1.27085537e-01
4.69179302e-01 -4.95496452e-01 -7.57883430e-01 -2.17676803e-01
5.74082732e-02 -7.64350355e-01 -2.95611322e-01 3.18408050e-02
-7.40912259e-01 -1.15298235e+00 -3.96935731e-01 -6.96700215e-01
7.26674855e-01 1.09416641e-01 6.85769916e-01 -4.99139950e-02
-6.09968752e-02 7.08567321e-01 7.91237969e-03 -6.90470263e-02
-7.26342380e-01 -7.93367147e-01 5.29496074e-01 2.85692930e-01
3.49215031e-01 -7.58692503e-01 -4.05159682e-01 4.33834434e-01
-9.19097006e-01 1.61789402e-01 -1.27258720e-02 6.83736622e-01
-3.15733477e-02 2.72200536e-03 3.59290749e-01 -3.09413970e-01
1.55798137e-01 -3.83755893e-01 -7.03472137e-01 3.08197141e-01
-3.16736847e-01 6.65705383e-01 8.52184892e-01 -1.09953722e-02
-1.22675896e+00 -6.72850847e-01 4.28918004e-01 5.52788854e-01
4.57648009e-01 -9.27900597e-02 -3.67287025e-02 -5.26358366e-01
9.94308233e-01 -7.24289268e-02 -5.09774163e-02 -2.86786795e-01
3.44418585e-01 4.37063217e-01 7.29711711e-01 -8.80520761e-01
6.68395042e-01 7.73841023e-01 9.60368216e-01 -9.51181173e-01
-7.81017900e-01 -3.63995880e-01 -8.96795750e-01 -1.09663054e-01
9.66064334e-01 -4.06667858e-01 -8.79734337e-01 8.81807283e-02
-1.13880718e+00 1.92225307e-01 -5.25128722e-01 6.39172614e-01
-1.20464861e+00 5.28568923e-01 -2.21085519e-01 -1.07199430e+00
1.33294770e-02 -7.35210180e-01 6.99257374e-01 -2.59409040e-01
-1.66988060e-01 -1.43671262e+00 3.38834912e-01 -1.19316727e-01
3.95499349e-01 9.38670561e-02 1.42660773e+00 -4.02911276e-01
-2.69044846e-01 -1.12452671e-01 -1.91308528e-01 1.37513906e-01
-1.71368688e-01 -6.44215643e-02 -5.50195694e-01 1.04809694e-01
6.87434256e-01 6.06380522e-01 -1.48309872e-01 8.19232687e-02
3.91282320e-01 -2.40307644e-01 1.48367032e-01 3.65150601e-01
1.61145961e+00 8.24370906e-02 6.96317494e-01 3.37994456e-01
1.17418535e-01 5.63678324e-01 2.93961078e-01 1.41636550e-01
2.11826459e-01 5.40421665e-01 3.09206545e-02 3.85069042e-01
1.54572025e-01 2.03570604e-01 3.56221199e-01 1.27380490e+00
-4.91704971e-01 7.98389912e-01 -6.60080373e-01 4.29464102e-01
-1.68534994e+00 -1.19921184e+00 -9.53681707e-01 2.28029060e+00
6.25666082e-01 -2.21449316e-01 1.33706123e-01 1.17990874e-01
8.21908832e-01 -7.52674937e-01 3.21775615e-01 -6.44766033e-01
-3.28269362e-01 4.88984615e-01 7.91047513e-01 1.02635825e+00
-5.58855832e-01 2.55334139e-01 6.64960814e+00 4.84643072e-01
-4.38552499e-01 6.34238303e-01 -4.67038393e-01 5.16930819e-01
-6.78794622e-01 7.04384625e-01 -4.50194061e-01 4.81477022e-01
7.68862545e-01 -6.77208364e-01 3.16939861e-01 4.13340837e-01
1.34878054e-01 -2.80173182e-01 -1.07322848e+00 6.49921536e-01
-2.31224876e-02 -1.13943362e+00 -2.85745680e-01 4.02276218e-01
7.54225016e-01 -2.71809489e-01 -2.30286896e-01 -1.71020105e-01
2.88837962e-02 -6.45210028e-01 9.53685522e-01 9.21305180e-01
6.73180401e-01 -7.93332756e-01 4.85341609e-01 1.67214811e-01
-1.18618917e+00 -5.63294888e-02 -4.05796736e-01 -4.15763497e-01
3.13191377e-02 3.69763881e-01 -9.18669328e-02 7.26959229e-01
2.23357856e-01 3.35492492e-01 -4.73885804e-01 5.96343756e-01
4.66972232e-01 -3.49750966e-02 -2.81260580e-01 2.87969530e-01
2.06770897e-01 -1.28668344e+00 3.57145876e-01 5.48407435e-01
5.24267972e-01 4.68998551e-01 -8.14684749e-01 1.06380618e+00
6.13276839e-01 3.96256238e-01 -6.23424053e-01 -1.34768626e-02
-1.96829438e-01 1.35247874e+00 -5.62005520e-01 -2.13684092e-04
-6.38791382e-01 2.44558856e-01 -2.42132559e-01 3.25358897e-01
-7.57502317e-01 -6.12568259e-01 4.57471877e-01 6.12811387e-01
1.41427040e-01 -3.60310733e-01 -2.77695894e-01 -1.09722149e+00
8.35176110e-02 -2.54254103e-01 3.43705676e-02 -7.73677528e-01
-1.09077251e+00 1.91114530e-01 6.52651608e-01 -1.12629664e+00
-5.73607609e-02 -9.30756509e-01 -7.71081924e-01 1.23252809e+00
-6.35633528e-01 -9.51938033e-01 2.14797184e-01 9.39666748e-01
-4.79518056e-01 -2.78271824e-01 9.35650051e-01 1.18881024e-01
3.97848308e-01 6.53699273e-03 4.60851550e-01 -2.05553040e-01
3.34043235e-01 -1.34625268e+00 1.24625370e-01 6.80992186e-01
-4.59831238e-01 7.77790725e-01 8.40769529e-01 -5.72899997e-01
-1.50146651e+00 -3.84700119e-01 1.41967356e+00 -5.68526685e-01
1.35356486e+00 -5.26960671e-01 -6.91945195e-01 6.69289708e-01
4.24429655e-01 -1.65922586e-02 5.96311688e-01 -8.66102353e-02
-1.61307558e-01 -3.44986320e-01 -1.03442740e+00 6.56733990e-01
9.99698639e-01 -5.74493647e-01 -7.76292801e-01 7.40126252e-01
4.91951764e-01 4.11920488e-01 -1.26150215e+00 3.21716011e-01
6.05966806e-01 -1.10926926e+00 7.74254978e-01 -6.51909411e-01
7.32478201e-02 -3.91046345e-01 -4.41767365e-01 -5.56860089e-01
-8.55697393e-01 -9.60973024e-01 4.54672992e-01 7.21152484e-01
1.40249671e-03 -9.89747465e-01 -3.08999211e-01 4.61871594e-01
4.53813896e-02 -2.10796222e-01 -1.08422101e+00 -9.08442259e-01
8.83090198e-01 2.13592201e-01 1.67663500e-01 9.27362144e-01
1.05925965e+00 7.76626348e-01 -4.35560122e-02 -2.19896957e-01
1.00258422e+00 -4.55006033e-01 6.04583509e-02 -1.61608589e+00
-2.12089568e-02 -2.50697583e-01 -7.93849766e-01 -3.90220247e-02
1.10601634e-01 -1.25439477e+00 -5.07355332e-01 -1.14629984e+00
-5.05919615e-03 -4.72514659e-01 6.57932088e-02 -1.43166915e-01
8.62375915e-01 -1.44113168e-01 3.29033546e-02 4.34666052e-02
-3.16009104e-01 3.03172112e-01 1.14639199e+00 3.70355040e-01
4.56880063e-01 -3.15406680e-01 -6.10333323e-01 8.32900047e-01
6.35429621e-01 -1.34311453e-01 -6.55648053e-01 8.31016079e-02
1.07283318e+00 1.28219068e-01 7.78401554e-01 -9.24451172e-01
2.01290146e-01 -2.05900773e-01 1.66593455e-02 -2.03306302e-01
-2.86893547e-02 -6.35672808e-01 5.05363762e-01 6.04293607e-02
-2.08505273e-01 4.74011302e-01 -3.21742207e-01 8.83609205e-02
2.14425504e-01 -7.66236186e-01 1.09862339e+00 7.85439461e-02
-2.74597228e-01 -1.00710459e-01 -2.44601861e-01 2.50884682e-01
1.10299754e+00 1.89604908e-01 -2.33680621e-01 -5.00132963e-02
-5.56405842e-01 -3.20269704e-01 4.85570580e-01 -7.58335590e-02
7.17191622e-02 -1.91115153e+00 -4.87080157e-01 5.47853351e-01
-8.22513700e-02 -5.59713662e-01 3.13025743e-01 1.50814152e+00
-7.55736411e-01 5.81044734e-01 -4.22806203e-01 -5.46767235e-01
-6.68692887e-01 5.87999701e-01 5.94427645e-01 4.89392459e-01
-7.26065576e-01 1.36701405e-01 9.93054688e-01 -1.57812938e-01
-4.95850652e-01 4.93778475e-03 1.54375449e-01 -3.45471323e-01
5.65095007e-01 5.07883728e-01 4.96822111e-02 -1.15634692e+00
-3.43905091e-01 8.83199930e-01 3.63183886e-01 -4.79906976e-01
9.79120433e-01 -2.64981449e-01 -9.18585002e-01 5.22403836e-01
1.30712128e+00 2.14870632e-01 -5.46134114e-01 -5.75029850e-02
-1.01445958e-01 -1.15898401e-01 -2.09728763e-01 -1.98727414e-01
-1.65902302e-02 7.92741358e-01 1.92318633e-01 9.53780234e-01
7.48971045e-01 2.98992008e-01 2.56768793e-01 3.87522280e-02
3.13701183e-01 -1.10526609e+00 -7.27086782e-01 5.14533162e-01
1.17760634e+00 -5.49483061e-01 -2.15256989e-01 -4.87532198e-01
-2.36917138e-01 1.42379069e+00 -1.62864681e-02 -5.01522183e-01
1.01552808e+00 2.06661284e-01 -4.51657146e-01 -1.79452002e-01
-5.33862650e-01 5.74655598e-03 4.31998581e-01 3.46614897e-01
6.04357123e-01 1.75792396e-01 -9.17632937e-01 1.85020119e-01
-5.35015464e-01 -2.42139488e-01 7.98651218e-01 7.39702225e-01
-4.97097999e-01 -1.05673492e+00 -5.77392817e-01 -2.64126416e-02
-2.81553537e-01 -2.03795761e-01 -2.76812129e-02 9.68793631e-01
4.45504814e-01 5.86463690e-01 4.59207088e-01 8.36759657e-02
9.27382931e-02 3.85086268e-01 1.14868748e+00 -5.39618097e-02
2.19647750e-01 -5.03933504e-02 -3.10897470e-01 -8.80223513e-02
-7.29727626e-01 -8.31995845e-01 -1.44173682e+00 -7.00972915e-01
-1.69695884e-01 9.55099821e-01 5.51363826e-01 1.24967825e+00
-1.28351539e-01 2.98033599e-02 6.20574832e-01 -1.65380538e-01
-9.30602729e-01 -9.36750650e-01 -1.21853781e+00 1.87484011e-01
4.00920510e-01 -8.08786571e-01 -5.56563616e-01 1.03646427e-01] | [6.4196014404296875, 4.730812072753906] |
83857e67-70cb-4358-a471-8124389e43cf | optimal-alphabet-for-single-text-compression | 2201.05234 | null | https://arxiv.org/abs/2201.05234v2 | https://arxiv.org/pdf/2201.05234v2.pdf | Optimal alphabet for single text compression | A text written using symbols from a given alphabet can be compressed using the Huffman code, which minimizes the length of the encoded text. It is necessary, however, to employ a text-specific codebook, i.e. the symbol-codeword dictionary, to decode the original text. Thus, the compression performance should be evaluated by the full code length, i.e. the length of the encoded text plus the length of the codebook. We studied several alphabets for compressing texts -- letters, n-grams of letters, syllables, words, and phrases. If only sufficiently short texts are retained, an alphabet of letters or two-grams of letters is optimal. For the majority of Project Gutenberg texts, the best alphabet (the one that minimizes the full code length) is given by syllables or words, depending on the representation of the codebook. Letter 3 and 4-grams, having on average comparable length to syllables/words, perform noticeably worse than syllables or words. Word 2-grams also are never the best alphabet, on the account of having a very large codebook. We also show that the codebook representation is important -- switching from a naive representation to a compact one significantly improves the matters for alphabets with large number of symbols, most notably the words. Thus, meaning-expressing elements of the language (syllables or words) provide the best compression alphabet. | ['Andranik Khachatryan', 'Armen E. Allahverdyan'] | 2022-01-13 | null | null | null | null | ['text-compression'] | ['natural-language-processing'] | [ 5.23396969e-01 9.60810035e-02 -3.35446209e-01 1.15733482e-02
-1.21335030e-01 -5.99635303e-01 5.97093046e-01 4.00016725e-01
-6.12864912e-01 6.90666676e-01 2.02995539e-01 -7.26760745e-01
8.66926536e-02 -9.42286670e-01 -4.24495429e-01 -8.68239999e-01
-4.32583801e-02 4.45532918e-01 1.14049934e-01 -4.49734330e-01
5.72178125e-01 5.67589104e-01 -1.94291711e+00 6.64958805e-02
5.79594612e-01 7.81269372e-01 5.27151823e-01 7.83761501e-01
-7.35332727e-01 5.11513948e-01 -5.75276375e-01 -3.41718554e-01
2.67466217e-01 -9.48916852e-01 -5.98645091e-01 1.25867873e-01
-1.54432163e-01 -2.01023206e-01 -5.09970963e-01 1.03281772e+00
2.49116924e-02 1.05152093e-01 8.12930763e-01 -5.20524502e-01
-3.70085955e-01 1.01832640e+00 -1.76236987e-01 -5.62946871e-02
4.67361242e-01 -3.46524924e-01 9.63220894e-01 -6.61474407e-01
5.46021104e-01 1.09937203e+00 2.87071735e-01 3.34180295e-01
-7.45325208e-01 -3.25441986e-01 -3.95935446e-01 -7.47694820e-02
-1.81440854e+00 -6.17493987e-01 3.10900569e-01 -1.60182133e-01
1.43052244e+00 6.62524939e-01 7.25776255e-01 6.46819830e-01
6.74462140e-01 1.59208149e-01 7.59309053e-01 -1.26440001e+00
2.25402534e-01 2.71519423e-01 2.33503636e-02 5.31508327e-01
6.49192572e-01 1.04437592e-02 -1.42528668e-01 -9.30750817e-02
5.59360623e-01 -1.31049871e-01 -1.76627085e-01 2.78438665e-02
-9.84181762e-01 1.00054777e+00 -4.89398092e-01 7.91157842e-01
-1.64318278e-01 2.62363911e-01 6.26801431e-01 6.22924387e-01
-1.46180153e-01 1.74357817e-01 -3.61564592e-03 -4.62231725e-01
-1.13224244e+00 6.20581098e-02 1.13091147e+00 1.40367830e+00
5.59605837e-01 2.64457524e-01 4.00483683e-02 7.88017690e-01
7.40133449e-02 7.66411722e-01 5.08351564e-01 -5.22440553e-01
4.42100286e-01 1.17464885e-01 -1.57380030e-01 -8.46540511e-01
-1.66525692e-01 -3.95903438e-01 -1.13853061e+00 -2.44023100e-01
2.25276276e-01 1.29462972e-01 -8.59044790e-01 1.40544546e+00
-3.66623729e-01 -6.37143195e-01 3.33583087e-01 3.81535351e-01
4.02786195e-01 9.74319756e-01 -2.96068460e-01 -8.01279485e-01
1.44757295e+00 -5.68486989e-01 -1.02393174e+00 -4.70026210e-02
9.63749468e-01 -9.31507349e-01 7.80932844e-01 4.32022303e-01
-1.28967941e+00 -4.22903121e-01 -1.17879832e+00 9.36834663e-02
-5.51303983e-01 -1.24168120e-01 2.93287426e-01 9.09709692e-01
-1.01394284e+00 6.10175848e-01 -3.65533799e-01 -3.81727546e-01
-2.49624357e-01 3.59171808e-01 -2.96758056e-01 1.85375381e-02
-1.22774565e+00 1.00197792e+00 9.39440489e-01 -4.69930977e-01
-3.83122146e-01 1.58742666e-01 -8.32953870e-01 3.51310462e-01
6.65213093e-02 -2.46331185e-01 9.02900457e-01 -7.53102779e-01
-1.24837315e+00 7.73674309e-01 -3.38601589e-01 -6.61003113e-01
1.23561777e-01 5.46108127e-01 -6.21441424e-01 4.04851317e-01
-2.44131714e-01 3.77678573e-01 1.02022839e+00 -8.57909083e-01
-5.84300041e-01 1.37856424e-01 -1.87459484e-01 -7.12688044e-02
-4.27255601e-01 2.58860171e-01 -4.61963713e-01 -8.23197365e-01
2.35821426e-01 -9.47904944e-01 -7.80347958e-02 -3.89127374e-01
-1.60997614e-01 -2.43451476e-01 6.51531219e-01 -5.52138269e-01
1.91505480e+00 -2.65331149e+00 4.83089760e-02 6.39813244e-01
-1.63415968e-01 1.01985909e-01 5.87081909e-02 8.69738638e-01
-1.32810012e-01 2.24397376e-01 -3.96747291e-01 -2.53087848e-01
-7.08700567e-02 9.93858993e-01 -5.57654083e-01 4.44870651e-01
-2.24001989e-01 5.98267496e-01 -6.21892571e-01 -6.62380636e-01
1.54666021e-01 2.29202822e-01 -4.22625244e-01 -2.73783267e-01
-9.45843309e-02 -1.42487302e-01 -2.82756805e-01 4.07470107e-01
5.19454718e-01 -8.51785392e-03 1.29442662e-01 3.56609881e-01
-3.73640269e-01 3.99747103e-01 -1.17017746e+00 1.22803891e+00
-4.72414792e-01 8.33888888e-01 -1.63988352e-01 -1.02399755e+00
1.28871655e+00 5.61828494e-01 3.39587778e-01 -6.57049119e-01
2.40053892e-01 6.76846921e-01 2.75479496e-01 -2.90573388e-01
8.93614292e-01 -3.12983423e-01 -1.78567261e-01 4.41435099e-01
-5.89969754e-02 -4.76574063e-01 9.10078108e-01 2.78097004e-01
9.52162087e-01 -5.69063008e-01 7.66176224e-01 -3.94217402e-01
5.90318382e-01 -2.91581899e-01 1.03684686e-01 8.73295307e-01
3.99050564e-01 7.17863381e-01 6.01897240e-01 -8.00164193e-02
-1.65531456e+00 -3.95793200e-01 -6.89379036e-01 9.29902434e-01
-1.76002458e-01 -8.83715987e-01 -6.24997020e-01 -2.61785295e-02
-4.96523976e-02 1.02535450e+00 -2.08705664e-01 -5.72363250e-02
-6.27932906e-01 -3.79835367e-01 7.27801621e-01 1.77302703e-01
1.17428057e-01 -1.21527541e+00 -7.75737166e-01 4.99932528e-01
1.55228926e-02 -1.01210284e+00 -3.53287935e-01 8.12113643e-01
-8.42468381e-01 -4.96784329e-01 -7.39411354e-01 -6.36346340e-01
7.91989207e-01 1.87075064e-01 8.00580025e-01 4.36067462e-01
1.22952886e-01 1.04486436e-01 -1.04373586e+00 -2.90937871e-01
-1.00722229e+00 -8.10531899e-02 1.54316470e-01 -3.21129829e-01
1.86222598e-01 -5.83099067e-01 2.13239416e-02 -2.18877625e-02
-1.26089323e+00 -1.06804289e-01 5.55986047e-01 6.67750359e-01
4.78265047e-01 3.36176693e-01 4.62304540e-02 -6.85959280e-01
5.32123148e-01 -1.42644808e-01 -3.68517190e-01 9.63285416e-02
-5.41017890e-01 2.67203093e-01 1.25042725e+00 -2.77555317e-01
-2.24932879e-01 -2.02509284e-01 -4.24251407e-01 1.47972787e-02
-1.16142467e-01 5.08109033e-01 5.34312911e-02 -2.17073083e-01
5.53176939e-01 7.71499097e-01 -1.72480062e-01 -4.09985840e-01
3.55711401e-01 1.15245271e+00 4.93724704e-01 -4.50340152e-01
4.46362197e-01 1.08746655e-01 2.63165504e-01 -1.35698175e+00
-1.31223843e-01 -5.16128719e-01 -6.73973203e-01 1.73053339e-01
5.65801740e-01 -6.59910321e-01 6.09587925e-03 2.82668620e-02
-1.28381264e+00 2.55120188e-01 -8.15267146e-01 5.49964905e-01
-7.48628139e-01 9.32414889e-01 -4.54846799e-01 -8.52535367e-01
-2.21681952e-01 -9.91367459e-01 9.06740904e-01 -2.56129175e-01
-4.86558855e-01 -7.85948157e-01 -2.08573729e-01 -4.83109623e-01
2.50090748e-01 -4.97455243e-04 1.41660798e+00 -7.45063722e-01
-6.89225048e-02 -4.40739125e-01 1.36083990e-01 4.70740676e-01
1.25179917e-01 1.06816649e-01 -3.22368473e-01 -4.25429761e-01
8.68945792e-02 5.62120900e-02 9.24126983e-01 1.88631549e-01
1.19416058e+00 -5.24989307e-01 -2.58619249e-01 6.40265167e-01
1.39188123e+00 7.20256507e-01 9.23038483e-01 2.29519352e-01
2.46978045e-01 2.79432267e-01 3.44858617e-01 8.50846410e-01
-1.86102018e-01 5.48157275e-01 7.05494806e-02 4.42832530e-01
-2.36432720e-02 -1.54331475e-01 4.01828915e-01 1.70113564e+00
-3.04050725e-02 -7.09920168e-01 -8.72010827e-01 4.86351162e-01
-1.24950254e+00 -1.08895218e+00 -3.36138189e-01 2.31189132e+00
1.06411147e+00 5.20909309e-01 -1.38691843e-01 9.21457589e-01
4.39879000e-01 2.45817348e-01 4.16394286e-02 -1.19944930e+00
-4.99959350e-01 5.26118934e-01 7.90536702e-01 6.77413225e-01
-7.05112159e-01 7.06824839e-01 7.66099930e+00 1.33541322e+00
-8.80991936e-01 -1.20052963e-01 -1.14630465e-03 1.29867926e-01
-5.01861155e-01 2.46587589e-01 -1.04979253e+00 9.09195065e-01
1.50361180e+00 -4.90598351e-01 5.76579988e-01 6.55355334e-01
1.06226556e-01 -1.80927277e-01 -9.18056250e-01 1.03211927e+00
2.62108922e-01 -1.20979476e+00 5.98449051e-01 2.75409967e-01
6.83551669e-01 -4.29036587e-01 -2.26445556e-01 2.71174908e-01
-2.42265135e-01 -1.11631429e+00 9.78802800e-01 4.67430174e-01
1.27673602e+00 -8.46110344e-01 7.06198752e-01 6.26956880e-01
-1.21412075e+00 2.18808278e-02 -9.36529338e-01 -1.46532774e-01
1.52813047e-01 6.11218572e-01 -2.69670784e-01 4.25665200e-01
7.82169774e-02 3.98461789e-01 -3.44466448e-01 8.20808053e-01
3.51711124e-01 5.70694864e-01 -5.83167911e-01 -3.06607008e-01
6.28174961e-01 -4.55902547e-01 5.65906703e-01 1.84308565e+00
7.55523682e-01 5.79463243e-01 5.94756864e-02 2.42882654e-01
1.96142361e-01 3.13897431e-01 -9.35992002e-01 -4.59073961e-01
6.54391229e-01 5.38372338e-01 -8.46412480e-01 -6.89391136e-01
-5.29252708e-01 6.58519089e-01 -3.65083426e-01 7.04696104e-02
-4.17307019e-01 -9.61847365e-01 1.43274441e-01 1.35134265e-01
5.84650695e-01 -5.13678551e-01 -3.94598752e-01 -6.66884542e-01
-7.45068938e-02 -1.04278874e+00 1.17971282e-02 -3.96007270e-01
-6.00619376e-01 7.06011057e-01 2.39965305e-01 -1.47006392e+00
-7.29540586e-01 -5.30338824e-01 -1.34504154e-01 8.87494445e-01
-9.75510955e-01 -3.64113331e-01 2.46982247e-01 4.48470592e-01
4.68697071e-01 -5.31129360e-01 1.06245184e+00 1.19356848e-01
-1.13588922e-01 6.93398297e-01 8.45346749e-01 -1.45028576e-01
1.66606024e-01 -9.67347264e-01 2.80841023e-01 6.13379538e-01
2.41506189e-01 6.99627817e-01 9.56724465e-01 -5.53000391e-01
-1.53637016e+00 -6.24346077e-01 1.42869782e+00 9.12495181e-02
4.80587274e-01 -3.79395455e-01 -9.72697437e-01 4.79910702e-01
2.14130297e-01 -5.56611300e-01 7.64602125e-01 -4.15782601e-01
-9.73563269e-02 2.33377725e-01 -8.53237271e-01 6.31132603e-01
9.30419266e-01 -6.40142739e-01 -7.67376721e-01 4.89718288e-01
8.54889393e-01 -3.23411942e-01 -6.66997612e-01 -1.44748231e-02
5.21734893e-01 -7.76230812e-01 6.10869348e-01 -1.45886809e-01
4.02526975e-01 -1.60425022e-01 -6.07759476e-01 -9.19616282e-01
-4.53636199e-01 -6.39206588e-01 -2.39579886e-01 7.90817499e-01
2.38589585e-01 -5.27759314e-01 4.34986055e-01 -3.65595259e-02
-1.39343366e-01 -6.00979269e-01 -1.10847831e+00 -1.17542136e+00
1.88690692e-01 -5.33211052e-01 6.80822611e-01 6.82941258e-01
3.40778649e-01 1.35179207e-01 -5.16442537e-01 -6.19507790e-01
3.53560746e-01 1.69819593e-02 3.65831941e-01 -1.14149916e+00
-2.67067283e-01 -5.99297464e-01 -4.20088589e-01 -1.41552210e+00
6.48380071e-02 -1.15941072e+00 -8.69130045e-02 -1.00584221e+00
-1.76186971e-02 -4.56973046e-01 1.95645496e-01 3.82910490e-01
5.83653450e-01 -1.31753115e-02 3.68132353e-01 3.44361991e-01
-1.30013883e-01 3.41642529e-01 1.19177532e+00 -6.15779534e-02
-1.03353530e-01 -1.80738851e-01 -4.49088722e-01 4.21147138e-01
6.80928349e-01 -5.92287421e-01 -2.89829046e-01 -1.48212671e-01
4.37864304e-01 3.67579013e-01 -2.19176978e-01 -9.56813037e-01
3.13908339e-01 -1.59339279e-01 -5.07945418e-02 -9.09279346e-01
2.16925263e-01 -1.11574841e+00 4.85064209e-01 5.90066552e-01
-3.76162589e-01 1.43861040e-01 4.33414429e-02 3.06499571e-01
-2.93482482e-01 -1.32306564e+00 6.37130499e-01 -2.29523197e-01
-4.51704323e-01 -1.22065820e-01 -9.00105178e-01 -2.42318884e-01
7.42927551e-01 -8.02540064e-01 1.03067912e-01 -4.44442749e-01
-5.88055074e-01 -3.04928958e-01 6.12254500e-01 1.13595486e-01
6.66792512e-01 -1.32247853e+00 -7.37615287e-01 7.00597703e-01
6.45468235e-02 -4.71415490e-01 -3.54662985e-01 5.30679882e-01
-1.02549827e+00 1.03780854e+00 -1.60423011e-01 -3.34179074e-01
-1.18448949e+00 5.82506239e-01 -9.13647041e-02 -9.72175375e-02
-8.80311966e-01 3.41560572e-01 -2.67797530e-01 3.46885443e-01
3.78789753e-01 -5.57017207e-01 -1.50856450e-01 9.78454351e-02
7.43321538e-01 2.24011153e-01 2.28442222e-01 -9.38710451e-01
-9.87776741e-02 7.87780225e-01 -2.63053793e-02 -1.79612651e-01
8.01711798e-01 -1.94841862e-01 -5.52493453e-01 4.54634905e-01
1.33332062e+00 1.71739563e-01 -3.14208984e-01 -1.69602204e-02
-1.91675633e-01 -5.85000157e-01 -9.24020708e-02 -9.90789756e-02
-4.82218593e-01 7.89514661e-01 8.26200917e-02 6.48146808e-01
1.14123976e+00 -2.96135426e-01 7.87001371e-01 6.29504740e-01
8.11462522e-01 -1.03676343e+00 -4.70599353e-01 1.05871272e+00
8.18123102e-01 -4.90104347e-01 1.61183909e-01 -2.51519799e-01
-3.96142185e-01 1.51306164e+00 -3.65777373e-01 -4.43324633e-02
6.64101064e-01 5.09290993e-01 -5.19257009e-01 2.50224054e-01
-6.80304289e-01 -3.11765403e-01 1.53067365e-01 3.05114239e-01
4.50949878e-01 2.63867646e-01 -1.02415097e+00 4.25455809e-01
-9.01431680e-01 -3.62929285e-01 7.49723732e-01 1.04532409e+00
-1.03361011e+00 -1.34089899e+00 -7.29235053e-01 7.48300076e-01
-3.72836947e-01 -4.13641781e-01 -2.89080322e-01 6.86744332e-01
3.18574846e-01 1.08882928e+00 3.60347450e-01 -3.58776450e-01
1.81263402e-01 3.49561721e-01 5.24289370e-01 -6.20191097e-01
-5.47511756e-01 1.44776702e-01 8.43006361e-04 -5.89160323e-02
-2.01760218e-01 -5.95790982e-01 -1.45190716e+00 -1.03472555e+00
-4.73827958e-01 4.56378132e-01 5.67882836e-01 9.11650002e-01
-2.00001881e-01 1.59708858e-01 6.05837822e-01 -4.20933604e-01
-5.78042388e-01 -7.51368523e-01 -1.17549396e+00 8.10615811e-03
3.26299071e-01 -1.85991496e-01 -4.90147829e-01 1.81237087e-01] | [12.156885147094727, 9.433152198791504] |
d40eadf9-bcca-4e20-b7bb-9214ce5c9e37 | lipstick-ain-t-enough-beyond-color-matching | 2104.01867 | null | https://arxiv.org/abs/2104.01867v1 | https://arxiv.org/pdf/2104.01867v1.pdf | Lipstick ain't enough: Beyond Color Matching for In-the-Wild Makeup Transfer | Makeup transfer is the task of applying on a source face the makeup style from a reference image. Real-life makeups are diverse and wild, which cover not only color-changing but also patterns, such as stickers, blushes, and jewelries. However, existing works overlooked the latter components and confined makeup transfer to color manipulation, focusing only on light makeup styles. In this work, we propose a holistic makeup transfer framework that can handle all the mentioned makeup components. It consists of an improved color transfer branch and a novel pattern transfer branch to learn all makeup properties, including color, shape, texture, and location. To train and evaluate such a system, we also introduce new makeup datasets for real and synthetic extreme makeup. Experimental results show that our framework achieves the state of the art performance on both light and extreme makeup styles. Code is available at https://github.com/VinAIResearch/CPM. | ['Minh Hoai', 'Anh Tran', 'Thao Nguyen'] | 2021-04-05 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Nguyen_Lipstick_Aint_Enough_Beyond_Color_Matching_for_In-the-Wild_Makeup_Transfer_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Nguyen_Lipstick_Aint_Enough_Beyond_Color_Matching_for_In-the-Wild_Makeup_Transfer_CVPR_2021_paper.pdf | cvpr-2021-1 | ['color-manipulation', 'facial-makeup-transfer'] | ['computer-vision', 'computer-vision'] | [ 1.71168789e-01 -6.08262777e-01 1.13031492e-01 -5.21118104e-01
-2.69757509e-01 -9.03086722e-01 4.37548280e-01 -5.81022859e-01
3.14160645e-01 5.33546865e-01 -2.60649085e-01 2.42362991e-02
2.87740529e-01 -1.07141864e+00 -9.88854110e-01 -5.33846974e-01
4.13997293e-01 3.17675650e-01 2.35939905e-01 -3.84262323e-01
3.77611071e-01 6.17292345e-01 -1.56266499e+00 5.94580114e-01
9.44852471e-01 9.52383339e-01 6.99226186e-02 5.89861989e-01
-4.40258563e-01 6.65163875e-01 -4.10463661e-01 -7.46724904e-01
5.05989850e-01 -4.67387706e-01 -2.72787035e-01 1.13922805e-01
9.65976775e-01 -4.48644370e-01 -6.13624901e-02 1.14357448e+00
6.02320075e-01 -6.82454258e-02 8.07482958e-01 -1.46085644e+00
-1.51195753e+00 2.51679033e-01 -9.75287735e-01 -3.33866239e-01
2.88293958e-01 3.04255635e-01 5.85880518e-01 -1.15034866e+00
6.05451882e-01 1.27159250e+00 7.99172103e-01 8.81994963e-01
-9.82760012e-01 -1.37119317e+00 2.32771635e-01 1.69408351e-01
-1.19784451e+00 -3.65465283e-01 9.94695127e-01 -4.33551311e-01
3.41749936e-01 3.53534520e-01 7.05133557e-01 1.08177376e+00
-3.38588394e-02 8.58200729e-01 1.61986303e+00 -4.27378237e-01
9.51706246e-02 1.24165125e-01 -1.09103911e-01 8.46455812e-01
1.77192455e-03 2.91913450e-02 -1.67107671e-01 2.37635851e-01
9.28346217e-01 1.63001403e-01 -3.99534017e-01 -6.26986086e-01
-1.25699615e+00 2.53395349e-01 6.31003320e-01 1.08648881e-01
1.28959477e-01 3.41082752e-01 -2.24386044e-02 3.65063608e-01
5.50501883e-01 2.55445868e-01 -4.09842044e-01 -1.07880309e-01
-7.51005411e-01 1.39026240e-01 7.25257516e-01 1.09425604e+00
1.20963311e+00 -1.56494230e-01 -2.40707248e-02 1.21496594e+00
2.50744559e-02 1.04407549e+00 2.19123647e-01 -5.88613391e-01
5.48170745e-01 5.72633982e-01 1.39744371e-01 -9.41025853e-01
-2.16126129e-01 1.32107839e-01 -9.13878798e-01 5.05667567e-01
3.41986120e-01 -4.19880033e-01 -1.30971527e+00 1.55904996e+00
5.95865250e-01 3.46420825e-01 -5.12342393e-01 9.53541756e-01
9.93598759e-01 7.60786831e-01 -2.70419419e-01 4.17939484e-01
1.08578336e+00 -1.29406583e+00 -4.78613019e-01 -1.40671283e-01
7.49336705e-02 -1.15029383e+00 1.51798034e+00 4.35710400e-01
-9.77555752e-01 -7.20300794e-01 -1.00949168e+00 -2.30162591e-01
-5.41928291e-01 3.99903089e-01 4.91958082e-01 6.60511553e-01
-1.00768101e+00 6.19720221e-01 -4.19144213e-01 -4.80937362e-01
3.26402843e-01 -1.02695577e-01 -4.74150479e-01 -4.06865895e-01
-7.97513723e-01 3.46020073e-01 1.58053890e-01 1.46351740e-01
-7.44045973e-01 -1.12235928e+00 -6.19785726e-01 -3.40762854e-01
2.92682081e-01 -7.42938221e-01 9.96551991e-01 -1.38091409e+00
-1.73880577e+00 1.15688372e+00 1.71569079e-01 4.16735142e-01
7.79842675e-01 -3.31818610e-01 -4.70517129e-01 -2.34782398e-01
-9.36611965e-02 5.17940223e-01 1.16487765e+00 -1.62729228e+00
-3.65423530e-01 -1.09324269e-01 1.15112595e-01 8.36728215e-02
-8.47359076e-02 -7.38563240e-02 -6.38108432e-01 -9.27399755e-01
-1.30125254e-01 -1.01599443e+00 4.51277316e-01 4.27588016e-01
-7.01523840e-01 1.57309353e-01 7.30744064e-01 -5.33898890e-01
9.40448821e-01 -2.25835586e+00 1.32987186e-01 3.33859921e-01
2.99654864e-02 2.31868729e-01 -4.02534038e-01 8.12122226e-01
-2.48028010e-01 2.01152951e-01 -3.43909532e-01 -9.78512391e-02
2.55895972e-01 -2.35472689e-03 -2.71588176e-01 1.43186733e-01
1.16451204e-01 9.15574312e-01 -7.98619807e-01 -2.79261678e-01
3.13923597e-01 5.87595344e-01 -4.46836680e-01 1.72516778e-01
-2.20864639e-01 4.75252688e-01 -3.80398661e-01 9.74597931e-01
1.29285514e+00 -4.72794250e-02 -1.70408636e-01 -3.11506063e-01
-4.48284484e-02 -5.62482893e-01 -1.20701718e+00 1.75517797e+00
-6.50331438e-01 3.02032381e-01 1.50676116e-01 -1.91388369e-01
1.14087367e+00 -3.89225066e-01 2.10558295e-01 -5.55980384e-01
1.22129567e-01 2.90505558e-01 -1.73167259e-01 -4.71640408e-01
4.47016180e-01 -2.96063900e-01 2.91838020e-01 4.14214551e-01
-3.18399489e-01 -4.94054914e-01 1.99269012e-01 -4.38518412e-02
6.34975016e-01 5.77513576e-01 -1.86095893e-01 -1.91703647e-01
3.11216414e-01 -3.98361593e-01 4.37619120e-01 1.13823466e-01
1.48420548e-02 1.19122958e+00 3.39461982e-01 -4.39703196e-01
-1.03438544e+00 -1.29601145e+00 -1.01233006e-01 1.24860489e+00
5.66810489e-01 -2.01183796e-01 -9.32934642e-01 -6.55387282e-01
2.84378201e-01 4.15198624e-01 -1.11752975e+00 1.39574319e-01
-5.30286491e-01 -6.12760544e-01 2.43663281e-01 4.83427465e-01
9.44938719e-01 -1.22577965e+00 -1.61592677e-01 -3.52429241e-01
-5.80462068e-02 -9.14345503e-01 -9.71475840e-01 -4.72955853e-01
-2.65955925e-01 -1.32959259e+00 -1.18019700e+00 -6.20587647e-01
7.01444268e-01 2.62785792e-01 1.22288620e+00 2.55342156e-01
-3.71479511e-01 4.56520259e-01 -6.12074733e-01 -6.17826760e-01
-2.68800408e-01 -5.11727929e-02 -2.73332238e-01 7.23056138e-01
1.20245181e-01 -6.91451967e-01 -9.70756710e-01 5.63997447e-01
-7.94184804e-01 6.19543493e-01 7.38217473e-01 5.85047126e-01
7.31517732e-01 -5.20690322e-01 3.38942826e-01 -1.15962398e+00
3.43142003e-01 -3.78709644e-01 -2.54867435e-01 6.19073391e-01
-2.49806792e-01 -2.91365236e-01 5.77433586e-01 -4.97556239e-01
-1.30615842e+00 -2.14360669e-01 -1.61515951e-01 -5.91354012e-01
-2.85713077e-01 -3.13751042e-01 -1.44620076e-01 -3.32024485e-01
6.22456312e-01 3.23977798e-01 -1.38509899e-01 -7.57374465e-01
7.53996372e-01 4.80995715e-01 3.75182658e-01 -8.49979162e-01
1.20558369e+00 5.97202837e-01 -1.26499429e-01 -7.38331735e-01
-4.89889026e-01 -2.47601047e-02 -6.01336360e-01 -3.99157852e-01
6.37389243e-01 -6.37216270e-01 -3.33594173e-01 8.60785007e-01
-7.00611949e-01 -9.82860565e-01 -2.84384400e-01 -9.67135355e-02
-5.62663972e-01 2.77884901e-01 -6.22727573e-01 -3.21078092e-01
-3.56306612e-01 -7.30240166e-01 1.25284362e+00 3.76489311e-01
2.79618442e-01 -6.63663328e-01 4.03401613e-01 2.92971544e-03
5.57017446e-01 7.41979480e-01 8.54111552e-01 3.18349481e-01
-5.44586420e-01 2.99915392e-02 -7.24182963e-01 4.34687763e-01
4.50057596e-01 5.75102270e-01 -1.07002103e+00 -2.48133734e-01
-7.29807675e-01 -4.06793535e-01 9.26928699e-01 1.71583537e-02
1.19478059e+00 -1.38076833e-02 -2.41308272e-01 1.19180489e+00
1.69261873e+00 2.75117040e-01 6.87418163e-01 1.62296206e-01
1.14022088e+00 6.92436397e-01 5.34271717e-01 2.86148190e-01
3.06915820e-01 6.40722990e-01 4.61440295e-01 -3.53222370e-01
-7.39017189e-01 -5.41879356e-01 3.42310429e-01 7.05155194e-01
-2.64940381e-01 -1.86609719e-02 -8.07552218e-01 3.75142276e-01
-1.58750403e+00 -7.28507578e-01 -4.98214364e-02 1.93238199e+00
7.89768934e-01 -4.82986540e-01 5.50451800e-02 -7.84045458e-02
8.16056609e-01 8.91428441e-02 -7.42827058e-01 -4.87246335e-01
-1.40583605e-01 1.87274113e-01 2.67224729e-01 7.38771036e-02
-9.98377085e-01 8.88100088e-01 5.42356777e+00 9.34759974e-01
-1.51519775e+00 1.15892299e-01 7.35294163e-01 -2.13111043e-01
-5.55400312e-01 -4.38350201e-01 -3.15504700e-01 7.98366785e-01
6.02628896e-03 3.83464247e-02 8.77883255e-01 7.26934433e-01
-1.97586671e-01 1.13056295e-01 -1.03851330e+00 1.41901731e+00
3.04834247e-01 -1.12565267e+00 9.96820703e-02 -3.47503573e-01
1.31012690e+00 5.20808287e-02 3.88947785e-01 1.28110558e-01
2.49794096e-01 -8.39689434e-01 1.05450583e+00 5.75799823e-01
1.49651861e+00 -5.24946094e-01 2.25467965e-01 -1.11255653e-01
-1.35988545e+00 5.61996587e-02 -3.31441730e-01 1.26924917e-01
-2.01725751e-01 5.45281887e-01 -1.51595607e-01 7.51809776e-01
8.74730945e-01 1.02954924e+00 -7.36143112e-01 1.08875465e+00
-4.65342015e-01 2.74930775e-01 -2.62335896e-01 -2.30730828e-02
3.43477428e-02 -5.51531792e-01 5.69081083e-02 1.28227615e+00
6.68992102e-01 -9.40320864e-02 1.90417562e-02 9.03937817e-01
-3.95458341e-01 5.00014246e-01 -6.99054778e-01 2.96421468e-01
3.61800760e-01 1.45876729e+00 -7.00446129e-01 -1.93100512e-01
-2.72566944e-01 1.31368482e+00 3.10203582e-01 5.45677602e-01
-9.56864715e-01 -5.31705499e-01 8.16299260e-01 1.24929324e-01
1.61892816e-01 -3.22403833e-02 -2.41798624e-01 -1.42434561e+00
1.54966116e-01 -6.78038180e-01 7.00414926e-02 -1.22723401e+00
-1.60232484e+00 4.83734965e-01 -5.01362346e-02 -1.45747387e+00
3.58878821e-01 -8.72765839e-01 -9.46898282e-01 7.84774721e-01
-1.79562092e+00 -1.72100544e+00 -8.73352289e-01 6.69372857e-01
6.00858629e-01 6.93608224e-02 4.66606647e-01 4.65444654e-01
-7.45509505e-01 5.61753035e-01 2.11351797e-01 1.01276465e-01
1.12567472e+00 -1.45379138e+00 5.44467390e-01 6.45549178e-01
1.41680822e-01 3.52131963e-01 2.94009089e-01 -3.49119961e-01
-1.56737185e+00 -1.33045948e+00 -4.82392162e-02 -4.98602122e-01
4.47108477e-01 -4.03966546e-01 -7.63787746e-01 7.38110423e-01
2.22428694e-01 4.27819826e-02 4.80281740e-01 -4.59959060e-02
-6.95920825e-01 -4.32957739e-01 -1.09326828e+00 5.72544456e-01
1.31491876e+00 -1.71908900e-01 -2.80050814e-01 1.99203745e-01
1.24184720e-01 -5.75415492e-01 -5.65022230e-01 2.72521108e-01
8.36435556e-01 -1.28422666e+00 9.82406557e-01 -1.69782653e-01
7.49146700e-01 -5.27864456e-01 8.00996646e-02 -1.59710801e+00
-4.41708386e-01 -6.32474303e-01 9.19533819e-02 1.56704056e+00
2.39484325e-01 -6.22204542e-01 5.40542245e-01 3.77078950e-01
-1.68034911e-01 -7.72805572e-01 -4.21371013e-01 -6.42522395e-01
2.96862781e-01 -1.70691952e-01 1.10652184e+00 1.09102857e+00
-6.18667603e-01 8.33326206e-02 -6.47585690e-01 -1.88283056e-01
5.37342131e-01 6.99682117e-01 1.20017016e+00 -8.05093884e-01
-2.84426063e-01 -5.22036135e-01 -2.70672012e-02 -8.07493508e-01
-1.34178698e-01 -9.29222286e-01 -7.89901540e-02 -1.64510977e+00
2.22555533e-01 -9.25778925e-01 -2.91746497e-01 6.49920404e-01
-2.84292847e-01 9.57121789e-01 5.12981117e-01 1.11815920e-02
-2.69416630e-01 4.58782285e-01 1.70763862e+00 -3.37107956e-01
3.58642600e-02 -3.12355280e-01 -6.77015185e-01 8.39855611e-01
1.06548250e+00 9.27921906e-02 -5.79992354e-01 -7.80612946e-01
3.75166774e-01 -6.40148938e-01 2.71893471e-01 -1.06231689e+00
-3.46143872e-01 -4.48690236e-01 4.22955900e-01 -2.05262184e-01
2.63704300e-01 -7.62005806e-01 4.46576387e-01 2.79140294e-01
1.39071956e-01 5.41418232e-03 2.28666082e-01 3.57003927e-01
-2.75015980e-02 1.32474437e-01 1.10282624e+00 -1.19368888e-01
-1.21708870e+00 6.47594690e-01 5.45940816e-01 1.20613120e-01
1.36145461e+00 -3.23420703e-01 -6.37643158e-01 -1.88897654e-01
-4.60808188e-01 8.74440148e-02 9.62437570e-01 7.12561607e-01
7.66090989e-01 -1.59105897e+00 -9.58682537e-01 5.52818656e-01
4.43549305e-01 -2.83585191e-01 7.39946604e-01 5.85640073e-01
-1.15209997e+00 -2.72621483e-01 -7.35032439e-01 -2.71729469e-01
-1.16422355e+00 8.26918244e-01 5.16779065e-01 2.99617261e-01
-7.76053488e-01 9.75046635e-01 9.02509987e-01 -6.02453589e-01
-2.67870873e-01 -6.29262447e-01 3.13051343e-02 -9.55748484e-02
7.13485658e-01 4.11451966e-01 -1.76054612e-01 -5.02714336e-01
-1.17228344e-01 1.58912909e+00 2.36788809e-01 5.30214489e-01
9.41272318e-01 -8.78430009e-02 -2.60761648e-01 5.51886737e-01
1.08821917e+00 3.27907056e-01 -1.50385857e+00 -6.54449314e-02
-6.75892949e-01 -8.91981423e-01 -5.17204881e-01 -1.16758978e+00
-1.37963414e+00 1.08758688e+00 5.73089004e-01 1.87880933e-01
1.32895029e+00 -2.24450007e-01 8.53978336e-01 -2.59269904e-02
6.59035623e-01 -7.87752509e-01 3.38689052e-02 2.20638230e-01
1.32260013e+00 -1.13694441e+00 -3.90802681e-01 -7.85290480e-01
-5.57885826e-01 1.13202012e+00 8.97996724e-01 -3.41424376e-01
8.18448782e-01 2.26215497e-01 4.34238821e-01 4.29851003e-02
-4.90367502e-01 -1.50604367e-01 3.41713220e-01 7.61895120e-01
2.82023519e-01 3.31183404e-01 3.81799601e-03 2.29040250e-01
-2.53115147e-01 2.77352422e-01 2.94223487e-01 6.01656556e-01
-1.17116660e-01 -1.18193543e+00 -4.46137905e-01 4.64941710e-01
-1.46920040e-01 -1.12289354e-01 -5.55458426e-01 8.41851711e-01
5.88035166e-01 5.52730262e-01 1.09420225e-01 -6.10083520e-01
6.86514318e-01 -4.63686734e-02 7.37896740e-01 -5.82158685e-01
-3.95615369e-01 -4.52907115e-01 -5.44152334e-02 -8.51973414e-01
-2.45709434e-01 -2.53648728e-01 -6.96843863e-01 -7.69068539e-01
1.97445810e-01 -2.56680638e-01 5.52676558e-01 2.55955607e-01
3.56875032e-01 5.56405127e-01 8.48809838e-01 -1.14217532e+00
7.29563693e-03 -1.08433735e+00 -9.61461186e-01 8.89859855e-01
2.70374328e-01 -7.48734355e-01 1.19072005e-01 2.70909607e-01] | [11.504873275756836, -0.8360395431518555] |
39e50e16-0c9e-4737-a550-91049b83b558 | semhint-md-learning-from-noisy-semantic | 2303.18219 | null | https://arxiv.org/abs/2303.18219v1 | https://arxiv.org/pdf/2303.18219v1.pdf | SemHint-MD: Learning from Noisy Semantic Labels for Self-Supervised Monocular Depth Estimation | Without ground truth supervision, self-supervised depth estimation can be trapped in a local minimum due to the gradient-locality issue of the photometric loss. In this paper, we present a framework to enhance depth by leveraging semantic segmentation to guide the network to jump out of the local minimum. Prior works have proposed to share encoders between these two tasks or explicitly align them based on priors like the consistency between edges in the depth and segmentation maps. Yet, these methods usually require ground truth or high-quality pseudo labels, which may not be easily accessible in real-world applications. In contrast, we investigate self-supervised depth estimation along with a segmentation branch that is supervised with noisy labels provided by models pre-trained with limited data. We extend parameter sharing from the encoder to the decoder and study the influence of different numbers of shared decoder parameters on model performance. Also, we propose to use cross-task information to refine current depth and segmentation predictions to generate pseudo-depth and semantic labels for training. The advantages of the proposed method are demonstrated through extensive experiments on the KITTI benchmark and a downstream task for endoscopic tissue deformation tracking. | ['Michael C. Yip', 'Yuheng Zhi', 'Shan Lin'] | 2023-03-31 | null | null | null | null | ['monocular-depth-estimation'] | ['computer-vision'] | [ 5.00573516e-01 7.11707056e-01 -3.25719625e-01 -7.76557267e-01
-1.08259940e+00 -4.55765128e-01 3.35767597e-01 1.59544542e-01
-5.13030112e-01 5.64032316e-01 1.22330971e-01 -6.06012158e-02
2.26161376e-01 -6.01757824e-01 -8.69473040e-01 -7.45004475e-01
3.30657601e-01 4.12505627e-01 3.59435081e-01 1.20734841e-01
1.45125270e-01 -1.03324838e-01 -9.53801572e-01 4.04890835e-01
1.08912206e+00 9.57551003e-01 4.15765226e-01 4.69731539e-01
9.43814497e-03 5.55773079e-01 -1.49270833e-01 -3.01292241e-01
4.72309798e-01 -4.80930477e-01 -8.33987892e-01 2.53813297e-01
4.34069782e-01 -4.90792185e-01 -7.38369953e-03 1.29427004e+00
4.41255063e-01 -1.80130571e-01 4.58707631e-01 -8.15213919e-01
-1.44530937e-01 5.60394466e-01 -5.79250038e-01 -1.54896334e-01
2.64403899e-03 2.63450384e-01 8.51998091e-01 -4.19284701e-01
7.45481253e-01 9.56179321e-01 7.53867269e-01 7.58594513e-01
-1.28895342e+00 -2.95296609e-01 4.38072115e-01 -2.18025982e-01
-1.07513416e+00 -3.68876725e-01 9.28200901e-01 -5.23089051e-01
6.45643473e-01 -2.16618210e-01 5.13567328e-01 9.37872529e-01
2.49346972e-01 8.14876080e-01 1.07485700e+00 -3.55019182e-01
2.33755291e-01 3.23744625e-01 -8.68742019e-02 9.55133557e-01
1.47704557e-01 2.29347020e-01 -5.36972880e-01 1.33651569e-01
8.53563011e-01 -3.96679848e-01 -4.95986074e-01 -6.88331485e-01
-1.08393431e+00 7.67713249e-01 8.49145591e-01 1.42211005e-01
-2.99363941e-01 1.93263113e-01 3.99546057e-01 7.36834183e-02
6.52242959e-01 4.42086995e-01 -5.13017952e-01 9.66757685e-02
-1.10395515e+00 -2.04413831e-01 7.72398293e-01 8.27803671e-01
9.30514991e-01 -4.29186881e-01 -1.46830648e-01 4.81362492e-01
5.67891955e-01 -1.63215265e-01 5.53044021e-01 -9.12304044e-01
5.91078281e-01 5.71406066e-01 5.14781326e-02 -6.68650210e-01
-4.55109328e-01 -4.99498874e-01 -5.15780032e-01 2.76905984e-01
8.39965284e-01 -2.05262810e-01 -1.19123721e+00 1.88643694e+00
5.32686234e-01 2.66544759e-01 3.52555178e-02 1.22609925e+00
6.28435493e-01 2.33468786e-01 8.28408226e-02 6.22882731e-02
1.07617712e+00 -1.30263484e+00 -5.40436864e-01 -6.62082136e-01
1.05197644e+00 -4.44029719e-01 1.03984082e+00 3.59856963e-01
-1.19755650e+00 -5.30783415e-01 -1.15125477e+00 -3.02383095e-01
2.06802469e-02 2.57154197e-01 4.39275652e-01 5.55331290e-01
-1.19510531e+00 9.17227864e-01 -1.38014889e+00 -9.90846157e-02
5.43045223e-01 4.67104524e-01 -6.16087951e-02 3.81089188e-02
-1.05396473e+00 7.90797055e-01 4.29470360e-01 3.64270926e-01
-8.26789200e-01 -7.62787700e-01 -1.00781429e+00 -3.17950755e-01
3.03837657e-01 -9.84534085e-01 1.21826088e+00 -1.38338923e+00
-1.63970327e+00 1.09200776e+00 -6.19096756e-02 -3.82631540e-01
8.57552409e-01 -1.27292007e-01 4.67250615e-01 1.92419484e-01
1.01150930e-01 1.23240411e+00 5.98921597e-01 -1.16659737e+00
-4.98890281e-01 -5.99319637e-01 2.57340491e-01 4.83447433e-01
-2.57161707e-01 -6.01814151e-01 -6.99583530e-01 -4.72207367e-01
4.66389000e-01 -1.12207854e+00 -5.79268456e-01 4.38741386e-01
-6.30295813e-01 1.99007586e-01 4.13218617e-01 -6.72734380e-01
7.73561060e-01 -2.04878950e+00 3.06594342e-01 2.38080695e-01
5.56399897e-02 -2.01072589e-01 5.78207225e-02 -2.57577509e-01
2.28543729e-01 -8.91322196e-02 -4.77057308e-01 -8.43550384e-01
-2.04393536e-01 3.14750433e-01 9.85027105e-02 6.90513730e-01
8.73301178e-02 8.21154773e-01 -9.86883223e-01 -6.46995068e-01
1.97216123e-01 5.27918458e-01 -9.15394783e-01 2.93677419e-01
-4.20038223e-01 1.11592245e+00 -5.08186042e-01 4.01716083e-01
5.15985906e-01 -5.42936385e-01 2.52592653e-01 -3.84300739e-01
6.19393252e-02 6.94719374e-01 -8.61175835e-01 2.47882438e+00
-6.43957853e-01 3.10239404e-01 2.32917801e-01 -9.20125902e-01
6.06434286e-01 2.01016709e-01 4.91556346e-01 -5.84760070e-01
2.98020482e-01 3.01599771e-01 -5.18457256e-02 -2.87952244e-01
2.16777518e-01 -3.75861496e-01 1.88141659e-01 2.40561664e-01
1.28793605e-02 -2.75789469e-01 -8.91646892e-02 -1.78641364e-01
8.29798579e-01 5.75055778e-01 -1.25206396e-01 -2.79884338e-01
4.40170765e-01 7.05499388e-03 6.38653994e-01 4.74800199e-01
-2.56084532e-01 9.91213143e-01 5.66969216e-01 -1.24485433e-01
-8.70040774e-01 -7.88133085e-01 -2.09832415e-01 6.66506708e-01
4.83065486e-01 -1.86095893e-01 -9.98235106e-01 -1.04795003e+00
-3.10450822e-01 5.14967442e-01 -7.27336586e-01 -2.64151663e-01
-4.71924007e-01 -8.43187153e-01 2.82858372e-01 5.75492978e-01
4.19434160e-01 -7.63204157e-01 -7.99152672e-01 3.68945479e-01
-3.16170782e-01 -1.39314246e+00 -6.44971430e-01 4.87264276e-01
-1.27221537e+00 -9.77253675e-01 -8.12704206e-01 -8.31110954e-01
1.12145650e+00 -2.32858464e-01 9.88993347e-01 8.44480991e-02
-1.16740935e-01 1.44644737e-01 -1.07427269e-01 -1.48397386e-01
-4.78393495e-01 4.87859517e-01 -5.59605896e-01 -3.83574255e-02
-6.19973503e-02 -2.97185093e-01 -9.90314782e-01 2.98245937e-01
-8.36456001e-01 5.72978497e-01 4.97018635e-01 8.68653536e-01
8.02008569e-01 -2.56095380e-01 4.33777720e-01 -1.12485337e+00
-9.19582546e-02 -1.51802272e-01 -7.38120198e-01 3.15771624e-02
-8.40951145e-01 4.41220850e-01 3.00606489e-01 -2.49011204e-01
-1.23974109e+00 5.97623289e-01 -3.13082278e-01 -4.30032820e-01
5.09890206e-02 5.06050348e-01 -1.52627006e-01 -1.59144387e-01
5.40835381e-01 -1.05483033e-01 1.84067085e-01 -3.02416295e-01
2.28532419e-01 3.06726873e-01 3.43217075e-01 -5.26894331e-01
4.84493047e-01 7.31122375e-01 -4.52780649e-02 -1.92902848e-01
-1.23689282e+00 -5.19866526e-01 -7.46274173e-01 -6.26176689e-03
1.07990825e+00 -1.08891177e+00 -1.89390719e-01 5.05033791e-01
-1.11001873e+00 -8.37142229e-01 -3.63734365e-01 4.92874205e-01
-7.04448521e-01 4.42479879e-01 -7.63901472e-01 -3.38723511e-01
-3.32719475e-01 -1.53490019e+00 1.47398019e+00 1.38343692e-01
-9.61095393e-02 -1.47605693e+00 3.08927726e-02 4.67163950e-01
2.27709219e-01 2.68204927e-01 7.06397295e-01 -3.02228421e-01
-7.44243205e-01 5.52332290e-02 -2.04505548e-01 3.84141088e-01
2.49014571e-01 -4.29630220e-01 -1.16759634e+00 -3.21284145e-01
1.87134326e-01 -4.27850872e-01 1.10821831e+00 7.19267428e-01
1.22224808e+00 -1.17284954e-01 -4.86794770e-01 1.04956818e+00
1.43548906e+00 -1.60541177e-01 5.61120808e-01 4.02637422e-01
9.85633135e-01 9.17937636e-01 5.77203691e-01 2.75242686e-01
6.17774189e-01 7.17979908e-01 6.15433633e-01 -4.31960523e-01
-2.25160405e-01 -4.36018348e-01 3.18292528e-01 5.89376032e-01
3.26272726e-01 -1.50868177e-01 -6.67419434e-01 4.44767416e-01
-1.60062420e+00 -1.84273645e-01 5.17707728e-02 2.31026459e+00
1.26600873e+00 4.04371083e-01 -4.13459241e-02 -1.51736245e-01
5.41513801e-01 3.49851772e-02 -6.33727968e-01 2.19824538e-02
2.49068722e-01 -6.92782477e-02 7.95272648e-01 8.94686639e-01
-1.19808877e+00 9.79678094e-01 5.13021517e+00 3.92769754e-01
-1.29800320e+00 2.20510289e-01 1.03272712e+00 -2.88762376e-02
-4.71438646e-01 1.99096009e-01 -8.78497601e-01 4.36286598e-01
7.01395452e-01 3.71728420e-01 1.16089620e-02 6.19506538e-01
2.56674826e-01 -3.07381034e-01 -1.52150750e+00 8.47917318e-01
-6.17427118e-02 -1.10372627e+00 -2.23980099e-01 3.09146140e-02
9.18328524e-01 1.26130641e-01 3.71758230e-02 -8.11432973e-02
1.48375288e-01 -8.60169411e-01 6.84368789e-01 2.80094266e-01
6.99934661e-01 -3.33135515e-01 6.61444366e-01 3.28969151e-01
-9.21437383e-01 1.71447828e-01 -1.23468854e-01 3.57791960e-01
3.20403606e-01 6.95817947e-01 -1.08818734e+00 5.29030919e-01
4.31267649e-01 9.14520442e-01 -4.27661985e-01 1.01608527e+00
-3.97720754e-01 3.62762362e-01 -3.82279247e-01 3.67302120e-01
3.26399803e-01 -2.37229496e-01 2.65393704e-01 9.07384217e-01
1.66246042e-01 -1.95571363e-01 2.54801750e-01 1.02573669e+00
6.28063977e-02 -1.26296058e-01 -1.76296741e-01 2.94196635e-01
-5.36531471e-02 1.07343137e+00 -1.04159808e+00 -1.87532216e-01
-3.78079712e-01 1.25076091e+00 1.76758334e-01 2.16170222e-01
-9.08994973e-01 1.33391529e-01 2.79903620e-01 4.52181369e-01
7.66326860e-02 -8.45064372e-02 -8.01360667e-01 -1.14890134e+00
4.74640504e-02 -4.24331427e-01 2.92806506e-01 -5.39159536e-01
-1.06316948e+00 5.32152772e-01 -1.63115248e-01 -1.23259234e+00
-4.05403852e-01 -5.70010424e-01 -4.09027457e-01 8.29805195e-01
-1.95992410e+00 -9.90769982e-01 -5.08122444e-01 2.50266612e-01
4.64418590e-01 4.81190205e-01 4.54161465e-01 3.75571668e-01
-4.76229489e-01 6.99809730e-01 -2.99837649e-01 -4.09945054e-03
8.71131182e-01 -1.41968060e+00 3.52093816e-01 5.67625165e-01
-2.91939583e-02 1.37765914e-01 6.39036477e-01 -7.06180811e-01
-9.90972638e-01 -1.18257844e+00 6.47655308e-01 -3.89428288e-01
2.48143658e-01 -2.65717268e-01 -8.97147536e-01 6.78852618e-01
-1.32804036e-01 1.23154990e-01 3.15945178e-01 -6.51519969e-02
-1.74441598e-02 4.56164181e-02 -1.21375442e+00 4.20828760e-01
9.45285857e-01 -4.03147995e-01 -1.16341241e-01 3.75332594e-01
6.14157915e-01 -9.50063288e-01 -7.11222529e-01 4.95382577e-01
4.03829724e-01 -9.71614599e-01 8.66238117e-01 -6.01943582e-02
5.43244243e-01 -2.47307166e-01 1.48666754e-01 -1.24003375e+00
2.35999569e-01 -4.32358056e-01 3.71468633e-01 9.37694132e-01
5.60965955e-01 -4.96665120e-01 1.34730256e+00 6.55168533e-01
-5.22265315e-01 -8.10174048e-01 -9.07027483e-01 -3.94547552e-01
1.27000690e-01 -3.41834515e-01 6.97445571e-02 8.28593493e-01
-2.14619592e-01 1.14261426e-01 -7.67793953e-02 3.05107147e-01
6.55494332e-01 7.43027404e-03 5.26638806e-01 -8.87614369e-01
-6.06441438e-01 -3.06534410e-01 -1.80357531e-01 -1.58206081e+00
2.37538725e-01 -1.11628091e+00 6.16082191e-01 -1.56337595e+00
9.40708891e-02 -7.63125360e-01 -1.91649899e-01 5.52784681e-01
-1.85518473e-01 2.23838702e-01 -2.16869339e-01 1.44321293e-01
-4.56329376e-01 6.00422144e-01 1.52325094e+00 -3.14160399e-02
-5.01441896e-01 4.94853407e-02 -5.20512223e-01 8.30132782e-01
6.09489679e-01 -5.51013768e-01 -6.63472295e-01 -6.74738705e-01
1.36891186e-01 2.63207704e-01 5.10650814e-01 -8.66644144e-01
3.78553540e-01 1.32895604e-01 2.50499845e-01 -2.74454594e-01
3.57032478e-01 -8.00183296e-01 -2.23736644e-01 4.33637798e-01
-5.48949242e-01 -4.55025673e-01 1.19189672e-01 6.11649990e-01
-2.65792549e-01 -4.02812392e-01 1.04532683e+00 -1.62861317e-01
-4.88970578e-01 2.83420384e-01 2.38890663e-01 1.85302883e-01
8.39125991e-01 -4.12368536e-01 1.06296383e-01 -3.08761805e-01
-9.02116299e-01 3.46016377e-01 7.41385162e-01 1.53445572e-01
5.47662854e-01 -9.15015399e-01 -4.71039027e-01 2.23149836e-01
1.08399345e-02 6.69417977e-01 2.53910989e-01 1.07316780e+00
-5.87603271e-01 2.30523199e-01 8.72470587e-02 -9.58923757e-01
-8.27835441e-01 1.97084725e-01 8.89608920e-01 -4.09184873e-01
-6.74039185e-01 1.00659323e+00 6.73909366e-01 -4.39775467e-01
4.00336981e-01 -5.92887044e-01 1.50996536e-01 -2.68800855e-01
6.83791563e-02 -2.34438390e-01 2.27123484e-01 -2.79127389e-01
-2.80354917e-01 6.66461825e-01 -1.36517003e-01 -1.32491976e-01
1.06281316e+00 -3.73896986e-01 2.04932362e-01 3.58305216e-01
1.29914296e+00 -3.04968536e-01 -1.98894203e+00 -2.39337265e-01
1.36517853e-01 -2.87599534e-01 3.67326856e-01 -7.56122053e-01
-1.42119455e+00 8.95004094e-01 7.98607230e-01 -4.29023981e-01
1.08210599e+00 5.02215512e-02 9.78882968e-01 -9.43570412e-05
2.62048811e-01 -1.13893211e+00 2.73835450e-01 1.63246587e-01
5.46185553e-01 -1.60396433e+00 -2.81013370e-01 -6.87081754e-01
-5.94514668e-01 9.07614768e-01 7.76202917e-01 -8.72214958e-02
5.11119664e-01 3.46661478e-01 2.77717531e-01 -6.26022667e-02
-5.81406295e-01 -1.14716008e-01 3.84663373e-01 3.53661984e-01
5.72776616e-01 -3.08716685e-01 -1.03581533e-01 3.38705122e-01
-7.11316764e-02 -2.23063162e-05 3.61873865e-01 9.23829317e-01
-1.38440549e-01 -1.25352752e+00 -2.13240031e-02 2.34527633e-01
-5.88733494e-01 -1.25647828e-01 -8.51803273e-02 5.44965208e-01
7.17465356e-02 6.48598731e-01 1.40160054e-01 -1.14407256e-01
9.36835036e-02 -5.49722984e-02 6.75140560e-01 -8.79261851e-01
-5.43340564e-01 4.70604122e-01 -1.24085927e-02 -5.71049988e-01
-4.85065341e-01 -6.22752249e-01 -1.44619763e+00 3.51596534e-01
-3.59375030e-01 -1.69954836e-01 7.73760140e-01 9.85458970e-01
1.29512519e-01 6.53547704e-01 4.70847785e-01 -9.07300353e-01
-5.71624756e-01 -6.85434401e-01 -2.20295116e-01 3.98444951e-01
3.96288365e-01 -4.90607232e-01 -5.05638182e-01 1.80367857e-01] | [14.609488487243652, -2.1058521270751953] |
2f85a552-98bf-4ef3-9156-e35ca19b208f | r2-trans-fine-grained-visual-categorization | 2204.10095 | null | https://arxiv.org/abs/2204.10095v1 | https://arxiv.org/pdf/2204.10095v1.pdf | R2-Trans:Fine-Grained Visual Categorization with Redundancy Reduction | Fine-grained visual categorization (FGVC) aims to discriminate similar subcategories, whose main challenge is the large intraclass diversities and subtle inter-class differences. Existing FGVC methods usually select discriminant regions found by a trained model, which is prone to neglect other potential discriminant information. On the other hand, the massive interactions between the sequence of image patches in ViT make the resulting class-token contain lots of redundant information, which may also impacts FGVC performance. In this paper, we present a novel approach for FGVC, which can simultaneously make use of partial yet sufficient discriminative information in environmental cues and also compress the redundant information in class-token with respect to the target. Specifically, our model calculates the ratio of high-weight regions in a batch, adaptively adjusts the masking threshold and achieves moderate extraction of background information in the input space. Moreover, we also use the Information Bottleneck~(IB) approach to guide our network to learn a minimum sufficient representations in the feature space. Experimental results on three widely-used benchmark datasets verify that our approach can achieve outperforming performance than other state-of-the-art approaches and baseline models. | ['Xinge You', 'Shujian Yu', 'Shuo Ye', 'Yu Wang'] | 2022-04-21 | null | null | null | null | ['fine-grained-visual-categorization'] | ['computer-vision'] | [ 2.66788512e-01 -6.17135882e-01 -2.64917105e-01 -4.60890621e-01
-4.78014052e-01 -5.64768910e-01 2.88906783e-01 1.73325554e-01
-2.89271176e-01 4.72017109e-01 6.86241314e-03 -6.31207451e-02
-1.05846897e-01 -6.02180183e-01 -4.17680740e-01 -9.90318060e-01
-1.01077196e-04 -7.88702816e-02 7.37410665e-01 -2.21402608e-02
6.94868505e-01 4.91587371e-01 -2.03742456e+00 7.98742712e-01
1.01990128e+00 1.42679346e+00 6.38129890e-01 2.08319694e-01
-4.04944241e-01 5.58139145e-01 -6.56246483e-01 4.06689122e-02
2.65582412e-01 -3.72750312e-01 -5.35498738e-01 1.52781934e-01
7.10696876e-01 -2.06476375e-02 -2.43298188e-01 1.20317638e+00
7.22391546e-01 1.60643578e-01 9.13732827e-01 -9.63036776e-01
-5.88956118e-01 2.47498274e-01 -7.94245422e-01 6.78550422e-01
-2.61970133e-01 1.34435102e-01 8.40318561e-01 -1.07592499e+00
4.62325394e-01 1.46716940e+00 3.64391476e-01 4.06564027e-01
-1.32746196e+00 -8.49444866e-01 8.73282254e-01 5.51307797e-01
-1.69209754e+00 -4.29071575e-01 8.61193180e-01 -2.95877397e-01
7.88374484e-01 3.98160756e-01 3.92859429e-01 7.13963866e-01
2.32958585e-01 7.39895403e-01 1.24982524e+00 -3.84260565e-01
3.73360276e-01 2.73870945e-01 2.54718751e-01 5.83596230e-01
1.22229807e-01 -3.49627696e-02 -5.87841034e-01 3.01996432e-02
6.11266673e-01 3.93165424e-02 -3.77828777e-01 -4.85497564e-01
-1.00308967e+00 6.22498930e-01 6.62854791e-01 4.05678719e-01
-6.40922338e-02 -5.28855659e-02 3.96244854e-01 3.08661997e-01
2.55168885e-01 8.09778199e-02 -3.53075862e-01 3.85703146e-01
-6.95450127e-01 -5.02058789e-02 2.46852234e-01 8.23400259e-01
9.27812934e-01 4.27921228e-02 -6.01197898e-01 1.00534654e+00
1.55668199e-01 2.07194820e-01 7.55690575e-01 -6.31567657e-01
7.44523346e-01 8.14304590e-01 -3.18921655e-01 -1.25800991e+00
-1.01282775e-01 -8.59282136e-01 -1.09735250e+00 3.00194323e-02
1.27223358e-01 3.25407624e-01 -9.80691791e-01 1.64711332e+00
3.38258535e-01 1.71734422e-01 -3.55478734e-01 1.00769687e+00
9.31668282e-01 6.03634536e-01 9.60658193e-02 -3.26725900e-01
1.10889351e+00 -8.06318223e-01 -3.80502343e-01 -6.38056099e-01
1.91086695e-01 -5.68846345e-01 1.09403980e+00 2.59415776e-01
-6.45139456e-01 -1.03236580e+00 -1.13470745e+00 1.43675998e-01
-4.75125939e-01 3.87743443e-01 4.56259280e-01 4.59129959e-01
-8.20434868e-01 5.99726856e-01 -3.55749905e-01 2.46432759e-02
5.37317455e-01 4.12436545e-01 -1.96509138e-01 -3.17206085e-01
-8.87984037e-01 3.70441675e-01 4.58220661e-01 3.04827064e-01
-9.26495731e-01 -4.82127190e-01 -6.23901486e-01 2.10485011e-01
2.59526819e-01 2.44220644e-02 7.02671170e-01 -1.23220992e+00
-9.68372405e-01 5.87038636e-01 -3.48272353e-01 -3.77619117e-02
3.96280587e-01 2.25016415e-01 -3.21111977e-01 3.07408482e-01
2.09597021e-01 8.02344739e-01 1.13608766e+00 -1.14831150e+00
-8.04394484e-01 -4.84446973e-01 -3.22260112e-01 2.22812802e-01
-5.44177949e-01 -5.23103774e-02 -6.56854272e-01 -8.68443429e-01
4.48196858e-01 -4.79306191e-01 -3.32993925e-01 1.93652734e-01
-1.96738109e-01 -3.63396704e-01 7.55714655e-01 -2.60790080e-01
1.43796396e+00 -2.61229682e+00 9.89730060e-02 2.85283417e-01
1.07351199e-01 2.17306703e-01 -3.23942691e-01 2.05650330e-01
1.99742988e-02 1.66227803e-01 -2.73669325e-02 1.42546505e-01
-2.22458854e-01 -1.77344866e-02 -3.39601070e-01 6.18354917e-01
3.74312431e-01 4.98428702e-01 -5.68801165e-01 -8.25840294e-01
7.41551965e-02 1.47797942e-01 -4.06414837e-01 1.45555452e-01
-7.13990852e-02 2.86832601e-01 -6.89396143e-01 6.96869314e-01
1.02312982e+00 -1.91147298e-01 2.38317400e-01 -3.79289299e-01
-2.97358423e-01 6.30384982e-02 -1.52135003e+00 1.38630354e+00
-2.32884958e-02 4.97723937e-01 -1.61044691e-02 -1.20663762e+00
9.36761439e-01 -2.63367772e-01 5.85118532e-02 -8.76493156e-01
1.83518723e-01 1.68738544e-01 9.62433964e-02 -2.79358298e-01
1.77245796e-01 4.35453840e-02 2.71231737e-02 1.08306125e-01
1.32996580e-02 2.98078209e-01 -4.57342295e-03 5.76468110e-02
6.34127855e-01 -2.60989144e-02 3.54553908e-01 -6.39130235e-01
7.35010386e-01 -1.29214436e-01 9.13140535e-01 8.50672126e-01
-5.98296285e-01 4.79748935e-01 2.81174719e-01 -2.87638783e-01
-3.86485785e-01 -8.91987264e-01 -2.63284147e-01 1.37123394e+00
7.50492632e-01 -2.46164843e-01 -5.28325200e-01 -8.13135266e-01
2.24992223e-02 4.08348888e-01 -7.39961028e-01 -4.23606426e-01
-6.09714627e-01 -8.95892382e-01 2.80549347e-01 6.64877117e-01
6.94175959e-01 -8.58049631e-01 -5.54887950e-01 1.96113363e-01
-2.99682289e-01 -6.07948720e-01 -5.05358458e-01 6.01434410e-01
-9.52970326e-01 -8.48560393e-01 -4.69050407e-01 -1.27148712e+00
7.67338991e-01 9.60052788e-01 8.69701505e-01 2.18547091e-01
-6.12144053e-01 -1.92111656e-01 -3.48648190e-01 -1.88823715e-01
1.98031396e-01 -5.80984056e-02 -1.30753055e-01 2.26581812e-01
4.92865473e-01 -2.58769751e-01 -9.11454022e-01 7.40646541e-01
-6.55624092e-01 -3.98014374e-02 7.28954971e-01 1.05113065e+00
1.02447248e+00 3.93539310e-01 6.91516399e-01 -3.53866309e-01
1.32785797e-01 -2.85937428e-01 -5.25765061e-01 3.47669870e-01
-5.11146843e-01 1.55814201e-01 7.46133983e-01 -7.06091464e-01
-1.14577293e+00 4.15790975e-02 1.45685181e-01 -3.71264607e-01
-1.41533822e-01 9.12984237e-02 -4.30778563e-01 -3.15631390e-01
6.10175908e-01 5.48571646e-01 -3.95802170e-01 -6.19230747e-01
1.41499266e-01 7.36972928e-01 2.63517320e-01 -4.76231545e-01
7.25444496e-01 4.29713488e-01 -1.72348872e-01 -6.47748411e-01
-6.42076492e-01 -7.34118938e-01 -8.48946512e-01 -1.77975088e-01
6.01504147e-01 -9.20935750e-01 -5.29688358e-01 2.33954370e-01
-6.76729143e-01 -6.34183139e-02 -9.46690887e-02 2.62619495e-01
-2.33385731e-02 4.18301135e-01 -4.75377887e-01 -7.92836010e-01
-3.92570980e-02 -1.14187622e+00 9.41939652e-01 3.43225747e-01
3.26691985e-01 -4.16325837e-01 -3.97087246e-01 -1.10898323e-01
3.16011906e-01 -1.66319266e-01 1.17339766e+00 -5.23103774e-01
-7.41424978e-01 7.17354193e-02 -6.81910694e-01 3.13821673e-01
1.04779221e-01 -1.78390145e-01 -1.19068277e+00 -4.36860263e-01
-1.74853712e-01 -3.92275482e-01 1.29347706e+00 4.10114825e-01
1.70302987e+00 -1.50365084e-01 -6.27084076e-01 6.42528474e-01
1.55380785e+00 3.57545137e-01 3.96852523e-01 2.46746540e-01
6.09284103e-01 8.02700818e-01 1.01377988e+00 3.79212201e-01
1.12855576e-01 6.67446136e-01 5.25415182e-01 -3.70528586e-02
-2.58055449e-01 -5.85469343e-02 1.86009347e-01 6.88694179e-01
1.24525569e-01 -1.17411524e-01 -4.83367056e-01 5.25822461e-01
-1.66960323e+00 -9.01689053e-01 2.41080783e-02 2.13377094e+00
7.81625032e-01 8.00037608e-02 -1.84215367e-01 1.82427585e-01
8.99977982e-01 2.93036163e-01 -8.30491900e-01 -1.95490673e-01
-3.18054318e-01 1.06268898e-02 4.51757997e-01 3.60298604e-02
-1.25554419e+00 9.36138511e-01 5.98279476e+00 1.30718613e+00
-1.25948095e+00 -8.15428197e-02 9.24614310e-01 -9.20462459e-02
-1.21656328e-01 -2.95081943e-01 -1.07392526e+00 4.31474030e-01
4.85471874e-01 2.54071146e-01 3.23855072e-01 8.29961658e-01
-7.55093619e-02 -1.47197738e-01 -9.48151171e-01 9.86162007e-01
1.59579024e-01 -1.18646681e+00 1.28746301e-01 -1.05838478e-01
5.15644848e-01 -1.57869473e-01 9.16899294e-02 3.52198333e-01
1.82640478e-02 -7.95336902e-01 1.03079736e+00 8.51664990e-02
9.83155727e-01 -7.12316394e-01 6.17869675e-01 4.00340408e-01
-1.69428134e+00 -4.95641410e-01 -1.04328990e+00 3.19979228e-02
-6.27337396e-01 4.15513337e-01 -5.60577214e-01 4.15871352e-01
9.40670431e-01 6.86433971e-01 -1.07798862e+00 1.11628222e+00
5.62070534e-02 5.59748411e-01 9.77387503e-02 -3.02235723e-01
2.46520698e-01 -8.71291570e-03 1.27279505e-01 1.34489501e+00
1.59878775e-01 1.94488287e-01 4.35589433e-01 4.70667362e-01
1.78465530e-01 3.18340838e-01 -2.82226890e-01 3.55655402e-01
5.19537270e-01 1.06821799e+00 -9.73226190e-01 -3.03471357e-01
-4.30918515e-01 1.03964901e+00 5.25117874e-01 3.32671195e-01
-6.62943423e-01 -4.90822613e-01 5.43074608e-01 -1.77914828e-01
7.26508737e-01 7.61271939e-02 -2.75672764e-01 -1.22073424e+00
9.48548168e-02 -8.81894708e-01 7.06992209e-01 -3.10781121e-01
-1.32381511e+00 7.27415442e-01 -2.18413711e-01 -1.58460855e+00
1.99016854e-01 -5.86976826e-01 -3.64064962e-01 8.97004485e-01
-1.65396059e+00 -8.57056141e-01 -4.45700049e-01 6.86268449e-01
9.40675676e-01 -7.42181987e-02 4.16199714e-01 2.85022944e-01
-8.75436127e-01 7.87496150e-01 3.38942438e-01 -5.87094575e-02
8.61584187e-01 -9.82011795e-01 5.84139340e-02 1.00128150e+00
-2.09607910e-02 6.01463497e-01 3.25918585e-01 -6.02652073e-01
-1.37076998e+00 -1.28835082e+00 4.93536204e-01 6.90383241e-02
3.11029732e-01 -5.43129921e-01 -1.05549121e+00 1.65251762e-01
-1.76995084e-01 3.07480067e-01 6.16080642e-01 -1.03976294e-01
-7.66216218e-01 -6.73209786e-01 -1.10940588e+00 4.76266623e-01
1.25539637e+00 -3.94861996e-01 -4.66461062e-01 8.47556368e-02
6.31140709e-01 5.79697043e-02 -2.61387914e-01 2.34113783e-01
6.76152289e-01 -1.00964427e+00 1.08716440e+00 -4.50924128e-01
1.39510021e-01 -6.48739159e-01 -4.48493809e-01 -1.16975820e+00
-8.84349823e-01 2.47018207e-02 3.43507946e-01 1.40944874e+00
1.30096450e-01 -5.61020494e-01 4.37483877e-01 1.02677993e-01
-1.07931055e-01 -6.47817612e-01 -1.03125405e+00 -7.53282487e-01
-1.45681620e-01 -2.59909958e-01 5.62666476e-01 6.70324087e-01
-2.45245025e-01 6.05668575e-02 -1.66223183e-01 2.81910241e-01
7.21468568e-01 6.92246318e-01 2.52151519e-01 -1.34873509e+00
-2.64846534e-01 -4.05762643e-01 -4.34820354e-01 -1.08532977e+00
1.04633253e-02 -8.38073313e-01 4.11438704e-01 -1.22705138e+00
5.02776623e-01 -6.87681913e-01 -7.58355141e-01 5.06260693e-01
-3.87597352e-01 4.36895549e-01 1.83320925e-01 4.06988919e-01
-6.95539415e-01 4.84233409e-01 1.20183897e+00 -4.34951693e-01
-2.03571156e-01 -1.49779096e-01 -9.66762900e-01 5.78831673e-01
7.02185810e-01 -4.18332011e-01 -5.19414663e-01 -3.14971000e-01
-3.46249312e-01 -3.24303418e-01 3.52594793e-01 -9.45345640e-01
1.75561473e-01 -5.01403511e-01 1.06395125e+00 -7.58704901e-01
4.22658324e-02 -7.55571902e-01 -1.47978649e-01 5.08407474e-01
-3.96014482e-01 -2.52343625e-01 3.40783417e-01 8.49119663e-01
-3.01912785e-01 -2.36530542e-01 9.82059658e-01 -1.31973475e-01
-1.17936766e+00 2.40720630e-01 -4.79010463e-01 6.71169534e-03
9.48143065e-01 -3.51042449e-01 -4.85468209e-01 1.18450150e-01
-2.24533364e-01 1.58982009e-01 2.75886178e-01 3.97303820e-01
7.42920101e-01 -1.19733500e+00 -4.28242236e-01 4.96799231e-01
3.75687093e-01 -2.36387819e-01 5.04570723e-01 5.42078316e-01
-1.08838603e-01 4.21019316e-01 -4.78900462e-01 -7.46197760e-01
-1.42179096e+00 9.85563338e-01 2.59176284e-01 7.05540106e-02
-4.54088628e-01 1.02239847e+00 8.85729611e-01 -1.10826530e-02
3.38121563e-01 -2.34166294e-01 -5.38604379e-01 3.34638476e-01
8.63057375e-01 2.97078401e-01 3.63866054e-02 -7.05644071e-01
-7.58851290e-01 8.10446262e-01 -2.09250048e-01 2.18843117e-01
1.11492825e+00 -3.81332070e-01 -3.39502580e-02 2.31400698e-01
1.27017772e+00 -3.70032596e-03 -1.46063447e+00 -3.84155273e-01
-1.38901308e-01 -8.63585532e-01 2.16150552e-01 -7.14341342e-01
-1.15044725e+00 1.17696762e+00 1.05462563e+00 -5.72404731e-03
1.60053062e+00 6.83161914e-02 1.01972498e-01 3.24005008e-01
4.37154889e-01 -1.18125224e+00 2.74518669e-01 1.29720464e-01
7.46809840e-01 -1.11451137e+00 5.56534790e-02 -7.86317289e-01
-7.08281219e-01 9.67993140e-01 9.58496332e-01 -5.58463708e-02
5.80312192e-01 5.64244390e-02 -2.25642957e-02 1.07941985e-01
-7.67572641e-01 -3.33743125e-01 4.83936995e-01 6.70101941e-01
1.72179461e-01 -6.30714558e-03 -2.28187695e-01 6.90072179e-01
4.08508658e-01 -3.59789550e-01 -3.55158350e-03 9.02688146e-01
-9.10034060e-01 -8.38271618e-01 -4.04568911e-01 6.09954178e-01
-4.09892291e-01 -1.47681758e-01 -3.38398784e-01 4.28501636e-01
4.42156553e-01 1.05394614e+00 3.00786309e-02 -4.84093726e-01
2.22875729e-01 -2.51166195e-01 2.81051576e-01 -4.20775771e-01
-6.73872471e-01 4.30128932e-01 -1.85871854e-01 -6.51241839e-01
-2.07489505e-01 -6.21809125e-01 -9.88668442e-01 1.36919320e-01
-4.11020905e-01 1.23498037e-01 3.01866323e-01 6.76072001e-01
5.19714832e-01 6.40952468e-01 8.80483329e-01 -9.13839459e-01
-6.38012171e-01 -7.74670780e-01 -6.98939562e-01 1.33905560e-01
3.90383482e-01 -9.47647691e-01 -4.42942351e-01 -7.01499879e-02] | [9.674469947814941, 1.9536728858947754] |
6cdf857b-95c8-4bf5-bece-8886547dfd7f | clause-based-discourse-segmentation-of-arabic | null | null | https://aclanthology.org/L12-1559 | https://aclanthology.org/L12-1559.pdf | Clause-based Discourse Segmentation of Arabic Texts | This paper describes a rule-based approach to segment Arabic texts into clauses. Our method relies on an extensive analysis of a large set of lexical cues as well as punctuation marks. Our analysis was carried out on two different corpus genres: news articles and elementary school textbooks. We propose a three steps segmentation algorithm: first by using only punctuation marks, then by relying only on lexical cues and finally by using both typology and lexical cues. The results were compared with manual segmentations elaborated by experts. | ['lamia hadrich belguith', 'ar', 'Isk Keskes', 'Farah Benamara'] | 2012-05-01 | null | null | null | lrec-2012-5 | ['discourse-segmentation'] | ['natural-language-processing'] | [ 2.02144787e-01 8.63399133e-02 -7.13318288e-02 -3.73241633e-01
-7.07101703e-01 -9.94734228e-01 6.97336078e-01 8.15005541e-01
-6.88466311e-01 9.24159408e-01 1.10032544e-01 -6.00524008e-01
-1.47485510e-01 -6.90100014e-01 -1.83822125e-01 -3.20912153e-01
2.68049836e-01 7.16832995e-01 6.52266741e-01 -5.27562499e-01
8.67132485e-01 3.69599938e-01 -1.43431330e+00 3.62544864e-01
1.25515056e+00 4.24858361e-01 2.36844540e-01 5.07192910e-01
-8.25930893e-01 7.38109231e-01 -7.95179427e-01 -4.75987583e-01
-5.19896334e-04 -7.24843860e-01 -1.21731639e+00 3.68524402e-01
-2.95370609e-01 1.45640314e-01 5.52350283e-01 1.04667163e+00
7.53979906e-02 -1.90341137e-02 7.35692084e-01 -5.33638239e-01
-1.94883749e-01 1.12052107e+00 -4.29241717e-01 1.55609384e-01
7.97122657e-01 -4.76779044e-01 9.29874301e-01 -7.98654437e-01
1.05049777e+00 7.29114830e-01 6.06792510e-01 9.83935446e-02
-8.96225870e-01 -1.81442738e-01 8.03601891e-02 2.71710813e-01
-1.44118905e+00 -1.65197223e-01 7.70036340e-01 -5.23595929e-01
1.04521585e+00 4.30360362e-02 7.06173241e-01 5.34563303e-01
7.45688006e-02 6.68029547e-01 1.58198369e+00 -1.36104989e+00
3.21370333e-01 4.11442459e-01 7.04786360e-01 3.74877572e-01
2.57309318e-01 -6.63739502e-01 -2.26533916e-02 5.02495095e-02
3.96966130e-01 -4.87859547e-01 -1.30033225e-01 2.96631753e-01
-7.85565138e-01 1.00711393e+00 -5.17741561e-01 1.19887495e+00
-4.43870753e-01 -5.63939631e-01 4.54494625e-01 2.80287385e-01
3.60797554e-01 5.95948994e-01 -4.79817063e-01 -3.06388676e-01
-1.40612555e+00 1.54845947e-02 1.14925230e+00 8.85736525e-01
8.13178122e-01 -3.00417960e-01 2.58876920e-01 9.44409311e-01
4.60461020e-01 4.23808903e-01 5.55991888e-01 -4.65453088e-01
5.13087749e-01 6.65655255e-01 3.36497009e-01 -1.01444316e+00
-2.78728634e-01 1.46087825e-01 -4.19304967e-02 -1.77978557e-02
7.63531148e-01 -4.52959955e-01 -1.14825547e+00 1.17214262e+00
3.88865322e-01 -8.26791883e-01 8.89527239e-03 4.82167691e-01
9.05394495e-01 7.67093897e-01 3.98598388e-02 -6.81887865e-01
1.60064709e+00 -6.61226273e-01 -1.24195623e+00 3.75872016e-01
7.03058243e-01 -1.32191586e+00 9.49632704e-01 8.03392291e-01
-1.30097461e+00 -3.04529607e-01 -1.01271105e+00 5.36541753e-02
-6.94170594e-01 2.97232211e-01 9.92297381e-02 9.14583564e-01
-8.60201955e-01 5.15340388e-01 -7.84250915e-01 -4.83670175e-01
-2.03039721e-01 2.94470042e-01 -2.19800502e-01 2.75902539e-01
-1.09111142e+00 9.30000186e-01 6.27993643e-01 4.93467525e-02
1.97768897e-01 2.75094777e-01 -8.63632083e-01 -1.89534694e-01
6.58472598e-01 1.87974632e-01 1.25761974e+00 -1.22941542e+00
-1.92860556e+00 1.09987342e+00 -1.75754815e-01 -2.94258654e-01
3.74563843e-01 -2.09130868e-01 -3.56822491e-01 6.12677932e-01
1.69106886e-01 5.89600354e-02 4.79125798e-01 -1.20105028e+00
-8.59493136e-01 -2.02625543e-01 1.17668860e-01 2.08644599e-01
-8.00378993e-02 6.83246374e-01 -5.42360067e-01 -7.35218704e-01
3.38904828e-01 -7.85510004e-01 6.73057931e-03 -9.88514841e-01
-3.45717609e-01 -5.05163133e-01 3.25231284e-01 -1.16667414e+00
1.60710287e+00 -1.56489170e+00 3.35654467e-02 7.30736077e-01
-1.78362988e-02 1.83087796e-01 4.00228858e-01 6.86132312e-01
1.03586920e-01 3.44810933e-01 -1.59498557e-01 8.21893290e-02
1.93662241e-01 1.81612074e-01 -1.59621164e-01 9.71718356e-02
1.62233338e-01 4.18096483e-01 -9.32658315e-01 -9.83725190e-01
3.68660316e-02 1.59424603e-01 -1.41624033e-01 -1.46272972e-01
-2.76985139e-01 2.57485285e-02 -4.00012463e-01 7.85235763e-01
3.08893830e-01 2.22496450e-01 6.37509286e-01 7.04130605e-02
-6.19120836e-01 7.20283806e-01 -1.38496757e+00 1.22704172e+00
-2.38828629e-01 3.15707713e-01 -6.39711767e-02 -7.87269592e-01
9.22747195e-01 5.40033221e-01 2.15252623e-01 -4.42771107e-01
5.95172644e-01 6.11341357e-01 1.00391515e-01 -5.11600077e-01
6.70184553e-01 -8.89383852e-02 -3.25787485e-01 6.99224055e-01
2.27499664e-01 -2.95775950e-01 8.59330833e-01 1.04435638e-01
5.43025315e-01 5.01676083e-01 8.22634459e-01 -6.35613799e-01
6.82848573e-01 4.97922063e-01 4.34549659e-01 5.09848893e-01
-1.44369662e-01 4.21533376e-01 8.12065423e-01 -2.71668911e-01
-9.57546532e-01 -5.25839746e-01 -2.73088455e-01 9.38999295e-01
-2.10166231e-01 -6.93848848e-01 -1.26355791e+00 -9.72138464e-01
-5.35064578e-01 7.15182424e-01 -8.02974463e-01 6.79730773e-01
-8.22468460e-01 -6.21412992e-01 3.45722795e-01 1.51570708e-01
1.74983397e-01 -1.29710519e+00 -5.85488737e-01 4.56150025e-01
-1.08822837e-01 -9.40978587e-01 -1.82746246e-01 4.92935926e-01
-4.72316921e-01 -1.04051006e+00 -6.10057414e-01 -8.09000015e-01
5.72171152e-01 -2.11205959e-01 1.03618991e+00 3.53091419e-01
2.13245645e-01 2.02704608e-01 -1.12352228e+00 -6.12989962e-01
-7.85313368e-01 4.12505835e-01 -3.49272072e-01 -2.97892034e-01
7.26417959e-01 -1.66446194e-01 5.97492009e-02 1.40851706e-01
-1.07980835e+00 -2.00215444e-01 7.49023020e-01 4.33212698e-01
5.71242750e-01 -8.22812393e-02 3.35574895e-01 -1.51264083e+00
6.11032426e-01 -2.85172433e-01 -6.33538127e-01 3.60944450e-01
-4.81270105e-01 -2.08941884e-02 6.40127897e-01 -2.70341456e-01
-1.13390899e+00 1.49048895e-01 -6.18527174e-01 3.41299415e-01
-5.59566975e-01 7.98234224e-01 -2.76694119e-01 -4.28751111e-03
5.74311614e-01 8.55909288e-02 -2.53240377e-01 -3.74342322e-01
1.02063499e-01 8.59694719e-01 3.14880788e-01 -6.73025846e-01
7.23231554e-01 -2.00879902e-01 -3.81821126e-01 -1.02953041e+00
-6.07814252e-01 -4.43622559e-01 -1.17785251e+00 -3.66577744e-01
1.01055646e+00 -2.97798425e-01 -1.27140656e-01 3.97474796e-01
-1.14816415e+00 -2.98574269e-01 -2.05130070e-01 5.56564987e-01
-1.64340571e-01 8.10621679e-01 -9.09519792e-01 -6.32890105e-01
-1.73506200e-01 -9.51163530e-01 5.60898781e-01 3.67958128e-01
-5.95084906e-01 -1.11612725e+00 1.66990340e-01 2.38278568e-01
-1.38155252e-01 3.52777839e-01 7.88435817e-01 -1.27527940e+00
3.31578813e-02 -1.22558959e-01 1.37218490e-01 1.56161897e-02
2.27358490e-01 5.89373767e-01 -4.28784043e-01 3.12944889e-01
8.29474479e-02 -1.19489469e-01 4.20047909e-01 1.96177632e-01
2.14167178e-01 -3.11647415e-01 -1.50438979e-01 -6.43305480e-02
1.62278926e+00 5.79604506e-01 6.91101372e-01 8.29501510e-01
5.40660858e-01 8.51159751e-01 8.74777257e-01 1.68495685e-01
3.89865130e-01 3.56284797e-01 -5.97332790e-02 1.44488290e-01
3.74793708e-01 2.82989055e-01 1.55193016e-01 1.39260185e+00
-3.21234435e-01 -2.31404305e-01 -1.42932272e+00 9.33630288e-01
-1.69073606e+00 -6.07118189e-01 -6.02576911e-01 1.88271952e+00
1.13556170e+00 5.13298869e-01 5.11881351e-01 6.45938754e-01
7.61607707e-01 -2.77978480e-01 4.59940016e-01 -9.17892575e-01
-2.31357828e-01 6.71400666e-01 3.04484427e-01 7.39056170e-01
-1.18530512e+00 1.18375731e+00 6.85233974e+00 9.53296244e-01
-9.73621428e-01 -9.13082808e-02 1.29697159e-01 4.36931103e-01
-3.67706239e-01 9.39687863e-02 -9.51641142e-01 4.95821059e-01
8.88033628e-01 4.54310281e-03 -2.32191421e-02 5.45594931e-01
-2.51155216e-02 -6.24286056e-01 -5.63052714e-01 3.01901042e-01
2.89795369e-01 -1.17889750e+00 -1.35065475e-02 -1.53661042e-01
6.86425269e-01 -3.76338750e-01 -6.59522712e-01 -5.61707355e-02
3.31353664e-01 -7.33682632e-01 9.92856383e-01 2.49281481e-01
3.69186103e-01 -7.95178652e-01 1.22067964e+00 2.45911852e-02
-9.09372032e-01 4.09700423e-01 9.22373980e-02 -2.26504907e-01
8.48373547e-02 4.00700420e-01 -7.73735344e-01 7.50789523e-01
4.21639502e-01 2.31922716e-01 -3.56363803e-01 9.76549566e-01
-7.97193885e-01 9.10302520e-01 -4.71806496e-01 -5.51250517e-01
5.18489540e-01 -4.16397452e-01 3.86906654e-01 1.91386747e+00
9.67856795e-02 4.74576443e-01 1.68647215e-01 3.32109123e-01
4.65717733e-01 1.05651462e+00 -1.71680555e-01 -3.11052576e-02
5.06897807e-01 1.36313248e+00 -1.86428893e+00 -4.78172511e-01
-3.88837516e-01 6.93528771e-01 -1.45569351e-02 1.92997128e-01
-5.13692260e-01 -1.06329501e+00 -3.24839532e-01 2.48612955e-01
7.48820245e-01 -4.13464904e-01 -3.91867042e-01 -9.55427587e-01
-6.13502897e-02 -9.67485428e-01 3.57615530e-01 -3.36822331e-01
-8.78107786e-01 9.70601439e-01 3.44819665e-01 -9.15043890e-01
-4.96700078e-01 -6.88199699e-01 -6.23601794e-01 8.92258704e-01
-1.46150136e+00 -8.80165517e-01 1.78971112e-01 3.32559884e-01
4.70161498e-01 -5.58289364e-02 7.49710381e-01 2.52060860e-01
-5.86906791e-01 2.97946513e-01 -5.73776811e-02 3.13454926e-01
5.77680886e-01 -1.51756942e+00 -2.59607583e-01 1.08587027e+00
2.50517696e-01 7.11974382e-01 1.01832068e+00 -5.58740675e-01
-6.60896897e-01 -3.16332370e-01 1.59562957e+00 -2.70106047e-01
8.09321404e-01 -2.84455240e-01 -1.07039475e+00 5.78703344e-01
7.23632932e-01 -8.71023297e-01 1.03230608e+00 6.61089271e-02
1.31560966e-01 2.59224087e-01 -1.03341794e+00 3.37120116e-01
1.58591256e-01 -2.17780590e-01 -1.30137837e+00 2.91115224e-01
2.42062673e-01 -5.55315852e-01 -8.16284537e-01 1.35745704e-02
4.69821304e-01 -6.52370036e-01 1.44185379e-01 -1.66911379e-01
4.48015720e-01 -4.05227363e-01 5.79439066e-02 -9.10611749e-01
-1.58032645e-02 -8.74172151e-01 5.72760165e-01 1.52856469e+00
7.28681803e-01 -5.24022996e-01 4.87492710e-01 2.91398644e-01
-2.62113690e-01 -5.05037546e-01 -4.82464284e-01 -2.81585872e-01
1.16906524e-01 -3.25561464e-01 2.54522204e-01 1.11309731e+00
5.67084670e-01 4.74237889e-01 2.65487134e-01 -2.33587295e-01
1.32532507e-01 1.64228186e-01 4.26931679e-01 -1.36834598e+00
2.57635489e-02 -7.01971948e-01 -1.45074055e-01 -4.97852623e-01
-5.10726161e-02 -3.76547158e-01 3.70445400e-01 -1.52816939e+00
-2.67870516e-01 -2.76913315e-01 9.07417834e-02 6.47688508e-01
-1.41694531e-01 5.48909903e-01 1.47085652e-01 3.01308662e-01
-6.50467753e-01 -1.54333264e-01 6.07067823e-01 3.21005046e-01
-7.13780046e-01 -1.50509760e-01 -5.01824796e-01 1.16273642e+00
8.78638625e-01 -5.74145973e-01 -3.43896374e-02 -3.84998284e-02
6.06620848e-01 -2.91658461e-01 -4.96695310e-01 -6.23099685e-01
1.03795126e-01 -3.99761170e-01 1.27350211e-01 -7.11160660e-01
-3.31460893e-01 -5.69895387e-01 -7.02080503e-02 2.77106017e-01
1.13191880e-01 3.14159811e-01 4.29939747e-01 -1.68515056e-01
-3.20419669e-01 -8.98383439e-01 7.11591959e-01 -2.97034562e-01
-6.66610301e-01 -6.36836290e-01 -1.06539643e+00 3.61841321e-02
1.13885999e+00 -7.18432844e-01 3.76261100e-02 -2.20934466e-01
-1.19231057e+00 8.28105137e-02 6.09909475e-01 -1.67356879e-01
2.11078092e-01 -9.39142585e-01 -5.99384367e-01 2.14379989e-02
-3.31973761e-01 -3.40466440e-01 -3.24597359e-01 9.77302551e-01
-1.16787970e+00 5.49372137e-01 -4.12894547e-01 -3.04872066e-01
-1.47102499e+00 3.73876274e-01 -2.07231820e-01 -5.86902142e-01
-1.51571095e-01 5.57827413e-01 -5.84568679e-01 -1.89501464e-01
6.20422289e-02 -6.42619073e-01 -1.07774353e+00 6.84343219e-01
2.89678305e-01 1.86755687e-01 1.78814828e-01 -1.12203777e+00
-4.22389716e-01 9.19820487e-01 -9.78144817e-03 -6.47324145e-01
1.20639277e+00 -4.04022753e-01 -3.81236821e-01 7.72865355e-01
4.69369680e-01 7.74061620e-01 -3.67533505e-01 -1.41254753e-01
6.23810112e-01 -1.37200683e-01 -2.68412381e-01 -8.55368257e-01
-3.02540600e-01 3.95100594e-01 -1.87820598e-01 8.62685323e-01
1.12956262e+00 -2.01564997e-01 6.48964167e-01 2.98577577e-01
1.96049839e-01 -1.76824081e+00 -2.94696778e-01 7.73680329e-01
5.48815489e-01 -9.07803178e-01 8.77029449e-02 -7.91986167e-01
-6.99178338e-01 1.40247178e+00 1.87068552e-01 -2.53190875e-01
7.78005004e-01 2.55472153e-01 4.09850866e-01 -6.45593479e-02
-3.06298137e-01 -5.88954926e-01 3.45576614e-01 3.56219977e-01
8.52826059e-01 -2.24617511e-01 -1.53604901e+00 9.42543626e-01
-4.07251805e-01 -5.72004244e-02 9.05976951e-01 1.39715827e+00
-7.30567813e-01 -1.54949808e+00 -7.10253358e-01 1.78340137e-01
-1.07784796e+00 -1.39343500e-01 -8.70748162e-01 1.32868826e+00
4.12341714e-01 1.06676090e+00 -1.10028572e-02 -9.45859700e-02
1.68355227e-01 2.90544689e-01 6.21247172e-01 -9.25823092e-01
-9.27684367e-01 6.77261889e-01 3.77567798e-01 4.62352410e-02
-8.63854766e-01 -1.19149840e+00 -1.36367977e+00 -1.11210890e-01
-5.28681874e-01 8.17738175e-01 8.16962004e-01 1.40025568e+00
-3.18571031e-01 3.38943988e-01 4.02816027e-01 -7.07747757e-01
-1.10243522e-01 -1.18354559e+00 -5.50075471e-01 1.57160982e-01
-7.09864050e-02 -4.10208344e-01 -2.95063257e-01 3.31363708e-01] | [10.385178565979004, 10.198904991149902] |
2ac59c35-841e-4cd6-852a-158e08342a44 | dissimilarity-mixture-autoencoder-for-deep | 2006.08177 | null | https://arxiv.org/abs/2006.08177v4 | https://arxiv.org/pdf/2006.08177v4.pdf | Dissimilarity Mixture Autoencoder for Deep Clustering | The dissimilarity mixture autoencoder (DMAE) is a neural network model for feature-based clustering that incorporates a flexible dissimilarity function and can be integrated into any kind of deep learning architecture. It internally represents a dissimilarity mixture model (DMM) that extends classical methods like K-Means, Gaussian mixture models, or Bregman clustering to any convex and differentiable dissimilarity function through the reinterpretation of probabilities as neural network representations. DMAE can be integrated with deep learning architectures into end-to-end models, allowing the simultaneous estimation of the clustering and neural network's parameters. Experimental evaluation was performed on image and text clustering benchmark datasets showing that DMAE is competitive in terms of unsupervised classification accuracy and normalized mutual information. The source code with the implementation of DMAE is publicly available at: https://github.com/juselara1/dmae | ['Fabio A. González', 'Juan S. Lara'] | 2020-06-15 | null | null | null | null | ['image-clustering', 'text-clustering'] | ['computer-vision', 'natural-language-processing'] | [-4.14014041e-01 9.54675004e-02 1.83124349e-01 -6.97510183e-01
-3.45146298e-01 -3.71126652e-01 1.04259825e+00 6.51778430e-02
-4.74796325e-01 -4.80576530e-02 2.01370746e-01 -2.19786003e-01
-3.90582711e-01 -6.58747911e-01 -5.90523541e-01 -8.77527058e-01
-8.04571509e-02 8.95773530e-01 -1.92280814e-01 1.20470040e-01
-1.16234913e-01 4.14494693e-01 -1.48324239e+00 2.81521082e-01
6.44472003e-01 7.10265040e-01 1.22535340e-01 8.17143381e-01
4.68894430e-02 6.18320763e-01 -5.00580370e-01 -3.88306230e-01
1.87252626e-01 -2.11077943e-01 -6.67895257e-01 3.98347676e-02
-3.94858308e-02 -2.06691742e-01 -5.15371501e-01 9.91124868e-01
4.40381706e-01 4.09468681e-01 1.24689901e+00 -1.55593753e+00
-8.04758310e-01 8.92118633e-01 -3.37216884e-01 -9.64422897e-02
-2.65647411e-01 -6.99439570e-02 6.66251183e-01 -1.01392400e+00
1.86531231e-01 1.45938361e+00 7.83682823e-01 4.03371453e-01
-1.48075855e+00 -3.97635847e-01 -8.26923624e-02 7.60400146e-02
-1.66608453e+00 -2.52077848e-01 5.75681508e-01 -7.40685821e-01
1.08817875e+00 2.07447514e-01 3.13288599e-01 1.12683415e+00
-2.93478798e-02 1.09072018e+00 6.33677125e-01 -3.66301388e-01
4.16750997e-01 1.26153037e-01 2.23513871e-01 4.32646185e-01
-1.48478430e-02 1.27511928e-02 1.52129322e-01 -3.10278267e-01
6.70278788e-01 3.34905475e-01 9.71789882e-02 -4.48331594e-01
-1.07661247e+00 1.12771010e+00 4.88696992e-01 4.48627204e-01
-3.43514860e-01 1.89562663e-01 3.76082361e-01 2.08191603e-01
4.85749811e-01 1.55161947e-01 -4.53920424e-01 4.25094320e-03
-1.00449395e+00 1.00076668e-01 8.84057224e-01 6.85616493e-01
5.92638135e-01 1.42876178e-01 1.53040364e-01 1.23796034e+00
6.77684486e-01 -6.58850297e-02 9.35189724e-01 -1.32345569e+00
-1.99278712e-01 5.79483390e-01 -1.98857903e-01 -7.99967170e-01
-5.43791890e-01 -2.08341286e-01 -1.23886549e+00 3.56081098e-01
1.78450167e-01 -2.80095190e-01 -8.56216848e-01 1.86078429e+00
3.04577440e-01 2.45481357e-01 1.06678456e-01 7.28637278e-01
9.69092071e-01 5.55662990e-01 4.39923070e-03 2.41096452e-01
1.04106998e+00 -8.99348617e-01 -5.64948976e-01 4.58496995e-02
4.62885350e-01 -5.36450088e-01 6.22664869e-01 5.02295196e-01
-1.03986394e+00 -7.81955123e-01 -9.74771023e-01 -3.48873734e-02
-7.58625567e-01 2.59467751e-01 4.92524415e-01 6.60901189e-01
-1.48155880e+00 8.15872788e-01 -1.29436648e+00 -2.99839199e-01
4.65668589e-01 7.51639843e-01 -4.23423380e-01 1.44763380e-01
-9.95613158e-01 6.57230377e-01 9.15414393e-01 1.48459211e-01
-8.71454179e-01 -3.89910996e-01 -9.05375123e-01 1.84549555e-01
-2.44303405e-01 -8.55598688e-01 1.17800117e+00 -1.16062558e+00
-1.74131155e+00 7.89891243e-01 2.47083426e-01 -5.10170102e-01
3.13765228e-01 -1.60347179e-01 -4.48645234e-01 -3.50987129e-02
-2.06449628e-01 1.33987069e+00 7.03337431e-01 -1.19612980e+00
-2.69663841e-01 -2.31916755e-01 -5.39019763e-01 1.39040649e-01
-3.09563667e-01 7.07439752e-03 -3.57857615e-01 -6.99993551e-01
-1.39407232e-01 -9.28957820e-01 -4.21627373e-01 -3.38100284e-01
-6.09648228e-01 -5.86160183e-01 8.43908906e-01 -5.42417228e-01
9.13435459e-01 -2.36133456e+00 4.28200573e-01 1.06893688e-01
4.34755683e-01 2.47645497e-01 -2.88552523e-01 4.94242787e-01
-5.14537573e-01 5.38882846e-03 -6.89019322e-01 -7.03623772e-01
5.96991062e-01 1.08839512e-01 2.80107737e-01 6.50427163e-01
2.96636194e-01 8.60347390e-01 -6.28795028e-01 -1.00778729e-01
5.68752348e-01 9.78737116e-01 -4.71423447e-01 2.70088166e-01
-2.34262228e-01 1.37854308e-01 2.50861496e-01 1.77175984e-01
8.21597040e-01 -2.95388222e-01 3.43236655e-01 1.09346015e-02
2.42780805e-01 8.91403705e-02 -1.30642903e+00 1.78530002e+00
-3.44429985e-02 7.49332190e-01 5.91173209e-02 -1.06490266e+00
7.86833584e-01 3.06233972e-01 6.21924102e-01 -2.99091320e-02
6.22337282e-01 -2.43704151e-02 2.33683735e-01 -1.26186743e-01
3.20755035e-01 1.55293554e-01 -1.52124077e-01 5.39283872e-01
6.64287984e-01 1.29587159e-01 2.26825520e-01 1.76793024e-01
8.53814006e-01 -4.61125337e-02 1.27678707e-01 -4.31838900e-01
2.87873000e-01 -2.53324628e-01 3.80252212e-01 5.36101758e-01
2.30848249e-02 5.48137188e-01 1.95482120e-01 -3.17475319e-01
-9.96707857e-01 -1.45316541e+00 -2.40002319e-01 1.11973369e+00
-4.22098756e-01 -4.10197765e-01 -1.01203394e+00 -2.78971583e-01
1.09795909e-02 7.75469661e-01 -7.23785400e-01 -3.47861528e-01
-2.50395685e-02 -1.01958084e+00 7.92856455e-01 5.81876099e-01
3.79370064e-01 -9.89716113e-01 -7.50403628e-02 4.63771187e-02
2.12858319e-02 -6.87563598e-01 -1.40474275e-01 6.19227231e-01
-8.68551910e-01 -9.48637068e-01 -5.80318987e-01 -9.54043567e-01
5.18377006e-01 -2.30724886e-01 1.12225735e+00 -2.87017822e-01
-1.90857828e-01 4.77690369e-01 -6.45691827e-02 -3.70200336e-01
-7.57588625e-01 5.91009036e-02 4.26092535e-01 -7.87643045e-02
1.04454052e+00 -9.45225716e-01 -6.70099080e-01 7.79049844e-02
-1.29332781e+00 -1.58269078e-01 3.70055467e-01 6.34187579e-01
5.16623378e-01 3.70921820e-01 4.50108588e-01 -5.30166626e-01
8.92737389e-01 -8.82852495e-01 -3.83350432e-01 -2.42608353e-01
-4.85146791e-01 -1.04525656e-01 4.55852479e-01 -5.22043049e-01
-7.23828614e-01 1.77554131e-01 -3.60663205e-01 -8.30442131e-01
-9.20284331e-01 4.44659948e-01 -3.01368594e-01 5.96453547e-01
5.37542820e-01 1.80951744e-01 2.70862013e-01 -6.37697697e-01
8.67144048e-01 9.68636870e-01 7.17483461e-01 -3.02570045e-01
3.30420375e-01 5.11922896e-01 -4.20237184e-01 -5.85487604e-01
-2.53845483e-01 -4.91634399e-01 -1.04896200e+00 3.66839059e-02
1.05615485e+00 -1.09586763e+00 -7.91040003e-01 7.47878790e-01
-8.28081369e-01 -6.31186366e-01 -1.25958771e-01 6.06703520e-01
-6.42237902e-01 3.14657807e-01 -8.53857100e-01 -4.57779646e-01
-3.10615510e-01 -1.07257831e+00 7.68273115e-01 2.26483569e-01
-4.67192411e-01 -1.45849967e+00 1.85220435e-01 -2.14022957e-02
2.24917665e-01 3.76260072e-01 9.60955083e-01 -1.29782367e+00
1.42261565e-01 -2.65296072e-01 1.37066737e-01 7.75984108e-01
1.61801279e-01 4.11227107e-01 -1.11184323e+00 -2.99088180e-01
1.11754902e-01 -1.09555312e-01 1.05452383e+00 7.67488599e-01
1.23438287e+00 -2.86932826e-01 -2.54334331e-01 5.98461807e-01
1.38165712e+00 1.29045218e-01 5.45629203e-01 2.37650037e-01
7.44890630e-01 5.59155583e-01 -1.17930532e-01 5.25084615e-01
6.19950593e-01 2.04620555e-01 5.64980209e-01 -3.34031843e-02
4.72119041e-02 1.98787376e-01 2.80002773e-01 1.15339208e+00
2.78989941e-01 -2.32304350e-01 -1.02545416e+00 6.44834578e-01
-2.04588389e+00 -1.03188145e+00 -2.99940586e-01 1.86034095e+00
6.96525693e-01 -1.10155247e-01 4.73689914e-01 2.21394166e-01
9.20734167e-01 -2.01103717e-01 -5.47867596e-01 -6.68296158e-01
-7.39930719e-02 7.86554068e-03 6.98560709e-03 3.96661431e-01
-1.46208525e+00 7.74277985e-01 6.48705626e+00 9.60963726e-01
-7.89628804e-01 1.88029557e-01 6.88299835e-01 -1.97393298e-01
-2.15030480e-02 -5.18793285e-01 -4.01691109e-01 6.09297276e-01
1.44703031e+00 2.02422753e-01 4.35317665e-01 7.92349517e-01
-3.26519236e-02 1.12637185e-01 -1.22788250e+00 8.37119162e-01
-2.28940248e-01 -1.19268322e+00 -2.29955316e-01 1.44303665e-01
6.55586183e-01 7.24171042e-01 3.79692376e-01 4.97139007e-01
1.06863582e+00 -1.15355253e+00 4.91355598e-01 6.48490071e-01
2.93371260e-01 -1.13401842e+00 6.76751435e-01 6.22695088e-01
-6.74639225e-01 -7.61535987e-02 -4.13381010e-01 1.15227196e-02
-2.96112508e-01 6.98257565e-01 -7.94363618e-01 2.34402359e-01
8.60160470e-01 8.29680860e-01 -6.66788936e-01 9.11717534e-01
1.21322289e-01 6.28758788e-01 -4.94007081e-01 2.47780636e-01
3.45443040e-01 -6.49924994e-01 3.87782931e-01 1.64080024e+00
1.49231270e-01 -3.72219861e-01 6.31350428e-02 1.11224496e+00
-1.66211471e-01 -2.46717185e-01 -4.57460761e-01 -7.19018504e-02
6.33551300e-01 1.36910927e+00 -9.04728889e-01 -4.65175092e-01
-1.66577145e-01 1.10408425e+00 4.00920987e-01 3.80663902e-01
-8.44176292e-01 -4.67178166e-01 9.25711513e-01 -3.63977879e-01
6.61669314e-01 -2.20378116e-01 -8.77369344e-02 -8.41924250e-01
-4.39423025e-01 -8.22011530e-01 4.92289722e-01 -6.25950336e-01
-1.67699647e+00 7.53265023e-01 -7.73480511e-04 -1.05034792e+00
-5.07442474e-01 -7.27991045e-01 -7.83544719e-01 7.46985257e-01
-7.91372359e-01 -8.48519444e-01 -2.57228762e-02 7.07317352e-01
2.84896076e-01 -4.01729226e-01 1.18347526e+00 2.75101811e-01
-9.04433191e-01 4.56910342e-01 9.78350103e-01 4.97560859e-01
5.10134339e-01 -1.61076546e+00 4.33690071e-01 6.94370866e-01
2.82557100e-01 7.40634084e-01 5.27301192e-01 -2.63956428e-01
-7.86942303e-01 -1.35403919e+00 4.05133843e-01 -6.70268655e-01
6.45718634e-01 -5.15410423e-01 -9.94949520e-01 8.33949089e-01
7.23921478e-01 -1.69767097e-01 1.37005305e+00 -6.87075080e-04
-4.95265692e-01 3.81213814e-01 -1.16817832e+00 5.37193775e-01
4.65493113e-01 -5.85145891e-01 -5.14909804e-01 2.84985244e-01
6.80333495e-01 -1.90451235e-01 -1.53387249e+00 2.49036372e-01
2.81077623e-01 -1.26539505e+00 1.09964263e+00 -5.88658750e-01
4.52119201e-01 -3.55343848e-01 -4.00999516e-01 -1.43911660e+00
-7.16568410e-01 -2.80673325e-01 -5.91062546e-01 1.28144526e+00
2.24772960e-01 -3.33718389e-01 4.46963221e-01 5.05109549e-01
-1.89369440e-01 -8.85523796e-01 -8.21649671e-01 -7.62551665e-01
5.75969875e-01 -6.31601870e-01 7.97607005e-01 1.33657551e+00
-1.09897174e-01 1.50914609e-01 1.04082100e-01 1.85442746e-01
6.00412905e-01 -9.83510613e-02 5.80375671e-01 -1.39285088e+00
-5.01203537e-01 -1.21610558e+00 -4.38764274e-01 -7.70199418e-01
4.67929363e-01 -1.37232018e+00 -1.82227977e-02 -1.68321085e+00
3.83369997e-02 -7.55574927e-02 -5.63008487e-01 3.89152706e-01
-5.74821010e-02 2.60534286e-01 -6.74630627e-02 3.00275058e-01
-7.71563649e-01 6.94398224e-01 3.54617953e-01 -2.05018118e-01
-2.73063362e-01 4.14799824e-02 -7.73217499e-01 9.74806726e-01
1.16777587e+00 -5.83471775e-01 -2.32304856e-01 -4.99537051e-01
-2.11024836e-01 -3.71512204e-01 2.81077743e-01 -1.00392342e+00
2.15937883e-01 3.52999359e-01 7.55711675e-01 -7.04077542e-01
3.48625600e-01 -7.08713055e-01 4.22070086e-01 3.88399363e-01
-3.57833266e-01 6.49498105e-02 3.02298725e-01 5.54798067e-01
-2.65700877e-01 -1.47212490e-01 8.44717622e-01 9.86781716e-02
-5.22115469e-01 2.62070447e-01 -9.48976815e-01 -3.92971486e-01
7.81761646e-01 -2.72875518e-01 -3.09471358e-02 -2.03914791e-01
-1.08401692e+00 1.09013624e-01 5.26929200e-01 6.05670273e-01
4.82846737e-01 -1.56880713e+00 -8.00593853e-01 7.47653991e-02
-1.58534572e-01 1.71286419e-01 1.32283628e-01 9.05629575e-01
-3.80126089e-01 2.33908534e-01 -9.83791649e-02 -7.48089850e-01
-8.98168147e-01 7.83997595e-01 7.27852404e-01 -4.11025248e-02
-3.22994113e-01 9.08333361e-01 9.09254849e-02 -1.17668545e+00
3.80670458e-01 -1.75589964e-01 -2.16889679e-01 -1.09944858e-01
3.84988487e-01 4.80224520e-01 -9.25289094e-03 -6.84395969e-01
-3.97482812e-01 -4.62427624e-02 -2.10225172e-02 -5.63467070e-02
1.62172759e+00 -1.57660812e-01 -4.32225943e-01 6.61960363e-01
1.56556845e+00 -6.45446420e-01 -1.10229325e+00 -2.91809559e-01
2.56000370e-01 2.74501145e-01 2.22475737e-01 -7.51327276e-01
-1.11395621e+00 8.86446655e-01 7.96407282e-01 2.61527985e-01
1.06907749e+00 9.60520282e-02 3.66301805e-01 5.34351707e-01
-4.22076136e-01 -1.07861352e+00 -6.61989301e-02 5.15131831e-01
6.20798290e-01 -1.34774435e+00 -3.08047384e-01 3.79517794e-01
-6.61279202e-01 9.95391190e-01 5.04307926e-01 -4.84703928e-01
1.35938370e+00 4.51697737e-01 1.49146885e-01 -2.93372929e-01
-8.28458548e-01 -1.13926508e-01 4.22283709e-01 8.37830186e-01
6.51864827e-01 3.00060987e-01 2.25776583e-01 9.65466201e-01
-3.06925178e-01 -3.33292395e-01 2.98671305e-01 4.82125849e-01
-2.01450929e-01 -1.05696249e+00 -3.55249584e-01 4.95419025e-01
-3.54379386e-01 -6.76243827e-02 -4.82464671e-01 7.35825181e-01
3.32534313e-01 1.02001929e+00 5.74002087e-01 -5.98850369e-01
-1.72731608e-01 2.92164117e-01 2.56006539e-01 -5.25148451e-01
-7.62938797e-01 3.56817931e-01 -3.50649536e-01 -3.00079823e-01
-6.00754142e-01 -5.78866124e-01 -1.40378511e+00 -4.97497737e-01
-2.81948745e-01 1.93890467e-01 7.83348382e-01 7.65338898e-01
5.97012937e-01 6.37386024e-01 4.94376957e-01 -1.05751836e+00
-2.27364659e-01 -1.41399956e+00 -6.78609312e-01 3.15472096e-01
1.56149462e-01 -3.42879266e-01 -3.23846519e-01 1.47410512e-01] | [9.119390487670898, 3.20656418800354] |
73695af7-84c4-4e70-89f8-d8bf0378cb8e | slot-filling-for-extracting-reskilling-and | 2207.04862 | null | https://arxiv.org/abs/2207.04862v1 | https://arxiv.org/pdf/2207.04862v1.pdf | Slot Filling for Extracting Reskilling and Upskilling Options from the Web | Disturbances in the job market such as advances in science and technology, crisis and increased competition have triggered a surge in reskilling and upskilling programs. Information on suitable continuing education options is distributed across many sites, rendering the search, comparison and selection of useful programs a cumbersome task. This paper, therefore, introduces a knowledge extraction system that integrates reskilling and upskilling options into a single knowledge graph. The system collects educational programs from 488 different providers and uses context extraction for identifying and contextualizing relevant content. Afterwards, entity recognition and entity linking methods draw upon a domain ontology to locate relevant entities such as skills, occupations and topics. Finally, slot filling integrates entities based on their context into the corresponding slots of the continuous education knowledge graph. We also introduce a German gold standard that comprises 169 documents and over 3800 annotations for benchmarking the necessary content extraction, entity linking, entity recognition and slot filling tasks, and provide an overview of the system's performance. | ['Philipp Kuntschik', 'Alexander van Schie', 'Andreas Fraefel', 'Roger Waldvogel', 'Albert Weichselbraun'] | 2022-07-11 | null | null | null | null | ['slot-filling'] | ['natural-language-processing'] | [-2.02415019e-01 3.00955087e-01 -5.57578087e-01 -1.37699291e-01
-6.55164361e-01 -8.47209513e-01 5.01380265e-01 1.26649237e+00
-6.57471895e-01 1.02487814e+00 2.51381814e-01 -4.24471855e-01
-7.18507230e-01 -1.09416270e+00 -5.37664711e-01 -3.84505279e-02
5.12735128e-01 7.97523320e-01 6.46366358e-01 -1.97011665e-01
2.70896882e-01 4.82136875e-01 -2.04490757e+00 8.26916769e-02
1.16597760e+00 5.56244135e-01 3.56116295e-01 2.96370238e-01
-8.82772863e-01 6.02451622e-01 -1.05596030e+00 -7.74751067e-01
-3.90462279e-01 -1.62500590e-01 -1.11732483e+00 -1.56266153e-01
4.37291056e-01 3.75231236e-01 -1.18709259e-01 1.10057449e+00
5.23083687e-01 4.13494289e-01 2.91557908e-01 -1.17060494e+00
-3.56518209e-01 8.87297928e-01 -1.23773310e-02 1.80679440e-01
9.43112195e-01 -8.54231179e-01 8.53427172e-01 -5.96189260e-01
1.34434557e+00 6.49241507e-01 5.32337546e-01 2.19264284e-01
-7.05782533e-01 -6.23477578e-01 -1.63908273e-01 4.04714763e-01
-1.36329305e+00 -9.08671170e-02 3.12261552e-01 -7.88023472e-01
7.91318536e-01 4.31106865e-01 8.30986798e-01 5.57176769e-01
-2.70810008e-01 3.50122839e-01 9.82884765e-01 -9.76610065e-01
1.92287564e-01 7.81794012e-01 3.50649834e-01 7.84709513e-01
7.76926994e-01 -5.79035401e-01 -8.05545330e-01 -6.52531162e-02
5.22492945e-01 -1.90803602e-01 1.49916913e-02 -3.91982615e-01
-9.57282245e-01 3.03155154e-01 -3.77861187e-02 6.60182714e-01
-2.05521405e-01 -6.79695785e-01 5.01216471e-01 2.22591206e-01
-3.87731791e-02 9.26724374e-01 -4.36785251e-01 -4.16139722e-01
-1.07760048e+00 5.47159910e-01 1.07913172e+00 1.30859673e+00
9.01735842e-01 -6.57248855e-01 -1.99216828e-01 8.82113636e-01
4.77444381e-02 7.52040818e-02 6.22646391e-01 -7.90798008e-01
4.38643545e-01 1.33278358e+00 2.42931589e-01 -1.22220671e+00
-4.44969445e-01 -4.24487919e-01 -7.52530769e-02 -2.08657444e-01
5.45255899e-01 -2.39764556e-01 -6.98157191e-01 1.23451602e+00
5.27285933e-01 -1.00647122e-01 1.12964520e-02 5.33602953e-01
1.56164670e+00 6.15882017e-02 4.46633965e-01 -6.14222437e-02
2.06676865e+00 -6.66911185e-01 -1.12988257e+00 1.00220077e-01
5.69694042e-01 -1.06619847e+00 5.22948623e-01 -1.61835179e-02
-1.19838059e+00 -5.35189927e-01 -6.57821476e-01 -2.23177075e-01
-1.10279095e+00 1.43793911e-01 6.02641404e-01 7.12430120e-01
-6.39885306e-01 6.11528277e-01 -4.57874715e-01 -5.45074642e-01
2.33934909e-01 3.54592979e-01 -6.94490254e-01 -5.23263030e-02
-1.23927462e+00 9.95012403e-01 5.72036386e-01 -6.10941529e-01
-2.09748633e-02 -9.68668401e-01 -1.04603529e+00 2.43441284e-01
6.68578625e-01 -6.18153453e-01 1.01720846e+00 -4.03073072e-01
-9.71177578e-01 1.24172390e+00 3.46768871e-02 -1.15417257e-01
2.39847466e-01 7.10841715e-02 -9.05961990e-01 8.25539231e-02
5.70810795e-01 2.41476610e-01 -8.77373964e-02 -6.88227415e-01
-1.20062864e+00 -3.05301756e-01 7.28770345e-02 4.11291212e-01
-4.12792057e-01 3.69641721e-01 -9.18891370e-01 -4.05454874e-01
2.53264785e-01 -3.83425206e-01 -1.04817882e-01 -9.20297384e-01
-1.90016642e-01 -5.72832823e-01 4.69283938e-01 -9.02723908e-01
1.87985408e+00 -1.65502322e+00 7.21319616e-02 4.02922213e-01
2.71434247e-01 -5.07593863e-02 5.03118217e-01 6.46001160e-01
-5.03342925e-03 1.57035321e-01 4.99153674e-01 2.17224807e-01
3.81801754e-01 2.37679437e-01 1.19975172e-01 -1.32086322e-01
-3.51328284e-01 6.00473404e-01 -1.18216729e+00 -9.99624789e-01
1.32268474e-01 9.11985412e-02 -3.10794026e-01 -1.29543155e-01
7.22251907e-02 2.01169029e-01 -6.91726685e-01 1.02988708e+00
1.12974092e-01 -1.19835809e-01 4.92281318e-01 3.34725492e-02
-5.93906760e-01 6.50625408e-01 -1.33470106e+00 1.89409947e+00
-5.00586569e-01 7.21503377e-01 -1.93521064e-02 -8.12654257e-01
8.83611798e-01 4.66304451e-01 7.32333481e-01 -6.82636201e-01
4.74745370e-02 6.13661289e-01 -4.84065115e-01 -8.12868476e-01
1.20733249e+00 3.09490353e-01 -4.74316299e-01 -1.57942399e-01
4.59894061e-01 -6.37272745e-02 8.39221239e-01 4.00547773e-01
1.15121651e+00 2.85685033e-01 2.99749881e-01 -3.48410845e-01
5.83119035e-01 4.08937663e-01 5.69499195e-01 4.82267708e-01
2.92118043e-02 -8.27752799e-02 5.75342119e-01 -2.69400835e-01
-9.86563683e-01 -6.00391150e-01 -3.57394844e-01 1.15215540e+00
8.83697644e-02 -8.05679500e-01 -8.53357196e-01 -7.28811145e-01
1.28814682e-01 3.54722798e-01 -1.56604782e-01 3.36663187e-01
-3.19623351e-01 -1.24600090e-01 1.94182381e-01 8.82516876e-02
4.15562004e-01 -9.74647701e-01 -6.49913192e-01 1.54880971e-01
-3.28513861e-01 -1.08245718e+00 1.60266995e-01 7.99034014e-02
-3.61370206e-01 -1.10813904e+00 -7.04572499e-01 -1.08695138e+00
8.76293957e-01 -3.10649067e-01 1.53437006e+00 1.60408348e-01
-3.95296425e-01 4.80518997e-01 -1.97349712e-01 -5.45576215e-01
-1.57981724e-01 5.87940872e-01 -2.14818716e-01 -9.26283896e-01
7.67608106e-01 -3.00462425e-01 -2.16108263e-01 2.95495778e-01
-7.63088167e-01 -1.95043504e-01 3.27379137e-01 6.21731043e-01
6.02354527e-01 4.92798865e-01 6.34373903e-01 -1.35279787e+00
6.58993959e-01 -6.02235973e-01 -8.31847072e-01 4.63874131e-01
-9.86435831e-01 -1.49460912e-01 -4.37435545e-02 -3.69845890e-02
-1.20672572e+00 5.28994191e-04 -2.08663523e-01 1.59300968e-01
-4.91079420e-01 8.93941760e-01 -2.79565901e-01 -2.91331802e-02
7.20830321e-01 -2.38492832e-01 -6.76182687e-01 -7.08847046e-01
2.62011051e-01 7.81354547e-01 6.83321297e-01 -9.12415922e-01
7.08593547e-01 -2.20245287e-01 7.22609982e-02 -6.47134185e-01
-7.88396716e-01 -8.90799940e-01 -7.43999720e-01 -3.80871236e-01
7.89484203e-01 -9.46766317e-01 -8.59071314e-01 3.88271436e-02
-7.94721305e-01 2.25676715e-01 -7.57547975e-01 5.77361286e-01
-8.62107351e-02 2.04680413e-01 -4.29029912e-01 -4.66979831e-01
2.85685897e-01 -8.22033346e-01 6.15223169e-01 1.10833275e+00
-3.50680053e-01 -8.73906374e-01 6.34543002e-02 8.86820436e-01
1.97115570e-01 1.79511979e-01 8.65060627e-01 -9.41295505e-01
-3.28665882e-01 -5.14766335e-01 -6.21825792e-02 -3.53159040e-01
-1.17198981e-01 -2.25042507e-01 -4.93975043e-01 1.82238594e-01
-8.52928221e-01 1.42700016e-03 3.16534847e-01 -8.37007761e-02
9.35897768e-01 -1.72533944e-01 -6.86628938e-01 1.77890003e-01
1.25490129e+00 3.56628120e-01 5.10012090e-01 8.65489841e-01
4.19187933e-01 9.13450480e-01 7.75404215e-01 3.42175007e-01
4.88848388e-01 8.34258854e-01 -7.83419013e-02 4.64130156e-02
-1.20483294e-01 -4.08197373e-01 -3.37108791e-01 8.15291405e-01
-3.59516114e-01 8.44164342e-02 -1.05501056e+00 9.25140560e-01
-1.58862936e+00 -9.87797022e-01 -9.20790881e-02 2.16426659e+00
1.19234049e+00 1.05457067e-01 7.40532652e-02 -1.64561830e-02
6.17630005e-01 -3.34842950e-01 -7.83732012e-02 -2.77337104e-01
1.15306437e-01 5.73883653e-01 8.57503891e-01 2.46865675e-01
-8.58260214e-01 9.77839828e-01 6.21140766e+00 1.07809615e+00
-3.61385524e-01 3.60652842e-02 2.18893260e-01 3.40782642e-01
-3.72815132e-01 6.77119941e-02 -1.31924927e+00 3.16585660e-01
9.31325912e-01 -6.27520442e-01 -1.29305723e-03 9.21793401e-01
-3.18028361e-01 -2.80372769e-01 -4.98597234e-01 7.12149262e-01
-9.21910703e-02 -1.66596627e+00 -5.19216657e-01 -5.32768443e-02
9.77867663e-01 -3.78078341e-01 -3.71834368e-01 6.47327542e-01
6.43117011e-01 -7.01400161e-01 5.72256684e-01 6.39523208e-01
5.99906445e-01 -7.81744003e-01 6.99505806e-01 2.95026097e-02
-1.21593392e+00 -6.99206665e-02 -2.43713945e-01 4.78044748e-02
-1.38720334e-01 6.10897660e-01 -8.90513659e-01 1.01120055e+00
7.99559057e-01 2.92977333e-01 -5.76881170e-01 1.37177515e+00
-3.75042140e-01 2.05933124e-01 -9.68913361e-02 -7.64871482e-03
-2.17914164e-01 -1.60945371e-01 4.99883056e-01 1.31433010e+00
5.86680710e-01 3.97646166e-02 1.31618723e-01 6.12028539e-01
-3.77630025e-01 5.02037823e-01 -3.68570983e-01 -3.53217781e-01
9.12739992e-01 1.51841271e+00 -1.02081347e+00 -5.47956645e-01
-4.10537988e-01 5.88110209e-01 -1.20021868e-02 2.17236921e-01
-5.09514630e-01 -1.05213284e+00 6.60212755e-01 3.35383624e-01
4.71579805e-02 -1.37720823e-01 1.05378814e-01 -9.12510872e-01
-1.11911394e-01 -7.26689100e-01 7.20707059e-01 -4.79892313e-01
-6.20370269e-01 3.94529849e-01 1.29922017e-01 -9.49742973e-01
-3.61674845e-01 -4.74905074e-01 -6.91409558e-02 9.68150675e-01
-1.45272529e+00 -6.68962300e-01 -4.25427586e-01 2.93120384e-01
2.68145204e-01 -1.83485270e-01 8.28300416e-01 9.59061742e-01
-7.16272652e-01 4.84653801e-01 -1.62215605e-02 2.68366069e-01
8.25433552e-01 -1.39900529e+00 -2.69466907e-01 4.94052052e-01
1.40993640e-01 5.91807306e-01 6.68201685e-01 -8.72587979e-01
-8.92445266e-01 -6.14653826e-01 1.69192910e+00 -5.46785235e-01
8.01244080e-01 3.72369066e-02 -6.93332613e-01 4.12251920e-01
2.03913867e-01 -3.94322693e-01 9.00992990e-01 3.59998494e-01
-1.08340830e-01 4.73064817e-02 -1.01082790e+00 3.68368208e-01
1.00385153e+00 -4.83213484e-01 -6.72654271e-01 4.97814149e-01
5.31425893e-01 -9.87879694e-01 -1.47790945e+00 1.09741740e-01
4.37376827e-01 -6.15838170e-01 1.05823195e+00 -7.65726268e-01
4.00464654e-01 -2.18336627e-01 2.40159005e-01 -1.02138937e+00
-9.60752815e-02 -4.63499248e-01 2.00634986e-01 1.67361212e+00
6.73676610e-01 -2.12515011e-01 1.00476694e+00 1.06660926e+00
-4.21871006e-01 -3.39225799e-01 -8.66902471e-01 -6.22592747e-01
-3.03297788e-01 1.01018481e-01 9.29751635e-01 1.51039624e+00
6.47502959e-01 1.60366610e-01 2.81553507e-01 -2.74586342e-02
2.48514935e-01 3.98603469e-01 4.70566064e-01 -1.75581706e+00
-6.73501864e-02 -5.64759672e-01 -6.27075732e-01 -3.58540535e-01
8.14199150e-02 -1.03469646e+00 -3.94696414e-01 -2.01341605e+00
-7.19671845e-02 -7.10652590e-01 -1.83017239e-01 7.18352199e-01
1.40382156e-01 -6.75873011e-02 -1.27113983e-01 3.43323350e-02
-9.04435039e-01 -2.85858065e-01 1.01251066e+00 2.14790493e-01
-3.85214061e-01 -5.40261045e-02 -6.50074184e-01 7.88601398e-01
4.03521448e-01 -3.15668762e-01 -2.47411560e-02 -3.83527055e-02
9.09048915e-01 1.39482409e-01 -1.80721909e-01 -1.15287209e+00
6.36285543e-01 -3.55147719e-01 3.98578405e-01 -4.02654320e-01
-4.41636629e-02 -1.06195617e+00 3.96979302e-01 1.37090415e-01
-3.22321832e-01 5.63011467e-02 4.39200789e-01 1.42051816e-01
-5.20252645e-01 -8.15594673e-01 1.66152090e-01 -2.14417264e-01
-9.86199379e-01 -1.89749569e-01 -4.29319203e-01 2.22161267e-04
1.09338772e+00 -1.66205004e-01 -4.11482155e-01 1.95031106e-01
-1.13973165e+00 2.74393082e-01 4.59473193e-01 4.38274920e-01
5.34666851e-02 -1.32762074e+00 -1.77430466e-01 3.29138748e-02
1.30401164e-01 -2.95683090e-02 4.66684520e-01 5.31544387e-01
-5.77082157e-01 6.59627259e-01 -4.37771231e-01 -5.11981826e-03
-1.71758580e+00 2.65613258e-01 -4.03386801e-02 -6.11224353e-01
-2.64208943e-01 7.19752848e-01 -5.92352927e-01 -6.32815957e-01
3.21201295e-01 -1.71498388e-01 -1.07119787e+00 6.60263717e-01
3.61181170e-01 6.42216623e-01 4.57502604e-01 -5.12697935e-01
-3.35466683e-01 2.89135516e-01 2.53174335e-01 -6.19949214e-02
1.13303232e+00 -1.28212497e-01 -1.84835449e-01 1.90933004e-01
4.04716343e-01 6.94493830e-01 -2.02014014e-01 -3.55657816e-01
5.92520714e-01 -5.29487312e-01 -3.34591754e-02 -1.04389238e+00
-7.46577442e-01 2.65938550e-01 2.57612526e-01 6.27502441e-01
1.14206731e+00 1.31716803e-01 3.91964763e-01 1.37067586e-01
2.75833160e-01 -1.63144338e+00 -5.40404916e-01 4.73351270e-01
1.22778937e-01 -9.45172548e-01 1.35263145e-01 -7.95325577e-01
-3.70954692e-01 1.00932992e+00 5.67783654e-01 6.74627900e-01
5.45347929e-01 1.57051504e-01 -1.37541085e-01 -5.12698293e-01
-4.52840984e-01 -3.94519091e-01 6.39778495e-01 3.97352546e-01
7.50204206e-01 1.97695643e-01 -9.13068593e-01 8.53385985e-01
-4.49824572e-01 1.06508955e-01 5.07957697e-01 1.04838741e+00
-7.46474445e-01 -1.65278733e+00 -4.13612813e-01 4.77633744e-01
-8.63468051e-01 8.68101567e-02 -3.34872305e-01 8.36044133e-01
6.25201881e-01 6.83107972e-01 1.05229177e-01 -5.95662370e-02
7.67435908e-01 3.22589606e-01 3.77913713e-01 -8.27810287e-01
-1.01665473e+00 -5.79181015e-02 4.84097421e-01 -1.42851502e-01
-4.74864930e-01 -6.48233473e-01 -1.37314701e+00 -1.64084911e-01
-4.47856426e-01 8.52499127e-01 1.06129503e+00 8.57102394e-01
6.71768785e-01 9.57015395e-01 -8.46138746e-02 -1.77944712e-02
2.23633815e-02 -5.35192430e-01 -7.07312405e-01 4.72050279e-01
-3.45973730e-01 -7.91717291e-01 2.80916184e-01 1.28471062e-01] | [9.334519386291504, 8.428871154785156] |
8785887e-bac9-47c9-8b40-ffafba697f38 | a-discourse-on-metods-meta-optimized | 2202.02363 | null | https://arxiv.org/abs/2202.02363v3 | https://arxiv.org/pdf/2202.02363v3.pdf | Meta-Reinforcement Learning with Self-Modifying Networks | Deep Reinforcement Learning has demonstrated the potential of neural networks tuned with gradient descent for solving complex tasks in well-delimited environments. However, these neural systems are slow learners producing specialized agents with no mechanism to continue learning beyond their training curriculum. On the contrary, biological synaptic plasticity is persistent and manifold, and has been hypothesized to play a key role in executive functions such as working memory and cognitive flexibility, potentially supporting more efficient and generic learning abilities. Inspired by this, we propose to build networks with dynamic weights, able to continually perform self-reflexive modification as a function of their current synaptic state and action-reward feedback, rather than a fixed network configuration. The resulting model, MetODS (for Meta-Optimized Dynamical Synapses) is a broadly applicable meta-reinforcement learning system able to learn efficient and powerful control rules in the agent policy space. A single layer with dynamic synapses can perform one-shot learning, generalizes navigation principles to unseen environments and manifests a strong ability to learn adaptive motor policies. | ['Rufin VanRullen', 'Thomas Serre', 'Mathieu Chalvidal'] | 2022-02-04 | null | null | null | null | ['one-shot-learning'] | ['methodology'] | [ 1.34070426e-01 4.75816391e-02 -8.00823048e-02 1.64618716e-01
6.28828168e-01 -4.23749059e-01 9.49483633e-01 -1.27413705e-01
-9.89941895e-01 1.11295569e+00 -3.70872825e-01 -6.73757121e-02
-5.67710698e-01 -9.48696554e-01 -8.32708001e-01 -9.72292006e-01
-2.83616066e-01 3.79288703e-01 5.56641996e-01 -1.02226508e+00
5.54075360e-01 5.32356679e-01 -1.94481218e+00 -1.40128791e-01
9.45954561e-01 4.78046507e-01 1.00881362e+00 8.01965654e-01
-1.35890897e-02 6.45376265e-01 -3.93321872e-01 2.04442143e-01
1.79658324e-01 -4.90199417e-01 -5.12670696e-01 -4.19270754e-01
-9.64727253e-02 2.08882332e-01 -4.31404293e-01 8.51326048e-01
4.76496279e-01 5.52014112e-01 6.99278831e-01 -4.91868407e-01
-8.80434513e-01 5.24750352e-01 1.93334490e-01 6.56054556e-01
-9.98780727e-02 5.75162172e-01 6.13070726e-01 -6.71652436e-01
6.68344140e-01 9.84372318e-01 4.37100857e-01 1.24999583e+00
-1.08296871e+00 -4.49457109e-01 1.51279017e-01 3.21370393e-01
-6.40492618e-01 -1.53387249e-01 2.63597637e-01 -2.14587167e-01
1.26073885e+00 -9.22435224e-02 1.47001946e+00 1.35981035e+00
8.45409513e-01 3.49190176e-01 1.13412726e+00 -3.50748152e-01
7.04794288e-01 -9.95498747e-02 -4.57178950e-01 8.33271205e-01
3.00577998e-01 7.92333364e-01 -7.42427588e-01 3.31898957e-01
9.76970196e-01 7.29750022e-02 -7.00556263e-02 -4.10198420e-01
-1.21894646e+00 6.78422928e-01 5.95821619e-01 5.35728455e-01
-5.57186127e-01 5.77685118e-01 2.09457695e-01 6.49849892e-01
-3.09096515e-01 1.27235031e+00 -9.30324435e-01 -2.78004855e-01
-4.94697154e-01 3.04209411e-01 4.79879439e-01 1.35199219e-01
6.74364209e-01 6.81037009e-01 -1.70004629e-02 9.42163944e-01
1.43057466e-01 6.85544908e-01 1.28803349e+00 -1.16211796e+00
-2.62732595e-01 5.74121654e-01 -1.63857862e-01 -5.60247540e-01
-6.68815732e-01 -7.39725471e-01 -5.94162226e-01 1.08359027e+00
1.68880150e-01 -3.26471180e-01 -8.42007041e-01 1.95203435e+00
2.65839219e-01 -9.54967216e-02 1.68531463e-01 6.76058531e-01
1.82865798e-01 5.22265077e-01 1.29870906e-01 -8.16795751e-02
8.45470846e-01 -9.08349872e-01 -4.12864000e-01 -3.83032143e-01
1.51669487e-01 1.58899605e-01 9.67611551e-01 3.67626697e-01
-1.27637482e+00 -5.86588740e-01 -1.29732573e+00 4.72346187e-01
-8.90467405e-01 -6.97889686e-01 7.65145421e-01 4.13692683e-01
-1.41495037e+00 1.32575572e+00 -9.09158051e-01 -6.28098786e-01
4.60653454e-01 6.58499777e-01 -1.21083684e-01 6.67966962e-01
-1.40622175e+00 1.35960937e+00 9.46841657e-01 -8.25803876e-02
-1.16328156e+00 -6.60716832e-01 -3.36292803e-01 3.88434790e-02
2.59947389e-01 -1.18117726e+00 1.07942152e+00 -1.03085375e+00
-2.43584132e+00 5.54375708e-01 2.34216630e-01 -6.54142380e-01
7.96919763e-02 4.65939604e-02 -2.32697785e-01 1.28419753e-02
-4.08204406e-01 9.57297325e-01 1.02100086e+00 -9.50395107e-01
-6.14631832e-01 -4.37224746e-01 1.09382369e-01 3.60573500e-01
-6.17674232e-01 -5.00165224e-01 3.13139468e-01 -5.60370922e-01
-6.91529512e-02 -1.07161593e+00 -5.52781820e-01 8.83445069e-02
3.32541406e-01 -1.67403981e-01 4.93112057e-01 1.57908976e-01
8.76387596e-01 -1.74982858e+00 6.83101833e-01 1.53839275e-01
5.56591116e-02 7.59632587e-01 -3.48316640e-01 4.46822584e-01
2.02236041e-01 -1.31463408e-01 -2.14228883e-01 5.86527348e-01
-6.52222261e-02 4.73726273e-01 -2.35670488e-02 -6.45994097e-02
1.08434066e-01 1.24434376e+00 -1.28842866e+00 1.84083268e-01
7.72045553e-02 3.06360811e-01 -3.84007722e-01 5.04052863e-02
-4.69080031e-01 5.55257440e-01 -4.83259559e-01 2.27495492e-01
3.57621349e-02 -1.04756348e-01 2.29026869e-01 5.85373759e-01
-4.78219390e-01 -5.36160693e-02 -7.70957589e-01 1.73685861e+00
-6.24193966e-01 6.28351748e-01 -2.78572855e-03 -1.14954734e+00
9.91325319e-01 1.03784665e-01 3.51695538e-01 -1.45913625e+00
1.46593511e-01 3.89227509e-01 6.28951669e-01 -4.44390208e-01
3.01041842e-01 9.39026754e-03 4.05298769e-01 3.73893917e-01
4.71391618e-01 -3.10402066e-01 4.38158303e-01 -1.94252282e-01
1.20187664e+00 4.01023149e-01 1.66916162e-01 -5.14326632e-01
6.63421869e-01 -6.60660863e-02 2.85004854e-01 9.98208940e-01
-2.01468766e-01 -8.10629129e-03 -7.63079226e-02 -6.08769417e-01
-1.10177016e+00 -1.22731960e+00 -3.57276052e-02 1.60806561e+00
9.22948644e-02 1.99022710e-01 -7.16596484e-01 1.30060315e-01
-1.85962319e-02 5.23224175e-01 -8.92141938e-01 -7.48398364e-01
-8.27704072e-01 -7.44438708e-01 3.19070905e-01 1.73208609e-01
5.69085360e-01 -2.01171970e+00 -1.35377002e+00 7.19188273e-01
7.62836039e-01 -2.78005362e-01 -1.10055231e-01 8.90149176e-01
-1.18819082e+00 -1.11996508e+00 -6.94128633e-01 -8.87365341e-01
4.96851891e-01 1.24009192e-01 9.85881627e-01 4.84726846e-01
-4.91076916e-01 4.41105723e-01 -6.99269101e-02 -4.06198204e-01
-4.69271421e-01 2.09138498e-01 5.45579433e-01 -5.83102465e-01
-2.06298947e-01 -1.09371853e+00 -8.80238712e-01 2.00323626e-01
-1.03094149e+00 -1.43004106e-02 7.13336229e-01 1.39083040e+00
6.30446017e-01 -2.33078718e-01 9.40120876e-01 -5.88947415e-01
1.10628295e+00 -4.82306391e-01 -3.65793377e-01 2.92420417e-01
-1.15134656e+00 6.30584478e-01 9.79141176e-01 -6.17231011e-01
-1.07200515e+00 -2.96672702e-01 -1.65103137e-01 2.77464747e-01
-9.38304067e-02 2.28246674e-01 3.35071951e-01 -3.97711605e-01
9.53963280e-01 7.83214688e-01 1.22147746e-01 6.56061806e-03
5.38718939e-01 -1.25895321e-01 4.72444117e-01 -8.55592608e-01
6.40913844e-01 2.29332089e-01 2.23357603e-01 -7.41273463e-01
-4.30260390e-01 2.23761857e-01 -7.39145815e-01 -4.03834641e-01
8.29708636e-01 -1.85219452e-01 -1.04039001e+00 6.26640618e-01
-5.13511896e-01 -9.87174749e-01 -7.81464696e-01 3.23000014e-01
-9.91620958e-01 -2.75743783e-01 -5.64457119e-01 -5.76440275e-01
-4.24163729e-01 -6.92297578e-01 1.13756999e-01 9.10450876e-01
3.87647524e-02 -1.19790602e+00 6.54149175e-01 -5.59140086e-01
8.78945649e-01 3.81389968e-02 1.01737499e+00 -2.50675797e-01
-2.84380466e-01 4.18434709e-01 4.87527698e-01 9.92822498e-02
4.30599041e-03 9.14485902e-02 -6.27950907e-01 -3.31580073e-01
-2.10454136e-01 -5.04015684e-01 1.16212821e+00 2.41662383e-01
9.64510739e-01 -7.89625049e-02 -2.49340415e-01 6.44178867e-01
1.19963801e+00 5.54595590e-01 5.11576414e-01 8.97725701e-01
2.13662721e-03 3.87938619e-01 -2.36737784e-02 1.19883023e-01
3.49853672e-02 3.42544407e-01 7.32518256e-01 4.26802635e-01
-1.05312422e-01 -1.40320510e-01 4.52442050e-01 8.72836709e-01
-6.28544748e-01 1.97954312e-01 -5.15649259e-01 1.95706189e-01
-1.86838293e+00 -1.45228672e+00 4.10029322e-01 2.04694200e+00
8.91565800e-01 4.55117971e-01 -2.74058469e-02 -1.10955805e-01
4.37001497e-01 -1.16923064e-01 -1.35423088e+00 -7.69053400e-01
-3.94071490e-01 6.75080478e-01 4.21378911e-01 4.20253277e-02
-4.83013511e-01 1.20310426e+00 7.11112595e+00 6.01212323e-01
-1.33946657e+00 -1.11807130e-01 1.16288245e-01 -2.53757745e-01
-2.13021666e-01 -3.84750962e-01 -4.74302649e-01 5.29713571e-01
1.44612885e+00 -1.32729486e-01 1.22137833e+00 5.46882749e-01
2.13195086e-01 -4.41315435e-02 -6.16732419e-01 5.34594595e-01
-4.40331668e-01 -1.77326441e+00 -6.17788173e-02 9.92910862e-02
9.08756196e-01 2.87822396e-01 6.30829215e-01 7.40069151e-01
4.84099060e-01 -1.10342360e+00 6.54437661e-01 1.05225706e+00
4.17807430e-01 -7.82270312e-01 5.15280962e-02 6.49955690e-01
-8.05774570e-01 -6.98777139e-01 -6.94830358e-01 -2.86982358e-01
-1.78605586e-01 -1.14122070e-01 -3.52986932e-01 -1.64216757e-01
5.22955239e-01 5.20900249e-01 -5.56313872e-01 1.10993433e+00
-2.88953066e-01 2.11878061e-01 -3.27802524e-02 -7.99956203e-01
3.47777873e-01 -3.00676554e-01 5.09523213e-01 7.57490695e-01
4.72257257e-01 4.54048254e-02 -2.48826057e-01 7.31609046e-01
8.38755071e-02 -1.86193869e-01 -5.73760927e-01 -1.04761340e-01
3.72803926e-01 1.22014093e+00 -8.85605395e-01 -1.56653568e-01
2.37416163e-01 8.90926182e-01 7.51242816e-01 4.18733269e-01
-5.84331691e-01 -2.94741511e-01 8.87087762e-01 -1.42151490e-01
4.85778481e-01 -5.99390626e-01 2.42667664e-02 -8.53309691e-01
-5.98455846e-01 -7.19874859e-01 -7.18876347e-02 -5.26885629e-01
-8.32154095e-01 5.82143068e-01 -5.84001660e-01 -7.08504379e-01
-4.39648181e-01 -9.63321149e-01 -7.84073770e-01 5.74531734e-01
-1.54573357e+00 -4.53389883e-01 1.13814741e-01 8.51219535e-01
5.70384741e-01 -7.92548239e-01 1.04575753e+00 -2.65117168e-01
-5.71447074e-01 3.52694124e-01 6.79974914e-01 -4.40395802e-01
2.47708857e-01 -1.57267678e+00 1.89592749e-01 3.37393701e-01
2.12563276e-01 8.82920086e-01 7.64678180e-01 -3.59619051e-01
-1.65505993e+00 -7.89433122e-01 -3.72440070e-02 -3.97936329e-02
7.04953790e-01 -1.73424974e-01 -9.95215476e-01 -5.25646508e-02
2.97378719e-01 -2.46882085e-02 3.79893303e-01 7.05500096e-02
-1.92245677e-01 -1.12344503e-01 -1.09031987e+00 9.04766798e-01
1.35914171e+00 -2.53310025e-01 -6.48255587e-01 1.54641688e-01
4.92206126e-01 -3.14660639e-01 -5.14537513e-01 8.69781822e-02
8.96251202e-01 -1.02698576e+00 9.42236662e-01 -1.09736705e+00
1.85496509e-01 -1.85132489e-01 3.09438497e-01 -1.86152565e+00
-7.84523964e-01 -6.39971077e-01 -5.24492562e-01 2.73362041e-01
5.39926589e-01 -8.23570609e-01 7.17389643e-01 8.99674371e-02
-4.92718637e-01 -1.00801504e+00 -9.27619040e-01 -9.30570185e-01
4.72993791e-01 5.27045988e-02 3.63483757e-01 5.35652518e-01
6.23389743e-02 -1.88351106e-02 4.84758392e-02 -2.87969232e-01
2.38664255e-01 -2.91907132e-01 2.75475293e-01 -1.31426883e+00
-5.19400895e-01 -1.14006519e+00 -1.04582958e-01 -6.69064820e-01
2.62348324e-01 -9.02608514e-01 1.46802649e-01 -1.45144188e+00
-1.88854262e-01 -1.68689713e-01 -8.05850029e-01 2.28230938e-01
-6.73042610e-02 -2.42490340e-02 1.03161246e-01 7.47591406e-02
-6.09999418e-01 6.58390582e-01 1.69140458e+00 1.09806089e-02
-3.59616637e-01 2.45741964e-03 -6.48665547e-01 6.01464570e-01
1.09870374e+00 -4.41967070e-01 -5.45127928e-01 -3.35219204e-01
7.28475630e-01 -2.53658324e-01 9.55394581e-02 -1.27133644e+00
3.85066599e-01 -5.37423372e-01 6.33770287e-01 2.21850112e-01
1.86224833e-01 -4.51064020e-01 -1.25563577e-01 1.25446582e+00
-5.55037916e-01 2.31054679e-01 8.81802216e-02 8.20430875e-01
3.14178854e-01 -4.43264276e-01 8.41648817e-01 -5.53985655e-01
-1.03059399e+00 3.67530346e-01 -1.15880871e+00 1.57652020e-01
9.89694834e-01 -6.76720679e-01 -4.44055855e-01 1.58356667e-01
-9.04552221e-01 2.78417498e-01 2.66269088e-01 6.60514951e-01
6.74631596e-01 -1.03083622e+00 -3.86178315e-01 6.01986237e-02
-3.00558627e-01 -5.56708992e-01 2.54882365e-01 3.53308499e-01
-5.26365459e-01 4.01965737e-01 -1.13049555e+00 -1.33011773e-01
-3.84874701e-01 5.25142789e-01 8.11770260e-01 -1.06024265e-01
-5.69206297e-01 8.98334801e-01 -1.79464102e-01 -5.54900467e-01
-1.37285173e-01 -1.79512143e-01 -6.23329937e-01 -1.62849128e-01
6.31351411e-01 4.05119121e-01 -2.75189020e-02 3.69548984e-02
-2.18891315e-02 4.44245279e-01 1.05444357e-01 -1.14410810e-01
1.63469076e+00 4.51338179e-02 -8.30364302e-02 5.14789999e-01
4.31683332e-01 -5.24352431e-01 -1.65888643e+00 5.83271645e-02
2.58348376e-01 9.06217620e-02 -2.54119277e-01 -1.01633739e+00
-8.77364516e-01 7.97687471e-01 8.22160542e-01 2.94110090e-01
7.55596340e-01 -3.76087278e-01 7.58709669e-01 1.26464188e+00
6.69752598e-01 -1.60874236e+00 8.29887331e-01 1.20686424e+00
7.51847863e-01 -9.91723716e-01 -3.64374399e-01 8.07973564e-01
-2.04195216e-01 1.63832986e+00 8.90364230e-01 -5.99181116e-01
5.08005679e-01 8.14154595e-02 -2.75407583e-01 -8.97961855e-02
-1.17358327e+00 -3.08265328e-01 5.76989837e-02 1.17497361e+00
3.81083250e-01 -1.37958109e-01 -4.26911116e-01 -4.53474298e-02
-1.87953025e-01 -1.96007863e-01 5.04980922e-01 8.26039732e-01
-1.34448719e+00 -1.28904331e+00 7.09621385e-02 4.11700100e-01
1.92969684e-02 1.26376674e-01 2.75499448e-02 7.23095655e-01
3.07524115e-01 4.49038118e-01 -1.61510110e-01 -1.59449235e-01
1.92818195e-01 9.56273973e-02 8.59358966e-01 -7.17543602e-01
-8.82394433e-01 -6.00575149e-01 -6.00919068e-01 -4.12697226e-01
-1.92252964e-01 -7.09170640e-01 -1.71917450e+00 -1.27228931e-01
5.65904006e-02 9.10643786e-02 8.50815058e-01 9.14273858e-01
3.09533030e-01 1.00701511e+00 5.17516553e-01 -1.01303434e+00
-7.70042956e-01 -6.81836784e-01 -5.25966108e-01 7.68563151e-02
1.19124867e-01 -9.87705231e-01 1.75411813e-02 -2.31341362e-01] | [4.255012035369873, 1.5443949699401855] |
b980ede0-759d-4b31-9fae-20d843560789 | video-question-answering-with-iterative-video | 2208.00934 | null | https://arxiv.org/abs/2208.00934v1 | https://arxiv.org/pdf/2208.00934v1.pdf | Video Question Answering with Iterative Video-Text Co-Tokenization | Video question answering is a challenging task that requires understanding jointly the language input, the visual information in individual video frames, as well as the temporal information about the events occurring in the video. In this paper, we propose a novel multi-stream video encoder for video question answering that uses multiple video inputs and a new video-text iterative co-tokenization approach to answer a variety of questions related to videos. We experimentally evaluate the model on several datasets, such as MSRVTT-QA, MSVD-QA, IVQA, outperforming the previous state-of-the-art by large margins. Simultaneously, our model reduces the required GFLOPs from 150-360 to only 67, producing a highly efficient video question answering model. | ['Anelia Angelova', 'Michael S. Ryoo', 'Weicheng Kuo', 'Kairo Morton', 'AJ Piergiovanni'] | 2022-08-01 | null | null | null | null | ['video-question-answering'] | ['computer-vision'] | [ 1.24300167e-01 -3.98220807e-01 -1.96429327e-01 -2.62700111e-01
-1.09340358e+00 -7.27058768e-01 2.86084950e-01 9.72936600e-02
-7.30924606e-01 5.11466324e-01 3.04031640e-01 -5.69236100e-01
2.33444646e-01 -2.81474024e-01 -9.05182481e-01 -2.17662588e-01
-1.45480447e-02 4.49552909e-02 7.25997210e-01 -2.27236189e-03
1.58507735e-01 -1.89281315e-01 -1.41074038e+00 6.43634617e-01
5.12457907e-01 1.17520642e+00 4.18552190e-01 1.41844201e+00
-3.94288391e-01 1.80566406e+00 -6.00756943e-01 -5.33604503e-01
-2.18731210e-01 -5.32264590e-01 -1.08558798e+00 2.41487831e-01
8.59188318e-01 -7.83846557e-01 -1.05266702e+00 7.38652587e-01
1.14751056e-01 2.39582285e-01 2.46313378e-01 -1.22726870e+00
-7.15624273e-01 8.93465057e-02 -3.38862568e-01 8.09864759e-01
8.24160159e-01 2.63244122e-01 1.16842377e+00 -7.58772492e-01
6.66551232e-01 1.21771538e+00 -2.19028704e-02 3.72882515e-01
-3.51960152e-01 -3.90992850e-01 3.45053017e-01 1.07896209e+00
-1.26690972e+00 -4.16568249e-01 2.65619755e-01 -2.37372383e-01
1.08241987e+00 2.26186723e-01 3.74860644e-01 7.93437362e-01
-5.03355451e-02 1.15664840e+00 3.78761023e-01 -2.33749926e-01
1.05746612e-01 -4.42280829e-01 3.77181806e-02 1.16190028e+00
-2.90899932e-01 -5.53720534e-01 -7.45603025e-01 -1.20582998e-01
5.13233125e-01 -1.20322825e-02 -3.15589130e-01 1.79108620e-01
-1.25161135e+00 6.88256204e-01 -6.17041923e-02 7.18039945e-02
-2.68916696e-01 6.10885322e-01 7.47008085e-01 5.30616045e-01
2.34636769e-01 -1.90797336e-02 -4.40144449e-01 -5.28795779e-01
-8.64703476e-01 2.13903502e-01 8.01528215e-01 1.08307850e+00
7.32801735e-01 -2.55205091e-02 -5.78189671e-01 3.18548322e-01
2.08661854e-01 6.10663831e-01 1.35160416e-01 -1.42329383e+00
9.86582041e-01 3.08162272e-01 1.83457494e-01 -9.03206050e-01
8.48169327e-02 2.87128001e-01 -3.58876050e-01 -7.86192298e-01
4.51236993e-01 -2.67024547e-01 -1.04531336e+00 1.42998326e+00
3.45684320e-01 5.77403367e-01 1.34530708e-01 9.72448707e-01
1.02926433e+00 1.35860407e+00 1.40182540e-01 -2.94360399e-01
1.72025681e+00 -1.35332870e+00 -9.71838474e-01 -1.73446208e-01
5.43335021e-01 -7.37787366e-01 9.91048396e-01 2.65422076e-01
-1.21042013e+00 -6.92984700e-01 -6.65426254e-01 -5.18226624e-01
6.78370446e-02 1.53739899e-01 3.62674534e-01 3.28377366e-01
-1.04159164e+00 -9.40815508e-02 -6.87654316e-01 -1.44856766e-01
4.03520316e-01 2.06424534e-01 -1.98913500e-01 -6.51798785e-01
-1.30969608e+00 3.69257689e-01 2.55323172e-01 -1.24380901e-01
-1.26097822e+00 -5.96797705e-01 -9.83045220e-01 1.30719513e-01
9.68269944e-01 -7.22586274e-01 1.60031796e+00 -9.44457412e-01
-1.47805035e+00 4.47915256e-01 -8.97674382e-01 -5.86060822e-01
8.98611546e-02 -4.84993130e-01 -4.95833993e-01 1.20276856e+00
3.35890353e-02 7.01193869e-01 1.10502231e+00 -5.90810895e-01
-9.01156187e-01 -7.04975724e-02 6.92999780e-01 2.95722365e-01
-1.82776406e-01 4.69250649e-01 -1.34398353e+00 -4.15358096e-01
-5.12557387e-01 -7.18670726e-01 -6.54012263e-02 1.03344776e-01
5.01710624e-02 -3.87763560e-01 1.10605264e+00 -1.06387150e+00
1.25487518e+00 -2.16421723e+00 2.25832269e-01 -2.59845644e-01
1.55478060e-01 2.52776295e-01 -3.57047081e-01 3.36898744e-01
2.62506872e-01 8.91283527e-03 5.98167367e-02 -1.27678350e-01
-1.74776733e-01 4.49245006e-01 -3.65826666e-01 3.71360004e-01
3.99923712e-01 1.13221622e+00 -9.99296069e-01 -8.92935812e-01
9.57195237e-02 2.50462651e-01 -6.63962305e-01 5.33947170e-01
-7.75390685e-01 2.07492277e-01 -5.29421866e-01 7.14240730e-01
2.26671606e-01 -5.23077250e-01 3.96678410e-02 -1.47016466e-01
1.40121236e-01 1.03792623e-01 -7.36599922e-01 1.93247473e+00
-2.91435599e-01 1.05655503e+00 -6.38886020e-02 -1.00317991e+00
3.05322140e-01 8.50243211e-01 5.04801929e-01 -9.83475327e-01
-9.97121260e-02 -8.03266689e-02 -4.84241635e-01 -1.21218538e+00
6.76139891e-01 2.28109866e-01 -7.00067356e-02 3.37183774e-01
4.23099011e-01 2.96620727e-01 7.35843837e-01 7.08243430e-01
1.21549666e+00 3.92851979e-02 8.06706473e-02 3.36835653e-01
8.05426836e-01 -1.35224219e-02 1.76882684e-01 6.98924661e-01
-2.80348003e-01 5.15810728e-01 7.73441672e-01 -2.92058527e-01
-8.99155676e-01 -8.06564510e-01 5.99586487e-01 1.27567112e+00
1.77255481e-01 -7.78558731e-01 -7.90404677e-01 -7.59773254e-01
-2.45604649e-01 3.52618873e-01 -3.56174171e-01 5.67666925e-02
-8.14170003e-01 -5.18350527e-02 6.15327656e-01 5.74026287e-01
5.21797359e-01 -7.60872781e-01 -5.52889347e-01 9.00648087e-02
-7.79188454e-01 -1.96200442e+00 -9.61552203e-01 -4.63536888e-01
-5.79836249e-01 -1.30194056e+00 -7.73829222e-01 -1.01181936e+00
4.13622797e-01 5.82300305e-01 1.27194357e+00 2.25944474e-01
-2.24446863e-01 8.83049190e-01 -7.19871342e-01 5.17487675e-02
-6.78163469e-02 -1.32438689e-01 -4.36642259e-01 2.06516743e-01
2.57955819e-01 1.90580726e-01 -5.38848758e-01 2.59573579e-01
-1.24539876e+00 -1.89836785e-01 3.26517105e-01 5.90832472e-01
6.99865401e-01 -5.65177724e-02 4.36041921e-01 -6.09145582e-01
2.27133796e-01 -7.77951837e-01 -3.86172473e-01 6.12236738e-01
2.04753622e-01 1.15821786e-01 7.29267895e-01 -4.90643859e-01
-1.05843270e+00 -1.30419210e-01 -2.22826466e-01 -8.68306994e-01
-1.24067597e-01 5.28126955e-01 2.98194345e-02 6.70151561e-02
1.83222778e-02 4.25364047e-01 -1.31242886e-01 -1.46256760e-01
4.58060145e-01 4.01095837e-01 9.09966409e-01 -3.63446057e-01
5.69316685e-01 5.21713674e-01 -2.35768452e-01 -1.00544691e+00
-9.05422390e-01 -9.76791203e-01 -2.04318345e-01 -4.22134995e-01
1.21695602e+00 -1.22028530e+00 -9.83973980e-01 2.30657533e-01
-1.41595662e+00 -2.34757379e-01 -2.79261153e-02 4.56390887e-01
-5.15569270e-01 9.20574129e-01 -8.37977946e-01 -5.88732719e-01
-3.52623492e-01 -1.11688614e+00 1.05922139e+00 1.26933187e-01
7.73228854e-02 -7.76266754e-01 -2.65225172e-01 8.77241194e-01
1.09258853e-01 -3.24233949e-01 8.80998731e-01 -3.78555536e-01
-1.07002175e+00 3.09906732e-02 -5.22831798e-01 3.14800709e-01
-9.01410878e-02 -4.00330946e-02 -4.64734703e-01 -2.00729147e-01
2.39645746e-02 -6.07460201e-01 1.04473901e+00 2.05199182e-01
1.51277339e+00 -4.00862485e-01 1.50007322e-01 3.81762058e-01
1.34305680e+00 3.12383384e-01 7.59788334e-01 -1.17233852e-02
7.32035518e-01 1.78724185e-01 8.68984759e-01 4.47312713e-01
8.54734719e-01 4.91787910e-01 5.33901274e-01 1.15400799e-01
-4.66133244e-02 -2.07832441e-01 7.53814518e-01 1.11125851e+00
1.50215432e-01 -7.79304981e-01 -6.19173288e-01 8.20779085e-01
-1.98359859e+00 -1.13872194e+00 -1.93193004e-01 1.63715136e+00
4.51462507e-01 -2.80755222e-01 1.03900753e-01 -5.89699820e-02
5.01207054e-01 5.15822053e-01 -4.69260246e-01 -2.08217308e-01
7.36075938e-02 2.49903843e-01 3.98049831e-01 4.60826010e-01
-1.16002846e+00 9.62510169e-01 7.06323910e+00 7.32551575e-01
-6.40416682e-01 2.24574134e-01 4.41588998e-01 -3.79101425e-01
-1.77570388e-01 1.23539204e-02 -6.25739872e-01 4.60495859e-01
1.26188135e+00 -1.07420504e-01 4.48575020e-01 4.24261630e-01
9.27396938e-02 -3.77651989e-01 -8.73079956e-01 1.19344413e+00
7.57786691e-01 -1.56330442e+00 3.57517689e-01 -6.06752753e-01
5.97856343e-01 3.95170832e-03 -1.21525444e-01 3.11388463e-01
-1.74906045e-01 -7.59472787e-01 8.14826071e-01 4.11069334e-01
8.95804226e-01 -6.31677270e-01 6.29534364e-01 1.28410980e-01
-1.52512920e+00 -8.84625912e-02 -2.56114513e-01 -1.90992951e-02
6.05337083e-01 1.76367536e-01 -5.20687759e-01 7.17600405e-01
8.59582901e-01 7.22197473e-01 -5.26185811e-01 7.72094727e-01
-9.90293995e-02 8.74801219e-01 -1.39689669e-01 -1.18136644e-01
4.76930380e-01 2.01940626e-01 2.47322053e-01 1.22460103e+00
2.89312541e-01 5.30767083e-01 2.92322755e-01 2.06971601e-01
-3.13623309e-01 2.18406972e-02 -1.13642424e-01 -3.83380622e-01
-8.45764484e-03 8.65348816e-01 -3.73160869e-01 -6.38503730e-01
-1.03991592e+00 1.24234593e+00 1.09795406e-01 6.40597165e-01
-1.13774741e+00 -2.36568749e-01 6.25787437e-01 -2.44891286e-01
9.32252884e-01 -5.57052791e-01 6.84558988e-01 -1.58925843e+00
3.34965974e-01 -1.10706139e+00 7.98216164e-01 -1.02918029e+00
-7.88058877e-01 4.79667068e-01 -6.11607358e-02 -9.58993316e-01
-3.58965009e-01 -6.05079770e-01 -1.74666777e-01 3.51041853e-01
-2.00085926e+00 -7.55931556e-01 -3.85125905e-01 1.10356998e+00
1.00874960e+00 -3.32420878e-02 4.00526375e-01 6.87583506e-01
-3.98505807e-01 3.72429192e-01 -1.58920541e-01 3.44665259e-01
7.65453696e-01 -8.42079520e-01 4.11573857e-01 9.86112475e-01
4.46803272e-01 1.07666194e-01 3.66197914e-01 -2.55566239e-01
-2.03328133e+00 -1.04146934e+00 9.15761709e-01 -2.65546679e-01
7.80717552e-01 -7.70932883e-02 -8.81292224e-01 6.81913733e-01
5.33766270e-01 1.25175193e-01 6.66152060e-01 -3.92722040e-01
-4.60469007e-01 -3.73026580e-02 -6.27918124e-01 4.82870579e-01
7.94859707e-01 -1.10528421e+00 -5.80678582e-01 3.28095347e-01
1.21048534e+00 -4.98922974e-01 -8.09099734e-01 8.26497301e-02
2.98826605e-01 -5.50873637e-01 9.29047644e-01 -8.61606836e-01
5.76299965e-01 -3.39001834e-01 -3.93008232e-01 -5.54949641e-01
1.28632160e-02 -6.97859704e-01 -7.08667457e-01 9.19676721e-01
1.46503806e-01 1.05768740e-01 6.49689913e-01 2.70103127e-01
6.99569704e-03 -5.61470449e-01 -1.02376080e+00 -3.65737706e-01
-5.71696937e-01 -7.39355147e-01 2.02583417e-01 4.53517646e-01
-1.30654410e-01 4.98972058e-01 -7.44139016e-01 2.77781963e-01
3.03958803e-01 6.82238191e-02 6.67104721e-01 -4.31043297e-01
-4.59002435e-01 9.29510966e-02 -3.84642571e-01 -1.91411352e+00
2.63029575e-01 -4.70681161e-01 3.41548398e-02 -1.59745693e+00
3.90558869e-01 4.71776932e-01 -7.02173337e-02 3.39427948e-01
-4.76437181e-01 1.44521698e-01 4.21680927e-01 -2.34665759e-02
-1.57880247e+00 4.91186619e-01 1.25543034e+00 -2.03419179e-01
3.01148891e-01 -4.32680011e-01 -3.47594142e-01 4.94738102e-01
3.51749927e-01 -2.60935903e-01 -5.13484955e-01 -1.07809937e+00
3.16514224e-01 6.55711651e-01 3.95529479e-01 -8.52014482e-01
4.31297719e-01 -1.15889564e-01 4.01229113e-02 -6.66449189e-01
4.97015595e-01 -6.61011934e-01 -4.52389717e-01 1.57696232e-01
-4.14172173e-01 4.51153219e-01 3.71641129e-01 8.59628618e-01
-6.96215212e-01 -1.68212414e-01 1.29998118e-01 -1.39433876e-01
-1.33132112e+00 7.10403264e-01 -6.98901355e-01 5.69230378e-01
1.10341692e+00 1.28241703e-01 -5.76724827e-01 -8.87792766e-01
-3.67905527e-01 7.33425081e-01 1.36609571e-02 7.10192502e-01
9.48691189e-01 -1.08705544e+00 -6.96233749e-01 -1.84024274e-01
1.05210416e-01 -2.45151520e-01 7.15903401e-01 7.29488075e-01
-7.62088120e-01 7.85247922e-01 1.82959139e-01 -6.52477443e-01
-1.52649164e+00 7.09649563e-01 9.07144696e-02 -1.42483577e-01
-3.75558019e-01 7.19009101e-01 1.45109907e-01 2.90967435e-01
3.74587297e-01 -4.76595104e-01 -3.18608165e-01 -4.04202789e-02
7.57275879e-01 2.20206559e-01 -2.45253950e-01 -6.19527102e-01
-2.13779703e-01 4.66521919e-01 -3.13139074e-02 -7.41878226e-02
8.14594030e-01 -3.37701946e-01 -1.88979972e-02 2.79147923e-01
1.52485573e+00 -3.00162733e-01 -1.28954208e+00 -4.22758430e-01
-1.42624229e-01 -4.98303026e-01 -5.47158159e-02 -2.31408030e-01
-1.08792675e+00 8.18584859e-01 2.04601198e-01 -3.78697515e-02
1.32336950e+00 3.04905951e-01 1.27822733e+00 5.69095671e-01
1.55149445e-01 -8.95312905e-01 6.86783075e-01 7.00622022e-01
5.07667363e-01 -1.19011629e+00 -5.25311790e-02 -3.64972264e-01
-9.17152107e-01 1.01574016e+00 4.50839221e-01 9.06296521e-02
3.11791927e-01 -1.35715798e-01 -8.55202600e-02 -1.55075625e-01
-1.35638368e+00 -3.85134697e-01 3.63724321e-01 1.76189378e-01
7.49846250e-02 -3.73437613e-01 -3.17884050e-02 2.62594342e-01
4.66179907e-01 3.41115564e-01 4.27908212e-01 1.03369963e+00
-4.31714982e-01 -8.96042705e-01 -2.36701250e-01 3.13180804e-01
-9.47953999e-01 -1.63270846e-01 3.30530317e-03 4.90257353e-01
-1.10062703e-01 1.21762061e+00 3.15906614e-01 -2.85754591e-01
1.47995725e-01 7.26219416e-02 4.73332763e-01 -3.69145006e-01
-3.43052596e-01 -1.49153978e-01 2.66686797e-01 -7.43532538e-01
-8.15718293e-01 -6.07876420e-01 -1.33356702e+00 -2.04969689e-01
-1.11546859e-01 2.11002767e-01 2.83310533e-01 1.37972808e+00
4.71092373e-01 4.21498418e-01 5.52807331e-01 -2.61384994e-01
-2.34645888e-01 -5.27892232e-01 -6.66228458e-02 4.37829345e-01
6.22248650e-01 -1.76230267e-01 -1.97829336e-01 5.38265884e-01] | [10.318291664123535, 0.9103173613548279] |
3e9209f7-1a52-401c-b77f-2d47ac521073 | latent-dynamics-networks-ldnets-learning-the | 2305.00094 | null | https://arxiv.org/abs/2305.00094v1 | https://arxiv.org/pdf/2305.00094v1.pdf | Latent Dynamics Networks (LDNets): learning the intrinsic dynamics of spatio-temporal processes | Predicting the evolution of systems that exhibit spatio-temporal dynamics in response to external stimuli is a key enabling technology fostering scientific innovation. Traditional equations-based approaches leverage first principles to yield predictions through the numerical approximation of high-dimensional systems of differential equations, thus calling for large-scale parallel computing platforms and requiring large computational costs. Data-driven approaches, instead, enable the description of systems evolution in low-dimensional latent spaces, by leveraging dimensionality reduction and deep learning algorithms. We propose a novel architecture, named Latent Dynamics Network (LDNet), which is able to discover low-dimensional intrinsic dynamics of possibly non-Markovian dynamical systems, thus predicting the time evolution of space-dependent fields in response to external inputs. Unlike popular approaches, in which the latent representation of the solution manifold is learned by means of auto-encoders that map a high-dimensional discretization of the system state into itself, LDNets automatically discover a low-dimensional manifold while learning the latent dynamics, without ever operating in the high-dimensional space. Furthermore, LDNets are meshless algorithms that do not reconstruct the output on a predetermined grid of points, but rather at any point of the domain, thus enabling weight-sharing across query-points. These features make LDNets lightweight and easy-to-train, with excellent accuracy and generalization properties, even in time-extrapolation regimes. We validate our method on several test cases and we show that, for a challenging highly-nonlinear problem, LDNets outperform state-of-the-art methods in terms of accuracy (normalized error 5 times smaller), by employing a dramatically smaller number of trainable parameters (more than 10 times fewer). | ['Alfio Quarteroni', "Luca Dede'", 'Matteo Salvador', 'Stefano Pagani', 'Francesco Regazzoni'] | 2023-04-28 | null | null | null | null | ['dimensionality-reduction'] | ['methodology'] | [-2.64698982e-01 -2.97719270e-01 1.05302423e-01 4.35501695e-01
-3.72590333e-01 -7.44116008e-01 8.66606295e-01 -2.08023358e-02
-2.51239270e-01 7.87256181e-01 -1.95085242e-01 -2.09094584e-01
-4.74257052e-01 -8.11987460e-01 -7.02226520e-01 -1.06595910e+00
-4.34089094e-01 7.63978660e-01 1.74550042e-01 -1.09052643e-01
2.64046211e-02 9.75493908e-01 -1.45352471e+00 -8.35062191e-02
5.85277677e-01 9.96348202e-01 -1.04798481e-01 8.48861396e-01
1.68168202e-01 5.03011703e-01 -1.04267769e-01 1.57540172e-01
2.65362889e-01 -3.88112009e-01 -7.57833660e-01 -1.28908247e-01
2.54324496e-01 -1.64392397e-01 -7.07727075e-01 6.08627915e-01
2.71789700e-01 6.98550940e-02 9.08041775e-01 -8.86197209e-01
-5.43943644e-01 -3.60214263e-02 8.89275670e-02 1.94215074e-01
1.03351073e-02 6.32298529e-01 8.26572120e-01 -8.65878344e-01
8.37566435e-01 9.77817237e-01 1.06673372e+00 4.50310171e-01
-2.11541581e+00 -3.50860745e-01 -2.80699916e-02 -1.11713475e-02
-1.52943850e+00 -3.27568203e-01 7.66349077e-01 -8.24525356e-01
1.13590777e+00 1.06741384e-01 9.36966777e-01 1.14421713e+00
7.54196823e-01 2.27229342e-01 1.07872951e+00 2.57829558e-02
6.22300982e-01 -1.84346020e-01 -2.65941530e-01 6.81649029e-01
-2.91469023e-02 2.76076674e-01 -4.90193367e-01 -4.67414886e-01
1.14975178e+00 1.54100239e-01 -1.03819802e-01 -7.63524830e-01
-1.34270692e+00 6.85201406e-01 3.54506642e-01 3.90088528e-01
-5.47269106e-01 4.52011049e-01 2.55772382e-01 5.77539980e-01
6.81783795e-01 7.43770242e-01 -7.01980591e-01 -2.51341045e-01
-1.00443327e+00 4.99456942e-01 9.34467077e-01 4.90365505e-01
8.74250531e-01 2.23404020e-01 1.01217933e-01 1.03991129e-01
-1.08211711e-01 4.24022913e-01 5.45277596e-01 -1.05896962e+00
-2.74405833e-02 7.66073346e-01 2.84071416e-01 -8.96676362e-01
-5.91982067e-01 -5.61534643e-01 -1.43903661e+00 2.37667844e-01
3.84618700e-01 -1.91966891e-01 -8.30896199e-01 1.73341727e+00
4.72461909e-01 3.60633790e-01 -6.93737995e-03 8.35602641e-01
-1.14729024e-01 9.68353748e-01 -2.93674946e-01 -4.66181219e-01
7.78423429e-01 -4.92301881e-01 -3.59863698e-01 3.16781014e-01
8.39144289e-01 -2.01075643e-01 1.17294240e+00 2.68331140e-01
-1.18773425e+00 -5.13885021e-01 -7.25201070e-01 2.05860808e-01
-4.50371802e-01 -1.81596026e-01 5.64548731e-01 -1.27306521e-01
-1.34551537e+00 1.27061319e+00 -1.26724374e+00 -2.58890837e-01
2.96892554e-01 5.56374907e-01 -4.10011947e-01 4.56014693e-01
-9.83936906e-01 7.74840415e-01 -3.18456963e-02 -3.29110562e-03
-1.43269455e+00 -1.27116585e+00 -4.07070249e-01 1.82097688e-01
-3.22353281e-02 -7.88166165e-01 8.31117511e-01 -5.87574661e-01
-1.64579177e+00 4.51058596e-01 -1.80003315e-01 -5.73784053e-01
5.84232867e-01 5.78492507e-02 -3.28023642e-01 1.47028521e-01
-1.30552337e-01 3.43917847e-01 1.04576349e+00 -9.20140326e-01
-3.65509465e-02 -2.14232221e-01 -1.39470339e-01 -1.34165555e-01
-5.79497397e-01 -7.54233539e-01 -5.11082541e-03 -4.38542664e-01
2.79435486e-01 -1.14360666e+00 -5.88091016e-01 3.42370927e-01
-1.01393148e-01 -1.75540566e-01 1.04037344e+00 -2.77525336e-01
1.19306314e+00 -1.88125205e+00 7.61490881e-01 -7.15858936e-02
7.13488579e-01 2.66532749e-01 1.13616467e-01 8.31217945e-01
-7.53898621e-02 1.08150311e-01 -4.43485916e-01 -3.83162230e-01
7.07615986e-02 2.61795402e-01 -7.47000337e-01 6.66447401e-01
3.32181334e-01 1.11064672e+00 -7.63909698e-01 -1.21353365e-01
1.76164165e-01 6.09917700e-01 -6.10668898e-01 1.90848649e-01
-3.53353202e-01 8.68409216e-01 -4.75187093e-01 4.11185808e-02
2.92574763e-01 -7.01902330e-01 3.80239300e-02 -1.62608717e-02
-3.95790517e-01 1.22842781e-01 -1.02953017e+00 1.77708662e+00
-7.76287138e-01 6.25214100e-01 -1.72699522e-02 -1.44229066e+00
7.45524764e-01 5.66485286e-01 9.26439881e-01 -6.97913885e-01
-8.32262784e-02 3.70839119e-01 -1.62630841e-01 -3.01810652e-01
1.41129140e-02 -1.87749892e-01 -2.57985145e-01 6.68080330e-01
2.12282017e-01 -1.54198408e-01 7.76710659e-02 3.99721973e-02
1.44349575e+00 1.46854326e-01 -3.61546427e-02 -6.09778404e-01
5.43346286e-01 8.29682574e-02 3.90138090e-01 6.39154434e-01
1.92307785e-01 1.22444391e-01 7.24582314e-01 -1.05190611e+00
-1.60765839e+00 -1.18721950e+00 -5.31812966e-01 4.60960269e-01
-2.34152097e-02 -3.71231049e-01 -6.39517069e-01 -1.48704350e-01
2.45049387e-01 2.33167037e-01 -7.96046615e-01 -3.58962506e-01
-7.18575239e-01 -6.46270871e-01 2.87348092e-01 2.49350488e-01
8.90636146e-02 -9.11899090e-01 -7.27650881e-01 5.54785550e-01
3.83177668e-01 -9.40511346e-01 -1.29873194e-02 2.33516291e-01
-1.21669042e+00 -7.55491257e-01 -7.37331033e-01 -4.21892285e-01
7.00006902e-01 -3.60699654e-01 1.00190747e+00 -1.04756422e-01
-6.85759306e-01 3.84110898e-01 2.90845066e-01 3.14704925e-01
-4.25415754e-01 3.89007151e-01 7.16978788e-01 4.48138937e-02
-9.04509798e-02 -1.35961258e+00 -6.76179409e-01 2.62953579e-01
-9.28681493e-01 2.30179399e-01 3.23662728e-01 9.09534752e-01
8.03882658e-01 8.10291767e-02 5.30011892e-01 -4.36830252e-01
4.97634947e-01 -5.06526113e-01 -1.06453240e+00 2.87272432e-03
-7.33615696e-01 5.37463903e-01 1.37122023e+00 -8.53761196e-01
-4.76660311e-01 2.27406368e-01 2.04649135e-01 -7.87666500e-01
-1.10629581e-01 2.51485318e-01 4.77210879e-01 -2.41331831e-01
6.65674806e-01 6.05218470e-01 -3.74482153e-03 -6.49284065e-01
3.35936129e-01 9.95966047e-02 3.19676578e-01 -7.12559342e-01
9.91707206e-01 6.53557241e-01 6.10852718e-01 -7.71966577e-01
-5.86029351e-01 -1.59327552e-01 -1.01480198e+00 -1.85655311e-01
6.29096031e-01 -6.44137681e-01 -1.11823475e+00 5.34073472e-01
-1.03245473e+00 -6.05780602e-01 -6.94412649e-01 3.26219410e-01
-8.80845964e-01 -2.07467735e-01 -7.96298683e-01 -7.15134919e-01
-1.13101192e-01 -8.61919761e-01 1.08751583e+00 -1.87787518e-01
-2.65215605e-01 -1.32755339e+00 5.45314908e-01 -5.76492727e-01
6.77339375e-01 5.44693649e-01 1.19222343e+00 -1.32033899e-01
-6.47251189e-01 -2.23270699e-01 9.54799131e-02 -1.73745342e-02
-4.26137224e-02 1.04902247e-02 -7.26698339e-01 -5.51245630e-01
2.09785342e-01 -2.12333366e-01 6.80284262e-01 4.16366428e-01
1.17553067e+00 -6.17366374e-01 -6.14347756e-01 9.21637118e-01
1.54833710e+00 -3.48117441e-01 1.89819649e-01 -1.80038363e-01
7.48396993e-01 4.04078007e-01 -1.88118339e-01 4.91291702e-01
1.78195462e-01 6.76022291e-01 3.21394682e-01 -1.10821277e-01
1.46212265e-01 -2.05098823e-01 3.30533534e-01 1.12810969e+00
-5.11060655e-02 6.88732862e-02 -1.23092031e+00 7.44274914e-01
-1.80172670e+00 -7.81004429e-01 -6.43244460e-02 2.21784663e+00
7.98497140e-01 9.40443426e-02 1.83705702e-01 9.37226880e-03
2.67055243e-01 5.29669365e-03 -1.22457957e+00 -2.41074175e-01
-5.59480973e-02 2.56633818e-01 4.18518245e-01 4.67369556e-01
-7.89385617e-01 8.71839404e-01 7.00817251e+00 4.75426883e-01
-1.58081651e+00 1.78395942e-01 4.79096740e-01 -1.97805882e-01
-2.54236728e-01 1.47330746e-01 -7.04543531e-01 4.48749036e-01
1.47690666e+00 -4.27453279e-01 7.57402122e-01 5.44521272e-01
2.95237094e-01 3.55726480e-01 -1.19545197e+00 7.89830446e-01
-5.71495712e-01 -1.93932056e+00 2.38792691e-02 4.21301335e-01
9.85286474e-01 1.90187991e-01 1.59316331e-01 8.35071430e-02
2.49233916e-01 -1.02206063e+00 5.25106907e-01 9.96276021e-01
9.51559663e-01 -5.11399269e-01 1.49250224e-01 7.23230600e-01
-1.16169596e+00 -1.32883877e-01 -4.65045869e-01 -4.01822686e-01
2.01529816e-01 6.37550056e-01 -3.60364199e-01 9.01482105e-02
6.68393850e-01 9.67867374e-01 -2.35976994e-01 6.03185594e-01
3.73152226e-01 7.93742120e-01 -6.54268086e-01 1.34928316e-01
2.57549256e-01 -3.50331008e-01 7.42607951e-01 7.04780221e-01
4.82305884e-01 2.36055050e-02 -1.72393266e-02 1.36800539e+00
6.60213083e-02 -3.43852818e-01 -8.29504251e-01 -2.50810772e-01
2.06819594e-01 1.21077919e+00 -6.34962261e-01 -1.47166625e-01
-2.98668128e-02 8.73770952e-01 4.72231269e-01 5.34356833e-01
-6.67741656e-01 -1.06037125e-01 8.62531602e-01 4.93990898e-01
5.58864713e-01 -8.83326471e-01 -9.55094323e-02 -1.23909175e+00
7.07012638e-02 -4.61139768e-01 -3.79478261e-02 -4.51075464e-01
-1.33468223e+00 7.36621916e-01 -1.39398277e-01 -1.47361195e+00
-4.93015379e-01 -7.20065713e-01 -4.97741878e-01 8.80773187e-01
-1.23424923e+00 -8.81124914e-01 -4.49627228e-02 5.43498278e-01
1.27799705e-01 -1.13165244e-01 1.24431121e+00 3.90043445e-02
-4.58319634e-01 1.52202785e-01 7.38294363e-01 -3.15191060e-01
4.91375104e-02 -1.19887030e+00 7.13751554e-01 5.03942490e-01
1.16221488e-01 4.95344520e-01 7.71872878e-01 -4.78878766e-01
-1.82194757e+00 -9.69691515e-01 6.20287180e-01 -7.02673793e-01
1.08497703e+00 -8.11927021e-01 -1.24400496e+00 3.64384770e-01
-3.18219811e-01 5.74546218e-01 2.76605338e-01 -3.44619714e-02
-1.32421806e-01 -1.37021944e-01 -8.46642017e-01 6.07998550e-01
1.12724674e+00 -8.82783055e-01 -7.35895857e-02 4.94748980e-01
6.84829116e-01 -3.17058831e-01 -1.16914451e+00 2.23907068e-01
5.38627267e-01 -6.75867319e-01 9.62340772e-01 -9.59416032e-01
5.54487884e-01 -2.97760159e-01 1.64706424e-01 -1.18194449e+00
-4.89526689e-01 -9.63799655e-01 -8.63961041e-01 7.03948736e-01
2.95044929e-01 -8.02513063e-01 7.08841920e-01 4.28363472e-01
5.94614968e-02 -1.12059844e+00 -1.12092364e+00 -8.96902323e-01
4.87519115e-01 -2.36195996e-01 5.15524685e-01 9.57148433e-01
-1.49631917e-01 1.21851236e-01 -2.41232082e-01 2.84776270e-01
5.62906921e-01 3.91154200e-01 5.87456763e-01 -1.36057484e+00
-3.35175157e-01 -5.08472323e-01 -5.64392090e-01 -9.68244851e-01
3.38961124e-01 -7.77653515e-01 -2.79714853e-01 -1.12766850e+00
-2.86386251e-01 -5.09251952e-01 -3.01796049e-01 1.76123053e-01
3.40639055e-01 3.10861945e-01 -2.07173586e-01 7.33628094e-01
-4.19417292e-01 9.11926627e-01 1.24808931e+00 1.40436843e-01
-2.72337198e-01 -1.57387629e-01 5.66483885e-02 5.80064058e-01
5.24875283e-01 -6.49698377e-01 -2.38631815e-01 -3.78915370e-01
3.82111788e-01 2.35675558e-01 7.81693816e-01 -1.35929573e+00
4.62334245e-01 -2.77598381e-01 4.75338250e-01 -4.61120695e-01
5.24599791e-01 -7.35577881e-01 4.85295147e-01 6.21598601e-01
-3.58773738e-01 2.72988826e-01 3.30568254e-01 5.92510462e-01
-5.87739684e-02 4.19229984e-01 6.11755371e-01 2.36621201e-02
-1.95399523e-01 8.15107644e-01 -6.93988442e-01 -8.73283222e-02
9.99932110e-01 -7.21860752e-02 -6.17003068e-03 -2.62375951e-01
-8.59812737e-01 -1.57520361e-02 7.57288694e-01 1.21763594e-01
1.42340690e-01 -1.46599817e+00 -4.56318110e-01 4.88319606e-01
-7.61644468e-02 8.81572813e-02 4.59625244e-01 8.61373425e-01
-5.06512046e-01 5.96878290e-01 -3.70544672e-01 -9.36611474e-01
-2.81150967e-01 5.35601079e-01 6.95003688e-01 -4.40043658e-01
-1.04023194e+00 4.21269447e-01 2.43735820e-01 -4.74848539e-01
-1.88403845e-01 -3.14470112e-01 3.33895952e-01 -1.53635338e-01
1.90983012e-01 2.62875885e-01 -9.19886604e-02 -5.30018806e-01
-9.32858735e-02 7.42261410e-01 2.39269376e-01 -1.00544058e-01
1.67584252e+00 1.24539338e-01 -2.18558922e-01 8.29206228e-01
1.36109710e+00 -5.46701968e-01 -1.87391210e+00 -4.48008537e-01
-9.98049304e-02 -8.59786272e-02 2.85751045e-01 -3.32574576e-01
-8.37395728e-01 1.17781746e+00 5.35487115e-01 6.72349691e-01
9.34516907e-01 -1.69560332e-02 8.01099181e-01 5.12995005e-01
3.85027617e-01 -9.77742434e-01 4.31015901e-02 6.04745448e-01
8.33669066e-01 -7.01575875e-01 -1.84791252e-01 3.16775709e-01
-1.97672267e-02 1.21765399e+00 3.54233027e-01 -5.73279440e-01
9.39286888e-01 4.04404908e-01 -3.31659347e-01 -3.96804184e-01
-1.33982778e+00 3.45674723e-01 4.12413955e-01 3.43632042e-01
1.48983240e-01 -1.02150984e-01 2.38982514e-01 2.24813849e-01
-1.36874482e-01 -4.86712530e-02 2.39437461e-01 6.11947775e-01
-8.22535753e-02 -1.03031993e+00 1.02415577e-01 5.17078042e-01
-3.33616929e-03 2.05126941e-01 -1.58598647e-01 6.49712801e-01
-1.13326617e-01 8.52651820e-02 3.32439512e-01 -1.89827248e-01
1.90475419e-01 1.64273739e-01 2.68072188e-01 -3.08023334e-01
-2.20756918e-01 -2.78527558e-01 -5.81056118e-01 -8.06077659e-01
-1.33856237e-01 -8.09272110e-01 -9.80237484e-01 -6.41030848e-01
3.60268913e-02 1.29526421e-01 4.86014664e-01 1.10581386e+00
8.55071902e-01 5.24680674e-01 6.93002284e-01 -1.27633607e+00
-7.14139938e-01 -6.58802271e-01 -4.98425931e-01 3.62371862e-01
7.80280948e-01 -6.20858729e-01 -5.12249172e-01 7.41562769e-02] | [6.541858196258545, 3.466791868209839] |
f51d2a1f-9696-486d-aa27-ed78dd718b92 | pre-train-and-plug-in-flexible-conditional | 1911.03882 | null | https://arxiv.org/abs/1911.03882v4 | https://arxiv.org/pdf/1911.03882v4.pdf | Pre-train and Plug-in: Flexible Conditional Text Generation with Variational Auto-Encoders | Conditional Text Generation has drawn much attention as a topic of Natural Language Generation (NLG) which provides the possibility for humans to control the properties of generated contents. Current conditional generation models cannot handle emerging conditions due to their joint end-to-end learning fashion. When a new condition added, these techniques require full retraining. In this paper, we present a new framework named Pre-train and Plug-in Variational Auto-Encoder (PPVAE) towards flexible conditional text generation. PPVAE decouples the text generation module from the condition representation module to allow "one-to-many" conditional generation. When a fresh condition emerges, only a lightweight network needs to be trained and works as a plug-in for PPVAE, which is efficient and desirable for real-world applications. Extensive experiments demonstrate the superiority of PPVAE against the existing alternatives with better conditionality and diversity but less training effort. | ['Yu Duan', 'Jiaxin Pei', 'Canwen Xu', 'Chenliang Li', 'Jialong Han'] | 2019-11-10 | pre-train-and-plug-in-flexible-conditional-1 | https://aclanthology.org/2020.acl-main.23 | https://aclanthology.org/2020.acl-main.23.pdf | acl-2020-6 | ['conditional-text-generation'] | ['natural-language-processing'] | [ 2.69642055e-01 1.84039921e-01 -1.69658646e-01 -9.20782387e-02
-7.39416957e-01 -4.69688833e-01 1.05892813e+00 -2.46115506e-01
-1.12420209e-01 1.18600225e+00 3.10787439e-01 -3.35125506e-01
4.42004830e-01 -9.33229566e-01 -6.15216315e-01 -7.45627344e-01
3.69364321e-01 5.44556558e-01 7.29032904e-02 -2.21772924e-01
-1.41972210e-02 3.67396534e-03 -1.51212335e+00 1.45234197e-01
1.43141174e+00 5.49668550e-01 6.37056231e-01 6.95173681e-01
-1.87865287e-01 5.14574885e-01 -6.00235939e-01 -6.13265336e-01
1.04168087e-01 -7.30481684e-01 -5.08014202e-01 -2.04816554e-02
1.48334010e-02 -3.51959914e-01 5.01990831e-03 7.29817510e-01
8.08926284e-01 3.54808897e-01 8.86258781e-01 -1.25530767e+00
-8.92033696e-01 8.33082020e-01 -4.23320949e-01 -2.74525076e-01
4.54068899e-01 3.65424275e-01 1.08547986e+00 -7.63749361e-01
7.59329557e-01 1.17681253e+00 3.37153792e-01 1.11725545e+00
-1.43281066e+00 -5.43233275e-01 4.12686974e-01 -1.46785587e-01
-1.01915431e+00 -3.13341320e-01 7.20789552e-01 -3.60247344e-01
8.03594053e-01 1.84394717e-01 7.04763055e-01 1.59594548e+00
2.19008610e-01 1.12261021e+00 6.81002498e-01 -4.31403100e-01
4.05867636e-01 1.69363961e-01 -4.42855656e-01 4.96288747e-01
1.53344765e-01 1.71213821e-01 -6.20585561e-01 -3.76445204e-02
6.34721160e-01 -1.90513238e-01 -3.66015285e-01 -1.17641449e-01
-1.22434735e+00 9.61758971e-01 1.64065868e-01 1.72967643e-01
-4.87535596e-01 3.10745239e-01 1.69507369e-01 5.27666733e-02
5.27137935e-01 4.72543895e-01 -3.23884130e-01 -2.60395616e-01
-1.21598649e+00 6.37648046e-01 7.05304265e-01 1.15633416e+00
6.84697568e-01 4.46492374e-01 -8.68686736e-01 7.76503503e-01
2.31022507e-01 5.29853404e-01 7.14844406e-01 -4.02807623e-01
4.14875418e-01 5.71741521e-01 1.54059812e-01 -3.84182304e-01
8.78259763e-02 -5.33986449e-01 -9.64645207e-01 1.44948857e-02
1.10620714e-03 -6.53340578e-01 -1.14219654e+00 2.01526117e+00
4.32455212e-01 1.96664795e-01 1.36569783e-01 5.51218867e-01
4.79236990e-01 1.10453522e+00 2.33879551e-01 -3.09909135e-01
1.16413593e+00 -9.69818771e-01 -8.62825811e-01 -1.31965190e-01
3.25376958e-01 -7.51673102e-01 1.25588298e+00 2.42325559e-01
-1.27746344e+00 -5.40451050e-01 -1.03705549e+00 -6.14279462e-03
-4.48980391e-01 1.50604859e-01 6.51821434e-01 4.55635399e-01
-1.10560477e+00 3.43072563e-01 -6.59417450e-01 5.12898192e-02
2.12686688e-01 7.77633190e-02 -1.16012814e-02 1.56503484e-01
-1.61003125e+00 6.86885476e-01 5.31436861e-01 -5.90816187e-03
-1.09957957e+00 -8.02989662e-01 -9.14875567e-01 1.47074103e-01
3.62428844e-01 -1.04508030e+00 1.48359561e+00 -7.98642576e-01
-2.05903935e+00 3.06067318e-01 -1.24944180e-01 -3.77591968e-01
7.82110333e-01 -3.98465008e-01 -3.66341472e-01 -1.11712277e-01
1.62793040e-01 8.77291977e-01 1.27116513e+00 -1.31809127e+00
-5.10355294e-01 2.26175353e-01 -5.27716950e-02 2.09788769e-01
-1.87492430e-01 -2.60905981e-01 -4.60023075e-01 -8.84106159e-01
-3.85832161e-01 -8.90518963e-01 -2.95517921e-01 -2.97977269e-01
-6.91472888e-01 -5.91374636e-01 8.11748147e-01 -3.49451870e-01
1.48708069e+00 -1.87726545e+00 3.23070228e-01 -1.38687924e-01
-1.38799191e-01 4.83521134e-01 -1.17866755e-01 6.85285449e-01
-3.21954004e-02 2.81611979e-01 -4.20408785e-01 -6.65199757e-01
3.08671147e-01 9.66984704e-02 -3.53484273e-01 -7.68511742e-02
5.33190489e-01 1.03980315e+00 -8.68906379e-01 -6.84738994e-01
9.51638818e-02 5.08612394e-01 -9.49577928e-01 5.45978427e-01
-9.48855042e-01 3.42816502e-01 -5.61456978e-01 4.11988974e-01
6.84255779e-01 -2.77477264e-01 6.14833683e-02 1.95212767e-01
-6.51715323e-02 1.36511505e-01 -9.98809457e-01 1.87697291e+00
-6.23262584e-01 2.79733300e-01 -1.00986660e-01 -5.95440209e-01
8.68143618e-01 5.67386687e-01 -1.62555743e-02 -3.59701216e-01
7.22459033e-02 7.77603535e-04 -1.91925347e-01 -4.07842368e-01
8.23970437e-01 -2.93177038e-01 -2.87387013e-01 5.12923241e-01
1.54718772e-01 -3.05100620e-01 5.98261058e-01 5.10525107e-01
6.05772555e-01 6.76474392e-01 1.81338608e-01 -1.63877741e-01
5.00220656e-01 -3.92067045e-01 6.69068396e-01 5.37248075e-01
2.97197312e-01 6.99124157e-01 4.16121185e-01 7.32564777e-02
-8.30632985e-01 -1.21413243e+00 4.40122522e-02 1.17536938e+00
-1.30213648e-01 -4.77076352e-01 -8.64688754e-01 -6.01615429e-01
-2.68707454e-01 1.18018019e+00 -5.95620036e-01 -2.41873667e-01
-5.56291044e-01 -7.75191844e-01 3.30438375e-01 5.03415048e-01
5.09903371e-01 -1.47274458e+00 -5.21720290e-01 3.60174328e-01
-4.76295799e-01 -8.44896197e-01 -6.75668180e-01 5.33314757e-02
-5.41498184e-01 -4.30689096e-01 -1.00233960e+00 -7.64933348e-01
6.31247699e-01 -2.26280004e-01 1.11721313e+00 -1.82081103e-01
-9.29008946e-02 1.31641999e-01 -4.98924673e-01 -4.76110309e-01
-7.17256010e-01 4.52304453e-01 -1.94450215e-01 1.45860940e-01
-2.04101607e-01 -5.75304329e-01 -6.44550800e-01 -1.52531818e-01
-1.09289944e+00 4.24583346e-01 6.52611852e-01 9.05014396e-01
4.52871382e-01 -1.92127019e-01 9.99825120e-01 -8.69229496e-01
1.11025047e+00 -4.69711810e-01 -6.39578879e-01 3.72065127e-01
-6.19827986e-01 3.96029621e-01 8.72357547e-01 -5.58419108e-01
-1.67391765e+00 -6.02078289e-02 -1.91758290e-01 -3.08424145e-01
1.61146540e-02 7.07891166e-01 -3.16752762e-01 8.85898113e-01
5.26566207e-01 5.30610621e-01 -1.95266888e-01 -3.47209215e-01
6.36000693e-01 5.68091571e-01 5.50093412e-01 -8.25059593e-01
8.75655890e-01 -7.36504793e-02 -2.04200163e-01 -4.47362870e-01
-7.03177989e-01 -2.44206730e-02 -3.42047632e-01 -1.62736207e-01
1.07503688e+00 -1.04030359e+00 -4.53963012e-01 4.02134567e-01
-1.17657566e+00 -5.88078141e-01 -4.68071312e-01 1.58619046e-01
-5.85906804e-01 1.10337853e-01 -3.52530092e-01 -8.51434410e-01
-7.73392737e-01 -9.65299308e-01 1.13581347e+00 4.46703404e-01
-1.90396532e-01 -1.04211771e+00 1.35117650e-01 5.45178726e-02
5.95034719e-01 2.90404916e-01 8.69241476e-01 -2.48079404e-01
-6.04778588e-01 -2.46768594e-01 1.23388372e-01 2.71756440e-01
1.28991574e-01 2.39417732e-01 -8.79342556e-01 -3.87175381e-01
-2.98601419e-01 -3.51887137e-01 8.69444430e-01 2.34924883e-01
1.04720938e+00 -4.86474544e-01 -2.92591691e-01 4.96103972e-01
1.33248818e+00 7.49441311e-02 6.95196867e-01 -1.83480099e-01
6.97351158e-01 3.18928987e-01 3.07625771e-01 6.43592179e-01
3.39021415e-01 3.43258083e-01 2.90981948e-01 -7.87125751e-02
-2.36533970e-01 -8.06538999e-01 6.36780679e-01 8.40785444e-01
1.05127282e-02 -8.67065430e-01 -5.23680806e-01 6.10014975e-01
-1.86120105e+00 -1.06647670e+00 1.07876889e-01 2.12709904e+00
1.28392637e+00 2.19244007e-02 -1.24738462e-01 -1.08512722e-01
8.14804018e-01 3.42567503e-01 -5.92044294e-01 -3.74986112e-01
5.27173765e-02 3.37866962e-01 -2.30413496e-01 4.18961585e-01
-8.19304645e-01 1.22506833e+00 6.15926266e+00 1.09532702e+00
-1.11965358e+00 -2.42647696e-02 5.29663086e-01 -4.31882776e-02
-7.40692317e-01 6.69351742e-02 -8.62730443e-01 7.25023210e-01
8.01437020e-01 -1.85198486e-01 2.37918228e-01 7.84757495e-01
1.55839384e-01 -5.28778881e-02 -1.17916107e+00 6.19682372e-01
9.49380163e-04 -1.35019314e+00 5.56308389e-01 -7.12098405e-02
9.91719842e-01 -4.45039392e-01 6.45042434e-02 5.85707366e-01
6.07795060e-01 -8.71806562e-01 8.14351499e-01 5.31412959e-01
1.08607721e+00 -7.48310149e-01 4.09711570e-01 5.46357572e-01
-1.12083650e+00 2.06659049e-01 -2.15337232e-01 1.56597406e-01
6.40476704e-01 5.36662400e-01 -9.15087521e-01 6.07802272e-01
1.72738299e-01 4.35576975e-01 -3.56930912e-01 5.06549299e-01
-6.20206833e-01 6.45797789e-01 -2.85918154e-02 -3.19747806e-01
2.40268499e-01 -2.64851511e-01 6.49209738e-01 1.21183085e+00
5.43196738e-01 -1.16203971e-01 2.23012239e-01 1.24566329e+00
-2.00138241e-01 -3.30166779e-02 -3.97624284e-01 -1.60939157e-01
3.83577406e-01 1.21652818e+00 -3.93270731e-01 -5.57473242e-01
-1.16116270e-01 1.20502317e+00 3.67364496e-01 5.03735363e-01
-9.48267162e-01 -3.09894532e-01 3.62077862e-01 1.50747910e-01
5.61216414e-01 2.77561937e-02 -7.72535130e-02 -1.43970263e+00
-5.65797351e-02 -8.54038477e-01 2.09026024e-01 -8.10853481e-01
-1.35228240e+00 6.94151819e-01 1.07321814e-01 -1.21049571e+00
-8.32904696e-01 -2.29991287e-01 -9.77331817e-01 9.28383231e-01
-1.36634183e+00 -1.39683592e+00 -6.44471794e-02 7.37196207e-01
8.46361101e-01 -1.97718605e-01 8.67055893e-01 -2.36195363e-02
-7.21433461e-01 7.82375991e-01 -1.60556406e-01 -1.90407157e-01
6.64704442e-01 -1.49187732e+00 4.41400290e-01 1.21510780e+00
-1.51961029e-01 6.39147937e-01 8.60329866e-01 -8.21369290e-01
-1.12770569e+00 -1.14672613e+00 1.07122171e+00 -1.33715332e-01
2.33861253e-01 -6.22972608e-01 -6.29291058e-01 5.25998771e-01
8.02921832e-01 -3.81154746e-01 5.75699329e-01 8.11818019e-02
-1.04219623e-01 8.78267363e-02 -9.47259903e-01 9.54879165e-01
8.34235311e-01 -2.55556643e-01 -3.56739759e-01 2.17977226e-01
1.09447861e+00 -3.37802231e-01 -4.48921889e-01 2.84234017e-01
2.80437589e-01 -7.80613482e-01 6.13929808e-01 -3.84448379e-01
7.77950048e-01 -3.43272567e-01 9.62058380e-02 -1.69313836e+00
-2.94593722e-01 -1.27357519e+00 -3.63364995e-01 1.44380343e+00
8.47932577e-01 -5.76253355e-01 5.91968715e-01 5.76487541e-01
-4.03372556e-01 -8.55421484e-01 -6.42620862e-01 -5.33734322e-01
2.90014118e-01 -3.09880257e-01 9.04122472e-01 7.67711103e-01
3.12807560e-02 7.17229843e-01 -7.04302669e-01 -2.34453723e-01
3.05556625e-01 3.05893153e-01 7.62872875e-01 -1.08776498e+00
-4.63601828e-01 -3.72827888e-01 2.67820895e-01 -1.30118430e+00
1.04971521e-01 -9.57131088e-01 3.96824569e-01 -1.77239418e+00
1.30353063e-01 -3.36754501e-01 1.12332692e-02 3.95921171e-01
-7.25688994e-01 -1.83759078e-01 3.08553249e-01 -1.68710768e-01
-4.14405733e-01 1.15008187e+00 1.59965932e+00 3.69015671e-02
-4.00725067e-01 1.15868494e-01 -8.47484529e-01 3.56172740e-01
1.00763631e+00 -2.57546335e-01 -8.22644770e-01 -3.79212052e-01
2.78369486e-01 2.49536172e-01 7.24937255e-03 -9.28450704e-01
2.03508168e-01 -1.91188604e-01 3.56809080e-01 -6.71004236e-01
2.88554281e-01 -4.63287264e-01 1.42333165e-01 2.63877600e-01
-4.15645480e-01 1.32987559e-01 5.95256593e-03 5.89879870e-01
-1.12665974e-01 -2.78345287e-01 5.47944307e-01 -1.23980746e-01
-2.96594620e-01 5.27429700e-01 -6.18616879e-01 2.66256303e-01
1.06963158e+00 -5.98230362e-02 -1.55564070e-01 -5.35098910e-01
-4.63512868e-01 3.62258404e-01 1.56713411e-01 4.72604930e-01
6.06541455e-01 -1.37161553e+00 -9.36275780e-01 3.15737247e-01
-9.31433926e-04 3.69156480e-01 3.08709264e-01 2.99241453e-01
-2.92354673e-01 1.96106747e-01 1.80754185e-01 -2.95394480e-01
-7.74837136e-01 5.39191306e-01 3.70755345e-02 -7.10317075e-01
-4.61714178e-01 8.98251951e-01 2.87125796e-01 -3.95894557e-01
9.76754948e-02 -2.88303614e-01 6.76439852e-02 1.28319308e-01
3.44105810e-01 4.25514579e-02 -2.98339903e-01 -2.82208234e-01
1.40041918e-01 2.90962942e-02 -2.86139816e-01 -3.86659175e-01
1.13811421e+00 1.18611559e-01 1.57885700e-01 3.07494044e-01
8.45470071e-01 -7.99845830e-02 -1.54939151e+00 -8.02418292e-02
-4.94125187e-01 -1.47064179e-01 -6.07437529e-02 -9.56322908e-01
-9.37427819e-01 8.65454614e-01 1.64414793e-01 1.83899686e-01
1.13787711e+00 -2.27534994e-01 9.18943942e-01 1.49692789e-01
1.42837420e-01 -1.33231461e+00 3.18543166e-01 4.29910690e-01
1.07729411e+00 -1.02955496e+00 -1.26456127e-01 -2.44657829e-01
-9.89199638e-01 6.72226846e-01 8.30920219e-01 1.07873473e-02
5.74374080e-01 2.16856986e-01 -8.26751739e-02 1.26017407e-01
-1.16030037e+00 -7.56586045e-02 2.89554149e-01 6.57193840e-01
6.03513539e-01 2.88383484e-01 -4.47467297e-01 5.33655882e-01
-5.41369617e-01 -6.90205470e-02 2.96583712e-01 7.52505779e-01
-1.54020414e-01 -1.38437426e+00 2.50217095e-02 3.89063984e-01
-3.86088699e-01 -4.03869450e-01 -1.55932412e-01 7.24023163e-01
1.88941121e-01 1.01666379e+00 -3.13293859e-02 -1.65716514e-01
2.72819214e-02 3.75275254e-01 2.67124265e-01 -6.87258959e-01
-6.51669443e-01 2.41398215e-01 6.44103438e-02 -1.92974642e-01
-2.98463970e-01 -6.63381398e-01 -1.21791148e+00 -1.61212474e-01
-5.98689139e-01 2.31066838e-01 4.30290520e-01 8.22193384e-01
5.92377245e-01 8.45985174e-01 5.45346916e-01 -7.23936975e-01
-5.96941531e-01 -1.18376875e+00 -3.74079853e-01 3.47289294e-01
3.16341519e-01 -4.17069703e-01 -1.99534744e-01 3.54949862e-01] | [11.89390754699707, 9.112628936767578] |
07281132-b0ca-42e6-821c-985eeca19b75 | re-2tal-rewiring-pretrained-video-backbones | 2211.14053 | null | https://arxiv.org/abs/2211.14053v2 | https://arxiv.org/pdf/2211.14053v2.pdf | Re^2TAL: Rewiring Pretrained Video Backbones for Reversible Temporal Action Localization | Temporal action localization (TAL) requires long-form reasoning to predict actions of various durations and complex content. Given limited GPU memory, training TAL end to end (i.e., from videos to predictions) on long videos is a significant challenge. Most methods can only train on pre-extracted features without optimizing them for the localization problem, consequently limiting localization performance. In this work, to extend the potential in TAL networks, we propose a novel end-to-end method Re2TAL, which rewires pretrained video backbones for reversible TAL. Re2TAL builds a backbone with reversible modules, where the input can be recovered from the output such that the bulky intermediate activations can be cleared from memory during training. Instead of designing one single type of reversible module, we propose a network rewiring mechanism, to transform any module with a residual connection to a reversible module without changing any parameters. This provides two benefits: (1) a large variety of reversible networks are easily obtained from existing and even future model designs, and (2) the reversible models require much less training effort as they reuse the pre-trained parameters of their original non-reversible versions. Re2TAL, only using the RGB modality, reaches 37.01% average mAP on ActivityNet-v1.3, a new state-of-the-art record, and mAP 64.9% at tIoU=0.5 on THUMOS-14, outperforming all other RGB-only methods. | ['Bernard Ghanem', 'Karttikeya Mangalam', 'Shuming Liu', 'Chen Zhao'] | 2022-11-25 | null | null | null | null | ['action-localization'] | ['computer-vision'] | [ 2.35975787e-01 -2.19252612e-02 -4.77236807e-01 -1.10609852e-01
-4.37596291e-01 -5.68824708e-01 3.96982521e-01 -5.65784752e-01
-6.41002893e-01 6.87073052e-01 3.14262033e-01 -6.49997266e-03
2.85643607e-01 -6.22895181e-01 -1.24415314e+00 -5.57101488e-01
-6.74995780e-02 1.29495293e-01 6.28390074e-01 1.92883164e-02
1.05765671e-01 3.87578338e-01 -1.17987692e+00 7.43100405e-01
4.86463100e-01 1.20392394e+00 4.38437372e-01 5.95000982e-01
2.27572143e-01 1.50945246e+00 -2.63328642e-01 -1.71811864e-01
2.54361957e-01 -4.66316700e-01 -7.39773870e-01 -2.39304543e-01
4.84575033e-01 -7.15787709e-01 -9.97668147e-01 5.04632294e-01
3.83986801e-01 1.41370282e-01 2.61824250e-01 -1.04160488e+00
-5.74268758e-01 7.12704778e-01 -2.69243300e-01 1.87103525e-01
1.01733342e-01 5.80000997e-01 8.85753274e-01 -1.00082147e+00
8.63632560e-01 1.05595267e+00 7.28835285e-01 7.93224752e-01
-1.21808410e+00 -7.39497125e-01 2.90901423e-01 6.11391962e-01
-1.38622832e+00 -6.35557950e-01 3.64606082e-01 -2.62169629e-01
1.57047451e+00 -9.32050869e-02 9.54165220e-01 1.37533438e+00
4.09772694e-01 9.33957756e-01 9.22710955e-01 1.66496024e-01
1.85760602e-01 -4.82060641e-01 -2.67128944e-01 1.00804484e+00
-4.65938002e-02 -1.80269301e-01 -1.01071846e+00 4.40881014e-01
1.01957655e+00 3.03136736e-01 -3.00401688e-01 -1.93495825e-01
-1.45361757e+00 3.00301015e-01 8.38202357e-01 1.43397868e-01
-3.87580633e-01 9.77097213e-01 3.58158231e-01 1.65338233e-01
2.21985117e-01 4.27536458e-01 -4.35184747e-01 -3.42301786e-01
-1.04560852e+00 -3.58366780e-02 6.18687212e-01 8.55450213e-01
9.14403617e-01 7.12399855e-02 -2.58877099e-01 3.36147100e-01
-3.22443969e-03 4.85469997e-01 5.02122939e-01 -1.17159700e+00
5.99925399e-01 6.18604124e-01 -3.49353179e-02 -6.24637842e-01
-4.47489917e-01 -3.79728079e-01 -7.13804841e-01 -1.24316826e-01
1.42073303e-01 -1.01338223e-01 -1.17229605e+00 1.83355856e+00
-1.11258626e-01 6.99157536e-01 5.16865365e-02 7.70404756e-01
7.04414666e-01 8.61125052e-01 3.79603952e-02 8.44627023e-02
8.11892807e-01 -1.50493073e+00 -3.39021415e-01 -6.60822392e-01
7.66962290e-01 -2.73864955e-01 1.09291661e+00 3.68549109e-01
-1.16734231e+00 -5.37016392e-01 -1.23046052e+00 -4.11028057e-01
-1.17373221e-01 1.80977613e-01 7.13207960e-01 4.25366201e-02
-1.35923803e+00 9.39701200e-01 -1.36262178e+00 -3.12726647e-01
6.37269795e-01 6.62252069e-01 -5.24320960e-01 -2.75965631e-01
-9.07613397e-01 7.85598397e-01 4.63370353e-01 3.59735042e-01
-1.32525349e+00 -6.29201829e-01 -7.28153944e-01 2.19615519e-01
4.36345279e-01 -7.22120106e-01 1.10464978e+00 -1.22649062e+00
-1.77238595e+00 3.79274666e-01 -2.83426732e-01 -6.46928549e-01
5.53418279e-01 -4.35377479e-01 -1.23039447e-01 4.38011050e-01
-1.69363573e-01 1.17053568e+00 8.28238428e-01 -6.50364161e-01
-4.57387894e-01 1.55022606e-01 2.25469604e-01 1.98459938e-01
-4.60381329e-01 -3.83655369e-01 -9.16517377e-01 -4.06770289e-01
6.95464090e-02 -1.17342460e+00 -9.10038948e-02 2.32908517e-01
-1.45247862e-01 -3.49281635e-03 7.34373450e-01 -6.86989844e-01
1.05963802e+00 -2.14514923e+00 4.64157462e-01 -6.74224049e-02
2.69165397e-01 2.17445657e-01 -4.96086091e-01 2.62121767e-01
5.15434295e-02 3.15863192e-02 -9.16291401e-02 -4.87878323e-01
-1.99539843e-03 3.15692902e-01 -3.51669282e-01 5.34127712e-01
3.12368810e-01 1.32787538e+00 -8.09444129e-01 -3.63666505e-01
1.55521527e-01 5.36917090e-01 -8.80997360e-01 -2.70447023e-02
-2.99913198e-01 3.84833604e-01 -1.60490319e-01 5.50643086e-01
3.00551474e-01 -4.40260977e-01 3.45161945e-01 -3.65413308e-01
-1.04964741e-01 4.85125661e-01 -7.45984375e-01 2.15594935e+00
-6.23933792e-01 8.57909977e-01 -2.89292514e-01 -9.32246268e-01
5.41476429e-01 1.29580483e-01 6.01432145e-01 -9.96780872e-01
8.20982177e-03 1.57423064e-01 -5.07322140e-02 -2.90747315e-01
3.64087075e-01 2.55861193e-01 -6.46393374e-02 4.83789235e-01
3.69471699e-01 3.56158674e-01 3.49282086e-01 1.15936495e-01
1.77433670e+00 5.98737001e-01 -1.87470049e-01 6.23609982e-02
1.02979153e-01 -1.48424804e-01 7.94602036e-01 7.33303607e-01
-1.10948645e-01 5.62679887e-01 6.85034931e-01 -5.94062567e-01
-1.02614021e+00 -1.14335608e+00 3.95759761e-01 1.23965752e+00
2.19845831e-01 -6.53274953e-01 -7.79463887e-01 -6.18452668e-01
-2.72049457e-01 4.04569864e-01 -6.13605857e-01 -4.68377858e-01
-1.09892464e+00 -3.94545555e-01 6.73087776e-01 9.14013267e-01
1.01823425e+00 -1.14459324e+00 -6.94578886e-01 3.79010201e-01
-3.54866654e-01 -1.19505537e+00 -5.15224516e-01 3.24559122e-01
-1.11082494e+00 -8.81065428e-01 -4.12388831e-01 -6.58171415e-01
6.61346793e-01 2.55982012e-01 1.03286707e+00 1.32360399e-01
6.67672008e-02 2.19093680e-01 -2.13327333e-01 3.91102314e-01
4.08574305e-02 5.18737853e-01 -1.20019592e-01 -8.95323530e-02
7.79963285e-02 -7.04236805e-01 -9.32804108e-01 4.49078947e-01
-7.68561959e-01 4.32194561e-01 9.54530597e-01 9.03442383e-01
7.11152315e-01 -2.57325768e-01 2.92385429e-01 -5.11473119e-01
-1.05743222e-01 -3.99353534e-01 -4.03963149e-01 2.38979653e-01
-3.09012830e-01 2.64580190e-01 6.98628366e-01 -7.49337792e-01
-8.71527433e-01 5.05235314e-01 -7.12613389e-02 -6.91989124e-01
3.06550473e-01 4.13436025e-01 -3.90095264e-02 -4.03120890e-02
6.08461559e-01 4.35212433e-01 -1.72990456e-01 -2.82706439e-01
3.46076339e-01 1.09861217e-01 6.62045240e-01 -4.32268500e-01
5.39061487e-01 5.47107935e-01 -2.44515035e-02 -4.81109202e-01
-8.33905995e-01 -9.22737345e-02 -7.55853057e-01 -2.43064120e-01
9.33939517e-01 -1.20248246e+00 -9.67917860e-01 5.47554255e-01
-1.10422730e+00 -1.12949598e+00 -2.43625447e-01 4.07147914e-01
-5.97740829e-01 -7.22195953e-02 -9.70240653e-01 -8.43572170e-02
-3.11963588e-01 -1.18562508e+00 7.52278030e-01 6.87861443e-02
-1.80495381e-01 -6.84550762e-01 -8.76022577e-02 3.63704830e-01
5.09748101e-01 -1.89201590e-02 6.14940107e-01 1.81721952e-02
-1.11998224e+00 -1.16974190e-01 -2.23286137e-01 3.42491865e-01
-3.45977634e-01 -1.47926047e-01 -8.84679496e-01 -4.86591548e-01
-2.67173678e-01 -6.79628551e-01 1.38907599e+00 1.10223398e-01
1.39376962e+00 -3.64605218e-01 -2.78575033e-01 9.49152708e-01
1.12727380e+00 1.27134044e-02 1.05075788e+00 3.96825790e-01
9.74564672e-01 -6.22130968e-02 2.70109981e-01 1.96715176e-01
3.28305900e-01 6.26639485e-01 5.69639921e-01 2.30599605e-02
-3.35590929e-01 -6.43032670e-01 1.10000110e+00 8.21694851e-01
-1.68269739e-01 -2.89801627e-01 -7.00505316e-01 3.22100103e-01
-2.10493135e+00 -1.07826686e+00 3.42430741e-01 1.97282732e+00
8.61008823e-01 3.02880824e-01 -8.41389373e-02 -3.12328666e-01
4.87323225e-01 4.14797425e-01 -9.01583910e-01 4.53081308e-03
-9.85301062e-02 4.92250711e-01 7.98699379e-01 2.72975087e-01
-9.89321053e-01 1.33498943e+00 5.88573456e+00 8.70029390e-01
-1.55747330e+00 3.81301463e-01 5.85591257e-01 -6.36602163e-01
-5.63712604e-02 2.24366993e-01 -7.99066722e-01 4.24391538e-01
1.14862931e+00 2.96398014e-01 7.28988111e-01 7.19597638e-01
1.94243088e-01 -1.33655116e-01 -1.34099221e+00 9.48260725e-01
-2.19849776e-03 -1.67326748e+00 -3.60722132e-02 -8.79791155e-02
7.47002959e-01 3.33444834e-01 7.23947734e-02 5.97418129e-01
1.33795157e-01 -1.17623127e+00 1.06480587e+00 6.53068900e-01
9.62419569e-01 -6.83882892e-01 5.08995950e-01 1.25711039e-01
-1.36069286e+00 -2.43481144e-01 -4.26790953e-01 -1.92410007e-01
1.58530280e-01 1.58571228e-01 -6.06577337e-01 -5.71721010e-02
6.72255635e-01 1.26494598e+00 -6.47621810e-01 9.02783513e-01
-6.23494625e-01 8.60621691e-01 -3.57704580e-01 3.52795832e-02
3.62147242e-01 2.68972874e-01 -2.03144550e-02 1.12832558e+00
4.58381385e-01 -7.82427564e-02 8.51208493e-02 6.67189956e-01
-5.89574456e-01 -3.70269537e-01 -3.51010919e-01 -1.32608488e-01
4.05205965e-01 1.12044001e+00 -7.01900899e-01 -3.17749918e-01
-3.85385096e-01 1.50550628e+00 6.52004242e-01 4.19177324e-01
-1.25072229e+00 -9.34140310e-02 4.49054778e-01 3.16665441e-01
4.96387810e-01 -4.51595694e-01 -1.45063419e-02 -1.28552842e+00
8.13840255e-02 -6.76974833e-01 6.20412454e-03 -8.53523374e-01
-7.20585108e-01 4.79942858e-01 -3.30255568e-01 -1.11181986e+00
-1.25722960e-01 -6.43181324e-01 -4.57297653e-01 3.17309141e-01
-1.18737888e+00 -1.21399415e+00 -4.31540817e-01 7.34095931e-01
5.00981927e-01 1.21392205e-01 6.12484694e-01 5.15010595e-01
-8.13621819e-01 6.89893305e-01 -5.23604602e-02 1.08239375e-01
7.30497479e-01 -8.98335278e-01 4.87336606e-01 1.04326069e+00
1.50318652e-01 5.49288034e-01 1.52100950e-01 -5.49412191e-01
-1.79289699e+00 -1.31143987e+00 5.10212421e-01 -2.27655187e-01
8.44868481e-01 -6.45049989e-01 -6.23774230e-01 1.06063032e+00
1.38920426e-01 3.62055451e-01 2.45788693e-01 -1.29216045e-01
-4.73594785e-01 -4.38212365e-01 -6.24456108e-01 8.26799452e-01
1.61253452e+00 -6.96857631e-01 -7.78800696e-02 3.09688240e-01
7.15928495e-01 -6.11208558e-01 -7.85486698e-01 1.58676043e-01
6.68120742e-01 -8.16008091e-01 9.37200248e-01 -3.43744576e-01
7.92708099e-01 -3.95968199e-01 -1.00523576e-01 -9.57470357e-01
-3.85924339e-01 -7.49931097e-01 -4.87716615e-01 7.19449162e-01
5.46018779e-01 -4.95517492e-01 1.01128721e+00 5.46126664e-01
-3.41502786e-01 -8.82859647e-01 -1.03527510e+00 -8.17709625e-01
-1.67493612e-01 -4.27340567e-01 2.17017770e-01 6.23789310e-01
-1.69294730e-01 4.07696992e-01 -7.93020010e-01 -5.47232181e-02
1.47323579e-01 -9.48997810e-02 7.50563920e-01 -5.06813586e-01
-6.50968134e-01 -1.66679978e-01 -4.84000891e-01 -1.72111046e+00
2.84322560e-01 -8.96463513e-01 2.10444972e-01 -1.41264641e+00
3.17047834e-01 -4.84411418e-01 -4.71773446e-01 1.14916730e+00
1.89625710e-01 5.51564455e-01 3.73264015e-01 4.23695117e-01
-1.02730954e+00 6.47980750e-01 1.31119943e+00 -2.27679461e-01
-1.87610909e-01 -4.14070457e-01 -3.82502466e-01 7.44365752e-01
7.70974517e-01 -5.88463485e-01 -5.48704028e-01 -7.80019403e-01
3.66326004e-01 1.32406861e-01 5.95306635e-01 -1.45016265e+00
3.72784346e-01 -1.27808601e-01 6.22618020e-01 -5.26399732e-01
5.76379240e-01 -6.03220999e-01 3.08807313e-01 6.34001076e-01
-2.72295147e-01 3.06816339e-01 1.39043123e-01 5.28152943e-01
-8.51617828e-02 -4.01361547e-02 6.77157581e-01 -1.51056930e-01
-1.23674870e+00 3.83158386e-01 -4.26875263e-01 -1.31612599e-01
1.00664926e+00 -1.64298400e-01 -7.03214228e-01 -1.49803340e-01
-7.59750903e-01 6.99930117e-02 6.47261083e-01 3.01916122e-01
6.41685545e-01 -1.23769248e+00 -2.67236203e-01 -1.09598346e-01
-2.04437301e-01 4.40925881e-02 4.39909279e-01 1.13190830e+00
-8.35091889e-01 2.59084076e-01 -4.10144836e-01 -6.15961671e-01
-7.87327409e-01 4.11298543e-01 4.93277580e-01 -4.16645914e-01
-8.44868004e-01 9.12842512e-01 1.99073613e-01 -7.02212900e-02
1.41551211e-01 -3.44556153e-01 2.43619204e-01 -1.01860575e-01
4.36140090e-01 2.57216722e-01 -6.14557751e-02 -4.36205506e-01
-4.01527822e-01 3.97556573e-01 -1.44144014e-01 -1.68712154e-01
1.55539417e+00 2.69300014e-01 -1.47011921e-01 2.12457940e-01
1.29847193e+00 -4.22293484e-01 -2.03983617e+00 -9.84254926e-02
-3.38703811e-01 -2.79957891e-01 1.63379461e-01 -7.76475966e-01
-1.53522837e+00 9.33679283e-01 5.99978805e-01 -6.07584357e-01
1.18799818e+00 -1.41188562e-01 9.57424760e-01 8.57120812e-01
4.04471070e-01 -1.07313430e+00 6.39146566e-01 6.95517778e-01
7.82972753e-01 -7.86797464e-01 3.32708769e-02 1.76880658e-02
-5.37673473e-01 1.13024187e+00 8.11062038e-01 -3.12755078e-01
1.78000987e-01 2.83772022e-01 -4.40428287e-01 -3.07314154e-02
-1.07766855e+00 2.22062841e-02 6.76948652e-02 1.96802512e-01
2.67246127e-01 -1.74927235e-01 1.01543926e-02 3.97090286e-01
1.23589844e-01 3.84272397e-01 4.66599464e-01 1.20068717e+00
-3.59910816e-01 -7.95038581e-01 2.29152724e-01 4.18305844e-01
-2.18963623e-01 -2.97100961e-01 -1.89811394e-01 6.79347277e-01
2.20733598e-01 5.73882222e-01 1.27378732e-01 -7.06256330e-01
2.04972804e-01 -1.36366054e-01 5.92576027e-01 -4.53580946e-01
-4.64141130e-01 -1.40096605e-01 1.42610013e-01 -1.19717014e+00
-2.92544723e-01 -5.12772262e-01 -1.47111809e+00 -6.72215044e-01
-3.42007428e-02 -4.43669647e-01 4.80166733e-01 9.71753240e-01
7.12260246e-01 6.82859600e-01 2.41256818e-01 -1.33770931e+00
-1.62568823e-01 -7.71683633e-01 -1.39583811e-01 8.28235298e-02
1.71395857e-02 -6.99165285e-01 -9.62232724e-02 1.50977775e-01] | [8.904682159423828, 0.5230687856674194] |
fc715c5c-b442-4134-927e-3f120c1ebb55 | in-silico-prediction-of-mozenavir-as | 2106.05440 | null | https://arxiv.org/abs/2106.05440v1 | https://arxiv.org/pdf/2106.05440v1.pdf | In silico Prediction of Mozenavir as potential drug for SARS-CoV-2 infection via Binding Multiple Drug Targets | Since the epidemic began in November 2019, no viable medicine against SARS-CoV-2 has been discovered. The typical medication discovery strategy requires several years of rigorous research and development as well as a significant financial commitment, which is not feasible in the face of the current epidemic. Through molecular docking and dynamic simulation studies, we used the FDA-approved drug mezonavir against the most important viral targets, including spike (S) glycoprotein, Transmembrane serine protease 2 (TMPRSS2), RNA-dependent RNA polymerase (RdRp), Main protease (Mpro), human angiotensin-converting enzyme 2 (ACE-2), and furin. These targets are critical for viral replication and infection propagation because they play a key role in replication/transcription and host cell recognition. Molecular docking revealed that the antiviral medication mozenavir showed a stronger affinity for SARS-CoV-2 target proteins than reference medicines in this investigation. We discovered that mozenavir increases the complex's stability and validates the molecular docking findings using molecular dynamics modelling. Furin, a target protein of COVID-19, has a greater binding affinity (-12.04 kcal/mol) than other COVID-19 target proteins, forming different hydrogen bonds and polar and hydrophobic interactions, suggesting that it might be used as an antiviral treatment against SARS-CoV-2. Overall, the present in silico results will be valuable in identifying crucial targets for subsequent experimental investigations that might help combat COVID-19 by blocking the protease furin's proteolytic activity. | ['Abhiav', 'Munipally Praveen Kumar', 'Swapna Gurrapu', 'Rakesh Davella', 'Estari Mamidalaa'] | 2021-06-10 | null | null | null | null | ['molecular-docking'] | ['medical'] | [ 1.21646665e-01 -5.11548400e-01 -2.94428229e-01 4.24303999e-03
-2.41583034e-01 -7.09987581e-01 -2.29754016e-01 3.97804677e-01
-3.86542290e-01 1.15268278e+00 -4.95149195e-02 -7.23977745e-01
2.80526429e-01 -3.57014269e-01 -4.44786996e-01 -8.56082380e-01
-4.36753035e-01 5.07715285e-01 -2.62178004e-01 -1.92572266e-01
-2.65299864e-02 9.04887080e-01 -5.01463413e-01 -7.83878714e-02
1.33048260e+00 1.62307799e-01 3.55062306e-01 2.07735151e-01
2.20611840e-01 -1.81961000e-01 -3.98882002e-01 -1.92623343e-02
1.24705032e-01 -4.89126086e-01 -6.21172301e-02 -5.96702158e-01
-2.95909613e-01 -4.69934553e-01 3.96380395e-01 6.20428979e-01
3.26863647e-01 -3.00106972e-01 9.43197429e-01 -7.51578629e-01
-3.72285008e-01 -3.89343888e-01 -6.32164657e-01 1.41950577e-01
2.98988014e-01 3.00705910e-01 6.72696769e-01 -1.03279006e+00
9.19054925e-01 9.77055073e-01 6.90159976e-01 6.09311342e-01
-1.29332066e+00 -7.07960904e-01 -2.50012726e-01 3.32745537e-02
-1.58549809e+00 -1.25855342e-01 1.53486431e-01 -8.94582391e-01
1.53970432e+00 2.62787580e-01 9.48539972e-01 6.77584469e-01
8.89604509e-01 1.09177113e-01 8.40118229e-01 3.12552810e-01
2.76865810e-01 -4.42013480e-02 -9.33273602e-03 4.86632586e-01
9.55890834e-01 -2.22432986e-02 2.21278727e-01 -1.10032368e+00
7.46740282e-01 6.41395330e-01 -6.57753587e-01 -4.27201718e-01
-8.11116040e-01 1.05562818e+00 -1.59394324e-01 3.12948883e-01
-5.98714471e-01 -2.58633465e-01 2.25183710e-01 -4.49182354e-02
-1.27173960e-01 3.75889033e-01 -9.28183019e-01 5.77001274e-02
-4.36145753e-01 3.31566542e-01 5.21094322e-01 -6.99457526e-02
3.36207807e-01 -6.72250912e-02 1.64004073e-01 3.56317520e-01
8.17636669e-01 9.34026062e-01 -1.96331635e-01 -5.81525154e-02
1.54076181e-02 5.01916409e-01 4.83342022e-01 -7.09692121e-01
-6.96859837e-01 -1.37945920e-01 -8.12257469e-01 1.62734035e-02
4.59216326e-01 -4.04277265e-01 -6.68105900e-01 1.90637863e+00
3.15446496e-01 1.02653615e-01 2.78904885e-01 9.90200937e-01
3.53848487e-01 8.44122052e-01 5.64980567e-01 -1.22207892e+00
1.73803127e+00 -2.48981938e-01 -4.97998178e-01 1.77808270e-01
3.86638403e-01 -7.76393652e-01 3.09669048e-01 5.80756813e-02
-7.08913922e-01 1.39734045e-01 -1.05796981e+00 6.49345100e-01
-6.29377924e-03 6.55956334e-04 4.08897072e-01 8.53982747e-01
-6.09222829e-01 4.84753519e-01 -1.18710566e+00 -6.68215811e-01
1.87851071e-01 6.37342989e-01 -3.01925719e-01 4.58674841e-02
-1.16800380e+00 9.54258442e-01 2.40833182e-02 7.03565553e-02
-5.38658619e-01 -5.13012946e-01 -4.60969001e-01 -5.81001602e-02
-3.02328795e-01 -1.16044474e+00 5.54932594e-01 -9.28331614e-02
-1.42889166e+00 5.70407748e-01 -4.87936616e-01 -3.21332008e-01
-2.21960500e-01 -1.58717528e-01 -3.63015085e-01 1.72440961e-01
-1.86039135e-01 1.02564879e-01 4.90395993e-01 -9.09025908e-01
-5.23316860e-02 -9.78464901e-01 -3.83808255e-01 1.76687792e-01
6.83899105e-01 6.08690262e-01 1.24665618e-01 -7.42947400e-01
-1.46758780e-01 -1.23366809e+00 -4.64469999e-01 -7.18962669e-01
-1.64417520e-01 -3.57761681e-02 6.89007878e-01 -3.72206271e-01
8.79656255e-01 -1.79653037e+00 2.19977880e-03 4.78833348e-01
1.91223800e-01 8.24278593e-01 3.89477722e-02 9.86862302e-01
-2.24608239e-02 4.72185552e-01 -7.31787831e-02 8.23790729e-01
-5.39729297e-01 -5.69418848e-01 -4.54311192e-01 1.00151324e+00
-1.79835930e-02 9.82876897e-01 -8.54725122e-01 2.00190693e-01
-2.56919563e-01 9.04436231e-01 -5.33688784e-01 -3.12332157e-02
-5.76841831e-01 9.57954466e-01 -9.87389326e-01 9.17109191e-01
1.13645279e+00 -3.11190933e-01 1.09434545e+00 -6.56983927e-02
-7.89211914e-02 1.28038302e-01 -3.64606440e-01 1.00538397e+00
4.71012980e-01 1.28551602e-01 -1.41762540e-01 -1.78616017e-01
7.05828667e-01 7.09889531e-01 5.50276816e-01 -4.28061306e-01
2.74537385e-01 2.75648803e-01 3.13577116e-01 -2.39714310e-01
-1.54848784e-01 -4.42172110e-01 4.48211759e-01 -1.27956625e-02
-7.81108439e-01 5.10332108e-01 -8.41017440e-03 -9.03281793e-02
7.89821148e-01 -8.97133946e-02 5.62761545e-01 -4.05581921e-01
4.83994693e-01 2.93304652e-01 9.14045811e-01 4.03749086e-02
-7.28799477e-02 -1.67351857e-01 3.99554163e-01 -2.62978852e-01
-8.69070947e-01 -9.51244950e-01 -4.53106582e-01 3.67078602e-01
2.33923018e-01 -3.88140291e-01 -6.43807113e-01 -1.46679357e-01
4.10680361e-02 6.17515206e-01 -1.89281538e-01 2.72210807e-01
-5.60850024e-01 -8.51265073e-01 5.18499017e-01 -1.50484622e-01
-7.80165941e-02 -4.91690874e-01 -5.57623446e-01 3.04723203e-01
-1.12094376e-02 -6.40074253e-01 -6.04058743e-01 -1.12167530e-01
-1.20417655e+00 -1.57054257e+00 -6.29149497e-01 -6.26777291e-01
5.70602119e-01 2.82201767e-01 1.76863208e-01 1.41765475e-01
-4.26834188e-02 -2.37917423e-01 4.47107032e-02 -5.61974585e-01
-3.58943075e-01 -4.67835397e-01 7.36236274e-01 -5.94752431e-01
7.59593487e-01 -4.65354294e-01 -9.79044437e-01 4.50138420e-01
-3.58910829e-01 -9.22472179e-02 4.96440440e-01 5.98913252e-01
6.49575770e-01 -4.11316633e-01 7.58655310e-01 -6.43959701e-01
7.27813601e-01 -7.67316163e-01 -8.20968509e-01 -7.81515315e-02
-6.69589996e-01 -4.69470412e-01 6.30892575e-01 -1.70301367e-02
-9.84064579e-01 -1.77544802e-01 -1.64971739e-01 2.45599784e-02
1.22761175e-01 6.07981324e-01 -2.50827760e-01 1.46020114e-01
4.19827700e-01 1.80698574e-01 3.42220813e-01 -5.22555768e-01
-3.92640948e-01 4.30823833e-01 -2.08613396e-01 -1.85340852e-01
7.58969128e-01 4.31491882e-01 1.44031510e-01 -1.05600739e+00
1.43442631e-01 -5.20649314e-01 3.11488450e-01 3.87526751e-01
1.33344495e+00 -1.02420390e+00 -1.52913284e+00 4.74103212e-01
-1.41099322e+00 -2.63128858e-02 7.49665201e-01 1.03729105e+00
-1.06783047e-01 4.14591461e-01 -9.20201123e-01 -4.62313563e-01
-7.04499722e-01 -1.43678558e+00 3.70460272e-01 2.05857247e-01
-3.24723452e-01 -5.88848114e-01 9.59211230e-01 4.75324780e-01
3.42475981e-01 5.16443133e-01 1.47013223e+00 -9.61385369e-01
-9.19085562e-01 1.84844941e-01 -1.25798911e-01 -9.27421544e-03
2.07846776e-01 2.41571262e-01 -1.50592119e-01 -5.67397356e-01
1.41253658e-02 -1.48718217e-02 5.02887189e-01 7.89297283e-01
-1.08565567e-02 -4.20968890e-01 -8.47094297e-01 6.66587174e-01
1.47470129e+00 1.32675731e+00 9.81944859e-01 1.85062289e-01
4.33186620e-01 3.20305005e-02 8.17091703e-01 4.80964720e-01
-3.07610892e-02 8.62909794e-01 4.05695796e-01 4.20072451e-02
8.85755777e-01 2.87595391e-01 4.42959636e-01 2.66085476e-01
-7.74065197e-01 -2.60560840e-01 -8.50312948e-01 -1.62169896e-02
-1.41387761e+00 -9.78621483e-01 -4.34366703e-01 2.34219170e+00
8.85820210e-01 -5.48593342e-01 3.83210033e-02 -8.73953760e-01
3.90794933e-01 -2.82684654e-01 -6.96348011e-01 -6.81123853e-01
-5.11225574e-02 1.98389962e-01 4.25680608e-01 2.79765666e-01
-6.27243459e-01 5.66827655e-01 6.35396004e+00 2.98287392e-01
-1.10140896e+00 -2.92339057e-01 1.79573715e-01 1.98187009e-01
-3.76658946e-01 4.47167844e-01 -6.60958767e-01 3.28070670e-01
8.10527802e-01 -1.98433816e-01 2.21644118e-01 5.08458734e-01
7.80726492e-01 -1.17599204e-01 -6.89659595e-01 8.33391130e-01
-3.73243719e-01 -1.71598637e+00 -1.51464731e-01 7.13742495e-01
5.20074904e-01 1.20268539e-01 -3.00305903e-01 -1.62892893e-01
-2.87628800e-01 -1.13913572e+00 -2.93259978e-01 4.50496703e-01
7.57457674e-01 -9.40824091e-01 9.47510540e-01 8.93461406e-02
-1.05079782e+00 4.32534248e-01 -2.09831536e-01 2.10915074e-01
4.11378473e-01 3.95211428e-01 -1.23767972e+00 -3.35313343e-02
1.41607866e-01 1.34150892e-01 2.88050950e-01 8.52071941e-01
-6.77001923e-02 3.49435061e-01 -1.26831651e-01 1.05462022e-01
-1.05103366e-01 -7.89232314e-01 8.19115698e-01 8.82045865e-01
-1.58557687e-02 8.63950014e-01 2.03719631e-01 5.72010279e-01
-3.10861208e-02 3.98311317e-01 -4.06283081e-01 -5.17904580e-01
2.34615862e-01 8.03876996e-01 -5.98819971e-01 -2.71096081e-01
-5.00101089e-01 8.04030418e-01 -3.59401673e-01 6.48756504e-01
-7.09869325e-01 -1.50467426e-01 1.71745563e+00 3.32046509e-01
3.49819899e-01 -1.95168123e-01 3.28244686e-01 -8.44599068e-01
-5.54322660e-01 -1.22554100e+00 1.43365428e-01 -3.20094556e-01
-7.96549261e-01 3.87541205e-01 -1.43896788e-01 -1.00934637e+00
-2.38639981e-01 -3.90110821e-01 -4.14469808e-01 1.00466573e+00
-1.15011597e+00 -8.89205933e-01 3.46426755e-01 2.48934895e-01
1.65320531e-01 -2.37066280e-02 1.26364744e+00 -1.11459404e-01
-6.24268651e-01 1.74242072e-02 6.62943721e-01 -2.53897369e-01
6.91162407e-01 -5.17889500e-01 1.83676988e-01 3.52447659e-01
-7.84300089e-01 1.58703864e+00 9.25912261e-01 -1.40557873e+00
-1.64468408e+00 -1.00956619e+00 8.91418636e-01 -2.34112054e-01
4.83685255e-01 -5.56320585e-02 -7.13467181e-01 8.01048875e-01
2.05568060e-01 -9.02976930e-01 1.51889682e+00 -3.63170892e-01
-3.04043233e-01 6.94409728e-01 -1.35906494e+00 8.72268379e-01
4.17280585e-01 -1.51708737e-01 -5.26782751e-01 3.72644961e-01
6.40972912e-01 -3.54663998e-01 -9.08031881e-01 7.90808737e-01
9.09822524e-01 -9.24663305e-01 9.42403734e-01 -7.72008002e-01
-4.38470721e-01 -9.03954268e-01 1.55179240e-02 -1.02274466e+00
-4.40151930e-01 -7.62357593e-01 1.31528825e-01 3.09113413e-01
3.30951601e-01 -7.71487772e-01 5.09362400e-01 3.15107703e-01
2.82804996e-01 -9.65626121e-01 -7.15694547e-01 -6.87596560e-01
-2.78348863e-01 1.19472213e-01 3.24042797e-01 7.85943866e-01
2.32931241e-01 5.24771988e-01 -3.83536458e-01 1.67504176e-01
4.00632590e-01 2.20823690e-01 8.14740777e-01 -1.35674298e+00
-1.37960970e-01 -1.84646502e-01 -2.30148852e-01 -5.58815658e-01
-3.52705300e-01 -5.58925509e-01 -7.62938201e-01 -1.62306535e+00
5.26152134e-01 -2.51705825e-01 -4.77676466e-03 2.18004748e-01
1.70083910e-01 2.74899658e-02 1.56402364e-01 5.17192960e-01
4.57594804e-02 3.08044672e-01 9.93104219e-01 -1.16835944e-02
-8.17739964e-01 7.17953593e-02 -6.58384502e-01 8.33803236e-01
9.11813736e-01 -5.49922287e-01 -4.93163198e-01 5.72919190e-01
4.82962161e-01 2.01656312e-01 -1.79505214e-01 1.80372640e-01
-4.50577766e-01 -8.56480896e-01 3.94928455e-01 -9.49112535e-01
3.99852544e-01 -8.04421067e-01 1.01261222e+00 1.10692489e+00
8.84128630e-01 7.10910140e-03 5.67808867e-01 4.21548635e-01
4.04260933e-01 2.93089569e-01 8.48000586e-01 5.77045120e-02
-1.14939049e-01 2.16711208e-01 -1.19178545e+00 -6.34606481e-02
1.21795976e+00 -7.41696000e-01 -3.15233469e-01 2.00604469e-01
-8.36253762e-01 -9.47360992e-02 1.09860051e+00 9.41820368e-02
6.64360285e-01 -5.87740660e-01 -8.20426404e-01 1.95963934e-01
2.67012566e-01 -6.71637952e-01 5.89358389e-01 1.29139316e+00
-1.05404878e+00 9.12594914e-01 -2.02073470e-01 -6.38276577e-01
-1.50919235e+00 9.09878671e-01 3.65086824e-01 -2.58649051e-01
-2.73119092e-01 3.21165442e-01 5.71245432e-01 2.68881351e-01
-1.08378321e-01 -1.45064756e-01 -4.50593174e-01 -1.27275571e-01
6.78832889e-01 3.67810696e-01 -1.09505139e-01 -8.60981822e-01
-1.06614268e+00 6.84968114e-01 -2.34666854e-01 8.78257692e-01
1.27756977e+00 3.37711461e-02 -6.03845239e-01 -2.24478140e-01
9.45418239e-01 6.48552597e-01 -5.32562733e-01 1.71138272e-01
-2.75759935e-01 -5.27518213e-01 -6.08222246e-01 -8.56149256e-01
-4.76363838e-01 -4.95970175e-02 8.11434388e-01 -5.14575601e-01
8.13882470e-01 4.92865145e-02 1.11372447e+00 2.05331743e-01
6.64135218e-01 -4.38836068e-01 -5.04993439e-01 3.16352069e-01
8.20297360e-01 -8.36153150e-01 -1.06705122e-01 -3.42498451e-01
-5.12465835e-01 7.11818457e-01 3.22786868e-01 2.59366810e-01
3.44771504e-01 2.72341352e-02 1.80624500e-01 -4.97665703e-01
-8.16824138e-01 2.37959474e-01 1.95571572e-01 5.71436822e-01
7.56541491e-01 2.60975868e-01 -1.19147372e+00 5.10396600e-01
4.78834063e-01 1.57110661e-01 3.52845013e-01 6.67137146e-01
-5.28508306e-01 -1.54442120e+00 -5.67237318e-01 2.14289770e-01
-6.40613794e-01 -4.00364935e-01 -3.64895612e-01 7.83107460e-01
1.75588503e-02 7.59677589e-01 -1.56975880e-01 1.08555667e-01
1.64019078e-01 2.15457812e-01 3.92724931e-01 -5.33992827e-01
-4.59006190e-01 7.42501020e-01 4.72296812e-02 -3.56776834e-01
-3.28261077e-01 -4.65798140e-01 -1.97461712e+00 -6.25091076e-01
-2.28034094e-01 5.23495078e-01 7.60464013e-01 8.08708906e-01
8.45982850e-01 -8.45672563e-02 5.19157529e-01 -2.96763957e-01
-1.59674063e-01 -6.46393538e-01 -7.93186724e-01 -3.12477220e-02
2.42763743e-01 -3.72822255e-01 -2.29395870e-02 -1.03440426e-01] | [4.691133975982666, 5.106047630310059] |
4f3c1497-0a8f-4955-b805-8d6d8291a60d | multi-modal-prompting-for-low-shot-temporal | 2303.11732 | null | https://arxiv.org/abs/2303.11732v1 | https://arxiv.org/pdf/2303.11732v1.pdf | Multi-modal Prompting for Low-Shot Temporal Action Localization | In this paper, we consider the problem of temporal action localization under low-shot (zero-shot & few-shot) scenario, with the goal of detecting and classifying the action instances from arbitrary categories within some untrimmed videos, even not seen at training time. We adopt a Transformer-based two-stage action localization architecture with class-agnostic action proposal, followed by open-vocabulary classification. We make the following contributions. First, to compensate image-text foundation models with temporal motions, we improve category-agnostic action proposal by explicitly aligning embeddings of optical flows, RGB and texts, which has largely been ignored in existing low-shot methods. Second, to improve open-vocabulary action classification, we construct classifiers with strong discriminative power, i.e., avoid lexical ambiguities. To be specific, we propose to prompt the pre-trained CLIP text encoder either with detailed action descriptions (acquired from large-scale language models), or visually-conditioned instance-specific prompt vectors. Third, we conduct thorough experiments and ablation studies on THUMOS14 and ActivityNet1.3, demonstrating the superior performance of our proposed model, outperforming existing state-of-the-art approaches by one significant margin. | ['Weidi Xie', 'Yanfeng Wang', 'Qi Tian', 'Xiaopeng Zhang', 'Ya zhang', 'Peisen Zhao', 'Zeqian Li', 'Chen Ju'] | 2023-03-21 | null | null | null | null | ['action-classification', 'action-localization'] | ['computer-vision', 'computer-vision'] | [ 4.47238415e-01 -1.53785929e-01 -5.14297664e-01 -1.51178852e-01
-7.58666039e-01 -4.57184821e-01 7.73804724e-01 -3.82511586e-01
-5.51012993e-01 5.76600611e-01 8.07154059e-01 2.63415664e-01
-3.40745747e-02 -2.20235631e-01 -5.87415218e-01 -7.29484200e-01
1.48330018e-01 1.41578913e-01 4.77364302e-01 8.49998277e-03
2.54293412e-01 6.76291287e-02 -1.41512954e+00 4.32639122e-01
5.18781900e-01 1.18363440e+00 1.11843847e-01 3.91753048e-01
1.13340944e-01 1.55099404e+00 -2.86328971e-01 -1.35055706e-01
1.69666395e-01 -5.77537060e-01 -9.03413117e-01 4.76453692e-01
5.77042818e-01 -7.10283875e-01 -8.17271471e-01 8.74910891e-01
3.40303659e-01 4.02005702e-01 6.22015476e-01 -1.40171528e+00
-8.48405123e-01 4.30763900e-01 -4.25276995e-01 4.12704349e-01
3.96074653e-01 7.78087795e-01 1.06950188e+00 -9.63506997e-01
7.93701351e-01 1.16346169e+00 1.74843132e-01 6.26245379e-01
-1.01402688e+00 -6.49067938e-01 4.71328616e-01 7.08566189e-01
-1.50159240e+00 -6.81598067e-01 8.86288941e-01 -5.88576376e-01
1.02084517e+00 -1.39937684e-01 7.73859084e-01 1.63849616e+00
4.89939414e-02 1.18552542e+00 9.22685325e-01 2.08910983e-02
3.24025035e-01 -2.39342958e-01 -2.26474330e-01 6.83877587e-01
-8.91171470e-02 -3.08181122e-02 -8.46430957e-01 2.67646790e-01
8.11961830e-01 3.37997705e-01 -4.36425865e-01 -6.93166733e-01
-1.63433862e+00 5.21139920e-01 3.09067249e-01 4.15239751e-01
-5.25646925e-01 4.46460634e-01 6.17947400e-01 9.90134254e-02
3.71080458e-01 1.81265891e-01 -4.21234369e-01 -6.45564437e-01
-9.00739193e-01 -1.45188540e-01 3.68299186e-01 1.14644039e+00
6.49979889e-01 2.11620763e-01 -7.24184334e-01 5.09382546e-01
7.05606565e-02 3.36436182e-01 8.27544391e-01 -9.22086298e-01
6.24158382e-01 5.16918182e-01 8.06970671e-02 -7.12883651e-01
-5.08025587e-02 -1.60549834e-01 -5.59070408e-01 -1.40895739e-01
3.84404391e-01 1.21357895e-01 -9.08060253e-01 1.63077772e+00
1.18365310e-01 6.91740572e-01 2.03959972e-01 1.09067905e+00
5.03721833e-01 4.12138671e-01 3.08839172e-01 -2.49411687e-01
1.28930247e+00 -1.45109761e+00 -8.33858371e-01 -3.99863303e-01
8.47662330e-01 -3.73187810e-01 1.26096058e+00 2.24793896e-01
-7.18095899e-01 -6.46209240e-01 -7.96959102e-01 -2.18618467e-01
-2.70519346e-01 3.53027463e-01 4.78751540e-01 8.86933282e-02
-6.78366899e-01 5.60023904e-01 -9.92946267e-01 -6.69069171e-01
8.10247302e-01 -4.07998301e-02 -5.24703562e-01 -1.34304017e-01
-1.02291703e+00 7.41472006e-01 5.08644402e-01 -1.86641857e-01
-1.52929533e+00 -5.04019022e-01 -1.01681268e+00 -1.79967866e-03
7.91780412e-01 -3.85706604e-01 1.15515614e+00 -1.23046863e+00
-1.51455426e+00 6.44873977e-01 -1.55140370e-01 -4.21758354e-01
5.44637084e-01 -4.20001924e-01 -3.16593289e-01 5.79012454e-01
2.50628293e-01 7.51612186e-01 1.01551390e+00 -7.88552999e-01
-6.38097286e-01 -2.09009156e-01 5.02335072e-01 3.44693601e-01
-6.59169555e-01 -5.56652173e-02 -5.24609029e-01 -7.63251245e-01
-1.52645120e-02 -8.06999028e-01 -1.08291455e-01 4.77261811e-01
-2.01168999e-01 -2.55088121e-01 8.64334226e-01 -6.83919370e-01
1.03946400e+00 -2.42416739e+00 2.97379583e-01 -5.51153004e-01
7.57510364e-02 2.10205838e-01 -3.04097176e-01 3.60530376e-01
-9.99298245e-02 -2.81542987e-01 -1.56771943e-01 -3.17079842e-01
4.09593619e-02 3.30354810e-01 -4.00606513e-01 6.65921867e-01
4.41784471e-01 1.11795795e+00 -1.15162480e+00 -6.54443324e-01
5.42519808e-01 2.94516802e-01 -5.10554612e-01 1.58682019e-01
-2.62104094e-01 7.78229952e-01 -7.22679377e-01 8.51274192e-01
1.18733600e-01 -3.38178098e-01 -5.12113683e-02 -5.63280582e-01
-6.25963360e-02 2.21012473e-01 -1.19589472e+00 2.33269954e+00
-5.02766073e-01 5.14432490e-01 -4.64257360e-01 -1.34134662e+00
3.44457597e-01 3.95318836e-01 8.37570369e-01 -7.11943328e-01
1.59893796e-01 -8.97366256e-02 -9.17664170e-02 -8.02913666e-01
2.26097688e-01 1.35804355e-01 -4.87366086e-03 4.02077705e-01
5.50960302e-01 3.75692844e-01 2.97324389e-01 3.02068144e-01
1.29576170e+00 7.96647310e-01 5.10319531e-01 2.92320829e-02
5.64757407e-01 -7.20268786e-02 4.82079178e-01 5.56993604e-01
-7.99832880e-01 6.70712590e-01 4.06116277e-01 -1.66288033e-01
-6.84521317e-01 -7.98203409e-01 1.67147443e-01 1.25250638e+00
3.78929287e-01 -5.21423757e-01 -4.13843364e-01 -9.07020032e-01
-2.46744737e-01 4.88241225e-01 -7.17180848e-01 -2.67584741e-01
-5.24289548e-01 -1.77540943e-01 4.84317452e-01 9.22495902e-01
7.46758878e-01 -1.13070059e+00 -8.62563312e-01 2.20811725e-01
-4.47818786e-01 -1.54030240e+00 -7.36443222e-01 1.19374365e-01
-6.95268869e-01 -1.23403382e+00 -7.80242443e-01 -6.49841666e-01
5.27079523e-01 3.70804518e-01 6.62763357e-01 -2.57718563e-01
-3.21132094e-01 8.71609926e-01 -7.49062717e-01 1.27140716e-01
2.57851034e-01 -3.96983564e-01 1.32720187e-01 5.15410006e-01
5.26762307e-01 -5.79445899e-01 -9.31902051e-01 2.99989849e-01
-7.40647256e-01 -1.85583869e-03 7.16209292e-01 7.22727537e-01
5.95342696e-01 -2.49648169e-01 2.40841433e-01 -2.47136295e-01
-3.53810308e-03 -3.30220401e-01 1.53085508e-03 2.99743861e-01
-2.16414124e-01 3.04138009e-02 6.70401096e-01 -8.92033935e-01
-9.52099025e-01 3.47798049e-01 2.22913831e-01 -9.93174911e-01
-2.99368262e-01 3.99274342e-02 -2.73054868e-01 -3.20316106e-02
4.92651939e-01 6.71533406e-01 -1.76758066e-01 -3.02793145e-01
6.42987490e-01 6.21901274e-01 4.65221643e-01 -3.59102190e-01
6.35694444e-01 7.97213495e-01 -3.30603004e-01 -6.39518023e-01
-1.06621504e+00 -7.92021334e-01 -9.50636685e-01 -3.73958379e-01
1.21713686e+00 -1.24476981e+00 -4.27392989e-01 3.34182084e-01
-9.78795409e-01 -4.72407132e-01 -4.64510024e-01 7.96790600e-01
-9.69825566e-01 5.07180274e-01 -5.05074561e-01 -6.19533181e-01
-5.87557666e-02 -1.03192055e+00 1.32575226e+00 1.39858830e-03
-2.13262662e-01 -7.27584898e-01 -1.00827180e-01 4.45153564e-01
1.45546794e-01 -2.48178728e-02 2.90000260e-01 -7.93061674e-01
-7.60899305e-01 -5.81465401e-02 -2.49992117e-01 3.91882271e-01
2.10254312e-01 -3.36379498e-01 -9.47621047e-01 -2.44051382e-01
-1.27896011e-01 -7.80360758e-01 9.74507093e-01 2.25960970e-01
1.24085915e+00 -1.93039089e-01 -3.98523718e-01 6.43255532e-01
1.18166637e+00 7.07590505e-02 6.57592893e-01 2.43332967e-01
8.44975054e-01 4.56174195e-01 1.02829778e+00 6.39068544e-01
3.12597275e-01 7.05962658e-01 4.25657600e-01 2.21721128e-01
-3.86772424e-01 -4.26225185e-01 6.51736140e-01 5.24150014e-01
-1.07334688e-01 -1.03683800e-01 -4.54239219e-01 7.05868006e-01
-2.05759859e+00 -1.14318347e+00 2.54948258e-01 1.79018831e+00
6.66158676e-01 1.17143519e-01 1.08696543e-01 -3.92780751e-02
4.72861916e-01 7.36165047e-01 -7.24979281e-01 2.99840808e-01
1.77209955e-02 7.45645389e-02 4.47003961e-01 7.83326775e-02
-1.44345903e+00 1.23039842e+00 4.86622715e+00 8.52360129e-01
-1.11543834e+00 4.63142157e-01 1.51261508e-01 -4.81111705e-01
3.26647669e-01 2.27399766e-01 -5.34975231e-01 5.09743631e-01
5.97280920e-01 -1.42704174e-01 2.93937027e-01 9.16703582e-01
4.29628879e-01 -8.06197152e-02 -1.44689560e+00 1.15241361e+00
4.72492188e-01 -1.15091300e+00 1.01946946e-02 -1.36709675e-01
5.37629604e-01 -8.05654973e-02 -2.36621633e-01 7.04739273e-01
7.50967115e-02 -6.25289261e-01 8.82673800e-01 5.61620772e-01
8.51766706e-01 -1.41976938e-01 2.69988239e-01 2.59928197e-01
-1.44407439e+00 -4.02733743e-01 -2.75123924e-01 -2.02454552e-01
3.90671790e-01 -2.21023406e-03 -8.22091177e-02 4.60558295e-01
5.89866161e-01 1.61179924e+00 -4.36842859e-01 7.91101933e-01
-3.57901603e-01 6.02015436e-01 8.02403595e-03 2.87831165e-02
5.14585376e-01 -6.29353244e-03 3.20903361e-01 9.06450152e-01
1.43050015e-01 3.91952425e-01 5.85745037e-01 6.41312838e-01
1.32693544e-01 7.23213255e-02 -6.37354076e-01 -3.39892149e-01
1.84247091e-01 1.10211766e+00 -5.38974881e-01 -5.54014742e-01
-7.42867827e-01 1.42787933e+00 2.14343876e-01 5.33640444e-01
-1.06429160e+00 -2.46327609e-01 6.93249822e-01 -6.58776984e-02
5.34688056e-01 -1.75811529e-01 2.51433522e-01 -1.53254378e+00
9.66602489e-02 -6.34783864e-01 3.80394936e-01 -9.22184050e-01
-9.44159329e-01 1.74638569e-01 4.65593040e-02 -1.83161271e+00
-1.65713474e-01 -5.98320305e-01 -5.33873498e-01 3.18378568e-01
-1.40323913e+00 -1.42824972e+00 -3.35682154e-01 8.45592201e-01
1.14011860e+00 -1.50242791e-01 5.29029727e-01 4.34132695e-01
-5.85203588e-01 3.64529192e-01 -9.29052755e-02 3.46379101e-01
9.47225034e-01 -8.67520750e-01 -6.69532791e-02 9.77142632e-01
3.66294831e-01 3.23534787e-01 4.26633716e-01 -5.97231269e-01
-1.54052567e+00 -1.37911522e+00 5.46549618e-01 -5.80748916e-01
1.17271221e+00 -3.21875930e-01 -6.85372233e-01 8.74879897e-01
-2.66065132e-02 4.68615502e-01 4.85438943e-01 -3.43576550e-01
-3.37665796e-01 -4.08603102e-02 -6.91982925e-01 7.18763590e-01
1.56596553e+00 -7.68944085e-01 -7.61671722e-01 5.83758175e-01
7.11589217e-01 -1.23975597e-01 -7.39472866e-01 2.34225839e-01
6.29118621e-01 -6.34248555e-01 9.45181489e-01 -8.48824978e-01
6.94190919e-01 -2.59623140e-01 -3.57685298e-01 -9.67898309e-01
-3.86454046e-01 -3.80812496e-01 -4.35166597e-01 1.04396594e+00
-1.90886229e-01 -2.16087118e-01 6.20632291e-01 1.71609446e-01
-2.02281088e-01 -6.19535148e-01 -9.88450348e-01 -9.68729258e-01
-2.72057325e-01 -5.18949449e-01 1.17601477e-01 9.20837939e-01
1.71227664e-01 3.29783767e-01 -8.73234510e-01 -4.38068882e-02
4.12423730e-01 5.02669103e-02 7.82470822e-01 -6.28918409e-01
-3.84563714e-01 -2.83899337e-01 -7.23748147e-01 -1.53597641e+00
4.14022237e-01 -6.46941304e-01 2.37314895e-01 -1.35509241e+00
3.37571353e-01 2.64957279e-01 -6.69876695e-01 8.42094064e-01
1.34353796e-02 4.33914304e-01 1.90212607e-01 3.88476133e-01
-1.18316793e+00 1.07795858e+00 1.46252108e+00 -4.12779361e-01
1.98968366e-01 -4.26445156e-01 -4.29636985e-01 7.13202596e-01
3.95217806e-01 -3.53080928e-01 -7.06490517e-01 -3.67714435e-01
-3.83629799e-01 -1.04769729e-02 8.26245904e-01 -1.20210779e+00
1.48800343e-01 -4.79349434e-01 1.27738446e-01 -3.89808148e-01
5.12523532e-01 -7.97562182e-01 -4.23359871e-01 4.15666729e-01
-5.64239860e-01 -4.83341306e-01 -1.84038252e-01 9.97196198e-01
-2.62188047e-01 1.03288153e-02 6.86803639e-01 -1.34889826e-01
-1.47158313e+00 6.03480160e-01 -3.10193479e-01 1.98238075e-01
1.32426775e+00 -3.64214420e-01 -3.97343427e-01 -3.19049329e-01
-7.70288408e-01 2.74188429e-01 4.42992926e-01 6.86858296e-01
6.11787081e-01 -1.40390193e+00 -3.90525520e-01 1.83705892e-02
7.41270542e-01 -4.12571549e-01 4.92498815e-01 1.33515716e+00
-9.14538652e-02 4.78928030e-01 -2.72453308e-01 -6.35054946e-01
-8.66400778e-01 8.24569762e-01 2.14688003e-01 5.64080738e-02
-1.08531117e+00 7.22950041e-01 4.87153381e-01 1.76070649e-02
5.02633750e-01 -2.94647217e-01 -2.64124662e-01 3.29359435e-02
5.57899952e-01 1.26133919e-01 -4.50408846e-01 -8.27845216e-01
-5.83861351e-01 6.13417804e-01 1.02154233e-01 1.65308267e-02
1.01479876e+00 -1.67762637e-01 3.81446779e-01 6.56545699e-01
1.18106902e+00 -5.33455729e-01 -1.71707070e+00 -4.53938365e-01
-2.03396350e-01 -6.92804277e-01 -4.44813780e-02 -6.21416986e-01
-1.09894907e+00 1.08476162e+00 7.53766060e-01 -4.20902848e-01
1.03999376e+00 1.89568549e-01 7.85204768e-01 4.54654545e-01
4.67339396e-01 -1.30722404e+00 7.83332586e-01 4.07556713e-01
7.90796697e-01 -1.44563854e+00 -1.35836348e-01 -5.42858429e-02
-9.81449008e-01 8.70270312e-01 9.55610991e-01 -5.52748181e-02
4.31767225e-01 -2.28943869e-01 -1.52975053e-01 -7.35795423e-02
-8.70870590e-01 -5.89953244e-01 2.25397617e-01 5.97703815e-01
1.40002742e-01 -2.45553732e-01 -3.01805019e-01 5.75485706e-01
5.10070980e-01 2.87725389e-01 4.64041531e-01 1.09266710e+00
-4.61969256e-01 -5.76479137e-01 7.77159631e-02 3.39764893e-01
-1.12731099e-01 -1.47580519e-01 -2.14558303e-01 6.87444866e-01
9.06275585e-02 7.81318009e-01 1.52403712e-01 -3.76825392e-01
2.37736136e-01 2.40210682e-01 4.79127944e-01 -6.93868637e-01
7.03546330e-02 4.67193127e-02 8.93032923e-03 -1.10212183e+00
-7.61994421e-01 -6.98040962e-01 -1.11498797e+00 3.17527145e-01
-1.12849884e-01 -2.85919935e-01 1.52241901e-01 1.26330853e+00
3.85913759e-01 6.43499315e-01 4.22468573e-01 -9.05716360e-01
-7.01431513e-01 -1.11572731e+00 -6.34010613e-01 7.12179065e-01
2.72958010e-01 -1.12993670e+00 -4.26498532e-01 4.03559357e-01] | [8.666393280029297, 0.7698109745979309] |
0434d656-4a7f-4b5a-8994-bf243d982f10 | wad-cmsn-wasserstein-distance-based-cross | 2202.05465 | null | https://arxiv.org/abs/2202.05465v1 | https://arxiv.org/pdf/2202.05465v1.pdf | WAD-CMSN: Wasserstein Distance based Cross-Modal Semantic Network for Zero-Shot Sketch-Based Image Retrieval | Zero-shot sketch-based image retrieval (ZSSBIR), as a popular studied branch of computer vision, attracts wide attention recently. Unlike sketch-based image retrieval (SBIR), the main aim of ZSSBIR is to retrieve natural images given free hand-drawn sketches that may not appear during training. Previous approaches used semantic aligned sketch-image pairs or utilized memory expensive fusion layer for projecting the visual information to a low dimensional subspace, which ignores the significant heterogeneous cross-domain discrepancy between highly abstract sketch and relevant image. This may yield poor performance in the training phase. To tackle this issue and overcome this drawback, we propose a Wasserstein distance based cross-modal semantic network (WAD-CMSN) for ZSSBIR. Specifically, it first projects the visual information of each branch (sketch, image) to a common low dimensional semantic subspace via Wasserstein distance in an adversarial training manner. Furthermore, identity matching loss is employed to select useful features, which can not only capture complete semantic knowledge, but also alleviate the over-fitting phenomenon caused by the WAD-CMSN model. Experimental results on the challenging Sketchy (Extended) and TU-Berlin (Extended) datasets indicate the effectiveness of the proposed WAD-CMSN model over several competitors. | ['Jia Cai', 'Zhensheng Hu', 'Guanglong Xu'] | 2022-02-11 | null | null | null | null | ['sketch-based-image-retrieval'] | ['computer-vision'] | [ 1.14129432e-01 -3.10492605e-01 -1.47086769e-01 -1.54373556e-01
-9.45230186e-01 -3.56289804e-01 7.03906357e-01 -3.55567276e-01
-2.55635142e-01 5.40779412e-01 2.49706935e-02 2.00914338e-01
-4.91286755e-01 -6.86710060e-01 -5.05220950e-01 -7.51185000e-01
5.16729712e-01 3.24638098e-01 1.37073457e-01 -2.00231969e-01
3.98372382e-01 5.04191637e-01 -1.43015981e+00 9.38287377e-02
6.58561707e-01 9.70286369e-01 3.61704826e-01 1.09110557e-01
-5.71043432e-01 3.44811678e-01 -5.07245243e-01 -4.98475283e-01
4.19550300e-01 -1.34840250e-01 -4.24492508e-01 7.42153078e-02
8.26557577e-01 -4.07286912e-01 -8.09654057e-01 1.24704981e+00
4.61097956e-01 3.67488354e-01 7.78473020e-01 -1.59877455e+00
-1.07173049e+00 -1.68900676e-02 -6.29304051e-01 -1.15144521e-01
2.54374534e-01 -2.23844573e-01 9.19431627e-01 -1.43703496e+00
1.06783378e+00 1.53421783e+00 4.66610342e-01 7.15709329e-01
-1.07353640e+00 -9.49881494e-01 1.81201026e-01 2.12646052e-01
-1.81326687e+00 -2.25296840e-01 1.35731518e+00 -1.95197076e-01
2.22343013e-01 1.92563534e-01 3.99207026e-01 1.22588885e+00
-2.58097947e-01 1.22448301e+00 8.33750129e-01 -2.24339411e-01
1.00922741e-01 2.66259104e-01 -1.22910738e-01 6.43552303e-01
5.44272698e-02 -2.47964426e-03 -6.69300377e-01 -3.74194026e-01
1.07953680e+00 5.62788308e-01 -3.10264647e-01 -9.65181410e-01
-1.14634430e+00 9.14218247e-01 5.47195971e-01 3.67521495e-01
-5.18082976e-02 -3.52785550e-02 4.44730580e-01 4.83161688e-01
3.75051528e-01 1.45370513e-01 -4.53979746e-02 3.54450792e-01
-1.08732200e+00 2.91046262e-01 4.10124063e-01 1.15820098e+00
6.77228153e-01 2.13542674e-02 -2.60163009e-01 1.26507020e+00
2.73488700e-01 7.24325538e-01 4.55639660e-01 -8.46064925e-01
5.92016816e-01 5.42376995e-01 3.49289365e-02 -1.37434375e+00
3.13831896e-01 -1.27130061e-01 -1.15532815e+00 1.66739658e-01
2.71273196e-01 5.69993973e-01 -7.33347058e-01 1.69195592e+00
7.43572190e-02 2.48999730e-01 1.23055845e-01 1.24237514e+00
1.04504395e+00 5.76143026e-01 4.51211855e-02 1.09497592e-01
1.02473116e+00 -9.77614641e-01 -5.39242327e-01 -5.30699454e-02
1.38470507e-03 -8.58774066e-01 1.17774832e+00 1.29188567e-01
-9.20661211e-01 -6.64687514e-01 -1.16545701e+00 -2.98282593e-01
-5.51407099e-01 2.56461471e-01 3.46952826e-01 2.40973100e-01
-8.91812742e-01 5.13426661e-01 -2.88531572e-01 -3.51473153e-01
6.33066773e-01 2.62848008e-03 -7.34416008e-01 -6.40503347e-01
-1.24164188e+00 5.35356998e-01 2.50499099e-01 1.49795324e-01
-6.74834251e-01 -7.43434906e-01 -8.67779136e-01 1.58878062e-02
5.01388669e-01 -6.63150787e-01 5.05598605e-01 -9.22339857e-01
-1.14582634e+00 7.82740295e-01 -2.62777805e-02 1.08089566e-01
7.27579832e-01 -4.10257950e-02 -3.77521008e-01 3.07054639e-01
1.79050550e-01 8.36074829e-01 1.31616724e+00 -1.47331977e+00
-1.02174222e-01 -5.76690376e-01 -1.35799319e-01 2.51295984e-01
-4.12757188e-01 -3.50305468e-01 -7.70576537e-01 -1.06175268e+00
3.87650013e-01 -8.42563033e-01 7.09506944e-02 6.02655172e-01
-8.01045522e-02 -2.85374075e-01 1.18777788e+00 -6.43857121e-01
9.07881081e-01 -2.26567698e+00 2.16950879e-01 2.48758391e-01
-9.12982151e-02 5.11020899e-01 -6.26044810e-01 5.24106562e-01
3.59679549e-03 -1.75396726e-01 -2.40659356e-01 -3.48356932e-01
1.78836975e-02 3.41425121e-01 -6.62936032e-01 3.71682495e-01
1.93198204e-01 1.06419551e+00 -1.05860615e+00 -6.99122310e-01
5.26868522e-01 7.13463426e-01 -2.81259000e-01 2.38383174e-01
6.09699897e-02 2.54357010e-01 -6.77982330e-01 8.57245862e-01
1.05743182e+00 -3.81016731e-02 -8.34482089e-02 -3.52784246e-01
3.50239009e-01 -5.09502172e-01 -1.23202121e+00 2.39745045e+00
-4.54882771e-01 4.77126628e-01 2.00660378e-02 -1.04942691e+00
1.23347247e+00 1.51807562e-01 4.30285066e-01 -9.13980365e-01
-1.66955814e-01 2.33074501e-01 -7.40811288e-01 -2.99524724e-01
3.35359305e-01 -2.95813918e-01 -3.86214517e-02 1.53113648e-01
1.93089232e-01 -4.89363782e-02 -2.16064513e-01 4.91263479e-01
5.87307036e-01 2.55268812e-01 -9.39065367e-02 -2.16616303e-01
9.20790434e-01 -2.82233775e-01 3.88061345e-01 6.10143542e-01
-1.77865848e-01 9.01523471e-01 1.32468268e-01 -4.85861421e-01
-1.04015756e+00 -1.34128511e+00 -4.30122204e-02 7.21234977e-01
7.40055084e-01 -2.43746296e-01 -3.98609042e-01 -7.04916418e-01
1.67853728e-01 4.35257494e-01 -4.97291863e-01 -3.42625767e-01
-3.02292645e-01 -1.32564548e-02 4.72094059e-01 3.59413624e-01
5.23213387e-01 -1.20448768e+00 -5.05579114e-02 -5.04984194e-03
-2.78609872e-01 -9.25823152e-01 -8.19917560e-01 -5.14616489e-01
-6.88097119e-01 -1.09551823e+00 -1.45138061e+00 -8.58130634e-01
8.55049491e-01 8.34018767e-01 8.42358470e-01 1.02814019e-01
-6.73688352e-01 7.20350742e-01 -1.33973792e-01 -4.12075482e-02
1.48637099e-02 -2.37533763e-01 -1.18236020e-01 2.84114748e-01
4.58473086e-01 -4.83571172e-01 -7.75944769e-01 4.12542582e-01
-1.16147041e+00 -2.16668658e-02 6.32696152e-01 1.35143757e+00
5.99048555e-01 -2.02491388e-01 7.19149649e-01 -6.03355825e-01
5.55609763e-01 -2.58288771e-01 -4.78870451e-01 6.81497097e-01
-5.24693012e-01 1.21492140e-01 5.49491704e-01 -5.43653965e-01
-1.04296386e+00 -1.44657996e-02 6.81753084e-02 -1.37079370e+00
9.13727060e-02 1.64522886e-01 -3.62239391e-01 -2.81719953e-01
1.51348189e-01 6.25079989e-01 3.37387055e-01 -4.63950574e-01
5.10611236e-01 5.39360702e-01 5.32714248e-01 -7.62581527e-01
1.09036624e+00 5.94925880e-01 9.14910138e-02 -8.95930529e-01
-7.46277273e-01 -8.00570428e-01 -5.85643947e-01 -2.91209985e-02
5.58341324e-01 -1.01569510e+00 -4.11807358e-01 3.00971955e-01
-1.19382226e+00 3.83227229e-01 -2.03805789e-01 2.21245989e-01
-5.08170664e-01 7.54414737e-01 -2.27184951e-01 -9.58734393e-01
-6.08118773e-01 -8.27692628e-01 1.51169050e+00 1.85769439e-01
1.25002474e-01 -9.76799786e-01 -9.76588875e-02 4.01890069e-01
2.80678600e-01 1.42661139e-01 8.64963472e-01 -4.95317429e-01
-7.17324555e-01 -4.20806348e-01 -7.17624009e-01 4.44218516e-01
-1.67501383e-02 -3.57744426e-01 -8.58404279e-01 -4.85776246e-01
-2.41754442e-01 -5.29038966e-01 9.89821553e-01 -2.87975576e-02
1.19476962e+00 -6.18595816e-02 -3.11671138e-01 5.32031000e-01
1.74476683e+00 1.15410667e-02 6.72144890e-01 1.23942837e-01
7.20954776e-01 5.06360114e-01 1.00165892e+00 2.30855450e-01
8.71910080e-02 7.46741474e-01 2.29444697e-01 -8.13561156e-02
-2.82077014e-01 -7.23908484e-01 8.97515267e-02 7.15211034e-01
1.89504668e-01 8.46588165e-02 -4.08169687e-01 5.81829727e-01
-1.95288610e+00 -9.88634527e-01 3.28876168e-01 2.40837812e+00
4.50448245e-01 -2.41903961e-01 -2.26244479e-01 3.82362753e-02
7.54085958e-01 3.91831130e-01 -5.66948593e-01 1.03367537e-01
-2.88915932e-01 2.41386965e-01 2.29747981e-01 1.92195341e-01
-8.97119582e-01 9.70422506e-01 4.74621868e+00 1.43235207e+00
-9.27393615e-01 1.34703010e-01 2.08910376e-01 2.26613656e-02
-4.21451896e-01 -1.20956346e-01 -3.20663571e-01 4.33349848e-01
6.08879188e-03 -1.76311925e-01 5.70306540e-01 9.64348137e-01
-2.76106626e-01 2.64323920e-01 -9.44799960e-01 1.56311798e+00
4.28834051e-01 -1.28371847e+00 5.66654086e-01 -1.18394062e-01
6.58897936e-01 -3.37276489e-01 1.35044947e-01 4.41658735e-01
-1.95666865e-01 -8.26095939e-01 4.60238338e-01 8.04209232e-01
1.20136845e+00 -8.64950538e-01 5.14879167e-01 2.78892726e-01
-1.27870476e+00 2.86986995e-02 -9.53521729e-01 4.78902608e-01
-2.40012825e-01 2.09133744e-01 -3.41502368e-01 9.22011733e-01
4.72271174e-01 9.85248387e-01 -4.33875114e-01 7.89327741e-01
1.70746490e-01 -1.45463571e-01 -2.33297888e-02 1.99321002e-01
3.89315724e-01 -5.91868103e-01 6.90651238e-01 8.76276135e-01
4.73736137e-01 5.58183379e-02 1.98958367e-01 1.10557640e+00
-9.20310169e-02 2.22669154e-01 -1.04421616e+00 -1.02263808e-01
4.04222041e-01 1.24264181e+00 -4.78357971e-01 -3.88454854e-01
-3.91740650e-01 1.45709407e+00 1.88389644e-01 5.95840096e-01
-5.02543151e-01 -5.42251706e-01 7.24864781e-01 -1.03738263e-01
3.15711617e-01 -1.38962958e-02 5.96752577e-02 -1.40839219e+00
2.75494337e-01 -5.72611988e-01 3.22698057e-01 -9.64970112e-01
-1.58706415e+00 2.95465887e-01 -3.35820079e-01 -1.56826460e+00
2.25303024e-01 -4.09570873e-01 -5.68310380e-01 1.04758227e+00
-1.71045828e+00 -1.51352906e+00 -4.93231386e-01 9.87490118e-01
8.73331845e-01 -4.99736011e-01 8.55671287e-01 4.90568101e-01
-2.36676916e-01 7.92268932e-01 3.13467562e-01 2.95344681e-01
9.80626941e-01 -9.08905149e-01 1.40033647e-01 3.69918972e-01
3.14830542e-01 7.77150154e-01 2.55364060e-01 -6.55306399e-01
-1.68635309e+00 -1.32096243e+00 5.44050574e-01 -2.97178954e-01
4.85088438e-01 -3.50079507e-01 -1.11447215e+00 1.81403369e-01
-2.12514922e-01 3.18937212e-01 3.26824963e-01 -2.86914945e-01
-8.21378767e-01 -2.25835040e-01 -1.18077338e+00 6.67518914e-01
1.17243159e+00 -9.36290920e-01 -6.37532353e-01 2.02636987e-01
4.91987616e-01 -6.58777431e-02 -7.95811713e-01 2.69939184e-01
8.99681807e-01 -7.94058025e-01 1.45804715e+00 -6.02840424e-01
4.29865122e-01 -2.37088859e-01 -3.49637210e-01 -1.12559354e+00
-1.48330748e-01 -2.03117013e-01 1.60745427e-01 1.26165581e+00
-2.48284549e-01 -3.81518096e-01 7.34723866e-01 4.21905994e-01
2.99452007e-01 -5.33928454e-01 -9.01100099e-01 -1.05884564e+00
3.15497778e-02 -1.75240442e-01 5.17526269e-01 9.79636371e-01
-3.06156993e-01 2.65033752e-01 -6.62112474e-01 -9.10475850e-02
1.07295191e+00 3.90065521e-01 8.25232029e-01 -1.34359419e+00
2.05023795e-01 -4.63715494e-01 -6.55224502e-01 -9.64178920e-01
5.13661504e-01 -1.14263356e+00 -8.58097896e-02 -1.47147000e+00
2.76170671e-01 -5.21529675e-01 -6.44840598e-01 1.23215452e-01
-1.97674468e-01 3.69976550e-01 4.74015623e-01 3.40125412e-01
-7.11827576e-01 9.31108773e-01 1.42998731e+00 -5.82489431e-01
2.25988463e-01 -1.39606938e-01 -4.61670697e-01 4.08052385e-01
2.57968932e-01 -2.97740817e-01 -7.50687063e-01 -2.95884132e-01
-1.74539804e-01 3.14379245e-01 7.64269471e-01 -7.03029037e-01
3.32990527e-01 -1.12727068e-01 4.76748407e-01 -7.76406586e-01
6.39699340e-01 -1.23620260e+00 1.19742408e-01 1.26928404e-01
-3.32081437e-01 -3.86820197e-01 -2.81550270e-02 1.09334004e+00
-5.91032088e-01 -1.94581717e-01 6.75003290e-01 -2.36649036e-01
-7.80558050e-01 7.37768710e-01 3.43557447e-01 -1.50638729e-01
8.76368582e-01 -4.02236134e-01 -1.30240455e-01 -2.20138416e-01
-4.89187002e-01 1.76027209e-01 5.09518802e-01 8.46291423e-01
1.19616425e+00 -1.82013905e+00 -5.31048596e-01 4.86193091e-01
6.17613018e-01 -6.57967031e-02 6.83443367e-01 4.78095949e-01
-2.28898063e-01 5.35719812e-01 -3.60653073e-01 -5.08249760e-01
-1.13193500e+00 9.26318765e-01 2.62615234e-02 6.31947443e-02
-9.15970862e-01 7.87976325e-01 5.83160520e-01 -5.27222157e-01
3.91825557e-01 4.02939856e-01 1.35202077e-03 2.35563740e-02
5.87443352e-01 3.30619127e-01 -8.89116749e-02 -6.24381483e-01
-3.64451826e-01 9.12367582e-01 -7.50651881e-02 -9.27482024e-02
1.11587334e+00 -1.79301247e-01 -7.04629067e-03 4.21001285e-01
1.59061456e+00 -3.21416020e-01 -1.16056490e+00 -7.36530542e-01
3.71311791e-02 -9.61753607e-01 7.86173493e-02 -4.66819972e-01
-1.17281759e+00 1.07301366e+00 6.44438982e-01 -2.69824237e-01
9.06852484e-01 -1.90564245e-01 1.13713205e+00 3.43979061e-01
5.60031116e-01 -1.11360610e+00 3.09342176e-01 5.42628095e-02
1.28089452e+00 -1.42959833e+00 -1.10405311e-01 -2.68688649e-01
-6.98308527e-01 1.22742856e+00 4.32305813e-01 -4.66465294e-01
7.07692444e-01 -4.48649198e-01 -3.12716179e-02 -1.79303005e-01
-4.26238954e-01 4.68628928e-02 5.05861461e-01 5.04413903e-01
-1.22734360e-01 -7.88038969e-02 -2.06463009e-01 5.98444581e-01
4.28404182e-01 8.43218938e-02 -2.33208854e-03 8.04251015e-01
-1.59289166e-01 -1.14643073e+00 -3.29458237e-01 2.77814090e-01
4.70535755e-02 -4.83004488e-02 -4.98431593e-01 7.13101447e-01
-1.59750327e-01 5.42658448e-01 -1.01084717e-01 -1.96041986e-01
3.36996198e-01 1.81722254e-01 5.85411251e-01 -2.03004241e-01
-1.02264479e-01 6.65517747e-02 -5.75454712e-01 -4.62003708e-01
-3.41710746e-01 -3.53829056e-01 -8.41408551e-01 -5.86342812e-03
-2.64056325e-01 1.26719519e-01 7.28852093e-01 6.86184466e-01
3.22390705e-01 1.73258975e-01 7.39804447e-01 -8.63071620e-01
-6.71532571e-01 -6.21753693e-01 -8.76050293e-01 7.98114777e-01
2.32616767e-01 -9.83413100e-01 -2.98483431e-01 -2.42684320e-01] | [11.591269493103027, 0.71086186170578] |
3db9c74a-a868-47f7-b7a1-40af1150c2e3 | arbitrary-conditional-distributions-with-1 | null | null | https://openreview.net/forum?id=IVxAlfGNKB | https://openreview.net/pdf?id=IVxAlfGNKB | Arbitrary Conditional Distributions with Energy | Modeling distributions of covariates, or density estimation, is a core challenge in unsupervised learning. However, the majority of work only considers the joint distribution, which has limited relevance to practical situations. A more general and useful problem is arbitrary conditional density estimation, which aims to model any possible conditional distribution over a set of covariates, reflecting the more realistic setting of inference based on prior knowledge. We propose a novel method, Arbitrary Conditioning with Energy (ACE), that can simultaneously estimate the distribution $p(\mathbf{x}_u \mid \mathbf{x}_o)$ for all possible subsets of unobserved features $\mathbf{x}_u$ and observed features $\mathbf{x}_o$. ACE is designed to avoid unnecessary bias and complexity --- we specify densities with a highly expressive energy function and reduce the problem to only learning one-dimensional conditionals (from which more complex distributions can be recovered during inference). This results in an approach that is both simpler and higher-performing than prior methods. We show that ACE achieves state-of-the-art for arbitrary conditional likelihood estimation and data imputation on standard benchmarks. | ['Junier Oliva', 'Ryan Strauss'] | 2021-05-21 | null | https://openreview.net/forum?id=_idcJrecij | https://openreview.net/pdf?id=_idcJrecij | neurips-2021-12 | ['arbitrary-conditional-density-estimation'] | ['methodology'] | [ 3.8509168e-02 4.1109990e-02 -2.8704467e-01 -6.6638708e-01
-1.1229237e+00 -2.1090826e-01 4.3529740e-01 4.8896998e-02
-3.4571487e-01 1.2638911e+00 -1.1504314e-01 -2.8108037e-01
-4.8710793e-01 -9.7872043e-01 -9.1990596e-01 -9.0835315e-01
-2.9985970e-01 8.6414874e-01 -3.1094959e-01 4.5021409e-01
-3.7102859e-02 2.8519395e-01 -1.6537150e+00 -3.6110759e-01
9.5918536e-01 8.5834914e-01 2.2036768e-01 3.9020810e-01
5.5236373e-02 7.3496395e-01 -4.2877731e-01 -6.9394624e-01
-2.5675011e-01 -3.3304283e-01 -6.1329055e-01 -1.3477111e-01
2.6807326e-01 -2.8335017e-01 -8.7572791e-02 1.0929036e+00
1.8188448e-01 2.1018787e-01 1.4647905e+00 -1.2485427e+00
-5.5148870e-01 6.4057773e-01 -6.7773443e-01 -1.5078202e-01
1.6469905e-01 -1.9204949e-01 8.9260387e-01 -7.2087240e-01
3.1716785e-01 1.1423641e+00 7.2875345e-01 3.0669174e-01
-1.8644837e+00 -7.7330720e-01 1.0127159e-01 7.4121531e-04
-1.8172868e+00 -2.2075340e-01 4.4107461e-01 -6.8667203e-01
6.3673002e-01 2.7955517e-01 1.4756505e-01 1.1120515e+00
2.7436644e-01 7.2959459e-01 1.1349765e+00 -3.9423269e-01
5.9898430e-01 1.9625697e-01 2.1635295e-01 5.2392942e-01
2.8323901e-01 6.2391706e-02 -4.7108114e-01 -4.1146114e-01
8.3320463e-01 1.7852961e-01 -3.1542751e-01 -3.8162515e-01
-9.9949253e-01 1.1718451e+00 -1.5979187e-01 -1.4347695e-01
-3.0172217e-01 3.6158669e-01 -1.0736329e-01 1.8473396e-02
5.0234568e-01 -6.4135730e-02 -7.1284807e-01 -2.6476032e-01
-1.1527107e+00 3.4848550e-01 1.0201992e+00 1.1866765e+00
1.2833205e+00 -2.1132246e-01 -2.0174725e-01 7.2828960e-01
3.8524452e-01 7.7932084e-01 -2.7592292e-01 -9.9479485e-01
3.0041236e-01 2.4916039e-01 5.3794914e-01 -3.2440761e-01
-1.7710713e-01 -2.4557190e-01 -1.2792453e+00 2.7800614e-01
7.5071156e-01 -4.5757505e-01 -9.2477822e-01 2.0995011e+00
2.7123097e-01 3.4900108e-01 -3.0226064e-01 5.2750313e-01
4.3189490e-01 5.9859687e-01 3.1494647e-01 -4.4811520e-01
1.1568067e+00 -1.8443002e-01 -6.6384768e-01 -2.1045856e-01
1.8487637e-01 -5.6008291e-01 1.0164844e+00 5.1461530e-01
-1.1556114e+00 -3.2168818e-01 -6.6138154e-01 -2.8092770e-02
-2.3841780e-01 1.8500705e-01 1.0302693e+00 9.0362418e-01
-7.7907443e-01 7.9084110e-01 -1.1088415e+00 6.7298055e-02
4.7272381e-01 6.1002612e-01 -3.2065955e-01 -1.8105271e-01
-1.0129802e+00 7.6549423e-01 1.4459723e-01 -1.5841800e-01
-9.1145152e-01 -1.0645690e+00 -1.1081871e+00 3.7487164e-01
4.2351833e-01 -6.6258663e-01 1.0048243e+00 -4.6518061e-01
-1.1810856e+00 6.2924939e-01 -5.3113568e-01 -2.0980214e-01
4.4325963e-01 -1.8218485e-01 -2.1665025e-01 -3.1994143e-01
6.2201545e-03 5.0015867e-01 8.8864732e-01 -1.1240777e+00
-2.7689564e-01 -4.6102929e-01 -2.4642524e-01 -2.2796512e-01
-1.7587614e-01 3.1804200e-02 -4.5079488e-01 -5.5677903e-01
-9.8335907e-02 -6.4154017e-01 -2.4182700e-01 -1.4347067e-01
-5.9660673e-01 -2.9267409e-01 2.2717634e-01 -5.8930916e-01
1.4289562e+00 -1.9708289e+00 2.5558361e-01 3.8106260e-01
6.1233208e-02 -2.7972868e-01 3.9678651e-01 3.8438782e-01
4.9583387e-02 1.5192611e-02 -8.0741727e-01 -7.4406958e-01
3.6518732e-01 3.8414392e-01 -2.1508476e-01 6.5553910e-01
3.4596574e-01 6.3486427e-01 -6.9489080e-01 -5.4914659e-01
3.8206694e-01 8.1370658e-01 -6.8355542e-01 3.3121580e-01
-4.5515448e-01 6.8295795e-01 -2.9385921e-01 6.8861777e-01
1.1668071e+00 -3.5886350e-01 1.8488844e-01 1.8734136e-01
-3.3945620e-02 3.2081738e-02 -1.6866575e+00 1.5273226e+00
-3.7055886e-01 1.6643991e-01 1.8553463e-01 -1.1877636e+00
9.1050529e-01 5.4864831e-02 4.5196226e-01 -7.2392330e-02
1.6090624e-01 -1.4940831e-02 -4.1598868e-01 -2.7774656e-01
2.4051461e-01 -4.8030391e-01 -3.9296716e-01 2.7455151e-01
4.1280422e-01 -5.9783101e-02 1.5681222e-01 4.3597486e-02
9.6733350e-01 5.1399863e-01 3.5091159e-01 -5.1383001e-01
1.4801577e-01 -4.1679218e-01 6.5887964e-01 1.0905762e+00
2.6673606e-01 6.4769214e-01 7.7272135e-01 7.3927112e-02
-7.6253271e-01 -1.7522441e+00 -7.9046458e-01 1.0499574e+00
-1.8181780e-01 -3.0177614e-01 -7.8643286e-01 -4.8017180e-01
3.3484676e-01 1.0056776e+00 -8.0664247e-01 2.2982124e-02
-2.4072596e-01 -1.2813063e+00 9.6609324e-02 6.4078617e-01
7.0145361e-02 -9.0282214e-01 -3.5988757e-01 2.2807889e-01
-1.6111070e-01 -6.0584325e-01 -3.1481710e-01 6.7729801e-01
-8.5562068e-01 -8.2089043e-01 -5.9221929e-01 -3.6505347e-01
6.5772009e-01 -7.2074020e-01 1.4919823e+00 -2.7284521e-01
-3.5585350e-01 4.1729596e-01 5.5793315e-02 -5.4975861e-01
-8.9179032e-02 -1.3430572e-01 -5.8601562e-02 5.8382202e-02
7.5433666e-01 -9.0573674e-01 -5.0323236e-01 2.3887028e-01
-8.6505651e-01 -2.9330653e-01 5.4084021e-01 8.8823217e-01
8.9532316e-01 1.0471930e-01 6.0347408e-01 -1.1769133e+00
1.1733510e-02 -8.3675390e-01 -7.6978660e-01 2.6947644e-01
-3.8620317e-01 2.4354185e-01 5.5601847e-01 -4.9346718e-01
-1.3195380e+00 1.5148550e-01 -3.1708953e-01 -2.6816085e-01
-5.0908142e-01 6.1644942e-01 -6.3460636e-01 5.8995432e-01
4.6552733e-01 1.6119674e-01 -1.9478738e-01 -1.0141037e+00
3.4068811e-01 6.0802323e-01 8.3660245e-01 -9.3223071e-01
5.1694179e-01 5.7054037e-01 3.5294828e-01 -6.5888309e-01
-9.8875582e-01 -3.2388362e-01 -8.2607085e-01 2.2787462e-01
7.7358890e-01 -1.0086451e+00 -1.0939500e+00 3.5862759e-01
-6.7480981e-01 -4.7054994e-01 -3.2177061e-01 7.4854290e-01
-7.4152458e-01 3.0158672e-01 -3.4414628e-01 -1.2178907e+00
-3.6994707e-02 -9.2194682e-01 1.1388066e+00 1.8228641e-01
-1.9086352e-01 -1.1505151e+00 6.3335836e-02 3.7860550e-02
2.3562205e-01 1.9282001e-01 1.0711765e+00 -3.2168078e-01
-6.0271436e-01 -6.3164257e-02 -2.5061902e-01 4.1369686e-01
-3.7504521e-03 -7.5511746e-02 -9.0945482e-01 -2.0558052e-01
1.1644984e-02 -3.2551625e-01 1.0047476e+00 7.6516068e-01
1.6252475e+00 -1.8012770e-01 -4.0753660e-01 5.2234018e-01
1.5571266e+00 -2.6375550e-01 9.4928116e-01 -3.8350907e-01
3.7558001e-01 3.6702278e-01 4.5195398e-01 9.9193937e-01
4.0510464e-01 4.6637887e-01 3.3348513e-01 5.3634018e-02
3.1226388e-01 -2.9987413e-01 1.8546988e-01 1.9877948e-01
1.2945478e-02 -2.3031786e-01 -6.7862636e-01 7.0626819e-01
-1.7239546e+00 -9.7103792e-01 -2.9008791e-01 2.6006536e+00
1.2057031e+00 -5.2010152e-02 1.8071204e-02 -4.0616695e-02
7.4469978e-01 -3.1836075e-01 -5.9519178e-01 4.5574017e-02
1.5241189e-01 9.8475987e-01 4.3148625e-01 5.0315893e-01
-1.0966882e+00 3.6644438e-01 6.2867756e+00 1.1089652e+00
-5.2210069e-01 1.1497628e-01 7.6713032e-01 -4.6202712e-02
-4.3822807e-01 1.5198016e-01 -1.0925740e+00 8.4969515e-01
1.0193996e+00 1.9192123e-01 3.8969323e-01 8.0342549e-01
-1.3061694e-02 -5.5686468e-01 -1.5001533e+00 7.5856698e-01
-1.9389074e-01 -9.9896950e-01 -3.3522028e-01 2.1390542e-01
6.9850528e-01 -2.0066075e-01 1.2212090e-01 5.6670064e-01
7.6263368e-01 -1.3582261e+00 4.9055913e-01 1.0427479e+00
1.1693949e+00 -9.7604281e-01 5.8467323e-01 6.6522473e-01
-9.2326808e-01 1.5453382e-01 -6.0087895e-01 -2.3979783e-01
3.5771367e-01 1.2481358e+00 -2.8057978e-01 5.0855064e-01
8.5174453e-01 5.3417277e-01 -1.3749866e-01 1.0215919e+00
-1.9213589e-01 9.5120943e-01 -6.2059039e-01 1.6012827e-01
-2.7533266e-01 -4.3837094e-01 1.1178374e-01 1.3717743e+00
5.4680842e-01 2.8181493e-02 1.4107466e-01 1.2237108e+00
-2.1045299e-01 -1.2800792e-01 -2.9716992e-01 3.9180112e-01
6.1200947e-01 1.0705098e+00 -4.4136140e-01 -2.2892083e-01
-4.5936736e-01 5.5516893e-01 7.3742747e-01 4.0379214e-01
-1.0887202e+00 -1.6799057e-01 7.5036061e-01 1.7802578e-01
5.7610649e-01 -2.0313647e-01 -2.6152077e-01 -1.3305938e+00
-7.9658888e-02 -3.1456181e-01 5.4606313e-01 -5.0701261e-01
-1.8617907e+00 -1.1784573e-01 5.0209653e-01 -8.1511939e-01
-2.6776776e-01 -5.7446325e-01 -5.1820010e-01 1.3241524e+00
-1.4844407e+00 -9.3243229e-01 -2.8714707e-02 8.2142520e-01
-2.1626391e-02 2.2139394e-01 1.1054155e+00 3.7749508e-01
-4.2633033e-01 6.3949144e-01 5.4443496e-01 -1.3092293e-01
6.8563592e-01 -1.7163086e+00 -2.3343910e-02 4.9076197e-01
3.8885970e-02 6.9669110e-01 6.8776977e-01 -7.1970284e-01
-1.3134241e+00 -1.1320902e+00 1.0023187e+00 -7.4702656e-01
4.6706462e-01 -6.3353956e-01 -9.3844467e-01 1.0016017e+00
-1.1443113e-02 -5.7996493e-02 9.6959502e-01 4.5911297e-01
-2.0595805e-01 1.1978936e-01 -1.4632118e+00 3.3204576e-01
8.8748205e-01 -2.5442731e-01 -3.9392784e-01 1.4979614e-01
1.5797777e-01 -2.4963973e-01 -1.2847530e+00 4.9050987e-01
4.2632657e-01 -9.6470565e-01 1.0959103e+00 -4.6362984e-01
2.3679905e-01 -2.4648467e-01 -3.0940604e-01 -9.2160523e-01
-1.7154758e-01 -6.5729094e-01 -4.0656662e-01 1.5058533e+00
3.8290843e-01 -5.6140983e-01 7.3327363e-01 1.0094790e+00
1.6927047e-01 -5.4261106e-01 -1.2437488e+00 -6.6969949e-01
4.4548124e-01 -7.7065170e-01 7.0027363e-01 8.1242847e-01
-1.6085200e-01 2.6523366e-04 -7.1480834e-01 1.5123624e-01
1.0150278e+00 1.8632127e-01 6.9200134e-01 -1.5288178e+00
-6.1636597e-01 -1.0583489e-02 -1.0996515e-01 -1.2637410e+00
4.0029645e-01 -7.8618872e-01 1.6836840e-01 -1.4183669e+00
6.4920962e-01 -7.1818119e-01 -2.7179772e-01 4.7060177e-01
-3.5305580e-01 1.3579333e-01 -3.2977307e-01 -1.4321278e-01
-4.8341376e-01 5.9594101e-01 9.5241994e-01 1.3469118e-01
9.1290534e-02 2.4728832e-01 -7.1693343e-01 8.0584770e-01
4.1511786e-01 -6.4841002e-01 -3.2166469e-01 -1.7353687e-01
1.2773201e-01 3.8930100e-01 6.5335900e-01 -7.4087495e-01
7.2549097e-02 -4.3216392e-01 6.9064939e-01 -8.1782758e-01
5.2675486e-01 -7.0777351e-01 4.4436941e-01 -1.1975133e-01
-5.3360522e-02 -3.7010872e-01 1.5375462e-02 5.9447801e-01
-8.6623132e-03 -5.1751471e-01 6.4423978e-01 -4.4215720e-02
-1.3221188e-01 3.4856910e-01 -3.3679059e-01 8.8431351e-02
7.7242255e-01 1.6160156e-01 5.1624168e-02 -4.2908850e-01
-1.1539235e+00 1.9463146e-01 3.8566417e-01 -2.5667289e-01
2.5471660e-01 -1.1727622e+00 -8.4326559e-01 2.6444104e-01
-1.3539343e-01 2.8382277e-01 4.6960333e-01 8.0226088e-01
1.4028379e-01 4.4842683e-02 2.6890665e-01 -7.3132485e-01
-6.0149527e-01 5.6727177e-01 4.2380162e-02 -5.0146896e-01
-3.8617364e-01 8.9216822e-01 1.3538891e-01 -5.2622515e-01
2.8509688e-01 -2.4178010e-01 1.0154914e-01 -5.4568749e-02
3.8203412e-01 4.8293170e-01 -1.0179809e-01 -3.3005401e-01
-2.2264586e-01 2.9619649e-01 1.3959289e-01 -4.0409859e-02
1.3071181e+00 -1.6815962e-01 -1.4122948e-01 5.6420249e-01
1.1178508e+00 -2.5441632e-02 -1.6948888e+00 -1.5828076e-01
-1.3360237e-01 -6.8385291e-01 -1.8131240e-01 -1.0547303e+00
-8.1851208e-01 9.7670507e-01 3.3183610e-01 4.6867900e-02
1.0811278e+00 2.9359552e-01 2.6471838e-01 1.7107080e-01
6.0935783e-01 -8.6291349e-01 -3.0246103e-01 3.5648653e-01
4.9266377e-01 -1.2586581e+00 7.0678122e-02 -3.7840137e-01
-4.5586675e-01 7.0721948e-01 4.0958211e-01 -7.1085043e-02
1.1397402e+00 5.7563490e-01 -4.4095170e-01 5.1645719e-02
-4.8646775e-01 -1.7542417e-01 3.8690832e-01 7.0340061e-01
5.6374019e-01 1.3874444e-01 -1.4063421e-01 1.0654995e+00
-2.1643752e-01 1.1404608e-01 4.0218824e-01 8.4671968e-01
-1.5218838e-01 -1.3973045e+00 -4.2079535e-01 9.6681482e-01
-6.5931737e-01 -1.8354771e-01 3.6724126e-01 8.3103412e-01
3.9638066e-01 9.0505445e-01 2.4852698e-01 3.7893677e-01
-3.6887001e-02 3.9248717e-01 6.3704789e-01 -6.6470110e-01
-5.4690775e-02 2.7655908e-01 -1.4118746e-01 -2.3357575e-01
-5.0434512e-01 -1.0285935e+00 -1.1271021e+00 -3.7994033e-01
-3.7263948e-01 2.4599470e-01 5.1938170e-01 9.1382426e-01
6.0996871e-02 2.3327442e-01 3.2504684e-01 -7.5175464e-01
-7.7078539e-01 -1.1180501e+00 -1.1841462e+00 4.6449271e-01
1.8635467e-01 -1.0967118e+00 -4.0866050e-01 2.2785318e-01] | [7.2665791511535645, 4.0636515617370605] |
6403619e-cdd8-4328-852e-c7c8d7dc3115 | efficient-video-generation-on-complex | 1907.06571 | null | https://arxiv.org/abs/1907.06571v2 | https://arxiv.org/pdf/1907.06571v2.pdf | Adversarial Video Generation on Complex Datasets | Generative models of natural images have progressed towards high fidelity samples by the strong leveraging of scale. We attempt to carry this success to the field of video modeling by showing that large Generative Adversarial Networks trained on the complex Kinetics-600 dataset are able to produce video samples of substantially higher complexity and fidelity than previous work. Our proposed model, Dual Video Discriminator GAN (DVD-GAN), scales to longer and higher resolution videos by leveraging a computationally efficient decomposition of its discriminator. We evaluate on the related tasks of video synthesis and video prediction, and achieve new state-of-the-art Fr\'echet Inception Distance for prediction for Kinetics-600, as well as state-of-the-art Inception Score for synthesis on the UCF-101 dataset, alongside establishing a strong baseline for synthesis on Kinetics-600. | ['Karen Simonyan', 'Aidan Clark', 'Jeff Donahue'] | 2019-07-15 | null | https://openreview.net/forum?id=Byx91R4twB | https://openreview.net/pdf?id=Byx91R4twB | null | ['3d-character-animation-from-a-single-photo'] | ['computer-vision'] | [ 3.88408870e-01 2.18411565e-01 -2.94109941e-01 -5.30079715e-02
-1.08485103e+00 -5.11054218e-01 1.07493734e+00 -8.87841880e-01
1.12547174e-01 7.50713229e-01 5.47220469e-01 -9.72038731e-02
3.68188947e-01 -8.91583800e-01 -1.29047585e+00 -5.31174839e-01
-5.39097711e-02 5.86774945e-01 -8.61492082e-02 -2.87554681e-01
-2.17389002e-01 3.56111616e-01 -1.24758792e+00 7.28416979e-01
5.03438890e-01 1.03461111e+00 -9.19507071e-02 1.19297957e+00
5.99990487e-01 1.23857856e+00 -5.56966007e-01 -7.56364405e-01
3.64045501e-01 -8.24536920e-01 -4.75612611e-01 1.67278305e-01
8.83725047e-01 -7.10243523e-01 -1.19476140e+00 5.68283975e-01
5.42360306e-01 2.08039489e-03 1.05191219e+00 -1.40976846e+00
-1.20006764e+00 7.06401885e-01 -1.30232945e-01 1.71522811e-01
3.33590209e-01 6.76747918e-01 9.86375570e-01 -9.58237767e-01
1.07050371e+00 1.29302645e+00 8.15718412e-01 1.02678287e+00
-1.44599247e+00 -7.23863363e-01 -1.69453934e-01 8.15687478e-02
-1.38654923e+00 -7.20606446e-01 5.17998457e-01 -5.43759882e-01
1.16309237e+00 9.21159238e-02 7.19561577e-01 2.16814351e+00
2.26804897e-01 9.19704914e-01 6.98411524e-01 -5.48561551e-02
6.08664714e-02 -2.71107525e-01 -9.37695861e-01 8.20595503e-01
-2.16920495e-01 6.42119646e-01 -5.80216527e-01 3.47415000e-01
1.23311114e+00 -2.88822800e-01 -3.69192868e-01 -4.69017774e-02
-1.15515351e+00 9.42744553e-01 2.71008790e-01 -9.52408090e-02
-2.34634444e-01 8.06179225e-01 4.23064590e-01 3.31851095e-01
4.49822873e-01 3.75093907e-01 -1.27474349e-02 -5.20182014e-01
-1.26005125e+00 6.14396811e-01 6.04031503e-01 1.42240775e+00
1.09822057e-01 7.59639919e-01 -6.42287731e-01 4.71801490e-01
9.43613574e-02 6.78478003e-01 3.82805735e-01 -1.34242237e+00
6.55373216e-01 -2.80847311e-01 -4.85738460e-03 -6.84588909e-01
1.43417493e-01 -3.51403296e-01 -9.85699475e-01 2.60153472e-01
1.86603531e-01 -2.35447273e-01 -1.16780853e+00 1.82374847e+00
-1.66276813e-01 6.12428248e-01 1.81591824e-01 5.99512339e-01
8.75837564e-01 1.08855534e+00 1.99272018e-03 3.86787136e-03
7.07069635e-01 -1.42097104e+00 -4.29457963e-01 -6.69138208e-02
3.37220997e-01 -5.51593781e-01 8.60985637e-01 4.51683313e-01
-1.56800640e+00 -1.05649173e+00 -1.08751988e+00 -1.87478945e-01
1.12871073e-01 5.44257555e-03 5.88465333e-01 6.42705262e-01
-1.41611373e+00 9.21572328e-01 -8.73935223e-01 -1.92894340e-02
7.68669844e-01 -3.42577137e-02 -4.21046227e-01 -2.97258168e-01
-1.05897021e+00 6.55179441e-01 2.86742061e-01 -2.60439426e-01
-1.87212408e+00 -1.19431221e+00 -9.32542741e-01 -4.55626063e-02
4.59602252e-02 -1.06537986e+00 1.13667679e+00 -1.04567766e+00
-1.69398057e+00 6.58081412e-01 1.19013235e-01 -8.46653342e-01
1.02868021e+00 -2.22963765e-01 -6.87538862e-01 2.73882419e-01
-1.28395870e-01 1.19228888e+00 1.27589321e+00 -1.22080338e+00
-5.06647408e-01 3.30897927e-01 -1.67135429e-02 -1.44246474e-01
4.49782014e-02 -3.33687395e-01 -3.87906939e-01 -1.17722130e+00
-8.57087255e-01 -1.04418337e+00 -8.35398734e-02 -7.34316707e-02
-2.18255877e-01 8.36685151e-02 8.05204570e-01 -1.00273323e+00
1.06110346e+00 -1.95389366e+00 6.13753676e-01 -3.42657745e-01
2.83979565e-01 3.18572611e-01 -5.64394593e-01 5.21996439e-01
-2.22507104e-01 3.65696877e-01 1.66656747e-02 -6.33841753e-01
1.16533488e-01 1.52478665e-01 -5.98621905e-01 1.73640966e-01
6.67267084e-01 1.30877924e+00 -7.73809612e-01 -3.61936867e-01
1.52711630e-01 7.46297717e-01 -1.06674910e+00 4.40536380e-01
-5.61375618e-01 3.92539799e-01 4.33988459e-02 5.41901827e-01
2.78685302e-01 -2.70054311e-01 -1.25458091e-01 -1.41896531e-01
4.09381658e-01 -1.04640238e-01 -4.79622990e-01 1.98169816e+00
-3.78640801e-01 8.71281624e-01 -3.84119183e-01 -7.99346685e-01
5.45794487e-01 6.39316916e-01 4.66014981e-01 -5.93995869e-01
-7.95671940e-02 1.37575790e-01 -1.68115407e-01 -2.47450039e-01
5.47251165e-01 -1.18991219e-01 4.31651808e-02 3.46142240e-02
5.70774555e-01 -3.45357150e-01 2.85393924e-01 3.56997967e-01
1.25192750e+00 7.69800007e-01 -2.05097049e-01 5.77136390e-02
1.68277975e-03 -1.83283970e-01 6.11606501e-02 7.06500530e-01
1.61368772e-01 9.58566368e-01 3.74331266e-01 -2.54758984e-01
-1.88192666e+00 -1.33034265e+00 1.11506209e-01 6.86622679e-01
-4.58370984e-01 -7.04931438e-01 -1.00623083e+00 -7.35059619e-01
-5.59976250e-02 7.89156139e-01 -9.38864768e-01 -4.07132745e-01
-7.05799222e-01 -3.02006066e-01 9.85392392e-01 9.43406284e-01
4.14511323e-01 -1.05305004e+00 -4.39766981e-02 3.36265862e-01
-1.37282774e-01 -1.44723654e+00 -5.19669294e-01 -2.02147081e-01
-7.30079770e-01 -5.86362123e-01 -1.02062988e+00 -6.97016478e-01
2.39961937e-01 -2.84429878e-01 1.71943641e+00 -1.33274511e-01
-1.72175035e-01 3.89853209e-01 -2.63915777e-01 1.74968820e-02
-1.20977342e+00 1.20193206e-01 -1.20773531e-01 -2.56060451e-01
-3.80073994e-01 -7.70067811e-01 -6.37546241e-01 2.12778196e-01
-9.81821358e-01 4.20727789e-01 4.58351254e-01 9.96265948e-01
6.23502135e-01 -1.81387648e-01 6.26752377e-01 -7.60341942e-01
2.35156462e-01 -5.65457404e-01 -4.48425025e-01 8.61474201e-02
-4.93379772e-01 4.72461767e-02 9.95997310e-01 -4.69723433e-01
-9.81853902e-01 -2.91896224e-01 -4.45650548e-01 -1.08311868e+00
1.15503063e-02 1.07905686e-01 -4.76613119e-02 -3.78955938e-02
9.29027259e-01 4.57804382e-01 -1.77820399e-01 -5.08746058e-02
7.19670832e-01 2.22295493e-01 9.66346622e-01 -6.93743885e-01
1.01353896e+00 3.07362050e-01 3.49323243e-01 -5.45531332e-01
-5.03826916e-01 5.17688096e-01 -1.63046554e-01 -1.32539213e-01
1.10334826e+00 -1.47306049e+00 -3.41589540e-01 3.96832407e-01
-9.32201803e-01 -7.81729817e-01 -5.79367340e-01 6.11051060e-02
-1.20785868e+00 8.63668323e-02 -9.88399148e-01 -2.29001731e-01
-2.17024967e-01 -1.15882969e+00 1.31032133e+00 -9.50038582e-02
-2.23676458e-01 -1.04354286e+00 1.35583088e-01 3.77901077e-01
4.28704172e-01 7.06573069e-01 6.84708774e-01 -8.82842615e-02
-1.00703609e+00 -1.33354902e-01 9.87439882e-03 7.75565386e-01
-2.51862109e-01 3.54771346e-01 -7.55934536e-01 -4.54561949e-01
-3.83609891e-01 -5.66869915e-01 1.03144825e+00 3.68871927e-01
1.15027106e+00 -3.71530116e-01 -1.18695013e-01 1.12142670e+00
1.44472468e+00 2.12959111e-01 1.14323282e+00 -1.72062188e-01
9.63562608e-01 -9.88074392e-02 2.26434723e-01 3.98220956e-01
5.64862452e-02 7.24861741e-01 3.61790806e-01 2.78165378e-02
-8.57240319e-01 -8.34585130e-01 7.78350830e-01 7.00550854e-01
-6.10005379e-01 -8.43312383e-01 -4.62941378e-01 4.73140687e-01
-1.61663115e+00 -1.35778320e+00 2.46448144e-01 1.66568303e+00
8.24260235e-01 2.18602821e-01 2.48606563e-01 -1.59768797e-02
4.18742508e-01 2.77873039e-01 -3.35982621e-01 -3.53557706e-01
-3.10108423e-01 6.02161288e-01 4.17696327e-01 3.47891927e-01
-1.01447010e+00 1.04055786e+00 7.60830212e+00 1.24194992e+00
-9.75895107e-01 1.85464263e-01 1.01373732e+00 -3.54953498e-01
-4.61727560e-01 -2.25465387e-01 -9.13661420e-01 7.07744062e-01
1.51717317e+00 -1.55962016e-02 8.84056926e-01 8.08519483e-01
-6.41375333e-02 5.00285089e-01 -1.59895825e+00 9.69724119e-01
1.96459562e-01 -2.12106228e+00 5.60764611e-01 2.35054672e-01
1.32119596e+00 4.98175919e-02 4.02694881e-01 5.35836101e-01
4.06979293e-01 -1.59546173e+00 1.12979198e+00 6.28315508e-01
1.42946255e+00 -7.76028693e-01 2.19358295e-01 5.04153781e-04
-1.01821303e+00 1.32182002e-01 -3.09971750e-01 2.32428625e-01
4.59168881e-01 -3.02166417e-02 -8.23143125e-01 4.68178451e-01
3.52494329e-01 1.07615793e+00 -5.32786191e-01 4.85624999e-01
-1.82553411e-01 7.49459326e-01 1.41719237e-01 4.89598572e-01
3.71051818e-01 7.79664591e-02 4.25717682e-01 1.15771019e+00
6.39912546e-01 -1.25493094e-01 -7.40309060e-02 9.42499995e-01
-5.97596169e-01 -4.68266964e-01 -7.22438455e-01 -6.43916249e-01
1.42937541e-01 7.46981919e-01 -1.76381916e-01 -6.84661269e-01
-1.90620944e-01 1.22448802e+00 2.76779741e-01 4.46905822e-01
-1.59353185e+00 1.82231843e-01 7.01477408e-01 2.40975991e-01
7.50029385e-01 -1.48148090e-01 4.45905291e-02 -1.35197103e+00
-3.07462960e-01 -1.37327445e+00 -5.40927872e-02 -1.10130250e+00
-1.07827663e+00 8.55051994e-01 -1.28429264e-01 -1.21923614e+00
-8.55647326e-01 -5.60533345e-01 -2.71643817e-01 7.58068681e-01
-1.00490475e+00 -1.60624826e+00 -2.06113100e-01 7.23318636e-01
8.68009567e-01 -6.16398931e-01 9.31940377e-01 4.85194534e-01
-3.30748856e-01 9.05860245e-01 2.48418137e-01 1.18203081e-01
4.55828547e-01 -9.54599619e-01 1.07711840e+00 1.01628304e+00
4.56858724e-01 -1.74969584e-02 7.08994269e-01 -6.20210826e-01
-1.45813489e+00 -1.33486366e+00 2.49697745e-01 -7.58660376e-01
5.91100752e-01 -4.80968893e-01 -2.97931135e-01 1.12602675e+00
4.63440299e-01 1.94490403e-01 2.52142549e-01 -5.07543266e-01
-5.73641658e-01 1.79072455e-01 -1.09994519e+00 7.64243186e-01
1.58829975e+00 -6.66948915e-01 -1.21810175e-01 3.85229498e-01
1.20702147e+00 -5.69073081e-01 -1.40807557e+00 3.31383497e-01
5.62882841e-01 -9.58983302e-01 1.25054848e+00 -9.02654767e-01
1.54719412e+00 6.66537285e-02 -3.68084878e-01 -1.34209979e+00
-5.29961526e-01 -1.16086912e+00 -5.25360584e-01 1.31478429e+00
3.51776600e-01 1.27051592e-01 9.82198477e-01 1.85445771e-01
-4.85389024e-01 -7.76264548e-01 -7.99151003e-01 -8.64155889e-01
2.70181000e-01 -3.63706231e-01 4.55688000e-01 6.38213575e-01
-6.59536541e-01 3.71400058e-01 -1.06970727e+00 -2.50222027e-01
5.78941762e-01 -2.82130331e-01 9.97676015e-01 -5.27083099e-01
-8.36842895e-01 -3.60813081e-01 -5.62861145e-01 -1.17547369e+00
2.46933863e-01 -9.12580192e-01 -1.75051004e-01 -1.34103286e+00
6.41274452e-02 -1.49917543e-01 1.20592594e-01 8.25539883e-03
1.02986105e-01 7.83212721e-01 4.57822055e-01 5.55208251e-02
-4.38828558e-01 7.83607364e-01 1.63899505e+00 -2.65216291e-01
4.33641970e-01 -4.30298388e-01 -4.47052866e-01 3.55003059e-01
2.44918481e-01 -2.35683247e-01 -5.84719837e-01 -5.99518001e-01
6.38926923e-02 5.37395060e-01 4.31112140e-01 -1.43328726e+00
-4.31554466e-01 1.82455797e-02 8.33698750e-01 -1.85845450e-01
8.04668486e-01 -5.40679276e-01 8.83560777e-01 3.61093789e-01
-4.71749753e-01 1.39900997e-01 1.81719705e-01 6.66785777e-01
-1.66594937e-01 1.95764691e-01 8.48705888e-01 -2.12009490e-01
-6.59672260e-01 7.07444787e-01 -7.21840113e-02 5.39935112e-01
9.67231870e-01 -2.01022580e-01 -5.45370996e-01 -6.47544801e-01
-7.89459050e-01 -2.66613990e-01 7.79518247e-01 6.31991744e-01
7.64248908e-01 -1.75702429e+00 -1.08940244e+00 2.76976854e-01
-1.86466768e-01 -2.94169366e-01 2.91520268e-01 2.70292461e-01
-9.03576195e-01 3.98640275e-01 -5.57199776e-01 -5.35984755e-01
-8.31551075e-01 8.33912969e-01 2.16688588e-01 -4.03999448e-01
-5.61625957e-01 1.14216697e+00 2.48634040e-01 2.53950030e-01
-4.09676842e-02 -2.98264295e-01 4.00177568e-01 -2.32122302e-01
4.61773992e-01 2.70508111e-01 -2.17058435e-01 -8.12536001e-01
1.14167303e-01 2.53890961e-01 1.98129967e-01 -1.93651840e-01
1.39970398e+00 3.62090856e-01 4.75645453e-01 -1.03822589e-01
1.40886354e+00 -1.44602686e-01 -2.01815772e+00 2.44743347e-01
-7.42587745e-01 -5.29225886e-01 -2.32778057e-01 -8.56761634e-01
-1.33741307e+00 8.60307992e-01 4.53223795e-01 -3.11974119e-02
9.56433177e-01 -6.93817660e-02 1.05984056e+00 -6.92201704e-02
3.88546854e-01 -5.47295988e-01 4.21267360e-01 2.96295285e-01
1.13862324e+00 -1.05267429e+00 -1.84041828e-01 -1.89087346e-01
-7.48184025e-01 8.89004230e-01 4.26488727e-01 -6.10658348e-01
3.60204250e-01 4.67383772e-01 -4.27059591e-01 2.40411028e-01
-1.03254557e+00 3.89663726e-01 3.50556672e-01 9.02574480e-01
3.21573138e-01 -3.51454876e-03 1.97602689e-01 3.33725274e-01
-4.07350987e-01 3.33976090e-01 4.64779913e-01 4.33704883e-01
1.31811827e-01 -1.01805460e+00 1.94940623e-02 2.43211821e-01
-6.83203459e-01 -4.31277305e-01 -4.06668074e-02 8.14369440e-01
9.17135626e-02 7.07704961e-01 1.41495988e-01 -6.39338195e-01
2.46415604e-02 -4.62197214e-02 1.04339945e+00 -1.81407660e-01
-4.53774363e-01 -3.07856966e-03 4.06645238e-01 -8.10581386e-01
-3.89259458e-01 -4.68271017e-01 -7.63107240e-01 -7.89874375e-01
1.65901735e-01 -1.31124929e-01 3.43986392e-01 5.58298886e-01
4.58816826e-01 8.92350137e-01 4.65270340e-01 -1.24314833e+00
-6.47296071e-01 -8.67576897e-01 -3.79929125e-01 6.16843641e-01
2.33597279e-01 -3.08615863e-01 -1.81240991e-01 5.87615311e-01] | [10.946781158447266, -0.5887318849563599] |
8f3a1e5a-ed1c-470a-b401-536e776857ee | avoiding-confusion-between-predictors-and | 1312.5714 | null | http://arxiv.org/abs/1312.5714v2 | http://arxiv.org/pdf/1312.5714v2.pdf | Avoiding Confusion between Predictors and Inhibitors in Value Function Approximation | In reinforcement learning, the goal is to seek rewards and avoid punishments.
A simple scalar captures the value of a state or of taking an action, where
expected future rewards increase and punishments decrease this quantity.
Naturally an agent should learn to predict this quantity to take beneficial
actions, and many value function approximators exist for this purpose. In the
present work, however, we show how value function approximators can cause
confusion between predictors of an outcome of one valence (e.g., a signal of
reward) and the inhibitor of the opposite valence (e.g., a signal canceling
expectation of punishment). We show this to be a problem for both linear and
non-linear value function approximators, especially when the amount of data (or
experience) is limited. We propose and evaluate a simple resolution: to instead
predict reward and punishment values separately, and rectify and add them to
get the value needed for decision making. We evaluate several function
approximators in this slightly different value function approximation
architecture and show that this approach is able to circumvent the confusion
and thereby achieve lower value-prediction errors. | ['Thomas P. Trappenberg', 'Patrick C. Connor'] | 2013-12-19 | null | null | null | null | ['value-prediction'] | ['computer-code'] | [ 2.32920542e-01 3.58272523e-01 -3.02526861e-01 -4.56472516e-01
-2.39210680e-01 -5.52705705e-01 3.71204972e-01 5.27050972e-01
-9.44432855e-01 1.25021160e+00 -1.33369908e-01 -1.14168145e-01
-2.79396117e-01 -8.02050710e-01 -6.60191536e-01 -1.00222003e+00
-2.42488444e-01 2.66391456e-01 2.17289969e-01 -5.25282979e-01
4.76281464e-01 5.74893117e-01 -1.18378448e+00 -7.30594397e-02
5.60730875e-01 1.04389536e+00 8.97881538e-02 6.13308668e-01
1.82082444e-01 1.05672336e+00 -8.62120807e-01 -1.22796871e-01
2.64119953e-01 -6.88898385e-01 -7.47742772e-01 -1.50246948e-01
-1.90772563e-01 -4.38020498e-01 -6.74048141e-02 1.13669109e+00
3.40347707e-01 2.78543532e-01 7.61584401e-01 -1.48481441e+00
-4.82558459e-01 7.26957798e-01 -3.85592610e-01 1.24747798e-01
1.17493831e-01 4.75300461e-01 9.84214008e-01 -1.15759082e-01
4.41067934e-01 1.03327560e+00 3.25423449e-01 9.77341473e-01
-1.57492828e+00 -1.79954797e-01 4.75613981e-01 1.13268182e-01
-9.51108813e-01 -1.70963734e-01 8.02786529e-01 -2.46429846e-01
9.54145253e-01 6.53216019e-02 8.67436647e-01 9.35796559e-01
2.41208434e-01 7.30308294e-01 1.22521126e+00 -2.63644814e-01
7.20320642e-01 3.14317763e-01 -9.30483490e-02 2.43096784e-01
-3.78220044e-02 5.20800114e-01 -2.28806227e-01 5.01896068e-02
8.08902204e-01 -1.51364014e-01 -2.95000821e-01 -4.04135168e-01
-7.17645526e-01 7.53585696e-01 7.80577421e-01 3.13533604e-01
-8.55611444e-01 5.76838136e-01 1.50666565e-01 8.54706228e-01
-8.27707276e-02 9.26303864e-01 -5.77575088e-01 -3.61297756e-01
-5.29890239e-01 4.70223993e-01 6.76880777e-01 2.02236190e-01
7.86462009e-01 2.08802983e-01 -3.49601179e-01 6.33570552e-01
-3.50285321e-02 2.80565113e-01 5.04075825e-01 -1.12044823e+00
2.01051399e-01 4.28577542e-01 6.29873335e-01 -7.20387578e-01
-5.83813727e-01 -3.23846638e-01 -3.61172974e-01 8.99554312e-01
8.23925078e-01 -4.51083839e-01 -6.61541879e-01 2.17523718e+00
2.79073920e-02 -1.04569331e-01 1.48932412e-01 1.10281897e+00
1.69630889e-02 5.67293048e-01 1.18711971e-01 -5.95466554e-01
9.60269451e-01 -5.52792907e-01 -5.73757708e-01 -4.77423221e-01
6.93057835e-01 -2.09087178e-01 9.82230246e-01 6.54922664e-01
-1.11928451e+00 -7.84275383e-02 -1.03172421e+00 3.70294899e-01
-1.36389762e-01 -2.75182098e-01 5.02204776e-01 2.41657659e-01
-9.49586391e-01 1.40604687e+00 -7.56784916e-01 1.22601338e-01
2.42763132e-01 6.21013403e-01 -7.27447793e-02 6.17682874e-01
-1.38368189e+00 1.18471479e+00 3.63527894e-01 9.47053879e-02
-8.52684200e-01 -3.75787139e-01 -3.54130387e-01 4.66532826e-01
6.00462258e-01 -1.53553829e-01 1.41930878e+00 -1.53722095e+00
-1.46060205e+00 5.58613837e-01 3.98998231e-01 -8.82253170e-01
4.46957082e-01 -5.76778315e-02 -7.29777887e-02 -2.41354004e-01
-2.99543798e-01 8.41361821e-01 9.86505389e-01 -1.01706159e+00
-3.96885037e-01 -3.16451341e-01 4.22732115e-01 2.84688503e-01
-3.86683762e-01 -1.29727349e-01 4.99825984e-01 -3.42292219e-01
-3.03113282e-01 -7.61402428e-01 -5.08712828e-01 1.99618608e-01
-6.36916757e-02 -2.43999451e-01 1.68669909e-01 -1.53931350e-01
1.20817161e+00 -2.06067061e+00 6.92706481e-02 3.27814460e-01
9.24934074e-02 2.37973839e-01 -2.24910825e-01 2.46706963e-01
-2.83783615e-01 -1.21960968e-01 -2.79494435e-01 3.03548723e-01
-7.48177245e-02 4.83623534e-01 -2.83776283e-01 4.65391845e-01
6.54109240e-01 7.48405337e-01 -1.07799029e+00 -1.16377629e-01
3.48358378e-02 3.06542575e-01 -6.60587549e-01 2.61331946e-01
-4.01521236e-01 1.50692731e-01 -6.77355766e-01 7.34327082e-03
2.53863752e-01 5.97328283e-02 3.46236289e-01 4.21060234e-01
-2.61459380e-01 6.12171769e-01 -1.53257573e+00 7.07740486e-01
-2.59639084e-01 4.69777703e-01 2.18683332e-01 -1.26238346e+00
1.22481322e+00 2.31414378e-01 4.03622806e-01 -8.62294197e-01
4.47869867e-01 3.59999388e-01 4.63637233e-01 -1.82699487e-01
4.45189148e-01 -5.35998583e-01 6.42117560e-02 5.73434770e-01
-1.31945193e-01 -3.15207124e-01 9.32860598e-02 -1.64205030e-01
1.17620468e+00 1.54831722e-01 4.45249349e-01 -5.06361388e-02
4.38988298e-01 -1.91381872e-01 5.96740544e-01 8.43043387e-01
-6.66054547e-01 9.95025486e-02 1.30157852e+00 -3.91867012e-01
-9.46043253e-01 -7.57736385e-01 1.01533711e-01 1.03167140e+00
6.19493201e-02 4.86678183e-02 -6.44410372e-01 -6.32228553e-01
2.84930348e-01 9.58368897e-01 -6.68378890e-01 -5.52863598e-01
-7.12384582e-01 -5.90370536e-01 7.57820159e-02 5.42624295e-01
-3.82083468e-02 -1.52437556e+00 -1.07274687e+00 3.75509948e-01
2.88487881e-01 -2.88250238e-01 -1.98243812e-01 9.40481424e-01
-9.22257900e-01 -7.84615457e-01 -6.32156432e-01 -3.78351420e-01
6.19078577e-01 -2.93885320e-01 9.77485001e-01 1.97803959e-01
2.03431636e-01 1.62419006e-01 -2.84450743e-02 -3.96587849e-01
-3.98733109e-01 -3.78266454e-01 2.02228993e-01 -4.48757932e-02
3.23403895e-01 -6.31096721e-01 -6.11987412e-01 9.10954624e-02
-6.65830076e-01 -3.54881376e-01 3.08689058e-01 1.10845089e+00
3.23222995e-01 -1.12689748e-01 9.08812523e-01 -5.56739986e-01
1.15558672e+00 -3.69575948e-01 -8.97204876e-01 3.87616009e-02
-8.87441874e-01 5.87567568e-01 1.03865039e+00 -5.67489028e-01
-6.91389322e-01 1.40027642e-01 -1.70306548e-01 -2.88105130e-01
3.66177447e-02 3.54802489e-01 2.90837705e-01 1.10694326e-01
8.54340255e-01 8.88809934e-02 2.08623484e-01 -2.60604650e-01
2.92200655e-01 4.02957320e-01 3.12726080e-01 -5.26607513e-01
3.51060539e-01 -7.46751949e-02 2.30041996e-01 -2.49005273e-01
-6.04828715e-01 5.34152314e-02 -1.76099852e-01 -5.44528604e-01
3.49636018e-01 -4.44261014e-01 -1.19910967e+00 1.65290341e-01
-1.00811791e+00 -5.23694456e-01 -8.77440572e-01 5.28285503e-01
-9.39522505e-01 -2.65325904e-01 -4.53902930e-01 -1.28315520e+00
9.19571519e-02 -1.10948706e+00 3.86879683e-01 4.85666037e-01
-2.40734875e-01 -7.17769384e-01 8.12626630e-02 -4.59180862e-01
3.94847840e-01 3.92215252e-02 8.40342164e-01 -7.16038883e-01
-3.61982077e-01 -1.21302195e-01 1.08204111e-01 6.00285769e-01
-1.54921755e-01 -1.40546203e-01 -8.74416173e-01 -2.04465955e-01
3.33535820e-01 -5.10805607e-01 8.25436711e-01 4.49527442e-01
8.99003267e-01 -5.17627954e-01 1.79648131e-01 6.49937764e-02
1.27471137e+00 5.02169311e-01 6.54101431e-01 5.46032965e-01
3.55738774e-02 7.42639720e-01 7.50137627e-01 6.62601650e-01
4.25613075e-02 8.04053366e-01 8.70660424e-01 1.77473918e-01
3.60850245e-01 -1.58101067e-01 5.35281718e-01 -9.08092945e-04
-2.14776427e-01 2.41381302e-02 -5.23905516e-01 4.80620444e-01
-2.00802708e+00 -1.11044717e+00 9.92962532e-03 2.51727438e+00
1.22606480e+00 4.98865694e-01 4.70450640e-01 2.67805010e-01
4.70925331e-01 3.91513780e-02 -9.67088580e-01 -1.04174423e+00
1.36951745e-01 8.46937671e-02 5.47612965e-01 6.80809140e-01
-7.81876147e-01 8.39062631e-01 6.60447979e+00 6.07467949e-01
-1.35299265e+00 -3.28493387e-01 7.12978542e-01 -2.64938623e-01
-2.95769513e-01 1.28169619e-02 -3.74298304e-01 3.71260077e-01
1.03263342e+00 -2.95219392e-01 9.48625982e-01 1.08958411e+00
3.62977356e-01 -3.13799500e-01 -1.23456216e+00 5.78965545e-01
-5.01275003e-01 -8.47706735e-01 -5.41132748e-01 -1.33815408e-01
3.68660927e-01 -3.47585559e-01 -1.66265041e-01 5.62187314e-01
3.09016049e-01 -1.06731200e+00 7.54018962e-01 6.18814707e-01
6.43008277e-02 -9.10934925e-01 5.76017499e-01 6.25768244e-01
-5.63392937e-01 -2.96684265e-01 -2.93393999e-01 -8.91294777e-01
-9.01997536e-02 2.69348234e-01 -7.57661700e-01 -1.63139462e-01
3.99531983e-02 3.11412692e-01 -2.32325301e-01 8.86517763e-01
-6.73667252e-01 3.75425428e-01 -4.47996289e-01 -7.13634968e-01
4.94608283e-01 -2.32677877e-01 4.89596307e-01 6.11364603e-01
5.62146045e-02 7.36398250e-03 -5.95651940e-02 1.15715969e+00
1.51491955e-01 -1.36965215e-02 -4.87020999e-01 -9.84066427e-02
2.99660504e-01 1.00881875e+00 -6.45667732e-01 -2.41989747e-01
-9.66569185e-02 7.43746936e-01 7.50613451e-01 3.19536865e-01
-7.29600608e-01 -4.22025114e-01 8.44895005e-01 3.38302925e-03
2.29332939e-01 6.97241649e-02 -4.62281078e-01 -7.66486049e-01
-6.12629801e-02 -6.22916281e-01 1.00344256e-01 -7.14218676e-01
-7.75413156e-01 2.11864874e-01 -1.82809174e-01 -1.13119876e+00
-7.73749769e-01 -6.41883790e-01 -6.14969611e-01 9.38473701e-01
-1.34466851e+00 -1.45533651e-01 5.23950517e-01 3.93780768e-01
-2.36938111e-02 2.55541593e-01 6.35429442e-01 3.13755833e-02
-2.82065630e-01 4.35434967e-01 1.37336344e-01 -2.39382103e-01
4.04336005e-01 -1.49341238e+00 -1.42887384e-01 4.03765112e-01
-8.36455002e-02 4.02530789e-01 1.13870060e+00 -2.96075583e-01
-1.29672837e+00 -5.19171476e-01 9.65827703e-01 -1.81493744e-01
7.48659134e-01 -2.44484581e-02 -1.00935125e+00 4.69756603e-01
-1.91404313e-01 -1.02415651e-01 6.46510497e-02 1.54651076e-01
-1.51286572e-01 -2.00508848e-01 -1.30408490e+00 9.10253227e-01
4.97173458e-01 -2.18994379e-01 -7.27102458e-01 -4.05883379e-02
3.60047668e-01 -1.29449442e-01 -5.65460980e-01 1.05737813e-01
7.01648831e-01 -1.19339049e+00 7.68814385e-01 -8.58945012e-01
5.52431524e-01 -2.83805937e-01 1.44174397e-01 -1.74714339e+00
-1.82210013e-01 -6.91435933e-01 -1.79859787e-01 6.98563039e-01
5.06875038e-01 -7.57599890e-01 7.15261221e-01 9.00515974e-01
1.21827841e-01 -1.18970048e+00 -1.09755087e+00 -9.43801284e-01
2.59053111e-01 -3.18098038e-01 4.63711470e-01 7.12943614e-01
4.57574666e-01 5.01845896e-01 -5.52433312e-01 -2.95855582e-01
1.93746209e-01 4.71106805e-02 3.80121589e-01 -8.31631541e-01
-5.10222733e-01 -8.64423275e-01 -4.04178053e-01 -1.15357471e+00
-9.01229009e-02 -6.40415728e-01 2.48421326e-01 -1.16830778e+00
-1.02697432e-01 -2.75198460e-01 -6.31556273e-01 8.60497117e-01
-1.01686217e-01 -1.87217206e-01 5.23752630e-01 -6.16978854e-02
-3.02723765e-01 5.91560066e-01 1.37613738e+00 8.23981594e-03
-6.03344858e-01 1.05605416e-01 -7.59094715e-01 7.96491623e-01
9.96653497e-01 -7.94437528e-01 -3.36645842e-01 7.15536401e-02
5.73339999e-01 3.60462219e-01 4.63431120e-01 -7.58791506e-01
4.26121242e-03 -5.95101833e-01 4.20047134e-01 3.64063047e-02
3.54200900e-01 -8.85668039e-01 -3.00929695e-01 7.97221422e-01
-8.82495522e-01 -1.81871727e-01 1.82265472e-02 3.47748697e-01
-2.28316206e-02 -7.17406571e-01 8.95548224e-01 3.34829576e-02
-3.85715544e-01 -2.06697494e-01 -5.67893505e-01 -4.33491953e-02
9.72661674e-01 -8.33124593e-02 -3.36161911e-01 -6.81531191e-01
-9.25225556e-01 5.28664589e-01 2.83654451e-01 2.70912051e-01
6.12843573e-01 -1.25582409e+00 -5.39472401e-01 1.63461670e-01
-1.51047871e-01 -5.99177241e-01 -3.66076119e-02 8.37959051e-01
1.25587657e-02 1.69046000e-01 -5.13421118e-01 -9.55953971e-02
-9.93428886e-01 5.78354061e-01 6.51944757e-01 -4.64871973e-01
-2.95361608e-01 7.42163718e-01 6.37148023e-02 -9.90633965e-02
3.46872538e-01 -4.03153539e-01 -4.77813989e-01 1.32427782e-01
6.43135905e-01 1.67899147e-01 -1.81687653e-01 -2.12049291e-01
-3.01694900e-01 8.65052491e-02 -1.32935300e-01 -2.05044881e-01
1.42232370e+00 1.37747645e-01 1.50329590e-01 5.90713382e-01
7.81659484e-01 -4.96333808e-01 -1.62800217e+00 9.33157802e-02
1.04075409e-01 -4.54733342e-01 2.01045778e-02 -1.12914717e+00
-9.39370513e-01 8.29363585e-01 5.81762612e-01 7.12277532e-01
1.28756106e+00 -1.92526057e-01 3.56814444e-01 4.87836629e-01
3.80341142e-01 -1.43725562e+00 1.32621210e-02 5.60329020e-01
1.09566438e+00 -1.05216253e+00 -7.75246248e-02 2.82848626e-01
-9.24108803e-01 1.12515247e+00 6.61946356e-01 -5.48263848e-01
3.02512467e-01 2.21555382e-01 -1.21072210e-01 -5.62911248e-03
-1.11406004e+00 -3.50434899e-01 -1.75117627e-01 6.44027114e-01
4.38760579e-01 3.00833523e-01 -6.81613386e-01 5.31194389e-01
-2.93785483e-02 1.13355488e-01 6.55026853e-01 9.00739253e-01
-7.03424215e-01 -1.22746503e+00 -3.32936585e-01 6.33139431e-01
-3.51419061e-01 -9.93428230e-02 -5.84689677e-01 6.87625945e-01
-1.56949282e-01 7.36230791e-01 2.03019947e-01 -2.12030172e-01
4.53593582e-01 6.70565367e-02 6.44115329e-01 -3.07914287e-01
-9.22613680e-01 4.90435958e-02 9.41494554e-02 -5.54640353e-01
-1.65696159e-01 -5.02468050e-01 -1.57018232e+00 -1.91026881e-01
-1.64862752e-01 2.50322111e-02 6.49179935e-01 8.20922673e-01
-4.51118723e-02 6.62093520e-01 8.32359910e-01 -6.17310226e-01
-1.34309328e+00 -6.70851350e-01 -9.08600271e-01 4.01100188e-01
5.92049360e-01 -6.76255047e-01 -4.52634215e-01 -4.10068929e-01] | [4.181521892547607, 2.0895655155181885] |
c9f40663-cd24-42ab-8114-ef9a0ace056d | star-shape-prior-in-fully-convolutional | 1806.08437 | null | http://arxiv.org/abs/1806.08437v1 | http://arxiv.org/pdf/1806.08437v1.pdf | Star Shape Prior in Fully Convolutional Networks for Skin Lesion Segmentation | Semantic segmentation is an important preliminary step towards automatic
medical image interpretation. Recently deep convolutional neural networks have
become the first choice for the task of pixel-wise class prediction. While
incorporating prior knowledge about the structure of target objects has proven
effective in traditional energy-based segmentation approaches, there has not
been a clear way for encoding prior knowledge into deep learning frameworks. In
this work, we propose a new loss term that encodes the star shape prior into
the loss function of an end-to-end trainable fully convolutional network (FCN)
framework. We penalize non-star shape segments in FCN prediction maps to
guarantee a global structure in segmentation results. Our experiments
demonstrate the advantage of regularizing FCN parameters by the star shape
prior and our results on the ISBI 2017 skin segmentation challenge data set
achieve the first rank in the segmentation task among $21$ participating teams. | ['Zahra Mirikharaji', 'Ghassan Hamarneh'] | 2018-06-21 | null | null | null | null | ['skin-lesion-segmentation'] | ['medical'] | [ 7.57933557e-01 5.63660204e-01 -2.51201034e-01 -8.83937061e-01
-1.03258729e+00 -2.64825076e-01 2.49540702e-01 1.69859424e-01
-7.21357286e-01 4.23564881e-01 -7.02196434e-02 -1.47779480e-01
1.32806078e-01 -5.56518376e-01 -9.36798632e-01 -5.81369042e-01
1.26942962e-01 5.54300487e-01 3.84947181e-01 8.83232057e-02
-1.69386733e-02 3.76544088e-01 -9.46121573e-01 3.31144363e-01
1.07069600e+00 1.24622130e+00 2.51919270e-01 4.14425403e-01
-1.08672850e-01 5.16144395e-01 -1.56757534e-01 -5.11594176e-01
3.18107277e-01 -5.52717507e-01 -1.29692888e+00 2.30839387e-01
5.41116297e-01 -1.80519640e-01 4.08435091e-02 1.14595914e+00
4.89020646e-01 1.74566284e-01 5.69782674e-01 -7.47323990e-01
-1.46639571e-01 6.44598782e-01 -5.97331464e-01 1.28705814e-01
-4.05485153e-01 7.82923922e-02 1.04440379e+00 -2.62978047e-01
7.01262474e-01 9.32928026e-01 7.98538208e-01 6.75427020e-01
-1.22362328e+00 -1.17309764e-01 7.90238753e-02 1.42840445e-01
-1.13783312e+00 -6.39647543e-02 8.91624033e-01 -3.59937161e-01
7.86729097e-01 3.30927223e-01 5.98660231e-01 6.72644198e-01
2.64388858e-03 1.27290344e+00 8.71187627e-01 -3.60936046e-01
9.63796675e-02 -1.73023984e-01 1.73151076e-01 9.25021589e-01
1.58632323e-01 -3.29847127e-01 -1.12752981e-01 7.29656219e-02
7.06555307e-01 -2.41364583e-01 -5.45167178e-02 -5.00846803e-01
-7.49592721e-01 7.79515624e-01 9.43429410e-01 2.82804608e-01
-3.90137762e-01 6.66381419e-01 3.72864723e-01 -4.64413553e-01
6.91860199e-01 3.80527854e-01 -4.48441148e-01 -3.11252102e-02
-1.01021683e+00 -3.44422733e-04 2.27547735e-01 2.65235901e-01
4.13563937e-01 -2.60199457e-01 -4.62923229e-01 9.41868722e-01
1.96214497e-01 -1.59457512e-02 2.04541683e-01 -1.19229472e+00
1.00849219e-01 8.05450201e-01 -2.58335304e-02 -5.18330932e-01
-5.83857238e-01 -7.47066855e-01 -5.19165277e-01 2.80687958e-01
6.88637972e-01 -1.90000713e-01 -1.45661008e+00 1.58059585e+00
5.19032061e-01 1.52788386e-01 -3.33237946e-01 1.16809261e+00
6.76111400e-01 3.24905545e-01 2.98884183e-01 1.63378298e-01
1.18572330e+00 -1.15093672e+00 -4.76038575e-01 -2.65726119e-01
5.84785998e-01 -5.10354638e-01 8.60018730e-01 4.40505981e-01
-1.13690412e+00 -3.27915400e-01 -9.44933891e-01 -3.25479776e-01
5.58860414e-02 -1.49497418e-02 7.64823914e-01 7.64880538e-01
-9.07750905e-01 6.84516191e-01 -1.27199745e+00 -2.55019665e-02
1.11508155e+00 5.76239347e-01 1.12837022e-02 -1.47151187e-01
-8.81183028e-01 7.94034541e-01 4.69231606e-01 2.03356087e-01
-6.24037325e-01 -8.10832262e-01 -7.38563895e-01 5.08527756e-02
3.60521317e-01 -6.53623998e-01 1.19063389e+00 -1.38379586e+00
-1.28856277e+00 1.26116502e+00 5.06128073e-02 -8.32924902e-01
8.02053869e-01 -9.70230773e-02 1.34655342e-01 2.59844214e-01
-9.28924605e-02 1.23407257e+00 6.51970744e-01 -1.14535940e+00
-5.69786072e-01 -1.86302915e-01 -6.59512952e-02 2.68628716e-01
-1.83890522e-01 -1.55827790e-01 -6.50690198e-01 -6.04197919e-01
1.22632170e-02 -8.62150013e-01 -5.54738045e-01 4.08321768e-01
-5.78710318e-01 -3.40299189e-01 7.20356762e-01 -8.62228453e-01
8.43689501e-01 -1.92624938e+00 -2.88445074e-02 3.38407993e-01
3.58598307e-02 2.72443265e-01 3.06027774e-02 -5.30627728e-01
9.24586281e-02 1.29949033e-01 -8.40368211e-01 -3.96732837e-01
-1.33504579e-03 1.75530091e-01 2.69851804e-01 4.08204675e-01
4.16256160e-01 1.10191536e+00 -7.76643753e-01 -5.44849634e-01
2.26594627e-01 9.05836999e-01 -6.83954000e-01 -8.29493701e-02
-5.24177492e-01 4.56466377e-01 -5.15938222e-01 5.12808025e-01
5.85391104e-01 -5.46723723e-01 3.44120823e-02 -2.97038734e-01
1.98206469e-01 1.79224655e-01 -6.31407917e-01 2.06661129e+00
-2.41622746e-01 5.06699800e-01 3.13061208e-01 -1.21038043e+00
4.77083176e-01 2.55504161e-01 9.61036861e-01 -7.44899213e-01
3.09996963e-01 1.49796382e-01 2.53987283e-01 -3.40583891e-01
1.18901417e-01 -3.45139593e-01 1.47968158e-01 7.14091137e-02
-3.29687335e-02 -6.70105144e-02 1.11053921e-01 -7.41379559e-02
8.93539250e-01 4.53241348e-01 -3.44512641e-01 -2.93979913e-01
2.10979789e-01 2.98324496e-01 7.11077332e-01 4.34562296e-01
-2.64136761e-01 1.15931416e+00 4.75036591e-01 -4.18118984e-01
-9.66926396e-01 -8.62597048e-01 -3.59692693e-01 9.55201745e-01
1.50897086e-01 9.49609950e-02 -1.24896085e+00 -1.00608981e+00
-2.93952435e-01 7.16747344e-01 -7.31927156e-01 1.28044812e-02
-7.28006423e-01 -1.02966964e+00 4.65228528e-01 6.31547511e-01
6.81084335e-01 -9.29828644e-01 -8.70891988e-01 3.60704154e-01
-4.13931966e-01 -1.20202494e+00 -6.83862984e-01 3.86365592e-01
-8.41059089e-01 -1.05856478e+00 -1.04467010e+00 -7.95381010e-01
9.71575499e-01 -3.23150605e-01 1.04192972e+00 1.09510921e-01
-8.45388412e-01 1.88987017e-01 -8.05702209e-02 -4.26208317e-01
-1.50840819e-01 5.30660041e-02 -7.88968682e-01 1.87257025e-02
5.94014302e-02 -8.31947476e-02 -1.11129570e+00 1.70104042e-01
-8.85130644e-01 3.90056342e-01 5.29309213e-01 7.20503867e-01
1.11932683e+00 7.68398691e-04 2.52260983e-01 -1.07516134e+00
9.85528529e-02 -3.73853706e-02 -4.65085298e-01 2.98282981e-01
-4.87035871e-01 1.80586889e-01 1.34170279e-01 4.82628345e-02
-1.10331273e+00 6.88581407e-01 -6.91912234e-01 -1.47621006e-01
-1.23899654e-01 4.42901880e-01 -1.36681661e-01 -1.87060267e-01
3.38025868e-01 -3.52207571e-02 3.49564455e-03 -4.94718432e-01
1.66437715e-01 1.17788032e-01 5.49772799e-01 -5.20209074e-01
2.51846582e-01 6.28294170e-01 2.62124985e-01 -4.26254451e-01
-1.12518811e+00 -5.19030392e-01 -5.72793782e-01 -1.11696891e-01
1.57324982e+00 -6.39560521e-01 -4.09840852e-01 4.53010827e-01
-1.00442278e+00 -7.62371957e-01 -3.20839554e-01 1.67366073e-01
-5.93360543e-01 4.12250340e-01 -4.21372712e-01 -4.74790066e-01
-6.20301545e-01 -1.32161319e+00 1.12692750e+00 3.57836187e-01
-1.47714794e-01 -1.07136393e+00 -9.65086743e-02 8.68366838e-01
2.84337461e-01 6.50980771e-01 8.86087954e-01 -5.00535965e-01
-4.53211904e-01 -1.53260022e-01 -3.08423042e-01 6.34113908e-01
-1.84368272e-03 -1.02424033e-01 -9.87015963e-01 -6.58409894e-02
-2.27333978e-01 -2.94569850e-01 1.37568247e+00 1.07055068e+00
1.74009597e+00 -4.06215191e-02 -3.31506133e-01 8.47860157e-01
1.54456854e+00 1.37390280e-02 5.16486526e-01 8.61099064e-02
1.03426862e+00 5.52836835e-01 3.31062853e-01 3.45893018e-02
2.95348555e-01 7.07535386e-01 6.67685091e-01 -6.29805326e-01
-4.99927908e-01 -1.60523094e-02 -1.15548603e-01 -6.31649196e-02
-1.15463790e-02 -3.25897992e-01 -1.05473578e+00 7.35046566e-01
-1.61260450e+00 -5.28131664e-01 -2.86901176e-01 1.92287683e+00
9.40002620e-01 4.56224889e-01 9.25561264e-02 -1.51474163e-01
6.42766535e-01 -1.18081287e-01 -7.78489828e-01 -3.24339300e-01
-2.12263186e-02 5.35651803e-01 7.16200590e-01 5.65653861e-01
-1.46632504e+00 8.61158967e-01 5.20775795e+00 1.05255902e+00
-1.22581637e+00 1.77487031e-01 1.29255664e+00 -4.10666913e-02
-3.56000543e-01 -2.27493346e-01 -5.26378870e-01 4.47935671e-01
6.08494937e-01 4.35491472e-01 -2.00912170e-02 6.17475808e-01
3.10792327e-01 -2.07196236e-01 -9.39305723e-01 7.50438690e-01
-2.67527461e-01 -1.37550282e+00 -9.71902162e-02 1.19925220e-03
9.79259670e-01 3.30793738e-01 2.23868385e-01 -1.43717617e-01
2.46795833e-01 -1.43024325e+00 6.31060898e-01 4.03738797e-01
6.79611921e-01 -7.97892928e-01 6.77871525e-01 7.54681751e-02
-7.66628981e-01 3.49345535e-01 -1.37286186e-01 4.14659947e-01
3.93706679e-01 7.32044995e-01 -9.04346645e-01 2.71304935e-01
5.21206975e-01 3.74646306e-01 -4.14934903e-01 1.35965872e+00
-4.24493924e-02 9.45622504e-01 -5.05254388e-01 2.88973182e-01
6.71435535e-01 -6.17825761e-02 3.24854553e-01 1.16365707e+00
-5.20705618e-02 1.08651683e-01 1.09281033e-01 9.36892450e-01
-3.32588702e-01 -1.21698931e-01 1.73891291e-01 -2.40730923e-02
-2.70054519e-01 1.25975096e+00 -1.34079456e+00 2.92884763e-02
-1.67772949e-01 1.16432798e+00 6.85172379e-02 1.67782426e-01
-7.92408705e-01 -1.63854107e-01 4.65085000e-01 2.86380589e-01
4.25111324e-01 9.00910199e-02 -9.83986080e-01 -5.95942497e-01
-5.83651699e-02 -4.18084174e-01 4.88097787e-01 -4.71693039e-01
-1.14251757e+00 4.05106992e-01 -2.67347693e-01 -6.71023786e-01
1.79057062e-01 -6.88774467e-01 -7.13487566e-01 8.18007886e-01
-1.61252451e+00 -1.17420757e+00 -2.20905825e-01 2.16870829e-01
5.95845282e-01 4.94690537e-01 5.41601837e-01 3.68925065e-01
-5.99811673e-01 5.71365952e-01 4.27791998e-02 2.95060247e-01
3.50227326e-01 -1.41445196e+00 3.32821578e-01 7.83409357e-01
1.66748445e-02 1.33951843e-01 7.19773412e-01 -5.14722884e-01
-8.17145944e-01 -1.16805029e+00 3.96867573e-01 -3.59024405e-01
1.30776078e-01 -1.30527496e-01 -7.86505103e-01 3.49686414e-01
1.66938201e-01 3.40746850e-01 6.18359029e-01 -1.28923327e-01
1.86103776e-01 3.40021551e-02 -1.33722150e+00 2.14583740e-01
7.02516794e-01 -1.04763485e-01 -2.55102143e-02 4.68280703e-01
6.07273519e-01 -6.46512747e-01 -7.63985097e-01 7.30788589e-01
3.13581914e-01 -5.92724025e-01 1.07378471e+00 -4.33438897e-01
5.47017515e-01 -1.33116022e-01 1.61293671e-01 -1.00365031e+00
-2.10124925e-02 -3.90115052e-01 2.19490096e-01 8.14179480e-01
5.39873838e-01 -1.02414768e-02 1.46241105e+00 1.01601374e+00
-5.50791323e-01 -1.18612123e+00 -1.04195333e+00 -4.56796169e-01
3.32087427e-01 -3.79564494e-01 2.32796129e-02 7.05767095e-01
-4.83564466e-01 -5.52589148e-02 -4.21922542e-02 -1.59832940e-01
8.83903861e-01 -8.90247896e-02 1.15170695e-01 -1.11611474e+00
-4.02147561e-01 -8.97356927e-01 -2.05615222e-01 -1.03288507e+00
8.31300095e-02 -1.10660410e+00 1.61782935e-01 -2.12215137e+00
4.90636170e-01 -6.14866853e-01 -5.25213003e-01 6.49635375e-01
-3.29042017e-01 5.78369796e-01 -7.68908188e-02 -2.67517447e-01
-7.16119230e-01 3.62188846e-01 1.45996976e+00 -3.92963439e-01
2.21893787e-02 1.06786199e-01 -5.25164664e-01 8.01061451e-01
7.83019066e-01 -4.89016622e-01 -3.93481076e-01 -6.66094542e-01
2.20768228e-02 -2.32482955e-01 5.89770257e-01 -8.30079257e-01
-2.51469892e-02 -1.07730269e-01 5.48219144e-01 -5.99493444e-01
2.58027643e-01 -8.10451150e-01 -6.56852424e-02 6.72293365e-01
-5.79502463e-01 -7.75490046e-01 2.83156902e-01 3.88683140e-01
-6.46139085e-02 -2.97699571e-01 1.14407527e+00 -1.70190990e-01
-6.60745561e-01 4.79318798e-01 7.01691806e-02 2.84284592e-01
9.19857025e-01 -4.10411090e-01 1.12146072e-01 4.61961664e-02
-9.45590079e-01 3.90251368e-01 3.00886571e-01 9.16556343e-02
4.48917478e-01 -8.99461031e-01 -6.86766803e-01 -1.95213318e-01
-2.52792597e-01 5.21673620e-01 3.93171817e-01 9.01277781e-01
-9.33496296e-01 3.02824229e-01 -2.05034271e-01 -7.41954744e-01
-1.18012881e+00 -1.72087383e-02 7.66664863e-01 -4.83219326e-01
-6.68679953e-01 1.52361631e+00 3.26301098e-01 -3.62949818e-01
3.27391297e-01 -4.07029718e-01 -2.51709055e-02 -3.69422227e-01
-8.14593397e-03 1.26944914e-01 2.99499243e-01 -5.12199640e-01
-4.65838701e-01 3.83716315e-01 -1.78138062e-01 -5.88304028e-02
1.40638697e+00 1.04450747e-01 6.07577004e-02 -3.48844267e-02
1.35119259e+00 -5.24763465e-01 -1.82069600e+00 -1.04080461e-01
6.76282048e-02 -5.74459881e-02 5.95142782e-01 -1.29451478e+00
-1.46241033e+00 9.29938912e-01 9.75017607e-01 -2.54388124e-01
1.10186446e+00 1.02727085e-01 1.20206666e+00 1.40138268e-02
3.36367823e-02 -1.30305886e+00 -1.15084916e-03 1.81089878e-01
4.22543436e-01 -1.40433860e+00 -9.76787880e-02 -5.92566669e-01
-6.62835419e-01 8.38502705e-01 6.15393162e-01 -9.80061293e-02
5.86178362e-01 1.49453774e-01 -4.50034998e-02 -2.27897644e-01
-2.85863847e-01 -2.35012442e-01 7.51026690e-01 3.72410685e-01
5.86771011e-01 1.01173401e-01 -4.39742595e-01 5.10405779e-01
1.75797284e-01 -1.15134632e-02 9.78598893e-02 7.42007852e-01
-5.10854781e-01 -1.15721524e+00 3.22496258e-02 5.06586015e-01
-9.86344695e-01 -2.18080170e-02 -3.88324648e-01 4.56001848e-01
3.92684340e-01 5.09218037e-01 1.79999843e-01 1.55957028e-01
1.16310000e-01 -7.70948902e-02 7.18488753e-01 -6.10565484e-01
-6.83809161e-01 1.81535020e-01 -2.78306250e-02 -5.24646521e-01
-3.44736606e-01 -6.34606302e-01 -1.97178125e+00 2.11772919e-01
-1.23751611e-01 2.78427973e-02 8.66712987e-01 1.03439415e+00
1.19387798e-01 7.82291234e-01 1.88994184e-01 -6.02343678e-01
-2.10811228e-01 -4.95973855e-01 -2.88059264e-01 5.53720236e-01
1.46268860e-01 -4.31486517e-01 1.95928246e-01 2.62683958e-01] | [14.557394027709961, -2.2799735069274902] |
c2e9dc57-fa61-486b-800d-3a819d304151 | mm-dfn-multimodal-dynamic-fusion-network-for | 2203.02385 | null | https://arxiv.org/abs/2203.02385v1 | https://arxiv.org/pdf/2203.02385v1.pdf | MM-DFN: Multimodal Dynamic Fusion Network for Emotion Recognition in Conversations | Emotion Recognition in Conversations (ERC) has considerable prospects for developing empathetic machines. For multimodal ERC, it is vital to understand context and fuse modality information in conversations. Recent graph-based fusion methods generally aggregate multimodal information by exploring unimodal and cross-modal interactions in a graph. However, they accumulate redundant information at each layer, limiting the context understanding between modalities. In this paper, we propose a novel Multimodal Dynamic Fusion Network (MM-DFN) to recognize emotions by fully understanding multimodal conversational context. Specifically, we design a new graph-based dynamic fusion module to fuse multimodal contextual features in a conversation. The module reduces redundancy and enhances complementarity between modalities by capturing the dynamics of contextual information in different semantic spaces. Extensive experiments on two public benchmark datasets demonstrate the effectiveness and superiority of MM-DFN. | ['Yang Mo', 'Lianxin Jiang', 'Lingwei Wei', 'Xiaolong Hou', 'Dou Hu'] | 2022-03-04 | null | null | null | null | ['emotion-recognition-in-conversation'] | ['natural-language-processing'] | [ 3.74086574e-02 -1.08779795e-01 5.36448881e-02 -4.18122619e-01
-5.37675977e-01 -5.27571023e-01 6.63978219e-01 1.79613695e-01
-2.06024036e-01 5.10034740e-01 9.27298009e-01 1.35200843e-01
-1.44692242e-01 -6.28840506e-01 9.45643187e-02 -7.35631824e-01
4.26088989e-01 5.75939938e-02 -3.91777128e-01 -7.24811375e-01
-5.61537109e-02 2.61143029e-01 -1.13190210e+00 9.68936861e-01
9.19764042e-01 9.04587507e-01 -6.53454214e-02 4.36432540e-01
-7.00422049e-01 9.52881515e-01 -3.80009472e-01 -1.11853921e+00
-4.36404824e-01 -7.03792691e-01 -7.97319353e-01 1.00257009e-01
-2.34265149e-01 3.53247732e-01 -5.78507483e-01 1.05658269e+00
4.75918323e-01 4.31626976e-01 2.86021829e-01 -1.69702768e+00
-3.80985349e-01 8.80074263e-01 -4.53879982e-01 -1.87855482e-01
9.34968531e-01 -1.11366548e-01 1.03764451e+00 -8.72518599e-01
6.64774776e-01 1.66167295e+00 6.42041445e-01 7.04808235e-01
-9.10209179e-01 -4.35040534e-01 5.25442719e-01 5.01669824e-01
-1.00758994e+00 -2.21449926e-01 1.32606578e+00 9.54670925e-03
8.16580176e-01 3.87948692e-01 8.40123355e-01 1.37802196e+00
1.22207906e-02 1.03138530e+00 1.00255573e+00 -2.50699043e-01
4.99872454e-02 -3.19897644e-02 2.77348906e-01 5.93791366e-01
-3.90128672e-01 -4.13979441e-01 -1.17000318e+00 -3.85012358e-01
7.07047358e-02 3.36808026e-01 -3.15909326e-01 1.80193465e-02
-1.19072187e+00 8.82697344e-01 5.14490128e-01 4.47300911e-01
-4.25296694e-01 -1.38293192e-01 6.12849474e-01 3.70160520e-01
3.35053772e-01 6.31016791e-02 1.24261238e-01 -3.70652348e-01
-9.03768465e-02 -1.27728790e-01 7.65977681e-01 5.85402548e-01
6.82348490e-01 -3.32977861e-01 -7.39818737e-02 1.18079031e+00
3.34849656e-01 3.75772327e-01 2.70375907e-01 -1.06478131e+00
5.80316186e-01 1.35271049e+00 -4.42088932e-01 -1.53870225e+00
-8.20667148e-01 5.88366836e-02 -1.07800603e+00 -4.62256104e-01
-1.46418795e-01 -5.00733554e-01 -3.35743129e-01 2.06289721e+00
3.97681624e-01 2.07214788e-01 5.50402999e-01 9.86138046e-01
1.33875811e+00 5.75112283e-01 3.80679458e-01 -2.39414930e-01
1.39555633e+00 -9.06338811e-01 -1.08922124e+00 -2.49292165e-01
6.24747396e-01 -6.70776904e-01 6.66966081e-01 1.34304687e-01
-8.44837427e-01 -1.41118079e-01 -6.30450487e-01 -1.04933150e-01
-5.90121329e-01 -2.41762921e-01 9.74879265e-01 3.99164051e-01
-8.78701150e-01 7.34900311e-02 -2.73913324e-01 -4.11499381e-01
1.05208606e-01 2.34282061e-01 -8.26537192e-01 -3.20046872e-01
-1.55298090e+00 8.11170101e-01 3.76071364e-01 4.78475511e-01
-4.33007888e-02 -1.34177133e-01 -1.13263083e+00 1.53750807e-01
3.81985366e-01 -7.81953216e-01 7.43195474e-01 -1.00505543e+00
-1.50989830e+00 4.36799645e-01 -3.42327237e-01 6.04933761e-02
1.05293535e-01 1.56810194e-01 -9.44409192e-01 3.82131368e-01
-4.02389020e-01 7.20795810e-01 5.19052327e-01 -1.58700383e+00
-7.13257134e-01 -6.17723048e-01 2.47915015e-01 7.00031221e-01
-3.61563176e-01 -1.74108684e-01 -4.87009257e-01 -3.49128187e-01
3.30675662e-01 -8.95858228e-01 -1.12788692e-01 -5.66610038e-01
-1.93631142e-01 -3.74563873e-01 1.06273794e+00 -4.63167489e-01
1.30869746e+00 -2.13760662e+00 7.52770662e-01 2.79403597e-01
4.85631824e-01 -9.85010192e-02 -4.85131770e-01 8.09085727e-01
1.63608164e-01 -1.43995941e-01 -2.92524546e-01 -8.82601619e-01
1.72938287e-01 4.51159924e-01 -5.07854670e-02 1.60703003e-01
3.54654603e-02 1.16888106e+00 -9.29289460e-01 -7.12245286e-01
3.65114599e-01 7.71027505e-01 -4.38980460e-01 1.84333753e-02
1.05115451e-01 4.95539278e-01 -2.58857042e-01 8.28975081e-01
7.14589894e-01 -8.95851627e-02 6.95719421e-01 -3.92619222e-01
3.38417411e-01 -4.13921267e-01 -9.32908237e-01 2.19819307e+00
-6.49619758e-01 6.09657109e-01 2.97588289e-01 -1.01021361e+00
8.91093254e-01 3.98729682e-01 5.99028885e-01 -6.41133130e-01
4.75218177e-01 -1.97523549e-01 -4.18787390e-01 -5.90596199e-01
8.02746296e-01 -4.87448126e-01 -7.64228404e-01 2.69399345e-01
4.00017023e-01 7.33163282e-02 -2.55022943e-02 6.56084359e-01
7.70146728e-01 -4.20491278e-01 5.57141230e-02 3.30597669e-01
8.51703227e-01 -2.15354145e-01 7.62970448e-01 2.36246705e-01
-3.79063517e-01 2.00820550e-01 6.84115648e-01 -1.64263129e-01
-3.75487804e-02 -9.37780559e-01 4.14208472e-01 1.22532475e+00
5.99359930e-01 -7.22621441e-01 -5.16312778e-01 -8.63769531e-01
-1.13507763e-01 5.98653674e-01 -7.19890118e-01 -4.82457906e-01
-1.67394012e-01 -6.58722937e-01 5.11447787e-01 3.76647383e-01
5.41915357e-01 -9.49464560e-01 4.56072465e-02 2.23423690e-01
-1.19742322e+00 -1.37545228e+00 -4.58127320e-01 -3.13625693e-01
-4.34948921e-01 -1.01941419e+00 -2.61626601e-01 -5.80811977e-01
3.04577440e-01 5.95466673e-01 9.26335514e-01 -6.14221208e-03
-2.88815959e-03 1.03317189e+00 -6.43530726e-01 -1.19598091e-01
-9.20005962e-02 -4.84673493e-02 -2.17567220e-01 6.07364178e-01
4.61276233e-01 -5.79462469e-01 -5.02141178e-01 3.27089608e-01
-8.20730805e-01 2.01300010e-01 2.50525326e-01 9.48230028e-01
4.72837240e-02 -1.19544744e-01 8.49891543e-01 -5.40689468e-01
1.06857705e+00 -8.55506241e-01 1.89643517e-01 8.10899496e-01
-4.09295075e-02 -1.72253981e-01 3.00675929e-01 -3.65841508e-01
-1.72831786e+00 -3.48653756e-02 7.22166523e-02 -3.66399765e-01
5.38098440e-02 8.33495975e-01 -5.31378686e-01 8.73903371e-03
4.86935638e-02 -9.27570313e-02 5.01171947e-02 -1.47838637e-01
7.11906314e-01 7.43868053e-01 7.45623827e-01 -7.27548122e-01
1.71304755e-02 6.48174167e-01 -4.34882045e-02 -6.50020897e-01
-6.71920657e-01 -7.61198878e-01 -3.45702976e-01 -8.59801531e-01
1.04710364e+00 -9.94397819e-01 -1.39130402e+00 4.55698252e-01
-1.23356652e+00 1.86978891e-01 1.35553077e-01 4.97116834e-01
-3.57958794e-01 5.39866745e-01 -5.93701899e-01 -9.16555822e-01
-2.99490988e-01 -9.05176342e-01 1.08416450e+00 7.14643836e-01
-1.59465447e-01 -1.49346256e+00 2.37689003e-01 8.84674191e-01
1.63060680e-01 4.40617770e-01 7.16207743e-01 -3.28782558e-01
1.43373311e-01 -1.60769612e-01 -3.80493402e-01 3.74956056e-02
1.27377436e-01 -1.71431392e-01 -9.13746297e-01 1.62306085e-01
-2.46414870e-01 -3.56936663e-01 9.79825974e-01 -7.24185929e-02
6.90727115e-01 -1.24090689e-03 -3.03755194e-01 1.94701478e-01
1.16947269e+00 1.85872570e-01 5.20881355e-01 -1.70395210e-01
8.36953223e-01 9.44636047e-01 3.17300379e-01 5.78242898e-01
9.23679411e-01 2.50643909e-01 3.84365380e-01 -2.95845605e-02
2.08880097e-01 -1.59586966e-01 4.45812404e-01 1.15340662e+00
-2.04546735e-01 -6.11642778e-01 -7.83030987e-01 3.81815642e-01
-2.39283371e+00 -1.20185494e+00 2.76645757e-02 1.34265065e+00
4.66102690e-01 -4.68150079e-01 1.41105521e-02 -2.37808913e-01
9.90419388e-01 2.44551033e-01 -2.18111113e-01 -5.35823405e-01
-7.46935070e-01 -3.00143361e-01 -3.61562163e-01 6.96487546e-01
-7.94580877e-01 9.34079468e-01 5.14532757e+00 7.48422027e-01
-7.57994354e-01 1.09854825e-01 3.24243248e-01 -1.20037161e-01
-8.36178541e-01 2.50891894e-02 -2.66969889e-01 2.72693276e-01
4.86061245e-01 -1.19568065e-01 5.96457720e-01 4.83836621e-01
-1.01226278e-01 -2.22829312e-01 -5.76727450e-01 1.57267559e+00
3.95037919e-01 -1.38035083e+00 1.45449355e-01 -1.94229543e-01
7.09343553e-01 -2.98733115e-01 -2.76077926e-01 5.61020017e-01
3.93530875e-01 -8.07005107e-01 1.99436069e-01 1.16493046e+00
2.78899968e-01 -1.23370969e+00 9.58422184e-01 4.08328921e-02
-1.41491807e+00 -2.58956462e-01 1.14894018e-01 9.37661231e-02
5.44546366e-01 2.74147034e-01 -1.37712017e-01 1.07911193e+00
3.95923287e-01 8.40481043e-01 -5.40472567e-01 3.74838680e-01
1.91519678e-01 -5.41447138e-04 -2.40067542e-01 -6.62502721e-02
3.50191474e-01 -5.00846863e-01 5.72743297e-01 1.42086613e+00
1.77260980e-01 6.16056621e-01 1.53418496e-01 3.26216966e-01
-2.44868711e-01 1.97083622e-01 -5.36468983e-01 -2.31866255e-01
4.52245235e-01 1.58636451e+00 -4.53832775e-01 -1.85677931e-01
-6.09050393e-01 1.12939858e+00 5.40602088e-01 3.51465136e-01
-8.74331951e-01 -2.43894607e-01 7.32684016e-01 -8.89976621e-01
-3.89542550e-01 -8.17267373e-02 -3.69288653e-01 -1.44851279e+00
-2.40664005e-01 -8.01562190e-01 9.58363712e-01 -7.90177882e-01
-1.52323890e+00 6.23550117e-01 -1.37270346e-01 -9.25313652e-01
8.95190835e-02 -4.16197896e-01 -7.79802442e-01 5.60269117e-01
-1.14879811e+00 -1.41872501e+00 -6.35277987e-01 1.16586971e+00
2.49654353e-01 -1.84861854e-01 7.75244117e-01 1.06034718e-01
-7.09779322e-01 5.63810349e-01 -2.94828027e-01 1.57298129e-02
6.89490139e-01 -9.31566358e-01 -4.89312500e-01 4.36904460e-01
-3.31583433e-02 5.45685530e-01 3.94905001e-01 -5.79093039e-01
-1.64019954e+00 -5.84907174e-01 9.13521111e-01 -1.78643405e-01
6.55880928e-01 -1.93809643e-01 -9.57197785e-01 3.49282563e-01
6.92938745e-01 -4.65707630e-01 1.23351145e+00 5.38514078e-01
-4.59792912e-01 2.03653686e-02 -1.17105103e+00 7.08688736e-01
1.12746489e+00 -1.02529490e+00 -6.96203232e-01 -3.56347440e-03
7.67714381e-01 -4.54711989e-02 -1.03019965e+00 5.17121673e-01
5.27311206e-01 -1.02209091e+00 7.17993975e-01 -5.88725984e-01
3.23653817e-01 -7.86583591e-03 -5.25932908e-01 -1.56539583e+00
1.16092972e-01 -7.12257504e-01 -5.85794337e-02 1.41121650e+00
2.29224563e-01 -6.11790895e-01 3.98760885e-01 7.57543385e-01
-1.07268624e-01 -5.14315963e-01 -8.58769059e-01 -7.70811141e-02
-2.85318404e-01 -6.18191779e-01 8.40805054e-01 1.43864346e+00
1.06070256e+00 6.84747100e-01 -4.74413306e-01 -4.56208996e-02
4.04046774e-01 2.30761513e-01 5.32719970e-01 -1.17952824e+00
1.45278126e-01 -6.36496365e-01 -3.97800595e-01 -3.68936390e-01
6.96322143e-01 -8.41383040e-01 -2.44178042e-01 -1.64345503e+00
2.85482973e-01 5.47470823e-02 -3.37910473e-01 5.01721501e-01
-3.65691632e-01 -2.75958162e-02 5.14396727e-01 -1.68537036e-01
-1.09095430e+00 9.83578026e-01 1.15834486e+00 -1.19654574e-01
-2.25143760e-01 -5.44733286e-01 -7.11692691e-01 7.29044735e-01
8.77501607e-01 1.16075147e-02 -4.77688283e-01 -1.71135679e-01
4.90724683e-01 6.24283135e-01 5.16430378e-01 -6.60721242e-01
5.54949522e-01 -3.90532881e-01 4.78643179e-02 -7.36922860e-01
7.65680611e-01 -8.71906757e-01 1.97979704e-01 -1.64600655e-01
-3.15677881e-01 2.01428741e-01 2.18532652e-01 8.56723309e-01
-7.09990442e-01 2.14016855e-01 4.28436697e-01 1.10729285e-01
-1.02724159e+00 -1.66008100e-02 -4.39617306e-01 -1.45772159e-01
7.80320525e-01 1.05991073e-01 -5.69943070e-01 -8.59216511e-01
-1.11356139e+00 6.16002977e-01 2.17769757e-01 7.12316394e-01
8.95189464e-01 -1.77912760e+00 -4.63397652e-01 -2.38419905e-01
4.69424546e-01 -4.99648869e-01 1.20795202e+00 1.01747870e+00
1.26398623e-01 9.75914672e-02 -2.06020519e-01 -3.51991683e-01
-1.54772437e+00 4.50503856e-01 3.65666360e-01 -1.50708482e-01
-1.29660711e-01 6.82688475e-01 1.04670353e-01 -7.36342192e-01
2.65101735e-02 2.34631225e-01 -4.26320165e-01 6.27791464e-01
4.15594727e-01 3.52778733e-01 -2.04889849e-01 -1.10020351e+00
-4.49079871e-01 4.09534216e-01 2.84532607e-01 -2.77810663e-01
1.06973779e+00 -5.51162004e-01 -3.98763627e-01 4.28994268e-01
1.34063780e+00 -7.10681453e-02 -8.27421784e-01 -5.40173650e-01
-1.59221232e-01 -1.98237821e-01 -4.11860929e-05 -8.73562813e-01
-1.10349452e+00 8.22318196e-01 1.89185664e-01 1.39643818e-01
1.47694051e+00 9.97137874e-02 8.77222717e-01 4.13529009e-01
7.97031671e-02 -1.35448766e+00 3.02790165e-01 5.58697939e-01
9.52671826e-01 -1.31857491e+00 -3.65061164e-01 -4.77326542e-01
-1.45770812e+00 1.24524367e+00 5.31190872e-01 4.98486251e-01
6.55143380e-01 6.25314340e-02 2.17497632e-01 -5.16372502e-01
-9.23676074e-01 -4.73608732e-01 4.09565032e-01 4.34626937e-01
8.19530487e-02 3.38181227e-01 -3.04517508e-01 9.84339237e-01
2.70893294e-02 -3.71446103e-01 2.26286992e-01 8.34178984e-01
-2.08094954e-01 -9.37226474e-01 -3.04990023e-01 -2.43301526e-01
-1.83243025e-02 7.91182593e-02 -9.60783243e-01 6.30249500e-01
1.54338673e-01 1.67360187e+00 -1.37438774e-01 -9.09502685e-01
3.05886656e-01 3.67655873e-01 3.11813802e-01 8.01477209e-02
-8.14714015e-01 1.24692805e-02 4.02303755e-01 -6.09475613e-01
-9.46049154e-01 -7.34687507e-01 -1.41208696e+00 -5.57977855e-01
-2.06907243e-01 3.30551326e-01 7.32721865e-01 1.13454282e+00
6.12058043e-01 3.77609015e-01 7.39873707e-01 -5.57722569e-01
3.15769762e-01 -7.61575341e-01 -3.84803206e-01 5.85663319e-01
1.22856917e-02 -8.49078000e-01 -5.08040905e-01 -2.29484886e-01] | [13.106938362121582, 5.353387832641602] |
b673a4ae-8ff6-4609-94c4-22e0931a0adc | hinglish-language-modeling-a-messy-code-mixed | 1912.13109 | null | https://arxiv.org/abs/1912.13109v1 | https://arxiv.org/pdf/1912.13109v1.pdf | "Hinglish" Language -- Modeling a Messy Code-Mixed Language | With a sharp rise in fluency and users of "Hinglish" in linguistically diverse country, India, it has increasingly become important to analyze social content written in this language in platforms such as Twitter, Reddit, Facebook. This project focuses on using deep learning techniques to tackle a classification problem in categorizing social content written in Hindi-English into Abusive, Hate-Inducing and Not offensive categories. We utilize bi-directional sequence models with easy text augmentation techniques such as synonym replacement, random insertion, random swap, and random deletion to produce a state of the art classifier that outperforms the previous work done on analyzing this dataset. | ['Vivek Kumar Gupta'] | 2019-12-30 | null | null | null | null | ['text-augmentation'] | ['natural-language-processing'] | [-3.24417472e-01 -8.87769386e-02 3.80135253e-02 -2.74788529e-01
-9.92245227e-02 -7.66740620e-01 9.59946692e-01 3.28959256e-01
-5.54205239e-01 1.00120807e+00 6.56317174e-01 -4.31160301e-01
3.96527946e-01 -5.32017469e-01 1.36319613e-02 -3.12022865e-01
1.43042162e-01 5.96457422e-01 -3.59274209e-01 -7.62859941e-01
4.92657751e-01 5.42951882e-01 -8.87398243e-01 3.53217572e-01
9.30488348e-01 2.05796942e-01 -4.56402659e-01 5.46229601e-01
-3.42551768e-01 1.35062265e+00 -9.08069611e-01 -1.02615106e+00
3.20872888e-02 -3.96061301e-01 -1.08255863e+00 -2.24094748e-01
6.14767790e-01 -3.39401126e-01 -5.71509600e-01 1.13237453e+00
7.31420338e-01 9.88083035e-02 9.44147885e-01 -1.10454547e+00
-1.10212505e+00 1.13952982e+00 -9.79688227e-01 5.73533952e-01
3.90260905e-01 -7.03173131e-02 9.51166213e-01 -6.51551366e-01
9.03154314e-01 1.31984329e+00 9.17328894e-01 4.93735582e-01
-1.00918782e+00 -9.27743971e-01 -5.02101719e-01 2.14148179e-01
-1.34023285e+00 -2.58193612e-01 9.43419993e-01 -7.97149420e-01
1.00569665e+00 1.85322046e-01 3.39004189e-01 1.78399587e+00
1.11733884e-01 9.63331640e-01 1.27525079e+00 -2.29177326e-01
-3.23578328e-01 5.35584629e-01 4.44582790e-01 4.43069309e-01
-8.80341455e-02 -4.78336573e-01 -4.36375529e-01 -3.25189024e-01
-2.05430493e-01 -2.76018530e-02 2.28673205e-01 3.94248605e-01
-6.91386938e-01 1.63345456e+00 4.46941495e-01 5.79258800e-01
-1.36949196e-01 -1.79324433e-01 1.15533960e+00 4.38294530e-01
1.11052132e+00 7.77510405e-01 -1.97598368e-01 -4.20674473e-01
-1.14763582e+00 2.33842611e-01 7.94108152e-01 5.63609600e-01
4.00001347e-01 1.13168918e-01 1.39845446e-01 1.49079359e+00
7.76323080e-02 3.32839042e-01 1.03147042e+00 -4.19687927e-01
4.97418791e-01 4.02745634e-01 -1.77473575e-01 -1.27559865e+00
-5.85122108e-01 -4.37340885e-01 -7.76284516e-01 -2.32238948e-01
-8.35299343e-02 -6.35786474e-01 -8.33391726e-01 1.58517444e+00
3.92508879e-02 -2.18370155e-01 -2.54553825e-01 2.50174761e-01
8.89219880e-01 7.76766837e-01 2.35139504e-01 -5.29566929e-02
8.50440443e-01 -8.44263554e-01 -9.21441078e-01 -1.98277030e-02
1.26460290e+00 -9.22281742e-01 1.13931966e+00 3.93431932e-01
-8.63404036e-01 -1.74116507e-01 -9.56332147e-01 -4.58340436e-01
-1.10427928e+00 -2.01520681e-01 3.25339258e-01 1.07169938e+00
-8.25641632e-01 6.28119349e-01 -1.09973297e-01 -7.61408210e-01
6.70893133e-01 1.40280694e-01 -6.61120772e-01 3.21851045e-01
-1.52010632e+00 1.39058173e+00 4.63470638e-01 -3.75181675e-01
-4.14370984e-01 -6.39460027e-01 -7.31594741e-01 -5.15841722e-01
-2.66740263e-01 3.09118569e-01 9.65714633e-01 -1.20386839e+00
-1.19501579e+00 1.52516937e+00 3.53377938e-01 -4.95285153e-01
5.02039790e-01 -6.59730613e-01 -7.25054443e-01 -3.90173584e-01
2.56959170e-01 5.29195249e-01 6.36060655e-01 -8.71231675e-01
-2.33143404e-01 -4.99479115e-01 -4.85679582e-02 1.48550972e-01
-8.52776587e-01 7.13913560e-01 5.42202353e-01 -1.04101443e+00
-5.74424863e-01 -1.16115081e+00 3.98222893e-01 -6.96550786e-01
-9.54392433e-01 -5.01933396e-01 1.37341833e+00 -1.37929559e+00
1.68898892e+00 -2.15329719e+00 3.11760008e-01 1.46191284e-01
4.90530133e-01 6.86320305e-01 2.92038828e-01 9.08435583e-01
-1.00281134e-01 6.90284729e-01 -9.53652635e-02 -3.42734993e-01
-1.61960781e-01 2.60414809e-01 -2.84481287e-01 7.34046578e-01
-7.68129230e-02 6.41982913e-01 -9.65700030e-01 -5.00117958e-01
1.19203284e-01 3.41497689e-01 -4.45725530e-01 2.04916224e-02
2.26134419e-01 1.40831575e-01 -2.95338556e-02 5.24413228e-01
5.75953066e-01 2.25797802e-01 6.00482821e-02 4.34002042e-01
-2.33158082e-01 4.21646386e-01 -3.46793324e-01 1.11884594e+00
-4.50259298e-01 1.29453003e+00 -3.84720489e-02 -6.39389217e-01
1.10184455e+00 -8.43309239e-02 3.52082625e-02 -4.52232122e-01
6.08492851e-01 1.37107134e-01 2.49810040e-01 -5.02807617e-01
7.61740208e-01 -2.20334068e-01 -4.11982119e-01 4.47331160e-01
3.63889709e-02 2.66354531e-02 1.28601775e-01 6.14570439e-01
9.19552565e-01 -1.36793882e-01 3.91977310e-01 -4.85545605e-01
6.29418790e-01 -1.73982769e-01 -1.25307059e-02 5.64328253e-01
-5.82661629e-01 3.90502721e-01 6.04030490e-01 -4.88399208e-01
-1.30319166e+00 -6.91381216e-01 -1.62457198e-01 1.70394373e+00
-6.47330642e-01 -3.94899160e-01 -8.24146867e-01 -8.13658118e-01
1.80774346e-01 1.01224267e+00 -8.72843146e-01 -3.57076854e-01
-5.60942829e-01 -7.42806971e-01 1.11476111e+00 2.32435137e-01
6.93007052e-01 -1.30723441e+00 1.19137131e-02 2.64642298e-01
-9.09379032e-03 -7.55935848e-01 -1.54493973e-01 2.62513936e-01
-9.07742605e-02 -5.45905650e-01 -6.09940886e-01 -6.68584824e-01
4.73641753e-02 -1.86749622e-01 1.03100467e+00 1.73854008e-01
-2.72847563e-01 -4.95134175e-01 -7.15260923e-01 -7.10937142e-01
-7.72298872e-01 6.87377036e-01 1.60702646e-01 -4.47991602e-02
7.48224974e-01 -4.42348510e-01 2.21763942e-02 -3.47513407e-01
-9.21068251e-01 -1.36910468e-01 1.59783348e-01 6.88823164e-01
-7.60147810e-01 -3.42540473e-01 4.87688363e-01 -1.69642675e+00
9.38179612e-01 -1.18144786e+00 2.64671654e-01 -1.29115805e-01
-5.07834703e-02 -3.17088157e-01 9.22422469e-01 -4.39075530e-01
-8.64652038e-01 -3.40619415e-01 -4.83706087e-01 2.60722265e-03
-7.89486170e-02 6.58778429e-01 1.99713349e-01 1.22281782e-01
8.44810963e-01 9.38155204e-02 -2.26368047e-02 -5.73267639e-01
3.93696517e-01 1.59656751e+00 1.24658756e-01 4.19521593e-02
8.77735138e-01 1.10841602e-01 -5.07124126e-01 -1.48314786e+00
-9.35022831e-01 -6.38499260e-01 -9.04781938e-01 -2.28545010e-01
8.99058759e-01 -6.57706380e-01 -4.77580547e-01 1.05172789e+00
-9.51172113e-01 -1.03124097e-01 4.00672406e-01 -7.07395598e-02
5.91408424e-02 5.16847730e-01 -1.07369637e+00 -6.97456956e-01
-4.66583729e-01 -5.36166072e-01 4.22742337e-01 -7.12909847e-02
-1.06115496e+00 -1.08354187e+00 4.32864159e-01 6.77219510e-01
6.51986122e-01 5.00002265e-01 1.13157213e+00 -1.52991712e+00
5.82807720e-01 -1.92104980e-01 -1.48256019e-01 6.15031481e-01
1.64873093e-01 5.27195573e-01 -7.29389966e-01 -1.89744458e-01
-3.21465313e-01 -1.00376070e+00 4.19915318e-01 -1.97638571e-01
1.09044528e+00 -6.96838498e-01 9.47641602e-05 2.98215508e-01
1.14786768e+00 6.69724494e-02 6.33508980e-01 5.79321265e-01
9.72050726e-01 5.80076516e-01 1.43709034e-01 6.52332842e-01
1.14099555e-01 4.35432076e-01 4.15430441e-02 7.96132386e-02
-9.00799967e-03 -3.38628262e-01 2.67308861e-01 9.60328102e-01
3.73373568e-01 -4.09272760e-01 -1.40068889e+00 4.91745234e-01
-1.44649386e+00 -1.33387828e+00 -1.47436395e-01 1.63037765e+00
1.15274882e+00 1.07316278e-01 4.29310471e-01 2.29332447e-01
8.49641562e-01 4.21952695e-01 -3.10670376e-01 -1.39576828e+00
-2.68768936e-01 1.27436876e-01 6.29238486e-01 6.88030243e-01
-1.29485655e+00 1.35329139e+00 6.27550507e+00 8.32676053e-01
-1.40290260e+00 3.12677264e-01 9.57834005e-01 -1.92580059e-01
-3.94069590e-03 -6.73928142e-01 -7.51115322e-01 9.94507074e-01
1.23256385e+00 -4.76522222e-02 3.36143166e-01 9.17746127e-01
4.50315177e-02 2.56659359e-01 -4.88014221e-01 8.47917736e-01
6.43778861e-01 -1.10500085e+00 8.92726555e-02 9.55910683e-02
7.72962809e-01 4.59634930e-01 2.71540314e-01 6.40098929e-01
7.52833307e-01 -1.24194562e+00 5.39369166e-01 -2.11073130e-01
3.59619498e-01 -1.04446733e+00 9.22260702e-01 3.12167376e-01
-1.28858134e-01 -2.59077489e-01 -1.95408240e-01 -4.12861168e-01
-8.86852667e-02 1.11823998e-01 -8.74020755e-01 -1.73127785e-01
6.51986659e-01 1.07826877e+00 -8.59210312e-01 4.87909496e-01
-3.67082730e-02 8.67799699e-01 -1.26167551e-01 -5.73620796e-01
6.77286386e-01 -1.63543612e-01 4.95128810e-01 1.74878597e+00
-2.39423260e-01 -1.32005259e-01 -6.87145144e-02 2.03458697e-01
-3.84259850e-01 4.14981484e-01 -1.07077646e+00 -5.70767760e-01
4.03246582e-01 1.22898889e+00 -5.55047691e-01 -3.53244543e-01
-4.17816460e-01 1.19785881e+00 9.32470202e-01 3.06132846e-02
-9.28213894e-01 -7.34210372e-01 5.55489659e-01 2.36174136e-01
-2.39349455e-01 -3.43499839e-01 -3.42636913e-01 -9.53858733e-01
-5.06143510e-01 -1.13768792e+00 4.83808577e-01 -8.47710013e-01
-1.67480195e+00 7.02304363e-01 -4.90318201e-02 -5.89808166e-01
-2.48440832e-01 -5.88769972e-01 -5.40246069e-01 8.36651444e-01
-8.15700293e-01 -1.46751332e+00 1.38188481e-01 5.04620552e-01
5.76920807e-01 -5.90557933e-01 4.95933414e-01 5.76464295e-01
-8.34365904e-01 7.80414462e-01 5.22815466e-01 6.33837163e-01
7.53207207e-01 -1.27344263e+00 2.64858246e-01 6.10797644e-01
-1.22123614e-01 6.87956929e-01 8.79381120e-01 -8.78108859e-01
-5.85340142e-01 -8.78713965e-01 1.31846869e+00 -6.80185616e-01
1.40188062e+00 -7.52536535e-01 -8.17005694e-01 9.24458981e-01
7.44000971e-01 -6.64593697e-01 1.20085025e+00 2.49802038e-01
-5.69365442e-01 2.91972548e-01 -1.31591880e+00 7.37136960e-01
8.98853660e-01 -8.36707115e-01 -8.25851083e-01 8.83750439e-01
3.33563417e-01 3.43151428e-02 -3.72350991e-01 -2.30417550e-01
2.54833192e-01 -7.56178796e-01 4.72814530e-01 -1.26127946e+00
1.09843004e+00 4.52098817e-01 9.56129655e-02 -1.49278736e+00
-5.41049182e-01 -8.50614250e-01 1.46827772e-01 1.59852374e+00
4.92994815e-01 -5.07679343e-01 8.30374956e-01 5.54214597e-01
-1.27782151e-01 -3.43773752e-01 -8.64745080e-01 -4.82964277e-01
6.55798018e-01 1.73730537e-01 -6.28633276e-02 1.78847706e+00
1.87875822e-01 7.41594911e-01 -1.00014246e+00 -6.11345232e-01
1.77468330e-01 -6.04251027e-01 6.57651544e-01 -1.24743760e+00
2.45054945e-01 -5.67575157e-01 -6.42709672e-01 -1.66892678e-01
6.15562916e-01 -1.05102479e+00 -5.01738548e-01 -1.18615150e+00
3.98721069e-01 -1.48948550e-01 5.64863086e-02 4.49848413e-01
3.77445072e-02 7.28548467e-01 1.76142305e-01 1.38045594e-01
-3.16478252e-01 3.63027215e-01 5.49207330e-01 -1.15827687e-01
6.40668906e-03 -5.01872659e-01 -7.72155404e-01 8.59083712e-01
9.84863043e-01 -4.48908210e-01 -1.22993685e-01 -2.92403519e-01
6.91984951e-01 -6.32115722e-01 -2.79040009e-01 -8.10364962e-01
-3.57883304e-01 3.39467488e-02 3.92408758e-01 -6.72529161e-01
3.55381072e-01 -5.27562022e-01 -2.42857441e-01 7.16889441e-01
-7.63435364e-01 4.75286603e-01 5.49708009e-02 5.62441424e-02
5.17906900e-03 -4.75574374e-01 1.11720419e+00 -2.25183249e-01
-5.56621373e-01 1.42149374e-01 -9.11590457e-01 4.30520386e-01
1.15053344e+00 -2.02866316e-01 -4.81557906e-01 -5.86285889e-01
-7.68928230e-01 -1.31382152e-01 4.14924443e-01 7.13595629e-01
1.13062240e-01 -1.14459074e+00 -1.20155656e+00 -7.04721361e-02
1.17329277e-01 -1.06542563e+00 5.01700155e-02 4.38357711e-01
-1.09263778e+00 1.47375718e-01 -7.38775671e-01 1.18860506e-01
-1.54260468e+00 4.24994349e-01 1.05533704e-01 -7.77755827e-02
-2.90758610e-01 1.02446425e+00 -3.10459375e-01 -7.58042991e-01
-3.76920141e-02 6.53901100e-01 -6.08357728e-01 6.45456731e-01
4.19736743e-01 5.65040231e-01 -1.47026926e-01 -1.37866831e+00
-3.39688122e-01 -1.44685403e-01 -7.29784966e-01 -3.94825488e-02
1.22590041e+00 -1.35229841e-01 -4.25607204e-01 7.04226196e-01
1.90036821e+00 2.84679264e-01 -1.59521565e-01 -4.85559832e-03
1.84544846e-01 -5.63423038e-01 2.67883167e-02 -1.04273283e+00
-6.23901486e-01 9.93622959e-01 3.08563352e-01 6.49733961e-01
3.56041104e-01 -1.73593298e-01 1.08186376e+00 4.16835099e-01
-2.37916976e-01 -1.52765000e+00 3.17185000e-02 1.17056227e+00
1.04486358e+00 -1.35809696e+00 1.23140350e-01 1.41263762e-02
-1.02102590e+00 7.25635707e-01 7.12523222e-01 -1.63265213e-01
8.46678734e-01 1.04985908e-01 2.70658851e-01 -1.17590167e-01
-2.58104324e-01 2.09571972e-01 -3.40415761e-02 5.75640678e-01
1.04289174e+00 2.19680369e-01 -7.01965392e-01 -7.93225989e-02
-8.39638054e-01 -5.49932659e-01 1.02711368e+00 8.46717834e-01
-2.70706117e-01 -8.38476717e-01 8.25050622e-02 8.22356343e-01
-1.28584123e+00 -4.98168588e-01 -1.24820995e+00 9.99130249e-01
-1.64587080e-01 8.13587666e-01 1.06430657e-01 -5.66422522e-01
-2.90977955e-01 3.28659147e-01 -1.27885759e-01 -8.15351427e-01
-1.31638086e+00 -4.64689046e-01 7.47942567e-01 9.78845507e-02
-1.36618271e-01 -6.66982710e-01 -7.53683746e-01 -1.09001446e+00
-4.16101469e-03 1.70867369e-02 8.02190304e-01 9.81936395e-01
1.33170992e-01 -2.35725254e-01 6.29313767e-01 -5.96098244e-01
-5.10590076e-01 -1.48286033e+00 -7.17009783e-01 1.17047513e+00
2.11377203e-01 -4.80368406e-01 -4.63678360e-01 -4.40352380e-01] | [8.825118064880371, 10.600881576538086] |
31c7ba51-0b42-4a1c-9f96-813f790b42f5 | self-training-via-metric-learning-for-source | 2212.04227 | null | https://arxiv.org/abs/2212.04227v1 | https://arxiv.org/pdf/2212.04227v1.pdf | Self-training via Metric Learning for Source-Free Domain Adaptation of Semantic Segmentation | Unsupervised source-free domain adaptation methods aim to train a model to be used in the target domain utilizing the pretrained source-domain model and unlabeled target-domain data, where the source data may not be accessible due to intellectual property or privacy issues. These methods frequently utilize self-training with pseudo-labeling thresholded by prediction confidence. In a source-free scenario, only supervision comes from target data, and thresholding limits the contribution of the self-training. In this study, we utilize self-training with a mean-teacher approach. The student network is trained with all predictions of the teacher network. Instead of thresholding the predictions, the gradients calculated from the pseudo-labels are weighted based on the reliability of the teacher's predictions. We propose a novel method that uses proxy-based metric learning to estimate reliability. We train a metric network on the encoder features of the teacher network. Since the teacher is updated with the moving average, the encoder feature space is slowly changing. Therefore, the metric network can be updated in training time, which enables end-to-end training. We also propose a metric-based online ClassMix method to augment the input of the student network where the patches to be mixed are decided based on the metric reliability. We evaluated our method in synthetic-to-real and cross-city scenarios. The benchmarks show that our method significantly outperforms the existing state-of-the-art methods. | ['Ugur Halici', 'Ibrahim Batuhan Akkaya'] | 2022-12-08 | null | null | null | null | ['source-free-domain-adaptation'] | ['computer-vision'] | [ 2.09261805e-01 1.58256337e-01 -4.49191630e-01 -7.08411038e-01
-1.09805167e+00 -3.82343888e-01 3.28987658e-01 -7.64343292e-02
-3.94519389e-01 9.30692673e-01 -2.41487294e-01 1.00090560e-02
-1.54710878e-02 -8.10046017e-01 -8.43309224e-01 -8.39178622e-01
2.26211950e-01 6.12574995e-01 3.74165326e-01 4.95654382e-02
-4.07855399e-02 3.82900573e-02 -1.41713178e+00 2.34060779e-01
1.34770775e+00 1.39349437e+00 9.63981077e-02 2.52277881e-01
-1.33708671e-01 7.14357316e-01 -8.64299893e-01 -2.44294807e-01
4.73913521e-01 -6.46228433e-01 -2.77124375e-01 2.57778317e-01
2.70355076e-01 -2.75694191e-01 -1.56400204e-01 1.04090035e+00
4.07107830e-01 2.89232403e-01 6.73456252e-01 -1.41549146e+00
-6.63642049e-01 4.00817275e-01 -5.49938798e-01 -2.31635034e-01
4.78729547e-04 -1.23015746e-01 7.05409884e-01 -8.53193462e-01
3.72783899e-01 8.72480989e-01 4.89299774e-01 5.15081048e-01
-1.23386061e+00 -1.07758701e+00 2.64531344e-01 3.36774468e-01
-1.44035637e+00 -4.89840418e-01 9.47316229e-01 -3.91344279e-01
3.64159226e-01 -1.92245141e-01 2.86336929e-01 9.65712905e-01
-2.04687610e-01 6.15339637e-01 9.87948775e-01 -5.23006201e-01
5.77773392e-01 7.13418126e-01 -1.30643710e-01 5.94673932e-01
-6.48182258e-02 2.55464107e-01 -6.57114565e-01 -2.98949659e-01
4.85230982e-01 1.85852125e-02 -7.33578429e-02 -7.11763024e-01
-9.59958375e-01 8.29309106e-01 3.54293197e-01 3.37625407e-02
-2.88586855e-01 -3.08317661e-01 2.05733508e-01 5.29821932e-01
8.50654721e-01 -5.89976227e-03 -7.76085734e-01 1.13317840e-01
-1.25379252e+00 -1.50899768e-01 7.93021441e-01 1.00002205e+00
1.03935766e+00 -4.28172722e-02 1.74494684e-02 1.19121695e+00
3.30998808e-01 4.19839889e-01 7.75845766e-01 -9.05893981e-01
5.67770898e-01 7.03460157e-01 6.46166727e-02 -4.88946795e-01
9.39963683e-02 -3.81374359e-01 -6.78822935e-01 2.70102292e-01
5.07303238e-01 -3.06665778e-01 -1.01343811e+00 1.77699935e+00
7.17631817e-01 5.53902268e-01 1.44786239e-01 7.51521885e-01
4.99880373e-01 5.60224354e-01 -8.70935172e-02 -3.38068634e-01
6.08635008e-01 -1.04392958e+00 -3.53850901e-01 -6.09723479e-02
6.01861596e-01 -5.17377317e-01 9.58178997e-01 4.56317902e-01
-6.30169630e-01 -6.67102575e-01 -1.28369689e+00 4.52667981e-01
-2.61759490e-01 3.33402693e-01 -1.16610162e-01 6.10070884e-01
-6.56117618e-01 8.53961349e-01 -7.03762233e-01 -1.23735145e-01
6.27096057e-01 1.97832167e-01 -3.10667813e-01 -2.34125510e-01
-1.18560350e+00 6.01480067e-01 3.79729420e-01 -4.41231757e-01
-1.12452602e+00 -7.01245964e-01 -6.58806562e-01 1.07942142e-04
3.81849140e-01 -2.27104157e-01 1.14773464e+00 -1.41302001e+00
-1.68152297e+00 5.22811413e-01 5.51606864e-02 -4.06216085e-01
6.24837458e-01 1.87100455e-01 -4.66519028e-01 -1.00767044e-02
2.50944495e-01 5.58595121e-01 1.10583711e+00 -1.13711321e+00
-9.38399792e-01 -2.17843026e-01 -2.77038932e-01 3.01910758e-01
-6.59247100e-01 -3.41826558e-01 -3.84127587e-01 -5.30092478e-01
4.97288071e-02 -8.05199802e-01 -2.38635838e-02 2.03611448e-01
-1.38691127e-01 -1.96858659e-01 9.35223758e-01 -5.57642698e-01
1.26165330e+00 -2.56292748e+00 -3.57826233e-01 4.07089829e-01
7.92762712e-02 1.37872085e-01 -3.24413121e-01 7.59089366e-02
-7.81717896e-02 -2.15267256e-01 -5.33818781e-01 -2.90375024e-01
-2.08965421e-01 2.03583002e-01 -2.28094846e-01 5.42978764e-01
1.27767801e-01 1.91334009e-01 -8.68909240e-01 -7.41804361e-01
3.34621370e-02 4.25224304e-01 -4.01371121e-01 5.95602214e-01
-1.67030677e-01 3.54081303e-01 -3.66855651e-01 5.49127162e-01
6.96828246e-01 -3.61102700e-01 9.24533382e-02 9.96455401e-02
2.25259766e-01 3.49709272e-01 -1.36444104e+00 1.66454625e+00
-6.09796643e-01 4.06062156e-01 -6.98177591e-02 -1.28675020e+00
1.18084431e+00 3.92571479e-01 6.52385354e-01 -6.05281472e-01
-1.53107509e-01 3.48801374e-01 -6.51597604e-02 -9.04158875e-02
-4.71325368e-02 1.50778100e-01 1.73278913e-01 5.68694949e-01
2.75647461e-01 -1.66412443e-02 1.89346075e-01 7.15581104e-02
1.11582983e+00 3.45445216e-01 1.12737134e-01 6.46626726e-02
4.52832639e-01 1.43666044e-01 8.67293656e-01 3.68353188e-01
-2.09277079e-01 6.28952026e-01 3.00735742e-01 -2.04432607e-01
-9.68906105e-01 -1.13863933e+00 -1.85217485e-01 1.28448367e+00
-2.55734678e-02 -9.32343956e-03 -8.93397748e-01 -1.33445454e+00
-4.05828953e-02 7.86186516e-01 -5.49590349e-01 -4.17762280e-01
-1.74139142e-01 -4.11919981e-01 1.79959208e-01 2.20019862e-01
6.55197322e-01 -7.29623437e-01 -2.59771168e-01 5.10097384e-01
-7.25371391e-02 -8.36590588e-01 -5.56729436e-01 3.35968763e-01
-9.51975882e-01 -1.09806180e+00 -7.65491843e-01 -7.71003127e-01
1.00078654e+00 1.25736311e-01 8.28583062e-01 -3.05088013e-01
2.21092463e-01 1.16087593e-01 -2.43596047e-01 -2.19446793e-01
-5.46856105e-01 1.54848352e-01 1.56562269e-01 2.29114339e-01
5.34006655e-01 -6.82277322e-01 -4.63136792e-01 5.91495216e-01
-5.96890211e-01 -1.73636138e-01 5.61552286e-01 1.14233446e+00
8.30152988e-01 1.99887499e-01 8.84709597e-01 -1.02774191e+00
3.69493335e-01 -8.06754589e-01 -7.82799780e-01 4.66795713e-01
-1.07434809e+00 1.17675796e-01 8.58388364e-01 -9.69841003e-01
-1.19662368e+00 2.94489652e-01 2.63259739e-01 -6.99903309e-01
-9.46567655e-02 2.90000051e-01 -4.01600271e-01 -1.85130127e-02
1.03658247e+00 4.93046902e-02 -1.09798647e-03 -4.30504590e-01
1.90727428e-01 1.03394771e+00 3.20969492e-01 -5.78943431e-01
8.26005459e-01 1.22213013e-01 -4.48172212e-01 -4.11022186e-01
-9.92059708e-01 -2.59884834e-01 -5.20425200e-01 -7.40242153e-02
2.40866229e-01 -1.08830225e+00 -6.48117959e-02 3.37129861e-01
-8.26394439e-01 -5.58886886e-01 -6.41401231e-01 5.66877186e-01
-4.63302463e-01 6.42021522e-02 -2.05003187e-01 -6.72454119e-01
-1.61903232e-01 -9.77906644e-01 6.83247328e-01 3.76339287e-01
7.47983158e-02 -9.00798082e-01 1.83310419e-01 1.92816138e-01
2.90806293e-01 1.99761782e-02 4.92111474e-01 -1.13554442e+00
-2.71061003e-01 -2.53535330e-01 -2.13272750e-01 8.68627071e-01
4.05773342e-01 -2.50641517e-02 -1.14223266e+00 -1.58943981e-01
1.26448795e-01 -4.25513506e-01 5.82671106e-01 2.17805073e-01
1.20222878e+00 -5.31081140e-01 -2.50093699e-01 5.09349823e-01
1.22769320e+00 1.82409331e-01 2.53282964e-01 2.42598072e-01
5.00494957e-01 6.70573294e-01 9.32607949e-01 6.02404892e-01
4.04169053e-01 4.15399402e-01 1.69323117e-01 5.17843217e-02
-1.42065370e-02 -4.84272510e-01 5.53170323e-01 5.76461613e-01
5.35989285e-01 -4.01230529e-02 -7.77151823e-01 6.69601679e-01
-1.76296878e+00 -7.16504991e-01 3.72580498e-01 2.60398912e+00
1.25979245e+00 1.85177952e-01 1.09763622e-01 3.45635638e-02
8.73978555e-01 -2.23798260e-01 -1.04074979e+00 -1.34171978e-01
1.56846479e-01 3.88257690e-02 5.03356695e-01 2.76668161e-01
-1.10377872e+00 7.12582886e-01 5.60220814e+00 1.07663703e+00
-1.20297158e+00 3.80994588e-01 9.13000345e-01 -1.90254271e-01
-1.27649665e-01 -8.22509900e-02 -7.55132318e-01 8.19433153e-01
1.22387803e+00 -2.05433324e-01 5.11619866e-01 1.30464053e+00
-7.08949380e-03 -2.35848829e-01 -1.13955438e+00 7.64933109e-01
4.99759763e-02 -8.48379791e-01 -4.95450228e-01 3.72426324e-02
9.83454585e-01 1.99802607e-01 4.25584987e-02 4.79913682e-01
6.35886252e-01 -5.22438288e-01 5.54167867e-01 3.79126638e-01
1.08579826e+00 -8.02606344e-01 6.39259338e-01 6.38111413e-01
-1.01939583e+00 -1.06686488e-01 -4.42121178e-01 3.16593885e-01
-2.49133781e-01 9.91353750e-01 -9.61179435e-01 3.05641949e-01
6.83005393e-01 7.31174350e-01 -3.90387863e-01 9.37786460e-01
-3.31116617e-01 9.85954881e-01 -5.65033674e-01 1.98481858e-01
-1.20998941e-01 -3.18593085e-01 1.67868748e-01 8.99740815e-01
6.30921125e-01 -1.67056218e-01 3.81578833e-01 7.44326711e-01
-2.32563347e-01 2.33338356e-01 -4.94657725e-01 5.34589291e-02
8.55027378e-01 1.10445631e+00 -2.90865004e-01 -4.97609258e-01
-5.82300007e-01 9.13247883e-01 4.92601037e-01 3.71140540e-01
-9.32004929e-01 -5.09067535e-01 4.99571383e-01 1.66186392e-01
2.98215240e-01 1.81892380e-01 -3.10408115e-01 -9.58356559e-01
1.35073990e-01 -7.90635526e-01 5.04041374e-01 -5.20130217e-01
-1.56863785e+00 5.57001829e-01 1.48688974e-02 -1.72621548e+00
-4.36137199e-01 -1.63660109e-01 -5.28541803e-01 9.05409515e-01
-1.47359478e+00 -7.51879215e-01 -2.43966371e-01 6.86956048e-01
2.89649606e-01 -5.13937593e-01 7.11433351e-01 3.97260308e-01
-6.07967138e-01 9.83695090e-01 4.69820768e-01 2.67239481e-01
1.23287022e+00 -1.05053139e+00 2.92986687e-02 6.33261263e-01
2.09386617e-01 1.52733147e-01 4.26588476e-01 -6.00113809e-01
-6.20518565e-01 -1.34386086e+00 7.24963248e-01 -1.07122079e-01
5.46776533e-01 -2.57427901e-01 -1.09963036e+00 3.15216511e-01
-1.18399434e-01 4.23281133e-01 8.88884783e-01 -1.19357757e-01
-5.95804513e-01 -7.29885340e-01 -1.60668600e+00 1.47706866e-01
6.71588242e-01 -3.04191440e-01 -3.59979063e-01 3.14960659e-01
7.69698977e-01 -2.81058282e-01 -8.35881889e-01 2.43262261e-01
3.96327853e-01 -7.60698855e-01 4.59772408e-01 -2.66688973e-01
3.15000772e-01 -4.05951381e-01 -1.12614915e-01 -1.62883377e+00
-5.54325543e-02 -1.84092999e-01 -3.48436385e-01 1.49889791e+00
6.46920562e-01 -7.02760160e-01 9.93846953e-01 8.09108853e-01
4.26752195e-02 -6.09447420e-01 -1.18286860e+00 -8.88153076e-01
-8.30652490e-02 -2.38806427e-01 7.16026902e-01 1.00788832e+00
-1.48455173e-01 2.00070053e-01 -1.48797244e-01 5.48927076e-02
8.25432122e-01 -5.27459718e-02 7.31688678e-01 -1.30334258e+00
-3.31539661e-01 -1.44196572e-02 -9.70055237e-02 -8.68181527e-01
2.26758018e-01 -6.64861202e-01 2.25485742e-01 -1.15332234e+00
2.62733847e-02 -7.75056422e-01 -5.80725789e-01 6.53744400e-01
-1.44704312e-01 5.72221875e-02 -1.60351291e-01 2.25400418e-01
-6.44243956e-01 6.73304915e-01 9.48388636e-01 -2.79648513e-01
-5.21962464e-01 2.31840178e-01 -5.08959532e-01 4.99890596e-01
9.31280375e-01 -9.73926485e-01 -5.88211656e-01 -1.06333286e-01
-1.84180021e-01 -9.66273695e-02 -5.62806102e-03 -1.20619714e+00
3.07389230e-01 -4.49395269e-01 4.78251129e-01 -3.54305357e-01
1.61962181e-01 -1.16157126e+00 -7.24681616e-02 7.25613311e-02
-4.38892365e-01 -4.90739405e-01 -1.53729707e-01 6.66185200e-01
-2.81333804e-01 -2.61952519e-01 1.06685317e+00 2.10027128e-01
-3.32213610e-01 4.97924268e-01 4.55457121e-02 1.96867824e-01
1.07915413e+00 -1.59439191e-01 -1.94095805e-01 -4.32531357e-01
-4.57912594e-01 3.10732692e-01 4.89767253e-01 1.91384748e-01
6.77150190e-01 -1.45373070e+00 -7.38764346e-01 3.61584574e-01
3.79587114e-01 2.65583277e-01 1.03512876e-01 5.07942617e-01
9.59817246e-02 -2.60689706e-01 -3.66464369e-02 -6.65970206e-01
-8.82951498e-01 7.12444663e-01 3.22580546e-01 -2.82268524e-01
-6.41984567e-02 7.00196505e-01 2.02525124e-01 -7.34927535e-01
4.90995198e-01 7.87338838e-02 -1.10670298e-01 2.14924272e-02
4.57432330e-01 3.65176648e-01 4.73508649e-02 -3.37612182e-01
-7.37352520e-02 1.96448356e-01 -1.65014133e-01 -3.30629200e-01
1.33506370e+00 -1.09754227e-01 2.41068184e-01 5.31969965e-01
1.28952205e+00 -1.62506416e-01 -1.66309214e+00 -8.00941169e-01
-1.14333153e-01 -4.37160969e-01 1.05379811e-02 -9.26011443e-01
-1.30135548e+00 7.32199192e-01 8.96593273e-01 -1.62577644e-01
1.34064317e+00 -2.12196663e-01 6.50560200e-01 2.99582720e-01
3.10621768e-01 -1.74009740e+00 1.38386965e-01 3.06991994e-01
4.99392629e-01 -1.59423363e+00 2.32077520e-02 4.71381471e-02
-7.69500434e-01 8.86633813e-01 9.22166348e-01 -6.98263720e-02
9.45224643e-01 1.70572996e-01 1.94892645e-01 4.74230945e-01
-8.54097009e-01 -5.09910434e-02 3.28869492e-01 7.73780346e-01
1.42725587e-01 6.47294000e-02 -4.39626165e-02 7.53055036e-01
1.98897868e-02 8.43098462e-02 1.82285011e-01 7.55046010e-01
-4.36913311e-01 -1.32731652e+00 -4.23752099e-01 6.21606052e-01
-7.23992959e-02 1.60024632e-02 -2.62094021e-01 3.21965009e-01
3.31185400e-01 1.15329647e+00 3.31091974e-03 -5.82989693e-01
2.63503492e-01 2.40146697e-01 2.84442361e-02 -7.22953022e-01
-3.38052243e-01 5.81388026e-02 -1.32856742e-01 -4.28817242e-01
-2.21961826e-01 -5.87536454e-01 -1.17575514e+00 -1.32849261e-01
-5.80548108e-01 4.18532401e-01 8.59290242e-01 7.82743335e-01
5.04412055e-01 3.61528724e-01 1.28543746e+00 -4.27346826e-01
-9.08907235e-01 -1.12599850e+00 -5.57610869e-01 2.00468749e-01
2.40992531e-01 -6.92354858e-01 -5.37130237e-01 1.41021371e-01] | [10.368896484375, 3.0999162197113037] |
68d72a48-d894-4139-a5d3-0174a2efe85e | rgb-depth-fusion-gan-for-indoor-depth | 2203.10856 | null | https://arxiv.org/abs/2203.10856v1 | https://arxiv.org/pdf/2203.10856v1.pdf | RGB-Depth Fusion GAN for Indoor Depth Completion | The raw depth image captured by the indoor depth sensor usually has an extensive range of missing depth values due to inherent limitations such as the inability to perceive transparent objects and limited distance range. The incomplete depth map burdens many downstream vision tasks, and a rising number of depth completion methods have been proposed to alleviate this issue. While most existing methods can generate accurate dense depth maps from sparse and uniformly sampled depth maps, they are not suitable for complementing the large contiguous regions of missing depth values, which is common and critical. In this paper, we design a novel two-branch end-to-end fusion network, which takes a pair of RGB and incomplete depth images as input to predict a dense and completed depth map. The first branch employs an encoder-decoder structure to regress the local dense depth values from the raw depth map, with the help of local guidance information extracted from the RGB image. In the other branch, we propose an RGB-depth fusion GAN to transfer the RGB image to the fine-grained textured depth map. We adopt adaptive fusion modules named W-AdaIN to propagate the features across the two branches, and we append a confidence fusion head to fuse the two outputs of the branches for the final depth map. Extensive experiments on NYU-Depth V2 and SUN RGB-D demonstrate that our proposed method clearly improves the depth completion performance, especially in a more realistic setting of indoor environments with the help of the pseudo depth map. | ['Jian Tang', 'Feifei Feng', 'Mengshi Qi', 'XIUQUAN QIAO', 'Zhiyuan Xu', 'Zhengping Che', 'Mingyuan Wang', 'Haowen Wang'] | 2022-03-21 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Wang_RGB-Depth_Fusion_GAN_for_Indoor_Depth_Completion_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Wang_RGB-Depth_Fusion_GAN_for_Indoor_Depth_Completion_CVPR_2022_paper.pdf | cvpr-2022-1 | ['transparent-objects', 'depth-completion'] | ['computer-vision', 'computer-vision'] | [ 4.97639686e-01 1.02305338e-01 8.37238505e-02 -7.36911595e-01
-7.20265210e-01 -1.67903587e-01 2.60503262e-01 -3.20923209e-01
-2.87383884e-01 7.92353153e-01 2.93077886e-01 -1.62417646e-02
2.06765905e-01 -1.20786154e+00 -6.12806261e-01 -7.79336393e-01
4.96158332e-01 1.23552665e-01 3.48466843e-01 -1.20861225e-01
9.98863354e-02 7.27708861e-02 -1.72195745e+00 1.99639097e-01
1.10242510e+00 1.59036422e+00 6.83986127e-01 6.34431899e-01
-3.74268115e-01 7.65423417e-01 -2.86652297e-01 -7.58765191e-02
5.25086820e-01 -3.99232268e-01 -3.46713305e-01 1.90577939e-01
2.79727519e-01 -1.07393324e+00 -6.14560485e-01 9.69537795e-01
5.25702655e-01 1.68849260e-03 1.41794413e-01 -1.05293584e+00
-2.28704900e-01 -1.74306147e-03 -7.18630254e-01 -1.33736134e-01
6.02491856e-01 2.03915909e-01 4.17293966e-01 -6.78220153e-01
3.04567575e-01 1.05528677e+00 4.21673566e-01 6.08094513e-01
-5.39767563e-01 -6.10006452e-01 2.74651974e-01 -1.15918837e-01
-1.07045627e+00 -2.75559306e-01 8.35520089e-01 -1.59631953e-01
7.48792171e-01 -1.19426884e-01 8.87667298e-01 9.85393703e-01
8.97218138e-02 5.52821219e-01 1.17773736e+00 -2.24831879e-01
4.56233710e-01 -3.73185992e-01 -4.00682241e-01 7.61278212e-01
1.66988000e-01 4.20826495e-01 -5.81675291e-01 2.51438707e-01
1.12062454e+00 3.98150325e-01 -6.25073135e-01 -2.16771707e-01
-1.10286844e+00 5.55028498e-01 9.44223821e-01 -1.08753711e-01
-3.91130745e-01 2.16492549e-01 -2.16531649e-01 5.40541559e-02
4.53141779e-01 -2.88328767e-01 -3.88857514e-01 -2.17509508e-01
-6.98991418e-01 7.64481500e-02 3.16205651e-01 9.42038536e-01
1.39943552e+00 -1.70606658e-01 2.84556657e-01 4.36752111e-01
6.28776193e-01 6.85080647e-01 4.85312164e-01 -1.05166399e+00
9.73495245e-01 6.87741935e-01 1.08392648e-01 -5.92369795e-01
-2.31276497e-01 -5.12427613e-02 -1.03591454e+00 5.46391129e-01
3.30506623e-01 -1.44813091e-01 -1.33946919e+00 1.44623661e+00
5.07868290e-01 6.42169267e-02 2.08471164e-01 1.27428305e+00
9.29022491e-01 7.02451825e-01 -3.42568278e-01 3.42253372e-02
7.78321028e-01 -8.68964732e-01 -6.03092253e-01 -7.11260319e-01
2.55813956e-01 -4.16385025e-01 9.59926546e-01 4.88844842e-01
-9.91128802e-01 -5.66691995e-01 -1.31664491e+00 -6.37170792e-01
-8.48778337e-02 -7.89878815e-02 8.98244381e-01 4.12523150e-01
-1.08492899e+00 2.93425798e-01 -1.04826379e+00 -2.96678282e-02
3.99941385e-01 2.28150249e-01 -4.61887896e-01 -8.41239274e-01
-9.91832912e-01 4.20441717e-01 3.41714650e-01 4.64645445e-01
-9.07855093e-01 -4.81936306e-01 -1.18574023e+00 -3.02534044e-01
1.22763529e-01 -9.95254099e-01 1.10852730e+00 -6.82885706e-01
-1.60196280e+00 4.85891223e-01 -3.22703570e-01 1.08853243e-01
6.26619935e-01 -2.00608477e-01 1.18565939e-01 1.42809600e-01
2.22337484e-01 9.79120135e-01 7.00894058e-01 -1.25448108e+00
-1.10249937e+00 -7.37907112e-01 2.91304976e-01 5.68547904e-01
-4.37889621e-02 -8.53928089e-01 -6.92732096e-01 -1.64374128e-01
8.95056129e-01 -3.61154646e-01 -3.83210123e-01 3.09436202e-01
-3.62297952e-01 3.99564087e-01 6.87856615e-01 -4.76971358e-01
7.94036150e-01 -2.04222679e+00 3.84025782e-01 8.49217400e-02
3.66974860e-01 -3.69133264e-01 1.92600414e-01 4.36906181e-02
3.91305894e-01 -2.79687405e-01 -5.01549065e-01 -8.34247768e-01
-3.50503415e-01 5.30820131e-01 -2.68792659e-01 2.99883693e-01
-6.08102866e-02 7.31597245e-01 -1.07840085e+00 -2.42013648e-01
5.63557625e-01 7.16397941e-01 -4.63325560e-01 5.50131917e-01
-2.97714829e-01 8.20052385e-01 -5.47299325e-01 9.62632298e-01
9.72384393e-01 -7.58739114e-02 -1.73228621e-01 -1.19986467e-01
-1.51238322e-01 2.34511435e-01 -1.00766468e+00 2.50054669e+00
-6.28019154e-01 3.79441351e-01 2.02343479e-01 -3.05533290e-01
1.01298380e+00 1.85896885e-02 3.41315061e-01 -9.93584812e-01
1.16615660e-01 3.12098175e-01 -4.12216187e-01 -3.54126722e-01
5.76017082e-01 -3.38733345e-01 -1.23272061e-01 2.31565133e-01
-1.35548130e-01 -7.50673115e-01 -3.86515200e-01 9.65896919e-02
1.24728608e+00 4.67111975e-01 1.15045672e-02 6.13753736e-01
3.64014626e-01 -2.24395260e-01 8.07114124e-01 3.69648099e-01
-1.65215321e-02 1.16057491e+00 1.48711637e-01 -4.05708194e-01
-9.52364385e-01 -1.33680677e+00 1.04269370e-01 2.83705741e-01
6.85410440e-01 -1.51273459e-01 -4.92966324e-01 -4.79540169e-01
-1.80833980e-01 1.40973374e-01 -6.12124860e-01 -1.29365548e-01
-2.75283068e-01 -5.06800592e-01 1.08043119e-01 6.39812946e-01
1.26943302e+00 -8.29647899e-01 -9.24866676e-01 1.71852246e-01
-4.08396900e-01 -1.38442361e+00 -8.07553753e-02 3.64623219e-01
-1.09662247e+00 -9.01968122e-01 -7.43133664e-01 -5.86335778e-01
6.66415274e-01 4.22287047e-01 8.78918588e-01 -4.77279909e-02
-1.29412087e-02 1.89389326e-02 -4.28998768e-01 -2.41714165e-01
1.58072621e-01 -1.22025140e-01 -2.82380342e-01 -1.09254360e-01
1.63083941e-01 -7.13608742e-01 -9.88424301e-01 2.27201700e-01
-1.00431645e+00 5.82745314e-01 6.86524868e-01 7.42147982e-01
8.34276259e-01 2.32441813e-01 1.31137013e-01 -3.66741866e-01
-1.26303127e-02 -3.29754442e-01 -7.19829321e-01 -2.19640344e-01
-3.15580040e-01 -5.19883111e-02 3.81846309e-01 6.53065443e-02
-1.25796759e+00 4.12107855e-01 -4.38107252e-01 -5.43868124e-01
-1.57284856e-01 2.56964266e-01 -7.05901682e-01 -1.40432641e-01
2.35644713e-01 2.19532117e-01 -6.19034506e-02 -3.22482169e-01
1.97867170e-01 6.53143406e-01 7.40839481e-01 -3.55575144e-01
6.80950224e-01 8.10654521e-01 -2.75060418e-03 -4.99254107e-01
-7.95265555e-01 -3.66001248e-01 -4.74858314e-01 -1.16958432e-01
9.78531182e-01 -1.31766367e+00 -5.25178313e-01 8.91305387e-01
-1.05939496e+00 -7.30734408e-01 -2.66776472e-01 6.10641420e-01
-5.14425099e-01 4.21379626e-01 -6.11914277e-01 -6.72552705e-01
-2.23876685e-01 -1.26709378e+00 1.46073365e+00 5.23119926e-01
4.90435451e-01 -7.50778079e-01 -8.44794735e-02 4.62248772e-01
2.08933830e-01 4.23975736e-01 5.30644178e-01 7.20241547e-01
-1.06586444e+00 -1.31064095e-02 -3.91356587e-01 3.80530953e-01
4.20528531e-01 -4.22326118e-01 -1.37796605e+00 -2.65393890e-02
2.19010234e-01 -4.62826550e-01 1.04362714e+00 4.14662093e-01
1.07272804e+00 1.75884724e-01 -1.63202628e-01 1.24627578e+00
1.73350036e+00 1.37201279e-01 9.86134231e-01 2.70361811e-01
1.00059104e+00 4.55915898e-01 7.61390448e-01 5.49747825e-01
7.12262392e-01 3.72438818e-01 9.60913658e-01 -1.98433310e-01
-2.26834387e-01 -5.28719008e-01 2.94672430e-01 6.74396336e-01
-7.02604204e-02 -2.97415882e-01 -7.69589424e-01 3.17501277e-01
-1.67239690e+00 -5.17765701e-01 1.33836031e-01 2.15067410e+00
7.32328534e-01 2.66981393e-01 -4.38835710e-01 3.13549936e-01
2.73376137e-01 1.44305855e-01 -7.50055075e-01 1.37546018e-01
-2.36958876e-01 2.48545557e-01 4.69556570e-01 8.10530066e-01
-6.49146259e-01 6.96457326e-01 5.23192453e+00 4.61813301e-01
-1.16145432e+00 -4.39613089e-02 6.06275797e-01 -3.23842406e-01
-6.90076530e-01 8.16429313e-03 -5.89873612e-01 5.46730101e-01
4.41365749e-01 4.23233449e-01 3.08444560e-01 7.59922087e-01
1.14223391e-01 -8.98975730e-01 -1.02446020e+00 1.48088849e+00
-1.23410679e-01 -1.01405632e+00 -2.02910826e-01 2.36177236e-01
9.30616379e-01 1.76804692e-01 -1.05804235e-01 -1.46941012e-02
2.88114011e-01 -9.54596102e-01 7.10441947e-01 5.26482224e-01
1.13949120e+00 -6.23607635e-01 8.48634601e-01 4.99207288e-01
-1.20884287e+00 -2.91192662e-02 -5.73192775e-01 -4.90005851e-01
4.09849584e-01 1.05855691e+00 -5.20498097e-01 8.14564645e-01
8.76710653e-01 8.69574964e-01 -3.33927512e-01 9.59001958e-01
-6.52074337e-01 -7.05430955e-02 -4.83050197e-01 3.75433505e-01
1.62387669e-01 -2.73298085e-01 -5.96310943e-02 3.01004291e-01
6.34522974e-01 3.46788645e-01 -5.29712737e-02 8.07564676e-01
5.00408523e-02 -5.95136702e-01 -6.45331264e-01 4.78056401e-01
3.29378039e-01 1.08244085e+00 -5.21423578e-01 -2.58988887e-01
-5.15232325e-01 1.30988228e+00 2.87482500e-01 4.78859305e-01
-6.30195737e-01 -3.69519264e-01 8.07650745e-01 3.33507881e-02
1.35857165e-01 -3.34092587e-01 -7.42364764e-01 -1.30300176e+00
3.60741615e-01 -2.88444877e-01 2.22479366e-02 -1.16891479e+00
-9.61256862e-01 6.85253561e-01 -2.73258299e-01 -1.23136199e+00
-2.93030888e-01 -5.01734793e-01 -3.67234021e-01 1.08181989e+00
-1.89349151e+00 -9.55544055e-01 -1.35965872e+00 7.76087403e-01
4.03669447e-01 4.29742396e-01 5.45761466e-01 4.58249539e-01
-2.95618176e-01 1.91439897e-01 -2.90894896e-01 -5.11801839e-02
3.98367375e-01 -1.13598394e+00 3.43192488e-01 9.15355742e-01
-4.01348442e-01 6.05356134e-02 3.30454469e-01 -7.89409161e-01
-1.30128169e+00 -9.90649104e-01 4.17291790e-01 -2.00828239e-01
-1.43656477e-01 -4.40766364e-01 -6.04615152e-01 5.07131875e-01
-2.78878003e-01 3.69641244e-01 8.33470300e-02 -5.30797124e-01
-1.39270887e-01 -3.64951551e-01 -1.34088421e+00 2.21777305e-01
1.27762723e+00 -5.10002434e-01 -2.61821538e-01 -5.59925064e-02
9.37885106e-01 -8.91873181e-01 -6.57504678e-01 5.82292199e-01
5.63571215e-01 -1.57481718e+00 9.23285484e-01 5.58787584e-01
8.91514063e-01 -5.99827766e-01 -6.38868988e-01 -1.16179872e+00
1.43770292e-01 -1.68591812e-01 -1.27626389e-01 9.70464826e-01
1.64708458e-02 -5.53740203e-01 1.23535419e+00 6.13541663e-01
-5.20681024e-01 -9.10577655e-01 -9.87064123e-01 -1.20420545e-01
-3.26533705e-01 -6.39813125e-01 8.23370636e-01 4.43686277e-01
-3.66533011e-01 1.66731507e-01 -2.27196112e-01 2.26443097e-01
8.01805794e-01 2.43157893e-01 9.18128133e-01 -9.10876334e-01
-3.62304807e-01 -5.61639890e-02 -6.08428121e-01 -1.79200768e+00
-3.50761086e-01 -3.53945851e-01 3.64249706e-01 -2.03348327e+00
-2.00380161e-01 -8.27919245e-01 1.76118791e-01 3.89891624e-01
-2.03068048e-01 4.19686347e-01 -2.00497329e-01 -8.46681558e-03
-2.89649814e-01 9.75320578e-01 1.51915646e+00 -2.93309353e-02
-1.96995631e-01 -1.37363955e-01 -5.24154484e-01 7.14681625e-01
5.52103162e-01 -1.02166072e-01 -7.81071782e-01 -9.52703416e-01
3.34325612e-01 4.11111832e-01 3.28655690e-01 -1.40024686e+00
2.71956772e-01 -1.55679360e-01 8.93575609e-01 -8.13912511e-01
8.35492611e-01 -8.56883585e-01 1.59952156e-02 1.48958400e-01
3.18872362e-01 -1.67698801e-01 -1.22290313e-01 6.31367922e-01
-4.38132674e-01 2.97604173e-01 5.14153481e-01 -3.14149767e-01
-1.05348623e+00 7.10200965e-01 2.34368056e-01 -2.57397145e-01
9.70557213e-01 -6.08148456e-01 -2.46834323e-01 -5.35756052e-01
-3.04198653e-01 2.90637463e-01 9.30120826e-01 2.02232674e-01
1.22104490e+00 -1.28695822e+00 -3.59794557e-01 8.37109566e-01
1.41921192e-01 1.19498515e+00 5.46919703e-01 5.40163577e-01
-8.33878517e-01 1.22819468e-01 -3.06880683e-01 -8.58659148e-01
-6.12148523e-01 2.51546681e-01 4.02950764e-01 -8.29523429e-02
-8.01280499e-01 1.03482997e+00 5.41545391e-01 -2.87858725e-01
2.35584572e-01 -8.23629975e-01 3.48947316e-01 -3.78556252e-01
5.79540491e-01 1.38770625e-01 2.22622827e-01 -3.60655040e-01
-1.37184173e-01 8.38482797e-01 2.89613724e-01 -3.49208355e-01
1.16751242e+00 -6.73879445e-01 6.10966608e-02 3.46637607e-01
1.17147160e+00 -1.53710067e-01 -1.96208405e+00 -1.47829533e-01
-7.38926828e-01 -9.20018852e-01 3.46336842e-01 -6.58692241e-01
-1.40614164e+00 1.06970406e+00 5.47580123e-01 -4.02527004e-01
1.59813690e+00 -8.87620673e-02 1.10255766e+00 -5.37639782e-02
9.88130152e-01 -6.44360244e-01 7.68276006e-02 4.39220488e-01
4.68538165e-01 -1.46517110e+00 2.21650433e-02 -5.61926603e-01
-3.59502673e-01 1.09730744e+00 9.55587506e-01 1.15803264e-01
4.15227920e-01 4.10918891e-01 1.91222981e-01 3.04377079e-02
-5.14452994e-01 -1.92214340e-01 -2.26873711e-01 8.23583424e-01
-1.17092701e-02 -1.85680419e-01 1.95894927e-01 3.45640004e-01
-3.25790197e-01 1.97271585e-01 5.41677654e-01 1.09412587e+00
-4.92814362e-01 -9.64467287e-01 -4.56581801e-01 3.07642788e-01
-3.78242247e-02 -1.25970105e-02 3.33010890e-02 4.82461125e-01
4.94186729e-01 1.01596427e+00 2.09375873e-01 -7.60229647e-01
1.60440534e-01 -4.36126024e-01 6.56737208e-01 -5.86601138e-01
-4.01341505e-02 -1.21004544e-02 -2.57361859e-01 -9.52024579e-01
-4.28264558e-01 -2.60938346e-01 -1.45403719e+00 -3.12946111e-01
-1.32328905e-02 -2.68488050e-01 8.43558788e-01 8.82128477e-01
6.15435056e-02 5.40929854e-01 5.64325392e-01 -1.27686846e+00
1.02250271e-01 -8.85043085e-01 -6.77539408e-01 9.39444676e-02
8.38518322e-01 -6.85493529e-01 -4.95592564e-01 -2.31526285e-01] | [8.933799743652344, -2.537264347076416] |
3c401135-f973-40b4-9524-f06f31658fab | hop-history-enhanced-and-order-aware-pre | null | null | https://ieeexplore.ieee.org/abstract/document/10006384 | https://ieeexplore.ieee.org/abstract/document/10006384 | HOP+: History-enhanced and Order-aware Pre-training for Vision-and-Language Navigation | Recent works attempt to employ pre-training in Vision-and-Language Navigation (VLN). However, these methods neglect the importance of historical contexts or ignore predicting future actions during pre-training, limiting the learning of visual-textual correspondence and the capability of decision-making. To address these problems, we present a history-enhanced and order-aware pre-training with the complementing fine-tuning paradigm (HOP+) for VLN. Specifically, besides the common Masked Language Modeling (MLM) and Trajectory-Instruction Matching (TIM) tasks, we design three novel VLN-specific proxy tasks: Action Prediction with History (APH) task, Trajectory Order Modeling (TOM) task and Group Order Modeling (GOM) task. APH task takes into account the visual perception trajectory to enhance the learning of historical knowledge as well as action prediction. The two temporal visual-textual alignment tasks, TOM and GOM further improve the agent’s ability to order reasoning. Moreover, we design a memory network to address the representation inconsistency of history context between the pre-training and the fine-tuning stages. The memory network effectively selects and summarizes historical information for action prediction during fine-tuning, without costing huge extra computation consumption for downstream VLN tasks. HOP+ achieves new state-of-the-art performance on four downstream VLN tasks (R2R, REVERIE, RxR, and NDH), which demonstrates the effectiveness of our proposed method. | ['and Qi Wu ̊', 'Peng Wang', 'Zheng Yu', 'Yicong Hong', 'Yuankai Qi', 'Yanyuan Qiao'] | 2023-03-20 | null | null | null | ieee-transactions-on-pattern-analysis-and-23 | ['vision-and-language-navigation'] | ['robots'] | [-3.20475027e-02 -3.01488072e-01 -3.56823891e-01 -3.42727661e-01
-2.70235002e-01 -3.27648908e-01 1.10057592e+00 -7.24514648e-02
-8.32697153e-01 4.47104394e-01 4.83572900e-01 -8.05665910e-01
-1.25762582e-01 -7.44838119e-01 -7.87745953e-01 -2.59669542e-01
-1.11819170e-01 4.83749688e-01 6.87650561e-01 -4.53295469e-01
5.44720709e-01 5.67518115e-01 -1.84083796e+00 6.87986016e-01
9.93381619e-01 1.02197516e+00 1.03417563e+00 5.65336108e-01
-2.92674661e-01 1.37220168e+00 -2.37981290e-01 1.42773269e-02
2.71667272e-01 -1.92565128e-01 -8.55853140e-01 -3.45635533e-01
3.80125970e-01 -6.92553997e-01 -7.06691861e-01 5.93557060e-01
4.59849745e-01 6.56611204e-01 5.02860129e-01 -1.17462528e+00
-7.71760702e-01 5.31077862e-01 -3.52436095e-01 3.03302437e-01
3.05385709e-01 5.77888012e-01 7.10717440e-01 -9.60758388e-01
7.65493095e-01 1.48247349e+00 5.88530898e-01 6.99853122e-01
-7.45164275e-01 -4.87284094e-01 6.80147946e-01 1.03598988e+00
-1.02542365e+00 -4.58959728e-01 4.72896844e-01 -6.42328203e-01
1.47789991e+00 -8.89574066e-02 7.68838823e-01 1.28267193e+00
3.24090987e-01 1.13994467e+00 1.04790664e+00 -1.96654484e-01
8.30714554e-02 -3.54660004e-01 1.25317136e-03 9.17776585e-01
-2.65307277e-01 8.54566813e-01 -9.12997246e-01 4.13338512e-01
9.31776643e-01 -6.13605343e-02 -7.40185454e-02 -2.84232616e-01
-1.47555912e+00 5.21592140e-01 5.48631430e-01 1.41809836e-01
-4.81739134e-01 2.34311730e-01 4.18207258e-01 2.37150714e-01
-2.98334152e-01 1.68721899e-01 -4.76804018e-01 -3.55009809e-02
-7.75627255e-01 3.73643450e-02 3.00127447e-01 1.16482782e+00
7.97281206e-01 1.55546010e-01 -6.59415007e-01 5.05004048e-01
4.69481915e-01 4.06005502e-01 7.96601474e-01 -1.11327457e+00
6.80889428e-01 5.37081122e-01 -2.15616846e-03 -7.31589496e-01
-7.75093436e-01 -3.64415884e-01 -6.12838328e-01 3.05254400e-01
4.88299012e-01 4.16070431e-01 -1.16445684e+00 1.92424834e+00
2.17653558e-01 1.02138951e-01 2.59592652e-01 9.00891721e-01
9.06028569e-01 6.58687711e-01 4.19334531e-01 -2.80247092e-01
1.15122664e+00 -1.68656766e+00 -6.45195484e-01 -6.19382501e-01
7.92229414e-01 -6.72251046e-01 1.49329102e+00 1.54034838e-01
-8.90992224e-01 -1.09976208e+00 -9.78108644e-01 -4.68067735e-01
-4.01122063e-01 1.38001159e-01 8.16533089e-01 -1.27424553e-01
-1.15731478e+00 6.21154249e-01 -8.35341454e-01 -4.27275509e-01
1.25453904e-01 3.03898603e-01 -3.27234894e-01 -8.08170959e-02
-1.35550845e+00 1.13162565e+00 5.57824969e-01 2.41064161e-01
-1.37846005e+00 -7.35931098e-01 -8.36337626e-01 -2.47448340e-01
4.38846678e-01 -8.48573446e-01 1.38919902e+00 -6.28909051e-01
-1.52322996e+00 7.55410373e-01 -3.49684477e-01 -5.59006512e-01
7.44801402e-01 -2.75840640e-01 -3.39226872e-01 -1.11705102e-01
2.25023881e-01 9.39211011e-01 6.96834445e-01 -1.12793458e+00
-9.66318369e-01 -3.14119220e-01 2.33547270e-01 4.34713334e-01
1.87085152e-01 -4.93989110e-01 -6.27847612e-01 -5.34222960e-01
1.92632556e-01 -6.79729283e-01 -1.69662043e-01 -6.01373911e-02
-9.78147462e-02 -4.33411479e-01 7.18091130e-01 -8.46680045e-01
1.37295675e+00 -1.96905828e+00 2.02100649e-01 -1.45646095e-01
-3.67280375e-03 2.64349312e-01 -6.21183872e-01 4.74659681e-01
3.29949588e-01 -3.31461757e-01 8.82821605e-02 -3.65455151e-01
3.32127027e-02 5.21783054e-01 -5.70083797e-01 6.65805815e-03
-3.85251075e-01 1.27174294e+00 -9.40731347e-01 -4.97450501e-01
3.30040306e-01 1.53893754e-02 -5.73731005e-01 2.93551505e-01
-8.32155049e-01 7.77723014e-01 -2.83181697e-01 4.74670112e-01
3.26382607e-01 -2.87044585e-01 2.11526737e-01 -3.78320038e-01
-4.55804676e-01 3.97943079e-01 -8.25583458e-01 2.06072402e+00
-7.30965614e-01 4.72369254e-01 -3.77415240e-01 -6.01126313e-01
5.85089505e-01 -1.19722433e-01 4.67640646e-02 -1.59551072e+00
-2.10125655e-01 7.72022679e-02 3.00291982e-02 -7.02763736e-01
5.91584206e-01 3.48949462e-01 1.92170501e-01 2.81906635e-01
-9.12322849e-02 4.45774168e-01 1.41517848e-01 1.09781928e-01
9.29719865e-01 8.08550239e-01 4.07402277e-01 1.42301172e-01
6.78295255e-01 2.83548266e-01 4.84716028e-01 9.98094738e-01
-3.77372891e-01 -8.33887532e-02 1.10818550e-01 -6.30230844e-01
-8.56952608e-01 -9.52539742e-01 4.11258698e-01 1.71327734e+00
5.08026719e-01 -3.54883462e-01 -4.51713234e-01 -7.53895581e-01
-8.34710225e-02 1.11912107e+00 -7.11597562e-01 -2.92898983e-01
-1.03958404e+00 -2.97713965e-01 4.30161178e-01 8.15779567e-01
8.99564385e-01 -1.56676316e+00 -8.34049821e-01 2.74331421e-01
-5.09457231e-01 -1.22836494e+00 -6.18513107e-01 -2.30122991e-02
-9.18482125e-01 -1.02840185e+00 -1.53240919e-01 -8.94500017e-01
3.87741387e-01 3.57205540e-01 9.99621451e-01 1.42288774e-01
6.58251569e-02 4.63375002e-01 -2.16703326e-01 1.54585794e-01
-2.81791002e-01 -1.07849613e-01 1.16133511e-01 -3.03958058e-01
3.49625140e-01 -4.98692870e-01 -6.85851455e-01 5.71800470e-01
-4.33714628e-01 6.86226606e-01 9.62998390e-01 8.83904457e-01
7.19575703e-01 -4.05632287e-01 3.06761056e-01 -4.65970010e-01
3.88444960e-01 -1.61839709e-01 -8.97282541e-01 5.51853061e-01
-9.21028137e-01 3.03487241e-01 6.53558791e-01 -7.29528964e-01
-1.35020792e+00 -7.48739019e-02 -1.22276254e-01 -4.09470826e-01
-1.01741441e-01 5.28598905e-01 -1.82909310e-01 -3.79251279e-02
6.13009334e-01 6.38127506e-01 -2.81207621e-01 -3.43082517e-01
6.09227061e-01 1.35862559e-01 8.34485531e-01 -5.54044247e-01
5.13790131e-01 3.99699122e-01 7.11818039e-02 -4.35103804e-01
-9.02674794e-01 -3.16639245e-01 -7.82360792e-01 -1.79885909e-01
1.00907564e+00 -9.17120218e-01 -1.19863653e+00 4.66591984e-01
-1.35089421e+00 -1.06006432e+00 -4.12813239e-02 6.15895152e-01
-8.92475665e-01 2.69575745e-01 -5.35684109e-01 -5.93695581e-01
-8.75329301e-02 -1.30750394e+00 7.31887996e-01 7.70200938e-02
4.63481955e-02 -9.72618222e-01 -1.76212728e-01 4.52173978e-01
4.20570642e-01 -1.92911193e-01 1.27976394e+00 -4.89562690e-01
-1.22440279e+00 4.38389331e-01 -4.29967731e-01 -2.71632582e-01
-3.60956043e-01 -7.13740289e-01 -7.91311562e-01 -7.92246237e-02
-1.41537145e-01 -3.06933016e-01 1.20088851e+00 4.37472165e-01
9.22580004e-01 -2.75609702e-01 -4.36911643e-01 8.02469552e-01
1.24359977e+00 7.24587977e-01 7.21782029e-01 8.29082012e-01
9.05947089e-01 7.95874238e-01 9.39563394e-01 3.27959538e-01
9.55498397e-01 9.17922735e-01 5.48616469e-01 3.46307784e-01
-5.57734609e-01 -7.41765261e-01 5.87516665e-01 6.83273733e-01
-2.65159786e-01 -1.10569462e-01 -9.23347831e-01 4.48450267e-01
-2.30053210e+00 -1.08743608e+00 5.93673177e-02 2.02104425e+00
6.58866167e-01 9.42701697e-02 -1.68180004e-01 -3.92351866e-01
3.00059468e-01 2.66329825e-01 -8.91073704e-01 -5.53105660e-02
-1.95726138e-02 -2.76773602e-01 3.72888356e-01 8.78039241e-01
-8.90898645e-01 1.65361297e+00 5.39951754e+00 9.05061960e-01
-8.78948152e-01 3.27811182e-01 1.00240640e-01 1.37410283e-01
-3.09823543e-01 4.13036272e-02 -1.14518368e+00 1.30214527e-01
6.00956976e-01 2.41445467e-01 7.35103190e-01 7.48511732e-01
3.85651529e-01 -3.30934674e-01 -1.42491674e+00 1.01753461e+00
-4.40014899e-02 -1.48268747e+00 4.54887718e-01 -1.35088131e-01
5.12366176e-01 6.59904256e-02 2.12034494e-01 8.25387597e-01
1.88432634e-01 -1.03834701e+00 1.07340801e+00 8.78716588e-01
6.14262402e-01 -4.00290161e-01 3.88572991e-01 5.74172974e-01
-1.50844157e+00 -5.54655075e-01 -2.91636616e-01 -1.34161845e-01
3.44940275e-01 -1.13562487e-01 -6.75189495e-01 4.72882807e-01
5.56644559e-01 6.65926754e-01 -6.69082344e-01 7.37352371e-01
-5.64254820e-01 1.29123986e-01 1.35351464e-01 -3.58844846e-02
5.72776496e-01 7.71887461e-03 4.43430364e-01 8.32462013e-01
2.05875233e-01 2.06324086e-01 4.41088319e-01 6.16456032e-01
2.99377084e-01 -1.21231541e-01 -4.08567786e-01 1.71669185e-01
5.85990429e-01 7.16202259e-01 -4.94786948e-01 -3.21383595e-01
-4.69239056e-01 1.03592837e+00 6.62527144e-01 5.62694848e-01
-8.31818461e-01 1.84517607e-01 5.93687296e-01 6.56945705e-02
3.39206427e-01 -6.03402317e-01 -2.21773699e-01 -8.57217371e-01
-4.88509089e-02 -7.15399921e-01 4.48425502e-01 -8.93239021e-01
-9.00534272e-01 5.97858548e-01 -3.35035138e-02 -1.13835502e+00
-5.20367980e-01 -7.50082195e-01 -3.74603659e-01 6.70965850e-01
-1.79728520e+00 -1.58912063e+00 -4.85114396e-01 9.07862604e-01
9.22248602e-01 -4.68399405e-01 5.24337709e-01 2.07024574e-01
-2.46497542e-01 4.57893252e-01 -3.55642021e-01 -1.32403001e-01
6.30169749e-01 -9.53444481e-01 4.19518739e-01 9.60255206e-01
-1.00016363e-01 7.44133234e-01 5.05697131e-01 -9.74315822e-01
-1.41730607e+00 -1.18740404e+00 1.02098525e+00 -4.19113815e-01
5.91564000e-01 -1.21779934e-01 -7.48677135e-01 8.94407749e-01
-2.44373898e-03 -2.72210956e-01 2.27379590e-01 -1.13073327e-01
-5.26743352e-01 -6.98735565e-02 -5.95414162e-01 1.04551673e+00
1.85216999e+00 -6.63880050e-01 -8.12542975e-01 4.34547603e-01
1.10000992e+00 -5.17871559e-01 -3.29088449e-01 4.41122234e-01
6.85994506e-01 -1.17967439e+00 1.23145187e+00 -6.14148200e-01
2.91850895e-01 -6.13935351e-01 -4.66454506e-01 -8.37419868e-01
-5.87603152e-01 -4.72233117e-01 -4.55201745e-01 8.53566289e-01
2.41813242e-01 -3.53461027e-01 5.53450048e-01 1.66898832e-01
-4.85729396e-01 -6.43446386e-01 -7.35909760e-01 -9.12675738e-01
-3.32336426e-01 -7.70119905e-01 5.63938618e-01 5.66854894e-01
-4.06388968e-01 4.38027829e-01 -5.71098030e-01 3.95741999e-01
4.87468004e-01 3.09862822e-01 7.38517702e-01 -9.99453962e-01
-3.41514260e-01 -5.96608162e-01 6.26438335e-02 -1.63658810e+00
3.35046142e-01 -9.97441888e-01 4.24055867e-02 -1.99084103e+00
-1.67728662e-01 -2.59809941e-01 -1.45815998e-01 6.55632555e-01
8.78149867e-02 -2.92364985e-01 5.02493322e-01 5.28296709e-01
-8.58388722e-01 6.42455816e-01 1.69547677e+00 -9.00160968e-02
-4.13646787e-01 -5.51457778e-02 -2.25820452e-01 7.30932117e-01
3.65710586e-01 -4.57463926e-03 -8.52829218e-01 -7.52489328e-01
1.25139236e-01 2.86427021e-01 5.89580357e-01 -1.04929817e+00
7.55583704e-01 -3.82332891e-01 5.45882404e-01 -1.15886617e+00
2.89799392e-01 -7.09976852e-01 -6.55079484e-02 7.62800753e-01
-3.89072508e-01 3.51210862e-01 3.82241696e-01 6.41032934e-01
-1.28275901e-02 -3.01959105e-02 5.37797809e-01 -3.87859613e-01
-1.98947406e+00 4.41979498e-01 -4.71756995e-01 -1.36108575e-02
8.32057118e-01 -3.13300759e-01 -6.81208491e-01 -2.00576276e-01
-9.02300775e-01 7.09653437e-01 1.52660862e-01 6.54378593e-01
9.64879274e-01 -1.42767060e+00 -1.35269761e-01 3.43645811e-01
2.82207668e-01 -1.45641208e-01 5.78192174e-01 1.14178467e+00
-4.28894967e-01 7.20536888e-01 -4.17947382e-01 -5.23155212e-01
-8.30041945e-01 9.73335624e-01 3.54693621e-01 -5.34879327e-01
-6.70717359e-01 8.13965857e-01 6.20782495e-01 -5.92335522e-01
6.00159824e-01 -2.54183233e-01 -6.16219282e-01 2.43314803e-02
6.32043302e-01 5.28663218e-01 -2.68517196e-01 -4.93374288e-01
-4.28628683e-01 6.85332060e-01 -1.14689052e-01 -2.88479775e-01
8.98618162e-01 -4.99839395e-01 -5.58669604e-02 4.28165793e-01
7.11816549e-01 -3.91794443e-01 -1.55189741e+00 -4.30027038e-01
3.52319479e-01 -7.89816082e-02 -1.65106118e-01 -1.16784239e+00
-5.56999207e-01 1.08618653e+00 6.42988503e-01 -6.14356637e-01
1.06048930e+00 -2.32357606e-01 8.57124746e-01 6.57484651e-01
6.01824284e-01 -1.13441288e+00 4.40426081e-01 1.15934670e+00
1.11948001e+00 -1.24308491e+00 -2.30954722e-01 5.70173413e-02
-7.50581801e-01 1.06610048e+00 1.28528368e+00 4.82003361e-01
3.08277696e-01 -2.38949612e-01 1.77906916e-01 4.35717478e-02
-9.98369396e-01 -5.59372783e-01 6.00924432e-01 8.29574108e-01
-7.86288381e-02 -2.05193147e-01 -1.11619711e-01 5.66154718e-01
-1.85419843e-01 -2.47248150e-02 -5.68746254e-02 8.72083545e-01
-6.04759753e-01 -8.02128315e-01 -3.35794315e-02 5.86245246e-02
3.69176567e-01 -3.84933203e-01 7.21560344e-02 9.31645751e-01
3.87430519e-01 7.87121892e-01 -1.43178934e-02 -6.29692078e-01
4.39570159e-01 8.12692195e-02 5.66512823e-01 -2.66473234e-01
-4.08676773e-01 9.18379724e-02 1.05567075e-01 -1.17407691e+00
-2.80348331e-01 -3.70934367e-01 -1.58442092e+00 -3.02065492e-01
3.20221573e-01 -2.94782668e-01 3.72061580e-01 1.34043932e+00
3.21217269e-01 7.95652628e-01 -1.29985260e-02 -9.17377710e-01
-5.32176316e-01 -7.63057113e-01 -5.93721820e-03 1.66301042e-01
4.39600945e-01 -9.70978796e-01 8.34961981e-02 -1.39800981e-01] | [4.475997447967529, 0.4651850759983063] |
21f0fb4d-6fc9-4007-837f-e2747b475082 | efficiently-predicting-high-resolution-mass | 2301.11419 | null | https://arxiv.org/abs/2301.11419v1 | https://arxiv.org/pdf/2301.11419v1.pdf | Efficiently predicting high resolution mass spectra with graph neural networks | Identifying a small molecule from its mass spectrum is the primary open problem in computational metabolomics. This is typically cast as information retrieval: an unknown spectrum is matched against spectra predicted computationally from a large database of chemical structures. However, current approaches to spectrum prediction model the output space in ways that force a tradeoff between capturing high resolution mass information and tractable learning. We resolve this tradeoff by casting spectrum prediction as a mapping from an input molecular graph to a probability distribution over molecular formulas. We discover that a large corpus of mass spectra can be closely approximated using a fixed vocabulary constituting only 2% of all observed formulas. This enables efficient spectrum prediction using an architecture similar to graph classification - GrAFF-MS - achieving significantly lower prediction error and orders-of-magnitude faster runtime than state-of-the-art methods. | ['Thomas Butler', 'David Healey', 'Tobias Kind', 'Ernest Fraenkel', 'Stefanie Jegelka', 'Michael Murphy'] | 2023-01-26 | null | null | null | null | ['graph-classification'] | ['graphs'] | [ 8.54440868e-01 1.47684991e-01 -7.77554452e-01 -2.86612213e-01
-8.64422560e-01 -1.00982332e+00 3.20933670e-01 9.41934109e-01
-3.69341932e-02 7.73756981e-01 -1.66595101e-01 -7.09032416e-01
-1.31100282e-01 -6.85529232e-01 -8.66695166e-01 -4.63378996e-01
-3.30765039e-01 9.56374288e-01 1.93806708e-01 1.51220605e-01
2.19220072e-01 4.71940666e-01 -1.20193982e+00 4.26932126e-01
8.26524436e-01 1.11201203e+00 1.68730214e-01 6.39688790e-01
-5.26951253e-01 7.07228959e-01 -2.79030025e-01 -5.21333933e-01
2.30972692e-01 -7.13032961e-01 -7.07874060e-01 -9.75252166e-02
6.15609467e-01 3.10844004e-01 -1.03414685e-01 1.46525156e+00
1.35277286e-01 1.69132557e-02 8.83542776e-01 -1.12240732e+00
-5.41536450e-01 5.80038786e-01 -4.69752312e-01 1.69341654e-01
9.00463641e-01 -1.43321410e-01 1.08333373e+00 -4.04545456e-01
1.07826018e+00 9.92527008e-01 6.15871251e-01 1.25069439e-01
-1.99200594e+00 -7.53873765e-01 -1.46312177e-01 2.57957608e-01
-1.50745165e+00 -2.87049264e-01 3.84138227e-01 -6.64974868e-01
1.25755656e+00 3.68654072e-01 7.46496379e-01 6.96802497e-01
5.17383695e-01 2.18289476e-02 7.82726705e-01 -2.58250833e-01
4.16524768e-01 3.21864262e-02 -2.67864158e-03 1.11078429e+00
2.97324985e-01 -9.39694270e-02 -6.98298991e-01 -8.70443642e-01
1.66698694e-01 1.38189882e-01 -3.93121630e-01 -7.08879828e-01
-1.17173314e+00 8.02681744e-01 1.08440936e-01 -3.03936124e-01
-2.19669253e-01 -1.42146170e-01 3.37898403e-01 6.79443419e-01
4.34814841e-01 8.53259206e-01 -7.74445176e-01 3.85021418e-01
-1.02837646e+00 1.23843618e-01 1.22295725e+00 8.99564981e-01
9.27466631e-01 -2.14566946e-01 6.49844050e-01 3.51034909e-01
4.10363227e-01 4.42249656e-01 5.95657051e-01 -3.94817799e-01
4.28587571e-02 5.49245715e-01 -2.67529339e-02 -6.30686641e-01
-6.09300196e-01 -2.78493732e-01 -4.06154245e-01 1.22848697e-01
8.70118856e-01 7.83269703e-02 -8.77680659e-01 1.47322965e+00
5.61692178e-01 5.25852740e-02 -1.99653760e-01 6.74721479e-01
3.10781896e-01 6.69664741e-01 2.86499739e-01 -6.86081886e-01
1.28110540e+00 -5.35793483e-01 -2.10506514e-01 -6.37667105e-02
3.36890221e-01 -9.37697530e-01 5.46927154e-01 7.04874337e-01
-7.51796186e-01 -1.13001160e-01 -1.27606952e+00 3.82505618e-02
-5.71455061e-01 -4.90780890e-01 9.54509914e-01 7.00665653e-01
-6.20683372e-01 1.20787156e+00 -6.42596304e-01 -2.92821884e-01
3.40955198e-01 6.10937834e-01 -6.08805478e-01 8.54812488e-02
-7.27608800e-01 7.24890590e-01 5.82340360e-01 -8.04670095e-01
-6.55706823e-01 -1.35967124e+00 -6.42308354e-01 4.95689474e-02
3.69752705e-01 -7.44211733e-01 1.40442169e+00 -9.28618848e-01
-1.29113925e+00 8.67535353e-01 -2.25635931e-01 -6.31846130e-01
1.29072681e-01 4.03138131e-01 -6.79639459e-01 5.38207233e-01
-1.16400952e-02 2.44909972e-01 8.57905030e-01 -6.27621114e-01
-4.35474753e-01 -4.62979704e-01 -2.04955772e-01 -1.97353646e-01
-5.11178784e-02 -1.32116124e-01 2.65846848e-01 -3.63113612e-01
3.25202793e-01 -9.02745545e-01 -4.15634423e-01 1.47643417e-01
-3.79613578e-01 1.13917626e-01 1.39733016e-01 -4.06258523e-01
1.03463364e+00 -1.42808473e+00 2.18197986e-01 5.86737812e-01
6.43860102e-01 -1.53080061e-01 -2.25320473e-01 7.57117510e-01
-6.39094412e-01 2.66972363e-01 1.08475082e-01 3.62552702e-01
1.35646850e-01 -5.22323310e-01 -4.51276720e-01 7.67875612e-01
-7.18603656e-02 5.45567036e-01 -1.18592858e+00 -3.10045332e-01
-1.71577930e-01 7.77241439e-02 -3.56616586e-01 -1.50333997e-02
-9.03686702e-01 -1.16425321e-01 -4.60488677e-01 7.83713102e-01
4.35852528e-01 -9.23707545e-01 1.00056612e+00 -3.48916531e-01
9.25130993e-02 4.40645516e-01 -8.25925887e-01 1.96969831e+00
1.43608302e-01 2.39423513e-02 -1.85217887e-01 -7.70977378e-01
7.05793023e-01 1.75778821e-01 8.72638285e-01 -2.80942678e-01
1.40302181e-01 2.08889619e-01 3.31698716e-01 -9.04181227e-02
1.32616296e-01 -7.24374235e-01 4.28081125e-01 6.58514202e-01
4.06978726e-01 2.13585664e-02 3.97314340e-01 1.77628949e-01
1.16079795e+00 2.34388664e-01 7.94902027e-01 -4.94763613e-01
2.42069319e-01 5.21839023e-01 3.11716974e-01 4.97074932e-01
-8.99156905e-04 1.58463698e-02 5.81997156e-01 -5.31737626e-01
-1.28218198e+00 -1.12791157e+00 -1.89563125e-01 1.26441395e+00
3.73678119e-03 -9.82697845e-01 -6.29672170e-01 -5.56153536e-01
3.05225492e-01 1.67028502e-01 -3.61075640e-01 -1.59465313e-01
2.40611866e-01 -9.18743849e-01 3.31258476e-01 -7.58623332e-02
-5.06519496e-01 -3.22128624e-01 -1.31509051e-01 3.69206756e-01
4.18686658e-01 -8.96030188e-01 -7.67536581e-01 6.71201527e-01
-8.27535987e-01 -1.66256547e+00 -1.54983714e-01 -5.95236480e-01
4.60379094e-01 2.17348814e-01 1.49988163e+00 -2.08333448e-01
-7.87324309e-01 -2.88477652e-02 -8.32686126e-02 -6.76771045e-01
-8.44476044e-01 2.47851703e-02 3.06820631e-01 -5.09805977e-02
5.98105550e-01 -9.75304246e-01 -6.87387168e-01 -1.31545931e-01
-8.22301388e-01 6.41628429e-02 6.86882794e-01 6.07879221e-01
1.14946866e+00 2.81147137e-02 6.41327679e-01 -1.20766532e+00
5.27268827e-01 -6.57318234e-01 -8.85000885e-01 5.57621300e-01
-9.63266730e-01 4.26890641e-01 1.02229905e+00 -7.42207825e-01
-4.26998854e-01 7.95699716e-01 2.58811772e-01 -3.93440396e-01
1.06086321e-02 7.64902949e-01 9.11827292e-03 -4.08030093e-01
8.80594254e-01 4.24995810e-01 3.25826436e-01 -3.80807489e-01
4.23852921e-01 3.30118835e-01 3.45931530e-01 -5.23590624e-01
5.12789547e-01 1.10508174e-01 6.94232225e-01 -6.53772652e-01
-9.21097100e-01 -6.19616568e-01 -5.64590156e-01 2.16610983e-01
6.55051827e-01 -8.62898290e-01 -1.21204710e+00 -4.04403299e-01
-6.11541092e-01 -1.31625116e-01 5.30550219e-02 5.15989959e-01
-5.70461333e-01 6.89279914e-01 -5.74487150e-01 -5.57626188e-01
-5.55583477e-01 -8.32277060e-01 7.86827326e-01 -1.34563133e-01
-4.52292591e-01 -7.98096478e-01 2.03148738e-01 2.03147069e-01
-3.75340916e-02 4.30502474e-01 1.53204966e+00 -1.04589081e+00
-5.85825801e-01 -2.07450733e-01 -4.48428504e-02 -4.06430632e-01
1.91012681e-01 -1.59817308e-01 -7.36777723e-01 -2.10213453e-01
-2.40289092e-01 -3.87528509e-01 6.71134949e-01 1.25375673e-01
9.88542736e-01 -1.24550007e-01 -6.45480871e-01 5.95486522e-01
1.50765049e+00 2.89386034e-01 1.54116645e-01 4.56271367e-03
4.87668842e-01 2.41179392e-01 3.01481277e-01 3.27014804e-01
-2.34386370e-01 2.96692938e-01 1.16291210e-01 2.83493638e-01
7.64918402e-02 -6.19446874e-01 2.05827311e-01 4.03399885e-01
2.57315755e-01 -1.61421999e-01 -8.44989240e-01 -1.46180913e-01
-1.65632057e+00 -1.22762895e+00 -7.59799257e-02 2.42358780e+00
1.26745713e+00 1.74029797e-01 4.85735714e-01 -1.18012689e-01
5.52701235e-01 -1.76408440e-01 -1.03635299e+00 -3.76829095e-02
8.84642899e-02 3.15390289e-01 7.62862206e-01 4.78925258e-01
-1.06634974e+00 5.55489957e-01 7.97649002e+00 8.21311235e-01
-1.20998192e+00 -3.10479939e-01 4.87151444e-01 -1.33814424e-01
-3.46945912e-01 -1.36422291e-02 -5.05816698e-01 3.65580171e-01
1.49831605e+00 -8.10945690e-01 8.57397079e-01 9.92391944e-01
-6.12777285e-02 1.15274884e-01 -1.71465123e+00 1.31638908e+00
1.63860209e-02 -1.80032551e+00 2.52594173e-01 3.11986208e-01
3.94084334e-01 3.73035789e-01 -2.78594315e-01 -2.04636648e-01
2.87715107e-01 -1.07473159e+00 7.40056872e-01 4.92192119e-01
9.94592786e-01 -7.46914327e-01 1.46253239e-02 2.60078996e-01
-1.27697742e+00 7.64866546e-02 -6.25355005e-01 2.92922463e-02
-2.32135370e-01 5.93998313e-01 -1.19569421e+00 5.97671986e-01
4.08113748e-02 5.71240842e-01 -5.01360774e-01 1.16175854e+00
2.95418561e-01 3.75138372e-01 -2.65945226e-01 -1.87961936e-01
-2.67961413e-01 -6.19508028e-01 1.92429274e-01 1.17150509e+00
2.78671175e-01 1.28208417e-02 8.19650948e-01 8.40788364e-01
-3.41656953e-01 3.96271408e-01 -6.74598396e-01 -7.99685717e-01
5.42908192e-01 1.15662432e+00 -9.66132581e-01 -5.28109491e-01
-6.39540315e-01 7.49500751e-01 4.42809582e-01 3.53030637e-02
-4.55668300e-01 -3.19069475e-01 8.85127366e-01 1.92154929e-01
2.04131335e-01 5.73570058e-02 2.03658357e-01 -1.25822520e+00
-4.65374708e-01 -1.05146503e+00 5.19477546e-01 -6.95050836e-01
-1.69469142e+00 4.67526823e-01 -3.47493231e-01 -1.09304750e+00
-1.77074581e-01 -1.04873824e+00 2.08900776e-02 8.36826444e-01
-1.34441483e+00 -6.90839171e-01 2.90597498e-01 1.74333185e-01
2.50848949e-01 -2.39918172e-01 1.38677955e+00 1.65483862e-01
-2.77249485e-01 3.00965250e-01 3.69971991e-01 -5.28758526e-01
7.73351371e-01 -1.59363449e+00 2.13320822e-01 3.13829571e-01
3.87835741e-01 7.75157273e-01 9.38876748e-01 -7.84341335e-01
-2.05145001e+00 -1.12903249e+00 5.75769901e-01 -4.38741684e-01
1.36187112e+00 -4.36026305e-01 -8.08124363e-01 5.26012897e-01
-2.34567091e-01 -9.75991189e-02 1.63002384e+00 1.50986210e-01
-1.03003955e+00 2.40459189e-01 -1.06863046e+00 4.11983460e-01
1.07427812e+00 -9.13806260e-01 -2.33543530e-01 9.35246408e-01
5.94854414e-01 -3.01281035e-01 -1.41416907e+00 -2.39382625e-01
8.71341288e-01 -4.64644164e-01 1.02464461e+00 -1.24239516e+00
1.87363639e-01 -4.27335769e-01 -2.05831468e-01 -1.27041817e+00
-3.30153704e-01 -1.06074524e+00 -3.50302726e-01 3.63918036e-01
7.62985706e-01 -4.15591985e-01 9.69628274e-01 5.17071247e-01
7.81927630e-02 -6.02030814e-01 -4.38969374e-01 -8.27705979e-01
-5.66194281e-02 -1.09668195e-01 7.74308443e-01 9.66135204e-01
8.11890423e-01 7.93237388e-01 -1.22301392e-02 2.39138454e-01
8.46522987e-01 7.89674520e-01 2.90356219e-01 -1.60219133e+00
-6.60313904e-01 -4.61939901e-01 -6.68332219e-01 -8.77691925e-01
2.83822566e-01 -1.54467058e+00 -2.32329965e-01 -1.01785803e+00
7.41577446e-01 1.75642431e-01 -4.93822306e-01 2.83161461e-01
1.23833731e-01 2.33741358e-01 -8.64409953e-02 1.13765575e-01
-7.50957906e-01 -1.27018332e-01 1.07647645e+00 -4.24051493e-01
6.18628822e-02 -4.43525493e-01 -7.64114857e-01 5.90090334e-01
6.14880145e-01 -7.53726184e-01 -7.04488814e-01 4.32520807e-01
6.90057755e-01 4.30397362e-01 4.87096757e-02 -6.90651178e-01
2.78946757e-01 -6.29697680e-01 5.49087822e-01 -3.95486951e-01
7.74148479e-02 -7.32141972e-01 6.60445333e-01 5.44894338e-01
-4.55886841e-01 -1.34915024e-01 1.05595991e-01 1.10411906e+00
1.68274283e-01 -1.64620563e-01 5.99789739e-01 -3.10008049e-01
-3.85917395e-01 6.93124294e-01 -3.84919584e-01 -2.02679634e-01
6.77717626e-01 -1.89403165e-02 -3.87102246e-01 -1.60410926e-01
-8.68205786e-01 -9.90946516e-02 6.70397878e-01 -6.31693453e-02
3.23771030e-01 -9.73151863e-01 -2.15369269e-01 -6.74582273e-02
2.92965889e-01 -6.81328893e-01 -1.64755434e-01 6.53605282e-01
-6.00253224e-01 4.65345711e-01 -1.46114662e-01 -2.74010032e-01
-1.46035743e+00 1.17428875e+00 4.68711346e-01 -1.03362359e-01
-5.17256677e-01 5.23732305e-01 1.33191124e-01 -3.22928458e-01
-3.21454951e-03 -2.00745210e-01 3.73852193e-01 -1.02818437e-01
8.68917644e-01 1.02615125e-01 1.19710267e-01 -1.46748036e-01
-3.96302521e-01 1.85687929e-01 -8.98384005e-02 5.33370972e-01
1.25235963e+00 1.77386254e-01 -3.31955999e-01 3.19617659e-01
1.29158294e+00 1.53864786e-01 -9.19712543e-01 -4.95880395e-02
2.39820704e-01 -2.46378183e-01 -2.92008463e-02 -1.01750624e+00
-3.85497361e-01 2.88693428e-01 5.16543925e-01 4.26710665e-01
6.99871659e-01 8.57950598e-02 6.46215975e-01 1.00111389e+00
4.37949866e-01 -7.42761731e-01 -8.60263258e-02 8.13475028e-02
4.24839377e-01 -1.19452822e+00 4.76484329e-01 -7.40233719e-01
-1.63850009e-01 1.57197940e+00 -9.84828696e-02 8.77579972e-02
7.18345523e-01 2.18632147e-01 -1.56049475e-01 -6.27145112e-01
-8.84650588e-01 -6.92510456e-02 5.38288414e-01 5.65202236e-01
6.86640203e-01 2.24841863e-01 3.03017497e-02 6.22254252e-01
-3.42835307e-01 3.43011878e-02 4.54586178e-01 7.85158694e-01
-7.55607307e-01 -1.49699152e+00 -1.41788408e-01 9.00569916e-01
-6.14362836e-01 -3.84758085e-01 -7.62007475e-01 4.08260137e-01
-2.06761882e-01 8.81689548e-01 -2.51566827e-01 -4.60938156e-01
-6.86384588e-02 7.33047664e-01 8.60038638e-01 -9.21766043e-01
-1.57519341e-01 2.47711703e-01 1.52089179e-01 -7.23256409e-01
-2.46921375e-01 -6.39049590e-01 -1.35910332e+00 -3.89580607e-01
-4.07104015e-01 4.44783598e-01 5.86720943e-01 6.82019472e-01
5.23436844e-01 3.29075128e-01 3.61843735e-01 -5.45131505e-01
-8.69358182e-01 -5.67979097e-01 -1.06010020e+00 3.21106136e-01
2.73881078e-01 -3.78409535e-01 -9.68873873e-02 4.74406421e-01] | [5.032797336578369, 5.737403869628906] |
d1d42152-8cc1-4c12-800a-27a26d58aba8 | weak-supervised-dysarthria-invariant-features | 2210.13144 | null | https://arxiv.org/abs/2210.13144v1 | https://arxiv.org/pdf/2210.13144v1.pdf | Weak-Supervised Dysarthria-invariant Features for Spoken Language Understanding using an FHVAE and Adversarial Training | The scarcity of training data and the large speaker variation in dysarthric speech lead to poor accuracy and poor speaker generalization of spoken language understanding systems for dysarthric speech. Through work on the speech features, we focus on improving the model generalization ability with limited dysarthric data. Factorized Hierarchical Variational Auto-Encoders (FHVAE) trained unsupervisedly have shown their advantage in disentangling content and speaker representations. Earlier work showed that the dysarthria shows in both feature vectors. Here, we add adversarial training to bridge the gap between the control and dysarthric speech data domains. We extract dysarthric and speaker invariant features using weak supervision. The extracted features are evaluated on a Spoken Language Understanding task and yield a higher accuracy on unseen speakers with more severe dysarthria compared to features from the basic FHVAE model or plain filterbanks. | ['Hugo Van hamme', 'Jinzi Qi'] | 2022-10-24 | null | null | null | null | ['spoken-language-understanding', 'spoken-language-understanding'] | ['natural-language-processing', 'speech'] | [ 2.97542904e-02 5.57123661e-01 2.42997229e-01 -6.17740273e-01
-1.04344225e+00 -4.81438249e-01 7.06581533e-01 -8.89562905e-01
-4.66463476e-01 6.09684169e-01 1.07945931e+00 -4.88294065e-02
3.72557789e-02 -2.50562668e-01 -3.14050704e-01 -5.05092680e-01
1.42739028e-01 5.44935048e-01 -1.73027441e-01 -6.06917620e-01
-3.79939497e-01 1.57738551e-01 -1.63138700e+00 2.87091762e-01
6.86654210e-01 3.03561658e-01 1.73204064e-01 1.00657701e+00
6.64172545e-02 5.05970836e-01 -8.59303296e-01 -8.08440298e-02
8.45271423e-02 -8.73103261e-01 -8.60970557e-01 2.85361684e-03
7.47653782e-01 -6.38055384e-01 -1.08757973e+00 9.63928044e-01
8.76260459e-01 1.56866804e-01 6.49343014e-01 -9.65589941e-01
-9.49113548e-01 7.75094688e-01 2.66243339e-01 6.75603211e-01
5.39555192e-01 1.05524786e-01 1.09363317e+00 -9.58101630e-01
4.77219582e-01 1.68071079e+00 3.85125965e-01 1.30814481e+00
-1.21191490e+00 -6.98135972e-01 -3.36108841e-02 3.48532885e-01
-1.14072049e+00 -1.24328279e+00 6.21473610e-01 -4.92089927e-01
1.17183149e+00 3.12287416e-02 6.88481033e-01 1.63739753e+00
-2.73442417e-01 6.90682650e-01 9.91662860e-01 -4.35921878e-01
8.28765854e-02 -9.86061152e-03 3.48459780e-01 5.77771783e-01
-3.55219513e-01 4.96409088e-01 -1.10764360e+00 -1.09771796e-01
5.41453362e-01 -4.12895292e-01 -7.34369457e-01 -4.78347316e-02
-9.45214629e-01 1.15749156e+00 -1.78044870e-01 2.16905698e-01
-1.40463129e-01 -1.51130021e-01 4.73263592e-01 1.08498728e+00
3.46625119e-01 2.89632469e-01 -6.20424390e-01 -3.67956907e-01
-7.13110685e-01 3.11425388e-01 8.52544725e-01 6.23369873e-01
1.47346988e-01 8.44604254e-01 6.49048835e-02 1.45180047e+00
2.89571077e-01 7.68313646e-01 1.23146462e+00 -9.76020873e-01
3.19540948e-01 7.65884891e-02 -4.15826410e-01 -2.84108818e-01
-3.01588118e-01 -2.68813912e-02 -1.44940212e-01 4.79430735e-01
4.00563002e-01 -2.38829419e-01 -1.19941723e+00 2.27862930e+00
-5.24577275e-02 -2.72241294e-01 5.36511719e-01 7.56487727e-01
8.13188612e-01 6.00716293e-01 1.17748789e-01 -7.56679177e-02
1.20276845e+00 -6.53300822e-01 -1.02187240e+00 -4.16930795e-01
5.47406912e-01 -4.60246980e-01 1.44914377e+00 3.80517721e-01
-1.15890646e+00 -3.99126142e-01 -1.16083896e+00 -2.53173262e-01
-3.54038805e-01 -4.08883691e-01 1.75645813e-01 1.08830774e+00
-1.23982477e+00 2.04224974e-01 -8.27125847e-01 -1.66158065e-01
3.29517633e-01 5.65545619e-01 -8.05881917e-01 -8.01225305e-02
-1.55563271e+00 1.18094766e+00 6.42078072e-02 -6.01068914e-01
-1.03444576e+00 -9.97413397e-01 -1.31419146e+00 1.47413239e-01
-2.12936461e-01 -3.52798492e-01 1.30990839e+00 -7.79940486e-01
-1.84628308e+00 9.49756682e-01 -5.42150326e-02 -6.21481657e-01
3.09935778e-01 -3.11488390e-01 -5.91380954e-01 2.55942017e-01
-3.02551478e-01 4.76304829e-01 1.12136686e+00 -7.28301764e-01
-3.10865581e-01 -6.81266189e-01 -4.10726428e-01 4.04188216e-01
-5.03111005e-01 5.17042726e-02 3.75050962e-01 -7.43619919e-01
-4.89564203e-02 -5.32829046e-01 5.76820254e-01 -3.33318323e-01
-1.77416235e-01 -4.28369552e-01 1.03915703e+00 -1.19279313e+00
7.24945247e-01 -2.11860657e+00 6.40850961e-01 -5.79121858e-02
1.89011395e-01 2.80775458e-01 -3.09923708e-01 2.37925753e-01
-2.11995870e-01 1.28656313e-01 -3.22263211e-01 -4.55748737e-01
3.36811513e-01 4.67350751e-01 -4.19840068e-01 6.21192396e-01
2.34445706e-01 6.24744475e-01 -5.47130883e-01 -1.50987029e-01
1.13660693e-01 6.29855752e-01 -8.63835037e-01 3.74133945e-01
-5.80966286e-03 3.68572593e-01 1.78094972e-02 4.18162972e-01
2.66894490e-01 4.67270941e-01 -1.21125109e-01 3.11982751e-01
2.31897339e-01 5.88399768e-01 -6.11826897e-01 1.70566595e+00
-6.46886349e-01 7.06770778e-01 6.00551724e-01 -9.25892949e-01
7.53570020e-01 6.99424684e-01 -1.02450721e-01 -7.13382781e-01
6.73032030e-02 2.66740561e-01 5.83832145e-01 -5.68577230e-01
-6.94757104e-02 -7.32878804e-01 -6.28398955e-02 3.84145796e-01
8.15802991e-01 -5.72965145e-01 -2.87173986e-01 1.68757141e-01
1.17178297e+00 -1.83959901e-01 1.35553598e-01 -3.31553757e-01
6.21650815e-01 -3.79622906e-01 2.40128726e-01 7.35251069e-01
-4.23202604e-01 8.04752648e-01 2.86325753e-01 1.20715193e-01
-1.17326117e+00 -1.55187607e+00 -3.40521902e-01 1.34322798e+00
-6.82011425e-01 -2.44599029e-01 -7.23643482e-01 -4.13310945e-01
3.10592111e-02 1.13105834e+00 -5.14527857e-01 -5.87595463e-01
-6.97328985e-01 -3.07114631e-01 9.37165856e-01 4.85673279e-01
7.34867677e-02 -1.15361202e+00 1.03461474e-01 7.64418915e-02
-7.40494281e-02 -1.04857433e+00 -4.69530970e-01 3.44983116e-02
-2.91139007e-01 -8.99201393e-01 -9.44736481e-01 -1.06806576e+00
4.11852933e-02 -5.34535527e-01 8.70751143e-01 -3.35422933e-01
-3.92261147e-01 7.87171781e-01 -4.49987888e-01 -5.36949158e-01
-1.13321388e+00 -5.63206859e-02 8.76807451e-01 -3.73374730e-01
6.71548605e-01 -7.55905867e-01 -9.71741974e-02 1.23437062e-01
-6.15052998e-01 -5.36496222e-01 1.53731033e-01 1.28699863e+00
1.11162603e-01 -5.03027618e-01 1.12918389e+00 -5.13204277e-01
8.81160557e-01 -3.59306335e-01 -2.01966166e-01 -2.89869219e-01
-2.72912413e-01 1.36295497e-01 7.50342071e-01 -3.91757637e-01
-1.14795125e+00 -1.69312239e-01 -7.14890480e-01 -3.94290984e-01
-4.25246418e-01 -6.24776185e-02 -3.62624705e-01 2.61623114e-01
7.61651576e-01 5.35041392e-01 6.96416676e-01 -6.06301427e-01
4.83174741e-01 1.08061576e+00 7.19235063e-01 -2.03305006e-01
6.68582261e-01 1.61086783e-01 -8.30964327e-01 -1.72003555e+00
-5.49909949e-01 -2.24749178e-01 -4.38884199e-01 2.32923135e-01
9.88549888e-01 -9.49139893e-01 -3.15702349e-01 4.93583471e-01
-8.48453045e-01 -4.36817050e-01 -6.13733768e-01 7.72126079e-01
-9.98282909e-01 3.13731104e-01 -4.85127240e-01 -7.58340240e-01
-3.27225208e-01 -1.25618660e+00 8.52271318e-01 -2.32615024e-01
-7.24099040e-01 -8.85658205e-01 4.59619343e-01 5.84121883e-01
5.85877776e-01 -3.22905600e-01 1.16068745e+00 -1.38521481e+00
1.77245945e-01 -2.12260574e-01 4.20281708e-01 9.71233070e-01
2.89233834e-01 -7.37313807e-01 -1.33232939e+00 -4.20839459e-01
4.39495981e-01 -8.08767378e-01 8.14825773e-01 1.85895711e-01
5.32194436e-01 -3.02191347e-01 2.34979138e-01 5.26546359e-01
5.24741411e-01 -3.62477526e-02 3.81672561e-01 -2.15065330e-01
4.63224113e-01 8.11590314e-01 -2.37853274e-01 2.99015939e-02
9.48927253e-02 5.79061210e-01 -1.64345875e-01 3.93850476e-01
-7.89617658e-01 -2.35617906e-01 8.48262131e-01 1.38121426e+00
2.74676979e-01 -1.67808101e-01 -8.82657528e-01 9.26944554e-01
-1.02438617e+00 -1.00939393e+00 4.31887597e-01 1.98532772e+00
9.62796569e-01 -1.66621208e-01 1.91256121e-01 2.04467729e-01
5.09569407e-01 3.48302811e-01 -5.56263447e-01 -4.61036772e-01
-2.72471160e-01 5.70794284e-01 -2.34966408e-02 1.05745077e+00
-6.51777565e-01 1.12103987e+00 7.09239149e+00 4.67023611e-01
-7.51675785e-01 3.84300113e-01 -3.27411201e-03 -6.47918761e-01
-1.60001010e-01 -6.24861956e-01 -6.93800628e-01 3.72457802e-01
1.60179424e+00 -2.78734714e-01 5.51436245e-01 7.65260577e-01
1.25255600e-01 6.93337679e-01 -1.33506787e+00 1.02917802e+00
5.64466774e-01 -9.05991733e-01 1.38337478e-01 -3.21961269e-02
2.46690094e-01 5.62162042e-01 2.47513190e-01 5.86737514e-01
5.64988017e-01 -1.51108050e+00 4.49609131e-01 3.19180131e-01
9.78621483e-01 -8.06714654e-01 5.49507678e-01 1.43417537e-01
-3.85896772e-01 -1.05150819e-01 -3.14599127e-01 2.98084058e-02
1.70075342e-01 -2.78104454e-01 -9.05526459e-01 -3.20430636e-01
5.32812834e-01 2.05885977e-01 2.29553152e-02 2.26039395e-01
-2.91275710e-01 9.96240497e-01 -5.06141782e-01 -4.66490909e-02
-3.11944168e-02 2.03768373e-01 1.02851760e+00 8.97334576e-01
6.60795197e-02 2.24756837e-01 -2.37548143e-01 7.46333241e-01
-9.79939252e-02 1.35715112e-01 -9.17252541e-01 -2.53367454e-01
3.08496237e-01 3.33291501e-01 3.41240972e-01 -1.48594841e-01
-6.88205421e-01 1.12228739e+00 4.15391773e-01 4.21463221e-01
-4.17341068e-02 -4.52786297e-01 1.15041995e+00 4.59223613e-02
3.94713551e-01 -2.52848059e-01 -1.97200682e-02 -1.18351567e+00
-2.30165586e-01 -1.27345729e+00 1.58461303e-01 -3.69691581e-01
-1.41375482e+00 6.84958696e-01 -7.12117478e-02 -4.55119729e-01
-9.68049049e-01 -8.22324097e-01 -6.64430082e-01 1.30404365e+00
-1.23087060e+00 -9.13581312e-01 1.57600805e-01 8.94773066e-01
1.14744353e+00 -8.96832108e-01 1.13836861e+00 -4.76574264e-02
-2.32822463e-01 8.17148387e-01 6.62946105e-02 2.01560721e-01
6.13635361e-01 -1.33150661e+00 4.97927517e-01 6.21479928e-01
1.41108468e-01 6.10164404e-01 8.87443244e-01 -4.47429687e-01
-1.05560613e+00 -5.96413374e-01 8.04208517e-01 -7.05612898e-01
7.45064795e-01 -7.76587367e-01 -1.14690208e+00 7.33310699e-01
3.39977682e-01 -2.74357975e-01 1.01884043e+00 4.99512181e-02
-5.66079140e-01 4.22206908e-01 -1.12611520e+00 5.33863127e-01
1.18332982e+00 -1.11138117e+00 -1.47859049e+00 4.72615689e-01
1.04690945e+00 7.20678717e-02 -9.68531013e-01 -1.34351566e-01
5.14783442e-01 -6.75621986e-01 1.03748357e+00 -1.10493660e+00
3.20368975e-01 4.85317022e-01 -5.77564716e-01 -1.88412714e+00
1.24004371e-02 -7.38632381e-01 -9.83928517e-02 1.03262532e+00
4.24237341e-01 -7.63925910e-01 7.70607471e-01 4.72013235e-01
-3.93464506e-01 -1.70775831e-01 -1.30744755e+00 -7.86044538e-01
8.68249059e-01 -2.54446507e-01 1.80040240e-01 7.62754440e-01
4.01149303e-01 4.23843801e-01 -2.56200492e-01 2.31724024e-01
5.40693283e-01 -5.05138874e-01 4.26684976e-01 -1.22274256e+00
-3.64871562e-01 -2.60046422e-01 -7.07448006e-01 -6.71757519e-01
9.24929321e-01 -1.12922215e+00 8.07579681e-02 -1.13778341e+00
-3.23174566e-01 3.33240449e-01 3.20229679e-02 4.71317083e-01
6.58883378e-02 8.80963355e-02 7.37691894e-02 -1.26084208e-01
1.67987451e-01 1.03191030e+00 9.41133499e-01 -1.59336194e-01
-3.63302886e-01 -1.12775266e-01 -5.55060327e-01 9.70307946e-01
9.34499025e-01 -4.03013140e-01 -6.41922832e-01 -3.30101013e-01
-4.65777040e-01 1.08591713e-01 1.93050146e-01 -8.15702081e-01
-4.88609122e-03 5.73683381e-01 3.24259311e-01 -1.00832447e-01
8.07813942e-01 -2.11134523e-01 -7.66686559e-01 4.71598089e-01
-5.90798318e-01 -2.87302792e-01 1.85545072e-01 4.99399185e-01
-3.37511927e-01 -2.49853790e-01 1.10982955e+00 -3.23908389e-01
-4.57557887e-01 5.87504804e-02 -8.29958975e-01 6.96723878e-01
3.17977309e-01 -3.52847688e-02 -1.30791321e-01 -8.96743596e-01
-1.53343487e+00 -3.23424302e-02 1.76661953e-01 7.45497823e-01
7.81330407e-01 -1.00629282e+00 -1.25310731e+00 8.54151487e-01
4.93658632e-02 -5.71490884e-01 3.28949630e-01 3.94857645e-01
-1.29214749e-01 4.30729836e-01 -3.58024418e-01 -2.88906872e-01
-1.42210913e+00 1.24799654e-01 4.77200449e-01 3.42619568e-01
-8.87378871e-01 1.09897697e+00 2.68988818e-01 -9.00932848e-01
2.87703604e-01 1.10436596e-01 -1.59139112e-01 7.88631737e-02
7.37958491e-01 4.14928973e-01 2.68658753e-02 -1.07798100e+00
-2.43761376e-01 2.02934161e-01 -5.61386645e-01 -6.17392302e-01
1.39954221e+00 -2.81649172e-01 3.96028906e-01 5.21093786e-01
1.49431062e+00 2.70002663e-01 -1.05772209e+00 -2.08080098e-01
-5.13624310e-01 -4.72906530e-02 4.01008129e-01 -7.24695921e-01
-6.73756421e-01 1.04663229e+00 9.20694709e-01 3.35318774e-01
3.82541806e-01 5.18839180e-01 7.38488555e-01 4.67560500e-01
8.43014102e-03 -1.05058777e+00 -1.60179317e-01 7.63852715e-01
1.33081794e+00 -1.09282291e+00 -4.21177030e-01 -7.79429972e-02
-6.17535591e-01 9.06672060e-01 4.05627847e-01 -3.24478626e-01
7.82866418e-01 3.92101645e-01 4.78417009e-01 -3.37946236e-01
-4.73815918e-01 -2.59666562e-01 1.79652303e-01 1.07169044e+00
2.58893490e-01 -1.01899199e-01 6.61772583e-03 9.13431227e-01
-1.28814697e+00 -5.00814497e-01 6.92501068e-01 4.71625805e-01
-4.71797228e-01 -8.65423024e-01 -3.94882709e-01 6.27793670e-01
-6.20355010e-01 -2.58004189e-01 -5.90448081e-01 9.43076611e-01
-3.94544691e-01 9.72068310e-01 3.41791183e-01 -8.44840705e-02
5.90599418e-01 9.45420444e-01 4.09645885e-01 -8.73843610e-01
-3.77450585e-01 -1.04543805e-01 3.35645586e-01 -5.85469484e-01
1.21948913e-01 -8.41566503e-01 -1.17625320e+00 2.17456490e-01
-1.09913044e-01 -2.86636379e-04 6.18976057e-01 1.06997538e+00
8.92900154e-02 7.11332262e-01 1.73876524e-01 -7.35534310e-01
-1.21935582e+00 -1.38788426e+00 -7.49961615e-01 6.25881732e-01
9.89048004e-01 -6.61636114e-01 -1.11441553e+00 -3.15827727e-02] | [14.581570625305176, 6.474009990692139] |
2405aa3b-3b21-4860-8fd8-e63f8693a77a | learning-sense-specific-word-embeddings-by | null | null | https://aclanthology.org/C14-1048 | https://aclanthology.org/C14-1048.pdf | Learning Sense-specific Word Embeddings By Exploiting Bilingual Resources | null | ['Ting Liu', 'Jiang Guo', 'Wanxiang Che', 'Haifeng Wang'] | 2014-08-01 | learning-sense-specific-word-embeddings-by-1 | https://aclanthology.org/C14-1048 | https://aclanthology.org/C14-1048.pdf | coling-2014-8 | ['learning-word-embeddings', 'chinese-named-entity-recognition'] | ['methodology', '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] | [-7.3983001708984375, 3.7880756855010986] |
12e9deee-93ca-468b-8280-e1d09aa6e97a | umuteam-lt-edi-acl2022-detecting-homophobic | null | null | https://aclanthology.org/2022.ltedi-1.16 | https://aclanthology.org/2022.ltedi-1.16.pdf | UMUTeam@LT-EDI-ACL2022: Detecting homophobic and transphobic comments in Tamil | This working-notes are about the participation of the UMUTeam in a LT-EDI shared task concerning the identification of homophobic and transphobic comments in YouTube. These comments are written in English, which has high availability to machine-learning resources; Tamil, which has fewer resources; and a transliteration from Tamil to Roman script combined with English sentences. To carry out this shared task, we train a neural network that combines several feature sets applying a knowledge integration strategy. These features are linguistic features extracted from a tool developed by our research group and contextual and non-contextual sentence embeddings. We ranked 7th for English subtask (macro f1-score of 45%), 3rd for Tamil subtask (macro f1-score of 82%), and 2nd for Tamil-English subtask (macro f1-score of 58%). | ['Rafael Valencia-García', 'Camilo Caparros-Laiz', 'José García-Díaz'] | null | null | null | null | ltedi-acl-2022-5 | ['transliteration'] | ['natural-language-processing'] | [-2.00432301e-01 8.69365782e-02 -3.19513887e-01 -3.14485073e-01
-9.37800229e-01 -4.81673151e-01 4.09613699e-01 4.13206756e-01
-1.00103331e+00 6.35359049e-01 5.78981876e-01 -5.50671577e-01
-2.10369706e-01 -4.65899169e-01 -1.71227604e-01 -5.55623136e-02
2.04729453e-01 4.29563314e-01 -1.15982458e-01 -3.69836926e-01
5.72216690e-01 1.72512218e-01 -1.13376284e+00 7.47904718e-01
1.01133740e+00 7.97443151e-01 2.92382427e-02 7.52960980e-01
5.58664165e-02 6.78057075e-01 -2.97062963e-01 -1.03872585e+00
8.46031308e-02 -3.01141828e-01 -1.02959585e+00 -7.78803468e-01
3.29029113e-01 -2.07202747e-01 -1.04872897e-01 1.08062863e+00
5.31762421e-01 2.64333576e-01 5.85357189e-01 -7.04431951e-01
-7.81070054e-01 3.79305869e-01 4.01865058e-02 3.58199507e-01
6.68170214e-01 -1.52846992e-01 1.21820235e+00 -1.21744061e+00
8.51385295e-01 1.21095622e+00 7.50615656e-01 4.12843794e-01
-8.52306366e-01 -9.34521854e-01 -2.97220021e-01 9.62584764e-02
-1.07851720e+00 -3.93500894e-01 4.50824499e-01 -4.57811356e-01
1.17760515e+00 2.24136367e-01 5.24706781e-01 1.45190942e+00
-3.49696390e-02 5.51520050e-01 1.32658160e+00 -3.12247187e-01
-6.18721358e-02 1.02754080e+00 8.20191428e-02 5.51673710e-01
2.07049269e-02 -1.55826151e-01 -7.07556605e-01 -2.63305455e-01
3.29108149e-01 -1.40790343e-01 4.32488956e-02 3.78082246e-01
-6.70753419e-01 1.41006303e+00 1.08478449e-01 6.25544012e-01
-1.54660314e-01 -5.48362195e-01 6.23565078e-01 5.98008156e-01
5.92117131e-01 6.95857942e-01 -7.03639627e-01 -7.82520354e-01
-6.46076679e-01 2.52466857e-01 9.05276418e-01 3.88853073e-01
6.39977515e-01 -2.46958449e-01 2.02870950e-01 1.49001014e+00
-2.65492257e-02 4.82044637e-01 1.04446626e+00 -6.11463904e-01
5.22544801e-01 7.74041533e-01 8.51583332e-02 -1.30439246e+00
-5.39554238e-01 -1.18167423e-01 -3.72854054e-01 -2.73355246e-01
2.97804385e-01 -6.07475936e-01 -1.92190647e-01 1.56582856e+00
-1.87594444e-02 -2.86956102e-01 4.31266129e-02 6.66344345e-01
1.09515834e+00 4.58664685e-01 1.96545556e-01 -1.34371340e-01
1.19107950e+00 -8.27537060e-01 -5.00499070e-01 -1.47491455e-01
8.68333340e-01 -1.01704204e+00 1.03142965e+00 3.24776351e-01
-1.24361026e+00 -7.43628800e-01 -6.94554627e-01 -1.15434356e-01
-8.26285839e-01 2.64303863e-01 2.74162263e-01 1.17201245e+00
-1.03023612e+00 5.00800967e-01 -1.04365937e-01 -8.28676581e-01
5.76553382e-02 4.02350187e-01 -7.12211132e-01 -2.24222876e-02
-1.42938924e+00 1.19906902e+00 2.79785991e-01 -4.46415246e-01
-3.56511652e-01 -7.90716410e-01 -9.24980581e-01 -3.00621390e-02
1.66733369e-01 -3.83926816e-02 9.51570928e-01 -1.02301693e+00
-1.35729039e+00 1.37530208e+00 -1.89411178e-01 -3.54447156e-01
4.04504806e-01 -3.76560003e-01 -8.60637724e-01 2.54839063e-01
4.76637632e-01 3.66825730e-01 6.87792659e-01 -4.22064394e-01
-8.49457264e-01 -5.04722655e-01 1.43041879e-01 2.91298509e-01
-1.05624950e+00 5.77112198e-01 -6.63440451e-02 -4.47648048e-01
-5.55191159e-01 -8.37877870e-01 7.62726665e-02 -4.38390106e-01
-9.39961597e-02 -5.08960485e-01 9.05402958e-01 -1.11457860e+00
1.41026175e+00 -2.10203695e+00 -4.60675471e-02 1.72726214e-01
2.57686768e-02 6.09473288e-01 -5.62631972e-02 8.11746538e-01
-1.69773832e-01 5.87161243e-01 4.21967685e-01 -4.52867806e-01
1.03574112e-01 -1.78521246e-01 -8.95481110e-02 -2.89438721e-02
1.73958108e-01 6.09716773e-01 -1.08837986e+00 -4.09127682e-01
2.14283586e-01 3.08568299e-01 -7.07562149e-01 -7.77400006e-03
5.27154922e-01 1.38340831e-01 -1.83773205e-01 2.90581197e-01
4.46303755e-01 1.92771912e-01 1.74099505e-01 5.40060028e-02
-3.70557338e-01 5.45364320e-01 -9.44822252e-01 1.10837674e+00
-9.69698846e-01 5.71984887e-01 1.17973536e-01 -6.60413504e-01
1.17636752e+00 1.86202511e-01 4.48175848e-01 -7.15100110e-01
3.20429713e-01 5.79706967e-01 -4.92296123e-04 -7.50555873e-01
9.02244985e-01 -4.83191311e-02 -1.94672197e-01 4.28181887e-01
1.43232435e-01 -2.35504583e-02 1.98982820e-01 4.37641144e-01
9.30443227e-01 -2.28279367e-01 2.07033679e-01 -4.00782138e-01
6.23247564e-01 -4.13412750e-01 4.72432852e-01 6.24455929e-01
-5.40797710e-01 3.27926457e-01 6.21740520e-01 -4.63125229e-01
-8.53145123e-01 -6.45197809e-01 -3.28521758e-01 1.32358086e+00
-3.90281588e-01 -7.47281551e-01 -5.52180707e-01 -7.23128855e-01
-1.65376887e-01 8.29152644e-01 -7.03519404e-01 -4.06329989e-01
-2.35033423e-01 -3.12206030e-01 5.41736186e-01 3.18392515e-01
5.70376039e-01 -1.04871237e+00 -5.50092936e-01 2.44148262e-02
-4.89607036e-01 -1.26789546e+00 -4.63074058e-01 9.42431297e-03
-2.72295862e-01 -9.83951509e-01 -6.04674518e-01 -9.33133006e-01
7.97705874e-02 -1.61460787e-01 6.66958988e-01 -3.74383442e-02
-3.38240623e-01 3.25518161e-01 -6.92863226e-01 -4.88719404e-01
-1.55141667e-01 2.57947624e-01 1.17953442e-01 1.03000030e-01
1.00800049e+00 -8.36222693e-02 -2.74408758e-01 1.30491674e-01
-6.61059856e-01 -7.17948526e-02 1.00667812e-01 7.66740739e-01
-1.54815763e-01 -1.05198853e-01 7.99875677e-01 -8.21446061e-01
8.42247784e-01 -7.27427006e-01 -3.19455005e-02 -1.87915280e-01
-1.36850983e-01 -7.42595077e-01 5.50647497e-01 -2.72978187e-01
-9.04905915e-01 -3.94409001e-01 -6.20703101e-01 1.09399129e-02
-1.69143513e-01 5.79273701e-01 1.74157768e-01 1.14581250e-01
5.52387893e-01 -1.80726796e-01 -1.68280885e-01 -3.16462755e-01
-2.60606736e-01 1.24404848e+00 1.31018758e-01 -4.80827510e-01
4.75067794e-01 8.89973436e-03 -5.32680273e-01 -1.35826576e+00
-9.12854373e-01 -7.88906276e-01 -4.72423226e-01 -1.04230277e-01
1.25685608e+00 -8.54122519e-01 -6.87681377e-01 3.63020509e-01
-8.84038925e-01 -2.65322924e-01 7.73372874e-02 7.92679846e-01
-1.57968223e-01 1.41578346e-01 -5.50189793e-01 -8.84178400e-01
-4.69353646e-01 -1.06525111e+00 6.11925721e-01 1.53477788e-01
-8.64010990e-01 -1.28397107e+00 1.44863963e-01 7.10224152e-01
5.41379571e-01 -4.85215485e-02 1.15294003e+00 -1.40438664e+00
6.35294139e-01 -3.11068237e-01 -5.47526836e-01 5.87868571e-01
1.93391263e-01 -1.44209027e-01 -7.33386159e-01 -3.68895739e-01
-1.11442879e-01 -7.95406401e-01 4.95104820e-01 -5.71452864e-02
9.18402612e-01 -2.21375957e-01 -1.89142115e-02 2.92568468e-02
1.24744499e+00 2.47172683e-01 4.55806017e-01 3.96606326e-01
3.54077965e-01 6.92509651e-01 3.28834504e-01 5.23893297e-01
6.63762987e-01 7.41104484e-01 6.63642436e-02 2.36276358e-01
1.29157722e-01 -5.23468018e-01 6.82912886e-01 9.08254385e-01
1.00740597e-01 2.90454417e-01 -9.01926041e-01 1.01425815e+00
-1.47804165e+00 -9.08022344e-01 -3.04950684e-01 2.17665076e+00
5.92896402e-01 1.55544236e-01 6.23379827e-01 6.41862899e-02
5.40545106e-01 5.84727824e-02 -1.11697510e-01 -1.20764041e+00
1.00461170e-01 6.30149066e-01 8.77487883e-02 6.74476564e-01
-1.12164998e+00 1.01064110e+00 5.74822617e+00 8.29929709e-01
-1.08168638e+00 2.62029260e-01 4.93791610e-01 7.37887397e-02
-1.65781543e-01 -2.09835526e-02 -9.85550940e-01 7.81170368e-01
1.52156603e+00 -3.43220651e-01 1.81855962e-01 7.89205730e-01
-1.47796585e-03 -1.15323126e-01 -7.34598815e-01 9.16491330e-01
4.50777441e-01 -1.13482010e+00 -1.78892300e-01 4.03430574e-02
8.29945505e-01 1.91271618e-01 3.11655492e-01 7.87543893e-01
2.28730783e-01 -1.04377079e+00 1.54910684e-01 2.03091711e-01
7.72025168e-01 -9.21636641e-01 1.07967222e+00 4.17062044e-01
-8.35212052e-01 -2.03403309e-01 -3.59308898e-01 -4.34925467e-01
-2.06843466e-01 3.84901464e-01 -8.91631305e-01 7.17674434e-01
9.84281182e-01 8.82233977e-01 -5.68709612e-01 7.19320893e-01
-1.61842257e-01 5.06865859e-01 -2.05522791e-01 -6.34548843e-01
3.07022452e-01 -1.47415653e-01 3.55094820e-01 1.52901757e+00
3.35967720e-01 -1.81964546e-01 2.90242404e-01 2.45283216e-01
-1.57721549e-01 7.20481098e-01 -9.25427973e-01 -2.59111047e-01
2.76290417e-01 1.45334959e+00 -3.47184628e-01 -2.05990478e-01
-4.81366158e-01 9.60789740e-01 4.91221398e-01 2.69674659e-02
-4.79609311e-01 -9.66918766e-01 4.77449000e-01 5.04839309e-02
2.35005274e-01 1.20899938e-01 -2.42175803e-01 -1.27896976e+00
-4.33214098e-01 -9.63867009e-01 3.26744288e-01 -7.24150956e-01
-1.35185742e+00 8.28831971e-01 -8.07144567e-02 -8.72293890e-01
-3.61068398e-01 -1.00600863e+00 -5.50177574e-01 1.02088237e+00
-9.92635727e-01 -1.01779461e+00 1.55528724e-01 6.37702942e-01
5.73192954e-01 -5.62790275e-01 1.11872709e+00 6.57084405e-01
-5.92155039e-01 7.82809436e-01 5.48876375e-02 3.27097178e-01
9.69546258e-01 -9.72464621e-01 -1.69067919e-01 4.40588504e-01
-2.89142430e-01 8.42767239e-01 4.14347529e-01 -6.98613226e-01
-7.26953924e-01 -1.15202355e+00 1.89193058e+00 -5.72076201e-01
1.09638441e+00 -1.52723834e-01 -6.16932094e-01 6.65697277e-01
3.24639827e-01 -6.92873180e-01 1.24842584e+00 4.00268525e-01
-3.14906210e-01 2.48692766e-01 -1.30070293e+00 6.67578518e-01
6.17116868e-01 -1.01109588e+00 -7.65899658e-01 5.27291894e-01
2.76562899e-01 -2.56736964e-01 -1.18964183e+00 9.38225165e-03
7.81376719e-01 -8.86465490e-01 5.61283469e-01 -8.01760495e-01
8.58727634e-01 4.98020828e-01 -3.00813377e-01 -1.36565447e+00
-8.30134824e-02 -5.39019763e-01 6.48525417e-01 1.16662562e+00
4.78897482e-01 -7.60272801e-01 5.48885047e-01 7.09650278e-01
-4.73212488e-02 -6.69879317e-01 -1.09027421e+00 -4.89291400e-01
4.31408912e-01 -4.44943100e-01 -8.63062590e-02 1.32392085e+00
3.48914534e-01 4.11465764e-01 -4.46033984e-01 -4.47088063e-01
-1.44601882e-01 -3.83598119e-01 6.48784757e-01 -1.23205483e+00
-2.76990235e-03 -5.88610768e-01 -1.57069311e-01 -3.46797079e-01
5.06399751e-01 -1.06315148e+00 -5.91237664e-01 -1.45933867e+00
2.24668697e-01 -6.13315627e-02 -1.21702619e-01 4.49282408e-01
-1.78097771e-03 2.74392426e-01 3.42803925e-01 -3.21884662e-01
-3.77332032e-01 3.50242287e-01 6.41399026e-01 8.83607268e-02
-2.00059190e-01 -2.83541624e-02 -1.00126481e+00 8.66953492e-01
9.35719430e-01 -4.84703809e-01 -2.49998465e-01 -1.66237727e-01
3.65791202e-01 -2.34411865e-01 1.27198875e-01 -8.30016255e-01
-4.81184237e-02 -2.96834968e-02 2.55048484e-01 -3.97896588e-01
5.76298177e-01 -5.82866907e-01 -3.86116058e-01 5.35997748e-01
-4.68565375e-01 3.22925448e-01 3.08702558e-01 -1.77055985e-01
-2.64365613e-01 -7.27683008e-01 5.82702994e-01 -1.29298389e-01
-3.83675903e-01 3.72173563e-02 -7.19303131e-01 3.45673800e-01
9.02870834e-01 -2.74166077e-01 -2.23417029e-01 -7.93538153e-01
-6.22305930e-01 -5.41863125e-03 1.29205346e-01 7.72182107e-01
6.82381928e-01 -1.02466893e+00 -6.69194520e-01 3.73992920e-01
8.08934718e-02 -1.05772448e+00 3.01028401e-01 1.01493752e+00
-3.07813823e-01 5.30561447e-01 -4.31750625e-01 2.75396168e-01
-1.32273722e+00 -6.61523864e-02 -3.09134950e-03 -6.09133840e-01
-5.67534193e-02 9.01862442e-01 -2.76242763e-01 -1.04478037e+00
-5.38196824e-02 2.78914988e-01 -4.84918028e-01 4.89175469e-01
6.71353042e-01 6.79026723e-01 3.00203145e-01 -9.51007783e-01
-2.80776471e-01 5.97812474e-01 -3.43890220e-01 -3.31297100e-01
1.27295935e+00 1.13128327e-01 -1.82854444e-01 5.29123843e-01
1.89347494e+00 4.63993788e-01 1.06538169e-01 -9.75248218e-02
-1.30714327e-01 -4.58793402e-01 -1.36014551e-01 -8.89587224e-01
-4.88999993e-01 1.09497273e+00 5.02636373e-01 3.54383469e-01
5.18617809e-01 -5.99926353e-01 8.73119354e-01 4.35586929e-01
1.38948500e-01 -1.49754202e+00 3.90567273e-01 1.09736049e+00
8.42583120e-01 -1.34328914e+00 -2.68865138e-01 2.76752301e-02
-9.93192434e-01 1.21785569e+00 7.27457941e-01 -1.43765464e-01
8.43182623e-01 -3.73861432e-01 3.56683098e-02 5.58164436e-03
-4.46890861e-01 -1.62182137e-01 2.91085154e-01 5.91364264e-01
5.13409376e-01 -6.23443648e-02 -7.90654719e-01 9.38035727e-01
-5.92296898e-01 -8.77810568e-02 3.85356337e-01 4.92375821e-01
-2.49544814e-01 -1.06698346e+00 1.41994596e-01 6.77092075e-01
-9.20402169e-01 -1.94356129e-01 -6.90323889e-01 8.74167919e-01
1.73763081e-01 1.19825518e+00 3.87470424e-02 -7.30466425e-01
2.07567632e-01 1.86713710e-01 -3.01133901e-01 -6.77579105e-01
-1.21547270e+00 -2.18838155e-01 6.73464894e-01 -2.63887852e-01
-3.37097757e-02 -7.91319668e-01 -8.09387386e-01 -3.53952974e-01
-2.48110276e-02 4.26797539e-01 7.40714371e-01 9.27131236e-01
1.83694586e-01 1.26128703e-01 6.41771793e-01 -4.43316460e-01
-2.46284232e-01 -1.17583013e+00 -2.84483045e-01 4.71234709e-01
3.58449817e-02 -3.20243806e-01 -4.35579300e-01 -5.37749767e-01] | [8.966151237487793, 10.672106742858887] |
78379d78-a44c-42bc-8a5a-ce93d675c1b2 | transductive-learning-for-zero-shot-object | null | null | http://openaccess.thecvf.com/content_ICCV_2019/html/Rahman_Transductive_Learning_for_Zero-Shot_Object_Detection_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Rahman_Transductive_Learning_for_Zero-Shot_Object_Detection_ICCV_2019_paper.pdf | Transductive Learning for Zero-Shot Object Detection | Zero-shot object detection (ZSD) is a relatively unexplored research problem as compared to the conventional zero-shot recognition task. ZSD aims to detect previously unseen objects during inference. Existing ZSD works suffer from two critical issues: (a) large domain-shift between the source (seen) and target (unseen) domains since the two distributions are highly mismatched. (b) the learned model is biased against unseen classes, therefore in generalized ZSD settings, where both seen and unseen objects co-occur during inference, the learned model tends to misclassify unseen to seen categories. This brings up an important question: How effectively can a transductive setting address the aforementioned problems? To the best of our knowledge, we are the first to propose a transductive zero-shot object detection approach that convincingly reduces the domain-shift and model-bias against unseen classes. Our approach is based on a self-learning mechanism that uses a novel hybrid pseudo-labeling technique. It progressively updates learned model parameters by associating unlabeled data samples to their corresponding classes. During this process, our technique makes sure that knowledge that was previously acquired on the source domain is not forgotten. We report significant 'relative' improvements of 34.9% and 77.1% in terms of mAP and recall rates over the previous best inductive models on MSCOCO dataset.
| [' Nick Barnes', ' Salman Khan', 'Shafin Rahman'] | 2019-10-01 | null | null | null | iccv-2019-10 | ['zero-shot-object-detection'] | ['computer-vision'] | [ 5.79707563e-01 2.93158144e-01 -3.33344877e-01 -3.43983233e-01
-7.90027738e-01 -3.53220522e-01 8.49916875e-01 2.01273873e-01
-2.97242224e-01 8.22776496e-01 -1.83673620e-01 -1.45635130e-02
-1.69623066e-02 -9.94331777e-01 -8.12812090e-01 -6.90881014e-01
2.74653077e-01 7.81823993e-01 7.93262124e-01 -1.50414392e-01
1.22436181e-01 1.12460665e-01 -2.03115439e+00 1.94483310e-01
7.42342174e-01 8.65729988e-01 9.83870402e-02 5.08100688e-01
-2.01789752e-01 7.30540454e-01 -6.52254879e-01 -3.17425251e-01
2.78070867e-01 -5.13014615e-01 -6.20088279e-01 5.40125184e-02
5.89699745e-01 -3.23243767e-01 -1.68084651e-01 1.34005415e+00
5.22142410e-01 2.30945662e-01 8.46539199e-01 -1.42691851e+00
-8.43445182e-01 4.57739472e-01 -5.84597528e-01 5.21330714e-01
6.15639761e-02 3.31846289e-02 6.80605948e-01 -1.17389107e+00
7.30892479e-01 1.32560515e+00 4.88300174e-01 8.56857300e-01
-1.36559057e+00 -6.81974947e-01 9.82330590e-02 2.85965085e-01
-1.40063667e+00 -4.79618400e-01 6.08763337e-01 -5.34770608e-01
7.26193011e-01 2.32561335e-01 2.72415012e-01 1.48312294e+00
-1.15715407e-01 7.49734938e-01 9.76812959e-01 -7.81693697e-01
7.83533156e-01 6.83109581e-01 5.53222001e-01 4.22728390e-01
5.65469801e-01 2.92589784e-01 -5.97537994e-01 -3.94973755e-02
4.16141063e-01 6.63503930e-02 3.67224328e-02 -7.05260694e-01
-1.00364375e+00 7.74879992e-01 2.82483578e-01 5.17471194e-01
-5.53222597e-02 -1.49410650e-01 1.77528322e-01 1.88458934e-01
5.91314316e-01 3.94898057e-01 -2.41130605e-01 1.72035709e-01
-7.64028728e-01 8.77058730e-02 6.01370335e-01 1.15867174e+00
8.90811026e-01 -5.00387326e-02 -4.53040749e-01 9.68074501e-01
2.33753636e-01 4.03761506e-01 5.11986852e-01 -4.47199494e-01
2.27331236e-01 7.28915036e-01 1.44525930e-01 -9.25762415e-01
8.63768347e-03 -4.94138718e-01 -5.15590250e-01 2.43220791e-01
3.63110244e-01 -7.50761479e-02 -1.31696641e+00 1.71957421e+00
6.61867440e-01 6.88692570e-01 3.00284684e-01 9.24139857e-01
9.93021309e-01 5.31713486e-01 3.57590824e-01 -2.03461662e-01
1.35901511e+00 -6.68959677e-01 -7.17817485e-01 -5.96404374e-01
4.03278530e-01 -4.94820505e-01 8.48807991e-01 1.09752454e-03
-7.43095219e-01 -6.62957907e-01 -1.32279468e+00 1.65844068e-01
-7.98310220e-01 -2.43471906e-01 3.93499166e-01 7.00596392e-01
-5.00387132e-01 3.29283357e-01 -5.22536814e-01 -8.10797215e-01
6.62614882e-01 8.64242315e-02 -8.58719274e-02 -3.42450082e-01
-1.42424750e+00 1.01495302e+00 7.41660297e-01 -3.89356941e-01
-1.29411125e+00 -8.15246999e-01 -7.02292860e-01 -9.44411606e-02
9.93260384e-01 -4.54664946e-01 1.31757092e+00 -8.39164674e-01
-9.91572797e-01 1.14890409e+00 -1.68850541e-01 -6.40609145e-01
5.07954955e-01 -1.23009123e-01 -6.48643613e-01 -8.50350931e-02
3.04007977e-01 4.83654797e-01 1.00427532e+00 -1.54032731e+00
-8.49252164e-01 -4.92161602e-01 -1.34825543e-01 1.26390710e-01
-2.99877852e-01 -3.87757003e-01 -3.79508317e-01 -6.30835831e-01
2.46160507e-01 -7.24549532e-01 2.02133760e-01 1.39563754e-01
-1.95812061e-01 -3.65203261e-01 1.14232016e+00 1.47381380e-01
9.74071741e-01 -2.13025928e+00 -2.30807602e-01 -1.98218375e-01
3.01639318e-01 7.09620237e-01 1.14428736e-01 1.41937479e-01
2.92923786e-02 -1.61805227e-01 -3.29099119e-01 -1.39556509e-02
1.58790220e-02 4.77131158e-01 -6.19336247e-01 2.69974262e-01
3.75246584e-01 8.79589200e-01 -1.42694581e+00 -6.34584785e-01
4.25390363e-01 2.54253238e-01 -8.86375457e-02 2.44674519e-01
-1.66848451e-01 -4.17356752e-02 -2.08829418e-01 8.50310087e-01
7.86215186e-01 -2.35294998e-01 2.89691389e-02 -7.30973631e-02
1.77387312e-01 -1.40028238e-01 -1.20896709e+00 1.39274204e+00
1.33509800e-01 4.97262359e-01 -4.79122490e-01 -1.10865128e+00
1.02674806e+00 2.51708418e-01 5.11782728e-02 -6.02479398e-01
1.89927965e-01 1.90500915e-01 -3.24001729e-01 -5.44433355e-01
2.23181888e-01 -7.91031599e-01 -8.69953334e-02 2.00894043e-01
5.99691868e-01 7.87393153e-02 -8.69484618e-02 3.41570884e-01
9.69948232e-01 -8.57252255e-02 4.52688217e-01 -1.88169166e-01
1.71033844e-01 1.92649648e-01 6.75949991e-01 1.28841293e+00
-4.65647459e-01 4.74870831e-01 3.57190549e-01 -2.71301925e-01
-8.65215421e-01 -1.50276160e+00 -3.90003413e-01 1.35938287e+00
5.29894888e-01 1.51741847e-01 -6.03370786e-01 -7.62843668e-01
1.23735100e-01 1.19164085e+00 -9.57488239e-01 -7.15458453e-01
4.84361909e-02 -8.11400473e-01 4.19212103e-01 5.05017221e-01
3.03652406e-01 -7.36722827e-01 -5.32312453e-01 2.14760885e-01
4.13057171e-02 -9.99874234e-01 5.80803528e-02 4.24918264e-01
-7.15377986e-01 -1.15405428e+00 -6.46386385e-01 -8.99574637e-01
7.42875695e-01 5.58916330e-01 1.01684117e+00 -2.17254966e-01
-4.65778023e-01 2.41961077e-01 -4.58330035e-01 -8.95377994e-01
-5.23187459e-01 -2.88905293e-01 1.54328853e-01 2.62357384e-01
1.06229842e+00 -2.05281690e-01 -3.17325652e-01 5.10451078e-01
-9.02922571e-01 -1.61062285e-01 5.06208777e-01 7.78515279e-01
4.10977423e-01 4.98577580e-02 8.49467993e-01 -1.18724036e+00
3.11355680e-01 -8.60360146e-01 -4.55331922e-01 5.17895699e-01
-7.34132707e-01 -5.26206195e-02 1.10298738e-01 -7.16723800e-01
-1.32883418e+00 -3.59303877e-02 3.98581535e-01 -7.38898456e-01
-4.26881850e-01 -9.60516110e-02 -1.93892177e-02 1.36984035e-01
1.04467773e+00 -6.83467463e-03 -2.28781134e-01 -4.30266351e-01
2.43330732e-01 8.36551964e-01 7.19315469e-01 -3.65133166e-01
9.29181457e-01 6.56457841e-01 -2.48654008e-01 -8.98554564e-01
-1.24906766e+00 -7.55176663e-01 -6.66860878e-01 -3.81837785e-01
8.07977974e-01 -9.50947344e-01 -2.05328658e-01 4.13641959e-01
-8.57016742e-01 -4.95467745e-02 -5.82271874e-01 4.56817925e-01
-2.11060062e-01 1.83230370e-01 -1.56396300e-01 -1.04479647e+00
-4.74678054e-02 -6.41426265e-01 1.09280336e+00 2.75117904e-01
-2.82342136e-01 -8.06711495e-01 3.65693420e-02 2.73167521e-01
2.08207756e-01 3.12117100e-01 8.46208692e-01 -1.02920485e+00
-3.33778203e-01 -4.16984558e-01 -3.00568998e-01 1.92015946e-01
1.12836748e-01 -2.38074929e-01 -1.35926712e+00 -2.75481522e-01
1.93324968e-01 -6.02606952e-01 9.08476949e-01 6.38949126e-02
7.25986242e-01 3.71113904e-02 -6.05957806e-01 -7.99337123e-03
1.37162626e+00 2.86313862e-01 4.15046275e-01 8.57476890e-02
4.05989498e-01 5.78857481e-01 9.44855392e-01 3.85669619e-01
-3.09003424e-02 4.57531035e-01 3.30262691e-01 -3.07651050e-02
-2.42117316e-01 -5.40373266e-01 -7.39821349e-04 3.03890079e-01
2.25585774e-01 -4.64931875e-01 -1.02934849e+00 8.20148170e-01
-1.92588043e+00 -1.08621979e+00 -7.74245262e-02 2.43759203e+00
6.60450518e-01 5.45618057e-01 -7.86519721e-02 2.08707839e-01
1.12730908e+00 -9.86431018e-02 -8.04997563e-01 8.84322971e-02
-6.99378848e-02 1.32481873e-01 2.76424408e-01 4.02936071e-01
-1.19036448e+00 9.44595873e-01 6.08955908e+00 8.88998628e-01
-8.56029689e-01 3.27171326e-01 3.59755635e-01 -2.22232223e-01
-2.01330949e-02 1.04095386e-02 -1.09348381e+00 5.01457512e-01
6.91682518e-01 -3.58812720e-01 -1.35327101e-01 1.06796420e+00
-3.76289368e-01 -2.53686786e-01 -1.23132634e+00 8.85121346e-01
3.81070614e-01 -1.19271398e+00 1.10634558e-01 -1.00472808e-01
7.28999019e-01 -1.38457507e-01 -6.61121979e-02 8.65258455e-01
3.22745502e-01 -6.53666556e-01 6.61758661e-01 5.23459971e-01
6.66320384e-01 -5.73247612e-01 6.96329415e-01 5.83071113e-01
-8.03822458e-01 -1.23174556e-01 -6.44855320e-01 -1.45975333e-02
-9.46892127e-02 6.51570678e-01 -1.24797189e+00 1.16588891e-01
6.26955807e-01 6.01667345e-01 -5.00380814e-01 9.21570718e-01
-1.85490027e-01 7.32251287e-01 -9.87254158e-02 -7.64184892e-02
1.19033679e-01 2.45580778e-01 8.24262023e-01 1.02524650e+00
1.01553559e-01 2.14956358e-01 5.19730859e-02 1.06361210e+00
-4.29720692e-02 -4.00208175e-01 -7.81417012e-01 1.31062835e-01
5.90022445e-01 9.09881532e-01 -8.32457185e-01 -8.31409156e-01
-2.54230350e-01 8.27244341e-01 2.70503849e-01 3.31290066e-01
-9.07953858e-01 -3.49275321e-01 3.49079847e-01 1.85902923e-01
3.86657059e-01 3.62872154e-01 -1.47699490e-01 -1.17018151e+00
-2.39539668e-01 -2.60172755e-01 7.15674698e-01 -6.35236621e-01
-1.59900570e+00 2.10491598e-01 1.38886943e-01 -1.32780933e+00
2.41134800e-02 -3.83345157e-01 -2.98625141e-01 4.06159103e-01
-1.34139562e+00 -9.07766461e-01 -3.30313206e-01 3.82666796e-01
7.41852582e-01 -1.73704118e-01 7.83628404e-01 3.70957345e-01
-4.74570900e-01 5.17366707e-01 1.10564068e-01 1.12764128e-01
8.17062497e-01 -1.22210538e+00 1.71742201e-01 8.55824471e-01
8.03210866e-03 4.49028492e-01 8.83476794e-01 -8.03436339e-01
-1.21479774e+00 -1.36518013e+00 8.50327313e-01 -7.42057502e-01
5.30803084e-01 -5.78861535e-01 -1.29962349e+00 6.68035030e-01
-3.09641987e-01 3.69216114e-01 6.98284924e-01 8.78967568e-02
-4.06135827e-01 -1.39468133e-01 -1.37662578e+00 3.62400979e-01
1.05187798e+00 -4.18308735e-01 -1.20141459e+00 3.15026522e-01
6.96151972e-01 -1.88649565e-01 -4.98910964e-01 5.96158206e-01
3.80373627e-01 -9.00914490e-01 1.04384196e+00 -7.19483078e-01
1.20917484e-01 -3.85161102e-01 -2.61228681e-01 -9.88900304e-01
-3.49783808e-01 -9.64341983e-02 -5.25742650e-01 1.34716082e+00
2.09407955e-01 -5.42699993e-01 6.63019240e-01 5.94503641e-01
4.30632532e-02 -3.97838324e-01 -1.08683193e+00 -1.07826436e+00
-2.84990430e-01 -3.43523353e-01 4.02982086e-01 1.20489848e+00
-1.65831596e-01 7.66462624e-01 -4.47370738e-01 2.75968105e-01
1.02021074e+00 -7.67295947e-03 6.07572198e-01 -1.47890854e+00
-8.21766481e-02 6.18886109e-03 -6.28356636e-01 -6.53712273e-01
-1.56817839e-01 -7.78110385e-01 3.62118334e-01 -1.16198456e+00
4.50959444e-01 -4.17966783e-01 -6.15461409e-01 5.66821098e-01
-2.82261550e-01 3.14837575e-01 -3.21662775e-03 1.13224626e-01
-8.28103960e-01 3.67021918e-01 9.52909291e-01 -1.68380603e-01
-1.49677709e-01 -1.33272475e-02 -7.31005847e-01 6.77920401e-01
6.61886036e-01 -7.60837257e-01 -6.56281829e-01 -2.07216680e-01
-1.97482958e-01 -1.81127951e-01 6.41731083e-01 -1.16214466e+00
3.13211262e-01 -1.05943404e-01 3.87685180e-01 -6.24396443e-01
4.27981853e-01 -7.42733300e-01 -3.29801962e-02 6.25993550e-01
-2.68465370e-01 -8.70034754e-01 1.26138061e-01 1.15536416e+00
9.08433124e-02 -3.72425497e-01 1.18672895e+00 -1.54784724e-01
-1.20937538e+00 5.57617322e-02 -3.74003083e-01 2.31567621e-01
1.68074358e+00 -4.91367012e-01 -4.28361356e-01 9.57419500e-02
-6.70961857e-01 9.91568416e-02 2.59524494e-01 7.67226934e-01
5.53857982e-01 -1.22950244e+00 -5.16686916e-01 2.86308676e-01
6.53913915e-01 1.63359419e-01 2.90737867e-01 4.76979733e-01
1.12148285e-01 3.01628113e-01 -1.46084754e-02 -8.68886530e-01
-1.20916593e+00 9.71515954e-01 2.14930773e-01 2.66461641e-01
-4.20436889e-01 9.86698925e-01 2.44370222e-01 -4.18774515e-01
3.99147511e-01 7.05070347e-02 4.86081205e-02 3.95286560e-01
5.83053052e-01 6.60641015e-01 9.68375355e-02 -4.42646921e-01
-4.36923474e-01 2.06214175e-01 -2.88030416e-01 1.88969284e-01
1.28091979e+00 2.09171642e-02 2.67031521e-01 9.78874445e-01
1.08475792e+00 -5.82294464e-01 -1.11787570e+00 -6.50956154e-01
8.08400363e-02 -6.10281169e-01 1.08781293e-01 -9.71627355e-01
-4.10668105e-01 8.53817940e-01 9.92406666e-01 1.83827162e-01
7.21167564e-01 4.50249523e-01 5.03331184e-01 2.77903974e-01
3.59088242e-01 -1.32104325e+00 3.48323613e-01 3.34988773e-01
4.68109578e-01 -1.53248894e+00 -1.42498821e-01 -4.40637201e-01
-6.08764052e-01 7.00446844e-01 9.00000572e-01 1.94667745e-03
3.97653729e-01 -3.25611085e-02 -1.06080480e-01 -4.18542564e-01
-7.88739204e-01 -4.24891472e-01 3.26274335e-01 6.46909893e-01
-1.58545852e-01 8.35399777e-02 7.12361857e-02 4.33233082e-01
3.30718040e-01 1.77844703e-01 1.61859974e-01 1.19764090e+00
-1.00772107e+00 -5.55083156e-01 -6.75753474e-01 5.77585280e-01
8.26870948e-02 1.95086777e-01 -5.61617076e-01 7.12381959e-01
4.72903430e-01 1.00813735e+00 1.83612883e-01 -2.43146181e-01
3.87061715e-01 3.00584495e-01 3.68053168e-01 -1.00811708e+00
2.16630790e-02 -1.43971369e-01 -2.36698419e-01 7.75686093e-03
-3.59102696e-01 -3.85417372e-01 -1.15296590e+00 1.05382241e-01
-6.49965465e-01 1.27135646e-02 3.24242026e-01 9.38158214e-01
3.06983948e-01 4.48398799e-01 4.18981493e-01 -5.48264503e-01
-6.53557241e-01 -8.54241908e-01 -6.25683784e-01 6.01702154e-01
3.57311487e-01 -1.27017415e+00 -5.78852952e-01 -5.30978814e-02] | [9.761503219604492, 2.2155601978302] |
ae4d9be8-d8d3-41d8-bf6b-9339f241c7b2 | a-systematic-survey-in-geometric-deep | 2306.11768 | null | https://arxiv.org/abs/2306.11768v3 | https://arxiv.org/pdf/2306.11768v3.pdf | A Systematic Survey in Geometric Deep Learning for Structure-based Drug Design | Structure-based drug design (SBDD), which utilizes the three-dimensional geometry of proteins to identify potential drug candidates, is becoming increasingly vital in drug discovery. However, traditional methods based on physiochemical modeling and experts' domain knowledge are time-consuming and laborious. The recent advancements in geometric deep learning, which integrates and processes 3D geometric data, coupled with the availability of accurate protein 3D structure predictions from tools like AlphaFold, have significantly propelled progress in structure-based drug design. In this paper, we systematically review the recent progress of geometric deep learning for structure-based drug design. We start with a brief discussion of the mainstream tasks in structure-based drug design, commonly used 3D protein representations and representative predictive/generative models. Then we delve into detailed reviews for each task (binding site prediction, binding pose generation, \emph{de novo} molecule generation, linker design, and binding affinity prediction), including the problem setup, representative methods, datasets, and evaluation metrics. Finally, we conclude this survey with the current challenges and highlight potential opportunities of geometric deep learning for structure-based drug design.We curate a GitHub repo containing the related papers \url{https://github.com/zaixizhang/Awesome-SBDD}. | ['Enhong Chen', 'Qi Liu', 'Jiaxian Yan', 'Zaixi Zhang'] | 2023-06-20 | null | null | null | null | ['drug-discovery'] | ['medical'] | [-3.35245058e-02 -1.86067045e-01 -5.22547603e-01 -3.22253048e-01
-9.22200978e-01 -7.11309612e-01 1.20832641e-02 6.04609072e-01
2.28555109e-05 1.32190263e+00 2.26756781e-02 -8.88354897e-01
-2.06705481e-01 -6.71367764e-01 -9.89237249e-01 -9.45740283e-01
-1.47980839e-01 8.22747886e-01 -2.63158232e-01 -2.39477709e-01
4.74891335e-01 1.10388517e+00 -6.64862752e-01 4.01913732e-01
9.70869601e-01 6.88512623e-01 2.03510344e-01 1.79244876e-01
-5.57813421e-02 3.72724533e-01 -4.09321547e-01 -2.12226480e-01
-2.25438684e-01 -6.58823252e-01 -9.82908130e-01 -4.50471282e-01
3.08170676e-01 -1.33895308e-01 -2.71841109e-01 4.31866735e-01
1.15551054e+00 3.79911973e-03 9.93356526e-01 -3.36917162e-01
-9.59193051e-01 1.01287536e-01 -3.28596532e-01 1.12771660e-01
4.49588388e-01 3.80780637e-01 1.01483142e+00 -1.36356235e+00
7.68411815e-01 9.78039563e-01 7.16104507e-01 8.25381815e-01
-1.31377566e+00 -7.02325761e-01 -1.02889255e-01 4.23136860e-01
-1.51843834e+00 -1.84441805e-01 6.85800850e-01 -6.35122418e-01
1.60476124e+00 -4.55024242e-02 5.07314980e-01 1.13309646e+00
7.33693957e-01 5.48400760e-01 7.01805413e-01 -1.49443597e-01
4.10601020e-01 -4.65084195e-01 -1.31707594e-01 7.14829385e-01
3.02712768e-01 2.19740272e-01 -3.67343754e-01 -5.73133290e-01
5.94187975e-01 7.55510330e-02 -2.03722611e-01 -7.53794909e-01
-6.30562365e-01 1.17523146e+00 6.57914877e-01 7.68136308e-02
-4.93220627e-01 8.75553712e-02 2.57368028e-01 -5.06750196e-02
2.33632579e-01 8.37707520e-01 -1.00523674e+00 3.03344846e-01
-5.77096283e-01 8.37860107e-01 5.06489813e-01 7.64247060e-01
4.85996962e-01 -1.94801763e-02 1.06394790e-01 5.45482576e-01
5.73504925e-01 2.46373326e-01 -1.44974470e-01 -1.92519322e-01
-5.01465872e-02 3.32760721e-01 7.72135183e-02 -6.42810106e-01
-8.11652184e-01 -2.12477937e-01 -6.51291490e-01 1.78202733e-01
1.81231827e-01 -1.17463015e-01 -9.81204450e-01 1.42731595e+00
4.33541268e-01 -3.08107257e-01 -1.89822748e-01 6.99940085e-01
1.31460631e+00 6.19621813e-01 5.87686718e-01 -3.51325244e-01
1.10695505e+00 -4.81904268e-01 -3.52379650e-01 4.48824912e-01
8.34081948e-01 -9.54810381e-01 3.79879534e-01 3.02466005e-01
-9.05424058e-01 -1.91725180e-01 -1.14531946e+00 -1.85474932e-01
-5.96648037e-01 5.03980927e-03 8.78919125e-01 4.99098033e-01
-5.54674327e-01 9.16197121e-01 -7.53061414e-01 5.38727781e-03
9.54764009e-01 9.25943851e-01 -4.26855534e-01 1.14617862e-01
-1.28484750e+00 1.16842103e+00 3.61093760e-01 -2.75382072e-01
-1.16438413e+00 -1.22862232e+00 -5.66895306e-01 -4.52625424e-01
1.21204361e-01 -9.65985119e-01 1.14422059e+00 -1.25379071e-01
-1.42552173e+00 8.45744133e-01 -2.34228745e-01 -4.07240152e-01
8.87764841e-02 -2.63552312e-02 6.65336195e-03 -1.51490077e-01
-1.49571791e-01 7.00319409e-01 1.13848932e-01 -9.45466876e-01
-9.58501920e-02 -7.23103642e-01 -2.44546875e-01 1.54587135e-01
5.62927127e-01 2.83437800e-02 2.26405308e-01 -6.00505590e-01
-1.99469194e-01 -6.97873712e-01 -7.20384598e-01 -1.82079047e-01
-5.59448361e-01 -4.82589394e-01 4.19665873e-01 -3.72436792e-01
1.12734151e+00 -1.36507392e+00 5.59122145e-01 2.45589957e-01
4.85302955e-01 5.64129949e-01 1.69158373e-02 1.03561580e+00
-7.48818755e-01 1.65655330e-01 4.09904756e-02 3.07892203e-01
-5.13034344e-01 -2.88542926e-01 -2.28053361e-01 7.14858651e-01
1.34517893e-01 1.21801054e+00 -8.38775873e-01 7.96347484e-02
3.13711882e-01 8.27360392e-01 -7.09414899e-01 7.52781034e-02
-8.09415162e-01 9.69581068e-01 -8.87459993e-01 8.95453870e-01
7.30642140e-01 -5.71646512e-01 4.22044903e-01 -2.94802070e-01
-1.26611561e-01 6.42556548e-01 -3.94331872e-01 1.68311703e+00
-1.89406686e-02 -7.75303692e-02 -6.00373983e-01 -8.94554734e-01
1.03120625e+00 3.78803611e-01 8.54296207e-01 -6.23554230e-01
1.61394760e-01 5.82336783e-01 1.29034191e-01 -2.17320129e-01
-4.97488469e-01 -3.84559870e-01 2.55879790e-01 1.44498810e-01
1.79234613e-02 4.15923782e-02 -2.38510698e-01 -1.74274504e-01
7.81079888e-01 5.80027282e-01 5.34737825e-01 -1.75990537e-01
6.53398514e-01 3.44924122e-01 3.64449650e-01 -2.07232255e-02
5.09155355e-02 3.13569367e-01 4.23807889e-01 -1.23847461e+00
-1.32467687e+00 -9.10121918e-01 -5.24561882e-01 8.76353383e-01
-1.57462865e-01 -7.53917754e-01 -6.81702316e-01 -6.72174215e-01
1.72735959e-01 1.81211427e-01 -7.52889037e-01 -1.80473477e-01
-8.52591813e-01 -1.31974471e+00 3.45325500e-01 2.14351624e-01
-1.95694834e-01 -9.60960984e-01 9.99279171e-02 6.30744576e-01
3.65054429e-01 -2.51888543e-01 -2.29256898e-01 4.25213605e-01
-1.00584805e+00 -1.47832620e+00 -9.02889967e-01 -9.15644765e-01
3.09317321e-01 -2.71846987e-02 1.09774244e+00 -1.24348223e-01
-7.19492912e-01 -5.17414570e-01 -5.95309585e-02 -7.19715357e-01
-2.24932805e-01 5.77729158e-02 2.11106598e-01 -7.26593912e-01
8.25069189e-01 -7.18238890e-01 -1.07094991e+00 1.30143210e-01
-4.26155597e-01 -9.17196944e-02 6.17496908e-01 8.21374536e-01
1.21150255e+00 -5.04381537e-01 8.48078907e-01 -1.03802037e+00
6.36382937e-01 -4.43450242e-01 -5.46868205e-01 8.62170011e-02
-7.18794167e-01 2.79811710e-01 5.18454969e-01 -3.78388390e-02
-4.69905704e-01 3.85344326e-01 -9.26329255e-01 -8.65993276e-02
-2.70483136e-01 5.68754256e-01 -3.27258140e-01 -3.80804211e-01
1.05693483e+00 1.48455739e-01 2.81421691e-01 -8.01405847e-01
3.32258224e-01 3.42270315e-01 -4.17751580e-01 -6.09050393e-01
3.38411391e-01 9.75040272e-02 3.98566604e-01 -6.55556381e-01
-6.92024291e-01 -1.79487914e-01 -6.29777789e-01 5.11827171e-01
9.80703652e-01 -6.93466127e-01 -1.28558505e+00 2.02848837e-01
-1.42961133e+00 -1.53294563e-01 2.01822326e-01 4.13767606e-01
-8.33072007e-01 3.69413346e-01 -5.64929664e-01 -1.28687859e-01
-8.15623760e-01 -1.68533933e+00 1.03843117e+00 -5.03172465e-02
-5.23406625e-01 -1.07272887e+00 5.24900794e-01 5.15222490e-01
8.27011988e-02 8.24682772e-01 1.55311370e+00 -8.50594282e-01
-7.56259739e-01 -8.47836211e-02 8.45686868e-02 -1.37189195e-01
2.99274713e-01 -2.82887846e-01 -6.87292278e-01 -8.94080400e-02
-4.74154443e-01 -3.85809451e-01 7.20149338e-01 9.31721807e-01
1.16896451e+00 -1.84624225e-01 -7.57455885e-01 7.24978864e-01
1.41948020e+00 8.90313208e-01 4.93673265e-01 1.92040995e-01
9.42702234e-01 3.99033219e-01 4.78617698e-01 4.63058710e-01
-5.82320057e-02 8.63069236e-01 4.52283472e-01 -3.77037436e-01
-9.27577391e-02 -2.69866914e-01 -2.25454852e-01 4.81974632e-02
-5.31078756e-01 -2.39586502e-01 -1.02138531e+00 -7.91033208e-02
-1.65739024e+00 -7.85437346e-01 -1.47030294e-01 1.98201811e+00
1.34108508e+00 -4.34715539e-01 4.29564804e-01 -3.26252073e-01
4.21648830e-01 -1.76286086e-01 -1.11301243e+00 -3.79835755e-01
-1.80040747e-01 9.17608619e-01 4.72262144e-01 5.92602432e-01
-1.21443415e+00 1.13783967e+00 6.92194843e+00 1.10078979e+00
-1.09550989e+00 -2.85585821e-01 1.04861808e+00 3.74227948e-02
-3.28137219e-01 -2.47460395e-01 -9.98590291e-01 1.89536005e-01
9.14333224e-01 -1.14587201e-02 -2.79479362e-02 8.08588564e-01
6.98723674e-01 3.99832278e-01 -1.30306542e+00 1.00969851e+00
-4.29914594e-01 -2.45256758e+00 4.89778578e-01 4.30132180e-01
6.76025212e-01 1.30704060e-01 3.58860999e-01 -2.58582324e-01
8.94417241e-02 -1.73575926e+00 -6.31482974e-02 2.56239951e-01
1.02639842e+00 -9.77461636e-01 4.67811346e-01 -1.18230544e-01
-9.45278704e-01 2.74552286e-01 -3.47679555e-01 3.58318746e-01
7.55421966e-02 3.86856198e-01 -8.62307429e-01 5.59824824e-01
3.06185633e-01 9.59294319e-01 -1.07664531e-02 1.12810516e+00
1.99997723e-02 4.92180660e-02 1.18917942e-01 -2.40851730e-01
3.19759518e-01 -2.36194298e-01 1.72308847e-01 9.46498036e-01
-1.62539080e-01 3.89044017e-01 3.17318231e-01 1.00889099e+00
-6.45063296e-02 4.88322288e-01 -4.50536430e-01 -8.34349915e-02
3.66465986e-01 5.99739432e-01 -2.22949088e-01 1.80392772e-01
-4.01993871e-01 5.22675455e-01 3.26254629e-02 3.25146019e-01
-9.82381284e-01 -3.72635901e-01 1.15569007e+00 4.86342818e-01
1.49434611e-01 -1.92082033e-01 -1.27430364e-01 -5.16672909e-01
-4.71850574e-01 -1.23501062e+00 2.50441700e-01 -4.02817845e-01
-1.18742347e+00 2.89857149e-01 -1.50729120e-01 -8.25367689e-01
3.51483613e-01 -1.10403109e+00 -3.14756393e-01 1.21181095e+00
-1.11082554e+00 -1.07312989e+00 5.52952111e-01 2.66549081e-01
5.64870298e-01 -3.08158189e-01 1.17202711e+00 4.26632017e-01
-5.08278430e-01 5.82122147e-01 5.14417529e-01 -2.66661108e-01
7.58886933e-01 -1.06095684e+00 5.88420093e-01 -4.53657359e-01
-3.75414193e-01 9.36152577e-01 7.30127633e-01 -8.60404015e-01
-1.59142435e+00 -1.02846766e+00 7.62335896e-01 -6.72705770e-01
4.39014852e-01 -2.31686145e-01 -8.39465320e-01 2.64535129e-01
-1.38162658e-01 -3.89825970e-01 1.47866166e+00 -9.89854857e-02
-2.65435070e-01 3.57993960e-01 -1.07865393e+00 4.46368754e-01
1.11999631e+00 -1.52937338e-01 -7.31289908e-02 9.82391357e-01
7.03933358e-01 -6.07030630e-01 -1.16786480e+00 6.59534276e-01
4.84207898e-01 -7.28314519e-01 1.39009094e+00 -1.43127990e+00
2.05850586e-01 -3.12340707e-01 1.85889795e-01 -9.15288985e-01
-7.61185944e-01 -6.72178924e-01 -1.33630157e-01 3.44186351e-02
8.88156652e-01 -2.78740257e-01 1.27810824e+00 1.95157036e-01
-3.30698133e-01 -1.66311908e+00 -9.08981025e-01 -5.00844121e-01
8.75412405e-01 1.31759629e-01 5.55250883e-01 6.37761176e-01
6.98271766e-02 9.04394507e-01 -2.88316756e-01 -3.32873702e-01
3.85710090e-01 2.71499962e-01 4.48881954e-01 -1.26178074e+00
-3.81660581e-01 -4.99453962e-01 -3.85475099e-01 -1.22432840e+00
-4.01568860e-02 -1.02657640e+00 -6.40731096e-01 -1.78818476e+00
2.40714684e-01 -2.45693520e-01 1.56111978e-02 3.50245595e-01
1.53845519e-01 -7.67744705e-02 -5.71862340e-01 1.50380179e-01
-2.84580320e-01 6.76912904e-01 1.68315542e+00 -3.00607413e-01
-4.80730683e-01 2.79719949e-01 -8.47516119e-01 2.61926144e-01
9.27925050e-01 -2.05247790e-01 -3.81248653e-01 2.78072320e-02
4.93044198e-01 -8.53603333e-02 -1.46878198e-01 -2.71541953e-01
-2.85188556e-01 -4.98595953e-01 8.22519898e-01 -8.23678315e-01
3.67761135e-01 -5.46596169e-01 2.92936146e-01 6.94300532e-01
-1.58852488e-02 -7.02254102e-02 4.09665316e-01 6.22095048e-01
1.69980228e-01 1.57519341e-01 1.14095092e+00 -3.14318091e-01
-2.07529306e-01 1.08196986e+00 -6.41990960e-01 -2.73326069e-01
1.14857662e+00 -4.10237193e-01 -8.85251909e-02 1.15376920e-01
-1.18217695e+00 -1.29512534e-01 3.08387607e-01 1.01108126e-01
9.56248403e-01 -1.25979722e+00 -5.69725096e-01 -7.74406120e-02
2.58761764e-01 4.98244166e-02 2.95976758e-01 7.44229317e-01
-1.01382387e+00 1.09681106e+00 -5.85461371e-02 -2.99827069e-01
-1.36577773e+00 8.58525872e-01 8.72851908e-01 -4.53480743e-02
-1.88418940e-01 1.00282657e+00 4.71468598e-01 -2.11989239e-01
1.66333750e-01 4.22353577e-03 -2.60719180e-01 -1.63865015e-01
4.93460655e-01 3.71079817e-02 3.93085569e-01 -4.95804906e-01
-7.34667718e-01 8.60016227e-01 -5.29942632e-01 8.23638499e-01
1.67062581e+00 5.20655453e-01 -1.09621741e-01 -3.38561416e-01
1.32702291e+00 -4.23507035e-01 -8.45361114e-01 -6.57262728e-02
1.59994401e-02 -4.21803556e-02 -1.94367096e-01 -1.00362003e+00
-5.68176508e-01 9.02690053e-01 6.48687661e-01 -6.83629215e-01
5.43918073e-01 3.80979359e-01 9.33535874e-01 5.39929450e-01
4.18256581e-01 -6.44050181e-01 1.92999884e-01 6.28553808e-01
1.02012408e+00 -1.17188609e+00 4.46799397e-01 -3.92158538e-01
-1.95098177e-01 1.21538246e+00 4.06122297e-01 6.08275309e-02
9.54903603e-01 -1.77855745e-01 -1.39415801e-01 -7.93315709e-01
-5.88704348e-01 9.65626463e-02 3.23496401e-01 8.46218050e-01
1.22371197e+00 -1.54448226e-02 -4.52854455e-01 4.71810162e-01
2.03088492e-01 -1.12611167e-01 -1.75632194e-01 9.75857437e-01
-7.07522929e-01 -2.01976895e+00 -2.19710171e-01 9.39222425e-02
-6.11502409e-01 -4.58538711e-01 -1.17871499e+00 5.04152954e-01
1.85755581e-01 5.35240591e-01 -4.09817517e-01 -1.87235177e-01
4.15299177e-01 -4.33718041e-02 9.91949439e-01 -8.00872147e-01
-3.35888684e-01 3.38366657e-01 -8.46024975e-02 -2.54460543e-01
-1.68102831e-01 -1.16687000e-01 -1.52074158e+00 -6.20728076e-01
-4.39322412e-01 3.44735026e-01 5.39208055e-01 6.59390628e-01
9.33276951e-01 1.91321746e-01 4.89067346e-01 -6.33204401e-01
-3.21066499e-01 -5.64587474e-01 -2.84193695e-01 -2.93762207e-01
1.86871991e-01 -7.77194917e-01 2.79298395e-01 5.52629083e-02] | [4.940145015716553, 5.805591106414795] |
6d300890-1f8d-4e0b-b0f6-100e26d41839 | sociallight-distributed-cooperation-learning | 2305.16145 | null | https://arxiv.org/abs/2305.16145v1 | https://arxiv.org/pdf/2305.16145v1.pdf | SocialLight: Distributed Cooperation Learning towards Network-Wide Traffic Signal Control | Many recent works have turned to multi-agent reinforcement learning (MARL) for adaptive traffic signal control to optimize the travel time of vehicles over large urban networks. However, achieving effective and scalable cooperation among junctions (agents) remains an open challenge, as existing methods often rely on extensive, non-generalizable reward shaping or on non-scalable centralized learning. To address these problems, we propose a new MARL method for traffic signal control, SocialLight, which learns cooperative traffic control policies by distributedly estimating the individual marginal contribution of agents on their local neighborhood. SocialLight relies on the Asynchronous Actor Critic (A3C) framework, and makes learning scalable by learning a locally-centralized critic conditioned over the states and actions of neighboring agents, used by agents to estimate individual contributions by counterfactual reasoning. We further introduce important modifications to the advantage calculation that help stabilize policy updates. These modifications decouple the impact of the neighbors' actions on the computed advantages, thereby reducing the variance in the gradient updates. We benchmark our trained network against state-of-the-art traffic signal control methods on standard benchmarks in two traffic simulators, SUMO and CityFlow. Our results show that SocialLight exhibits improved scalability to larger road networks and better performance across usual traffic metrics. | ['Guillaume Sartoretti', 'Mehul Damani', 'Yifeng Zhang', 'Harsh Goel'] | 2023-04-20 | null | null | null | null | ['multi-agent-reinforcement-learning'] | ['methodology'] | [-4.01763260e-01 2.36736864e-01 -4.36362565e-01 -2.15184122e-01
-9.27613735e-01 -3.25553656e-01 6.82942092e-01 -9.65987798e-03
-7.12622285e-01 1.38112056e+00 6.23086579e-02 -6.06438518e-01
-2.36620858e-01 -9.68986809e-01 -7.25621045e-01 -8.92497599e-01
-3.81944865e-01 7.41211236e-01 7.15942383e-01 -6.34645343e-01
-3.31853367e-02 5.25769591e-01 -1.32662392e+00 -4.20415223e-01
8.95267248e-01 6.67068839e-01 -3.07990789e-01 9.06088889e-01
3.71361002e-02 1.20150411e+00 -6.28769577e-01 -2.86933720e-01
3.52751613e-01 -2.00243428e-01 -2.94839442e-01 -1.89163238e-01
2.45693833e-01 -4.45618033e-01 -5.09719789e-01 6.95012808e-01
5.41688621e-01 4.32611853e-01 5.29615521e-01 -1.86546218e+00
1.08973861e-01 8.98597538e-01 -6.25190675e-01 2.86405623e-01
-2.51149237e-01 7.84169793e-01 9.37790394e-01 8.88571218e-02
3.79909903e-01 1.55972326e+00 5.11395156e-01 5.91240585e-01
-1.29148161e+00 -7.95701027e-01 6.98394299e-01 3.77985418e-01
-9.91657376e-01 -5.97373307e-01 6.72198772e-01 -4.35853489e-02
8.14805329e-01 2.00556502e-01 6.16569817e-01 7.27389812e-01
1.57289818e-01 8.38461697e-01 8.94570708e-01 7.32765272e-02
6.41147912e-01 -4.33315746e-02 -2.83292443e-01 5.80742896e-01
1.59876168e-01 3.71414691e-01 -1.00112133e-01 -3.39825749e-01
5.93035221e-01 -4.31514353e-01 3.16336989e-01 -4.56681639e-01
-9.24485564e-01 8.97612691e-01 4.44314361e-01 -3.39538187e-01
-8.18768919e-01 9.23259437e-01 4.62097287e-01 5.36864996e-01
3.80312115e-01 -2.97792978e-03 -3.01932544e-01 -3.74509782e-01
-4.58623290e-01 5.42052865e-01 8.62808585e-01 6.74206734e-01
1.02056766e+00 3.87470812e-01 -4.80824977e-01 5.14190853e-01
3.62387002e-01 1.01128757e+00 -3.86674762e-01 -1.83140922e+00
6.59024596e-01 1.35493875e-01 4.55231935e-01 -7.38817930e-01
-3.53706449e-01 -2.73246795e-01 -4.80580240e-01 7.56230414e-01
6.29959285e-01 -9.05618668e-01 -4.07450467e-01 1.84293711e+00
5.63491046e-01 5.64490557e-01 6.32364377e-02 7.14133263e-01
-6.99852556e-02 6.31607294e-01 3.33672702e-01 -7.95710012e-02
6.04352951e-01 -1.09952939e+00 -3.25621873e-01 -5.30208461e-02
7.17266798e-01 -2.64880270e-01 4.47109014e-01 -4.44577448e-02
-1.34839380e+00 3.90512533e-02 -6.42915845e-01 5.58502376e-01
-3.14944565e-01 -2.59681433e-01 3.96914810e-01 8.36270869e-01
-1.08660507e+00 3.87339503e-01 -7.99940467e-01 7.08053559e-02
6.60314023e-01 5.15193641e-01 3.65291119e-01 1.55964881e-01
-1.23824644e+00 9.93937612e-01 -3.08753163e-01 -1.43994719e-01
-1.25571740e+00 -1.03831971e+00 -5.59135973e-01 7.94081166e-02
9.99169409e-01 -6.49740577e-01 1.49550807e+00 -8.72297287e-01
-2.18608356e+00 1.00762174e-02 1.27849415e-01 -9.12844419e-01
9.09341037e-01 1.84028313e-01 -3.64133090e-01 2.78432481e-02
3.86597365e-01 7.36563623e-01 6.39463127e-01 -1.24972939e+00
-1.06506002e+00 1.81079701e-01 4.09246624e-01 2.29886159e-01
1.62211955e-01 -2.63797224e-01 -1.43767849e-01 -2.31697977e-01
-1.05208611e+00 -1.11208355e+00 -8.71586502e-01 1.55188367e-01
-7.85476193e-02 -5.18773854e-01 7.19058335e-01 -3.51703912e-02
9.32634950e-01 -1.81749606e+00 -1.79350838e-01 6.47685766e-01
2.68259376e-01 3.61635029e-01 -4.32850569e-01 4.87304032e-01
5.90603411e-01 -9.57101882e-02 9.87839773e-02 -1.70368433e-01
3.72922003e-01 3.69332224e-01 -1.41185910e-01 5.41902900e-01
-1.12827485e-02 8.59303474e-01 -1.24504316e+00 -5.19649148e-01
6.07884824e-01 2.31868088e-01 -7.35786736e-01 -7.26715103e-02
-3.69824469e-01 4.05139595e-01 -8.64120483e-01 1.49480730e-01
5.20907164e-01 9.03308317e-02 1.32383689e-01 3.06840986e-01
-4.11397129e-01 1.02503195e-01 -1.36786306e+00 8.23850274e-01
-8.06695759e-01 6.66387737e-01 5.80052555e-01 -1.03905189e+00
4.82399642e-01 2.15174183e-01 7.88394094e-01 -9.36108947e-01
1.45803034e-01 2.04139128e-01 4.40434702e-02 -2.30019793e-01
1.55039385e-01 2.72124112e-01 3.29208523e-02 5.54674506e-01
-4.65501070e-01 -1.03756841e-02 6.16268098e-01 3.32464248e-01
1.34010065e+00 -2.73126423e-01 -1.72018573e-01 -1.03200927e-01
6.87748849e-01 4.11098227e-02 6.43769026e-01 1.10781848e+00
-6.17703557e-01 -4.66834188e-01 8.45363498e-01 -3.09433609e-01
-9.21347439e-01 -1.11462522e+00 6.02865279e-01 1.32625890e+00
2.28167921e-01 7.29786009e-02 -5.46519935e-01 -7.62991428e-01
4.51716989e-01 8.86142492e-01 -4.49182838e-01 -1.53466249e-02
-9.39240336e-01 -4.93340433e-01 6.24148369e-01 3.31327558e-01
5.81266940e-01 -8.10516357e-01 -4.60395455e-01 6.26840413e-01
3.15575868e-01 -1.05999589e+00 -5.47818780e-01 -4.00958151e-01
-2.27669507e-01 -1.14365816e+00 -5.19303799e-01 -2.15346605e-01
5.76832950e-01 3.44960064e-01 8.32624078e-01 1.17481157e-01
1.84927478e-01 7.41444647e-01 2.10560068e-01 -4.10831362e-01
-6.61629975e-01 3.03198993e-01 -4.52042464e-03 2.89153218e-01
-1.66022167e-01 -6.68700755e-01 -7.39780724e-01 6.34559035e-01
-3.53177279e-01 -1.51275486e-01 5.68545222e-01 4.71526295e-01
3.25071752e-01 -1.58586875e-01 1.05215681e+00 -5.88881254e-01
9.08137858e-01 -6.18617237e-01 -1.21894813e+00 1.08649895e-01
-4.90324646e-01 3.60622585e-01 8.73464763e-01 -3.23750585e-01
-1.12864113e+00 -1.26779780e-01 -1.74365155e-02 -8.48388374e-02
8.76174495e-02 -3.27091385e-03 1.57837987e-01 -3.46896827e-01
4.16164547e-01 -1.48698047e-01 4.37539846e-01 2.29398146e-01
7.78972745e-01 3.73771161e-01 7.01612458e-02 -9.17897761e-01
9.40036833e-01 6.13089859e-01 4.72519159e-01 -7.35231459e-01
-4.08247054e-01 -1.28861472e-01 7.72417262e-02 -7.23572493e-01
4.99877632e-01 -9.65470552e-01 -1.70759904e+00 2.28605688e-01
-8.81772280e-01 -9.46438849e-01 -6.30395353e-01 4.24188763e-01
-7.33688235e-01 2.27198489e-02 -3.83578181e-01 -9.81522143e-01
-6.78311586e-02 -1.17282093e+00 7.18942046e-01 2.70108044e-01
3.30121189e-01 -1.28566539e+00 3.77346665e-01 1.42131925e-01
1.07056189e+00 2.11012915e-01 5.63117683e-01 -1.64545536e-01
-9.85863268e-01 6.44849166e-02 -1.49897993e-01 4.52624336e-02
-2.65508622e-01 1.41840100e-01 -4.85821635e-01 -2.20856398e-01
-9.47836936e-01 -2.72404939e-01 7.39295185e-01 6.53629959e-01
6.59887612e-01 -4.04131651e-01 -6.32298768e-01 -4.56837788e-02
1.11976111e+00 5.00734011e-03 4.63173687e-01 3.96086901e-01
3.80158126e-01 4.48355138e-01 4.93020058e-01 5.42578399e-01
9.31569517e-01 7.74586260e-01 5.94354451e-01 -5.78894503e-02
-1.54303610e-01 -1.31765604e-01 6.92980528e-01 3.61882359e-01
-1.22622997e-01 -2.14537278e-01 -5.63721776e-01 5.61108828e-01
-2.24146557e+00 -1.23483777e+00 -1.41231567e-01 2.02030015e+00
6.76839888e-01 3.82085472e-01 6.23029947e-01 -2.95728266e-01
6.07513130e-01 2.47031271e-01 -9.20115530e-01 -4.70534682e-01
4.34019975e-02 -1.24147698e-01 1.31587279e+00 9.64101732e-01
-6.05265677e-01 9.65778410e-01 6.51816845e+00 1.07987726e+00
-9.14980173e-01 1.81231305e-01 5.69763899e-01 -3.46241504e-01
-1.66185975e-01 6.88145682e-02 -7.67212391e-01 3.84840816e-01
1.34433031e+00 -5.10841072e-01 8.72598708e-01 7.47347414e-01
9.80622888e-01 -1.86646402e-01 -6.44699216e-01 4.54978913e-01
-6.35208845e-01 -1.63167155e+00 -2.50358462e-01 1.39755562e-01
9.04004514e-01 5.88286877e-01 -5.13810813e-02 6.51279330e-01
1.30056643e+00 -4.54559922e-01 4.76952732e-01 5.25852919e-01
1.82516500e-01 -1.12415552e+00 3.92852098e-01 1.95413277e-01
-1.36110318e+00 -4.04529601e-01 -2.73561366e-02 -1.46856025e-01
5.65217316e-01 3.58149737e-01 -6.83905780e-01 1.25392407e-01
1.75495237e-01 6.37555897e-01 -2.09767178e-01 1.00258422e+00
-2.96229124e-01 8.37068319e-01 -5.29387951e-01 -4.99676645e-01
6.64595366e-01 -1.64470434e-01 8.08316648e-01 7.87069201e-01
-2.87180632e-01 -3.10807407e-01 5.35968125e-01 7.44684696e-01
-1.47232920e-01 -1.07313760e-01 -4.20252591e-01 5.17382145e-01
6.45101786e-01 1.26305056e+00 -4.25189912e-01 -4.93867129e-01
-4.38646764e-01 9.08642411e-02 3.80033731e-01 7.42834747e-01
-1.32059872e+00 -3.53713751e-01 1.32225657e+00 1.88808858e-01
3.46466243e-01 -3.26870233e-01 9.37283561e-02 -5.89078069e-01
-2.02419370e-01 -6.15236938e-01 1.11276850e-01 -3.50159705e-01
-1.05431044e+00 -8.16809665e-03 3.31044793e-02 -1.06526971e+00
-2.82630056e-01 -2.46755734e-01 -9.81374085e-01 2.65060902e-01
-2.02705336e+00 -7.35538244e-01 2.80536145e-01 6.57675207e-01
4.27897215e-01 -3.75744194e-01 9.64100286e-02 5.36969185e-01
-8.30996871e-01 5.32713115e-01 2.94364572e-01 4.07060087e-02
4.79169816e-01 -1.14424348e+00 3.30554932e-01 4.29831445e-01
-5.08435071e-01 -2.18403459e-01 6.86416805e-01 -3.73164982e-01
-1.41524589e+00 -1.22670245e+00 3.47278804e-01 -1.73650458e-01
1.02469289e+00 -1.50555102e-02 -1.86779723e-01 4.71382797e-01
4.37244326e-01 2.18833581e-01 -2.41480954e-02 -1.86172187e-01
1.78674057e-01 -7.63607383e-01 -1.03442001e+00 1.02296972e+00
7.75249422e-01 -5.59173077e-02 3.28639835e-01 2.80991882e-01
6.52537644e-01 -1.13272615e-01 -6.12462938e-01 -2.66853243e-01
3.54991049e-01 -5.70529222e-01 7.83861041e-01 -6.65375113e-01
-3.86644989e-01 -4.66759235e-01 1.40508994e-01 -1.73350644e+00
-8.34058598e-02 -1.11128581e+00 2.71977601e-03 9.65207636e-01
6.02385640e-01 -1.05996442e+00 9.51480985e-01 6.82537556e-01
-5.38106374e-02 -5.76566458e-01 -1.45996428e+00 -7.93712258e-01
1.82934254e-01 -4.65051472e-01 7.39061952e-01 3.89140576e-01
-2.43054897e-01 4.07995045e-01 -5.29202819e-01 1.68008670e-01
1.02565551e+00 -3.58714759e-01 1.14667702e+00 -8.42164934e-01
-4.09556068e-02 -8.34478378e-01 -1.70922250e-01 -1.11218643e+00
6.67043507e-01 -5.82688332e-01 9.31299999e-02 -1.32038701e+00
-3.40917885e-01 -7.15060711e-01 -1.71795134e-02 4.52778786e-01
1.44058034e-01 -2.20204458e-01 1.95984349e-01 -2.82606483e-01
-1.36837816e+00 8.84846509e-01 1.28736985e+00 -3.67900223e-01
-3.72872680e-01 4.74040776e-01 -4.02834266e-01 4.82533455e-01
1.06754029e+00 -6.60564065e-01 -4.09829557e-01 -3.20424587e-01
2.37875864e-01 3.18742305e-01 4.86461252e-01 -1.02976394e+00
6.37143731e-01 -7.38346517e-01 -4.42717910e-01 -3.86065453e-01
9.98170748e-02 -8.92074108e-01 -1.65430889e-01 6.62801623e-01
-5.92164755e-01 4.97491993e-02 8.54206830e-02 9.33718383e-01
1.54371336e-01 3.93189639e-01 8.11577797e-01 6.10425770e-02
-5.57940066e-01 4.30063665e-01 -1.12040770e+00 4.35486406e-01
1.30972862e+00 2.27610677e-01 -5.15967309e-01 -8.47497046e-01
-5.08427799e-01 1.16248047e+00 -4.44965363e-02 2.06916004e-01
2.80197501e-01 -1.37888718e+00 -8.95036638e-01 -2.49675304e-01
-3.25831652e-01 -4.92357761e-01 1.09405853e-01 1.09230506e+00
-2.56053239e-01 2.94083297e-01 -1.24663211e-01 -3.27182263e-01
-7.87878692e-01 1.34409562e-01 8.83845985e-01 -4.63580757e-01
-2.33999848e-01 1.01947263e-01 -4.21611696e-01 -5.34800112e-01
2.50162750e-01 -2.55078673e-01 1.17218792e-01 -2.23622546e-01
3.19661587e-01 1.18791246e+00 -3.15067261e-01 -4.76803660e-01
-2.51772195e-01 3.05180252e-01 1.83165092e-02 -3.28683496e-01
1.26404667e+00 -3.92179012e-01 2.91262656e-01 5.03409356e-02
7.71227002e-01 -2.89450437e-02 -1.70443404e+00 -3.14633280e-01
-2.16797993e-01 -1.70322508e-01 2.19993144e-01 -6.04918361e-01
-1.43256903e+00 3.78941745e-01 4.46887016e-01 2.39911810e-01
5.69061399e-01 -2.39477411e-01 8.03413212e-01 7.47399986e-01
4.68914300e-01 -1.56576312e+00 -9.18457136e-02 6.15037441e-01
3.74767274e-01 -1.30082214e+00 -6.29144683e-02 -8.77936408e-02
-7.01520562e-01 8.24652076e-01 7.59071529e-01 -3.44941854e-01
8.40687752e-01 2.90504426e-01 8.31680745e-02 2.93374006e-02
-1.23953581e+00 -5.73738754e-01 -4.69905257e-01 5.44297338e-01
-2.15005293e-01 2.92174548e-01 -1.07291080e-01 -1.98369175e-01
4.14804965e-01 -3.61817842e-03 6.57816291e-01 8.38661909e-01
-6.67373061e-01 -1.13682914e+00 -2.12025687e-01 4.47458833e-01
-5.54485321e-02 4.32608813e-01 6.54990375e-02 8.69023502e-01
-2.28344023e-01 1.16060674e+00 7.15821311e-02 1.52105436e-01
5.24150968e-01 -6.15224898e-01 7.51241371e-02 1.12613000e-01
-4.52848375e-01 -2.63771594e-01 3.92371893e-01 -7.81196058e-01
-5.20040751e-01 -6.98108137e-01 -1.47456586e+00 -8.25962126e-01
-2.36663851e-03 3.76249909e-01 4.68114823e-01 1.00850165e+00
6.31247640e-01 5.66778839e-01 1.06219745e+00 -9.26218867e-01
-7.34161735e-01 -3.71410698e-01 -2.56850719e-01 1.89484674e-02
5.71067870e-01 -7.61071801e-01 -3.76107842e-01 -6.92016602e-01] | [5.218703269958496, 1.4901231527328491] |
57d9643b-4bcc-49b3-a125-dc4233b33aa0 | on-parsing-as-tagging | 2211.07344 | null | https://arxiv.org/abs/2211.07344v2 | https://arxiv.org/pdf/2211.07344v2.pdf | On Parsing as Tagging | There have been many proposals to reduce constituency parsing to tagging in the literature. To better understand what these approaches have in common, we cast several existing proposals into a unifying pipeline consisting of three steps: linearization, learning, and decoding. In particular, we show how to reduce tetratagging, a state-of-the-art constituency tagger, to shift--reduce parsing by performing a right-corner transformation on the grammar and making a specific independence assumption. Furthermore, we empirically evaluate our taxonomy of tagging pipelines with different choices of linearizers, learners, and decoders. Based on the results in English and a set of 8 typologically diverse languages, we conclude that the linearization of the derivation tree and its alignment with the input sequence is the most critical factor in achieving accurate taggers. | ['Ryan Cotterell', 'Afra Amini'] | 2022-11-14 | null | null | null | null | ['constituency-parsing'] | ['natural-language-processing'] | [ 2.05393165e-01 4.19626355e-01 -3.22303355e-01 -5.83117485e-01
-1.22816145e+00 -1.23921967e+00 3.35744888e-01 3.08492690e-01
-4.58534241e-01 6.53843641e-01 6.78839803e-01 -8.43094051e-01
4.02983874e-01 -5.40089965e-01 -5.82633972e-01 -1.71802193e-01
2.86831737e-01 5.65580606e-01 3.62329364e-01 -3.14395159e-01
1.19608708e-01 -7.09346533e-02 -9.55090702e-01 5.70532560e-01
9.34908330e-01 1.76948413e-01 2.60070652e-01 6.20609403e-01
-3.40151995e-01 7.32425332e-01 -4.15249318e-01 -9.77936149e-01
4.10752073e-02 -2.81628251e-01 -1.02516472e+00 -6.10471070e-01
5.19797921e-01 7.63219967e-02 1.75014213e-01 1.04667270e+00
3.66762310e-01 -3.68345678e-01 3.85388285e-01 -5.30567765e-01
-5.44033766e-01 1.51658523e+00 -1.90993026e-01 3.10284376e-01
3.06572527e-01 -2.97807992e-01 1.46274590e+00 -5.09692669e-01
7.94560432e-01 1.20910585e+00 1.07113755e+00 6.88808858e-01
-1.22723651e+00 -6.29026830e-01 3.72442603e-01 -3.80617827e-01
-9.38387871e-01 -5.50504088e-01 4.89952952e-01 -4.90642726e-01
1.30990505e+00 1.14229649e-01 3.39121431e-01 7.51239479e-01
2.97284514e-01 5.86156070e-01 1.12510371e+00 -9.86570954e-01
-1.20397359e-01 -1.23914368e-02 6.33709848e-01 7.93055177e-01
4.49366927e-01 -2.38030180e-02 -4.35411155e-01 -2.09698960e-01
3.11320007e-01 -7.32026160e-01 1.65474325e-01 -5.16161993e-02
-9.43778515e-01 8.91261458e-01 -8.45568702e-02 6.28246427e-01
1.35469154e-01 2.81797796e-02 5.20842314e-01 1.79065228e-01
6.33851171e-01 5.90484440e-01 -9.22157407e-01 1.39250094e-03
-6.97303295e-01 1.29274391e-02 1.03250635e+00 1.08736610e+00
7.03338861e-01 -2.83324033e-01 1.11152671e-01 5.91223300e-01
3.10043693e-01 2.54704446e-01 2.71220803e-01 -8.42316568e-01
9.17223811e-01 4.74104494e-01 -1.05933227e-01 -2.25343958e-01
-5.34141183e-01 -4.68837678e-01 4.88022603e-02 -3.57303023e-01
8.58394027e-01 -5.66143811e-01 -7.68766820e-01 2.16056180e+00
2.50714213e-01 -2.35620245e-01 2.63154924e-01 2.71801770e-01
6.25663877e-01 5.40440917e-01 7.61887074e-01 -1.93175167e-01
1.71114802e+00 -6.89846754e-01 -6.03536785e-01 -6.55046403e-01
1.39953065e+00 -9.93126810e-01 1.03424585e+00 5.35281654e-03
-1.17599952e+00 -4.20687526e-01 -9.45701718e-01 -5.39826155e-01
-2.07064152e-01 1.34186044e-01 9.07461166e-01 9.41237569e-01
-9.76077437e-01 6.60637558e-01 -1.07064497e+00 -4.92252767e-01
-3.70345116e-01 3.44010055e-01 -3.65270019e-01 5.58256246e-02
-1.21986127e+00 1.12159240e+00 6.90401495e-01 -4.51240897e-01
2.72307992e-02 -6.35726631e-01 -8.50849748e-01 -4.68109995e-02
-8.53096619e-02 -5.80775738e-01 1.68630362e+00 -6.66281462e-01
-1.30952430e+00 1.28289700e+00 -1.80751547e-01 -1.61915948e-03
2.11489461e-02 -3.61919969e-01 -2.46639580e-01 -6.01953983e-01
3.38169307e-01 4.65743333e-01 6.70470297e-02 -1.05820930e+00
-1.29396355e+00 -4.52296436e-01 6.48689792e-02 2.50783056e-01
-1.08590916e-01 6.52083039e-01 -3.74780893e-01 -5.92207015e-01
3.72964829e-01 -1.13451362e+00 -9.31685939e-02 -1.10820580e+00
-2.17824504e-01 -4.40478683e-01 4.43499684e-02 -1.07973635e+00
1.54670978e+00 -2.15624213e+00 2.53661722e-03 -2.31689829e-02
-3.83128040e-02 1.76159918e-01 -1.05218530e-01 6.14871025e-01
-3.18832472e-02 4.45343673e-01 -5.28230965e-02 -6.18123591e-01
1.10282302e-01 4.62166518e-01 -2.49722973e-01 2.29819790e-01
-1.58098251e-01 9.43683147e-01 -1.07770658e+00 -5.56124330e-01
-7.51614422e-02 2.52386242e-01 -6.79618835e-01 3.94986160e-02
-1.07644141e-01 5.04736245e-01 -3.45461279e-01 4.63308454e-01
4.34679002e-01 3.30823153e-01 1.00348759e+00 6.06865762e-03
-6.02250457e-01 1.36520219e+00 -8.19490373e-01 1.70706844e+00
-4.59282517e-01 2.56107777e-01 1.88093752e-01 -5.29349983e-01
6.06944561e-01 3.65929514e-01 2.10258439e-01 -4.42366779e-01
1.82378709e-01 5.67542315e-01 3.88543636e-01 -6.70970008e-02
6.21133626e-01 -3.51764530e-01 -8.30079973e-01 3.33173722e-01
3.92370939e-01 9.53850150e-02 3.29306096e-01 3.29324640e-02
1.20501244e+00 3.17056298e-01 5.77753127e-01 -6.69685483e-01
8.34240019e-02 2.93857723e-01 1.08575034e+00 7.42741108e-01
-2.95499098e-02 3.16871583e-01 4.86266762e-01 -3.59590769e-01
-1.10996020e+00 -1.08287227e+00 -1.42844766e-01 1.80793488e+00
-3.15208584e-01 -7.24747181e-01 -1.15273476e+00 -8.27614844e-01
-3.33848894e-01 1.14306700e+00 -4.10637259e-01 2.90905237e-01
-1.39523005e+00 -8.27834368e-01 9.48047936e-01 6.80524111e-01
-4.26890515e-02 -9.54084575e-01 -2.63968468e-01 5.04323125e-01
-5.68494856e-01 -1.16969061e+00 -4.58732009e-01 7.36308515e-01
-8.72239053e-01 -8.69433224e-01 6.89896122e-02 -1.24816310e+00
3.27835262e-01 -1.31031543e-01 1.51176691e+00 1.14473760e-01
5.93300402e-01 -1.04770645e-01 -5.30857563e-01 -3.28978032e-01
-1.03338933e+00 9.34735298e-01 -2.41257861e-01 -8.65423024e-01
5.92862844e-01 -4.24883246e-01 6.50588796e-02 -1.28221035e-01
-4.44210827e-01 7.95596614e-02 5.39490104e-01 4.51448828e-01
4.29921359e-01 -4.04826522e-01 1.26611963e-01 -1.66450644e+00
3.94235253e-01 -3.96723688e-01 -6.08153999e-01 5.05582094e-01
-4.02540952e-01 3.86585504e-01 6.86720312e-01 2.46493500e-02
-1.09065604e+00 4.43946391e-01 -7.88690567e-01 6.28291607e-01
-2.00303227e-01 4.62002575e-01 -6.43047810e-01 1.96600646e-01
5.21032572e-01 -3.03813130e-01 -5.77908456e-01 -7.72751868e-01
7.92640686e-01 5.21644056e-01 7.00565815e-01 -9.59111392e-01
4.68635380e-01 -1.42091885e-01 -3.77478093e-01 -3.72887254e-01
-1.18827879e+00 -4.24339741e-01 -9.90165889e-01 3.50700051e-01
9.97508526e-01 -8.89852345e-01 -3.42776090e-01 3.77844870e-01
-1.43814147e+00 -5.00017583e-01 -8.44954476e-02 4.24213350e-01
-3.28083396e-01 4.98106539e-01 -1.14788210e+00 -2.72950083e-01
-3.92701954e-01 -1.00330853e+00 8.42370927e-01 -1.30842943e-02
-3.98970187e-01 -1.21940207e+00 4.78088111e-01 3.36405516e-01
2.57240981e-02 -9.70530137e-02 1.34391201e+00 -9.04661357e-01
-1.38852686e-01 1.61034673e-01 9.92610082e-02 5.55303544e-02
1.17335826e-01 -1.27095357e-01 -8.27743709e-01 -2.01566309e-01
-2.01025337e-01 9.07736719e-02 7.18334496e-01 3.45569462e-01
1.18776873e-01 -2.53678411e-01 -4.70719427e-01 8.44042003e-01
1.33431256e+00 1.43334210e-01 5.08149862e-01 5.14139175e-01
7.97102988e-01 6.61286533e-01 5.04836321e-01 -2.74755776e-01
8.84581506e-01 5.78203022e-01 1.67171154e-02 6.91670924e-02
-2.13772923e-01 -7.14022458e-01 5.84316194e-01 1.46243834e+00
2.35372572e-03 -3.55748236e-01 -9.96815801e-01 5.04558384e-01
-1.74486971e+00 -5.49864888e-01 -3.01486105e-01 1.99938345e+00
1.07110310e+00 2.32679322e-01 1.30452171e-01 -9.75316316e-02
8.80354345e-01 3.73802930e-02 1.58755690e-01 -6.76601171e-01
-4.93304571e-03 3.84449869e-01 9.48955476e-01 1.00329888e+00
-1.17836368e+00 1.87312627e+00 7.38992643e+00 4.54374522e-01
-1.07727087e+00 4.96402234e-01 2.01597840e-01 5.07182002e-01
-5.92766285e-01 6.05601490e-01 -1.48807073e+00 4.26603109e-01
1.09887540e+00 2.84605771e-01 1.32371828e-01 7.99823463e-01
-2.30255127e-01 1.12037495e-01 -1.11053503e+00 2.46823326e-01
-2.40896821e-01 -9.41031933e-01 -3.06799233e-01 -5.65573573e-02
4.36719239e-01 3.44801247e-01 -2.65380412e-01 4.72740352e-01
1.14746118e+00 -5.12164772e-01 9.95476604e-01 -1.70417875e-01
8.24408114e-01 -4.47223306e-01 7.43066072e-01 4.13311243e-01
-1.28502953e+00 5.68762049e-02 -3.17259490e-01 -3.68489683e-01
4.89798665e-01 4.96543288e-01 -8.08231831e-01 2.97886550e-01
3.50053221e-01 2.65991867e-01 -4.08160836e-01 5.22116303e-01
-9.13693547e-01 1.21217430e+00 -3.71580094e-01 1.10828020e-01
1.81636050e-01 -4.77074943e-02 4.26263660e-01 1.81421947e+00
3.05211812e-01 2.22642824e-01 3.29688936e-01 1.16032809e-01
-1.27330432e-02 4.33762550e-01 -3.51454347e-01 3.10437083e-01
8.64469230e-01 1.12745237e+00 -8.91265869e-01 -2.63050735e-01
-5.91763794e-01 6.96411610e-01 8.97037566e-01 -1.72749534e-01
-5.93207002e-01 -1.30642466e-02 6.75862253e-01 1.26364931e-01
4.02738094e-01 -4.66211081e-01 -5.14654100e-01 -1.14173305e+00
-1.01857863e-01 -8.06634009e-01 7.36282229e-01 -2.77445793e-01
-8.98106039e-01 6.55681729e-01 1.93686202e-01 -5.42011142e-01
-4.12153602e-01 -6.51313424e-01 -5.28484106e-01 8.26238215e-01
-1.46814919e+00 -1.57343102e+00 3.37118775e-01 4.45642285e-02
2.41460830e-01 1.41211390e-01 1.02289534e+00 3.79372001e-01
-5.54921508e-01 8.78837764e-01 -7.42932111e-02 5.50415695e-01
6.80049598e-01 -1.48661768e+00 1.27281189e+00 1.34395242e+00
3.68969053e-01 8.16088319e-01 6.56772435e-01 -8.91430974e-01
-9.75779235e-01 -1.04121637e+00 1.80742967e+00 -8.24443340e-01
7.62691796e-01 -7.85484970e-01 -7.42770135e-01 1.26870191e+00
3.13880473e-01 -3.88485104e-01 8.17431092e-01 7.94501603e-01
-3.93155426e-01 2.72634625e-01 -8.34399283e-01 4.12040800e-01
1.36772907e+00 -3.82083803e-01 -9.77080762e-01 3.04215606e-02
7.11809814e-01 -6.54926598e-01 -1.02140272e+00 1.82728738e-01
5.92642188e-01 -8.21526766e-01 4.55800444e-01 -7.03427553e-01
7.41995350e-02 -6.63945526e-02 -3.39336902e-01 -1.39138389e+00
-9.73842025e-01 -7.65861809e-01 4.24726009e-01 1.70074666e+00
8.51017416e-01 -6.31149948e-01 6.73460901e-01 5.83274186e-01
-7.85928726e-01 -2.58905768e-01 -8.95577610e-01 -5.03477991e-01
6.92699790e-01 -4.74498898e-01 4.96389359e-01 8.57902050e-01
3.10621023e-01 7.33960330e-01 -1.28103212e-01 2.81192720e-01
4.19410437e-01 -1.20431691e-01 4.51755106e-01 -1.34106970e+00
-4.24508512e-01 -5.62679954e-02 -8.08716416e-02 -1.11885655e+00
4.11999583e-01 -1.19630265e+00 3.05843264e-01 -1.55173719e+00
6.69717342e-02 -1.02922022e+00 -2.59417832e-01 8.26027393e-01
-4.28944141e-01 1.03861019e-01 2.70174980e-01 2.49571219e-01
-4.05923039e-01 -3.56640965e-01 5.51455379e-01 3.82538617e-01
-2.89378434e-01 -5.61328270e-02 -1.08565593e+00 1.01145351e+00
8.71140361e-01 -9.59491432e-01 6.13418035e-02 -9.56414580e-01
6.49474382e-01 -1.54684365e-01 -5.04569888e-01 -7.66531765e-01
5.36946319e-02 -3.01284958e-02 -1.62481934e-01 -3.54141891e-01
-4.32219841e-02 -2.79240936e-01 7.68948868e-02 4.03056651e-01
-2.58231103e-01 4.68802035e-01 2.20654964e-01 -1.91297293e-01
7.43030682e-02 -5.83816230e-01 7.37447381e-01 -3.65858376e-01
-8.19639087e-01 -1.85914651e-01 -4.29564148e-01 7.68902481e-01
3.87714058e-01 -6.50848448e-02 -4.37483311e-01 1.58834621e-01
-5.27196705e-01 -2.43745595e-01 5.98414302e-01 2.74422616e-01
-4.19408739e-01 -8.29287589e-01 -8.74355435e-01 1.98954657e-01
-6.76569268e-02 -4.78551000e-01 -3.37857813e-01 3.88054937e-01
-5.13347030e-01 4.77622390e-01 -5.42097315e-02 -1.14665240e-01
-1.21790457e+00 3.06683004e-01 1.09702349e-01 -7.78973520e-01
-3.68816137e-01 7.96835363e-01 1.26321509e-01 -8.50554287e-01
-1.63221538e-01 -6.85357034e-01 -1.55803308e-01 -1.44216433e-01
8.51555690e-02 7.42578730e-02 3.92110258e-01 -8.70053232e-01
-6.31033957e-01 6.46114230e-01 -2.38497809e-01 -1.51188195e-01
1.12234175e+00 -1.62101686e-01 -9.39143226e-02 2.22811088e-01
7.73685217e-01 8.56554866e-01 -8.59690905e-01 -1.77116677e-01
4.75557983e-01 1.73680529e-01 -8.91226456e-02 -7.12844670e-01
-6.60081744e-01 6.02527022e-01 6.57334700e-02 2.53219336e-01
6.30438268e-01 3.33419740e-01 9.53061402e-01 1.25304595e-01
3.76914352e-01 -1.26956820e+00 -8.45711231e-01 1.29168797e+00
8.21579546e-02 -8.82946312e-01 -2.19836503e-01 -1.02423096e+00
-4.28951323e-01 7.82148421e-01 3.41375381e-01 -5.85811585e-02
6.29173577e-01 7.38424718e-01 4.68437552e-01 1.18191406e-01
-5.93705654e-01 -2.66394347e-01 -2.65623797e-02 5.74278414e-01
1.19331062e+00 3.49051476e-01 -8.63184810e-01 9.53950822e-01
-8.38306129e-01 -5.21415353e-01 2.78970659e-01 8.34648252e-01
-5.71796119e-01 -1.89776874e+00 -1.44579455e-01 5.98895401e-02
-1.12343729e+00 -6.73899472e-01 -4.33014482e-01 9.57307220e-01
4.82577622e-01 9.89918232e-01 -2.09525730e-02 -4.27310348e-01
3.86863142e-01 3.27746838e-01 6.63214087e-01 -1.22163463e+00
-1.06013310e+00 -2.26184465e-02 6.91238284e-01 -2.08609477e-01
-1.53748482e-01 -9.76900280e-01 -1.32725549e+00 -2.44881868e-01
-5.05451202e-01 3.51035208e-01 6.50546253e-01 1.22564399e+00
2.57609248e-01 1.94933817e-01 1.04061745e-01 -4.81013745e-01
-4.83138293e-01 -1.00129533e+00 -3.25298458e-01 3.03429097e-01
-2.85005361e-01 -3.94966125e-01 -1.82963401e-01 1.31738158e-02] | [10.372177124023438, 9.806989669799805] |
5ade4b63-71f7-4716-b8e8-ac44be8ea869 | a-closer-look-at-few-shot-3d-point-cloud | 2303.18210 | null | https://arxiv.org/abs/2303.18210v1 | https://arxiv.org/pdf/2303.18210v1.pdf | A Closer Look at Few-Shot 3D Point Cloud Classification | In recent years, research on few-shot learning (FSL) has been fast-growing in the 2D image domain due to the less requirement for labeled training data and greater generalization for novel classes. However, its application in 3D point cloud data is relatively under-explored. Not only need to distinguish unseen classes as in the 2D domain, 3D FSL is more challenging in terms of irregular structures, subtle inter-class differences, and high intra-class variances {when trained on a low number of data.} Moreover, different architectures and learning algorithms make it difficult to study the effectiveness of existing 2D FSL algorithms when migrating to the 3D domain. In this work, for the first time, we perform systematic and extensive investigations of directly applying recent 2D FSL works to 3D point cloud related backbone networks and thus suggest a strong learning baseline for few-shot 3D point cloud classification. Furthermore, we propose a new network, Point-cloud Correlation Interaction (PCIA), with three novel plug-and-play components called Salient-Part Fusion (SPF) module, Self-Channel Interaction Plus (SCI+) module, and Cross-Instance Fusion Plus (CIF+) module to obtain more representative embeddings and improve the feature distinction. These modules can be inserted into most FSL algorithms with minor changes and significantly improve the performance. Experimental results on three benchmark datasets, ModelNet40-FS, ShapeNet70-FS, and ScanObjectNN-FS, demonstrate that our method achieves state-of-the-art performance for the 3D FSL task. Code and datasets are available at https://github.com/cgye96/A_Closer_Look_At_3DFSL. | ['Tao Chen', 'Bo Zhang', 'Hongyuan Zhu', 'Chuangguan Ye'] | 2023-03-31 | null | null | null | null | ['3d-point-cloud-classification', 'few-shot-3d-point-cloud-classification', 'point-cloud-classification'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [-7.87301585e-02 -2.38933668e-01 -3.24815631e-01 -3.62061143e-01
-7.00555444e-01 -3.35163176e-01 6.55120075e-01 1.29211709e-01
-3.43283266e-03 2.35442072e-01 -2.73845464e-01 -1.74618766e-01
-1.90248013e-01 -8.42079043e-01 -6.19367898e-01 -5.50165057e-01
-2.17497736e-01 4.16739255e-01 7.01576829e-01 -2.08531424e-01
-1.70419997e-04 1.05038762e+00 -1.84719121e+00 4.52067181e-02
6.37416005e-01 1.11925924e+00 3.21173906e-01 9.70131010e-02
-5.58526933e-01 4.69396502e-01 -2.62946218e-01 -1.27879784e-01
5.18141091e-01 8.76584947e-02 -4.96459097e-01 1.47949494e-02
5.28562129e-01 -2.40162507e-01 -4.27807003e-01 9.64945853e-01
7.86180019e-01 2.08268285e-01 6.69927597e-01 -1.55802739e+00
-6.47989273e-01 1.13792546e-01 -7.37595439e-01 3.13283384e-01
1.00567237e-01 2.36278519e-01 8.69244814e-01 -1.40865064e+00
7.36389756e-01 1.18261552e+00 6.46321535e-01 5.86769998e-01
-9.50610816e-01 -9.90444064e-01 2.02643439e-01 3.65293890e-01
-1.42836320e+00 -2.08672956e-01 1.11529756e+00 -5.48326313e-01
1.16839671e+00 -1.93151087e-02 7.65622556e-01 1.16227353e+00
-1.39276072e-01 9.07392979e-01 7.80186176e-01 -1.77998930e-01
3.60734224e-01 7.77194798e-02 2.62298375e-01 5.19951582e-01
2.34109953e-01 2.68711627e-01 -3.46409559e-01 -2.68496238e-02
8.18281710e-01 3.63948703e-01 -3.38293523e-01 -8.81475866e-01
-8.19572806e-01 8.01705718e-01 7.67043233e-01 3.96474063e-01
-9.62219685e-02 -1.75490141e-01 3.94069523e-01 2.61034757e-01
8.83380055e-01 1.19842760e-01 -4.21250165e-01 -7.39772171e-02
-8.05285037e-01 3.64810467e-01 6.00629747e-01 1.25349557e+00
9.47883844e-01 5.21128029e-02 -3.95788476e-02 1.08475971e+00
8.87570605e-02 3.73824745e-01 2.62546599e-01 -6.91883981e-01
3.34108979e-01 7.43971169e-01 -4.02369738e-01 -8.53504121e-01
-4.16330010e-01 -6.11078441e-01 -8.66337121e-01 5.92624187e-01
-8.54915231e-02 1.68296814e-01 -1.13008523e+00 1.37473476e+00
5.53801239e-01 7.20939875e-01 -1.31590798e-01 9.78873134e-01
1.34620762e+00 4.90978569e-01 -2.78279215e-01 1.22766577e-01
1.00880742e+00 -9.36049283e-01 -1.88749582e-01 -7.36530051e-02
6.55904830e-01 -5.81533790e-01 1.07648122e+00 -4.97774407e-02
-8.42792809e-01 -7.44137883e-01 -1.11307526e+00 -2.18621958e-02
-7.35407650e-01 -3.84049088e-01 7.33005643e-01 3.18743378e-01
-7.38871813e-01 7.14346886e-01 -8.10652673e-01 -5.27469218e-01
8.86879265e-01 1.75381750e-01 -3.94964516e-01 -4.01058525e-01
-1.06808603e+00 6.84130251e-01 2.37312526e-01 -3.02626938e-01
-7.61125267e-01 -1.17703438e+00 -9.54727411e-01 1.23716578e-01
4.34441775e-01 -5.64954638e-01 1.09971654e+00 -3.36845875e-01
-1.18170702e+00 1.03488827e+00 1.23602137e-01 -1.84803307e-01
3.06757003e-01 7.23840371e-02 -4.00230139e-01 1.67913944e-01
2.21362278e-01 8.30336273e-01 7.47931480e-01 -1.55817437e+00
-6.33925557e-01 -4.61788327e-01 1.02778725e-01 -1.13479383e-02
-9.06010717e-02 -2.89187074e-01 -6.64885819e-01 -4.81479585e-01
1.35855153e-01 -8.93956363e-01 -7.33387917e-02 4.18452501e-01
-3.13676149e-02 -5.10868967e-01 1.18818521e+00 1.05923958e-01
9.06317830e-01 -2.48080301e+00 -2.10187450e-01 7.80517235e-02
4.06357318e-01 6.00885034e-01 -2.74634480e-01 3.18089843e-01
-3.56256247e-01 7.54841566e-02 -2.55716324e-01 -3.81093621e-01
5.76047190e-02 2.34996706e-01 -7.34819770e-02 3.57728034e-01
5.15685678e-01 9.83440042e-01 -9.59721565e-01 -5.09299099e-01
7.04324901e-01 4.83474731e-01 -4.17757779e-01 -7.37689855e-03
3.06128953e-02 2.14456916e-01 -4.37001795e-01 1.16297102e+00
1.01181459e+00 -4.86612499e-01 -5.59416592e-01 -1.45138443e-01
-1.63284853e-01 -1.63080722e-01 -1.10039425e+00 2.10529780e+00
-4.68355149e-01 3.86648864e-01 -2.28267774e-01 -9.51720536e-01
1.16673017e+00 1.44252613e-01 7.70564735e-01 -6.63646698e-01
2.40969181e-01 1.86898559e-01 -8.36364403e-02 -5.18587172e-01
1.50049642e-01 -1.26508757e-01 1.11581624e-01 8.51956606e-02
5.03332138e-01 -3.74154627e-01 1.33943502e-02 1.88725427e-01
1.10287130e+00 1.29423821e-02 1.28009513e-01 -1.95328612e-03
2.83253640e-01 -1.17081501e-01 6.13298178e-01 7.26890683e-01
-4.61977869e-01 8.82522941e-01 1.24682091e-01 -3.83580118e-01
-8.18701565e-01 -1.14780343e+00 -4.70876247e-01 6.53364062e-01
5.16134918e-01 -3.53532642e-01 -1.43605039e-01 -7.42311001e-01
4.59854305e-01 6.56698644e-01 -5.12347221e-01 -2.11846352e-01
-3.20243150e-01 -4.46763098e-01 2.20451057e-01 5.75644493e-01
5.27577460e-01 -7.46453643e-01 -3.55563402e-01 2.88389083e-02
3.02253842e-01 -1.20123780e+00 4.41045687e-02 3.43118966e-01
-9.75846767e-01 -1.13214862e+00 -8.00478637e-01 -9.33909297e-01
3.94791543e-01 8.60189438e-01 1.05915141e+00 -5.75270317e-02
-3.45495850e-01 4.14425731e-01 -6.24871969e-01 -6.75474226e-01
2.12456688e-01 7.65270814e-02 -7.69483820e-02 -2.31884316e-01
7.95463443e-01 -9.38306987e-01 -5.08984447e-01 3.83329719e-01
-7.45082736e-01 -1.15320943e-01 5.37656903e-01 7.35958934e-01
7.28564441e-01 -7.30803907e-02 3.77202630e-01 -6.90078199e-01
2.71872282e-01 -6.95362091e-01 -2.68011272e-01 -5.46999015e-02
-5.31984746e-01 -4.09564346e-01 3.76632184e-01 -2.92447507e-01
-7.81727433e-01 -1.80967271e-01 -2.15402573e-01 -1.42636478e+00
-4.21737760e-01 2.80792981e-01 -1.17075488e-01 -4.03648764e-01
6.52400732e-01 4.23664898e-02 3.02209090e-02 -6.42798841e-01
2.84990191e-01 7.41145134e-01 2.69399643e-01 -4.67375934e-01
1.12935185e+00 6.98660672e-01 1.01136759e-01 -1.05430293e+00
-8.15181792e-01 -8.42419863e-01 -7.46346831e-01 -3.73276621e-01
6.80891573e-01 -1.13238704e+00 -2.79866368e-01 4.27958965e-01
-1.01533079e+00 -1.07293557e-02 -6.50888205e-01 4.54106987e-01
-4.61009502e-01 2.87736624e-01 -2.46424198e-01 -6.42440975e-01
-3.03212643e-01 -1.18648183e+00 1.20234370e+00 2.66794294e-01
9.65227336e-02 -7.77529299e-01 9.10036117e-02 1.97358623e-01
3.09475213e-01 4.25343007e-01 1.01116848e+00 -7.10501909e-01
-7.91810274e-01 -3.44193906e-01 -4.62616771e-01 5.37508249e-01
1.76806271e-01 -3.34473811e-02 -1.16932666e+00 -4.11911607e-01
-2.15595160e-02 -3.49492043e-01 9.35093641e-01 3.07870299e-01
1.33931828e+00 5.42581677e-01 -4.80376601e-01 8.86018217e-01
1.45565569e+00 1.34276703e-01 4.65137213e-01 1.91854760e-01
7.56446004e-01 3.28596354e-01 6.26980245e-01 4.18940634e-01
2.87387222e-01 6.52593255e-01 6.75695717e-01 -8.94587561e-02
-3.78903002e-01 -1.28838435e-01 -7.59771839e-02 9.77180719e-01
-2.49629870e-01 -2.00270846e-01 -1.02719557e+00 6.23089612e-01
-1.85698879e+00 -9.35080945e-01 -5.53200245e-02 1.97027218e+00
2.44613960e-01 3.35192323e-01 -1.00723349e-01 5.60955815e-02
6.82153046e-01 4.16415334e-01 -6.61631763e-01 5.11635728e-02
-7.78164044e-02 5.24281740e-01 3.87140960e-01 -2.52402741e-02
-1.21428323e+00 9.48450267e-01 4.62103224e+00 1.18573940e+00
-1.17614698e+00 2.69649684e-01 2.50379950e-01 -2.83613414e-01
-1.38055369e-01 -5.11066802e-02 -9.43158865e-01 3.98171693e-01
4.10870075e-01 -9.29615945e-02 2.83977352e-02 1.09216940e+00
-5.39231580e-03 2.15215847e-01 -1.10093451e+00 1.45622838e+00
1.30827412e-01 -1.59128392e+00 9.86944959e-02 -8.93664360e-02
6.46419823e-01 7.16495872e-01 -1.23281114e-01 5.88142157e-01
-1.90136760e-01 -6.66570425e-01 5.05850017e-01 3.68746251e-01
8.63440752e-01 -6.60868526e-01 6.54104114e-01 3.95355046e-01
-1.41119742e+00 -2.79837567e-02 -6.65934384e-01 -6.77272230e-02
2.46271953e-01 6.09110415e-01 -5.22653401e-01 7.76541710e-01
1.04068708e+00 1.27593279e+00 -4.82813567e-01 1.42833316e+00
-9.15761739e-02 2.59832323e-01 -3.63768667e-01 -1.84841491e-02
4.41171020e-01 -5.92274480e-02 8.64391148e-01 9.42508519e-01
3.84470433e-01 1.26305297e-01 2.21209869e-01 9.15507674e-01
-7.98666626e-02 -3.74359898e-02 -8.69293392e-01 1.55300006e-01
4.25615788e-01 1.28565967e+00 -7.38632023e-01 -2.17940435e-01
-8.48160625e-01 5.97180307e-01 2.99143434e-01 2.73669213e-01
-7.84436643e-01 -5.54707527e-01 1.06915593e+00 2.83633858e-01
5.66060603e-01 -2.13226572e-01 -8.95390287e-02 -1.23593986e+00
4.11700755e-02 -3.71387661e-01 1.45082355e-01 -7.63200104e-01
-1.87619424e+00 5.05939901e-01 1.54443383e-01 -1.81511891e+00
3.07436079e-01 -6.02771282e-01 -8.03667963e-01 5.85585535e-01
-1.81614399e+00 -1.14599001e+00 -6.53440833e-01 7.43156075e-01
7.63148308e-01 -2.73962915e-01 6.97709143e-01 6.90557957e-01
-3.60393673e-01 5.61737299e-01 1.11781076e-01 4.74910140e-02
5.64184248e-01 -9.43886280e-01 6.35895610e-01 4.50786650e-01
1.56801119e-01 2.01800853e-01 2.36298889e-01 -5.89065015e-01
-1.13695943e+00 -1.31427908e+00 6.49074972e-01 -4.40674543e-01
5.63212752e-01 -7.13040769e-01 -1.23350894e+00 3.04963678e-01
-2.33918220e-01 3.61478835e-01 6.37723386e-01 4.98180874e-02
-3.84618104e-01 -5.71735315e-02 -1.09244418e+00 2.64579892e-01
1.52246642e+00 -4.93605852e-01 -5.86933672e-01 4.30033863e-01
1.01281798e+00 -4.24639732e-01 -9.25380409e-01 7.96821475e-01
2.35246837e-01 -1.01583314e+00 1.18527734e+00 -5.43132603e-01
4.18374002e-01 -3.81176233e-01 -3.23330641e-01 -1.13039577e+00
-6.60993218e-01 -1.19156778e-01 -2.48877540e-01 1.21497226e+00
2.21696228e-01 -4.79304254e-01 8.54667425e-01 2.29631349e-01
-6.11427844e-01 -1.13012064e+00 -1.03103566e+00 -1.18027425e+00
1.08272485e-01 -7.81757891e-01 8.05279076e-01 1.17841816e+00
-3.51128608e-01 3.02928180e-01 -7.45172799e-02 2.14479432e-01
7.32262135e-01 3.46647173e-01 9.30528522e-01 -1.61509538e+00
-7.84140751e-02 -4.94646907e-01 -9.29882228e-01 -1.00067902e+00
1.23490699e-01 -1.33226633e+00 -3.39071035e-01 -1.51081479e+00
8.79102722e-02 -5.69591701e-01 -4.06222373e-01 4.30212468e-01
1.52558103e-01 8.15057307e-02 4.67364341e-01 2.94523537e-01
-5.32379925e-01 9.64190602e-01 1.33257389e+00 -4.25482482e-01
-1.85678750e-01 1.59214929e-01 -2.86703825e-01 6.43984735e-01
5.87635398e-01 -4.12386179e-01 -5.18983603e-01 -3.76149684e-01
-2.10182384e-01 -3.41933250e-01 5.02586901e-01 -1.38408566e+00
2.48388112e-01 1.66943595e-01 4.14472342e-01 -1.03473067e+00
5.57167768e-01 -9.75341558e-01 3.16942371e-02 6.30906224e-02
3.08723539e-01 -2.88588107e-01 3.40002090e-01 7.29822278e-01
-3.37381214e-01 -2.25586444e-01 9.26093757e-01 -3.01038623e-01
-1.18741560e+00 8.81876230e-01 1.65642679e-01 1.21546037e-01
1.27324152e+00 -6.43151462e-01 -3.28257054e-01 4.81881294e-03
-6.82851195e-01 4.05860513e-01 5.82753122e-01 8.72261524e-01
9.08943713e-01 -1.56293404e+00 -4.38102931e-01 4.21621442e-01
6.64021134e-01 4.25626069e-01 6.01522088e-01 7.23043263e-01
-3.73017222e-01 2.14799047e-01 -1.73971117e-01 -1.01116848e+00
-1.02751744e+00 5.56845129e-01 2.20968872e-01 1.32938176e-01
-1.00822258e+00 1.07861853e+00 2.49881878e-01 -6.88842833e-01
3.87914062e-01 -3.26369733e-01 1.85284745e-02 1.31568104e-01
3.03036749e-01 4.09473300e-01 1.99459523e-01 -5.05085409e-01
-5.55451334e-01 8.16524029e-01 -1.22135013e-01 5.52134395e-01
1.70273972e+00 1.63464859e-01 2.33200029e-01 6.61924303e-01
1.45785177e+00 -5.72198033e-01 -1.20080483e+00 -5.26271403e-01
-1.96088433e-01 -7.34464824e-01 2.15284720e-01 -5.15954137e-01
-1.28589463e+00 1.08217025e+00 9.28607464e-01 8.62187147e-02
8.97339404e-01 4.02173936e-01 7.76869595e-01 1.33165643e-01
5.75052500e-01 -8.86967897e-01 3.00431233e-02 6.68083847e-01
6.42968476e-01 -1.48476052e+00 -1.15521424e-01 -4.38898087e-01
-3.66114855e-01 9.78291094e-01 7.64123559e-01 -2.57114589e-01
1.25043786e+00 4.33796383e-02 -1.31548643e-01 -6.17805839e-01
-6.52100801e-01 -3.95484060e-01 1.61708623e-01 7.03715980e-01
9.90628749e-02 -1.83949962e-01 2.44875774e-01 5.35931230e-01
-2.49195355e-03 7.88073912e-02 2.31124580e-01 1.13117337e+00
-5.00926316e-01 -9.20585573e-01 3.06774713e-02 6.41541123e-01
1.72601506e-01 9.80494320e-02 -2.80425996e-01 1.06666493e+00
4.58366364e-01 6.63506210e-01 2.82693624e-01 -6.03407204e-01
6.36749387e-01 9.14870948e-02 3.95833433e-01 -8.77340496e-01
-2.64666528e-01 -3.93053144e-02 -3.65521431e-01 -7.13450491e-01
-6.11530006e-01 -5.17309606e-01 -1.18737209e+00 -3.08070511e-01
-5.39852142e-01 -8.22363347e-02 5.91860473e-01 6.70281529e-01
6.69098496e-01 5.50709665e-01 6.86758697e-01 -1.29478073e+00
-2.53273159e-01 -7.52750158e-01 -7.82333910e-01 3.36940944e-01
1.96079150e-01 -1.19978905e+00 -5.76503694e-01 -5.88370204e-01] | [7.965387344360352, -3.2819056510925293] |
6eea4111-bbb7-478a-8284-dff564321963 | simple-applications-of-bert-for-ad-hoc | 1903.10972 | null | http://arxiv.org/abs/1903.10972v1 | http://arxiv.org/pdf/1903.10972v1.pdf | Simple Applications of BERT for Ad Hoc Document Retrieval | Following recent successes in applying BERT to question answering, we explore
simple applications to ad hoc document retrieval. This required confronting the
challenge posed by documents that are typically longer than the length of input
BERT was designed to handle. We address this issue by applying inference on
sentences individually, and then aggregating sentence scores to produce
document scores. Experiments on TREC microblog and newswire test collections
show that our approach is simple yet effective, as we report the highest
average precision on these datasets by neural approaches that we are aware of. | ['Wei Yang', 'Haotian Zhang', 'Jimmy Lin'] | 2019-03-26 | null | null | null | null | ['ad-hoc-information-retrieval'] | ['natural-language-processing'] | [ 1.87558591e-01 2.13285331e-02 -1.72329191e-02 -5.69629014e-01
-1.42283940e+00 -8.55263889e-01 7.80730486e-01 4.79796082e-01
-1.01292253e+00 8.77151072e-01 4.25582051e-01 -5.88331461e-01
-5.96200287e-01 -7.91865766e-01 -6.45014703e-01 -1.64657563e-01
-4.43379939e-01 7.54318357e-01 5.71968377e-01 -5.37971377e-01
4.71260130e-01 7.68944398e-02 -1.45134032e+00 6.98871076e-01
4.52921242e-01 9.65647042e-01 -1.84943125e-01 1.21282303e+00
-2.48661742e-01 9.31827903e-01 -1.20199049e+00 -6.25927806e-01
-1.28677045e-03 -1.04851805e-01 -1.36518073e+00 -4.17784929e-01
6.18044257e-01 -5.14408827e-01 -1.25635818e-01 7.03482330e-01
5.20770252e-01 3.33769470e-01 7.15902388e-01 -4.95018780e-01
-7.49821246e-01 8.88327420e-01 5.40072136e-02 8.76135647e-01
7.15355217e-01 -1.61782816e-01 1.49008858e+00 -6.28632605e-01
5.46182573e-01 1.15251708e+00 4.57320333e-01 6.59608185e-01
-9.71099198e-01 -9.45160091e-02 2.84452707e-01 2.53397107e-01
-1.20824516e+00 -5.02026439e-01 3.19267064e-01 -3.26615646e-02
1.45806825e+00 5.78686595e-01 2.46801630e-01 9.11968946e-01
-4.31246217e-03 1.00071990e+00 5.69514930e-01 -5.96391439e-01
3.25110018e-01 -1.04737230e-01 6.91934705e-01 2.16452703e-01
8.48847777e-02 -3.47997606e-01 -2.31176734e-01 -4.49618369e-01
-6.84762821e-02 -4.99728978e-01 -2.42617533e-01 4.20435160e-01
-7.75197923e-01 1.04542935e+00 3.80286992e-01 3.69639516e-01
-4.63146567e-01 3.66047055e-01 5.42444110e-01 5.54445922e-01
6.83158755e-01 8.08289111e-01 -6.65451348e-01 -1.20106168e-01
-9.35075402e-01 8.41834009e-01 1.30557084e+00 7.20745385e-01
2.45456487e-01 -4.97933716e-01 -3.64371717e-01 8.54783356e-01
1.16315424e-01 3.32463503e-01 5.50305784e-01 -9.36756015e-01
6.62885427e-01 1.93122298e-01 2.74522096e-01 -8.75242949e-01
-3.45572680e-01 -1.89375043e-01 -1.85584709e-01 -4.59155113e-01
5.38971663e-01 -2.14975327e-01 -8.87601614e-01 1.52031183e+00
8.50991085e-02 -4.22091424e-01 1.51213989e-01 5.78370512e-01
9.56674099e-01 8.12801778e-01 1.38194814e-01 -2.26691023e-01
1.36734939e+00 -7.14569807e-01 -6.86344683e-01 -3.79626900e-01
8.46395969e-01 -5.50749779e-01 9.33885515e-01 4.42953378e-01
-1.28843999e+00 5.65553494e-02 -1.11541390e+00 -2.72488803e-01
-5.89611471e-01 -4.45971251e-01 6.63414657e-01 4.54646260e-01
-1.16130650e+00 4.92301226e-01 -3.94961178e-01 -3.28497320e-01
1.67136401e-01 4.26429033e-01 1.07088156e-01 -5.87563254e-02
-1.58718979e+00 1.02583587e+00 4.92755890e-01 8.08406398e-02
-5.24284065e-01 -2.57969320e-01 -5.54980338e-01 1.81206107e-01
5.99353015e-01 -4.91553247e-01 2.02274132e+00 -6.21366262e-01
-1.20568144e+00 5.12439728e-01 -2.36982435e-01 -9.94260490e-01
2.35250294e-01 -4.10279602e-01 -3.12335342e-01 4.63803560e-01
7.72428745e-03 6.21189892e-01 2.79045314e-01 -6.84838593e-01
-5.78906178e-01 -8.36259425e-02 6.16270423e-01 2.66523689e-01
-5.83209753e-01 4.30558026e-01 -6.29192770e-01 -2.16182590e-01
-1.33397758e-01 -4.49327976e-01 -3.98912504e-02 -4.84376341e-01
-2.54464328e-01 -8.72874677e-01 4.28525716e-01 -7.03613341e-01
1.52965403e+00 -1.63719201e+00 -1.51542127e-01 3.88852000e-01
1.41524747e-01 3.82573664e-01 -3.60067010e-01 6.89316928e-01
4.55997318e-01 4.68191296e-01 -7.53057301e-02 -1.89908788e-01
9.90945995e-02 3.65664929e-01 -6.34052694e-01 -2.75151599e-02
2.96304077e-01 1.14791393e+00 -8.94965768e-01 -7.49542773e-01
-5.25071442e-01 7.37194493e-02 -6.31084442e-01 -1.97923984e-02
-8.26452255e-01 -2.58756131e-01 -7.41490424e-01 5.63495755e-01
1.04498751e-01 -4.19297904e-01 2.14046929e-02 3.61613601e-01
2.23246321e-01 9.52450335e-01 -5.00102162e-01 1.33845150e+00
-1.75621808e-01 6.14818990e-01 1.92813165e-02 -9.45486128e-01
5.47282517e-01 3.26491743e-01 1.93750247e-01 -7.09328473e-01
1.24990813e-01 1.65679961e-01 1.51183538e-03 -8.61923993e-01
7.91784763e-01 -6.11833781e-02 -8.44810084e-02 6.91080809e-01
7.61392619e-03 -3.83215457e-01 6.91955626e-01 5.85751593e-01
1.50435102e+00 -5.89238822e-01 -2.78472044e-02 -2.20435724e-01
3.83965284e-01 2.73470998e-01 -1.06351957e-01 1.30247605e+00
-1.84138402e-01 5.86428344e-01 6.14500940e-01 -3.31688434e-01
-9.69718099e-01 -9.31330979e-01 -2.96533793e-01 1.25729835e+00
-3.51824254e-01 -5.81628501e-01 -8.56208384e-01 -9.04566884e-01
1.57993529e-02 7.53537357e-01 -6.32204711e-01 1.61764741e-01
-7.47178555e-01 -6.81852400e-01 7.39558458e-01 5.34002662e-01
9.06380489e-02 -1.20744836e+00 -2.87163317e-01 5.69071591e-01
-1.70499414e-01 -1.01475048e+00 -3.64951462e-01 2.10972041e-01
-6.95362329e-01 -9.85453725e-01 -7.39829719e-01 -7.35875726e-01
1.63673133e-01 -2.56632324e-02 1.46650481e+00 5.08069754e-01
-1.69901639e-01 5.94171941e-01 -4.32741106e-01 -5.65156281e-01
-1.65577590e-01 5.33282876e-01 -1.83933124e-01 -6.54561937e-01
7.10003436e-01 -2.02906996e-01 -6.04450881e-01 -1.48549571e-03
-1.24709213e+00 -4.98990208e-01 3.15493792e-01 7.94537723e-01
-6.68240264e-02 -1.68681547e-01 9.81203377e-01 -1.08402646e+00
1.38377047e+00 -6.65274441e-01 -4.14203078e-01 4.73857045e-01
-5.99061489e-01 -2.91569568e-02 4.33845639e-01 -3.19886088e-01
-6.61212802e-01 -5.34515262e-01 -6.07072353e-01 4.55882460e-01
-9.48490575e-02 1.10185409e+00 5.36150455e-01 3.09686393e-01
1.05480051e+00 -7.24366382e-02 -1.10302538e-01 -2.52245933e-01
2.75153369e-01 9.20622230e-01 4.24408674e-01 -6.60334408e-01
4.32887584e-01 6.44887611e-02 -4.30642962e-01 -7.08085537e-01
-1.28852105e+00 -6.80580020e-01 -2.01282218e-01 -1.35687187e-01
8.25311661e-01 -4.87915546e-01 -7.90243089e-01 -5.44736050e-02
-1.52460563e+00 1.59121603e-02 -1.46916121e-01 2.66082525e-01
-2.73549676e-01 3.26084286e-01 -1.01219785e+00 -8.62566113e-01
-6.80516839e-01 -6.86000466e-01 9.23870504e-01 -8.13344587e-03
-4.59364146e-01 -1.07326841e+00 2.50604749e-01 3.97822767e-01
7.76708245e-01 -3.36275369e-01 1.07142019e+00 -1.21880794e+00
-4.25466031e-01 -6.80505633e-01 -1.20408289e-01 4.10166979e-01
-4.03697163e-01 -1.35137171e-01 -9.54614758e-01 -1.34877071e-01
-6.01367727e-02 -7.82198608e-01 9.95149314e-01 2.03830227e-01
1.24536920e+00 -6.06257677e-01 3.16309370e-02 -3.29223961e-01
1.18265104e+00 1.58486933e-01 3.95160466e-01 6.08362556e-01
-1.31333666e-02 8.12960744e-01 3.40535194e-01 5.11316285e-02
4.18227792e-01 4.38332826e-01 5.50929606e-02 3.70552212e-01
3.69824946e-01 -2.52059265e-03 8.60583857e-02 7.65231967e-01
2.20536545e-01 -8.64790380e-01 -1.05667782e+00 6.84494734e-01
-2.01439691e+00 -1.17175329e+00 1.44405216e-01 1.75890195e+00
1.03616524e+00 5.80456853e-01 3.65840015e-03 9.42096412e-02
3.21162730e-01 2.46629581e-01 -4.25351471e-01 -7.54280865e-01
-8.72689709e-02 5.33048093e-01 2.62965083e-01 6.97786033e-01
-1.17476368e+00 8.22157145e-01 8.39302254e+00 6.98833585e-01
-7.11038291e-01 -1.39026746e-01 5.04403770e-01 -4.60963011e-01
-4.57537770e-01 -1.86017424e-01 -8.40506673e-01 1.75969288e-01
1.56208277e+00 -3.59440416e-01 4.53613311e-01 6.25437856e-01
-1.72562912e-01 -1.91318542e-01 -1.29376781e+00 3.85306120e-01
3.71240824e-01 -1.33212769e+00 9.07293707e-02 -3.48219842e-01
4.11453873e-01 2.12665021e-01 -6.39247596e-02 7.14124739e-01
4.87298131e-01 -1.01967788e+00 3.70670468e-01 3.64269465e-01
1.60798728e-01 -4.24698770e-01 9.41807508e-01 6.06669724e-01
-3.92710239e-01 -2.46677756e-01 -3.83850604e-01 -2.37646043e-01
2.13875592e-01 4.80220586e-01 -1.24475348e+00 3.06088775e-01
6.03244722e-01 3.35448444e-01 -8.13781023e-01 1.17144966e+00
-1.70506090e-01 8.69789958e-01 -5.94734490e-01 -8.63809109e-01
5.80874145e-01 5.54821730e-01 2.27688611e-01 1.39889681e+00
-1.20014399e-01 2.06088796e-01 7.00776279e-02 4.11332518e-01
-2.86629319e-01 2.13801250e-01 -5.87911367e-01 -1.76870525e-01
4.22749281e-01 9.50530231e-01 -4.97438967e-01 -7.33802795e-01
-3.91606092e-01 6.30885065e-01 6.49038315e-01 5.09488821e-01
-4.01286751e-01 -6.86573744e-01 -5.80373080e-03 -1.97764963e-01
2.30737716e-01 -2.14529708e-01 -6.70846105e-02 -9.38607097e-01
3.01192909e-01 -9.08901334e-01 8.11210096e-01 -8.04070950e-01
-1.38100362e+00 7.58132696e-01 4.36444163e-01 -6.38332367e-01
-8.85916591e-01 -7.42936969e-01 -5.78319073e-01 6.94424808e-01
-1.49673688e+00 -4.84758943e-01 2.98501372e-01 1.62891835e-01
5.57370126e-01 7.13227242e-02 9.45928335e-01 3.93173933e-01
-2.24032968e-01 4.91413116e-01 3.23887110e-01 1.37447357e-01
4.82266039e-01 -1.29583883e+00 5.34554541e-01 5.42249918e-01
3.87402505e-01 1.01658010e+00 7.30283082e-01 -2.59005815e-01
-1.34370852e+00 -5.38700819e-01 1.40453506e+00 -8.33504856e-01
9.01203096e-01 -3.49214166e-01 -9.26236868e-01 6.69164598e-01
4.02470946e-01 -4.57264453e-01 6.42159224e-01 4.49503154e-01
-2.79091090e-01 1.24198953e-02 -9.01880682e-01 4.57871705e-01
6.67635620e-01 -7.77945101e-01 -1.05067921e+00 7.86305130e-01
9.23253775e-01 -3.28597128e-01 -6.74217880e-01 4.77841616e-01
4.08071965e-01 -4.75268692e-01 8.83004665e-01 -1.16359496e+00
5.86246133e-01 7.59732276e-02 -2.01503009e-01 -1.02024662e+00
-4.68581356e-02 -1.97572708e-01 -1.79968521e-01 1.02761137e+00
9.47609782e-01 -5.09670198e-01 8.90708089e-01 9.25899506e-01
-2.71750577e-02 -1.00321102e+00 -1.01978779e+00 -5.96732616e-01
5.45892835e-01 -5.83025277e-01 5.36838591e-01 4.86394733e-01
2.43616641e-01 5.31418681e-01 7.60666328e-03 -1.49061427e-01
2.60667026e-01 -1.42951176e-01 4.74360138e-01 -1.00906515e+00
-5.79678595e-01 -4.58906025e-01 4.74260524e-02 -1.32101977e+00
1.51055545e-01 -8.65420759e-01 4.23875093e-01 -1.70369446e+00
1.30774528e-01 -1.40491307e-01 -4.13781196e-01 3.37552756e-01
-2.00020254e-01 3.04219306e-01 -1.23197310e-01 1.60659790e-01
-1.05137837e+00 2.38778710e-01 9.57587361e-01 -4.75016981e-01
1.64946482e-01 -1.15235426e-01 -8.82137835e-01 4.44844961e-01
7.78272927e-01 -4.34690744e-01 -3.18431139e-01 -9.27718580e-01
6.75660849e-01 1.00981057e-01 1.13414019e-01 -6.51047528e-01
6.38234735e-01 1.56140417e-01 1.93403840e-01 -7.39606321e-01
3.98478746e-01 -3.73109460e-01 -6.87278390e-01 1.82023868e-01
-1.11001039e+00 3.36254537e-01 1.91502780e-01 4.87606615e-01
-4.66803521e-01 -7.73686886e-01 2.44772971e-01 -1.97066873e-01
-4.21994388e-01 9.23231691e-02 -7.42573798e-01 3.74297470e-01
5.43488085e-01 3.12981367e-01 -7.28142083e-01 -8.72470856e-01
-6.41505778e-01 6.53509855e-01 -3.01366389e-01 4.09840912e-01
5.69247842e-01 -9.01211798e-01 -7.00905561e-01 -1.69948965e-01
1.47035867e-01 5.52725866e-02 -1.92978725e-01 4.62910533e-01
-5.18912852e-01 9.66651797e-01 5.69762051e-01 -3.31344783e-01
-1.33160102e+00 4.64689761e-01 8.40342864e-02 -4.94160891e-01
-3.28867346e-01 1.08088231e+00 -4.54840988e-01 -3.51080477e-01
5.92073083e-01 -2.61832029e-01 -5.20499766e-01 1.24534890e-01
9.25650716e-01 -2.61540543e-02 5.11082590e-01 6.60779476e-02
-2.78298318e-01 1.71467081e-01 -6.83375061e-01 -5.82493424e-01
1.24275386e+00 1.00885376e-01 -1.17796667e-01 3.14528972e-01
1.47692776e+00 -2.18801364e-01 -3.47730845e-01 -3.99437547e-01
5.14929950e-01 -1.30586565e-01 5.41072572e-03 -9.94162083e-01
-1.34101212e-01 6.12211287e-01 1.92924976e-01 8.75394523e-01
8.22458029e-01 2.18098164e-01 9.06142473e-01 1.49620497e+00
1.69582009e-01 -1.16695547e+00 -2.63618845e-02 1.02169502e+00
9.13254440e-01 -9.80097592e-01 6.67149350e-02 1.73703477e-01
-2.73286581e-01 9.73320663e-01 3.59127671e-01 -1.87040865e-01
5.87336063e-01 3.91430371e-02 2.21373081e-01 -4.79940414e-01
-1.32144475e+00 -8.36659521e-02 4.31755781e-01 7.22857937e-02
5.68690062e-01 -3.99885595e-01 -8.93137753e-01 1.50331855e-01
-4.67491359e-01 -7.14613497e-02 3.40732425e-01 1.29359031e+00
-8.34636390e-01 -9.60532069e-01 -2.62369156e-01 8.58195722e-01
-1.09807444e+00 -3.84215593e-01 -8.07949066e-01 5.50862372e-01
-7.24820614e-01 1.26262152e+00 2.33166814e-01 -1.72019690e-01
2.44339332e-01 3.09682339e-01 3.20439875e-01 -6.43094778e-01
-7.15912998e-01 -2.75230557e-01 7.82835066e-01 -3.49297255e-01
-2.56483138e-01 -5.03670573e-01 -1.03842640e+00 -1.79559708e-01
-4.03564900e-01 8.39373291e-01 7.97964036e-01 1.06435573e+00
2.29963019e-01 2.77687967e-01 4.15883452e-01 -4.59915996e-01
-1.24919891e+00 -1.23957372e+00 -1.17511503e-01 3.94831270e-01
6.36456549e-01 -8.32747519e-02 -6.17614269e-01 -3.32102329e-01] | [11.295520782470703, 8.0549898147583] |
30a2c9cb-68b7-4198-b945-38702be6ab0c | learning-sparse-causal-models-is-not-np-hard | 1309.6824 | null | http://arxiv.org/abs/1309.6824v1 | http://arxiv.org/pdf/1309.6824v1.pdf | Learning Sparse Causal Models is not NP-hard | This paper shows that causal model discovery is not an NP-hard problem, in
the sense that for sparse graphs bounded by node degree k the sound and
complete causal model can be obtained in worst case order N^{2(k+2)}
independence tests, even when latent variables and selection bias may be
present. We present a modification of the well-known FCI algorithm that
implements the method for an independence oracle, and suggest improvements for
sample/real-world data versions. It does not contradict any known hardness
results, and does not solve an NP-hard problem: it just proves that sparse
causal discovery is perhaps more complicated, but not as hard as learning
minimal Bayesian networks. | ['Joris Mooij', 'Tom Heskes', 'Tom Claassen'] | 2013-09-26 | null | null | null | null | ['model-discovery'] | ['miscellaneous'] | [ 4.86032695e-01 8.89167786e-01 -6.31112039e-01 -3.53471935e-01
-5.69532633e-01 -7.01010227e-01 3.76170576e-01 -6.96815476e-02
4.88427607e-03 1.45774698e+00 9.63316634e-02 -6.45515978e-01
-1.23156381e+00 -1.00918531e+00 -8.43647778e-01 -8.98843706e-01
-8.61647666e-01 1.33516371e+00 4.30013448e-01 4.20337290e-01
5.07431617e-03 4.02918994e-01 -1.01010847e+00 -2.64723867e-01
6.30285501e-01 2.47661099e-01 -3.50498438e-01 9.48634803e-01
5.04898310e-01 8.26921403e-01 -1.85336396e-01 -5.09168267e-01
1.97648123e-01 -6.94399297e-01 -1.06995738e+00 -2.74524093e-01
4.73315060e-01 -1.90649658e-01 -6.78346872e-01 1.28824854e+00
2.06238717e-01 -3.09862942e-01 6.90022469e-01 -1.67581654e+00
-4.34021473e-01 1.49613142e+00 -6.36316001e-01 3.97509277e-01
4.61884648e-01 -4.18441385e-01 1.24151516e+00 4.66532521e-02
6.45149469e-01 1.60955393e+00 5.09757876e-01 1.76386505e-01
-1.61785579e+00 -8.94814432e-01 -1.97696146e-02 4.03926075e-01
-1.40442598e+00 -2.56949693e-01 1.78637370e-01 -2.00017199e-01
7.07760990e-01 7.25603402e-01 5.34252644e-01 1.04146862e+00
-2.66418792e-02 4.63830143e-01 1.44659257e+00 -4.98816133e-01
4.99172330e-01 -3.36579412e-01 3.84257436e-01 7.83257604e-01
1.15323961e+00 4.12741840e-01 -6.39459550e-01 -8.30247283e-01
1.03573716e+00 -4.51651573e-01 -2.80541211e-01 -2.70849973e-01
-1.11352587e+00 9.22292650e-01 -1.26301363e-01 9.63981226e-02
7.02201948e-02 5.13448060e-01 -1.34548917e-01 4.84617501e-01
-1.23681746e-01 2.80759603e-01 -5.35514295e-01 5.75835966e-02
-8.94252479e-01 2.98059225e-01 1.57140243e+00 1.24650836e+00
4.00203854e-01 -7.50793070e-02 2.31144056e-01 -5.41772246e-02
3.29697251e-01 8.31125081e-01 -4.92911577e-01 -1.17610300e+00
1.35875627e-01 -2.02277936e-02 2.85622448e-01 -9.33018148e-01
-4.30088758e-01 -3.85893196e-01 -8.82576644e-01 -9.36282948e-02
6.89742148e-01 -3.31082106e-01 -5.28124511e-01 1.98576117e+00
2.92973787e-01 3.94814193e-01 -3.07489306e-01 5.39659381e-01
5.49453139e-01 2.03515276e-01 1.98120568e-02 -7.66862512e-01
9.74786580e-01 -1.00426562e-01 -8.68016124e-01 -1.44412816e-01
1.40704289e-01 -3.43209177e-01 1.89282328e-01 7.64711559e-01
-1.16564608e+00 1.43577516e-01 -8.84081900e-01 3.43912899e-01
1.15211077e-01 -6.28605783e-01 1.51897144e+00 1.15811062e+00
-8.86190414e-01 4.25102174e-01 -7.06785023e-01 -4.18135166e-01
1.44830078e-01 5.79187572e-01 -5.81983626e-01 -5.31946123e-01
-1.22343171e+00 6.06499553e-01 5.38085043e-01 -1.72360361e-01
-1.25710642e+00 -4.48324144e-01 -5.99531233e-01 1.65896699e-01
9.69819546e-01 -7.62019217e-01 7.34170854e-01 -4.32687104e-01
-8.74910951e-01 4.91714478e-01 -5.15284538e-01 -3.85582566e-01
3.27442795e-01 2.48213574e-01 -5.85309386e-01 1.52601600e-01
1.03203364e-01 -1.34916201e-01 4.25892889e-01 -1.34348524e+00
-7.24963307e-01 -6.46509767e-01 2.85250157e-01 -2.99218744e-01
9.57484618e-02 1.24043792e-01 -1.45058677e-01 -6.30678311e-02
4.55721587e-01 -8.76184404e-01 -5.38541734e-01 -4.81946468e-01
-1.02955687e+00 -2.37334058e-01 1.60848424e-01 -4.59574223e-01
1.20565796e+00 -1.63785756e+00 9.44683477e-02 9.33656991e-01
5.32103598e-01 -4.67459649e-01 7.49354856e-03 5.19284248e-01
-3.17066133e-01 4.93680358e-01 -2.24002227e-01 4.01785821e-01
1.65917724e-01 5.65768898e-01 -1.07356466e-01 8.74417305e-01
-3.45681250e-01 6.12324536e-01 -1.01213646e+00 -7.71438658e-01
-2.55406201e-01 -1.84771806e-01 -4.59094673e-01 -2.99324632e-01
-1.54490069e-01 7.77365714e-02 -4.39564586e-01 5.77059388e-01
8.39817703e-01 -6.62267804e-01 1.02113020e+00 2.90333688e-01
8.26609805e-02 2.14435279e-01 -1.79008961e+00 1.02731872e+00
2.10243225e-01 4.91600573e-01 2.77763218e-01 -1.02396500e+00
2.98827201e-01 5.74687064e-01 2.68944860e-01 -2.41411969e-01
-4.60146070e-02 1.18493952e-01 3.22561413e-01 -5.45151711e-01
-2.06252873e-01 -2.45099664e-01 -2.17850968e-01 4.15982604e-01
6.09971322e-02 1.33602202e-01 3.44597638e-01 5.91620862e-01
1.79542196e+00 -4.66616988e-01 3.17932338e-01 -5.07292271e-01
-1.96773201e-01 1.66667208e-01 8.88478279e-01 1.53461969e+00
1.88456759e-01 1.89537182e-01 1.16339958e+00 8.15721229e-02
-7.52708852e-01 -1.35678709e+00 -3.82683069e-01 8.12182188e-01
1.70363650e-01 -4.10556734e-01 -4.08405453e-01 -5.64804018e-01
5.85966706e-02 7.92282224e-01 -9.24378395e-01 3.23850244e-01
-1.31280974e-01 -1.30104506e+00 5.13949037e-01 4.29756969e-01
2.39676749e-03 -3.50483179e-01 2.12405130e-01 1.64383247e-01
-3.15929241e-02 -1.04782629e+00 2.80224264e-01 8.32152665e-01
-9.95421290e-01 -1.67850006e+00 1.83477610e-01 -3.22900712e-01
8.00430596e-01 9.50906649e-02 1.21664798e+00 -1.63633507e-02
-1.38311267e-01 2.43029892e-01 -4.72605564e-02 -3.88904631e-01
-3.16251516e-01 -3.08926910e-01 -3.91770117e-02 -6.70641482e-01
5.42318285e-01 -9.56776321e-01 -5.23306280e-02 2.93425709e-01
-7.23111749e-01 -4.75229204e-01 4.63065326e-01 8.54786992e-01
6.11124992e-01 8.73351216e-01 4.61745113e-01 -1.36026764e+00
2.70483226e-01 -8.49600792e-01 -1.02414727e+00 4.74349201e-01
-5.66081882e-01 1.81375397e-03 2.99175799e-01 3.19420844e-02
-9.79820013e-01 2.69908041e-01 3.15648258e-01 2.34757587e-01
-2.14939684e-01 8.22063148e-01 -5.46306789e-01 -4.05456834e-02
7.19952583e-01 -2.48651668e-01 -3.88575435e-01 -3.25139821e-01
2.93741375e-01 1.62466437e-01 5.20690620e-01 -4.87098098e-01
1.11887729e+00 7.95554340e-01 9.07194972e-01 -6.60473883e-01
-8.54747474e-01 -2.09892169e-01 -5.10305941e-01 3.09683114e-01
5.13766050e-01 -7.38503575e-01 -9.72213089e-01 -4.12880927e-01
-8.19003820e-01 -3.40424061e-01 -6.74145669e-02 7.93941617e-01
-6.24953330e-01 4.52751428e-01 -3.62607509e-01 -9.43205595e-01
4.09519166e-01 -4.61575091e-01 2.72236794e-01 -1.67539015e-01
-2.03676134e-01 -1.06480718e+00 1.49244234e-01 2.53067881e-01
-2.72781461e-01 3.66350770e-01 1.27985775e+00 -9.44656610e-01
-7.48582244e-01 -3.68252337e-01 -2.94290841e-01 -3.23214382e-01
-2.12261751e-01 2.17629790e-01 -6.51469827e-01 -1.88764930e-01
-2.31808588e-01 -1.73841953e-01 8.14190388e-01 6.36179984e-01
9.64136839e-01 -5.90276241e-01 -8.69165599e-01 2.24339053e-01
1.60586572e+00 -7.52073824e-02 7.15179682e-01 -2.41029412e-01
5.07490635e-01 3.51018131e-01 2.00155810e-01 4.52786773e-01
3.66615355e-01 2.01290503e-01 6.82884514e-01 1.78338420e-02
1.16403282e-01 -1.77116349e-01 -2.05716223e-01 2.79232025e-01
-4.38948691e-01 -4.15143996e-01 -7.66509056e-01 5.71958482e-01
-1.92614424e+00 -1.47109568e+00 -1.10978913e+00 2.41259027e+00
1.14052749e+00 2.52957344e-01 1.42487183e-01 3.19920838e-01
7.05000639e-01 -5.21335661e-01 -1.58529893e-01 9.39539517e-04
-4.06716436e-01 5.17690063e-01 1.26634824e+00 8.63265812e-01
-7.93467820e-01 5.34528136e-01 8.51297665e+00 9.01766121e-01
-1.38145378e-02 3.78020763e-01 2.35638186e-01 -1.04650892e-01
-8.70189011e-01 6.76572859e-01 -7.31006682e-01 1.49488941e-01
1.02544785e+00 -3.59420896e-01 4.74054456e-01 7.30212510e-01
1.21223845e-01 -5.32024801e-01 -1.34825289e+00 4.18392956e-01
-3.42905492e-01 -1.14416885e+00 -4.28303838e-01 4.06432658e-01
1.01633668e+00 -1.89344928e-01 -5.19998193e-01 -2.22303405e-01
1.44816208e+00 -1.38336694e+00 3.31870258e-01 3.16876233e-01
6.03044808e-01 -8.17305982e-01 6.26556337e-01 5.07576227e-01
-6.91560209e-01 -1.62826866e-01 -6.51525974e-01 -3.01050246e-01
1.63895711e-01 1.17758238e+00 -8.12971234e-01 7.67802715e-01
7.25063324e-01 2.22042143e-01 -3.16230386e-01 1.51763487e+00
-7.07206428e-01 1.37862504e+00 -8.18467140e-01 2.85853073e-02
-2.26572692e-01 2.89436549e-01 6.92256570e-01 1.13908195e+00
1.63527563e-01 4.88186359e-01 3.77310142e-02 6.32062733e-01
2.81570852e-01 -4.33664769e-01 -5.71764350e-01 -4.78885099e-02
8.64082217e-01 8.27456653e-01 -9.97270763e-01 -2.22388491e-01
-2.12399915e-01 4.65419918e-01 -7.89510086e-02 3.01633179e-01
-7.11851656e-01 -1.57904148e-01 6.86535910e-02 1.22885674e-01
3.40004086e-01 -1.62758976e-01 -3.86606693e-01 -9.33574796e-01
-6.09506667e-01 -8.11633945e-01 1.01223588e+00 -4.14743036e-01
-1.51527929e+00 -1.03707261e-01 4.16713327e-01 -3.48460346e-01
-3.90374243e-01 -4.38703805e-01 -3.68324250e-01 6.82559252e-01
-9.70436990e-01 -8.42164040e-01 6.62291348e-02 8.66789401e-01
-4.31841612e-01 2.44930819e-01 1.00043142e+00 -7.74100330e-03
-3.30251694e-01 4.01740551e-01 1.31876394e-01 -2.73942411e-01
3.66341829e-01 -1.69882107e+00 -2.27173999e-01 1.04978776e+00
9.20748562e-02 6.98650241e-01 1.19107211e+00 -9.57929790e-01
-1.68661392e+00 -6.86798096e-01 1.12369227e+00 -4.05032307e-01
9.57613170e-01 -3.62959892e-01 -3.87334973e-01 1.00897217e+00
1.41985416e-01 -1.78860407e-02 9.55882072e-01 8.41972232e-01
-4.86782402e-01 4.68752980e-02 -1.07656503e+00 4.01446640e-01
1.54613078e+00 -1.39816418e-01 -5.32141924e-01 7.76354492e-01
4.88139391e-01 9.18576941e-02 -8.66120100e-01 5.36432862e-01
3.76635283e-01 -8.48664939e-01 8.13149095e-01 -7.79823542e-01
7.63205290e-02 -3.84833366e-01 -2.84239084e-01 -6.96106851e-01
-8.68060648e-01 -8.38388681e-01 -2.14370981e-01 9.84391689e-01
6.82228148e-01 -5.30159950e-01 8.63921225e-01 4.77380425e-01
4.25305218e-01 7.51723051e-02 -1.25311852e+00 -9.93269801e-01
-1.29164338e-01 -5.12878537e-01 4.93642688e-01 1.29659510e+00
2.16253117e-01 4.03666854e-01 -4.61271495e-01 7.46663094e-01
1.46070254e+00 2.25436285e-01 6.15880191e-01 -1.80824089e+00
-8.76777768e-01 -1.38208568e-01 -3.00233215e-01 -5.91861963e-01
1.48065656e-01 -7.91231453e-01 -2.11022452e-01 -1.67173362e+00
6.84336662e-01 -7.72435427e-01 -3.18599381e-02 6.40291810e-01
6.44983947e-02 6.86308295e-02 -4.61673081e-01 -1.05593666e-01
-7.12069809e-01 -1.98986188e-01 1.12552929e+00 -8.43187422e-02
2.74005562e-01 2.33690619e-01 -9.40400243e-01 7.83976972e-01
5.30720472e-01 -9.18021083e-01 -6.19487524e-01 -2.91901059e-03
9.72744942e-01 7.48099983e-01 8.94757092e-01 -7.83719063e-01
4.17144507e-01 -6.53916597e-01 3.95710737e-01 -6.35765314e-01
-2.12365583e-01 -8.28740478e-01 8.61314952e-01 5.89247763e-01
-4.08392012e-01 -4.04603958e-01 -1.63561061e-01 9.06292558e-01
2.56344914e-01 -6.95059299e-01 2.83744365e-01 -3.10125411e-01
-2.64519930e-01 1.87311038e-01 -2.72906005e-01 -6.32381141e-02
7.08409309e-01 -6.39034808e-02 -5.18310189e-01 -6.22661114e-01
-1.09027946e+00 2.22663715e-01 1.50415659e-01 -4.09282953e-01
2.22473666e-01 -9.01686251e-01 -9.64171469e-01 -3.61445755e-01
-2.12160602e-01 -1.25548914e-01 1.58039808e-01 1.00888693e+00
1.56499762e-02 4.38580602e-01 2.41440281e-01 -2.92764187e-01
-1.17407978e+00 7.56031394e-01 -5.84541336e-02 -2.34236360e-01
-1.61124870e-01 9.41604853e-01 6.10189177e-02 3.23874205e-02
2.29761750e-01 2.97247916e-01 1.77470744e-01 -2.19038799e-01
4.80882794e-01 7.03538537e-01 -3.36264789e-01 7.88215324e-02
-3.68431062e-01 -9.00531933e-02 2.29931176e-01 -2.65057534e-01
1.49098015e+00 -1.96464613e-01 -7.67513514e-01 2.64310688e-01
5.33352792e-01 4.34538692e-01 -7.29912102e-01 -1.96684733e-01
2.43553184e-02 -4.27033991e-01 -5.63446730e-02 -1.02794588e+00
-8.01377118e-01 3.09721321e-01 1.11983858e-01 8.44276965e-01
9.35071468e-01 5.41731358e-01 -1.78380534e-01 6.08683586e-01
7.12900877e-01 -7.69175529e-01 -5.78059494e-01 1.86007828e-01
6.15981162e-01 -8.01524162e-01 4.75134432e-01 -9.53745365e-01
1.39043212e-01 8.03476036e-01 2.35987589e-01 -3.81784588e-02
7.84472108e-01 6.97539330e-01 -7.89073527e-01 -4.93591249e-01
-1.02024233e+00 -2.53335148e-01 -2.44409502e-01 8.04768682e-01
1.52394101e-01 4.27853167e-01 -6.03043795e-01 7.86169648e-01
-2.80604124e-01 2.54115701e-01 7.37859547e-01 6.64553821e-01
-3.94699246e-01 -1.07568645e+00 -5.13794720e-01 7.32239068e-01
-5.82551718e-01 -3.28344405e-01 -4.07724470e-01 9.99160349e-01
2.22872108e-01 1.26491940e+00 -2.25371033e-01 -5.75419962e-02
-1.88649029e-01 -2.85525829e-01 9.85051095e-01 -4.38464165e-01
8.84840637e-02 1.62223235e-01 5.56810856e-01 -4.23668683e-01
-6.46659076e-01 -7.97287285e-01 -1.07658780e+00 -8.77514005e-01
-6.17013991e-01 4.10529315e-01 2.15263799e-01 9.42468941e-01
-2.29040548e-01 2.34959796e-01 3.88042539e-01 -6.91706613e-02
-5.87457716e-01 -7.68605232e-01 -1.21376967e+00 -3.47110093e-01
7.22450204e-03 -6.03976130e-01 -8.38809490e-01 1.45549119e-01] | [7.695193290710449, 5.279510974884033] |
4d366686-36f0-43c2-9498-681c4a8e4a2f | multilingual-text-to-speech-synthesis-for | 2305.15749 | null | https://arxiv.org/abs/2305.15749v1 | https://arxiv.org/pdf/2305.15749v1.pdf | Multilingual Text-to-Speech Synthesis for Turkic Languages Using Transliteration | This work aims to build a multilingual text-to-speech (TTS) synthesis system for ten lower-resourced Turkic languages: Azerbaijani, Bashkir, Kazakh, Kyrgyz, Sakha, Tatar, Turkish, Turkmen, Uyghur, and Uzbek. We specifically target the zero-shot learning scenario, where a TTS model trained using the data of one language is applied to synthesise speech for other, unseen languages. An end-to-end TTS system based on the Tacotron 2 architecture was trained using only the available data of the Kazakh language. To generate speech for the other Turkic languages, we first mapped the letters of the Turkic alphabets onto the symbols of the International Phonetic Alphabet (IPA), which were then converted to the Kazakh alphabet letters. To demonstrate the feasibility of the proposed approach, we evaluated the multilingual Turkic TTS model subjectively and obtained promising results. To enable replication of the experiments, we make our code and dataset publicly available in our GitHub repository. | ['Yerbolat Khassanov', 'Saida Mussakhojayeva', 'Rustem Yeshpanov'] | 2023-05-25 | null | null | null | null | ['transliteration', 'text-to-speech-synthesis', 'speech-synthesis'] | ['natural-language-processing', 'speech', 'speech'] | [-1.93077564e-01 1.11080967e-01 2.41934732e-01 -3.02740514e-01
-1.09489632e+00 -9.04625595e-01 8.16896021e-01 -6.92095399e-01
-2.68003613e-01 8.95677388e-01 1.64801329e-01 -7.81813323e-01
4.34325218e-01 -3.16706896e-01 -5.04100025e-01 -6.33914411e-01
3.97322983e-01 8.90335858e-01 -7.97011778e-02 -4.33240175e-01
-1.04892187e-01 3.31871271e-01 -1.34086621e+00 9.01559964e-02
1.06891882e+00 3.81039351e-01 2.26514503e-01 6.63840294e-01
2.46771336e-01 7.15506017e-01 -8.02457392e-01 -4.32607174e-01
3.06035459e-01 -6.47586524e-01 -8.29717994e-01 -2.56207973e-01
3.84913236e-01 -2.14235142e-01 -6.26179993e-01 6.98430598e-01
5.49896479e-01 3.98442030e-01 8.62000108e-01 -7.66901672e-01
-8.21970940e-01 9.60865438e-01 1.39012247e-01 -1.16767392e-01
1.03277065e-01 5.88239245e-02 8.59784126e-01 -1.15551627e+00
9.41153109e-01 1.42154431e+00 -7.29942471e-02 4.99091566e-01
-6.90149307e-01 -7.12647915e-01 -2.09672228e-01 4.07973319e-01
-1.48844337e+00 -1.08583629e+00 3.95191848e-01 -4.45697039e-01
8.82818699e-01 2.03806534e-01 3.52710247e-01 1.36924362e+00
-2.02431604e-01 7.38005459e-01 1.15473664e+00 -8.25216413e-01
1.96618006e-01 3.09437692e-01 -3.10664833e-01 6.72163844e-01
-3.42821419e-01 8.22640136e-02 -6.09756172e-01 1.27023578e-01
3.52483779e-01 -7.61820197e-01 -9.62652359e-03 3.09273481e-01
-1.08636129e+00 7.24869192e-01 -2.75923721e-02 4.27918047e-01
-1.46057814e-01 -2.51134336e-01 4.15802419e-01 5.24683416e-01
6.51345551e-01 4.21285555e-02 -5.18062055e-01 -3.16068232e-01
-9.00936723e-01 -3.73221822e-02 5.39804399e-01 1.10735595e+00
5.14469683e-01 7.77885497e-01 5.45062609e-02 1.47048104e+00
2.04279706e-01 8.95421445e-01 7.14623749e-01 -5.68348944e-01
7.63611615e-01 -1.38436913e-01 2.60594249e-01 -1.64730281e-01
1.39549658e-01 6.99102879e-02 -3.00785989e-01 -9.46336836e-02
3.63953322e-01 -6.58055007e-01 -1.43778348e+00 1.37946343e+00
2.96981007e-01 -2.81421244e-02 6.87313914e-01 9.72846210e-01
7.55530894e-01 1.10329652e+00 -2.09477589e-01 -3.63802135e-01
1.18737638e+00 -1.28173161e+00 -6.99756026e-01 2.25674547e-02
6.69295967e-01 -1.24756789e+00 1.24415016e+00 2.17656404e-01
-9.26372170e-01 -6.75325394e-01 -9.06711876e-01 -1.59292459e-01
-4.97043669e-01 5.62423825e-01 -4.55347374e-02 6.25519753e-01
-8.44281614e-01 5.76089695e-02 -7.32976556e-01 -4.80557472e-01
-4.29745048e-01 -2.25088567e-01 -1.90837890e-01 9.47855189e-02
-1.57430184e+00 9.18631136e-01 6.46575928e-01 9.61709097e-02
-9.92677748e-01 -1.14446029e-01 -9.46882129e-01 -1.57060698e-01
1.47717699e-01 3.33284698e-02 1.34563935e+00 -9.08366680e-01
-2.18497062e+00 6.53424561e-01 -8.87692645e-02 -1.96401462e-01
6.01090908e-01 -2.64130384e-01 -7.67695248e-01 -9.82730929e-03
5.76518737e-02 2.85909921e-01 7.99833953e-01 -1.06172180e+00
-6.30850971e-01 -2.21572354e-01 -5.40642142e-01 4.79556978e-01
5.33986501e-02 3.58562171e-01 -4.81653959e-01 -8.10163796e-01
1.02088384e-01 -1.38288510e+00 2.43308380e-01 -1.20535624e+00
-5.40013492e-01 -1.05876975e-01 6.80925250e-01 -1.33184445e+00
9.42188621e-01 -2.31825376e+00 6.45079985e-02 2.16572240e-01
-7.66505778e-01 5.14508426e-01 -2.00317457e-01 6.86184227e-01
1.32973284e-01 -1.95722684e-01 2.12960839e-02 -7.80944824e-02
-1.47196740e-01 1.81739971e-01 -4.84938592e-01 2.76933074e-01
-2.69356407e-02 7.38867640e-01 -9.51926172e-01 -4.00026292e-01
4.94065523e-01 3.44731957e-01 5.11936024e-02 1.60495684e-01
-1.23831667e-01 7.14984238e-01 -1.82582229e-01 6.41706169e-01
2.72938490e-01 5.90912282e-01 3.43738645e-01 1.51713774e-01
-3.72051626e-01 5.49233794e-01 -6.72697484e-01 1.33439445e+00
-7.48824239e-01 7.03701854e-01 -1.66936114e-01 -5.39579570e-01
1.21701646e+00 8.16361547e-01 3.87421250e-02 -8.82025540e-01
2.28198603e-01 8.23874056e-01 1.64305612e-01 -4.75391358e-01
7.49346793e-01 -6.55532405e-02 -2.17430800e-01 2.65564799e-01
3.38997036e-01 -5.90222538e-01 1.83502197e-01 -1.60085276e-01
3.52270186e-01 3.36028010e-01 1.27313420e-01 -1.35660052e-01
4.42304343e-01 -1.84582383e-03 3.51566643e-01 4.54161108e-01
-7.36215487e-02 7.51901329e-01 -2.92770602e-02 -3.83234292e-01
-1.21161044e+00 -1.38869357e+00 -1.05071375e-02 1.21052372e+00
-4.81590241e-01 -3.16011488e-01 -8.77296984e-01 -4.92640585e-01
-5.34430087e-01 1.27712822e+00 -1.24394879e-01 2.79864937e-01
-7.90746570e-01 -1.59203798e-01 8.58242333e-01 -1.75070897e-01
4.17094857e-01 -1.25695205e+00 -3.03999502e-02 3.46755475e-01
-6.41626954e-01 -1.34739220e+00 -5.31006932e-01 -1.31475348e-02
-3.56253833e-01 -6.57055795e-01 -1.06072807e+00 -1.04330564e+00
1.97616130e-01 -1.36241302e-01 4.37786579e-01 -7.01033056e-01
1.55721083e-01 -1.59406140e-01 -5.46117187e-01 -5.68921447e-01
-9.57078040e-01 1.49293274e-01 4.59782332e-01 1.12746902e-01
9.78253037e-02 -7.12083951e-02 9.45914835e-02 2.48321936e-01
-4.39219058e-01 1.92524761e-01 1.45727679e-01 5.38140714e-01
4.59363043e-01 2.83706896e-02 5.69113493e-01 -9.13513184e-01
3.01123202e-01 -2.51687825e-01 -7.68172681e-01 3.12450200e-01
-1.33317322e-01 -1.34267896e-01 9.38714325e-01 -3.51085335e-01
-1.42352486e+00 1.74181536e-01 -2.80734152e-01 -3.81368816e-01
-1.86362844e-02 5.51698625e-01 -8.07170793e-02 4.50024545e-01
5.57381988e-01 5.81109703e-01 -4.72137034e-01 -4.69395489e-01
6.88396990e-01 1.53171909e+00 7.02735364e-01 -6.52446151e-01
6.69559419e-01 -1.22721680e-01 -6.03724182e-01 -1.32101095e+00
-3.95854682e-01 -7.47340247e-02 -6.50541425e-01 -3.81326824e-01
6.10605121e-01 -1.19577157e+00 -1.24703780e-01 8.06359291e-01
-1.23043501e+00 -4.18017477e-01 8.58722404e-02 6.76861703e-01
-6.30098820e-01 7.60458261e-02 -8.33261073e-01 -1.08965230e+00
-3.56122524e-01 -1.18131053e+00 9.78410780e-01 1.26505673e-01
-3.18756580e-01 -8.99521172e-01 2.30555028e-01 6.15023196e-01
2.08422348e-01 -1.93805650e-01 1.17294288e+00 -8.37241113e-01
-9.10616666e-02 5.59862927e-02 1.94007263e-01 6.73339427e-01
3.25166911e-01 3.09542477e-01 -1.12736154e+00 -1.89591035e-01
-1.95637390e-01 -4.18126643e-01 3.37773740e-01 1.22677699e-01
3.17249238e-01 -3.32396030e-01 2.55628169e-01 2.74565369e-01
9.35378134e-01 6.45332873e-01 3.57544452e-01 3.42224129e-02
7.06315637e-01 6.30340159e-01 6.96181118e-01 1.26244888e-01
4.95965332e-01 8.41099203e-01 -4.40809101e-01 7.70983696e-02
-3.56840640e-01 -3.83805841e-01 7.69295514e-01 1.80472922e+00
2.50658812e-03 -2.82059878e-01 -1.02278018e+00 7.76831090e-01
-1.64156783e+00 -5.04067779e-01 -6.47811070e-02 2.35794234e+00
1.00278199e+00 -1.97328165e-01 -6.69400245e-02 -2.40310162e-01
7.46710718e-01 -7.84330219e-02 -9.87966731e-02 -7.58214414e-01
-2.06621647e-01 3.82619441e-01 2.51779199e-01 9.16519284e-01
-8.49073350e-01 1.86084294e+00 5.47618580e+00 1.08343530e+00
-1.50513923e+00 1.63074821e-01 5.55429757e-01 -2.14928970e-01
-9.47283655e-02 6.07836805e-02 -5.26300192e-01 5.67240775e-01
1.75087702e+00 -2.53336400e-01 8.20726693e-01 5.77522933e-01
7.05835879e-01 6.07020333e-02 -8.93151283e-01 9.42027688e-01
-3.17412689e-02 -9.45934832e-01 1.45707205e-01 -1.90177679e-01
1.07095551e+00 5.36139607e-01 1.54553607e-01 6.63303912e-01
6.69872820e-01 -7.92796135e-01 1.09477198e+00 1.78726941e-01
1.01659954e+00 -8.04578900e-01 4.89237875e-01 3.10598701e-01
-8.05412173e-01 2.45442286e-01 -3.34485054e-01 2.37940386e-01
1.97631553e-01 1.76487431e-01 -1.50392556e+00 8.54139268e-01
2.21776411e-01 5.58354370e-02 -2.83539474e-01 4.96858537e-01
-4.69064444e-01 1.20578229e+00 -2.05636069e-01 5.69884256e-02
5.96781611e-01 -3.99220884e-01 3.97619724e-01 1.15367293e+00
7.20589578e-01 -1.49659082e-01 1.59362201e-02 2.54286200e-01
-1.34185746e-01 7.84197628e-01 -7.72387207e-01 -3.21775436e-01
7.12498844e-01 8.24506342e-01 -4.60463107e-01 -6.17803216e-01
-3.14658493e-01 1.01043785e+00 2.28826955e-01 8.70845675e-01
-5.54984987e-01 -6.04419589e-01 3.91334951e-01 -3.33981991e-01
8.67042392e-02 -4.68743533e-01 -8.18270445e-02 -1.25244796e+00
-1.67982608e-01 -1.23426664e+00 8.84150267e-02 -1.00275791e+00
-8.46032798e-01 1.00441563e+00 -3.62113975e-02 -1.10775399e+00
-9.34104860e-01 -4.34191853e-01 -1.65592298e-01 1.37275195e+00
-1.05401385e+00 -1.24751794e+00 4.96798813e-01 2.64800906e-01
9.54512417e-01 -7.86921382e-01 9.40223634e-01 1.81335047e-01
-4.31809962e-01 4.20484483e-01 6.31871343e-01 3.35057884e-01
8.36274683e-01 -8.13166440e-01 7.88614154e-01 9.51708376e-01
5.01294076e-01 3.94041598e-01 6.10809743e-01 -7.54455864e-01
-9.47763026e-01 -1.20105362e+00 1.41166484e+00 -2.19611779e-01
7.40795016e-01 -5.81188500e-01 -5.87548554e-01 9.06214356e-01
3.79125267e-01 -4.95887548e-01 4.70864087e-01 2.79154647e-02
-2.75724381e-01 1.00110032e-01 -6.76875114e-01 1.14657593e+00
4.52396333e-01 -8.03657413e-01 -5.40060222e-01 4.08283561e-01
6.55033767e-01 -4.55068320e-01 -4.89474505e-01 -5.62413298e-02
6.75498068e-01 -4.48238164e-01 3.58151019e-01 -6.71234787e-01
1.47619650e-01 -2.10944891e-01 -5.26869297e-01 -1.90550721e+00
-1.81913078e-02 -8.40760529e-01 4.27051395e-01 1.24356449e+00
5.67700326e-01 -5.64105630e-01 2.91798055e-01 5.65157235e-02
-4.38141286e-01 1.44239664e-01 -1.42907906e+00 -1.02484834e+00
2.97308445e-01 -4.21101600e-01 5.15942931e-01 9.98221934e-01
-2.96774041e-02 5.48343956e-01 -7.41764367e-01 1.97587594e-01
1.39331728e-01 -1.83184862e-01 7.51311481e-01 -7.03337789e-01
-2.08959714e-01 -7.36301988e-02 -1.82471238e-02 -3.71182114e-01
4.98508692e-01 -1.00257587e+00 3.99621695e-01 -1.19411743e+00
-5.06798446e-01 -3.68659496e-01 -9.37009007e-02 4.48649734e-01
-8.96106288e-02 3.27550620e-01 4.97444361e-01 2.20369607e-01
-1.40895778e-02 6.54828608e-01 1.22692001e+00 -5.29166721e-02
-3.96180898e-01 -1.02270842e-01 -1.04672141e-01 4.93867904e-01
6.87379003e-01 -2.89143980e-01 -5.06683469e-01 -7.33505845e-01
-4.25235361e-01 3.53140622e-01 -3.91722262e-01 -9.52804923e-01
-3.04538123e-02 -3.39164466e-01 8.38575289e-02 -4.33181494e-01
5.83161891e-01 -3.82615596e-01 2.45789275e-01 2.89179087e-01
-1.19061634e-01 -2.67940253e-01 2.08357379e-01 -2.75321156e-01
-2.52152771e-01 -2.08525538e-01 9.94794786e-01 -1.38516417e-02
-5.54479957e-01 -2.05478538e-02 -8.34471583e-01 -4.37630750e-02
7.95386016e-01 9.52658281e-02 -2.20919847e-01 -5.62724888e-01
-6.18331075e-01 -4.10392880e-02 1.98672384e-01 8.05836141e-01
4.24375147e-01 -1.33245730e+00 -1.23249090e+00 5.66361606e-01
3.39002281e-01 -6.63090706e-01 1.56713516e-01 2.85886437e-01
-5.88498116e-01 8.92799795e-01 -2.49806911e-01 -1.75509453e-01
-1.05015552e+00 1.63537368e-01 1.28170818e-01 2.36750141e-01
-3.83394092e-01 4.89153385e-01 1.28352195e-01 -1.16362596e+00
-1.23768104e-02 1.26458928e-01 1.16347328e-01 1.35051804e-02
2.56225932e-02 3.27821165e-01 3.29322666e-01 -1.18111610e+00
-2.43117467e-01 1.56256899e-01 -1.14704482e-01 -9.63249624e-01
8.03994715e-01 -3.14324468e-01 1.38040081e-01 8.46779943e-01
9.68553960e-01 3.89778614e-01 -6.13982737e-01 -1.79447412e-01
-6.70036450e-02 -2.05150604e-01 -2.25896329e-01 -9.41064775e-01
-5.35517275e-01 7.58794785e-01 2.80391484e-01 -7.81223625e-02
7.34621406e-01 -1.55290574e-01 7.85984516e-01 4.35989618e-01
6.08001053e-01 -1.52767956e+00 -5.77045918e-01 1.30261695e+00
7.51510978e-01 -8.34443271e-01 -7.59877086e-01 -5.90440631e-02
-9.46366847e-01 1.22926271e+00 2.49721244e-01 2.14079157e-01
3.55746329e-01 -1.89367637e-01 8.75374675e-01 4.12783593e-01
-8.62921178e-01 -1.79623559e-01 2.35533848e-01 6.39700770e-01
8.40377510e-01 4.62432116e-01 -2.75982767e-01 2.77142495e-01
-1.04895890e+00 -1.20974079e-01 8.65872264e-01 4.96163130e-01
-1.75914854e-01 -1.05582166e+00 -5.26615560e-01 1.77958086e-01
-4.36174393e-01 -4.13165241e-01 -4.42199022e-01 6.90533102e-01
-1.22164965e-01 1.26072061e+00 -4.87932265e-02 -3.43294829e-01
1.56208858e-01 5.47529817e-01 2.36528292e-01 -7.11327434e-01
-2.66252339e-01 6.26226544e-01 5.76376319e-01 3.74149857e-03
3.76226336e-01 -6.50286555e-01 -1.10561001e+00 -1.27319902e-01
7.98081681e-02 5.29333293e-01 9.70310628e-01 1.10998905e+00
2.42929272e-02 1.90801755e-01 8.91550243e-01 -5.55150330e-01
-3.89788836e-01 -1.26838148e+00 -6.20805860e-01 1.98067218e-01
2.16452926e-01 -2.81363070e-01 -1.53839514e-01 3.57396334e-01] | [14.512297630310059, 7.051295757293701] |
d6cb044a-5daf-4a3c-a5b4-96f20cc40f6f | increasing-trustworthiness-of-deep-neural | 2007.01472 | null | https://arxiv.org/abs/2007.01472v1 | https://arxiv.org/pdf/2007.01472v1.pdf | Increasing Trustworthiness of Deep Neural Networks via Accuracy Monitoring | Inference accuracy of deep neural networks (DNNs) is a crucial performance metric, but can vary greatly in practice subject to actual test datasets and is typically unknown due to the lack of ground truth labels. This has raised significant concerns with trustworthiness of DNNs, especially in safety-critical applications. In this paper, we address trustworthiness of DNNs by using post-hoc processing to monitor the true inference accuracy on a user's dataset. Concretely, we propose a neural network-based accuracy monitor model, which only takes the deployed DNN's softmax probability output as its input and directly predicts if the DNN's prediction result is correct or not, thus leading to an estimate of the true inference accuracy. The accuracy monitor model can be pre-trained on a dataset relevant to the target application of interest, and only needs to actively label a small portion (1% in our experiments) of the user's dataset for model transfer. For estimation robustness, we further employ an ensemble of monitor models based on the Monte-Carlo dropout method. We evaluate our approach on different deployed DNN models for image classification and traffic sign detection over multiple datasets (including adversarial samples). The result shows that our accuracy monitor model provides a close-to-true accuracy estimation and outperforms the existing baseline methods. | ['Zhihui Shao', 'Shaolei Ren', 'Jianyi Yang'] | 2020-07-03 | null | null | null | null | ['traffic-sign-detection'] | ['computer-vision'] | [ 2.29881734e-01 1.27864763e-01 -1.31635368e-01 -8.13538849e-01
-9.17477846e-01 -6.66657269e-01 4.60904509e-01 -2.29326814e-01
-8.32207203e-01 9.30711865e-01 -6.40545547e-01 -7.05221534e-01
1.23517692e-01 -8.17407131e-01 -1.28594220e+00 -6.99970603e-01
8.88647884e-02 4.46495622e-01 5.31274736e-01 4.57967550e-01
5.23813516e-02 4.73885685e-01 -1.25055087e+00 2.90978521e-01
8.14354599e-01 1.67088938e+00 -4.11992192e-01 8.95446599e-01
3.77784312e-01 1.11324108e+00 -1.14885700e+00 -9.54442680e-01
2.22613618e-01 -1.57692894e-01 -4.97141570e-01 -4.10779804e-01
8.44372094e-01 -9.35610116e-01 -3.47018361e-01 1.53026032e+00
3.93788487e-01 -1.90974876e-01 5.07115841e-01 -1.72779393e+00
-4.47844148e-01 7.18656957e-01 -4.04688343e-02 2.52220690e-01
-4.29447830e-01 5.88855743e-01 8.69828045e-01 -2.65263975e-01
2.20673248e-01 9.00357008e-01 7.06791043e-01 8.13462138e-01
-1.01291025e+00 -1.12096334e+00 7.53177926e-02 3.72600317e-01
-1.27026129e+00 -6.39905155e-01 4.71037269e-01 -2.31690794e-01
4.07182604e-01 -3.41191851e-02 1.82112277e-01 1.70787680e+00
1.96132995e-02 7.37230003e-01 1.06638336e+00 1.12749487e-02
7.92860568e-01 4.53843266e-01 2.54991084e-01 4.29148197e-01
4.48945612e-01 3.43193203e-01 -3.79720777e-01 -2.67029732e-01
6.14032507e-01 -3.67308706e-01 -2.35398263e-01 -2.75400039e-02
-7.90564597e-01 6.26948535e-01 3.94159764e-01 -2.25750163e-01
-3.54128718e-01 7.36736655e-01 4.78194445e-01 3.30799967e-01
3.11524123e-01 -7.42328018e-02 -6.11127496e-01 -4.14612085e-01
-1.03813934e+00 2.80714422e-01 9.76203263e-01 1.00088680e+00
4.12781209e-01 1.24044664e-01 -3.55258405e-01 3.65876377e-01
2.70617902e-01 6.61716461e-01 1.25194788e-01 -9.37371969e-01
6.07499361e-01 1.78010657e-01 1.95770577e-01 -7.43410707e-01
2.49992713e-01 -4.45160240e-01 -7.14584589e-01 6.89593852e-01
7.22600043e-01 -3.85746539e-01 -9.66027498e-01 1.94995809e+00
2.16977119e-01 5.33949733e-01 1.04376502e-01 9.31414247e-01
6.21563077e-01 3.87910247e-01 2.72233188e-01 2.95539051e-01
1.14739645e+00 -4.59339559e-01 -4.35720205e-01 -1.75949410e-01
5.03389716e-01 -1.33519933e-01 9.84080851e-01 5.75906277e-01
-7.77036130e-01 -3.91043484e-01 -1.14617765e+00 2.12158173e-01
-2.83827871e-01 3.14288288e-01 1.83938101e-01 1.10206711e+00
-9.08254385e-01 7.41910577e-01 -9.51416254e-01 1.21020786e-01
1.11071241e+00 4.78318989e-01 -2.59665847e-01 -5.90504631e-02
-1.26520181e+00 9.60365832e-01 3.85576457e-01 3.60019296e-01
-1.44827247e+00 -6.34513021e-01 -3.67806435e-01 -8.04780647e-02
3.12560469e-01 -4.31970954e-01 1.44483244e+00 -1.02092063e+00
-1.43071973e+00 5.70816755e-01 1.99894756e-01 -8.19329262e-01
1.09142339e+00 -1.74371794e-01 -5.14235854e-01 6.27686232e-02
-3.37243289e-01 6.06203437e-01 1.10457361e+00 -1.20632625e+00
-7.16843426e-01 -8.36050063e-02 3.59990329e-01 -4.45650369e-01
-3.51618588e-01 -1.99473407e-02 -3.06589216e-01 -2.85046101e-01
-4.90085632e-01 -8.13129842e-01 1.21398069e-01 5.65410852e-01
-5.79089403e-01 -3.93572636e-02 9.39206004e-01 -6.67358756e-01
1.00833726e+00 -2.09311128e+00 -7.01169193e-01 3.73434335e-01
2.51438528e-01 6.79229856e-01 -1.65568292e-02 -2.83145636e-01
1.97741166e-01 1.90384477e-01 -3.70775610e-01 -4.40029114e-01
1.79239675e-01 5.14454126e-01 -4.08031553e-01 5.46463788e-01
4.56348807e-01 6.73193455e-01 -7.98748016e-01 -5.42583823e-01
1.40056834e-01 6.27604723e-01 -3.99435580e-01 3.02176058e-01
-2.02066839e-01 3.06217432e-01 -4.63920712e-01 4.38421190e-01
8.59890282e-01 -1.79792449e-01 -1.71273351e-01 -3.95492941e-01
4.38128769e-01 3.18789095e-01 -1.08251190e+00 9.47513878e-01
-6.00414515e-01 8.57977808e-01 -1.57768816e-01 -8.53794873e-01
8.40412199e-01 3.54037642e-01 -2.31283918e-01 -4.53378052e-01
4.71172303e-01 1.90304905e-01 1.54817134e-01 -3.84748012e-01
1.36115015e-01 3.20877917e-02 1.20720796e-01 4.42522049e-01
1.75636977e-01 3.44764709e-01 -2.20788702e-01 4.67262603e-02
1.40906417e+00 -7.78118521e-02 -3.33282083e-01 8.83250982e-02
3.21691394e-01 -3.35219473e-01 6.30089343e-01 1.05073619e+00
-6.71011806e-01 5.85658967e-01 8.61489892e-01 -2.40354165e-01
-1.05748665e+00 -1.12691271e+00 -3.15857232e-01 6.41491711e-01
-1.06789909e-01 2.85149008e-01 -1.01569724e+00 -1.29460061e+00
-2.50529405e-02 1.12322152e+00 -8.51086557e-01 -5.43728530e-01
-2.42270157e-01 -6.44152999e-01 1.19093478e+00 7.57969320e-01
9.40541446e-01 -9.85544741e-01 -6.24721348e-01 -9.53133702e-02
7.11115301e-02 -1.42795718e+00 -1.60265878e-01 5.54271117e-02
-6.93775713e-01 -1.12081611e+00 -3.28916579e-01 -1.53840482e-01
7.91975915e-01 -5.07610619e-01 1.07354307e+00 7.44369924e-02
2.00780228e-01 -7.75893182e-02 -1.29739195e-01 -5.24765134e-01
-8.86138797e-01 2.55899969e-02 6.79231808e-02 3.54899347e-01
4.79861200e-01 -5.38900971e-01 -5.81949830e-01 4.11446005e-01
-9.50080395e-01 -4.92414892e-01 5.38865864e-01 6.03105843e-01
2.47041911e-01 1.62976235e-01 5.93640089e-01 -1.08285677e+00
5.80174148e-01 -3.44977558e-01 -9.86073673e-01 3.18539798e-01
-6.35450959e-01 1.51057363e-01 7.58386791e-01 -8.82834733e-01
-7.23456740e-01 -2.61482656e-01 -2.99368769e-01 -7.26977706e-01
-3.24355006e-01 1.11341298e-01 -4.34340000e-01 -2.24776700e-01
7.11456597e-01 -7.85333887e-02 -8.68920013e-02 -2.57669210e-01
9.84842852e-02 9.23495829e-01 7.06477106e-01 -5.30439436e-01
7.09382117e-01 2.98126787e-01 -3.21852379e-02 -2.52685189e-01
-1.01925790e+00 3.17852527e-01 -2.72233486e-01 -4.02020693e-01
4.99896735e-01 -7.22430348e-01 -1.33007395e+00 8.63627076e-01
-1.29692614e+00 -6.07698441e-01 -2.21645430e-01 5.70887029e-01
-9.54361409e-02 1.08664855e-01 -5.83577752e-01 -1.03314066e+00
-3.46269011e-01 -1.14966238e+00 7.98703074e-01 -9.76573117e-03
-1.09332427e-01 -1.03060412e+00 -3.82490605e-01 2.41125613e-01
5.55998564e-01 3.43105167e-01 5.51637590e-01 -9.40326869e-01
-5.93609869e-01 -7.68732131e-01 -4.80376840e-01 1.24818933e+00
-3.42295140e-01 2.04088777e-01 -1.52735281e+00 -1.14842415e-01
2.14799773e-02 -5.38193107e-01 7.40737557e-01 3.34142417e-01
1.43276560e+00 -6.64426744e-01 1.26417145e-01 4.04165089e-01
1.26017320e+00 -1.46913543e-01 7.27273464e-01 1.28738463e-01
5.93950868e-01 2.52757400e-01 3.53287429e-01 2.97384292e-01
1.09203756e-01 3.28990132e-01 8.90274644e-01 2.16736674e-01
1.67242736e-01 -3.24393332e-01 5.51460743e-01 -3.19060385e-02
1.02502793e-01 -5.61311424e-01 -5.98652184e-01 2.83713579e-01
-1.66524374e+00 -8.52527916e-01 -2.31626093e-01 2.52669644e+00
9.80482280e-01 7.26813197e-01 -7.02576386e-03 3.41229260e-01
8.58071327e-01 -6.98551536e-02 -9.05234814e-01 -2.69906223e-01
1.61817789e-01 1.61613494e-01 9.89512265e-01 3.81176651e-01
-9.80664492e-01 6.75213099e-01 6.09832096e+00 1.01046431e+00
-1.10645270e+00 3.94400597e-01 9.39217687e-01 -1.34854928e-01
4.74463813e-02 -2.67687235e-02 -9.24555719e-01 9.54955876e-01
1.32330596e+00 2.66119540e-01 1.85651168e-01 9.54603732e-01
2.57761121e-01 -1.70253322e-01 -1.32228816e+00 7.93164015e-01
-1.70739695e-01 -1.06460798e+00 1.63266268e-02 8.42075869e-02
4.85185772e-01 1.34774745e-01 2.56257892e-01 3.10803056e-01
6.06375873e-01 -1.05500329e+00 1.03510630e+00 5.50715446e-01
1.03422415e+00 -9.38377082e-01 1.12465060e+00 5.58178842e-01
-5.07637978e-01 3.52267437e-02 -3.52791876e-01 2.17906669e-01
7.79288635e-02 9.37471509e-01 -9.17144358e-01 -6.18401431e-02
6.54472649e-01 2.28744879e-01 -4.94802892e-01 9.43198264e-01
-7.28275299e-01 1.30937874e+00 -4.90352899e-01 -2.48057723e-01
1.23949699e-01 2.12798461e-01 2.56439120e-01 9.05901313e-01
1.01222180e-01 -2.96027392e-01 -3.93727124e-01 1.31558180e+00
-5.53700209e-01 -6.38087809e-01 -2.51427740e-01 1.70578197e-01
6.82029724e-01 1.17160928e+00 -2.91720361e-01 -5.86539984e-01
-5.87112457e-02 8.72026622e-01 3.29003066e-01 4.10416692e-01
-1.37358356e+00 -3.59733194e-01 5.92061758e-01 -9.60453749e-02
3.79472762e-01 2.17216432e-01 -3.35775107e-01 -7.45749474e-01
4.53380108e-01 -7.57805169e-01 1.39656633e-01 -7.45519638e-01
-1.31464291e+00 6.93861306e-01 -1.40662327e-01 -1.24567413e+00
-1.56286865e-01 -8.93488228e-01 -7.09398746e-01 8.65810871e-01
-1.52837777e+00 -1.20041251e+00 -2.49193668e-01 4.84283835e-01
-1.20296359e-01 -9.88408402e-02 5.73336601e-01 5.36980152e-01
-8.00311506e-01 1.31816649e+00 -4.55822609e-03 6.69856966e-01
5.89490414e-01 -1.03244042e+00 5.55788100e-01 1.10929346e+00
-1.63774326e-01 2.17006624e-01 8.16875458e-01 -3.89339834e-01
-9.58619893e-01 -1.53212726e+00 6.04135036e-01 -8.39223683e-01
7.18708575e-01 -3.13099563e-01 -1.01863956e+00 7.44948149e-01
-4.08824563e-01 6.84437811e-01 4.51748878e-01 -2.19089761e-01
-5.25042057e-01 -5.30960977e-01 -1.62259185e+00 3.37086946e-01
7.62830436e-01 -6.11149192e-01 -2.31783107e-01 -6.84378762e-03
4.51310545e-01 -4.26518440e-01 -7.74823785e-01 2.04752326e-01
7.20701277e-01 -9.29120541e-01 5.72465181e-01 -7.05619752e-01
2.59872198e-01 -2.28037015e-01 -7.89062455e-02 -9.93118167e-01
3.19417447e-01 -2.99992323e-01 -4.21140730e-01 1.49060619e+00
6.35408998e-01 -7.33159065e-01 1.03240991e+00 1.20044088e+00
2.62024909e-01 -3.77130896e-01 -1.27591431e+00 -8.87166381e-01
-2.44560037e-02 -1.11952555e+00 7.24137485e-01 7.11811721e-01
-7.24403203e-01 -1.05575703e-01 -4.22979444e-01 7.36362994e-01
9.05263126e-01 -5.25937319e-01 7.77624071e-01 -9.43816125e-01
-4.34052765e-01 -3.04719716e-01 -7.60774672e-01 -9.07718897e-01
3.12984735e-01 -5.37900329e-01 2.35033378e-01 -8.80597889e-01
-2.18849257e-02 -4.36909944e-01 -3.82557213e-01 6.00958824e-01
2.34774463e-02 3.25273991e-01 -2.11569786e-01 -2.63067223e-02
-5.70450604e-01 3.36097419e-01 8.29247355e-01 5.93377976e-03
2.02984244e-01 6.28709733e-01 -4.82400417e-01 5.34170091e-01
6.78293526e-01 -8.90687644e-01 -2.93386996e-01 -4.59804386e-01
1.93225667e-01 -2.62511849e-01 1.17935324e+00 -1.26270676e+00
3.81004632e-01 9.74015519e-02 2.68337399e-01 -2.34971985e-01
1.41782805e-01 -1.04764390e+00 -8.69318619e-02 3.66886348e-01
-5.11218905e-01 -3.63938540e-01 1.13389745e-01 6.95718050e-01
-2.85067060e-03 -4.45681334e-01 9.72310483e-01 2.85872161e-01
-4.12206024e-01 3.94228369e-01 -1.88111037e-01 1.46237671e-01
7.63837397e-01 -1.68905795e-01 -3.42266828e-01 -3.99266779e-01
-4.25955921e-01 1.58252060e-01 2.27671847e-01 -2.28753574e-02
5.96191823e-01 -1.28626037e+00 -7.10510492e-01 8.04351196e-02
1.36396378e-01 1.60622820e-01 7.60890022e-02 6.66333437e-01
-4.46899772e-01 -1.93412960e-01 6.95365667e-02 -6.65968597e-01
-1.01053762e+00 3.33297163e-01 6.82124078e-01 -5.09188585e-02
-2.43581638e-01 8.82484794e-01 -3.22100371e-01 -2.97171354e-01
6.43023491e-01 -6.51387095e-01 1.79599240e-01 -4.30597395e-01
6.15614712e-01 3.84623975e-01 2.13283762e-01 -2.85214245e-01
-2.95640767e-01 -6.25440036e-04 -4.66183908e-02 -2.85684675e-01
1.02141678e+00 2.99986273e-01 2.94797421e-01 4.35084194e-01
1.26985335e+00 -4.91854280e-01 -1.73720193e+00 -1.96354941e-01
-2.81519383e-01 -5.27402043e-01 4.28921133e-01 -9.72129881e-01
-1.49352407e+00 1.16124773e+00 7.95424640e-01 5.51499240e-02
9.10025597e-01 -2.30491832e-01 7.33170867e-01 5.13501108e-01
2.59618014e-01 -1.10987341e+00 -3.50282252e-01 9.90229622e-02
4.18751866e-01 -1.47388840e+00 -3.08451027e-01 1.17239945e-01
-6.21155798e-01 9.55799758e-01 6.75635099e-01 -5.57300076e-02
7.41947234e-01 3.86994958e-01 1.94629610e-01 1.64767012e-01
-8.29602420e-01 2.48631418e-01 9.36925188e-02 6.67781353e-01
-8.98792595e-02 -4.68204953e-02 3.85040343e-01 6.83869064e-01
-1.02727346e-01 3.41561288e-01 3.94526720e-01 7.25788236e-01
-2.16418222e-01 -7.36245871e-01 -2.37755731e-01 6.62608683e-01
-5.50933242e-01 -1.52230263e-01 -1.62835330e-01 5.98385930e-01
7.65866712e-02 9.11281288e-01 2.19827816e-01 -4.79124069e-01
4.55799073e-01 -2.68325005e-02 1.43593892e-01 -7.43898153e-02
-4.94313985e-01 -7.71093488e-01 2.85841793e-01 -5.07819831e-01
-1.39946252e-01 -5.20068169e-01 -9.95908141e-01 -7.31552482e-01
-4.40029740e-01 -1.56440705e-01 7.80272543e-01 1.07935226e+00
1.98471636e-01 3.43090445e-01 6.61952674e-01 -4.19583380e-01
-1.27395427e+00 -1.11795056e+00 -4.25333291e-01 3.67331684e-01
4.62913781e-01 -4.08491731e-01 -7.50987589e-01 -2.88455188e-03] | [5.658976078033447, 7.603898525238037] |
1647bf0d-aa97-4920-82bd-0d3db8d1f251 | m-genseg-domain-adaptation-for-target | 2212.07276 | null | https://arxiv.org/abs/2212.07276v1 | https://arxiv.org/pdf/2212.07276v1.pdf | M-GenSeg: Domain Adaptation For Target Modality Tumor Segmentation With Annotation-Efficient Supervision | Automated medical image segmentation using deep neural networks typically requires substantial supervised training. However, these models fail to generalize well across different imaging modalities. This shortcoming, amplified by the limited availability of annotated data, has been hampering the deployment of such methods at a larger scale across modalities. To address these issues, we propose M-GenSeg, a new semi-supervised training strategy for accurate cross-modality tumor segmentation on unpaired bi-modal datasets. Based on image-level labels, a first unsupervised objective encourages the model to perform diseased to healthy translation by disentangling tumors from the background, which encompasses the segmentation task. Then, teaching the model to translate between image modalities enables the synthesis of target images from a source modality, thus leveraging the pixel-level annotations from the source modality to enforce generalization to the target modality images. We evaluated the performance on a brain tumor segmentation datasets composed of four different contrast sequences from the public BraTS 2020 challenge dataset. We report consistent improvement in Dice scores on both source and unannotated target modalities. On all twelve distinct domain adaptation experiments, the proposed model shows a clear improvement over state-of-the-art domain-adaptive baselines, with absolute Dice gains on the target modality reaching 0.15. | ['Samuel Kadoury', 'Eugene Vorontsov', "Malo Alefsen de Boisredon d'Assier"] | 2022-12-14 | null | null | null | null | ['tumor-segmentation', 'brain-tumor-segmentation'] | ['computer-vision', 'medical'] | [ 8.57923865e-01 2.64697462e-01 -3.42894375e-01 -6.02898955e-01
-1.38696241e+00 -6.51915193e-01 6.08108103e-01 -1.59464985e-01
-6.49945796e-01 8.39655936e-01 8.16762596e-02 -1.85868219e-01
2.81927347e-01 -3.01489711e-01 -5.80516338e-01 -1.07513440e+00
3.19346875e-01 5.27363062e-01 2.98472494e-01 9.78487134e-02
-1.77580640e-01 3.15361083e-01 -6.97881579e-01 4.41773653e-01
1.05034626e+00 1.03954542e+00 3.62611771e-01 5.64738512e-01
-5.32937422e-03 4.90596235e-01 -2.59894311e-01 -2.38283113e-01
2.83173203e-01 -7.73443639e-01 -9.32201505e-01 3.26389223e-01
7.89255559e-01 -4.00750875e-01 -3.59752089e-01 1.26615560e+00
5.15361607e-01 -1.97519958e-01 9.74287868e-01 -1.12073255e+00
-6.01589203e-01 5.05493402e-01 -6.77349567e-01 2.27152914e-01
-2.57276326e-01 4.56186175e-01 6.92579091e-01 -4.15199697e-01
9.52215254e-01 6.55185759e-01 6.73405051e-01 9.07955468e-01
-1.48057508e+00 -6.81895792e-01 -1.49849892e-01 -1.34092674e-01
-1.06294155e+00 -3.50763649e-01 6.45850182e-01 -6.13114953e-01
6.01933718e-01 1.50711417e-01 3.14905912e-01 1.29047668e+00
1.36987314e-01 8.97031069e-01 1.48068285e+00 -2.30396956e-01
3.74091528e-02 -1.01936311e-01 7.30222790e-03 6.48797154e-01
-6.88004121e-03 2.22546622e-01 -2.75726318e-01 7.48406723e-02
6.17496729e-01 -3.51198852e-01 -3.89895260e-01 -4.89654124e-01
-1.62209213e+00 6.01332664e-01 7.32034326e-01 3.83422971e-01
-3.74524057e-01 -1.15597717e-01 6.15459681e-01 6.45035459e-03
4.97943252e-01 4.08422530e-01 -3.86864126e-01 1.98781192e-01
-1.25442123e+00 -1.24216065e-01 3.01404625e-01 6.60825312e-01
1.30837634e-01 -7.54454583e-02 -4.38577324e-01 9.14968550e-01
1.71078295e-01 5.43026626e-01 7.05734849e-01 -8.97795677e-01
3.73913586e-01 4.38531727e-01 -2.83418417e-01 -3.90831947e-01
-6.23950183e-01 -6.02232158e-01 -1.13623655e+00 3.67711037e-01
1.00799632e+00 -1.72113225e-01 -1.65175831e+00 2.10382795e+00
3.73740196e-01 1.57089502e-01 1.09149262e-01 9.73745644e-01
8.73714447e-01 1.09375291e-01 4.42984879e-01 -1.83086276e-01
1.18506444e+00 -1.16113639e+00 -5.36817908e-01 -2.34607279e-01
7.70980000e-01 -5.63784480e-01 1.04284811e+00 2.32683912e-01
-9.70329821e-01 -2.57446408e-01 -1.02784550e+00 8.80507678e-02
-2.95204282e-01 8.93975496e-02 4.80391741e-01 6.77914977e-01
-1.05406857e+00 2.73989826e-01 -1.16997564e+00 -3.15007478e-01
9.85541642e-01 2.99252450e-01 -5.88640690e-01 -3.30571443e-01
-1.05416393e+00 9.57095027e-01 5.61952353e-01 -7.56614432e-02
-1.16426480e+00 -1.10511088e+00 -6.62116587e-01 -3.40458691e-01
7.03903586e-02 -7.60303378e-01 1.18842900e+00 -1.46459770e+00
-1.24533153e+00 1.38413584e+00 1.56915799e-01 -5.76167405e-01
9.78800058e-01 3.54937822e-01 -4.01426584e-01 5.39908946e-01
2.45311856e-01 1.24825108e+00 7.26486683e-01 -1.26065135e+00
-5.32879353e-01 -4.54516292e-01 -5.59310436e-01 3.76233459e-01
-5.13249598e-02 -2.47640774e-01 -4.27026480e-01 -7.56103575e-01
3.28158289e-02 -1.11500120e+00 -1.99944764e-01 2.41880253e-01
-6.72351718e-01 2.70568848e-01 7.74003625e-01 -9.53941166e-01
5.25496364e-01 -2.21360636e+00 2.06560075e-01 8.24664310e-02
3.63332868e-01 7.29626343e-02 -3.35687637e-01 -5.37558258e-01
-3.95844817e-01 -1.18584462e-01 -8.81593227e-01 -4.16721702e-01
-2.32529834e-01 2.40622818e-01 1.39704794e-01 6.77234530e-01
1.20380461e-01 9.84085560e-01 -8.69719148e-01 -6.97570443e-01
1.09389357e-01 2.87176996e-01 -3.39699626e-01 1.17866658e-01
-1.74685434e-01 1.13099992e+00 -2.00313002e-01 6.90643966e-01
7.39329338e-01 -3.48434031e-01 8.05917531e-02 -3.31333518e-01
4.94467914e-01 -1.38460189e-01 -3.58338892e-01 2.15094304e+00
-1.99149251e-01 5.92790246e-01 1.38733864e-01 -1.30465925e+00
4.19917494e-01 4.55901235e-01 9.14267838e-01 -8.01657259e-01
1.48572803e-01 4.32294577e-01 3.44660074e-01 -4.33212608e-01
-1.39851039e-02 -4.89059657e-01 -6.30972683e-02 2.35548988e-01
2.67173320e-01 -3.53502542e-01 1.66714683e-01 2.05620557e-01
1.04893863e+00 -6.93367980e-03 -5.75093403e-02 -1.83122292e-01
3.57411444e-01 3.43720675e-01 5.45145690e-01 6.75384760e-01
-7.41049230e-01 1.08107364e+00 4.08610642e-01 -4.73174043e-02
-1.03806186e+00 -1.43120694e+00 -2.91515261e-01 8.88303578e-01
-7.50596449e-02 5.39790392e-01 -9.63808894e-01 -1.27136612e+00
-1.92022443e-01 7.04763949e-01 -1.02920020e+00 -2.77125746e-01
-3.24074388e-01 -1.10416925e+00 9.33418810e-01 5.78324378e-01
6.71289206e-01 -8.27269197e-01 -4.83810633e-01 -1.77380759e-02
-5.18805444e-01 -1.52511263e+00 -6.65322721e-01 2.80037224e-01
-1.02778649e+00 -1.06284058e+00 -1.36628032e+00 -8.91567588e-01
9.44249213e-01 -2.53487647e-01 1.19862485e+00 -1.72164589e-01
-3.25662374e-01 2.88531721e-01 -1.37160480e-01 -8.84165987e-02
-8.42974782e-01 2.18246132e-01 -2.26558119e-01 -1.00630932e-01
9.18590799e-02 -3.09888989e-01 -6.72969520e-01 3.64369869e-01
-1.09632051e+00 5.78286171e-01 8.40634465e-01 1.12962043e+00
7.81654060e-01 -2.95649827e-01 6.89302027e-01 -9.28072512e-01
3.51289630e-01 -3.66612047e-01 -3.39184612e-01 3.84122282e-01
-3.73637557e-01 -1.45312443e-01 3.50402474e-01 -6.31702244e-01
-1.17787504e+00 3.61763686e-01 -9.10501927e-03 -2.80608624e-01
-4.09869224e-01 4.73774344e-01 -4.50630076e-02 -1.09034684e-03
8.09436023e-01 4.57062304e-01 4.12306696e-01 6.26758933e-02
4.26963180e-01 4.85206395e-01 9.75630283e-01 -4.59664106e-01
5.47625244e-01 6.80817306e-01 1.52958343e-02 -4.92312551e-01
-8.16641033e-01 -2.49514267e-01 -8.30347955e-01 -2.75726020e-01
1.39129734e+00 -7.70805001e-01 -5.64348698e-02 8.78245056e-01
-7.67448962e-01 -8.37763309e-01 -2.41108298e-01 5.10881424e-01
-7.10514128e-01 1.62267312e-01 -4.98965830e-01 7.13560581e-02
-3.09916109e-01 -1.61576724e+00 9.90727842e-01 2.60745764e-01
-2.67908484e-01 -1.21185863e+00 8.10887516e-02 7.71596730e-01
4.59339112e-01 6.24373794e-01 9.09162581e-01 -1.00652456e+00
-3.28945965e-01 -1.01800829e-01 -4.62128550e-01 5.72346568e-01
3.91003937e-01 -2.78675437e-01 -8.32226694e-01 -3.11056137e-01
-3.64571661e-01 -6.37401760e-01 9.95772719e-01 6.97767496e-01
1.20020592e+00 3.20924759e-01 -4.64241713e-01 7.16882527e-01
1.01669145e+00 -1.00748911e-02 4.61120784e-01 1.94923803e-01
6.78154647e-01 4.08672780e-01 3.82195294e-01 -1.42334461e-01
1.63361222e-01 4.62096095e-01 3.53713989e-01 -5.96450090e-01
-7.42812157e-01 -3.66246291e-02 -1.30073987e-02 4.09395665e-01
3.47464889e-01 -3.35169137e-02 -1.19764483e+00 7.85406291e-01
-1.38565087e+00 -5.62529087e-01 9.18895155e-02 1.92936909e+00
1.24371910e+00 8.16991180e-02 1.18157238e-01 -2.84306049e-01
6.22223735e-01 -9.40240771e-02 -1.02793205e+00 1.62964344e-01
-2.67726034e-01 -1.71760358e-02 7.65085042e-01 2.97539771e-01
-1.34151065e+00 7.86768854e-01 6.19703102e+00 9.76645172e-01
-1.47125518e+00 4.36877251e-01 1.24432933e+00 -1.62748843e-02
2.46110130e-02 -5.19218147e-01 -2.29065105e-01 4.81617093e-01
8.41470122e-01 1.23604499e-01 1.58887699e-01 4.70127195e-01
1.27520099e-01 -2.54141271e-01 -1.21060300e+00 7.60242462e-01
5.03028370e-02 -1.29391479e+00 -1.02121547e-01 8.46879482e-02
1.21165752e+00 4.71807718e-01 4.56879556e-01 1.92602441e-01
2.53972501e-01 -1.22428143e+00 4.55918789e-01 4.54940379e-01
1.16058600e+00 -3.66622567e-01 6.84466422e-01 1.04788944e-01
-5.98806143e-01 4.28407878e-01 1.67764559e-01 7.50181675e-01
2.16821641e-01 4.96283114e-01 -1.08456016e+00 4.38019902e-01
4.95408624e-01 6.41626894e-01 -6.09087229e-01 9.95325148e-01
-2.07563154e-02 6.56839252e-01 -3.21710169e-01 6.22814298e-01
3.70081872e-01 4.95213866e-02 3.76264274e-01 1.24667227e+00
1.00384839e-01 -1.26732484e-01 1.22309789e-01 8.42694461e-01
-3.72598201e-01 -1.91871673e-01 -2.24808753e-01 -1.93193421e-01
2.65498571e-02 1.38320971e+00 -1.02886200e+00 -6.02009892e-01
-3.69118392e-01 1.08578992e+00 7.70699885e-03 4.32645947e-01
-1.08073938e+00 -1.54514592e-02 3.00519943e-01 -5.20398654e-02
6.61179945e-02 2.76899748e-02 -8.53690207e-01 -9.91543472e-01
-3.09726596e-01 -1.03463864e+00 6.20004475e-01 -7.29136050e-01
-1.50849950e+00 7.05617845e-01 5.71773108e-03 -1.21073544e+00
-1.05366446e-01 -6.34599805e-01 -4.05773669e-01 9.70031381e-01
-1.49860072e+00 -1.35988128e+00 -4.31971490e-01 6.82345271e-01
3.94245952e-01 -2.71495461e-01 7.38311410e-01 3.77009571e-01
-4.22270387e-01 8.83572638e-01 1.58927232e-01 3.66291851e-01
1.11625493e+00 -1.30084586e+00 -2.48886477e-02 8.76490831e-01
-1.34247541e-01 4.51892912e-02 5.50407112e-01 -6.47873580e-01
-7.14039803e-01 -1.11422801e+00 1.88889518e-01 -4.61925656e-01
7.38496065e-01 1.29714429e-01 -9.92134213e-01 6.60032928e-01
4.09004718e-01 5.08380055e-01 7.20826983e-01 -4.45390522e-01
-2.69254088e-01 1.84242219e-01 -1.54510212e+00 5.44267356e-01
7.07126796e-01 -4.69553828e-01 -4.63745803e-01 4.90351707e-01
6.12277150e-01 -8.25601876e-01 -1.02242756e+00 5.66869855e-01
4.21806693e-01 -6.39218926e-01 1.01220608e+00 -7.02681363e-01
7.86406279e-01 -1.04638591e-01 -1.23843074e-01 -1.46733356e+00
1.31522015e-01 -9.18044895e-02 3.03726375e-01 8.83059263e-01
7.08733380e-01 -6.04087055e-01 9.05338526e-01 7.07401931e-01
-3.58148456e-01 -6.40868962e-01 -1.11135674e+00 -6.01177216e-01
7.47554362e-01 -3.46625268e-01 1.53418139e-01 1.17660141e+00
-1.58799574e-01 -2.00274047e-02 -4.94573079e-02 1.20582312e-01
8.10141325e-01 -1.82101488e-01 3.68413985e-01 -6.90695226e-01
-3.51429254e-01 -8.12003136e-01 -2.13316292e-01 -7.81275451e-01
2.87896782e-01 -1.35294294e+00 2.79043198e-01 -1.34071839e+00
5.36160350e-01 -5.19185364e-01 -5.12171388e-01 5.77900052e-01
-2.20210463e-01 7.87342072e-01 7.51032215e-03 1.42347574e-01
-4.46244955e-01 3.32020730e-01 1.84123433e+00 -5.73053896e-01
9.87179354e-02 -2.12132290e-01 -6.36858940e-01 6.30173147e-01
9.30741012e-01 -4.17335093e-01 -3.69721681e-01 -5.29403210e-01
-5.76358795e-01 1.44733086e-01 4.52576578e-01 -9.79083061e-01
5.32893501e-02 -9.73325744e-02 6.58924282e-01 -3.04378420e-01
1.66806936e-01 -7.75388002e-01 -2.28886232e-02 5.47224045e-01
-6.75138116e-01 -3.47484261e-01 4.64423925e-01 4.40996110e-01
-2.04977870e-01 -7.94381741e-03 1.30028701e+00 2.33239606e-02
-5.29285550e-01 4.53184724e-01 -1.41534939e-01 4.14810419e-01
1.05737495e+00 -1.70468241e-01 -3.50322962e-01 -3.22668143e-02
-1.22688639e+00 1.43718541e-01 4.08416420e-01 2.83531725e-01
3.81891906e-01 -1.20316613e+00 -7.85854280e-01 1.88780017e-03
2.41353139e-01 -2.41927672e-02 4.98697311e-01 1.56143951e+00
-4.56095517e-01 2.69711763e-01 -4.98264700e-01 -1.07110608e+00
-1.22981310e+00 1.91208661e-01 8.64749551e-01 -5.87863445e-01
-4.77383405e-01 9.05956268e-01 5.30690968e-01 -6.76846385e-01
5.02284206e-02 -3.95604163e-01 2.16427311e-01 -2.52026498e-01
3.44064862e-01 -1.67648867e-01 1.70558363e-01 -8.09907317e-01
-3.63341510e-01 2.55859256e-01 -3.25659811e-01 -2.78482527e-01
9.92390931e-01 -7.25549236e-02 1.52782708e-01 7.66921714e-02
1.42082739e+00 -5.06109059e-01 -1.58230805e+00 -4.31446373e-01
-2.24687383e-01 -1.75512239e-01 3.32003623e-01 -1.51098788e+00
-1.49584794e+00 8.08921635e-01 1.12875319e+00 -2.53823102e-01
1.31156647e+00 6.52546138e-02 9.07398999e-01 -1.17702708e-01
-2.65239067e-02 -9.03310597e-01 2.25668103e-02 2.42119819e-01
6.09733284e-01 -1.74568522e+00 -3.88557792e-01 -2.59526759e-01
-9.99442875e-01 7.61202991e-01 6.11917317e-01 1.76830828e-01
3.29420328e-01 3.14445972e-01 4.37863022e-01 -7.08581507e-02
-2.95386165e-01 -1.65319275e-02 5.31694531e-01 8.55323672e-01
4.09650147e-01 2.66688049e-01 1.36431539e-02 4.29323524e-01
1.29017442e-01 1.67276740e-01 3.08325708e-01 6.57709062e-01
-5.98199144e-02 -9.61881042e-01 -3.31586510e-01 5.68535864e-01
-5.34676611e-01 -2.15143822e-02 -2.84934044e-01 8.64022911e-01
1.03731096e-01 4.72026795e-01 5.76675963e-03 4.02571335e-02
9.49049443e-02 5.88909211e-03 6.32601202e-01 -3.48220944e-01
-3.12921166e-01 3.70297618e-02 -1.70044094e-01 -4.19758976e-01
-5.31675935e-01 -7.90878654e-01 -1.44622231e+00 1.14507027e-01
1.00044414e-01 -2.62859613e-01 5.84753573e-01 1.04025865e+00
9.93861482e-02 8.88881147e-01 2.20857367e-01 -7.37579107e-01
-4.21553016e-01 -9.63238120e-01 -3.56209517e-01 7.26292551e-01
3.08157980e-01 -7.03427315e-01 -8.35426822e-02 4.25975293e-01] | [14.558452606201172, -2.101393222808838] |
4f86edd1-96a5-4563-8e10-c4483499896c | a-new-level-set-based-protocol-for-accurate | 1505.03093 | null | http://arxiv.org/abs/1505.03093v1 | http://arxiv.org/pdf/1505.03093v1.pdf | A new Level-set based Protocol for Accurate Bone Segmentation from CT Imaging | In this work it is proposed a medical image segmentation pipeline for
accurate bone segmentation from CT imaging. It is a two-step methodology, with
a pre-segmentation step and a segmentation refinement step. First, the user
performs a rough segmenting of the desired region of interest. Next, a fully
automatic refinement step is applied to the pre-segmented data. The automatic
segmentation refinement is composed by several sub-stpng, namely image
deconvolution, image cropping and interpolation. The user-defined
pre-segmentation is then refined over the deconvolved, cropped, and up-sampled
version of the image. The algorithm is applied in the segmentation of CT images
of a composite femur bone, reconstructed with different reconstruction
protocols. Segmentation outcomes are validated against a gold standard model
obtained with coordinate measuring machine Nikon Metris LK V20 with a digital
line scanner LC60-D that guarantees an accuracy of 28 $\mu m$. High sub-pixel
accuracy models were obtained for all tested Datasets. The algorithm is able to
produce high quality segmentation of the composite femur regardless of the
surface meshing strategy used. | ['Manuel Pinheiro', 'J. L. Alves'] | 2015-05-12 | null | null | null | null | ['image-deconvolution', 'image-cropping'] | ['computer-vision', 'computer-vision'] | [ 5.79517365e-01 3.10554355e-01 5.84077299e-01 -3.14327925e-01
-8.18548858e-01 6.48792312e-02 9.20756906e-02 5.47020316e-01
-8.97828937e-01 5.17471254e-01 -2.78011054e-01 -1.78585857e-01
-5.15289009e-02 -9.50596333e-01 -3.34108353e-01 -5.64411044e-01
1.80611163e-01 1.24552178e+00 8.21033180e-01 -1.14959981e-02
2.59824187e-01 8.35508049e-01 -1.43106914e+00 2.64611896e-02
6.62817538e-01 7.66008914e-01 4.99933213e-01 7.93925405e-01
-1.64098293e-01 2.96177208e-01 -8.01689252e-02 5.95384799e-02
2.07575768e-01 -4.55707282e-01 -1.06249487e+00 6.10645771e-01
-3.07281762e-01 -1.11649320e-01 6.11140907e-01 8.80193591e-01
3.59937251e-01 -2.97171205e-01 7.78250694e-01 -4.90564972e-01
3.78122866e-01 3.58564168e-01 -5.82303345e-01 -6.17650338e-02
1.33811966e-01 1.20667510e-01 3.48740906e-01 -8.66681099e-01
6.80156291e-01 8.71571124e-01 6.13190532e-01 3.42487603e-01
-1.60867500e+00 -3.53596032e-01 -8.64926875e-01 -2.35000238e-01
-1.37708294e+00 9.63298529e-02 5.40493906e-01 -9.85758781e-01
4.71868515e-01 1.19254492e-01 9.45930064e-01 -4.16918658e-02
4.74756658e-01 3.20586972e-02 1.52935994e+00 -6.53131604e-01
5.82681775e-01 -2.76688635e-02 3.87380660e-01 5.24928391e-01
2.91902602e-01 -1.27354845e-01 1.58265784e-01 3.18609290e-02
1.04885328e+00 -1.19198553e-01 -2.59805262e-01 -2.28036880e-01
-9.98729289e-01 5.31070888e-01 3.33245099e-01 7.79977381e-01
-8.49536359e-01 -9.99481902e-02 2.92951524e-01 4.58906293e-02
4.64284360e-01 1.28027320e-01 -1.77285716e-01 1.46114707e-01
-1.38517094e+00 6.55473694e-02 5.98649919e-01 3.33845645e-01
9.84456718e-01 -2.08805561e-01 2.43451968e-01 6.43577993e-01
5.66410184e-01 1.86915770e-01 5.50270498e-01 -7.23593056e-01
-2.67545789e-01 8.39372158e-01 5.80042833e-03 -5.50756037e-01
-5.88552594e-01 -8.00828561e-02 -6.20922029e-01 8.01838994e-01
6.69482529e-01 7.21602067e-02 -1.35477912e+00 7.93794572e-01
7.71718800e-01 -1.68898746e-01 -2.52093405e-01 7.95627117e-01
5.21902025e-01 3.59227866e-01 1.31415710e-01 -3.31646383e-01
1.63663352e+00 -4.20864880e-01 -5.03105462e-01 2.24167347e-01
4.87074643e-01 -9.37385142e-01 9.22963440e-01 5.72233200e-01
-1.32681477e+00 -6.15528703e-01 -1.05229974e+00 5.05922101e-02
5.45180626e-02 2.16250140e-02 -1.26298770e-01 6.25119746e-01
-1.05622005e+00 6.83671594e-01 -1.14937186e+00 -2.70567209e-01
3.37147743e-01 5.71271777e-01 -4.94547158e-01 3.52387801e-02
-7.22996175e-01 7.15666234e-01 5.84515154e-01 1.27442285e-01
-5.92304647e-01 -7.08968222e-01 -6.64360464e-01 -1.06865950e-01
3.95412780e-02 -6.25543892e-01 1.04321992e+00 -1.12773836e+00
-1.64898264e+00 1.38247550e+00 4.88886498e-02 -4.79789406e-01
8.29741359e-01 1.28337264e-01 -3.05668917e-02 5.22315979e-01
3.03789288e-01 4.22735006e-01 8.29498947e-01 -1.58319831e+00
-5.96767962e-01 -5.46745181e-01 -5.32776713e-01 6.04639910e-02
5.14969230e-01 1.75651144e-02 -5.91240227e-01 -3.10025692e-01
5.43166757e-01 -7.52564847e-01 -5.58537900e-01 -2.08635941e-01
-2.69352794e-01 2.71067172e-01 6.55009925e-01 -1.06780148e+00
9.75732684e-01 -1.94218278e+00 -8.81605670e-02 7.19249070e-01
1.63339123e-01 2.20197048e-02 4.18793768e-01 1.42973155e-01
-2.92895645e-01 -1.92444965e-01 -8.69928777e-01 -3.91071826e-01
-4.83505398e-01 1.27703637e-01 4.18598890e-01 6.43603921e-01
-4.67953205e-01 3.97463739e-01 -6.09329998e-01 -9.30576921e-01
3.96847934e-01 6.06806219e-01 -2.60823876e-01 1.58163622e-01
-2.18590572e-01 8.45332444e-01 -2.77631015e-01 3.84252578e-01
7.57716835e-01 -1.96867976e-02 2.45487187e-02 -4.40668650e-02
-4.88481849e-01 -3.19628716e-01 -1.34656692e+00 1.70535171e+00
-2.86773771e-01 9.92294550e-02 5.76419234e-01 -8.79780412e-01
9.54634607e-01 6.75362408e-01 9.38979268e-01 -3.91122133e-01
3.94415051e-01 7.28889108e-01 -5.33637740e-02 -4.93888676e-01
3.83713573e-01 -6.39102757e-01 5.99376976e-01 6.27644539e-01
-1.32791981e-01 -6.99441850e-01 1.63485497e-01 -1.02916889e-01
5.79893529e-01 2.00130299e-01 2.82488137e-01 -7.99463868e-01
8.97896349e-01 4.78031695e-01 1.75983682e-01 1.37644321e-01
1.21666774e-01 9.65966344e-01 3.17633629e-01 -2.79848754e-01
-1.22225595e+00 -7.48324275e-01 -6.69082940e-01 2.38561362e-01
-9.72682051e-03 1.67436600e-01 -1.52953851e+00 -1.46424606e-01
-2.29519397e-01 3.99656355e-01 -5.56137443e-01 4.43730682e-01
-6.64863467e-01 -6.97645545e-01 4.92742844e-02 -3.98951881e-02
3.79981786e-01 -1.14701343e+00 -1.14764762e+00 5.36784768e-01
3.32042985e-02 -6.82443500e-01 -7.00486219e-03 9.21960324e-02
-1.38395011e+00 -1.15524995e+00 -9.76256251e-01 -9.04267967e-01
8.74634445e-01 -3.03300560e-01 1.02279246e+00 5.60025990e-01
-4.46960777e-01 3.30346048e-01 -1.45225137e-01 -1.04353629e-01
-8.27196777e-01 -1.30173013e-01 -4.54087198e-01 2.98253484e-02
1.22824591e-02 -4.76150811e-01 -8.17121327e-01 3.02985400e-01
-1.14961040e+00 2.96475947e-01 6.30336940e-01 5.25792778e-01
1.24640048e+00 2.75803030e-01 -6.28005480e-03 -1.10529816e+00
3.35649282e-01 -1.73186064e-01 -7.13049054e-01 4.48572896e-02
-7.80843377e-01 -2.20055114e-02 2.80902624e-01 7.84996971e-02
-1.03451562e+00 4.60466474e-01 -5.91875196e-01 5.61394468e-02
-6.42318726e-01 5.44012606e-01 -2.06714850e-02 -8.28414559e-02
5.65946221e-01 1.66137695e-01 4.96871680e-01 -6.40945554e-01
-1.01616435e-01 4.94998395e-01 6.45775378e-01 -4.02992398e-01
5.75339437e-01 7.68799901e-01 2.72648185e-01 -1.09198475e+00
2.01643094e-01 -5.99369347e-01 -1.08918369e+00 -6.05260432e-01
1.32738793e+00 -3.07978570e-01 -2.46662781e-01 7.49724984e-01
-8.73362005e-01 -6.43355429e-01 -4.36252654e-01 5.71787536e-01
-5.30235946e-01 5.34122407e-01 -7.24622846e-01 -5.00774086e-01
-5.66759050e-01 -1.53875375e+00 9.30393100e-01 2.19467357e-01
-1.37334198e-01 -8.95874798e-01 2.90649414e-01 4.50239956e-01
2.32261255e-01 4.83368933e-01 8.01015317e-01 -2.95318633e-01
-3.47316772e-01 -2.43590489e-01 7.73326755e-02 3.33835065e-01
1.05461940e-01 1.62529916e-01 -6.99029863e-01 1.00168251e-01
3.67004722e-01 3.61213721e-02 5.92402875e-01 7.50133216e-01
7.98180640e-01 4.73029613e-01 -4.10555422e-01 3.16477001e-01
1.99531937e+00 2.61169702e-01 8.71850908e-01 4.49304521e-01
3.21386278e-01 6.05114937e-01 8.22621286e-01 1.28266156e-01
8.40746686e-02 5.30261636e-01 4.46517408e-01 -4.51373488e-01
-1.58279970e-01 3.34152281e-01 -1.75082818e-01 3.96888047e-01
-3.65353435e-01 4.03886795e-01 -1.30903792e+00 6.64138854e-01
-1.23871613e+00 -4.83601362e-01 -9.81427550e-01 2.37811995e+00
9.25205827e-01 3.43819469e-01 2.40458205e-01 6.85945928e-01
7.52911091e-01 -5.68265915e-01 -1.29819037e-02 -3.55665654e-01
4.18574750e-01 9.95179594e-01 4.47395653e-01 1.07029271e+00
-5.72853506e-01 6.56709015e-01 5.51602983e+00 6.38515472e-01
-1.40736127e+00 2.30030924e-01 6.89404905e-01 4.57956791e-01
-1.52637452e-01 -7.73903504e-02 -3.41191918e-01 4.25976187e-01
8.41697574e-01 2.08163932e-01 -6.26706854e-02 5.92770040e-01
3.72945517e-01 -1.02100015e+00 -5.89438975e-01 5.33559918e-01
-3.63444805e-01 -1.09617209e+00 -2.33249456e-01 1.48773521e-01
4.48542446e-01 -7.18015060e-02 -6.14854515e-01 -3.85302395e-01
6.59867525e-02 -1.01502478e+00 7.91966975e-01 6.34614110e-01
8.32603574e-01 -7.05106914e-01 7.34029830e-01 3.35745871e-01
-1.15239847e+00 4.61303532e-01 5.20872287e-02 2.13286757e-01
5.66752553e-01 8.91707242e-01 -9.64416027e-01 6.41768396e-01
5.48024654e-01 1.56826363e-03 -2.89250702e-01 1.02545106e+00
-6.38251677e-02 5.48065305e-01 -4.74558860e-01 5.34625411e-01
7.34805390e-02 -7.20644832e-01 3.56328458e-01 1.17177606e+00
1.96279988e-01 3.67443502e-01 -1.03278659e-01 7.38832653e-01
5.82772374e-01 4.42468703e-01 1.16014376e-01 5.97521484e-01
-1.57423802e-02 1.23759973e+00 -1.57270467e+00 -4.23684448e-01
-1.53238267e-01 9.29455101e-01 -1.81841105e-01 8.42107162e-02
-4.35410440e-01 -2.62217782e-02 -7.16506764e-02 7.81958997e-01
5.93416318e-02 -1.82666868e-01 -5.13352752e-01 -4.25261766e-01
-3.16151857e-01 -4.13535118e-01 3.04683387e-01 -7.86109090e-01
-5.11470377e-01 7.20172822e-01 1.96022853e-01 -8.33042502e-01
-1.27715871e-01 -5.06420255e-01 -5.13132215e-01 1.14044094e+00
-9.63143766e-01 -9.53630626e-01 -2.54921526e-01 4.83520359e-01
4.26527023e-01 4.02918190e-01 7.17625201e-01 2.91353226e-01
-2.74084687e-01 -3.71200144e-01 -7.69038796e-02 9.88217369e-02
1.04502313e-01 -1.50429809e+00 -2.40541086e-01 9.45887864e-01
-5.18649936e-01 1.30086109e-01 8.82312357e-01 -9.43889201e-01
-7.37196982e-01 -8.57584059e-01 7.75487781e-01 1.85986683e-01
2.40297005e-01 1.54737860e-01 -1.09225190e+00 5.29169977e-01
4.33858000e-02 1.46422395e-02 4.01515275e-01 -7.87904680e-01
6.65366173e-01 1.09430723e-01 -1.53009033e+00 1.62099481e-01
6.17811866e-02 -1.15226824e-02 -4.95002121e-01 -3.71699408e-02
4.42104861e-02 -7.09234118e-01 -1.38547635e+00 2.96444029e-01
5.72531819e-01 -1.11812198e+00 8.99564862e-01 2.23905057e-01
4.38548833e-01 -5.14151156e-01 2.80968040e-01 -8.43013644e-01
-1.85812265e-02 -2.45417729e-01 7.93412507e-01 9.39056993e-01
3.15511554e-01 -5.40215373e-01 8.02590370e-01 6.91806674e-01
-1.94626063e-01 -6.74461186e-01 -1.06487870e+00 -2.63168663e-01
1.62993327e-01 -3.96911681e-01 1.60263777e-01 6.09683871e-01
-1.69249937e-01 -9.82762352e-02 5.06891131e-01 1.76059484e-01
9.95792270e-01 -8.47321749e-02 6.40817642e-01 -1.50371718e+00
-3.99253611e-03 -3.68956119e-01 -4.41950321e-01 -4.33275491e-01
-3.87943059e-01 -5.84433675e-01 4.88622822e-02 -1.79677880e+00
-1.53029129e-01 -5.39086461e-01 2.27748960e-01 -1.88356824e-02
-3.03851776e-02 3.42375696e-01 -3.45979124e-01 5.46080172e-01
3.62806469e-01 -1.99473903e-01 1.34642529e+00 3.22572857e-01
-3.40586483e-01 1.20575547e-01 6.64169481e-03 7.25710213e-01
7.15549290e-01 -5.29155850e-01 -2.99638808e-01 -5.89409992e-02
-3.13757509e-01 2.89973319e-01 4.35539335e-01 -1.32638538e+00
5.68636926e-03 1.16106972e-01 4.50059175e-02 -8.78342271e-01
1.08696602e-01 -1.40461743e+00 1.04971278e+00 1.13989556e+00
2.46295668e-02 -2.51695842e-01 -1.18065186e-01 1.52379587e-01
-1.84240386e-01 -5.50998449e-01 1.51050889e+00 -4.86537188e-01
-4.14599389e-01 6.98540509e-02 -5.13913393e-01 -4.35321540e-01
1.30208325e+00 -7.95311570e-01 6.07696474e-01 3.32938060e-02
-1.17208862e+00 -1.23091683e-01 8.43388975e-01 -7.36114621e-01
3.86216313e-01 -9.09312665e-01 -8.49770546e-01 2.43143007e-01
-2.87771940e-01 5.72363734e-01 4.00943488e-01 1.26023316e+00
-1.38372695e+00 5.67232398e-03 -2.98465788e-01 -1.05522060e+00
-1.39839065e+00 3.36461216e-01 7.74361193e-01 -5.32601357e-01
-8.36617172e-01 4.93356735e-01 -1.37313679e-01 2.55560372e-02
-3.94432485e-01 -8.03572118e-01 -4.03084099e-01 -1.70580279e-02
2.03219205e-01 6.00652814e-01 3.97431701e-01 -1.16051090e+00
-1.78747654e-01 9.34517443e-01 3.76786292e-01 -6.76933408e-01
1.40314019e+00 -2.00665206e-01 -3.92756760e-01 3.90377223e-01
8.58342648e-01 -6.96599036e-02 -1.17220461e+00 -4.27248552e-02
1.78574294e-01 -1.30534276e-01 2.65228420e-01 -5.67226052e-01
-1.21559441e+00 7.10009694e-01 7.64866054e-01 3.04512173e-01
1.12063575e+00 -1.18236028e-01 7.22596705e-01 -5.84551632e-01
3.78571361e-01 -1.20706451e+00 -4.01346475e-01 1.28222600e-01
8.26784253e-01 -7.98121870e-01 2.70150155e-01 -7.49886394e-01
-2.49685600e-01 1.38327765e+00 1.24197915e-01 -1.72823802e-01
1.05418217e+00 4.04249340e-01 2.38490328e-01 -6.46232128e-01
1.41611770e-01 -3.29519957e-01 1.82241887e-01 3.90150815e-01
6.37307286e-01 3.82973775e-02 -9.72834587e-01 2.70197421e-01
-1.08296104e-01 4.11234647e-01 4.00075972e-01 1.02380335e+00
-6.82726324e-01 -1.07010186e+00 -9.67144966e-01 2.25755110e-01
-6.57569051e-01 1.52582675e-01 3.96590903e-02 1.12047470e+00
2.37808853e-01 7.52586544e-01 1.72940835e-01 4.55812812e-02
4.15275753e-01 3.10139684e-03 3.99227917e-01 -6.83552265e-01
-8.34093869e-01 4.82357502e-01 -1.84276685e-01 -3.95577371e-01
-5.73779285e-01 -8.76748502e-01 -1.90238357e+00 8.36537555e-02
-2.18167856e-01 2.49369621e-01 1.02061152e+00 9.59642231e-01
-2.18072340e-01 4.49095309e-01 4.54273969e-01 -1.12110174e+00
9.64815617e-02 -9.55642343e-01 -9.75965083e-01 5.68525195e-01
3.57428975e-02 -4.88596946e-01 -2.46929094e-01 5.46856999e-01] | [14.031861305236816, -2.669954538345337] |
05431761-488a-460a-a7e6-dce96e89f31e | the-emotion-is-not-one-hot-encoding-learning | 2206.07359 | null | https://arxiv.org/abs/2206.07359v2 | https://arxiv.org/pdf/2206.07359v2.pdf | The Emotion is Not One-hot Encoding: Learning with Grayscale Label for Emotion Recognition in Conversation | In emotion recognition in conversation (ERC), the emotion of the current utterance is predicted by considering the previous context, which can be utilized in many natural language processing tasks. Although multiple emotions can coexist in a given sentence, most previous approaches take the perspective of a classification task to predict only a given label. However, it is expensive and difficult to label the emotion of a sentence with confidence or multi-label. In this paper, we automatically construct a grayscale label considering the correlation between emotions and use it for learning. That is, instead of using a given label as a one-hot encoding, we construct a grayscale label by measuring scores for different emotions. We introduce several methods for constructing grayscale labels and confirm that each method improves the emotion recognition performance. Our method is simple, effective, and universally applicable to previous systems. The experiments show a significant improvement in the performance of baselines. | ['Joosung Lee'] | 2022-06-15 | null | null | null | null | ['emotion-recognition-in-conversation'] | ['natural-language-processing'] | [ 2.74691880e-01 -2.04195172e-01 -1.29189074e-01 -9.37170327e-01
-6.94494963e-01 -4.61732209e-01 2.64189839e-01 3.20698798e-01
-4.00112629e-01 5.79790950e-01 2.19160795e-01 8.59492868e-02
5.34861028e-01 -6.21192157e-01 -1.71972737e-01 -6.96915746e-01
1.86819613e-01 7.20378458e-02 -1.61396742e-01 -9.23587903e-02
2.07730964e-01 3.28950793e-01 -1.45221663e+00 8.15520227e-01
7.36401916e-01 1.43578100e+00 -1.35988733e-02 5.93241215e-01
-4.81954426e-01 1.09250891e+00 -9.66591835e-01 -4.02437687e-01
-1.73471883e-01 -8.75923872e-01 -9.62552369e-01 2.61204362e-01
4.23934124e-02 1.03866465e-01 3.43965083e-01 1.08621418e+00
2.93673754e-01 1.76401109e-01 7.63172925e-01 -1.29407156e+00
-3.02760899e-01 3.28506857e-01 -5.55646062e-01 -1.56662673e-01
6.09749019e-01 -5.65621316e-01 1.10382259e+00 -8.34667742e-01
4.07420516e-01 1.24322212e+00 4.98366982e-01 5.67788064e-01
-7.68624306e-01 -6.13273323e-01 4.65269744e-01 2.89346665e-01
-1.18443620e+00 -1.63240761e-01 1.04882252e+00 -2.31275842e-01
7.84430027e-01 3.09958875e-01 5.11644065e-01 8.58475149e-01
2.18100980e-01 8.24753046e-01 1.60078311e+00 -7.48007298e-01
4.57665116e-01 3.49205285e-01 2.36440107e-01 7.47905314e-01
-6.23867750e-01 -5.25147378e-01 -4.81317818e-01 -4.86691706e-02
3.77775878e-02 9.08245891e-02 -3.61561403e-02 1.07504919e-01
-1.02083480e+00 7.65596747e-01 2.68884152e-01 4.73863602e-01
-3.82001281e-01 -1.28950834e-01 4.64847505e-01 5.54283977e-01
7.71868289e-01 3.24417979e-01 -4.33781058e-01 -4.55378145e-01
-7.88297534e-01 -3.25421482e-01 8.95095587e-01 4.56820995e-01
1.06547999e+00 -3.17974955e-01 -8.99568796e-02 1.23858762e+00
1.03890777e-01 2.59821713e-01 6.51984394e-01 -9.49977517e-01
1.18187949e-01 7.12056339e-01 6.86201081e-02 -1.31208622e+00
-5.56785882e-01 -3.65145877e-02 -6.59653127e-01 2.96945333e-01
1.58757254e-01 -3.93593699e-01 -7.52691984e-01 1.80865991e+00
2.52800822e-01 2.36999556e-01 3.31211478e-01 8.17451358e-01
6.98033988e-01 9.27336335e-01 2.81115711e-01 -4.78476971e-01
1.37849832e+00 -1.27119434e+00 -1.01128960e+00 -3.81252795e-01
8.77093554e-01 -8.07553589e-01 1.09543502e+00 7.06161916e-01
-5.58744907e-01 -4.68812048e-01 -1.11530733e+00 8.64258558e-02
-6.92259133e-01 3.52309734e-01 7.50021160e-01 7.63328493e-01
-9.23419893e-01 3.69586706e-01 -4.39485312e-01 -2.97559917e-01
-2.07923636e-01 1.42360955e-01 -2.54821479e-01 -1.31673161e-02
-1.27336860e+00 1.13308334e+00 2.18459919e-01 8.50962773e-02
-2.43873999e-01 1.27163544e-01 -7.68730521e-01 3.25639322e-02
1.75396219e-01 1.64748311e-01 1.27636039e+00 -1.59623408e+00
-1.86819947e+00 8.72662485e-01 -5.34027696e-01 -4.63821329e-02
-3.45128104e-02 1.46316234e-02 -7.97189653e-01 3.16993892e-01
-1.54894888e-01 8.39328110e-01 5.96501768e-01 -1.36501169e+00
-7.80620277e-01 -8.29519704e-02 1.39516875e-01 5.05792022e-01
-3.97941977e-01 3.94399256e-01 -4.77626354e-01 -4.21452343e-01
2.70084471e-01 -8.65905702e-01 -1.41905576e-01 -2.68575013e-01
-1.56893022e-02 -4.56920177e-01 7.64262438e-01 -5.39636314e-01
1.22234666e+00 -2.37069750e+00 -9.83464569e-02 2.22340465e-01
-1.74610674e-01 -8.84071589e-02 -1.02418035e-01 1.24691173e-01
-1.20208956e-01 2.68938780e-01 -1.97662622e-01 -4.82089013e-01
1.64451972e-01 2.56384075e-01 -2.49298140e-01 1.32756278e-01
3.70570064e-01 5.81231117e-01 -8.77191842e-01 -8.58457386e-01
5.05177006e-02 3.70763153e-01 -2.84834236e-01 3.75412583e-01
-5.09187877e-02 1.44626155e-01 -3.71474713e-01 6.79726422e-01
4.66227978e-01 -1.36708573e-01 4.65513557e-01 -1.43372059e-01
1.16681218e-01 2.85109252e-01 -1.16347158e+00 1.48215270e+00
-8.05931389e-01 5.13713658e-01 -1.68321103e-01 -1.10645568e+00
1.27891576e+00 5.52381516e-01 4.06374395e-01 -5.12078702e-01
1.90994009e-01 1.89264268e-01 -2.05402732e-01 -4.65108395e-01
6.07908547e-01 -5.34000337e-01 -6.23682916e-01 6.87416255e-01
3.10105495e-02 -2.33121201e-01 1.81310833e-01 3.04031372e-03
9.59877729e-01 -1.49985120e-01 4.39298004e-01 2.20095530e-01
5.63878059e-01 -1.77494362e-01 7.59389222e-01 3.62701952e-01
-3.69514763e-01 5.00562429e-01 8.52621436e-01 -3.88962924e-01
-5.57383478e-01 -6.29755199e-01 -3.07536107e-02 1.34514272e+00
1.61771730e-01 -5.25317073e-01 -7.45396852e-01 -1.00552249e+00
-5.28986037e-01 4.65839773e-01 -6.40699625e-01 -1.62286133e-01
-2.54331291e-01 -6.81032836e-01 3.80918413e-01 3.88410389e-01
5.03636122e-01 -1.35438025e+00 -5.93267381e-01 2.19994724e-01
-6.02138162e-01 -1.08750713e+00 -3.99039090e-01 3.90110672e-01
-5.67539275e-01 -6.07322156e-01 -2.81904459e-01 -1.01121783e+00
6.88185334e-01 2.36572936e-01 1.13089001e+00 2.87986517e-01
6.60829321e-02 2.55319327e-01 -7.87664354e-01 -2.62038797e-01
-3.29692364e-01 -1.85001150e-01 -1.09688975e-01 3.10490936e-01
6.08581305e-01 -1.03200734e-01 -3.38599563e-01 2.23595604e-01
-7.40596771e-01 3.25356668e-04 3.94849271e-01 9.07399356e-01
4.83463317e-01 3.65298331e-01 8.34010303e-01 -8.57177258e-01
8.51270437e-01 -3.64854157e-01 6.68378919e-02 6.57157898e-01
-5.11937141e-01 -1.40675902e-01 6.48247957e-01 -4.43847656e-01
-1.25738025e+00 3.36759031e-01 -2.95386046e-01 -7.47603402e-02
-4.25659567e-01 6.03933871e-01 -2.35256001e-01 9.45315957e-02
1.27168640e-01 -2.07834125e-01 -2.33459681e-01 -9.01982263e-02
3.70811135e-01 1.14377797e+00 3.21193993e-01 -6.95107818e-01
-4.36973050e-02 1.08110614e-01 -3.98368388e-01 -5.21825910e-01
-1.11617672e+00 -4.79234040e-01 -5.14728248e-01 -6.96174800e-01
9.94471788e-01 -7.40568757e-01 -7.72470415e-01 3.86043936e-01
-1.15255702e+00 -1.53076902e-01 7.02341646e-02 5.72065115e-01
-2.37605318e-01 3.63807976e-01 -7.79560208e-01 -1.05346310e+00
-1.26291871e-01 -1.17336142e+00 1.00906789e+00 4.06669170e-01
-4.49726552e-01 -1.10908139e+00 -5.25178388e-02 2.17472538e-01
2.02841982e-01 1.15970045e-01 8.35636139e-01 -5.66903472e-01
3.03294957e-01 -3.39124829e-01 -1.52140319e-01 4.72075075e-01
1.98320821e-01 7.81920999e-02 -1.16031051e+00 2.92734131e-02
2.49503568e-01 -8.48226607e-01 7.17431903e-01 -2.21120745e-01
1.27084589e+00 -1.14138663e-01 -6.53070733e-02 3.23964730e-02
1.19913793e+00 6.39473855e-01 6.98196113e-01 9.04040709e-02
2.82632500e-01 8.84481907e-01 9.08631444e-01 4.14419949e-01
5.19122064e-01 5.33938885e-01 4.43860255e-02 -2.83407360e-01
3.89560044e-01 5.46203367e-02 5.21148086e-01 1.11280847e+00
1.46709278e-01 -3.57754827e-01 -6.51891053e-01 1.55661300e-01
-1.74045336e+00 -9.17953551e-01 2.99984962e-01 1.78075671e+00
1.22213030e+00 5.05193807e-02 -1.71938717e-01 1.77619562e-01
8.40977073e-01 2.02496931e-01 -3.46386552e-01 -1.04443014e+00
-1.41247600e-01 2.11360320e-01 -2.76106209e-01 5.57369471e-01
-1.16785598e+00 1.06767094e+00 6.89171839e+00 5.92445910e-01
-1.55869234e+00 7.88743496e-02 1.15563762e+00 -9.00039151e-02
-8.91344547e-02 -1.65080830e-01 -4.31776524e-01 4.31509018e-01
9.79903400e-01 1.39890283e-01 3.18898797e-01 8.85589838e-01
-3.06510478e-02 -4.20756608e-01 -8.78686011e-01 1.06283200e+00
4.77458030e-01 -6.61722124e-01 -3.76087368e-01 -4.09341186e-01
5.46719432e-01 -5.09083331e-01 -1.69167459e-01 6.33753896e-01
2.21068338e-01 -8.68957877e-01 5.49909890e-01 5.22311628e-01
5.77212095e-01 -8.63628924e-01 9.72310066e-01 1.48553535e-01
-1.01686120e+00 1.64772391e-01 -3.05257887e-01 -3.40616167e-01
-7.94583187e-03 5.44502437e-01 -6.57670736e-01 3.41184050e-01
6.80353284e-01 5.70643604e-01 -5.64003527e-01 4.64264244e-01
-3.86194497e-01 6.44464016e-01 -1.88581169e-01 -4.26391482e-01
2.00482249e-01 -2.34730452e-01 -2.92067260e-01 1.59961200e+00
4.06902045e-01 2.00998560e-01 6.48782253e-01 2.23048896e-01
-3.11320722e-01 6.06984556e-01 -2.67852753e-01 -1.51921099e-03
3.84526014e-01 1.66488898e+00 -1.03618968e+00 -6.75116301e-01
-5.73727965e-01 1.38138628e+00 5.61263561e-01 1.66328058e-01
-8.32296431e-01 -6.11773133e-01 3.63553822e-01 -8.92594814e-01
4.08602096e-02 1.55057153e-02 -4.03381258e-01 -1.08959508e+00
1.44760638e-01 -8.55008006e-01 4.15132463e-01 -8.58851671e-01
-1.23379529e+00 9.33900595e-01 -4.74127710e-01 -1.36067522e+00
-3.08522284e-01 -6.94989324e-01 -6.76119268e-01 6.72632754e-01
-1.46213770e+00 -6.21314406e-01 -3.58380497e-01 3.80060405e-01
3.94892961e-01 1.14280485e-01 1.28416002e+00 1.48031324e-01
-5.95908105e-01 5.46150684e-01 -8.46452117e-02 6.81607053e-02
1.20817304e+00 -1.41213953e+00 -3.83943349e-01 6.41662300e-01
2.55927145e-01 4.00958627e-01 5.57391286e-01 -4.18232739e-01
-9.19597328e-01 -6.55465722e-01 1.27147639e+00 9.51725245e-02
5.69623709e-01 -3.47953588e-01 -9.20535207e-01 2.32198283e-01
5.46364725e-01 3.21312845e-02 1.19490957e+00 4.40556258e-01
-5.51627755e-01 -2.14839950e-01 -1.18920350e+00 4.67290491e-01
4.63297397e-01 -7.85165071e-01 -5.58162868e-01 1.56246766e-01
5.25755823e-01 -2.55935580e-01 -8.94261777e-01 2.74454713e-01
5.69519460e-01 -8.88591766e-01 4.20246720e-01 -4.30707723e-01
5.86012483e-01 -3.35618347e-01 -2.80703843e-01 -1.59418261e+00
-4.45359349e-02 -1.51611760e-01 2.85828561e-01 1.26940989e+00
6.53051555e-01 -4.21019375e-01 4.98570770e-01 8.19671154e-01
-1.08701549e-01 -8.47464323e-01 -7.45412946e-01 -2.90484816e-01
-8.20496008e-02 -4.97156769e-01 6.05409861e-01 1.21387184e+00
7.60980487e-01 5.27068079e-01 -5.85962474e-01 6.16381504e-02
1.07151305e-03 5.16227484e-01 3.53212774e-01 -1.04845428e+00
-1.12321116e-01 -3.81742299e-01 -2.79570043e-01 -8.58746886e-01
4.93742138e-01 -7.79000342e-01 5.62390029e-01 -1.47191012e+00
1.75124705e-01 -3.70546311e-01 -7.77331769e-01 7.56426096e-01
-3.55826944e-01 5.27552962e-01 2.44187817e-01 -4.11324129e-02
-1.05606842e+00 3.95673156e-01 9.56111729e-01 -2.10962862e-01
2.14835536e-02 -2.73403227e-01 -6.19391799e-01 7.75530219e-01
1.05935752e+00 -4.42296773e-01 -2.21611634e-01 -5.70679493e-02
2.80521035e-01 1.43129155e-01 -1.37210891e-01 -9.32821989e-01
2.03884110e-01 -3.11024636e-01 4.28501099e-01 -4.56563890e-01
5.25892913e-01 -9.01900470e-01 -2.05188066e-01 6.33774400e-02
-5.64002395e-01 1.60195366e-01 -1.31462350e-01 2.10247397e-01
-7.24541306e-01 -5.50905585e-01 7.74156749e-01 -9.48425457e-02
-1.00610757e+00 -2.20130086e-01 -5.23966908e-01 -1.97483048e-01
9.28682089e-01 7.58662224e-02 -2.12721691e-01 -5.75551152e-01
-7.49908805e-01 1.43846169e-01 3.55406970e-01 4.70043659e-01
6.84946418e-01 -1.42490447e+00 -2.83782989e-01 -2.25295443e-02
2.29590520e-01 -4.79464918e-01 2.09894031e-01 4.64447230e-01
-1.33586779e-01 1.42270168e-02 -2.15518132e-01 -2.56443858e-01
-1.49424767e+00 4.32940096e-01 3.17596018e-01 -3.71083289e-01
-1.30691722e-01 5.93629837e-01 -9.50383581e-03 -5.27904212e-01
3.45366627e-01 -1.71724409e-01 -6.18992686e-01 3.65412027e-01
6.69400930e-01 -1.38954282e-01 1.10645525e-01 -7.86893845e-01
-3.01681548e-01 6.37333453e-01 2.40954477e-03 -2.41168752e-01
1.07118046e+00 -2.33297959e-01 -4.66871113e-01 1.07470632e+00
1.32906413e+00 5.19154519e-02 -9.75547731e-01 -2.43500292e-01
1.29161477e-01 -1.85892910e-01 -7.55300885e-03 -1.07635713e+00
-1.03790498e+00 8.93584609e-01 5.73578358e-01 3.98442745e-01
1.46283126e+00 -1.15358971e-01 6.41836107e-01 5.22547007e-01
3.18917811e-01 -1.48744440e+00 3.26516658e-01 7.17075229e-01
6.72633767e-01 -1.35526800e+00 -2.21994802e-01 -3.32228750e-01
-1.20995212e+00 1.25383961e+00 6.80114150e-01 2.42671609e-01
5.86255372e-01 3.94052058e-01 7.35087335e-01 3.01738549e-02
-8.67195010e-01 -1.87703848e-01 2.06349060e-01 1.63208023e-01
9.19732928e-01 3.45425695e-01 -5.10657370e-01 5.89276195e-01
-2.02639684e-01 -1.78065404e-01 4.06810731e-01 1.02228343e+00
-6.42260551e-01 -1.41838753e+00 -2.38282293e-01 3.72249901e-01
-6.03537321e-01 4.75819744e-02 -5.84722459e-01 1.26958936e-01
2.31176987e-01 1.51196420e+00 6.14037551e-02 -8.32058191e-01
1.51556134e-01 6.66535258e-01 7.91778788e-02 -5.98396897e-01
-6.13623977e-01 1.15532175e-01 3.40487838e-01 -3.42997015e-01
-7.87443757e-01 -5.81860125e-01 -1.52138662e+00 4.53137085e-02
-1.99219912e-01 5.55511236e-01 7.63187706e-01 1.02897131e+00
1.76762998e-01 4.79120255e-01 1.01999927e+00 -4.73335564e-01
-1.09582350e-01 -1.07114303e+00 -7.20424473e-01 7.22127259e-01
-4.71722335e-02 -6.13646626e-01 -4.25941348e-01 1.23958565e-01] | [13.137177467346191, 5.88057804107666] |
14958995-d7cb-43e6-b993-a85e30f93ef4 | disparate-censorship-undertesting-a-source-of | 2208.01127 | null | https://arxiv.org/abs/2208.01127v1 | https://arxiv.org/pdf/2208.01127v1.pdf | Disparate Censorship & Undertesting: A Source of Label Bias in Clinical Machine Learning | As machine learning (ML) models gain traction in clinical applications, understanding the impact of clinician and societal biases on ML models is increasingly important. While biases can arise in the labels used for model training, the many sources from which these biases arise are not yet well-studied. In this paper, we highlight disparate censorship (i.e., differences in testing rates across patient groups) as a source of label bias that clinical ML models may amplify, potentially causing harm. Many patient risk-stratification models are trained using the results of clinician-ordered diagnostic and laboratory tests of labels. Patients without test results are often assigned a negative label, which assumes that untested patients do not experience the outcome. Since orders are affected by clinical and resource considerations, testing may not be uniform in patient populations, giving rise to disparate censorship. Disparate censorship in patients of equivalent risk leads to undertesting in certain groups, and in turn, more biased labels for such groups. Using such biased labels in standard ML pipelines could contribute to gaps in model performance across patient groups. Here, we theoretically and empirically characterize conditions in which disparate censorship or undertesting affect model performance across subgroups. Our findings call attention to disparate censorship as a source of label bias in clinical ML models. | ['Jenna Wiens', 'Michael W. Sjoding', 'Trenton Chang'] | 2022-08-01 | null | null | null | null | ['machine-learning', 'machine-learning'] | ['methodology', 'miscellaneous'] | [ 3.64122391e-01 2.73850173e-01 -9.15261626e-01 -8.49463284e-01
-6.33311927e-01 -9.07284677e-01 2.70014375e-01 8.42839897e-01
-3.11793417e-01 7.10715652e-01 5.80247939e-01 -1.05967224e+00
-2.99223810e-01 -4.40957725e-01 -5.87381482e-01 -3.64248157e-01
-1.27153397e-01 6.91694736e-01 -6.54599488e-01 6.76381767e-01
3.25313210e-02 4.79883432e-01 -6.37130499e-01 4.80093718e-01
9.84462738e-01 4.18704301e-01 -7.36131072e-01 5.11274695e-01
7.34626427e-02 1.01474023e+00 -7.14527547e-01 -3.95992845e-01
6.40209019e-02 -7.66546011e-01 -3.17417771e-01 -1.70942949e-04
4.06456798e-01 -4.40960020e-01 -5.73552027e-02 6.53971314e-01
5.07798016e-01 -4.56163824e-01 1.17060208e+00 -1.28855455e+00
-7.42578745e-01 9.12490547e-01 -2.76951939e-01 -1.14525855e-01
-3.52063752e-03 4.26636636e-01 9.24288988e-01 -3.55799079e-01
5.81033111e-01 1.15915322e+00 9.90432203e-01 5.65430164e-01
-1.67006767e+00 -7.09590673e-01 7.73604289e-02 -3.16844672e-01
-1.07687461e+00 -4.18792546e-01 3.26238722e-01 -1.05088913e+00
1.19562492e-01 5.69190323e-01 5.76401770e-01 1.02972627e+00
8.25908124e-01 2.54336208e-01 1.16619468e+00 -1.12937517e-01
5.60493946e-01 3.26563150e-01 5.35222173e-01 3.35465640e-01
9.19101834e-01 4.48109746e-01 -1.98053166e-01 -1.15820444e+00
6.45642161e-01 6.11416817e-01 -2.49273404e-01 -3.00059110e-01
-1.17313766e+00 1.02592564e+00 3.21302921e-01 -2.67400593e-01
-1.99220106e-01 -9.57585499e-02 3.41218561e-01 3.05547528e-02
2.62215436e-01 5.54411232e-01 -5.75945079e-01 3.77761960e-01
-1.10896039e+00 6.79376051e-02 8.91854584e-01 6.95962310e-01
1.36904344e-01 -5.11335373e-01 -6.46889210e-01 6.76435351e-01
2.18985036e-01 3.68507415e-01 3.93630683e-01 -8.52304518e-01
1.00055940e-01 6.45061731e-01 3.56754124e-01 -3.98872256e-01
-7.09415674e-01 -8.38184714e-01 -7.06517577e-01 4.14425105e-01
7.88918376e-01 -3.99255425e-01 -9.46110547e-01 1.77911234e+00
-6.48506656e-02 -2.23921314e-01 -4.61019836e-02 7.97627032e-01
3.00473630e-01 -1.29392296e-01 7.34600484e-01 -1.75932959e-01
1.46560907e+00 -3.70365113e-01 -5.65882027e-01 -3.93646002e-01
1.08290756e+00 -5.62537789e-01 8.83384049e-01 2.14477777e-01
-9.70547378e-01 2.51852691e-01 -7.37301767e-01 1.37863651e-01
4.78631668e-02 -1.59670904e-01 6.34001672e-01 6.41526341e-01
-4.86016095e-01 8.72138858e-01 -6.58633828e-01 -4.86889094e-01
1.00192392e+00 1.59158573e-01 -4.33556698e-02 -1.24448709e-01
-8.76103818e-01 7.66877174e-01 -1.84950039e-01 6.56981096e-02
-9.40182507e-01 -1.19484925e+00 -4.99590844e-01 8.56880620e-02
1.04736067e-01 -9.36442196e-01 1.07738423e+00 -1.01178443e+00
-4.37839240e-01 9.06721473e-01 -2.93435752e-01 -3.98528695e-01
8.79354656e-01 3.14905308e-02 -4.35097516e-01 -3.78495604e-01
1.06061600e-01 1.55714184e-01 5.03019035e-01 -1.00446558e+00
-5.08179665e-01 -7.07825363e-01 -5.86680233e-01 -4.37622564e-03
1.85261101e-01 7.09259734e-02 5.75678289e-01 -6.27315044e-01
1.12009779e-01 -9.94873047e-01 -4.83248979e-01 2.03588769e-01
-3.95229131e-01 2.48214118e-02 1.59588248e-01 -5.02063692e-01
1.19144547e+00 -2.08674216e+00 -5.80521584e-01 3.41838412e-02
7.36705720e-01 -1.22173123e-01 2.80738920e-01 2.54591316e-01
-1.76953167e-01 7.61552453e-01 -9.33762565e-02 2.17463411e-02
-1.53331399e-01 7.64912087e-03 -2.36104593e-01 7.80247450e-01
2.40659714e-01 9.60716665e-01 -7.78545439e-01 -5.48804462e-01
2.22504996e-02 2.22111523e-01 -7.72070646e-01 1.30516782e-01
1.95258688e-02 6.07829630e-01 -2.74075866e-01 7.85504520e-01
5.25720537e-01 -5.52015185e-01 3.50458801e-01 4.29591089e-01
1.39914110e-01 3.85307699e-01 -3.50015044e-01 7.57674158e-01
-1.54482806e-02 3.94780248e-01 -1.02157518e-01 -3.91455352e-01
5.68902194e-01 2.84239888e-01 1.39093906e-01 1.70618445e-01
1.59400970e-01 4.48331535e-01 5.91385305e-01 -2.98585206e-01
-1.83366016e-01 -8.61372471e-01 6.59307986e-02 6.91276073e-01
-7.16333747e-01 5.10686226e-02 -6.57184184e-01 5.64992838e-02
1.18796575e+00 -5.18154562e-01 2.35926822e-01 -3.07073295e-01
-2.24236995e-01 3.86555493e-01 1.31011546e+00 1.11198914e+00
-6.89232767e-01 8.73609543e-01 9.57872510e-01 -3.75963211e-01
-1.01556361e+00 -1.21329975e+00 -9.24124718e-01 7.01896548e-01
-4.63251442e-01 2.34102845e-01 -1.24932960e-01 -8.17678928e-01
6.13677502e-01 1.03956389e+00 -9.49403226e-01 -5.41180730e-01
-8.77077878e-02 -1.07208061e+00 7.94226944e-01 4.50874627e-01
-5.15384734e-01 -6.02891028e-01 -5.95122397e-01 2.93726474e-01
3.58351946e-01 -3.43977004e-01 -6.70729518e-01 2.91685998e-01
-1.26323426e+00 -1.55971301e+00 -9.27335083e-01 -4.01835084e-01
9.19408023e-01 -1.56204432e-01 1.32940114e+00 1.98816001e-01
-2.26445094e-01 5.35100698e-02 -7.18491226e-02 -9.93280888e-01
-8.66336703e-01 -1.38084248e-01 -5.59695438e-02 -1.05306566e-01
6.89410686e-01 -3.49791385e-02 -1.10440350e+00 2.21420109e-01
-7.83304751e-01 -8.27878490e-02 6.07227802e-01 7.67841816e-01
1.44892827e-01 -6.36789858e-01 1.02546525e+00 -1.73696911e+00
4.38695818e-01 -8.59472692e-01 -4.45107460e-01 2.43023187e-01
-1.30242014e+00 -1.48759842e-01 3.97717327e-01 -6.77748740e-01
-7.14867353e-01 -3.35512429e-01 5.71071267e-01 2.05626637e-02
-2.17740431e-01 4.73239332e-01 1.73043117e-01 3.61117035e-01
7.76997864e-01 -5.09236515e-01 3.95446539e-01 -4.38622534e-01
-2.06018254e-01 9.84674811e-01 -1.70326233e-01 -2.51128107e-01
1.84980705e-01 6.33562326e-01 1.07443519e-02 -2.36964114e-02
-1.00495410e+00 -1.06938004e-01 -1.18129626e-01 1.57029212e-01
7.04300344e-01 -9.23330903e-01 -6.33007884e-01 9.86895561e-02
-5.89964509e-01 -5.54900706e-01 -3.43181908e-01 9.81441855e-01
-1.68258414e-01 -1.43950984e-01 -6.18049741e-01 -7.96494842e-01
-2.30409920e-01 -1.21803164e+00 4.74002331e-01 1.86937138e-01
-8.68186295e-01 -1.27902794e+00 -6.23012753e-03 3.40377122e-01
3.82502675e-01 5.35819232e-01 1.57266867e+00 -1.28269839e+00
-2.80546248e-01 -5.85537970e-01 -2.15999112e-02 9.19259116e-02
3.44174504e-01 1.33734765e-02 -7.01267242e-01 -2.63992727e-01
1.68969110e-02 -1.14074565e-01 5.04008234e-01 9.45474088e-01
9.60023105e-01 -2.00172573e-01 -7.38411307e-01 5.88939846e-01
1.02475429e+00 4.11842555e-01 -1.07079223e-02 -1.53019167e-02
4.70175445e-01 9.25668478e-01 2.89081365e-01 4.46572244e-01
2.13947535e-01 2.66480625e-01 1.01243228e-01 -4.08431947e-01
1.31014407e-01 -4.69916046e-01 4.51670252e-02 -6.19939715e-03
6.53790951e-01 1.57812908e-02 -1.24568105e+00 5.62341869e-01
-1.51516104e+00 -5.03767133e-01 -6.00694537e-01 2.84368777e+00
9.40553963e-01 1.26898065e-01 -2.90266182e-02 -3.84688467e-01
9.64327216e-01 -2.93270141e-01 -1.27770400e+00 -4.99548674e-01
-3.80240232e-01 -5.50387383e-01 9.16002035e-01 6.34729087e-01
-4.53626245e-01 1.68527633e-01 7.57277966e+00 -1.20452203e-01
-1.03433537e+00 5.88182881e-02 1.37079227e+00 -5.57352781e-01
-5.56464732e-01 3.46698582e-01 -5.56143284e-01 5.13078213e-01
1.00441611e+00 -4.59653229e-01 -4.53102849e-02 5.93765080e-01
8.05970728e-01 5.03862239e-02 -1.77855194e+00 5.32003462e-01
-4.31469902e-02 -1.20476604e+00 -2.07194820e-01 4.35219258e-01
8.05348992e-01 2.00827303e-03 1.63396329e-01 8.36517289e-02
7.81202197e-01 -1.34116828e+00 7.88170993e-01 3.62543046e-01
1.06659734e+00 -3.60876739e-01 1.02263224e+00 1.62354857e-01
-6.20901175e-02 -1.57993048e-01 -3.23690981e-01 -1.26490772e-01
2.00021431e-01 1.18758738e+00 -1.28294277e+00 -9.38711539e-02
1.66986167e-01 4.30599123e-01 -6.23919725e-01 1.06107092e+00
6.29696995e-03 1.21857715e+00 1.17534017e-02 2.49749705e-01
-7.19177769e-03 2.31259558e-02 2.49382883e-01 8.28767121e-01
5.35194539e-02 7.27918446e-02 -2.97970563e-01 9.53347683e-01
-2.09911719e-01 5.74768558e-02 -5.29576540e-01 -2.88425624e-01
5.43304563e-01 8.48688841e-01 -5.01756370e-01 -4.94654089e-01
-4.65303719e-01 4.85017508e-01 9.97106582e-02 4.76274908e-01
-6.52413726e-01 6.27152473e-02 6.78082883e-01 6.15791857e-01
-5.34609139e-01 3.67924958e-01 -1.08427536e+00 -1.13126457e+00
-3.56902808e-01 -7.94596076e-01 1.03311753e+00 -4.36677486e-01
-1.62556005e+00 4.10092697e-02 -4.02312249e-01 -9.67873454e-01
2.25047488e-02 -3.20416272e-01 -4.03290719e-01 1.22037387e+00
-1.31053710e+00 -7.09131420e-01 6.49788156e-02 8.69761128e-03
9.26990137e-02 1.04415253e-01 6.55078948e-01 -1.41777825e-02
-5.71050167e-01 7.96131909e-01 3.96153033e-01 9.46292430e-02
1.42520916e+00 -1.13367355e+00 1.63273931e-01 1.15252919e-01
-4.68438685e-01 8.87988627e-01 4.90544945e-01 -1.04276264e+00
-9.09043550e-01 -1.24943435e+00 1.11067665e+00 -1.05916655e+00
4.83015686e-01 -1.54227108e-01 -1.12757695e+00 1.14421999e+00
-6.18231833e-01 -2.02937454e-01 1.56330085e+00 5.09719551e-01
-4.58230555e-01 1.41053155e-01 -1.50142539e+00 7.95023322e-01
7.79979110e-01 -3.75852406e-01 -5.80656826e-01 3.96749943e-01
4.68445748e-01 9.05468166e-02 -9.76728201e-01 2.02061266e-01
7.79423952e-01 -8.14127445e-01 5.99977911e-01 -1.35795891e+00
4.71984327e-01 -3.83855663e-02 3.32514912e-01 -1.12415135e+00
-5.73717594e-01 -3.09356213e-01 5.65866768e-01 1.02287757e+00
9.19684052e-01 -1.20727885e+00 8.22024167e-01 1.48708093e+00
1.75440207e-01 -7.28510737e-01 -7.25831449e-01 -5.30097365e-01
8.08060646e-01 -2.22671121e-01 8.66982222e-01 1.19018435e+00
2.49682292e-01 -2.29891203e-02 6.34294143e-03 3.66097003e-01
8.41337383e-01 1.46739319e-01 8.22554156e-02 -1.48449779e+00
-2.01790377e-01 -3.26312512e-01 -3.77963632e-02 -2.19526723e-01
-3.14899124e-02 -1.02084804e+00 -2.63564438e-01 -1.41486120e+00
7.79850423e-01 -1.10999835e+00 -4.33314860e-01 5.05168915e-01
-5.99534750e-01 -1.22366436e-01 1.31930038e-01 7.63463914e-01
5.00889588e-03 3.53489555e-02 8.66135716e-01 9.07340720e-02
-3.12156588e-01 1.91018641e-01 -1.15843129e+00 7.60994673e-01
7.64438093e-01 -9.62088943e-01 -2.84900039e-01 -3.69740695e-01
1.80121019e-01 2.55370140e-01 5.61264932e-01 -4.25628603e-01
3.04659996e-02 -4.99580979e-01 6.57309890e-01 -1.65977046e-01
-2.66896516e-01 -8.38196337e-01 5.90882540e-01 1.14110088e+00
-8.65135252e-01 -4.68970910e-02 -9.10296440e-02 5.92131615e-01
3.95094782e-01 -2.65887141e-01 7.78335035e-01 -4.34937440e-02
5.69801927e-01 7.38034844e-02 -7.03106463e-01 3.10978651e-01
8.91117930e-01 -2.03100797e-02 -5.02998590e-01 -2.62614846e-01
-8.49547744e-01 2.99386233e-01 1.04031301e+00 2.18994409e-01
2.29201950e-02 -1.06789136e+00 -9.53331292e-01 1.02518033e-02
3.16043735e-01 -2.23273858e-01 2.55575329e-01 9.32582140e-01
-4.74841893e-01 3.37479174e-01 2.82477945e-01 -3.47294122e-01
-1.12801445e+00 5.69943070e-01 5.83757877e-01 -5.67707717e-02
-3.31721693e-01 4.42050725e-01 6.06905460e-01 -2.69153506e-01
1.08936153e-01 -2.57961690e-01 3.59195054e-01 -1.19246403e-02
4.72434163e-01 5.22510111e-01 -1.78566471e-01 -1.93336904e-01
-5.91166317e-01 1.37915641e-01 -3.89402390e-01 -1.25835076e-01
8.74145150e-01 2.04391088e-02 -6.06324375e-02 8.17576647e-01
8.13668966e-01 9.56684127e-02 -1.26211393e+00 7.01980442e-02
1.08107729e-02 -3.79062414e-01 -1.53510094e-01 -1.27104342e+00
-4.78505045e-01 6.92583203e-01 5.76809168e-01 -4.80485670e-02
5.83689809e-01 1.46387130e-01 2.60891229e-01 -3.34937155e-01
2.38182500e-01 -6.49703801e-01 -6.65265679e-01 -2.74262339e-01
6.75125599e-01 -1.49275208e+00 1.15728736e-01 -1.50553271e-01
-7.09379911e-01 5.85880280e-01 1.23512454e-01 2.79219896e-01
4.51441437e-01 -3.16686369e-02 4.61781114e-01 -1.84911545e-02
-9.49663758e-01 5.03159523e-01 -7.65964463e-02 5.31842113e-01
8.80976081e-01 7.23603427e-01 -8.17010283e-01 1.03338325e+00
-3.15782316e-02 4.16323572e-01 9.02352571e-01 8.76935363e-01
-9.29942168e-03 -1.12039733e+00 -8.20812166e-01 1.26054811e+00
-8.22215855e-01 -1.57032907e-01 -5.56207478e-01 3.26209188e-01
1.18359886e-01 1.10568917e+00 1.67263806e-01 1.77349105e-01
2.80081868e-01 4.68685120e-01 -1.77412525e-01 -1.01155996e+00
-8.32062066e-01 -1.03874959e-01 9.56313983e-02 -1.04911149e-01
4.13601965e-01 -9.85689461e-01 -1.14221132e+00 -3.72217864e-01
-2.70934463e-01 2.59959340e-01 3.92405212e-01 5.73896646e-01
8.45887721e-01 4.99248326e-01 4.16622698e-01 4.06268463e-02
-9.49220121e-01 -7.82907188e-01 -8.57327282e-01 5.79154968e-01
6.24692142e-01 -3.25504214e-01 -9.30567265e-01 -4.20470163e-02] | [8.046419143676758, 5.513938903808594] |
f2f471d4-4e97-4813-91d9-9ed67f490c69 | slicenstitch-continuous-cp-decomposition-of | 2102.11517 | null | https://arxiv.org/abs/2102.11517v2 | https://arxiv.org/pdf/2102.11517v2.pdf | SliceNStitch: Continuous CP Decomposition of Sparse Tensor Streams | Consider traffic data (i.e., triplets in the form of source-destination-timestamp) that grow over time. Tensors (i.e., multi-dimensional arrays) with a time mode are widely used for modeling and analyzing such multi-aspect data streams. In such tensors, however, new entries are added only once per period, which is often an hour, a day, or even a year. This discreteness of tensors has limited their usage for real-time applications, where new data should be analyzed instantly as it arrives. How can we analyze time-evolving multi-aspect sparse data 'continuously' using tensors where time is'discrete'? We propose SLICENSTITCH for continuous CANDECOMP/PARAFAC (CP) decomposition, which has numerous time-critical applications, including anomaly detection, recommender systems, and stock market prediction. SLICENSTITCH changes the starting point of each period adaptively, based on the current time, and updates factor matrices (i.e., outputs of CP decomposition) instantly as new data arrives. We show, theoretically and experimentally, that SLICENSTITCH is (1) 'Any time': updating factor matrices immediately without having to wait until the current time period ends, (2) Fast: with constant-time updates up to 464x faster than online methods, and (3) Accurate: with fitness comparable (specifically, 72 ~ 100%) to offline methods. | ['Kijung Shin', 'Dongjin Lee', 'Inkyu Park', 'Taehyung Kwon'] | 2021-02-23 | null | null | null | null | ['stock-market-prediction'] | ['time-series'] | [-3.07567030e-01 -8.07771564e-01 4.68525216e-02 -1.59019232e-01
-1.19522989e-01 -8.75737786e-01 3.90251070e-01 2.14589700e-01
-1.73520848e-01 3.41538370e-01 -3.61059001e-03 -5.10536432e-01
-4.00889009e-01 -8.26482236e-01 -5.03974438e-01 -6.72343314e-01
-6.05547249e-01 9.29878056e-01 3.71297300e-01 -1.83140412e-01
5.69955930e-02 7.05138862e-01 -1.53046358e+00 1.08047888e-01
4.87533629e-01 1.37851000e+00 -4.21478212e-01 6.41237020e-01
-3.45631331e-01 6.56266630e-01 -6.06622398e-01 -3.41970682e-01
5.32985747e-01 2.71997750e-01 -2.18192711e-01 4.65106726e-01
-2.39724945e-02 -1.89070329e-01 -5.36950767e-01 6.57157362e-01
-2.07789272e-01 2.84747839e-01 9.61257592e-02 -1.70035148e+00
-1.91638827e-01 4.33507919e-01 -9.39889193e-01 7.06326544e-01
-2.92231701e-02 9.87536907e-02 8.99659336e-01 -8.56053770e-01
3.02950442e-01 1.11910093e+00 3.77126813e-01 -3.04819763e-01
-1.17275846e+00 -7.87766993e-01 6.91187739e-01 4.56877977e-01
-1.13365483e+00 -3.87868941e-01 8.53873253e-01 -4.56887037e-01
6.82724357e-01 6.79271042e-01 7.59051919e-01 7.90587962e-01
5.06256670e-02 6.55353308e-01 6.55799925e-01 2.26541504e-01
4.04534250e-01 -3.52695674e-01 3.04802537e-01 8.99804905e-02
3.51927400e-01 -1.92826480e-01 -6.84410512e-01 -5.99367917e-01
5.44270873e-01 5.25153995e-01 5.29997870e-02 -1.82235166e-02
-1.57119179e+00 5.65999925e-01 -1.92390725e-01 2.88414419e-01
-9.50716674e-01 3.64631489e-02 6.40218735e-01 5.91783643e-01
5.58152080e-01 1.08478948e-01 -5.67773640e-01 -6.99003816e-01
-9.18513119e-01 2.07815781e-01 8.18897426e-01 1.04101133e+00
6.45976841e-01 5.93250632e-01 3.19473557e-02 6.35508299e-01
-5.20464219e-02 9.22257781e-01 2.70005018e-01 -9.28239822e-01
6.81689143e-01 3.92127037e-01 2.48270839e-01 -1.37953389e+00
-4.48634475e-01 -7.02981591e-01 -1.12803996e+00 -4.91735935e-01
4.06219780e-01 -2.61944532e-01 -3.66219729e-01 1.45327020e+00
5.83743513e-01 8.28305125e-01 -3.01552206e-01 7.05921650e-01
1.64120563e-03 8.96989644e-01 -4.15194780e-01 -7.89129972e-01
1.51970947e+00 -4.19445187e-01 -7.48384356e-01 1.46500811e-01
2.33239144e-01 -8.90804052e-01 5.14900982e-01 9.44746315e-01
-8.14237952e-01 -2.50431657e-01 -5.12575388e-01 5.63340187e-01
-2.77989119e-01 -1.53719857e-01 8.14479172e-01 5.49231768e-01
-6.11229718e-01 2.23088726e-01 -1.02186191e+00 8.45928676e-03
-1.04307979e-01 1.24187820e-01 -2.56545961e-01 -5.49737643e-03
-9.79243159e-01 4.24924307e-02 -1.09022096e-01 2.38110036e-01
-6.02262259e-01 -1.17481029e+00 -2.15201572e-01 2.05382302e-01
6.07613385e-01 -4.40654069e-01 8.98795664e-01 -4.75392073e-01
-1.19438076e+00 6.78413361e-02 -4.27530020e-01 -5.28320312e-01
1.70105904e-01 -7.24487230e-02 -1.19892251e+00 2.06060126e-01
5.46084717e-02 -4.07724351e-01 1.16076517e+00 -8.09352279e-01
-7.91854322e-01 -5.37751794e-01 -5.53914579e-03 -2.11238667e-01
-5.49513996e-01 2.22888798e-01 -7.03203201e-01 -9.30323899e-01
4.05022740e-01 -1.11884868e+00 -2.46918470e-01 -3.65674794e-01
-1.81781545e-01 -7.47186244e-02 1.20854974e+00 -7.52459884e-01
1.55518878e+00 -2.26471925e+00 -3.69591117e-02 7.24588871e-01
4.77581739e-01 1.70898974e-01 -8.67505744e-02 7.98161805e-01
-2.43929014e-01 -1.05689369e-01 -1.18778192e-03 -2.37783939e-01
7.67946104e-03 4.56483990e-01 -7.17550099e-01 4.77208942e-01
-2.25476354e-01 2.54594296e-01 -8.90671432e-01 1.45751223e-01
7.15285018e-02 8.53773132e-02 -5.73923111e-01 -8.07384681e-03
-1.04870655e-01 4.16018039e-01 -3.08643907e-01 7.59424150e-01
8.07584107e-01 -3.94394070e-01 1.58706203e-01 -3.99047524e-01
-5.28695345e-01 -1.68544054e-01 -1.56498361e+00 1.03153944e+00
-4.83516097e-01 7.51523733e-01 2.32031904e-02 -1.08719611e+00
7.31941819e-01 5.32110155e-01 1.31921327e+00 -6.96721315e-01
-6.52920604e-02 1.18218169e-01 -2.71045417e-02 -1.78055719e-01
5.19057989e-01 2.39844263e-01 7.33417273e-02 7.39361763e-01
-3.22529882e-01 5.69728017e-01 7.00551987e-01 3.24711144e-01
1.42879498e+00 -6.74440622e-01 -9.53098312e-02 3.95694487e-02
5.53633332e-01 -2.37924993e-01 7.81612515e-01 3.12886953e-01
6.05881438e-02 1.90489531e-01 7.83932626e-01 -7.16572881e-01
-9.74947453e-01 -1.01253664e+00 6.86571449e-02 1.06598306e+00
-2.23417997e-01 -6.85827494e-01 -1.81340307e-01 -1.74599290e-01
4.38107044e-01 5.64394057e-01 -3.99825871e-01 1.27293736e-01
-6.93840444e-01 -8.92272770e-01 9.36457738e-02 4.02713805e-01
2.83203214e-01 -9.53181386e-02 -2.87371010e-01 7.86438286e-01
-3.47933561e-01 -1.54637969e+00 -6.37543023e-01 -2.15972871e-01
-1.06387758e+00 -8.92971337e-01 -2.88864046e-01 3.05704772e-01
5.73097527e-01 8.38352919e-01 1.01978850e+00 -1.79645032e-01
9.27600414e-02 7.98854291e-01 -4.06540781e-01 -2.65703499e-01
2.26484314e-01 -4.28868949e-01 4.95861501e-01 1.08738863e+00
3.04870129e-01 -1.18250668e+00 -6.85168386e-01 5.83265841e-01
-1.18790340e+00 -3.00982445e-01 3.60457301e-01 4.78899360e-01
6.46161199e-01 4.33938384e-01 4.96346056e-01 -7.06511080e-01
5.93786061e-01 -1.05331206e+00 -9.05998528e-01 1.21421658e-01
-6.13659978e-01 -2.89617807e-01 8.68207097e-01 -7.04970062e-01
-5.29992461e-01 -5.66202641e-01 3.18762660e-01 -1.09904659e+00
2.22019568e-01 7.11796522e-01 2.38317817e-01 1.39105797e-01
2.01908007e-01 3.74443203e-01 1.56379845e-02 -5.86560249e-01
2.43923739e-01 4.22899812e-01 3.76831234e-01 -9.16456342e-01
1.14063704e+00 7.68725038e-01 3.13372225e-01 -1.13300920e+00
-3.47518563e-01 -9.40502107e-01 -4.58660066e-01 -4.44778800e-01
1.68792441e-01 -9.67477977e-01 -1.06919694e+00 4.73874480e-01
-7.05706120e-01 3.00188065e-01 -2.58051068e-01 6.35592639e-01
-8.67161229e-02 4.91965592e-01 -5.49723148e-01 -6.45396829e-01
-7.05433264e-02 -6.56932950e-01 6.82434678e-01 -1.14419751e-01
8.08475018e-02 -9.81301546e-01 -7.10164309e-02 3.21235627e-01
7.36243963e-01 3.34888101e-01 7.26901531e-01 -7.22612202e-01
-7.58184612e-01 -4.93450731e-01 -8.71416554e-02 1.81370214e-01
1.69789255e-01 2.42889300e-01 -6.00057840e-01 -4.19221133e-01
-3.31707373e-02 5.92906952e-01 1.07250094e-01 1.54108837e-01
1.25072515e+00 -7.43470848e-01 -9.51866135e-02 4.85737234e-01
9.02332366e-01 3.61271679e-01 3.92790049e-01 5.02963252e-02
5.75227499e-01 3.00320804e-01 6.94852173e-01 1.12267387e+00
6.26725018e-01 7.27944076e-01 3.36345077e-01 2.87016898e-01
3.73985827e-01 2.02704594e-01 4.68516320e-01 1.67497122e+00
-3.51399124e-01 -9.34004933e-02 -7.89041460e-01 6.26628101e-01
-1.79980755e+00 -1.12792730e+00 -4.67859715e-01 2.52639008e+00
4.26733643e-01 1.48606792e-01 5.43661594e-01 3.16904664e-01
3.86616409e-01 3.11348587e-01 -6.83036804e-01 -2.06457213e-01
-2.67517734e-02 9.87883210e-02 5.26985288e-01 1.82536080e-01
-8.57039809e-01 3.63746941e-01 4.90051651e+00 5.11508048e-01
-1.39116430e+00 1.13959476e-01 1.97360769e-01 -3.31911057e-01
-3.07009816e-01 2.05760896e-01 -4.97022569e-01 8.80748093e-01
1.52011907e+00 -5.69584489e-01 8.50828350e-01 4.69429463e-01
5.05078435e-01 1.49994135e-01 -9.80437636e-01 1.25452638e+00
-2.56621838e-01 -1.11398828e+00 8.36102739e-02 3.50922108e-01
4.48451787e-01 9.58329886e-02 1.13037422e-01 2.78361648e-01
1.95073128e-01 -1.84019342e-01 3.93117040e-01 6.86747849e-01
2.27492049e-01 -7.82184720e-01 2.43296742e-01 2.61646360e-01
-1.67900336e+00 -2.40509868e-01 -2.41221458e-01 7.31795877e-02
3.41458350e-01 1.36328185e+00 -3.74411345e-01 9.21280146e-01
7.06202209e-01 7.83734262e-01 -3.44617218e-01 1.06813323e+00
4.57822859e-01 1.03637910e+00 -7.43308604e-01 3.64276707e-01
2.51400441e-01 -6.01989269e-01 9.71369743e-01 8.76944959e-01
7.47698843e-01 2.27151439e-01 3.46925229e-01 2.50013798e-01
2.01873913e-01 4.53116372e-02 -3.54902893e-01 -1.05479002e-01
6.93862140e-01 1.16820145e+00 -6.15778029e-01 -4.51800734e-01
-6.65250897e-01 5.19605935e-01 -4.38584626e-01 7.79223204e-01
-8.46704006e-01 -2.33803540e-01 1.20065236e+00 2.99964458e-01
5.17519653e-01 -7.62540162e-01 6.34595454e-02 -1.47369492e+00
3.34195167e-01 -9.39839602e-01 5.57399154e-01 -3.85448605e-01
-1.35739315e+00 3.42201263e-01 8.09128210e-02 -1.85835111e+00
-2.84131020e-01 -3.94997954e-01 -3.19237292e-01 4.14355814e-01
-1.15355742e+00 -6.82271421e-01 -5.55211455e-02 9.25734103e-01
2.71451890e-01 -2.73367196e-01 5.40154517e-01 8.45897377e-01
-7.28735566e-01 3.17662954e-01 4.30685043e-01 1.40261520e-02
3.32379818e-01 -9.91958857e-01 5.18408418e-01 9.81070638e-01
3.70949328e-01 6.83682621e-01 7.13650644e-01 -4.28498298e-01
-2.02072048e+00 -1.00758719e+00 5.58625102e-01 -2.29918092e-01
1.30777824e+00 -2.24756047e-01 -9.85944927e-01 6.86255634e-01
-1.31196290e-01 2.56308049e-01 8.96116912e-01 3.40726912e-01
-5.78790367e-01 -9.23959732e-01 -8.36138606e-01 4.59655672e-01
7.17434049e-01 -3.60112756e-01 -5.50243296e-02 5.59626937e-01
5.63536346e-01 -4.54710782e-01 -1.36676800e+00 1.42299563e-01
4.38891500e-01 -6.55862093e-01 1.02189219e+00 -7.60022104e-01
-3.98273170e-01 -6.86970413e-01 -4.14234668e-01 -1.20118558e+00
-4.82851595e-01 -1.03030264e+00 -7.76665509e-01 1.03734505e+00
1.20010480e-01 -1.14451540e+00 2.60607123e-01 5.01623631e-01
1.92144625e-02 -5.36430240e-01 -1.21205056e+00 -9.53795433e-01
-6.87716186e-01 -8.16115379e-01 1.14956307e+00 1.09287953e+00
-3.71022254e-01 2.51527261e-02 -5.27679324e-01 7.27797568e-01
8.78862500e-01 4.62281108e-01 8.40880036e-01 -1.50080955e+00
-3.97484064e-01 -3.11741978e-01 -5.35364807e-01 -1.04445183e+00
-3.45823139e-01 -3.85277957e-01 -7.14779377e-01 -6.46129906e-01
-2.10470274e-01 -5.28908432e-01 -5.59071243e-01 4.20578629e-01
2.39672005e-01 -1.52327269e-01 2.07103580e-01 5.61700463e-01
-4.78563786e-01 4.36416149e-01 8.34297478e-01 -6.42085895e-02
-5.61641119e-02 4.22496796e-01 -4.25080389e-01 3.49651963e-01
5.58579683e-01 -2.92463154e-01 -6.43225849e-01 -3.43711197e-01
3.62346619e-01 4.22529101e-01 8.97759385e-03 -1.07295382e+00
3.03199947e-01 -4.35592920e-01 -2.24406198e-02 -8.70596468e-01
5.68887532e-01 -1.07787073e+00 6.14685237e-01 3.01023126e-02
4.47018385e-01 7.04657555e-01 1.61144227e-01 9.08816159e-01
-4.36285824e-01 3.38602692e-01 8.48316699e-02 2.97903568e-01
-4.61221308e-01 8.23514163e-01 -5.37352264e-01 1.05386220e-01
1.03105950e+00 5.67948408e-02 -3.36202383e-01 -6.08123362e-01
-8.46092641e-01 4.31872725e-01 -5.63513860e-02 4.69241500e-01
4.16265935e-01 -1.43963373e+00 -5.37943900e-01 2.76165605e-01
-1.76069494e-02 -1.92684039e-01 6.29544437e-01 1.20688927e+00
-1.58327252e-01 5.24725556e-01 9.35979187e-02 -8.61467659e-01
-8.97093594e-01 7.05924451e-01 -1.05146505e-01 -2.88827062e-01
-7.05907404e-01 2.00797141e-01 -8.30297694e-02 2.03615110e-02
8.32748562e-02 -5.09097159e-01 -8.12632516e-02 5.35842597e-01
6.95059180e-01 6.96548820e-01 4.08187032e-01 -4.89088982e-01
-2.89856672e-01 4.96044517e-01 -3.61259207e-02 -1.58926621e-02
1.54065049e+00 -1.02351539e-01 -5.36981285e-01 8.76602650e-01
9.78331506e-01 3.12306851e-01 -8.34533870e-01 -5.92828274e-01
2.10653394e-02 -8.40968966e-01 -1.37400091e-01 -1.21615432e-01
-1.62967443e+00 6.25359535e-01 2.62255549e-01 7.66164899e-01
1.31590819e+00 -4.66037661e-01 1.12531698e+00 3.77590328e-01
5.23218989e-01 -8.26888561e-01 -1.61907785e-02 5.20307422e-01
7.95082510e-01 -5.80556810e-01 -2.92373192e-03 -3.06972891e-01
-4.42899436e-01 1.16015434e+00 1.47926867e-01 1.80326104e-01
1.25537527e+00 1.25555217e-01 -1.46431506e-01 -1.92985475e-01
-1.29616868e+00 6.26609772e-02 4.97320071e-02 1.87788516e-01
-1.10575527e-01 4.94226784e-01 1.06754154e-01 4.17908847e-01
-1.86778575e-01 -3.12254965e-01 6.01579547e-01 8.82429123e-01
-8.77781957e-02 -1.13025796e+00 -7.07817137e-01 7.29164541e-01
-3.81668866e-01 3.41939121e-01 3.73821795e-01 3.38856637e-01
-2.34010115e-01 1.05879521e+00 3.49506348e-01 -4.37142581e-01
5.60624242e-01 -1.00776494e-01 -1.55994102e-01 -1.66001782e-01
-3.06880295e-01 2.01129317e-01 4.65145744e-02 -8.60076904e-01
-1.96302533e-01 -1.14542007e+00 -8.76778603e-01 -8.95794809e-01
9.23329294e-02 3.80204201e-01 6.83738291e-01 8.19697082e-01
7.26374030e-01 4.55500394e-01 1.37648761e+00 -4.71745491e-01
-4.04194146e-01 -7.57756710e-01 -6.97654128e-01 2.25660890e-01
3.18732679e-01 -7.37904608e-01 -4.46207553e-01 6.44485354e-02] | [7.206669330596924, 2.9109768867492676] |
0e3e83b9-5f4c-44f3-a73c-caef017361d7 | efficient-gpt-model-pre-training-using-tensor | 2306.02697 | null | https://arxiv.org/abs/2306.02697v1 | https://arxiv.org/pdf/2306.02697v1.pdf | Efficient GPT Model Pre-training using Tensor Train Matrix Representation | Large-scale transformer models have shown remarkable performance in language modelling tasks. However, such models feature billions of parameters, leading to difficulties in their deployment and prohibitive training costs from scratch. To reduce the number of the parameters in the GPT-2 architecture, we replace the matrices of fully-connected layers with the corresponding Tensor Train Matrix~(TTM) structure. Finally, we customize forward and backward operations through the TTM-based layer for simplicity and the stableness of further training. % The resulting GPT-2-based model stores up to 40% fewer parameters, showing the perplexity comparable to the original model. On the downstream tasks, including language understanding and text summarization, the model performs similarly to the original GPT-2 model. The proposed tensorized layers could be used to efficiently pre-training other Transformer models. | ['Alexander Panchenko', 'Ivan Oseledets', 'Julia Gusak', 'Georgii Novikov', 'Viktoriia Chekalina'] | 2023-06-05 | null | null | null | null | ['text-summarization'] | ['natural-language-processing'] | [-6.34804368e-03 3.66913944e-01 9.66043174e-02 -3.31022203e-01
-5.90645909e-01 -5.02637267e-01 5.17984271e-01 1.25282386e-03
-3.51627469e-01 3.71870428e-01 3.77692968e-01 -9.59598720e-01
2.63621151e-01 -5.33082247e-01 -6.30177319e-01 -6.02893770e-01
-1.90698117e-01 6.83810532e-01 1.55541286e-01 -2.93648630e-01
1.36046484e-01 1.44979045e-01 -8.52183163e-01 4.29760754e-01
9.32910621e-01 1.03519332e+00 2.49888524e-01 6.97668195e-01
-4.43835914e-01 9.17352140e-01 -2.64190972e-01 -8.18257987e-01
-8.48895162e-02 -1.13636978e-01 -1.00407493e+00 -3.61359157e-02
4.14944470e-01 -3.90090913e-01 -7.24493384e-01 7.81531215e-01
3.14969689e-01 -5.75566143e-02 3.94169480e-01 -9.48548436e-01
-6.44014835e-01 1.09083927e+00 -3.43550146e-01 9.42431837e-02
-5.41598722e-02 -1.26338333e-01 1.32891476e+00 -9.55633044e-01
1.67566791e-01 1.45098329e+00 7.97500193e-01 3.52299541e-01
-1.33143628e+00 -5.30244410e-01 4.27386552e-01 2.10688174e-01
-1.00624681e+00 -5.59042156e-01 5.05005240e-01 -3.04346770e-01
1.64033782e+00 2.28327483e-01 7.49856412e-01 9.21187758e-01
1.83248669e-01 1.02568388e+00 6.84939325e-01 -8.57599303e-02
-1.55354947e-01 -1.12329066e-01 2.47099042e-01 1.06687188e+00
1.59831628e-01 -5.67502201e-01 -4.27328110e-01 -1.45543471e-01
6.19597673e-01 -1.27314880e-01 -9.27966982e-02 -9.77278799e-02
-1.52026260e+00 8.60793352e-01 4.09226507e-01 3.17361087e-01
-2.16092229e-01 5.45956492e-01 6.46425247e-01 4.16869670e-01
6.69255257e-01 3.24912578e-01 -6.13557994e-01 -3.13970387e-01
-9.38261211e-01 1.39083946e-02 8.03714871e-01 1.18626606e+00
6.46343887e-01 4.03589427e-01 -1.08713403e-01 9.78135228e-01
1.38797715e-01 2.87063837e-01 5.65828443e-01 -7.88627923e-01
9.48520005e-01 6.60486639e-01 -3.16531688e-01 -8.90306234e-01
-3.28127801e-01 -6.74606264e-01 -1.15783083e+00 -8.51840973e-01
1.55339032e-01 -1.09729089e-01 -9.03079331e-01 1.48049355e+00
6.24591159e-03 -3.86582166e-02 -8.76039118e-02 4.36983079e-01
5.04201770e-01 9.40795541e-01 -5.86151071e-02 1.01251036e-01
1.35134411e+00 -1.21732247e+00 -6.50044978e-01 -4.34829623e-01
1.23002934e+00 -8.59731019e-01 1.04163337e+00 1.82341173e-01
-1.30664659e+00 -1.33810446e-01 -7.48284936e-01 -6.06674552e-01
-1.85095087e-01 5.00678241e-01 1.10519063e+00 3.32564861e-01
-1.30604887e+00 8.36481035e-01 -1.16344249e+00 -4.51811075e-01
2.85876244e-01 5.21832526e-01 -3.65132362e-01 -3.22318450e-02
-9.90087450e-01 9.91144836e-01 4.65569794e-01 5.60735881e-01
-8.61050427e-01 -6.38110757e-01 -9.42155778e-01 4.47363198e-01
-7.16634840e-02 -1.07547629e+00 1.40673125e+00 -2.03048974e-01
-1.74938893e+00 4.62876856e-01 -4.92580861e-01 -6.98749542e-01
2.91915715e-01 -1.94565564e-01 1.30817126e-02 -6.45621633e-03
-6.40799105e-02 5.62030137e-01 7.27670789e-01 -6.00303411e-01
-4.21891868e-01 -1.14929698e-01 1.83497623e-01 1.81879386e-01
-6.98912442e-01 5.79534005e-03 -7.46502519e-01 -7.29727030e-01
3.46961617e-01 -1.01470590e+00 -2.46039540e-01 -4.36322987e-01
-8.52079749e-01 -2.73322552e-01 5.54083943e-01 -9.19134498e-01
1.55372906e+00 -1.81383741e+00 4.26541656e-01 -1.38459817e-01
3.45902145e-01 3.43177825e-01 -4.40329760e-01 7.20098257e-01
-1.72719866e-01 2.93026835e-01 -1.98130161e-01 -7.83747196e-01
1.12583898e-01 5.04947424e-01 -4.61115211e-01 1.06355332e-01
3.12056154e-01 1.05077851e+00 -6.12947881e-01 -4.57607150e-01
1.22243967e-02 3.11414301e-01 -8.17588747e-01 5.47618531e-02
-2.46026963e-01 1.73808828e-01 -3.88723284e-01 2.61887074e-01
4.81369406e-01 -4.46138233e-01 3.76082987e-01 -4.22093153e-01
6.36675134e-02 1.19278860e+00 -6.78563654e-01 1.60845709e+00
-7.47064054e-01 6.80455565e-01 -6.63568303e-02 -1.24707854e+00
5.57814240e-01 4.06019032e-01 2.93884307e-01 -4.71120954e-01
1.24609441e-01 1.70976594e-01 1.80252984e-01 -3.08938384e-01
5.91307104e-01 -9.61040854e-02 -5.46135530e-02 8.54685903e-01
3.05385649e-01 -2.34044850e-01 3.74413401e-01 5.61187327e-01
8.82739604e-01 -5.62922060e-02 -2.14319691e-01 -2.36610398e-01
4.30218339e-01 -1.51179418e-01 4.23475385e-01 3.99375618e-01
4.01650310e-01 2.22275570e-01 8.27971697e-01 -5.03621936e-01
-1.32956088e+00 -6.44044876e-01 1.65672809e-01 1.21117330e+00
-6.53116524e-01 -9.59094703e-01 -8.78560543e-01 -3.74297827e-01
-1.24258883e-01 8.14650893e-01 -4.05996323e-01 -1.51797533e-01
-9.80952680e-01 -9.13925231e-01 8.48645866e-01 5.75012922e-01
5.48926711e-01 -4.39533681e-01 1.44045725e-01 4.14262474e-01
-5.55580616e-01 -1.32075906e+00 -6.22446716e-01 1.93093136e-01
-1.50290322e+00 -5.97631991e-01 -5.75924933e-01 -8.64768922e-01
7.04701304e-01 3.01612109e-01 1.07984006e+00 2.29802445e-01
4.56429958e-01 -1.14571206e-01 4.24807481e-02 -1.17096819e-01
-4.54025447e-01 5.64280808e-01 -3.02015468e-02 -7.37231001e-02
-4.22471389e-02 -7.94244230e-01 -4.37145919e-01 1.55390918e-01
-6.12565696e-01 5.56868672e-01 5.93810022e-01 9.86875534e-01
1.34875134e-01 -8.09079930e-02 1.58076823e-01 -7.85809875e-01
8.93579364e-01 -2.18868330e-02 -5.02857566e-01 3.50328505e-01
-4.96289760e-01 4.02328491e-01 7.16154814e-01 -5.61160445e-01
-7.60863185e-01 -4.42783475e-01 -1.75498366e-01 -1.74745068e-01
6.41236484e-01 7.10643589e-01 1.02206446e-01 -3.65223736e-02
4.19594795e-02 4.56077278e-01 -1.39224663e-01 -8.81088495e-01
5.33747315e-01 4.11468625e-01 1.49059400e-01 -5.63637376e-01
9.02364671e-01 4.37744074e-02 -2.91942745e-01 -7.40864754e-01
-9.70424116e-01 -7.92045966e-02 -5.25069118e-01 5.19664466e-01
5.22317588e-01 -9.73694026e-01 -7.57096708e-01 5.58243871e-01
-1.51149452e+00 -2.78814286e-01 1.64387841e-02 5.57125151e-01
-2.73728728e-01 6.56129777e-01 -1.17353415e+00 -4.61976767e-01
-7.54930615e-01 -1.16103387e+00 1.07327497e+00 -3.63384426e-01
-6.03495017e-02 -1.27751887e+00 -2.12940633e-01 5.30097663e-01
6.79560542e-01 -6.63555205e-01 1.48754418e+00 -6.10391617e-01
-6.31362438e-01 -2.63183326e-01 -4.00199980e-01 6.43667281e-01
-2.06663728e-01 -2.43535817e-01 -8.42878401e-01 -3.67633849e-01
6.01235777e-03 -6.22941107e-02 1.01902938e+00 2.24763244e-01
1.23581386e+00 -6.63585722e-01 -2.72534966e-01 8.22388232e-01
8.97304773e-01 -1.56165451e-01 4.29559499e-01 1.92034647e-01
1.03997743e+00 3.20088536e-01 -6.28693476e-02 -2.87313126e-02
8.01127553e-01 3.82973731e-01 6.07789494e-02 -2.64171481e-01
-4.56055924e-02 -4.60851789e-01 5.41942358e-01 1.92286658e+00
-1.00474104e-01 -2.28408769e-01 -8.85032356e-01 5.60616791e-01
-1.81992853e+00 -7.32234955e-01 -2.59645700e-01 1.98653269e+00
7.45032609e-01 2.44812965e-01 -1.74288526e-01 9.58372951e-02
3.22576225e-01 2.30235025e-01 -3.92248362e-01 -7.52926707e-01
-1.01208277e-01 1.34641856e-01 5.59543133e-01 6.23892903e-01
-7.81496584e-01 1.32718229e+00 7.56692934e+00 7.50099659e-01
-1.26207149e+00 2.34963313e-01 5.23704469e-01 -1.80557266e-01
-5.26834607e-01 3.47839355e-01 -6.93683207e-01 2.47553900e-01
1.26612604e+00 -1.96457788e-01 6.51889443e-01 4.01840866e-01
3.71458560e-01 1.68658331e-01 -1.24593663e+00 8.62656593e-01
-1.41363621e-01 -1.37721682e+00 5.09956956e-01 1.03003249e-01
3.46000582e-01 4.72865701e-01 -2.33786646e-02 5.44052899e-01
3.59569848e-01 -8.91465902e-01 6.17721438e-01 4.72492427e-02
6.31541371e-01 -5.04462302e-01 4.85269874e-01 3.32958758e-01
-1.19013286e+00 -1.13383025e-01 -6.72722816e-01 -2.11268544e-01
2.74863750e-01 6.40960395e-01 -1.20473635e+00 4.27218795e-01
5.10661423e-01 6.24901831e-01 -5.52143812e-01 6.54328465e-01
-1.88648552e-01 9.16583955e-01 -5.32662451e-01 -1.12876426e-02
5.65201581e-01 -4.60315377e-01 2.86881268e-01 1.35309958e+00
3.52563232e-01 -2.65962005e-01 -1.26567289e-01 6.27744257e-01
-2.94643193e-01 -1.74963952e-03 -4.32183534e-01 -4.86196905e-01
2.55513489e-01 1.21835899e+00 -2.94084072e-01 -6.82304621e-01
-4.14633960e-01 1.09424043e+00 7.45839834e-01 5.09218335e-01
-7.70778954e-01 -2.48237431e-01 7.44314134e-01 7.54394010e-02
6.15067959e-01 -6.68759048e-01 -1.78260446e-01 -1.68964255e+00
1.56432539e-01 -9.56487179e-01 1.55765265e-01 -6.43872797e-01
-1.03011000e+00 8.90572011e-01 -7.30117708e-02 -7.57858694e-01
-3.08132172e-01 -7.44721532e-01 -6.34683430e-01 1.01521611e+00
-1.54348373e+00 -1.44085240e+00 3.65592748e-01 5.82219005e-01
3.59756768e-01 2.33049691e-02 9.46946561e-01 4.59031552e-01
-1.13148975e+00 6.70595109e-01 2.63552248e-01 2.52665788e-01
3.53722960e-01 -1.15041745e+00 1.11998510e+00 9.45609331e-01
-1.30398676e-01 1.14790022e+00 5.15017092e-01 -3.37716460e-01
-1.62750733e+00 -1.04932344e+00 1.61491632e+00 -2.44446978e-01
1.27790940e+00 -8.36780250e-01 -9.19664562e-01 1.16431153e+00
2.56678104e-01 -4.58612233e-01 6.10375941e-01 5.02846658e-01
-5.52911758e-01 -9.27263498e-02 -4.54967469e-01 7.79002666e-01
1.01964986e+00 -8.96836460e-01 -5.77127099e-01 6.05878890e-01
9.11561966e-01 -3.10967982e-01 -1.10772634e+00 -5.44394776e-02
4.40210521e-01 -6.42154455e-01 8.94425213e-01 -8.15417290e-01
2.76572585e-01 -3.27534117e-02 7.21957311e-02 -1.36611867e+00
-5.47038078e-01 -9.26436245e-01 -1.74576029e-01 1.06215262e+00
7.67980456e-01 -8.80889893e-01 5.17316639e-01 5.40896237e-01
-3.41212362e-01 -8.45447659e-01 -8.82671058e-01 -5.39149344e-01
4.45842594e-01 -5.57306230e-01 6.52093709e-01 6.96726263e-01
1.67837128e-01 9.32032287e-01 -4.43517983e-01 -6.72853738e-02
3.64620775e-01 1.30425498e-01 6.16254866e-01 -1.14817464e+00
-3.86509389e-01 -4.63542998e-01 1.42859295e-01 -1.66108453e+00
1.17179863e-01 -1.23902762e+00 -3.73240083e-01 -1.73647809e+00
3.11209649e-01 -5.37454724e-01 7.39434455e-03 7.57165074e-01
-1.90394476e-01 9.17187259e-02 2.87357569e-01 3.47506106e-01
-3.67406219e-01 7.46423185e-01 1.31744492e+00 -2.62634099e-01
1.11949421e-01 -2.43448243e-01 -8.65571976e-01 6.17517114e-01
6.59534991e-01 -3.69241148e-01 -4.25157696e-01 -1.23881483e+00
4.01077926e-01 3.42106670e-02 5.20099103e-02 -4.99263793e-01
3.15587819e-01 2.61610329e-01 2.44891495e-02 -6.83038592e-01
5.30538619e-01 -4.39285547e-01 -4.64978591e-02 3.20640743e-01
-2.21908525e-01 4.77451861e-01 2.41294444e-01 2.06283867e-01
-3.25505465e-01 -7.23070279e-02 2.87008703e-01 -1.65185437e-01
-6.51237443e-02 4.26913917e-01 -5.47287166e-01 -2.71649003e-01
2.34071940e-01 -1.68664247e-01 -1.89121723e-01 -4.09202844e-01
-6.29811585e-01 2.83643365e-01 2.15642497e-01 2.32871622e-01
4.66461182e-01 -1.17889106e+00 -6.64135218e-01 2.20758542e-01
-2.25299209e-01 2.16170385e-01 2.32725129e-01 1.28591251e+00
-5.08057833e-01 1.01315498e+00 1.42958567e-01 -5.12034655e-01
-9.51591790e-01 5.33642292e-01 4.42371249e-01 -7.62297451e-01
-7.43802667e-01 7.94247866e-01 3.94679397e-01 -7.23829329e-01
2.30589926e-01 -8.81327450e-01 4.82400842e-02 -3.99772488e-02
3.37733358e-01 2.75196046e-01 4.31545675e-01 -4.45665717e-01
-2.60086566e-01 4.43575084e-01 -3.36860538e-01 1.75333768e-03
1.54318357e+00 -1.36092797e-01 -5.68035722e-01 3.27821702e-01
1.39059126e+00 -3.56689811e-01 -6.51651323e-01 -3.36425096e-01
1.02494977e-01 1.56323925e-01 2.53061175e-01 -3.85132104e-01
-1.02387738e+00 1.41408896e+00 -6.10762127e-02 7.81505257e-02
8.62163067e-01 -3.04183424e-01 1.08257878e+00 8.56176257e-01
2.98170030e-01 -8.27544808e-01 -2.08898053e-01 1.02241814e+00
8.05015981e-01 -7.11985648e-01 -5.00669219e-02 -3.93834651e-01
-4.93579745e-01 1.18223596e+00 9.60276127e-02 9.63342711e-02
5.25018334e-01 2.70387590e-01 -1.11992121e-01 -1.43889531e-01
-1.31170690e+00 3.07788193e-01 3.29937190e-01 2.58706719e-01
8.52491736e-01 8.27550068e-02 -1.83917418e-01 2.55975574e-01
-6.20023429e-01 -1.47149086e-01 3.91797662e-01 6.58714712e-01
-7.79142901e-02 -1.31128621e+00 -1.34270966e-01 5.50309300e-01
-3.96950454e-01 -7.07149923e-01 -1.44664168e-01 4.51332539e-01
-5.43980002e-01 9.25866365e-01 1.06420569e-01 -5.47677696e-01
2.39003405e-01 2.78502315e-01 4.51188356e-01 -4.74240422e-01
-7.64368176e-01 2.30295494e-01 6.38550341e-01 -6.32971525e-01
-1.21720389e-01 -3.92232239e-01 -1.04107904e+00 -9.18756425e-01
-4.76086557e-01 2.32854769e-01 7.39241064e-01 8.87480736e-01
6.94925368e-01 4.54080194e-01 3.34173054e-01 -8.13861549e-01
-8.08254659e-01 -1.44346046e+00 -2.35977754e-01 2.44232826e-02
1.70374691e-01 -3.06737900e-01 -1.27449974e-01 -7.65917525e-02] | [10.800053596496582, 7.578328609466553] |
0f77c132-bc81-47b8-aa71-0722bfab053d | gemini-controlling-the-sentence-level-writing | 2304.03548 | null | https://arxiv.org/abs/2304.03548v1 | https://arxiv.org/pdf/2304.03548v1.pdf | GEMINI: Controlling the Sentence-level Writing Style for Abstractive Text Summarization | Human experts write summaries using different techniques, including rewriting a sentence in the document or fusing multiple sentences to generate a summary sentence. These techniques are flexible and thus difficult to be imitated by any single method. To address this issue, we propose an adaptive model, GEMINI, that integrates a rewriter and a fuser to mimic the sentence rewriting and fusion techniques, respectively. GEMINI adaptively chooses to rewrite a specific document sentence or generate a summary sentence from scratch. Experiments demonstrate that our adaptive approach outperforms the pure abstractive and rewriting baselines on various benchmark datasets, especially when the dataset has a balanced distribution of styles. Interestingly, empirical results show that the human writing style of each summary sentence is consistently predictable given its context. | ['Yue Zhang', 'Zebin Ou', 'Guangsheng Bao'] | 2023-04-07 | null | null | null | null | ['abstractive-text-summarization'] | ['natural-language-processing'] | [ 4.80798095e-01 6.25557527e-02 -1.59135610e-01 -5.97594917e-01
-9.09725606e-01 -1.04476130e+00 7.86006927e-01 -6.50077267e-03
-5.50417192e-02 7.69739747e-01 5.92276573e-01 -1.69438079e-01
3.02617818e-01 -4.19857293e-01 -3.79558146e-01 -4.59560126e-01
8.96466732e-01 4.25075561e-01 8.31101388e-02 -3.44565332e-01
6.48973703e-01 3.60267907e-01 -1.19831395e+00 4.38728988e-01
1.32283378e+00 4.32600498e-01 6.04460955e-01 1.04914856e+00
-4.51054960e-01 8.57684970e-01 -1.44932187e+00 -9.89825249e-01
1.20716609e-01 -6.21963680e-01 -5.65264404e-01 2.52845407e-01
8.15582871e-01 -3.52393568e-01 -1.22272469e-01 1.05892825e+00
5.05922139e-01 2.92320669e-01 7.09263802e-01 -7.54678488e-01
-9.26510274e-01 1.10846913e+00 -2.89798439e-01 1.90235585e-01
6.92980409e-01 1.09717652e-01 9.19159889e-01 -6.46964729e-01
6.99518383e-01 1.35430992e+00 2.95375049e-01 8.06173921e-01
-1.17659533e+00 -2.41292819e-01 4.76093084e-01 -3.31584603e-01
-9.56857085e-01 -7.10838556e-01 7.92248070e-01 -2.06752107e-01
1.15127671e+00 5.67789137e-01 2.07470477e-01 1.79484081e+00
6.96937203e-01 9.64101970e-01 8.45845819e-01 -3.72887164e-01
1.54441029e-01 1.01949342e-01 4.54433352e-01 4.18511152e-01
3.42830122e-01 -3.83309513e-01 -5.42870581e-01 -2.58418083e-01
2.41517559e-01 2.30698958e-02 -3.87247473e-01 3.92723113e-01
-1.09663224e+00 6.21308625e-01 -2.14233220e-01 2.15198591e-01
-2.91458666e-01 -1.54985785e-01 5.68754137e-01 5.62318325e-01
3.76254410e-01 8.04191649e-01 -1.44982055e-01 -2.18218476e-01
-1.15471756e+00 6.19049847e-01 1.13431811e+00 1.32824421e+00
3.92956764e-01 2.25098226e-02 -9.34156656e-01 8.25059414e-01
-2.04479396e-01 7.27329373e-01 6.05945647e-01 -1.12016773e+00
8.76584947e-01 5.76898098e-01 2.67145544e-01 -9.90222394e-01
7.20882555e-03 -3.37542504e-01 -7.99581110e-01 -2.02125654e-01
-2.17979811e-02 -3.37046444e-01 -7.35240757e-01 1.56380129e+00
-3.17181677e-01 -2.48889685e-01 2.72652179e-01 5.23359478e-01
9.74412441e-01 6.24070168e-01 -2.35711664e-01 -4.16334331e-01
1.11772132e+00 -1.50377500e+00 -1.03571510e+00 -6.74542725e-01
2.05777556e-01 -1.07218301e+00 1.26266372e+00 4.48617369e-01
-1.43954861e+00 -7.67874241e-01 -1.34891045e+00 -3.07415903e-01
-7.06391856e-02 5.22832870e-01 3.72869782e-02 3.81504327e-01
-9.60303664e-01 7.32981443e-01 -6.91300869e-01 -3.71695608e-01
6.62969751e-03 -7.83351623e-03 -7.43035451e-02 2.44001627e-01
-9.36709225e-01 1.02900183e+00 2.80281126e-01 -2.41303727e-01
-5.42783558e-01 -4.75689977e-01 -7.09911644e-01 2.12749004e-01
4.42940027e-01 -1.27618980e+00 1.80093884e+00 -9.93530273e-01
-1.78186238e+00 7.10737288e-01 -6.05663478e-01 -3.51183236e-01
5.58097780e-01 -5.11222541e-01 -4.53485221e-01 -4.11080848e-03
2.73388743e-01 2.35320061e-01 1.06007349e+00 -1.07437015e+00
-5.22982240e-01 -1.89486668e-01 1.61163688e-01 2.04738766e-01
-2.49583840e-01 1.84781507e-01 -4.31080222e-01 -1.04527736e+00
-6.00921869e-01 -7.75710464e-01 -2.18173981e-01 -7.46553600e-01
-7.53875911e-01 -3.65754455e-01 6.81385815e-01 -5.55960536e-01
2.02102995e+00 -1.91128480e+00 6.03058398e-01 -3.72935683e-01
1.58874437e-01 3.35441887e-01 -2.38689184e-01 6.91999257e-01
3.65586877e-01 2.45104641e-01 -2.01563492e-01 -9.50776577e-01
2.41842084e-02 3.24839652e-01 -7.73869693e-01 -2.86565691e-01
8.82538557e-02 8.15855205e-01 -1.09698319e+00 -4.95136976e-01
-2.65421629e-01 -2.74358373e-02 -2.67691791e-01 4.75777090e-01
-4.32036877e-01 -2.29438092e-03 -5.09163320e-01 4.58762288e-01
3.99723053e-01 -3.17956418e-01 2.43989453e-01 7.20572844e-02
-1.09663241e-01 4.91345584e-01 -7.43862450e-01 1.89615893e+00
-4.11374599e-01 5.64642429e-01 -3.36188436e-01 -4.45933878e-01
1.18151367e+00 2.81331539e-01 -4.07858253e-01 -3.51089180e-01
7.56308883e-02 2.41687208e-01 -3.16503823e-01 -6.35381877e-01
1.01784861e+00 4.26410101e-02 -5.31260908e-01 7.71517992e-01
1.20960549e-01 -4.44940865e-01 8.32179427e-01 5.82943261e-01
1.33048487e+00 -9.53027010e-02 5.66466749e-01 -1.35280982e-01
6.38088048e-01 -2.30980903e-01 5.84000468e-01 1.25370371e+00
-3.69350170e-03 9.85529542e-01 7.13112056e-01 -2.32874125e-01
-9.65146244e-01 -1.14590073e+00 3.83895487e-01 9.99809682e-01
6.61160871e-02 -8.95840406e-01 -9.86185670e-01 -1.01460350e+00
-2.23935097e-01 1.22622871e+00 -5.84690392e-01 -2.90599287e-01
-7.58167565e-01 -1.96004763e-01 4.48875427e-01 6.60582066e-01
3.92045051e-01 -1.07591987e+00 -7.44009912e-01 1.89853936e-01
-4.48120952e-01 -1.06087828e+00 -1.18310940e+00 -4.62120064e-02
-5.05980909e-01 -7.13182807e-01 -4.38607603e-01 -5.13022065e-01
5.20324111e-01 3.71766090e-01 1.50102580e+00 5.49771674e-02
3.88944715e-01 1.11944132e-01 -3.59645575e-01 -5.77348948e-01
-9.20334458e-01 6.03893399e-01 -3.06608286e-02 -2.58139074e-01
1.67727858e-01 -4.08646673e-01 -4.13171440e-01 -1.20893319e-03
-1.02039850e+00 1.53898746e-01 4.99118596e-01 7.83899605e-01
1.37375399e-01 -4.21588123e-01 7.47823119e-01 -1.13042283e+00
1.43384409e+00 -2.35315681e-01 -1.22253634e-01 9.60425377e-01
-5.19454241e-01 2.14539334e-01 1.18815863e+00 -5.27856231e-01
-1.47783864e+00 -2.85327405e-01 2.00187892e-01 -3.27745497e-01
-1.19796664e-01 4.30520028e-01 -1.82413742e-01 7.14759767e-01
7.55951226e-01 2.99115509e-01 -1.34397000e-01 -5.14846623e-01
2.10842952e-01 8.12764883e-01 8.55019212e-01 -5.57858825e-01
6.12011790e-01 8.38875473e-02 -5.94358504e-01 -4.59305853e-01
-1.33865857e+00 -6.81246072e-02 -4.58704978e-01 -7.80104026e-02
6.46175086e-01 -7.25028038e-01 -1.65784061e-01 2.80162781e-01
-1.66669953e+00 -3.42546105e-02 -1.72965795e-01 -1.19147725e-01
-4.40120786e-01 4.71118867e-01 -5.74411333e-01 -5.26248515e-01
-7.97181845e-01 -1.34544647e+00 1.29738450e+00 7.46538818e-01
-1.02526045e+00 -1.02671278e+00 1.95475891e-01 2.18433902e-01
4.39389557e-01 1.32944360e-01 7.40899861e-01 -9.30493236e-01
-1.52175963e-01 -1.99739695e-01 2.06870630e-01 4.60782856e-01
5.93386590e-01 6.28594279e-01 -6.57727242e-01 -1.52269766e-01
1.95420891e-01 -1.14152648e-01 9.06407535e-01 -3.63299027e-02
8.79675686e-01 -5.08661628e-01 -2.46925205e-01 4.41086948e-01
9.10377204e-01 2.59522855e-01 3.33894312e-01 1.24000035e-01
4.91047889e-01 2.53353149e-01 4.11542684e-01 4.55575585e-01
3.63469064e-01 4.72588569e-01 -1.48789823e-01 4.77403283e-01
-2.39746764e-01 -4.16613817e-01 7.16505647e-01 9.93939698e-01
6.86305165e-02 -8.69988620e-01 -5.58967531e-01 2.74873376e-01
-1.97630882e+00 -1.25032783e+00 1.77732527e-01 1.74925208e+00
1.02982092e+00 3.71930599e-01 -1.10042736e-01 -2.05026254e-01
6.55649006e-01 5.09029090e-01 -6.02480292e-01 -9.67606604e-01
-4.59525138e-01 1.96019843e-01 -9.49337408e-02 5.50623417e-01
-8.53102982e-01 9.98976529e-01 7.42910528e+00 7.22039223e-01
-1.17855954e+00 -3.59920204e-01 3.91526312e-01 -3.97231549e-01
-7.03053772e-01 -6.33917376e-02 -1.06495440e+00 8.55407417e-01
9.28514838e-01 -7.32390940e-01 3.43588442e-01 5.90720832e-01
2.68581569e-01 -3.90051119e-02 -1.29671288e+00 5.80706298e-01
6.24803722e-01 -1.41511369e+00 5.79370856e-01 -4.34410185e-01
1.00134754e+00 -3.74214113e-01 4.90978137e-02 4.26145077e-01
4.83427137e-01 -8.01737905e-01 9.70240831e-01 9.83540297e-01
3.61152351e-01 -6.60819471e-01 5.45045614e-01 6.02144480e-01
-6.62780225e-01 -9.29285306e-03 -2.42969468e-01 -6.01231344e-02
3.54356050e-01 4.14167345e-01 -4.80017900e-01 7.32210159e-01
3.18423152e-01 6.36204123e-01 -1.07064438e+00 4.42783147e-01
-5.22558689e-01 4.22776818e-01 1.04767695e-01 -2.41694689e-01
8.90621617e-02 -2.51045525e-01 8.71941030e-01 1.36715770e+00
4.92467046e-01 -7.04248101e-02 2.09044129e-01 8.59085083e-01
-2.06124842e-01 -1.23941585e-01 -6.29133701e-01 -2.26711318e-01
6.72907054e-01 1.34461403e+00 -4.79845107e-01 -7.86073029e-01
-3.31609994e-01 1.28288710e+00 5.19643962e-01 5.36309063e-01
-7.39659607e-01 -6.51675344e-01 5.14468193e-01 -2.85808921e-01
3.70916307e-01 -9.90278460e-03 -6.94703877e-01 -1.59434819e+00
3.48342091e-01 -1.26092124e+00 2.00604469e-01 -9.52661455e-01
-1.34376895e+00 1.00247359e+00 1.35383248e-01 -1.15322340e+00
-5.97923577e-01 -2.29918435e-01 -1.18454897e+00 8.14875960e-01
-1.04875052e+00 -8.48739624e-01 -2.89511442e-01 5.19204214e-02
1.13519967e+00 -4.16986734e-01 7.03386068e-01 -1.93583310e-01
-9.36225593e-01 7.90505767e-01 -1.22212477e-01 -1.67071536e-01
1.10545015e+00 -1.49466956e+00 7.69938231e-01 1.04764366e+00
-1.88061260e-02 1.19411802e+00 1.12326849e+00 -7.75827348e-01
-1.10337460e+00 -8.42760086e-01 1.32695937e+00 -7.20173717e-01
6.93053901e-01 -2.70483375e-01 -1.04079294e+00 7.20552802e-01
9.50935483e-01 -6.90494120e-01 7.24491179e-01 -1.33365616e-01
-2.85653591e-01 -4.32208888e-02 -8.00633788e-01 1.05152190e+00
1.06799316e+00 -4.89577413e-01 -1.18587482e+00 2.06570551e-01
8.52869630e-01 -6.27030134e-01 -4.93553370e-01 4.67978925e-01
4.77529466e-01 -1.08641541e+00 4.96483624e-01 -6.76224828e-01
1.11041284e+00 -2.38981873e-01 -1.28418356e-01 -1.65756130e+00
-2.79270232e-01 -9.14321184e-01 -5.29690564e-01 1.68218386e+00
4.23251063e-01 -2.85872370e-01 1.72385290e-01 7.21121430e-01
-3.95248055e-01 -7.77042985e-01 -2.85481870e-01 -8.35111380e-01
3.00154500e-02 2.38120601e-01 8.39745641e-01 4.38012004e-01
7.93658271e-02 9.38755333e-01 -2.65437573e-01 -2.50401378e-01
2.05164388e-01 4.57977384e-01 8.82406473e-01 -9.06194508e-01
-4.44067001e-01 -8.83925676e-01 2.93257117e-01 -1.31269050e+00
5.36941886e-01 -7.63013363e-01 -1.17210541e-02 -1.47835624e+00
3.43179077e-01 3.35995376e-01 1.06335245e-01 1.08758062e-01
-9.12072659e-01 -3.03844154e-01 3.54631335e-01 1.34636149e-01
-7.58012056e-01 5.16807735e-01 1.32405078e+00 -3.82562548e-01
-2.44515315e-01 2.57624745e-01 -1.34070504e+00 6.32992387e-01
7.85717845e-01 -1.80399716e-01 -6.14584625e-01 -7.05868304e-01
2.55026698e-01 3.27067822e-01 -1.00094959e-01 -6.72563434e-01
4.16502297e-01 -3.00410926e-01 2.16891468e-01 -7.21423805e-01
-1.24197416e-01 -1.27630621e-01 8.58471468e-02 -3.47946025e-02
-7.11162806e-01 5.79620898e-01 -4.81794775e-02 4.39802945e-01
-1.78676561e-01 -5.48089147e-01 5.23425221e-01 -2.07743019e-01
-1.24554083e-01 -1.11486614e-01 -5.69895387e-01 5.85731864e-02
8.03768277e-01 -1.42924577e-01 -6.82476044e-01 -3.98987055e-01
-4.92796600e-01 3.31239432e-01 7.87700951e-01 7.52882779e-01
4.13187027e-01 -1.20803833e+00 -7.39621282e-01 -1.92430727e-02
7.46022165e-02 -3.35116312e-02 1.61466748e-01 4.27927941e-01
-3.01858217e-01 3.21857184e-01 4.33588549e-02 -3.09486330e-01
-1.26825511e+00 5.35540640e-01 1.84308827e-01 -6.57360435e-01
-4.52789456e-01 7.16014445e-01 1.47970870e-01 -1.24577448e-01
1.83476090e-01 -4.26700413e-01 -9.02597532e-02 1.31578371e-01
9.32601392e-01 4.46193397e-01 1.36290312e-01 -2.47612327e-01
-4.96438257e-02 3.45370978e-01 -6.08853459e-01 -6.67534620e-02
9.02376413e-01 -3.89159411e-01 -1.45373449e-01 7.99593031e-01
8.75522912e-01 2.61033386e-01 -1.08605421e+00 -2.19337493e-01
9.36055928e-02 -2.82558709e-01 -5.68998516e-01 -1.00030541e+00
-5.14140844e-01 4.85394716e-01 -4.98539746e-01 4.27370697e-01
1.07409728e+00 -1.21615924e-01 8.77641082e-01 6.08203709e-01
5.81649691e-02 -1.15537369e+00 2.73747325e-01 7.39443481e-01
1.43659747e+00 -1.04900098e+00 1.63021803e-01 -1.69777200e-01
-1.09208798e+00 1.29002130e+00 8.62623692e-01 -1.99884519e-01
6.31763712e-02 5.46721280e-01 2.76178509e-01 9.58008021e-02
-1.45830286e+00 4.01506096e-01 4.55235660e-01 1.53722957e-01
5.55094481e-01 2.00663302e-02 -4.95098114e-01 9.77850199e-01
-6.82286382e-01 -1.64453655e-01 7.70264804e-01 9.04119015e-01
-2.46971443e-01 -1.35285330e+00 -2.17676446e-01 5.51505864e-01
-3.86761189e-01 -9.66260117e-03 -1.04437447e+00 2.25487247e-01
-3.45733315e-01 1.12552786e+00 3.80881969e-03 -4.02598053e-01
5.24792612e-01 2.50855923e-01 6.10295355e-01 -8.72028649e-01
-1.15783620e+00 1.49589628e-02 1.64967954e-01 -2.29419082e-01
-1.99077353e-01 -7.95892715e-01 -9.16537225e-01 -3.91244411e-01
-5.45506142e-02 1.03812248e-01 1.27818346e-01 1.09695601e+00
5.86171329e-01 6.87342286e-01 7.37001836e-01 -6.84144378e-01
-9.74503160e-01 -1.11810756e+00 -1.84410840e-01 4.27189231e-01
4.21652049e-01 -2.10811675e-01 -3.38086575e-01 2.01070949e-01] | [12.209059715270996, 9.243310928344727] |
f3299660-6140-4155-bc2e-cac29e65a705 | arena-rosnav-2-0-a-development-and | 2302.10023 | null | https://arxiv.org/abs/2302.10023v1 | https://arxiv.org/pdf/2302.10023v1.pdf | Arena-Rosnav 2.0: A Development and Benchmarking Platform for Robot Navigation in Highly Dynamic Environments | Following up on our previous works, in this paper, we present Arena-Rosnav 2.0 an extension to our previous works Arena-Bench and Arena-Rosnav, which adds a variety of additional modules for developing and benchmarking robotic navigation approaches. The platform is fundamentally restructured and provides unified APIs to add additional functionalities such as planning algorithms, simulators, or evaluation functionalities. We have included more realistic simulation and pedestrian behavior and provide a profound documentation to lower the entry barrier. We evaluated our system by first, conducting a user study in which we asked experienced researchers as well as new practitioners and students to test our system. The feedback was mostly positive and a high number of participants are utilizing our system for other research endeavors. Finally, we demonstrate the feasibility of our system by integrating two new simulators and a variety of state of the art navigation approaches and benchmark them against one another. The platform is openly available at https://github.com/Arena-Rosnav. | ['Jens Lambrecht', 'Boris Meinardus', 'Teham Bhuiyan', 'Tuan Anh Le', 'Jacek Kmiecik', 'Huajian Zeng', 'Reyk Carstens', 'Linh Kästner'] | 2023-02-20 | null | null | null | null | ['robot-navigation'] | ['robots'] | [-7.77615964e-01 5.17958552e-02 2.26492658e-01 -2.88619041e-01
-5.50729513e-01 -1.09849000e+00 5.16655922e-01 -5.14739752e-02
-4.95015800e-01 8.50392401e-01 2.12321743e-01 -8.65143597e-01
1.07600629e-01 -7.94026732e-01 -6.44024670e-01 -3.43978256e-02
-3.10888886e-01 5.12040734e-01 7.80705690e-01 -1.13854516e+00
5.78999639e-01 4.76709843e-01 -1.93106890e+00 -2.02579945e-01
9.95924711e-01 1.74407259e-01 4.03527766e-01 8.69186997e-01
3.07042539e-01 6.64794445e-01 -5.27793705e-01 -1.70586050e-01
3.92617434e-01 -2.68937051e-02 -9.09828246e-01 -8.11039567e-01
3.19205642e-01 -3.67811292e-01 -4.89996016e-01 5.62196612e-01
9.13781047e-01 6.85068130e-01 -7.96307176e-02 -1.58407128e+00
-2.13582352e-01 4.72232550e-01 -2.22649574e-02 9.27978531e-02
1.15487123e+00 6.14614308e-01 3.45575005e-01 -5.00728309e-01
9.59190905e-01 1.10740328e+00 9.63861883e-01 4.67110276e-01
-1.04537785e+00 -6.70434952e-01 4.12431955e-02 1.27180830e-01
-1.37853777e+00 -3.39498699e-01 2.75894463e-01 -3.87077481e-01
1.02133572e+00 2.61475623e-01 8.27991903e-01 1.44904959e+00
2.70928234e-01 4.87489998e-01 8.88666451e-01 -1.47776306e-01
6.37009263e-01 3.35229218e-01 3.28561097e-01 7.48595715e-01
2.21205115e-01 3.66164744e-01 -3.02903820e-02 -1.16877943e-01
8.84379625e-01 -4.03829962e-01 -1.57324195e-01 -9.04680848e-01
-1.34358549e+00 5.51285863e-01 6.38981640e-01 -6.21814048e-03
-9.58228707e-02 2.39428699e-01 3.33184391e-01 8.14053882e-03
-5.14854252e-01 5.01184285e-01 -1.55418888e-01 -7.82769382e-01
-1.85050890e-01 9.78202164e-01 1.13053727e+00 1.35427642e+00
3.55850935e-01 -2.50405014e-01 -2.07693074e-02 5.53573191e-01
4.04430211e-01 1.69564307e-01 2.94732660e-01 -1.61709476e+00
2.33631685e-01 3.37337255e-01 8.13808262e-01 -7.89676011e-01
-5.69546282e-01 -9.01262611e-02 1.34220213e-01 9.00844753e-01
4.68028694e-01 -4.65267897e-01 -7.11401403e-01 1.36326766e+00
4.16379720e-01 -5.75097576e-02 5.11996523e-02 1.13734317e+00
1.16357017e+00 4.31396931e-01 1.25919074e-01 8.58796120e-01
1.30406928e+00 -1.49942923e+00 -4.27734852e-01 -1.41799808e-01
8.63711953e-01 -9.00066793e-01 1.71953166e+00 4.77509916e-01
-1.09301651e+00 -6.55251205e-01 -1.26907158e+00 9.63022560e-03
-7.98461139e-01 -6.22073673e-02 8.39126110e-01 9.07417059e-01
-1.27554154e+00 6.98820591e-01 -1.17824066e+00 -9.68928099e-01
-1.38762593e-01 2.16758162e-01 -4.09286052e-01 -8.20076186e-03
-8.96279991e-01 1.26188111e+00 1.11278789e-02 -3.19524586e-01
-9.06091213e-01 -4.48052585e-01 -1.00007045e+00 -4.06083614e-01
1.82767630e-01 -7.17388213e-01 2.06641769e+00 2.42750496e-01
-1.78896284e+00 4.84096855e-01 1.15013748e-01 -2.14514777e-01
8.35174859e-01 -3.82241666e-01 -1.80273637e-01 -3.21252942e-01
6.00622475e-01 7.13965416e-01 -7.23068893e-01 -1.29926944e+00
-6.74806714e-01 -1.29011855e-01 6.29053831e-01 4.46547359e-01
5.59836566e-01 -4.85910848e-02 -5.61685801e-01 -2.00028747e-01
-2.04059526e-01 -1.11289907e+00 -8.33739161e-01 -1.33925416e-02
-5.86894862e-02 2.41088450e-01 6.84276760e-01 -2.64107198e-01
7.75726497e-01 -1.81021690e+00 -2.05562100e-01 1.22376867e-01
-3.10534298e-01 -1.29248966e-02 5.20199828e-04 1.04575133e+00
2.78882861e-01 3.53338644e-02 1.13289617e-01 -7.20723718e-02
2.49843970e-01 1.30468071e-01 -1.53083533e-01 2.79774308e-01
-6.30898118e-01 5.29798269e-01 -1.24881399e+00 -1.58585429e-01
7.33969569e-01 6.42926753e-01 -4.85145330e-01 -1.13914564e-01
4.08018604e-02 7.99280286e-01 -3.75841945e-01 6.71456754e-01
7.09568262e-01 9.58406776e-02 -1.04637243e-01 3.61428052e-01
-7.95397162e-01 5.97175777e-01 -1.56608284e+00 2.40333200e+00
-5.50740898e-01 5.08231223e-01 6.21971376e-02 -2.46473327e-01
5.40042102e-01 -3.68389413e-02 1.00390144e-01 -6.54716015e-01
3.07721198e-01 2.72538096e-01 -1.11176610e-01 -6.36446536e-01
1.01804292e+00 7.58476257e-01 -2.20999569e-01 1.85977489e-01
-1.42995492e-01 -4.48495239e-01 5.76158345e-01 3.17635030e-01
1.30283201e+00 9.05768514e-01 3.01546305e-01 -3.89029324e-01
3.37390572e-01 8.61920714e-01 -1.00575231e-01 1.05905032e+00
-5.79733312e-01 3.68895739e-01 2.94795651e-02 -3.22513878e-01
-9.44682181e-01 -1.46970809e+00 -8.59681889e-02 1.14086008e+00
8.96512926e-01 -9.59024727e-01 -5.87110102e-01 -3.93854082e-01
-7.93277659e-03 9.65172768e-01 -6.12750709e-01 4.85855252e-01
-4.40811247e-01 -4.89199124e-02 7.57226348e-01 7.24819243e-01
4.88747030e-01 -1.09998059e+00 -1.17532122e+00 2.41454337e-02
-1.72450155e-01 -1.01031220e+00 -8.55900049e-02 -1.31895900e-01
-4.60141689e-01 -9.64785457e-01 -3.29452515e-01 -9.06292140e-01
2.92009324e-01 6.81984305e-01 9.52360928e-01 4.82674465e-02
-4.16180611e-01 6.81890488e-01 -5.83536923e-01 -4.47282106e-01
-2.44390473e-01 1.90798879e-01 -2.03789100e-01 -1.56333792e+00
-1.62658080e-01 -7.22107410e-01 -8.38307738e-01 8.67474973e-01
-3.80590618e-01 9.83064100e-02 1.40869364e-01 2.34968185e-01
2.21981749e-01 -5.44488728e-01 5.63123710e-02 -3.79284084e-01
1.02833211e+00 -4.49520975e-01 -9.42041516e-01 -2.98680782e-01
-2.42387936e-01 -2.64430910e-01 3.01842004e-01 -2.04497427e-01
-7.49275565e-01 3.09147891e-02 -6.09551549e-01 3.60614300e-01
-4.54528511e-01 4.16219175e-01 1.52184650e-01 -5.24929464e-01
1.13672912e+00 -2.43153915e-01 1.31008342e-01 -1.34280637e-01
6.29445493e-01 4.82196540e-01 7.50316501e-01 -8.63284707e-01
4.36664253e-01 4.86432910e-01 -3.57374579e-01 -4.34217840e-01
1.39399424e-01 -4.31884915e-01 -3.79694879e-01 -4.01231140e-01
4.34917271e-01 -8.33709717e-01 -1.28828621e+00 1.27828375e-01
-7.78986871e-01 -8.80926609e-01 -3.63925070e-01 6.77305162e-01
-6.28058076e-01 1.66281104e-01 -4.02778924e-01 -5.78598619e-01
4.16718461e-02 -1.65925705e+00 8.47082913e-01 6.88497424e-01
-4.27810520e-01 -7.28200793e-01 7.57718384e-01 8.60252082e-02
7.70086765e-01 3.28786075e-01 -5.61544634e-02 -4.27785367e-01
-5.07958055e-01 -3.88276845e-01 3.75715755e-02 -4.93725002e-01
-2.75434732e-01 7.61721209e-02 -6.17570996e-01 -1.57073900e-01
-8.46747816e-01 -2.78780043e-01 7.43596926e-02 1.43484622e-01
6.91155434e-01 3.22741628e-01 -7.92087674e-01 2.05983832e-01
1.21958578e+00 4.79234993e-01 9.12733376e-01 1.41563463e+00
2.47415483e-01 5.10402381e-01 1.02861261e+00 3.08316976e-01
1.01010406e+00 9.26369667e-01 5.52323222e-01 1.24811433e-01
2.91294515e-01 -3.08378339e-01 2.94318289e-01 1.45975128e-01
-5.27116477e-01 -9.93704870e-02 -1.25906336e+00 4.70677823e-01
-2.03650451e+00 -9.32242692e-01 -4.71089363e-01 2.22938848e+00
2.26140469e-01 -5.73239848e-02 5.61320066e-01 -3.15856218e-01
2.69802123e-01 -2.22972944e-01 -2.77431369e-01 -5.51145375e-01
5.03643930e-01 3.67281735e-01 4.32289600e-01 6.74670994e-01
-9.55369830e-01 1.18984950e+00 6.99700022e+00 1.65916860e-01
-9.77982521e-01 -1.10890329e-01 -3.38646770e-01 8.31482038e-02
1.40237018e-01 3.74973208e-01 -5.36034465e-01 1.50786772e-01
8.67173374e-01 -4.32527512e-01 4.92503166e-01 1.44608033e+00
4.39351171e-01 -4.55148250e-01 -7.48049915e-01 4.95117426e-01
-3.79904419e-01 -1.28888071e+00 -4.61095452e-01 3.82177555e-03
2.63578415e-01 4.76366729e-01 -1.53693572e-01 9.85413969e-01
1.00569427e+00 -1.09133506e+00 8.45854282e-01 1.43478975e-01
1.61425203e-01 -5.49623847e-01 7.93841422e-01 2.29081601e-01
-1.24090755e+00 9.54572111e-02 -1.53825909e-01 -6.16120040e-01
4.34360236e-01 -6.41318202e-01 -7.66626358e-01 6.52989388e-01
1.19856334e+00 3.00264865e-01 -6.13904357e-01 1.73576307e+00
-2.85288662e-01 -2.12059878e-02 -3.40955228e-01 -3.00591409e-01
2.42105037e-01 -2.60182291e-01 5.60543954e-01 1.15300739e+00
3.39498281e-01 -9.45649669e-02 3.38337839e-01 6.05376542e-01
5.62967002e-01 1.61547258e-01 -8.50644112e-01 6.14380240e-01
6.30486429e-01 1.41132092e+00 -8.41782987e-01 -1.08524069e-01
-1.26280591e-01 7.41259277e-01 6.29707202e-02 1.92036852e-01
-1.21197152e+00 -7.17973292e-01 9.76550400e-01 2.40461275e-01
7.07966387e-02 -7.37744033e-01 -3.54769230e-01 -7.77121067e-01
-7.63409659e-02 -6.88379467e-01 7.66684264e-02 -1.28936315e+00
-6.30811751e-01 7.38115311e-01 2.17896476e-01 -1.54717386e+00
-2.99221039e-01 -5.79057574e-01 -8.43164265e-01 7.08567977e-01
-1.11266279e+00 -9.81484294e-01 -1.15454113e+00 3.08097601e-01
4.51275975e-01 1.05651759e-01 1.01614702e+00 4.16205823e-01
-3.50359917e-01 3.64069074e-01 -4.14046198e-02 -1.60424337e-01
8.09678912e-01 -1.20037329e+00 1.00305843e+00 6.75066173e-01
-5.32933474e-01 1.20100224e+00 1.16347539e+00 -5.23833275e-01
-1.36064732e+00 -5.56263030e-01 -5.99304885e-02 -7.56809652e-01
7.44006336e-01 -3.60717863e-01 -3.62543732e-01 1.10545540e+00
6.16725087e-01 -1.55938610e-01 4.80965644e-01 9.93983001e-02
3.50964554e-02 3.59878987e-01 -1.35806668e+00 1.27630067e+00
1.41329837e+00 1.66113630e-01 -5.45305431e-01 -7.26565346e-03
5.65130949e-01 -1.27416766e+00 -6.10506952e-01 4.97975975e-01
1.16166818e+00 -1.38680065e+00 1.07782483e+00 -2.33525246e-01
3.31722796e-01 -6.84498191e-01 -1.82061180e-01 -1.46954310e+00
-1.23325408e-01 -6.16893768e-01 4.57801640e-01 8.84266794e-01
5.16607344e-01 -7.86073208e-01 6.64861262e-01 5.01912177e-01
-5.77885747e-01 -4.98326749e-01 -4.26166624e-01 -7.58268237e-01
1.72934998e-02 -6.02870882e-01 6.21464252e-01 5.82109272e-01
5.54204881e-01 1.46432117e-01 9.00902152e-02 4.56993133e-01
2.03074679e-01 -2.55106091e-01 1.77778757e+00 -8.67657363e-01
-1.14079162e-01 -4.89386797e-01 -5.63732207e-01 -1.09151947e+00
-4.59442973e-01 -6.92125499e-01 2.01126099e-01 -2.23979616e+00
-3.82754743e-01 -4.64017272e-01 3.24568570e-01 3.19564819e-01
4.63440537e-01 9.28124264e-02 1.72387749e-01 1.04152067e-02
-7.70029068e-01 3.79504830e-01 1.08224535e+00 3.97161365e-01
-6.83517158e-01 1.49830878e-01 -8.50764334e-01 7.00792372e-01
1.05522287e+00 -9.18532237e-02 -5.25084853e-01 -2.27389216e-01
-7.41739869e-02 -8.52694660e-02 7.00757682e-01 -1.71354461e+00
4.51951295e-01 -3.16582583e-02 1.19453296e-01 -6.49368107e-01
4.46824700e-01 -5.34353852e-01 2.61484623e-01 4.91739601e-01
8.45097452e-02 5.81687689e-01 8.52100313e-01 4.58306000e-02
2.50073522e-01 -7.87638500e-02 3.05381954e-01 -3.63627285e-01
-1.09869289e+00 -3.57387871e-01 -8.39325249e-01 -1.27139136e-01
1.45145011e+00 -5.25330663e-01 -8.60000670e-01 -6.04889214e-01
-7.55156636e-01 7.20396996e-01 8.76228511e-01 7.54031777e-01
4.71101403e-01 -9.61817622e-01 -1.46525219e-01 -5.81865609e-02
3.95363003e-01 -1.63625017e-01 2.39313826e-01 6.97525501e-01
-1.40427446e+00 2.41582319e-01 -7.63365090e-01 -6.24551177e-01
-1.19673014e+00 2.03782901e-01 3.43551666e-01 1.74946845e-01
-7.56948352e-01 4.14297432e-01 -3.91335577e-01 -1.49037004e+00
3.28809768e-01 -4.25617129e-01 -2.10646719e-01 -7.92651474e-01
4.25451159e-01 7.33793557e-01 3.99088152e-02 -4.12164867e-01
-5.87902009e-01 3.18512261e-01 1.62303030e-01 -6.15975499e-01
1.17457223e+00 -1.90140843e-01 4.17117387e-01 3.93331319e-01
3.46720845e-01 4.76072550e-01 -9.64247525e-01 6.59742236e-01
-2.42399216e-01 -5.10910213e-01 -4.51352149e-01 -1.11353827e+00
-2.20558524e-01 4.37563300e-01 8.50431740e-01 -1.80318467e-02
3.59130025e-01 -4.49503288e-02 3.39844167e-01 5.83316684e-01
1.16607273e+00 -7.88745999e-01 -2.66131788e-01 7.47104526e-01
9.60773706e-01 -1.02142286e+00 -1.42679334e-01 -5.42862594e-01
-6.83737874e-01 8.86157751e-01 1.16494370e+00 -4.04835910e-01
5.02217829e-01 5.99020600e-01 3.93312246e-01 -2.53675759e-01
-3.32494438e-01 -3.40523660e-01 -3.82777601e-01 1.10284197e+00
5.85143149e-01 -6.00569993e-02 -5.35172045e-01 6.73625469e-01
-9.98665512e-01 3.18898767e-01 9.90398765e-01 1.65785539e+00
-3.03345472e-01 -1.20023441e+00 -5.19437909e-01 -3.09702933e-01
6.18490651e-02 3.38563859e-01 -1.78508699e-01 1.53067708e+00
-1.68570757e-01 1.03149796e+00 -3.07504416e-01 -6.20476544e-01
1.06904984e+00 -4.65173930e-01 5.85842371e-01 -3.78049552e-01
-6.73154175e-01 -5.89736760e-01 4.13781822e-01 -1.04960644e+00
1.03653446e-01 -5.84955812e-01 -1.58326483e+00 -7.50236988e-01
6.42571300e-02 2.81107634e-01 1.08344305e+00 2.40670383e-01
7.32027709e-01 6.30059302e-01 -8.31243992e-02 -1.54945374e+00
-1.89804360e-01 -7.99902260e-01 -6.39465004e-02 -1.80502504e-01
-1.36580214e-01 -1.04808700e+00 4.84131649e-02 -4.33172315e-01] | [4.577202320098877, 0.9560837745666504] |
fe6b633f-c7bc-4fa1-95f0-ae1620fcca1f | neural-wave-functions-for-superfluids | 2305.06989 | null | https://arxiv.org/abs/2305.06989v2 | https://arxiv.org/pdf/2305.06989v2.pdf | Neural Wave Functions for Superfluids | Understanding superfluidity remains a major goal of condensed matter physics. Here we tackle this challenge utilizing the recently developed Fermionic neural network (FermiNet) wave function Ansatz for variational Monte Carlo calculations. We study the unitary Fermi gas, a system with strong, short-range, two-body interactions known to possess a superfluid ground state but difficult to describe quantitatively. We demonstrate key limitations of the FermiNet Ansatz in studying the unitary Fermi gas and propose a simple modification that outperforms the original FermiNet significantly, giving highly accurate results. We prove mathematically that the new Ansatz, which only differs from the original Ansatz by the method of antisymmetrization, is a strict generalization of the original FermiNet architecture, despite the use of fewer parameters. Our approach shares several advantages with the FermiNet: the use of a neural network removes the need for an underlying basis set; and the flexibility of the network yields extremely accurate results within a variational quantum Monte Carlo framework that provides access to unbiased estimates of arbitrary ground-state expectation values. We discuss how the method can be extended to study other superfluids. | ['James S. Spencer', 'David Pfau', 'Johannes Knolle', 'W. M. C. Foulkes', 'Gino Cassella', 'Halvard Sutterud', 'Wan Tong Lou'] | 2023-05-11 | null | null | null | null | ['variational-monte-carlo'] | ['miscellaneous'] | [ 1.25550792e-01 -2.61427790e-01 1.78541988e-01 -2.70085752e-01
-7.78174624e-02 -4.95525658e-01 8.15174937e-01 -4.84936565e-01
-5.20488739e-01 1.20108211e+00 -1.18959032e-01 -5.63214600e-01
-1.70643419e-01 -9.00694013e-01 -5.25838494e-01 -1.36604941e+00
-3.04132879e-01 6.99164271e-01 3.19206476e-01 -8.87903154e-01
2.89285719e-01 3.21031034e-01 -1.26094425e+00 -6.35982305e-02
9.67641056e-01 7.19463646e-01 -2.34695241e-01 6.77741945e-01
5.21765232e-01 5.92259943e-01 -8.97564590e-02 2.67720339e-03
5.44823468e-01 -7.51192093e-01 -9.88678336e-01 -7.24472761e-01
5.29842436e-01 -1.94080308e-01 -6.95495486e-01 1.36140513e+00
2.14282647e-01 6.90905273e-01 6.26600742e-01 -6.18622780e-01
-7.09866941e-01 5.10600746e-01 1.45198464e-01 3.63280296e-01
-1.69198647e-01 2.55587757e-01 1.27368617e+00 -4.86510605e-01
5.02658546e-01 8.38039041e-01 7.22605765e-01 1.00412917e+00
-1.51010203e+00 -4.64884907e-01 -6.95614934e-01 1.05549455e-01
-1.43987978e+00 -3.95069391e-01 5.05254745e-01 -4.02147353e-01
1.37926579e+00 3.06532741e-01 8.94159496e-01 5.52266479e-01
5.66303849e-01 -2.63384432e-02 1.23301625e+00 -8.46919000e-01
5.22725701e-01 -2.99688548e-01 6.15689456e-01 6.14572525e-01
6.23646438e-01 6.01120949e-01 -2.84424692e-01 -3.63713950e-01
6.24110699e-01 -1.65776074e-01 -5.09716749e-01 -7.55976737e-01
-9.56801534e-01 1.02118027e+00 4.92465258e-01 8.01656187e-01
4.59773578e-02 6.76941335e-01 2.40431800e-01 1.99927762e-01
2.08414689e-01 8.46204579e-01 -1.07016586e-01 -1.04256764e-01
-1.16275394e+00 7.25320518e-01 1.33522975e+00 2.12530959e-02
7.59341061e-01 4.19145554e-01 1.31930798e-01 -1.51163474e-01
1.21072985e-01 5.90848207e-01 1.58978075e-01 -1.23194265e+00
-4.73338924e-02 -1.27648920e-01 4.64739054e-01 -4.52848762e-01
-4.04947340e-01 -2.00387761e-01 -1.00228906e+00 3.30367565e-01
3.42152774e-01 -2.81012990e-02 -7.31598735e-01 1.65482593e+00
1.00033462e-01 -6.28508106e-02 3.98610681e-02 8.60332906e-01
6.35144293e-01 6.81517065e-01 -1.72522455e-01 -2.45421126e-01
8.00984502e-01 -5.62529564e-01 -5.74738145e-01 4.22080278e-01
5.41147888e-01 -2.94641387e-02 4.82889801e-01 3.61415267e-01
-1.13507986e+00 -2.66432315e-01 -1.31547654e+00 -4.68756258e-02
-3.90498042e-01 -5.53233922e-01 1.26124525e+00 1.02917385e+00
-1.43390191e+00 1.50649095e+00 -1.24132276e+00 -1.02367766e-01
-2.15591848e-01 9.32531476e-01 -4.24935251e-01 4.33552742e-01
-1.61433268e+00 1.10166299e+00 7.94913113e-01 1.80495922e-02
-5.69922328e-01 -5.17364562e-01 -7.57313192e-01 -1.53592639e-02
7.81176612e-02 -6.93916976e-01 1.31348968e+00 -5.67056000e-01
-1.58033419e+00 4.97901738e-01 -3.36093962e-01 -4.72454876e-01
4.41364348e-02 3.06526899e-01 -2.26215139e-01 1.49959713e-01
-7.91386217e-02 1.84347779e-01 1.82262018e-01 -6.91749752e-01
2.37064317e-01 -7.93126822e-02 2.23525271e-01 -2.92085826e-01
-2.21754521e-01 -2.67384619e-01 2.82208383e-01 -1.88130420e-02
2.05781579e-01 -9.92777109e-01 -5.21132529e-01 -7.77964830e-01
-9.88386869e-02 -3.42334896e-01 4.56351727e-01 -2.27373049e-01
1.14145052e+00 -1.56885195e+00 4.13122326e-01 5.05735695e-01
6.23378038e-01 4.13385004e-01 2.86481410e-01 7.21083283e-01
-3.47171724e-01 1.27657428e-01 -4.93694723e-01 2.12072864e-01
1.28291905e-01 2.86155164e-01 -7.31307268e-02 6.68211222e-01
-9.47047248e-02 1.12860608e+00 -1.00731027e+00 4.91037741e-02
4.16457713e-01 4.96954262e-01 -1.11437726e+00 -2.77731508e-01
-1.70108333e-01 7.86013722e-01 -3.14550400e-01 2.59598158e-02
8.46517026e-01 -4.99360502e-01 5.84324479e-01 -2.78932452e-01
-6.22376263e-01 5.07252336e-01 -1.07807946e+00 1.39407754e+00
-1.25741391e-02 3.90064061e-01 1.83449224e-01 -1.37189329e+00
4.93977338e-01 3.30988437e-01 4.91209209e-01 -6.39688849e-01
3.34032744e-01 6.92631185e-01 7.69236267e-01 -3.39519233e-01
6.46745920e-01 -8.92920911e-01 -3.84780578e-02 6.08245432e-01
4.36055481e-01 -4.45414811e-01 6.28628135e-01 2.74020642e-01
1.16024923e+00 7.45889619e-02 4.49541390e-01 -1.09092581e+00
5.15479445e-01 -1.91791341e-01 2.62804270e-01 1.12066734e+00
-2.44166031e-01 2.09927276e-01 4.31097358e-01 -8.52154195e-01
-1.33388078e+00 -1.03365040e+00 -5.67048013e-01 5.78717530e-01
2.12508261e-01 -7.18989670e-01 -7.56680727e-01 -1.31321903e-02
4.61467691e-02 6.68334842e-01 -7.92390645e-01 -2.08968550e-01
-8.22504282e-01 -1.34239066e+00 3.91923875e-01 2.49031097e-01
4.63258743e-01 -9.35289204e-01 -5.92898726e-01 8.05884674e-02
2.42522880e-01 -6.86343014e-01 8.54529291e-02 6.37705922e-01
-7.08865762e-01 -1.04538047e+00 -3.09433132e-01 -9.17421728e-02
3.16977084e-01 -1.76661223e-01 1.13663876e+00 3.75912070e-01
-3.78208250e-01 -1.08280957e-01 -4.90113087e-02 5.25262415e-01
-7.15540528e-01 1.89531073e-02 5.78342557e-01 -5.35736799e-01
7.35036492e-01 -8.17171097e-01 -4.62858826e-01 -3.76144677e-01
-8.65125418e-01 -1.76564798e-01 -7.35098049e-02 9.93739188e-01
6.30103871e-02 2.08873078e-01 -4.02931683e-02 -8.53809714e-01
5.40661275e-01 -4.02906448e-01 -8.02602410e-01 -1.63913518e-01
-3.87777209e-01 5.91606438e-01 8.81218016e-01 2.75538921e-01
-8.02562714e-01 -3.82231086e-01 -4.28273320e-01 1.74062878e-01
2.35428900e-01 1.23641320e-01 2.84355104e-01 -1.00085306e+00
6.22833788e-01 2.17561156e-01 -2.98488915e-01 -5.10456622e-01
1.86632246e-01 4.25524265e-01 4.94122684e-01 -8.57604921e-01
9.24599051e-01 5.35563946e-01 7.65346527e-01 -7.41558790e-01
-6.27712309e-01 -1.85030565e-01 -1.01933038e+00 1.17360383e-01
9.51890051e-01 -4.63177413e-01 -1.47872162e+00 2.26146415e-01
-1.15137875e+00 -1.74851701e-01 -3.65284801e-01 6.36975467e-01
-5.43385327e-01 5.38702667e-01 -8.56472671e-01 -9.73330975e-01
-4.08478886e-01 -1.13679087e+00 5.78539789e-01 1.02533057e-01
-1.84422806e-01 -1.36733055e+00 5.46522439e-01 5.08273765e-02
8.34674239e-01 2.80709773e-01 1.03919053e+00 -5.18152356e-01
-6.08913720e-01 -5.34256212e-02 -4.47583795e-02 4.02386039e-01
-1.86013550e-01 -4.18962687e-02 -6.52896583e-01 -5.40198028e-01
2.89729446e-01 -3.55400562e-01 1.56550694e+00 3.88784796e-01
9.08725083e-01 8.85963347e-03 -1.66523784e-01 7.08110154e-01
1.60751832e+00 1.99662581e-01 3.45509917e-01 -1.15376212e-01
7.76022255e-01 7.58045912e-02 -5.33202589e-01 3.70370075e-02
2.09919699e-02 5.92586875e-01 3.22622389e-01 4.67291415e-01
2.78738111e-01 1.26235858e-01 2.10044682e-01 1.18029785e+00
-9.87107813e-01 6.19714614e-03 -1.01075709e+00 1.56603500e-01
-1.74720371e+00 -1.31345701e+00 -4.24876928e-01 2.28557324e+00
7.72055864e-01 -6.18166476e-03 -1.44335866e-01 5.52807860e-02
3.98647875e-01 2.82622248e-01 -3.97661388e-01 -4.44162488e-01
-1.96592771e-02 1.14136136e+00 6.46648169e-01 9.16545868e-01
-1.20138180e+00 8.87020767e-01 7.92730427e+00 7.26586640e-01
-1.04652882e+00 5.96109927e-01 1.25806540e-01 -3.29422623e-01
-3.83846223e-01 3.30015928e-01 -6.78322375e-01 5.08794129e-01
1.31955242e+00 -3.50770019e-02 9.06460464e-01 4.06758577e-01
1.26793027e-01 -1.17765583e-01 -1.06158328e+00 6.28022611e-01
-7.00955912e-02 -1.86213648e+00 -6.77790046e-02 4.13075387e-02
1.08998835e+00 3.93727154e-01 -2.52445281e-01 1.23768389e-01
4.89446312e-01 -1.26573765e+00 5.01744330e-01 3.40380609e-01
8.82755101e-01 -5.76065242e-01 6.69548631e-01 1.42838597e-01
-8.96566689e-01 2.24907830e-01 -4.76786226e-01 -7.71455109e-01
2.31775478e-01 4.97167796e-01 5.24386056e-02 3.73183578e-01
5.04546523e-01 4.68580544e-01 -1.87672824e-01 6.64521217e-01
7.17072040e-02 5.74929297e-01 -4.41816509e-01 -4.35580432e-01
4.84757066e-01 -6.71504140e-01 5.09776473e-01 8.53385806e-01
1.75594360e-01 3.23950112e-01 8.10600072e-02 1.30321193e+00
-2.86410041e-02 -2.40545139e-01 -4.82519805e-01 -2.38756299e-01
-6.10102899e-03 1.08989000e+00 -6.64192855e-01 -2.65607178e-01
-1.35482594e-01 5.78823745e-01 3.40860844e-01 3.23068678e-01
-3.02711010e-01 -4.90980238e-01 4.31888312e-01 8.31003636e-02
4.67020661e-01 -4.85772312e-01 1.05584167e-01 -1.69127536e+00
-2.01019838e-01 -5.30197263e-01 -2.09417686e-01 -2.84707189e-01
-1.10066855e+00 5.39122939e-01 1.90074846e-01 -5.20020008e-01
-4.24355060e-01 -1.30141449e+00 -8.41092348e-01 9.87903833e-01
-1.19354320e+00 -8.25083733e-01 2.02977851e-01 3.57036859e-01
-5.11528075e-01 -5.18940538e-02 9.57528353e-01 1.71143234e-01
-3.35701495e-01 7.78001472e-02 7.95069993e-01 -1.32332489e-01
2.80738063e-02 -1.66905272e+00 5.24470866e-01 8.29592764e-01
-1.45511612e-01 1.25546002e+00 1.15931857e+00 -6.50932491e-01
-1.52212954e+00 -5.74018180e-01 8.67949486e-01 -4.91537839e-01
1.03242850e+00 -5.64326406e-01 -6.89872682e-01 5.96595705e-01
2.90198863e-01 2.31168509e-01 5.85025549e-01 1.90723509e-01
-9.17454734e-02 2.52659172e-01 -1.06243265e+00 3.13588768e-01
9.60141301e-01 -6.32080019e-01 -5.76295137e-01 6.19228661e-01
3.27755988e-01 -1.45688504e-01 -6.20513916e-01 3.58614981e-01
6.53119743e-01 -1.39917171e+00 7.22781360e-01 -8.23017657e-01
2.76172817e-01 -1.86467499e-01 -2.04527915e-01 -1.07180870e+00
-5.84130526e-01 -1.07781589e+00 -2.97601938e-01 2.41121113e-01
-2.24501248e-02 -8.14862549e-01 6.98037624e-01 6.46506190e-01
-2.94176452e-02 -5.29127002e-01 -1.52240753e+00 -8.02664161e-01
8.94848168e-01 -1.52731135e-01 4.13534731e-01 6.97921813e-01
4.36913818e-01 1.13565937e-01 -5.59030235e-01 -3.06703180e-01
6.94920003e-01 1.71580359e-01 1.26076207e-01 -1.40569341e+00
-5.84038019e-01 -3.40711415e-01 -4.17839855e-01 -8.10117841e-01
5.21334708e-01 -1.20784461e+00 2.67875399e-02 -8.95323515e-01
5.13325810e-01 -1.81368545e-01 -4.05891031e-01 1.59714714e-01
2.11609244e-01 8.60177577e-01 -1.48216654e-02 2.82223314e-01
-6.50674701e-01 5.55285573e-01 1.07311201e+00 1.40261635e-01
8.24246854e-02 -2.43831754e-01 -3.47675085e-01 8.15127850e-01
5.82768142e-01 -4.10482973e-01 1.85186088e-01 6.80552656e-03
7.03448415e-01 -2.55813658e-01 6.98707461e-01 -1.25277805e+00
2.22301722e-01 2.63687689e-02 2.03882307e-01 -1.63128451e-01
3.95632863e-01 -1.71091527e-01 1.79741561e-01 9.68250751e-01
4.95708920e-02 -3.42044234e-01 -1.33412272e-01 1.97157741e-01
4.35160892e-03 -7.44801462e-01 1.14036334e+00 -7.17952192e-01
-5.25799811e-01 1.93341419e-01 -4.26278651e-01 -1.25046477e-01
7.54706204e-01 1.34898312e-02 -2.24011973e-01 -3.75912450e-02
-7.67152071e-01 -2.75212228e-01 7.14694977e-01 -3.95056546e-01
-1.15553871e-01 -1.33077824e+00 -3.28764558e-01 5.35609305e-01
-3.40975583e-01 -6.62075043e-01 2.47888044e-01 9.07940984e-01
-1.08914375e+00 1.03651726e+00 -1.93828568e-01 -2.42736682e-01
-6.27016246e-01 1.82643846e-01 8.92061591e-01 -5.69277585e-01
-4.45424020e-01 7.14226902e-01 1.85802206e-01 -6.64818883e-01
-7.93913662e-01 -2.79013455e-01 9.36521962e-02 -5.46757400e-01
4.91420209e-01 3.79191279e-01 1.57322705e-01 -8.90763223e-01
-2.98858583e-01 4.73028392e-01 1.09927408e-01 2.00140756e-03
1.41867423e+00 3.51905316e-01 -8.17922056e-01 4.03354228e-01
9.52991664e-01 -3.59514821e-03 -6.78978622e-01 4.11979631e-02
-4.22869831e-01 -1.85667664e-01 2.99262643e-01 -3.61936539e-01
-7.23783970e-01 1.10444021e+00 2.05075741e-01 5.55627167e-01
2.96542346e-01 -3.90009359e-02 9.05501842e-01 9.87948716e-01
6.16324723e-01 -8.06556821e-01 -5.22583544e-01 8.77019465e-01
3.28747153e-01 -1.13035059e+00 1.98111292e-02 2.38000005e-02
2.02010468e-01 1.44935596e+00 5.18562496e-01 -5.61795771e-01
8.13174486e-01 9.57386941e-02 -5.71646690e-01 -5.03128827e-01
-4.54955220e-01 -5.54904249e-03 2.63124496e-01 2.57633448e-01
6.05713010e-01 2.77055532e-01 -5.02194524e-01 4.62788790e-01
-5.06052017e-01 2.87056919e-02 7.43142366e-01 6.79045737e-01
-6.36765003e-01 -1.32459164e+00 -7.99328238e-02 5.34710586e-01
-5.65021455e-01 -5.84481061e-01 9.49152373e-03 7.48917937e-01
3.65417689e-01 5.66105783e-01 1.36393160e-01 -1.81674510e-01
-3.08865428e-01 3.56850743e-01 1.03378177e+00 -6.77805722e-01
-3.78664136e-01 -5.46799123e-01 -2.70387799e-01 -5.50926685e-01
-7.64245629e-01 -3.98227215e-01 -1.15278816e+00 -9.44535077e-01
-5.60934842e-01 9.58627164e-01 2.64824986e-01 1.45653927e+00
-3.91680151e-02 2.63077587e-01 1.11055873e-01 -1.32395709e+00
-8.91257167e-01 -8.97999048e-01 -1.10562909e+00 3.28857839e-01
5.97588718e-01 -8.53588879e-01 -6.68491125e-01 -4.33974504e-01] | [5.448435306549072, 5.068500518798828] |
2a131007-7475-4172-ac0d-d4dba3635cd8 | effects-of-differential-privacy-and-data | 1911.09777 | null | https://arxiv.org/abs/1911.09777v1 | https://arxiv.org/pdf/1911.09777v1.pdf | Effects of Differential Privacy and Data Skewness on Membership Inference Vulnerability | Membership inference attacks seek to infer the membership of individual training instances of a privately trained model. This paper presents a membership privacy analysis and evaluation system, called MPLens, with three unique contributions. First, through MPLens, we demonstrate how membership inference attack methods can be leveraged in adversarial machine learning. Second, through MPLens, we highlight how the vulnerability of pre-trained models under membership inference attack is not uniform across all classes, particularly when the training data itself is skewed. We show that risk from membership inference attacks is routinely increased when models use skewed training data. Finally, we investigate the effectiveness of differential privacy as a mitigation technique against membership inference attacks. We discuss the trade-offs of implementing such a mitigation strategy with respect to the model complexity, the learning task complexity, the dataset complexity and the privacy parameter settings. Our empirical results reveal that (1) minority groups within skewed datasets display increased risk for membership inference and (2) differential privacy presents many challenging trade-offs as a mitigation technique to membership inference risk. | ['Stacey Truex', 'Mehmet Emre Gursoy', 'Lei Yu', 'Ling Liu', 'Wenqi Wei'] | 2019-11-21 | null | null | null | null | ['membership-inference-attack'] | ['computer-vision'] | [ 2.67199814e-01 3.02099347e-01 -1.11516669e-01 -4.93155926e-01
-9.09339428e-01 -1.15240717e+00 5.68831980e-01 9.13169086e-02
-3.21614921e-01 7.03657746e-01 -1.44906104e-01 -5.72725117e-01
-2.29453862e-01 -8.75168085e-01 -1.00605357e+00 -7.37023890e-01
-2.93084800e-01 2.60377735e-01 -3.07825178e-01 1.50835454e-01
1.61650255e-01 6.46334767e-01 -1.23829305e+00 2.75302768e-01
7.72914946e-01 9.90242660e-01 -9.28540766e-01 4.39797103e-01
4.35433656e-01 4.47441846e-01 -9.73886847e-01 -1.07587528e+00
9.42689836e-01 -5.15859239e-02 -5.90560555e-01 -3.26754779e-01
9.87087488e-01 -6.29644573e-01 -4.54312414e-01 1.26020288e+00
4.56192762e-01 3.30926781e-03 8.43830168e-01 -1.89533997e+00
-2.55453259e-01 8.01250577e-01 -2.41617143e-01 5.35967946e-02
6.46733418e-02 3.61842781e-01 7.77096987e-01 -3.70079249e-01
4.52732950e-01 1.15663171e+00 7.05494702e-01 4.87033844e-01
-1.56650591e+00 -1.31717575e+00 -4.42396267e-04 -2.86833793e-01
-1.71281457e+00 -7.00089097e-01 4.72783446e-01 -5.43318629e-01
2.22932830e-01 8.09577465e-01 7.30997510e-03 1.20562255e+00
3.90913337e-02 4.80348349e-01 1.25983167e+00 -9.76138469e-03
4.39681023e-01 5.43563068e-01 2.92397439e-01 3.82892996e-01
5.99376380e-01 3.10073823e-01 -4.30176616e-01 -1.16435671e+00
2.88599521e-01 2.16422100e-02 -2.78774232e-01 -4.47002560e-01
-5.48485219e-01 8.36846888e-01 8.51021856e-02 -3.84030432e-01
8.15887973e-02 2.37040773e-01 4.53824431e-01 6.77818358e-01
4.95382160e-01 5.61533868e-01 -4.86433059e-01 2.84568429e-01
-8.10545564e-01 6.40241444e-01 1.21426880e+00 1.07129288e+00
7.69589007e-01 -1.97173819e-01 -3.10576409e-01 3.26265126e-01
3.94819565e-02 4.05502528e-01 -3.87618355e-02 -1.05338609e+00
6.94078088e-01 1.12442091e-01 4.15052287e-02 -1.02882481e+00
2.04099357e-01 -2.00987563e-01 -6.66146040e-01 4.76051867e-01
6.54512763e-01 -6.73674047e-01 -3.11612934e-01 2.18714738e+00
4.52396363e-01 -3.91373672e-02 2.27789044e-01 2.89621711e-01
2.78546900e-01 1.94394365e-01 2.37184271e-01 -9.90746543e-02
1.27102602e+00 -2.81128228e-01 -4.78259563e-01 1.83112733e-02
5.58863640e-01 -2.59664804e-01 9.22528148e-01 1.73234686e-01
-1.03666914e+00 6.99533299e-02 -1.06393576e+00 1.86416909e-01
-6.77372932e-01 -4.88471329e-01 7.95252979e-01 1.33649576e+00
-8.00629616e-01 6.79832816e-01 -5.22750318e-01 1.67441741e-01
9.99858558e-01 7.04687238e-01 -5.14719605e-01 1.97146490e-01
-1.42404377e+00 4.07914072e-01 2.55850673e-01 -5.34165084e-01
-7.41703629e-01 -1.36380517e+00 -6.54992521e-01 1.68170691e-01
3.44790310e-01 -6.93682969e-01 7.90042639e-01 -6.87329590e-01
-9.24133956e-01 1.15489578e+00 1.76285818e-01 -7.61689901e-01
1.15784252e+00 3.51423863e-03 -3.23840499e-01 -7.28001222e-02
-3.29366297e-01 1.92198783e-01 1.09365356e+00 -1.26238585e+00
-4.98560399e-01 -7.49579549e-01 2.42708176e-01 8.82998854e-02
-5.54938972e-01 1.84135716e-02 1.29294202e-01 -7.79118359e-01
-2.38188118e-01 -9.03685510e-01 -2.73537546e-01 2.46161312e-01
-7.06597447e-01 2.63844073e-01 1.06459653e+00 -3.79200131e-01
1.24047351e+00 -2.58905268e+00 -2.73410261e-01 8.02131236e-01
3.76472116e-01 1.99596062e-01 4.58069742e-01 2.28374749e-01
1.17361106e-01 6.24243617e-01 -3.79760087e-01 -5.35167634e-01
3.40481728e-01 -2.30275001e-02 -8.06138754e-01 6.85646534e-01
-1.06145047e-01 5.86346447e-01 -3.49460125e-01 -2.57208943e-01
-1.98015034e-01 2.29274407e-01 -7.49559700e-01 3.10326040e-01
-2.62248605e-01 2.80888468e-01 -2.48185530e-01 7.24496722e-01
1.22928882e+00 -1.70479212e-02 1.82582244e-01 2.43773386e-02
3.91391277e-01 1.72775552e-01 -1.10889637e+00 8.88794959e-01
-3.13862488e-02 4.06772554e-01 4.49216962e-01 -3.47021669e-01
6.28141224e-01 2.30292574e-01 1.52036026e-02 1.47312701e-01
6.32563233e-02 3.39343622e-02 -1.52089089e-01 -1.70139953e-01
3.52779537e-01 -6.47380622e-03 -2.99590021e-01 8.18222880e-01
-2.95635730e-01 1.67473912e-01 -4.05343026e-01 1.60900116e-01
1.05522370e+00 -5.81829906e-01 8.15493837e-02 -4.39596951e-01
3.11156720e-01 -4.80432957e-01 5.92402577e-01 1.02001023e+00
-4.11758631e-01 3.86987060e-01 8.22781444e-01 -2.00698763e-01
-7.11891651e-01 -1.35114014e+00 -6.12646341e-01 1.12626016e+00
-3.35936225e-03 -3.65232080e-01 -1.01995039e+00 -1.10188639e+00
8.68036866e-01 6.43510461e-01 -7.57071853e-01 -3.60880375e-01
-2.58225560e-01 -8.58569086e-01 1.30999660e+00 1.76206008e-01
4.00104910e-01 -5.72332561e-01 -1.77746981e-01 -5.69506228e-01
1.39758736e-01 -1.00969815e+00 -7.38055646e-01 2.97137089e-02
-8.13553870e-01 -1.07843912e+00 -6.16703294e-02 -3.53023559e-01
9.66139317e-01 -5.80152795e-02 8.87814045e-01 8.43241215e-02
-1.72310442e-01 5.56273997e-01 2.52574801e-01 -7.10021317e-01
-7.42305279e-01 2.05519527e-01 3.46864015e-01 2.40935713e-01
4.57984954e-01 -9.16893721e-01 -5.70233643e-01 2.84618050e-01
-8.98231387e-01 -5.73193014e-01 1.89203456e-01 3.40352684e-01
3.82932067e-01 1.63576335e-01 5.94661057e-01 -1.78728962e+00
7.65977919e-01 -6.84434474e-01 -7.04755425e-01 3.11648279e-01
-7.19617307e-01 -2.77475417e-01 6.81153357e-01 -6.89970374e-01
-7.93828547e-01 -6.75189942e-02 3.36262025e-02 -6.56918168e-01
-1.71068549e-01 2.32815184e-03 -6.39306903e-01 -4.97143954e-01
8.02927017e-01 7.93523341e-02 4.67890173e-01 -3.01755250e-01
3.52054745e-01 6.87285066e-01 5.22358537e-01 -9.85369742e-01
1.00109375e+00 6.64603412e-01 2.34886110e-01 -6.15110159e-01
-4.93055135e-01 1.38894171e-01 -4.34904754e-01 2.17281833e-01
3.52145046e-01 -9.14493680e-01 -1.08421063e+00 7.19898164e-01
-5.02193093e-01 -2.72825748e-01 -3.53786528e-01 -5.03755407e-03
-5.36984324e-01 6.48215652e-01 -5.42497218e-01 -8.74407053e-01
-5.18114269e-01 -1.01949215e+00 5.37945569e-01 -9.67442617e-02
-3.51143181e-01 -1.08449256e+00 -4.41338152e-01 6.29254878e-01
3.84225875e-01 8.11857581e-01 9.55710530e-01 -1.27916169e+00
-6.38756752e-01 -4.94944572e-01 4.03513527e-03 4.23528969e-01
6.15343340e-02 1.90099441e-02 -1.40951777e+00 -6.78612113e-01
1.70208335e-01 -3.43867719e-01 5.40889800e-01 4.35527228e-02
1.77545440e+00 -1.12382746e+00 -1.37660906e-01 1.30854416e+00
1.15404058e+00 -2.27163479e-01 5.99450529e-01 1.47127122e-01
6.16411567e-01 7.37289310e-01 4.68065411e-01 6.99163914e-01
4.80240323e-02 3.38243037e-01 4.30498153e-01 3.89502430e-03
5.59796333e-01 -4.41824079e-01 3.58642370e-01 -1.89953491e-01
5.90892613e-01 1.11835584e-01 -5.62680304e-01 4.81710471e-02
-1.50489855e+00 -1.07997835e+00 2.74519593e-01 2.72201395e+00
1.23571002e+00 4.49529998e-02 2.99076080e-01 -7.34328255e-02
5.94563663e-01 1.47301987e-01 -7.62960494e-01 -4.54356223e-01
-2.53012925e-01 3.34121436e-02 9.69071865e-01 5.59532583e-01
-1.35758781e+00 7.29966044e-01 7.12268591e+00 8.81302476e-01
-7.86860645e-01 -2.63362024e-02 1.06054890e+00 -4.26273137e-01
-3.95610571e-01 2.78511699e-02 -8.09910655e-01 8.37420285e-01
8.18834960e-01 -7.13072956e-01 5.30059576e-01 1.01273572e+00
-4.80693936e-01 3.95584255e-01 -1.61749339e+00 6.90791070e-01
-8.36223289e-02 -1.29291880e+00 -7.52618490e-03 6.65161252e-01
5.90369701e-01 -2.48977095e-01 8.33043635e-01 2.96067744e-01
5.89365900e-01 -1.14269304e+00 4.56107050e-01 2.69605309e-01
9.71651375e-01 -1.21059799e+00 3.20090920e-01 3.53657514e-01
-4.86032188e-01 -2.35049412e-01 -3.04797441e-01 1.79535910e-01
-2.90650159e-01 4.56752062e-01 -8.15861762e-01 1.67866319e-01
5.59066474e-01 2.04204279e-03 -5.87553501e-01 4.59932268e-01
1.22994788e-01 6.16150022e-01 -6.32324517e-01 5.40463984e-01
-1.82098910e-01 -3.25131744e-01 5.95025837e-01 1.15538013e+00
-1.44715771e-01 1.66125502e-02 4.52416949e-03 1.19806933e+00
-5.53219140e-01 -2.09643528e-01 -9.60398018e-01 5.93828857e-02
1.05632246e+00 9.16467130e-01 7.06617683e-02 -5.89170009e-02
7.84272775e-02 7.28662074e-01 3.15167964e-01 4.17529583e-01
-6.45758569e-01 -2.40055799e-01 1.58690774e+00 2.74693310e-01
2.18336061e-02 2.20779851e-01 -5.55383325e-01 -9.92202163e-01
5.59354760e-02 -1.21318805e+00 8.52780640e-01 5.34732938e-02
-1.71761429e+00 -1.37507298e-03 1.59507290e-01 -8.87505054e-01
-1.16845876e-01 -2.79092669e-01 -6.97758973e-01 1.03024626e+00
-1.13206208e+00 -1.05023694e+00 3.05684984e-01 6.53308988e-01
-3.37865800e-01 -4.27620769e-01 9.61069703e-01 1.71022058e-01
-6.82492554e-01 1.79286170e+00 2.81753659e-01 3.90655279e-01
7.60360599e-01 -1.13531590e+00 6.25245214e-01 8.46090555e-01
-1.19711995e-01 1.12534142e+00 6.23517394e-01 -5.17570436e-01
-1.42041838e+00 -1.30628192e+00 5.38172483e-01 -9.67721820e-01
4.28417802e-01 -7.76323080e-01 -1.06251764e+00 1.03892863e+00
-3.57825011e-01 2.63973624e-01 1.39790940e+00 2.48212650e-01
-9.77975368e-01 -2.08999053e-01 -2.12299657e+00 7.61536419e-01
7.59192348e-01 -1.00492156e+00 -9.33070853e-02 3.25641930e-01
1.07885075e+00 -5.15467286e-01 -1.15730596e+00 3.50791872e-01
6.68851197e-01 -9.52787578e-01 1.21522689e+00 -9.63627756e-01
-6.16505891e-02 -3.73370685e-02 -4.72161233e-01 -6.55495465e-01
5.29844426e-02 -1.08610356e+00 -7.82140374e-01 1.46733379e+00
5.37083685e-01 -1.09250832e+00 1.12973738e+00 1.43652439e+00
6.81266427e-01 -5.60635149e-01 -1.11600161e+00 -7.46852815e-01
4.91613507e-01 -4.32105213e-02 1.05889452e+00 1.41663826e+00
-2.31868401e-01 -3.84771556e-01 -3.68256956e-01 5.25771439e-01
1.12417030e+00 -1.11524276e-01 1.10794103e+00 -1.10943580e+00
-5.33014178e-01 -2.50489414e-01 -3.77801478e-01 -3.68384957e-01
4.16396379e-01 -8.45461845e-01 -6.00623429e-01 -2.92617977e-01
4.82530743e-02 -7.93135583e-01 -2.07039729e-01 4.62475657e-01
-2.94984579e-01 -1.18922917e-02 4.35143173e-01 2.43851811e-01
-1.41203821e-01 6.18618876e-02 6.32648230e-01 -6.82032779e-02
-1.21688917e-01 6.19900167e-01 -1.14365339e+00 4.32962269e-01
8.88455033e-01 -5.82742989e-01 -6.11300051e-01 1.03183560e-01
2.80879647e-01 -1.66386113e-01 5.07430613e-01 -8.77960622e-01
3.70622724e-01 -1.74795166e-01 1.72404677e-01 -3.58887583e-01
2.87698507e-01 -1.15197039e+00 3.80874842e-01 3.62169534e-01
-5.89383662e-01 -1.61952868e-01 2.03948885e-01 8.03750575e-01
3.01834315e-01 1.91802867e-02 9.13479090e-01 2.05557011e-02
4.16326374e-02 6.71616793e-01 -1.60160720e-01 5.82402766e-01
1.37493694e+00 -2.74650633e-01 -4.89052951e-01 -2.16617838e-01
-5.73581755e-01 3.88773978e-01 9.28913534e-01 1.33833572e-01
2.55799234e-01 -1.00963771e+00 -5.93428314e-01 5.77533305e-01
1.21720858e-01 9.30907764e-03 1.54590085e-01 4.24847126e-01
-2.98164576e-01 -2.27317423e-01 -8.03033337e-02 -2.24397168e-01
-1.74855423e+00 6.53659165e-01 3.89618486e-01 4.44143303e-02
-4.65430915e-01 1.01466620e+00 3.43178302e-01 -6.90431654e-01
7.38373995e-01 2.14636579e-01 4.44372118e-01 -1.11884261e-02
5.94982624e-01 4.10262883e-01 -1.61199700e-02 -3.24962139e-01
-2.35626161e-01 -1.55387446e-01 -3.96096915e-01 6.34950101e-02
8.33100379e-01 3.15148532e-02 -2.82416791e-01 1.44075602e-01
1.45589066e+00 1.85878932e-01 -1.36975217e+00 -2.23605826e-01
-2.59611249e-01 -9.24510479e-01 -3.37266445e-01 -5.77835202e-01
-1.06690300e+00 5.46200991e-01 4.10646588e-01 2.77590185e-01
8.26477110e-01 -3.10493469e-01 6.33269429e-01 3.47640842e-01
3.65916520e-01 -9.50542271e-01 -5.83045483e-01 7.35458732e-02
4.93058741e-01 -1.15937376e+00 3.18822831e-01 -6.38743758e-01
-6.11590683e-01 4.38695461e-01 5.08701622e-01 4.93524782e-02
1.10421062e+00 4.49598253e-01 3.26248892e-02 -9.25061014e-03
-7.75728464e-01 6.81551516e-01 -8.79944041e-02 8.06016743e-01
-1.50889158e-01 2.97502369e-01 1.66434586e-01 8.70704174e-01
-5.74908912e-01 -6.26857877e-01 3.71250719e-01 8.23608816e-01
7.59609938e-02 -1.11078513e+00 -5.44573963e-01 6.83640897e-01
-1.08638096e+00 3.14004347e-02 -7.39651561e-01 5.88795245e-01
2.26150919e-02 5.82086742e-01 2.92854160e-02 -3.95230740e-01
1.21217191e-01 1.63502455e-01 8.96521807e-02 -4.73178923e-01
-1.22611642e+00 -8.26443315e-01 -3.73393856e-02 -4.39917952e-01
3.97175342e-01 -8.04091752e-01 -8.14349651e-01 -1.03879440e+00
-1.43649742e-01 5.20692840e-02 4.51497287e-01 5.11787891e-01
7.27047145e-01 -1.62158329e-02 1.07077563e+00 -2.55242139e-01
-1.52514935e+00 -3.76540780e-01 -8.82729411e-01 6.68100953e-01
4.88750428e-01 -2.15272233e-01 -9.55083430e-01 -2.64246464e-01] | [5.937662601470947, 7.08242130279541] |
b427c1b5-d185-4fda-8969-bd03338c3bec | unsupervised-artifact-detection-for-whole | null | null | https://openreview.net/forum?id=j9JTX5IwPC | https://openreview.net/pdf?id=j9JTX5IwPC | Unsupervised Artifact Detection for Whole Slide Images of Prostate Biopsies | High-quality image digitisation of histological slides is essential for digital pathology to facilitate diagnosis and to develop reliable computer-aided assistance systems. Currently, image quality control (QC) to identify artefacts that result from slide preparation, staining, or scanning is mainly conducted manually, which is time-consuming and prone to errors. The few existing automated approaches are limited to specific artefacts, require manual parameter tuning, or depend on large amounts of annotated data for training.
Here, we propose an unsupervised and training-free method for artefact detection of digitised histological slides based on one-class classification. By assuming that deep representations from normal (in-distribution) tissue samples are dense and compact in feature space, artefacts can be identified by their distance to the mean in-distribution representation. The use of representations from convolutional neural networks pre-trained on natural images (ImageNet) avoids the need for the collection of annotated data and the costs of training.
We demonstrate the efficacy of our approach for the detection of various types of artefacts (out-of-focus, pen marks, tissue folds, etc.) in an open dataset of prostate biopsy whole-slide images. We compare the proposed distance-based artefact detection with conventional task-specific artefact-detection methods as well as with approaches based on uncertainty estimation. Our method achieved a mean AUROC value of 0.82 on our test dataset consisting of 78 prostate biopsy whole-slide images annotated with five varieties of artefacts. Our proposed method can be utilised for slide-level QC as well as the identification of artefacts within a slide. It is a simple, efficient, yet robust method for automated QC of digitized histological slides. | ['Arto Järvinen', 'Andrew Janowczyk', 'Kristian Eurén', 'Yijiang Chen', 'Nadieh Khalili', 'Walter de Back', 'Amit Suveer'] | 2021-09-29 | null | null | null | iclr-2022 | ['one-class-classification'] | ['miscellaneous'] | [ 6.80342436e-01 1.83418527e-01 4.99972522e-01 -3.27702463e-01
-1.47304666e+00 -7.04043150e-01 6.31186485e-01 8.33216608e-01
-5.98754227e-01 5.03812850e-01 -2.63502330e-01 -2.63412327e-01
-3.22239786e-01 -5.92706025e-01 -5.01209378e-01 -1.05887759e+00
-1.21353425e-01 4.30976331e-01 2.21566617e-01 2.93264091e-01
3.47940356e-01 8.46652925e-01 -1.20423067e+00 1.30521998e-01
6.46639585e-01 9.77017999e-01 2.47229144e-01 1.05656445e+00
-8.03561658e-02 3.37404817e-01 -7.98273146e-01 -3.93979847e-01
-9.15216953e-02 -4.28475857e-01 -7.38862693e-01 2.55276412e-01
4.46163803e-01 -1.09495513e-01 1.10994399e-01 1.15273094e+00
6.94348514e-01 -3.53440613e-01 1.04399145e+00 -6.47478819e-01
-3.07834864e-01 1.02440462e-01 -3.88154894e-01 5.23710489e-01
-3.17431167e-02 3.52655768e-01 3.39044303e-01 -5.76712608e-01
7.91855514e-01 7.07455575e-01 9.74620104e-01 3.89199823e-01
-1.43117249e+00 -3.65873724e-01 -8.91181231e-01 -4.33823727e-02
-1.26395082e+00 -5.22232890e-01 4.87764001e-01 -6.44758523e-01
6.58057749e-01 5.43184638e-01 5.15161514e-01 7.13120878e-01
6.30261481e-01 4.28879052e-01 1.11169851e+00 -5.37746668e-01
5.64098954e-01 1.88268751e-01 -1.81550130e-01 6.46488845e-01
3.91080201e-01 -7.09824935e-02 2.55087644e-01 -1.45231783e-01
9.05632377e-01 -1.12630976e-02 -2.25092366e-01 -4.33964163e-01
-9.36025977e-01 6.42325759e-01 5.64528286e-01 7.37809122e-01
-1.12259947e-01 1.79063171e-01 6.32252753e-01 -8.02406743e-02
3.44273865e-01 6.33676529e-01 1.53830215e-01 2.95866560e-02
-1.29463661e+00 -2.10014641e-01 4.42385048e-01 4.41007107e-01
5.37407517e-01 -3.09889346e-01 -2.32886568e-01 6.98219657e-01
2.64885724e-01 3.08436781e-01 9.69768107e-01 -4.01374757e-01
-9.03275982e-02 9.11857486e-01 7.27282837e-02 -9.98836875e-01
-7.52968788e-01 -4.31402147e-01 -1.20680714e+00 3.44374925e-01
6.48214042e-01 1.97033420e-01 -1.36348152e+00 8.74853492e-01
2.53840923e-01 -1.73042953e-01 -7.25281015e-02 7.57677376e-01
5.29366314e-01 8.46768767e-02 -2.87078992e-02 -1.74776703e-01
1.42154872e+00 -3.57985586e-01 -6.85288072e-01 1.39750928e-01
9.09820676e-01 -6.82024360e-01 7.95509696e-01 4.04444873e-01
-8.37083399e-01 -2.65847176e-01 -1.27930176e+00 -4.37520370e-02
-5.30277431e-01 4.78526324e-01 5.00365019e-01 9.60963368e-01
-1.09868395e+00 7.98798144e-01 -1.15875137e+00 -4.55815732e-01
8.80326092e-01 6.02257907e-01 -7.90944934e-01 -7.44753629e-02
-5.37596285e-01 8.69652390e-01 1.99582413e-01 4.39979315e-01
-7.88720131e-01 -5.84677815e-01 -9.00300741e-01 -1.17648862e-01
-1.75920740e-01 -2.15259016e-01 9.68038738e-01 -1.21641266e+00
-1.36036015e+00 1.15535045e+00 1.83679089e-01 -4.68104035e-01
9.08607423e-01 1.87325418e-01 -2.43290141e-01 2.94959724e-01
-1.23534881e-01 3.35961133e-01 6.63942754e-01 -1.06642354e+00
-2.79183000e-01 -4.30795342e-01 -4.51014996e-01 -2.19823465e-01
-7.43093938e-02 -1.44914776e-01 -3.35975230e-01 -3.79868567e-01
2.33198032e-02 -7.37452149e-01 -4.00720090e-01 2.95666605e-01
-3.99850041e-01 3.13810781e-02 4.79778677e-01 -7.31027365e-01
9.60386932e-01 -2.31867433e+00 -3.22303265e-01 4.94600952e-01
7.67134279e-02 4.62878317e-01 1.18344888e-01 6.80767447e-02
-2.51208041e-02 3.47120017e-01 -5.53872466e-01 -2.20464110e-01
-6.80043474e-02 -2.26377323e-02 4.00531143e-01 9.70483303e-01
6.46022677e-01 6.91405892e-01 -9.88632441e-01 -7.77097046e-01
4.75788265e-01 6.30653203e-01 5.15612252e-02 2.51377672e-01
2.10472196e-01 2.97317863e-01 -1.20935710e-02 6.88140988e-01
7.25755453e-01 -2.48095915e-01 2.61683315e-01 -3.33784997e-01
9.99817476e-02 -1.07837975e-01 -1.03311694e+00 1.56376112e+00
-5.35091281e-01 7.96816349e-01 -8.00629109e-02 -5.90622008e-01
7.82318771e-01 2.90894687e-01 8.51405039e-02 -2.91230053e-01
3.42336386e-01 5.14768600e-01 1.50551915e-01 -6.49779737e-01
1.49798885e-01 -3.86971354e-01 1.01364739e-01 1.97875381e-01
3.89447212e-01 -3.79135281e-01 2.73659915e-01 -1.64022252e-01
1.48935473e+00 -2.67306417e-01 6.83006942e-01 -4.40444112e-01
5.41839898e-01 -1.33232981e-01 5.08063138e-02 4.63560879e-01
-4.33130324e-01 8.15388322e-01 4.49713916e-01 -5.17184794e-01
-1.02589250e+00 -8.06648731e-01 -5.98373711e-01 1.90132767e-01
-1.73615828e-01 9.45444927e-02 -7.30009258e-01 -8.77675712e-01
-1.09428905e-01 2.35892996e-01 -9.98081148e-01 1.13992207e-01
-2.95563668e-01 -8.99592996e-01 5.77772558e-01 3.22325259e-01
1.63420439e-01 -1.10249722e+00 -7.44305611e-01 2.86058486e-01
1.75662637e-01 -5.03306270e-01 -1.26453429e-01 5.23161948e-01
-9.53070521e-01 -1.32048333e+00 -8.43811929e-01 -6.34546697e-01
1.31032383e+00 -1.95682928e-01 8.89391363e-01 2.85137862e-01
-1.08806849e+00 2.14540455e-02 5.02678603e-02 -2.30241373e-01
-6.29795015e-01 -1.89050198e-01 -4.26462471e-01 -7.35102743e-02
3.21508974e-01 -2.02141598e-01 -8.01338136e-01 2.30915919e-01
-1.47543275e+00 -4.78341550e-01 1.12123549e+00 1.01191795e+00
9.04264927e-01 1.20730020e-01 3.24458957e-01 -1.09723210e+00
4.68557596e-01 -2.30255052e-01 -6.40386403e-01 8.00176263e-02
-2.03923926e-01 1.24711312e-01 4.65755552e-01 -2.39518553e-01
-7.69751966e-01 3.83101583e-01 -1.18896373e-01 -2.14724049e-01
-5.06578445e-01 4.29710329e-01 -1.34140523e-02 -4.96771872e-01
9.77651834e-01 8.01996812e-02 1.33904621e-01 -1.58219874e-01
-1.27127275e-01 7.37009048e-01 6.24504209e-01 9.78877991e-02
5.39151609e-01 6.35253072e-01 2.87612587e-01 -7.73935258e-01
-2.09531233e-01 -8.74661386e-01 -6.67967200e-01 -1.42808944e-01
6.09605134e-01 -5.24058580e-01 -3.66679847e-01 4.09229159e-01
-7.78653562e-01 -3.67366731e-01 -1.56332031e-01 2.94614613e-01
-4.44820017e-01 5.99678874e-01 -5.34666717e-01 -8.43963623e-01
-5.21594882e-01 -1.10940456e+00 1.25456691e+00 3.41146618e-01
-3.13932359e-01 -1.02475345e+00 2.16354683e-01 1.19087331e-01
4.42639619e-01 5.93931079e-01 7.14635313e-01 -6.02585077e-01
-2.85996974e-01 -9.08539474e-01 -2.72603184e-01 3.34811717e-01
4.98611361e-01 1.75582722e-01 -1.06792355e+00 -3.62092763e-01
-3.11763566e-02 -3.91424060e-01 8.27539265e-01 3.81365240e-01
1.29051387e+00 -2.24658772e-01 -4.73468482e-01 4.54645783e-01
1.93739629e+00 -3.08229420e-02 9.18013752e-01 4.64314342e-01
1.85715660e-01 4.43968713e-01 3.94880146e-01 1.06279559e-01
-3.60469103e-01 3.93980116e-01 5.52869141e-01 -2.07831055e-01
-5.56354336e-02 3.33317667e-01 -2.47829724e-02 1.14034109e-01
-1.38150543e-01 -1.10565424e-01 -1.03978431e+00 1.03100383e+00
-1.18085003e+00 -7.05983043e-01 -2.51629442e-01 2.24597287e+00
1.01072288e+00 2.60452598e-01 -6.78016022e-02 4.76803899e-01
7.51509726e-01 -4.54100877e-01 -2.90325612e-01 -3.85309011e-01
2.77051866e-01 4.05969113e-01 7.12279618e-01 2.71868706e-01
-1.42320824e+00 2.35751793e-01 5.67323303e+00 8.69297147e-01
-1.29990876e+00 -7.90219083e-02 8.57584774e-01 4.99779955e-02
1.39990017e-01 -6.17585659e-01 -4.46185768e-01 5.13416231e-01
1.01004839e+00 2.88929433e-01 -1.00419983e-01 6.71451330e-01
1.13730542e-01 -4.51111555e-01 -1.12543380e+00 9.02883232e-01
1.04032911e-01 -1.37331307e+00 -3.97137254e-01 1.00506611e-01
5.67320883e-01 -1.76308334e-01 -1.96295716e-02 -1.30240545e-01
3.67564596e-02 -1.24464083e+00 1.86647832e-01 4.88038033e-01
1.13910472e+00 -7.13558435e-01 1.51988232e+00 -1.13803886e-01
-6.51929140e-01 2.57256657e-01 -4.09716845e-01 4.58965003e-01
-3.70796740e-01 8.40850532e-01 -1.42710125e+00 5.04989445e-01
4.82403368e-01 2.29995385e-01 -9.02324080e-01 1.50791955e+00
1.04130507e-01 5.01404405e-01 -3.63165051e-01 -1.18895724e-01
1.92205861e-01 4.38325882e-01 1.07443184e-01 1.73709512e+00
5.09339035e-01 -4.04415399e-01 -4.43049639e-01 6.38939917e-01
8.71616453e-02 1.44859150e-01 -3.77859324e-01 -8.97194222e-02
5.49783669e-02 1.77052784e+00 -1.44210100e+00 -1.06087640e-01
8.12707543e-02 9.20905352e-01 1.11906864e-01 -1.88960388e-01
-3.31076562e-01 -6.59387112e-01 -3.67080905e-02 1.04609922e-01
3.97722304e-01 2.28740588e-01 -1.82385474e-01 -5.04075408e-01
-4.47115749e-02 -6.67620420e-01 2.72089064e-01 -3.85745287e-01
-1.39990532e+00 6.91597700e-01 -4.95769531e-01 -1.25293410e+00
-3.18070471e-01 -9.11946297e-01 -6.27966404e-01 6.85915053e-01
-1.39866948e+00 -1.01268566e+00 -3.97168428e-01 1.88725799e-01
3.42492759e-01 3.50532681e-01 1.27942336e+00 1.23675242e-01
-2.47206256e-01 5.04198968e-01 4.55797613e-01 2.89606124e-01
7.12857306e-01 -1.69300520e+00 -1.78368986e-01 6.22735023e-01
-2.04897434e-01 3.99265856e-01 7.18357444e-01 -3.83293778e-01
-1.10847199e+00 -1.23443615e+00 6.57289326e-01 -3.95804346e-01
5.89931011e-01 -1.72023356e-01 -7.69494057e-01 3.16546142e-01
4.03931402e-02 5.72072327e-01 1.11026907e+00 -2.48935670e-01
-2.29611304e-02 1.04682408e-02 -1.63694429e+00 1.38878003e-01
2.00990871e-01 -2.97312200e-01 -3.75157565e-01 3.25593740e-01
-2.96645790e-01 -3.80606413e-01 -9.15248752e-01 1.67145550e-01
6.11333132e-01 -1.02698314e+00 5.84818363e-01 -4.25454564e-02
3.94730121e-01 -3.79876614e-01 3.29566240e-01 -1.26746786e+00
-2.66582310e-01 -2.60691524e-01 4.02658939e-01 9.99252081e-01
1.95980057e-01 -1.57472208e-01 7.77379274e-01 3.24122429e-01
-2.10546061e-01 -5.64791501e-01 -1.25434828e+00 -5.09310305e-01
-9.55778584e-02 -1.16551891e-01 2.16468215e-01 6.99495196e-01
1.45494059e-01 -3.73383015e-01 3.89259577e-01 2.91354597e-01
5.66296935e-01 -2.31395617e-01 5.18565536e-01 -1.15848112e+00
-9.72797871e-02 -5.62493324e-01 -1.21353245e+00 1.73690766e-01
-4.62647945e-01 -8.68492007e-01 4.49427307e-01 -1.52126372e+00
1.94456384e-01 -3.95731091e-01 -5.09530544e-01 2.66878098e-01
-1.41329020e-01 7.65047431e-01 -2.60707617e-01 2.91741222e-01
-4.98799711e-01 -1.37293652e-01 1.07950747e+00 -4.03992057e-01
6.64098114e-02 -9.49913114e-02 -3.78533572e-01 7.22051919e-01
7.07565069e-01 -5.59024572e-01 -2.86860969e-02 7.50842690e-02
2.01792344e-01 -2.53344804e-01 5.61821461e-01 -1.31860030e+00
-1.26231564e-02 2.49828696e-01 8.64365578e-01 -4.45177734e-01
-7.37466514e-02 -9.48869467e-01 3.11831474e-01 7.56026208e-01
-3.45058113e-01 -4.45034474e-01 4.14606839e-01 7.74016500e-01
-1.56491116e-01 -5.72554946e-01 1.04857993e+00 -4.42795247e-01
-2.19267294e-01 -2.76866525e-01 -5.58658063e-01 -6.08181655e-01
1.00717282e+00 -4.01981354e-01 -2.60467589e-01 8.20033625e-02
-6.95706248e-01 -1.88087776e-01 5.50215900e-01 -1.93098411e-01
6.23034716e-01 -1.08777606e+00 -5.41024327e-01 1.31869435e-01
3.64308417e-01 4.33768421e-01 5.87346017e-01 9.93287206e-01
-1.09751213e+00 4.09565121e-01 -3.42344552e-01 -8.54880989e-01
-1.45821440e+00 3.74708205e-01 5.17819107e-01 -8.06839049e-01
-3.84224474e-01 9.77826357e-01 -7.52516240e-02 -1.56754971e-01
6.80061653e-02 -6.79959953e-01 -3.73354495e-01 -6.00383468e-02
6.91086769e-01 8.67785215e-02 6.22441351e-01 -2.87216157e-01
-3.90985757e-01 2.89720505e-01 -2.75820792e-01 2.47767359e-01
1.31606710e+00 2.10889339e-01 -5.88986874e-02 4.63960558e-01
1.17869186e+00 -1.33664623e-01 -1.10120308e+00 3.03938419e-01
2.75577933e-01 -4.42593515e-01 3.48076403e-01 -1.01753390e+00
-9.02982891e-01 8.97368371e-01 1.27725315e+00 3.79116058e-01
1.10945249e+00 -1.09929748e-01 1.21206827e-01 9.90524516e-02
2.05502361e-01 -9.37662065e-01 -1.30935207e-01 -2.88414448e-01
8.59296262e-01 -1.42339265e+00 1.59453779e-01 -2.54208297e-01
-2.08163083e-01 1.48071110e+00 2.05458775e-01 -3.14824939e-01
3.91677082e-01 3.16885978e-01 2.51369774e-01 -3.67792487e-01
-3.55062902e-01 -4.66097984e-03 2.53360271e-01 8.18208873e-01
6.76159024e-01 9.32006463e-02 -3.24182183e-01 4.53214288e-01
1.66543692e-01 3.27559114e-01 7.16993332e-01 1.04581916e+00
-2.70254463e-01 -7.33818412e-01 -3.95843327e-01 7.53596544e-01
-7.70113885e-01 1.91453710e-01 -3.84782970e-01 9.35011029e-01
2.04508200e-01 6.02464557e-01 2.14613736e-01 1.57517090e-01
2.21888777e-02 2.79802009e-02 6.47299170e-01 -5.49072623e-01
-6.84081137e-01 2.53260314e-01 -7.15158135e-02 -3.18268031e-01
-5.84515929e-01 -6.15024149e-01 -1.06345046e+00 1.84379011e-01
-7.47032046e-01 -6.88225031e-02 9.87524688e-01 8.46165001e-01
2.04682350e-01 7.84216106e-01 3.42757851e-01 -1.01126599e+00
-3.67363781e-01 -1.30025268e+00 -7.06636548e-01 4.21613902e-01
5.42346358e-01 -5.93851507e-01 -6.55307353e-01 3.74259710e-01] | [15.099752426147461, -2.9673826694488525] |
608ce8e3-8537-417f-a798-d954e7970413 | towards-controllable-agent-in-moba-games-with | 2112.08093 | null | https://arxiv.org/abs/2112.08093v1 | https://arxiv.org/pdf/2112.08093v1.pdf | Towards Controllable Agent in MOBA Games with Generative Modeling | We propose novel methods to develop action controllable agent that behaves like a human and has the ability to align with human players in Multiplayer Online Battle Arena (MOBA) games. By modeling the control problem as an action generation process, we devise a deep latent alignment neural network model for training agent, and a corresponding sampling algorithm for controlling an agent's action. Particularly, we propose deterministic and stochastic attention implementations of the core latent alignment model. Both simulated and online experiments in the game Honor of Kings demonstrate the efficacy of the proposed methods. | ['Shubao Zhang'] | 2021-12-15 | null | null | null | null | ['action-generation'] | ['computer-vision'] | [-0.02820425 0.3135816 -0.03295505 -0.07626566 -0.41386008 -0.46079046
0.79489285 -0.6466322 -0.79352397 0.8979067 0.44944784 -0.00909883
-0.11242223 -0.88298213 -0.43031833 -0.84936345 -0.1502408 1.1864558
-0.1241602 -0.48642048 0.13734731 0.07001867 -0.9812556 -0.07907677
0.71911126 0.6162148 0.19428322 0.9964933 0.5058263 1.576711
-0.907153 -0.42456707 0.6119703 -0.6134235 -0.98083276 0.09686388
-0.26606032 -0.84988195 -0.52773786 1.0407788 0.65525854 0.42986435
0.6719115 -1.4010963 -0.41998142 0.8923175 -0.28182444 0.18095642
0.34207338 0.70662093 1.0750357 0.1065051 0.5181818 1.2244316
0.2285575 0.95082575 -0.8513876 -0.6727227 0.2003892 0.17458828
-0.80741304 -0.25851476 0.56584895 -0.25941563 0.89616174 0.08757758
1.1946588 1.5140015 0.46882093 1.2104002 0.94718164 -0.127857
0.5437835 -0.72662383 -0.45329982 0.4639185 -0.07806167 0.1693507
-0.31041962 -0.35354146 1.3202146 -0.19619112 0.21361722 -0.4615592
-1.1423001 0.9983008 0.10637567 -0.23855191 -1.1797403 0.6423641
0.25072744 0.31763884 0.21385062 1.0189561 -0.0887738 -0.67401403
-0.47558284 0.7254828 0.81235224 0.77957827 0.02344598 0.5310454
-0.5975727 0.38061395 0.0502746 0.37864152 0.7507016 -1.5894332
0.18379067 0.29904103 0.66539854 -1.0696537 -0.297959 -0.28708437
-0.7784818 0.6196567 0.18363775 -0.5025665 -0.6533001 2.1076553
0.23358275 0.45948407 0.39891124 1.0436478 0.14220068 0.56759906
0.06682032 -0.45522562 1.0399812 -1.3837084 -0.8229666 -0.5677296
0.03486145 -0.25771224 0.92665553 0.2572352 -1.7361844 -0.340087
-1.0066295 0.37504807 0.35408846 -0.33865103 1.0661689 0.08586752
-0.8983675 0.6308848 -1.11238 0.05394556 0.29287967 0.63375825
-0.28299248 0.6055989 -1.2836754 0.85544705 0.3790211 0.11837789
-1.768743 -0.06795393 -0.80730665 0.18400718 0.6531617 -1.1200265
1.8278072 -1.141702 -2.5303411 0.74835867 0.46506375 -0.97747785
0.8434615 -0.52485627 0.28434715 0.01496888 0.31656078 0.6285951
0.8580035 -0.6524855 -0.65006095 0.05720571 0.78599143 0.8536012
-0.13648278 -0.03500223 -0.03868606 -0.57030004 -0.40713274 -0.88304806
-0.85404557 -0.6592642 -0.27436444 -0.1660117 0.07266755 -0.31472793
0.933162 -1.7624942 0.83683115 -0.17713757 0.6337064 0.17494282
-0.23442477 0.40555918 0.13734096 -0.32879713 0.2990278 -0.21981774
0.41436493 0.31208465 -0.4226581 0.40983328 -0.4150924 1.2191608
-1.0057112 -0.1510383 -0.14153841 -0.26611468 -0.8611897 0.85245866
-0.49316132 0.6155275 -0.8443899 0.09331606 0.11501656 -0.3091026
0.46671924 0.49850875 -0.08580753 0.28673044 -0.8562831 1.4942665
-0.17550212 0.21046571 0.35150623 -0.69513714 0.55879486 0.4388003
0.54934573 -0.65744597 0.66152585 -0.22708951 0.4495286 -0.26158682
0.48432475 -0.0789507 -0.6040542 0.9005588 -0.05668541 -0.1462269
0.13762401 0.28079107 1.2877648 0.19629896 0.6591082 -0.03673334
0.1766906 0.14581433 0.67726254 1.4293933 -0.5207905 -0.07636642
0.7105597 -0.858089 -1.151761 -0.86251175 0.8959487 1.3698434
0.19668981 -0.335309 -0.9551453 -0.24989927 -0.28062305 0.7218453
-0.95560575 -0.45659375 -0.9501446 -0.4580267 0.46246243 0.62044877
0.9322085 -1.5497928 -1.0856783 0.4617982 -0.3408028 -0.8883594
-0.71234393 -0.04933381 -0.31078187 -0.9294658 -0.2771723 -0.67595565
0.14273843 -0.07930599 1.1419362 -0.24708584 0.00859684 0.44487536
-0.07119969 -0.26604435 -0.3983754 0.16225825 0.5205955 -0.19852293
0.2223077 -0.71800965 -0.76362693 0.22292434 -0.5744892 0.70038337
0.556554 1.0877931 0.06199588 -0.22193053 0.24483748 -0.47737527
1.4548872 -0.29816526 -0.92115015 -0.01460868 -0.12102641 -0.02550263
0.5727335 -0.88313544 -0.77751845 -0.03706803 0.18787576 -0.36354947
0.1061088 0.29511818 -0.20485891 0.414093 0.42978957 0.36896452
0.34204412 -0.04197422 0.2863033 0.4968425 0.66651326 -0.7686453
0.6697326 0.3996958 -0.29513124 -0.12802875 -0.6611507 0.1210448
-0.35865173 -0.41036338 1.2445558 -0.86010665 -1.7068185 0.841247
-1.034793 -0.87957996 -0.48871985 0.47057873 -1.3808703 -0.07742545
-0.95378274 -0.76479346 -0.38115448 -1.4052753 0.80240583 0.44811115
-0.4153793 -0.70387226 0.61037284 0.40166456 0.5926162 0.16791227
0.47957134 -0.73205 -0.597011 -0.24163602 0.4855344 -0.17680526
0.065455 -0.4388251 -0.29329976 -0.39925724 0.39971167 -0.7096652
0.17838563 0.6732658 0.6369172 -0.6689261 -0.085462 0.6031453
0.6430652 0.46745846 0.563683 0.67831093 0.39167666 0.19244535
0.6478926 0.8819323 0.2963681 0.8832556 0.763903 0.2513671
0.5690511 -0.3767123 0.5962449 0.87217206 -0.5335757 -0.21152751
-0.69418436 0.26690945 -2.3753076 -1.4520614 0.5796377 1.6623671
1.006738 0.36755118 0.34748736 -0.5004145 0.692736 0.26277354
-0.7251009 -0.32081717 0.03116422 0.21440633 0.35284486 0.81913984
-1.2610747 1.626741 7.3985105 0.8694442 -0.6855959 0.07011186
0.63499707 -0.44980243 0.15240522 -0.07215118 -0.44124416 0.27492473
0.6759237 -0.679693 0.8199504 1.093383 0.45457977 0.10978037
-0.9354512 0.76539654 -0.13337108 -1.5076425 0.04031949 0.18763788
0.76470184 -0.3270545 0.2709036 0.7227083 1.5087324 -1.4007248
0.92001253 0.40558472 0.5072475 -0.64037865 0.7166268 0.68546957
-0.72862315 -0.3504765 -0.33349422 -0.941898 0.1857062 -0.76762116
-0.71970737 -0.14386767 0.20756528 0.2464191 0.01330742 0.7039872
-0.6054644 0.45231444 -0.07789961 -0.25519565 0.7482579 -0.5231333
0.7188022 0.33704108 -0.23824328 0.5824966 0.54462665 0.6801835
-0.01965835 -0.12526175 -0.40776482 -0.2725472 0.31799805 1.010484
-0.57428133 -0.36794502 0.20293161 1.3024839 0.7426067 0.1877794
-1.168923 0.05938819 1.0706985 -0.26042968 0.05835767 -0.23574895
0.08838934 -1.1370733 -0.42367277 -1.379582 0.29408884 -0.8909854
-1.0851791 0.7506896 -0.01629279 -1.0527602 -0.7909652 -0.3173434
-0.94548666 0.46546653 -0.495397 -0.87106586 -0.18123256 0.56586
0.6797126 -0.66526866 0.91837966 -0.15833527 -0.6268817 0.31546995
-0.05191831 0.39068118 0.2456862 -1.3364378 0.64493054 0.6253354
-0.14245565 0.5655324 0.93782896 -0.63833183 -1.1959155 -0.5614554
0.01286372 -0.26466033 0.9069407 -0.43672884 -0.12643388 1.090449
0.7236863 -0.6916716 0.60236126 -0.12801473 0.3386073 0.33530873
-0.68182975 1.3559406 1.1170408 -0.23542665 -0.8405533 0.49234688
0.76148146 -0.75082606 -0.24217987 0.09087432 0.49783716 -0.6811616
0.8886423 -1.4843835 0.6105429 0.03873595 0.10818055 -1.52358
-0.8243752 -1.3024224 -0.01186577 0.35973546 -0.16374798 -0.3863014
1.176258 0.75465816 0.05267001 -0.52794534 -0.99966854 -0.3990908
0.22436345 -0.04716312 0.5378067 0.83012074 0.16901034 0.4964508
-1.103553 0.15289378 0.45251536 0.15346111 1.22607 -0.67822844
-0.9917634 -0.61564374 -0.36335364 -1.5248798 0.41122842 -0.44523683
0.12731409 -1.347472 0.3309413 0.06990264 -0.06808028 0.35651675
-0.29363683 -0.07678927 0.29539594 0.28924203 -1.1799924 0.9060416
1.3693345 -0.1544997 -0.15543878 0.22061566 -0.63614583 0.9545064
0.85333884 -0.33647275 -0.44487128 -0.39660153 0.45305157 0.6353871
0.10672925 -0.93879503 0.5360682 -0.8656903 -0.02589859 0.01561653
0.5101548 -0.39650163 0.06573939 0.99252677 -0.6313987 0.49119163
-0.12334083 0.45232272 -0.10441212 -0.04422165 0.6718232 -0.30727726
-0.39211294 0.47103202 -1.0655727 0.19174777 1.3467917 0.09021264
-0.09377119 -1.1549485 -0.97461504 0.7244565 0.23902917 0.19647561
0.40734938 -1.3576058 -0.93657845 -0.04657491 -0.49399984 -0.22633815
0.16243647 0.43144152 -0.9605258 0.07383147 -0.8813323 -0.04795558
-1.093148 0.4002381 0.68971676 -0.83866656 -0.5290583 0.9494193
0.314381 -0.45551038 0.33028078 0.29064995 -0.46108478 -0.3863085
0.5859598 0.1525854 -0.9217558 -0.32934663 0.03667393 -0.22619045
-0.28314826 -0.5423722 1.5628643 0.18082452 -0.1296128 0.17267036
0.19448361 -0.02968655 -1.5614638 0.02159757 -0.34861267 -0.52996075
-0.45112947 -0.45249012 -0.7180052 0.36694193 0.21083009 0.13518701
0.8267271 -0.35430792 0.6312582 0.5087742 0.57718706 -1.1902062
0.6530288 0.8614498 0.7950785 -0.93947136 -0.17372747 0.3199008
-1.1213411 0.8368085 1.0701722 -0.72852135 0.1874598 0.48562452
0.22572419 -0.36569485 -1.2920198 -0.04242763 -0.44165176 0.50899583
-0.23481008 0.2341565 -0.28797692 0.72102726 -0.7143749 -0.13194199
0.80372274 0.9378402 -0.43110222 -1.0878774 -0.0645823 0.12049925
-0.41643104 0.0311597 -0.4731446 0.63106614 -0.25592175 0.6103752
0.29310387 -0.37052563 0.12213124 -0.17482504 0.47311696 -0.5796263
-0.74535376 0.05157196 -0.0731006 -0.7984329 -0.2798387 -0.62335575
-0.91867006 -0.54881 0.0273105 0.48578298 -0.06234677 0.99642736
0.09646747 0.70308584 0.60409635 -0.97920996 -1.0327632 -1.2645813
-0.41115192 0.49504253 -0.05104862 -0.70769507 -0.08677201 -0.39850992] | [3.7143454551696777, 1.5431870222091675] |
426549e2-69c4-4fb1-b781-4f33979b8c79 | always-strengthen-your-strengths-a-drift | 2304.09062 | null | https://arxiv.org/abs/2304.09062v1 | https://arxiv.org/pdf/2304.09062v1.pdf | Always Strengthen Your Strengths: A Drift-Aware Incremental Learning Framework for CTR Prediction | Click-through rate (CTR) prediction is of great importance in recommendation systems and online advertising platforms. When served in industrial scenarios, the user-generated data observed by the CTR model typically arrives as a stream. Streaming data has the characteristic that the underlying distribution drifts over time and may recur. This can lead to catastrophic forgetting if the model simply adapts to new data distribution all the time. Also, it's inefficient to relearn distribution that has been occurred. Due to memory constraints and diversity of data distributions in large-scale industrial applications, conventional strategies for catastrophic forgetting such as replay, parameter isolation, and knowledge distillation are difficult to be deployed. In this work, we design a novel drift-aware incremental learning framework based on ensemble learning to address catastrophic forgetting in CTR prediction. With explicit error-based drift detection on streaming data, the framework further strengthens well-adapted ensembles and freezes ensembles that do not match the input distribution avoiding catastrophic interference. Both evaluations on offline experiments and A/B test shows that our method outperforms all baselines considered. | ['Jingping Shao', 'Jinghe Hu', 'Zhangang Lin', 'Xiwei Zhao', 'Fei Teng', 'Congcong Liu'] | 2023-04-17 | null | null | null | null | ['click-through-rate-prediction'] | ['miscellaneous'] | [ 1.05333827e-01 -5.01962304e-01 -3.40382308e-01 -2.30932117e-01
-3.12096685e-01 -4.34641898e-01 2.26398572e-01 2.78104931e-01
-1.31397784e-01 8.88997853e-01 -2.97602981e-01 -3.04250509e-01
-1.17835984e-01 -6.97503984e-01 -8.36512506e-01 -8.22889090e-01
-2.08463356e-01 5.98753095e-01 5.79225421e-01 -3.67396206e-01
4.51241553e-01 1.27303481e-01 -1.71468449e+00 4.42158580e-01
9.04643118e-01 1.11190915e+00 3.58292431e-01 7.07712233e-01
-4.13676918e-01 7.04063892e-01 -1.04772604e+00 -2.17042208e-01
2.20276877e-01 -2.64293969e-01 -1.18566593e-02 -2.25500494e-01
7.58927241e-02 -3.91820610e-01 -2.33873487e-01 6.84204996e-01
3.27703506e-01 1.34223357e-01 3.95973891e-01 -1.20473993e+00
-4.27563637e-01 7.71889031e-01 -7.01508105e-01 5.66685736e-01
1.95254683e-01 2.58464158e-01 4.61522788e-01 -4.38893884e-01
4.50169593e-01 9.53173876e-01 6.47861123e-01 5.66985488e-01
-1.27391922e+00 -8.79176617e-01 8.06369007e-01 2.73038775e-01
-1.06358099e+00 -1.71076238e-01 7.99573183e-01 -3.22302669e-01
6.80082560e-01 1.19108938e-01 6.08280301e-01 1.57621098e+00
8.88208985e-01 1.00461304e+00 4.95377719e-01 -1.46462455e-01
8.37318003e-01 3.24330598e-01 -4.07127999e-02 -1.03797741e-01
5.56058943e-01 2.37274975e-01 -8.22645426e-01 -3.12712133e-01
5.73829412e-01 4.85345393e-01 1.09111644e-01 -4.45267558e-01
-7.76556730e-01 6.12929583e-01 -8.19554329e-02 9.01013389e-02
-3.72397274e-01 1.21009173e-02 5.48702836e-01 7.34436333e-01
4.67873514e-01 2.23620385e-01 -8.10173750e-01 -2.97168314e-01
-9.68736291e-01 5.53045571e-01 9.75250781e-01 1.11654496e+00
4.89608705e-01 3.62822562e-01 -3.36182535e-01 5.95073223e-01
-1.39720798e-01 5.76899946e-01 9.60151553e-01 -4.79998887e-01
4.32247192e-01 5.01844347e-01 4.63168710e-01 -8.19563150e-01
-9.60301831e-02 -7.60529101e-01 -8.55360270e-01 1.99701622e-01
3.54284674e-01 -3.85726243e-02 -8.45033705e-01 1.60900295e+00
4.01024848e-01 4.79959577e-01 -2.13842720e-01 5.47661364e-01
-2.55261928e-01 3.13074350e-01 -6.46848306e-02 -6.64114654e-01
7.26201057e-01 -7.68424273e-01 -8.54924977e-01 -1.35772098e-02
1.79701284e-01 -6.42704606e-01 1.15110242e+00 1.00351071e+00
-7.65046418e-01 -7.21721768e-01 -1.29041696e+00 8.28939795e-01
-1.28461406e-01 -3.71209383e-01 4.74493802e-01 7.64085293e-01
-3.52715433e-01 1.01181507e+00 -9.33778644e-01 -5.19141369e-02
1.95070356e-01 2.24013403e-01 5.37099481e-01 -1.67847108e-02
-1.10834754e+00 4.90203500e-01 3.21932942e-01 -6.79736361e-02
-1.08685362e+00 -1.03070831e+00 -4.32101376e-02 -2.20521167e-01
6.66402221e-01 -6.66520536e-01 1.66930151e+00 -9.66021240e-01
-1.65412879e+00 -2.68804342e-01 -1.17011189e-01 -8.27882469e-01
9.04964983e-01 -4.79116708e-01 -1.09796035e+00 -4.71252590e-01
-3.45847219e-01 -3.56444836e-01 1.60514832e+00 -1.09791911e+00
-9.71933067e-01 -3.85040015e-01 -4.06416655e-01 -2.16198936e-01
-4.61683452e-01 -6.81481659e-01 7.34306276e-02 -7.64149725e-01
-9.98759717e-02 -7.25614071e-01 -3.20081800e-01 -6.09643817e-01
-1.34674743e-01 3.90979238e-02 1.09896457e+00 -3.33162576e-01
1.96832454e+00 -2.09591746e+00 -2.38475189e-01 1.88918322e-01
-4.91330326e-02 2.90757656e-01 3.85420881e-02 6.78940654e-01
1.18070342e-01 -1.55941129e-01 1.17971137e-01 -1.69026971e-01
-1.25628904e-01 3.20241362e-01 -7.13553846e-01 2.10640561e-02
-9.45030153e-02 2.50459611e-01 -9.93359625e-01 1.02128044e-01
1.07415207e-01 1.68012902e-01 -6.02791548e-01 4.36012112e-02
-7.78702497e-01 5.49543917e-01 -3.98578495e-01 6.54597878e-01
8.24988902e-01 -2.82325715e-01 2.62413949e-01 6.76151961e-02
-1.97387263e-02 2.94428878e-02 -1.38149285e+00 1.53260410e+00
-7.59828448e-01 -1.27755985e-01 -2.79023111e-01 -8.24352622e-01
1.09488153e+00 -6.44408911e-02 5.66207707e-01 -8.36862624e-01
-6.69080466e-02 2.71199405e-01 -4.55232561e-02 -2.76546955e-01
7.38996804e-01 2.51730442e-01 -1.32356346e-01 6.22431099e-01
-1.59305751e-01 3.10172647e-01 2.12776050e-01 1.57532126e-01
1.58423877e+00 6.33004606e-02 6.38276115e-02 1.35188356e-01
3.51496398e-01 -1.01756670e-01 8.22322309e-01 1.15803790e+00
-2.90572103e-02 2.23887548e-01 2.35462680e-01 -7.66784668e-01
-1.07965934e+00 -1.32606494e+00 5.59691526e-02 1.28046870e+00
1.12395667e-01 -5.21924317e-01 -1.75157651e-01 -7.75467753e-01
3.98341298e-01 1.17882633e+00 -5.11832058e-01 -6.81539893e-01
-5.10795712e-01 -7.47684419e-01 -1.67232960e-01 2.68574297e-01
1.16862938e-01 -8.12410533e-01 -4.85954523e-01 9.33084846e-01
1.79997027e-01 -6.54849648e-01 -5.07686436e-01 2.72756279e-01
-1.26964772e+00 -9.61509407e-01 -3.57621431e-01 -2.48803422e-01
3.05948049e-01 2.01821312e-01 1.18532956e+00 -3.75513315e-01
-3.56052220e-01 3.52070123e-01 -4.78244841e-01 -8.02746415e-01
-6.46649778e-01 2.40390748e-01 3.21230173e-01 1.99339822e-01
3.97611529e-01 -9.00395691e-01 -1.00542903e+00 4.68017012e-01
-9.73968983e-01 -6.02854311e-01 5.38849711e-01 8.21926177e-01
5.44723451e-01 5.47821522e-01 9.40978289e-01 -1.12988794e+00
7.90206373e-01 -7.38675416e-01 -5.97582161e-01 3.43213052e-01
-1.20053291e+00 1.49652828e-02 8.82542610e-01 -1.17506909e+00
-1.05196238e+00 -1.76129509e-02 3.45140547e-01 -5.76996386e-01
9.45778564e-02 1.87707484e-01 2.00326696e-01 5.81139565e-01
6.91803455e-01 3.95599633e-01 -7.86788687e-02 -6.36543691e-01
1.89654663e-01 7.04437971e-01 3.88216674e-01 -5.42587161e-01
4.80366588e-01 2.93664545e-01 -4.25280482e-01 -7.22767472e-01
-7.67680347e-01 -3.78527939e-01 -3.64866346e-01 -2.98244745e-01
-9.15339589e-02 -8.07378471e-01 -7.03578353e-01 4.60001439e-01
-7.70459414e-01 -7.56336004e-02 -7.13538289e-01 2.54532427e-01
-5.53123891e-01 2.10283771e-01 -2.85603285e-01 -1.20616114e+00
-3.67461026e-01 -5.82230866e-01 6.88540399e-01 1.20267928e-01
-2.74270505e-01 -6.15423024e-01 3.45651597e-01 -1.73448637e-01
8.10421765e-01 9.97360051e-02 8.29035044e-01 -7.71337569e-01
-5.80201566e-01 -5.11477947e-01 5.71970165e-01 4.08057392e-01
3.23659182e-01 -1.71642751e-01 -6.65352046e-01 -6.23715222e-01
1.10583924e-01 1.56101093e-01 8.76710653e-01 1.94013357e-01
1.29076314e+00 -4.63756084e-01 -3.76540542e-01 8.64780322e-02
1.32631564e+00 4.22299832e-01 4.35017884e-01 2.74868608e-01
1.35533378e-01 9.91163328e-02 8.06202054e-01 1.09399509e+00
1.05345137e-02 5.54295778e-01 5.21168053e-01 7.18029797e-01
1.27465963e-01 -6.89324141e-01 6.17755949e-01 8.97180736e-01
2.47800425e-01 -3.22462976e-01 -2.93907583e-01 5.24212480e-01
-2.12246490e+00 -1.08332539e+00 2.17901513e-01 3.01472950e+00
8.78851116e-01 6.94504917e-01 3.59329671e-01 2.21824542e-01
7.82028913e-01 -2.60157228e-01 -1.33632457e+00 -2.34662786e-01
1.55110568e-01 5.00954427e-02 5.33871710e-01 6.52561262e-02
-6.76338077e-01 4.75817919e-01 6.01054573e+00 7.65698075e-01
-1.04303920e+00 2.75697857e-01 4.21792805e-01 -6.12136126e-01
-1.67086750e-01 -2.57837087e-01 -1.19144797e+00 1.01339018e+00
1.23912311e+00 -4.26535547e-01 4.93354857e-01 1.01981258e+00
1.91393971e-01 -1.16919503e-01 -1.15550506e+00 9.85864401e-01
-9.58816409e-02 -1.09160841e+00 9.86930449e-04 -1.59005314e-01
8.44792724e-01 -4.12178673e-02 3.02464604e-01 6.97379649e-01
6.43680394e-01 -3.51937801e-01 6.18009627e-01 8.34767282e-01
3.40142578e-01 -8.38162065e-01 6.38775885e-01 5.84213197e-01
-8.94714892e-01 -6.69266582e-01 -4.35549855e-01 1.51997209e-01
3.21037978e-01 1.15161800e+00 -8.04649711e-01 4.94464517e-01
8.19806516e-01 4.59024221e-01 -4.03854311e-01 1.28336406e+00
9.48161855e-02 8.99703979e-01 -5.19787133e-01 -2.21465409e-01
-4.47038472e-01 1.52340904e-01 8.10859203e-01 7.81754375e-01
5.74117124e-01 -6.21620595e-01 1.90475360e-01 5.38485527e-01
2.97716171e-01 6.00201217e-03 -3.90594751e-01 9.65977386e-02
6.94569826e-01 8.19877088e-01 -4.54744995e-01 -2.96491802e-01
-1.29362077e-01 9.83202517e-01 1.32415136e-02 4.13718134e-01
-9.49532986e-01 -1.94811612e-01 6.09240770e-01 4.64509249e-01
7.32055783e-01 -2.95812011e-01 -3.38884629e-02 -1.04693913e+00
1.04185931e-01 -6.61729574e-01 4.64168519e-01 -1.57372400e-01
-1.72147357e+00 5.66568196e-01 -2.46031657e-01 -1.46978664e+00
-4.77785200e-01 -6.93407059e-02 -5.60475230e-01 3.60047370e-01
-1.19977558e+00 -5.74261725e-01 -1.17669865e-01 7.59386539e-01
9.47481215e-01 -3.95935535e-01 5.17241120e-01 3.52332622e-01
-5.74684322e-01 8.35500598e-01 5.87394714e-01 -8.73600185e-01
9.48892295e-01 -1.11346161e+00 1.61631271e-01 4.23439384e-01
-1.15164652e-01 5.86656868e-01 1.14749825e+00 -9.51231182e-01
-1.54189897e+00 -1.29233062e+00 2.49987781e-01 -3.53824437e-01
8.45003188e-01 -3.94817561e-01 -1.12828028e+00 3.21503371e-01
-1.28701404e-01 -2.92951614e-01 5.11684239e-01 2.72465736e-01
-3.07531208e-01 -8.32180858e-01 -1.06922936e+00 3.99729371e-01
7.87983298e-01 8.46113190e-02 -2.64122814e-01 4.19636607e-01
7.20771372e-01 -1.89551547e-01 -6.94928527e-01 1.87424406e-01
6.29991412e-01 -9.06436026e-01 7.10313976e-01 -8.10812712e-01
-2.13450305e-02 -2.29173675e-01 -1.93601057e-01 -1.45590770e+00
-3.57688904e-01 -1.05865908e+00 -9.42505181e-01 1.27107000e+00
4.11677271e-01 -7.77621090e-01 7.49369144e-01 4.15353119e-01
1.67853206e-01 -2.97804326e-01 -9.49664712e-01 -1.18371642e+00
-2.49674886e-01 -4.10738587e-01 9.25080776e-01 3.97202104e-01
-1.38680160e-01 3.40151489e-01 -6.83244586e-01 1.17453560e-01
8.78427148e-01 1.78273499e-01 9.53244567e-01 -1.26975536e+00
-6.61376476e-01 -1.56565353e-01 -7.14962697e-03 -1.00861502e+00
-4.58013296e-01 -3.68274331e-01 -4.42176573e-02 -7.54740775e-01
-2.01011553e-01 -6.27689421e-01 -9.51872408e-01 5.81459291e-02
1.17999362e-02 -3.96402001e-01 -6.76424652e-02 4.52435941e-01
-9.95523930e-01 6.28812730e-01 9.72670138e-01 -1.17488861e-01
-7.26415753e-01 7.69072354e-01 -4.37044293e-01 2.87349194e-01
1.08234596e+00 -6.04384124e-01 -6.63035452e-01 8.86333138e-02
4.19260234e-01 9.15341824e-02 -1.02317974e-01 -1.11962974e+00
3.95059556e-01 -5.95006570e-02 4.29011524e-01 -9.67421055e-01
-2.85248850e-02 -9.54450548e-01 4.21944350e-01 5.11609435e-01
-2.25232527e-01 2.24197894e-01 1.96187362e-01 1.41099560e+00
1.53680816e-01 -8.81573260e-02 4.51587111e-01 -7.18787760e-02
-4.64863598e-01 3.07579726e-01 -3.80351633e-01 -1.42302707e-01
1.16473448e+00 -1.06594726e-01 -1.76228508e-01 -4.31141466e-01
-8.72943103e-01 2.18373910e-01 1.29164472e-01 9.56546426e-01
3.21243554e-01 -1.27360511e+00 -4.09760803e-01 3.29363763e-01
9.80570689e-02 -1.81614414e-01 6.03862643e-01 6.69451714e-01
1.74818382e-01 1.59212947e-01 1.29253387e-01 -7.26570904e-01
-7.78524637e-01 9.87886310e-01 -8.42543319e-02 -4.79338348e-01
-6.27732992e-01 4.48269308e-01 -2.59338588e-01 6.74264356e-02
4.94604766e-01 -3.38757664e-01 9.86647159e-02 3.09923310e-02
9.47215259e-01 6.21845305e-01 4.72042441e-01 4.16507244e-01
-5.79774007e-02 -4.56749797e-02 -8.45440030e-01 2.79268205e-01
1.30860484e+00 -3.72606128e-01 3.47986311e-01 1.07186437e+00
4.94933218e-01 -1.47883430e-01 -1.45928311e+00 -2.78310269e-01
2.90609580e-02 -5.71045816e-01 -7.09973350e-02 -9.85330880e-01
-9.77080345e-01 4.67164755e-01 9.00173128e-01 4.77604687e-01
1.06628609e+00 -4.84875143e-01 1.04555118e+00 2.93968022e-01
7.20950484e-01 -1.63116789e+00 4.40596223e-01 3.02146077e-01
7.46966958e-01 -8.98219883e-01 -6.33803457e-02 1.06663287e-01
-5.40950418e-01 1.02073300e+00 7.59542704e-01 -4.72334039e-04
1.01478028e+00 4.59000498e-01 -2.39880815e-01 3.94528627e-01
-1.32966769e+00 2.20787808e-01 -3.85474920e-01 6.39258802e-01
-9.02869105e-02 2.58772653e-02 -2.46223018e-01 8.26969147e-01
-2.98886281e-02 4.45339203e-01 4.58658457e-01 1.07797134e+00
-6.25422001e-01 -1.33518183e+00 -3.12148362e-01 6.92184985e-01
-2.68101990e-01 2.01001778e-01 1.21920049e-01 4.31500256e-01
6.81198239e-02 9.57680881e-01 1.18997745e-01 -7.31760979e-01
5.58544338e-01 1.31107286e-01 4.04999673e-01 -3.07155520e-01
-5.02569318e-01 2.15214685e-01 -3.09080929e-01 -5.21800101e-01
1.85927719e-01 -9.30688739e-01 -8.04769635e-01 -3.57098132e-01
-4.47634041e-01 7.16315731e-02 7.28189290e-01 5.42456746e-01
7.08459675e-01 8.38364661e-01 9.84919190e-01 -4.58523214e-01
-1.27729404e+00 -1.03421128e+00 -8.00212860e-01 2.42983907e-01
2.96696126e-01 -8.91947448e-01 -5.55887580e-01 -4.17886488e-02] | [7.5717082023620605, 3.0970537662506104] |
2637487f-7dc3-457e-a78e-8b65a089b5b7 | a-warm-start-and-a-clean-crawled-corpus-a-1 | null | null | https://aclanthology.org/2022.lrec-1.464 | https://aclanthology.org/2022.lrec-1.464.pdf | A Warm Start and a Clean Crawled Corpus - A Recipe for Good Language Models | We train several language models for Icelandic, including IceBERT, that achieve state-of-the-art performance in a variety of downstream tasks, including part-of-speech tagging, named entity recognition, grammatical error detection and constituency parsing. To train the models we introduce a new corpus of Icelandic text, the Icelandic Common Crawl Corpus (IC3), a collection of high quality texts found online by targeting the Icelandic top-level-domain .is. Several other public data sources are also collected for a total of 16GB of Icelandic text. To enhance the evaluation of model performance and to raise the bar in baselines for Icelandic, we manually translate and adapt the WinoGrande commonsense reasoning dataset. Through these efforts we demonstrate that a properly cleaned crawled corpus is sufficient to achieve state-of-the-art results in NLP applications for low to medium resource languages, by comparison with models trained on a curated corpus. We further show that initializing models using existing multilingual models can lead to state-of-the-art results for some downstream tasks. | ['Hafsteinn Einarsson', 'Vilhjalmur THorsteinsson', 'Haukur Jónsson', 'Svanhvít Lilja Ingólfsdóttir', 'Pétur Orri Ragnarsson', 'Haukur Barri Símonarson', 'Vésteinn Snæbjarnarson'] | null | null | null | null | lrec-2022-6 | ['grammatical-error-detection', 'constituency-parsing'] | ['natural-language-processing', 'natural-language-processing'] | [-1.69976711e-01 3.27215165e-01 -1.04377523e-01 -4.01048034e-01
-1.64801908e+00 -1.07701027e+00 6.04299903e-01 3.42840940e-01
-7.68315017e-01 9.25647676e-01 6.31425321e-01 -5.97006083e-01
3.07772934e-01 -5.46196282e-01 -8.87422740e-01 -9.82942656e-02
-2.60226503e-02 9.52541590e-01 1.94153674e-02 -4.77193117e-01
3.72965932e-02 1.14623740e-01 -9.15072381e-01 7.46160150e-01
1.11270118e+00 1.82340309e-01 3.55309486e-01 7.71305978e-01
-1.16176538e-01 9.08123732e-01 -6.54145539e-01 -6.98574841e-01
2.59774864e-01 -1.22804634e-01 -1.28949106e+00 -5.55688977e-01
4.77737039e-01 1.03730649e-01 -1.61720738e-01 9.42177534e-01
3.76155645e-01 -2.48591572e-01 4.25160915e-01 -7.47998118e-01
-4.52122867e-01 1.13807368e+00 1.33553058e-01 3.46271932e-01
3.71509433e-01 8.83787274e-02 1.42690551e+00 -7.51375556e-01
1.35160816e+00 1.22737765e+00 7.76892722e-01 4.48579848e-01
-8.98540676e-01 -6.54798746e-01 1.48432612e-01 -5.46355024e-02
-1.04914689e+00 -6.67234242e-01 3.28039914e-01 -2.74924189e-01
1.75283003e+00 5.28080687e-02 2.47748107e-01 1.28525281e+00
3.36605161e-01 8.55338812e-01 1.04711473e+00 -7.86522090e-01
-1.95664257e-01 -4.78143524e-03 1.87251732e-01 7.16335297e-01
4.75797981e-01 -9.81478915e-02 -5.75484097e-01 -1.86299309e-01
-7.12640956e-02 -7.40807533e-01 -1.22776493e-01 2.79231071e-01
-1.23671436e+00 8.15968573e-01 -1.29585177e-01 5.95863163e-01
-2.81983286e-01 9.56775472e-02 7.87124991e-01 5.37093997e-01
9.35095370e-01 5.30039251e-01 -1.29868138e+00 -5.40874720e-01
-9.71169412e-01 3.06715637e-01 1.30774653e+00 1.04931533e+00
4.97174680e-01 -3.73480320e-01 2.84688622e-01 9.69826460e-01
2.79889733e-01 8.25349569e-01 3.61533105e-01 -9.57841754e-01
1.49285734e+00 5.08884430e-01 1.07021816e-01 -2.87362754e-01
-5.22244751e-01 -2.59911448e-01 -8.25343281e-02 -1.55143827e-01
6.56640768e-01 -4.94920284e-01 -9.89996850e-01 1.65714204e+00
2.85492033e-01 -6.79437399e-01 7.17477620e-01 4.19461757e-01
8.63846660e-01 8.73888791e-01 4.66832340e-01 6.91896677e-02
1.59478676e+00 -9.07718956e-01 -4.98918861e-01 -8.23864818e-01
1.25204098e+00 -9.39830959e-01 9.51226056e-01 2.18290657e-01
-1.04592824e+00 -1.56078324e-01 -1.03935027e+00 -4.40021932e-01
-6.29926026e-01 1.64919779e-01 5.21058083e-01 5.22894919e-01
-7.62619019e-01 5.17655194e-01 -1.24385822e+00 -7.17695236e-01
-7.48406425e-02 -1.35846764e-01 -7.48190224e-01 -2.19484240e-01
-1.49406505e+00 1.22988975e+00 6.61870241e-01 -3.12411785e-01
-8.10994446e-01 -7.68338799e-01 -1.22522438e+00 -1.90154269e-01
2.24362593e-02 -1.07881270e-01 1.42344677e+00 -3.99916470e-01
-1.03900301e+00 1.53414059e+00 2.11959779e-02 -4.67842728e-01
2.77624160e-01 -6.89655304e-01 -5.78286707e-01 7.88697377e-02
6.99128330e-01 5.24150968e-01 -2.87225959e-03 -9.12317038e-01
-9.37392354e-01 -3.80532891e-01 -1.22468859e-01 1.74882337e-01
-3.08760107e-02 5.84250629e-01 -2.92276889e-01 -5.11458278e-01
-1.01170335e-02 -1.07437503e+00 -1.14948802e-01 -8.83757770e-01
-2.27511227e-01 -2.18080834e-01 4.94660884e-01 -1.55482280e+00
1.10630774e+00 -1.85193241e+00 -2.05921724e-01 -2.10073739e-01
-3.50832492e-01 3.95267546e-01 -3.21121216e-01 7.87222087e-01
1.80964872e-01 5.19086719e-01 -4.33482260e-01 -3.71210307e-01
-1.82794202e-02 4.22920227e-01 1.76968928e-02 3.76770496e-01
5.85447490e-01 8.51137757e-01 -1.04806793e+00 -5.74438334e-01
-9.77746546e-02 2.16646627e-01 -4.68383163e-01 1.68853298e-01
-3.65527272e-01 3.87244910e-01 -3.17159921e-01 6.34694576e-01
4.99264121e-01 1.81079090e-01 5.80781996e-01 3.83182913e-01
-3.06636482e-01 1.40351546e+00 -5.48910618e-01 2.05518484e+00
-8.75548005e-01 5.31286597e-01 4.22849059e-01 -6.37896180e-01
6.91244364e-01 3.58622938e-01 2.59565543e-02 -6.04929984e-01
-1.91736326e-01 6.37471735e-01 8.94677639e-02 -6.07359111e-01
7.58200586e-01 -1.29648030e-01 -7.94742584e-01 3.15875292e-01
4.53454405e-01 -2.99748331e-01 5.03837407e-01 2.26556346e-01
1.47334921e+00 5.14387369e-01 3.76050651e-01 -6.46912575e-01
4.47250336e-01 8.01745117e-01 8.97904694e-01 5.05101740e-01
-9.98843983e-02 4.07920957e-01 3.32592100e-01 -3.80351663e-01
-1.26386869e+00 -7.77980506e-01 -2.37032309e-01 1.35106289e+00
-5.35416305e-01 -6.28053784e-01 -8.86608839e-01 -1.14943838e+00
-2.90641189e-01 1.09714520e+00 -3.58365774e-01 4.68286157e-01
-1.39427519e+00 -9.78569567e-01 1.33847427e+00 3.54751706e-01
6.06099665e-01 -1.26554024e+00 -4.41720545e-01 6.12179458e-01
-7.96636820e-01 -1.59716451e+00 -1.35267451e-01 5.86055219e-01
-6.10839784e-01 -1.15191841e+00 -3.80947232e-01 -1.07210135e+00
1.97958257e-02 -5.47610879e-01 1.72636926e+00 -2.13370055e-01
8.96561146e-02 8.90505537e-02 -6.35781288e-01 -3.08415562e-01
-1.30431557e+00 6.25281870e-01 -3.67131561e-01 -9.43021715e-01
6.04347944e-01 -3.41305822e-01 8.01762938e-02 -2.93563366e-01
-6.24619365e-01 -1.36231363e-01 5.26963413e-01 6.76208735e-01
1.26840264e-01 -5.88704169e-01 4.23404902e-01 -1.25678182e+00
2.95209944e-01 -4.42072034e-01 -4.84291613e-01 3.27350646e-01
-2.98729241e-01 3.61917078e-01 9.07009661e-01 2.41266623e-01
-1.52807403e+00 -5.56900054e-02 -6.09919071e-01 6.10115111e-01
-3.73861879e-01 6.12625182e-01 -3.40970188e-01 4.78700846e-01
7.85867989e-01 -4.32074703e-02 -4.58935946e-01 -7.46464610e-01
7.50253141e-01 1.07429981e+00 6.89775050e-01 -8.63244653e-01
4.49873477e-01 7.95188323e-02 -4.42197502e-01 -7.27636456e-01
-1.09454143e+00 -6.19026482e-01 -7.96440780e-01 3.73740256e-01
9.81201589e-01 -1.35301530e+00 -1.46402314e-01 3.32386702e-01
-1.48003674e+00 -6.87461317e-01 3.16659547e-02 3.44884694e-01
-3.72043431e-01 2.34963074e-01 -1.38782322e+00 -5.49737573e-01
-7.58642018e-01 -7.37118483e-01 1.16413045e+00 -2.95246750e-01
-2.96443909e-01 -1.21786165e+00 6.03817761e-01 5.20679593e-01
-1.72851726e-01 1.26083568e-01 1.30107701e+00 -1.13792384e+00
-5.74836135e-02 1.29960418e-01 -5.21511622e-02 4.51327890e-01
-2.15846732e-01 -2.94924319e-01 -6.39998734e-01 -2.11485952e-01
-1.47700652e-01 -5.22060275e-01 7.77020216e-01 -2.01850589e-02
-8.52708146e-02 -3.24484229e-01 -3.97751778e-01 3.73710960e-01
1.47367036e+00 1.45526424e-01 6.26783788e-01 9.58063185e-01
4.95741814e-01 6.15869105e-01 6.98275924e-01 8.51106495e-02
7.39489615e-01 1.53686926e-01 -5.31933345e-02 8.31583515e-02
-4.03406054e-01 -5.13003588e-01 6.75802708e-01 1.14291263e+00
2.82561988e-01 -3.88092697e-01 -1.53800642e+00 1.12521136e+00
-1.78612423e+00 -6.52607620e-01 -2.30325073e-01 1.78529048e+00
1.21525788e+00 7.82333314e-02 -3.46265316e-01 -3.48101914e-01
7.27268517e-01 1.94842830e-01 7.29186013e-02 -5.12088358e-01
-2.97496468e-01 4.53810751e-01 7.93277383e-01 6.65663183e-01
-1.29470205e+00 1.65326798e+00 6.76564550e+00 7.08895504e-01
-7.28392482e-01 5.52321136e-01 5.44266582e-01 1.10493056e-01
-2.14738131e-01 2.16775909e-01 -1.36814475e+00 2.86307901e-01
1.65839612e+00 2.54504681e-01 3.17089021e-01 7.72371113e-01
2.81205684e-01 4.91379499e-02 -9.46801007e-01 5.19956291e-01
8.26310664e-02 -1.24599707e+00 -3.58503550e-01 -1.20738655e-01
8.55433285e-01 1.02193809e+00 -7.10300744e-01 7.26935923e-01
1.00468659e+00 -7.38549113e-01 8.65739942e-01 -1.39063984e-01
8.15603137e-01 -7.07903385e-01 1.01128852e+00 4.49085951e-01
-8.90950680e-01 2.35289037e-01 -3.23781222e-01 1.25466175e-02
4.12472814e-01 6.21583760e-01 -9.07466769e-01 6.73726797e-01
7.56935954e-01 5.91730237e-01 -5.87755859e-01 4.28159982e-01
-7.62321889e-01 1.15586436e+00 -6.12545669e-01 -9.95227620e-02
6.06806755e-01 6.20472170e-02 7.04172790e-01 1.92721844e+00
9.95816514e-02 2.66176164e-02 2.46283427e-01 4.84446645e-01
-4.86932337e-01 5.19018412e-01 -5.04051507e-01 -1.16389334e-01
5.08831084e-01 1.24833882e+00 -4.22378480e-01 -5.34028769e-01
-5.34423053e-01 8.76848042e-01 8.38239133e-01 -2.75999252e-02
-5.18408954e-01 -7.07008541e-01 5.92656672e-01 4.98536974e-02
2.17097446e-01 -4.66880172e-01 -2.33382255e-01 -1.27925217e+00
1.86370090e-01 -1.11645103e+00 6.58445954e-01 -5.78824282e-01
-1.30652058e+00 7.58535147e-01 -2.27730367e-02 -6.52368307e-01
-6.52515471e-01 -8.95145953e-01 -3.16097409e-01 7.73222148e-01
-1.56040633e+00 -1.51097023e+00 3.11430961e-01 1.53175071e-01
5.79717457e-01 -1.36115387e-01 9.65599000e-01 3.69775146e-01
-1.79810241e-01 2.43733793e-01 2.48008102e-01 7.31848240e-01
8.62108111e-01 -1.27161360e+00 1.10563755e+00 1.17960107e+00
2.51374722e-01 3.80617946e-01 6.20510995e-01 -1.03123736e+00
-1.08080590e+00 -1.26719415e+00 1.74999440e+00 -7.00911880e-01
1.07158804e+00 -6.45177960e-01 -7.06884861e-01 1.12265587e+00
4.34942722e-01 -3.27018559e-01 5.15915573e-01 4.98461306e-01
-3.12288582e-01 3.93582642e-01 -1.14334333e+00 3.21968168e-01
1.25861096e+00 -4.32933480e-01 -1.21967673e+00 5.83354473e-01
8.39453518e-01 -4.40635115e-01 -9.92163479e-01 2.46742442e-01
2.94141561e-01 -6.21936560e-01 2.95413584e-01 -8.56328368e-01
5.74324548e-01 1.28251553e-01 -3.42408419e-01 -1.53387296e+00
-1.07226416e-01 -7.58130014e-01 5.60343981e-01 1.71922266e+00
1.05194199e+00 -6.59077406e-01 5.75439572e-01 5.08516252e-01
-6.66051567e-01 -4.36048210e-02 -1.16798878e+00 -9.24583077e-01
7.61436403e-01 -6.15076900e-01 1.67577371e-01 9.16039169e-01
4.94879782e-01 7.29047060e-01 1.46391049e-01 1.13921344e-01
3.60627860e-01 -1.10823192e-01 4.71455306e-01 -1.01457632e+00
-1.99400693e-01 2.45985448e-01 -1.16752550e-01 -8.13985169e-01
6.78189754e-01 -1.30769300e+00 5.55050671e-01 -1.76651943e+00
1.38424367e-01 -7.16853023e-01 3.65077198e-01 8.49887788e-01
-9.54136848e-02 9.54059064e-02 2.87777692e-01 1.60109237e-01
-5.23790896e-01 1.31021336e-01 6.14928246e-01 1.26232490e-01
-2.46648848e-01 -6.54742479e-01 -6.50450349e-01 8.34351897e-01
7.10757434e-01 -1.02189779e+00 4.53190327e-01 -7.52332509e-01
4.06387925e-01 1.23715717e-02 -1.66373312e-01 -9.02899802e-01
-1.33961126e-01 7.25150257e-02 9.32241976e-02 -6.05139434e-01
-1.93814281e-03 -3.37402016e-01 -1.61170796e-01 5.79645514e-01
-2.35546082e-01 2.77110428e-01 3.65475744e-01 5.87153099e-02
-2.39375517e-01 -5.31238496e-01 6.81605220e-01 -6.51060879e-01
-7.93804765e-01 -3.24832827e-01 -7.58689582e-01 9.78482723e-01
3.91968578e-01 5.80492735e-01 -7.43286133e-01 -2.97064222e-02
-4.56803232e-01 7.99225196e-02 3.38737190e-01 3.70001853e-01
-3.10175240e-01 -7.54148304e-01 -1.27923167e+00 4.90646474e-02
2.46480867e-01 -2.43000373e-01 -3.77291054e-01 2.84801662e-01
-9.74344075e-01 8.86588156e-01 -6.81214221e-03 -2.74362266e-01
-9.43615258e-01 2.06091180e-01 -7.84954727e-02 -1.14299083e+00
-7.05352306e-01 4.45111424e-01 -2.56319106e-01 -1.14211142e+00
-5.03845751e-01 -3.75137657e-01 1.30275160e-01 -2.32168227e-01
2.95760751e-01 1.21606387e-01 5.59482336e-01 -7.52856791e-01
-7.03701377e-01 2.11161584e-01 -1.38053760e-01 -5.59799910e-01
1.48960483e+00 -1.24757953e-01 -1.92091167e-02 2.87379473e-01
1.10613966e+00 4.21866864e-01 -8.07852685e-01 1.81085989e-01
4.06374216e-01 1.72854871e-01 -1.73474163e-01 -1.38765788e+00
-4.00909036e-01 6.45253897e-01 -1.80947818e-02 -1.83728904e-01
5.76773643e-01 2.99383372e-01 1.06860185e+00 7.08380222e-01
6.53235912e-01 -1.40619254e+00 -6.64268851e-01 1.22908771e+00
5.51249564e-01 -1.25463021e+00 -2.51051039e-01 -5.26001215e-01
-6.81978285e-01 8.46430540e-01 2.19859824e-01 -2.06776768e-01
5.50769508e-01 7.49769211e-01 4.87818867e-01 -1.14712402e-01
-7.53874004e-01 -2.96186000e-01 -1.18150771e-01 5.54151416e-01
8.73552740e-01 1.53499290e-01 -8.42672288e-01 7.93129921e-01
-6.36525095e-01 -3.97276670e-01 4.48093981e-01 1.02096403e+00
-4.00945097e-01 -1.40170193e+00 -2.27782860e-01 1.59401596e-01
-1.26057732e+00 -8.92339110e-01 -2.74384528e-01 1.37356889e+00
-5.50233163e-02 1.24623895e+00 4.57728194e-04 6.58416897e-02
2.68819302e-01 5.29416561e-01 3.74122262e-01 -8.99870336e-01
-1.02762020e+00 -1.81707535e-02 1.15284395e+00 -2.97440559e-01
-2.72877991e-01 -6.88852012e-01 -1.46559691e+00 -2.11127505e-01
-1.99144408e-01 4.68931228e-01 9.50302482e-01 1.14506269e+00
4.26027209e-01 5.03008701e-02 -8.81453976e-02 -6.08108461e-01
-3.49666953e-01 -1.35829616e+00 -3.52539718e-01 5.19849896e-01
-3.01458389e-01 -2.30099484e-02 -4.32209730e-01 2.93437332e-01] | [10.471439361572266, 9.838172912597656] |
9e8e6d6f-1f60-4fa5-99f0-c6403810c6cf | occlusion-aware-networks-for-3d-human-pose | null | null | http://openaccess.thecvf.com/content_ICCV_2019/html/Cheng_Occlusion-Aware_Networks_for_3D_Human_Pose_Estimation_in_Video_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Cheng_Occlusion-Aware_Networks_for_3D_Human_Pose_Estimation_in_Video_ICCV_2019_paper.pdf | Occlusion-Aware Networks for 3D Human Pose Estimation in Video | Occlusion is a key problem in 3D human pose estimation from a monocular video. To address this problem, we introduce an occlusion-aware deep-learning framework. By employing estimated 2D confidence heatmaps of keypoints and an optical-flow consistency constraint, we filter out the unreliable estimations of occluded keypoints. When occlusion occurs, we have incomplete 2D keypoints and feed them to our 2D and 3D temporal convolutional networks (2D and 3D TCNs) that enforce temporal smoothness to produce a complete 3D pose. By using incomplete 2D keypoints, instead of complete but incorrect ones, our networks are less affected by the error-prone estimations of occluded keypoints. Training the occlusion-aware 3D TCN requires pairs of a 3D pose and a 2D pose with occlusion labels. As no such a dataset is available, we introduce a "Cylinder Man Model" to approximate the occupation of body parts in 3D space. By projecting the model onto a 2D plane in different viewing angles, we obtain and label the occluded keypoints, providing us plenty of training data. In addition, we use this model to create a pose regularization constraint, preferring the 2D estimations of unreliable keypoints to be occluded. Our method outperforms state-of-the-art methods on Human 3.6M and HumanEva-I datasets.
| [' Robby T. Tan', ' Wending Yan', ' Bo Wang', ' Bo Yang', 'Yu Cheng'] | 2019-10-01 | null | null | null | iccv-2019-10 | ['monocular-3d-human-pose-estimation'] | ['computer-vision'] | [-4.77414489e-01 4.37594540e-02 -2.87788749e-01 -3.00235361e-01
-2.55783349e-01 -3.71380627e-01 3.02439302e-01 -3.50262821e-01
-6.14069700e-01 6.70713186e-01 1.67171448e-01 3.13723207e-01
2.69178420e-01 -3.85113239e-01 -9.55311894e-01 -1.79603934e-01
-2.05809236e-01 5.90979874e-01 3.08860481e-01 1.12886086e-01
-9.93266851e-02 6.36189282e-01 -1.50528371e+00 -3.78541090e-02
5.67607522e-01 1.06219220e+00 -2.08171308e-01 5.78322828e-01
2.76351213e-01 2.78505474e-01 -7.03639328e-01 -2.85594106e-01
6.22789800e-01 -4.15069461e-02 -5.69579840e-01 1.52546361e-01
1.03712881e+00 -7.87441909e-01 -6.18181229e-01 8.04183185e-01
4.77731824e-01 2.72911668e-01 3.08525801e-01 -1.45776868e+00
-2.28099287e-01 -3.02539676e-01 -7.17653871e-01 -1.30530357e-01
8.17798555e-01 2.31871799e-01 5.89139104e-01 -1.11225235e+00
1.12231946e+00 1.60378408e+00 7.27507472e-01 5.74572086e-01
-1.06897700e+00 -5.07192194e-01 4.34677273e-01 1.13928042e-01
-1.55653441e+00 -2.23774254e-01 8.56349945e-01 -5.62908053e-01
6.91858292e-01 1.64305881e-01 1.19378078e+00 1.22594500e+00
3.24828178e-01 9.96572673e-01 5.97162664e-01 -9.13776755e-02
7.09340200e-02 -1.97542697e-01 -3.03052757e-02 7.89249718e-01
1.92210123e-01 3.85611683e-01 -7.41629779e-01 -5.80776818e-02
1.07290864e+00 3.23118687e-01 -3.07571292e-01 -1.03357887e+00
-1.26125586e+00 5.73208153e-01 6.97567284e-01 -2.82811582e-01
-4.38029915e-01 2.88901716e-01 1.60805926e-01 -1.31922895e-02
6.17300749e-01 1.20981596e-01 -6.17139339e-01 2.76497360e-02
-8.31227422e-01 6.77478909e-01 4.68634367e-01 1.19687569e+00
7.30616510e-01 -2.32078254e-01 -1.75756544e-01 4.47943658e-01
3.52715105e-01 5.49951792e-01 -1.38127282e-02 -1.35285485e+00
5.41871905e-01 7.52178073e-01 5.36397755e-01 -1.03505504e+00
-7.29468942e-01 -5.13772011e-01 -6.52035713e-01 4.62208241e-01
5.48505604e-01 1.91403609e-02 -1.12727618e+00 1.68342268e+00
8.23010981e-01 -5.83449490e-02 -4.64052677e-01 1.53540146e+00
5.26046634e-01 3.07348102e-01 -1.97555482e-01 9.48615894e-02
1.25755632e+00 -1.01465583e+00 -6.81444705e-01 -2.74070263e-01
4.13385928e-01 -5.54034889e-01 6.77730322e-01 5.25595665e-01
-1.05862451e+00 -1.00840056e+00 -1.00427735e+00 -3.41056496e-01
-2.15884387e-01 3.02648336e-01 5.65250933e-01 1.63377300e-01
-6.75406635e-01 6.32438958e-01 -1.13245010e+00 -1.95429698e-01
2.20389858e-01 2.20383182e-01 -7.83946276e-01 -1.35106146e-01
-1.14881802e+00 9.51531112e-01 1.29812419e-01 5.99329889e-01
-8.64727020e-01 -6.12839520e-01 -1.34064531e+00 -3.80777955e-01
5.73512435e-01 -9.79602933e-01 9.60622966e-01 -3.45323652e-01
-1.26711750e+00 9.38587725e-01 -2.28161737e-01 -2.66332418e-01
1.11815774e+00 -1.12732291e+00 2.85088853e-03 2.74061918e-01
1.49939001e-01 1.08623898e+00 9.28447604e-01 -1.16729164e+00
-3.99486035e-01 -5.49523175e-01 6.12822138e-02 4.71534610e-01
2.95832098e-01 -4.73038435e-01 -9.15228248e-01 -5.02491057e-01
6.15096509e-01 -1.15617085e+00 -1.15007728e-01 8.11033428e-01
-6.28680348e-01 -1.53375328e-01 6.98127925e-01 -7.59608865e-01
9.03389275e-01 -1.92587197e+00 5.59254885e-01 1.46895379e-01
3.80730271e-01 1.33852407e-01 3.36324163e-02 -1.54064730e-01
1.08147688e-01 -2.66183972e-01 1.76683366e-01 -8.65231991e-01
6.08956628e-02 2.71071553e-01 7.55425394e-02 8.93172801e-01
2.47883826e-01 8.36062968e-01 -9.94466305e-01 -4.69605148e-01
6.71382070e-01 9.12181199e-01 -6.47715390e-01 4.36008722e-01
-2.61706650e-01 7.65974998e-01 -4.10492808e-01 7.07396507e-01
8.76521766e-01 -4.91705127e-02 -2.00440183e-01 -4.44804341e-01
-8.19375813e-02 1.11804113e-01 -1.45787537e+00 2.30666733e+00
-1.89736634e-01 4.10253972e-01 -1.52778953e-01 -4.16259527e-01
7.50915766e-01 2.97232509e-01 6.21227503e-01 -4.29784566e-01
1.43644765e-01 -5.07217124e-02 -5.24846554e-01 -4.34143543e-01
4.09154415e-01 4.39025939e-01 1.06757402e-01 6.62526861e-02
1.44412994e-01 -8.63103494e-02 -2.32441425e-02 1.93513930e-02
8.11004102e-01 1.04859483e+00 6.70776367e-02 1.45841017e-01
4.07790244e-01 -2.69405007e-01 8.52554560e-01 4.75492418e-01
-4.83000606e-01 9.97429788e-01 4.49328154e-01 -1.13022411e+00
-1.09374928e+00 -1.27322435e+00 -2.16802713e-02 5.28976738e-01
3.27269524e-01 -6.01047814e-01 -5.43351114e-01 -7.92363644e-01
2.62437969e-01 1.05878018e-01 -6.87534750e-01 -3.37970108e-02
-9.04954612e-01 -9.06789228e-02 8.85433331e-02 5.30759454e-01
5.23787260e-01 -6.27117276e-01 -1.02713704e+00 5.10840900e-02
-4.68787163e-01 -1.13735461e+00 -6.94754720e-01 7.43106157e-02
-7.84335315e-01 -1.36935079e+00 -1.03759348e+00 -3.89052719e-01
8.69779825e-01 5.76770306e-02 1.14493310e+00 9.21166837e-02
-3.82706821e-01 1.95608854e-01 -1.83562681e-01 -5.04505709e-02
1.24538146e-01 -1.38840288e-01 4.28185642e-01 -1.95591003e-01
3.39071482e-01 -3.04775178e-01 -9.31117356e-01 5.49350202e-01
-2.72209942e-01 1.33571491e-01 1.65004566e-01 7.63122261e-01
7.63501883e-01 -2.59475857e-01 -1.67890623e-01 -3.36481899e-01
-1.57333404e-01 2.31388286e-02 -8.11427772e-01 1.62533168e-02
-5.86468577e-02 5.07426150e-02 1.37038395e-01 -6.13376915e-01
-7.68745542e-01 6.30294740e-01 -3.81960720e-02 -1.14374924e+00
-2.32268825e-01 1.03742825e-02 -6.67113960e-02 4.99491282e-02
6.88084960e-01 -3.27747792e-01 2.46882532e-02 -6.77786827e-01
4.20588672e-01 1.06577493e-01 7.58800685e-01 -5.53657413e-01
8.07860196e-01 5.91001213e-01 1.27578601e-01 -5.00918686e-01
-1.06130230e+00 -6.65495753e-01 -1.21095383e+00 -4.46910858e-01
1.17772722e+00 -1.20017159e+00 -8.48818779e-01 5.38107216e-01
-1.47298670e+00 -2.68265419e-02 -1.99814439e-01 7.65848517e-01
-6.78334832e-01 4.43981767e-01 -5.12064040e-01 -9.35846388e-01
-1.32254526e-01 -1.26315212e+00 1.56529748e+00 -5.66400439e-02
-5.21607399e-01 -4.86106664e-01 9.28254649e-02 1.01518072e-01
-2.39465252e-01 5.58202922e-01 2.24192142e-01 -9.77561548e-02
-6.07597768e-01 -5.12266338e-01 4.24093939e-02 2.72599965e-01
-7.78208151e-02 -1.55511022e-01 -9.98246133e-01 -4.17651653e-01
-1.26177073e-01 -4.21771646e-01 7.02404797e-01 6.70989037e-01
9.68645155e-01 -4.63050529e-02 -5.86043239e-01 9.68069017e-01
7.48973191e-01 -1.96219921e-01 5.03082275e-01 2.34280512e-01
8.80031347e-01 7.31601179e-01 7.98993349e-01 4.95260447e-01
4.98846650e-01 9.77265179e-01 5.09072006e-01 -8.29659924e-02
-1.57381132e-01 -7.38437772e-01 3.03330552e-02 3.31001759e-01
-2.36074120e-01 4.40619476e-02 -5.99641562e-01 2.64751285e-01
-1.95836353e+00 -5.82948506e-01 7.18747005e-02 2.51280713e+00
6.42951667e-01 3.57164651e-01 2.56123304e-01 -6.46298472e-03
6.49172544e-01 1.58173159e-01 -7.98366487e-01 3.49512488e-01
1.95443690e-01 -1.32904604e-01 2.74180204e-01 6.66784704e-01
-1.34327352e+00 8.59800160e-01 5.86091852e+00 1.14119843e-01
-7.83583522e-01 -2.03793123e-01 3.74954551e-01 -4.79576319e-01
-2.07607653e-02 1.91823933e-02 -1.01061249e+00 3.47647011e-01
1.71378821e-01 6.07882380e-01 3.77655439e-02 9.21009600e-01
1.05851486e-01 -3.23079795e-01 -1.33230019e+00 1.34469092e+00
6.20059529e-03 -8.51276815e-01 -2.53996491e-01 9.01177377e-02
6.10773504e-01 -1.98998868e-01 -1.49815664e-01 1.28592536e-01
1.74542312e-02 -7.86700368e-01 9.69479561e-01 7.24968791e-01
7.00016439e-01 -6.10875666e-01 5.71703255e-01 4.22207087e-01
-1.29643011e+00 2.00071603e-01 -4.65164125e-01 -1.83004841e-01
4.14707839e-01 5.75267315e-01 -2.17848152e-01 4.67461407e-01
9.84397352e-01 8.04493845e-01 -3.86164695e-01 1.17950022e+00
-6.36098385e-01 -3.45433801e-01 -6.47877276e-01 4.03804302e-01
-1.23880453e-01 -5.60410023e-02 5.99501431e-01 6.47640526e-01
1.55635595e-01 -5.08305524e-03 3.89627486e-01 9.39047813e-01
2.96321698e-02 -3.96485835e-01 -5.22933185e-01 6.65172338e-01
4.10361499e-01 1.12179661e+00 -4.99508947e-01 -4.01226491e-01
-2.36897394e-01 1.37110877e+00 3.75365436e-01 5.06240547e-01
-8.24369729e-01 -9.10591781e-02 9.15369987e-01 1.51910454e-01
7.58871511e-02 -5.26086092e-01 2.47799344e-02 -1.49584854e+00
5.30395448e-01 -5.20446420e-01 3.44151199e-01 -9.72777009e-01
-1.04537177e+00 5.85229099e-01 2.99296945e-01 -1.42929006e+00
-3.68422210e-01 -5.49050450e-01 -1.50967613e-01 8.84746432e-01
-1.22051871e+00 -9.54733193e-01 -5.41079283e-01 5.01360416e-01
3.36098701e-01 2.85783917e-01 4.82885569e-01 2.36516893e-01
-1.85292765e-01 4.39596117e-01 -7.16255903e-01 1.77503332e-01
9.84739482e-01 -1.16930938e+00 8.43286514e-01 7.26194501e-01
3.70737202e-02 6.19395435e-01 6.67580664e-01 -9.23756957e-01
-1.21145344e+00 -8.04720044e-01 9.29706454e-01 -9.87589478e-01
-9.82650183e-03 -6.78233683e-01 -7.14307070e-01 8.50050330e-01
-4.96467382e-01 5.83296120e-01 -6.61473116e-03 6.13809638e-02
-3.54620218e-01 9.29491874e-03 -1.11211348e+00 5.72538674e-01
1.52856791e+00 -3.99075150e-01 -6.56038463e-01 4.26387489e-01
9.36277866e-01 -9.91357744e-01 -8.16834450e-01 4.13601398e-01
8.98207247e-01 -9.42636967e-01 1.32114387e+00 -5.42522252e-01
1.51456017e-02 -4.18541551e-01 6.19422495e-02 -1.15207732e+00
-1.27611175e-01 -5.71591079e-01 -5.00391364e-01 4.15563732e-01
-1.41418949e-01 -2.12844729e-01 1.14720368e+00 7.80845940e-01
7.61388475e-03 -6.63780808e-01 -1.19065750e+00 -8.10717463e-01
-1.77793756e-01 -4.21976298e-01 3.37610334e-01 5.12662530e-01
-1.95761785e-01 -9.71994400e-02 -8.16263378e-01 2.45216101e-01
9.42280233e-01 -6.38965368e-02 1.18549156e+00 -1.28439653e+00
2.17336304e-02 8.49479437e-02 -6.49887562e-01 -1.62268639e+00
1.62279591e-01 -2.47829825e-01 7.35917240e-02 -1.05470431e+00
-1.58377662e-01 -2.55551398e-01 1.58681929e-01 4.31625366e-01
-1.21719442e-01 3.18052828e-01 2.31476784e-01 1.57109976e-01
-6.09829485e-01 6.70790076e-01 1.46013224e+00 1.52337020e-02
-3.38506669e-01 1.22549333e-01 1.43406942e-01 9.80223298e-01
3.37783664e-01 -4.18282509e-01 -2.75183976e-01 -5.17261386e-01
2.46728793e-01 1.87490270e-01 9.36186194e-01 -1.19332433e+00
2.05313161e-01 1.28507102e-03 1.18703628e+00 -1.31828856e+00
8.75625849e-01 -8.71589839e-01 3.46702904e-01 6.13715887e-01
-1.83149800e-01 8.80037844e-02 -5.13557568e-02 4.45102751e-01
6.64765462e-02 1.76873311e-01 5.75854599e-01 -4.18025494e-01
-6.43276334e-01 8.26709628e-01 1.66847333e-01 -3.16678658e-02
8.53080094e-01 -3.51087034e-01 6.83676600e-02 -2.83688456e-01
-1.00228012e+00 3.99345726e-01 7.17901587e-01 7.14371979e-01
8.63550544e-01 -1.53670716e+00 -4.33196902e-01 6.68248296e-01
1.22656725e-01 5.88352859e-01 3.45088840e-01 6.38236880e-01
-5.98616660e-01 4.19240922e-01 -2.57059962e-01 -1.10852623e+00
-9.09009695e-01 7.04582512e-01 5.31729698e-01 6.40878305e-02
-9.67141926e-01 7.55857110e-01 2.97128946e-01 -6.51767969e-01
6.26968682e-01 -4.02096242e-01 1.15340911e-01 -1.19800784e-01
5.76064289e-01 4.66981500e-01 -1.31186888e-01 -6.42983079e-01
-4.59430814e-01 9.71933961e-01 7.20526800e-02 -1.02497645e-01
9.94796216e-01 -3.38541955e-01 2.19630212e-01 2.84823895e-01
1.36064851e+00 -2.26381868e-01 -1.93608940e+00 -2.53857136e-01
-3.86322796e-01 -8.18177164e-01 -2.84775525e-01 -5.44569194e-01
-1.07204354e+00 8.97243738e-01 6.49837613e-01 -5.29081345e-01
6.56888366e-01 -9.69034731e-02 7.44090915e-01 3.05152506e-01
4.74100649e-01 -1.19670331e+00 1.55878380e-01 4.77976829e-01
1.01803982e+00 -1.35459077e+00 2.99110025e-01 -3.59144062e-01
-2.82807589e-01 1.06497133e+00 1.00070107e+00 -1.81350842e-01
5.75137079e-01 -1.20510839e-01 1.91707700e-01 -2.50275701e-01
-4.75759655e-01 -2.53887866e-02 6.85487449e-01 7.01418042e-01
1.28801465e-01 -6.71767071e-02 -1.00052685e-01 1.28126115e-01
-2.23401695e-01 4.20491882e-02 9.77904126e-02 9.11613762e-01
-2.08596796e-01 -8.29054117e-01 -5.62522054e-01 -9.59503576e-02
-1.49910048e-01 3.56198668e-01 -1.67219162e-01 9.72518802e-01
3.30437571e-01 5.46681643e-01 2.81609744e-01 -3.89583439e-01
5.24838269e-01 -1.30974716e-02 5.96180260e-01 -4.51419234e-01
-9.71188173e-02 2.85888255e-01 -4.73340563e-02 -1.19130468e+00
-3.35719556e-01 -4.36120331e-01 -1.09987009e+00 -3.65598589e-01
-2.79468864e-01 -1.98717862e-01 6.14465654e-01 8.49202812e-01
3.41800481e-01 2.89467186e-01 2.26353854e-01 -1.43848681e+00
-4.62399513e-01 -7.16399729e-01 -3.34125131e-01 6.12025499e-01
6.13994181e-01 -1.20287299e+00 -2.79311180e-01 -8.72480050e-02] | [7.002685070037842, -0.9564274549484253] |
a0723f9b-c645-4fd3-990e-7c0bf233f349 | group-activity-recognition-using-joint | null | null | http://www.mva-org.jp/Proceedings/2021/papers/O1-2-2.pdf | http://www.mva-org.jp/Proceedings/2021/papers/O1-2-2.pdf | Group Activity Recognition Using Joint Learning of Individual Action Recognition and People Grouping | This paper proposes joint learning of individual action recognition and people grouping for improving group activity recognition. By sharing the information between two similar tasks (i.e., individual action recognition and people grouping) through joint learning, errors of these two tasks are mutually corrected. This joint learning also improves the accuracy of group activity recognition. Our proposed method is designed to consist of any individual action recognition methods as a component. The effectiveness is validated with various IAR methods. By employing existing group activity recognition methods for ensembling with the proposed method, we achieved the best performance compared to the similar SOTA group activity recognition methods. | ['Norimichi Ukita', 'Kohei Sendo', 'Chihiro Nakatani'] | 2021-07-17 | null | null | null | mva-2021-7 | ['group-activity-recognition'] | ['computer-vision'] | [ 4.10004586e-01 -9.16270763e-02 -3.95650864e-01 -2.68072307e-01
-7.70660460e-01 5.90235516e-02 7.61419117e-01 -2.97861844e-01
-5.42398989e-01 8.87118638e-01 4.09321457e-01 2.49957055e-01
-5.47620766e-02 -4.78819579e-01 -3.05242538e-01 -8.42470288e-01
-9.18195918e-02 2.47692510e-01 2.17743382e-01 4.44036245e-01
4.49296832e-01 3.39785188e-01 -1.70185184e+00 3.92272532e-01
1.00214469e+00 8.55867565e-01 -1.10124961e-01 7.64239907e-01
-2.41541378e-02 1.27919006e+00 -1.10744309e+00 2.79230863e-01
1.37420982e-01 -7.57715166e-01 -8.39305997e-01 7.93957889e-01
4.68180329e-01 -4.37881470e-01 -3.22398752e-01 5.76875448e-01
5.67518055e-01 7.68465579e-01 7.43017137e-01 -1.27432716e+00
-5.39848387e-01 4.12952602e-01 -5.43636560e-01 3.07989955e-01
6.27464414e-01 -7.53173530e-02 8.02088320e-01 -6.06215835e-01
-1.15856521e-01 1.38720834e+00 6.62529826e-01 6.91204607e-01
-7.51825571e-01 -8.52473557e-01 3.71794075e-01 4.99593735e-01
-1.53798473e+00 -5.33800304e-01 5.35213351e-01 -3.57285589e-01
1.18823195e+00 1.63984358e-01 6.06445789e-01 7.11189508e-01
-7.89211020e-02 1.34090543e+00 9.66858506e-01 -4.95179951e-01
1.41022310e-01 1.39232993e-03 4.88443762e-01 8.14233124e-01
3.32217544e-01 -2.83064783e-01 -6.01325154e-01 -1.66060299e-01
8.07369351e-01 3.85851562e-01 1.15721487e-01 -2.19202396e-02
-1.37794447e+00 2.31618837e-01 6.16918989e-02 6.27931893e-01
-3.78002733e-01 4.20603633e-01 1.55531332e-01 2.30118670e-02
3.05389106e-01 7.73500353e-02 -2.04988211e-01 -2.80464917e-01
-6.10110998e-01 -8.40039700e-02 8.23618531e-01 5.31629145e-01
7.13099301e-01 1.32905394e-01 -6.96963906e-01 1.03035438e+00
6.17072165e-01 5.71676970e-01 8.56168926e-01 -9.08888638e-01
3.86887044e-01 1.11717618e+00 3.48755687e-01 -7.38235116e-01
-2.79823899e-01 -1.46914750e-01 -7.59134471e-01 -1.47926575e-02
4.25138891e-01 -3.37451309e-01 -9.07268763e-01 1.48005784e+00
1.95700988e-01 8.94976974e-01 2.97005326e-01 4.42444474e-01
6.76214993e-01 6.78664029e-01 4.49064761e-01 -3.87483895e-01
9.67035890e-01 -1.46192372e+00 -1.12617791e+00 -1.99104458e-01
1.02679241e+00 -5.36788225e-01 3.13377142e-01 4.50459421e-01
-7.90450156e-01 -1.25487351e+00 -1.12940192e+00 4.62826788e-01
-1.53353497e-01 7.68161774e-01 9.77021754e-01 1.03917348e+00
-6.59286201e-01 5.85221112e-01 -9.74323392e-01 -6.46929860e-01
5.33103049e-01 7.62377024e-01 -3.61879766e-01 3.14546287e-01
-8.81814837e-01 5.67743957e-01 4.21759218e-01 1.70281500e-01
-8.38660359e-01 6.13318793e-02 -4.81115490e-01 -3.10373574e-01
3.96797299e-01 -3.64084423e-01 1.24648678e+00 -9.89964128e-01
-1.54932141e+00 4.63273197e-01 -1.53397426e-01 -4.71787721e-01
3.64740074e-01 -6.50358140e-01 -8.20237577e-01 -1.28288358e-01
-7.23147690e-02 3.69192511e-01 6.43068194e-01 -7.71762490e-01
-9.64349687e-01 -4.84305471e-01 -2.69711733e-01 7.32554853e-01
-5.43158352e-01 9.16595161e-02 -3.57994378e-01 -5.46674669e-01
6.08189218e-02 -7.19343603e-01 -3.97297218e-02 -2.54585117e-01
1.82964936e-01 -8.66118550e-01 9.90457356e-01 -9.27274764e-01
1.40029001e+00 -1.95393741e+00 -7.89168775e-02 1.53863356e-01
-1.92854851e-01 5.84688485e-01 -3.68664898e-02 2.43076980e-01
4.23701815e-02 -3.38101052e-02 1.63225189e-01 -5.12450099e-01
-2.47581050e-01 4.01929885e-01 2.97521591e-01 5.76935589e-01
-2.97506303e-01 8.26908231e-01 -7.49982595e-01 -6.07832849e-01
4.07727957e-01 1.22768313e-01 -2.10998729e-01 6.38224185e-01
3.69388849e-01 3.91841322e-01 -6.94299579e-01 6.92763090e-01
2.88484782e-01 -2.49649063e-01 4.05309916e-01 -1.59339476e-02
1.90244675e-01 -7.86183923e-02 -1.68394768e+00 1.58678436e+00
-9.32211727e-02 1.46563813e-01 -4.12159890e-01 -1.26561224e+00
1.21420455e+00 5.88561296e-01 9.60532010e-01 -3.99172515e-01
9.89674106e-02 2.13723499e-02 1.07087523e-01 -3.69722545e-01
1.99112773e-01 2.40399480e-01 -3.78516503e-02 8.78281236e-01
2.84180254e-01 4.98423815e-01 1.29124299e-01 -2.74828839e-04
1.05900586e+00 2.59089619e-01 9.44913208e-01 2.23575473e-01
1.08924067e+00 -4.23741072e-01 6.57035947e-01 1.14833200e+00
-7.47955918e-01 8.29117820e-02 -2.19721541e-01 -3.07416558e-01
-5.00636339e-01 -9.58110154e-01 3.62700701e-01 1.17313337e+00
2.57900923e-01 -3.23557824e-01 -5.68055034e-01 -1.29607713e+00
-7.59249032e-02 2.19466984e-01 -7.21317768e-01 -3.65134448e-01
-6.52738273e-01 -8.23022902e-01 8.68638158e-01 1.05850768e+00
1.32665563e+00 -1.04562509e+00 1.39342770e-01 7.06924275e-02
3.13818641e-02 -6.42607331e-01 -5.92242360e-01 -1.91003725e-01
-9.32974339e-01 -1.21150005e+00 -6.83207631e-01 -9.20384288e-01
6.34325147e-01 5.90578914e-01 4.57457513e-01 3.62978637e-01
-2.30846945e-02 7.29274988e-01 -5.31505525e-01 -1.93390593e-01
-2.64870495e-01 -2.41304085e-01 4.04472977e-01 6.24732852e-01
6.47616208e-01 -3.61518264e-01 -4.81675982e-01 8.10081720e-01
-5.08187652e-01 -1.32246912e-01 8.93990636e-01 5.66511452e-01
3.17641050e-01 1.44780174e-01 5.70190310e-01 -6.44013822e-01
5.10340571e-01 -8.90747681e-02 8.78972746e-03 8.12995136e-01
-5.62601507e-01 -1.85597241e-01 8.90195891e-02 -5.97729385e-01
-1.54061985e+00 1.95833310e-01 2.59926170e-01 1.51551023e-01
-6.38034821e-01 -6.55183643e-02 -4.93268907e-01 -1.97779790e-01
5.90447307e-01 4.47670609e-01 1.29911110e-01 -6.71182692e-01
-2.90463287e-02 1.24050975e+00 3.34309548e-01 -2.49775454e-01
3.03350329e-01 2.37831920e-01 -2.34326944e-01 -7.97045529e-01
-7.86225677e-01 -9.76114571e-01 -9.09952343e-01 -5.05564332e-01
1.15484154e+00 -1.16456735e+00 -7.67407775e-01 1.20551038e+00
-8.19473088e-01 -2.71779954e-01 -1.41422693e-02 8.73874307e-01
-5.34672856e-01 9.76985931e-01 -3.96045625e-01 -1.27292466e+00
-1.54188097e-01 -6.16789579e-01 1.00571012e+00 4.43532288e-01
-2.89960742e-01 -1.15983748e+00 2.98288345e-01 9.34598327e-01
-7.77286068e-02 -3.18478048e-01 -1.44945666e-01 -1.25576198e+00
-5.02801955e-01 -4.97057885e-01 -1.03892982e-02 6.55911624e-01
8.25946212e-01 -2.87227362e-01 -6.71094120e-01 -2.22589642e-01
-2.03002870e-01 -3.09310287e-01 6.97247684e-01 8.65438059e-02
1.04662108e+00 -3.07322532e-01 -5.94908237e-01 7.44733028e-03
7.22303510e-01 8.38478684e-01 9.21609581e-01 -3.46376598e-02
9.47282135e-01 3.19596171e-01 9.25322890e-01 5.47606289e-01
-6.75763935e-02 9.84014452e-01 -2.89796829e-01 3.04971784e-02
-4.09929514e-01 -1.15944572e-01 7.09738970e-01 7.98515439e-01
-5.67744613e-01 -3.87592047e-01 -8.20515692e-01 1.46512851e-01
-2.42672586e+00 -1.44142771e+00 -3.64329549e-04 2.17241430e+00
5.35962224e-01 -3.19903046e-01 4.18958485e-01 2.35292181e-01
7.36721814e-01 3.49729866e-01 -2.73863882e-01 2.30933070e-01
9.97441337e-02 -1.00360826e-01 3.72321695e-01 4.04973000e-01
-1.50334883e+00 6.84879184e-01 7.45981455e+00 9.97222364e-01
-1.71926782e-01 1.39204279e-01 3.50423247e-01 1.21571355e-01
4.04205531e-01 -8.78815874e-02 -1.17389882e+00 3.76532674e-01
6.88549757e-01 3.54983173e-02 2.25360528e-01 6.09253526e-01
1.84614062e-01 -3.88158232e-01 -9.60195959e-01 1.20727062e+00
6.69592559e-01 -7.94112086e-01 2.91050643e-01 -1.30031794e-01
9.26733673e-01 -7.93458641e-01 -4.85910654e-01 5.33786297e-01
5.39503336e-01 -8.00465405e-01 -3.70778283e-03 9.88289356e-01
-5.05338237e-02 -5.25583565e-01 8.75730276e-01 3.48508835e-01
-1.39736891e+00 -3.73534799e-01 8.69632214e-02 -4.24458444e-01
3.18548203e-01 1.63452059e-01 -7.85633564e-01 7.37007618e-01
4.09401447e-01 1.32204235e+00 -9.09854293e-01 1.19350386e+00
-1.84721380e-01 7.42762029e-01 -1.80288866e-01 -1.14351116e-01
-1.64077342e-01 -3.02012116e-01 2.56985873e-01 1.02151012e+00
2.41369247e-01 3.54740061e-02 7.31127799e-01 -7.53264758e-04
1.46126330e-01 1.87940046e-01 -4.21418905e-01 -3.18171412e-01
3.36825937e-01 8.53194714e-01 -7.24418759e-01 -8.57629955e-01
-6.33910239e-01 1.02055502e+00 2.47793972e-01 2.05487669e-01
-9.28245068e-01 -2.64069766e-01 3.76229048e-01 -3.22643906e-01
1.05420910e-01 -2.56866366e-01 -1.95674852e-01 -1.10371029e+00
-3.24279785e-01 -7.77657628e-01 8.67385983e-01 -6.19723439e-01
-1.07835102e+00 -1.36300340e-01 4.41104285e-02 -1.44312966e+00
-2.11216867e-01 -3.67572516e-01 -7.40528822e-01 4.48055178e-01
-4.27698314e-01 -1.19545996e+00 -5.36255479e-01 7.89587319e-01
5.37417233e-01 -7.55126059e-01 6.42056584e-01 4.10637677e-01
-9.42288160e-01 6.58911288e-01 -3.23178709e-01 2.98414111e-01
9.03275907e-01 -1.06689703e+00 -1.24516182e-01 1.15302885e+00
2.75020033e-01 6.07618749e-01 1.90995798e-01 -1.14200449e+00
-9.94566619e-01 -1.00535369e+00 7.48248875e-01 -5.37247777e-01
3.87432277e-01 -1.86895087e-01 -7.26806164e-01 9.06926751e-01
6.94213137e-02 -2.98490942e-01 1.11339808e+00 2.29444355e-01
1.18545881e-02 -3.73478681e-01 -8.64081740e-01 5.24752319e-01
1.43607366e+00 -4.86422688e-01 -9.34613109e-01 4.37878042e-01
3.02534759e-01 2.11300164e-01 -8.16476285e-01 4.37978923e-01
5.77252567e-01 -6.53724968e-01 9.68412697e-01 -5.84799349e-01
-4.29535151e-01 -7.79548168e-01 -2.03541577e-01 -8.37396920e-01
-4.68438476e-01 -3.32443804e-01 -6.29303575e-01 1.28865099e+00
5.97552350e-03 -6.72606111e-01 9.43847060e-01 6.00179076e-01
8.61788243e-02 -4.14047614e-02 -6.33844256e-01 -1.15628946e+00
-8.71178269e-01 3.34029994e-03 3.63878667e-01 6.61276579e-01
1.11352138e-01 4.14120853e-01 -9.57123458e-01 6.18018098e-02
6.73783839e-01 -1.47859678e-01 1.11662614e+00 -1.07875538e+00
-7.36798882e-01 -1.72931731e-01 -5.71276903e-01 -1.19267547e+00
-2.31524073e-02 -6.04375422e-01 1.04466751e-01 -1.67322254e+00
3.37361127e-01 -4.44926135e-02 -6.43876135e-01 8.56822789e-01
-4.34685588e-01 2.95067877e-01 -1.37278676e-01 4.06493843e-01
-1.39158237e+00 5.25635183e-01 9.75452900e-01 -1.32848337e-01
-6.76077425e-01 3.93329442e-01 -4.72054631e-01 6.25536025e-01
9.97395039e-01 -1.28513590e-01 -4.51892078e-01 1.19057097e-01
-6.71589255e-01 -1.01797163e-01 1.61342070e-01 -1.73864579e+00
4.46193606e-01 -2.60228395e-01 5.01640737e-01 -5.54551244e-01
3.75658959e-01 -6.50845766e-01 3.42603475e-01 7.08660364e-01
-3.82532567e-01 -6.27681732e-01 -2.25766152e-01 8.44025671e-01
-2.52788305e-01 -1.26534268e-01 4.95526850e-01 -1.11573942e-01
-1.31905770e+00 1.58654302e-01 -6.84114397e-01 -6.68994188e-01
1.48541975e+00 -6.67539179e-01 -2.62695730e-01 -4.88474220e-01
-1.02375340e+00 1.93813130e-01 -1.15065359e-01 3.99693549e-01
4.41771954e-01 -1.54276526e+00 -3.47776234e-01 1.68931544e-01
1.39090151e-01 -5.35208523e-01 4.77496743e-01 1.06714523e+00
-7.33768567e-02 3.80776137e-01 -4.64223564e-01 -3.63107294e-01
-1.75225198e+00 2.30491415e-01 4.35882807e-01 -5.57164371e-01
-4.45699170e-02 5.57595372e-01 -9.99716818e-02 -4.44211423e-01
6.62385523e-01 2.41444290e-01 -9.11349535e-01 1.19385958e-01
9.33120191e-01 9.09397066e-01 -3.12466025e-01 -5.27740240e-01
-5.20223916e-01 5.14114797e-01 -2.05675796e-01 3.50877307e-02
8.86825204e-01 -2.19310150e-01 1.65035278e-02 5.68595409e-01
7.35243618e-01 -1.54016122e-01 -1.24559236e+00 -1.39713228e-01
-3.63321230e-02 -5.79594493e-01 -2.04502255e-01 -8.06411624e-01
-7.98634231e-01 5.04097044e-01 9.97107446e-01 -1.95059314e-01
9.30679321e-01 -1.63014948e-01 4.47596699e-01 7.12455332e-01
5.63778639e-01 -1.44219768e+00 7.32131422e-01 2.86397368e-01
4.94015485e-01 -1.27066600e+00 5.43341875e-01 -2.74154097e-01
-6.70350730e-01 7.06401348e-01 1.03379774e+00 -2.39187315e-01
5.97806334e-01 -2.38138810e-01 -1.64709818e-02 2.08657756e-01
-6.65761232e-01 -3.76531214e-01 5.92487276e-01 9.30878401e-01
2.85204560e-01 1.18267149e-01 -7.38649845e-01 5.74827015e-01
7.37071455e-01 3.26443464e-01 1.61553562e-01 1.11234081e+00
-7.96406209e-01 -1.25341475e+00 -4.81532454e-01 6.80516005e-01
-2.01010212e-01 6.59935236e-01 -5.57227075e-01 5.83244741e-01
2.01845005e-01 1.20707381e+00 8.20862725e-02 -6.91685855e-01
1.83675021e-01 2.62537926e-01 4.82897729e-01 -6.75612271e-01
-5.58269560e-01 -3.43211666e-02 3.80304605e-01 -7.52843201e-01
-1.16142094e+00 -9.13455546e-01 -1.08100522e+00 -6.95803110e-03
-1.10892400e-01 1.39816314e-01 4.20626365e-02 1.25694656e+00
8.13444033e-02 5.19746125e-01 6.17498457e-01 -5.43150127e-01
-3.06252122e-01 -1.19167781e+00 -9.40392375e-01 5.82895219e-01
-2.96419859e-01 -7.49638855e-01 -3.59299660e-01 1.62085861e-01] | [8.163925170898438, 0.6534997224807739] |
f9df5c85-d408-48f8-b663-7662d789f71a | a-survey-of-surface-defect-detection-of | 2203.05733 | null | https://arxiv.org/abs/2203.05733v1 | https://arxiv.org/pdf/2203.05733v1.pdf | A Survey of Surface Defect Detection of Industrial Products Based on A Small Number of Labeled Data | The surface defect detection method based on visual perception has been widely used in industrial quality inspection. Because defect data are not easy to obtain and the annotation of a large number of defect data will waste a lot of manpower and material resources. Therefore, this paper reviews the methods of surface defect detection of industrial products based on a small number of labeled data, and this method is divided into traditional image processing-based industrial product surface defect detection methods and deep learning-based industrial product surface defect detection methods suitable for a small number of labeled data. The traditional image processing-based industrial product surface defect detection methods are divided into statistical methods, spectral methods and model methods. Deep learning-based industrial product surface defect detection methods suitable for a small number of labeled data are divided into based on data augmentation, based on transfer learning, model-based fine-tuning, semi-supervised, weak supervised and unsupervised. | ['Li Chen', 'Qifan Jin'] | 2022-03-11 | null | null | null | null | ['defect-detection'] | ['computer-vision'] | [ 4.99608368e-01 5.33498004e-02 9.70301703e-02 -3.00787151e-01
-2.91265994e-01 -2.13011187e-02 7.18613043e-02 4.34533775e-01
1.63885251e-01 1.90183222e-01 -5.86057007e-01 -1.69333890e-01
-7.27255642e-02 -1.05273068e+00 1.81179624e-02 -6.45690084e-01
-7.62973651e-02 7.09263384e-01 4.02058661e-01 -2.27638736e-01
5.28891087e-01 3.29860866e-01 -1.69220185e+00 4.02575314e-01
9.39430654e-01 1.27590263e+00 7.39531457e-01 3.01498175e-01
-2.90777802e-01 2.78771639e-01 -7.27079749e-01 4.50131893e-01
3.79630715e-01 -4.92976964e-01 -7.77256668e-01 9.55513000e-01
3.52475047e-02 -1.04175337e-01 2.31718332e-01 1.37353146e+00
5.31237483e-01 -5.82037084e-02 9.66803670e-01 -1.12706208e+00
-8.35282624e-01 -3.08555484e-01 -7.99882650e-01 8.02234709e-02
-5.96788479e-03 1.55913755e-01 4.59999770e-01 -1.07575476e+00
3.90571892e-01 1.40308928e+00 7.27925777e-01 2.57157177e-01
-1.01203096e+00 -1.52145594e-01 -2.30938420e-01 2.30637416e-01
-1.02386963e+00 3.15052360e-01 1.10477746e+00 -7.52904236e-01
9.00685728e-01 -1.26646593e-01 6.81461275e-01 3.81382763e-01
3.11173767e-01 5.73736489e-01 9.54786599e-01 -8.11792850e-01
3.78367305e-01 -1.16889074e-01 1.78837717e-01 1.01149571e+00
4.61963743e-01 4.41915542e-01 -1.82360560e-01 2.56749213e-01
1.28073895e+00 4.75136727e-01 2.57967323e-01 -4.28224176e-01
-9.40906465e-01 1.19481230e+00 2.35076159e-01 4.65146542e-01
-1.47284210e-01 -5.33585250e-01 8.16303492e-01 8.77337396e-01
6.83509648e-01 3.73996824e-01 -9.75107312e-01 2.96831310e-01
-7.69497514e-01 -1.33989781e-01 2.23885983e-01 7.40444362e-01
1.06534469e+00 3.65367353e-01 2.80395597e-01 1.33231521e+00
7.78361738e-01 6.98966980e-01 4.80754763e-01 -8.38807821e-01
1.72593608e-01 1.10922992e+00 -8.77583101e-02 -9.18894112e-01
-6.28954887e-01 2.85071462e-01 -1.11953425e+00 1.16465449e+00
5.76836802e-02 -8.71490501e-03 -1.45843029e+00 5.09040296e-01
2.48818591e-01 -6.62725091e-01 -1.07695259e-01 7.89280891e-01
8.97422135e-01 7.83225060e-01 -1.96958289e-01 -3.11393917e-01
1.50632989e+00 -1.23016369e+00 -7.65304327e-01 -3.26470554e-01
7.68165052e-01 -1.11074221e+00 1.20494318e+00 1.11691713e+00
-8.21493387e-01 -1.16509902e+00 -1.29011559e+00 3.90675813e-02
-5.45982480e-01 4.97465909e-01 7.08585858e-01 7.17407048e-01
-5.27246118e-01 7.62275159e-01 -9.16222632e-01 -5.03330410e-01
6.49040282e-01 5.14847815e-01 -5.25637865e-01 -2.05865413e-01
-6.96053386e-01 7.63709962e-01 4.25221592e-01 1.41481668e-01
-9.90130603e-01 -2.30241984e-01 -1.11515307e+00 -3.62673193e-01
1.75843686e-01 -3.28525543e-01 1.13649786e+00 -1.01295900e+00
-1.59103549e+00 1.12432480e+00 2.67062932e-01 2.24328600e-02
7.78850019e-02 -7.02181412e-03 -4.98962313e-01 2.12110102e-01
1.99958339e-01 3.59144926e-01 1.12792754e+00 -1.49968874e+00
-8.87256563e-01 -7.03256071e-01 -4.43429261e-01 -2.80072272e-01
-1.77648515e-01 -5.80321029e-02 1.28172457e-01 -5.42064309e-01
7.25110531e-01 -5.29604733e-01 -2.63539433e-01 6.38149828e-02
-3.31706613e-01 -7.42245317e-01 1.33373940e+00 -6.89460695e-01
4.92438883e-01 -2.15784335e+00 -3.18879426e-01 2.12277666e-01
4.45410199e-02 3.88300657e-01 -2.51377523e-01 4.68919039e-01
-2.15378106e-01 -4.63440828e-02 -3.75971317e-01 -8.34355056e-02
-2.12490037e-01 1.21522307e-01 4.23309743e-01 5.31971276e-01
3.18965733e-01 3.07839632e-01 -8.38332057e-01 -5.34650505e-01
6.62799478e-01 -1.44230679e-01 -2.80800253e-01 3.20632935e-01
5.00956699e-02 4.76934940e-01 -4.86331940e-01 1.21952605e+00
9.72115099e-01 3.91669795e-02 -2.22575843e-01 -4.44217563e-01
-1.19308896e-01 -6.61779121e-02 -1.17077076e+00 1.87410891e+00
-6.09401107e-01 2.74659634e-01 2.98806936e-01 -1.49565899e+00
1.61382043e+00 4.32294875e-01 6.64716661e-01 -6.03924572e-01
6.57105744e-02 4.47692096e-01 -2.54902601e-01 -9.82913554e-01
-3.90853621e-02 -6.11706614e-01 1.74625129e-01 1.91473901e-01
2.49786362e-01 -7.30748177e-01 3.00924540e-01 -4.32343721e-01
9.00200069e-01 5.31135760e-02 -1.77963544e-02 -1.42127678e-01
4.45879877e-01 2.67233223e-01 6.88458979e-01 2.58718014e-01
-3.43392849e-01 8.88023198e-01 1.38987958e-01 -7.55677104e-01
-1.28676474e+00 -1.06280982e+00 -2.16108412e-01 6.49770975e-01
4.21330005e-01 1.08013883e-01 -4.17551130e-01 -9.09169495e-01
2.05255657e-01 -4.55308929e-02 -4.27870095e-01 -4.49507475e-01
-3.04574333e-02 -7.61257708e-01 -1.87401921e-01 4.36237067e-01
8.47809076e-01 -1.73938537e+00 -2.23941252e-01 5.50218150e-02
3.00861955e-01 -4.06346947e-01 -4.74831201e-02 3.01071495e-01
-1.36103046e+00 -1.36592019e+00 -5.78883767e-01 -1.58554125e+00
1.01350439e+00 3.53829384e-01 9.57482219e-01 3.54835093e-01
-8.37607324e-01 1.34255618e-01 -5.58204472e-01 -8.03214490e-01
-5.19695044e-01 -4.21157002e-01 1.11018248e-01 -2.51204967e-01
6.85320199e-01 -3.12140942e-01 -6.00571036e-01 3.72102261e-01
-7.52778769e-01 -3.95725936e-01 8.76996040e-01 1.27198160e+00
5.73691666e-01 8.77817988e-01 7.52554834e-01 -9.54801261e-01
4.16481405e-01 -4.82819714e-02 -4.62942094e-01 -9.04805064e-02
-9.71578538e-01 -3.40995222e-01 1.95133597e-01 -1.26929343e-01
-1.30235040e+00 7.01475590e-02 -4.75767702e-02 -2.41113529e-01
-6.05304897e-01 4.51721996e-01 -2.01218590e-01 1.07181124e-01
9.70318079e-01 -1.47566348e-01 4.99411106e-01 -9.57642615e-01
-7.91199878e-02 1.08938789e+00 8.93262103e-02 4.29019034e-02
7.14440465e-01 4.42488700e-01 -1.70663372e-01 -9.54107642e-01
-5.20425916e-01 -9.43389237e-01 -1.11306310e+00 -2.74816990e-01
1.00754499e+00 -5.61895669e-01 -5.16361296e-01 9.44775224e-01
-1.11194718e+00 -3.81233305e-01 -8.06062996e-01 6.55374527e-01
-6.32198870e-01 8.46485555e-01 -8.92837107e-01 -8.79054010e-01
-4.40966368e-01 -1.11361325e+00 1.36934495e+00 -1.82902351e-01
1.84586331e-01 -1.13850236e+00 7.01565295e-02 6.78695023e-01
-7.05861002e-02 8.50865245e-02 1.17024541e+00 -3.25852334e-01
-3.58666033e-02 -4.25055027e-01 -3.22997212e-01 1.13825977e+00
8.71278226e-01 -1.27785817e-01 -6.78540051e-01 -3.40846032e-01
3.24157923e-01 -5.06637514e-01 1.01139879e+00 6.43844664e-01
9.61772323e-01 2.92080790e-01 -3.58241707e-01 -1.23687927e-02
1.40979111e+00 6.33589685e-01 7.20435739e-01 4.16425824e-01
7.00186014e-01 8.05883706e-01 1.17954135e+00 3.13233972e-01
-2.87923396e-01 1.61278933e-01 7.33245194e-01 -7.19968259e-01
-1.46580026e-01 9.11559239e-02 2.59414196e-01 1.06611323e+00
-3.21652770e-01 2.06314340e-01 -5.95778406e-01 7.28347123e-01
-1.51364183e+00 -6.31663918e-01 -5.29854536e-01 1.93876100e+00
6.57711089e-01 3.33084255e-01 2.31281996e-01 8.77679586e-01
7.34860420e-01 -2.88023531e-01 -4.64911044e-01 -5.16841054e-01
2.82955438e-01 2.03502059e-01 3.94746751e-01 1.07997164e-01
-1.31642544e+00 6.64759219e-01 6.57288265e+00 9.35185432e-01
-9.05803442e-01 2.69842863e-01 3.14568698e-01 6.70862556e-01
1.72754526e-01 -2.53858358e-01 -2.14675054e-01 3.22541386e-01
2.03484342e-01 7.03782082e-01 3.78510989e-02 1.08828723e+00
3.57988685e-01 -4.30184722e-01 -9.53517377e-01 1.24942100e+00
-1.65670291e-02 -1.03669500e+00 -3.66229303e-02 1.20044366e-01
8.32873225e-01 -1.50268331e-01 -1.52074516e-01 2.71258920e-01
1.56103879e-01 -7.72261500e-01 3.18497628e-01 2.34643221e-01
6.04715407e-01 -3.82425010e-01 1.24153447e+00 1.04130376e-02
-9.44826603e-01 -1.87584937e-01 -8.91749740e-01 -3.34795505e-01
4.46478352e-02 1.15067184e+00 -5.89952528e-01 6.08987987e-01
1.06945908e+00 9.21390593e-01 -3.20334166e-01 9.50652421e-01
-1.64294988e-01 6.20201349e-01 1.30823642e-01 3.16984206e-01
-7.99004361e-02 -5.35512269e-01 -8.96532030e-04 7.26610661e-01
4.55670625e-01 -5.26010275e-01 7.50678480e-01 9.06924665e-01
5.42689145e-01 5.11037186e-02 -7.06673503e-01 -6.01759239e-04
-3.19712684e-02 1.31749570e+00 -8.81570160e-01 -1.64437130e-01
-6.37279570e-01 1.01020062e+00 -5.01040161e-01 9.69024673e-02
-1.78841241e-02 -7.41959453e-01 1.59897968e-01 2.35407427e-01
2.80179799e-01 -3.93178642e-01 -6.76369250e-01 -5.31826317e-01
-2.94156253e-01 -7.23790586e-01 4.00538892e-01 -8.09706271e-01
-1.87154591e+00 1.91647172e-01 -2.20948458e-01 -1.50532019e+00
3.13504487e-01 -1.34397399e+00 -7.07341671e-01 6.97854340e-01
-1.23721921e+00 -1.23144782e+00 -3.26755345e-01 3.17110687e-01
1.05635011e+00 -6.45778239e-01 7.06682980e-01 2.42724955e-01
-3.65373999e-01 -2.06666306e-01 6.01809137e-02 3.61963571e-03
5.62467039e-01 -1.46390104e+00 -1.41913258e-02 2.70424783e-01
-1.60916165e-01 -7.13870153e-02 3.49138260e-01 -7.89283395e-01
-9.82364416e-01 -1.04162049e+00 5.03010392e-01 -2.04146095e-02
2.41618142e-01 1.06081255e-01 -1.01304293e+00 3.48015487e-01
1.38365999e-01 1.83864638e-01 4.54627514e-01 1.54156417e-01
1.66463852e-01 -3.68998975e-01 -1.51130509e+00 -1.41047135e-01
8.09881747e-01 -2.82570988e-01 -8.52620125e-01 8.18243146e-01
1.74295783e-01 4.16502059e-02 -1.01060927e+00 6.88050449e-01
-1.76416021e-02 -8.28730524e-01 8.41326654e-01 -1.95654929e-01
2.77722001e-01 -6.56633914e-01 3.51057887e-01 -1.39319038e+00
-5.53789377e-01 1.32464200e-01 6.03612959e-01 1.05579543e+00
2.47910812e-01 -6.20757341e-01 9.16675806e-01 -2.93783486e-01
-8.21808040e-01 -5.15880883e-01 -6.55289292e-01 -9.74366009e-01
-1.54566050e-01 -9.75564867e-02 4.47645225e-02 9.32050765e-01
2.01011151e-01 3.42038989e-01 -4.20927629e-02 1.12870462e-01
4.09791529e-01 2.20673293e-01 4.22887683e-01 -1.91497123e+00
9.95118916e-02 -1.63005665e-01 -9.09780562e-01 -7.07171321e-01
-3.79882216e-01 -6.33837461e-01 4.04849708e-01 -2.12717819e+00
3.42761055e-02 -5.62318683e-01 -2.33691022e-01 7.55888283e-01
1.93667144e-01 6.09795570e-01 -4.61020499e-01 2.26478308e-01
-3.08366001e-01 5.30302465e-01 1.72375965e+00 -8.29264224e-01
-3.27208221e-01 2.93783486e-01 -1.03432149e-01 8.27819586e-01
8.87305677e-01 -1.54169098e-01 -4.35529411e-01 -3.71002071e-02
1.27509966e-01 -2.82553613e-01 1.17055662e-01 -1.00453460e+00
-3.25982869e-01 -1.04579039e-01 4.00331318e-01 -7.86679089e-01
-2.93138884e-02 -8.69448423e-01 -6.01002336e-01 8.76772881e-01
3.59015733e-01 -3.01954299e-01 2.09778976e-02 4.61317480e-01
-8.55371952e-01 -7.63091445e-01 9.63399410e-01 -5.63684404e-01
-1.07319832e+00 2.73165882e-01 -6.14748538e-01 -4.08112288e-01
1.08453465e+00 -7.22109973e-01 7.88458213e-02 8.70393962e-02
-1.15885818e+00 -8.89521465e-02 1.59801617e-01 4.66948479e-01
7.86793828e-01 -1.42124152e+00 -4.11141366e-01 7.99682081e-01
2.97393322e-01 4.08359021e-01 4.70646083e-01 6.33261263e-01
-9.39061403e-01 2.26403803e-01 -5.41454077e-01 -8.36595654e-01
-1.16307485e+00 8.87926579e-01 2.37896815e-01 -1.71785116e-01
-5.95077336e-01 8.84269714e-01 1.67720452e-01 -6.46333694e-01
-1.45497799e-01 -6.45503640e-01 -3.01518917e-01 -1.76499262e-01
1.14139304e-01 6.79908693e-01 6.68647110e-01 -3.80256534e-01
-1.69693660e-02 9.15729523e-01 2.67854273e-01 5.29797971e-01
1.23232400e+00 -6.13197638e-03 -4.25912440e-01 6.88031495e-01
1.05037200e+00 -3.72695148e-01 -1.21293688e+00 4.78406437e-02
8.05482827e-03 -4.31310177e-01 3.51667792e-01 -8.63454938e-01
-1.32831502e+00 1.30204582e+00 1.40206218e+00 7.62707174e-01
1.27333939e+00 -8.49333778e-03 8.02030921e-01 3.61615241e-01
6.01544380e-01 -1.56158328e+00 6.96395814e-01 1.74551755e-01
1.00315249e+00 -1.92028797e+00 6.33115098e-02 -8.95774305e-01
-4.07958210e-01 1.39140058e+00 7.11251497e-01 -2.68798798e-01
1.19976056e+00 1.28333852e-01 4.41966504e-01 -7.03392923e-01
-5.41342003e-03 -2.27921620e-01 2.31605291e-01 1.20509088e+00
3.85435522e-01 -1.15324251e-01 -2.42750913e-01 2.88584381e-01
2.98858613e-01 -6.12670928e-03 -2.86396616e-03 1.14428711e+00
-1.07354188e+00 -1.47542846e+00 -5.88325918e-01 6.97476625e-01
-2.36154348e-01 2.43565366e-01 -4.41715866e-02 5.29109240e-01
7.67524719e-01 1.38003337e+00 3.58976841e-01 -4.61209774e-01
5.24901807e-01 -2.29172245e-01 7.10477471e-01 -1.21237314e+00
-1.45865381e-01 -4.38588634e-02 -1.74419641e-01 -5.39690666e-02
-6.32186472e-01 -5.43808699e-01 -1.48717737e+00 1.96818545e-01
-6.94657683e-01 -1.29936919e-01 7.43505001e-01 9.62334037e-01
1.23966243e-02 5.74943900e-01 8.68822396e-01 -1.04623199e+00
-2.88783997e-01 -1.60333145e+00 -1.50332046e+00 5.04959524e-01
2.33452857e-01 -1.11253309e+00 -3.57831776e-01 6.66599810e-01] | [7.408583641052246, 1.8989673852920532] |
1d80f23f-d08e-4cc8-af80-5120361508c5 | virtual-screening-of-plant-metabolites | 2005.11254 | null | https://arxiv.org/abs/2005.11254v1 | https://arxiv.org/pdf/2005.11254v1.pdf | Virtual Screening of Plant Metabolites against Main protease, RNA-dependent RNA polymerase and Spike protein of SARS-CoV-2: Therapeutics option of COVID-19 | Covid-19, a serious respiratory complications caused by SARS-CoV-2 has become one of the global threat to human healthcare system. The present study evaluated the possibility of plant originated approved 117 therapeutics against the main protease protein (MPP), RNA-dependent RNA polymerase (RdRp) and spike protein (S) of SARS-CoV-2 including drug surface analysis by using molecular docking through drug repurposing approaches. The molecular interaction study revealed that Rifampin (-16.3 kcal/mol) were topmost inhibitor of MPP where Azobechalcone were found most potent plant therapeutics for blocking the RdRp (-15.9 kcal /mol) and S (-14.4 kcal/mol) protein of SARS-CoV-2. After the comparative analysis of all docking results, Azobechalcone, Rifampin, Isolophirachalcone, Tetrandrine and Fangchinoline were exhibited as the most potential inhibitory plant compounds for targeting the key proteins of SARS-CoV-2. However, amino acid positions; H41, C145, and M165 of MPP played crucial roles for the drug surface interaction where F368, L371, L372, A375, W509, L514, Y515 were pivotal for RdRP. In addition, the drug interaction surface of S proteins also showed similar patterns with all of its maximum inhibitors. ADME analysis also strengthened the possibility of screened plant therapeutics as the potent drug candidates against SARS-C with the highest drug friendliness. | ['Mahmudul Hasan', 'Farhana Rumzum Bhuiyan', 'Sabbir Howlader', 'Tasfia Saiyara Shammi', 'Aklima Begum', 'Topu Raihan', 'Abdus Shukur Imran', 'Kazi Faizul Azim', 'Md Sorwer Alam Parvez'] | 2020-05-22 | null | null | null | null | ['molecular-docking'] | ['medical'] | [ 1.29046381e-01 -2.95212537e-01 -1.52283132e-01 3.14410180e-01
-2.04209968e-01 -8.43307674e-01 3.61892246e-02 4.85226661e-01
-1.04071416e-01 1.18621600e+00 -2.90772449e-02 -6.01307273e-01
-9.77828056e-02 -3.64372224e-01 -3.88095737e-01 -1.07957780e+00
-1.52760684e-01 6.25433251e-02 -7.39154220e-02 -3.16975892e-01
1.98440105e-01 8.35772574e-01 -5.25286257e-01 1.05966553e-01
1.38774049e+00 2.31328189e-01 4.46456641e-01 4.67589110e-01
3.16622704e-01 1.75551802e-01 -4.07712668e-01 9.94587988e-02
-6.92613646e-02 -3.93162429e-01 -3.76849860e-01 -5.64134538e-01
-1.97570607e-01 -2.95863062e-01 7.07500160e-01 7.59399831e-01
3.73945683e-01 -2.54323423e-01 1.02103961e+00 -1.00050676e+00
-6.55577302e-01 -3.59571069e-01 -9.25360978e-01 -9.74133015e-02
3.99998516e-01 2.55144507e-01 5.95766366e-01 -1.18578029e+00
7.29011893e-01 1.04412544e+00 6.28190994e-01 4.96735930e-01
-9.08687711e-01 -5.81500947e-01 -4.16034758e-01 2.59737372e-01
-1.37129724e+00 3.94685417e-02 2.33880356e-01 -7.03870893e-01
1.40344489e+00 3.42298508e-01 4.73123491e-01 3.88661057e-01
9.35468435e-01 1.92389354e-01 1.00877714e+00 1.43998712e-01
1.08928703e-01 9.89966374e-03 1.46145552e-01 5.30779600e-01
7.32487619e-01 -3.14877957e-01 1.80432498e-01 -9.81035292e-01
4.53185767e-01 3.98077041e-01 -6.63897872e-01 -2.43319243e-01
-4.71219391e-01 8.57559323e-01 2.92761236e-01 2.55365849e-01
-5.66948593e-01 -4.37791556e-01 4.04633760e-01 -3.01888883e-01
-8.43662098e-02 1.50617912e-01 -1.17289889e+00 2.47291535e-01
2.08120309e-02 2.18507066e-01 3.80603880e-01 1.18438132e-01
3.65364552e-01 -1.43070603e-02 -3.14371660e-02 3.57532412e-01
8.20031643e-01 8.01266730e-01 -2.48951197e-01 -1.35964856e-01
3.96821424e-02 7.26603687e-01 6.23771012e-01 -8.47119570e-01
-5.69547117e-01 -4.56528477e-02 -6.86130464e-01 -9.07410681e-02
5.29335082e-01 -4.11567628e-01 -7.51842618e-01 1.87472045e+00
4.65146840e-01 1.04829483e-01 2.69049793e-01 7.85893798e-01
5.43978751e-01 1.11448002e+00 6.77183390e-01 -1.15025175e+00
1.91370475e+00 -3.60599935e-01 -4.52743351e-01 4.55389082e-01
3.78023475e-01 -9.13044035e-01 6.16823971e-01 1.56557292e-01
-5.76891422e-01 -1.40131071e-01 -9.01493967e-01 5.53359330e-01
-1.35501295e-01 1.32582650e-01 2.66553551e-01 7.76167393e-01
-5.02613962e-01 6.77377641e-01 -9.11470115e-01 -5.45517981e-01
1.22589983e-01 5.57968140e-01 -3.42033863e-01 1.41458154e-01
-1.08716500e+00 1.09964252e+00 1.55369446e-01 1.11573763e-01
-7.76544392e-01 -7.50478923e-01 -3.41439039e-01 -6.12849854e-02
-1.05231546e-01 -7.73816705e-01 6.11975968e-01 -2.87472874e-01
-1.28860986e+00 4.09372717e-01 -3.29326242e-01 -3.05001140e-01
-3.97910804e-01 -4.40576434e-01 -4.04200226e-01 3.72970641e-01
-3.68238315e-02 1.37296677e-01 6.51249737e-02 -1.29210949e+00
-6.95329458e-02 -8.87022793e-01 -3.72068286e-01 2.45028928e-01
4.46518689e-01 6.17953002e-01 5.82393587e-01 -5.04550159e-01
-1.54650956e-01 -1.01322925e+00 -4.20337439e-01 -5.71601272e-01
-5.57809830e-01 -2.52888680e-01 9.29122269e-01 -6.48154438e-01
7.96298265e-01 -1.71282578e+00 -8.33593830e-02 4.08491194e-01
-2.35173535e-02 9.38268185e-01 1.20193198e-01 1.25001609e+00
-1.32109061e-01 5.17781079e-01 -1.42982200e-01 1.05447781e+00
-6.11895800e-01 -4.54814285e-01 -2.82037050e-01 9.47167277e-01
1.48599222e-01 6.98218644e-01 -7.66353488e-01 -3.93549204e-02
-5.21845035e-02 8.75656128e-01 -3.61913294e-01 -4.62370254e-02
-5.34877837e-01 8.60993683e-01 -1.17420697e+00 8.04756641e-01
1.20536852e+00 -1.44578621e-01 6.17599607e-01 -2.89214700e-01
-2.79506892e-01 1.09801538e-01 -5.28868556e-01 8.67866397e-01
4.89148140e-01 -1.68764964e-01 -9.54843536e-02 -6.64809868e-02
6.96029365e-01 6.99645638e-01 6.69773817e-01 -3.45450163e-01
1.93501666e-01 2.40339726e-01 1.53053254e-01 -4.17221576e-01
-1.93687439e-01 -1.96643740e-01 5.71178079e-01 -1.92373633e-01
-5.89870036e-01 4.95223820e-01 -3.44570316e-02 -4.71684374e-02
7.76639581e-01 2.39872068e-01 6.50752485e-01 -6.74185216e-01
9.86161411e-01 2.59979635e-01 8.41098428e-01 -7.89819658e-02
-2.01975971e-01 -6.03699386e-02 5.80712616e-01 -1.57169357e-01
-8.26643229e-01 -9.57594335e-01 -5.62377512e-01 4.20071304e-01
1.39389604e-01 -3.82272363e-01 -8.82404864e-01 -2.68424511e-01
-3.02173227e-01 7.23236799e-01 -9.98487696e-02 2.56997257e-01
-4.77728844e-01 -6.67595029e-01 5.85092843e-01 -1.84788227e-01
4.56471384e-01 -7.19688773e-01 -4.86838907e-01 1.60241947e-01
1.66733101e-01 -4.96645182e-01 -3.77381504e-01 8.58114883e-02
-9.78277326e-01 -1.55760741e+00 -5.22711396e-01 -6.09204412e-01
3.56417328e-01 2.43954182e-01 1.48144051e-01 1.59756169e-01
-2.17158660e-01 -3.64603311e-01 -2.18617901e-01 -7.84267724e-01
-2.44496197e-01 -5.94868481e-01 4.08701807e-01 -5.06626904e-01
5.29489875e-01 -6.34630680e-01 -9.64511991e-01 3.04709882e-01
-2.30027869e-01 -2.50335425e-01 3.03633422e-01 3.60544831e-01
5.23169339e-01 -3.73917103e-01 7.46822298e-01 -8.52761567e-01
6.23506606e-01 -7.53952086e-01 -5.04391313e-01 2.64817566e-01
-5.18024325e-01 -3.99446309e-01 8.83255661e-01 8.05518776e-02
-1.05470216e+00 -8.53906423e-02 -3.35849710e-02 1.96940631e-01
-1.41332388e-01 2.43562952e-01 -5.92755079e-01 1.88288048e-01
4.86593187e-01 1.27824694e-01 9.36872661e-02 -5.85393548e-01
-3.49033177e-01 2.96616018e-01 9.67484899e-03 -3.72212082e-01
7.27691650e-01 3.22656900e-01 2.79203624e-01 -1.04052246e+00
-1.14346206e-01 -4.83053327e-01 1.86349005e-02 2.93890148e-01
1.54843736e+00 -9.53786969e-01 -1.48420191e+00 5.64413428e-01
-1.44168270e+00 3.58990073e-01 7.88951695e-01 7.13683128e-01
-8.08782727e-02 5.78383088e-01 -8.05839062e-01 -7.90751338e-01
-8.83519888e-01 -1.04761386e+00 4.44860190e-01 3.98206264e-01
-1.02586582e-01 -5.18878698e-01 7.42536783e-01 4.07511115e-01
1.30889386e-01 7.64680386e-01 1.70576215e+00 -1.14497507e+00
-6.11080587e-01 3.41117501e-01 -2.74722070e-01 -3.68232206e-02
2.56814301e-01 5.14809668e-01 -6.12141013e-01 -1.93163976e-01
-1.81222633e-02 -6.91113546e-02 1.25976205e-01 6.55685902e-01
1.87558234e-01 -4.11826849e-01 -6.27990007e-01 4.83817220e-01
1.77600658e+00 1.31769347e+00 9.98960435e-01 2.22092316e-01
5.21017730e-01 1.27122268e-01 9.24259901e-01 5.98988414e-01
-1.48656085e-01 6.90124571e-01 7.10964322e-01 2.18470134e-02
6.03490412e-01 3.56775932e-02 5.32900214e-01 9.37440321e-02
-8.40398729e-01 -3.58692497e-01 -7.90183008e-01 3.35961692e-02
-1.31219053e+00 -9.27091718e-01 -8.92478108e-01 2.15424538e+00
9.49074090e-01 -5.90253651e-01 1.68970749e-02 -4.83776420e-01
4.25276965e-01 -2.59126157e-01 -6.43714011e-01 -5.78152835e-01
-3.24402638e-02 4.02063817e-01 5.49854636e-01 3.11889797e-01
-6.66634738e-01 5.86550057e-01 4.90337753e+00 5.29465079e-01
-6.01529419e-01 -4.28603292e-01 1.98548183e-01 4.44078058e-01
-2.44475901e-01 6.34130359e-01 -6.85224175e-01 3.29093039e-01
8.43219042e-01 4.29525552e-03 2.44653262e-02 6.29606068e-01
9.08587039e-01 -3.58910352e-01 -6.34102404e-01 5.69038749e-01
-4.54843223e-01 -1.40237677e+00 7.04602003e-02 5.02894759e-01
6.77857459e-01 -1.00760378e-01 -3.40634108e-01 -3.27997178e-01
1.00137532e-01 -1.04962945e+00 -3.26251924e-01 1.55299827e-01
4.12507474e-01 -1.15376353e+00 6.55989051e-01 3.04028869e-01
-1.17360806e+00 2.09808111e-01 -4.40223306e-01 3.52472246e-01
2.29894772e-01 3.03958714e-01 -1.01949060e+00 4.81947064e-01
3.57317358e-01 2.01142654e-01 -1.67881370e-01 6.73268080e-01
-3.20975572e-01 3.14021409e-01 4.81283292e-02 -3.75649966e-02
-5.63237164e-03 -7.10886598e-01 5.66201031e-01 8.00176382e-01
3.18296166e-04 8.14377785e-01 2.87505388e-01 3.17007720e-01
1.70325056e-01 5.95076084e-01 -5.52628577e-01 -1.82263345e-01
4.64050680e-01 8.90200675e-01 -5.63487232e-01 -8.89216736e-02
-2.81358361e-01 5.80438375e-01 -2.88277239e-01 5.71139932e-01
-8.84971619e-01 -1.91971645e-01 1.40619814e+00 2.15074182e-01
2.11134076e-01 -6.50309846e-02 2.43431494e-01 -4.55464602e-01
-6.91691935e-01 -1.16424406e+00 3.28870565e-01 -6.77761674e-01
-7.42981791e-01 2.11624593e-01 1.41323462e-01 -9.39152658e-01
2.86621213e-01 -4.90600973e-01 -7.96337485e-01 1.25966930e+00
-9.71126020e-01 -1.00310802e+00 2.27041036e-01 5.16308248e-01
3.00489724e-01 -6.27297312e-02 1.27214253e+00 -2.36553386e-01
-6.61750436e-01 -1.76637739e-01 5.47261477e-01 -3.50651145e-01
6.05559647e-01 -6.54292047e-01 -2.14834418e-02 3.32450449e-01
-7.35003889e-01 1.31989408e+00 8.94092202e-01 -1.25075793e+00
-1.52250242e+00 -8.31076324e-01 9.22362506e-01 -1.49570554e-01
4.04797316e-01 2.52538413e-01 -6.97600842e-01 4.87813979e-01
6.23749018e-01 -9.36500013e-01 1.30111587e+00 -4.74767148e-01
-2.15677425e-01 6.74643457e-01 -1.49096966e+00 9.68004405e-01
2.97371566e-01 -2.37808511e-01 -2.84567237e-01 5.94794869e-01
7.90417254e-01 -1.58500955e-01 -1.05038702e+00 6.60894632e-01
6.79824948e-01 -9.60860431e-01 1.00026786e+00 -7.85099447e-01
-5.47537804e-02 -9.83127713e-01 -1.50338143e-01 -8.55604172e-01
-4.07259911e-01 -5.75058281e-01 2.73311108e-01 9.20249820e-01
2.72642910e-01 -5.63260734e-01 3.79796863e-01 1.70461684e-01
1.71717182e-01 -7.20486045e-01 -6.96411312e-01 -4.75345463e-01
-4.11071852e-02 4.06480104e-01 1.68873370e-01 7.39087880e-01
-5.51274717e-02 8.25848401e-01 -4.56906915e-01 2.89607942e-01
3.68438542e-01 6.02972284e-02 6.13621235e-01 -1.18959379e+00
-1.28311068e-01 1.32854044e-01 4.07320559e-02 -3.85761052e-01
-5.19924819e-01 -5.17188728e-01 -9.60529923e-01 -1.74943936e+00
5.05870759e-01 1.89375460e-01 -4.73927930e-02 2.55747646e-01
2.09815465e-02 -2.75023609e-01 2.91996717e-01 3.36929142e-01
3.11214060e-01 1.55955657e-01 1.31419158e+00 1.31814629e-01
-6.91128850e-01 1.64142385e-01 -6.57339096e-01 8.63999248e-01
9.75505710e-01 -5.22630692e-01 -8.91239107e-01 3.49189490e-01
2.38738686e-01 1.88900396e-01 -1.25643194e-01 1.01856247e-01
-4.60514784e-01 -9.51006472e-01 5.20230234e-01 -1.04556108e+00
4.36439931e-01 -8.13250959e-01 8.18001330e-01 1.23651183e+00
7.01719522e-01 1.24818012e-01 4.38770920e-01 4.12311614e-01
6.02167130e-01 -1.25429541e-01 1.02664721e+00 -1.14616655e-01
-1.81028098e-01 9.82885137e-02 -9.42758739e-01 -7.95926750e-02
1.29655969e+00 -8.23631167e-01 -4.36940283e-01 3.01915437e-01
-7.22372532e-01 -8.07136521e-02 6.69312060e-01 1.76926423e-02
5.34908712e-01 -6.53517246e-01 -7.22289681e-01 -8.35997984e-02
4.08524238e-02 -4.84876007e-01 5.07575393e-01 1.14622867e+00
-1.16866815e+00 7.54500508e-01 -5.00844121e-01 -4.94839907e-01
-1.75634241e+00 9.46009874e-01 2.96750098e-01 -2.44936034e-01
-1.36745572e-01 4.91944671e-01 5.45826197e-01 1.96003512e-01
5.49495779e-03 -2.46661052e-01 -7.12724686e-01 -8.46568942e-02
3.39834273e-01 6.62519872e-01 -4.39876579e-02 -8.07947516e-01
-1.19544065e+00 7.35421658e-01 -1.96220636e-01 9.06427622e-01
1.32311583e+00 1.11300953e-01 -4.96120125e-01 -1.03090003e-01
9.44603980e-01 5.67305386e-01 -6.40541852e-01 3.15535307e-01
-2.90619582e-02 -3.29241663e-01 -7.61287034e-01 -1.06773341e+00
-2.23399878e-01 1.70547843e-01 7.50219822e-01 -4.83170182e-01
9.38835323e-01 -1.99822739e-01 7.37437248e-01 1.28564239e-01
6.72340393e-01 -5.79225183e-01 -2.54077286e-01 2.43548498e-01
8.17402005e-01 -6.67293191e-01 1.14222929e-01 -6.37172282e-01
-7.61082947e-01 8.02089930e-01 6.09832108e-01 -1.17487451e-02
4.36085403e-01 -1.12652592e-01 -5.71318762e-03 -4.38191026e-01
-9.93616819e-01 3.38354796e-01 2.74607778e-01 6.57322168e-01
9.87509906e-01 1.43941924e-01 -1.33706212e+00 5.10871053e-01
4.25894797e-01 -1.70910284e-01 4.29133087e-01 7.31049955e-01
-7.90745795e-01 -1.42895174e+00 -6.03905499e-01 -1.38138458e-01
-6.47888720e-01 -2.86927253e-01 -6.81271493e-01 9.20919359e-01
1.96023807e-01 1.02162743e+00 -5.05832791e-01 7.32532665e-02
1.97060093e-01 2.04677716e-01 3.69753152e-01 -4.82319683e-01
-8.40375006e-01 9.12436903e-01 5.80019876e-02 -1.03768364e-01
-4.08152401e-01 -2.85634488e-01 -1.98514056e+00 -5.28313279e-01
-5.20766854e-01 5.59879720e-01 6.20159924e-01 8.45996439e-01
4.92092192e-01 -1.59669578e-01 8.15901577e-01 2.63583809e-02
-4.62595046e-01 -8.08489561e-01 -8.08861732e-01 -1.09421775e-01
1.79370701e-01 -2.56502539e-01 1.23034157e-01 1.08775571e-01] | [4.665122032165527, 5.097651958465576] |
12945276-aebe-4aa4-a325-60c35c937c5c | dynaeval-unifying-turn-and-dialogue-level | 2106.01112 | null | https://arxiv.org/abs/2106.01112v3 | https://arxiv.org/pdf/2106.01112v3.pdf | DynaEval: Unifying Turn and Dialogue Level Evaluation | A dialogue is essentially a multi-turn interaction among interlocutors. Effective evaluation metrics should reflect the dynamics of such interaction. Existing automatic metrics are focused very much on the turn-level quality, while ignoring such dynamics. To this end, we propose DynaEval, a unified automatic evaluation framework which is not only capable of performing turn-level evaluation, but also holistically considers the quality of the entire dialogue. In DynaEval, the graph convolutional network (GCN) is adopted to model a dialogue in totality, where the graph nodes denote each individual utterance and the edges represent the dependency between pairs of utterances. A contrastive loss is then applied to distinguish well-formed dialogues from carefully constructed negative samples. Experiments show that DynaEval significantly outperforms the state-of-the-art dialogue coherence model, and correlates strongly with human judgements across multiple dialogue evaluation aspects at both turn and dialogue level. | ['Haizhou Li', 'Grandee Lee', 'Thomas Friedrichs', 'Yan Zhang', "Luis Fernando D'Haro", 'Yiming Chen', 'Chen Zhang'] | 2021-06-02 | null | https://aclanthology.org/2021.acl-long.441 | https://aclanthology.org/2021.acl-long.441.pdf | acl-2021-5 | ['dialogue-evaluation'] | ['natural-language-processing'] | [-3.85759205e-01 4.50607061e-01 6.37905672e-02 -6.73822165e-01
-5.47447026e-01 -6.83164239e-01 9.99234617e-01 2.23008201e-01
-1.10462740e-01 5.56735933e-01 6.70873761e-01 -6.53975382e-02
2.00282544e-01 -7.20259488e-01 8.81924778e-02 -4.41870570e-01
1.03802420e-02 7.22013652e-01 -4.74616028e-02 -1.03331995e+00
9.98236537e-02 -8.86360705e-02 -6.93944693e-01 3.54997396e-01
1.11863148e+00 8.32951367e-01 -3.21180642e-01 1.10146761e+00
-2.51379848e-01 1.23979330e+00 -8.84518683e-01 -9.68092561e-01
-4.18268174e-01 -1.04747283e+00 -1.36096680e+00 2.06903413e-01
2.07497463e-01 -3.41717184e-01 -4.61456627e-01 9.57088649e-01
5.90340972e-01 2.18316495e-01 4.98475760e-01 -1.19150054e+00
-3.59114170e-01 9.96793330e-01 -3.84695865e-02 1.70573755e-03
8.77615988e-01 4.44811374e-01 1.64232743e+00 -6.45509303e-01
6.16009295e-01 1.67118621e+00 5.56067944e-01 4.21269894e-01
-1.15890646e+00 -1.93994150e-01 1.17725274e-02 4.56355773e-02
-6.01190031e-01 -4.67130512e-01 9.55403805e-01 -3.76874298e-01
9.75810826e-01 3.29418749e-01 7.29358315e-01 1.05710804e+00
1.69739351e-01 1.13545918e+00 9.24531698e-01 -2.47414082e-01
-2.30307266e-01 -6.29010424e-02 6.98597908e-01 6.85886919e-01
-7.37537265e-01 -3.28851521e-01 -5.99374831e-01 -5.80235571e-02
2.47143373e-01 -6.99067175e-01 -4.01498854e-01 -1.26230881e-01
-1.13896608e+00 1.05236030e+00 4.77937013e-01 5.69721282e-01
-2.25419030e-01 -2.77839631e-01 9.08234060e-01 7.40002871e-01
5.99305332e-01 6.70328438e-01 -2.50586241e-01 -7.08886743e-01
-4.37135667e-01 3.76543432e-01 1.48096049e+00 6.56596243e-01
6.40931964e-01 -1.74284250e-01 -7.71848559e-01 1.16369641e+00
7.22151324e-02 5.69240050e-03 3.27589512e-01 -1.07889676e+00
6.01167262e-01 1.03397691e+00 -1.03295028e-01 -1.17518628e+00
-7.69175291e-01 -7.88725242e-02 -9.35877919e-01 -3.69180739e-01
5.54345667e-01 -4.33167249e-01 2.48338487e-02 1.83850014e+00
3.93666208e-01 -6.48520172e-01 1.60855860e-01 8.18062544e-01
1.56611621e+00 5.64960539e-01 -3.54776531e-01 -4.46243823e-01
1.17166674e+00 -1.42633843e+00 -1.27458835e+00 1.63989469e-01
8.65088940e-01 -7.82056570e-01 1.24073172e+00 3.02889287e-01
-1.25622344e+00 -6.02600694e-01 -8.49436104e-01 -1.62999362e-01
-1.46784976e-01 -6.32047206e-02 5.27097344e-01 4.83099401e-01
-1.20948458e+00 5.85965574e-01 -2.98868537e-01 -8.10526088e-02
-2.84685403e-01 2.47825801e-01 -3.41897160e-01 2.32352018e-01
-1.57451868e+00 1.17238700e+00 1.91070125e-01 4.05858517e-01
-5.38243234e-01 -9.56854671e-02 -1.02387869e+00 6.54896498e-02
2.99282104e-01 -4.10427541e-01 1.75592089e+00 -8.27602565e-01
-2.17327952e+00 8.63719463e-01 4.87212762e-02 -2.59930968e-01
5.90310872e-01 -1.52988851e-01 -3.74620348e-01 -8.53213072e-02
-1.38398841e-01 2.97263741e-01 3.06885272e-01 -9.92525160e-01
-2.71270424e-01 -1.33798227e-01 6.00861609e-01 5.86830258e-01
-1.95226341e-01 -9.31483880e-02 -5.55926442e-01 -5.28537072e-02
-1.39336184e-01 -9.40856755e-01 6.67911693e-02 -6.62332714e-01
-8.78541052e-01 -8.66494238e-01 6.52413785e-01 -6.41130567e-01
1.64664841e+00 -1.81745660e+00 5.32971621e-01 -1.79279655e-01
6.39954686e-01 3.39065194e-01 -3.75756830e-01 8.99233937e-01
2.84870833e-01 -5.47472015e-02 -5.23554645e-02 -5.29351652e-01
2.04359695e-01 -8.51136744e-02 1.83460832e-01 3.44666541e-01
6.02673106e-02 1.19922578e+00 -1.16949606e+00 -3.31771880e-01
2.38635272e-01 2.01170444e-01 -3.98421288e-01 7.75414228e-01
-2.81406224e-01 5.21249652e-01 -2.52047747e-01 2.09064275e-01
4.16967541e-01 -3.82625222e-01 6.86073959e-01 -1.89029604e-01
5.47555201e-02 8.25246394e-01 -4.78230685e-01 1.63170528e+00
-5.81628561e-01 9.44267511e-01 3.90556425e-01 -6.44729435e-01
9.99939799e-01 4.34899330e-01 1.90381557e-01 -8.63664508e-01
4.03651237e-01 -2.42863479e-03 5.16237319e-01 -4.76602137e-01
1.04037595e+00 2.60999531e-01 -5.04693151e-01 8.00942957e-01
4.28473741e-01 -5.24533808e-01 4.97108221e-01 6.37591839e-01
9.91928577e-01 -3.47736150e-01 5.19950092e-01 -7.01356828e-02
7.85524607e-01 -2.85835147e-01 3.53549838e-01 5.77167511e-01
-6.65132284e-01 2.82937497e-01 1.23382962e+00 -2.06909925e-01
-7.12103784e-01 -6.42603695e-01 1.18223317e-01 1.38961589e+00
2.24373534e-01 -7.09140718e-01 -9.58001018e-01 -1.08309269e+00
-1.63285911e-01 6.44718111e-01 -7.20246673e-01 -1.24844909e-01
-5.83160400e-01 -5.81641436e-01 5.37946045e-01 -6.89520091e-02
6.71767056e-01 -1.23974013e+00 7.48414844e-02 4.23867464e-01
-8.44065368e-01 -1.09964895e+00 -6.20162845e-01 -1.62250623e-01
-3.39034379e-01 -1.34173405e+00 -2.58500308e-01 -4.56720382e-01
-2.32461095e-02 2.01335564e-01 1.73222625e+00 1.87158972e-01
2.19412431e-01 3.72396499e-01 -6.43912375e-01 1.50499299e-01
-1.01348042e+00 2.22288087e-01 -1.81149960e-01 5.39709218e-02
2.52273858e-01 -3.56564432e-01 -4.89288986e-01 3.68227929e-01
-4.08009529e-01 8.50159153e-02 4.48409580e-02 1.26893818e+00
-3.09901237e-01 -3.58263969e-01 7.38524675e-01 -1.02745366e+00
1.65811586e+00 -3.62997025e-01 -6.30097985e-02 4.49101120e-01
-4.63980556e-01 -2.06792235e-01 5.31051755e-01 -1.17270678e-01
-1.01640320e+00 -6.98437572e-01 -3.02534461e-01 3.19345742e-01
4.47258800e-02 6.39750540e-01 -2.22405061e-01 1.64102718e-01
5.19649863e-01 -1.50701869e-02 1.26824185e-01 -5.19972146e-02
6.52654409e-01 8.68157268e-01 2.96847939e-01 -4.49022800e-01
3.54836285e-02 -1.78381011e-01 -4.77431357e-01 -9.66606975e-01
-9.19140041e-01 -5.70754111e-01 -8.24286044e-01 -8.16818714e-01
7.28929818e-01 -6.23607576e-01 -1.19981933e+00 6.90894246e-01
-1.53218126e+00 -6.42527282e-01 6.79441169e-02 2.52169311e-01
-5.58202088e-01 5.20779848e-01 -1.06152463e+00 -9.07090545e-01
-4.56096977e-01 -1.20483387e+00 7.72466600e-01 2.15197444e-01
-6.30046964e-01 -1.37129009e+00 4.62392986e-01 6.43754184e-01
2.79961526e-01 2.97836870e-01 7.32269943e-01 -9.06935513e-01
-1.49710953e-01 -1.99528724e-01 -2.39428818e-01 5.39870024e-01
1.41050398e-01 2.75459826e-01 -1.00565040e+00 -1.64544612e-01
9.50987265e-03 -9.80382264e-01 6.50637567e-01 2.58611590e-01
4.23539370e-01 -4.72560763e-01 2.07309335e-01 -1.68905873e-02
6.46782756e-01 8.43881667e-02 5.21089911e-01 -3.74420360e-02
7.19654799e-01 1.02498066e+00 7.20016956e-01 5.00920534e-01
8.29676807e-01 7.98441410e-01 4.42105293e-01 -6.43553510e-02
4.09342013e-02 3.17475833e-02 4.38937306e-01 1.63285494e+00
9.39733833e-02 -7.61059821e-01 -8.33529472e-01 4.26669538e-01
-2.08156586e+00 -9.41131592e-01 -6.39699519e-01 1.66460454e+00
1.07866466e+00 1.35776415e-01 3.94641310e-01 -1.15413859e-01
7.29028523e-01 7.83871233e-01 -2.58602828e-01 -8.73296916e-01
-2.21347153e-01 -2.43100673e-01 -4.80181217e-01 9.39313173e-01
-9.52206194e-01 1.05474043e+00 6.38550711e+00 6.55712664e-01
-7.87157953e-01 3.12154088e-02 7.58581519e-01 1.28037900e-01
-2.70530194e-01 -2.09432498e-01 -3.47199500e-01 1.94332287e-01
8.28981638e-01 -1.59758776e-01 3.77071947e-01 5.17443419e-01
2.35053897e-01 -1.27720028e-01 -1.24367392e+00 7.24487960e-01
-2.82381801e-03 -1.03668618e+00 -2.49821618e-01 -8.01726133e-02
6.92893147e-01 -1.58347800e-01 -1.96155712e-01 7.46359646e-01
7.82580853e-01 -1.06247616e+00 2.78803021e-01 5.62903762e-01
3.42634678e-01 -9.01399493e-01 1.08030355e+00 4.45291102e-01
-1.10410905e+00 3.13106954e-01 5.65601476e-02 -1.08145081e-01
2.18936756e-01 5.65350175e-01 -1.10946655e+00 6.24628127e-01
1.95788890e-01 6.44962192e-01 -5.00627756e-01 4.26202178e-01
-3.36118072e-01 5.93786359e-01 2.46310025e-01 -4.16223824e-01
6.03186309e-01 -4.70294505e-01 6.19253278e-01 1.66903031e+00
-3.90900731e-01 1.43609866e-02 3.59849334e-01 6.99565113e-01
-4.02749121e-01 3.72994214e-01 -5.92958868e-01 -3.33031952e-01
3.99945080e-01 1.73306608e+00 -2.57645011e-01 -3.22570294e-01
-2.12348193e-01 1.02405238e+00 8.08505714e-01 6.17531538e-02
-5.70642889e-01 -4.31381226e-01 6.01487756e-01 -6.87925875e-01
-2.12556854e-01 -6.22592345e-02 -1.08125523e-01 -9.82867658e-01
-9.27113071e-02 -1.17566657e+00 1.41676947e-01 -3.46895307e-01
-1.43449020e+00 9.58982229e-01 -5.62999129e-01 -1.05903113e+00
-7.13780880e-01 -2.97569126e-01 -8.75543594e-01 8.70783508e-01
-1.12263203e+00 -1.05644202e+00 -4.80979353e-01 4.50761944e-01
6.54613435e-01 -1.61533311e-01 9.89148378e-01 -2.02699348e-01
-7.88440347e-01 7.40712583e-01 -1.30773753e-01 5.36016583e-01
8.39333415e-01 -1.57785046e+00 3.99838001e-01 2.11993948e-01
-1.14474513e-01 5.00391066e-01 6.90830946e-01 -3.10616195e-01
-9.13070023e-01 -7.12515414e-01 1.24081588e+00 -4.49379981e-01
8.76706183e-01 -3.15158635e-01 -1.04481506e+00 1.67738855e-01
8.97295535e-01 -4.70349550e-01 8.32758009e-01 8.04863036e-01
-2.99993217e-01 2.23652437e-01 -8.07670355e-01 5.32964408e-01
7.38427758e-01 -9.42012608e-01 -7.20695555e-01 4.75380719e-01
8.04790854e-01 -6.58755124e-01 -1.17540359e+00 4.08182055e-01
5.25522888e-01 -1.56817961e+00 4.73868698e-01 -5.72782516e-01
8.25355768e-01 3.10057104e-01 8.38018954e-02 -1.65326524e+00
-1.35876283e-01 -1.04421961e+00 -2.62776822e-01 1.37911463e+00
3.88359815e-01 -2.04162642e-01 3.63701522e-01 5.00145853e-01
-2.67838478e-01 -8.58415246e-01 -6.36588335e-01 -3.89006823e-01
4.90189418e-02 -5.82925938e-02 4.23537374e-01 1.32087755e+00
7.72921324e-01 1.22530425e+00 -5.60999155e-01 -5.67630410e-01
1.21809475e-01 2.63472646e-01 1.08780134e+00 -1.21775329e+00
-2.05683619e-01 -9.59155202e-01 -1.19319476e-01 -1.23164296e+00
4.84166265e-01 -6.37207210e-01 3.06271315e-01 -1.57389355e+00
1.63416266e-01 -1.44751891e-01 1.66750044e-01 -9.00105089e-02
-5.19382298e-01 -7.73760602e-02 2.06006989e-01 7.79424757e-02
-1.10947442e+00 1.02278101e+00 1.73998761e+00 -2.66983718e-01
-4.18048948e-01 1.44156709e-01 -5.19356370e-01 6.39281571e-01
7.41252959e-01 9.14912075e-02 -2.93955922e-01 -8.12226757e-02
1.45577028e-01 7.32182920e-01 -1.57538790e-03 -5.45739472e-01
2.44040415e-01 -8.80099926e-03 -3.78415793e-01 -6.78501487e-01
4.74597514e-01 -3.59431915e-02 -5.01598418e-01 1.10880941e-01
-9.24950540e-01 3.20047396e-03 -2.97122985e-01 4.97725934e-01
-7.85468161e-01 -9.64589715e-02 7.16057658e-01 -1.56265140e-01
-1.30670696e-01 1.07259959e-01 -2.55095422e-01 4.32392657e-01
5.18786669e-01 3.60540122e-01 -5.03940940e-01 -9.73691225e-01
-7.73574412e-01 5.15165508e-01 1.44976035e-01 4.74127531e-01
2.95240104e-01 -1.32883811e+00 -8.63402486e-01 -4.13033873e-01
1.72256485e-01 -1.93657622e-01 3.96560550e-01 9.87497628e-01
-2.98514068e-01 6.23592436e-01 -1.01208137e-02 -6.74365520e-01
-1.53370392e+00 -8.71928409e-02 5.11001945e-01 -9.84952211e-01
-1.75762445e-01 1.08095634e+00 1.64032593e-01 -1.14338386e+00
4.36576277e-01 -1.52696595e-01 -6.57049954e-01 5.70747316e-01
3.01285952e-01 3.56487572e-01 8.48305896e-02 -9.13872838e-01
6.67366758e-02 -6.30047470e-02 -1.58085257e-01 -1.46016985e-01
9.95384276e-01 -4.19796884e-01 -4.85695660e-01 9.68249142e-01
1.36402905e+00 -9.21138003e-02 -9.90276814e-01 -3.90081555e-01
-8.79219845e-02 -9.72053558e-02 -2.99460024e-01 -8.76104057e-01
-7.94769466e-01 1.03034449e+00 -5.30465804e-02 1.05003285e+00
5.45926273e-01 -2.00405717e-01 7.19009340e-01 6.70475125e-01
2.09833197e-02 -1.31937730e+00 6.82461083e-01 1.31465721e+00
1.19151211e+00 -1.60888767e+00 -1.27311617e-01 -2.84928769e-01
-1.11660230e+00 1.31479883e+00 8.18815827e-01 4.46309410e-02
2.31698334e-01 -2.64686793e-01 1.68084592e-01 -5.43794751e-01
-9.98737931e-01 -2.10184008e-01 5.18748760e-01 2.50580907e-01
1.14362979e+00 3.66929412e-01 -4.67409879e-01 4.70848203e-01
-5.60746074e-01 -5.32694638e-01 5.41578472e-01 3.60285044e-01
-2.94277072e-01 -1.19510090e+00 1.60331711e-01 3.14026207e-01
-1.10288873e-01 -7.76010081e-02 -1.29870272e+00 8.48891139e-01
-4.00056660e-01 1.46274722e+00 -1.07925311e-01 -8.83481503e-01
4.50044572e-01 9.98482928e-02 2.59230942e-01 -5.91499329e-01
-1.42929280e+00 -5.58061823e-02 9.27362442e-01 -6.69313967e-01
-5.78546524e-01 -4.37243521e-01 -9.27559078e-01 -8.44790161e-01
-6.71907842e-01 5.07436097e-01 3.10743481e-01 1.00548255e+00
4.97155637e-02 7.56801784e-01 1.18292356e+00 -6.80910110e-01
-5.99486589e-01 -1.59361875e+00 -4.75547284e-01 4.67550218e-01
4.73704427e-01 -4.65129107e-01 -3.52663219e-01 -5.03842831e-01] | [12.726475715637207, 8.143411636352539] |
2059603a-8649-4846-b2c7-332c11a4acd6 | systematic-generalization-for-predictive | 2102.05602 | null | https://arxiv.org/abs/2102.05602v2 | https://arxiv.org/pdf/2102.05602v2.pdf | Systematic Generalization in Neural Networks-based Multivariate Time Series Forecasting Models | Systematic generalization aims to evaluate reasoning about novel combinations from known components, an intrinsic property of human cognition. In this work, we study systematic generalization of NNs in forecasting future time series of dependent variables in a dynamical system, conditioned on past time series of dependent variables, and past and future control variables. We focus on systematic generalization wherein the NN-based forecasting model should perform well on previously unseen combinations or regimes of control variables after being trained on a limited set of the possible regimes. For NNs to depict such out-of-distribution generalization, they should be able to disentangle the various dependencies between control variables and dependent variables. We hypothesize that a modular NN architecture guided by the readily-available knowledge of independence of control variables as a potentially useful inductive bias to this end. Through extensive empirical evaluation on a toy dataset and a simulated electric motor dataset, we show that our proposed modular NN architecture serves as a simple yet highly effective inductive bias that enabling better forecasting of the dependent variables up to large horizons in contrast to standard NNs, and indeed capture the true dependency relations between the dependent and the control variables. | ['Prathosh A. P', 'Pankaj Malhotra', 'Gantavya Bhatt', 'Hritik Bansal'] | 2021-02-10 | null | null | null | null | ['systematic-generalization'] | ['reasoning'] | [ 1.12915024e-01 1.60395861e-01 -1.74547747e-01 -5.04716337e-01
5.38228601e-02 -8.41088057e-01 1.02186894e+00 -5.07965349e-02
-1.28028886e-02 1.07315254e+00 1.61006972e-01 -6.02665186e-01
-5.67099035e-01 -1.03193843e+00 -8.45795512e-01 -7.78998196e-01
-6.07915998e-01 5.66262841e-01 -8.22974294e-02 -6.17770195e-01
4.59749848e-02 7.69158721e-01 -1.72637069e+00 -4.43725325e-02
6.72274292e-01 1.02340174e+00 6.88294321e-02 7.00668991e-01
3.49725276e-01 1.05411756e+00 -5.50216258e-01 -9.73309800e-02
2.79012233e-01 -4.63213563e-01 -6.80456460e-01 -4.60463047e-01
5.74056506e-02 -8.35044086e-02 -4.99070048e-01 8.74527454e-01
1.64031073e-01 5.15099943e-01 9.88500595e-01 -1.22113931e+00
-1.10144997e+00 9.08890128e-01 2.72402138e-01 6.40511334e-01
1.04001880e-01 4.37282920e-01 8.50535035e-01 -3.49092215e-01
5.69686830e-01 1.31640589e+00 8.02163780e-01 7.09988117e-01
-1.42883408e+00 -4.90490109e-01 4.95925963e-01 2.82928973e-01
-8.41004491e-01 -3.85738254e-01 1.09966743e+00 -5.70104182e-01
9.44713891e-01 1.66621536e-01 6.26199007e-01 1.61888731e+00
9.54026163e-01 2.91290343e-01 1.12129569e+00 -3.37900043e-01
7.79436350e-01 7.99650773e-02 4.08026814e-01 2.16984391e-01
8.72367322e-02 8.24951291e-01 -4.55786884e-01 -2.13606596e-01
6.44625068e-01 8.83732215e-02 -2.41461515e-01 -1.14469588e-01
-1.26222336e+00 6.58322155e-01 6.05049849e-01 5.36368906e-01
-5.87972403e-01 2.54077733e-01 3.08167219e-01 6.19171262e-01
5.58422744e-01 9.04974759e-01 -1.04200923e+00 2.18940631e-01
-4.78128701e-01 3.00559014e-01 8.86673629e-01 8.78438473e-01
6.72396839e-01 3.59825313e-01 -8.01258162e-02 2.02779710e-01
-2.32465193e-01 7.56086946e-01 7.21165478e-01 -7.59131432e-01
1.57220796e-01 4.02660102e-01 1.10312417e-01 -9.68386769e-01
-8.24701190e-01 -8.24001193e-01 -1.20723975e+00 2.71459401e-01
3.22321743e-01 -5.76408446e-01 -1.02188671e+00 2.29853725e+00
-1.40578821e-02 1.22916333e-01 4.72741693e-01 6.06032193e-01
2.18502700e-01 6.00977898e-01 3.66372876e-02 -3.02698851e-01
7.37491250e-01 -3.28550965e-01 -6.46666646e-01 -2.64212489e-01
6.19203925e-01 2.97418106e-02 5.51144183e-01 7.73213357e-02
-8.52372169e-01 -9.06620026e-01 -1.14529586e+00 2.45499164e-01
-7.74974644e-01 -4.31616753e-02 8.12822402e-01 1.17049858e-01
-1.15829968e+00 1.08304346e+00 -9.41664636e-01 -2.75947988e-01
4.34266962e-02 3.77382487e-01 -7.89885297e-02 4.15323853e-01
-1.82664931e+00 1.31187367e+00 6.53405786e-01 4.17313218e-01
-1.09473097e+00 -8.30152452e-01 -7.37688005e-01 1.76732317e-01
2.13501260e-01 -8.97796988e-01 1.13236642e+00 -1.15327096e+00
-1.47388470e+00 3.69179957e-02 -7.29751438e-02 -8.11663270e-01
8.47874805e-02 -9.67647582e-02 -8.15750718e-01 -1.73712328e-01
-2.03987688e-01 3.66064578e-01 1.02039075e+00 -1.02048838e+00
-4.08014655e-01 -5.29650748e-01 4.17192698e-01 -3.20064388e-02
-1.93811476e-01 -9.02308345e-01 8.83170784e-01 -5.46695888e-01
5.46906926e-02 -9.75936055e-01 -3.96069169e-01 -3.50212395e-01
-1.57842368e-01 -3.03051263e-01 5.05302131e-01 -3.10670465e-01
1.16931045e+00 -1.97544682e+00 4.92239386e-01 1.65897414e-01
3.21164280e-01 -1.32278241e-02 3.62714119e-02 5.66121876e-01
-5.32961309e-01 -1.21389858e-01 -1.47650763e-01 2.31520906e-01
2.60873623e-02 5.57806313e-01 -9.35960710e-01 4.62491661e-01
4.18247402e-01 1.02623403e+00 -1.06765997e+00 1.48391694e-01
1.48275137e-01 3.46914291e-01 -5.69501296e-02 1.92619428e-01
-6.68590844e-01 5.38417161e-01 -6.24983966e-01 1.85232103e-01
4.46414262e-01 -1.73895404e-01 1.58026814e-01 9.61932167e-03
4.43427525e-02 2.52416700e-01 -6.58546209e-01 1.28417289e+00
-4.36581433e-01 6.57467723e-01 -6.92972302e-01 -1.45372045e+00
1.16186857e+00 3.01864058e-01 -8.96864384e-03 -7.78866827e-01
2.55874425e-01 1.08091727e-01 3.52485746e-01 -5.04416168e-01
1.90038249e-01 -7.02714682e-01 -3.19753557e-01 2.42598742e-01
4.68993336e-01 -1.11749180e-01 1.15238756e-01 5.85997440e-02
1.08218157e+00 1.81101114e-01 4.11090970e-01 -6.97140038e-01
7.55801201e-01 -1.70073360e-01 5.70040822e-01 8.46162677e-01
4.02115546e-02 -1.31349668e-01 8.03031027e-01 -8.86324584e-01
-9.17817771e-01 -1.08893085e+00 -2.81133771e-01 1.01053405e+00
-5.99489138e-02 4.93870974e-02 -1.26750588e-01 -4.65922862e-01
5.23917899e-02 1.26761246e+00 -1.17573273e+00 -5.35520971e-01
-5.77887714e-01 -6.93868518e-01 3.77191603e-01 7.87656128e-01
1.76749945e-01 -1.10806382e+00 -7.57897377e-01 1.53549939e-01
3.53766978e-01 -5.94926596e-01 3.23573142e-01 8.47073913e-01
-9.64631319e-01 -1.00398839e+00 -5.50748944e-01 -6.10416949e-01
5.00883043e-01 -5.49233139e-01 1.20938694e+00 -4.09793049e-01
2.36741126e-01 3.79041016e-01 1.60189867e-02 -4.05410469e-01
-7.86190748e-01 9.61144194e-02 5.16826689e-01 -3.06273997e-01
1.94498122e-01 -1.09730840e+00 -3.56555820e-01 6.98214248e-02
-8.47408473e-01 -1.90526135e-02 3.75807345e-01 9.29311812e-01
2.53763725e-03 1.31604061e-01 1.04239714e+00 -7.38123775e-01
6.50412738e-01 -9.35627043e-01 -6.59535885e-01 4.43341672e-01
-6.87366128e-01 5.63960850e-01 1.36695993e+00 -9.28952992e-01
-1.21863151e+00 -2.43128002e-01 2.41389766e-01 -3.92026037e-01
-2.54085988e-01 8.71343732e-01 1.72717303e-01 4.42863256e-01
9.78516817e-01 4.20160532e-01 -1.54915303e-01 -2.96243668e-01
5.37463307e-01 6.12080470e-02 7.51821101e-01 -7.37136543e-01
6.73175037e-01 2.67386109e-01 6.49155796e-01 -5.12757719e-01
-8.34711790e-01 1.98035806e-01 -9.52704668e-01 -2.80654222e-01
5.13279557e-01 -5.66619456e-01 -6.57790780e-01 4.31398451e-01
-1.14577639e+00 -4.11028683e-01 -5.66058636e-01 4.75970328e-01
-6.32014573e-01 -1.73353925e-01 -4.15922910e-01 -1.08276451e+00
3.58778611e-02 -6.90485716e-01 5.65995276e-01 3.12572569e-01
-3.56683046e-01 -1.70934975e+00 4.13617700e-01 -7.49916971e-01
5.19682884e-01 5.27045846e-01 1.19310653e+00 -1.01870835e+00
-1.80425987e-01 -2.68179864e-01 3.61175388e-01 4.09491509e-01
1.24976210e-01 -8.61690864e-02 -8.90722513e-01 -2.38075390e-01
5.00963330e-01 -5.25179543e-02 8.57784152e-01 2.99824774e-01
7.59186566e-01 -4.56248939e-01 -4.51089412e-01 3.60933512e-01
1.31117463e+00 5.43671429e-01 3.71146381e-01 2.92949984e-03
4.01537478e-01 6.23328388e-01 9.66681838e-02 3.41762424e-01
2.08847746e-01 2.77813543e-02 2.78175235e-01 3.31423014e-01
3.12744409e-01 -1.56159058e-01 3.98393393e-01 8.63157153e-01
-1.39143392e-01 -4.44504350e-01 -8.84763658e-01 5.24932981e-01
-1.77432287e+00 -9.95364428e-01 1.79713175e-01 1.95665526e+00
7.18277574e-01 3.14867139e-01 -1.20175399e-01 2.41468288e-02
6.96058035e-01 9.00945887e-02 -1.26069331e+00 -3.93228561e-01
-3.79286915e-01 1.49918362e-01 2.02889964e-01 4.27198827e-01
-8.62047255e-01 4.07352090e-01 7.01007748e+00 4.80797201e-01
-1.27427518e+00 -2.66681194e-01 6.98238730e-01 9.58005711e-02
-5.66835761e-01 8.59191716e-02 -6.42280877e-01 3.79349500e-01
1.49560618e+00 -6.73424840e-01 5.24592757e-01 6.14136696e-01
-3.01420629e-01 1.21634163e-01 -1.67827368e+00 4.52558398e-01
-9.36146379e-02 -1.34799421e+00 2.70678431e-01 -2.74753869e-01
8.74400437e-01 -8.59211385e-02 2.28789315e-01 7.31480718e-01
4.92763191e-01 -9.79551852e-01 7.29092717e-01 1.24808753e+00
3.44050407e-01 -6.25659466e-01 8.05876315e-01 6.60779119e-01
-8.13319206e-01 -5.38777173e-01 -2.78428674e-01 -7.51716435e-01
1.14639634e-02 6.27894044e-01 -4.47222769e-01 5.88588715e-01
3.08990657e-01 1.05258286e+00 -5.40026546e-01 2.84992397e-01
-2.28156596e-01 5.34901202e-01 -2.92341620e-01 -2.22073540e-01
5.72351813e-02 6.54830486e-02 5.65711319e-01 8.03384662e-01
2.71624297e-01 4.33295935e-01 -2.28454605e-01 1.12346113e+00
2.66159564e-01 -5.56040466e-01 -8.98024201e-01 -1.65077806e-01
2.06277609e-01 8.90640199e-01 -7.00899661e-01 -6.72390699e-01
9.01888758e-02 4.38821763e-01 5.77268064e-01 7.30575383e-01
-8.48488569e-01 -1.24717459e-01 4.99349296e-01 -4.15965855e-01
4.99272525e-01 -2.78206795e-01 -3.10128301e-01 -1.25605547e+00
3.73111293e-02 -4.82627481e-01 3.53655010e-01 -8.34073484e-01
-1.71481037e+00 8.93187702e-01 3.31167251e-01 -1.19170332e+00
-8.78204346e-01 -7.98590422e-01 -5.53775072e-01 1.07995677e+00
-1.19763756e+00 -8.59144688e-01 2.57686794e-01 5.06192803e-01
2.38452524e-01 -2.31393576e-01 1.02918899e+00 -4.99668717e-01
-3.81512940e-01 2.12846339e-01 3.18272352e-01 -3.10212553e-01
2.65317112e-01 -1.21761203e+00 4.51847136e-01 7.56077111e-01
-1.19721025e-01 8.91747177e-01 1.07054663e+00 -7.75328219e-01
-1.31997859e+00 -9.73266542e-01 6.72732055e-01 -7.37455726e-01
1.07378078e+00 -4.37963814e-01 -1.02359140e+00 9.39189017e-01
2.54044384e-02 -6.25597015e-02 2.43981495e-01 2.09449276e-01
-2.85200536e-01 -1.81021035e-01 -8.68577063e-01 7.49564350e-01
1.02809572e+00 -6.15346253e-01 -1.17245471e+00 2.28378817e-01
9.30125475e-01 -3.25182348e-01 -9.73654270e-01 5.65931618e-01
6.60004139e-01 -9.51840520e-01 8.64223361e-01 -1.00980031e+00
7.55239666e-01 -9.90961865e-02 -2.95270294e-01 -1.56011510e+00
-5.64556658e-01 -6.10740125e-01 -3.88851285e-01 9.81100619e-01
5.55702031e-01 -1.14025664e+00 7.75617957e-02 8.61810386e-01
-3.54434289e-02 -8.41960430e-01 -9.91402984e-01 -1.01212704e+00
5.57020545e-01 -5.87403536e-01 7.92118073e-01 8.72105956e-01
1.64663598e-01 5.12867272e-01 -1.69341832e-01 4.03503686e-01
3.31760436e-01 2.31088817e-01 3.14583808e-01 -1.31697154e+00
-3.01359922e-01 -3.39995325e-01 -6.37524843e-01 -9.22081113e-01
6.09756827e-01 -9.23813760e-01 -9.46946368e-02 -1.01427543e+00
-1.92689449e-01 -3.34410608e-01 -5.38602412e-01 3.03760231e-01
-5.30764498e-02 -3.40446442e-01 4.15203609e-02 1.64947614e-01
-1.83333814e-01 8.98442388e-01 1.22307503e+00 -1.51165083e-01
-1.81019157e-01 1.50083855e-01 -5.55069804e-01 8.33317935e-01
7.98398316e-01 -2.46107504e-01 -7.99935102e-01 -7.37071633e-02
5.04680157e-01 4.12990242e-01 6.77857935e-01 -1.04956162e+00
3.28296870e-01 -3.08626890e-01 7.07194030e-01 -5.36438465e-01
2.40448371e-01 -8.08273971e-01 1.71120703e-01 4.51386005e-01
-7.77879775e-01 3.20939541e-01 4.42550659e-01 7.06412435e-01
7.14313239e-02 1.19735762e-01 3.22067320e-01 -9.92494375e-02
-7.42127240e-01 2.12639812e-02 -5.87105155e-01 4.32239100e-02
8.88079941e-01 1.73866376e-01 -4.21931833e-01 -4.27375823e-01
-1.10420251e+00 1.97506800e-01 8.30382928e-02 5.39183795e-01
3.99286985e-01 -1.19292748e+00 -3.58097255e-01 4.34532017e-01
-2.30668262e-01 -3.49930972e-01 3.78696084e-01 5.11565030e-01
1.88008755e-01 5.89844644e-01 -4.19654071e-01 -5.90320051e-01
-3.63607496e-01 1.20881999e+00 5.26198745e-01 -7.61157423e-02
-5.95736086e-01 5.29135764e-01 4.47278410e-01 -4.20286506e-01
-3.73423100e-02 -1.06379819e+00 -2.20934510e-01 3.90940532e-02
4.25149977e-01 4.46081012e-01 -2.41548941e-01 -3.83135706e-01
-3.60316038e-01 2.29813963e-01 2.35724121e-01 -5.59727363e-02
1.39895344e+00 -2.86663949e-01 -1.48826256e-01 1.37651134e+00
1.04498267e+00 -5.62433839e-01 -1.27482641e+00 -2.88440526e-01
-1.05343901e-01 1.73673004e-01 -2.22115189e-01 -1.12911117e+00
-6.71723604e-01 6.70745969e-01 3.83289188e-01 7.26689339e-01
1.19008803e+00 6.56522587e-02 3.61961842e-01 8.00944567e-01
4.76177335e-01 -7.07222641e-01 -4.21431512e-01 7.42455065e-01
1.10800552e+00 -9.90019560e-01 -5.62526919e-02 2.72353590e-01
-3.20025086e-01 1.58322299e+00 5.27397454e-01 -6.15617692e-01
1.01653397e+00 1.72988549e-01 -1.44703746e-01 -1.93846628e-01
-1.38374472e+00 -2.45059039e-02 5.26136220e-01 6.71261787e-01
1.76351264e-01 2.34256744e-01 1.19646303e-01 6.75983608e-01
-5.34623504e-01 1.24099754e-01 4.55999017e-01 7.06028581e-01
5.85971493e-03 -5.47398090e-01 -2.06686378e-01 4.72794116e-01
4.60728630e-02 -1.10878637e-02 -4.69396710e-02 9.74680662e-01
2.14835644e-01 6.46736324e-01 1.49722964e-01 -4.77719396e-01
2.71423131e-01 3.00126344e-01 4.34023172e-01 -2.69769490e-01
-3.81841093e-01 -5.46207070e-01 -8.40851367e-02 -4.04400945e-01
-4.95458931e-01 -6.84604108e-01 -1.06957805e+00 -1.81945041e-01
-2.19605505e-01 -1.76276699e-01 3.25432062e-01 1.50304973e+00
3.40469360e-01 8.74895573e-01 6.57582998e-01 -1.08244848e+00
-8.39736938e-01 -1.11053109e+00 -5.23693860e-01 3.32499415e-01
9.54410374e-01 -9.45001900e-01 -9.32010770e-01 1.47470459e-01] | [7.082828521728516, 3.182791233062744] |
59927c91-ab5c-4844-8456-8784b9bc0eeb | confidence-guided-stereo-3d-object-detection | 2003.05505 | null | https://arxiv.org/abs/2003.05505v1 | https://arxiv.org/pdf/2003.05505v1.pdf | Confidence Guided Stereo 3D Object Detection with Split Depth Estimation | Accurate and reliable 3D object detection is vital to safe autonomous driving. Despite recent developments, the performance gap between stereo-based methods and LiDAR-based methods is still considerable. Accurate depth estimation is crucial to the performance of stereo-based 3D object detection methods, particularly for those pixels associated with objects in the foreground. Moreover, stereo-based methods suffer from high variance in the depth estimation accuracy, which is often not considered in the object detection pipeline. To tackle these two issues, we propose CG-Stereo, a confidence-guided stereo 3D object detection pipeline that uses separate decoders for foreground and background pixels during depth estimation, and leverages the confidence estimation from the depth estimation network as a soft attention mechanism in the 3D object detector. Our approach outperforms all state-of-the-art stereo-based 3D detectors on the KITTI benchmark. | ['Jason Ku', 'Steven L. Waslander', 'Chengyao Li'] | 2020-03-11 | null | null | null | null | ['3d-object-detection-from-stereo-images'] | ['computer-vision'] | [ 1.62261903e-01 -1.24970727e-01 -2.48634331e-02 -3.87884736e-01
-8.53229821e-01 -4.27335560e-01 6.24862254e-01 2.24118352e-01
-6.19474888e-01 2.46833935e-01 -3.45795512e-01 -4.44024652e-01
5.48975408e-01 -6.85941458e-01 -8.42897058e-01 -6.83292031e-01
3.07004601e-01 6.50765598e-01 1.30857146e+00 2.48726934e-01
6.70084119e-01 7.09334791e-01 -2.04983115e+00 -1.48819145e-02
7.46690154e-01 1.31246483e+00 5.63148975e-01 8.27409863e-01
-2.65025079e-01 5.90433419e-01 -4.15525675e-01 -1.62709981e-01
3.95009071e-01 6.80036563e-03 -1.61920354e-01 1.22306190e-01
9.71978128e-01 -7.88886011e-01 -1.63547382e-01 1.15980351e+00
5.40622056e-01 -3.73118997e-01 8.21843863e-01 -1.18628299e+00
4.28634554e-01 -1.18773974e-01 -8.91087294e-01 6.44589424e-01
1.83489755e-01 4.92857009e-01 6.31147444e-01 -1.10297084e+00
3.99634659e-01 1.44834495e+00 4.62216526e-01 5.03243148e-01
-9.77414906e-01 -8.09726059e-01 2.01732218e-01 2.88092911e-01
-1.40336025e+00 -5.06099939e-01 5.34174442e-01 -7.70854294e-01
1.19781113e+00 -1.09226912e-01 7.48668134e-01 8.98461461e-01
3.00199538e-01 1.07098460e+00 1.19742942e+00 -5.66819161e-02
2.78179765e-01 1.08739577e-01 1.98474690e-01 6.45628989e-01
6.41033649e-01 5.81702769e-01 -8.42472076e-01 2.53593683e-01
5.90374470e-01 -4.57307905e-01 2.65308142e-01 -6.90347195e-01
-9.35806215e-01 7.31954873e-01 2.63012707e-01 -2.86343664e-01
-1.17826186e-01 2.25904107e-01 3.20004821e-01 -2.91918278e-01
6.31470621e-01 -8.71703699e-02 -1.61729559e-01 -1.35197148e-01
-1.10681033e+00 4.74417329e-01 3.15613478e-01 1.10456693e+00
7.12259173e-01 -1.58118770e-01 -2.60353923e-01 2.32118666e-01
6.68342113e-01 7.87238479e-01 -2.94731915e-01 -9.69089329e-01
6.23510540e-01 8.78396332e-01 1.42334560e-02 -5.50743163e-01
-2.81650305e-01 -4.16811347e-01 -2.13808656e-01 8.62997532e-01
5.31937242e-01 4.68865454e-01 -1.18703997e+00 1.06996655e+00
8.33322763e-01 3.02988708e-01 -1.28556326e-01 1.23296559e+00
9.66586232e-01 3.38802785e-01 -5.13248220e-02 3.08983535e-01
1.16518307e+00 -6.25351727e-01 -9.51949209e-02 -9.28559422e-01
3.38331312e-01 -7.28829920e-01 4.19928074e-01 1.56888217e-01
-1.03617835e+00 -5.71811616e-01 -1.32658076e+00 -3.45430464e-01
-9.11031812e-02 -1.01933964e-01 4.59779590e-01 9.14109826e-01
-8.17092180e-01 1.19828850e-01 -8.88736546e-01 -3.95492107e-01
7.66346157e-01 1.92253590e-01 -1.03027865e-01 -1.50559917e-01
-7.12797999e-01 1.27133858e+00 5.40282130e-01 -4.41749729e-02
-1.30282879e+00 -6.17612898e-01 -9.33405042e-01 -2.91647345e-01
5.46138108e-01 -8.26399565e-01 1.35753274e+00 -2.88412958e-01
-1.15555000e+00 1.62396550e+00 -3.98194939e-01 -7.82506108e-01
9.57314253e-01 -2.95658499e-01 1.17680438e-01 9.68897790e-02
3.25721771e-01 9.47440028e-01 1.11033189e+00 -1.23280883e+00
-1.49857581e+00 -6.59718454e-01 -1.17057003e-01 3.39848638e-01
5.77955127e-01 -8.12705904e-02 -9.08466637e-01 2.06973359e-01
3.48359168e-01 -5.92228055e-01 -2.73469955e-01 3.35222781e-01
-5.21088362e-01 -1.93217710e-01 9.44479108e-01 -1.63892895e-01
5.21375775e-01 -2.02506113e+00 -1.43649876e-01 -7.00418577e-02
2.65909076e-01 2.99710095e-01 2.40694582e-01 -3.10194582e-01
5.14728367e-01 -4.38587606e-01 -1.09818235e-01 -4.90246445e-01
8.02677777e-03 9.97013226e-02 -1.52327016e-01 7.03128874e-01
6.83936357e-01 5.70635736e-01 -9.42659020e-01 -7.62455821e-01
8.79847407e-01 5.22006333e-01 -5.13495505e-01 2.54038662e-01
-3.11696023e-01 5.27230918e-01 -5.97900748e-01 9.46176469e-01
1.05484009e+00 2.73818433e-01 -3.39416713e-01 6.70248596e-03
-5.60365617e-01 7.43111551e-01 -1.19315064e+00 1.41146362e+00
8.64350470e-04 7.87007987e-01 2.56473839e-01 -5.84723949e-01
7.37597942e-01 -1.42511725e-01 9.35922265e-02 -8.03077817e-01
3.71283174e-01 5.17567873e-01 -2.15083703e-01 -2.41058111e-01
7.34638870e-01 -6.89110067e-03 1.56690136e-01 -1.33959547e-01
-1.45943761e-01 -7.01399565e-01 1.61978483e-01 2.20563665e-01
1.16060090e+00 3.63122433e-01 9.64788273e-02 -1.93636686e-01
5.39381087e-01 1.95113137e-01 6.88393533e-01 8.88497651e-01
-4.86811161e-01 7.76787639e-01 4.33888078e-01 -1.79510385e-01
-1.14091969e+00 -1.15738297e+00 -5.68885565e-01 3.82559776e-01
7.09753752e-01 -5.11515923e-02 -6.42489791e-01 -6.81356966e-01
5.67242146e-01 6.89468801e-01 -4.55321401e-01 -8.80459771e-02
-3.98389816e-01 -4.75798547e-01 3.91094804e-01 5.33336282e-01
4.06407416e-01 -3.80642533e-01 -1.33443642e+00 3.35914731e-01
1.37766749e-01 -1.51367366e+00 -4.46653105e-02 6.73768640e-01
-9.98025715e-01 -1.17404544e+00 -4.70145881e-01 -2.80617505e-01
4.22325253e-01 8.05870891e-01 1.38483846e+00 -1.86539307e-01
-4.71141696e-01 2.42909387e-01 -1.06050566e-01 -8.18176627e-01
-5.24766803e-01 -7.74611160e-02 -1.71201125e-01 -2.81907707e-01
6.99442089e-01 -5.63236289e-02 -7.44241238e-01 4.30275440e-01
-4.48037684e-01 2.62149364e-01 5.11809111e-01 4.65540946e-01
6.73172593e-01 -1.68707937e-01 -2.07454488e-01 -4.85068500e-01
-2.22785816e-01 -1.27804011e-01 -1.43125117e+00 -5.21431208e-01
-4.36099231e-01 2.42402423e-02 -3.07770878e-01 1.18252888e-01
-8.09073687e-01 5.35681665e-01 -2.93285906e-01 -4.88386303e-01
-2.64010668e-01 -2.19832420e-01 -2.21678331e-01 -1.56111658e-01
5.22655189e-01 9.73474979e-02 -1.18874535e-01 -4.19574767e-01
1.39951734e-02 7.21954823e-01 6.49056196e-01 -1.39755920e-01
8.98526788e-01 8.83532286e-01 2.38110095e-01 -9.34030771e-01
-1.03591514e+00 -1.01249218e+00 -7.10001171e-01 -5.30625463e-01
9.83304441e-01 -1.36448157e+00 -6.16242349e-01 8.00852597e-01
-1.42605150e+00 -1.91493273e-01 1.18219880e-02 3.71456474e-01
-4.01469976e-01 2.55835474e-01 -2.58135647e-01 -1.35878110e+00
-6.28478592e-03 -1.40704310e+00 1.86353219e+00 1.25296801e-01
1.55157551e-01 -3.48667502e-01 -4.43696469e-01 5.47970474e-01
6.66285381e-02 1.14993796e-01 4.24994677e-01 -2.19436157e-02
-1.30668926e+00 -4.36859220e-01 -6.41828239e-01 2.19741240e-01
-4.72381860e-01 5.77391386e-02 -1.53633964e+00 9.61298496e-02
-1.19124562e-01 -1.35438263e-01 1.35481608e+00 5.60737312e-01
7.53918469e-01 6.41214430e-01 -6.14507318e-01 5.02795517e-01
1.15303421e+00 -6.05769567e-02 5.73642731e-01 4.71631557e-01
5.85251451e-01 5.34281552e-01 1.07345784e+00 4.60654527e-01
5.18418193e-01 6.84095740e-01 1.03129411e+00 7.68536255e-02
-3.15778881e-01 -9.64074358e-02 4.62939709e-01 8.85312036e-02
2.12038755e-01 -2.04753503e-01 -9.78139997e-01 5.60197949e-01
-1.57032502e+00 -9.43033457e-01 -6.76268756e-01 2.40862751e+00
5.89418113e-01 9.41647947e-01 1.53475389e-01 3.88778418e-01
7.44016528e-01 3.04227937e-02 -8.22089434e-01 -2.14261990e-02
-2.07431510e-01 -1.50187135e-01 1.05623889e+00 4.76025641e-01
-1.46605551e+00 1.09071112e+00 5.93541002e+00 7.52028108e-01
-8.79638731e-01 3.09330858e-02 4.23602283e-01 -2.39552990e-01
6.95921406e-02 -9.52074081e-02 -1.85232925e+00 3.59606057e-01
6.90977812e-01 2.50793278e-01 -2.56159008e-01 9.84698117e-01
3.42847824e-01 -9.60264742e-01 -1.20719373e+00 1.15158546e+00
-3.04974876e-02 -1.13019979e+00 -3.13179821e-01 5.16622812e-02
5.26806116e-01 6.58250630e-01 -3.20637748e-02 3.41232866e-02
1.63339674e-01 -6.31093025e-01 1.41036355e+00 -5.60938716e-02
5.00950754e-01 -7.25041270e-01 6.73800647e-01 6.30593359e-01
-1.26773489e+00 1.37359187e-01 -5.50765097e-01 -6.70195520e-02
3.36010039e-01 1.16925704e+00 -1.13178360e+00 2.99498960e-02
8.56366992e-01 6.97574735e-01 -6.54529274e-01 1.54981327e+00
-5.52677572e-01 3.96423519e-01 -5.89698255e-01 -7.64790131e-03
4.28847820e-01 1.29197687e-01 9.94833291e-01 1.12801456e+00
2.68886983e-01 -3.48142385e-01 1.01191394e-01 1.13951743e+00
1.56920925e-01 -5.24781048e-01 -6.33745551e-01 2.67530084e-01
4.14913028e-01 1.06890786e+00 -8.79231215e-01 -3.35032076e-01
-4.98890460e-01 8.11696410e-01 6.73813149e-02 -1.89072073e-01
-9.13592517e-01 5.07628210e-02 1.00353682e+00 5.19120574e-01
5.55672228e-01 -5.50333560e-01 -4.10832703e-01 -8.29385877e-01
1.31663442e-01 -5.39901376e-01 8.65056962e-02 -7.04069376e-01
-9.85035896e-01 1.25399590e-01 -4.65104282e-02 -1.32192123e+00
-7.42140934e-02 -8.96662235e-01 -1.29056752e-01 9.12486196e-01
-2.21900344e+00 -8.08097601e-01 -5.48121572e-01 2.28347123e-01
6.18360400e-01 3.16224843e-01 1.01996407e-01 5.82038164e-01
-3.40690613e-01 1.86814535e-02 -3.04732978e-01 -1.88872442e-01
5.58062613e-01 -1.20486307e+00 9.94424343e-01 1.05271733e+00
-2.44445756e-01 -1.08862229e-01 7.04275429e-01 -7.53878713e-01
-1.69317698e+00 -1.18932915e+00 8.48274708e-01 -9.02937531e-01
3.04754853e-01 -5.61854839e-01 -6.94638669e-01 1.62214920e-01
-3.52935106e-01 1.05113655e-01 -3.67884748e-02 -3.29300225e-01
-2.92629033e-01 5.67163415e-02 -9.75621879e-01 4.56801236e-01
1.20275211e+00 -5.16142488e-01 -5.22287667e-01 6.67097941e-02
5.97644210e-01 -1.02652442e+00 -9.56880685e-04 5.80493748e-01
5.29968619e-01 -1.44820392e+00 1.25608575e+00 1.13308921e-01
7.41865262e-02 -7.99598813e-01 -1.49286240e-01 -6.13804519e-01
7.93401990e-03 -1.67916343e-02 -2.33416185e-01 8.51620018e-01
2.19127640e-01 -3.12308013e-01 1.15539336e+00 2.44041070e-01
-4.38743889e-01 -4.52712476e-02 -1.43490553e+00 -9.92891252e-01
-3.20851296e-01 -1.07216251e+00 2.07302809e-01 1.44479841e-01
-7.20991373e-01 3.88966233e-01 -2.65610497e-03 5.50875783e-01
1.15417862e+00 1.35797754e-01 1.12430513e+00 -1.48276651e+00
-7.07574263e-02 -6.75119340e-01 -9.78539526e-01 -1.44181812e+00
-7.04508722e-02 -6.74689591e-01 5.93868673e-01 -1.46111286e+00
2.36597791e-01 -3.73889774e-01 9.13579687e-02 -1.15029119e-01
-3.27228308e-01 4.45517331e-01 -7.50638768e-02 -4.21312302e-02
-6.35446548e-01 4.34661657e-01 1.00274432e+00 -1.83211103e-01
1.39843719e-02 5.60820922e-02 -9.74626467e-02 9.72407758e-01
3.62090141e-01 -6.95023656e-01 6.76018000e-02 -4.75792527e-01
1.95061028e-01 -1.12668686e-01 8.93347681e-01 -1.39230442e+00
2.53750622e-01 8.47681239e-02 4.13015664e-01 -1.62279999e+00
5.94837368e-01 -8.50413144e-01 -4.00862306e-01 5.96876681e-01
1.65572941e-01 -4.86698955e-01 3.76617312e-01 6.31102741e-01
-3.52357663e-02 -1.85807541e-01 1.28217089e+00 -1.02622546e-01
-1.19544744e+00 3.35306764e-01 -5.49882114e-01 1.36009067e-01
1.06891322e+00 -7.83858120e-01 -9.19488743e-02 2.00622886e-01
-4.40893359e-02 4.02946860e-01 5.27265966e-01 5.08484483e-01
7.06946254e-01 -9.90738869e-01 -7.43013144e-01 4.18948025e-01
5.11191010e-01 6.55450463e-01 -1.68895528e-01 9.23268616e-01
-6.54837966e-01 4.60044652e-01 1.46523258e-02 -1.31700253e+00
-1.29455900e+00 2.66097575e-01 4.67099398e-01 2.66744673e-01
-6.38588786e-01 1.15974534e+00 3.11646432e-01 -1.15268841e-01
6.15915120e-01 -6.22757912e-01 1.05390839e-01 -2.71956194e-02
5.90894103e-01 4.46703404e-01 3.04249167e-01 -6.43983781e-01
-7.15974092e-01 6.72154546e-01 1.15685448e-01 -3.93488333e-02
9.19188738e-01 -2.64023036e-01 3.55010360e-01 5.67174852e-01
7.64444947e-01 -3.23368520e-01 -1.75307298e+00 -1.39485538e-01
6.87280521e-02 -8.10488641e-01 7.81101942e-01 -4.16993588e-01
-8.62189770e-01 1.28145051e+00 8.06541026e-01 -1.99537352e-01
5.54231822e-01 9.14883018e-02 5.80851018e-01 1.58646092e-01
5.92613935e-01 -1.21664965e+00 -1.10986270e-01 8.12311113e-01
3.40818703e-01 -1.88922083e+00 1.43455520e-01 -6.84682369e-01
-1.92295790e-01 9.09822106e-01 7.04498649e-01 1.43628125e-03
2.64468640e-01 5.94595134e-01 -6.98183030e-02 -1.39352977e-01
-5.90768099e-01 -8.23759854e-01 3.47332507e-01 7.08536685e-01
1.36450395e-01 -2.74109304e-01 2.05196440e-01 4.66752835e-02
3.44823182e-01 -1.90854147e-01 1.81646660e-01 1.06690466e+00
-1.07207155e+00 -8.42120886e-01 -6.31847203e-01 2.51062930e-01
-2.24748597e-01 -1.13291800e-01 -4.23609883e-01 7.17347562e-01
3.66616726e-01 1.01012385e+00 2.45792806e-01 -6.95644990e-02
3.93466502e-01 -3.02780475e-02 6.34937763e-01 -7.06914604e-01
-2.75767267e-01 1.38531014e-01 1.39324337e-01 -9.11266327e-01
-3.42728049e-01 -9.55629706e-01 -1.20636964e+00 -1.98880836e-01
-6.65871620e-01 -4.53483790e-01 9.92613435e-01 8.39352429e-01
2.55537152e-01 3.81215483e-01 3.32503110e-01 -1.41992390e+00
-3.03127170e-01 -4.88044441e-01 -6.23917043e-01 -7.63283595e-02
4.42317188e-01 -1.04923987e+00 -5.19184470e-01 -2.95346886e-01] | [7.773555755615234, -2.5683035850524902] |
3437ab95-6d13-4763-8282-2225a5dc94aa | multi-task-adversarial-network-for | null | null | http://openaccess.thecvf.com/content_cvpr_2018/html/Liu_Multi-Task_Adversarial_Network_CVPR_2018_paper.html | http://openaccess.thecvf.com/content_cvpr_2018/papers/Liu_Multi-Task_Adversarial_Network_CVPR_2018_paper.pdf | Multi-Task Adversarial Network for Disentangled Feature Learning | We address the problem of image feature learning for the applications where multiple factors exist in the image generation process and only some factors are of our interest. We present a novel multi-task adversarial network based on an encoder-discriminator-generator architecture. The encoder extracts a disentangled feature representation for the factors of interest. The discriminators classify each of the factors as individual tasks. The encoder and the discriminators are trained cooperatively on factors of interest, but in an adversarial way on factors of distraction. The generator provides further regularization on the learned feature by reconstructing images with shared factors as the input image. We design a new optimization scheme to stabilize the adversarial optimization process when multiple distributions need to be aligned. The experiments on face recognition and font recognition tasks show that our method outperforms the state-of-the-art methods in terms of both recognizing the factors of interest and generalization to images with unseen variations. | ['Zhaowen Wang', 'Ian Wassell', 'Yang Liu', 'Hailin Jin'] | 2018-06-01 | null | null | null | cvpr-2018-6 | ['font-recognition'] | ['computer-vision'] | [ 3.53379846e-01 2.09323838e-01 -4.03364860e-02 -3.23812395e-01
-8.32282603e-01 -7.10236073e-01 5.47320724e-01 -9.71561015e-01
-2.15855107e-01 7.14483202e-01 1.98196307e-01 1.36252612e-01
-1.01066409e-02 -4.89835709e-01 -1.09795928e+00 -1.14255071e+00
8.33013952e-02 3.99572253e-01 -3.93826902e-01 -1.21955447e-01
1.12463487e-02 4.77618754e-01 -1.51195347e+00 3.18219721e-01
8.49193156e-01 1.04909122e+00 2.88001299e-02 9.62974250e-01
2.90315092e-01 6.91193104e-01 -1.10399771e+00 -7.01782405e-01
9.48840857e-01 -6.03967190e-01 -4.17459130e-01 5.50118387e-01
5.98251820e-01 -3.78134638e-01 -7.17242122e-01 1.06562209e+00
6.45895958e-01 -4.95485738e-02 9.46734011e-01 -1.65005076e+00
-1.52647662e+00 2.65775084e-01 -5.20871699e-01 7.80907497e-02
2.95407325e-01 1.60601348e-01 7.67857194e-01 -1.09069359e+00
3.50081414e-01 1.39394295e+00 9.57863107e-02 9.87243176e-01
-1.38460326e+00 -8.37194383e-01 1.48015514e-01 -2.05712542e-02
-1.46325099e+00 -6.54685199e-01 7.72065997e-01 -5.60803592e-01
6.14480376e-01 1.26008540e-01 1.90509796e-01 1.58734679e+00
6.07061803e-01 6.62516475e-01 1.04895699e+00 -3.70504290e-01
-4.43223165e-03 5.27936816e-01 -5.54756105e-01 8.08989644e-01
-2.12700162e-02 4.03210819e-01 -7.39823222e-01 -2.57058859e-01
1.01881087e+00 -1.65573414e-02 -4.30830866e-01 -3.93728465e-01
-1.08625567e+00 1.11013126e+00 2.34590128e-01 -6.73076734e-02
-4.43028986e-01 2.17347309e-01 -8.25470686e-02 7.64756083e-01
6.17857397e-01 5.39440453e-01 -4.74060208e-01 4.92140830e-01
-5.25702417e-01 1.96422055e-01 7.39021540e-01 1.18085945e+00
8.32264781e-01 4.97438103e-01 -4.37370330e-01 7.41838753e-01
1.16354138e-01 9.23332274e-01 6.00051522e-01 -8.41381073e-01
5.22233963e-01 2.77617306e-01 7.59159178e-02 -7.66990602e-01
2.88030326e-01 -3.51761609e-01 -9.08006608e-01 6.82068348e-01
4.02882934e-01 -5.29207110e-01 -1.10387576e+00 2.17949724e+00
5.06077185e-02 8.80310014e-02 2.34779567e-01 7.73670256e-01
3.96759808e-01 5.45014143e-01 -2.42999703e-01 -1.99700251e-01
1.17270648e+00 -1.09094274e+00 -9.78265464e-01 -4.44224745e-01
-2.48666704e-01 -1.00887012e+00 8.22650969e-01 3.50602210e-01
-1.02057278e+00 -8.81373525e-01 -1.20024598e+00 1.48380488e-01
-3.18627059e-01 3.75323892e-01 6.78660512e-01 8.33761990e-01
-9.49061453e-01 3.97295922e-01 -4.18639332e-01 2.82608300e-01
6.07701242e-01 5.15428901e-01 -7.15835094e-01 -1.22731410e-01
-1.08746088e+00 8.21201026e-01 -6.81315288e-02 -1.29416868e-01
-1.59053504e+00 -6.64682209e-01 -9.24349248e-01 6.19472377e-02
2.81642616e-01 -9.42903936e-01 7.50834763e-01 -1.57531011e+00
-1.50616598e+00 9.16129351e-01 -5.70588410e-02 8.54242295e-02
5.68589568e-01 -2.04113334e-01 -5.20267904e-01 5.41650802e-02
2.81218559e-01 5.88566303e-01 1.83969891e+00 -1.30837667e+00
-1.92091271e-01 -4.51261103e-01 9.01140049e-02 1.77349970e-01
-4.56351340e-01 6.41751662e-02 -1.06300071e-01 -9.35575366e-01
-4.05749053e-01 -9.56715941e-01 -1.64840683e-01 3.46051395e-01
-3.91311407e-01 5.81029914e-02 8.08498800e-01 -5.38993895e-01
7.55597234e-01 -2.26060700e+00 6.95726335e-01 -1.26194462e-01
3.86984617e-01 -6.62689516e-03 -5.23993611e-01 2.12108836e-01
-4.09015059e-01 1.06802516e-01 8.40241686e-02 -4.64053124e-01
-7.18565509e-02 3.62810820e-01 -4.66316342e-01 5.61287224e-01
7.29911506e-01 8.27123225e-01 -6.23566151e-01 -2.61091173e-01
-1.35956243e-01 4.85734165e-01 -5.58229744e-01 1.04057050e+00
-8.02204758e-02 5.43284774e-01 -5.75403333e-01 5.21848977e-01
8.27856660e-01 -2.06735238e-01 -3.12851310e-01 -8.83528143e-02
4.99469012e-01 -5.42342544e-01 -1.28876078e+00 1.75254786e+00
-4.10775155e-01 2.70676464e-01 1.70222834e-01 -8.40339184e-01
8.20453346e-01 5.48262060e-01 3.33497971e-01 -1.67890057e-01
5.29990382e-02 6.31007506e-03 1.18714660e-01 -5.92239201e-01
1.83349103e-01 -3.05038899e-01 -2.27140576e-01 6.09511912e-01
8.06682825e-01 -2.65743673e-01 -1.83397695e-01 2.22160742e-01
1.04996336e+00 8.94830748e-02 3.90187114e-01 -2.19406173e-01
6.62567973e-01 -6.84044719e-01 6.31965995e-01 6.05595350e-01
-2.32095957e-01 7.94080555e-01 5.93329132e-01 -3.96477401e-01
-1.26253772e+00 -1.27030265e+00 7.94612169e-02 1.13369799e+00
-2.25817859e-01 -2.91435644e-02 -6.10591948e-01 -9.99619007e-01
2.16955259e-01 4.67912793e-01 -1.10289264e+00 -3.00943911e-01
-1.98060200e-01 -6.36156201e-01 5.34355819e-01 4.83070940e-01
3.09921950e-01 -9.43483531e-01 -7.38818338e-03 -1.29634708e-01
2.03808725e-01 -1.14706194e+00 -1.02288735e+00 5.06754935e-01
-7.61692300e-02 -1.13739693e+00 -8.70027602e-01 -6.92652166e-01
1.16471410e+00 1.16077140e-01 9.99409795e-01 -3.32358390e-01
-3.49716783e-01 3.77536774e-01 -1.37867443e-02 -5.81263006e-01
-6.88297689e-01 -2.71721303e-01 1.19364858e-01 6.20383382e-01
1.12806015e-01 -7.36463964e-01 -5.69435716e-01 6.19589388e-01
-1.04625165e+00 -1.76478535e-01 6.14323258e-01 1.28380120e+00
2.13920072e-01 -1.70893237e-01 7.30550051e-01 -9.61837471e-01
7.55056322e-01 -4.84118044e-01 -4.59939718e-01 2.72108406e-01
-4.56630290e-01 5.58693767e-01 9.94618058e-01 -1.03088713e+00
-1.06490767e+00 3.45794111e-01 2.93549955e-01 -1.04041553e+00
-4.63007800e-02 -3.24221909e-01 -6.97018504e-01 -3.54992777e-01
1.03543234e+00 2.39555508e-01 2.90959567e-01 -8.40175897e-02
5.81132710e-01 4.93895024e-01 4.17861253e-01 -6.40894890e-01
1.28173482e+00 2.03583702e-01 1.50447860e-01 -2.33439893e-01
-9.32871878e-01 4.28879671e-02 -6.32819474e-01 -1.54886708e-01
8.51315498e-01 -1.14790583e+00 -5.16201913e-01 6.02252245e-01
-1.22337592e+00 2.34047011e-01 -4.67348874e-01 2.51484156e-01
-8.48837495e-01 6.07221574e-03 -3.62500399e-01 -5.53368032e-01
-7.54636573e-03 -1.49934351e+00 1.12464964e+00 3.19803387e-01
2.59104103e-01 -8.37974072e-01 1.11046191e-02 2.98139602e-01
4.94120866e-01 3.90726835e-01 7.13368595e-01 -8.74749362e-01
-7.59573758e-01 -8.42358917e-02 9.37398523e-02 8.99861753e-01
3.56258988e-01 -3.93595308e-01 -1.23965323e+00 -3.81650120e-01
2.46346772e-01 -7.53893971e-01 6.94679737e-01 -1.09689264e-02
1.22208750e+00 -5.93030930e-01 -5.77939302e-02 9.14099097e-01
1.32350028e+00 2.00785860e-01 6.39096797e-01 -1.75797239e-01
7.27254510e-01 3.65648478e-01 2.19436720e-01 3.81440192e-01
-2.40437072e-02 3.88492048e-01 5.10721564e-01 -1.67590156e-01
-1.80182561e-01 -2.86046147e-01 5.54041743e-01 4.41787243e-01
-6.05738424e-02 -5.00694096e-01 -3.04775536e-01 3.09985578e-01
-1.68799293e+00 -9.86885309e-01 5.82662225e-01 2.14368200e+00
9.12215352e-01 -2.44124129e-01 -3.03545356e-01 -1.12760104e-01
8.21361780e-01 5.69586515e-01 -8.21649909e-01 -3.57322037e-01
-4.13476199e-01 5.07846773e-01 4.41141725e-01 5.09081244e-01
-1.15192604e+00 8.09633315e-01 7.07245779e+00 7.04123080e-01
-8.61410260e-01 1.94419786e-01 7.36867368e-01 -9.28048491e-02
-3.33823681e-01 -2.22652361e-01 -6.60524547e-01 3.52650344e-01
6.83091283e-01 -3.41466606e-01 9.90030110e-01 8.98269176e-01
-2.23282993e-01 2.75268793e-01 -1.45707381e+00 9.65731382e-01
6.42845631e-01 -8.23956847e-01 2.79839367e-01 2.34117880e-01
9.58115935e-01 -5.55504322e-01 5.71821630e-01 2.31205761e-01
3.31329882e-01 -1.40560591e+00 6.25808239e-01 6.87198222e-01
1.04723370e+00 -7.81007826e-01 3.88146013e-01 3.59430552e-01
-6.50530636e-01 -6.43286481e-02 -4.97027129e-01 1.52894706e-01
-9.84954014e-02 2.48089969e-01 -5.15126646e-01 3.82401109e-01
7.59675726e-02 3.51437062e-01 -5.75911283e-01 2.25285187e-01
-5.17142415e-01 2.16263473e-01 1.43201277e-01 3.50374669e-01
-1.47707552e-01 -1.58960164e-01 5.04519343e-01 6.99085414e-01
3.79198521e-01 -1.33356517e-02 2.99896121e-01 1.09428442e+00
-4.35006559e-01 -9.29314867e-02 -9.51383770e-01 -1.00724921e-01
2.05388904e-01 1.33240151e+00 3.55716818e-03 -1.00978240e-01
-6.22108161e-01 1.52684367e+00 5.68017244e-01 4.64389890e-01
-8.62084508e-01 -4.03870046e-01 9.74494457e-01 -2.33389452e-01
4.27808404e-01 -7.69414008e-02 1.50285019e-02 -1.39720321e+00
1.53444752e-01 -1.17480099e+00 6.67734742e-02 -7.25380182e-01
-1.61474371e+00 9.79300797e-01 -2.32978165e-01 -1.29287994e+00
-6.34911537e-01 -7.63778687e-01 -5.88838995e-01 1.35721397e+00
-1.35844326e+00 -1.32251692e+00 -1.66500971e-01 1.10287058e+00
5.40954530e-01 -9.99224901e-01 1.17614841e+00 5.80887645e-02
-3.19134176e-01 9.77118134e-01 -4.96885218e-02 1.02271892e-01
1.02143919e+00 -1.23192036e+00 2.06828952e-01 1.00371337e+00
1.49562314e-01 5.23861229e-01 6.26903594e-01 -3.44850510e-01
-1.46769309e+00 -1.07287955e+00 6.20251060e-01 -6.09052777e-01
5.42942345e-01 -8.69978011e-01 -5.50880611e-01 8.37447524e-01
5.69799244e-01 5.12116015e-01 9.45352495e-01 -2.12976530e-01
-7.90343285e-01 -1.13000058e-01 -1.38508511e+00 3.26704383e-01
1.03003049e+00 -7.14333713e-01 -4.57700700e-01 5.95158994e-01
8.11017931e-01 -4.60708141e-01 -8.09670985e-01 -6.26443923e-02
4.47817177e-01 -8.21281910e-01 1.03863263e+00 -9.03627515e-01
6.83651447e-01 -1.14129879e-01 -2.61196941e-01 -1.91857016e+00
-5.71144760e-01 -1.00030482e+00 -1.83124751e-01 1.13067067e+00
4.63838607e-01 -7.28879869e-01 6.11604989e-01 5.42237043e-01
4.33605164e-02 -5.83670080e-01 -1.04583275e+00 -7.34059453e-01
8.88200328e-02 5.99753968e-02 5.97015798e-01 6.48694694e-01
-6.12916231e-01 6.37881696e-01 -1.10818994e+00 3.58737856e-01
7.89485991e-01 -6.82960153e-02 8.32663655e-01 -7.33710051e-01
-7.57440448e-01 1.70311153e-01 -4.66095448e-01 -8.08688283e-01
6.58329070e-01 -7.92578340e-01 -8.62841308e-02 -5.97191095e-01
4.37131345e-01 1.61029980e-01 -2.08605751e-01 4.28505093e-01
-5.55455506e-01 6.09632507e-02 2.54522026e-01 7.92991091e-03
-1.92847803e-01 8.52049291e-01 1.74525905e+00 -2.63661832e-01
2.58199632e-01 2.16307789e-01 -1.17481697e+00 3.17739099e-01
5.16367197e-01 -5.30732810e-01 -6.11972392e-01 -4.98796254e-01
-1.58916488e-01 1.69228181e-01 2.74667948e-01 -7.17986107e-01
1.47673711e-01 -3.53654087e-01 9.26327825e-01 1.39374062e-01
5.96078753e-01 -8.49928677e-01 1.33363456e-01 1.10100441e-01
-4.37337756e-01 1.90961555e-01 -1.69649899e-01 6.49776578e-01
-2.07704052e-01 -2.13321969e-01 9.73932385e-01 -1.56393081e-01
-4.18346561e-02 6.92507625e-01 -1.75299346e-01 1.57784849e-01
1.17335069e+00 1.52765378e-01 -3.71358842e-01 -6.30593300e-01
-8.31805289e-01 -1.93599965e-02 1.99232176e-01 6.92869246e-01
7.89995492e-01 -1.75011230e+00 -9.65579748e-01 7.85661280e-01
2.08568022e-01 -2.95134127e-01 -4.56998311e-02 1.32861454e-02
1.27897993e-01 1.90281153e-01 -5.34542084e-01 -2.49285430e-01
-1.08914065e+00 1.07717466e+00 3.29436719e-01 -3.32883179e-01
1.28531277e-01 8.29616129e-01 7.16945648e-01 -1.99835882e-01
9.69604552e-02 2.14598835e-01 1.26360194e-03 -1.07531428e-01
5.83851814e-01 1.96360517e-02 -3.29534054e-01 -6.71776712e-01
-1.83275994e-02 4.76107746e-01 -2.27532625e-01 -1.40609860e-01
1.34321165e+00 3.80001776e-02 9.31847841e-02 1.75389871e-01
1.48138463e+00 2.09011286e-01 -1.57606471e+00 -3.36190373e-01
-7.11628675e-01 -9.93728340e-01 -2.14498073e-01 -8.05580974e-01
-1.36437953e+00 6.79708898e-01 6.64550602e-01 -1.61358584e-02
1.33781040e+00 -5.82361929e-02 1.86451137e-01 2.30626087e-03
3.47811282e-01 -7.51529098e-01 6.74431920e-01 2.92965293e-01
1.47416246e+00 -1.28961658e+00 -1.64149970e-01 -4.34412539e-01
-8.63409817e-01 1.04029202e+00 8.49653542e-01 -5.07653952e-01
7.58180678e-01 5.83413661e-01 9.39908326e-02 -7.96502177e-03
-9.67604399e-01 -3.32841501e-02 6.06080294e-01 9.49681818e-01
1.11512072e-01 6.60466822e-03 -1.08022131e-02 6.64623201e-01
-1.52973726e-01 -4.06723708e-01 3.51410955e-01 5.95060647e-01
-5.50747616e-03 -1.38249707e+00 -5.55963337e-01 2.68740147e-01
-6.45527363e-01 -1.56708807e-02 -2.93532491e-01 6.12076879e-01
3.90584350e-01 7.34751105e-01 -1.65684938e-01 -4.85735536e-01
3.11223060e-01 2.21265867e-01 7.28546321e-01 -6.86392426e-01
-7.21540093e-01 3.09581449e-03 -2.31760427e-01 -6.11279368e-01
-9.15397108e-02 -6.23543382e-01 -3.59238982e-01 8.81213546e-02
-2.63835341e-01 8.18161014e-03 5.77201903e-01 5.81864357e-01
4.31797832e-01 6.46896064e-01 1.54465306e+00 -8.92641008e-01
-1.06633627e+00 -1.00730944e+00 -7.02181458e-01 7.40508080e-01
5.67344725e-01 -6.84032619e-01 -7.42776990e-01 2.97462016e-01] | [12.807409286499023, 0.0918397456407547] |
f36ec4c1-c06d-4d61-8fc9-8d1b6d2f1961 | geometric-algebra-based-embeddings-for | 2202.09464 | null | https://arxiv.org/abs/2202.09464v3 | https://arxiv.org/pdf/2202.09464v3.pdf | Geometric Algebra based Embeddings for Static and Temporal Knowledge Graph Completion | Recent years, Knowledge Graph Embeddings (KGEs) have shown promising performance on link prediction tasks by mapping the entities and relations from a Knowledge Graph (KG) into a geometric space and thus have gained increasing attentions. In addition, many recent Knowledge Graphs involve evolving data, e.g., the fact (\textit{Obama}, \textit{PresidentOf}, \textit{USA}) is valid only from 2009 to 2017. This introduces important challenges for knowledge representation learning since such temporal KGs change over time. In this work, we strive to move beyond the complex or hypercomplex space for KGE and propose a novel geometric algebra based embedding approach, GeomE, which uses multivector representations and the geometric product to model entities and relations. GeomE subsumes several state-of-the-art KGE models and is able to model diverse relations patterns. On top of this, we extend GeomE to TGeomE for temporal KGE, which performs 4th-order tensor factorization of a temporal KG and devises a new linear temporal regularization for time representation learning. Moreover, we study the effect of time granularity on the performance of TGeomE models. Experimental results show that our proposed models achieve the state-of-the-art performances on link prediction over four commonly-used static KG datasets and four well-established temporal KG datasets across various metrics. | ['Jens Lehmann', 'Yung-Yu Chen', 'Mojtaba Nayyeri', 'Chengjin Xu'] | 2022-02-18 | null | null | null | null | ['knowledge-graph-embeddings', 'temporal-knowledge-graph-completion', 'knowledge-graph-embeddings'] | ['graphs', 'knowledge-base', 'methodology'] | [-6.32212043e-01 -1.33742481e-01 -5.61099291e-01 -3.30774374e-02
7.79775158e-02 -4.77406770e-01 7.01210201e-01 6.02102220e-01
-2.65082300e-01 6.59704924e-01 1.11300997e-01 -5.15292585e-01
-8.45385611e-01 -1.22132838e+00 -7.10491538e-01 -4.24878985e-01
-8.11093628e-01 5.08199155e-01 3.09994280e-01 -4.25250798e-01
-3.75298649e-01 3.66096646e-01 -1.10737014e+00 -1.30631059e-01
8.65926802e-01 1.04450750e+00 -3.96056622e-01 2.92265713e-01
-1.23443685e-01 7.55319536e-01 -1.96246728e-01 -1.06854606e+00
8.19071680e-02 1.86128408e-01 -1.02233505e+00 -4.60101932e-01
3.97177726e-01 5.10240272e-02 -9.80269074e-01 7.01743782e-01
1.03230692e-01 2.47676685e-01 6.46440744e-01 -1.62143350e+00
-1.07328832e+00 8.04756939e-01 -5.18773735e-01 5.19318938e-01
3.27213854e-01 -3.77018064e-01 1.46112740e+00 -8.15579653e-01
8.72037351e-01 1.14994276e+00 7.86368251e-01 -2.06572875e-01
-1.22373617e+00 -5.99202633e-01 5.10365367e-01 9.59284365e-01
-1.67268300e+00 2.06792355e-01 8.18997920e-01 -6.51259780e-01
1.05631471e+00 8.32548514e-02 8.52202237e-01 9.34252262e-01
3.40406358e-01 5.41029394e-01 7.33145833e-01 -2.16416344e-01
-1.34680137e-01 -2.18155608e-01 3.93079430e-01 9.03091669e-01
4.53611881e-01 -1.18920982e-01 -6.53819740e-01 -2.32269228e-01
5.57286739e-01 1.20485881e-02 -4.40033644e-01 -6.56620622e-01
-1.38874173e+00 8.13640296e-01 8.23198795e-01 4.39132094e-01
-3.67970586e-01 3.07465881e-01 4.89251375e-01 2.99596429e-01
6.15088284e-01 5.43029942e-02 -6.15226865e-01 -1.06071912e-01
-4.07103539e-01 1.18288204e-01 8.81100893e-01 7.94285715e-01
6.48720741e-01 -1.18867472e-01 3.36147174e-02 7.48067856e-01
1.80236131e-01 7.39644244e-02 1.26169890e-01 -2.07734868e-01
7.78393269e-01 9.18471634e-01 -8.05776343e-02 -1.57184911e+00
-5.90133607e-01 -7.09772348e-01 -8.30693483e-01 -6.41389549e-01
3.16801488e-01 2.72181600e-01 -6.51964724e-01 1.54835320e+00
4.62958634e-01 6.35313630e-01 -7.76476087e-03 6.16548777e-01
8.58238518e-01 6.40499890e-01 1.72870994e-01 -2.37687558e-01
1.30290842e+00 -7.36090541e-01 -7.84094572e-01 3.60445201e-01
8.81563962e-01 -2.87754804e-01 6.75391018e-01 1.83427498e-01
-6.67226374e-01 -3.08094889e-01 -9.17507470e-01 -4.52361815e-03
-1.11891770e+00 2.36316901e-02 1.26542366e+00 2.58582145e-01
-8.86227667e-01 7.94354558e-01 -9.41779137e-01 -5.58609545e-01
2.75919139e-01 1.76063254e-01 -6.07835770e-01 -2.44314298e-01
-1.75101447e+00 8.35468233e-01 8.46542001e-01 2.39235520e-01
-2.16154829e-01 -1.05657935e+00 -9.94877219e-01 1.56451285e-01
7.28523672e-01 -8.24274600e-01 4.26655293e-01 7.87973031e-02
-1.03554142e+00 4.80924189e-01 2.29880452e-01 -5.78377843e-01
3.53569001e-01 -2.36834303e-01 -1.19794929e+00 4.45979908e-02
-1.32350817e-01 9.75964367e-02 5.39852560e-01 -8.88084352e-01
-4.38410252e-01 -4.03830022e-01 2.90268153e-01 -1.96531773e-01
-8.61575782e-01 -3.90344113e-01 -6.37180328e-01 -8.19474101e-01
1.25902250e-01 -9.40744638e-01 2.52504319e-01 -2.92949706e-01
-4.08451736e-01 -6.55089915e-01 8.75323892e-01 -8.01196158e-01
1.92151761e+00 -1.85504365e+00 6.76004112e-01 2.31152758e-01
3.67978275e-01 2.21854836e-01 1.18077263e-01 8.59017074e-01
-2.88980663e-01 2.13285714e-01 1.94836274e-01 1.77181512e-02
1.96318299e-01 4.52290535e-01 -3.37920278e-01 4.47657764e-01
-4.12755087e-02 1.26734698e+00 -1.12792003e+00 -5.07384121e-01
-5.41457999e-03 6.40344083e-01 -3.40085298e-01 -3.41696143e-01
-1.63445130e-01 -2.57609561e-02 -4.40141171e-01 5.72648466e-01
4.68387872e-01 -5.75904608e-01 4.09988195e-01 -7.28583574e-01
-2.64674015e-02 1.57109171e-01 -1.03242195e+00 1.65647411e+00
-2.79307902e-01 1.71609238e-01 -5.08960843e-01 -1.05148780e+00
7.51640499e-01 1.95464820e-01 8.38535607e-01 -5.09486735e-01
3.82822528e-02 1.33791313e-01 5.42351454e-02 -2.64366895e-01
5.82208455e-01 2.73189694e-01 -1.43327326e-01 1.41806230e-01
1.93587363e-01 2.99511313e-01 5.02489448e-01 6.74056709e-01
1.17135680e+00 1.51142225e-01 1.79858193e-01 -1.46143124e-01
5.22321343e-01 -1.83755144e-01 6.66243851e-01 4.61924709e-02
-3.58459577e-02 -2.65814096e-01 6.99117422e-01 -6.37610614e-01
-5.89759290e-01 -1.26224649e+00 -1.93063706e-01 7.79840708e-01
1.23823822e-01 -9.41367984e-01 9.50887948e-02 -1.00299561e+00
6.59609854e-01 5.27408957e-01 -8.83013606e-01 -3.20613027e-01
-3.32821369e-01 -7.27213144e-01 3.34092468e-01 6.64755940e-01
4.07847852e-01 -3.44956696e-01 2.36752152e-01 3.49336654e-01
-1.29666671e-01 -1.33641267e+00 -3.15632403e-01 -1.14572026e-01
-7.11624086e-01 -1.27056670e+00 -3.81770164e-01 -4.99453276e-01
2.61794567e-01 1.01626366e-01 1.07973814e+00 -2.34876171e-01
-2.67647207e-01 7.93138862e-01 -6.70381129e-01 -3.18191350e-02
2.98450172e-01 2.19898269e-01 4.18777585e-01 3.97943467e-01
2.27072641e-01 -9.33323383e-01 -4.99762267e-01 3.27276111e-01
-8.68994892e-01 -1.66850984e-01 3.41570854e-01 8.16514432e-01
6.19938195e-01 4.54317808e-01 7.05459714e-01 -7.89157689e-01
5.09016871e-01 -9.19853747e-01 -4.88415629e-01 5.70831597e-01
-8.67835164e-01 1.12888575e-01 5.47830045e-01 -5.76914132e-01
-6.46483958e-01 -6.04067981e-01 5.17521739e-01 -8.54510665e-01
7.22482085e-01 1.31032729e+00 1.46896899e-01 -1.35201097e-01
3.79005581e-01 8.96065384e-02 -3.99547428e-01 -4.74900246e-01
7.03725934e-01 9.44642257e-03 4.51754928e-01 -7.24196076e-01
1.23108387e+00 4.50013697e-01 2.81160504e-01 -5.98255932e-01
-6.86167479e-01 -4.96666491e-01 -6.90982521e-01 -7.97450244e-02
6.33457482e-01 -8.74478519e-01 -8.81319225e-01 1.40895948e-01
-9.21586275e-01 -4.67684828e-02 -5.79473600e-02 7.00045705e-01
-1.32717609e-01 5.05749166e-01 -6.70620859e-01 -4.77583349e-01
-1.61647409e-01 -3.23442906e-01 5.47128916e-01 -4.00951132e-02
3.98428412e-03 -1.38082361e+00 1.19237602e-01 3.41696441e-01
2.57419169e-01 6.04946315e-01 1.34150267e+00 -4.18752372e-01
-6.71457231e-01 -1.39868528e-01 -3.88213456e-01 -1.00118695e-02
1.73741445e-01 1.10970959e-01 -2.92298019e-01 -3.55277687e-01
-8.27899158e-01 -3.14002633e-02 1.00326884e+00 -5.83443418e-03
1.00084913e+00 -2.78165728e-01 -7.13596523e-01 6.69460893e-01
1.49209750e+00 -6.56741112e-02 5.09548008e-01 3.75985146e-01
1.06759727e+00 3.94634694e-01 5.95511854e-01 4.03996795e-01
1.03930187e+00 9.36645746e-01 4.04431462e-01 3.89630705e-01
3.33397016e-02 -4.30538446e-01 1.41872004e-01 1.21922100e+00
-5.05726755e-01 -1.76773235e-01 -9.82493997e-01 9.73859966e-01
-2.07939386e+00 -9.27299798e-01 -3.73851180e-01 2.07801318e+00
6.93367243e-01 1.32723942e-01 8.28525573e-02 1.81859761e-01
3.48435849e-01 3.73590708e-01 -4.18128312e-01 -1.26297429e-01
-6.59406558e-02 2.42511809e-01 5.88397145e-01 1.81372732e-01
-1.19762588e+00 1.01352978e+00 4.64834833e+00 7.16914833e-01
-8.83778989e-01 2.00535372e-01 -1.02643184e-02 1.65896341e-01
-3.62839699e-01 2.35620514e-02 -6.74827814e-01 3.21492761e-01
8.68131340e-01 -5.54943919e-01 4.32619005e-01 5.97199380e-01
-1.47561133e-01 1.92779630e-01 -1.08895493e+00 1.08996665e+00
-3.60196829e-02 -1.31309795e+00 3.97979952e-02 2.20161974e-01
8.23225021e-01 -1.17362872e-01 6.00388832e-02 7.84519553e-01
4.00202364e-01 -8.87038648e-01 2.89754391e-01 6.91888273e-01
7.36951828e-01 -7.23381579e-01 5.85233688e-01 -1.59099236e-01
-1.84844923e+00 -6.41764775e-02 -9.27907750e-02 1.17330134e-01
1.80416659e-01 8.23096037e-01 -8.04628313e-01 1.62503767e+00
8.47202659e-01 1.22954619e+00 -7.53387034e-01 9.10145164e-01
-2.62622148e-01 4.79365289e-01 -2.99958944e-01 1.91930309e-01
9.10592377e-02 -3.26049209e-01 3.93390507e-01 8.84953618e-01
3.87570232e-01 7.71037489e-02 2.91711897e-01 6.28667355e-01
-3.50711107e-01 2.62245506e-01 -6.05068624e-01 -4.88986641e-01
5.39249718e-01 1.21905398e+00 -4.29634035e-01 -3.71180288e-02
-6.12497449e-01 7.82353163e-01 6.67706490e-01 6.59475803e-01
-1.03263879e+00 -3.15373808e-01 7.92483807e-01 2.12882608e-01
6.26343966e-01 -5.40912688e-01 4.14643019e-01 -1.42172647e+00
2.57393241e-01 -2.43162885e-01 9.46574986e-01 -5.28699160e-01
-1.50273192e+00 3.36335838e-01 2.34897345e-01 -1.01815200e+00
6.09593913e-02 -5.41171968e-01 -2.76082903e-01 5.23259342e-01
-1.70605493e+00 -1.53255379e+00 -1.44521922e-01 8.57995689e-01
-2.98473507e-01 -2.36736797e-02 7.51616776e-01 6.04219437e-01
-7.29501188e-01 7.18554616e-01 1.26483902e-01 3.33969414e-01
4.89430428e-01 -1.21919405e+00 1.90526351e-01 5.53514063e-01
4.06579942e-01 7.00185239e-01 2.38908038e-01 -8.44352067e-01
-1.84022415e+00 -1.34346282e+00 1.03073049e+00 -5.54546952e-01
1.49887991e+00 -2.29026183e-01 -9.47087228e-01 1.07609189e+00
-2.45181367e-01 5.90457380e-01 6.11953855e-01 8.32811892e-01
-9.01548386e-01 -4.36470091e-01 -7.12281466e-01 5.68938851e-01
1.43237245e+00 -6.52530491e-01 -4.47198987e-01 3.72395754e-01
7.88741827e-01 -2.08248973e-01 -1.83286512e+00 6.69353247e-01
5.07075012e-01 -4.43463862e-01 1.18735087e+00 -8.53969872e-01
1.34564072e-01 -4.32710767e-01 -1.04594067e-01 -1.43112636e+00
-5.37602067e-01 -5.41254342e-01 -1.04658449e+00 1.26045120e+00
2.50185698e-01 -9.50238705e-01 6.74106658e-01 1.56839877e-01
7.15968758e-02 -1.13152730e+00 -1.13428533e+00 -1.24489033e+00
-1.06676765e-01 -3.34436119e-01 5.75419307e-01 1.53284168e+00
2.14647263e-01 3.54531974e-01 -3.26083213e-01 3.47585857e-01
5.43776095e-01 2.76701152e-01 6.21153414e-01 -1.56692958e+00
-2.75224864e-01 -4.20120239e-01 -1.05750740e+00 -6.26955807e-01
2.43860379e-01 -1.29231513e+00 -1.02078354e+00 -1.80303824e+00
-1.01916902e-01 -5.99103630e-01 -5.69052100e-01 6.62343264e-01
-9.60514173e-02 -1.58026755e-01 9.71961394e-02 3.87615860e-02
-6.78128421e-01 9.07463312e-01 1.10432100e+00 -4.01467681e-01
-6.89450204e-02 -3.82080019e-01 -4.08959121e-01 2.85393357e-01
3.65211219e-01 -1.33084774e-01 -6.74774110e-01 -2.76034683e-01
7.19754398e-01 -2.62610875e-02 4.17334825e-01 -7.84447551e-01
3.34617734e-01 5.57056516e-02 8.71607512e-02 -5.40907621e-01
4.01529640e-01 -9.39302444e-01 5.54511309e-01 1.27531752e-01
3.58215392e-01 5.19599486e-03 2.29451463e-01 1.25268710e+00
-3.78600121e-01 4.37909186e-01 -3.40085966e-03 5.01895189e-01
-9.90252733e-01 8.11455965e-01 4.34652805e-01 9.87417810e-03
1.23618913e+00 4.24429812e-02 -7.09748149e-01 -1.86201021e-01
-9.85941410e-01 6.21677399e-01 8.24972913e-02 7.65229046e-01
5.14212787e-01 -1.79746878e+00 -5.82788467e-01 -3.30109239e-01
5.27230501e-01 -1.74740359e-01 3.73633116e-01 1.25512934e+00
-1.68650180e-01 5.41712046e-01 1.71896964e-01 -3.66198123e-01
-9.72025931e-01 8.34617853e-01 7.27082789e-02 -8.39504540e-01
-6.15743697e-01 7.99433529e-01 -1.95307046e-01 -3.06692183e-01
1.44104853e-01 -4.87378836e-01 -3.46960574e-01 4.89273340e-01
9.05867666e-02 5.07152200e-01 1.44754484e-01 -6.09080911e-01
-5.49998105e-01 6.17630303e-01 -2.55419821e-01 4.11240757e-01
1.49443889e+00 1.87624171e-01 -2.77067006e-01 6.74336910e-01
1.25722194e+00 -3.66936065e-02 -6.83738291e-01 -5.48599482e-01
1.85817108e-01 -3.47184360e-01 5.45955785e-02 -6.40661836e-01
-1.28410208e+00 4.12305176e-01 2.83334106e-01 4.40949678e-01
9.10465539e-01 1.78001106e-01 8.78044963e-01 2.71647125e-01
8.17474008e-01 -9.18024182e-01 3.52284610e-02 5.96543729e-01
8.49184930e-01 -9.97047961e-01 3.87154609e-01 -7.70409226e-01
-2.86964238e-01 9.87678409e-01 4.18833941e-01 1.24357708e-01
1.18696833e+00 -5.13582885e-01 -5.98958790e-01 -5.92945576e-01
-7.49615312e-01 -3.62697005e-01 6.42470300e-01 3.70253503e-01
1.44441441e-01 3.87571484e-01 -5.66469133e-01 6.26781940e-01
-2.09108323e-01 -4.98900078e-02 1.30611449e-01 7.23035276e-01
2.74923772e-01 -1.23122227e+00 -2.52201781e-02 5.38857341e-01
-1.62285924e-01 -5.93082979e-04 -2.38187179e-01 1.14884245e+00
1.44462883e-01 6.39122307e-01 -2.48440534e-01 -7.74805427e-01
3.81170630e-01 2.83315361e-01 5.60447335e-01 -4.06214982e-01
-2.17161521e-01 -4.66506690e-01 3.13514024e-01 -4.03795004e-01
-3.39557499e-01 -6.30930960e-01 -1.11128402e+00 -5.17768085e-01
-4.09006923e-01 2.89423436e-01 3.91155511e-01 6.61339939e-01
6.24120772e-01 7.88892508e-01 4.37398136e-01 -3.23502541e-01
-1.39613181e-01 -8.88892353e-01 -8.93674076e-01 5.10742307e-01
-1.52829081e-01 -1.31984508e+00 -1.68763787e-01 -1.41124740e-01] | [8.557193756103516, 7.905128002166748] |
969eb610-1528-47ff-8cf0-02098cfc8d42 | pix2nerf-unsupervised-conditional-p-gan-for-1 | null | null | http://openaccess.thecvf.com//content/CVPR2022/html/Cai_Pix2NeRF_Unsupervised_Conditional_p-GAN_for_Single_Image_to_Neural_Radiance_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Cai_Pix2NeRF_Unsupervised_Conditional_p-GAN_for_Single_Image_to_Neural_Radiance_CVPR_2022_paper.pdf | Pix2NeRF: Unsupervised Conditional p-GAN for Single Image to Neural Radiance Fields Translation | We propose a pipeline to generate Neural Radiance Fields (NeRF) of an object or a scene of a specific class, conditioned on a single input image. This is a challenging task, as training NeRF requires multiple views of the same scene, coupled with corresponding poses, which are hard to obtain. Our method is based on pi-GAN, a generative model for unconditional 3D-aware image synthesis, which maps random latent codes to radiance fields of a class of objects. We jointly optimize (1) the pi-GAN objective to utilize its high-fidelity 3D-aware generation and (2) a carefully designed reconstruction objective. The latter includes an encoder coupled with pi-GAN generator to form an auto-encoder. Unlike previous few-shot NeRF approaches, our pipeline is unsupervised, capable of being trained with independent images without 3D, multi-view, or pose supervision. Applications of our pipeline include 3d avatar generation, object-centric novel view synthesis with a single input image, and 3d-aware super-resolution, to name a few. | ['Luc van Gool', 'Dengxin Dai', 'Anton Obukhov', 'Shengqu Cai'] | 2022-01-01 | null | null | null | cvpr-2022-1 | ['3d-aware-image-synthesis'] | ['computer-vision'] | [ 5.62690318e-01 2.38277674e-01 2.42791638e-01 -3.79224688e-01
-1.05954814e+00 -6.31754100e-01 9.12171006e-01 -7.55784452e-01
1.94361255e-01 5.19725204e-01 4.16601449e-01 1.83906570e-01
4.31548566e-01 -1.07243037e+00 -1.23399580e+00 -6.88003242e-01
6.51861131e-01 6.58396542e-01 -2.37102769e-02 -9.67564210e-02
-4.24154773e-02 6.95866764e-01 -1.63914895e+00 1.55896068e-01
6.08707309e-01 8.54046166e-01 7.19908834e-01 1.01979899e+00
1.65525764e-01 9.00161684e-01 -4.29392099e-01 -3.51816416e-01
4.62657720e-01 -7.54511535e-01 -5.43690443e-01 4.82175231e-01
7.56834805e-01 -7.46335268e-01 -2.96041578e-01 7.38001764e-01
3.97829920e-01 1.83874056e-01 8.69658113e-01 -9.93964732e-01
-9.00610626e-01 1.76228270e-01 -5.56917071e-01 -3.78694147e-01
3.76422405e-01 3.62206995e-01 8.91605198e-01 -8.41582716e-01
8.75083208e-01 1.19736505e+00 2.96326280e-01 7.84157515e-01
-1.60598731e+00 -2.40921438e-01 -1.49132058e-01 -3.52673680e-01
-1.21331429e+00 -3.76181573e-01 9.92297530e-01 -6.12153292e-01
8.26437652e-01 -8.50387011e-03 7.57236660e-01 1.42211902e+00
1.53926071e-02 4.75747377e-01 1.11336291e+00 -2.35372007e-01
4.61624563e-01 9.02062505e-02 -7.71344960e-01 6.29225433e-01
-2.79857993e-01 7.77783841e-02 -6.40638173e-01 1.45898998e-01
1.36096251e+00 1.31258354e-01 -2.33403191e-01 -4.76290762e-01
-1.28786016e+00 7.78192341e-01 5.79096615e-01 -3.35279375e-01
-5.75620174e-01 4.60531265e-01 -2.89516419e-01 -1.91336170e-01
5.64478576e-01 4.91207510e-01 -4.29520309e-01 1.27364665e-01
-7.62235880e-01 4.61353660e-01 3.90805483e-01 1.10792387e+00
1.07427180e+00 4.16336119e-01 -1.43865436e-01 6.25988066e-01
1.90644413e-01 1.00195289e+00 4.07430045e-02 -1.18460608e+00
3.20053458e-01 2.52625853e-01 1.33525491e-01 -4.90557283e-01
6.88433200e-02 -3.90902787e-01 -7.68563449e-01 5.38035393e-01
-2.30463464e-02 -7.08238408e-02 -1.24305880e+00 1.97200799e+00
5.25056958e-01 1.74879506e-01 2.26010889e-01 1.07372510e+00
8.19516122e-01 1.01904941e+00 -3.97521406e-01 4.25675139e-03
1.02690959e+00 -1.14699697e+00 -3.05165112e-01 -4.51870620e-01
3.20972502e-02 -7.01563835e-01 1.14566195e+00 2.19165996e-01
-1.34386051e+00 -6.61207736e-01 -7.52401531e-01 -4.92158204e-01
-2.57703494e-02 8.57157260e-03 6.32055938e-01 2.36189410e-01
-9.35330033e-01 2.47340605e-01 -6.59385800e-01 -3.86597365e-02
3.99994314e-01 -2.08495051e-01 -5.20491600e-01 -3.49517971e-01
-7.61534512e-01 8.11141431e-01 4.55739610e-02 -1.55781657e-01
-1.65215218e+00 -9.74681795e-01 -1.02148807e+00 -1.04261622e-01
2.16440335e-01 -1.25949097e+00 1.18049765e+00 -9.40502584e-01
-1.73971033e+00 9.77595925e-01 -9.37516838e-02 -1.15188219e-01
3.33446711e-01 -2.02526763e-01 -6.54016994e-03 2.93658942e-01
3.34454060e-01 9.49130058e-01 1.01973152e+00 -1.64170003e+00
-1.36779636e-01 -2.52557158e-01 3.04922551e-01 5.46931684e-01
3.21722329e-01 -3.83729160e-01 -3.56502175e-01 -5.81656277e-01
-4.24287878e-02 -7.05061555e-01 -3.96169066e-01 1.14867941e-01
-5.21332443e-01 3.81264806e-01 6.77597046e-01 -6.80625081e-01
3.10175836e-01 -2.06571484e+00 6.22163832e-01 -1.42615139e-01
-5.56321181e-02 -3.25068086e-01 -2.24814877e-01 3.51252496e-01
6.63714483e-03 -1.08569719e-01 -3.70815217e-01 -5.56410432e-01
-6.56945258e-02 1.43643320e-01 -3.76247346e-01 2.38151729e-01
4.59311128e-01 9.20526385e-01 -9.67447340e-01 -1.15418769e-01
4.38373834e-01 8.87553155e-01 -8.42788160e-01 7.62182355e-01
-7.76305616e-01 1.02898467e+00 -3.39336872e-01 4.62867379e-01
7.04366326e-01 -4.88497436e-01 -1.47869855e-01 -5.92745543e-01
-3.58372182e-02 2.32047439e-01 -9.61127877e-01 2.23351383e+00
-8.29806566e-01 3.25770855e-01 -1.23619959e-01 -6.28251731e-01
9.80187476e-01 2.05525950e-01 4.88469481e-01 -4.18234676e-01
3.02241519e-02 4.06770408e-02 -6.23011529e-01 -3.81056517e-01
5.17134488e-01 -3.53583068e-01 -2.13563293e-01 5.86550772e-01
6.42585039e-01 -8.95154595e-01 -1.66477740e-01 3.46786350e-01
9.38591063e-01 7.53340662e-01 1.43949315e-01 2.56674409e-01
-7.51098664e-03 -2.33868957e-01 4.34316248e-01 4.62201059e-01
7.63172746e-01 1.47194600e+00 2.38697991e-01 -1.88598618e-01
-1.62247884e+00 -1.42062140e+00 8.78131688e-02 5.31589806e-01
1.37017548e-01 -2.19462469e-01 -6.69467330e-01 -5.71929991e-01
-2.76198804e-01 9.29071724e-01 -6.59735262e-01 -4.96045947e-02
-3.84033293e-01 -2.50964701e-01 6.96732327e-02 4.08734828e-01
4.49396849e-01 -9.31422889e-01 -8.18152606e-01 -3.13912705e-02
-3.62148255e-01 -1.31697261e+00 -5.83126247e-01 1.99615940e-01
-6.88809216e-01 -7.97126889e-01 -9.54282939e-01 -5.77787936e-01
8.72049153e-01 4.35615331e-01 1.42899084e+00 -4.15468693e-01
-2.89021730e-01 5.41792274e-01 -3.00173789e-01 -1.61152989e-01
-5.40899336e-01 -2.53623992e-01 -2.03681991e-01 1.82310641e-01
-3.80156994e-01 -1.05263638e+00 -7.63616502e-01 1.73565984e-01
-9.59980607e-01 7.28622437e-01 6.13437176e-01 8.19129229e-01
1.02683842e+00 -2.27429420e-01 1.95003629e-01 -8.61239612e-01
-2.72738129e-01 -4.83851135e-01 -6.46815240e-01 1.29239798e-01
-1.03855558e-01 -5.07204421e-03 6.87394917e-01 -3.69053543e-01
-1.43150985e+00 4.71621215e-01 -1.49290055e-01 -9.54874098e-01
-4.12851363e-01 4.27215686e-03 -6.02382243e-01 1.76448792e-01
9.22784686e-01 3.88504684e-01 -2.32409433e-01 -5.31531513e-01
9.05288398e-01 2.74289846e-01 8.27248394e-01 -4.86965418e-01
1.06163669e+00 5.92518330e-01 1.42404884e-01 -6.03103042e-01
-1.26749611e+00 1.62584949e-02 -7.45899081e-01 -2.67060459e-01
1.40649021e+00 -1.37491345e+00 -2.25579664e-01 4.01437193e-01
-1.23242748e+00 -7.64497042e-01 -6.42479658e-01 4.63470310e-01
-1.00900042e+00 -2.98157424e-01 -3.55801612e-01 -6.28076077e-01
-1.89081475e-01 -1.16572714e+00 1.69169950e+00 4.35597599e-01
1.27556413e-01 -7.00623930e-01 1.21280789e-01 7.48696506e-01
2.16121927e-01 6.20857060e-01 5.63635826e-01 3.10113966e-01
-1.08604193e+00 1.15185387e-01 -4.77503762e-02 5.06402731e-01
1.58619121e-01 -1.18556134e-01 -1.25829971e+00 -9.91631001e-02
-5.94571643e-02 -7.45757103e-01 5.92781901e-01 4.29875195e-01
1.24159181e+00 -3.87440950e-01 8.30646083e-02 1.07498276e+00
1.59317672e+00 -1.07447140e-01 6.54881835e-01 -1.74004987e-01
1.09563541e+00 3.87008309e-01 2.73282140e-01 5.24576306e-01
4.81556952e-01 8.23007345e-01 8.53955567e-01 -2.56050020e-01
-5.29125154e-01 -7.79950202e-01 4.35425639e-01 4.65358257e-01
-2.72031754e-01 -3.13603103e-01 -4.01809394e-01 4.72503930e-01
-1.37552404e+00 -1.05710709e+00 4.96893898e-02 2.09829259e+00
6.98839068e-01 -3.03889394e-01 -8.38253424e-02 -3.32066685e-01
3.49252611e-01 3.53775620e-01 -8.44611347e-01 -4.78568412e-02
-2.62881458e-01 2.88671970e-01 2.12052494e-01 5.21022260e-01
-7.30851233e-01 9.19242322e-01 5.76597881e+00 3.62807751e-01
-1.18657112e+00 2.86054045e-01 6.45275176e-01 -2.86262661e-01
-1.04171062e+00 2.16987610e-01 -6.69648349e-01 2.62964845e-01
5.85328400e-01 2.07069606e-01 6.56857252e-01 9.72866118e-01
-6.05859049e-02 -1.02789275e-01 -1.07432759e+00 1.04113150e+00
4.19989973e-01 -1.53661942e+00 1.81991503e-01 2.25648850e-01
1.31474030e+00 6.57954544e-04 9.93232355e-02 -3.96622829e-02
5.73842406e-01 -1.02414620e+00 1.07276928e+00 8.40807021e-01
1.23462617e+00 -7.45350063e-01 5.97466994e-03 3.95462751e-01
-8.84768486e-01 2.74986923e-01 -2.81593174e-01 4.70677942e-01
7.34931827e-01 6.94358766e-01 -6.00536823e-01 7.28914917e-01
8.08640242e-01 7.55433500e-01 -1.84922576e-01 4.10587013e-01
-6.62592530e-01 2.08345518e-01 -2.56429225e-01 5.12280822e-01
-6.27462193e-02 -2.64862686e-01 5.22457659e-01 6.66789949e-01
5.95840156e-01 2.50600100e-01 7.92075135e-03 1.43248594e+00
-1.99614927e-01 -3.67257416e-01 -7.84774661e-01 1.24985501e-01
2.61554927e-01 1.46515679e+00 -4.00445431e-01 -1.95494041e-01
-3.52579355e-01 1.43937778e+00 3.48510593e-01 4.55440968e-01
-9.03179944e-01 -5.23472279e-02 5.66076338e-01 2.78203219e-01
6.33471847e-01 -2.26748139e-01 -1.56998172e-01 -1.47340822e+00
-2.62920633e-02 -6.86704874e-01 -1.20761439e-01 -1.67184591e+00
-1.23122132e+00 6.50963247e-01 -1.83999110e-02 -1.43526995e+00
-5.51071644e-01 -2.56190270e-01 -5.81919432e-01 1.15679669e+00
-1.37799406e+00 -1.58241773e+00 -6.35087848e-01 6.29804790e-01
6.12022519e-01 -1.75362155e-01 9.39820707e-01 9.58297402e-02
-7.38137513e-02 1.92806587e-01 -2.93434650e-01 -2.16191545e-01
5.80588102e-01 -1.21435404e+00 5.32590747e-01 9.18145478e-01
3.12916011e-01 9.41047147e-02 6.22640073e-01 -5.88677704e-01
-1.51053858e+00 -1.37602246e+00 4.35970098e-01 -6.54822826e-01
6.46123523e-03 -6.81753099e-01 -5.50347507e-01 7.36487865e-01
1.55195549e-01 3.09244186e-01 3.92345905e-01 -2.32011661e-01
-4.61260438e-01 -9.51230377e-02 -1.15100789e+00 5.41113019e-01
1.22252393e+00 -6.63040698e-01 -1.26685336e-01 4.22000885e-01
9.44039702e-01 -7.77158678e-01 -1.08363998e+00 2.26821259e-01
2.79404074e-01 -1.18736577e+00 1.18408680e+00 -2.30009392e-01
1.16541946e+00 -6.33077741e-01 -4.99914885e-01 -1.54574895e+00
-3.13548833e-01 -4.09269482e-01 -1.77476913e-01 1.12129343e+00
3.76774848e-01 -7.68446848e-02 6.68013811e-01 3.08419794e-01
-3.55752140e-01 -4.08209532e-01 -3.90381753e-01 -4.64828700e-01
-2.13450179e-01 -2.65656382e-01 8.00835073e-01 8.16527486e-01
-8.22674572e-01 7.73540676e-01 -8.29932868e-01 4.20408279e-01
9.24361348e-01 4.83164430e-01 1.33827603e+00 -8.36798906e-01
-8.85890603e-01 -3.83946523e-02 -1.56302020e-01 -1.35687315e+00
-4.11318708e-03 -8.67689073e-01 2.98405319e-01 -1.76196015e+00
2.39554763e-01 -3.02960336e-01 5.34868360e-01 2.99042106e-01
7.32767507e-02 4.64492619e-01 7.31647238e-02 1.18061431e-01
-2.35787809e-01 1.08052373e+00 1.63456142e+00 6.48228899e-02
-1.96206179e-02 -1.16652124e-01 -7.65966117e-01 4.97657388e-01
4.42594171e-01 -2.96573400e-01 -7.78354287e-01 -7.42237210e-01
4.63993222e-01 4.58309948e-01 7.50304222e-01 -8.37577820e-01
-1.69359729e-01 -3.29144776e-01 6.95259213e-01 -7.25091875e-01
8.66445482e-01 -6.72901511e-01 7.80696273e-01 -2.24624038e-01
-2.75180995e-01 -3.12106252e-01 -2.44538099e-01 5.38858056e-01
-4.83186617e-02 -4.23599221e-03 8.04077148e-01 -5.38197279e-01
-5.79417646e-01 6.61506772e-01 7.43920058e-02 7.15461746e-02
7.97251821e-01 -7.71265179e-02 -9.19030309e-02 -6.00717783e-01
-5.69985449e-01 -1.46570057e-01 8.71828854e-01 3.76504928e-01
7.45730162e-01 -1.64898789e+00 -7.97795951e-01 3.68671685e-01
2.28038490e-01 9.68418598e-01 6.71071529e-01 2.09658012e-01
-4.80652839e-01 -1.78593650e-01 -2.64827013e-01 -8.21863234e-01
-8.58787239e-01 4.78877842e-01 4.03805017e-01 2.11517252e-02
-9.24086809e-01 8.79580975e-01 8.20573449e-01 -8.18570197e-01
-2.27812126e-01 1.41269296e-01 1.71483651e-01 -4.73567903e-01
4.97690111e-01 -1.40167013e-01 -1.49189934e-01 -6.39551699e-01
7.64982998e-02 7.35621810e-01 3.20264995e-01 -5.58954597e-01
1.62734342e+00 -6.94446117e-02 1.59726173e-01 5.79268277e-01
1.30671418e+00 -4.07948755e-02 -2.01198626e+00 -3.23559284e-01
-1.03881013e+00 -7.05737948e-01 1.76149130e-01 -7.56057739e-01
-1.36921406e+00 8.25320184e-01 1.46514609e-01 -3.48113716e-01
1.21121287e+00 3.32694530e-01 6.74590528e-01 -4.41334993e-02
6.18889689e-01 -7.12690949e-01 4.44340795e-01 3.39429706e-01
1.13675261e+00 -1.04244173e+00 -5.32427207e-02 -2.91600198e-01
-8.33829403e-01 9.45372283e-01 8.33452940e-01 -2.88109839e-01
5.34413636e-01 2.27631792e-01 -1.08470812e-01 -1.26396701e-01
-8.87690306e-01 -1.51923329e-01 3.55650038e-01 7.03850210e-01
2.36326948e-01 -1.32891508e-02 6.35673523e-01 2.45616004e-01
-4.71962899e-01 -1.57629564e-01 5.87404132e-01 6.96133077e-01
2.90301293e-02 -8.49827290e-01 -1.56047687e-01 1.62649050e-01
-4.77426238e-02 -4.66660671e-02 -1.39304519e-01 2.45348573e-01
1.35609210e-01 4.58296329e-01 2.88081676e-01 -4.96093929e-01
2.93727249e-01 -2.42527738e-01 9.18184042e-01 -8.45761120e-01
-9.89682600e-02 2.23875016e-01 -4.14128378e-02 -6.81409299e-01
-5.29592037e-01 -7.18182445e-01 -8.59587789e-01 -1.17173299e-01
-6.43322393e-02 -1.88428521e-01 7.08246350e-01 6.76290691e-01
4.11046952e-01 6.45033419e-01 1.08545530e+00 -1.44475532e+00
-1.04850784e-01 -7.90489018e-01 -6.09237552e-01 4.44768041e-01
3.93000931e-01 -5.87304592e-01 -2.88648844e-01 4.58458364e-01] | [9.260529518127441, -3.132289409637451] |
574b3131-688e-43b6-95fa-3aefc15d9dc1 | self-distillation-amplifies-regularization-in | 2002.05715 | null | https://arxiv.org/abs/2002.05715v3 | https://arxiv.org/pdf/2002.05715v3.pdf | Self-Distillation Amplifies Regularization in Hilbert Space | Knowledge distillation introduced in the deep learning context is a method to transfer knowledge from one architecture to another. In particular, when the architectures are identical, this is called self-distillation. The idea is to feed in predictions of the trained model as new target values for retraining (and iterate this loop possibly a few times). It has been empirically observed that the self-distilled model often achieves higher accuracy on held out data. Why this happens, however, has been a mystery: the self-distillation dynamics does not receive any new information about the task and solely evolves by looping over training. To the best of our knowledge, there is no rigorous understanding of this phenomenon. This work provides the first theoretical analysis of self-distillation. We focus on fitting a nonlinear function to training data, where the model space is Hilbert space and fitting is subject to $\ell_2$ regularization in this function space. We show that self-distillation iterations modify regularization by progressively limiting the number of basis functions that can be used to represent the solution. This implies (as we also verify empirically) that while a few rounds of self-distillation may reduce over-fitting, further rounds may lead to under-fitting and thus worse performance. | ['Peter L. Bartlett', 'Mehrdad Farajtabar', 'Hossein Mobahi'] | 2020-02-13 | null | http://proceedings.neurips.cc/paper/2020/hash/2288f691b58edecadcc9a8691762b4fd-Abstract.html | http://proceedings.neurips.cc/paper/2020/file/2288f691b58edecadcc9a8691762b4fd-Paper.pdf | neurips-2020-12 | ['l2-regularization'] | ['methodology'] | [ 8.27898011e-02 5.80908358e-01 1.46277443e-01 -2.63079226e-01
-2.23636240e-01 -4.73907173e-01 6.02776408e-01 -1.23156987e-01
-5.88469803e-01 1.08726501e+00 -3.31575908e-02 -2.91526020e-01
6.79824501e-02 -7.18692958e-01 -1.03671944e+00 -9.28793371e-01
-2.63846572e-02 4.31089312e-01 3.57757173e-02 -3.06278169e-01
2.33472884e-02 4.30978686e-01 -1.17161429e+00 -3.88753451e-02
8.50670338e-01 6.16675854e-01 6.40803948e-02 7.15029418e-01
3.14120129e-02 6.93541527e-01 -4.70836312e-01 -3.68617684e-01
4.82051224e-01 -5.53886175e-01 -1.04698825e+00 -1.55533955e-01
2.88327515e-01 -1.76917151e-01 -5.34111142e-01 9.80373979e-01
1.50642887e-01 4.70519751e-01 6.16766989e-01 -7.60858178e-01
-6.16751015e-01 6.75677657e-01 -7.52364844e-02 3.77845109e-01
-2.76681274e-01 3.20074379e-01 8.99869382e-01 -7.14581788e-01
4.11020100e-01 1.01699150e+00 8.72327983e-01 7.03414857e-01
-1.74551964e+00 -6.12469256e-01 3.55414338e-02 -1.21723749e-01
-1.27900028e+00 -3.74516755e-01 7.74583936e-01 -5.91708183e-01
1.00844967e+00 2.16549143e-01 7.29546666e-01 9.93635535e-01
1.39786243e-01 5.51165581e-01 1.14938569e+00 -5.90332866e-01
3.38942438e-01 4.89795506e-01 3.48921746e-01 7.42698014e-01
3.00501645e-01 4.73579794e-01 -2.19637230e-01 -1.09247945e-01
9.58960295e-01 -1.73692450e-01 -4.85441387e-01 -4.95149285e-01
-9.46322083e-01 9.71195042e-01 5.59078097e-01 6.91297829e-01
-3.63319457e-01 3.09593946e-01 2.98812836e-01 9.11808550e-01
5.15952468e-01 1.02121568e+00 -6.40396178e-01 -3.05506680e-02
-1.00221503e+00 3.38160574e-01 9.33001041e-01 3.77605170e-01
1.11048019e+00 3.54418129e-01 1.41877159e-01 6.19303286e-01
-1.23497963e-01 -2.53286362e-02 8.79825950e-01 -1.08727837e+00
1.15200877e-02 3.01934391e-01 1.35327935e-01 -6.47922933e-01
-4.90153313e-01 -8.45084846e-01 -7.89522171e-01 3.84046078e-01
6.30938292e-01 -4.79458690e-01 -8.40462625e-01 2.03970647e+00
-3.98172662e-02 4.82631683e-01 1.85569793e-01 8.69831443e-01
2.43709847e-01 6.44962132e-01 -1.33615538e-01 -4.00485605e-01
7.17213273e-01 -8.62030387e-01 -5.10602355e-01 -2.76996940e-01
8.58467460e-01 -2.59721041e-01 1.05729401e+00 3.68176579e-01
-1.22095573e+00 -7.23268688e-01 -1.17419195e+00 1.88797832e-01
-3.60245556e-01 -3.23127240e-01 6.30782664e-01 4.17220652e-01
-1.11373127e+00 1.23451447e+00 -1.04737616e+00 -3.17090601e-01
2.50453085e-01 4.91476536e-01 -3.86750817e-01 2.23872632e-01
-1.35806775e+00 1.36473095e+00 7.72924066e-01 5.75093031e-02
-6.47122204e-01 -7.70333230e-01 -6.36058390e-01 1.80103444e-02
2.75926739e-01 -7.81259954e-01 1.39479399e+00 -1.30022681e+00
-1.68112099e+00 8.20444524e-01 -1.52839020e-01 -8.51907074e-01
4.41991001e-01 -2.33307824e-01 -1.21694103e-01 -3.78613561e-01
-4.01468158e-01 4.33199823e-01 1.04065549e+00 -1.25454164e+00
-1.38411269e-01 -2.74269491e-01 1.20848835e-01 1.39395818e-01
-2.78915346e-01 -5.50617635e-01 -7.99911171e-02 -5.09895921e-01
2.83080727e-01 -1.00829720e+00 -3.96657169e-01 -2.84566492e-01
-1.54473588e-01 -2.44985800e-02 4.91130620e-01 -4.00253832e-01
1.28156662e+00 -2.09850121e+00 2.98486799e-01 2.00983778e-01
3.88028681e-01 4.23856646e-01 -7.79098868e-02 1.78066030e-01
-3.71731877e-01 3.20565775e-02 -5.14164269e-01 -2.68665463e-01
-5.44184484e-02 5.38120389e-01 -5.92779756e-01 6.28914654e-01
8.33435953e-02 1.20509338e+00 -7.75935709e-01 -7.30058923e-02
5.93763962e-02 5.03659427e-01 -8.19662869e-01 1.63398296e-01
-3.48925501e-01 7.22454607e-01 -1.99477851e-01 -1.69183776e-01
5.27657807e-01 -4.16688919e-01 1.20107181e-01 -1.29288509e-01
-1.37201414e-01 2.70194560e-01 -1.04998147e+00 1.51978362e+00
-3.54374737e-01 9.04781878e-01 -5.92432991e-02 -1.60785520e+00
8.84461164e-01 2.21664876e-01 4.03626800e-01 -4.86137211e-01
2.79291421e-01 3.06253731e-01 3.40147316e-01 -4.18801188e-01
1.98154792e-01 -6.14321828e-01 1.83745608e-01 3.92183602e-01
2.73224324e-01 -2.49948725e-01 -1.33912973e-02 -1.83334470e-01
1.06022787e+00 1.58947855e-01 2.72956222e-01 -2.77366936e-01
4.51192707e-01 1.19139478e-01 4.97495443e-01 9.35111225e-01
-1.07561948e-03 2.60357440e-01 4.20203865e-01 -6.27815902e-01
-1.27030265e+00 -9.57843304e-01 -3.53025019e-01 1.03925991e+00
-3.54908437e-01 -9.54604447e-02 -9.15009558e-01 -4.00316328e-01
1.23983756e-01 8.91970754e-01 -9.37657237e-01 -6.09756887e-01
-8.19403410e-01 -6.80881381e-01 4.50731218e-01 3.63521725e-01
5.54787993e-01 -9.59834397e-01 -5.91656089e-01 4.04362351e-01
3.11826706e-01 -8.03697824e-01 -1.73807263e-01 6.05138898e-01
-1.35492098e+00 -8.21509123e-01 -8.64848733e-01 -6.32959247e-01
7.46522903e-01 -2.64189839e-01 1.07991111e+00 9.02899951e-02
-9.22670439e-02 2.82666087e-01 9.49364156e-02 -1.13509648e-01
-7.05984414e-01 1.41307980e-01 1.88680321e-01 -1.60091877e-01
3.38035524e-01 -8.14236999e-01 -5.17910421e-01 2.63189506e-02
-7.83568323e-01 -2.65965331e-03 5.45580864e-01 1.07995176e+00
2.84194052e-01 6.74484968e-02 4.71719623e-01 -1.12877524e+00
8.08987617e-01 -4.04973477e-01 -5.55308282e-01 -2.90695690e-02
-8.55932653e-01 5.99280119e-01 8.21326911e-01 -6.18008971e-01
-8.53638947e-01 -3.00299469e-02 2.59973537e-02 -7.03630447e-01
-2.35652607e-02 6.91046953e-01 4.63901907e-01 -1.70324162e-01
1.04740143e+00 2.72170156e-01 1.39691398e-01 -5.71335196e-01
5.65470397e-01 1.74264759e-01 5.83454609e-01 -5.86665928e-01
8.72105658e-01 2.55273163e-01 -3.65343690e-02 -7.22952187e-01
-1.10805798e+00 3.70550975e-02 -7.01443255e-01 1.98463481e-02
6.32985294e-01 -5.33889830e-01 -7.59313226e-01 2.56869674e-01
-9.76530552e-01 -8.41678321e-01 -8.07681143e-01 5.12660444e-01
-6.63776457e-01 1.24550335e-01 -4.78661716e-01 -7.74752438e-01
-5.63400984e-02 -8.62656891e-01 2.09916249e-01 1.81350768e-01
-2.77964383e-01 -1.30640888e+00 4.00529265e-01 -6.69104308e-02
6.76901340e-01 1.21329546e-01 8.13911617e-01 -7.62285650e-01
-2.05945447e-01 -1.40999064e-01 1.53026357e-01 7.01508164e-01
-1.87543184e-02 -4.15634692e-01 -1.18800902e+00 -3.29751730e-01
4.93095726e-01 -3.65346730e-01 1.15006769e+00 3.05783272e-01
1.16892231e+00 -4.83874500e-01 -1.36234060e-01 5.55575669e-01
1.29341888e+00 6.72719404e-02 5.59731662e-01 3.87550443e-01
4.49062407e-01 3.83321941e-01 3.17635201e-02 1.42740563e-01
2.41476111e-02 4.72791910e-01 2.46595830e-01 -4.37399261e-02
9.90127921e-02 -9.20739025e-02 3.96121085e-01 8.98862183e-01
-3.01433980e-01 3.01317722e-01 -9.28770423e-01 3.48841548e-01
-1.80675209e+00 -9.85025823e-01 1.92271963e-01 2.34497213e+00
1.20157480e+00 4.02392030e-01 -1.82306115e-02 7.72863105e-02
4.92973000e-01 -3.36944032e-03 -9.62184668e-01 -5.61265528e-01
-1.14291064e-01 5.27481675e-01 5.27563453e-01 8.44525158e-01
-8.85628283e-01 1.02360666e+00 7.05077648e+00 3.40595931e-01
-1.47354639e+00 2.01139599e-01 5.66955686e-01 -4.21495922e-02
-9.41402838e-02 1.84180424e-01 -5.01980007e-01 3.21895182e-01
1.12364078e+00 -3.93212616e-01 7.62086511e-01 7.73122489e-01
-6.53296709e-02 5.17182574e-02 -1.33283746e+00 8.53270173e-01
-6.81754872e-02 -1.47030699e+00 -3.60857099e-01 -7.22766146e-02
8.30122828e-01 2.26298511e-01 5.94816208e-02 7.93867767e-01
4.11824077e-01 -1.07757878e+00 4.16438311e-01 8.57287288e-01
3.89814228e-01 -5.48603833e-01 4.79949772e-01 6.52228117e-01
-6.24364436e-01 -2.22905964e-01 -4.60176706e-01 -3.01454514e-01
-8.95561129e-02 5.11547327e-01 -8.33632350e-01 1.63365565e-02
2.53855258e-01 4.55932409e-01 -3.33344787e-01 9.04882252e-01
1.37546714e-02 8.64861369e-01 -2.13726044e-01 1.95369676e-01
2.73740381e-01 -3.42826098e-01 5.84404171e-01 1.11120260e+00
1.71590596e-01 2.54664749e-01 -7.33621567e-02 1.12896097e+00
-1.34384558e-02 -2.14525044e-01 -6.48112595e-01 -6.65684044e-02
3.14875618e-02 8.79024804e-01 -3.21650535e-01 -4.56085831e-01
-3.39636415e-01 9.88291800e-01 6.34962976e-01 7.29457378e-01
-6.19092345e-01 -3.01047295e-01 5.10093987e-01 2.73982793e-01
3.45181525e-01 -4.15967345e-01 -2.78081536e-01 -1.35397470e+00
-2.54153281e-01 -5.46089351e-01 4.89332713e-02 -6.58649445e-01
-1.19044673e+00 5.10319591e-01 -3.62304114e-02 -6.20602965e-01
-4.73513961e-01 -6.56144321e-01 -6.22887790e-01 9.38780665e-01
-1.39521265e+00 -4.56267029e-01 -5.00342287e-02 4.44313824e-01
2.78356969e-01 -1.33902773e-01 8.98233175e-01 -2.91551165e-02
-4.94892120e-01 5.50593674e-01 4.11421210e-01 6.72712326e-02
4.29559082e-01 -1.25409734e+00 1.94548056e-01 4.80295748e-01
9.85609964e-02 7.82669604e-01 1.07937765e+00 -3.89945567e-01
-1.18421042e+00 -8.91638875e-01 5.72664797e-01 -4.15793002e-01
7.96371043e-01 -2.85436869e-01 -1.56511211e+00 9.41710114e-01
9.14231092e-02 1.31400570e-01 4.54625785e-01 1.08034581e-01
-2.40836278e-01 -3.59008508e-03 -1.04029322e+00 6.28649473e-01
7.18840003e-01 -5.20187676e-01 -7.89195716e-01 2.20328346e-01
7.43865728e-01 -3.56119722e-01 -9.05022264e-01 3.06063205e-01
3.31541955e-01 -9.03100669e-01 7.98234165e-01 -8.90483260e-01
4.59433831e-02 1.97608415e-02 -4.20371890e-02 -1.57085574e+00
-4.61702615e-01 -8.79469395e-01 -5.50949156e-01 6.66234016e-01
5.13972580e-01 -1.05979717e+00 9.24306989e-01 9.83480394e-01
-1.60881102e-01 -8.74206185e-01 -9.18060064e-01 -1.01742029e+00
7.11329937e-01 -3.89032632e-01 3.08814377e-01 1.13938928e+00
1.67190880e-01 3.51095378e-01 -9.18330178e-02 -1.23092666e-01
4.86163974e-01 -4.79187779e-02 6.71222508e-01 -1.40326238e+00
-5.63894331e-01 -6.38307035e-01 -3.15113306e-01 -1.14692724e+00
4.19386119e-01 -1.02149796e+00 -1.40401080e-01 -1.11882341e+00
-6.10847585e-02 -5.54293215e-01 -4.32047278e-01 6.02063775e-01
6.39908900e-03 9.09380801e-03 1.08727463e-01 4.25612241e-01
-1.02869123e-01 4.84493613e-01 1.37996769e+00 9.85665172e-02
-5.04953086e-01 6.17962554e-02 -6.47804737e-01 7.09847867e-01
1.05451357e+00 -3.93559724e-01 -3.54155749e-01 -2.44906053e-01
2.44796857e-01 -4.24828473e-03 4.40853477e-01 -9.54065263e-01
1.95456997e-01 -2.13579923e-01 3.21108192e-01 9.71120149e-02
4.10410672e-01 -6.27989471e-01 3.07672583e-02 6.80714786e-01
-5.92213035e-01 -3.27839166e-01 4.30998445e-01 4.24160182e-01
8.36350545e-02 -5.86711109e-01 1.05389369e+00 -3.17822367e-01
-4.22507197e-01 2.19937135e-02 -3.21935475e-01 1.19653888e-01
8.51319909e-01 -2.15280309e-01 5.16706612e-03 -4.21476573e-01
-1.24067545e+00 -8.01805481e-02 4.06184703e-01 -2.23376323e-02
2.32112244e-01 -1.19953179e+00 -5.97259998e-01 4.77560282e-01
-4.36224103e-01 1.45234033e-01 -3.80780399e-02 8.00523281e-01
-1.67862356e-01 5.13796747e-01 2.59551052e-02 -5.15798151e-01
-5.77388942e-01 5.81561744e-01 8.18612278e-01 -3.47686797e-01
-8.28736544e-01 8.64696860e-01 3.32565695e-01 -3.70331794e-01
2.45599851e-01 -4.53586727e-01 1.10421658e-01 -2.25947186e-01
2.26832420e-01 1.81084931e-01 3.51436846e-02 -3.69614691e-01
6.15481324e-02 4.46529061e-01 -3.63703698e-01 -3.45600635e-01
1.46894598e+00 6.09779619e-02 6.45049363e-02 8.72123063e-01
1.24265563e+00 -4.10022289e-01 -1.59139514e+00 -5.23907721e-01
-1.09333135e-01 -9.11504552e-02 1.69365138e-01 -6.51586473e-01
-1.03708994e+00 9.56739068e-01 4.19835746e-01 5.02902687e-01
8.14775109e-01 3.58312987e-02 5.02531707e-01 8.25323105e-01
2.53429748e-02 -1.24125290e+00 9.77618769e-02 7.91502059e-01
1.04191339e+00 -1.08027589e+00 -1.93398714e-01 3.03196937e-01
-5.00226736e-01 1.14561141e+00 7.53489554e-01 -6.25063300e-01
9.11188185e-01 8.51455182e-02 -2.49528468e-01 -6.76843524e-02
-7.56068766e-01 -8.73865187e-02 2.08910301e-01 3.37411076e-01
4.64131087e-01 -1.55664176e-01 -7.49628758e-03 4.83017743e-01
-4.73667771e-01 1.87089041e-01 3.75857770e-01 8.32174242e-01
-6.54388666e-01 -9.26816165e-01 -3.86663467e-01 5.09867013e-01
-1.76652223e-01 1.60927847e-02 -1.23681881e-01 8.23462903e-01
-2.15428621e-02 4.24401969e-01 3.97608243e-02 -1.99181944e-01
2.08457425e-01 4.63897079e-01 7.52345681e-01 -4.94808376e-01
-5.66648781e-01 -1.01500332e-01 -2.92008162e-01 -3.02697897e-01
-1.59399033e-01 -3.92667234e-01 -1.43924177e+00 -3.43797833e-01
-3.52047414e-01 3.45247865e-01 6.61748469e-01 1.07853031e+00
2.49941945e-01 6.43355191e-01 4.31876183e-01 -9.17602420e-01
-9.30237651e-01 -1.03451359e+00 -4.26542550e-01 3.02469909e-01
7.12078154e-01 -7.34879911e-01 -6.70854807e-01 1.21096201e-01] | [7.830048084259033, 3.563427686691284] |
2b000fb1-b7a1-4d28-92e7-9ba9b5b0b421 | real-or-fake-spoofing-state-of-the-art-face | 1911.05351 | null | https://arxiv.org/abs/1911.05351v4 | https://arxiv.org/pdf/1911.05351v4.pdf | GANprintR: Improved Fakes and Evaluation of the State of the Art in Face Manipulation Detection | The availability of large-scale facial databases, together with the remarkable progresses of deep learning technologies, in particular Generative Adversarial Networks (GANs), have led to the generation of extremely realistic fake facial content, raising obvious concerns about the potential for misuse. Such concerns have fostered the research on manipulation detection methods that, contrary to humans, have already achieved astonishing results in various scenarios. In this study, we focus on the synthesis of entire facial images, which is a specific type of facial manipulation. The main contributions of this study are four-fold: i) a novel strategy to remove GAN "fingerprints" from synthetic fake images based on autoencoders is described, in order to spoof facial manipulation detection systems while keeping the visual quality of the resulting images; ii) an in-depth analysis of the recent literature in facial manipulation detection; iii) a complete experimental assessment of this type of facial manipulation, considering the state-of-the-art fake detection systems (based on holistic deep networks, steganalysis, and local artifacts), remarking how challenging is this task in unconstrained scenarios; and finally iv) we announce a novel public database, named iFakeFaceDB, yielding from the application of our proposed GAN-fingerprint Removal approach (GANprintR) to already very realistic synthetic fake images. The results obtained in our empirical evaluation show that additional efforts are required to develop robust facial manipulation detection systems against unseen conditions and spoof techniques, such as the one proposed in this study. | ['Hugo Proença', 'Ruben Vera-Rodriguez', 'João C. Neves', 'Vasco Lopes', 'Ruben Tolosana', 'Julian Fierrez'] | 2019-11-13 | null | null | null | null | ['steganalysis'] | ['computer-vision'] | [ 7.26974070e-01 2.54502714e-01 3.43739808e-01 -1.64082069e-02
-2.76918828e-01 -3.61842662e-01 8.06637287e-01 -8.07935596e-01
3.30912229e-03 6.77558362e-01 -3.78586724e-02 3.09954910e-03
6.14044443e-02 -8.68825972e-01 -9.54978108e-01 -8.36284816e-01
-1.20336123e-01 -4.00814749e-02 -2.02014104e-01 -6.92625880e-01
2.27187544e-01 9.89684880e-01 -1.97550559e+00 4.52193916e-01
4.80876565e-01 1.05380821e+00 -6.14207983e-01 5.24323046e-01
4.18815613e-01 6.20568871e-01 -1.01904202e+00 -1.04054737e+00
4.40746129e-01 -6.78528428e-01 -5.27501047e-01 3.13759714e-01
8.09716761e-01 -7.27402210e-01 -4.14783299e-01 1.21475983e+00
5.85631549e-01 -3.72284442e-01 5.28923750e-01 -1.57780039e+00
-7.62910604e-01 2.42946848e-01 -2.14221194e-01 -2.19557956e-01
3.00876945e-01 4.20883775e-01 1.96232632e-01 -8.00403714e-01
8.22709560e-01 1.41404128e+00 7.44823158e-01 9.91475642e-01
-8.72461557e-01 -1.09612620e+00 -5.14413655e-01 1.60177618e-01
-1.26918769e+00 -8.42754543e-01 1.14931524e+00 -2.93056279e-01
3.68945241e-01 2.94568956e-01 6.47399783e-01 1.99449229e+00
1.26493469e-01 4.53394294e-01 1.53264701e+00 -4.99009937e-01
-2.95580961e-02 3.40669543e-01 -6.20021701e-01 8.42235923e-01
5.40517032e-01 5.73064864e-01 -4.62987095e-01 -1.54128775e-01
8.28677654e-01 -3.24900806e-01 -2.36223117e-01 -2.18905419e-01
-6.60479844e-01 8.43466878e-01 1.20942242e-01 5.36115408e-01
-5.30679047e-01 4.04602110e-01 4.99750495e-01 5.18071532e-01
3.86029512e-01 2.75709003e-01 7.35865310e-02 1.03586197e-01
-1.07360053e+00 3.13316703e-01 7.59888291e-01 8.39500010e-01
4.57162857e-01 5.96697628e-01 1.67298615e-01 5.72378337e-01
8.61818120e-02 6.95352554e-01 3.27780962e-01 -5.47725379e-01
3.06661040e-01 1.03897341e-01 2.48668604e-02 -1.77915430e+00
5.85240126e-03 -1.49805084e-01 -8.78676772e-01 4.14163560e-01
3.42620075e-01 -2.59470969e-01 -8.88841391e-01 1.67711186e+00
1.99746951e-01 2.55351484e-01 4.22008894e-02 6.29877031e-01
7.42237508e-01 1.31710768e-01 -1.75656423e-01 -9.84691177e-03
1.27205551e+00 -4.90559548e-01 -8.80324543e-01 2.55480260e-01
2.67960876e-01 -8.14926624e-01 5.98319411e-01 6.87850475e-01
-8.60320628e-01 -5.44909000e-01 -1.16421127e+00 3.23463142e-01
-7.30333269e-01 7.05231130e-02 7.24866390e-01 1.68787444e+00
-1.14628172e+00 6.72114193e-01 -4.87523764e-01 -3.77586901e-01
8.96855891e-01 4.81724501e-01 -6.70873165e-01 -7.59739205e-02
-1.30052280e+00 9.74183917e-01 1.56379804e-01 4.75925952e-01
-1.35647726e+00 -3.32678974e-01 -6.63355112e-01 -1.01528637e-01
3.85488778e-01 -1.98093325e-01 6.19119048e-01 -1.67766547e+00
-1.70774651e+00 1.04881597e+00 2.43248478e-01 -3.69176418e-01
9.18691337e-01 -1.10434942e-01 -9.20074642e-01 3.77176136e-01
-4.57108378e-01 5.08739173e-01 1.66359198e+00 -1.59978282e+00
8.04357231e-02 -3.76769125e-01 -1.37148529e-01 -5.60848594e-01
-4.97418910e-01 3.13397795e-01 9.86882970e-02 -8.95330608e-01
-2.97383755e-01 -9.88940716e-01 3.28460246e-01 -6.02067634e-02
-4.91690159e-01 3.21064144e-01 1.14588499e+00 -9.83202517e-01
8.02508533e-01 -2.00567460e+00 -1.49135021e-02 2.88435489e-01
-3.79708782e-02 7.86459744e-01 -1.89361259e-01 6.14374220e-01
-4.44017470e-01 3.13527644e-01 -2.26711959e-01 -3.81597936e-01
-1.12544820e-01 2.80119702e-02 -4.73236233e-01 9.28825736e-01
4.57972467e-01 8.96645248e-01 -7.34018743e-01 -3.02323729e-01
2.87010193e-01 7.53385961e-01 -3.97591352e-01 1.35269448e-01
4.68051098e-02 4.45757747e-01 -7.81791881e-02 1.14389575e+00
1.08682954e+00 4.83639240e-01 -6.16048090e-02 -1.88393593e-01
3.28479648e-01 -3.52030516e-01 -8.17957520e-01 1.05549359e+00
-2.58457124e-01 6.88910484e-01 2.28153080e-01 -1.03372586e+00
9.33743596e-01 5.68686008e-01 4.50534105e-01 -5.34411550e-01
5.67980587e-01 3.84672791e-01 -7.70237446e-02 -8.00384223e-01
3.01217526e-01 -2.07089081e-01 8.26281160e-02 1.83607355e-01
4.10599053e-01 -8.40668008e-02 -3.29479873e-01 -2.89433122e-01
8.77913058e-01 3.41115057e-01 -1.06235698e-01 -1.63265895e-02
6.80389583e-01 -3.98815930e-01 3.05935927e-02 6.46401823e-01
-3.38587075e-01 4.61703092e-01 6.83566272e-01 -3.28395426e-01
-1.03764260e+00 -5.96270680e-01 6.30016327e-02 4.09675002e-01
-1.20907947e-01 9.60141569e-02 -1.27565801e+00 -8.70275497e-01
1.84503719e-01 2.96563148e-01 -8.83717597e-01 -3.81033361e-01
-7.72606790e-01 -7.28383541e-01 1.43312609e+00 -8.61098319e-02
8.02070379e-01 -1.30776155e+00 -4.63161498e-01 -8.53153914e-02
3.24928239e-02 -1.20611107e+00 2.33315393e-01 -4.12311077e-01
-5.87710440e-01 -1.24504352e+00 -7.45465219e-01 -5.34674287e-01
6.52287483e-01 -7.57691041e-02 7.39919305e-01 3.67914140e-01
-4.51319307e-01 2.74719179e-01 -4.72992629e-01 -5.11871457e-01
-1.10060036e+00 -3.43266159e-01 4.19751890e-02 5.78650117e-01
1.96621329e-01 -5.75842679e-01 -3.62383693e-01 3.70992929e-01
-1.25367594e+00 -3.56026322e-01 6.73746705e-01 9.02625680e-01
-2.38868326e-01 8.89346302e-02 6.00334883e-01 -1.04007053e+00
5.02626657e-01 -4.87293184e-01 -5.70291638e-01 1.28170609e-01
-4.67008084e-01 -2.77736455e-01 5.40582895e-01 -5.12268066e-01
-9.91954803e-01 -2.94777691e-01 -5.38231313e-01 -6.32202089e-01
-4.85583186e-01 -1.60584047e-01 -4.00700569e-01 -9.07682657e-01
7.27886081e-01 3.90227824e-01 4.37174946e-01 -2.17224374e-01
1.88615352e-01 7.23722100e-01 5.14244735e-01 -2.46781826e-01
1.12853587e+00 8.45034480e-01 4.02601391e-01 -1.19103646e+00
-1.70121878e-01 1.99948758e-01 -4.73064780e-01 -5.57826161e-01
4.70723599e-01 -5.47055304e-01 -9.37987685e-01 1.26886463e+00
-1.14430153e+00 -1.40048489e-02 1.62881926e-01 -7.21611008e-02
-6.69042885e-01 7.65666723e-01 -4.71735656e-01 -9.72398639e-01
-2.33370900e-01 -1.20367169e+00 1.09640849e+00 -6.35775402e-02
8.93287435e-02 -8.04689705e-01 -2.98006475e-01 6.33468330e-01
6.90674543e-01 1.09759068e+00 5.41895688e-01 -4.26731318e-01
-5.43052733e-01 -4.57948744e-01 -2.06094943e-02 9.53586400e-01
3.12393486e-01 8.56097341e-02 -1.40384209e+00 -3.68603528e-01
2.09765598e-01 -4.53383118e-01 7.09718645e-01 -1.75402850e-01
1.14332712e+00 -5.30782819e-01 -3.25997025e-02 7.14739203e-01
1.52376664e+00 7.59918764e-02 1.19848204e+00 9.36202779e-02
6.81837916e-01 8.07062268e-01 3.26768607e-01 3.04563940e-01
-4.39701110e-01 8.05829167e-01 1.06233525e+00 -1.67364821e-01
-4.77058470e-01 -2.15626180e-01 5.75835586e-01 1.52924761e-01
-4.23043042e-01 -4.81442362e-01 -2.78208047e-01 3.80042315e-01
-1.29590249e+00 -1.20158887e+00 -1.73485857e-02 2.13006496e+00
2.79576272e-01 -1.92258865e-01 1.10123195e-01 4.63128656e-01
8.01538467e-01 3.47375691e-01 -1.36624470e-01 -6.42258644e-01
-3.79747123e-01 6.69299901e-01 5.76636553e-01 1.51995137e-01
-1.24024427e+00 1.14190471e+00 5.98315811e+00 8.95283639e-01
-1.40803671e+00 3.58675838e-01 3.92984122e-01 3.04734498e-01
5.98263592e-02 -4.20074672e-01 -5.22967696e-01 7.62577593e-01
9.98072863e-01 6.29640222e-01 6.41069293e-01 6.56372547e-01
1.39050931e-01 1.98578328e-01 -5.84166288e-01 1.00622773e+00
7.94256389e-01 -1.16937041e+00 2.57893384e-01 4.00800139e-01
6.33948267e-01 -5.11888862e-01 4.09095049e-01 9.32796113e-03
-1.25894889e-01 -1.32743061e+00 6.74951971e-01 4.65931863e-01
9.74451005e-01 -9.36322629e-01 8.91587079e-01 2.90178917e-02
-3.17744851e-01 -8.46157372e-02 -3.26416939e-01 3.03688437e-01
6.05269521e-02 4.82985079e-01 -8.47732663e-01 6.51673377e-01
4.05282825e-01 3.05416137e-01 -4.87626195e-01 4.63772833e-01
-4.36333328e-01 6.90621078e-01 -1.06333613e-01 -9.22264345e-03
1.10064700e-01 4.37848717e-02 6.22474790e-01 1.15913785e+00
3.72221321e-01 -3.22453648e-01 -7.49307275e-01 9.27662909e-01
-3.19457769e-01 -2.02790424e-02 -1.01599157e+00 -2.55758077e-01
8.29931721e-02 1.07448018e+00 -4.79035407e-01 -7.38903508e-02
-1.44539058e-01 1.19203150e+00 -1.69739634e-01 1.17889747e-01
-1.06190455e+00 -2.62499690e-01 7.44329214e-01 1.94633409e-01
3.23194921e-01 7.35002607e-02 9.41293165e-02 -1.09953678e+00
1.53533414e-01 -1.32492888e+00 -4.53805253e-02 -6.66625619e-01
-1.01870799e+00 6.28203094e-01 -9.95356664e-02 -1.04254627e+00
-1.95852920e-01 -7.76068389e-01 -1.79351777e-01 6.92957997e-01
-1.58137345e+00 -1.53359616e+00 -4.83381301e-01 7.09629297e-01
1.42667368e-01 -3.98829311e-01 9.43645477e-01 4.95504469e-01
-2.57035196e-01 1.01065564e+00 -1.29250437e-01 3.31660897e-01
5.97348213e-01 -4.64367092e-01 3.62898737e-01 8.70474756e-01
-6.17743470e-02 4.67515320e-01 7.53686011e-01 -6.52038395e-01
-1.50694823e+00 -9.49109852e-01 5.85129797e-01 -4.01047915e-01
4.87013847e-01 -6.63680255e-01 -6.22039855e-01 4.77636278e-01
1.91663563e-01 -4.31489758e-02 3.63348991e-01 -7.07707703e-01
-4.31949973e-01 -5.08797392e-02 -1.85846436e+00 4.50287253e-01
1.07778358e+00 -5.68530142e-01 -2.56527305e-01 4.23592538e-01
1.84345454e-01 -1.87689677e-01 -6.73905313e-01 5.46041369e-01
9.29958522e-01 -1.48103535e+00 9.40692544e-01 -6.28346503e-01
5.82201540e-01 7.44929612e-02 3.06709688e-02 -1.04786193e+00
2.57848501e-01 -9.99004126e-01 -1.46872073e-01 1.18178642e+00
-1.43482968e-01 -8.58377397e-01 9.98270750e-01 -1.17390545e-03
1.66512370e-01 -6.27267361e-01 -9.79591250e-01 -8.14128935e-01
-1.03675820e-01 -2.69314826e-01 7.99264908e-01 1.05658770e+00
-6.15476966e-01 -6.86141431e-01 -1.07056808e+00 3.14227462e-01
7.12963641e-01 -4.36323315e-01 1.08653450e+00 -9.63268459e-01
-1.28861144e-01 -2.40187123e-01 -9.94493425e-01 -3.01332146e-01
3.92078042e-01 -3.60480040e-01 -3.04434687e-01 -3.99084806e-01
-2.05221355e-01 -4.69819829e-02 1.18537277e-01 3.42102051e-01
3.58503312e-01 8.77719998e-01 1.78305022e-02 -5.57373986e-02
3.38798285e-01 3.30139607e-01 1.29038799e+00 -2.92006712e-02
4.68378872e-01 -5.28727435e-02 -5.13880253e-01 6.26548588e-01
5.37478566e-01 -5.78045309e-01 -1.38858795e-01 -1.32691475e-05
5.03395386e-02 -1.34256727e-03 8.38514388e-01 -1.23435771e+00
-2.77453303e-01 1.24314189e-01 4.26042169e-01 -1.16450183e-01
6.46946847e-01 -9.74176407e-01 3.65755141e-01 7.04990506e-01
-2.37044264e-02 -2.09214091e-01 1.96641833e-01 2.94512898e-01
-2.16821343e-01 -1.95529461e-01 1.19184899e+00 -2.08416134e-01
-4.58739787e-01 7.81054720e-02 -2.87606269e-01 -4.94517177e-01
1.13627589e+00 -6.05360389e-01 -3.78247172e-01 -5.84511161e-01
-6.27397537e-01 -7.75319576e-01 6.53092921e-01 5.60872734e-01
6.64883614e-01 -1.07984936e+00 -7.46958554e-01 5.94014287e-01
-6.05152138e-02 -7.49194324e-01 3.31620485e-01 6.28751695e-01
-8.76669526e-01 2.72748649e-01 -7.04964042e-01 -1.32172391e-01
-1.49259436e+00 8.17656875e-01 3.92758578e-01 7.14029893e-02
-2.35133454e-01 7.10813165e-01 -2.81281024e-01 -1.17881581e-01
-5.11111915e-02 1.05031192e-01 -6.87935352e-02 -4.19493951e-02
3.27803493e-01 4.75552529e-01 4.28242654e-01 -1.14498818e+00
-1.85277477e-01 4.35637146e-01 4.05670702e-01 8.12611580e-02
1.31899047e+00 1.40097305e-01 -3.16866338e-01 -3.80939335e-01
1.32425213e+00 2.50063390e-01 -8.43352854e-01 3.05866867e-01
-2.44927287e-01 -7.73023665e-01 -1.84301853e-01 -7.85474241e-01
-1.43119442e+00 7.70017445e-01 7.41044879e-01 3.43499154e-01
1.23602664e+00 -3.58474672e-01 7.56646812e-01 1.71491224e-02
6.36886299e-01 -7.13845313e-01 2.21483290e-01 9.40963030e-02
1.04793274e+00 -9.66062605e-01 -6.43394589e-02 -7.30753660e-01
-2.39067271e-01 1.17895257e+00 1.82369709e-01 -4.21710014e-01
4.80913758e-01 9.25780684e-02 -5.95396981e-02 -3.76265645e-01
-3.64788659e-02 1.89428389e-01 -2.18004603e-02 8.15371990e-01
-4.97678779e-02 -1.35046646e-01 -1.59193993e-01 -1.35081662e-02
-2.82194376e-01 2.31290907e-01 7.33262360e-01 8.83826971e-01
1.48447648e-01 -1.18255007e+00 -8.24374855e-01 1.10620722e-01
-7.73876727e-01 1.37320057e-01 -8.04021716e-01 1.16060376e+00
6.01688564e-01 1.00224674e+00 -4.59801167e-01 -4.96873856e-01
2.02706978e-01 1.10289000e-01 7.91506469e-01 -8.71128663e-02
-8.09808433e-01 -3.81174266e-01 9.93905440e-02 -7.16338634e-01
-6.28846526e-01 -5.72379649e-01 -2.04282224e-01 -5.93999624e-01
-5.71587920e-01 -3.16946089e-01 1.13083935e+00 8.16888094e-01
3.53846192e-01 2.09934428e-01 8.31249297e-01 -1.07463396e+00
-6.32556379e-01 -1.06564307e+00 -7.00469375e-01 7.44856358e-01
3.99347365e-01 -9.03914809e-01 -6.87702298e-01 5.85353114e-02] | [12.620269775390625, 1.077177882194519] |
3cbaf27a-44f7-459d-9512-7c1ba8e59fe2 | joint-2d-3d-multi-task-learning-on-cityscapes | 2304.00971 | null | https://arxiv.org/abs/2304.00971v3 | https://arxiv.org/pdf/2304.00971v3.pdf | Joint 2D-3D Multi-Task Learning on Cityscapes-3D: 3D Detection, Segmentation, and Depth Estimation | This report serves as a supplementary document for TaskPrompter, detailing its implementation on a new joint 2D-3D multi-task learning benchmark based on Cityscapes-3D. TaskPrompter presents an innovative multi-task prompting framework that unifies the learning of (i) task-generic representations, (ii) task-specific representations, and (iii) cross-task interactions, as opposed to previous approaches that separate these learning objectives into different network modules. This unified approach not only reduces the need for meticulous empirical structure design but also significantly enhances the multi-task network's representation learning capability, as the entire model capacity is devoted to optimizing the three objectives simultaneously. TaskPrompter introduces a new multi-task benchmark based on Cityscapes-3D dataset, which requires the multi-task model to concurrently generate predictions for monocular 3D vehicle detection, semantic segmentation, and monocular depth estimation. These tasks are essential for achieving a joint 2D-3D understanding of visual scenes, particularly in the development of autonomous driving systems. On this challenging benchmark, our multi-task model demonstrates strong performance compared to single-task state-of-the-art methods and establishes new state-of-the-art results on the challenging 3D detection and depth estimation tasks. | ['Dan Xu', 'Hanrong Ye'] | 2023-04-03 | null | null | null | null | ['monocular-depth-estimation'] | ['computer-vision'] | [ 1.69809029e-01 4.74364683e-02 -2.11484641e-01 -3.84299517e-01
-1.14966261e+00 -5.96614838e-01 9.79216158e-01 -1.23158775e-01
-4.25748259e-01 1.68412611e-01 -1.84852734e-01 -4.35324252e-01
-2.07211003e-02 -4.51513052e-01 -6.40439272e-01 -7.00488985e-01
5.78909479e-02 7.60896385e-01 4.48250473e-01 -1.18292250e-01
9.53586847e-02 5.66258252e-01 -1.85839510e+00 4.03912425e-01
6.51719928e-01 1.09199739e+00 7.21916914e-01 7.78601408e-01
3.39487717e-02 5.80815792e-01 -4.74593550e-01 -4.61641280e-03
4.63731676e-01 3.90135974e-01 -6.17339373e-01 3.46191645e-01
9.25052106e-01 -2.60460734e-01 -1.74239844e-01 7.27825463e-01
5.61147690e-01 2.82879233e-01 9.68064845e-01 -1.54317069e+00
-3.35383505e-01 -3.29629779e-01 -7.00831652e-01 2.80756623e-01
-3.18472907e-02 4.58766550e-01 9.82912302e-01 -9.31970656e-01
4.48285222e-01 1.40896523e+00 5.23787558e-01 4.84910756e-01
-1.10937726e+00 -4.68841732e-01 6.27043724e-01 3.34277213e-01
-1.21818650e+00 -3.02734852e-01 6.34351134e-01 -8.33041966e-01
1.74984515e+00 -2.36276641e-01 4.64544237e-01 1.34454346e+00
2.83195198e-01 1.41873693e+00 9.54899192e-01 1.18090934e-03
-1.11571133e-01 1.60902426e-01 7.77071938e-02 6.38652563e-01
7.68021643e-02 6.51511729e-01 -3.45645338e-01 5.08366823e-01
5.61996043e-01 -2.61928201e-01 2.98255473e-01 -6.45022035e-01
-1.10519934e+00 7.91700244e-01 3.11704665e-01 -3.44091281e-02
-2.06325591e-01 1.21228732e-01 5.67029119e-01 4.98551652e-02
8.88683498e-01 1.41309097e-01 -7.25445926e-01 1.00100286e-01
-7.86666632e-01 5.17059803e-01 3.23480904e-01 9.97733116e-01
8.86891484e-01 3.63359302e-01 -3.94320995e-01 9.83773470e-01
3.66440654e-01 8.24065745e-01 5.38496599e-02 -8.75909328e-01
8.79310608e-01 5.23600817e-01 7.39747882e-02 -6.11008048e-01
-8.05813670e-01 -4.65071499e-01 -6.96764112e-01 6.36282444e-01
2.62361586e-01 -1.50120467e-01 -1.11219442e+00 1.61758900e+00
2.88694412e-01 -8.35065357e-03 1.07505165e-01 8.96846771e-01
1.28045583e+00 7.68529117e-01 2.50962317e-01 4.19092000e-01
1.24911773e+00 -1.38408875e+00 -3.70081037e-01 -1.05677366e+00
8.31600904e-01 -5.58277130e-01 7.72836864e-01 2.49945074e-01
-9.42335486e-01 -1.13013363e+00 -9.73703802e-01 -3.66140932e-01
-7.82799482e-01 3.44248861e-01 6.71962738e-01 5.37737250e-01
-1.16333938e+00 -1.82490081e-01 -4.07753617e-01 -2.25492522e-01
4.74943936e-01 2.12029204e-01 -3.69286329e-01 -1.95231706e-01
-9.86807823e-01 1.43522251e+00 4.54122961e-01 -5.48529476e-02
-1.70259941e+00 -7.45579600e-01 -1.29535055e+00 -2.24978313e-01
5.13105392e-01 -8.06377053e-01 1.47658587e+00 -4.59782183e-01
-1.07082856e+00 1.55818248e+00 -2.31788680e-01 -4.27477926e-01
5.08209705e-01 -2.41843328e-01 -1.48664549e-01 -5.46501651e-02
3.76890987e-01 1.36192107e+00 1.03800738e+00 -1.45329368e+00
-1.14460111e+00 -4.47095424e-01 4.71154042e-02 6.18523121e-01
2.45586708e-01 -3.33092064e-01 -6.44275367e-01 -4.26124573e-01
-1.13420323e-01 -8.12115848e-01 -3.12695354e-01 8.40331614e-02
-2.46646971e-01 -5.87246954e-01 1.01035416e+00 -2.18083560e-01
4.06865329e-01 -2.22263432e+00 2.02732995e-01 -3.89291376e-01
4.85641748e-01 3.19168091e-01 -5.17713428e-01 -1.12727964e-02
-1.30187139e-01 -2.05477431e-01 -1.77314617e-02 -1.01074541e+00
1.09665096e-01 1.67616099e-01 -9.38722864e-02 4.17482913e-01
4.02152181e-01 9.94775474e-01 -8.47139716e-01 -4.18993533e-01
8.66330445e-01 2.29992449e-01 -2.71293193e-01 1.95337445e-01
-4.09035295e-01 3.88043314e-01 -4.21420246e-01 7.06562817e-01
8.07061851e-01 -1.39928132e-01 -3.98026019e-01 -1.15294471e-01
-3.16643119e-01 2.89953172e-01 -8.56311083e-01 1.76232183e+00
-7.62690783e-01 9.24103260e-01 1.28511563e-01 -1.18909109e+00
1.12447381e+00 1.80313990e-01 5.87028921e-01 -1.08109701e+00
7.40161389e-02 8.76255855e-02 -4.04033124e-01 -5.76427579e-01
5.83516777e-01 1.18136339e-01 -2.71260172e-01 2.83332080e-01
1.87045231e-01 -7.53879070e-01 1.25808001e-01 -2.51163840e-02
6.68607831e-01 3.61037821e-01 1.48028806e-01 -2.85711795e-01
4.30061519e-01 1.64706528e-01 2.94423521e-01 7.48790503e-01
-6.03773296e-01 4.35815841e-01 3.49686742e-01 -7.53285289e-01
-9.96560752e-01 -1.14896679e+00 -1.58026457e-01 1.16380775e+00
3.13854873e-01 1.52769852e-02 -2.54132688e-01 -9.06152785e-01
4.50939506e-01 7.88868845e-01 -9.48291123e-01 -6.07128628e-02
-2.46878982e-01 -7.36606836e-01 6.24172211e-01 5.96825242e-01
6.89485967e-01 -8.72542202e-01 -7.80775666e-01 4.65487130e-02
-2.33959198e-01 -1.67352903e+00 -2.08983585e-01 7.54384160e-01
-7.01703727e-01 -1.16689682e+00 -6.74233973e-01 -9.59928393e-01
2.46664912e-01 8.82950187e-01 1.23461103e+00 -5.02638698e-01
-2.21566409e-01 5.22353351e-01 -1.07869156e-01 -7.13358700e-01
-1.08175546e-01 3.08891177e-01 -2.91210115e-01 -2.17987508e-01
5.63602448e-01 -1.37795538e-01 -2.76102304e-01 4.86912936e-01
-5.30012608e-01 4.40882295e-01 7.11806715e-01 5.37712276e-01
7.63196468e-01 -2.18857661e-01 5.63336730e-01 -3.18750739e-01
1.56801805e-01 -4.78492051e-01 -9.00133967e-01 -1.45435771e-02
-4.11788493e-01 -1.21575318e-01 9.42412615e-02 -2.28578314e-01
-1.13720143e+00 3.55598629e-01 -2.53492296e-01 -5.32641947e-01
-5.62121212e-01 2.21066326e-02 -2.64042951e-02 -3.91181968e-02
6.44556344e-01 3.49334180e-01 -3.86275463e-02 -3.44372332e-01
3.15713376e-01 4.69423920e-01 4.08777773e-01 -4.20839548e-01
7.11132228e-01 4.84890401e-01 2.15030923e-01 -1.04017818e+00
-1.09768808e+00 -8.07246625e-01 -9.03260052e-01 -4.32131857e-01
1.26929390e+00 -1.51891387e+00 -5.74066579e-01 7.63968885e-01
-1.40198004e+00 -8.42242539e-01 -4.92523275e-02 3.09104651e-01
-8.68804157e-01 1.61870674e-03 -7.96605572e-02 -7.71346748e-01
7.72699341e-02 -1.49518108e+00 1.75750566e+00 -3.25278968e-01
5.52426241e-02 -1.22710013e+00 -5.04994839e-02 6.34100020e-01
1.67262569e-01 1.98228404e-01 9.23356354e-01 -4.44718182e-01
-8.84534717e-01 -3.49639426e-03 -6.93793893e-01 3.44947726e-01
-1.20529063e-01 -3.75145227e-01 -1.34519339e+00 -1.69355452e-01
-1.78804114e-01 -6.56744182e-01 1.35455227e+00 8.37076724e-01
1.06628096e+00 4.89374459e-01 -3.69910508e-01 7.53802359e-01
1.16474783e+00 8.33515301e-02 3.90502363e-01 5.46269238e-01
8.22072566e-01 9.25904095e-01 7.66668260e-01 2.28894055e-01
9.94931817e-01 9.10406053e-01 9.70704913e-01 -3.94356102e-01
-4.65057135e-01 -1.08059876e-01 3.12011003e-01 1.67194113e-01
2.69911945e-01 -3.20250183e-01 -1.07630384e+00 8.95064533e-01
-1.88445735e+00 -6.82322025e-01 -3.80046874e-01 1.82526803e+00
1.39329672e-01 3.30363125e-01 1.67649865e-01 -1.63088720e-02
3.51339042e-01 7.04168677e-01 -8.16240072e-01 -3.81426990e-01
-2.64790267e-01 -3.27210784e-01 6.04008079e-01 5.10540307e-01
-1.64603770e+00 1.30792761e+00 6.66919851e+00 8.54038775e-01
-8.63244832e-01 3.31146419e-01 5.49261034e-01 -3.20014954e-01
-5.58735393e-02 -5.26104808e-01 -1.36873806e+00 -7.59967119e-02
5.75975835e-01 3.89622934e-02 -1.73593052e-02 1.16234922e+00
2.67959088e-01 -3.71389598e-01 -1.18492746e+00 1.30510700e+00
2.53664494e-01 -1.22002387e+00 1.39697462e-01 1.78426802e-01
9.33803856e-01 6.94318056e-01 3.14380854e-01 5.58197796e-01
3.29570770e-01 -1.05209219e+00 9.45983350e-01 -8.61671567e-03
7.50889599e-01 -5.82473457e-01 5.66409707e-01 5.44550300e-01
-1.48193908e+00 -2.78646052e-01 -1.82363510e-01 2.28926744e-02
2.83233941e-01 5.99779963e-01 -6.90935493e-01 3.83022904e-01
8.11256289e-01 1.09978354e+00 -6.01873040e-01 9.92274880e-01
-3.25797647e-01 -2.31199995e-01 -1.06876455e-02 1.63872659e-01
7.86765516e-01 1.41548470e-01 6.36077344e-01 1.38679814e+00
1.02411143e-01 -4.31320697e-01 5.47174752e-01 7.54075170e-01
1.47569388e-01 -2.77644277e-01 -1.05863559e+00 3.98486525e-01
1.91068783e-01 1.21247375e+00 -5.05377293e-01 -3.70651007e-01
-6.05068207e-01 6.64752841e-01 3.36998999e-01 3.73813152e-01
-8.22660983e-01 1.14219815e-01 1.08032358e+00 -8.49855915e-02
4.24832195e-01 -6.65953934e-01 -6.34595037e-01 -7.28412211e-01
-1.96406975e-01 -6.42380595e-01 1.72048301e-01 -8.58381391e-01
-9.80396867e-01 5.62830865e-01 3.14437032e-01 -1.09576404e+00
-3.23681295e-01 -1.01003122e+00 -2.96160817e-01 9.22642529e-01
-2.33171058e+00 -1.47338068e+00 -6.04810953e-01 6.57523513e-01
1.19030833e+00 -3.40608329e-01 5.44985235e-01 3.84560585e-01
-5.30397117e-01 2.65198082e-01 -3.72937918e-01 -3.45253438e-01
6.36090755e-01 -1.24448335e+00 8.83264661e-01 6.15613580e-01
-7.18423650e-02 -5.84837675e-01 2.72644103e-01 -4.05203253e-01
-1.03847659e+00 -1.56400967e+00 9.10557926e-01 -9.19491589e-01
4.06525671e-01 -5.93002319e-01 -5.04609168e-01 7.36432135e-01
-1.72201216e-01 -1.39151230e-01 2.61176586e-01 8.03274512e-02
-3.22611302e-01 -1.12337589e-01 -8.91100049e-01 3.35946083e-01
1.20030165e+00 -6.22362316e-01 -5.41525364e-01 6.27479970e-01
7.59465337e-01 -5.86166739e-01 -4.13544953e-01 5.64312160e-01
3.62155855e-01 -1.00296152e+00 1.30024886e+00 -2.61462182e-01
4.56063688e-01 -1.17570512e-01 -4.17878419e-01 -1.30469465e+00
-3.20406646e-01 -9.25376639e-02 -1.30664200e-01 5.96266747e-01
6.25895739e-01 -4.82345015e-01 7.91908085e-01 -1.35569656e-02
-8.33474755e-01 -6.41031206e-01 -1.03783071e+00 -9.31426942e-01
1.39344305e-01 -9.52312768e-01 3.94212276e-01 6.12696588e-01
-8.50171804e-01 4.27596748e-01 -3.41482162e-01 3.12634617e-01
7.55745113e-01 -1.22303464e-01 1.10010493e+00 -1.46618533e+00
1.39358928e-02 -6.83151364e-01 -2.14494109e-01 -1.65904701e+00
3.92112285e-01 -9.76755440e-01 2.44417265e-01 -1.82897055e+00
-1.01614133e-01 -3.86714667e-01 7.05341026e-02 6.08157873e-01
6.62226751e-02 5.64477295e-02 2.23627776e-01 1.24826133e-01
-7.85914660e-01 8.10264409e-01 1.57076263e+00 -3.47800940e-01
-2.20022395e-01 2.13476390e-01 -5.78089476e-01 7.02309370e-01
5.53999245e-01 -3.11793327e-01 -7.40037024e-01 -9.89452541e-01
-2.34022131e-03 -2.88071912e-02 7.07309723e-01 -1.08219075e+00
1.07044563e-01 -1.23620264e-01 4.07667488e-01 -1.19154871e+00
8.19296122e-01 -7.37773240e-01 -5.26332438e-01 2.49519184e-01
-1.11739911e-01 -7.47390166e-02 7.60257185e-01 3.62917125e-01
-2.55459249e-01 6.90210387e-02 8.08658600e-01 -2.73617834e-01
-1.55033207e+00 6.24947786e-01 -5.15476406e-01 5.72714582e-02
1.29650152e+00 -5.87758601e-01 -5.61799705e-01 -9.63029936e-02
-7.28849411e-01 7.52474725e-01 -7.05796182e-02 1.08189428e+00
7.18410492e-01 -1.16149783e+00 -9.08603489e-01 4.55384851e-01
5.05624056e-01 3.94488364e-01 5.50881743e-01 5.17789721e-01
-4.94987220e-02 9.14160728e-01 -2.49512002e-01 -1.20152819e+00
-1.09371448e+00 2.80319870e-01 5.49379349e-01 -4.57624614e-01
-4.71318960e-01 8.66894960e-01 8.19903016e-01 -7.48967052e-01
3.57216895e-01 -2.75587469e-01 -3.41043323e-01 3.18091512e-01
2.19009116e-01 2.13124871e-01 2.63657868e-01 -6.66447639e-01
-3.30190092e-01 7.58076310e-01 -7.31903315e-02 8.38855505e-02
1.21384692e+00 -3.35342824e-01 5.22902191e-01 4.98071164e-01
1.15953076e+00 -8.96374822e-01 -1.90493882e+00 -2.33714998e-01
-2.22216621e-01 -3.21375638e-01 5.61350465e-01 -8.85141671e-01
-9.31875169e-01 1.32109046e+00 7.17772663e-01 -1.41571108e-02
9.04720366e-01 2.06067413e-01 5.41333616e-01 4.51923639e-01
3.40634167e-01 -1.16545904e+00 5.54791152e-01 1.06569242e+00
9.70641196e-01 -1.75552523e+00 -2.34109834e-01 -3.01349938e-01
-8.78886402e-01 6.71267033e-01 1.06375921e+00 7.20927641e-02
6.41637206e-01 9.26650837e-02 5.19843213e-02 -4.54612732e-01
-9.73809600e-01 -7.81780362e-01 6.22099459e-01 1.21567428e+00
6.84098676e-02 3.71851586e-02 6.73952699e-01 2.62834340e-01
4.86676730e-02 -3.83180350e-01 2.12471336e-02 5.64616501e-01
-5.43209493e-01 -7.89823115e-01 -1.84887692e-01 2.24051833e-01
1.99147463e-01 9.40769678e-04 -1.52345017e-01 1.09932542e+00
4.59354550e-01 9.99206424e-01 3.55020136e-01 -3.75715584e-01
5.76563179e-01 -8.65420420e-03 5.05636930e-01 -8.22921753e-01
-4.08849359e-01 -5.63740619e-02 3.36604506e-01 -6.22121036e-01
-3.89630079e-01 -6.89992011e-01 -7.35380590e-01 -6.04334809e-02
9.52423960e-02 -5.25367498e-01 1.04476690e+00 1.19696379e+00
4.04820174e-01 8.17600191e-01 6.28832877e-01 -1.57889593e+00
-2.04459459e-01 -8.46925616e-01 -3.59065980e-01 3.25134061e-02
6.39212191e-01 -1.09318089e+00 -2.33796015e-01 -2.35784844e-01] | [8.133723258972168, -1.7041726112365723] |
42458c88-c3e8-4850-9192-e92f94a55fed | fewer-is-more-efficient-object-detection-in | 2212.13136 | null | https://arxiv.org/abs/2212.13136v2 | https://arxiv.org/pdf/2212.13136v2.pdf | Fewer is More: Efficient Object Detection in Large Aerial Images | Current mainstream object detection methods for large aerial images usually divide large images into patches and then exhaustively detect the objects of interest on all patches, no matter whether there exist objects or not. This paradigm, although effective, is inefficient because the detectors have to go through all patches, severely hindering the inference speed. This paper presents an Objectness Activation Network (OAN) to help detectors focus on fewer patches but achieve more efficient inference and more accurate results, enabling a simple and effective solution to object detection in large images. In brief, OAN is a light fully-convolutional network for judging whether each patch contains objects or not, which can be easily integrated into many object detectors and jointly trained with them end-to-end. We extensively evaluate our OAN with five advanced detectors. Using OAN, all five detectors acquire more than 30.0% speed-up on three large-scale aerial image datasets, meanwhile with consistent accuracy improvements. On extremely large Gaofen-2 images (29200$\times$27620 pixels), our OAN improves the detection speed by 70.5%. Moreover, we extend our OAN to driving-scene object detection and 4K video object detection, boosting the detection speed by 112.1% and 75.0%, respectively, without sacrificing the accuracy. Code is available at https://github.com/Ranchosky/OAN. | ['Junwei Han', 'Ke Li', 'Shicheng Miao', 'Qingyang Li', 'Gong Cheng', 'Xingxing Xie'] | 2022-12-26 | null | null | null | null | ['video-object-detection'] | ['computer-vision'] | [ 1.28312796e-01 -4.25088435e-01 -1.28075048e-01 -1.83705352e-02
-3.51219475e-01 -5.87357819e-01 8.63044895e-03 -1.03684865e-01
-4.70125616e-01 2.16282338e-01 -4.36358988e-01 -1.92766607e-01
1.53574571e-01 -9.92210209e-01 -7.59192944e-01 -5.43255150e-01
-2.53095925e-01 -1.15659349e-01 8.27088296e-01 2.30854517e-03
-1.29392162e-01 3.78959715e-01 -1.57302439e+00 2.06237510e-01
7.21498549e-01 1.29518843e+00 6.02656662e-01 7.84450889e-01
3.79979372e-01 6.45833075e-01 -5.91091394e-01 -2.36615986e-02
6.06876850e-01 -1.66241694e-02 -5.28547764e-01 1.77439928e-01
9.04704034e-01 -1.09942985e+00 -4.84779060e-01 1.33028090e+00
4.90645558e-01 -2.93593138e-01 2.89959610e-01 -1.00555944e+00
-5.72305739e-01 5.41426599e-01 -1.04870462e+00 5.94702601e-01
-5.30001558e-02 5.07695556e-01 1.00539625e+00 -1.07790029e+00
-6.41175732e-02 1.16392231e+00 6.08842134e-01 2.21709862e-01
-1.17167938e+00 -9.77166116e-01 3.69144857e-01 -1.49657605e-02
-1.77601171e+00 -2.13840261e-01 2.15771213e-01 -3.05840343e-01
7.82281101e-01 3.55195820e-01 7.39912808e-01 5.02131641e-01
-1.88581586e-01 1.11071074e+00 5.82607448e-01 -5.72727546e-02
-1.98464423e-01 -2.40715429e-01 1.41455010e-01 8.93224359e-01
7.51743436e-01 2.27458805e-01 -4.15975451e-02 8.06607828e-02
9.79491115e-01 4.75021303e-01 -4.08623308e-01 6.75293654e-02
-1.29947603e+00 5.94969153e-01 1.15241170e+00 -9.02075320e-02
-4.91269112e-01 3.06068838e-01 2.96490788e-01 -7.05967695e-02
8.73886421e-02 3.86244029e-01 -4.43510681e-01 6.29492760e-01
-8.36574018e-01 1.99038714e-01 2.16146871e-01 1.13641059e+00
7.21568048e-01 4.75751273e-02 -3.88930678e-01 6.96034253e-01
4.14060533e-01 1.07680023e+00 2.31122039e-02 -7.50928104e-01
4.58620906e-01 1.00805366e+00 1.89498231e-01 -1.14106369e+00
-3.07883382e-01 -5.24317205e-01 -1.01168585e+00 2.96516687e-01
2.85703897e-01 -2.25461274e-01 -9.97706413e-01 1.38761067e+00
3.73917371e-01 8.89447853e-02 -2.23109663e-01 1.25056410e+00
9.78411794e-01 9.09155726e-01 2.05831870e-01 1.95869535e-01
1.69745779e+00 -9.80014801e-01 -2.75490046e-01 -6.35676801e-01
3.82588953e-01 -7.40213037e-01 9.46039855e-01 3.27301592e-01
-8.29651535e-01 -8.90163422e-01 -1.10854411e+00 2.72744577e-02
-1.68268085e-01 9.59933698e-01 6.80936337e-01 1.86995879e-01
-8.44340384e-01 3.05074155e-01 -7.95674026e-01 -3.29118013e-01
8.96130621e-01 4.68921185e-01 6.68939352e-02 -2.30307817e-01
-8.50588560e-01 3.51778865e-01 7.63298571e-01 4.59952265e-01
-1.27516437e+00 -4.79383200e-01 -6.97995961e-01 3.14053982e-01
7.94438362e-01 -6.22468412e-01 1.21843135e+00 -7.84763217e-01
-9.18876886e-01 6.89468622e-01 1.44586638e-01 -4.77704376e-01
2.97320902e-01 -4.09668982e-01 -4.22274351e-01 1.96350157e-01
3.20179403e-01 1.18056023e+00 9.46775854e-01 -6.42083585e-01
-1.20375693e+00 -2.94242501e-01 3.51860851e-01 1.05233252e-01
-5.71831942e-01 4.60756570e-02 -8.09101164e-01 -5.70016086e-01
8.72719288e-02 -8.45815063e-01 -1.95594579e-01 6.57505155e-01
-4.67197835e-01 -2.77735680e-01 9.88120317e-01 -3.84774506e-01
1.32605946e+00 -2.22217011e+00 -1.76161513e-01 -3.69411528e-01
4.14822519e-01 7.38948822e-01 -3.83147538e-01 -3.86129059e-02
1.78381085e-01 -1.29653430e-02 -3.67466182e-01 2.38161206e-01
-2.10138261e-01 -1.37735859e-01 -4.60661829e-01 4.68975872e-01
5.80917954e-01 1.04615247e+00 -7.88057208e-01 -4.14255053e-01
3.80234241e-01 2.31839508e-01 -4.93739575e-01 1.51343539e-01
-1.00017048e-01 -2.57599413e-01 -5.76027453e-01 1.24137223e+00
9.48872507e-01 -4.83518600e-01 -1.23608917e-01 -3.45923543e-01
-2.11677581e-01 -1.84524015e-01 -1.33029592e+00 1.38097155e+00
-7.02098683e-02 7.46700466e-01 2.27927208e-01 -7.12516248e-01
8.53563249e-01 -1.36301845e-01 1.67398766e-01 -5.90935647e-01
1.67992815e-01 1.69866800e-01 1.03291921e-01 -2.79944867e-01
4.17642444e-01 3.13384742e-01 -9.70145464e-02 -1.74859077e-01
-9.93542001e-02 7.84534365e-02 4.77853596e-01 9.89585444e-02
9.95408416e-01 -1.30969703e-01 4.94897664e-01 -3.76175761e-01
3.65408927e-01 1.88936308e-01 6.79372847e-01 8.92684996e-01
-3.26275617e-01 4.63637143e-01 1.82337746e-01 -7.33581543e-01
-6.86203837e-01 -9.41693902e-01 -4.20732915e-01 9.81198192e-01
6.35125101e-01 -3.93501848e-01 -5.37773192e-01 -6.51568353e-01
3.40240777e-01 1.67514518e-01 -5.04190624e-01 -1.50215060e-01
-2.34679699e-01 -1.07279134e+00 6.72733009e-01 8.17448378e-01
1.04405546e+00 -1.04707623e+00 -1.01015043e+00 9.94234234e-02
-8.74554962e-02 -1.13672447e+00 -3.74781877e-01 1.92318968e-02
-9.84335721e-01 -1.15583599e+00 -6.93487287e-01 -7.11739421e-01
6.89074039e-01 9.84885573e-01 1.04063869e+00 4.52965975e-01
-8.64297390e-01 -1.45407528e-01 -1.99846715e-01 -5.10596991e-01
2.20660418e-01 -4.58265096e-02 6.03664443e-02 -2.94930547e-01
5.22868097e-01 -7.84025788e-02 -9.44877028e-01 6.74375474e-01
-9.56108093e-01 -6.37030080e-02 1.03503025e+00 7.71009088e-01
5.13737738e-01 1.36807919e-01 2.12546930e-01 -3.93350899e-01
-1.25248864e-01 -3.10929775e-01 -1.13625646e+00 1.64177626e-01
-4.34192061e-01 -4.66566145e-01 5.59067011e-01 -4.14771080e-01
-4.96897668e-01 3.41875076e-01 3.32473475e-03 -4.78636891e-01
-3.13229084e-01 1.48472697e-01 -9.05889347e-02 -1.54079244e-01
7.19248950e-01 2.16520369e-01 -3.28667134e-01 -2.24244535e-01
6.13526329e-02 6.70775235e-01 5.50032675e-01 -1.07659236e-01
9.61820304e-01 5.60906649e-01 -2.92810529e-01 -8.16881180e-01
-9.79983449e-01 -7.83084810e-01 -6.43247426e-01 -1.34152189e-01
7.06022680e-01 -1.55807543e+00 -8.37393343e-01 5.97760737e-01
-9.73288774e-01 -4.39564675e-01 -3.01277786e-02 4.22527552e-01
2.20539808e-01 2.82709002e-01 -6.76512837e-01 -5.82192659e-01
-5.02578616e-01 -1.03252029e+00 1.25848103e+00 5.06177783e-01
3.50177765e-01 -2.89493442e-01 -5.42163253e-01 1.29050447e-03
3.81856769e-01 2.82638948e-02 2.48739541e-01 2.56681070e-02
-9.68108594e-01 -6.33206904e-01 -8.58495414e-01 4.37305421e-01
3.31970714e-02 1.77488208e-01 -9.58036542e-01 -5.18065274e-01
-6.33217037e-01 -4.14751381e-01 1.42848110e+00 4.89207029e-01
1.38108492e+00 -1.28308997e-01 -5.71042836e-01 8.14384937e-01
1.60215151e+00 5.98830916e-02 6.18922591e-01 1.46856174e-01
7.23295093e-01 4.13126573e-02 8.35996866e-01 5.24897933e-01
2.29354063e-03 6.21104658e-01 8.64085674e-01 -4.58909929e-01
-2.21273690e-01 -1.31607994e-01 6.43175662e-01 1.99539497e-01
-1.56446159e-01 -4.55913395e-01 -7.14316130e-01 5.36447346e-01
-1.65337813e+00 -9.23018396e-01 -3.21049511e-01 1.94442213e+00
4.91963267e-01 2.66176313e-01 1.21689148e-01 -1.76792830e-01
9.08248961e-01 3.29262316e-01 -9.21786070e-01 4.94174093e-01
-1.42451301e-01 -1.56132802e-01 9.51163530e-01 9.08284336e-02
-1.65867305e+00 1.06376529e+00 5.43294811e+00 9.25774515e-01
-1.05033255e+00 -2.21029267e-01 4.85518843e-01 -1.97318003e-01
6.09785914e-01 -2.62910165e-02 -1.29695332e+00 5.05252302e-01
2.93227464e-01 1.50951773e-01 1.97215632e-01 1.33148885e+00
8.31097960e-02 -2.93824464e-01 -6.13458097e-01 9.93995547e-01
-1.01222731e-01 -1.19442439e+00 7.54417777e-02 5.13487868e-02
5.65667748e-01 1.68620959e-01 -8.57186466e-02 4.28236961e-01
2.47191384e-01 -7.16324449e-01 7.80009031e-01 -6.57167956e-02
9.17836726e-01 -4.69313353e-01 9.01406467e-01 4.60756302e-01
-1.65737224e+00 -4.97456282e-01 -1.05276203e+00 -2.94251800e-01
-6.91739619e-02 5.43928921e-01 -5.49848437e-01 1.99167266e-01
1.20609021e+00 9.44144785e-01 -9.02169943e-01 1.30810058e+00
-4.45927382e-01 7.12833226e-01 -5.04522085e-01 -3.93064618e-02
3.42845827e-01 2.00607907e-02 3.59631270e-01 1.24305642e+00
4.09468412e-01 5.38463056e-01 6.52540207e-01 9.19837296e-01
-1.92128122e-01 -2.78250664e-01 -4.27851945e-01 4.71011102e-02
5.94727635e-01 1.68282139e+00 -9.47535336e-01 -5.33079982e-01
-4.41706210e-01 1.12499177e+00 1.76288038e-01 9.64428559e-02
-1.04424822e+00 -7.20885396e-01 7.06776559e-01 9.21658725e-02
7.09880233e-01 -8.75429884e-02 1.79579437e-01 -1.12542868e+00
9.23162401e-02 -7.87378728e-01 4.84461576e-01 -6.80892766e-01
-1.15428507e+00 5.56697905e-01 -2.29953274e-01 -1.46465373e+00
5.72944164e-01 -1.08399832e+00 -5.44511020e-01 5.89770555e-01
-1.60822821e+00 -1.05901682e+00 -9.68160570e-01 2.78609544e-01
6.82500899e-01 1.16178520e-01 6.03863597e-01 5.84406435e-01
-1.06334186e+00 4.02045071e-01 -1.61736459e-01 6.54672444e-01
5.62569678e-01 -1.01082850e+00 5.58499992e-01 1.32759011e+00
8.26307982e-02 4.41942424e-01 8.60511810e-02 -5.37087739e-01
-1.41637039e+00 -1.58081472e+00 1.90550193e-01 -3.89533132e-01
4.22425061e-01 -3.61436546e-01 -8.67214739e-01 4.91377413e-01
-7.13291466e-02 5.56570709e-01 1.13176443e-01 -2.07912773e-01
-2.94712216e-01 -3.59696925e-01 -7.10582137e-01 3.15236330e-01
1.16015577e+00 -1.32785648e-01 -2.10742682e-01 4.65278387e-01
7.32014000e-01 -4.01420593e-01 -5.66772103e-01 7.30034173e-01
4.34402794e-01 -8.88403118e-01 1.20853627e+00 -2.86591679e-01
3.88165742e-01 -1.00595665e+00 -2.58217245e-01 -6.71952426e-01
-8.29626799e-01 -1.21273264e-01 -1.69900477e-01 8.48807156e-01
3.29435080e-01 -4.57867056e-01 4.54953730e-01 5.93139566e-02
-1.92624018e-01 -5.27920127e-01 -3.40995759e-01 -9.61241603e-01
-4.49451685e-01 -3.27229947e-01 5.15545070e-01 5.84817767e-01
-5.53122461e-01 4.09876168e-01 -3.71112645e-01 8.95419955e-01
5.75532675e-01 5.11591077e-01 7.40950346e-01 -1.19368649e+00
-2.20487207e-01 -5.09383440e-01 -5.58042467e-01 -1.69901729e+00
-5.15810788e-01 -6.74290299e-01 3.24276209e-01 -1.48589122e+00
4.53935772e-01 -3.70018661e-01 -3.02803218e-01 8.24606597e-01
-5.51297605e-01 7.56029904e-01 2.06972674e-01 3.67078483e-01
-8.81024659e-01 4.67024535e-01 1.00271738e+00 -4.12213534e-01
-9.07505117e-03 8.02537333e-03 -6.45883381e-01 8.25654984e-01
7.08977520e-01 -4.28097218e-01 6.55630650e-03 -7.99256444e-01
-1.00257315e-01 -3.38455826e-01 9.89051342e-01 -1.43116331e+00
2.24290684e-01 1.01380929e-01 9.37461972e-01 -8.26742232e-01
1.04682378e-01 -6.88073277e-01 -2.93438882e-01 9.16677952e-01
4.23795842e-02 -5.47827035e-02 6.18134260e-01 7.35324323e-01
-2.19421178e-01 -4.45184447e-02 9.29172933e-01 -1.53763130e-01
-1.43166053e+00 6.91341281e-01 -1.46310344e-01 -2.40802377e-01
1.31776524e+00 -1.50696233e-01 -3.53574693e-01 2.39528358e-01
-2.07691610e-01 5.81715107e-01 3.21215957e-01 3.32369953e-01
7.23131657e-01 -1.12435997e+00 -9.08210933e-01 3.68660331e-01
3.97390336e-01 4.02560651e-01 3.88244092e-01 7.99502850e-01
-7.79676795e-01 4.71374452e-01 -9.81953368e-02 -9.31755722e-01
-1.25645077e+00 6.20594501e-01 3.51239383e-01 2.09620684e-01
-6.96734428e-01 1.15627360e+00 5.10906279e-01 -1.71086892e-01
2.01677635e-01 -6.47976756e-01 -1.20354861e-01 -1.77760012e-02
8.58211219e-01 3.53487432e-01 -2.64771551e-01 -2.38373205e-01
-5.22936881e-01 7.84023404e-01 -7.89724290e-02 6.80565774e-01
1.01402605e+00 1.00600369e-01 4.81948182e-02 -1.40722871e-01
7.12289095e-01 -2.84598589e-01 -1.60407746e+00 -2.72540778e-01
-3.17451328e-01 -7.13618219e-01 3.52575153e-01 -7.02206016e-01
-1.24700391e+00 9.44044054e-01 8.79497111e-01 2.50703573e-01
1.45453608e+00 4.13170531e-02 7.60537565e-01 6.25494003e-01
3.32315743e-01 -7.31796682e-01 1.27581686e-01 3.22297633e-01
6.34909868e-01 -1.67598271e+00 3.60683322e-01 -4.41924304e-01
-4.27553058e-01 1.16281235e+00 1.00793338e+00 -3.55099618e-01
3.41095209e-01 2.57215798e-01 -1.43425196e-01 -3.54606241e-01
-5.24111211e-01 -5.42774796e-01 2.99498051e-01 2.22385600e-01
1.17478305e-02 3.19419980e-01 1.80434257e-01 3.74014556e-01
4.61938351e-01 -9.52008143e-02 2.26016983e-01 6.89064682e-01
-8.90607476e-01 -3.77369851e-01 -3.81339848e-01 5.46928763e-01
-2.61108041e-01 -3.19056302e-01 -2.23335341e-01 9.17551458e-01
3.63979071e-01 8.01071048e-01 1.53587937e-01 -4.30050343e-01
4.77271587e-01 -6.67670786e-01 4.26033512e-02 -7.56058097e-01
-2.05668330e-01 8.43190029e-02 -3.24028581e-01 -5.91058314e-01
-2.29401872e-01 -4.88921911e-01 -1.02901578e+00 -2.13693142e-01
-9.37346458e-01 -2.65201569e-01 1.76651999e-01 4.07265633e-01
4.69276547e-01 6.31168783e-01 4.75419194e-01 -7.70317435e-01
-7.44262040e-01 -9.81972456e-01 -6.19853199e-01 -8.44314843e-02
2.38808662e-01 -5.79287946e-01 -1.83620527e-01 4.57662530e-02] | [8.761281967163086, -0.7196851372718811] |
b25a92b5-000e-4cdc-81e5-71c67050b1ed | unsupervised-feature-learning-with-emergent | 2307.01421 | null | https://arxiv.org/abs/2307.01421v1 | https://arxiv.org/pdf/2307.01421v1.pdf | Unsupervised Feature Learning with Emergent Data-Driven Prototypicality | Given an image set without any labels, our goal is to train a model that maps each image to a point in a feature space such that, not only proximity indicates visual similarity, but where it is located directly encodes how prototypical the image is according to the dataset. Our key insight is to perform unsupervised feature learning in hyperbolic instead of Euclidean space, where the distance between points still reflect image similarity, and yet we gain additional capacity for representing prototypicality with the location of the point: The closer it is to the origin, the more prototypical it is. The latter property is simply emergent from optimizing the usual metric learning objective: The image similar to many training instances is best placed at the center of corresponding points in Euclidean space, but closer to the origin in hyperbolic space. We propose an unsupervised feature learning algorithm in Hyperbolic space with sphere pACKing. HACK first generates uniformly packed particles in the Poincar\'e ball of hyperbolic space and then assigns each image uniquely to each particle. Images after congealing are regarded more typical of the dataset it belongs to. With our feature mapper simply trained to spread out training instances in hyperbolic space, we observe that images move closer to the origin with congealing, validating our idea of unsupervised prototypicality discovery. We demonstrate that our data-driven prototypicality provides an easy and superior unsupervised instance selection to reduce sample complexity, increase model generalization with atypical instances and robustness with typical ones. | ['Stella X. Yu', 'Yubei Chen', 'Youren Zhang', 'Yunhui Guo'] | 2023-07-04 | null | null | null | null | ['metric-learning', 'metric-learning'] | ['computer-vision', 'methodology'] | [ 1.11559674e-01 1.23096608e-01 1.88468188e-01 -3.45059931e-01
-2.59356678e-01 -8.95122170e-01 6.63215339e-01 4.29091454e-01
-5.24338722e-01 3.53665799e-01 1.04925767e-01 1.28731072e-01
-6.29457712e-01 -8.77052844e-01 -7.67965972e-01 -1.12440920e+00
-1.40664861e-01 7.69145429e-01 1.62276775e-01 -1.24198370e-01
3.93871307e-01 5.89565575e-01 -1.88396764e+00 1.59497708e-01
4.86824483e-01 9.17279184e-01 3.80936146e-01 7.72985339e-01
4.17754918e-01 4.05205667e-01 -5.71523368e-01 -1.80513799e-01
4.97797608e-01 -5.34832478e-01 -9.10374939e-01 5.21405458e-01
4.43300724e-01 3.77553076e-01 -3.84142011e-01 9.82944369e-01
4.91064563e-02 3.89256507e-01 1.11510932e+00 -1.35748816e+00
-9.93352532e-01 2.63869882e-01 -2.59795010e-01 1.70671225e-01
2.72305697e-01 1.94890514e-01 1.18988836e+00 -7.20263302e-01
9.39200997e-01 8.04908335e-01 5.61684966e-01 2.16179788e-01
-1.46975303e+00 4.53296341e-02 -3.25261980e-01 3.07005286e-01
-1.67517376e+00 -1.09976560e-01 6.89336181e-01 -4.80964750e-01
3.95233810e-01 5.54704964e-01 9.13888991e-01 7.33040214e-01
2.39653200e-01 3.86029303e-01 1.23867273e+00 -3.30540806e-01
5.80524266e-01 3.66407901e-01 -3.39843482e-02 5.41003942e-01
-3.31212580e-02 1.13074794e-01 -2.95138299e-01 1.50285974e-01
5.12442350e-01 5.11857927e-01 -2.13237703e-01 -1.01386094e+00
-1.33845687e+00 8.17174256e-01 9.80152071e-01 2.80769050e-01
-1.56483486e-01 -3.03531736e-01 8.38279724e-03 3.36193532e-01
-8.00074451e-03 1.23103476e+00 -1.85272783e-01 -1.44609902e-02
-8.19087267e-01 4.18206096e-01 6.23010457e-01 1.01634920e+00
1.15352285e+00 -7.37136424e-01 2.95927115e-02 6.68757081e-01
-8.23616832e-02 5.42084754e-01 6.65607989e-01 -1.08853269e+00
-1.79099202e-01 9.33359027e-01 -1.78445101e-01 -1.29219294e+00
-5.88704228e-01 -3.66876066e-01 -6.94539368e-01 3.24535400e-01
5.37167311e-01 2.21643940e-01 -4.83494341e-01 1.67702639e+00
2.17393368e-01 -2.85781026e-01 1.66538805e-01 1.02966726e+00
5.48541367e-01 6.35169744e-01 -4.73059714e-01 4.30980846e-02
1.40941286e+00 -4.33384269e-01 1.08909324e-01 1.30469054e-01
7.42039680e-01 -5.38010836e-01 1.48267806e+00 1.44508228e-01
-7.56592751e-01 -3.87667030e-01 -1.25627255e+00 3.97284962e-02
-6.50213659e-01 -1.69609472e-01 2.19903767e-01 3.71468484e-01
-9.88955736e-01 9.23996568e-01 -3.03137451e-01 -6.40850663e-01
2.13053435e-01 2.16588482e-01 -6.81180894e-01 1.15593225e-01
-6.37618184e-01 7.55596101e-01 4.90732282e-01 -5.50040185e-01
-5.21855950e-01 -5.68114817e-01 -8.30199480e-01 -3.27962711e-02
1.12491593e-01 -3.95503640e-01 7.93680072e-01 -1.01812208e+00
-1.06124556e+00 1.05491698e+00 7.55406022e-02 -4.82488871e-01
2.09145144e-01 6.54163241e-01 -3.42897415e-01 2.47322679e-01
2.74891287e-01 8.45685780e-01 7.74719834e-01 -1.51546955e+00
-6.86047554e-01 -6.44340515e-01 2.50904560e-01 3.73634577e-01
-3.42803687e-01 -3.64249021e-01 -3.69027890e-02 -4.26209182e-01
4.61559862e-01 -1.27659142e+00 5.15220538e-02 6.92689642e-02
-4.94623929e-01 -2.17441499e-01 1.03651512e+00 1.53880984e-01
1.05376470e+00 -2.31024814e+00 1.17222764e-01 6.72388375e-01
6.34127676e-01 -2.58806586e-01 3.16797234e-02 7.00019300e-01
-2.08901674e-01 -9.39048361e-03 -3.30812097e-01 1.11666143e-01
1.04566351e-01 2.88816422e-01 -6.31705150e-02 8.36374998e-01
1.08861476e-01 7.33606994e-01 -1.13101888e+00 -4.48489636e-01
2.01940343e-01 1.30535632e-01 -4.93710637e-01 9.99505594e-02
7.16739818e-02 4.28709388e-01 5.81136271e-02 1.58766761e-01
4.57512379e-01 -6.73865080e-01 -2.36963749e-01 -1.40527368e-01
-1.97111949e-01 -1.35956541e-01 -1.08755279e+00 1.11203909e+00
-1.71668023e-01 8.66214156e-01 -7.92207360e-01 -1.01776361e+00
1.02821195e+00 -2.20552832e-01 3.52656841e-01 -7.34511554e-01
5.03689144e-03 -1.73997417e-01 2.97088325e-01 -6.12527013e-01
4.96135116e-01 -9.44100171e-02 -7.63825849e-02 6.88626349e-01
-1.44711643e-01 -1.41300738e-01 1.75778225e-01 3.44031155e-01
1.08754730e+00 -6.09584212e-01 2.22562492e-01 -7.35624135e-01
1.80700496e-01 5.69719523e-02 2.61088699e-01 7.00064421e-01
-6.07799217e-02 1.07138097e+00 6.82940960e-01 -6.12855077e-01
-1.53080559e+00 -1.47071683e+00 -6.49604440e-01 1.04436517e+00
5.99506855e-01 -4.55656111e-01 -6.58063769e-01 -6.76510513e-01
1.74891349e-04 4.78722066e-01 -1.23959732e+00 -4.22531843e-01
-1.74458981e-01 -4.77279186e-01 9.61022973e-02 1.54264539e-01
2.84947395e-01 -1.15008521e+00 -9.28768277e-01 -2.80493557e-01
2.66951591e-01 -5.00528336e-01 -5.77428281e-01 4.47303385e-01
-3.79834563e-01 -1.41519904e+00 -6.66282535e-01 -9.65592325e-01
1.06470108e+00 4.44691390e-01 1.07545269e+00 -9.74706784e-02
-6.09007180e-01 3.96862090e-01 -6.33065879e-01 -5.14138713e-02
-3.30409169e-01 -9.97178927e-02 1.45423636e-01 2.63133794e-01
1.83148503e-01 -4.15954798e-01 -9.78360951e-01 6.14396811e-01
-9.10981357e-01 -1.94014490e-01 3.17005783e-01 6.94311321e-01
7.07654893e-01 2.26248682e-01 1.83535665e-02 -7.93176055e-01
2.97709584e-01 -5.49703419e-01 -1.51794642e-01 1.84479550e-01
-4.47657168e-01 3.51896405e-01 9.06816363e-01 -4.67451602e-01
-1.86484545e-01 -7.62247655e-04 5.37032068e-01 -2.96634853e-01
-1.96543187e-01 1.67417541e-01 -7.27507174e-02 -1.06365025e-01
1.12861001e+00 4.88765925e-01 3.47943723e-01 -7.25960592e-03
5.00173569e-01 4.93582219e-01 5.42791605e-01 -3.94967496e-01
7.69624949e-01 7.23988891e-01 2.09730029e-01 -1.05727804e+00
-4.48303223e-01 -4.45278615e-01 -8.20780814e-01 -2.10430473e-01
9.31023061e-01 -3.53592604e-01 -8.88145924e-01 3.13522033e-02
-5.45638680e-01 -2.16048554e-01 -8.65570903e-01 2.70540148e-01
-8.22508693e-01 2.90663749e-01 -4.56253923e-02 -4.88890558e-01
2.71438718e-01 -1.07102752e+00 9.75978613e-01 2.23221600e-01
-4.48602408e-01 -1.02815270e+00 3.04833621e-01 -3.12234368e-02
2.40083295e-03 1.55457109e-01 1.20022953e+00 -8.31301749e-01
-3.15113336e-01 -3.38474691e-01 -3.16027880e-01 1.29565656e-01
2.53883272e-01 2.70976722e-01 -6.11568928e-01 -4.19269741e-01
6.20023832e-02 -1.65122956e-01 4.62395787e-01 3.33071858e-01
1.26121283e+00 -5.45883417e-01 -2.77631015e-01 7.50803411e-01
1.38086379e+00 2.19984666e-01 4.84713346e-01 4.93293524e-01
5.48569918e-01 7.79833853e-01 4.08539385e-01 3.05362314e-01
3.50308448e-01 7.61033177e-01 4.09718722e-01 -1.95963070e-01
1.23080730e-01 -1.17302947e-01 5.97861409e-02 3.33718032e-01
1.02304429e-01 6.27955869e-02 -8.58832419e-01 3.83094966e-01
-1.64876091e+00 -1.15371060e+00 2.31844820e-02 2.57848668e+00
5.80776453e-01 -2.38590539e-02 6.80017292e-01 2.29177728e-01
8.23511362e-01 -8.14767703e-02 -5.28562248e-01 -1.60325989e-01
-4.86363083e-01 -1.40620843e-01 4.45912093e-01 3.08754921e-01
-1.18783927e+00 5.15962124e-01 6.07663059e+00 8.04350972e-01
-1.10502684e+00 -2.98798800e-01 9.16018844e-01 -2.61535078e-01
-2.51906216e-01 3.55178211e-03 -3.93192202e-01 7.50147641e-01
4.86384660e-01 -4.01843816e-01 4.80655879e-01 8.04130733e-01
1.47095397e-01 -8.78739655e-02 -1.49930763e+00 1.04482675e+00
3.14483166e-01 -1.47402596e+00 1.33632228e-01 3.71066481e-01
8.31993520e-01 -1.81356475e-01 5.54018259e-01 -2.05188885e-01
2.05307618e-01 -1.23827708e+00 6.98200107e-01 4.38479215e-01
5.69621146e-01 -8.75910580e-01 3.51339221e-01 2.63677210e-01
-1.01880109e+00 -2.35588983e-01 -6.56289458e-01 3.45741324e-02
-3.89092565e-01 1.50326177e-01 -9.90703225e-01 -3.55204120e-02
7.70977139e-01 6.39269590e-01 -1.11023545e+00 1.17439747e+00
3.09850723e-01 -5.83052039e-02 -1.72149867e-01 -1.80354118e-01
4.90769118e-01 -4.96511787e-01 6.50992692e-01 1.01622725e+00
3.75670701e-01 1.25010997e-01 9.55592841e-02 8.36854398e-01
1.34867057e-01 1.91950500e-01 -1.14151132e+00 3.61807466e-01
5.92483699e-01 1.33790290e+00 -1.13114142e+00 -3.56084853e-01
-1.96964711e-01 9.19393957e-01 3.66032749e-01 2.25175858e-01
-7.21918046e-01 -6.45059109e-01 4.98620182e-01 5.54826617e-01
1.50987327e-01 -4.57649790e-02 -1.27297044e-01 -6.81502342e-01
-6.94602076e-03 -6.75732374e-01 3.56785208e-01 -1.00224435e+00
-1.33083487e+00 7.07516968e-01 -1.66660145e-01 -1.58549416e+00
-5.84805310e-02 -7.36035347e-01 -6.75266147e-01 3.87412488e-01
-4.86604124e-01 -7.13831246e-01 -3.39493454e-01 7.35115170e-01
1.78106442e-01 -1.16052605e-01 7.63654113e-01 -3.11342537e-01
-1.44639865e-01 6.72748208e-01 7.40614235e-01 2.41771325e-01
5.73050439e-01 -1.76618028e+00 1.12260304e-01 3.96596014e-01
6.75063431e-01 6.28549993e-01 8.21132660e-01 -1.72837630e-01
-1.20011425e+00 -1.12680972e+00 7.80565739e-01 -7.78172255e-01
8.53871882e-01 -4.54736948e-01 -8.29648793e-01 4.01448131e-01
-1.55607343e-01 5.53976037e-02 7.66353846e-01 -6.00777380e-02
-6.48684323e-01 -2.69937754e-01 -1.24403906e+00 9.33087468e-01
1.03351152e+00 -6.11656725e-01 -4.93747771e-01 6.28125131e-01
4.15757835e-01 1.37409508e-01 -9.61347759e-01 -2.74591967e-02
3.90652150e-01 -1.16138828e+00 9.04401541e-01 -6.85329318e-01
5.77711642e-01 -4.56214100e-01 -3.58965248e-01 -1.51742589e+00
-6.15772903e-01 -3.87960821e-01 4.04994905e-01 7.79049218e-01
5.91918588e-01 -5.20522296e-01 9.53195155e-01 4.57433462e-01
-6.07186258e-02 -9.71458495e-01 -1.06999803e+00 -1.01068127e+00
2.39396647e-01 3.75906192e-02 5.76972008e-01 8.56813073e-01
5.18105209e-01 2.59813070e-01 2.30817631e-01 5.86854145e-02
5.97687840e-01 5.24488926e-01 7.03554928e-01 -1.34902954e+00
-2.35597610e-01 -6.24583006e-01 -1.00936699e+00 -8.92048657e-01
-1.47560224e-01 -1.22782981e+00 1.52922347e-01 -9.82824087e-01
5.47789395e-01 -6.42623127e-01 -1.12865433e-01 1.87899023e-01
1.46836147e-01 4.92978901e-01 2.00121731e-01 5.10201275e-01
-9.90637839e-01 3.89708698e-01 1.20885372e+00 -1.32403746e-01
-1.54379234e-01 -1.65536344e-01 -6.29617691e-01 5.70561707e-01
5.25111079e-01 -3.36833149e-01 -4.38040584e-01 -9.04055461e-02
2.21412450e-01 -2.31919989e-01 7.00308084e-01 -1.22708094e+00
2.99320996e-01 3.59880552e-02 6.69176102e-01 -1.20313071e-01
2.84198821e-01 -9.86982226e-01 2.45814696e-01 2.49033675e-01
-5.76551974e-01 2.76237816e-01 -4.10050690e-01 4.24186200e-01
9.78251845e-02 -3.43806654e-01 1.09750462e+00 -9.44659710e-02
-5.50898671e-01 3.33174676e-01 -3.54192615e-01 2.60346711e-01
1.35611784e+00 -6.15142822e-01 -3.79613727e-01 -3.43904167e-01
-9.15493488e-01 -1.28123730e-01 1.20914912e+00 1.54793754e-01
5.00837564e-01 -1.47093856e+00 -5.55332601e-01 5.28111160e-01
7.47282147e-01 -1.63345575e-01 1.37807086e-01 6.66574895e-01
-4.93333727e-01 1.24417551e-01 -2.53303528e-01 -1.03900170e+00
-1.09378088e+00 8.87434602e-01 2.58848250e-01 2.86826432e-01
-9.98879671e-01 7.15618491e-01 6.48041368e-01 -2.69962102e-01
-3.49893160e-02 -1.87420368e-01 -1.23054504e-01 1.08936995e-01
3.13850731e-01 3.18563700e-01 6.84253231e-04 -9.23709989e-01
-2.15168342e-01 6.77324533e-01 -1.35128334e-01 -1.38336476e-02
1.07813215e+00 -5.76033257e-02 -6.20567910e-02 6.64375484e-01
1.69812000e+00 -3.32192123e-01 -1.49082220e+00 -5.54638095e-02
2.19567679e-02 -4.99998242e-01 -4.03027683e-01 -4.99487728e-01
-8.82012129e-01 4.62958276e-01 6.73084915e-01 9.27187979e-01
7.46047795e-01 6.06569290e-01 6.78445175e-02 6.62403882e-01
3.38409901e-01 -1.05671382e+00 3.37010264e-01 4.51303720e-01
8.77476811e-01 -1.23385048e+00 -1.22386999e-01 8.39392319e-02
-8.53723824e-01 9.55351532e-01 2.34629944e-01 -7.21809924e-01
6.08528137e-01 -1.53152704e-01 -9.02931765e-02 -6.27139807e-01
-4.11144972e-01 -7.17981085e-02 2.38873810e-01 8.01942348e-01
7.07278997e-02 1.19850777e-01 1.72494709e-01 2.91915089e-01
-9.71490145e-01 -8.35797906e-01 5.24616003e-01 4.91027743e-01
-7.47660995e-01 -5.14094472e-01 -3.78432631e-01 4.70439285e-01
2.10657120e-01 -1.30425049e-02 -6.05249763e-01 9.37537551e-01
2.98772842e-01 6.80238068e-01 8.18563700e-01 -4.56778407e-01
3.90164822e-01 -2.98644334e-01 5.72562814e-01 -5.56945622e-01
-2.21564054e-01 -3.59435350e-01 -6.03886366e-01 -3.75571966e-01
4.69547547e-02 -8.82381737e-01 -1.35942078e+00 -4.42838132e-01
9.32368264e-02 4.79948699e-01 3.41853291e-01 7.31490314e-01
3.95926625e-01 1.02941729e-02 1.23749661e+00 -9.61113572e-01
-2.09932834e-01 -4.68221486e-01 -9.66001153e-01 9.34515476e-01
3.99983674e-01 -5.46354055e-01 -7.73105681e-01 2.26974979e-01] | [9.166852951049805, 2.983635663986206] |
efc883ed-6c46-4259-b5e8-9322ae27669e | neural-group-recommendation-based-on-a | 2303.07001 | null | https://arxiv.org/abs/2303.07001v1 | https://arxiv.org/pdf/2303.07001v1.pdf | Neural Group Recommendation Based on a Probabilistic Semantic Aggregation | Recommendation to groups of users is a challenging subfield of recommendation systems. Its key concept is how and where to make the aggregation of each set of user information into an individual entity, such as a ranked recommendation list, a virtual user, or a multi-hot input vector encoding. This paper proposes an innovative strategy where aggregation is made in the multi-hot vector that feeds the neural network model. The aggregation provides a probabilistic semantic, and the resulting input vectors feed a model that is able to conveniently generalize the group recommendation from the individual predictions. Furthermore, using the proposed architecture, group recommendations can be obtained by simply feedforwarding the pre-trained model with individual ratings; that is, without the need to obtain datasets containing group of user information, and without the need of running two separate trainings (individual and group). This approach also avoids maintaining two different models to support both individual and group learning. Experiments have tested the proposed architecture using three representative collaborative filtering datasets and a series of baselines; results show suitable accuracy improvements compared to the state-of-the-art. | ['Jesús Bobadilla', 'Fernando Ortega', 'Raúl Lara-Cabrera', 'Jorge Dueñas-Lerín'] | 2023-03-13 | null | null | null | null | ['collaborative-filtering'] | ['miscellaneous'] | [ 1.38151452e-01 6.88474625e-02 -2.48317942e-01 -9.23527896e-01
-3.31191152e-01 -4.75090832e-01 6.61943734e-01 3.28407884e-01
-4.48141247e-01 4.65383112e-01 4.05291975e-01 -4.33878005e-01
-3.52163345e-01 -1.09028447e+00 -6.76543057e-01 -4.79319781e-01
9.50552449e-02 6.33672237e-01 2.54576474e-01 -1.40405595e-01
2.66912401e-01 2.16398776e-01 -1.97134054e+00 8.62718999e-01
9.12996650e-01 1.43732643e+00 2.77190953e-01 5.34127176e-01
-2.41158038e-01 5.21997154e-01 -4.04804796e-01 -3.18052083e-01
3.73685062e-01 -3.55755925e-01 -5.58272243e-01 1.38819680e-01
6.19350195e-01 -4.00034368e-01 1.97138652e-01 5.76071382e-01
3.41193587e-01 8.64721715e-01 8.15415084e-01 -6.85528755e-01
-9.07480419e-01 1.15129042e+00 8.58419910e-02 -1.56470373e-01
1.93807617e-01 -6.33090794e-01 1.33619010e+00 -9.30931330e-01
4.27859753e-01 9.60604310e-01 4.65767026e-01 2.87950128e-01
-1.10071135e+00 -5.37507594e-01 6.25087500e-01 5.24797216e-02
-8.97286713e-01 -1.39737278e-01 3.33838493e-01 -6.06466532e-01
9.12972867e-01 4.16029632e-01 5.65296173e-01 7.33032346e-01
3.37588675e-02 6.20278060e-01 6.53808296e-01 -3.31670076e-01
5.44332445e-01 6.59544468e-01 7.03633308e-01 2.93428093e-01
1.79852247e-01 -8.51878226e-02 -1.66618422e-01 -3.95601690e-01
5.11285663e-01 7.24925578e-01 -2.31226310e-01 -3.92528772e-01
-8.38617444e-01 9.41144884e-01 6.44411862e-01 2.95416206e-01
-6.69816971e-01 -1.97948441e-01 1.60604849e-01 5.36001980e-01
6.30317390e-01 5.63151002e-01 -6.16783202e-01 6.77039206e-01
-1.10433042e+00 3.06597233e-01 9.40069377e-01 5.52937865e-01
8.26156139e-01 -1.28288060e-01 -2.17582807e-01 8.13804924e-01
5.10655761e-01 6.06200472e-02 7.06898093e-01 -5.27570009e-01
3.10826868e-01 6.77252352e-01 2.78379917e-01 -7.50703573e-01
-3.33479077e-01 -8.89228344e-01 -6.68880880e-01 1.32797465e-01
3.65591377e-01 -3.76246572e-01 -8.36399317e-01 1.35144138e+00
3.23412597e-01 3.90728414e-01 2.91413739e-02 7.70018876e-01
8.34517896e-01 8.88326883e-01 -2.67725773e-02 -2.83252656e-01
9.50333357e-01 -1.30189967e+00 -4.18375015e-01 -1.07520826e-01
6.45834565e-01 -6.58612788e-01 6.72247589e-01 6.69719279e-01
-9.19417560e-01 -1.11594105e+00 -9.07609284e-01 2.14378729e-01
-7.15938866e-01 4.37249750e-01 7.78094411e-01 5.76533556e-01
-1.07460046e+00 9.96143460e-01 -4.09689695e-01 -3.83157581e-01
-1.25292568e-02 7.38481939e-01 -1.40872017e-01 4.07081246e-02
-1.19053841e+00 7.32195020e-01 5.42103767e-01 -5.87056875e-02
-2.07703888e-01 -7.51333356e-01 -2.95295745e-01 4.35053945e-01
1.13406464e-01 -7.85741866e-01 1.04025662e+00 -1.17521727e+00
-1.59742141e+00 -1.14949733e-01 2.05615219e-02 -4.56695914e-01
7.98822865e-02 -4.31585282e-01 -7.07855284e-01 -4.28104669e-01
-4.00546849e-01 2.61291623e-01 8.03267419e-01 -1.10868454e+00
-1.19307292e+00 -3.18611085e-01 1.89463437e-01 2.81075060e-01
-6.32801950e-01 7.84061626e-02 -2.53805667e-01 -5.25425553e-01
1.34030238e-01 -9.38971341e-01 -5.54459393e-01 -5.58252692e-01
1.35347977e-01 -6.35027468e-01 3.72964531e-01 -6.48058534e-01
1.49295461e+00 -1.83354867e+00 2.68264592e-01 6.49422228e-01
-8.28065500e-02 3.01315427e-01 -1.39998317e-01 3.49950612e-01
1.60244629e-02 -3.45171914e-02 4.03619945e-01 -3.94850671e-01
5.12284935e-02 7.46914744e-02 -5.70869923e-01 1.35341018e-01
-3.65071267e-01 5.72305024e-01 -7.49249876e-01 1.09767511e-01
2.52685010e-01 4.22435492e-01 -1.10788095e+00 4.21285242e-01
-3.95095050e-01 3.49443883e-01 -4.66547519e-01 -8.59575570e-02
4.54730809e-01 -3.25741529e-01 4.58404541e-01 -3.62071365e-01
-1.52813867e-01 5.52775145e-01 -1.52083445e+00 1.52000213e+00
-6.15650773e-01 -1.63229406e-01 -4.09119606e-01 -9.40409482e-01
1.19126928e+00 3.34448159e-01 2.24351868e-01 -2.17231810e-01
2.81748384e-01 1.64696977e-01 1.00356247e-02 -2.37744171e-02
8.42493534e-01 1.44717574e-01 -6.67390600e-02 1.05369353e+00
5.05261779e-01 5.05448520e-01 6.44202307e-02 1.51331753e-01
6.07267439e-01 1.88525856e-01 1.68791935e-01 -7.14425743e-02
5.11816323e-01 -5.64296901e-01 3.33810866e-01 9.92096364e-01
7.01125264e-01 3.49147767e-01 -1.22423999e-01 -6.50099695e-01
-6.91892445e-01 -7.85528660e-01 1.67957887e-01 1.88094437e+00
-1.28979534e-01 -7.10546911e-01 -2.20966607e-01 -9.54581976e-01
7.33498409e-02 9.77134466e-01 -7.24425793e-01 -2.58627776e-02
-3.57141078e-01 -5.57380199e-01 -4.13447291e-01 6.20853066e-01
-2.41255835e-01 -8.73997569e-01 -8.42146575e-02 3.89597952e-01
1.12507984e-01 -5.09308159e-01 -2.59559482e-01 4.24288839e-01
-1.10954189e+00 -5.27616084e-01 -4.35478657e-01 -7.02948809e-01
7.15730250e-01 2.95869470e-01 1.01622176e+00 2.82413185e-01
5.95466018e-01 -5.11683850e-03 -7.60808825e-01 -1.64968714e-01
-1.07443586e-01 2.30250061e-01 2.56383777e-01 5.34586787e-01
3.72618020e-01 -6.29412293e-01 -5.83431721e-01 5.56585908e-01
-6.62808776e-01 -9.13425386e-02 4.14726496e-01 7.41680384e-01
5.45079172e-01 3.77205536e-02 8.60986590e-01 -1.33181190e+00
6.43997669e-01 -7.69704580e-01 -3.64210546e-01 4.43746746e-01
-8.07557583e-01 2.92162243e-02 1.09441483e+00 -5.54488122e-01
-1.10436594e+00 -2.03493610e-03 -1.70632109e-01 -1.47146150e-01
-2.59411514e-01 7.26133823e-01 4.45937105e-02 2.05395043e-01
8.56024265e-01 -3.96060683e-02 -3.14000905e-01 -1.05573153e+00
8.99722219e-01 1.10239136e+00 2.25397423e-01 -3.13579023e-01
4.09615785e-01 8.98317713e-03 -5.86008430e-01 -3.85654360e-01
-1.02886581e+00 -8.30582380e-01 -7.30994701e-01 -1.25947952e-01
4.86084610e-01 -7.12316513e-01 -3.25153202e-01 -3.91225964e-02
-7.18528807e-01 -1.33141607e-01 -3.31592679e-01 7.95974672e-01
-2.28505954e-01 9.66659933e-02 -5.15856206e-01 -5.99788904e-01
-5.02980709e-01 -9.42931473e-01 3.80793840e-01 3.09464842e-01
-2.54830599e-01 -1.11024249e+00 -9.60996076e-02 2.64655977e-01
7.08059967e-01 -4.91367608e-01 7.49129176e-01 -1.62166452e+00
-3.10600042e-01 -6.21686816e-01 3.55899185e-02 5.96702397e-01
1.09757833e-01 -1.12667434e-01 -8.22856724e-01 -3.82165581e-01
-9.77803543e-02 -4.90126237e-02 9.78548110e-01 2.26586163e-01
1.24330914e+00 -3.62862080e-01 -3.85629952e-01 4.52442288e-01
1.20017159e+00 2.91178524e-01 1.66516170e-01 2.49467529e-02
6.11979723e-01 4.21012282e-01 5.46894610e-01 4.55822319e-01
3.36602628e-01 6.81601822e-01 1.84462517e-01 2.06204176e-01
2.04217974e-02 -2.87967384e-01 2.46032327e-01 1.01850903e+00
-5.10607302e-01 -4.65530545e-01 -1.77311540e-01 1.42273232e-01
-2.17698669e+00 -9.89976406e-01 1.16856761e-01 2.53029060e+00
3.05742443e-01 -1.67927042e-01 1.59756884e-01 1.15341721e-02
5.17411113e-01 -1.01290971e-01 -2.74486661e-01 -5.67000210e-01
3.13871562e-01 3.54835719e-01 3.02751958e-01 5.18488407e-01
-1.25199604e+00 7.94336140e-01 6.00149298e+00 6.13373101e-01
-1.15140700e+00 7.89281055e-02 3.82930428e-01 -1.54119506e-01
-4.64646995e-01 -6.47424459e-02 -1.16330576e+00 4.29440200e-01
1.19547713e+00 9.81907453e-03 4.88341928e-01 8.55517209e-01
-2.67556589e-02 2.56993234e-01 -1.32002866e+00 4.75688994e-01
2.88926989e-01 -1.32979631e+00 3.74403983e-01 2.75867432e-03
9.15601969e-01 1.88779905e-02 6.50588647e-02 6.10187709e-01
6.78126752e-01 -7.51402974e-01 4.94001508e-01 8.35065246e-01
1.49476141e-01 -6.57375693e-01 9.53360021e-01 4.89760965e-01
-9.15975273e-01 -5.09817302e-01 -7.11877644e-01 -2.08873034e-01
1.58499703e-01 4.29889023e-01 -7.48007059e-01 8.49989116e-01
6.40321016e-01 8.94588888e-01 -3.94324005e-01 1.13555646e+00
-7.57597238e-02 7.35050201e-01 -3.21164221e-01 -2.21647173e-01
7.98469037e-02 -5.12192786e-01 7.50290677e-02 1.10621476e+00
7.95216918e-01 2.66480763e-02 5.49729109e-01 1.69525012e-01
1.88084282e-02 6.27287805e-01 -1.39578968e-01 2.35188276e-01
2.89144218e-01 1.35853434e+00 -5.76729536e-01 -6.49611115e-01
-6.43546939e-01 7.57060587e-01 6.05403662e-01 3.58351678e-01
-4.44790035e-01 -9.99054536e-02 4.69180495e-01 1.43293813e-01
9.06837583e-01 2.39156395e-01 -9.28786322e-02 -1.19994700e+00
-4.14025784e-01 -6.28579438e-01 6.44044697e-01 -4.56648231e-01
-1.44127452e+00 8.24047029e-01 -1.99209884e-01 -1.26334620e+00
-5.48559368e-01 -5.17359376e-01 -6.51609898e-01 1.08874929e+00
-9.96430337e-01 -1.02369440e+00 -1.47478342e-01 5.90801656e-01
3.73757422e-01 -5.36708415e-01 1.08839428e+00 4.45226490e-01
-2.20237046e-01 7.72391558e-01 4.33137029e-01 -2.54216969e-01
8.13428462e-01 -1.19283295e+00 8.12760070e-02 6.03320718e-01
3.06702763e-01 9.38938618e-01 5.33738434e-01 -4.31550413e-01
-1.04439604e+00 -1.30222249e+00 1.13914478e+00 -4.39512193e-01
5.29002786e-01 -2.47631788e-01 -8.91372263e-01 8.56668055e-01
1.55620594e-02 -7.56778270e-02 1.34832454e+00 7.36384094e-01
-4.15982991e-01 -3.55800867e-01 -1.00901520e+00 3.68144155e-01
7.24182665e-01 -1.97430640e-01 -7.56757379e-01 3.43625814e-01
7.24131525e-01 1.19378887e-01 -1.11108112e+00 1.72629431e-01
8.98914516e-01 -1.01531410e+00 9.96519744e-01 -9.57770288e-01
2.58704990e-01 -3.94193083e-01 -4.76111621e-01 -1.76885283e+00
-9.70310032e-01 -5.84475771e-02 -3.53650659e-01 1.02083099e+00
7.56261110e-01 -6.86700940e-01 6.15333378e-01 7.01220274e-01
-2.70258963e-01 -7.25269496e-01 -6.93047047e-02 -3.12515438e-01
-2.02276975e-01 -2.05915213e-01 8.98193359e-01 8.33780885e-01
1.87442020e-01 5.90469420e-01 -7.16380954e-01 1.72935173e-01
2.53409535e-01 5.57995379e-01 7.52241850e-01 -1.73722887e+00
-7.78834343e-01 -2.89770097e-01 -1.17945552e-01 -1.26466024e+00
-1.22694418e-01 -1.32811904e+00 -1.80105254e-01 -1.85523641e+00
-2.39761606e-01 -8.09397340e-01 -1.01606226e+00 5.77512622e-01
-1.52665079e-01 1.99638471e-01 3.36022228e-01 1.42462909e-01
-7.42134035e-01 1.74667075e-01 9.44642186e-01 1.63517475e-01
-6.02057517e-01 6.17157459e-01 -1.02763736e+00 6.68222189e-01
6.78632796e-01 -3.49567473e-01 -6.22704268e-01 -2.02855557e-01
3.56507927e-01 -2.96452865e-02 -2.57157415e-01 -9.59443331e-01
5.00165045e-01 -9.49436873e-02 6.33149326e-01 -6.45557046e-01
2.65338063e-01 -9.66090977e-01 5.37311316e-01 6.81045353e-02
-6.92457497e-01 -3.40897322e-01 -3.50205809e-01 6.03643060e-01
-2.25444473e-02 -3.80633891e-01 5.04489958e-01 -1.11285038e-02
-6.56900167e-01 3.46910268e-01 -1.29564419e-01 -8.54417384e-01
7.92187512e-01 -9.87879634e-02 -7.36200288e-02 -3.32657665e-01
-1.19888222e+00 1.14542963e-02 1.94805577e-01 6.70622647e-01
3.27646971e-01 -1.21415770e+00 -5.73395133e-01 4.66515690e-01
2.29621306e-02 -3.99464488e-01 3.44458818e-01 4.42039102e-01
1.14585884e-01 3.24395865e-01 -1.43051177e-01 -2.07466990e-01
-1.08939588e+00 7.39267111e-01 8.49824175e-02 -4.44458634e-01
-4.76337552e-01 9.28919733e-01 1.19167782e-01 -7.70368814e-01
3.51747930e-01 -3.40617388e-01 -1.01328707e+00 4.81937826e-01
8.76676083e-01 1.67265192e-01 3.06260854e-01 -4.79178756e-01
1.19116068e-01 2.87460178e-01 -4.63911176e-01 2.06036910e-01
1.55040801e+00 -1.40049279e-01 -6.79681748e-02 5.45190096e-01
9.54810143e-01 -4.89531048e-02 -8.03599536e-01 -5.75512230e-01
-2.71915466e-01 -3.70715231e-01 2.88943201e-01 -9.96158719e-01
-1.12365460e+00 5.96691847e-01 5.51015198e-01 4.86037374e-01
9.27792728e-01 -1.52717948e-01 5.20708442e-01 4.53503311e-01
4.80574638e-01 -9.34163928e-01 -4.42573190e-01 5.84213793e-01
6.96679652e-01 -8.83536160e-01 5.50198592e-02 -1.50554433e-01
-5.14639080e-01 1.04189444e+00 3.99997890e-01 -4.81100976e-01
1.25795507e+00 -3.23572189e-01 2.31006909e-02 2.13342905e-01
-9.84370768e-01 -3.75066930e-03 8.53149712e-01 1.89172074e-01
8.00544143e-01 3.48134160e-01 -5.22354186e-01 1.38542581e+00
-1.99295297e-01 2.98254132e-01 1.65108547e-01 6.05316699e-01
-7.90066063e-01 -1.50364792e+00 3.09813581e-02 1.28128541e+00
-3.72835249e-01 -4.37574610e-02 4.37997617e-02 5.71198016e-02
4.42589134e-01 9.71239448e-01 1.76190600e-01 -9.74498272e-01
2.11624503e-01 2.86920667e-01 7.52084628e-02 -9.73142505e-01
-1.18678558e+00 4.13681604e-02 9.80569348e-02 -3.42109293e-01
-2.64128715e-01 -4.62237358e-01 -8.06814373e-01 1.58514716e-02
-6.69776559e-01 6.50533497e-01 7.54853189e-01 9.55309331e-01
4.71802324e-01 4.95670974e-01 8.27582955e-01 -1.20711982e+00
-6.13038540e-01 -1.17042494e+00 -7.01689184e-01 3.98776293e-01
5.68459556e-03 -5.97118437e-01 -4.35002297e-01 -9.29909721e-02] | [10.141571998596191, 5.702173709869385] |
d705caf7-573e-48f4-967c-71cbdb72f3b7 | occlusion-aware-instance-segmentation-via | 2208.04438 | null | https://arxiv.org/abs/2208.04438v2 | https://arxiv.org/pdf/2208.04438v2.pdf | Occlusion-Aware Instance Segmentation via BiLayer Network Architectures | Segmenting highly-overlapping image objects is challenging, because there is typically no distinction between real object contours and occlusion boundaries on images. Unlike previous instance segmentation methods, we model image formation as a composition of two overlapping layers, and propose Bilayer Convolutional Network (BCNet), where the top layer detects occluding objects (occluders) and the bottom layer infers partially occluded instances (occludees). The explicit modeling of occlusion relationship with bilayer structure naturally decouples the boundaries of both the occluding and occluded instances, and considers the interaction between them during mask regression. We investigate the efficacy of bilayer structure using two popular convolutional network designs, namely, Fully Convolutional Network (FCN) and Graph Convolutional Network (GCN). Further, we formulate bilayer decoupling using the vision transformer (ViT), by representing instances in the image as separate learnable occluder and occludee queries. Large and consistent improvements using one/two-stage and query-based object detectors with various backbones and network layer choices validate the generalization ability of bilayer decoupling, as shown by extensive experiments on image instance segmentation benchmarks (COCO, KINS, COCOA) and video instance segmentation benchmarks (YTVIS, OVIS, BDD100K MOTS), especially for heavy occlusion cases. Code and data are available at https://github.com/lkeab/BCNet. | ['Chi-Keung Tang', 'Yu-Wing Tai', 'Lei Ke'] | 2022-08-08 | null | null | null | null | ['video-instance-segmentation'] | ['computer-vision'] | [ 3.06581795e-01 4.03167188e-01 -2.90070623e-01 -2.43268251e-01
-4.70498443e-01 -6.83928967e-01 4.96408999e-01 -4.51663956e-02
-2.20546365e-01 3.82923961e-01 -3.21021885e-01 -2.44349048e-01
5.06377667e-02 -7.20647693e-01 -1.06822371e+00 -5.14028311e-01
-6.98387623e-02 4.83998448e-01 8.99393737e-01 -1.26429228e-02
2.20901873e-02 4.21418846e-01 -1.42050731e+00 5.74905813e-01
8.62716436e-01 1.19979298e+00 6.14321157e-02 7.17881083e-01
-2.37008840e-01 6.83453083e-01 -6.44269526e-01 -5.10981977e-01
6.14393115e-01 4.38591279e-03 -9.09310937e-01 3.96223962e-01
9.52624619e-01 -3.53642225e-01 -2.27441415e-01 9.83109057e-01
1.91060022e-01 -1.30700678e-01 6.05223119e-01 -1.39155638e+00
-6.85221195e-01 5.23051918e-01 -9.96642292e-01 2.13526160e-01
-4.80280146e-02 4.13109958e-01 8.69626522e-01 -8.92496645e-01
7.42319286e-01 1.33884704e+00 7.09443688e-01 3.63913655e-01
-1.47264957e+00 -5.75135231e-01 6.55564129e-01 9.13880244e-02
-1.41349149e+00 -2.97034085e-01 6.46524310e-01 -6.12485647e-01
6.30504727e-01 2.71066755e-01 5.87499976e-01 9.93926466e-01
-1.51393833e-02 1.14465594e+00 9.10950363e-01 -9.83482972e-02
5.20048961e-02 1.51778147e-01 4.84014958e-01 7.33413696e-01
4.28590775e-01 -1.97934099e-02 -6.75178319e-02 8.87513384e-02
9.52088594e-01 -1.35455588e-02 -3.58784467e-01 -5.75957000e-01
-7.51966894e-01 6.81641281e-01 4.94293213e-01 -6.97340220e-02
-1.69554755e-01 1.39487743e-01 2.04903603e-01 1.50233462e-01
3.98451090e-01 7.52359927e-02 -5.22704005e-01 3.65738779e-01
-1.02316773e+00 2.24580020e-01 8.65791619e-01 1.24797750e+00
1.03574979e+00 -6.39688224e-02 -3.47529590e-01 8.51484418e-01
1.70499876e-01 1.03834033e-01 5.54750022e-03 -1.10086894e+00
3.34748775e-01 8.98299575e-01 -1.18115500e-01 -8.56413186e-01
-3.45387578e-01 -4.35463488e-01 -7.04361856e-01 3.01576823e-01
5.60280144e-01 -1.87769756e-01 -1.54914689e+00 1.52884412e+00
6.19973779e-01 5.55552423e-01 -1.98545828e-01 9.63900506e-01
1.19442892e+00 4.58433747e-01 1.60813838e-01 2.02460676e-01
1.65393817e+00 -1.30425894e+00 -4.32486862e-01 -4.21015263e-01
3.93651485e-01 -6.21002972e-01 9.80048418e-01 2.93543667e-01
-1.25992048e+00 -7.54481494e-01 -9.38141763e-01 -4.62548316e-01
-4.39043015e-01 7.19035789e-02 6.53798342e-01 3.03271502e-01
-1.15967083e+00 2.74034649e-01 -8.26670647e-01 2.21565403e-02
9.17975366e-01 5.42101204e-01 -1.89991727e-01 -8.66287425e-02
-9.22151268e-01 2.33638689e-01 3.32701564e-01 3.12277436e-01
-1.09507108e+00 -9.46512520e-01 -1.01174664e+00 1.01494744e-01
6.86008394e-01 -6.16258979e-01 1.00901330e+00 -1.16410792e+00
-1.11218739e+00 1.11579144e+00 -7.64045119e-02 -5.94066083e-01
8.33330393e-01 -2.55297482e-01 6.14052303e-02 2.55135059e-01
1.34279594e-01 1.22582281e+00 9.47396576e-01 -1.61357880e+00
-7.20831335e-01 -2.72463381e-01 4.49914604e-01 1.13910109e-01
1.92775518e-01 -1.71428293e-01 -1.10314691e+00 -4.47115272e-01
1.93029612e-01 -7.58233607e-01 -2.20383987e-01 2.23864704e-01
-9.54503477e-01 -2.80047625e-01 1.21212292e+00 -5.96734941e-01
1.13413143e+00 -2.16157126e+00 1.34278029e-01 5.79857938e-02
4.84404325e-01 3.66395980e-01 -3.06689620e-01 3.31434757e-02
-4.22271602e-02 2.40233868e-01 -5.39209843e-01 -6.09266520e-01
-1.46814123e-01 5.08530974e-01 -2.31409162e-01 3.40430498e-01
5.48040211e-01 1.12902653e+00 -5.53250611e-01 -6.61813319e-01
3.67373765e-01 5.08773983e-01 -6.07111692e-01 8.96039456e-02
-5.11434674e-01 3.03466797e-01 -3.56283724e-01 9.42143857e-01
9.24249828e-01 -5.52474678e-01 6.60571381e-02 -4.32677448e-01
-1.71499047e-03 4.22062986e-02 -1.42267942e+00 1.48982179e+00
7.64942840e-02 6.56800032e-01 4.83072281e-01 -8.82699311e-01
5.48830152e-01 9.73843113e-02 4.54829007e-01 -7.65610158e-01
9.25454870e-02 -1.72868505e-01 -1.14085391e-01 -5.85958958e-01
3.66278619e-01 3.97192001e-01 2.10487127e-01 2.13412307e-02
6.29601181e-02 -5.40434457e-02 5.38213253e-01 3.93603712e-01
9.70295787e-01 2.59078145e-01 -1.31590202e-01 -2.84891397e-01
3.71338069e-01 4.91034649e-02 8.52992952e-01 9.12609339e-01
-3.52676123e-01 9.70761538e-01 9.85829234e-01 -5.22338331e-01
-7.16547549e-01 -1.04350996e+00 -5.21548212e-01 9.71481025e-01
7.87308812e-01 -3.29089016e-01 -1.02641737e+00 -7.29448617e-01
2.76462555e-01 3.62581521e-01 -7.62391031e-01 1.85381278e-01
-8.83484423e-01 -7.31307328e-01 5.06784379e-01 6.10540271e-01
5.35638988e-01 -1.26162660e+00 -6.40716851e-01 8.01191479e-02
-1.48765787e-01 -1.57698202e+00 -4.76050615e-01 1.94164351e-01
-6.21574104e-01 -1.48062730e+00 -4.25741702e-01 -8.52888584e-01
5.84128499e-01 2.40239441e-01 1.40766728e+00 4.58521098e-01
-6.30075455e-01 2.56377101e-01 2.07986161e-02 -2.28787690e-01
-9.17968377e-02 1.65600196e-01 -5.42553544e-01 5.11997715e-02
8.62859190e-02 -4.77615803e-01 -8.64009321e-01 5.21046996e-01
-1.06005085e+00 4.59683686e-01 4.14234191e-01 5.65448344e-01
9.10317719e-01 -2.82326221e-01 2.09799990e-01 -1.28069985e+00
1.94207489e-01 -4.55299437e-01 -8.56386065e-01 2.48905078e-01
-2.90683627e-01 -2.91343391e-01 1.56499922e-01 -5.67185402e-01
-8.36539507e-01 1.48421481e-01 -5.77948689e-02 -7.41203666e-01
-4.98856992e-01 3.10623161e-02 -2.10225284e-01 2.99270172e-02
4.51952398e-01 -5.45877628e-02 -1.85902715e-01 -5.03021121e-01
4.54907477e-01 2.67788172e-01 6.28714085e-01 -7.87465572e-01
5.91215014e-01 6.68190837e-01 -4.24541026e-01 -8.12901020e-01
-7.77125716e-01 -4.16179806e-01 -7.18006968e-01 -1.04487531e-01
1.12978733e+00 -8.49512219e-01 -7.16676533e-01 4.81548756e-01
-1.27983367e+00 -8.77965748e-01 -4.55119342e-01 -1.10090747e-01
-2.36072674e-01 1.38713688e-01 -8.90478909e-01 -5.86417735e-01
-1.03152387e-01 -1.56665254e+00 1.48384941e+00 4.35449302e-01
7.96949714e-02 -8.75183344e-01 -3.99671078e-01 5.54639220e-01
3.21431942e-02 6.84345126e-01 8.25012803e-01 -3.83095384e-01
-1.11720681e+00 1.17091566e-01 -6.31377101e-01 3.50495398e-01
1.02542825e-02 4.21561867e-01 -1.03211570e+00 -2.42848873e-01
-3.34359825e-01 -3.48880410e-01 1.13380599e+00 7.86173522e-01
1.36599708e+00 -1.39612511e-01 -5.20904005e-01 9.82753217e-01
1.44725692e+00 9.76059213e-02 8.10635328e-01 7.97035396e-02
1.13668418e+00 5.82598090e-01 9.85439960e-03 -2.46269889e-02
4.61463720e-01 5.44035494e-01 6.93841457e-01 -6.30311787e-01
-5.90919435e-01 7.57963732e-02 2.71890266e-03 2.26298198e-01
1.31557897e-01 -4.25602317e-01 -8.13714504e-01 5.73528647e-01
-1.85942602e+00 -5.08888841e-01 -4.10973132e-01 1.79108989e+00
8.46520364e-01 4.20108765e-01 1.34693518e-01 -1.32534161e-01
6.83740199e-01 2.64545530e-01 -8.43280613e-01 -1.25602245e-01
-2.46869355e-01 1.85717747e-01 5.61849892e-01 6.50854349e-01
-1.38267136e+00 1.22977877e+00 5.48853683e+00 8.37094724e-01
-1.02861524e+00 9.68783349e-02 1.14228857e+00 -2.08000597e-02
-9.03604254e-02 2.43630316e-02 -8.69178414e-01 4.08799678e-01
1.63828477e-01 5.28501868e-01 1.79767504e-01 7.60591209e-01
-6.02460094e-02 -1.28896266e-01 -1.16419542e+00 6.01304114e-01
-3.60682607e-01 -1.50965953e+00 1.00853741e-01 1.47119816e-02
8.17799151e-01 2.35888734e-01 5.99397682e-02 2.46626422e-01
3.50161999e-01 -9.84553456e-01 9.40256119e-01 2.00268194e-01
7.53752530e-01 -2.80263722e-01 3.42007309e-01 1.03471242e-01
-1.38801360e+00 6.64155651e-03 -2.06681356e-01 1.75884023e-01
1.89642552e-02 4.90692586e-01 -4.72721726e-01 5.09919405e-01
9.01587665e-01 7.20980644e-01 -6.32641256e-01 9.55477238e-01
-2.73064345e-01 7.04745710e-01 -4.43640321e-01 7.27392256e-01
3.99612904e-01 -3.49933833e-01 5.10933757e-01 1.24312150e+00
-4.86677557e-01 4.82658595e-02 4.43792462e-01 1.28396308e+00
-3.35504621e-01 -1.81153476e-01 -2.86697268e-01 3.88646662e-01
3.27653766e-01 1.41055536e+00 -1.19805586e+00 -4.53883260e-01
-5.19730210e-01 8.05489480e-01 4.67871636e-01 7.86890328e-01
-1.09311914e+00 8.07490498e-02 8.99471641e-01 3.60761702e-01
6.16681933e-01 -4.59455736e-02 -3.69290411e-01 -1.02749920e+00
1.04654059e-01 -8.79215121e-01 5.36701977e-01 -4.41197872e-01
-1.13965797e+00 4.91586894e-01 1.49817094e-01 -8.29762638e-01
3.77130181e-01 -6.90919518e-01 -7.40119040e-01 6.41220510e-01
-1.67074156e+00 -1.18348658e+00 -5.67985237e-01 5.15781999e-01
8.18386197e-01 5.37841320e-01 1.66784093e-01 5.08983612e-01
-1.02430785e+00 4.97541219e-01 -5.04685879e-01 4.99881834e-01
2.82633185e-01 -1.40627706e+00 5.04198194e-01 7.40337431e-01
2.78814375e-01 3.55378300e-01 2.87658483e-01 -6.60381377e-01
-1.09455597e+00 -1.34103847e+00 8.68423730e-02 -4.68445480e-01
3.30417126e-01 -6.67275608e-01 -1.23978937e+00 7.46937096e-01
2.00811401e-01 4.71803606e-01 7.26844594e-02 -1.82589754e-01
-3.95401806e-01 6.51966408e-02 -9.01925266e-01 6.37595057e-01
1.28563094e+00 -2.35449716e-01 -2.30843887e-01 5.24952471e-01
1.09843540e+00 -7.47720480e-01 -6.58134103e-01 6.72225118e-01
5.00230014e-01 -1.24357462e+00 1.24681211e+00 -6.36745691e-01
5.16145408e-01 -3.89117718e-01 9.60165560e-02 -5.96041858e-01
-4.21137884e-02 -4.73241806e-01 -2.89445788e-01 1.06234932e+00
3.84751618e-01 -4.90036488e-01 1.01723671e+00 7.13222206e-01
-2.60308027e-01 -1.22506237e+00 -7.01868832e-01 -5.09605765e-01
8.80034119e-02 -3.59800637e-01 5.66870689e-01 9.15396810e-01
-8.80967915e-01 1.97906479e-01 -6.54542670e-02 3.77234161e-01
5.86591482e-01 9.81521681e-02 8.34358633e-01 -1.14520955e+00
-4.06405121e-01 -6.38357580e-01 -2.25708872e-01 -1.27180195e+00
-3.43241505e-02 -7.88157523e-01 -1.36714637e-01 -1.63900685e+00
2.42446270e-02 -7.90898621e-01 -1.73960254e-01 7.09037840e-01
-3.69019151e-01 5.20143688e-01 2.17569172e-01 2.81217456e-01
-6.83770359e-01 1.43475235e-01 1.48590028e+00 -3.74455452e-01
-4.40757006e-01 -8.67756903e-02 -5.57336807e-01 9.55757976e-01
6.83004797e-01 -3.91146809e-01 -5.83875954e-01 -7.22225189e-01
-2.15278432e-01 -3.31940018e-02 7.90965259e-01 -8.28998864e-01
1.66971162e-01 1.78901870e-02 4.06300724e-01 -5.74051321e-01
2.59850919e-01 -6.76047981e-01 2.43050158e-01 3.01184803e-01
-2.25750685e-01 -1.25438765e-01 5.00036657e-01 4.99143451e-01
-1.12143196e-01 -1.75607111e-02 7.83403635e-01 -2.04665303e-01
-8.00480902e-01 6.71661973e-01 1.05378471e-01 4.35574949e-01
1.10425019e+00 -6.35804713e-01 -4.94168282e-01 9.39444825e-02
-7.89561629e-01 5.92316210e-01 3.67019683e-01 5.63030720e-01
5.56964278e-01 -7.76822567e-01 -5.69192886e-01 4.50460434e-01
-8.91709179e-02 8.04707944e-01 4.01694238e-01 9.29509819e-01
-6.96037233e-01 -9.97093692e-02 1.62127361e-01 -1.00468206e+00
-1.11479330e+00 4.49293971e-01 6.75629020e-01 -1.97056517e-01
-9.04588580e-01 1.07612634e+00 1.06349361e+00 -2.39794284e-01
5.70390105e-01 -6.12574577e-01 -7.97979981e-02 -1.73953548e-02
1.83960259e-01 1.61209285e-01 -1.21123865e-01 -5.09741127e-01
-2.50368595e-01 5.21553695e-01 -4.07127053e-01 3.04714978e-01
1.08367550e+00 3.14377956e-02 -1.81624234e-01 1.02275297e-01
1.09519899e+00 -4.17038888e-01 -1.80384123e+00 -2.11088985e-01
1.36553928e-01 -1.63299054e-01 -1.32441968e-01 -5.31754136e-01
-1.43341625e+00 7.83309639e-01 4.97407734e-01 3.89821142e-01
8.77737820e-01 1.96187273e-01 8.01908553e-01 -7.04665631e-02
4.34807874e-02 -9.51074600e-01 4.08910662e-02 2.25107446e-01
7.37345874e-01 -1.19377434e+00 -1.46949232e-01 -8.58965635e-01
-4.46894735e-01 7.90290415e-01 1.11237097e+00 -2.69539237e-01
6.90914929e-01 6.16039634e-01 2.38621265e-01 -5.33096611e-01
-6.25520766e-01 -2.03667223e-01 4.65266943e-01 3.57169807e-01
1.33529648e-01 6.76178634e-02 -8.27195048e-02 4.72463012e-01
1.39148653e-01 -3.39387178e-01 3.06472927e-01 8.65940511e-01
-1.90316543e-01 -9.20161903e-01 -2.01220378e-01 5.34374833e-01
-3.95928144e-01 -1.22858547e-01 -4.52587217e-01 1.29937148e+00
7.28183031e-01 6.73114836e-01 3.84487569e-01 -6.49357364e-02
3.46406937e-01 -1.40912205e-01 2.43054956e-01 -7.28227973e-01
-8.03401887e-01 3.10902387e-01 -1.59492999e-01 -8.25657427e-01
-3.47217411e-01 -3.52322847e-01 -1.37556565e+00 7.45752081e-02
-4.31929201e-01 -9.72218737e-02 3.54964942e-01 7.78755963e-01
5.64792871e-01 8.93485546e-01 5.90516776e-02 -9.97540176e-01
-5.92139103e-02 -6.28942490e-01 -4.03091669e-01 6.33836329e-01
3.80046606e-01 -7.60014951e-01 -2.21923962e-01 3.06851387e-01] | [9.455224990844727, 0.18913201987743378] |
bdf3343a-c9c2-48d1-8ce7-85994be49041 | edge-aided-sensor-data-sharing-in-vehicular | 2206.08882 | null | https://arxiv.org/abs/2206.08882v1 | https://arxiv.org/pdf/2206.08882v1.pdf | Edge-Aided Sensor Data Sharing in Vehicular Communication Networks | Sensor data sharing in vehicular networks can significantly improve the range and accuracy of environmental perception for connected automated vehicles. Different concepts and schemes for dissemination and fusion of sensor data have been developed. It is common to these schemes that measurement errors of the sensors impair the perception quality and can result in road traffic accidents. Specifically, when the measurement error from the sensors (also referred as measurement noise) is unknown and time varying, the performance of the data fusion process is restricted, which represents a major challenge in the calibration of sensors. In this paper, we consider sensor data sharing and fusion in a vehicular network with both, vehicle-to-infrastructure and vehicle-to-vehicle communication. We propose a method, named Bidirectional Feedback Noise Estimation (BiFNoE), in which an edge server collects and caches sensor measurement data from vehicles. The edge estimates the noise and the targets alternately in double dynamic sliding time windows and enhances the distributed cooperative environment sensing at each vehicle with low communication costs. We evaluate the proposed algorithm and data dissemination strategy in an application scenario by simulation and show that the perception accuracy is on average improved by around 80 % with only 12 kbps uplink and 28 kbps downlink bandwidth. | ['Andreas Festag', 'Alois Knoll', 'Numan Senel', 'Anupama Hegde', 'Rui Song'] | 2022-06-17 | null | null | null | null | ['noise-estimation'] | ['medical'] | [ 1.10959843e-01 1.49578020e-01 -5.20700812e-02 -3.49310666e-01
-4.49822336e-01 -5.30942857e-01 4.79583174e-01 4.86960143e-01
-6.06818020e-01 8.37538719e-01 -2.59011954e-01 -3.37854028e-01
-7.62410164e-02 -9.89825904e-01 -8.80140424e-01 -9.92211461e-01
-1.60022542e-01 -9.70669463e-02 8.86885762e-01 -1.51270509e-01
-1.63254380e-01 3.67579371e-01 -1.78457808e+00 -4.18969274e-01
8.87124062e-01 1.51944184e+00 5.50302148e-01 9.10688400e-01
6.52673393e-02 3.00226510e-01 -7.33063936e-01 1.33500963e-01
7.72411823e-02 2.62180030e-01 3.16792250e-01 -1.89427540e-01
-1.60808295e-01 -4.11130905e-01 -2.50119925e-01 1.16922367e+00
5.74393451e-01 -1.57787964e-01 4.12922621e-01 -1.93081963e+00
-3.55890803e-02 3.64904463e-01 -6.98323250e-01 8.06910768e-02
-9.28631276e-02 -1.57265306e-01 -9.82380658e-02 -6.35595739e-01
3.79806459e-01 1.20188439e+00 6.73739254e-01 4.12494689e-01
-6.24229968e-01 -1.11372256e+00 1.18491001e-01 6.60045505e-01
-1.55181217e+00 -9.13771093e-01 6.74659431e-01 -1.71023890e-01
4.70832169e-01 3.49042088e-01 6.58907592e-01 4.53759402e-01
5.35831749e-01 3.29108685e-01 8.27266574e-01 -1.05513632e-02
6.43414438e-01 3.77995193e-01 1.16596006e-01 -6.44541606e-02
1.17321813e+00 4.13729429e-01 -3.34704846e-01 -9.30065662e-02
-2.64098477e-02 -1.77769020e-01 -6.47599176e-02 -2.52127558e-01
-7.87846565e-01 4.80496049e-01 1.23704806e-01 -1.01632744e-01
-8.99977326e-01 5.37923157e-01 5.79619169e-01 6.16158247e-01
3.48418862e-01 -9.54547822e-01 -3.09194595e-01 7.00408872e-03
-4.69661772e-01 1.20424509e-01 6.37008071e-01 1.27334023e+00
7.22483873e-01 1.59094036e-01 1.11813456e-01 3.27345848e-01
7.77355433e-01 1.69320393e+00 -2.67596900e-01 -1.20662189e+00
3.18253130e-01 8.89815167e-02 5.03196359e-01 -1.23369181e+00
-4.49771523e-01 -1.19393274e-01 -1.01146281e+00 7.13215709e-01
-1.80697888e-01 -8.56132209e-01 -6.91580057e-01 1.82813764e+00
9.87757623e-01 5.46079874e-01 9.08280730e-01 7.65612602e-01
7.89985061e-01 7.05858052e-01 2.27540851e-01 -8.29051495e-01
9.27840233e-01 -2.26116359e-01 -1.31963241e+00 -1.81467891e-01
3.53029698e-01 -7.37122476e-01 -2.85915047e-01 3.15043181e-01
-9.70852256e-01 -5.99221349e-01 -1.44574821e+00 6.18507624e-01
-4.11068320e-01 -1.40313268e-01 -4.54918966e-02 8.95026445e-01
-1.11388910e+00 -2.21324906e-01 -7.53994346e-01 -4.68979448e-01
9.76509526e-02 4.41100180e-01 8.10218975e-03 -2.66604304e-01
-1.36219919e+00 8.24967027e-01 1.19569674e-01 1.64677292e-01
-8.97001028e-01 -7.09397554e-01 -5.98884404e-01 -4.14657414e-01
5.16029596e-01 -2.92411625e-01 1.09083188e+00 -5.95400751e-01
-1.30547333e+00 -1.42314300e-01 -3.85912895e-01 -7.17070520e-01
4.68220413e-01 1.31617010e-01 -7.76309133e-01 -1.23430170e-01
-1.53335817e-02 3.28787297e-01 4.12516356e-01 -1.70644271e+00
-1.12358713e+00 -5.06014407e-01 -1.81411117e-01 1.54227048e-01
1.77543640e-01 -2.98245013e-01 -3.52961332e-01 2.97014624e-01
1.34299785e-01 -9.10601377e-01 -5.04719198e-01 3.02503377e-01
6.07950501e-02 -2.93351740e-01 1.48124444e+00 -3.16802114e-01
9.72568572e-01 -2.28765130e+00 -6.59582794e-01 5.74332237e-01
1.80441201e-01 3.81604195e-01 -2.11739298e-02 3.50410670e-01
6.98904634e-01 -2.89019346e-01 2.67705172e-01 -3.32778811e-01
-2.57526219e-01 6.10272765e-01 4.18777615e-02 7.83493876e-01
-5.53973198e-01 3.08481395e-01 -9.28675711e-01 -3.58043075e-01
5.93528688e-01 7.94167638e-01 2.86788084e-02 1.92598999e-02
1.65275097e-01 1.75043821e-01 -7.14348316e-01 3.21113467e-01
1.49191070e+00 3.89116555e-01 1.77659869e-01 -3.30618113e-01
-5.15757263e-01 -2.02650800e-01 -1.76296973e+00 9.94647980e-01
-1.96292788e-01 8.24614465e-01 1.08248866e+00 -9.34652269e-01
7.95931458e-01 6.47384465e-01 7.48595178e-01 -8.48091900e-01
3.17188025e-01 2.21437469e-01 -2.55972207e-01 -5.75494707e-01
5.48379540e-01 5.18745482e-01 1.51330745e-02 -5.66324852e-02
-7.30904162e-01 1.03446588e-01 -5.26028909e-02 3.29766244e-01
1.16447234e+00 -5.99249959e-01 2.07617149e-01 -4.75731399e-03
5.48736870e-01 -2.53333952e-02 8.66712272e-01 7.10240185e-01
-5.18384635e-01 -5.41301906e-01 -2.65192091e-01 3.49347919e-01
-1.03221893e+00 -1.03143871e+00 4.56987731e-02 5.97518086e-01
1.03145123e+00 -1.28172040e-01 -5.70282221e-01 1.29084706e-01
5.00845015e-01 5.20391405e-01 -2.72193968e-01 -2.68898159e-01
-1.32568300e-01 -2.60230631e-01 3.24368566e-01 3.66776586e-01
6.23142421e-01 -1.99602395e-01 -7.70977318e-01 6.07275903e-01
3.92785557e-02 -1.33079064e+00 -3.33727412e-02 -1.34478480e-01
-2.62438297e-01 -1.19295049e+00 7.90849105e-02 -3.69841456e-01
4.70616132e-01 1.26611793e+00 3.38805705e-01 7.93991387e-02
2.65737474e-01 8.58539760e-01 -3.03481430e-01 -1.36932957e+00
-5.01277685e-01 -7.38968730e-01 4.18295920e-01 1.09313428e-01
6.14074707e-01 -2.51816660e-01 -8.08075726e-01 2.62632549e-01
-5.77277422e-01 -3.03363204e-01 3.60188812e-01 7.29977936e-02
6.23226702e-01 3.53659302e-01 9.83251929e-01 -2.82012731e-01
5.80451310e-01 -9.98767078e-01 -9.66677904e-01 -1.47529274e-01
-6.59892678e-01 -5.66414833e-01 1.25200421e-01 -1.60019800e-01
-9.70178902e-01 1.96794897e-01 8.92530456e-02 -2.03275159e-01
-5.25433663e-03 2.25997388e-01 -3.95469844e-01 -5.36003470e-01
3.39594990e-01 -3.20361648e-03 3.16650510e-01 4.17274795e-02
3.30345750e-01 1.20464420e+00 4.01803881e-01 1.48823008e-01
8.27594578e-01 6.80745065e-01 2.82701045e-01 -1.34041166e+00
2.57276237e-01 -6.50623083e-01 1.63851023e-01 -9.18237984e-01
6.86938226e-01 -1.51190388e+00 -1.21305871e+00 5.34124315e-01
-1.44455373e+00 2.35668831e-02 -4.00102623e-02 7.54744172e-01
-9.63819101e-02 3.79478067e-01 3.98059860e-02 -1.51666069e+00
-1.65197492e-01 -1.00804424e+00 6.90664351e-01 3.01277250e-01
3.27378064e-01 -6.68411851e-01 -2.33721346e-01 -1.65270105e-01
7.10754514e-01 3.79464626e-01 -6.39372319e-02 -1.86481446e-01
-9.85579431e-01 -1.27373189e-01 -3.73768926e-01 2.89180219e-01
2.63693426e-02 -1.40160084e-01 -8.75885904e-01 -4.51890916e-01
-1.85508266e-01 9.96970236e-02 5.56682110e-01 6.98059082e-01
5.07291913e-01 2.71989275e-02 -9.10618007e-01 1.92021243e-02
1.60464513e+00 6.16666913e-01 4.75389749e-01 -2.07932949e-01
2.56982833e-01 4.04699177e-01 1.02300096e+00 5.73161602e-01
8.54339242e-01 4.19383466e-01 1.01909721e+00 1.25877246e-01
-1.50651455e-01 2.59839267e-01 4.65897173e-01 7.44172931e-01
2.23767385e-01 -8.45976651e-01 -1.87035799e-01 8.50386500e-01
-1.98706210e+00 -7.39851415e-01 -6.10601664e-01 2.20481110e+00
2.72220284e-01 3.88548039e-02 -2.20913813e-01 3.54805499e-01
9.03249562e-01 5.79680204e-02 -7.38406718e-01 -2.42488205e-01
-3.19400430e-02 -6.11885488e-01 1.38991976e+00 8.64886045e-01
-5.69242060e-01 3.90897095e-01 5.75616169e+00 7.43020296e-01
-9.07009721e-01 5.03093302e-01 -6.86287880e-02 2.75800377e-01
-3.14921379e-01 -3.91602159e-01 -7.99263537e-01 7.08403647e-01
1.37962687e+00 -3.85349929e-01 -7.69904465e-04 6.37825429e-01
9.77158785e-01 -7.30750680e-01 -5.85248768e-01 8.42894077e-01
-2.28163049e-01 -1.22669172e+00 -5.26333869e-01 1.62926927e-01
6.17341638e-01 4.67113912e-01 -3.42562377e-01 -2.08318397e-01
6.57837689e-01 -2.92062104e-01 6.10303700e-01 6.09931529e-01
4.96629924e-01 -8.52620780e-01 8.56108606e-01 3.91665250e-01
-1.71570671e+00 -2.73571223e-01 -5.36516845e-01 -3.88468653e-02
5.77277482e-01 9.02224600e-01 -5.76780558e-01 5.68866014e-01
6.99496210e-01 3.73291016e-01 1.82305828e-01 1.14781356e+00
2.28130519e-01 6.44322991e-01 -8.66288662e-01 -4.73503739e-01
-6.57443628e-02 -8.52467194e-02 1.05216539e+00 1.01155078e+00
7.12815881e-01 5.22923589e-01 3.63253087e-01 8.35684314e-02
1.12923205e-01 -1.93172082e-01 -9.32390034e-01 6.86602354e-01
1.37615168e+00 1.01339591e+00 -1.21906184e-01 -5.84780574e-01
-5.49662173e-01 1.59959167e-01 -4.29816812e-01 7.18475461e-01
-8.05241346e-01 -3.98721993e-01 1.23886347e+00 -2.60588694e-02
4.28296715e-01 -4.93856400e-01 -2.74597257e-01 -7.62510747e-02
6.10641800e-02 -2.16568448e-02 -2.32003003e-01 -5.76149523e-01
-7.84137905e-01 1.57957047e-01 -1.17614500e-01 -1.18961072e+00
8.17698389e-02 -6.76941946e-02 -3.41639638e-01 5.07772863e-01
-1.79181170e+00 -7.45655060e-01 -8.09753776e-01 8.08901548e-01
1.92157239e-01 -2.89363060e-02 1.22695133e-01 7.19525337e-01
-1.37487113e-01 2.42934898e-01 8.11319411e-01 -4.59584028e-01
4.26779330e-01 -4.24974889e-01 -1.97953254e-01 8.81660044e-01
-8.22705150e-01 -2.61617512e-01 1.25597048e+00 -9.30948555e-01
-1.85987020e+00 -1.40876031e+00 6.04514003e-01 1.29537806e-01
3.27874601e-01 -1.90196082e-01 -5.78695834e-01 1.73339367e-01
4.85027373e-01 1.14879511e-01 3.28851253e-01 -7.67739296e-01
3.77024829e-01 -8.72027993e-01 -1.60085440e+00 3.03133458e-01
7.37001836e-01 -3.52317914e-02 1.16975784e-01 1.54488727e-01
9.15887058e-01 -2.08025962e-01 -6.39366388e-01 4.29535359e-01
5.46660602e-01 -5.69463372e-01 6.17790401e-01 1.45806193e-01
-9.62221026e-01 -8.09335113e-01 -6.34629309e-01 -1.36525178e+00
1.24810711e-02 -5.39294899e-01 9.99591872e-02 1.26074755e+00
1.96162328e-01 -9.22566831e-01 6.69031143e-01 5.10233343e-01
-2.02692777e-01 5.02717383e-02 -1.46212637e+00 -5.51070213e-01
-5.49683034e-01 -1.05988741e+00 6.88664019e-01 1.38083071e-01
-1.88208610e-01 5.90526946e-02 -3.08398366e-01 9.71790969e-01
1.26035798e+00 -7.21530497e-01 1.07837832e+00 -1.37428916e+00
7.13110626e-01 3.25538039e-01 -6.54236078e-01 -1.29580081e+00
2.90234480e-02 -1.28759369e-01 6.02828681e-01 -1.70169640e+00
-3.71905774e-01 -6.92499280e-01 -1.42552942e-01 -2.02467918e-01
3.04228067e-01 1.04631476e-01 -2.71356851e-02 -3.08564961e-01
-7.92043209e-01 5.40596187e-01 8.17323327e-01 -1.37516007e-01
-6.87798411e-02 2.05622748e-01 -3.64084065e-01 5.72357833e-01
8.44866574e-01 -3.71256202e-01 -8.56168985e-01 -4.51771140e-01
7.60659873e-02 4.22461629e-01 4.66775805e-01 -1.31394887e+00
9.30791259e-01 -2.47319892e-01 3.25500928e-02 -1.06573570e+00
5.98797917e-01 -1.75564253e+00 6.24458849e-01 6.33838356e-01
2.04600662e-01 -1.14604935e-01 2.72487789e-01 1.39144349e+00
-1.02379872e-02 3.77495587e-01 7.33664811e-01 4.65185851e-01
-9.68387485e-01 3.12329262e-01 -1.04512501e+00 -6.65541887e-01
1.73108935e+00 -4.04651582e-01 -3.98439556e-01 -7.10927963e-01
-4.18395191e-01 7.60281503e-01 9.65960100e-02 4.02915806e-01
8.21842074e-01 -1.34189010e+00 -7.70164192e-01 1.38251528e-01
1.07820183e-01 -2.80496985e-01 5.78507185e-01 8.92715752e-01
1.64237440e-01 5.06897271e-01 1.12773322e-01 -9.65369582e-01
-1.63909018e+00 3.33987355e-01 1.75044864e-01 6.89277112e-01
4.45333421e-02 4.09058899e-01 -2.25524992e-01 8.19804892e-02
4.21702743e-01 -2.60195076e-01 -3.43086869e-01 -1.86632648e-01
6.26208305e-01 8.95897508e-01 -2.73205191e-02 -9.80095506e-01
-3.96587968e-01 6.07679188e-01 3.82038951e-01 -9.31404531e-02
6.63185775e-01 -1.31194341e+00 1.25302508e-01 2.23319456e-01
8.69168222e-01 4.39919271e-02 -1.34757423e+00 -4.03727919e-01
-3.97454739e-01 -3.20123315e-01 5.30536354e-01 -4.05872166e-01
-1.06685674e+00 2.03296334e-01 1.25792372e+00 4.61436689e-01
9.99297678e-01 -2.10178122e-01 7.83744633e-01 3.09128881e-01
8.15770388e-01 -1.43915856e+00 -5.67823410e-01 2.92876124e-01
3.87938827e-01 -1.35149682e+00 -1.77081853e-01 -7.48050511e-01
-2.08074555e-01 6.17877424e-01 5.79440653e-01 3.64187993e-02
1.21787155e+00 9.02227402e-01 2.97758728e-01 3.36989239e-02
-1.05136061e+00 -4.14598227e-01 -4.14400905e-01 1.14710367e+00
-3.31088632e-01 3.34057659e-01 -5.63006163e-01 3.13102186e-01
2.50391513e-01 1.95724651e-01 6.53822064e-01 8.65462363e-01
-1.28559172e+00 -7.51792610e-01 -6.88354373e-01 3.43003958e-01
-1.15215860e-01 4.79750782e-01 3.27085942e-01 5.54306746e-01
4.67861384e-01 2.01450825e+00 1.65295646e-01 -4.93250430e-01
4.56266463e-01 -5.97032428e-01 -1.67170286e-01 2.09652126e-01
2.79782802e-01 -2.53887828e-02 4.92898077e-01 -7.01669991e-01
-8.60251904e-01 -5.83484232e-01 -1.54681027e+00 -7.08217621e-01
-2.80782729e-01 6.23689950e-01 1.36474407e+00 8.76869082e-01
7.47087121e-01 6.41308546e-01 8.70157182e-01 -5.92142045e-01
-2.76727825e-01 -6.65207982e-01 -6.61797404e-01 -2.48023257e-01
8.69814694e-01 -8.09865952e-01 -3.39681596e-01 -1.79943636e-01] | [5.716083526611328, 1.5058810710906982] |
c3081861-ec50-4e28-863b-b66d47214528 | gradient-informed-quality-diversity-for-the | 2306.05138 | null | https://arxiv.org/abs/2306.05138v1 | https://arxiv.org/pdf/2306.05138v1.pdf | Gradient-Informed Quality Diversity for the Illumination of Discrete Spaces | Quality Diversity (QD) algorithms have been proposed to search for a large collection of both diverse and high-performing solutions instead of a single set of local optima. While early QD algorithms view the objective and descriptor functions as black-box functions, novel tools have been introduced to use gradient information to accelerate the search and improve overall performance of those algorithms over continuous input spaces. However a broad range of applications involve discrete spaces, such as drug discovery or image generation. Exploring those spaces is challenging as they are combinatorially large and gradients cannot be used in the same manner as in continuous spaces. We introduce map-elites with a Gradient-Informed Discrete Emitter (ME-GIDE), which extends QD optimisation with differentiable functions over discrete search spaces. ME-GIDE leverages the gradient information of the objective and descriptor functions with respect to its discrete inputs to propose gradient-informed updates that guide the search towards a diverse set of high quality solutions. We evaluate our method on challenging benchmarks including protein design and discrete latent space illumination and find that our method outperforms state-of-the-art QD algorithms in all benchmarks. | ['Antoine Cully', 'Thomas Pierrot', 'Jérémie Dona', 'Guillaume Richard', 'Raphael Boige'] | 2023-06-08 | null | null | null | null | ['drug-discovery', 'protein-design'] | ['medical', 'medical'] | [ 2.98866838e-01 -3.88865888e-01 -2.07117841e-01 -2.53691733e-01
-1.34749591e+00 -8.95011127e-01 4.58126426e-01 1.12698726e-01
-2.97298789e-01 1.06566751e+00 2.98006713e-01 -1.10296853e-01
-4.62389380e-01 -4.64080751e-01 -8.01961482e-01 -9.74253058e-01
6.07799888e-02 6.60054684e-01 -2.19531171e-03 -1.70893863e-01
4.68475908e-01 4.82094496e-01 -1.39393556e+00 2.03568023e-02
1.10607433e+00 8.25618625e-01 2.31556296e-01 4.66835797e-01
1.21524977e-02 -1.87656358e-01 -6.01564884e-01 -4.20902520e-02
4.23727304e-01 -6.32168770e-01 -6.96252525e-01 -5.20724542e-02
3.25542152e-01 2.55557209e-01 2.65192568e-01 1.01560307e+00
1.09149742e+00 3.72364104e-01 4.40033227e-01 -1.07180119e+00
-7.65441775e-01 -1.50770381e-01 -4.79433775e-01 -6.81893975e-02
3.73579741e-01 5.50361931e-01 1.28877807e+00 -1.13069868e+00
9.73655462e-01 1.30464518e+00 4.83066618e-01 2.45135576e-01
-1.93214929e+00 -2.82748461e-01 2.38393992e-01 1.03051234e-02
-1.35683346e+00 -3.52096349e-01 6.04357719e-01 -2.14529842e-01
1.21746194e+00 4.42259848e-01 6.36171341e-01 8.51729512e-01
2.89084077e-01 7.51193166e-01 9.69116986e-01 -1.87720314e-01
7.17158854e-01 -1.05677925e-01 -9.87331495e-02 7.06422508e-01
5.25078326e-02 1.14863537e-01 -7.75393724e-01 -7.60841131e-01
4.15488124e-01 -6.31056502e-02 -3.55163008e-01 -1.02979326e+00
-1.34515953e+00 1.10420656e+00 3.72645974e-01 -2.14959994e-01
-6.81477606e-01 -1.00997388e-01 2.06474334e-01 1.72581360e-01
2.70187676e-01 1.34626329e+00 -7.02017844e-01 -4.36649531e-01
-7.87242413e-01 7.36214519e-01 7.94737160e-01 4.91132110e-01
7.76737034e-01 -4.02410746e-01 -5.20002425e-01 9.36206877e-01
6.73493892e-02 3.00469667e-01 5.54933548e-02 -1.15497327e+00
4.15385485e-01 7.84437537e-01 4.68342781e-01 -9.30925965e-01
-4.11275327e-01 -7.77331293e-01 -4.36440587e-01 3.82547736e-01
3.42870116e-01 -1.74914554e-01 -9.77185071e-01 1.81031549e+00
7.75034130e-01 -3.46117973e-01 -5.98367639e-02 1.07116795e+00
3.38405520e-01 8.87475371e-01 -1.76797390e-01 -3.35266054e-01
9.12007093e-01 -9.20525372e-01 -5.10438919e-01 -2.29001433e-01
5.74390411e-01 -7.17039347e-01 1.19903886e+00 5.49583137e-01
-1.33987522e+00 -1.62958086e-01 -1.01335657e+00 -2.27688909e-01
-1.96551368e-01 -5.22813760e-02 5.57583809e-01 3.08918029e-01
-1.07995749e+00 8.89618993e-01 -6.75388455e-01 -4.04740013e-02
5.79640329e-01 7.28106201e-01 -7.66723529e-02 -6.14848547e-02
-8.52424324e-01 8.30731809e-01 2.32186809e-01 -2.72528958e-02
-7.94712663e-01 -9.34947073e-01 -5.15391231e-01 -2.60541558e-01
6.51646614e-01 -8.44629467e-01 8.13487470e-01 -6.44307852e-01
-1.69259262e+00 4.68793005e-01 -2.40853846e-01 -2.28910390e-02
4.48855758e-01 -2.09635243e-01 1.29086599e-01 -2.08838359e-01
1.27286896e-01 8.62458408e-01 6.83798432e-01 -8.85360360e-01
-4.34909433e-01 -2.58173198e-01 -6.89574331e-02 5.68217337e-01
-1.58943176e-01 -2.02896699e-01 -3.59784126e-01 -3.82474869e-01
2.33135685e-01 -1.05657673e+00 -4.07270342e-01 3.90569121e-01
-4.74371284e-01 -2.53583699e-01 6.62025273e-01 -3.51484537e-01
1.33853948e+00 -1.73865259e+00 9.43252921e-01 4.02242362e-01
2.64579475e-01 1.52042285e-01 -2.57944584e-01 5.53597927e-01
1.50585204e-01 -2.77250241e-02 -3.37909281e-01 -2.55173475e-01
2.52320319e-01 3.24473202e-01 -6.19392879e-02 5.94233930e-01
5.33243060e-01 9.94770110e-01 -1.18104267e+00 -1.43026292e-01
-1.60637394e-01 4.80856806e-01 -9.80903983e-01 1.06658190e-02
-5.80343962e-01 5.02263904e-01 -6.29226327e-01 6.18690789e-01
5.63466132e-01 -6.34820819e-01 8.68140161e-02 -1.48713678e-01
-3.69439423e-01 3.68576735e-01 -1.29021001e+00 2.01925921e+00
1.65143125e-02 1.19229533e-01 -1.06201701e-01 -8.00979435e-01
9.89105999e-01 -1.45257637e-01 7.19452441e-01 -6.04327798e-01
-2.06507891e-01 3.69773835e-01 -5.63853458e-02 4.14461503e-03
4.15783137e-01 -4.10778560e-02 7.41626918e-02 4.72153664e-01
-2.13583589e-01 -1.37099147e-01 3.03143442e-01 -1.38831362e-01
1.25915360e+00 5.29453695e-01 8.95813704e-02 -3.06743979e-01
2.68232852e-01 1.86157942e-01 8.34386528e-01 6.21482790e-01
1.02377690e-01 6.70222163e-01 7.22257376e-01 -5.36247611e-01
-1.27769351e+00 -1.09369254e+00 -2.38423496e-01 1.17622447e+00
1.69240817e-01 -4.41033661e-01 -4.13190782e-01 -5.94282210e-01
1.04613811e-01 4.33035284e-01 -4.49268818e-01 -3.32840741e-01
-5.21274030e-01 -1.18032253e+00 6.29595071e-02 1.32825091e-01
-6.90958463e-03 -8.91415834e-01 -8.07983398e-01 6.27279282e-01
1.45154476e-01 -6.17425740e-01 -8.52136791e-01 5.89877009e-01
-7.30756283e-01 -7.46705890e-01 -8.52822959e-01 -6.49949968e-01
6.14955902e-01 -1.32494807e-01 1.34461117e+00 -2.22399697e-01
-6.54072821e-01 -5.39908186e-02 -2.02031657e-02 -1.40976310e-01
-1.68644145e-01 5.28285056e-02 -3.32123525e-02 -1.00019276e-01
1.01910092e-01 -3.54108483e-01 -9.46403205e-01 5.70344210e-01
-7.23423719e-01 -2.99907252e-02 4.40146416e-01 1.29120147e+00
1.14256060e+00 -3.12503546e-01 5.60557663e-01 -3.73749018e-01
9.21635747e-01 -5.05231142e-01 -8.68678629e-01 3.63702118e-01
-8.20316017e-01 8.47604990e-01 4.46392328e-01 -6.19388819e-01
-4.79102373e-01 2.38483995e-01 -1.22649614e-02 -3.85500938e-01
5.01500726e-01 4.37753260e-01 -9.64486599e-02 -1.72237724e-01
8.20784509e-01 1.78103000e-01 1.06914081e-01 -4.86869305e-01
4.40596998e-01 1.76196352e-01 3.06766599e-01 -1.01408267e+00
3.59831870e-01 2.10777134e-01 2.76730269e-01 -4.09508675e-01
-6.53655052e-01 -5.78894734e-01 -1.51141137e-01 2.79827505e-01
5.79475760e-01 -5.22575557e-01 -7.70767748e-01 1.79078013e-01
-9.15035963e-01 -2.08790123e-01 -5.60917974e-01 2.13047743e-01
-6.61210418e-01 1.96751654e-01 1.58428531e-02 -4.89383221e-01
-3.93399358e-01 -1.79435468e+00 1.55155361e+00 3.78251486e-02
-5.62312186e-01 -8.99962723e-01 4.29242611e-01 -6.17453502e-03
3.81661683e-01 5.74730635e-01 9.63107646e-01 -3.69180262e-01
-7.52792060e-01 5.75426482e-02 8.96773189e-02 -2.66918610e-03
1.69032902e-01 -3.37575749e-02 -5.36962032e-01 -5.56894243e-01
-3.64372671e-01 -4.84455705e-01 5.62674165e-01 6.18335664e-01
9.37859237e-01 -2.92670161e-01 -4.97243136e-01 1.11673856e+00
1.29001844e+00 2.01346144e-01 3.76061618e-01 5.67317367e-01
3.06136996e-01 1.79939806e-01 6.84814990e-01 5.98682404e-01
8.53495076e-02 1.14859295e+00 4.24340248e-01 -2.76975632e-01
2.24890828e-01 -9.49259177e-02 1.94616288e-01 1.16834782e-01
2.74320573e-01 -2.82426387e-01 -8.18331301e-01 4.13932234e-01
-1.99496329e+00 -7.92009950e-01 2.79516608e-01 2.24047184e+00
1.43338573e+00 5.84663935e-02 4.42667603e-01 -2.23402217e-01
5.40009022e-01 -4.13038060e-02 -1.33212125e+00 -4.72851992e-01
-2.60164738e-01 3.60567063e-01 4.18987751e-01 3.88321251e-01
-8.58344376e-01 6.12802863e-01 6.50516176e+00 7.98104525e-01
-1.20680285e+00 -1.93637639e-01 7.91545391e-01 -5.92570126e-01
-5.86361110e-01 -4.57277410e-02 -9.39267874e-01 4.12245274e-01
6.62339211e-01 -1.82217911e-01 6.97944880e-01 6.48429513e-01
2.47906819e-01 8.06622058e-02 -1.09441757e+00 1.05435240e+00
-4.62187439e-01 -1.65878689e+00 -2.60687858e-01 3.66904020e-01
1.27837932e+00 5.18544540e-02 4.58645493e-01 -5.92046529e-02
4.61987585e-01 -1.15269995e+00 5.95388710e-01 5.03129482e-01
6.89772367e-01 -8.40275466e-01 2.11776048e-01 1.83972955e-01
-7.60377884e-01 -1.11507379e-01 -3.90631109e-01 3.41692418e-01
8.83000940e-02 4.64747488e-01 -7.52582312e-01 1.83075920e-01
4.83113617e-01 5.84025204e-01 -2.78816909e-01 1.24453056e+00
1.34117976e-01 7.68162459e-02 -5.89512825e-01 -3.28694522e-01
6.48515463e-01 -5.54540873e-01 8.19129884e-01 7.69200981e-01
1.74826458e-01 -9.61811915e-02 2.86903650e-01 1.28844941e+00
3.17398310e-02 1.14772292e-02 -2.14619502e-01 -3.96818787e-01
3.92622709e-01 9.22776282e-01 -5.87528944e-01 7.69191757e-02
6.29861131e-02 1.02473164e+00 3.03279012e-01 4.58092719e-01
-7.80271649e-01 -3.67143452e-01 1.12736738e+00 -6.66217431e-02
4.11185980e-01 -2.81472266e-01 -3.18603486e-01 -8.30786526e-01
1.01734877e-01 -1.22388637e+00 2.76720047e-01 -4.66676027e-01
-1.20755732e+00 3.76641780e-01 -2.02217504e-01 -9.65392649e-01
-1.74207032e-01 -4.84752297e-01 -2.70105481e-01 1.13637519e+00
-1.36449873e+00 -5.75705171e-01 1.40680134e-01 2.44073838e-01
5.31061172e-01 -1.11026034e-01 7.18042135e-01 1.06765732e-01
-3.31133455e-01 5.35481215e-01 9.30704355e-01 -7.51027465e-01
6.95228279e-01 -1.32540071e+00 4.78016168e-01 3.64161789e-01
-8.27006325e-02 7.51632214e-01 8.42885494e-01 -6.65154934e-01
-1.88194335e+00 -6.99715495e-01 6.12897992e-01 -5.80402493e-01
4.39646542e-01 -2.58969456e-01 -7.57315159e-01 -1.37014300e-01
-2.51846492e-01 1.41401067e-01 3.11673105e-01 1.42091826e-01
2.37551834e-02 1.92716916e-03 -1.05567837e+00 6.35989368e-01
1.37022936e+00 -2.82850921e-01 -2.84824166e-02 6.44792497e-01
6.47920728e-01 -6.86014354e-01 -9.08057272e-01 3.40667963e-01
4.80461389e-01 -6.40015066e-01 1.39361382e+00 -8.50509822e-01
9.42738205e-02 -5.60732841e-01 -1.77581713e-01 -1.46917892e+00
-3.46711099e-01 -1.18603027e+00 -1.64025560e-01 6.04004979e-01
5.28016329e-01 -7.21878409e-01 6.69934988e-01 6.26073837e-01
-2.39949375e-01 -1.49931097e+00 -8.96378040e-01 -7.28558421e-01
1.25832134e-03 2.38374174e-01 8.80003810e-01 4.89632308e-01
-3.01933885e-01 3.84500355e-01 -1.91155568e-01 -7.02053085e-02
6.52677894e-01 4.34732705e-01 7.08109021e-01 -1.00369978e+00
-5.79344511e-01 -5.97518802e-01 -2.21927479e-01 -1.28625953e+00
-2.87207901e-01 -8.03021371e-01 1.54883713e-01 -1.58799398e+00
1.66017860e-01 -6.72390521e-01 -1.88180164e-01 4.33950275e-01
-4.08215284e-01 -3.89354974e-02 -6.27925694e-02 1.46828204e-01
-8.53294790e-01 9.21911359e-01 1.46845484e+00 -1.12091288e-01
-6.17687762e-01 -7.24932551e-02 -8.12477052e-01 1.63530156e-01
4.33625698e-01 -5.61927080e-01 -4.68538612e-01 -1.94506690e-01
6.75460398e-01 -7.19054639e-02 2.35270053e-01 -6.92263007e-01
3.70947011e-02 -4.59308892e-01 4.04363394e-01 -5.84205031e-01
5.45691967e-01 -3.94634515e-01 4.01113659e-01 3.42484325e-01
-4.62958276e-01 1.79107890e-01 1.84574146e-02 5.92418373e-01
1.02350302e-01 6.91772550e-02 8.12903225e-01 -1.74944140e-02
-2.80866235e-01 3.56821895e-01 2.82796621e-01 3.85735989e-01
9.31569815e-01 -4.71173525e-01 -1.62072197e-01 -1.59756899e-01
-6.50811911e-01 4.70269024e-01 7.70202219e-01 3.09507161e-01
6.01520181e-01 -1.22845781e+00 -6.13141000e-01 1.76939636e-01
-2.66865660e-02 2.55578458e-01 -3.40003639e-01 6.94315970e-01
-3.33122939e-01 3.94343287e-01 -1.18121102e-01 -9.10343707e-01
-1.08338761e+00 3.03175837e-01 4.05752689e-01 -3.67927253e-01
-4.22881097e-01 1.06617558e+00 1.25304759e-01 -5.35435319e-01
2.17633888e-01 -1.85547963e-01 4.02239978e-01 -4.47403379e-02
1.84070364e-01 4.64478135e-01 1.19913973e-01 -2.01774031e-01
-4.89912808e-01 6.42452121e-01 2.88637057e-02 -6.93889409e-02
1.79813039e+00 5.41016646e-02 1.23482216e-02 1.44440100e-01
1.56393695e+00 -3.36562097e-01 -1.86457586e+00 -2.57815957e-01
1.64835244e-01 -6.72107399e-01 1.30358458e-01 -1.07862055e+00
-4.52018291e-01 4.63117778e-01 6.68890953e-01 -1.62918061e-01
1.14835906e+00 -1.50303375e-02 8.03628445e-01 4.65221077e-01
2.82970607e-01 -1.30076277e+00 4.66503441e-01 3.70359629e-01
1.02169073e+00 -1.13364887e+00 3.34513754e-01 1.75329417e-01
-7.44430363e-01 9.43116248e-01 2.55535781e-01 1.16767429e-01
2.09172770e-01 5.91673255e-02 -2.84093231e-01 -4.27628696e-01
-9.60017025e-01 -6.45837113e-02 6.33578777e-01 2.91905880e-01
3.90069187e-01 -2.93203861e-01 -3.33657086e-01 -1.28033549e-01
3.78784426e-02 -1.38358980e-01 -1.65416017e-01 1.23150933e+00
-4.96246070e-01 -1.52533472e+00 -4.42933559e-01 4.39560741e-01
-2.20736504e-01 -8.25222433e-02 -3.24276119e-01 3.54997307e-01
2.59449184e-02 7.19037592e-01 -3.18609565e-01 1.01809882e-01
3.45204741e-01 1.50847360e-01 4.47359473e-01 -5.06350756e-01
-5.71960568e-01 2.25095019e-01 5.16600646e-02 -8.45976293e-01
-7.14922771e-02 -8.87733817e-01 -1.29919267e+00 -8.33046660e-02
-3.66344750e-01 3.40299875e-01 8.39545012e-01 7.66163349e-01
7.92304099e-01 3.33469033e-01 6.89138234e-01 -7.24432945e-01
-1.03798068e+00 -1.57564491e-01 -8.59611481e-02 2.74483263e-01
6.11682951e-01 -7.37391114e-01 1.55414026e-02 -3.02158684e-01] | [4.934123992919922, 5.43230676651001] |
e6b4d683-9304-41ed-b200-b996d740f5cc | cardiac-arrhythmia-detection-from-ecg | 1801.10033 | null | http://arxiv.org/abs/1801.10033v1 | http://arxiv.org/pdf/1801.10033v1.pdf | Cardiac Arrhythmia Detection from ECG Combining Convolutional and Long Short-Term Memory Networks | Objectives: Atrial fibrillation (AF) is a common heart rhythm disorder
associated with deadly and debilitating consequences including heart failure,
stroke, poor mental health, reduced quality of life and death. Having an
automatic system that diagnoses various types of cardiac arrhythmias would
assist cardiologists to initiate appropriate preventive measures and to improve
the analysis of cardiac disease. To this end, this paper introduces a new
approach to detect and classify automatically cardiac arrhythmias in
electrocardiograms (ECG) recordings.
Methods: The proposed approach used a combination of Convolution Neural
Networks (CNNs) and a sequence of Long Short-Term Memory (LSTM) units, with
pooling, dropout and normalization techniques to improve their accuracy. The
network predicted a classification at every 18th input sample and we selected
the final prediction for classification. Results were cross-validated on the
Physionet Challenge 2017 training dataset, which contains 8,528 single lead ECG
recordings lasting from 9s to just over 60s.
Results: Using the proposed structure and no explicit feature selection,
10-fold stratified cross-validation gave an overall F-measure of 0.83.10-0.015
on the held-out test data (mean-standard deviation over all folds) and 0.80 on
the hidden dataset of the Challenge entry server. | ['Masun Nabhan Homsi', 'Philip Warrick'] | 2018-01-30 | null | null | null | null | ['arrhythmia-detection'] | ['medical'] | [ 4.08083975e-01 -1.36710539e-01 3.17825302e-02 -4.43816841e-01
-5.07109463e-01 -3.98386717e-01 -1.26432374e-01 3.75032991e-01
-5.87291479e-01 1.22500038e+00 -1.49691552e-01 -5.34300566e-01
-3.56406093e-01 -5.59513688e-01 -3.29671472e-01 -6.34332657e-01
-6.08342826e-01 2.63076782e-01 -3.56133491e-01 3.70359004e-01
9.77270156e-02 7.41074383e-01 -1.13056707e+00 6.42129719e-01
8.48134875e-01 1.25372267e+00 -4.06690359e-01 1.12306285e+00
3.06565434e-01 7.49656558e-01 -7.97080755e-01 -1.53326005e-01
-4.69615534e-02 -4.76401508e-01 -8.36984038e-01 -3.51988465e-01
3.73483934e-02 3.55911329e-02 8.46318603e-02 4.60397065e-01
1.21097612e+00 -2.91898489e-01 3.75509888e-01 -8.70828807e-01
-1.22320302e-01 4.19682264e-01 -3.43811549e-02 7.42834330e-01
-1.28918067e-01 1.35939166e-01 4.27552849e-01 -9.26712513e-01
2.46494681e-01 5.18526435e-01 1.26307309e+00 6.17991984e-01
-1.18659532e+00 -7.43064880e-01 -5.32306314e-01 2.86075413e-01
-1.39155567e+00 -4.37212259e-01 3.99770230e-01 -5.25207162e-01
1.21247840e+00 2.53825337e-01 1.01953220e+00 1.05316401e+00
7.72204161e-01 1.94160759e-01 8.84673059e-01 -2.46419862e-01
1.13566250e-01 1.73565134e-01 4.58053768e-01 5.84482372e-01
3.87468755e-01 -7.76614621e-02 -5.13019025e-01 -6.35619223e-01
5.50526798e-01 1.53217331e-01 -3.90198678e-01 1.92216977e-01
-1.37278867e+00 5.87354660e-01 2.22432259e-02 5.06373703e-01
-8.00645471e-01 -6.32214099e-02 7.09183156e-01 5.91837406e-01
4.15492505e-01 5.49423456e-01 -1.07064509e+00 -3.50339293e-01
-8.52589130e-01 -3.53005566e-02 7.01604784e-01 1.73481345e-01
2.64113009e-01 1.89077228e-01 -3.49547684e-01 7.21704543e-01
-7.13816583e-02 4.10467327e-01 8.69680047e-01 -4.71483320e-01
3.86565179e-01 6.39397919e-01 -1.12034865e-01 -7.68719673e-01
-8.34958375e-01 -1.16659749e+00 -1.56007099e+00 1.28669441e-01
3.05728853e-01 -6.60023749e-01 -9.27854896e-01 1.49698043e+00
-3.00656289e-01 2.49886900e-01 2.29125217e-01 5.66219509e-01
8.25296283e-01 3.75302196e-01 1.97137833e-01 -4.49864417e-01
1.31426656e+00 -2.90005624e-01 -6.92163825e-01 9.14632827e-02
7.05174208e-01 -3.91418129e-01 7.90232003e-01 7.57585406e-01
-9.57881093e-01 -6.01118684e-01 -1.06156754e+00 4.09851283e-01
-3.10788214e-01 5.60363054e-01 2.25903422e-01 9.26724672e-01
-1.03205001e+00 1.03210187e+00 -7.13621855e-01 -4.00865763e-01
9.75713253e-01 6.87092066e-01 -2.76791215e-01 3.44617009e-01
-1.27942920e+00 8.30053091e-01 3.45006347e-01 4.43361938e-01
-5.71539044e-01 -8.59849274e-01 -2.67355919e-01 -1.77704673e-02
-2.79536039e-01 -9.51559722e-01 7.04234064e-01 -1.13833892e+00
-1.21267927e+00 8.86969447e-01 -2.02154949e-01 -9.34857488e-01
4.06463802e-01 -4.31101233e-01 -5.05689323e-01 -6.16709962e-02
-2.02960238e-01 1.81616977e-01 7.24948168e-01 -3.41530412e-01
-5.05314887e-01 -6.22507691e-01 -6.02285385e-01 4.32648212e-02
-3.90821338e-01 7.10286722e-02 4.56519723e-01 -6.83812797e-01
-6.10744208e-02 -6.56595409e-01 -3.05691570e-01 -4.36349094e-01
-3.38050157e-01 1.06547363e-01 5.84992409e-01 -1.05015814e+00
1.34941697e+00 -2.10035849e+00 7.26250336e-02 2.37427190e-01
3.57564926e-01 6.46476328e-01 2.69791573e-01 1.76934257e-01
-4.14190531e-01 3.75850976e-01 -3.62207860e-01 -2.57964674e-02
-6.74905121e-01 -8.33819956e-02 -3.50094074e-03 4.57913309e-01
2.41592452e-01 1.04360271e+00 -3.39623958e-01 -1.24829695e-01
9.68157873e-02 8.23730290e-01 5.77585362e-02 1.03623234e-01
3.57866257e-01 6.38838887e-01 -1.90078154e-01 5.20755351e-01
3.16548973e-01 -2.63840735e-01 -3.04994900e-02 -1.20798782e-01
1.89790860e-01 1.91193357e-01 -8.31249833e-01 1.62141669e+00
-2.08998665e-01 6.42536998e-01 -4.84723151e-01 -1.11133051e+00
1.10350955e+00 8.77466679e-01 4.50399280e-01 -4.03649747e-01
3.66611481e-01 3.56110334e-01 3.91532481e-01 -6.32902503e-01
-5.40537119e-01 -2.26165652e-01 2.91768044e-01 4.07421708e-01
1.93789061e-02 6.77549362e-01 -1.71781257e-01 -2.14152023e-01
1.22580206e+00 -3.26416492e-01 4.33405846e-01 -3.51672530e-01
6.11840129e-01 -2.79265404e-01 8.09963465e-01 8.41373920e-01
-2.60304064e-01 6.14283800e-01 4.12449807e-01 -1.37884283e+00
-7.27381408e-01 -7.71108568e-01 -2.49526978e-01 3.41364503e-01
-7.92069912e-01 -2.29327947e-01 -6.39268458e-01 -6.10225976e-01
-2.21589476e-01 3.59024465e-01 -6.13139808e-01 -4.21621978e-01
-5.88057935e-01 -1.21856809e+00 1.20840538e+00 5.79490185e-01
6.03787839e-01 -1.57455587e+00 -1.29915619e+00 4.67590123e-01
-2.02221513e-01 -5.93691170e-01 8.44161361e-02 5.46102881e-01
-1.43524277e+00 -1.14690316e+00 -9.13109541e-01 -6.44411027e-01
2.02645808e-01 -7.86674500e-01 1.13104212e+00 9.61718336e-02
-7.23570049e-01 -1.74529999e-01 3.35976891e-02 -7.07809150e-01
-7.56212175e-02 3.65776837e-01 1.23433843e-01 1.75700068e-01
5.43971062e-01 -8.12100947e-01 -7.20189810e-01 -2.42117718e-01
-2.60161281e-01 -1.18643180e-01 8.27776074e-01 9.42359865e-01
3.70932013e-01 -2.72816032e-01 1.20791519e+00 -8.43974710e-01
8.23704660e-01 -3.07343900e-01 -2.07779139e-01 1.18389286e-01
-1.06686366e+00 -3.17980528e-01 7.50758290e-01 -2.67449021e-01
-4.54222977e-01 1.52739897e-01 -9.04100388e-02 -3.14735025e-01
-3.85209650e-01 4.86125201e-01 -9.77450311e-02 3.16201001e-01
8.33275497e-01 3.83529305e-01 -1.88409556e-02 -3.24888706e-01
-4.04783487e-01 7.94120610e-01 4.01351333e-01 -4.20492776e-02
1.22594066e-01 5.79297207e-02 1.01548702e-01 -9.19938087e-01
-5.88442981e-01 -9.43147764e-02 -7.74483204e-01 -1.09025734e-02
9.09661710e-01 -8.33880424e-01 -7.46837974e-01 7.66496480e-01
-1.21908045e+00 -1.78502724e-01 -2.24882528e-01 6.04313433e-01
-1.96371585e-01 2.69613955e-02 -6.99398041e-01 -8.15570295e-01
-1.29301608e+00 -5.67424655e-01 3.75928849e-01 2.17616055e-02
-5.91636121e-01 -8.05931091e-01 1.45130724e-01 2.96579320e-02
7.37015426e-01 5.96513391e-01 9.38152611e-01 -7.80498147e-01
1.79974347e-01 -3.77431184e-01 3.06351157e-03 7.97010779e-01
3.79135311e-01 -1.83655903e-01 -1.27291858e+00 -2.36279979e-01
9.72654521e-02 -9.97827649e-02 8.25693667e-01 7.72376895e-01
1.20803094e+00 -1.83091953e-01 -1.10742316e-01 6.54333889e-01
1.20806396e+00 7.04172194e-01 9.14614856e-01 1.79008797e-01
5.57281971e-01 -2.35177130e-02 -3.48140462e-03 4.63874131e-01
-1.08451404e-01 9.27015916e-02 -2.54723970e-02 -4.77770776e-01
1.47684097e-01 5.29496491e-01 4.92502712e-02 4.20867473e-01
-3.70348066e-01 9.69638582e-03 -1.11805618e+00 5.23411512e-01
-1.60642147e+00 -9.76920187e-01 -4.67195958e-01 2.48104787e+00
8.14747095e-01 2.74208337e-01 5.22773154e-02 7.33004928e-01
6.67690098e-01 -4.59208429e-01 -5.92762053e-01 -7.94516563e-01
-3.29088002e-01 9.01377618e-01 2.97200054e-01 2.01713800e-01
-1.21557987e+00 3.53160530e-01 5.97330618e+00 1.57603309e-01
-1.55561936e+00 2.95545273e-02 1.17371011e+00 -3.88812050e-02
5.46228230e-01 -4.45987344e-01 -3.23905915e-01 4.74366963e-01
1.54868960e+00 -5.22040017e-02 4.80782576e-02 3.50434512e-01
4.96370286e-01 1.76795408e-01 -8.71086359e-01 1.12201214e+00
6.21320941e-02 -1.40191793e+00 -1.63787633e-01 -2.28575200e-01
3.54182631e-01 1.26366407e-01 -1.69125423e-01 7.11490437e-02
-6.11475468e-01 -1.50885630e+00 1.56119183e-01 9.51274991e-01
1.15218782e+00 -1.01052821e+00 1.23114717e+00 2.54540265e-01
-8.31181228e-01 -2.92527765e-01 5.00043109e-02 -4.00919348e-01
-1.24881528e-01 1.11596608e+00 -8.78809869e-01 4.01932716e-01
1.03059828e+00 8.33550215e-01 -5.68721473e-01 1.05479622e+00
1.85253590e-01 1.09862888e+00 -1.65810317e-01 8.88182819e-02
-1.90327376e-01 1.71861321e-01 3.77404273e-01 1.16891599e+00
3.39166641e-01 8.87415186e-02 -2.09264934e-01 4.48116601e-01
1.53334355e-02 2.21192867e-01 -4.84532028e-01 5.48445955e-02
4.85572033e-02 1.11412394e+00 -7.02280581e-01 -6.54188931e-01
-4.10770718e-03 8.59554827e-01 -1.13102570e-01 3.35351437e-01
-6.69900000e-01 -8.61371636e-01 1.25822127e-01 3.12212080e-01
-8.33138674e-02 4.25805926e-01 -8.48119080e-01 -1.06187260e+00
2.38207459e-01 -1.00587416e+00 5.73445261e-01 -2.61607736e-01
-9.83670890e-01 9.40225542e-01 -4.88704860e-01 -1.03689384e+00
-1.22089379e-01 -3.38093936e-01 -7.03999281e-01 1.33129299e+00
-1.15028441e+00 -7.78010726e-01 -2.51851290e-01 4.19061005e-01
2.39822820e-01 -4.86423194e-01 1.51283586e+00 6.74135864e-01
-6.25628591e-01 4.07992959e-01 -3.47320139e-01 2.56189615e-01
7.24694371e-01 -1.17381334e+00 1.90214977e-01 7.46892035e-01
-1.48760363e-01 6.07763708e-01 3.98219466e-01 -4.62469161e-01
-6.90063775e-01 -1.38569474e+00 1.47728837e+00 -2.12836608e-01
-3.89999710e-02 -1.28338588e-02 -8.74819934e-01 4.11584288e-01
1.24117866e-01 2.71201879e-01 9.51709151e-01 3.72436754e-02
2.01062471e-01 -3.92372102e-01 -1.21272969e+00 -1.86333880e-02
5.54418445e-01 -2.94716090e-01 -4.65873778e-01 3.69317345e-02
-1.07631452e-01 -1.11150004e-01 -9.62195277e-01 7.60368884e-01
9.75198328e-01 -1.07061136e+00 7.83882141e-01 -9.01574314e-01
2.73563951e-01 -2.71987952e-02 3.76609057e-01 -1.13059235e+00
-2.64495283e-01 -6.76883638e-01 -1.72568247e-01 6.92617238e-01
7.49393523e-01 -8.39178979e-01 8.90721679e-01 1.82679161e-01
-1.33825734e-01 -1.32438052e+00 -8.40552866e-01 -4.15977150e-01
1.90268699e-02 -3.80399436e-01 2.01909155e-01 8.33059490e-01
-2.79013157e-01 5.23587286e-01 -5.21948695e-01 6.08922727e-03
4.93820250e-01 -2.01968163e-01 2.71958262e-01 -1.71709967e+00
-1.95387796e-01 -8.16664994e-02 -5.77307880e-01 8.79600495e-02
-2.89281666e-01 -9.29029226e-01 -4.79355842e-01 -1.45890605e+00
1.10962562e-01 -3.16628069e-01 -9.67724681e-01 8.10229897e-01
-2.30431892e-02 4.84330535e-01 -3.72901410e-01 -1.59406126e-01
-8.47271159e-02 7.37652034e-02 5.72061241e-01 -2.02662945e-02
-5.42219937e-01 5.13871014e-01 -6.04639649e-01 6.59897029e-01
1.42277181e+00 -7.10859835e-01 -4.15512741e-01 -1.27794489e-01
1.80878118e-01 2.19367817e-01 4.62756395e-01 -1.43078196e+00
-6.53433204e-02 3.53973627e-01 1.15809572e+00 -6.33462012e-01
1.91425875e-01 -4.44203019e-01 4.32426780e-01 1.28392851e+00
-3.92446160e-01 2.94065326e-01 3.12661588e-01 2.49001518e-01
-1.26989871e-01 1.50736019e-01 6.49083376e-01 -6.09659664e-02
3.92687805e-02 3.33640277e-01 -7.25351572e-01 -7.07402304e-02
9.75427508e-01 -3.41643929e-01 2.46971488e-01 -5.62394150e-02
-1.27899432e+00 -1.31152466e-01 -3.24800402e-01 2.35568866e-01
8.05921733e-01 -1.06080580e+00 -1.04576838e+00 5.06448805e-01
-2.91123837e-02 -1.78149983e-01 2.10710019e-01 1.20759952e+00
-5.60433805e-01 3.98580074e-01 -4.08278793e-01 -6.47720098e-01
-1.70605636e+00 -1.55471027e-01 8.09700429e-01 -3.04077387e-01
-9.17669773e-01 8.04212213e-01 -7.14262545e-01 -1.05126090e-01
4.64774340e-01 -5.72308183e-01 -5.18494248e-01 1.76310286e-01
6.16688669e-01 4.53537047e-01 6.53946698e-01 -2.49408305e-01
-6.59146011e-01 3.55071485e-01 5.84138818e-02 7.53582940e-02
1.44062829e+00 3.37760806e-01 -2.33830258e-01 8.69804800e-01
1.05226386e+00 -4.64202762e-01 -4.64878649e-01 2.38995314e-01
4.31833602e-02 7.31853545e-02 -1.13107311e-02 -1.30321074e+00
-1.15346420e+00 1.17503238e+00 1.40603650e+00 1.45647377e-01
1.23864460e+00 -6.30213022e-01 7.44827688e-01 6.86389983e-01
-6.53181970e-02 -7.64076233e-01 -4.29192841e-01 2.29233041e-01
6.02062881e-01 -8.02821815e-01 -2.50017911e-01 1.83291554e-01
-4.52727109e-01 1.52822745e+00 8.86552036e-02 -1.18784949e-01
1.00182784e+00 1.69930443e-01 3.79324049e-01 -2.37014130e-01
-8.78785312e-01 4.98297274e-01 1.08735725e-01 4.78370309e-01
8.08458328e-01 8.81855637e-02 -5.20714641e-01 9.85637546e-01
1.67156994e-01 6.97078049e-01 3.29568803e-01 1.02260804e+00
-2.36581609e-01 -1.04720378e+00 -1.20339490e-01 1.08924389e+00
-1.08933496e+00 -2.24167362e-01 -3.79595786e-01 2.06075415e-01
4.93008435e-01 8.57845187e-01 -2.03776926e-01 -2.32471421e-01
2.05622584e-01 7.29762256e-01 2.87108600e-01 -4.39430445e-01
-1.04573941e+00 -1.92925319e-01 1.07424311e-01 -2.91902304e-01
-4.26297665e-01 -5.43932974e-01 -1.20832431e+00 1.37204945e-01
-9.82668996e-02 1.43253610e-01 5.48536420e-01 9.00752366e-01
7.78941572e-01 9.47739422e-01 4.61905599e-01 -2.67173260e-01
-3.47398907e-01 -1.36897993e+00 -5.37263632e-01 1.48368225e-01
5.26404619e-01 -1.01502657e-01 -3.34634274e-01 4.30772573e-01] | [14.304343223571777, 3.292480945587158] |
0ecd3d50-4f64-4c36-aaa6-b50e434c4c14 | generalization-algorithm-of-multimodal-pre | 2302.10315 | null | https://arxiv.org/abs/2302.10315v1 | https://arxiv.org/pdf/2302.10315v1.pdf | Generalization algorithm of multimodal pre-training model based on graph-text self-supervised training | Recently, a large number of studies have shown that the introduction of visual information can effectively improve the effect of neural machine translation (NMT). Its effectiveness largely depends on the availability of a large number of bilingual parallel sentence pairs and manual image annotation. The lack of images and the effectiveness of images have been difficult to solve. In this paper, a multimodal pre-training generalization algorithm for self-supervised training is proposed, which overcomes the lack of visual information and inaccuracy, and thus extends the applicability of images on NMT. Specifically, we will search for many pictures from the existing sentences through the search engine, and then through the relationship between visual information and text, do the self-supervised training task of graphics and text to obtain more effective visual information for text. We show that when the filtered information is used as multimodal machine translation for fine-tuning, the effect of translation in the global voice dataset is 0.5 BLEU higher than the baseline. | ['Fuxianghua', 'Longzi', 'Tangzhenhao', 'Zhangxiaobing'] | 2023-02-16 | null | null | null | null | ['nmt', 'multimodal-machine-translation'] | ['computer-code', 'natural-language-processing'] | [ 2.76375562e-01 -1.65792421e-01 -2.57154375e-01 -2.28824317e-01
-1.06454849e+00 -7.41456270e-01 6.19995415e-01 -2.99925089e-01
-6.53791130e-01 7.47774243e-01 1.49083748e-01 -4.98405993e-01
5.58842957e-01 -4.15153414e-01 -1.06826079e+00 -5.90602517e-01
5.81194162e-01 5.21580160e-01 2.09148601e-01 -2.10730329e-01
1.78172007e-01 2.84042150e-01 -1.41464496e+00 5.17267227e-01
1.03669071e+00 5.40455639e-01 7.57192075e-01 4.21037883e-01
-2.21023798e-01 1.62862167e-01 -7.87587464e-01 -5.46587050e-01
1.19623370e-01 -7.38037944e-01 -6.09619379e-01 2.17055291e-01
9.16795015e-01 -5.96581221e-01 -1.50460184e-01 9.86758053e-01
9.11151826e-01 -4.19549607e-02 6.05477035e-01 -1.07222164e+00
-7.57250428e-01 4.22118813e-01 -7.42056489e-01 1.15745254e-02
2.22219884e-01 2.89360464e-01 6.66218042e-01 -1.31200719e+00
9.37523782e-01 1.40686750e+00 2.67733872e-01 9.28567886e-01
-1.15210855e+00 -7.26115227e-01 -1.69211835e-01 1.83219910e-01
-1.14171934e+00 -5.74189305e-01 6.08842969e-01 -1.81113556e-01
8.73295724e-01 3.39788467e-01 6.84392691e-01 1.26138294e+00
9.87934768e-02 7.10141540e-01 1.18140185e+00 -8.41666818e-01
-7.08289295e-02 5.67338288e-01 -5.46790063e-01 9.15936768e-01
-1.28129035e-01 1.31170243e-01 -6.68641925e-01 3.78351718e-01
8.80231202e-01 -3.30219448e-01 -3.75684977e-01 -3.44566107e-01
-1.38180804e+00 6.41222119e-01 4.19392586e-01 5.24041176e-01
8.71492922e-02 -1.89842671e-01 2.54002154e-01 3.19707572e-01
4.23855662e-01 5.20857513e-01 -3.49259138e-01 -8.41894075e-02
-1.09894288e+00 -4.29410428e-01 4.25187886e-01 8.65637600e-01
7.11147964e-01 6.89564571e-02 -7.53192678e-02 1.21021497e+00
1.85787037e-01 1.12830031e+00 5.81793249e-01 -8.37505817e-01
8.27472568e-01 5.85223019e-01 -1.47029921e-01 -9.08837676e-01
-1.17584780e-01 -1.20399810e-01 -6.08129799e-01 2.38356605e-01
6.64330363e-01 -2.63047606e-01 -1.11636901e+00 1.82707548e+00
1.92238331e-01 -4.63615954e-01 1.48681372e-01 1.13789189e+00
8.25974345e-01 7.50880718e-01 1.79971673e-03 -4.48360026e-01
1.09845221e+00 -1.21397436e+00 -8.39013040e-01 -4.33924973e-01
5.85527062e-01 -1.22183335e+00 1.61684883e+00 4.87519689e-02
-1.17637300e+00 -6.44038379e-01 -1.11597228e+00 -1.91843435e-01
-3.33498448e-01 5.18436909e-01 3.07586014e-01 5.42214334e-01
-1.30648017e+00 2.92224586e-01 -5.90110898e-01 -7.33387470e-01
2.38453165e-01 4.76640999e-01 -5.68068385e-01 -2.19971031e-01
-1.16316354e+00 1.31786311e+00 4.24984060e-02 -7.10840747e-02
-8.56250703e-01 -2.03045279e-01 -7.44854569e-01 -1.76863208e-01
1.55838490e-01 -8.76516700e-01 9.24247324e-01 -1.56746268e+00
-1.57453358e+00 1.05751073e+00 -2.38714442e-01 -5.67892706e-03
5.92213035e-01 7.60204718e-02 -1.23431966e-01 5.66628218e-01
-1.69916935e-02 1.35028219e+00 9.98929203e-01 -1.36763191e+00
-4.49119925e-01 -4.45815504e-01 -1.06068887e-01 6.90586507e-01
-5.02304137e-01 2.92184241e-02 -8.99481535e-01 -6.09687984e-01
-2.39416976e-02 -1.03410280e+00 1.04981191e-01 -4.37009409e-02
4.99147847e-02 3.93430255e-02 7.93427169e-01 -8.97639811e-01
8.42650414e-01 -2.16830325e+00 3.19614381e-01 -3.27814706e-02
-5.83496578e-02 8.62305090e-02 -5.63131511e-01 2.07758501e-01
5.99735007e-02 1.99386343e-01 -4.93295826e-02 -1.48075193e-01
-3.60314786e-01 1.52280837e-01 4.16307785e-02 2.95047730e-01
6.84199929e-02 1.20567751e+00 -6.83109403e-01 -9.17587280e-01
1.71513781e-01 4.71550107e-01 -3.20725739e-01 1.36506008e-02
-2.11664468e-01 6.26413405e-01 -2.34657992e-02 6.20031595e-01
5.79771042e-01 -1.51889414e-01 2.97966212e-01 -3.94851536e-01
-8.42577145e-02 -1.05151786e-02 -5.30655563e-01 1.80263376e+00
-4.05207932e-01 1.00136936e+00 -2.29449961e-02 -6.31091475e-01
6.56327605e-01 2.55131185e-01 2.71014124e-01 -1.13270295e+00
1.83689699e-01 3.84723067e-01 3.46962623e-02 -6.68408513e-01
2.72393286e-01 -1.90872982e-01 2.56303847e-01 4.32838857e-01
1.37953103e-01 -3.10781360e-01 1.64615065e-01 1.40834913e-01
4.73888993e-01 1.44448534e-01 -1.74331561e-01 -4.18155380e-02
3.28927338e-01 3.80890608e-01 1.07322633e-01 4.75427389e-01
-3.84238698e-02 6.87446773e-01 1.11077219e-01 -2.94339471e-03
-1.43231070e+00 -9.49206412e-01 -5.99390194e-02 1.12058449e+00
2.66772658e-01 -1.86498493e-01 -1.19825256e+00 -6.92360878e-01
-4.33992505e-01 5.99292696e-01 -3.17741632e-01 -1.07768565e-01
-5.69234371e-01 -6.96559429e-01 3.12850475e-01 4.52723145e-01
6.47665262e-01 -9.39062476e-01 -2.94143409e-01 -2.84184366e-01
-6.80780709e-01 -1.13725352e+00 -7.71612227e-01 -2.39318684e-01
-9.98113990e-01 -6.85333908e-01 -1.16706860e+00 -1.14677453e+00
1.04679644e+00 4.80604231e-01 8.98541391e-01 1.17813006e-01
-8.38782266e-02 4.62177873e-01 -9.77371782e-02 -7.93161839e-02
-6.84895992e-01 -2.64240615e-02 7.65878707e-03 -2.21171856e-01
4.53349277e-02 -1.51882172e-01 -4.88463938e-01 6.33651435e-01
-6.42679870e-01 4.64007676e-01 1.02829397e+00 1.14058852e+00
5.09036899e-01 -3.59932780e-01 3.41156870e-01 -4.58348364e-01
6.00368500e-01 2.65075266e-01 -4.78564948e-01 5.11273563e-01
-5.83362579e-01 -8.43353756e-03 5.80185726e-02 -8.83866072e-01
-1.05475044e+00 1.86714247e-01 8.58838037e-02 -5.01819074e-01
1.55776873e-01 3.03149372e-01 -2.56818593e-01 -3.90293300e-01
6.11666977e-01 2.66730189e-01 1.88308403e-01 -1.74612179e-01
5.13382196e-01 7.58964002e-01 3.16945195e-01 -4.10899639e-01
6.58413589e-01 1.31558686e-01 -2.77652770e-01 -8.54099751e-01
-3.79469872e-01 -8.56813118e-02 -6.53866589e-01 -5.14456689e-01
1.00368321e+00 -8.84259343e-01 -3.00341100e-01 2.70969450e-01
-1.21717525e+00 -4.95690793e-01 1.15668230e-01 7.98287868e-01
-4.54545408e-01 5.84303141e-01 -5.57781279e-01 -5.80416024e-01
-3.23355496e-01 -1.46913433e+00 1.06701386e+00 5.01446649e-02
-9.65498686e-02 -7.87810445e-01 -2.09807530e-01 8.85901511e-01
1.78743064e-01 -3.89589548e-01 1.06881762e+00 -2.69614309e-01
-7.73930490e-01 1.45439124e-02 -4.26034510e-01 4.50187564e-01
5.36151556e-03 -1.36228547e-01 -8.77577782e-01 -3.90702397e-01
1.90476067e-02 -4.13534462e-01 7.40936339e-01 3.86381835e-01
5.25914371e-01 -1.70519531e-01 -2.74973392e-01 3.47754538e-01
1.16197526e+00 2.66573787e-01 6.79434597e-01 2.20008582e-01
7.04339206e-01 6.82222426e-01 6.24454856e-01 -3.45509857e-01
3.76790166e-01 7.80334949e-01 1.32970333e-01 -5.93769670e-01
-4.82338846e-01 -2.45172635e-01 5.20202100e-01 1.05183804e+00
-2.72940397e-01 -1.72423586e-01 -7.36694574e-01 3.66692454e-01
-1.54689407e+00 -7.55474746e-01 4.61397059e-02 2.25487351e+00
1.04009807e+00 -1.50889084e-01 8.14998597e-02 -2.56730378e-01
1.00500739e+00 -3.37957531e-01 -2.70380378e-01 -2.21993864e-01
-4.57442373e-01 -2.53271490e-01 5.23285925e-01 6.47220016e-01
-8.14715445e-01 1.41048706e+00 7.09029388e+00 8.80259097e-01
-1.31636310e+00 2.09154680e-01 7.37271965e-01 -5.31108677e-02
-2.48012304e-01 -2.63126254e-01 -6.33064628e-01 3.03771317e-01
7.91291773e-01 2.42962986e-01 7.34242618e-01 4.27973658e-01
2.92005241e-01 -4.05274838e-01 -1.01983333e+00 1.08640540e+00
5.79004824e-01 -1.05934262e+00 2.45707482e-01 4.04248908e-02
8.94840360e-01 -7.96886757e-02 3.36419731e-01 1.57655373e-01
-2.58287251e-01 -8.28620017e-01 4.97348875e-01 2.60405004e-01
1.13755155e+00 -5.63682139e-01 6.59400582e-01 2.85464138e-01
-6.14163280e-01 1.01117842e-01 -3.47445846e-01 3.29052240e-01
-3.20284739e-02 2.18799300e-02 -1.09422171e+00 1.90949723e-01
5.20043612e-01 2.93436855e-01 -8.94811869e-01 7.82675087e-01
-3.33336920e-01 4.21223164e-01 -2.41516113e-01 -2.61473268e-01
5.60416393e-02 -1.78705111e-01 4.12043184e-01 1.05036461e+00
3.96863639e-01 -2.29605272e-01 -1.14542603e-01 6.22335970e-01
-2.15314820e-01 6.25445962e-01 -6.87854648e-01 -2.90571868e-01
2.08412960e-01 1.13933849e+00 -6.74308479e-01 -4.48240846e-01
-6.49870038e-01 1.36935329e+00 2.91327149e-01 6.92438543e-01
-7.57824481e-01 6.42777700e-03 -1.18198790e-01 6.02157004e-02
1.15967505e-01 -2.15125456e-01 -4.37032580e-01 -1.16987133e+00
1.23352401e-01 -1.23980379e+00 4.44033705e-02 -1.26729190e+00
-8.70665073e-01 6.38672590e-01 -1.79151520e-01 -1.22966003e+00
-2.69428164e-01 -6.93992496e-01 -1.74574599e-01 7.67875314e-01
-8.99895668e-01 -1.35696125e+00 -8.62093419e-02 5.59600949e-01
7.51909614e-01 -3.59644771e-01 7.32882857e-01 3.51698607e-01
-4.15457457e-01 8.39671314e-01 1.24398759e-02 2.28833221e-02
1.24811447e+00 -8.67319107e-01 -4.08897214e-02 7.68963397e-01
2.25071773e-01 5.89882970e-01 5.77874362e-01 -8.46548796e-01
-1.45920348e+00 -6.30779147e-01 9.76807296e-01 -6.38751209e-01
4.59534645e-01 -3.93397212e-01 -6.19054675e-01 4.88858819e-01
7.11778045e-01 -6.22603357e-01 4.00414646e-01 -1.30673394e-01
-1.55760631e-01 -2.30815589e-01 -9.30385888e-01 1.03436744e+00
8.51602077e-01 -5.68608224e-01 -5.49525201e-01 4.28339183e-01
6.32796705e-01 -4.53885227e-01 -5.30610681e-01 3.77135485e-01
5.91234803e-01 -5.74947298e-01 1.04671550e+00 -3.60655636e-01
6.37678504e-01 -1.68901429e-01 -1.42807841e-01 -1.38543952e+00
1.49790332e-01 -3.10936034e-01 4.74032789e-01 1.06218874e+00
9.91669893e-01 -2.68331945e-01 5.17265975e-01 3.39807868e-01
1.99565641e-03 -4.37589169e-01 -9.44935203e-01 -4.52969134e-01
9.37948748e-02 6.18719049e-02 1.91188965e-03 9.95691776e-01
2.37065360e-01 8.93653035e-01 -5.61845601e-01 -5.11817150e-02
3.79882723e-01 1.24761619e-01 8.08859110e-01 -7.82766104e-01
-2.02329040e-01 -4.04418796e-01 -5.49620502e-02 -1.02835250e+00
7.29467943e-02 -9.38256025e-01 8.76925513e-02 -1.64754844e+00
6.72919571e-01 1.86744139e-01 1.09007321e-01 5.44686556e-01
-1.67740434e-01 5.71911514e-01 2.95880318e-01 3.72217506e-01
-4.66169238e-01 4.11751688e-01 2.07605100e+00 -3.96120816e-01
-2.10523158e-01 -2.65104026e-01 -3.05626243e-01 4.55427587e-01
7.22536981e-01 -2.08750054e-01 -3.45869482e-01 -7.40393400e-01
4.70920764e-02 2.01582134e-01 2.65595078e-01 -6.93927169e-01
1.89230263e-01 1.89298689e-02 8.71474743e-01 -7.16303349e-01
5.18416166e-01 -8.83651018e-01 -2.42211908e-01 3.41500312e-01
-2.28121206e-01 3.20017844e-01 4.37515527e-01 2.59967089e-01
-1.51675165e-01 -2.59124190e-01 8.02533448e-01 -2.09504932e-01
-5.29400110e-01 -1.40888005e-01 -3.98811072e-01 -1.17423035e-01
6.55264020e-01 -1.49218991e-01 -3.61602783e-01 -6.11420631e-01
-6.99598432e-01 1.50195003e-01 7.51659036e-01 4.00627017e-01
6.71124041e-01 -1.42258453e+00 -5.01720726e-01 1.54678389e-01
-1.82575602e-02 -5.85251331e-01 1.80349737e-01 1.15591371e+00
-4.82717961e-01 3.38630468e-01 -5.62543333e-01 -8.56138647e-01
-1.64890516e+00 7.19426811e-01 3.15828383e-01 2.07484469e-01
-3.93028647e-01 4.82976198e-01 3.95642400e-01 -2.70309210e-01
2.96422362e-01 3.86336297e-02 -1.31776512e-01 -3.36188413e-02
3.93479407e-01 1.05693415e-01 -1.85336232e-01 -7.49048650e-01
-1.88613594e-01 8.54549766e-01 1.25731215e-01 -8.20567131e-01
8.97191107e-01 -4.67640132e-01 -1.32615492e-01 4.26525652e-01
1.13272524e+00 2.04311579e-01 -1.11537600e+00 -3.19123529e-02
-3.69879037e-01 -4.44111973e-01 -3.91738117e-02 -1.23174143e+00
-1.11447370e+00 1.21662712e+00 9.90899324e-01 -3.46377969e-01
1.22994232e+00 1.23741686e-01 5.69351256e-01 4.70617801e-01
7.10892156e-02 -1.27743816e+00 3.89334172e-01 3.80704761e-01
1.04688001e+00 -1.50564563e+00 -1.96234599e-01 -3.60010087e-01
-8.61223102e-01 1.03309119e+00 7.84453511e-01 4.73737031e-01
-1.83339775e-01 1.05395682e-01 5.65364778e-01 1.16404831e-01
-5.00217080e-01 -2.18503118e-01 6.35313392e-01 5.87590396e-01
4.36279058e-01 -1.03214271e-01 -5.36565006e-01 2.11089879e-01
-1.64033130e-01 -3.22085530e-01 1.07658692e-01 4.60506260e-01
-5.19746006e-01 -1.12075913e+00 -5.03374875e-01 8.11127294e-03
-2.53030539e-01 -4.46647853e-01 -8.04314077e-01 8.89267266e-01
-4.76041660e-02 9.19360638e-01 7.12412149e-02 -4.35526282e-01
2.14430377e-01 2.34089196e-01 9.70491707e-01 -2.95985848e-01
-2.37697840e-01 7.70209968e-01 2.41621047e-01 -2.49075890e-01
-4.87861156e-01 -5.14445901e-01 -1.02096689e+00 -1.16642840e-01
-3.66851151e-01 -9.47113857e-02 9.99634266e-01 8.98606360e-01
1.99877813e-01 2.61832327e-01 4.91829962e-01 -8.02068651e-01
-2.52109677e-01 -1.15765738e+00 -8.96388292e-03 3.60629588e-01
3.65821943e-02 -5.52177191e-01 -4.51821029e-01 2.68351793e-01] | [11.4533052444458, 1.4707562923431396] |
f673b0ad-358c-4ca1-9237-959d2a8cb2d7 | an-integrated-multi-time-scale-modeling-for | 1905.02616 | null | https://arxiv.org/abs/1905.02616v2 | https://arxiv.org/pdf/1905.02616v2.pdf | An Integrated Multi-Time-Scale Modeling for Solar Irradiance Forecasting Using Deep Learning | For short-term solar irradiance forecasting, the traditional point forecasting methods are rendered less useful due to the non-stationary characteristic of solar power. The amount of operating reserves required to maintain reliable operation of the electric grid rises due to the variability of solar energy. The higher the uncertainty in the generation, the greater the operating-reserve requirements, which translates to an increased cost of operation. In this research work, we propose a unified architecture for multi-time-scale predictions for intra-day solar irradiance forecasting using recurrent neural networks (RNN) and long-short-term memory networks (LSTMs). This paper also lays out a framework for extending this modeling approach to intra-hour forecasting horizons thus, making it a multi-time-horizon forecasting approach, capable of predicting intra-hour as well as intra-day solar irradiance. We develop an end-to-end pipeline to effectuate the proposed architecture. The performance of the prediction model is tested and validated by the methodical implementation. The robustness of the approach is demonstrated with case studies conducted for geographically scattered sites across the United States. The predictions demonstrate that our proposed unified architecture-based approach is effective for multi-time-scale solar forecasts and achieves a lower root-mean-square prediction error when benchmarked against the best-performing methods documented in the literature that use separate models for each time-scale during the day. Our proposed method results in a 71.5% reduction in the mean RMSE averaged across all the test sites compared to the ML-based best-performing method reported in the literature. Additionally, the proposed method enables multi-time-horizon forecasts with real-time inputs, which have a significant potential for practical industry applications in the evolving grid. | ['Sakshi Mishra', 'Praveen Palanisamy'] | 2019-05-07 | null | null | null | null | ['solar-irradiance-forecasting'] | ['time-series'] | [-3.42208035e-02 -3.63060236e-01 2.43408427e-01 -2.71478087e-01
-3.51399451e-01 -6.96512401e-01 6.19846880e-01 3.26630101e-02
2.81340986e-01 9.07020748e-01 -5.35120964e-02 -7.54131317e-01
-2.85322487e-01 -9.84182596e-01 -3.58237684e-01 -8.37889910e-01
-1.34009004e-01 -3.09517443e-01 -4.60588247e-01 -3.11517924e-01
2.39127964e-01 6.22785270e-01 -1.78631508e+00 -1.51139097e-02
1.15646529e+00 1.20175433e+00 4.58693117e-01 7.42434442e-01
2.34224379e-01 5.18290937e-01 -8.09647620e-01 3.60671490e-01
4.46115226e-01 -4.16765004e-01 5.34142442e-02 -1.03496693e-01
2.39805430e-01 -3.08277041e-01 2.10816562e-01 6.74617112e-01
6.64308488e-01 3.93141389e-01 4.37721938e-01 -8.59361351e-01
-2.90827811e-01 8.43710974e-02 -3.21379900e-01 2.63917536e-01
4.27339040e-02 2.40761358e-02 5.07557929e-01 -7.20589280e-01
-2.36575723e-01 4.45621610e-01 8.06788921e-01 -1.83221325e-01
-9.59639490e-01 -3.03810984e-01 -1.81613788e-02 5.65151870e-02
-1.11683059e+00 -2.11371437e-01 6.35861576e-01 -6.27205729e-01
1.66876078e+00 5.15381277e-01 7.38632619e-01 5.48837245e-01
8.11400473e-01 -7.78312758e-02 1.47108197e+00 -4.70460117e-01
4.72364008e-01 -7.37337694e-02 -1.72391355e-01 1.93150956e-02
8.69611055e-02 7.11667418e-01 -3.04899752e-01 2.07874075e-01
4.54215586e-01 1.19059719e-01 -4.21182573e-01 4.48327869e-01
-8.51235092e-01 5.79769433e-01 6.27841175e-01 5.31239331e-01
-8.83938491e-01 -2.05059066e-01 4.18849468e-01 2.92478234e-01
9.19242442e-01 6.83113337e-02 -7.24911571e-01 -1.61335915e-01
-1.55206561e+00 -1.55323893e-01 9.76468563e-01 5.07745624e-01
3.53759736e-01 9.85396922e-01 -1.45549431e-01 6.79458559e-01
2.51909316e-01 7.83115149e-01 7.85729706e-01 -4.47199613e-01
1.63694620e-01 4.33647126e-01 3.83643299e-01 -6.66976690e-01
-7.03157246e-01 -1.02337384e+00 -1.25236011e+00 6.40578330e-01
-9.52634811e-02 -6.44226015e-01 -9.22130406e-01 1.36970031e+00
7.81371891e-02 2.82957941e-01 4.71873283e-01 6.22846842e-01
2.54087240e-01 1.26470935e+00 -1.29155934e-01 -8.33786607e-01
1.04638672e+00 -1.03493857e+00 -6.63572669e-01 -4.21223119e-02
3.49467129e-01 -9.23280716e-01 4.78807211e-01 3.15562040e-01
-8.36439133e-01 -7.73604512e-01 -1.24659884e+00 4.39835340e-01
-7.37692952e-01 4.54438686e-01 -4.77335304e-02 7.70650685e-01
-1.33361900e+00 7.35020757e-01 -7.03169882e-01 -4.27910388e-01
-4.42309231e-01 3.01806405e-02 2.59992570e-01 7.63667703e-01
-1.18482304e+00 1.30210745e+00 6.14807129e-01 8.65189016e-01
-6.17238641e-01 -8.17592502e-01 -5.94436467e-01 4.92579997e-01
-1.65395632e-01 -6.46882951e-01 1.40082753e+00 -1.04430091e+00
-1.96255624e+00 -1.99390799e-01 -4.47044224e-01 -8.41563582e-01
2.84416825e-01 6.39399216e-02 -8.28158557e-01 -4.62944150e-01
-4.30819660e-01 -1.27825692e-01 6.38555944e-01 -9.44413781e-01
-9.21835482e-01 -2.92369545e-01 -2.43531585e-01 3.54032189e-01
-3.34290296e-01 -3.93133044e-01 5.34434319e-01 -6.20763123e-01
-2.75806844e-01 -8.57667089e-01 -2.36929968e-01 -8.37272823e-01
-3.98345888e-02 -9.85228941e-02 8.55898082e-01 -1.17902112e+00
1.38164830e+00 -1.77317619e+00 -2.61574477e-01 2.52208620e-01
-6.58056140e-01 3.17216456e-01 3.63220811e-01 9.64295566e-01
-3.88269454e-01 -1.04908504e-01 -1.23306468e-01 -2.84715623e-01
-7.14589376e-03 3.10323477e-01 -7.48608410e-01 8.44713971e-02
-1.91575453e-01 6.93636596e-01 -4.37858075e-01 5.30600667e-01
7.86619067e-01 7.32285738e-01 5.92322648e-01 1.83020130e-01
-3.06884110e-01 2.78417021e-01 7.52060562e-02 4.03618306e-01
5.76901138e-01 -2.04833373e-01 2.03881964e-01 -1.02344416e-02
-8.94870877e-01 1.06097095e-01 -9.19532955e-01 1.26293314e+00
-1.14265156e+00 7.21282244e-01 -3.66798609e-01 -6.86825216e-01
1.20372379e+00 5.90522587e-01 1.36227682e-01 -9.52453792e-01
-1.85646936e-01 4.67255414e-01 -1.55068904e-01 -1.44307867e-01
5.53086579e-01 -1.56245068e-01 6.08709693e-01 3.94266695e-01
-4.30730641e-01 -5.55722266e-02 3.73060256e-01 -5.73793948e-01
4.49553490e-01 3.01737785e-01 5.09393275e-01 -4.88309324e-01
7.66524196e-01 -3.84188712e-01 5.88184714e-01 1.34279400e-01
2.95864850e-01 6.86489418e-02 -1.27697706e-01 -5.82819402e-01
-1.10315537e+00 -5.23425281e-01 -3.74103874e-01 7.06309974e-01
-3.96566093e-01 -3.33900005e-02 -5.40090144e-01 -7.19153360e-02
3.31553956e-03 1.49916089e+00 -5.23111880e-01 4.27891910e-01
-1.04226463e-01 -9.76003885e-01 -5.49709890e-03 5.08213341e-01
5.09318829e-01 -9.76573110e-01 -1.12366164e+00 4.45329905e-01
2.58307546e-01 -7.40645349e-01 5.46974577e-02 3.22321773e-01
-1.04974508e+00 -5.04495561e-01 -9.08746660e-01 -3.54076385e-01
4.29161370e-01 1.10225953e-01 1.21672809e+00 -2.06978917e-01
9.55594517e-03 2.05760002e-01 -1.25729188e-01 -7.42650330e-01
-3.13799739e-01 -3.95605676e-02 6.68227300e-02 -2.14959607e-01
9.50357616e-02 -7.20509052e-01 -7.69827008e-01 2.77457505e-01
-7.62745321e-01 1.32332236e-01 3.74851555e-01 7.82708466e-01
4.32229608e-01 4.34882343e-01 1.05436432e+00 -3.96384776e-01
7.32626319e-01 -5.77111006e-01 -1.50155818e+00 5.85260451e-01
-1.41715896e+00 -2.67992049e-01 1.07677698e+00 6.10549301e-02
-1.37469316e+00 -2.13291556e-01 8.49854276e-02 -1.41202569e-01
-9.53513943e-03 1.08521342e+00 5.28027833e-01 -6.57941252e-02
4.35885370e-01 5.12781620e-01 -2.99287081e-01 -4.91516650e-01
4.08788137e-02 7.32636511e-01 2.28711367e-01 4.62894840e-03
6.86776817e-01 -1.33034782e-02 4.15721685e-01 -9.82082963e-01
-5.59300482e-01 -2.30865389e-01 -4.31103975e-01 -4.86766636e-01
5.54597199e-01 -1.18599534e+00 -8.01828146e-01 7.18717992e-01
-9.90798533e-01 -5.28706908e-01 -1.58512786e-01 3.42640817e-01
-3.44850123e-01 2.10081175e-01 -2.76969463e-01 -1.36516464e+00
-1.04663992e+00 -8.44605744e-01 6.62530363e-01 7.06015050e-01
1.64992854e-01 -1.51364529e+00 2.10859939e-01 -8.98278132e-02
1.03316963e+00 4.89947557e-01 6.87900484e-01 -2.50821620e-01
-3.65981966e-01 -2.01340109e-01 -2.78767701e-02 8.84892225e-01
2.83750832e-01 2.93454975e-01 -1.14978111e+00 -5.88533521e-01
2.30006635e-01 1.08861320e-01 3.26384634e-01 6.11812472e-01
7.95918524e-01 -3.11793894e-01 2.16768265e-01 4.56361085e-01
2.07201457e+00 5.79413056e-01 5.50203264e-01 4.22616243e-01
6.64184168e-02 3.14443737e-01 5.15486121e-01 7.05124974e-01
3.23019594e-01 2.29894072e-01 5.66075683e-01 -2.71593988e-01
3.25473249e-01 1.50495954e-02 6.02547169e-01 1.04917526e+00
-2.82614410e-01 -4.04785782e-01 -7.57523775e-01 6.71056986e-01
-1.76340985e+00 -1.03294897e+00 -8.40367004e-02 2.68735194e+00
3.29343855e-01 -7.65137300e-02 -2.16201127e-01 1.51914060e-01
3.77985656e-01 3.74375999e-01 -7.22821534e-01 -1.01928365e+00
-7.22411349e-02 2.39770725e-01 6.85297549e-01 6.06622040e-01
-7.53660202e-01 1.55002490e-01 6.16977644e+00 4.59205985e-01
-1.47291505e+00 -4.18565087e-02 7.33198464e-01 -5.40931597e-02
6.93553034e-03 2.04800460e-02 -7.13979304e-01 7.53481984e-01
1.88753629e+00 -6.93445504e-01 6.08117640e-01 7.56842911e-01
1.02903032e+00 -4.74603951e-01 -8.39918494e-01 5.62254429e-01
1.00708820e-01 -1.22303414e+00 -3.98071647e-01 1.47886410e-01
1.26407146e+00 5.14950335e-01 -1.98740065e-01 2.33151317e-01
-2.52490461e-01 -9.11500275e-01 3.23467910e-01 1.05548501e+00
5.44447064e-01 -8.30366433e-01 1.04066980e+00 7.53679574e-01
-1.41959286e+00 -4.46974874e-01 -4.37214941e-01 -5.72762370e-01
4.36368406e-01 1.09837484e+00 -7.85793364e-01 1.26349187e+00
6.20764613e-01 7.70936131e-01 -2.52953619e-01 9.80952322e-01
-1.21311091e-01 6.73357844e-01 -5.87797046e-01 -1.57120693e-02
1.78067148e-01 -6.74187422e-01 2.50061512e-01 1.00842285e+00
1.05502439e+00 -7.69120008e-02 5.43684848e-02 4.63479340e-01
3.11569721e-01 4.35742885e-02 -6.46493554e-01 9.53727439e-02
3.92444521e-01 1.47195339e+00 -2.63773441e-01 -5.17464161e-01
-4.48874921e-01 5.65918684e-01 -2.74832249e-01 6.70618296e-01
-6.55995786e-01 -1.99494898e-01 3.94137412e-01 -3.17151278e-01
3.54939163e-01 -2.69096494e-01 -6.08403862e-01 -8.80918920e-01
3.12584162e-01 -5.84562421e-01 1.90711185e-01 -1.17101359e+00
-1.32584381e+00 9.13775146e-01 -1.82736263e-01 -1.38313866e+00
-9.81814921e-01 -4.63778526e-01 -1.03421748e+00 1.80031681e+00
-2.07330132e+00 -1.07776058e+00 -4.81850624e-01 2.27083415e-02
8.38051021e-01 -2.67455161e-01 1.22809172e+00 3.02258078e-02
-6.73697531e-01 -3.26712206e-02 8.57869029e-01 -7.08984613e-01
1.44036993e-01 -1.34370863e+00 3.47182870e-01 1.26364231e+00
-2.03534052e-01 3.12982500e-01 7.47688353e-01 -5.12788713e-01
-1.07768738e+00 -1.04246509e+00 9.35873210e-01 1.34173870e-01
6.97432578e-01 1.86211392e-01 -8.43557596e-01 6.67936862e-01
9.54746366e-01 -3.79907817e-01 6.62867844e-01 -1.78771410e-02
1.06379032e-01 -5.41107893e-01 -1.02259398e+00 9.41769183e-02
-4.85269167e-02 -4.92816925e-01 -3.86101216e-01 3.74279827e-01
2.10228473e-01 -4.17953521e-01 -1.44005167e+00 7.20370948e-01
6.58123851e-01 -1.26726389e+00 4.31034386e-01 1.45717353e-01
2.28881449e-01 -5.34681618e-01 -1.62357122e-01 -1.84269476e+00
-1.90991774e-01 -6.15100086e-01 -3.75860304e-01 1.15095735e+00
5.79726815e-01 -1.17044318e+00 2.59284198e-01 6.36104763e-01
-2.81102538e-01 -9.37350452e-01 -9.80531514e-01 -7.30005622e-01
-2.22681947e-02 -7.99473077e-02 6.95054352e-01 7.55177677e-01
-2.20691055e-01 8.44774395e-02 -2.61161089e-01 6.50803506e-01
3.91829967e-01 5.66517234e-01 5.10219634e-01 -1.07831681e+00
-7.64575824e-02 -3.18160474e-01 -7.12787881e-02 -6.12572193e-01
-1.24009416e-01 -3.47215027e-01 -3.92208062e-02 -1.78544414e+00
-2.96889573e-01 3.65200415e-02 -6.47812247e-01 5.65279782e-01
1.17753081e-01 2.29929779e-02 2.42867023e-01 6.55181929e-02
5.35464168e-01 7.47617185e-01 4.95996594e-01 2.42829248e-01
-2.25862384e-01 3.55327010e-01 6.05603531e-02 3.92061740e-01
1.14662695e+00 2.00554535e-01 -6.63529336e-01 -2.71047503e-01
3.15438837e-01 2.69091487e-01 1.78163096e-01 -1.28644598e+00
2.53243774e-01 -6.78733438e-02 6.25230789e-01 -1.02340877e+00
1.66067153e-01 -1.07225883e+00 7.39575028e-01 5.02808869e-01
3.02887857e-01 5.80589592e-01 5.54877281e-01 3.67178708e-01
-3.38905990e-01 -1.23255119e-01 5.47754824e-01 2.65705884e-01
-6.80783749e-01 -1.23354152e-01 -3.64012063e-01 -8.38550270e-01
1.20367122e+00 -1.62113219e-01 -6.51013315e-01 -4.00333345e-01
-5.83463550e-01 3.06602985e-01 3.70536655e-01 3.39522839e-01
1.39207259e-01 -8.43001485e-01 -7.30128109e-01 2.90468246e-01
-2.85434902e-01 -3.65853608e-01 5.86317182e-01 7.55124152e-01
-5.25875211e-01 7.30139434e-01 -1.34537771e-01 -5.34496963e-01
-1.01881838e+00 3.39807034e-01 8.83108258e-01 -2.75319040e-01
-4.47130531e-01 1.12624563e-01 -3.11848193e-01 -2.25652084e-01
-1.01083063e-01 -6.42610967e-01 -2.32161403e-01 2.27156147e-01
4.24617678e-01 6.20440066e-01 6.64096594e-01 -4.95080113e-01
7.92294815e-02 4.96521950e-01 5.87841749e-01 2.23482758e-01
1.39964402e+00 -2.75346458e-01 -4.03761715e-01 1.08030784e+00
7.31742263e-01 -2.73372978e-01 -1.20123601e+00 2.11067721e-01
-2.58931480e-02 -1.24580733e-01 5.53005159e-01 -1.56318140e+00
-9.49747264e-01 7.17248023e-01 1.05072498e+00 7.75531471e-01
1.66912246e+00 -1.19637096e+00 6.37207031e-01 3.30562055e-01
3.30164611e-01 -1.19082367e+00 -1.00440729e+00 5.37610292e-01
9.77294147e-01 -1.01822031e+00 1.57193422e-01 2.01437935e-01
-3.99251163e-01 1.43020940e+00 2.07302079e-01 1.83267146e-01
7.71138370e-01 2.32483253e-01 4.25779670e-01 2.56412327e-01
-1.15338624e+00 3.17665702e-03 3.46429378e-01 1.24003487e-02
6.78655863e-01 3.79474461e-01 -3.94346207e-01 4.68681741e-04
-2.33090743e-01 3.16360950e-01 5.03098011e-01 6.75619006e-01
-2.39066884e-01 -8.35814714e-01 -5.10697484e-01 4.13991153e-01
-2.76144117e-01 -1.77318498e-01 2.59753793e-01 5.00514448e-01
-1.05187699e-01 1.25852013e+00 1.40867442e-01 7.38084093e-02
3.58982861e-01 3.69606912e-01 -1.17921986e-01 -1.19626090e-01
-8.89938593e-01 1.78915396e-01 5.01233293e-03 -2.76405782e-01
-4.51138377e-01 -4.95578766e-01 -8.90878916e-01 -4.58775431e-01
-4.10036415e-01 2.02573493e-01 1.45059788e+00 1.02545273e+00
5.79597712e-01 8.16390336e-01 1.24040926e+00 -1.16132724e+00
-8.14476728e-01 -1.42322338e+00 -6.64658427e-01 -2.66148239e-01
2.96947837e-01 -2.30431154e-01 -5.75018585e-01 -3.67487781e-02] | [6.236677646636963, 2.7851836681365967] |
ca54bf93-e330-4a73-9c81-cd15feb4c7b4 | temporal-point-cloud-completion-with-pose | 2202.03084 | null | https://arxiv.org/abs/2202.03084v1 | https://arxiv.org/pdf/2202.03084v1.pdf | Temporal Point Cloud Completion with Pose Disturbance | Point clouds collected by real-world sensors are always unaligned and sparse, which makes it hard to reconstruct the complete shape of object from a single frame of data. In this work, we manage to provide complete point clouds from sparse input with pose disturbance by limited translation and rotation. We also use temporal information to enhance the completion model, refining the output with a sequence of inputs. With the help of gated recovery units(GRU) and attention mechanisms as temporal units, we propose a point cloud completion framework that accepts a sequence of unaligned and sparse inputs, and outputs consistent and aligned point clouds. Our network performs in an online manner and presents a refined point cloud for each frame, which enables it to be integrated into any SLAM or reconstruction pipeline. As far as we know, our framework is the first to utilize temporal information and ensure temporal consistency with limited transformation. Through experiments in ShapeNet and KITTI, we prove that our framework is effective in both synthetic and real-world datasets. | ['Shaojie Shen', 'Xiaozhi Chen', 'Peiliang Li', 'Lingyun Xu', 'Jieqi Shi'] | 2022-02-07 | null | null | null | null | ['point-cloud-completion'] | ['computer-vision'] | [ 6.08074442e-02 -1.14157982e-01 6.41847774e-02 -2.61982650e-01
-4.43799853e-01 -6.97163999e-01 5.24894476e-01 -4.38536853e-02
-2.46754318e-01 3.42526436e-01 -1.38798356e-01 -9.00044367e-02
9.02861804e-02 -7.02866375e-01 -1.08076799e+00 -3.56980413e-01
2.10361425e-02 7.69777358e-01 4.81305867e-01 -1.17624477e-01
1.62814096e-01 9.97791290e-01 -1.68178380e+00 -2.61440165e-02
6.18477404e-01 8.62413466e-01 6.31529212e-01 5.57276845e-01
-8.46495777e-02 4.13949966e-01 -1.82632282e-01 5.11864051e-02
5.47968447e-01 2.65666127e-01 -6.65340960e-01 3.80075395e-01
5.74967504e-01 -6.01391137e-01 -5.20813227e-01 7.66664565e-01
3.20216745e-01 1.61638856e-01 7.07372674e-04 -1.11214125e+00
-2.93917209e-01 1.78355664e-01 -6.19630694e-01 -2.63376415e-01
4.28408772e-01 4.08726126e-01 4.97402519e-01 -1.24188280e+00
9.52287018e-01 1.15159845e+00 7.57643998e-01 3.23208779e-01
-1.14278305e+00 -7.34616458e-01 4.63225722e-01 -7.50201643e-02
-1.46968126e+00 -6.98016703e-01 7.93970466e-01 -2.18677133e-01
1.03736603e+00 2.80694254e-02 9.12724078e-01 8.17444146e-01
-1.45894811e-01 4.53177631e-01 4.27790523e-01 3.36352829e-03
4.14256573e-01 -5.73274791e-01 -3.15393150e-01 4.80303526e-01
2.49791294e-01 2.66869485e-01 -6.95674956e-01 -1.45278946e-01
1.17436552e+00 5.66803157e-01 -3.33087802e-01 -7.55082428e-01
-1.75071919e+00 3.94022584e-01 8.04969251e-01 8.58646911e-03
-6.54964805e-01 5.25391221e-01 -1.43659934e-01 2.82251209e-01
4.20501590e-01 7.71937221e-02 -4.93067771e-01 -1.11047052e-01
-1.04397786e+00 1.48079872e-01 5.88719726e-01 1.48288035e+00
9.76963639e-01 1.14674516e-01 4.64174241e-01 2.54971027e-01
2.44483307e-01 9.50164855e-01 5.70521457e-03 -1.29778361e+00
6.22364342e-01 4.45962697e-01 3.45000088e-01 -9.78346646e-01
-3.92769903e-01 -1.81500360e-01 -8.68413210e-01 4.19213355e-01
9.66888200e-03 1.80644289e-01 -9.96852636e-01 1.63045645e+00
6.15761340e-01 8.23624492e-01 -1.21154673e-01 1.20935655e+00
5.24394572e-01 6.43952906e-01 -2.20286593e-01 -2.74831831e-01
9.48011756e-01 -6.47071123e-01 -4.29014087e-01 -4.19980794e-01
2.14999437e-01 -7.49529958e-01 6.34530485e-01 3.97025198e-01
-1.14024615e+00 -5.82322598e-01 -9.86910343e-01 -2.95924783e-01
1.03202201e-01 -1.85095504e-01 7.26994991e-01 -2.61658192e-01
-1.04025626e+00 8.41797829e-01 -1.62458241e+00 -4.20577914e-01
2.29075953e-01 4.61191952e-01 -8.11147630e-01 -5.16769171e-01
-2.88894415e-01 7.11230934e-01 2.00215757e-01 3.90721440e-01
-9.15699482e-01 -6.94633305e-01 -1.01313245e+00 -2.27232873e-01
3.94060969e-01 -1.13212395e+00 1.16041279e+00 -6.48411512e-01
-1.18639386e+00 3.68663460e-01 -5.09530067e-01 -2.95654535e-01
4.32910383e-01 -3.05298835e-01 1.27206743e-01 1.69217154e-01
1.10817291e-01 9.17726576e-01 9.05629575e-01 -1.46000707e+00
-5.01419306e-01 -6.68400109e-01 -1.35901779e-01 2.69881248e-01
3.27831715e-01 -3.32535028e-01 -8.38383198e-01 -2.99154013e-01
1.07675254e+00 -1.11143482e+00 -5.14765143e-01 4.21550363e-01
-9.03289318e-02 1.22096449e-01 9.16822433e-01 -5.31299472e-01
4.41612363e-01 -2.12753892e+00 3.26635718e-01 2.40680203e-01
2.37271085e-01 -1.35511160e-01 -3.19215447e-01 2.84124315e-01
-5.33880852e-03 -1.69923216e-01 -3.48840863e-01 -1.06704652e+00
-2.46660501e-01 7.55262733e-01 -6.31345272e-01 8.02519679e-01
3.54991227e-01 9.28715765e-01 -1.09940386e+00 -1.63083404e-01
5.72864771e-01 7.59100318e-01 -6.03508234e-01 2.00168967e-01
-3.99598569e-01 9.00519431e-01 -4.76524234e-01 8.48224819e-01
9.96886551e-01 -2.02013075e-01 -8.21127295e-02 -2.54881054e-01
-4.67169732e-01 2.38319695e-01 -1.46967936e+00 2.56717587e+00
-4.88605052e-01 3.92724931e-01 2.82855898e-01 -4.00059998e-01
7.95311332e-01 3.30283403e-01 6.68519497e-01 -4.82183158e-01
5.51470648e-03 2.05070272e-01 -3.28028589e-01 -1.63358673e-01
7.39481270e-01 1.26549974e-01 1.46235436e-01 3.57880235e-01
5.32333516e-02 -5.28342366e-01 -1.76431358e-01 2.22812027e-01
1.26124966e+00 6.52542114e-01 -3.68896648e-02 3.22670758e-01
2.09733881e-02 1.20509423e-01 6.97372675e-01 5.49678385e-01
2.00775564e-01 1.15727794e+00 -1.74992546e-01 -6.89844191e-01
-1.40268338e+00 -1.14566696e+00 8.24914500e-02 5.25452495e-01
5.06533265e-01 -2.87024558e-01 -2.32555028e-02 -1.42422214e-01
1.15502760e-01 2.86970198e-01 -1.65742964e-01 1.57655537e-01
-8.06704581e-01 -3.31357680e-02 -5.49829826e-02 6.24672830e-01
1.08486883e-01 -9.62742090e-01 -8.21609795e-01 3.83173436e-01
-2.97698677e-01 -1.37597668e+00 -4.11553025e-01 1.59317896e-01
-1.32775283e+00 -1.01248324e+00 -3.24098170e-01 -5.36360919e-01
9.66530621e-01 7.56673515e-01 1.12596273e+00 3.05468023e-01
7.33741745e-02 3.52096409e-01 -3.08863342e-01 -2.24960342e-01
2.17441674e-02 -7.68912956e-02 2.67789036e-01 -9.53414440e-02
-1.31760821e-01 -1.17154276e+00 -5.37146330e-01 1.90866739e-01
-9.76422250e-01 3.28795850e-01 5.02968550e-01 4.95876908e-01
1.06384337e+00 -3.67218196e-01 2.93712094e-02 -2.99606144e-01
-1.29982293e-01 -4.07894939e-01 -9.02089059e-01 -2.00097084e-01
-2.91846186e-01 -1.41679486e-02 3.99571270e-01 -4.25588191e-01
-7.18372524e-01 7.63688207e-01 -2.82729547e-02 -1.31163168e+00
-1.27182290e-01 3.92033011e-01 -9.13058966e-02 -1.16532661e-01
4.46108729e-01 1.12180769e-01 1.78785231e-02 -6.82161748e-01
5.82400084e-01 1.43326983e-01 1.06391549e+00 -5.42920351e-01
1.17348957e+00 9.90013182e-01 1.66347083e-02 -6.38824940e-01
-4.72200394e-01 -6.88404977e-01 -9.61348057e-01 -1.19266145e-01
3.97355616e-01 -1.20423150e+00 -8.35786283e-01 2.52835423e-01
-1.53336310e+00 -1.34411886e-01 -3.79205495e-01 5.18132925e-01
-5.87575316e-01 5.09710252e-01 -5.08886755e-01 -7.25738466e-01
-3.09431493e-01 -1.02858961e+00 1.53602111e+00 -3.22443247e-02
9.55500007e-02 -3.26289654e-01 -3.36567983e-02 -1.21605463e-01
1.78226113e-01 3.80056232e-01 -2.71360930e-02 -6.02685250e-02
-1.33595610e+00 -6.87879995e-02 2.72543095e-02 -4.88024838e-02
-8.62655044e-03 5.07701077e-02 -8.02411258e-01 -6.10044479e-01
2.70050108e-01 -1.50856897e-01 7.87581742e-01 2.89249867e-01
1.03312814e+00 -2.08086610e-01 -3.74943048e-01 1.16143382e+00
1.52438104e+00 -1.59112155e-01 5.46894848e-01 1.37782678e-01
8.54984224e-01 1.98770463e-01 7.86074698e-01 6.43680871e-01
5.23712516e-01 4.92188543e-01 1.10176730e+00 2.09466387e-02
1.45282866e-02 -4.80525702e-01 1.04750268e-01 1.03872991e+00
-2.47228429e-01 3.80715095e-02 -9.50860441e-01 6.95301831e-01
-2.05048442e+00 -8.25083077e-01 -5.04467227e-02 2.25479102e+00
5.12910485e-01 -4.78581823e-02 -3.86057168e-01 -9.44859907e-02
4.95420337e-01 2.18310997e-01 -7.23900557e-01 3.17706466e-01
5.89900501e-02 3.35349083e-01 5.94949782e-01 6.22926056e-01
-7.50028849e-01 1.05404711e+00 5.34998226e+00 1.68401320e-02
-1.23010433e+00 9.88782048e-02 -2.85447296e-02 -4.16336298e-01
-4.83012676e-01 4.68696058e-01 -3.42322052e-01 1.79240018e-01
6.51867807e-01 7.21104816e-02 6.37355685e-01 8.01744342e-01
2.08144128e-01 -6.57056272e-02 -1.35499704e+00 1.27802050e+00
-1.53095588e-01 -1.43587112e+00 -2.49610513e-01 -6.56073019e-02
6.79984987e-01 7.54678369e-01 -3.03517699e-01 -9.41147730e-02
5.10073006e-01 -8.03271055e-01 9.75528955e-01 6.56543732e-01
8.32810581e-01 -6.34239733e-01 4.99537289e-01 7.19686151e-01
-1.37930810e+00 1.88186452e-01 -6.26811564e-01 -2.19351590e-01
4.86775488e-01 6.83376074e-01 -8.70093882e-01 8.63354623e-01
6.07617974e-01 1.02413464e+00 -2.26191103e-01 1.12484062e+00
-1.87650979e-01 1.94617942e-01 -9.23155546e-01 5.99801004e-01
-9.80776772e-02 -2.84218073e-01 5.88143766e-01 5.17357469e-01
7.28511572e-01 5.33637583e-01 5.18105924e-01 8.80921364e-01
1.27465814e-01 -4.45219815e-01 -8.36017489e-01 3.71362954e-01
8.14478993e-01 1.28370702e+00 -6.58647358e-01 -2.25411847e-01
-3.43401015e-01 9.96355653e-01 4.08914655e-01 3.23502630e-01
-5.65056562e-01 3.00446242e-01 7.19477594e-01 -7.72892684e-02
3.21706980e-01 -7.72774994e-01 -5.31452000e-01 -1.44955730e+00
3.31233293e-01 -4.83823121e-01 -5.49976490e-02 -1.16311073e+00
-1.03209054e+00 6.79865658e-01 -1.93398908e-01 -1.70030189e+00
-2.40923420e-01 -2.75544729e-02 -2.75859058e-01 9.32013750e-01
-1.56318676e+00 -1.31500065e+00 -8.04663718e-01 6.89046502e-01
3.96111846e-01 3.19158107e-01 5.51507890e-01 2.78289348e-01
-1.52203843e-01 -2.39045486e-01 -3.74719262e-01 -8.03820416e-02
4.81105775e-01 -8.00872207e-01 8.55810106e-01 1.07853246e+00
2.48747319e-01 6.91002965e-01 7.18069136e-01 -8.76439154e-01
-2.04888535e+00 -1.21076739e+00 7.51492143e-01 -5.37488163e-01
3.95786196e-01 -3.03212255e-01 -1.00884151e+00 1.01070201e+00
-2.00395748e-01 5.62164187e-01 -2.01154932e-01 -2.32146844e-01
-2.93424159e-01 -9.18430761e-02 -9.39272046e-01 3.99870843e-01
1.46963096e+00 -4.17522639e-01 -5.15409648e-01 3.50548744e-01
1.12641311e+00 -1.09071255e+00 -7.61851192e-01 6.27620757e-01
2.79152602e-01 -7.49534726e-01 1.02417111e+00 -1.98824272e-01
2.48212516e-01 -8.35806847e-01 -2.28204846e-01 -1.07846332e+00
-3.08979630e-01 -5.32418549e-01 -2.09790125e-01 9.40183461e-01
4.49177139e-02 -5.10325909e-01 8.69943619e-01 5.89328229e-01
-5.29051006e-01 -3.60217124e-01 -1.09135580e+00 -5.96434057e-01
-4.64529663e-01 -9.75308180e-01 9.17629659e-01 8.29229116e-01
-5.13881803e-01 8.03804174e-02 -3.53193134e-01 8.81639898e-01
6.50905669e-01 5.30282021e-01 1.17874479e+00 -1.16528475e+00
-4.00566123e-02 1.24149792e-01 -4.21734482e-01 -1.35976434e+00
1.11498058e-01 -6.83038116e-01 3.78715634e-01 -1.57117951e+00
-2.00767308e-01 -7.53405035e-01 1.16221331e-01 7.63407111e-01
2.22462952e-01 2.32804045e-01 4.51121747e-01 5.76518834e-01
-4.53523457e-01 7.43551552e-01 1.07190323e+00 -5.56294732e-02
-2.61643171e-01 -4.27291840e-02 -2.29460537e-01 6.12695873e-01
3.89139622e-01 -5.89494705e-01 -3.24438840e-01 -1.06539750e+00
2.67825812e-01 4.59267795e-01 5.70336699e-01 -1.14881778e+00
5.36753058e-01 -3.62926573e-01 4.57349002e-01 -1.21930802e+00
7.11709082e-01 -1.41105974e+00 8.35638344e-01 2.78587252e-01
2.84480423e-01 4.41254050e-01 1.28210872e-01 5.85973442e-01
-1.43186316e-01 2.12625161e-01 4.45609808e-01 -2.10359335e-01
-7.62524784e-01 1.02390230e+00 4.71794814e-01 -4.77372766e-01
8.29293966e-01 -2.45308518e-01 -1.40810072e-01 -2.86482275e-01
-7.28857338e-01 4.86093462e-01 1.16601741e+00 5.61788142e-01
9.43806112e-01 -1.38387060e+00 -6.08896255e-01 4.83670384e-01
1.91713646e-01 1.05602753e+00 2.93269396e-01 7.54399776e-01
-7.38300264e-01 1.11501761e-01 -8.16988479e-03 -1.24260318e+00
-9.08242643e-01 6.88910127e-01 7.96670392e-02 1.83445409e-01
-1.04080558e+00 4.29784119e-01 -3.69052403e-02 -6.13184035e-01
2.27149591e-01 -6.40464067e-01 3.16495359e-01 -3.41757834e-01
4.23668742e-01 1.92354023e-02 3.47303569e-01 -6.89341664e-01
-3.46149415e-01 7.21617520e-01 3.78403753e-01 -3.23467523e-01
1.62113237e+00 -1.96806252e-01 -2.46684670e-01 4.81716812e-01
8.69930923e-01 -8.54627229e-04 -1.65921903e+00 -3.92145753e-01
-2.38779411e-01 -7.25466728e-01 -1.00559384e-01 -1.92866474e-01
-1.20600331e+00 6.51083529e-01 3.11658859e-01 -3.21293384e-01
1.00995755e+00 3.48020531e-02 7.86489844e-01 4.76035655e-01
9.16840851e-01 -6.27862930e-01 -2.33354911e-01 6.88779056e-01
1.06098509e+00 -1.12194169e+00 4.79494184e-02 -4.11871552e-01
-2.46382564e-01 9.88736033e-01 5.58375716e-01 -2.98962206e-01
4.31044757e-01 4.54398245e-01 -2.05245852e-01 -8.72170478e-02
-1.12070119e+00 -2.03687549e-01 -3.10898479e-02 6.29718244e-01
-2.00204045e-01 -1.23780765e-01 1.62693828e-01 1.24440178e-01
-3.09653193e-01 2.48955682e-01 3.82578731e-01 1.27677941e+00
-4.81045395e-01 -9.92569029e-01 -5.97506046e-01 1.10935099e-01
1.93797782e-01 9.02460590e-02 -3.69257890e-02 5.52482486e-01
-1.06974868e-02 7.29856610e-01 3.36450517e-01 -4.00415719e-01
4.97642100e-01 -2.28554681e-01 4.38474804e-01 -7.45201349e-01
-2.31846511e-01 2.06780970e-01 -2.68966854e-01 -1.10114408e+00
-5.17153382e-01 -6.77865982e-01 -1.56932616e+00 -5.19123495e-01
-1.34795770e-01 -1.73568502e-01 8.77767801e-01 8.38184357e-01
6.72876358e-01 3.46140683e-01 6.28222406e-01 -1.59985518e+00
-2.97389746e-01 -8.16026866e-01 -2.57406622e-01 4.54544663e-01
6.71642005e-01 -4.61303264e-01 -1.94609419e-01 1.84290707e-01] | [8.30034065246582, -3.2644920349121094] |
428437a2-47d4-44e6-bb23-afcd10c16155 | fast-algorithm-for-overcomplete-order-3 | 2202.06442 | null | https://arxiv.org/abs/2202.06442v2 | https://arxiv.org/pdf/2202.06442v2.pdf | Fast algorithm for overcomplete order-3 tensor decomposition | We develop the first fast spectral algorithm to decompose a random third-order tensor over $\mathbb{R}^d$ of rank up to $O(d^{3/2}/\text{polylog}(d))$. Our algorithm only involves simple linear algebra operations and can recover all components in time $O(d^{6.05})$ under the current matrix multiplication time. Prior to this work, comparable guarantees could only be achieved via sum-of-squares [Ma, Shi, Steurer 2016]. In contrast, fast algorithms [Hopkins, Schramm, Shi, Steurer 2016] could only decompose tensors of rank at most $O(d^{4/3}/\text{polylog}(d))$. Our algorithmic result rests on two key ingredients. A clean lifting of the third-order tensor to a sixth-order tensor, which can be expressed in the language of tensor networks. A careful decomposition of the tensor network into a sequence of rectangular matrix multiplications, which allows us to have a fast implementation of the algorithm. | ['David Steurer', 'Stefan Tiegel', 'Chih-Hung Liu', "Tommaso d'Orsi", 'Jingqiu Ding'] | 2022-02-14 | null | null | null | null | ['tensor-networks'] | ['methodology'] | [ 7.62534514e-02 -8.38782359e-03 1.79603055e-01 -3.69633436e-02
-6.98293447e-01 -1.06218743e+00 1.03845224e-01 1.34636521e-01
-4.69670743e-01 2.29531154e-01 -2.97171623e-03 -9.57641423e-01
-4.18892622e-01 -7.73822844e-01 -8.39627862e-01 -6.67019248e-01
-1.06807208e+00 4.61162925e-01 2.90450454e-01 -5.49510777e-01
2.55133986e-01 4.64844286e-01 -1.34782875e+00 4.34993915e-02
4.60024416e-01 1.32136464e+00 -4.96055901e-01 1.11481583e+00
6.16069447e-05 6.47499263e-01 3.04406099e-02 -8.42488647e-01
8.83201122e-01 -1.63488612e-01 -1.25741315e+00 1.36608183e-01
6.47858441e-01 -4.07280654e-01 -6.36909306e-01 1.30008256e+00
1.12302624e-01 2.50739306e-02 3.40882659e-01 -1.09119260e+00
-1.96727261e-01 9.25767958e-01 -7.97783792e-01 1.64211869e-01
2.89910644e-01 -4.94459048e-02 1.45692456e+00 -9.05918002e-01
4.36558574e-01 9.93096709e-01 4.33102340e-01 -3.78696509e-02
-1.47871566e+00 -6.61794186e-01 -2.89723799e-02 8.98244139e-03
-1.50755084e+00 -3.28683794e-01 6.16292179e-01 -4.72086549e-01
6.53424740e-01 6.16790116e-01 5.85092723e-01 3.91349941e-02
-1.80038080e-01 4.83889878e-01 1.18476975e+00 -3.02259833e-01
-2.93348655e-02 -4.04022485e-01 4.92907405e-01 1.14354956e+00
7.08851099e-01 -2.18105406e-01 -5.24921417e-01 -4.60660875e-01
7.97989428e-01 1.76263116e-02 -9.52677131e-02 -1.26924604e-01
-1.53850150e+00 7.58421659e-01 2.13809624e-01 2.00615883e-01
2.48957761e-02 5.04395366e-01 5.29405177e-01 5.32131970e-01
2.23419353e-01 1.36160702e-01 -4.99291658e-01 -1.85794681e-01
-9.16895330e-01 3.04015309e-01 9.75273609e-01 1.16282272e+00
1.01480031e+00 9.89454836e-02 3.43450963e-01 3.92005831e-01
1.24604821e-01 8.46039593e-01 -4.44021434e-01 -1.27879298e+00
7.79269934e-01 1.61168188e-01 -2.33274209e-03 -1.12492502e+00
-3.99284273e-01 -1.92648157e-01 -1.23884320e+00 -8.41530785e-02
6.90750003e-01 -2.15745434e-01 -2.81778127e-01 1.54639328e+00
2.56124169e-01 -4.16310161e-01 -3.21322143e-01 9.57230985e-01
6.75369427e-02 5.13579130e-01 -7.59022653e-01 -1.95037827e-01
1.69693899e+00 -6.55645072e-01 -1.79752201e-01 1.76321678e-02
9.08203363e-01 -1.01695800e+00 6.32407069e-01 7.06305563e-01
-1.51671910e+00 -4.87561412e-02 -1.09476614e+00 -1.66948497e-01
3.01762465e-02 -8.96803439e-02 1.17727256e+00 6.63050532e-01
-1.30164576e+00 7.24666953e-01 -7.47150898e-01 1.59468561e-01
1.71569977e-02 5.17842472e-01 -5.36794007e-01 -2.81431913e-01
-8.31057608e-01 1.80806398e-01 -1.05985254e-01 2.99338371e-01
-6.09263837e-01 -6.88644946e-01 -6.25073671e-01 -1.99327976e-01
5.30977905e-01 -4.94636834e-01 9.74386990e-01 -2.28104413e-01
-1.24511504e+00 8.80011559e-01 -1.44002259e-01 -5.62175870e-01
2.36225471e-01 5.52751534e-02 -1.23659214e-02 4.19192463e-01
-5.74584231e-02 -1.52328774e-01 6.27423048e-01 -8.83247912e-01
-4.21848506e-01 -6.80765033e-01 5.35240591e-01 -2.62983620e-01
-1.67784691e-01 2.25907624e-01 -2.50452816e-01 -3.97074699e-01
7.54994988e-01 -1.29886889e+00 -4.88472164e-01 -2.17207402e-01
-4.56082642e-01 -6.82562515e-02 -1.47299811e-01 -5.12285173e-01
1.14359570e+00 -2.10254097e+00 3.40812773e-01 6.11232698e-01
9.37731743e-01 -7.71149322e-02 1.33658305e-01 6.74821854e-01
-2.00943187e-01 3.21407378e-01 -2.12427571e-01 -2.03509197e-01
3.34198534e-01 -4.28937236e-03 -4.28154081e-01 1.00146151e+00
-2.69622236e-01 2.25288033e-01 -6.93423986e-01 -1.29574254e-01
-3.32403809e-01 2.02375576e-01 -8.31315339e-01 -2.37270564e-01
1.75114982e-02 1.38123920e-02 -3.24668139e-01 6.06562078e-01
1.18042743e+00 -3.81450534e-01 5.59887707e-01 -6.23378038e-01
-3.00707787e-01 4.63260204e-01 -1.67373276e+00 1.54827058e+00
-2.67879546e-01 3.04150254e-01 8.16581309e-01 -9.78626192e-01
4.39372480e-01 7.23496005e-02 7.50853539e-01 -4.27703202e-01
2.56110430e-01 5.53981543e-01 1.04124822e-01 -1.39385268e-01
5.74852288e-01 -3.75325978e-01 -3.34150195e-01 9.59078968e-01
-2.32753709e-01 -8.73285756e-02 5.64310908e-01 6.25267088e-01
1.59946907e+00 -4.11912531e-01 -4.30294186e-01 -3.90307277e-01
3.47412348e-01 -1.77916229e-01 3.79085898e-01 4.09255952e-01
-4.52788882e-02 1.32834285e-01 9.98141885e-01 -6.08980894e-01
-1.29050684e+00 -1.07925940e+00 3.62918489e-02 1.43612170e+00
-1.49763510e-01 -8.26491833e-01 -5.73150694e-01 -3.04650575e-01
4.92435098e-02 5.64783253e-02 -5.66475570e-01 1.59226820e-01
-6.29601061e-01 -7.02173412e-01 7.69389212e-01 3.75216186e-01
2.73926944e-01 -5.26007712e-02 -1.31999806e-01 -1.17895417e-01
1.24903247e-01 -1.10066926e+00 -8.09159398e-01 3.97703350e-01
-1.10114574e+00 -1.01103187e+00 -3.76578927e-01 -4.91954148e-01
9.84601974e-01 6.47661746e-01 8.05176437e-01 7.34319463e-02
-1.05737597e-01 1.87473997e-01 -1.97632879e-01 1.20709926e-01
-6.21621870e-02 -1.77921176e-01 4.81259495e-01 8.92634410e-03
7.17410520e-02 -8.92626226e-01 -7.63621449e-01 1.45194426e-01
-1.15936792e+00 -1.21748880e-01 6.17709577e-01 6.71100020e-01
5.04678309e-01 2.51089752e-01 -5.22333622e-01 -6.94387853e-01
2.45326936e-01 -2.13502198e-01 -1.08265233e+00 -2.15426981e-01
-4.93658990e-01 6.04796290e-01 9.41763461e-01 -1.75080627e-01
-1.59292877e-01 8.89726281e-02 2.60999035e-02 -4.48357582e-01
7.83313572e-01 5.79622746e-01 2.77952969e-01 -3.07770431e-01
5.21533668e-01 2.62139827e-01 8.94541666e-02 -6.69985414e-01
6.59101605e-01 2.09852591e-01 5.39656222e-01 -1.04396212e+00
1.26952040e+00 6.62460387e-01 7.09022641e-01 -5.74159443e-01
-9.43783164e-01 -4.09169614e-01 -5.93336105e-01 2.88282573e-01
3.18212867e-01 -7.93245018e-01 -1.59400165e+00 1.25622466e-01
-8.91427100e-01 -1.42306387e-01 2.26978324e-02 4.22896415e-01
-3.96344692e-01 6.24132037e-01 -7.87396312e-01 -9.48953032e-01
-2.27576196e-01 -1.00783789e+00 7.81401813e-01 -5.48569620e-01
1.23881176e-01 -5.79088509e-01 -1.76638708e-01 6.54537857e-01
2.96319962e-01 -6.74850633e-03 9.70159352e-01 -3.76296759e-01
-1.14693570e+00 -2.95531631e-01 -6.29769325e-01 5.87751269e-01
-4.21649605e-01 -4.69936430e-02 -2.65253216e-01 -3.93945426e-01
-3.11094057e-02 -1.28779888e-01 6.95742786e-01 -3.19772243e-01
1.16214585e+00 -8.59646201e-01 2.44663656e-01 7.28229403e-01
1.40794396e+00 -4.68599439e-01 5.77327728e-01 -1.49246350e-01
7.99730480e-01 1.44571185e-01 2.38534600e-01 6.45213187e-01
6.59224093e-01 3.86962086e-01 3.62784088e-01 2.23100722e-01
1.55270264e-01 6.66751415e-02 4.69450802e-01 1.51559985e+00
-5.67152560e-01 6.32853627e-01 -9.04549778e-01 6.29968047e-01
-1.40094316e+00 -7.57577777e-01 -5.48933089e-01 2.52475762e+00
1.09876788e+00 2.77970910e-01 3.54981959e-01 4.74878937e-01
2.89328247e-01 1.10706612e-01 -6.74643889e-02 -7.84717858e-01
-6.82323128e-02 8.54315221e-01 1.43762946e+00 6.41309440e-01
-7.94935107e-01 6.67190611e-01 5.91088009e+00 6.92525148e-01
-9.23856974e-01 1.25569329e-01 1.53563172e-01 -2.90191233e-01
-4.94765073e-01 4.15427864e-01 -4.93542612e-01 2.38103390e-01
1.05598891e+00 -4.10709530e-01 1.00533986e+00 7.41195858e-01
-2.14446709e-01 -5.86934648e-02 -1.19383013e+00 1.17875516e+00
-5.04671074e-02 -1.20812964e+00 -2.72440851e-01 5.54718256e-01
8.03200006e-01 -1.19386248e-01 1.84361115e-01 -1.56052127e-01
7.81598628e-01 -1.02054524e+00 6.18753493e-01 6.64967597e-02
8.90709817e-01 -7.53874004e-01 4.26615119e-01 9.28805321e-02
-1.42147422e+00 -5.44162989e-02 -4.50258821e-01 -2.99710214e-01
-6.03277469e-04 1.09918678e+00 -6.65413857e-01 6.67793334e-01
7.72084832e-01 3.01849637e-02 -1.18384445e-02 4.96061593e-01
2.06582174e-01 6.23699605e-01 -8.40775907e-01 -1.27628118e-01
2.40337715e-01 -5.09099066e-01 4.49552923e-01 1.10114992e+00
3.86118889e-01 5.77745795e-01 3.06286573e-01 2.82458276e-01
-4.30499852e-01 1.67786956e-01 -2.60700792e-01 -3.11846286e-01
2.27617651e-01 1.23074913e+00 -6.14011288e-01 -2.66565263e-01
-2.32347071e-01 8.52647662e-01 1.77743420e-01 9.46511626e-02
-5.74019969e-01 -7.29474723e-01 9.14987147e-01 4.08008426e-01
7.92139471e-01 -1.13909721e+00 -5.15195668e-01 -1.29884386e+00
4.22959715e-01 -9.21608329e-01 1.97550863e-01 -6.81624562e-02
-1.26541698e+00 3.55533630e-01 -6.42747357e-02 -1.06658638e+00
6.60383403e-02 -8.93890798e-01 1.37433290e-01 8.90578628e-01
-9.07033920e-01 -8.14471364e-01 1.95379689e-01 6.50396168e-01
-4.39950764e-01 2.12847590e-01 8.37652683e-01 5.09523153e-01
-3.51738393e-01 7.31018186e-01 2.17650518e-01 2.89890766e-01
4.78621304e-01 -1.16557121e+00 2.74818033e-01 8.28259885e-01
-5.05284071e-02 9.68134344e-01 7.51318455e-01 -2.41055846e-01
-2.58266115e+00 -5.97476244e-01 8.83118093e-01 -2.47709259e-01
1.37406397e+00 -6.09524369e-01 -4.14014459e-01 7.99065113e-01
-1.50179118e-01 2.39835605e-01 8.52981746e-01 1.98430344e-01
-1.13303554e+00 -6.25034869e-01 -7.77367771e-01 6.72228158e-01
1.29732203e+00 -9.23810184e-01 7.09740967e-02 5.58231115e-01
6.36343241e-01 -5.23023367e-01 -1.43485367e+00 2.11563319e-01
5.47381401e-01 -8.88281882e-01 9.39553261e-01 -5.56496322e-01
1.67856261e-01 -5.76053977e-01 -7.08027720e-01 -5.66109657e-01
-2.96894103e-01 -1.34922576e+00 -3.15954566e-01 5.42367220e-01
6.04077518e-01 -5.64442754e-01 5.22588015e-01 4.98128653e-01
1.00712933e-01 -6.53055370e-01 -1.05457020e+00 -6.43476784e-01
3.13910276e-01 -7.53894389e-01 4.29501146e-01 8.43625128e-01
2.29970157e-01 2.91975021e-01 -4.17366534e-01 2.63521701e-01
7.82635689e-01 9.21252519e-02 9.31282461e-01 -9.58356142e-01
-7.03028440e-01 -4.81153429e-01 -5.30304313e-01 -1.28622067e+00
-3.45576078e-01 -1.17851460e+00 -3.54231328e-01 -1.13317955e+00
3.44943762e-01 -8.60243797e-01 -1.79857582e-01 5.35317779e-01
4.19692606e-01 4.85047489e-01 2.58253217e-01 8.41888189e-02
-7.97533333e-01 1.00409672e-01 1.18934441e+00 7.66573427e-03
3.32057267e-01 -1.59222022e-01 -1.04136586e+00 6.38708055e-01
2.94740230e-01 -3.62779945e-01 -6.64819106e-02 -6.05781734e-01
1.10058498e+00 4.87713754e-01 1.40106291e-01 -8.43737781e-01
4.03281778e-01 -3.63116972e-02 -3.15649748e-01 -6.43423259e-01
2.64984995e-01 -5.48800886e-01 6.93004355e-02 4.97868121e-01
-7.32680038e-02 4.23074961e-01 9.25838128e-02 3.44664127e-01
1.88415110e-01 -6.49752989e-02 5.32832146e-01 -1.58402137e-02
1.03704065e-01 5.89501083e-01 -1.54790804e-01 8.74781087e-02
5.84260345e-01 2.49450997e-01 -4.52081114e-01 -3.69032532e-01
-5.91663957e-01 -3.87564339e-02 3.30624908e-01 -3.30365807e-01
2.59673953e-01 -1.39117777e+00 -7.69029438e-01 -8.12955424e-02
-3.38968337e-01 3.16667169e-01 4.46885467e-01 1.40190935e+00
-1.00823283e+00 4.60263610e-01 8.74677151e-02 -3.93767327e-01
-1.01764321e+00 6.47138655e-01 -4.58707213e-02 -5.17339766e-01
-9.07436684e-02 1.08050287e+00 -2.94498116e-01 -1.44229010e-01
2.63242330e-03 -3.59314322e-01 9.84889984e-01 -8.04627463e-02
6.86007917e-01 6.29767060e-01 2.04531953e-01 -5.47949910e-01
-4.49968636e-01 5.33195436e-01 -1.48319870e-01 -4.11935240e-01
1.45474494e+00 -8.45211595e-02 -1.03901255e+00 3.53930146e-01
1.75448227e+00 5.96412122e-01 -5.92993677e-01 -3.64999175e-01
-2.53043622e-01 -2.90667057e-01 -2.78121978e-01 -2.18248084e-01
-1.07077551e+00 8.84440124e-01 1.36613876e-01 6.45216882e-01
1.26036239e+00 9.36056003e-02 1.08612370e+00 7.24923849e-01
7.87490010e-01 -8.93716037e-01 1.20376788e-01 7.38919497e-01
5.11299253e-01 -6.41740561e-01 3.20094258e-01 -5.88935494e-01
-1.75884336e-01 1.30928886e+00 -1.94507957e-01 -3.77587795e-01
1.06606698e+00 3.08503330e-01 -4.77367252e-01 -2.24206701e-01
-7.23834813e-01 -2.50151783e-01 2.44510829e-01 3.31071988e-02
4.77064431e-01 3.73389542e-01 -4.40308452e-01 4.70370889e-01
-8.00516009e-01 -2.50904709e-01 7.70266056e-01 9.64076340e-01
-4.96730536e-01 -1.37858760e+00 -5.49142599e-01 3.93375486e-01
-7.69359171e-01 -3.52878451e-01 -1.40211713e-02 3.04940552e-01
-1.99061036e-02 8.79162192e-01 -3.69806141e-02 -7.30177164e-01
4.83575687e-02 5.01792058e-02 8.03295791e-01 -3.96014065e-01
-2.25142315e-01 1.63407072e-01 2.31935903e-01 -6.51892245e-01
-7.31484890e-02 -6.82395160e-01 -1.35735106e+00 -1.40404069e+00
4.83749025e-02 8.33902508e-02 9.67908561e-01 7.59485662e-01
2.34983861e-01 2.09582206e-02 9.99967098e-01 -4.10864443e-01
-7.46499479e-01 -7.00578690e-01 -8.76171052e-01 1.95444182e-01
3.10531348e-01 -2.89785862e-01 -5.10106266e-01 1.04041800e-01] | [6.625377178192139, 4.783371925354004] |
e78f0b12-c1d6-4ceb-9ffc-546ab26d3601 | lightweight-real-time-semantic-segmentation | 2302.10484 | null | https://arxiv.org/abs/2302.10484v1 | https://arxiv.org/pdf/2302.10484v1.pdf | Lightweight Real-time Semantic Segmentation Network with Efficient Transformer and CNN | In the past decade, convolutional neural networks (CNNs) have shown prominence for semantic segmentation. Although CNN models have very impressive performance, the ability to capture global representation is still insufficient, which results in suboptimal results. Recently, Transformer achieved huge success in NLP tasks, demonstrating its advantages in modeling long-range dependency. Recently, Transformer has also attracted tremendous attention from computer vision researchers who reformulate the image processing tasks as a sequence-to-sequence prediction but resulted in deteriorating local feature details. In this work, we propose a lightweight real-time semantic segmentation network called LETNet. LETNet combines a U-shaped CNN with Transformer effectively in a capsule embedding style to compensate for respective deficiencies. Meanwhile, the elaborately designed Lightweight Dilated Bottleneck (LDB) module and Feature Enhancement (FE) module cultivate a positive impact on training from scratch simultaneously. Extensive experiments performed on challenging datasets demonstrate that LETNet achieves superior performances in accuracy and efficiency balance. Specifically, It only contains 0.95M parameters and 13.6G FLOPs but yields 72.8\% mIoU at 120 FPS on the Cityscapes test set and 70.5\% mIoU at 250 FPS on the CamVid test dataset using a single RTX 3090 GPU. The source code will be available at https://github.com/IVIPLab/LETNet. | ['Dong Yue', 'Jian Yang', 'Huimin Lu', 'Guangwei Gao', 'Juncheng Li', 'Guoan Xu'] | 2023-02-21 | null | null | null | null | ['real-time-semantic-segmentation'] | ['computer-vision'] | [-1.80663764e-02 -2.46464625e-01 -2.61999130e-01 -3.82276684e-01
-4.27791327e-01 -3.01357746e-01 2.49824330e-01 -1.45601884e-01
-6.58895671e-01 3.88645262e-01 -2.43020296e-01 -4.44902271e-01
2.08103716e-01 -8.85887921e-01 -7.39967942e-01 -6.41081452e-01
1.43911596e-02 -3.67993978e-03 4.64866132e-01 -1.31426632e-01
-9.55417678e-02 2.65901238e-01 -1.30001342e+00 7.50971511e-02
9.63527679e-01 1.21788836e+00 4.31979477e-01 3.73862594e-01
-2.04837680e-01 6.70621634e-01 -5.17032802e-01 -4.94330853e-01
2.87812412e-01 -1.47270530e-01 -5.86616158e-01 -1.82924524e-01
2.92975605e-01 -3.05827618e-01 -5.43190479e-01 1.05999994e+00
6.77667022e-01 -1.99448824e-01 9.19020176e-02 -1.19833553e+00
-4.69069391e-01 3.88613582e-01 -6.93017006e-01 2.32306346e-01
-1.68506607e-01 2.64602333e-01 8.87951434e-01 -6.89305902e-01
3.24429870e-01 1.06294441e+00 7.48346627e-01 4.80352312e-01
-8.29280615e-01 -9.05096769e-01 4.56509888e-02 1.96910858e-01
-1.40856004e+00 -1.01278186e-01 5.59269547e-01 -7.74749815e-02
9.69656944e-01 1.40435129e-01 7.59526432e-01 8.26793492e-01
1.39389202e-01 1.05624831e+00 8.29021871e-01 1.88820526e-01
-5.52866943e-02 -2.00903729e-01 6.66177943e-02 9.39559340e-01
1.95076138e-01 -3.09982579e-02 -3.28018337e-01 3.45173031e-01
9.36541677e-01 1.79460108e-01 -2.60925829e-01 8.83377492e-02
-9.29298937e-01 8.57664406e-01 8.68787408e-01 3.01596910e-01
-1.43535793e-01 3.06717008e-01 6.82002842e-01 1.59760207e-01
5.31679988e-01 -3.42234336e-02 -6.62185192e-01 -2.96429485e-01
-6.96812332e-01 5.60681373e-02 5.15793860e-01 1.05851424e+00
6.42501950e-01 1.57118887e-01 -7.27285519e-02 1.04030967e+00
-3.19303721e-02 4.90442634e-01 5.17708540e-01 -7.41166353e-01
5.08773327e-01 6.66893363e-01 -3.38411927e-01 -1.19198883e+00
-5.73938906e-01 -8.01091015e-01 -1.19906926e+00 -1.92580849e-01
3.65896255e-01 -1.76642016e-01 -9.33527827e-01 1.58744097e+00
3.93060058e-01 3.72991115e-01 -1.09088622e-01 1.14941692e+00
9.65052247e-01 8.43514264e-01 1.79522961e-01 3.21781278e-01
1.49173701e+00 -1.38104069e+00 -3.89550328e-01 -2.79874116e-01
8.14250112e-01 -8.16036046e-01 1.19801366e+00 4.34336841e-01
-9.22221541e-01 -7.87987351e-01 -1.02892375e+00 -2.62551546e-01
-2.26061523e-01 3.51563960e-01 6.69719398e-01 6.04131222e-01
-1.01135099e+00 5.00612497e-01 -9.56818581e-01 -1.71181560e-01
7.93215930e-01 4.76257771e-01 -1.36584230e-02 -1.58456787e-01
-1.09829295e+00 3.40601236e-01 4.90753829e-01 2.84740299e-01
-4.87413079e-01 -7.34061003e-01 -7.44304776e-01 1.53829172e-01
4.22933400e-01 -5.12155473e-01 1.12461376e+00 -9.57531273e-01
-1.47188723e+00 5.67414105e-01 8.85811225e-02 -5.64658582e-01
5.84769249e-01 -3.51677418e-01 -3.42290580e-01 2.63634801e-01
1.03608266e-01 9.59262609e-01 6.27586424e-01 -7.19049096e-01
-7.54125357e-01 -3.22639912e-01 -1.15596376e-01 5.70239276e-02
-5.27414262e-01 -7.77195469e-02 -9.64419961e-01 -8.40722978e-01
1.85321093e-01 -9.51150894e-01 -2.49952942e-01 2.85880174e-02
-2.71585166e-01 -1.98236272e-01 1.04676211e+00 -5.21738827e-01
1.13418829e+00 -2.10039401e+00 -3.67783487e-01 -9.64412466e-02
1.60539672e-01 8.21873248e-01 -1.34916037e-01 1.08742565e-01
6.81112707e-02 2.23386422e-01 -2.33088270e-01 -2.23742902e-01
-1.04904786e-01 2.22284555e-01 -2.19636112e-02 4.02524889e-01
2.26837322e-01 1.27029562e+00 -7.82528579e-01 -5.62594593e-01
2.92497635e-01 8.11332703e-01 -6.84033036e-01 2.53158733e-02
-8.72916058e-02 2.57682383e-01 -6.30792558e-01 8.33222091e-01
1.01919198e+00 -5.03991067e-01 5.60608786e-03 -2.80065209e-01
-1.65038809e-01 1.54416133e-02 -6.75909638e-01 1.82919586e+00
-4.26400989e-01 5.90053916e-01 8.53422731e-02 -1.22226465e+00
9.64509666e-01 1.54160693e-01 5.69722891e-01 -1.17870331e+00
4.26076084e-01 3.05169523e-01 -1.03051439e-01 -4.70914453e-01
4.14051652e-01 9.02263299e-02 -6.42256737e-02 1.06527917e-01
-1.34924695e-01 9.04804617e-02 1.83949560e-01 2.94776317e-02
1.07208526e+00 1.47039413e-01 -1.04335055e-01 -2.42821634e-01
4.98520434e-01 1.14889383e-01 7.47730494e-01 4.22481447e-01
-2.57299304e-01 6.47597730e-01 3.62922549e-01 -6.01358533e-01
-9.43804026e-01 -8.73531878e-01 -2.27526516e-01 8.34042907e-01
4.59496111e-01 -5.16773701e-01 -7.96382964e-01 -4.75745469e-01
-2.85227180e-01 2.08221883e-01 -2.41024405e-01 -1.48542270e-01
-7.30140865e-01 -9.14325774e-01 8.13499391e-01 7.46711969e-01
1.30914056e+00 -1.02844763e+00 -8.13495040e-01 3.08198720e-01
-1.74852148e-01 -1.24238169e+00 -5.26839852e-01 -1.17937475e-01
-9.66491580e-01 -9.23882663e-01 -8.64308417e-01 -1.14395094e+00
5.13613760e-01 4.52771157e-01 9.38459277e-01 2.65837669e-01
-4.83810157e-01 -2.24389017e-01 -4.56186086e-01 -3.13344449e-01
2.52181351e-01 4.44287270e-01 -3.86748165e-01 -1.98112622e-01
3.72949064e-01 -4.71013844e-01 -8.58537197e-01 4.74595577e-01
-1.01760292e+00 3.72778803e-01 5.75151265e-01 9.31184471e-01
5.31617284e-01 8.91804788e-03 4.45416719e-01 -6.65775478e-01
2.62143284e-01 -3.09481740e-01 -7.11038351e-01 -9.42551419e-02
-4.60635036e-01 -3.63931090e-01 9.12146509e-01 -3.05814862e-01
-9.64193642e-01 -2.21059974e-02 -6.11442268e-01 -4.55617338e-01
-3.36586349e-02 4.62948292e-01 -6.61382638e-03 8.71119872e-02
2.05043107e-01 3.95213991e-01 1.23423345e-01 -4.75751191e-01
-3.26146856e-02 6.21071577e-01 3.61966103e-01 -5.41579008e-01
5.31321347e-01 4.64832723e-01 -2.22883940e-01 -9.23368514e-01
-8.50281000e-01 -5.03930807e-01 -2.52317101e-01 -2.38732807e-02
8.69917929e-01 -1.20785129e+00 -9.95835602e-01 7.83857226e-01
-1.03059268e+00 -5.17316818e-01 -3.48310033e-03 2.67919689e-01
-2.13102132e-01 2.85423309e-01 -7.84472823e-01 -3.21394622e-01
-6.61182165e-01 -1.33136547e+00 1.06288433e+00 4.67601299e-01
1.76655650e-01 -8.88927460e-01 -5.37782729e-01 5.66944122e-01
5.53525031e-01 1.50018707e-01 5.56877136e-01 -3.31441283e-01
-7.57382452e-01 -1.74925596e-01 -7.76298463e-01 5.44947922e-01
-6.04429692e-02 -2.18104035e-01 -8.18320692e-01 -3.62791002e-01
-8.31870064e-02 -2.44883522e-01 9.79086995e-01 4.18763995e-01
1.55856383e+00 -4.00853455e-02 -2.77319908e-01 1.11025679e+00
1.70977712e+00 3.81433040e-01 6.19624197e-01 3.34528863e-01
8.94402146e-01 1.17067859e-01 6.74045563e-01 3.57355714e-01
3.55171323e-01 5.30370474e-01 5.18717349e-01 -4.64461803e-01
-2.15921655e-01 -1.30974114e-01 1.61781535e-01 9.73830342e-01
1.01925276e-01 -3.40604156e-01 -9.99270201e-01 4.56504077e-01
-1.81016052e+00 -6.39792740e-01 -2.98213214e-01 1.76589048e+00
6.73300982e-01 2.70714551e-01 -3.03103658e-03 8.51481687e-03
5.68249881e-01 2.64949530e-01 -6.27249420e-01 -3.50565583e-01
-8.74869302e-02 4.21911567e-01 7.39650249e-01 1.72903925e-01
-1.10332274e+00 1.26382172e+00 4.71073580e+00 1.24936414e+00
-1.47831869e+00 2.40019292e-01 1.04544473e+00 1.58484820e-02
5.09654395e-02 -2.77335167e-01 -7.81866670e-01 6.86285377e-01
6.93188906e-01 1.69877246e-01 5.77945560e-02 7.90342689e-01
2.47425795e-01 -8.61004442e-02 -3.62368047e-01 1.04567206e+00
-2.11840406e-01 -1.35980427e+00 -4.76278290e-02 1.03832759e-01
6.02892637e-01 3.94942462e-01 1.15614854e-01 3.38608980e-01
-4.87477109e-02 -1.12921965e+00 6.54372036e-01 -3.67934592e-02
8.25228691e-01 -9.88531053e-01 9.92538273e-01 3.79172385e-01
-1.42242372e+00 -4.61104400e-02 -6.01406097e-01 -1.54085964e-01
1.04068622e-01 7.08682895e-01 -7.48121679e-01 5.83179235e-01
9.46807086e-01 9.06334341e-01 -4.83401150e-01 9.82334435e-01
-1.94295228e-01 7.99924731e-01 -4.50337559e-01 -1.06094807e-01
7.35525072e-01 -3.93627554e-01 1.60956442e-01 1.20925570e+00
3.82641912e-01 1.92607805e-01 1.42230913e-01 6.47118092e-01
-1.55767173e-01 2.03389302e-01 -2.42355511e-01 1.11400150e-01
1.30561203e-01 1.37879419e+00 -1.01969349e+00 -2.90120631e-01
-5.46985209e-01 9.61475730e-01 2.32111678e-01 2.90116280e-01
-1.32626939e+00 -4.34020847e-01 8.25172067e-01 -6.95585608e-02
5.97288787e-01 -3.14046323e-01 -4.38567460e-01 -1.17331553e+00
8.82435963e-02 -6.39731228e-01 1.85323715e-01 -5.77076137e-01
-8.78119051e-01 7.00814068e-01 -5.20979285e-01 -1.12401319e+00
5.08747876e-01 -7.32056439e-01 -5.29109418e-01 6.97358489e-01
-1.72620475e+00 -1.12678611e+00 -5.83914280e-01 5.50626159e-01
7.62435615e-01 1.47650480e-01 3.94098550e-01 7.37973034e-01
-1.04353249e+00 7.06321180e-01 1.06851541e-01 4.03218180e-01
4.70390886e-01 -9.47225869e-01 5.89352250e-01 6.61470652e-01
-2.39523619e-01 4.14720893e-01 3.44769746e-01 -4.17668611e-01
-1.31183016e+00 -1.50018263e+00 7.17849016e-01 1.57399461e-01
5.05239367e-01 -3.92122537e-01 -8.79433095e-01 4.19728607e-01
1.47732764e-01 2.91155994e-01 2.58339852e-01 -2.33631343e-01
-2.04979509e-01 -2.57413447e-01 -9.20764089e-01 6.39639378e-01
1.16940749e+00 -1.68996572e-01 3.35429274e-02 2.77911305e-01
8.82143497e-01 -5.81943631e-01 -8.27476561e-01 4.62359220e-01
4.09809530e-01 -1.02411997e+00 1.06502247e+00 -5.91480844e-02
6.46332681e-01 -2.17964575e-01 3.94913889e-02 -9.18064952e-01
-1.74956679e-01 -4.56522435e-01 2.25951254e-01 1.01119161e+00
1.45076752e-01 -8.32438290e-01 1.07445347e+00 1.84126437e-01
-3.46429348e-01 -1.32055855e+00 -8.54382634e-01 -8.68399680e-01
1.00176679e-02 -5.24909735e-01 6.33166373e-01 6.78160071e-01
-4.18213218e-01 4.10003215e-01 -2.14404970e-01 -1.38424663e-02
4.18666959e-01 1.33366555e-01 6.96517646e-01 -9.74151373e-01
4.07825410e-03 -5.17589629e-01 -3.89549315e-01 -1.44775724e+00
2.84289494e-02 -8.93585861e-01 -9.95783657e-02 -1.28177702e+00
6.48209900e-02 -6.68285191e-01 -2.31693119e-01 6.67351246e-01
-1.53632730e-01 5.97353160e-01 2.34470457e-01 4.01826613e-02
-6.46543503e-01 6.80655897e-01 1.57192504e+00 -1.45727307e-01
-1.92767158e-02 -1.35382310e-01 -5.92266619e-01 6.25304341e-01
1.29249108e+00 -2.81893581e-01 -5.57677567e-01 -8.25725019e-01
3.81484330e-02 -4.02746350e-02 5.05704284e-01 -1.20484531e+00
2.42660135e-01 2.47159183e-01 2.84499079e-01 -6.42357230e-01
3.59017581e-01 -6.73376620e-01 1.10836653e-02 7.84906805e-01
7.51535296e-02 1.58961922e-01 4.31654096e-01 2.50309527e-01
-5.10004163e-01 3.68274152e-02 7.51967549e-01 -6.00054264e-02
-9.63041484e-01 8.17582846e-01 -6.87023401e-02 1.09788552e-01
9.47379231e-01 -3.78842890e-01 -2.61511147e-01 -7.75767714e-02
-2.96518147e-01 3.86263430e-01 3.95648420e-01 4.74596113e-01
6.76667869e-01 -1.14659917e+00 -5.56426048e-01 2.70644248e-01
-1.60769910e-01 3.89049768e-01 5.47097147e-01 9.16135311e-01
-9.38874483e-01 6.24587715e-01 -8.63071531e-02 -7.78784513e-01
-1.18090963e+00 2.75920421e-01 2.19430566e-01 -2.23006666e-01
-8.13670099e-01 1.04290915e+00 4.08862770e-01 -4.22901958e-01
1.30708843e-01 -3.71555805e-01 4.86633517e-02 1.33835245e-02
3.52741539e-01 3.44114214e-01 1.16137169e-01 -5.46357393e-01
-3.13424826e-01 6.20548904e-01 -1.31437913e-01 3.20315093e-01
1.44584167e+00 -2.24758983e-02 1.18241869e-02 -1.59368277e-01
1.59091246e+00 -5.15539229e-01 -1.49636114e+00 -2.17442364e-01
-2.45508432e-01 -4.40037906e-01 2.68232912e-01 -7.01647580e-01
-1.67454314e+00 1.03632569e+00 6.00014448e-01 -1.48064252e-02
1.37980807e+00 -1.88847378e-01 1.55927551e+00 8.58280584e-02
3.79166543e-01 -1.04026282e+00 1.05160400e-01 7.22408533e-01
4.54061449e-01 -1.27517390e+00 -1.05424576e-01 -6.93790615e-01
-4.79496300e-01 9.99721110e-01 7.51560390e-01 -2.28202879e-01
6.00471437e-01 2.53065944e-01 1.06757104e-01 -2.21335679e-01
-5.02534926e-01 -1.52526259e-01 -5.23512959e-02 2.42867902e-01
5.23099065e-01 5.93755469e-02 -3.76950413e-01 5.69886506e-01
-3.22481960e-01 4.39768471e-02 2.29389772e-01 9.09460008e-01
-2.63416290e-01 -9.67602491e-01 1.49803385e-02 4.29035693e-01
-6.32373393e-01 -2.38520890e-01 2.65613288e-01 9.27443862e-01
3.53729695e-01 7.54927814e-01 2.98993915e-01 -4.82362479e-01
2.06735790e-01 -3.72140139e-01 1.42826751e-01 -3.13861698e-01
-7.43211985e-01 2.14815721e-01 -1.34847999e-01 -7.61414766e-01
-3.17234129e-01 -4.99495149e-01 -1.44702089e+00 -4.85505790e-01
-1.65862352e-01 7.63771981e-02 6.48106158e-01 6.30534232e-01
4.90600318e-01 7.70498276e-01 4.16529477e-01 -5.98784864e-01
-2.71281064e-01 -7.39127338e-01 -4.06531185e-01 1.39970541e-01
1.01794146e-01 -3.29516351e-01 -2.49491725e-02 -3.50238569e-02] | [9.32580280303955, -0.41749945282936096] |
89840b12-6e25-4c69-8894-85c99da285f2 | computing-a-partition-function-of-a | 2305.17526 | null | https://arxiv.org/abs/2305.17526v1 | https://arxiv.org/pdf/2305.17526v1.pdf | Computing a partition function of a generalized pattern-based energy over a semiring | Valued constraint satisfaction problems with ordered variables (VCSPO) are a special case of Valued CSPs in which variables are totally ordered and soft constraints are imposed on tuples of variables that do not violate the order. We study a restriction of VCSPO, in which soft constraints are imposed on a segment of adjacent variables and a constraint language $\Gamma$ consists of $\{0,1\}$-valued characteristic functions of predicates. This kind of potentials generalizes the so-called pattern-based potentials, which were applied in many tasks of structured prediction. For a constraint language $\Gamma$ we introduce a closure operator, $ \overline{\Gamma^{\cap}}\supseteq \Gamma$, and give examples of constraint languages for which $|\overline{\Gamma^{\cap}}|$ is small. If all predicates in $\Gamma$ are cartesian products, we show that the minimization of a generalized pattern-based potential (or, the computation of its partition function) can be made in ${\mathcal O}(|V|\cdot |D|^2 \cdot |\overline{\Gamma^{\cap}}|^2 )$ time, where $V$ is a set of variables, $D$ is a domain set. If, additionally, only non-positive weights of constraints are allowed, the complexity of the minimization task drops to ${\mathcal O}(|V|\cdot |\overline{\Gamma^{\cap}}| \cdot |D| \cdot \max_{\rho\in \Gamma}\|\rho\|^2 )$ where $\|\rho\|$ is the arity of $\rho\in \Gamma$. For a general language $\Gamma$ and non-positive weights, the minimization task can be carried out in ${\mathcal O}(|V|\cdot |\overline{\Gamma^{\cap}}|^2)$ time. We argue that in many natural cases $\overline{\Gamma^{\cap}}$ is of moderate size, though in the worst case $|\overline{\Gamma^{\cap}}|$ can blow up and depend exponentially on $\max_{\rho\in \Gamma}\|\rho\|$. | ['Rustem Takhanov'] | 2023-05-27 | null | null | null | null | ['structured-prediction'] | ['methodology'] | [ 5.04321635e-01 6.17607415e-01 -1.38101116e-01 -1.64995119e-01
-4.98335958e-01 -8.08832586e-01 -7.90491924e-02 4.43100005e-01
-5.58060884e-01 9.72429991e-01 -7.46034801e-01 -6.58420920e-01
-7.31456339e-01 -1.27353239e+00 -5.77094316e-01 -6.70489788e-01
-7.15763867e-01 7.98824131e-01 5.10001719e-01 -5.40685773e-01
9.63862911e-02 2.07453728e-01 -2.03253961e+00 -6.00567833e-02
9.55098450e-01 1.59124267e+00 1.89069107e-01 1.80757031e-01
-3.96128416e-01 1.49531960e-01 -7.30108798e-01 -2.26869449e-01
3.81037027e-01 -5.92413008e-01 -8.32005441e-01 4.49570827e-02
-2.78376192e-01 6.32673383e-01 3.23006570e-01 1.47650182e+00
-9.02828649e-02 1.78856835e-01 6.26990557e-01 -1.43682730e+00
-1.84667572e-01 4.08760041e-01 -5.47929764e-01 -1.10783324e-01
4.75388229e-01 1.16745792e-01 1.35031509e+00 -4.83830899e-01
7.60264337e-01 7.01319098e-01 2.32000947e-01 2.86526382e-01
-1.27467501e+00 -7.99706936e-01 4.97986943e-01 -5.12696743e-01
-1.54237401e+00 3.38235289e-01 3.48199785e-01 -4.58309799e-01
1.20509171e+00 7.88879573e-01 5.46571255e-01 -3.10424920e-02
8.48829001e-02 -5.20748161e-02 1.13343823e+00 -7.57265568e-01
5.88868856e-01 -1.26372522e-03 3.19959044e-01 7.66749322e-01
2.72900850e-01 1.82332069e-01 -2.67614335e-01 -2.70330220e-01
6.01102829e-01 -4.29320008e-01 -2.04399869e-01 -2.61507779e-02
-7.26091385e-01 1.31462598e+00 -2.07683474e-01 4.74513739e-01
2.34655321e-01 6.89940378e-02 1.36779040e-01 4.62795377e-01
5.30610979e-02 6.04043603e-01 -6.71572268e-01 -5.10137230e-02
-6.27823472e-01 5.27313530e-01 8.83335710e-01 1.61261189e+00
8.97245109e-01 -3.69252898e-02 2.15382949e-01 7.32834935e-01
-1.48717910e-01 6.86821401e-01 -3.43759060e-01 -1.17059433e+00
5.99524617e-01 1.02110064e+00 2.44294867e-01 -2.12707520e-01
-6.79075181e-01 -5.47108091e-02 -7.18516886e-01 6.29724503e-01
7.12998092e-01 -1.50501862e-01 -8.91610444e-01 2.08202648e+00
-5.73071279e-02 -7.34841704e-01 -2.23284975e-01 5.56151271e-01
3.24541271e-01 7.86578059e-01 -4.91525978e-02 -1.10635030e+00
1.46398687e+00 -2.36730486e-01 -2.03513667e-01 -2.84807533e-01
7.24742830e-01 -6.23446226e-01 1.14230466e+00 7.32636809e-01
-1.63060880e+00 -5.24786413e-02 -1.22881901e+00 3.29907566e-01
-4.80641186e-01 -7.82330871e-01 9.30545092e-01 7.63975382e-01
-7.31232524e-01 4.48451251e-01 -2.79704034e-01 3.45996320e-01
1.00249328e-01 1.01788390e+00 -5.05548716e-01 -2.87558973e-01
-1.21094906e+00 6.84870124e-01 3.01071376e-01 -1.51351228e-01
-3.06909114e-01 -3.73738587e-01 -1.07161021e+00 9.38368738e-02
9.02269602e-01 9.97688323e-02 7.99376369e-01 -5.35840571e-01
-8.02804828e-01 8.54966223e-01 -1.05366178e-01 -2.76673157e-02
2.48164803e-01 7.60038674e-01 -4.62437212e-01 -1.51397467e-01
2.87269384e-01 2.30647817e-01 4.93595451e-02 -1.08664393e+00
-8.02504838e-01 -6.23125672e-01 5.51358879e-01 -1.49227992e-01
3.63941133e-01 4.10978019e-01 -5.02183795e-01 -2.67983913e-01
6.32133484e-01 -1.07412875e+00 -5.91158628e-01 -4.60519522e-01
-2.39374667e-01 -3.70012522e-01 2.49314010e-01 1.44437477e-01
1.62549174e+00 -1.95583427e+00 3.00072283e-01 1.05774140e+00
2.42106318e-01 -3.16458195e-02 4.41984147e-01 3.85813981e-01
-1.90036491e-01 4.57149982e-01 -9.99908149e-01 2.67418474e-01
2.00287223e-01 6.33756399e-01 3.01248003e-02 1.90374956e-01
-5.43477200e-02 -6.14097342e-02 -5.21410227e-01 -1.87720582e-01
-1.77621946e-01 9.31347385e-02 -7.41273999e-01 -1.00580454e-01
-1.00867808e+00 1.94863304e-01 -4.88913029e-01 5.68598568e-01
9.14745092e-01 -9.30550620e-02 4.78878975e-01 5.83288550e-01
-4.52639401e-01 1.38862431e-01 -2.00494623e+00 1.32597613e+00
-3.97959799e-02 -1.35450989e-01 4.99158919e-01 -1.04034626e+00
1.05721259e+00 -3.27550131e-03 7.65800357e-01 -9.67573822e-01
3.09308976e-01 4.62626964e-01 1.04251415e-01 1.04919747e-01
1.90988094e-01 -6.71147704e-01 -8.04942966e-01 5.40297449e-01
-3.77883822e-01 -7.40730107e-01 6.10960603e-01 -2.08864990e-03
1.37865305e+00 -6.00343980e-02 1.51219875e-01 -5.75718582e-01
6.96603715e-01 2.17893630e-01 1.18596137e+00 6.13848567e-01
1.91146478e-01 3.70611638e-01 1.56454802e+00 -3.18376869e-01
-8.76306057e-01 -9.61627066e-01 -6.11498773e-01 1.25197470e+00
4.24272925e-01 -8.25624824e-01 -8.15257192e-01 -4.19032782e-01
-1.77239999e-01 6.59543216e-01 -4.41069305e-01 2.34260932e-01
-9.34966803e-01 -9.39806521e-01 3.37114662e-01 6.38411522e-01
2.12448955e-01 -1.41682911e+00 -8.75883639e-01 3.84861171e-01
1.82949916e-01 -8.08763385e-01 -1.37998328e-01 1.16280413e+00
-5.26590407e-01 -1.00685108e+00 -1.39723405e-01 -7.30516493e-01
1.01587665e+00 -6.09710276e-01 1.08068776e+00 4.27263439e-01
-4.46629077e-01 -2.89700568e-01 -3.83748561e-01 -6.77505970e-01
7.05844313e-02 -4.17367458e-01 3.90105471e-02 -7.68317282e-01
2.80469686e-01 -7.17619359e-01 -2.51039118e-01 5.40735364e-01
-1.19337833e+00 -3.27774853e-01 -1.23182751e-01 9.31562006e-01
1.30817330e+00 3.73896360e-01 2.99275666e-01 -1.22584772e+00
2.66973019e-01 -9.13538709e-02 -1.43051100e+00 -3.53760347e-02
-7.82828450e-01 2.28750840e-01 7.95098424e-01 -2.39504158e-01
-2.03034267e-01 -7.73750916e-02 -8.28630775e-02 -5.83298504e-02
2.11519748e-01 9.01582837e-01 -4.81454670e-01 2.64081866e-01
6.03851557e-01 -3.07875220e-02 -3.15797508e-01 -3.09182227e-01
2.85211820e-02 1.36037380e-01 4.05197412e-01 -9.75156724e-01
4.32430953e-01 1.83955699e-01 6.50404990e-01 -1.97117954e-01
-1.90765798e-01 3.12541693e-01 -3.15453947e-01 4.26738292e-01
6.80919468e-01 -3.45837712e-01 -1.35474038e+00 -3.57179254e-01
-6.14809334e-01 -3.60260487e-01 -7.27343202e-01 1.76294416e-01
-7.88432419e-01 1.13316312e-01 3.57399285e-02 -1.12056136e+00
-4.82432470e-02 -1.47319055e+00 4.88990396e-01 1.68706566e-01
-3.80738407e-01 -4.47764158e-01 -5.70291996e-01 1.82416990e-01
1.93458870e-01 5.41496813e-01 1.73811436e+00 -7.55652070e-01
-5.11565745e-01 -4.34944272e-01 -1.59892142e-01 5.31833112e-01
-3.77568424e-01 -1.84175566e-01 -2.49343380e-01 -1.54298797e-01
3.28385420e-02 2.14696586e-01 3.05916071e-01 9.24546495e-02
1.16881728e+00 -2.47992724e-01 -3.99203867e-01 1.96242288e-01
1.62780166e+00 7.36519873e-01 6.49330795e-01 6.95974156e-02
-1.90226492e-02 6.67941749e-01 8.15069377e-01 7.89859772e-01
-1.94443297e-02 7.83055127e-01 6.42378926e-01 2.15572909e-01
4.59652573e-01 5.98944962e-01 1.14360731e-03 6.07960001e-02
-3.83540481e-01 -3.52461904e-01 -1.13565624e+00 6.06859326e-01
-1.36635983e+00 -6.26233041e-01 -2.06218839e-01 2.73373890e+00
9.78503883e-01 8.73147845e-01 1.02420077e-01 5.46156585e-01
6.59660578e-01 -4.88834567e-02 -1.06252581e-01 -1.26969385e+00
-2.81381249e-01 1.52295327e+00 7.21570432e-01 8.38508904e-01
-6.49832904e-01 8.70426655e-01 3.59804130e+00 8.55385482e-01
-7.88057506e-01 -8.12124088e-03 2.19076440e-01 -5.15072227e-01
-4.64944452e-01 5.27895033e-01 -7.51529753e-01 8.99886549e-01
6.58513963e-01 -1.27168328e-01 4.16954070e-01 8.59774888e-01
-2.00063542e-01 -8.21847439e-01 -1.32924378e+00 6.29448712e-01
-2.21475050e-01 -1.11569655e+00 -8.85836601e-01 4.47328150e-01
6.90713644e-01 -4.53696191e-01 5.82859293e-02 2.96776682e-01
5.54786980e-01 -1.43233585e+00 7.15317011e-01 -3.12136531e-01
1.39073682e+00 -1.08821905e+00 4.79379207e-01 4.54922885e-01
-1.51195502e+00 -2.99054444e-01 -8.77005532e-02 -4.16866571e-01
4.07455057e-01 5.49721301e-01 -1.72607839e-01 6.50451779e-01
8.13868582e-01 -5.04680395e-01 4.04738337e-01 5.96643031e-01
-1.22150883e-01 -4.46468405e-02 -7.07805276e-01 2.12918893e-01
3.28095168e-01 -4.09157634e-01 4.56172645e-01 8.43885124e-01
1.73184872e-01 8.65150869e-01 5.82722545e-01 7.92120636e-01
1.30711822e-02 -4.18693908e-02 -3.75141889e-01 4.00830835e-01
4.17780936e-01 4.57237184e-01 -9.81090546e-01 -6.91931173e-02
-3.24457020e-01 1.34184510e-01 -1.21199209e-02 2.94667572e-01
-9.92459118e-01 -8.77502739e-01 7.53293157e-01 6.17592454e-01
4.41352189e-01 -2.20965281e-01 -7.46082723e-01 -5.43634653e-01
5.47478735e-01 -7.28975296e-01 7.41325319e-01 -3.60471845e-01
-5.76551139e-01 7.75701344e-01 4.65584368e-01 -1.04773808e+00
-2.20532879e-01 -6.18151724e-01 -2.79951870e-01 1.28108430e+00
-8.93472016e-01 -4.52791601e-01 1.03431523e-01 8.56788099e-01
-1.24659881e-01 3.95659804e-02 1.23076057e+00 2.97794789e-02
-1.67434335e-01 6.48954570e-01 -1.84322789e-01 -2.01314315e-01
6.64570276e-03 -8.57464790e-01 -5.29192209e-01 5.14104962e-01
-4.62162405e-01 5.93789399e-01 8.56581569e-01 -5.80001712e-01
-1.19430351e+00 -5.99886000e-01 1.12374890e+00 -8.17829221e-02
1.07575499e-01 -6.23058677e-01 -8.40245187e-01 8.75395060e-01
-9.68212113e-02 3.04644018e-01 6.34180188e-01 2.91061085e-02
-5.06508231e-01 -1.26155317e-01 -1.70882523e+00 4.88003939e-01
1.19434738e+00 -1.37854293e-01 -3.22178036e-01 5.08224666e-01
5.58292985e-01 -5.61661720e-01 -1.37870681e+00 8.56877983e-01
3.26822281e-01 -9.64603186e-01 4.20002788e-01 -2.46015966e-01
-6.14442043e-02 -3.98752689e-01 -5.56125224e-01 -4.43067521e-01
-5.46560921e-02 -8.28763783e-01 2.93447882e-01 7.62279570e-01
8.95262361e-01 -8.26039076e-01 6.96537495e-01 1.19134963e+00
-5.47698498e-01 -1.13649464e+00 -1.63629854e+00 -8.91015232e-01
6.77008510e-01 -6.33606315e-01 6.33129120e-01 7.21915305e-01
6.77423954e-01 -2.58315355e-01 1.75933033e-01 -9.70457494e-02
1.50345564e-01 6.42669052e-02 -7.50843212e-02 -1.40951240e+00
-7.03243434e-01 -5.34353375e-01 -1.91149116e-01 -3.23863685e-01
-2.25088120e-01 -8.62658858e-01 -1.19827569e-01 -1.24431467e+00
-3.29818614e-02 -1.32832265e+00 -9.80194211e-02 9.86599743e-01
7.47700334e-01 1.30434614e-02 2.78249323e-01 -1.76502153e-01
-1.80587873e-01 -2.57909864e-01 1.28789103e+00 -1.06300563e-01
-3.80861044e-01 -1.56466112e-01 -7.99463034e-01 7.54577339e-01
4.18486655e-01 -5.71738660e-01 -3.46613258e-01 -2.72131294e-01
9.67414856e-01 7.73717642e-01 -1.16235919e-01 -6.71375930e-01
8.34335834e-02 -7.80466497e-01 -1.39659598e-01 -3.05473596e-01
6.54766917e-01 -6.34705961e-01 4.80511487e-01 2.89352864e-01
-1.20708654e-02 3.97670209e-01 3.36045533e-01 7.06952363e-02
-1.76012024e-01 -3.86449993e-01 8.52658629e-01 -3.72629315e-01
-2.29892656e-01 -2.46698949e-02 -3.22946250e-01 -5.36793619e-02
1.42982137e+00 -4.85611647e-01 -4.56087917e-01 -1.64121781e-02
-1.02108848e+00 4.31529284e-01 4.88891751e-01 -6.60362542e-02
2.72334129e-01 -9.73736763e-01 -2.43910030e-01 3.91455084e-01
1.52932033e-01 9.97391105e-01 1.00814663e-01 7.82575607e-01
-4.90709484e-01 4.72592503e-01 2.96216388e-03 -3.87366146e-01
-9.15606737e-01 7.14593351e-01 1.21291712e-01 -4.91384208e-01
-2.28297055e-01 1.35447037e+00 -3.50911617e-02 7.45674521e-02
1.90444082e-01 -6.49132967e-01 2.50536174e-01 4.41844836e-02
1.54336572e-01 1.67288184e-01 3.12122911e-01 -5.11866748e-01
-7.61955380e-01 4.72350717e-01 2.22832207e-02 -4.92492199e-01
1.23593783e+00 4.00256276e-01 -6.24338090e-01 -1.31671742e-01
8.27706635e-01 1.82723254e-01 -7.54855216e-01 1.95913628e-01
-1.69296801e-01 -1.26365766e-01 -7.40570784e-01 -9.76705551e-01
-8.57964337e-01 5.45856833e-01 3.89704436e-01 4.03557390e-01
1.42795956e+00 3.43113482e-01 3.30516547e-01 -4.94197011e-02
1.11457980e+00 -1.61971974e+00 -3.33733380e-01 9.02098358e-01
6.02771640e-01 -7.56752908e-01 1.62058026e-01 -8.72673690e-01
-6.72285736e-01 6.66991591e-01 6.12540543e-01 -1.51151642e-01
7.02821314e-01 5.76130986e-01 -4.30683672e-01 -2.55886167e-01
-4.27312672e-01 -3.31789851e-01 -1.69645160e-01 1.62580416e-01
4.30562735e-01 2.39906222e-01 -1.12104094e+00 1.37005162e+00
-6.72080934e-01 -2.83701479e-01 6.38214111e-01 1.56486094e+00
-6.21402621e-01 -1.93487251e+00 -3.61980140e-01 4.77117330e-01
-4.18207288e-01 -1.74286485e-01 -9.35121672e-04 9.41398740e-01
1.08311892e+00 1.14665759e+00 1.80367470e-01 -1.59624621e-01
4.14593995e-01 3.84957135e-01 6.65182412e-01 -9.69546080e-01
-9.23559666e-01 3.99708569e-01 3.55902344e-01 -3.03691804e-01
1.13119714e-01 -5.75749159e-01 -2.27688146e+00 -2.87381381e-01
-2.82474697e-01 4.73183990e-01 4.37624812e-01 9.53769743e-01
6.68997550e-03 2.00268358e-01 2.80802190e-01 -9.69096273e-02
-7.76963308e-02 -5.71661651e-01 -1.06043577e+00 2.05630079e-01
-4.09997970e-01 -1.02841485e+00 -3.87361795e-01 -3.03455651e-01] | [6.458448886871338, 4.632206439971924] |
68c122d5-2180-4da6-840f-8bea2d28394f | linguistically-driven-multi-task-pre-training | 2201.08070 | null | https://arxiv.org/abs/2201.08070v1 | https://arxiv.org/pdf/2201.08070v1.pdf | Linguistically-driven Multi-task Pre-training for Low-resource Neural Machine Translation | In the present study, we propose novel sequence-to-sequence pre-training objectives for low-resource machine translation (NMT): Japanese-specific sequence to sequence (JASS) for language pairs involving Japanese as the source or target language, and English-specific sequence to sequence (ENSS) for language pairs involving English. JASS focuses on masking and reordering Japanese linguistic units known as bunsetsu, whereas ENSS is proposed based on phrase structure masking and reordering tasks. Experiments on ASPEC Japanese--English & Japanese--Chinese, Wikipedia Japanese--Chinese, News English--Korean corpora demonstrate that JASS and ENSS outperform MASS and other existing language-agnostic pre-training methods by up to +2.9 BLEU points for the Japanese--English tasks, up to +7.0 BLEU points for the Japanese--Chinese tasks and up to +1.3 BLEU points for English--Korean tasks. Empirical analysis, which focuses on the relationship between individual parts in JASS and ENSS, reveals the complementary nature of the subtasks of JASS and ENSS. Adequacy evaluation using LASER, human evaluation, and case studies reveals that our proposed methods significantly outperform pre-training methods without injected linguistic knowledge and they have a larger positive impact on the adequacy as compared to the fluency. We release codes here: https://github.com/Mao-KU/JASS/tree/master/linguistically-driven-pretraining. | ['Sadao Kurohashi', 'Chenhui Chu', 'Zhuoyuan Mao'] | 2022-01-20 | null | null | null | null | ['low-resource-neural-machine-translation'] | ['natural-language-processing'] | [ 1.11318618e-01 -2.55948484e-01 -3.91801029e-01 -3.52175772e-01
-1.27368224e+00 -6.59266233e-01 4.10531968e-01 -2.97520697e-01
-6.29526138e-01 1.24492574e+00 4.60517615e-01 -1.01949382e+00
4.00823921e-01 -2.00860590e-01 -7.83501744e-01 -3.28334212e-01
3.25282156e-01 6.49013400e-01 -6.61385879e-02 -6.77220523e-01
2.50135720e-01 -1.42628133e-01 -7.49699175e-01 4.35501635e-01
1.33222234e+00 3.07218224e-01 7.77443528e-01 6.45686567e-01
-8.61859694e-02 6.21382535e-01 -4.10359174e-01 -4.86529917e-01
3.47464442e-01 -6.20794356e-01 -9.92877364e-01 -2.31965199e-01
8.69191960e-02 -1.90452844e-01 1.17676713e-01 1.08880556e+00
6.48555756e-01 6.41556382e-02 5.19939423e-01 -7.10341930e-01
-1.17463815e+00 1.09039867e+00 -6.15162432e-01 4.57949549e-01
4.64329213e-01 2.01907754e-01 1.02629268e+00 -1.23825538e+00
5.68258762e-01 1.06446707e+00 7.25185275e-01 8.38864207e-01
-1.01938081e+00 -6.52588546e-01 -7.33525157e-02 -5.29505592e-03
-1.20442796e+00 -5.38441300e-01 2.27958798e-01 -2.74114311e-01
1.54453397e+00 3.75461489e-01 1.69871807e-01 1.16054642e+00
3.37394983e-01 8.26624513e-01 1.37544358e+00 -8.89986932e-01
-1.01454042e-01 2.03745440e-01 1.54066205e-01 6.01074219e-01
9.69191715e-02 2.54615396e-01 -4.36002046e-01 2.78612703e-01
6.90613031e-01 -5.84616661e-01 -2.76070416e-01 5.00954866e-01
-1.67093110e+00 6.37991726e-01 -8.56011212e-02 6.67961955e-01
-3.22768271e-01 -1.73405707e-01 6.16779983e-01 7.38776505e-01
4.98925984e-01 6.27156436e-01 -9.81546879e-01 -4.25240785e-01
-8.31484795e-01 -1.27371207e-01 7.37907529e-01 1.47604084e+00
6.80398524e-01 3.21775526e-01 -3.09840709e-01 9.41271961e-01
-4.63264473e-02 8.58329952e-01 9.24471498e-01 -6.29572809e-01
9.18939769e-01 2.08806694e-01 1.49006382e-01 -2.97771305e-01
-2.42804706e-01 -4.27583426e-01 -7.53341019e-01 -4.31520849e-01
1.87916026e-01 -4.30128068e-01 -8.96357954e-01 1.97280490e+00
7.84007162e-02 -2.11291015e-01 3.60783815e-01 9.12110269e-01
7.05032706e-01 9.62418258e-01 4.32557799e-02 -5.76530218e-01
1.49234593e+00 -1.46091020e+00 -7.43052006e-01 -4.38287109e-01
1.09851658e+00 -1.16892040e+00 1.86396217e+00 -1.55956969e-01
-1.48845875e+00 -7.55077600e-01 -7.40143716e-01 -1.18623540e-01
-2.11893141e-01 5.11586070e-01 2.53600657e-01 6.99137509e-01
-1.24118412e+00 3.79286140e-01 -6.53169513e-01 -5.44560730e-01
-4.21380252e-01 3.25171590e-01 -3.13191712e-01 -3.97027843e-02
-1.57139921e+00 1.21999991e+00 6.70154810e-01 6.21938184e-02
-6.59421802e-01 -6.66288614e-01 -8.31604064e-01 -1.77599147e-01
2.60378003e-01 -7.89605379e-01 1.54429972e+00 -1.11578488e+00
-1.80655992e+00 7.56490231e-01 -3.15408528e-01 -3.31397176e-01
2.46813625e-01 -4.23833251e-01 -4.86425102e-01 -1.84460849e-01
4.71244097e-01 6.47564530e-01 2.94783801e-01 -9.81035173e-01
-8.82806361e-01 1.02596715e-01 -2.22638234e-01 7.80846596e-01
-3.71696204e-01 6.12076998e-01 -4.73317593e-01 -9.30117726e-01
-2.23609433e-01 -1.07811928e+00 -1.24798417e-01 -1.04198050e+00
-5.30131698e-01 -1.70610830e-01 2.67248064e-01 -1.35432994e+00
1.67886162e+00 -1.77602899e+00 3.93621236e-01 -2.80744165e-01
-3.90833944e-01 3.00415695e-01 -5.37196457e-01 6.69521451e-01
-4.06013280e-02 4.18856144e-01 -3.93831968e-01 -2.77442604e-01
-6.78298622e-02 2.09752485e-01 -3.74109223e-02 -1.40510067e-01
2.17380732e-01 1.15037549e+00 -9.93463457e-01 -3.67337912e-01
-9.41841602e-02 -1.04155846e-01 -4.25991952e-01 3.01740021e-01
-1.51562318e-01 6.78869247e-01 -9.68942419e-02 8.35059166e-01
3.18730056e-01 -4.68755886e-02 3.90813619e-01 1.37693524e-01
-3.02622706e-01 8.89820635e-01 -4.96552408e-01 1.67975473e+00
-7.62192130e-01 2.41709158e-01 -1.22687995e-01 -7.41356194e-01
5.85432827e-01 6.02415681e-01 -7.11990297e-02 -1.03603327e+00
4.10356224e-02 1.01019979e+00 4.45387065e-01 -4.64868098e-01
8.07132959e-01 -4.06297356e-01 -3.41233552e-01 4.50213313e-01
5.71822114e-02 -7.21094236e-02 7.10579991e-01 -1.04692234e-02
7.74784684e-01 3.86298597e-01 5.55599928e-01 -6.31041050e-01
6.76357329e-01 2.49974906e-01 7.37136304e-01 4.07726735e-01
-1.02132615e-02 4.56395894e-01 1.12651266e-01 2.46462841e-02
-1.34666646e+00 -8.64240885e-01 1.47822753e-01 1.32849205e+00
-7.23428503e-02 -2.54202068e-01 -1.05174625e+00 -7.37646163e-01
-4.95156854e-01 1.16316807e+00 -2.11342305e-01 2.06899885e-02
-1.05906010e+00 -7.33041763e-01 6.26906693e-01 6.32318258e-01
4.55326736e-01 -1.21416128e+00 -3.90640460e-03 3.69969398e-01
-8.59788418e-01 -1.39353526e+00 -1.27156007e+00 8.67500529e-02
-7.90111005e-01 -4.69111353e-01 -8.88695061e-01 -1.20365250e+00
5.61313272e-01 8.51074085e-02 1.34266269e+00 -6.81657046e-02
3.84126633e-01 -4.09694836e-02 -5.95104218e-01 -1.50915295e-01
-8.41812789e-01 4.56731230e-01 5.20395696e-01 -4.88992065e-01
5.43093085e-01 -3.89116466e-01 -1.39513195e-01 4.71343040e-01
-5.66128612e-01 4.38208103e-01 1.00423443e+00 1.12917042e+00
3.63793314e-01 -4.80995417e-01 6.47057176e-01 -7.57044971e-01
7.91267335e-01 -4.04146910e-01 -2.89115518e-01 4.67115372e-01
-6.85005128e-01 -1.30128476e-03 8.54833364e-01 -5.42276859e-01
-9.64563847e-01 -3.21671546e-01 -3.31097931e-01 -1.43856287e-01
-5.05093262e-02 8.57262254e-01 -2.69354552e-01 3.73626500e-01
6.15353286e-01 6.34201407e-01 -3.13925117e-01 -5.50954700e-01
2.19661832e-01 9.12996590e-01 4.18898255e-01 -8.07868958e-01
7.19080925e-01 -2.96929598e-01 -7.72641242e-01 -6.00875497e-01
-5.31497538e-01 -4.68826830e-01 -7.05410063e-01 2.69024640e-01
8.38363886e-01 -1.06671882e+00 -1.91111676e-02 4.28486973e-01
-1.25673139e+00 -6.32915080e-01 -8.19073245e-02 7.04736292e-01
-5.13525724e-01 5.07074654e-01 -1.21993923e+00 -4.10291731e-01
-8.07584167e-01 -1.36872327e+00 8.77943397e-01 -5.16940206e-02
-4.12325174e-01 -1.04649353e+00 -4.30554673e-02 6.84268117e-01
4.23401684e-01 -4.83812422e-01 1.18396330e+00 -8.48005354e-01
-1.79412156e-01 2.70510942e-01 -1.34462059e-01 6.74889982e-01
2.41801307e-01 -3.43891591e-01 -1.46834821e-01 -4.11100090e-01
-1.03196979e-01 -3.88270825e-01 3.74894112e-01 2.49781311e-01
2.46461704e-01 -6.82851195e-01 1.12520970e-01 4.00480866e-01
1.32202423e+00 3.98720741e-01 6.44242644e-01 5.04893124e-01
6.41945481e-01 4.78234500e-01 7.78284907e-01 -1.74934827e-02
6.22463226e-01 6.13237679e-01 -3.33767384e-01 -1.40140727e-01
-2.45735645e-01 -2.89801836e-01 1.07092404e+00 1.96834636e+00
-1.55511409e-01 -2.58132726e-01 -9.97763753e-01 6.93058670e-01
-1.65553772e+00 -6.23972416e-01 -1.92486927e-01 2.11597872e+00
1.59984934e+00 9.66956541e-02 1.00072585e-01 -2.96366632e-01
8.73744726e-01 -2.66538203e-01 -9.90046784e-02 -7.64618814e-01
-7.06758946e-02 1.02702215e-01 4.08537626e-01 8.40766549e-01
-6.88060939e-01 1.70181286e+00 5.12128258e+00 1.16639280e+00
-1.13511395e+00 4.48228955e-01 6.18657529e-01 8.79251286e-02
-4.08733189e-01 2.34111473e-01 -1.02197385e+00 6.32234633e-01
1.36371946e+00 -2.42619023e-01 5.85244775e-01 6.47372186e-01
3.19098324e-01 9.44391936e-02 -9.97067213e-01 5.53163409e-01
-3.92303653e-02 -1.19061613e+00 3.44154388e-01 -1.46503240e-01
1.02079594e+00 2.75796086e-01 -1.76844954e-01 6.69306278e-01
3.44650239e-01 -8.40001225e-01 9.24056053e-01 6.00848980e-02
1.28976035e+00 -5.25026083e-01 9.32780802e-01 5.36438584e-01
-1.04169333e+00 2.61550069e-01 -4.57510084e-01 -7.96760321e-02
3.72329921e-01 3.69183421e-01 -9.02705491e-01 9.36751246e-01
4.43727165e-01 3.43128294e-01 -2.04836220e-01 3.92698675e-01
-4.00581360e-01 1.05104625e+00 -2.06715204e-02 -2.18138009e-01
4.59489167e-01 -3.87252897e-01 6.49669826e-01 1.74662340e+00
5.44145346e-01 1.16295107e-01 2.72145629e-01 5.02559125e-01
-1.48809984e-01 5.43824077e-01 -2.89321393e-01 -3.44265729e-01
5.33775091e-01 1.00297165e+00 -4.68264312e-01 -5.46472490e-01
-4.36264575e-01 1.14487541e+00 3.90247822e-01 4.66483653e-01
-8.70800316e-01 -4.65025753e-01 3.93566549e-01 -2.10396871e-02
1.74508899e-01 -4.10813749e-01 -4.04502690e-01 -1.21658039e+00
7.25678131e-02 -1.48380399e+00 4.61182147e-02 -7.01382875e-01
-1.20786667e+00 1.04442501e+00 -1.44158229e-01 -1.41285706e+00
-5.22246838e-01 -6.18339181e-01 -3.57549548e-01 1.30290532e+00
-1.64184237e+00 -1.25097311e+00 4.99489605e-01 3.45076978e-01
1.07205677e+00 -4.54690456e-01 7.63528407e-01 3.67371351e-01
-6.31174564e-01 7.63484776e-01 2.75407910e-01 2.34499708e-01
7.66790926e-01 -1.18025553e+00 8.15181494e-01 1.12223160e+00
9.71121714e-02 9.60408092e-01 5.84076524e-01 -8.83701801e-01
-1.17174351e+00 -8.86897504e-01 1.80582988e+00 -4.45293605e-01
8.36049438e-01 -3.79849225e-01 -7.66933799e-01 7.35902846e-01
6.93157971e-01 -6.23635113e-01 6.32588267e-01 1.61501735e-01
-1.38574347e-01 3.08497310e-01 -7.20744252e-01 8.92265975e-01
1.20433044e+00 -6.54134214e-01 -7.86741495e-01 3.80120605e-01
1.11205459e+00 -6.61481142e-01 -7.89613426e-01 5.65913737e-01
3.01884741e-01 -6.50963187e-01 5.20952880e-01 -8.09266806e-01
8.41323078e-01 -1.58467725e-01 -4.41925704e-01 -1.57521856e+00
-5.91169357e-01 -9.04421151e-01 2.12669447e-01 1.08329475e+00
1.04176378e+00 -5.09303093e-01 2.53846616e-01 3.01438309e-02
-7.42291272e-01 -6.42839372e-01 -9.57033038e-01 -1.12344503e+00
4.18042809e-01 -1.42878696e-01 3.70010763e-01 1.07680523e+00
1.10183790e-01 8.27481687e-01 -5.85094571e-01 1.11311488e-01
6.59517571e-02 -1.21740848e-01 4.44846243e-01 -3.93568546e-01
-5.73875725e-01 -4.21476573e-01 3.52620274e-01 -1.09738874e+00
1.93920642e-01 -1.06374180e+00 1.81000561e-01 -1.54678011e+00
2.65305400e-01 -2.79954940e-01 -2.33519763e-01 6.14195406e-01
-6.68669939e-01 2.74136722e-01 2.53681988e-01 3.60565871e-01
-2.90413707e-01 5.10552406e-01 1.34455144e+00 1.64047882e-01
-2.90666759e-01 -1.00705296e-01 -7.33278632e-01 4.25574988e-01
8.65551889e-01 -4.58500355e-01 -2.22095773e-01 -8.51622164e-01
1.62436068e-01 1.27288640e-01 -3.29439580e-01 -6.66221917e-01
-5.65518290e-02 -4.04582202e-01 -1.75420851e-01 -3.42771411e-01
-8.39124247e-02 -3.87537867e-01 -1.53996974e-01 5.98699808e-01
-2.55049199e-01 5.80690265e-01 4.55655873e-01 -6.37464151e-02
-1.86241597e-01 -4.36096579e-01 5.14476597e-01 -3.48800272e-01
-7.60986209e-01 -4.60166447e-02 -4.74604845e-01 4.19857651e-01
6.59680903e-01 -3.04053992e-01 -2.50800103e-01 -3.12492311e-01
-3.63560081e-01 3.17125082e-01 3.02837253e-01 6.18768454e-01
3.57369751e-01 -1.28171301e+00 -1.24099803e+00 4.35603186e-02
2.00433377e-02 -3.84239614e-01 1.34792387e-01 1.08738303e+00
-5.40283620e-01 7.36510754e-01 -1.94357291e-01 -3.23357254e-01
-1.24341846e+00 3.94434273e-01 -1.95196865e-03 -6.86895728e-01
-8.40904936e-02 9.88507509e-01 2.24140674e-01 -1.06982124e+00
-6.04948662e-02 -3.74714494e-01 1.05380453e-01 -3.76559466e-01
2.41404623e-01 3.13946545e-01 1.20692573e-01 -7.59500682e-01
-3.23625207e-01 3.11800867e-01 -3.96543682e-01 -4.78547841e-01
8.64247859e-01 -3.95276487e-01 -3.78542066e-01 4.07332778e-01
9.02990580e-01 3.67754281e-01 -6.04988515e-01 -3.07832748e-01
4.70473737e-01 -3.91989797e-02 -3.58436227e-01 -1.29285491e+00
-4.25516337e-01 8.96906435e-01 9.84528437e-02 -3.04707766e-01
1.15211570e+00 -2.29849845e-01 1.16778541e+00 2.99783468e-01
5.24527073e-01 -1.33569431e+00 -9.81871113e-02 1.17479002e+00
1.02009749e+00 -1.12352049e+00 -5.02442002e-01 -3.59232008e-01
-9.97503698e-01 8.13616753e-01 7.21143305e-01 3.43315899e-01
1.56786069e-01 3.21141630e-01 2.15367466e-01 5.36072493e-01
-8.20599377e-01 -2.52838820e-01 3.72766942e-01 3.62299681e-01
9.64716911e-01 2.08157673e-01 -9.74857271e-01 6.63593411e-01
-4.61646259e-01 1.78555977e-02 3.50187778e-01 9.16170895e-01
-2.59391516e-01 -1.56942201e+00 -3.01119894e-01 1.13886558e-01
-5.71135461e-01 -1.12145638e+00 -2.73200870e-01 7.85710156e-01
-9.39569473e-02 8.86629105e-01 -1.71589464e-01 -5.45062482e-01
2.70188004e-01 2.78042436e-01 3.51901859e-01 -9.79002118e-01
-8.79767716e-01 5.16246200e-01 4.96237636e-01 -1.83624044e-01
-2.12116808e-01 -5.41923881e-01 -1.32655716e+00 -3.98982584e-01
-4.08998519e-01 4.77999419e-01 5.35457432e-01 1.02259254e+00
3.74207795e-01 3.57922822e-01 6.95323527e-01 -5.25892138e-01
-9.29648340e-01 -1.41517138e+00 -2.28841126e-01 1.74481675e-01
-2.82808021e-02 3.17310654e-02 -2.68864781e-01 8.81224424e-02] | [11.629557609558105, 10.32180404663086] |
202e7aa1-23e9-41be-8712-ea4c3d8bf176 | gist-distributed-training-for-large-scale | 2102.10424 | null | https://arxiv.org/abs/2102.10424v4 | https://arxiv.org/pdf/2102.10424v4.pdf | GIST: Distributed Training for Large-Scale Graph Convolutional Networks | The graph convolutional network (GCN) is a go-to solution for machine learning on graphs, but its training is notoriously difficult to scale both in terms of graph size and the number of model parameters. Although some work has explored training on large-scale graphs (e.g., GraphSAGE, ClusterGCN, etc.), we pioneer efficient training of large-scale GCN models (i.e., ultra-wide, overparameterized models) with the proposal of a novel, distributed training framework. Our proposed training methodology, called GIST, disjointly partitions the parameters of a GCN model into several, smaller sub-GCNs that are trained independently and in parallel. In addition to being compatible with all GCN architectures and existing sampling techniques for efficient GCN training, GIST i) improves model performance, ii) scales to training on arbitrarily large graphs, iii) decreases wall-clock training time, and iv) enables the training of markedly overparameterized GCN models. Remarkably, with GIST, we train an astonishgly-wide 32,768-dimensional GraphSAGE model, which exceeds the capacity of a single GPU by a factor of 8x, to SOTA performance on the Amazon2M dataset. | ['Anastasios Kyrillidis', 'Santiago Segarra', 'Artun Bayer', 'Chen Dun', 'Arindam Chowdhury', 'Jingkang Yang', 'Cameron R. Wolfe'] | 2021-02-20 | null | null | null | null | ['graph-sampling'] | ['graphs'] | [-9.42158923e-02 2.80770630e-01 -1.70251027e-01 3.05675529e-02
-2.96290278e-01 -6.70322955e-01 3.36250097e-01 1.31468162e-01
-2.28858963e-01 4.58508462e-01 -3.80068094e-01 -8.99887741e-01
-1.78553015e-01 -1.18296671e+00 -8.43164146e-01 -6.34322226e-01
-5.06600976e-01 8.88261139e-01 3.37493658e-01 -2.53905705e-03
-1.20456457e-01 6.97192788e-01 -1.30770338e+00 -1.52287170e-01
6.81997299e-01 8.72116089e-01 -3.32872383e-02 1.25343668e+00
3.79579328e-02 5.45016348e-01 -2.63746887e-01 -5.08217216e-01
3.61225188e-01 -4.63154688e-02 -8.63987446e-01 9.37845632e-02
7.61831164e-01 -1.64629877e-01 -6.26621783e-01 8.60145390e-01
2.83339769e-01 1.34699658e-01 1.97421685e-01 -1.45606923e+00
-3.56949508e-01 7.43878782e-01 -4.65080172e-01 3.68773490e-02
-3.02263558e-01 2.53764242e-01 8.15144658e-01 -4.33661789e-01
6.05769396e-01 1.22960603e+00 1.00387120e+00 3.30459774e-01
-1.11739755e+00 -5.51785529e-01 1.31992042e-01 -2.01274186e-01
-1.52929401e+00 -5.25879525e-02 3.02887768e-01 -4.69039716e-02
1.38788736e+00 3.08839768e-01 9.91006732e-01 7.78069139e-01
3.20779651e-01 3.35965723e-01 6.67423427e-01 -3.28442842e-01
2.58882970e-01 -2.60468692e-01 1.29988417e-01 1.02608371e+00
7.27908015e-01 -3.72407609e-03 -1.17664509e-01 -3.83075684e-01
1.14644206e+00 -1.04389422e-01 1.16195939e-02 -4.17393714e-01
-9.55759645e-01 9.94718075e-01 7.13630259e-01 2.19047219e-01
-8.45796987e-02 7.87950695e-01 7.15708673e-01 3.37227970e-01
3.95678043e-01 2.47843921e-01 -5.55161774e-01 -1.58486322e-01
-7.75123298e-01 1.71170682e-01 1.06032586e+00 1.33921087e+00
1.12887740e+00 1.55914679e-01 4.20570821e-01 5.48065066e-01
1.19154036e-01 4.11033064e-01 2.59927899e-01 -5.66047490e-01
5.15940905e-01 8.02633047e-01 -4.49910879e-01 -1.22366607e+00
-7.63851285e-01 -4.47618693e-01 -1.27682102e+00 -2.14178696e-01
3.47089410e-01 -2.43100032e-01 -1.02872324e+00 1.47288013e+00
5.88960946e-01 3.04824650e-01 -3.12544256e-01 6.49459779e-01
7.61282086e-01 5.92322528e-01 1.41531736e-01 3.19252640e-01
1.12152898e+00 -1.21457148e+00 -5.06743230e-02 -3.00138265e-01
1.29858899e+00 -2.98521876e-01 9.72878456e-01 3.67934853e-01
-9.21790183e-01 -4.12854522e-01 -8.28081667e-01 -1.62332028e-01
-6.73206866e-01 -2.35621631e-01 1.38559473e+00 9.65520203e-01
-1.83134055e+00 6.94044769e-01 -1.05829930e+00 -7.64732897e-01
3.88344556e-01 6.99255407e-01 -5.46632171e-01 -3.65046680e-01
-8.68284345e-01 5.96326888e-01 8.68829429e-01 -2.91571748e-02
-9.05604899e-01 -5.11417508e-01 -8.84649813e-01 3.74350458e-01
5.67674518e-01 -8.10660601e-01 7.90651560e-01 -6.22829616e-01
-1.31110275e+00 6.48955405e-01 3.37432444e-01 -5.64980865e-01
2.35026300e-01 2.48115301e-01 -3.68377984e-01 1.58459798e-01
-4.40826595e-01 7.00033724e-01 6.74226522e-01 -8.45218539e-01
-1.45788893e-01 -3.21092069e-01 1.37628064e-01 2.56659500e-02
-5.41912675e-01 -3.81203508e-03 -7.90817976e-01 -2.87632763e-01
-4.40832600e-02 -1.28992248e+00 -6.67929828e-01 -2.03908592e-01
-5.43873191e-01 -9.35832113e-02 7.56566226e-01 -3.28151166e-01
1.29547989e+00 -1.93945730e+00 1.43044302e-02 6.60526693e-01
8.95638347e-01 4.79099393e-01 -2.85268754e-01 6.78102791e-01
-6.83905855e-02 2.82434314e-01 -7.22495243e-02 -1.68882385e-01
1.03302421e-02 5.96193850e-01 2.54519075e-01 4.67262805e-01
-1.75568581e-01 1.20303750e+00 -7.16966808e-01 -5.51092803e-01
1.50482982e-01 3.99017155e-01 -8.34190547e-01 1.38532668e-02
-2.44425237e-01 -6.77637681e-02 -3.78338009e-01 4.71347183e-01
9.21232581e-01 -9.72634196e-01 6.25825286e-01 2.04044655e-01
1.86520308e-01 3.57857309e-02 -1.23830771e+00 1.64306235e+00
-4.26735491e-01 3.38365942e-01 1.72556341e-01 -1.03300476e+00
8.36275518e-01 1.38005763e-01 3.57755870e-01 -2.96913594e-01
1.67668566e-01 4.32263762e-01 -2.21333257e-03 -6.69430569e-02
4.68257964e-01 2.29683369e-01 8.75482485e-02 5.21595776e-01
4.16585177e-01 -8.77301618e-02 4.63197976e-01 4.58815336e-01
1.44467390e+00 -1.58727363e-01 3.15571666e-01 -5.97025514e-01
1.82304919e-01 -7.74058327e-02 2.51557468e-03 8.32677662e-01
4.51165847e-02 2.23738730e-01 7.81886935e-01 -9.67284083e-01
-1.17592978e+00 -5.51085055e-01 1.21498935e-01 1.25788438e+00
-8.16667154e-02 -9.55820799e-01 -9.59783137e-01 -8.04914296e-01
2.32251108e-01 1.57602966e-01 -4.94006872e-01 -1.10159971e-01
-7.62064636e-01 -9.47118342e-01 6.32300675e-01 5.66543519e-01
4.34868842e-01 -9.27617013e-01 -2.58447349e-01 2.55270451e-01
5.83091378e-01 -1.19543004e+00 -4.89293545e-01 3.23095441e-01
-1.13638484e+00 -1.33510125e+00 -2.50916898e-01 -8.43472838e-01
6.78104937e-01 4.82813835e-01 1.54434037e+00 7.53817976e-01
-3.04454446e-01 3.34124774e-01 -1.87058568e-01 -8.59263390e-02
-4.04751986e-01 6.37900829e-01 -2.08860412e-01 -3.70068729e-01
2.20837533e-01 -9.53273594e-01 -5.08443773e-01 1.68412387e-01
-9.28950608e-01 1.94695666e-01 5.26762426e-01 5.91845393e-01
4.63856876e-01 2.23463997e-01 2.91454762e-01 -1.32104087e+00
4.17885870e-01 -4.54972297e-01 -8.90742362e-01 2.97722995e-01
-9.94559169e-01 -1.17337771e-01 9.19001222e-01 -4.27666634e-01
-4.04980212e-01 -2.99746633e-01 -1.51379555e-01 -5.27497649e-01
-3.86731327e-02 6.10638022e-01 1.18146151e-01 -7.54977047e-01
7.56344855e-01 -3.14430781e-02 6.46681190e-02 -1.39988348e-01
5.78998208e-01 1.83097199e-01 3.42886895e-02 -6.63497627e-01
8.74195576e-01 1.02252796e-01 4.04967964e-01 -8.28806520e-01
-1.52964160e-01 -3.47728252e-01 -6.36799693e-01 -3.23255882e-02
5.19755363e-01 -8.54319334e-01 -8.51924479e-01 6.32191658e-01
-8.30441177e-01 -9.81653571e-01 -1.83761539e-03 1.98121116e-01
-3.47881317e-01 4.44524676e-01 -9.18604195e-01 -4.22126323e-01
-5.50553143e-01 -7.99227118e-01 1.10731041e+00 5.89974448e-02
2.26618454e-01 -1.50881088e+00 -8.05627778e-02 -1.02883019e-01
8.40151131e-01 3.22926849e-01 1.08618033e+00 -6.70181036e-01
-8.25091839e-01 -4.76808190e-01 -4.76993799e-01 1.64755866e-01
-3.49391192e-01 2.95824371e-02 -4.83556032e-01 -8.22009444e-01
-4.56178010e-01 -4.46083575e-01 6.46393061e-01 3.22771281e-01
1.43943095e+00 -3.61997157e-01 -5.79957664e-01 1.07287955e+00
1.84744728e+00 -1.49224907e-01 7.32302368e-01 1.49496064e-01
1.17794967e+00 -1.21368440e-02 -2.60322932e-02 1.45455942e-01
4.23854351e-01 3.27604860e-01 6.58041060e-01 -4.36572999e-01
-1.39906421e-01 -2.32163653e-01 -1.30790591e-01 1.17469490e+00
-2.72276253e-01 -6.58255458e-01 -1.13928926e+00 2.82524854e-01
-1.81786823e+00 -3.68821591e-01 -3.30963850e-01 2.06766796e+00
2.08442673e-01 1.43480793e-01 2.40768895e-01 -1.52555197e-01
6.34801090e-01 1.91193834e-01 -5.50816953e-01 -5.61667979e-01
5.07129617e-02 5.67275465e-01 8.48320425e-01 2.64366060e-01
-9.62578118e-01 1.21616840e+00 6.75081873e+00 9.73787725e-01
-1.13510239e+00 7.54554868e-02 7.52499819e-01 7.93365836e-02
-1.28755152e-01 7.18101263e-02 -7.08534956e-01 2.05777943e-01
1.47403789e+00 -1.49041489e-01 9.73039687e-01 1.13828874e+00
-2.85382986e-01 1.48076236e-01 -8.83655608e-01 9.96132851e-01
-6.45487905e-02 -1.38669801e+00 1.69264451e-01 5.64664841e-01
7.35092163e-01 4.59142566e-01 -2.91811675e-01 6.08933508e-01
7.00056255e-01 -1.20789540e+00 1.94466040e-01 -2.27996007e-01
9.78565335e-01 -9.85465229e-01 6.64003968e-01 2.87001550e-01
-1.46593869e+00 2.69500483e-02 -7.24845111e-01 4.76179682e-02
-2.70420611e-01 5.56183994e-01 -9.15163457e-01 6.88232064e-01
6.64267421e-01 5.25434017e-01 -9.88991618e-01 9.36479568e-01
9.26124901e-02 7.96797752e-01 -6.35796130e-01 -1.62974730e-01
6.06841743e-01 -5.01726627e-01 -5.09159006e-02 1.15987372e+00
4.52641517e-01 1.93046145e-02 2.65067548e-01 4.29191351e-01
-3.54956865e-01 2.42397353e-01 -6.64228261e-01 -3.19865465e-01
4.27218139e-01 1.48710740e+00 -9.89819527e-01 -4.39891160e-01
-4.76983219e-01 5.25859833e-01 9.27545428e-01 3.25852215e-01
-9.14918005e-01 -2.96604395e-01 3.46651226e-01 1.82990447e-01
5.17824233e-01 -4.73685712e-01 6.37536198e-02 -1.17861831e+00
-1.30630583e-01 -1.12145853e+00 6.09627843e-01 -5.47224224e-01
-1.12401140e+00 8.36338937e-01 -2.02055186e-01 -7.01726377e-01
-7.71449432e-02 -9.16548789e-01 -6.52790666e-01 7.85097539e-01
-1.29193795e+00 -1.33687973e+00 -5.70272803e-01 7.58635640e-01
-1.93172053e-01 1.04103021e-01 1.05106235e+00 2.81887084e-01
-5.70716202e-01 6.33934200e-01 1.27726451e-01 1.14597499e-01
3.02291453e-01 -1.30121982e+00 1.05965173e+00 6.84929967e-01
5.91619425e-02 5.40894151e-01 3.43291044e-01 -5.87154210e-01
-1.77249837e+00 -1.30905545e+00 6.18291974e-01 -1.03612736e-01
8.52878749e-01 -7.42022097e-01 -9.95116234e-01 1.06824958e+00
-3.93268391e-02 4.91186738e-01 4.67718691e-01 6.23495877e-01
-4.78017539e-01 -1.03820413e-01 -1.04948437e+00 5.96168637e-01
1.36214221e+00 -3.33109736e-01 4.62103486e-01 8.84254396e-01
8.22970212e-01 -7.94399977e-01 -1.16498494e+00 3.01893950e-02
2.36244455e-01 -9.07975852e-01 8.71163368e-01 -8.28071177e-01
-6.91045970e-02 1.10611588e-01 6.89435974e-02 -1.13399851e+00
-5.18996418e-01 -9.72669899e-01 -4.50927675e-01 6.02490127e-01
2.61980861e-01 -1.01100636e+00 1.12000227e+00 5.01362801e-01
-2.20920250e-01 -9.60875869e-01 -8.51007998e-01 -8.53589475e-01
2.18381807e-01 -3.07028115e-01 1.12563074e+00 8.43151212e-01
-2.03432530e-01 2.92454183e-01 -2.75800794e-01 1.81780279e-01
5.27173817e-01 9.46362242e-02 1.20067596e+00 -1.32550550e+00
-5.10639191e-01 -5.12812972e-01 -7.41174400e-01 -1.02372181e+00
-8.74995217e-02 -1.20240855e+00 -4.48057801e-01 -1.41030824e+00
1.41504660e-01 -7.54759312e-01 5.36940545e-02 7.01432824e-01
-1.43967167e-01 1.82340845e-01 1.28432617e-01 -3.23028862e-02
-7.82400131e-01 2.35133562e-02 1.29366493e+00 2.46013463e-01
3.11203301e-03 -2.11656392e-01 -5.43940485e-01 3.22871804e-01
6.96461916e-01 -3.07995349e-01 -5.64731419e-01 -5.29475987e-01
3.90612334e-01 5.54175340e-02 2.83383161e-01 -1.13852191e+00
2.19039470e-01 7.84879737e-03 1.75896689e-01 -3.17533284e-01
-4.33769152e-02 -6.43764853e-01 5.28416336e-01 4.09084499e-01
1.59703583e-01 3.60955656e-01 3.64095598e-01 5.86171508e-01
1.84822909e-03 1.36232432e-02 7.38936722e-01 -3.88998181e-01
-5.00936866e-01 1.02942014e+00 5.92054166e-02 4.77913730e-02
9.11227405e-01 -1.57783031e-01 -6.74247921e-01 -2.79034615e-01
-5.07012963e-01 2.41628975e-01 6.90438986e-01 5.39336130e-02
2.61395425e-01 -1.19685280e+00 -4.95772779e-01 3.70142519e-01
-6.20891899e-02 4.56509620e-01 3.93475622e-01 7.25037873e-01
-1.10184765e+00 5.30807376e-01 1.80688009e-01 -6.45041645e-01
-1.15511084e+00 9.32545900e-01 4.65911627e-01 -8.54766607e-01
-9.86540496e-01 8.67514968e-01 8.47082362e-02 -6.43208504e-01
-3.24529819e-02 -3.74951184e-01 1.71536416e-01 -5.25550127e-01
2.43514135e-01 3.53922844e-01 4.39105004e-01 -2.82575607e-01
-1.26525074e-01 4.09213781e-01 -8.98208320e-02 6.50471866e-01
1.44609869e+00 3.05801928e-01 -4.34289724e-01 -1.44390970e-01
1.32577217e+00 -3.05776894e-01 -1.01787901e+00 -5.87792173e-02
-8.07851404e-02 -3.08557808e-01 -1.91900076e-03 -4.18040276e-01
-1.31391382e+00 7.19874561e-01 1.06501855e-01 6.94062710e-01
1.01667249e+00 -8.29842612e-02 9.33521509e-01 4.45295244e-01
7.29770720e-01 -8.10784578e-01 -6.32948279e-02 6.97816014e-01
4.95256215e-01 -1.09425402e+00 1.80254102e-01 -6.27711236e-01
-1.69354290e-01 1.32950091e+00 7.66063809e-01 -4.14489269e-01
8.07649076e-01 2.42248401e-01 -3.43245864e-01 -6.37350321e-01
-9.49795127e-01 1.40625298e-01 7.70109147e-02 6.18040323e-01
1.29493237e-01 3.22736770e-01 2.16573998e-01 5.36608100e-02
-4.25836504e-01 -1.94901213e-01 3.81212205e-01 6.75512910e-01
-3.08456451e-01 -9.51352835e-01 -1.08340504e-02 7.39323318e-01
2.75404472e-02 -3.20810080e-01 -8.24452564e-02 1.36331844e+00
-9.94793475e-02 6.88220739e-01 6.75718412e-02 -3.84775162e-01
9.65346694e-02 -8.28822404e-02 5.41415155e-01 -5.59602380e-01
-7.42536902e-01 -1.35033190e-01 2.37684458e-01 -7.40000486e-01
7.04442039e-02 -9.13207307e-02 -8.98052692e-01 -1.08620727e+00
-4.57259178e-01 9.16819423e-02 7.52268910e-01 6.52009070e-01
7.10884452e-01 3.71413529e-01 3.12720686e-01 -1.00251102e+00
-3.79891068e-01 -9.66557205e-01 -9.01564538e-01 1.02255054e-01
-1.54776201e-01 -2.29983166e-01 -4.74834949e-01 -5.28391242e-01] | [6.97326135635376, 5.837663650512695] |
2fd25b3c-ff21-4b2b-99b7-419e58877a8b | graphfit-learning-multi-scale-graph | 2207.11484 | null | https://arxiv.org/abs/2207.11484v1 | https://arxiv.org/pdf/2207.11484v1.pdf | GraphFit: Learning Multi-scale Graph-Convolutional Representation for Point Cloud Normal Estimation | We propose a precise and efficient normal estimation method that can deal with noise and nonuniform density for unstructured 3D point clouds. Unlike existing approaches that directly take patches and ignore the local neighborhood relationships, which make them susceptible to challenging regions such as sharp edges, we propose to learn graph convolutional feature representation for normal estimation, which emphasizes more local neighborhood geometry and effectively encodes intrinsic relationships. Additionally, we design a novel adaptive module based on the attention mechanism to integrate point features with their neighboring features, hence further enhancing the robustness of the proposed normal estimator against point density variations. To make it more distinguishable, we introduce a multi-scale architecture in the graph block to learn richer geometric features. Our method outperforms competitors with the state-of-the-art accuracy on various benchmark datasets, and is quite robust against noise, outliers, as well as the density variations. | ['Gang Xiong', 'Fei-Yue Wang', 'Zhen Shen', 'Dong-Ming Yan', 'Huaiyu Wu', 'Mingyang Zhao', 'Keqiang Li'] | 2022-07-23 | null | null | null | null | ['surface-normals-estimation'] | ['computer-vision'] | [-3.81833613e-01 -2.60442287e-01 1.13002084e-01 -3.43613416e-01
-2.88304090e-01 -1.76583648e-01 3.93595546e-01 3.98101091e-01
-2.44015381e-01 2.43148625e-01 -9.98389274e-02 6.14649020e-02
-4.45450172e-02 -1.20885324e+00 -8.08019102e-01 -5.87946177e-01
-5.90469949e-02 4.72194105e-01 5.51353753e-01 -8.03256705e-02
3.04056168e-01 9.55998659e-01 -1.29891491e+00 -2.59656250e-01
1.12789345e+00 1.16595340e+00 6.82624876e-02 1.99620381e-01
-3.64212096e-01 2.66402006e-01 -4.70848739e-01 -1.83892697e-01
3.19486737e-01 2.30475336e-01 -9.92424712e-02 1.98960692e-01
7.45026588e-01 -4.34299231e-01 -6.52499020e-01 1.17172372e+00
4.60855216e-01 1.60004258e-01 8.74735355e-01 -9.43511069e-01
-1.04735243e+00 -1.08381271e-01 -1.07729387e+00 1.27351478e-01
2.35043287e-01 3.89056832e-01 8.10396791e-01 -1.06112349e+00
2.50937253e-01 1.43061781e+00 8.77547979e-01 5.68572134e-02
-9.38021660e-01 -7.27049589e-01 5.96943080e-01 -6.90793395e-02
-1.70775938e+00 7.55694211e-02 1.04740274e+00 -2.99745470e-01
9.14985597e-01 -1.30905136e-01 6.72004282e-01 7.72315145e-01
9.44872424e-02 6.02799952e-01 4.30738270e-01 7.64734447e-02
1.43856868e-01 -4.50438559e-01 -2.34392196e-01 6.26916289e-01
6.78546309e-01 -1.48014307e-01 5.20135984e-02 -2.04970777e-01
1.30948770e+00 4.58375484e-01 -4.23408210e-01 -6.71384931e-01
-1.12996221e+00 8.77561629e-01 9.79115903e-01 1.23331949e-01
-3.71465236e-01 1.92038551e-01 2.15214729e-01 -2.03694794e-02
6.85774267e-01 1.65220499e-01 -3.70448232e-01 6.52699172e-02
-6.20130956e-01 3.50157857e-01 4.79958117e-01 1.11243725e+00
1.18100381e+00 2.01098621e-02 -1.24529809e-01 9.20071006e-01
6.47528708e-01 8.25181186e-01 9.87992734e-02 -6.73800707e-01
5.54253101e-01 1.03742707e+00 -4.34814543e-02 -1.40439677e+00
-4.65678215e-01 -6.40444458e-01 -1.11210144e+00 1.20339468e-01
5.06141670e-02 9.47716683e-02 -1.08193135e+00 1.32687676e+00
5.55042446e-01 6.03783667e-01 -3.34821254e-01 9.79662418e-01
1.05874825e+00 6.03618979e-01 -3.37339997e-01 4.61889133e-02
7.70780265e-01 -7.13880479e-01 -4.00940120e-01 -2.60083467e-01
3.84403437e-01 -6.12058103e-01 9.93210196e-01 6.96623400e-02
-9.84305918e-01 -5.69059491e-01 -9.79145765e-01 -2.41448000e-01
-8.49443302e-02 -1.21941827e-01 6.11145794e-01 1.45637646e-01
-8.03962409e-01 6.89804852e-01 -1.09055853e+00 -2.69059569e-01
8.90978515e-01 3.03232282e-01 -2.85235643e-01 -3.05202544e-01
-6.52788997e-01 3.43469173e-01 -3.97796556e-02 1.53883666e-01
-3.57163727e-01 -8.24673116e-01 -1.06960022e+00 2.85257548e-01
2.26860270e-01 -8.23439419e-01 7.30035305e-01 -3.10711384e-01
-1.34271908e+00 4.42755282e-01 -1.62073240e-01 3.07226321e-03
6.19578302e-01 -2.46656761e-01 -1.60718814e-01 1.30241990e-01
1.06637299e-01 4.13710833e-01 7.60385811e-01 -1.25625861e+00
-2.39765495e-01 -7.19466984e-01 -1.02711841e-01 2.62087971e-01
-2.04712495e-01 -5.69884181e-01 -7.23510504e-01 -6.97547078e-01
5.80765903e-01 -5.69965184e-01 -4.52236503e-01 3.65321308e-01
-3.48659128e-01 -2.61647731e-01 1.06974733e+00 -1.03953868e-01
9.71640587e-01 -2.33861470e+00 -1.54868737e-01 4.92167294e-01
6.50488257e-01 9.24723148e-02 -1.81796312e-01 1.87298417e-01
2.74921000e-01 9.63364840e-02 -2.11495310e-01 -4.15206552e-01
-8.40070564e-03 2.48776689e-01 -1.03153279e-02 6.57065392e-01
5.46554923e-01 9.63597000e-01 -8.92275155e-01 -2.37262532e-01
5.69290459e-01 7.76031852e-01 -6.76699102e-01 2.94482440e-01
-2.09420905e-01 3.31237823e-01 -9.00683105e-01 6.62693799e-01
1.32459116e+00 -4.62895274e-01 -5.55635870e-01 -3.19502145e-01
1.65883556e-01 9.73424166e-02 -1.32894683e+00 1.55874407e+00
-3.33376497e-01 1.02000542e-01 6.30904660e-02 -7.57799804e-01
1.39351201e+00 -1.82743773e-01 5.10273635e-01 -3.46916854e-01
4.34485078e-02 2.79924929e-01 -9.77147520e-02 -2.04005957e-01
3.87882441e-01 1.37132257e-01 3.17390889e-01 -4.02638949e-02
-1.21263705e-01 -2.79502988e-01 -3.42748821e-01 2.38265589e-01
1.23602188e+00 -8.60777944e-02 1.59989536e-01 -2.23248571e-01
5.08904338e-01 -6.36771381e-01 6.73667073e-01 4.57714587e-01
-9.60024372e-02 9.55435574e-01 4.45568651e-01 -5.15456736e-01
-8.11685920e-01 -1.24729443e+00 -2.81699598e-01 2.96611041e-01
6.34053826e-01 -3.95197541e-01 -2.86233574e-01 -7.07088470e-01
5.36842585e-01 1.08535863e-01 -5.43053389e-01 -1.88017815e-01
-5.24057329e-01 -5.42959213e-01 6.00502826e-02 8.09117913e-01
7.06557214e-01 -5.73115587e-01 6.34902790e-02 1.57940090e-01
2.42228821e-01 -1.28312397e+00 -5.41413009e-01 -1.69430152e-01
-1.09848523e+00 -9.40958858e-01 -6.89750791e-01 -6.12187326e-01
8.15757990e-01 3.87100220e-01 1.02844119e+00 2.85688251e-01
-5.42438067e-02 2.32728347e-01 -3.13192457e-01 -2.65238881e-01
2.10637540e-01 3.15063000e-01 -1.97525829e-01 -5.41550927e-02
5.98418891e-01 -9.88548934e-01 -8.87643993e-01 2.59502381e-01
-7.47694433e-01 -3.08233082e-01 6.32806301e-01 7.49307811e-01
8.10278177e-01 -3.48158106e-02 3.58430833e-01 -8.03701222e-01
3.71573836e-01 -6.43895984e-01 -6.53917015e-01 -1.76612258e-01
-3.13216239e-01 -3.74261774e-02 6.64539814e-01 -1.99999720e-01
-6.18065715e-01 -3.49234752e-02 -3.67727935e-01 -9.26308990e-01
-2.97045082e-01 2.62179643e-01 -4.41387057e-01 -6.86337531e-01
4.49534804e-01 1.06282912e-01 -1.54912174e-01 -5.99158823e-01
2.09907576e-01 4.56223607e-01 3.28219146e-01 -5.88680863e-01
1.22337651e+00 5.81958175e-01 2.53130317e-01 -8.52606952e-01
-5.98705769e-01 -5.34736872e-01 -5.10183930e-01 4.80850860e-02
5.24124920e-01 -1.09739029e+00 -5.92532635e-01 5.76914549e-01
-1.20315886e+00 -9.86938030e-02 -1.71397969e-01 4.54699248e-01
-3.36163878e-01 6.86438560e-01 -7.87233770e-01 -5.25453150e-01
-2.52727389e-01 -1.12477982e+00 1.45519578e+00 4.42006528e-01
3.21796179e-01 -1.03260541e+00 -8.15511961e-03 -8.57551396e-02
5.73733151e-01 4.35586810e-01 8.36287320e-01 -5.30449390e-01
-1.03049421e+00 -2.89328784e-01 -6.49889827e-01 3.53684664e-01
3.80099237e-01 2.42898911e-01 -7.69447386e-01 -4.56478685e-01
1.11144921e-02 2.49352120e-02 9.05571938e-01 3.63507926e-01
1.46113503e+00 -9.24919620e-02 -4.81989712e-01 1.09974253e+00
1.52376497e+00 -3.54148418e-01 7.67039180e-01 -7.49623636e-03
1.26771545e+00 1.15741685e-01 3.97216976e-01 4.76471752e-01
5.79893887e-01 3.15081537e-01 7.80435145e-01 -3.04858059e-01
1.35357708e-01 -2.41914690e-01 -1.60119042e-01 9.77947772e-01
-1.18941523e-01 -2.43530080e-01 -7.64165044e-01 4.84762013e-01
-1.78499281e+00 -5.57414830e-01 -1.28117964e-01 1.95748174e+00
2.32459158e-01 3.11716676e-01 -1.88275144e-01 -1.74726725e-01
7.12532818e-01 4.86199170e-01 -7.36918867e-01 7.48745864e-03
-1.31529748e-01 2.72681594e-01 5.75601339e-01 3.35500836e-01
-1.07065725e+00 8.43378007e-01 5.94080591e+00 9.16033804e-01
-1.08765519e+00 -2.80107379e-01 6.06404245e-01 7.75722340e-02
-6.19520187e-01 -2.75349855e-01 -6.44449353e-01 5.34425497e-01
1.81316987e-01 6.19844124e-02 2.33389914e-01 1.08158541e+00
-2.40117013e-02 2.26936832e-01 -7.12798417e-01 1.16719699e+00
3.39111648e-02 -1.30652213e+00 3.18505138e-01 1.71798185e-01
9.42215145e-01 5.92223823e-01 -9.07494407e-03 2.05725625e-01
2.65541345e-01 -9.56635654e-01 2.88316011e-01 5.14995694e-01
6.75107539e-01 -9.57852185e-01 8.25138927e-01 2.30051026e-01
-1.58693588e+00 2.31199965e-01 -8.12084019e-01 -1.02173820e-01
4.75635082e-02 1.01243460e+00 -3.01071048e-01 6.98476076e-01
6.39482379e-01 1.26573658e+00 -6.88143790e-01 1.40962410e+00
-1.10314891e-01 1.84597656e-01 -6.81120038e-01 2.39424203e-02
2.24366665e-01 -3.34995270e-01 5.54618061e-01 9.85088706e-01
6.01281524e-01 9.33614559e-03 5.26547253e-01 1.00739288e+00
-2.58295298e-01 3.40095967e-01 -6.98089242e-01 2.25491956e-01
6.78150713e-01 1.23447800e+00 -7.20730424e-01 -4.81802300e-02
-7.41254866e-01 7.69210637e-01 6.16466165e-01 5.22884965e-01
-5.92805445e-01 -5.18208623e-01 9.03696716e-01 4.62990999e-01
6.44870818e-01 -4.64993924e-01 -3.26102644e-01 -1.34580362e+00
2.79362082e-01 -4.10115868e-01 3.94248590e-02 -7.19015360e-01
-1.82731616e+00 4.81150746e-01 -2.81214327e-01 -1.21650779e+00
3.05443883e-01 -6.65852070e-01 -1.05554819e+00 8.29764724e-01
-1.67127407e+00 -1.04490578e+00 -8.54753077e-01 5.48481166e-01
7.77367353e-02 8.69629160e-02 3.06054264e-01 4.60338742e-01
-3.68265688e-01 5.55753887e-01 8.48764256e-02 2.88568765e-01
6.16006136e-01 -1.17644143e+00 8.40954065e-01 6.99679136e-01
-9.74293202e-02 6.54073000e-01 6.04924001e-02 -8.33563566e-01
-1.19375169e+00 -1.34210277e+00 2.42965028e-01 -2.31344298e-01
5.52018285e-01 -4.47753519e-01 -1.34719431e+00 5.86844981e-01
-5.00165820e-01 6.74311936e-01 2.63091058e-01 2.80710459e-02
-4.92105991e-01 -2.22302839e-01 -1.17529225e+00 3.91108334e-01
1.35721767e+00 -4.26789224e-01 -3.25808913e-01 2.23921522e-01
8.71842027e-01 -7.31660724e-01 -9.43616211e-01 8.60615969e-01
3.85524407e-02 -9.47370231e-01 9.26554978e-01 -7.33796135e-02
2.54454374e-01 -5.14282823e-01 -8.03709626e-02 -1.19335127e+00
-7.36739397e-01 -4.50658649e-01 -3.07422251e-01 1.19118559e+00
6.66467100e-02 -8.01604152e-01 9.31276917e-01 1.57491580e-01
-3.57653171e-01 -1.07200146e+00 -8.79408300e-01 -6.77788258e-01
9.73811895e-02 -3.75314802e-01 1.12337565e+00 9.08984005e-01
-5.40639341e-01 2.75858998e-01 4.34508286e-02 4.32507128e-01
5.68492055e-01 5.18450588e-02 1.03455210e+00 -1.52815557e+00
9.92629939e-05 -5.72075307e-01 -1.13588929e+00 -1.55871439e+00
2.84423139e-02 -7.35992432e-01 -5.75271919e-02 -1.62771058e+00
-7.50290975e-02 -6.33245707e-01 -2.06029877e-01 5.96884340e-02
-4.52058673e-01 2.77428418e-01 -1.49115309e-01 2.07171172e-01
-6.31149888e-01 1.12705874e+00 1.55236566e+00 -6.11366481e-02
-8.66771713e-02 8.09699763e-03 -6.70367897e-01 8.14969718e-01
5.59801817e-01 -2.52913982e-01 -3.10160309e-01 -7.41762996e-01
1.54587150e-01 -4.47556525e-01 2.96920151e-01 -1.25874984e+00
1.48819000e-01 -1.19221836e-01 7.56888926e-01 -8.30396175e-01
2.44272351e-01 -7.88361311e-01 -2.13387296e-01 -1.53758815e-02
2.75622636e-01 1.79558218e-01 2.47724548e-01 1.01226878e+00
-2.05325618e-01 1.83537051e-01 8.14898014e-01 -8.18457752e-02
-2.99348831e-01 1.17250288e+00 2.36113161e-01 7.82985538e-02
9.37538385e-01 -1.80806682e-01 -3.15772504e-01 -3.49087238e-01
-2.44981512e-01 4.73722309e-01 1.03550410e+00 2.54261762e-01
7.95777142e-01 -1.52188301e+00 -5.92431545e-01 5.68864703e-01
2.87004232e-01 8.64234686e-01 4.05545861e-01 6.52278185e-01
-9.41087067e-01 -1.20896526e-01 -3.09037673e-03 -1.01352942e+00
-4.80656505e-01 5.03745258e-01 4.24911469e-01 6.85940683e-02
-9.75708365e-01 8.54828060e-01 4.21414018e-01 -4.98376489e-01
1.34748578e-01 -8.20071220e-01 5.12684286e-02 -6.07111871e-01
3.71808469e-01 1.93161428e-01 8.82928520e-02 -6.13504589e-01
-4.55595315e-01 1.09813666e+00 -1.31775841e-01 6.21236265e-01
1.24266672e+00 -1.18076660e-01 -5.49016148e-02 3.57555896e-01
1.26919925e+00 2.62871206e-01 -1.42536819e+00 -4.22382742e-01
-2.91878998e-01 -9.10340130e-01 1.86105520e-01 -6.82418346e-02
-1.49763584e+00 9.82362449e-01 4.07583356e-01 1.39292911e-01
7.71790087e-01 -3.13615948e-02 1.03344011e+00 3.48118871e-01
2.31701806e-01 -6.08768404e-01 1.14449151e-01 6.70480847e-01
6.80660963e-01 -1.46027696e+00 2.94328421e-01 -7.06256986e-01
-1.73355341e-01 1.10505629e+00 9.19774890e-01 -7.22674787e-01
9.40577030e-01 2.45066598e-01 -4.03930359e-02 -3.61709058e-01
-4.62063134e-01 -1.05708696e-01 3.00933808e-01 8.06695938e-01
2.53114253e-01 -1.94764659e-01 4.25438434e-02 4.36944246e-01
8.28536823e-02 -1.97381780e-01 1.22026019e-01 4.88960564e-01
-5.21139383e-01 -7.43953764e-01 -1.58553705e-01 7.56312668e-01
-1.82021141e-01 -2.18281825e-03 -1.43995926e-01 7.86834717e-01
6.86984137e-02 3.37554753e-01 4.83759910e-01 -3.90624434e-01
4.94858325e-01 -5.73814809e-01 2.87849635e-01 -6.89159572e-01
-1.95086256e-01 1.63939118e-01 -4.93115604e-01 -6.89695418e-01
-2.72076935e-01 -4.38078314e-01 -1.23516357e+00 -4.44812089e-01
-4.24409837e-01 -5.02039604e-02 2.33329996e-01 7.72720516e-01
6.51555896e-01 4.91019279e-01 7.19197512e-01 -1.16491044e+00
-3.25136632e-01 -9.21414733e-01 -6.73698425e-01 4.81836289e-01
3.74435514e-01 -9.84984815e-01 -6.29547656e-01 -7.62797654e-01] | [7.996511936187744, -3.5471701622009277] |
1fd4bdaa-2d30-4434-83c5-3fbe77b48a2c | unifying-cross-lingual-semantic-role-labeling | null | null | https://aclanthology.org/2021.naacl-main.31 | https://aclanthology.org/2021.naacl-main.31.pdf | Unifying Cross-Lingual Semantic Role Labeling with Heterogeneous Linguistic Resources | While cross-lingual techniques are finding increasing success in a wide range of Natural Language Processing tasks, their application to Semantic Role Labeling (SRL) has been strongly limited by the fact that each language adopts its own linguistic formalism, from PropBank for English to AnCora for Spanish and PDT-Vallex for Czech, inter alia. In this work, we address this issue and present a unified model to perform cross-lingual SRL over heterogeneous linguistic resources. Our model implicitly learns a high-quality mapping for different formalisms across diverse languages without resorting to word alignment and/or translation techniques. We find that, not only is our cross-lingual system competitive with the current state of the art but that it is also robust to low-data scenarios. Most interestingly, our unified model is able to annotate a sentence in a single forward pass with all the inventories it was trained with, providing a tool for the analysis and comparison of linguistic theories across different languages. We release our code and model at https://github.com/SapienzaNLP/unify-srl. | ['Roberto Navigli', 'Andrea Bacciu', 'Simone Conia'] | 2021-06-01 | null | null | null | naacl-2021-4 | ['semantic-role-labeling'] | ['natural-language-processing'] | [-3.90850492e-02 8.95529762e-02 -5.25950432e-01 -5.22290707e-01
-9.33549881e-01 -1.11601985e+00 7.63250768e-01 4.99400765e-01
-7.31568336e-01 9.13159966e-01 3.93228382e-01 -4.49493200e-01
-3.06461118e-02 -5.32985449e-01 -4.33987111e-01 -1.78016663e-01
5.68160832e-01 7.76097834e-01 2.68576711e-01 -5.65989017e-01
-1.87504143e-02 1.95286259e-01 -1.20538676e+00 3.79133612e-01
7.03863025e-01 5.47059000e-01 3.96549284e-01 1.23684272e-01
-3.21010560e-01 8.70235920e-01 -2.69760638e-01 -7.07661510e-01
4.55389060e-02 -2.47972965e-01 -1.26825488e+00 -3.40645939e-01
3.36633354e-01 3.84329885e-01 8.84733573e-02 1.06619954e+00
2.78468847e-01 3.22794132e-02 3.89464587e-01 -9.89312708e-01
-6.04864001e-01 1.03899205e+00 -1.76173508e-01 2.20921114e-01
5.11603236e-01 -1.61322653e-01 1.56995666e+00 -6.65318072e-01
1.09836745e+00 1.50099385e+00 4.65829581e-01 7.32087314e-01
-1.40061224e+00 -4.80062991e-01 2.53155649e-01 2.06560761e-01
-1.31674612e+00 -6.10232294e-01 7.26421356e-01 -3.95014524e-01
1.06271780e+00 -1.56840995e-01 1.40649676e-01 9.95641232e-01
-2.13998497e-01 6.33705378e-01 1.30315661e+00 -8.26940954e-01
-5.11315539e-02 2.55048633e-01 2.34773800e-01 6.87500238e-01
2.35988528e-01 -4.10561830e-01 -6.85001016e-01 6.99636191e-02
5.21089613e-01 -6.67827070e-01 -1.02444895e-01 -9.18025896e-02
-1.28423464e+00 7.96444356e-01 -2.28248835e-02 9.71667945e-01
-5.45970201e-02 -2.69721877e-02 7.78626978e-01 5.57311952e-01
5.08826971e-01 7.86375225e-01 -9.70649600e-01 -1.02585070e-01
-5.93728006e-01 3.07735801e-01 8.03412139e-01 7.21891582e-01
5.16529322e-01 -2.05561444e-01 3.22211981e-01 1.18234932e+00
2.92557597e-01 1.90289900e-01 6.31447196e-01 -1.13703489e+00
6.23340845e-01 6.52232230e-01 1.50738865e-01 -6.84836626e-01
-4.38340455e-01 -1.50833458e-01 -2.59546280e-01 -2.07448438e-01
7.63400435e-01 9.15073231e-02 -3.43202412e-01 2.09045815e+00
2.93557107e-01 -3.71411592e-01 3.82796049e-01 6.68891013e-01
6.58374131e-01 3.85454506e-01 4.90570456e-01 -7.39893243e-02
1.71640503e+00 -7.42759347e-01 -5.90804279e-01 -6.69480979e-01
9.37320888e-01 -6.78847790e-01 1.42867541e+00 2.98489034e-01
-1.12920880e+00 -4.32207078e-01 -8.92604887e-01 -6.43335462e-01
-6.26579762e-01 1.50054425e-01 6.29869998e-01 4.26007122e-01
-1.04665196e+00 2.69918740e-01 -7.15336204e-01 -6.68136656e-01
1.59193143e-01 8.15551132e-02 -6.70279801e-01 -2.15066701e-01
-1.56823325e+00 1.41322029e+00 7.65687168e-01 -2.77197748e-01
-3.35609496e-01 -6.37278438e-01 -1.15093362e+00 -3.17591935e-01
6.01685226e-01 -5.34593463e-01 1.43479669e+00 -9.69034731e-01
-1.42542684e+00 1.43325675e+00 -3.13087076e-01 -5.64860642e-01
2.53932178e-01 -2.65366554e-01 -4.55678076e-01 -1.40515089e-01
6.30491316e-01 5.52401900e-01 1.46924615e-01 -1.01009965e+00
-7.35660553e-01 -3.77448738e-01 5.14196634e-01 3.76031965e-01
-1.28847554e-01 6.92435920e-01 -2.43489757e-01 -5.56191444e-01
7.99160451e-02 -7.53154576e-01 -1.12940826e-01 -2.67531842e-01
1.90716773e-01 -7.53160238e-01 2.34791890e-01 -6.13206565e-01
1.16831434e+00 -1.83875525e+00 3.37665915e-01 -2.81819254e-01
-2.55723953e-01 1.70953825e-01 1.96915939e-02 5.38476229e-01
-1.67928413e-02 2.98527211e-01 -5.43357968e-01 -4.53931153e-01
1.94274142e-01 6.69547319e-01 -2.76143372e-01 4.54651892e-01
1.72014922e-01 8.66281986e-01 -1.15246391e+00 -5.38235307e-01
2.77160257e-01 1.95371732e-01 -4.24123377e-01 -1.37881771e-01
-4.77204561e-01 6.02968454e-01 -3.18114221e-01 4.81927782e-01
1.08327374e-01 -1.32959560e-01 7.41778851e-01 1.75154451e-02
-3.84723127e-01 1.24225378e+00 -1.05244994e+00 2.21157980e+00
-9.47763562e-01 2.64443547e-01 1.71619147e-01 -1.10013926e+00
6.35926664e-01 6.33351266e-01 2.92236209e-01 -6.65507615e-01
1.44511431e-01 8.41785669e-01 1.82223126e-01 -2.42997319e-01
3.36620688e-01 -5.23945808e-01 -5.04320741e-01 5.47894835e-01
4.97382194e-01 -3.56895417e-01 7.31584966e-01 3.83183770e-02
6.98644817e-01 4.06583518e-01 9.41687822e-01 -6.77994132e-01
1.11041105e+00 1.92363396e-01 6.01485550e-01 2.07898155e-01
-4.86647636e-02 1.43466040e-01 4.75272685e-01 -2.00731784e-01
-9.43706274e-01 -8.03834379e-01 -5.12365282e-01 1.38500023e+00
-3.26685309e-02 -5.71521342e-01 -5.21607637e-01 -7.70381927e-01
-1.75486505e-01 1.02703738e+00 -2.70184815e-01 3.21824908e-01
-8.00643504e-01 -4.25562650e-01 7.64537811e-01 1.88837707e-01
3.25951397e-01 -1.25667846e+00 -4.34131414e-01 4.75809813e-01
-2.11837709e-01 -1.54191279e+00 -2.64407285e-02 9.13049430e-02
-4.64411438e-01 -1.16922927e+00 -1.17843509e-01 -7.56447613e-01
1.39394641e-01 -1.66365743e-01 1.55984545e+00 -1.28084540e-01
2.00381190e-01 1.78379238e-01 -4.23866719e-01 -3.22495848e-01
-7.42258012e-01 3.57116997e-01 1.34995505e-01 -1.89489439e-01
5.47841549e-01 -4.48044360e-01 1.91138480e-02 8.49112943e-02
-7.73413539e-01 -4.29965630e-02 -1.62071455e-02 5.12887061e-01
5.62124372e-01 -1.61501020e-01 5.55642188e-01 -1.32531369e+00
6.41144335e-01 -5.98398805e-01 -6.27700865e-01 2.66408086e-01
-4.58333969e-01 1.90483421e-01 6.76732242e-01 1.19392172e-01
-1.16506171e+00 -1.73098937e-01 -4.64733154e-01 2.77159661e-01
-4.09124047e-01 7.46394992e-01 -5.11647940e-01 3.43702644e-01
6.42619371e-01 3.52368178e-03 -2.53282577e-01 -7.95880020e-01
6.95371211e-01 4.78150427e-01 5.07465959e-01 -1.07495546e+00
6.24061465e-01 3.37509692e-01 -1.59300119e-01 -7.83468306e-01
-1.42420554e+00 -5.56842744e-01 -1.02465641e+00 1.58687577e-01
9.74633873e-01 -1.00783277e+00 -3.24488312e-01 2.05792204e-01
-1.35714114e+00 -4.35322732e-01 -2.48373419e-01 3.52822781e-01
-4.63479310e-01 2.32596919e-01 -6.57558680e-01 -3.94385785e-01
-1.55284792e-01 -9.53422070e-01 9.21702266e-01 -1.51386306e-01
-6.25660717e-01 -1.54982018e+00 2.67207235e-01 5.05318165e-01
1.48795277e-01 3.62581722e-02 1.07526720e+00 -8.72841120e-01
-1.07649323e-02 1.02260940e-01 -8.85242224e-02 4.69759107e-01
3.13269138e-01 -4.05021161e-01 -7.06566036e-01 -1.12445563e-01
4.07848246e-02 -6.02920890e-01 3.97656560e-01 -4.82004182e-03
5.57418644e-01 -2.52199590e-01 7.59623153e-03 2.53728092e-01
1.61452639e+00 -1.44071415e-01 1.91267088e-01 6.52164161e-01
6.45803928e-01 9.40293908e-01 7.83301830e-01 -2.98566133e-01
8.98262143e-01 8.71634007e-01 -2.57698838e-02 1.87601492e-01
-1.09889150e-01 -1.40758500e-01 5.39451778e-01 9.49891090e-01
-2.40294617e-02 -7.29285926e-02 -1.26975989e+00 8.47666502e-01
-2.01644087e+00 -6.78798616e-01 -1.61302730e-01 1.95624781e+00
1.30317628e+00 8.22493508e-02 3.21826339e-02 -6.75448924e-02
4.22533333e-01 3.79325062e-01 -1.29923180e-01 -4.98116493e-01
-3.68059307e-01 5.14593184e-01 3.01437855e-01 9.13158953e-01
-1.07407022e+00 1.66945624e+00 5.33684683e+00 7.54620671e-01
-9.65085089e-01 6.46078825e-01 1.29366577e-01 1.66235447e-01
-3.17958325e-01 1.72333464e-01 -9.48486507e-01 3.20453882e-01
1.01161349e+00 -3.79573584e-01 4.02393669e-01 6.57580674e-01
2.42003649e-02 -4.63136286e-02 -1.16221225e+00 7.51797140e-01
6.68574944e-02 -1.26510870e+00 3.64243574e-02 -3.16069305e-01
5.16752660e-01 4.72838342e-01 -4.74929035e-01 1.44054219e-01
6.48024559e-01 -9.41535771e-01 1.05976057e+00 -7.06855208e-02
8.72259140e-01 -5.58981001e-01 7.21293032e-01 3.83355886e-01
-1.16494024e+00 2.65899450e-01 -2.10674748e-01 -2.46547922e-01
3.87988240e-01 2.33031631e-01 -4.31527138e-01 7.43938088e-01
3.13813716e-01 8.44093204e-01 -5.03481865e-01 2.01009855e-01
-8.71894956e-01 5.69769800e-01 -1.99084625e-01 2.40979895e-01
2.53048807e-01 -2.05646113e-01 5.47341406e-01 1.39594817e+00
-2.13192888e-02 -8.01658481e-02 5.24878979e-01 6.30791783e-01
-2.09504738e-01 5.48635840e-01 -6.90683424e-01 -1.07670158e-01
5.13370275e-01 1.11349642e+00 -5.83582997e-01 -3.77408683e-01
-6.80168867e-01 7.64410317e-01 7.14725018e-01 -1.06004193e-01
-4.60282624e-01 -7.27041513e-02 7.67487884e-01 2.44057044e-01
-2.77412772e-01 -5.49844861e-01 -3.24455321e-01 -1.41853309e+00
1.07018232e-01 -9.42130923e-01 7.28365779e-01 -5.94931126e-01
-1.47747397e+00 5.92379153e-01 2.15496674e-01 -6.72139645e-01
-5.31923652e-01 -1.01389277e+00 -3.46124843e-02 1.02370250e+00
-1.84566689e+00 -1.41927373e+00 4.78728473e-01 6.35241926e-01
7.32549548e-01 -2.08722383e-01 1.01086688e+00 3.81642133e-01
-2.71028042e-01 2.25301564e-01 -4.45945174e-01 1.80749521e-01
9.36826885e-01 -1.29961789e+00 4.56358671e-01 9.54714596e-01
5.05233109e-01 8.32583249e-01 8.23606670e-01 -3.53022009e-01
-9.87945676e-01 -8.69422436e-01 1.62535727e+00 -6.55486643e-01
1.34919596e+00 -5.10686934e-01 -1.14530313e+00 9.38398361e-01
4.02360082e-01 -6.08037710e-02 7.47912407e-01 4.93820399e-01
-5.96573234e-01 7.31007159e-02 -9.95542884e-01 5.17941177e-01
1.15034795e+00 -8.11781824e-01 -9.68726158e-01 3.94329101e-01
7.92331219e-01 -4.07991678e-01 -1.00294030e+00 1.03569813e-01
2.73709387e-01 -6.13346756e-01 6.53722644e-01 -7.20354021e-01
4.18868750e-01 -5.75202107e-01 -4.03166980e-01 -1.18897295e+00
-2.01291382e-01 -4.27698940e-01 3.75985622e-01 1.46657622e+00
7.79380560e-01 -8.93593788e-01 1.06652759e-01 5.08936107e-01
-2.23733962e-01 -2.51253992e-01 -1.09000790e+00 -7.76728868e-01
4.54913557e-01 -8.87143254e-01 3.70385587e-01 1.30496526e+00
1.70367494e-01 9.39991176e-01 6.26310930e-02 7.46044070e-02
4.41596329e-01 1.31290197e-01 5.05277812e-01 -1.46698177e+00
-3.74896377e-01 -4.09628093e-01 -7.16831684e-02 -5.79882622e-01
8.65483463e-01 -1.46651840e+00 -1.46236032e-01 -1.67587531e+00
-9.94509906e-02 -6.61778271e-01 -2.63951063e-01 8.68723512e-01
1.21416569e-01 3.61692190e-01 2.15724498e-01 1.87585562e-01
-6.01542532e-01 2.06853226e-01 8.29428017e-01 2.79926479e-01
-6.78656548e-02 -3.96483243e-01 -8.84162068e-01 1.03681672e+00
9.37476039e-01 -6.82886004e-01 -1.62878990e-01 -8.29316735e-01
5.81285655e-01 -2.53056765e-01 2.06569493e-01 -7.01698959e-01
8.69060010e-02 -2.92565048e-01 -3.88860404e-01 2.11132959e-01
1.28269017e-01 -5.99381268e-01 -1.28692746e-01 1.54470414e-01
-3.30336481e-01 9.13153961e-02 3.29611152e-01 4.73458916e-02
-3.43879879e-01 -6.03014708e-01 7.97484100e-01 -5.12983322e-01
-9.16474164e-01 -9.99504846e-05 -2.10919216e-01 5.63213110e-01
6.29127800e-01 2.20616564e-01 -2.37164721e-01 9.69626084e-02
-5.22554040e-01 1.80430755e-01 7.90443122e-01 7.03855097e-01
-1.60873085e-01 -1.07413602e+00 -8.40260983e-01 -1.37942508e-01
3.68857652e-01 -2.58385371e-02 -1.73486799e-01 4.95404392e-01
-3.72739196e-01 7.85813451e-01 -1.36898771e-01 -1.88185796e-01
-1.06402695e+00 3.68392676e-01 2.99421340e-01 -5.94882905e-01
-3.17841053e-01 7.01641917e-01 4.91951853e-02 -8.63487124e-01
-3.22365493e-01 1.25032077e-02 -3.93404305e-01 1.56461552e-01
2.11416602e-01 -1.10536128e-01 5.92470802e-02 -1.13439035e+00
-6.63892329e-01 5.43496251e-01 3.53747830e-02 -4.91412014e-01
1.35457087e+00 -2.46258527e-01 -6.75681174e-01 9.32324529e-01
8.22720587e-01 3.96219105e-01 -6.74478590e-01 -5.11483848e-01
7.25678802e-01 -8.49653482e-02 -9.66840461e-02 -9.21826601e-01
-6.19364500e-01 6.11636877e-01 -2.49623001e-01 1.09798871e-01
6.83274865e-01 4.38591540e-01 5.32647967e-01 2.82637715e-01
7.20158279e-01 -1.15995860e+00 -5.01103640e-01 1.00265348e+00
9.00927186e-01 -1.18449831e+00 -8.38539600e-02 -4.73362684e-01
-8.04099441e-01 8.07792306e-01 3.05907547e-01 -1.29920021e-01
5.60960531e-01 2.05537751e-01 3.04595321e-01 -3.43098879e-01
-8.73727441e-01 -4.94241387e-01 2.42416710e-01 4.18247879e-01
1.15409970e+00 2.17185736e-01 -8.65905523e-01 6.11148238e-01
-5.35131514e-01 -9.27193761e-02 3.41204941e-01 7.40690649e-01
-1.99245900e-01 -1.89005339e+00 8.89345333e-02 -1.48048744e-01
-8.61469984e-01 -3.69375825e-01 -4.09853876e-01 1.04309511e+00
2.86868036e-01 8.15028191e-01 -2.52236649e-02 3.53692204e-01
3.50137591e-01 5.30919790e-01 6.66471779e-01 -1.32793760e+00
-6.81224465e-01 -1.32843748e-01 6.46112978e-01 -4.42579985e-01
-8.56126904e-01 -1.02092183e+00 -1.39282429e+00 -3.81265680e-04
1.78336218e-01 2.97791392e-01 6.27417266e-01 1.34787488e+00
4.95177396e-02 2.52689511e-01 -7.49121746e-03 -4.61099684e-01
-7.76908621e-02 -8.51703882e-01 -4.16995525e-01 3.12788248e-01
4.17039497e-03 -6.67988718e-01 -1.83476880e-01 1.23117790e-01] | [10.411953926086426, 9.550265312194824] |
93cecec7-471b-4328-b780-e6785c115f59 | progressively-generating-better-initial | 2203.16051 | null | https://arxiv.org/abs/2203.16051v1 | https://arxiv.org/pdf/2203.16051v1.pdf | Progressively Generating Better Initial Guesses Towards Next Stages for High-Quality Human Motion Prediction | This paper presents a high-quality human motion prediction method that accurately predicts future human poses given observed ones. Our method is based on the observation that a good initial guess of the future poses is very helpful in improving the forecasting accuracy. This motivates us to propose a novel two-stage prediction framework, including an init-prediction network that just computes the good guess and then a formal-prediction network that predicts the target future poses based on the guess. More importantly, we extend this idea further and design a multi-stage prediction framework where each stage predicts initial guess for the next stage, which brings more performance gain. To fulfill the prediction task at each stage, we propose a network comprising Spatial Dense Graph Convolutional Networks (S-DGCN) and Temporal Dense Graph Convolutional Networks (T-DGCN). Alternatively executing the two networks helps extract spatiotemporal features over the global receptive field of the whole pose sequence. All the above design choices cooperating together make our method outperform previous approaches by large margins: 6%-7% on Human3.6M, 5%-10% on CMU-MoCap, and 13%-16% on 3DPW. | ['Guiqing Li', 'Qing Zhang', 'Chengjiang Long', 'Yongwei Nie', 'Tiezheng Ma'] | 2022-03-30 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Ma_Progressively_Generating_Better_Initial_Guesses_Towards_Next_Stages_for_High-Quality_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Ma_Progressively_Generating_Better_Initial_Guesses_Towards_Next_Stages_for_High-Quality_CVPR_2022_paper.pdf | cvpr-2022-1 | ['human-pose-forecasting'] | ['computer-vision'] | [-4.74056005e-02 3.16227019e-01 -2.76903391e-01 -2.42498085e-01
-3.59878719e-01 -9.82743502e-03 5.51512539e-01 -7.01185465e-02
-4.76676673e-01 4.30400997e-01 5.26783884e-01 -8.88542272e-03
3.28409880e-01 -7.92554915e-01 -7.08667755e-01 -3.88400763e-01
-3.54147166e-01 3.86109620e-01 7.88710535e-01 -3.14872801e-01
1.54908091e-01 4.03797746e-01 -1.34662068e+00 3.12019497e-01
4.69298750e-01 1.18717980e+00 3.27654481e-01 7.39723563e-01
3.79665434e-01 1.24669242e+00 3.63915823e-02 -3.34958106e-01
2.64762938e-01 -2.85723180e-01 -8.49874258e-01 -6.58972487e-02
7.54442662e-02 -6.13708377e-01 -6.67193532e-01 4.76204783e-01
3.83974344e-01 5.49228728e-01 5.09407640e-01 -1.19042253e+00
-1.47012606e-01 3.71021390e-01 -4.42873359e-01 2.21770912e-01
3.91079634e-01 5.11466861e-01 1.01928031e+00 -7.57908940e-01
5.61880589e-01 1.12829149e+00 8.05910647e-01 6.88526094e-01
-6.61645710e-01 -4.01949346e-01 4.29658413e-01 5.12060761e-01
-1.32116365e+00 -2.56547332e-01 7.62088358e-01 -4.93959218e-01
1.21050894e+00 -5.84469400e-02 8.00399482e-01 1.10065913e+00
5.10324240e-01 1.08303559e+00 5.05187869e-01 2.87731867e-02
1.67668730e-01 -6.29115999e-01 -3.12951654e-02 8.67907286e-01
-2.23703220e-01 2.77647763e-01 -5.17762184e-01 6.57779574e-02
7.53361702e-01 3.25207680e-01 -3.59548956e-01 -2.11258128e-01
-1.33129430e+00 6.13129973e-01 8.88595283e-01 1.24242440e-01
-6.73542261e-01 5.75149059e-01 2.80390859e-01 -1.83036461e-01
3.24736685e-01 2.09073082e-01 -3.42032492e-01 -1.98975503e-01
-1.09307408e+00 3.44214141e-01 3.76754194e-01 7.71415472e-01
7.93102622e-01 -8.17979127e-02 -2.61125982e-01 2.55249709e-01
4.47122544e-01 3.45528930e-01 4.41361248e-01 -8.49296987e-01
6.03248775e-01 5.06859899e-01 3.71280223e-01 -1.38438654e+00
-7.03376234e-01 -4.54680473e-01 -8.02053690e-01 -1.14604868e-01
2.73469239e-01 -2.36821249e-01 -1.06392217e+00 1.71953058e+00
3.72520953e-01 6.85404718e-01 -1.89808428e-01 1.27525043e+00
5.12480617e-01 8.93771529e-01 3.02393794e-01 -3.34032737e-02
1.14790154e+00 -1.46489418e+00 -4.33652282e-01 -4.43577796e-01
7.39387691e-01 -3.51139694e-01 7.57987440e-01 1.93691686e-01
-8.78017187e-01 -9.32149887e-01 -1.00469422e+00 -1.05897173e-01
6.78916574e-02 2.72971869e-01 6.68736339e-01 1.58452228e-01
-1.22692263e+00 9.25189197e-01 -1.29229200e+00 -4.09632802e-01
2.52979547e-01 4.37016666e-01 -3.08722317e-01 -4.75863777e-02
-1.06268311e+00 8.53768229e-01 3.69989067e-01 4.80146945e-01
-8.96190763e-01 -2.56600678e-01 -8.52411807e-01 -9.57597658e-05
3.08384985e-01 -8.22512329e-01 1.06780446e+00 -7.85489440e-01
-1.53720617e+00 3.63928020e-01 -3.61356556e-01 -7.74544001e-01
6.76607966e-01 -5.11124194e-01 -2.45132253e-01 2.12467566e-01
2.97215451e-02 9.93240654e-01 6.86462581e-01 -8.12814534e-01
-7.23706484e-01 -2.79676884e-01 -2.22898889e-02 2.98534423e-01
8.46744552e-02 -4.27596807e-01 -8.50390971e-01 -5.87386370e-01
2.23137096e-01 -1.12874115e+00 -7.90017903e-01 -1.67545632e-01
-2.45265305e-01 -3.51595193e-01 5.60678065e-01 -7.30708778e-01
1.28024447e+00 -1.91313231e+00 3.72104764e-01 1.30070344e-01
3.14911395e-01 2.47738168e-01 -1.57077029e-01 4.50241596e-01
1.65287286e-01 -3.25228602e-01 3.78741398e-02 -6.25737727e-01
-1.09887779e-01 2.34515890e-01 -3.96570772e-01 4.97094542e-01
2.13559791e-01 1.29582965e+00 -8.32932770e-01 -2.95592636e-01
3.22639555e-01 4.49385703e-01 -8.27776551e-01 3.84634465e-01
-3.70770901e-01 5.86874127e-01 -5.59221208e-01 2.02749074e-01
4.02981907e-01 -5.67538500e-01 2.03357950e-01 -1.60445362e-01
-2.74547618e-02 2.82363892e-01 -1.00662851e+00 1.84305775e+00
-1.19101703e-01 3.89699489e-01 -4.82634127e-01 -8.81697595e-01
7.91080952e-01 2.24057421e-01 6.45708799e-01 -6.77760899e-01
1.49076149e-01 5.13029918e-02 -1.52064055e-01 -4.58720684e-01
6.25106096e-01 4.13211957e-02 3.67251262e-02 1.59434527e-01
-3.38280387e-02 5.07098377e-01 -1.70927405e-01 7.93062821e-02
1.15384853e+00 4.91716534e-01 1.82726070e-01 1.73533950e-02
6.16513669e-01 -5.60063962e-03 8.60117316e-01 4.96138960e-01
-5.28553128e-01 7.76530743e-01 4.23102140e-01 -9.36123729e-01
-9.05950487e-01 -6.68217719e-01 6.77941263e-01 1.08287013e+00
3.55610192e-01 -5.48025250e-01 -6.17649853e-01 -7.71887898e-01
-2.29129136e-01 3.44157130e-01 -8.25683355e-01 -1.13939360e-01
-1.04634476e+00 -3.76837641e-01 2.92495281e-01 9.60977674e-01
7.46729672e-01 -1.07526374e+00 -9.12939012e-01 2.55627722e-01
-4.48408812e-01 -1.17173254e+00 -5.49168944e-01 -1.16947219e-01
-8.34643424e-01 -1.01170218e+00 -9.27004039e-01 -6.72814667e-01
4.70330536e-01 1.51733562e-01 9.08674181e-01 3.20389748e-01
2.70046979e-01 5.14921583e-02 -5.49022675e-01 1.66206047e-01
6.85841292e-02 1.72012880e-01 2.74911076e-02 2.07398221e-01
2.11829990e-01 -5.99709213e-01 -1.03669620e+00 2.45781034e-01
-3.39629948e-01 5.60283303e-01 6.74362481e-01 6.71348870e-01
5.60592592e-01 -2.02676192e-01 1.86295047e-01 -6.24408185e-01
5.23538664e-02 -5.50352871e-01 -4.72144902e-01 8.29359517e-02
-4.50483114e-01 2.06063032e-01 7.10065007e-01 -3.30334604e-01
-9.39601123e-01 5.94860315e-01 -5.65254033e-01 -6.30347729e-01
-3.92883383e-02 4.50114220e-01 -1.22979833e-02 1.91449255e-01
5.76334417e-01 2.65433520e-01 -2.11370841e-01 -4.36311483e-01
3.78628612e-01 2.14879870e-01 6.68050766e-01 -5.88573217e-02
6.40850365e-01 5.36048174e-01 7.09189922e-02 -3.83673966e-01
-7.08165169e-01 -5.31286597e-01 -9.32390511e-01 -3.95949870e-01
1.16836238e+00 -1.09020483e+00 -8.81125927e-01 4.25479352e-01
-1.23749483e+00 -7.14945734e-01 1.43894926e-01 6.54846013e-01
-6.66719198e-01 4.83658940e-01 -7.79757500e-01 -7.85215855e-01
-3.76012415e-01 -1.14077413e+00 1.25893998e+00 2.94626169e-02
-1.77984864e-01 -8.11763227e-01 5.38640805e-02 3.20008427e-01
2.40057409e-01 4.13026065e-01 4.36523050e-01 -3.80574733e-01
-8.99774373e-01 -1.37541190e-01 -1.39745429e-01 -1.39947921e-01
-3.13190401e-01 -3.26901793e-01 -7.83881128e-01 -3.26581717e-01
-2.48693883e-01 -1.20020397e-01 1.17155325e+00 5.65273523e-01
9.19189095e-01 -2.48277649e-01 -6.80390954e-01 6.32556021e-01
1.25544226e+00 3.99725735e-02 8.50462377e-01 1.57538429e-01
1.02624583e+00 4.62955207e-01 7.61738479e-01 5.30204654e-01
7.96230912e-01 8.73045325e-01 5.13426900e-01 1.32089391e-01
-1.45394251e-01 -5.74466884e-01 4.80531514e-01 7.80632198e-01
-3.56453627e-01 -3.66057128e-01 -1.06892836e+00 5.96800745e-01
-2.31761432e+00 -9.71410096e-01 -7.90618733e-02 1.87056196e+00
3.78493428e-01 3.04549664e-01 4.21414465e-01 -2.17518583e-02
4.32285279e-01 5.04339695e-01 -3.51752102e-01 3.11855040e-02
3.41693312e-01 1.51939718e-02 4.62898195e-01 6.08533144e-01
-1.23490548e+00 1.25474751e+00 5.67117929e+00 6.13813341e-01
-1.27494609e+00 -3.93577591e-02 7.53700674e-01 4.90132757e-02
-1.16670266e-01 -4.07151319e-03 -7.27071941e-01 3.71659458e-01
9.16141093e-01 2.59653300e-01 2.43411139e-01 9.35374856e-01
3.38827431e-01 -1.99399337e-01 -1.08594239e+00 9.41310942e-01
-2.74273921e-02 -1.48421335e+00 7.52350688e-02 -4.56770766e-04
4.71502185e-01 3.15308690e-01 -1.01689115e-01 3.91650707e-01
1.23685315e-01 -1.06618226e+00 7.65046060e-01 7.74816155e-01
3.95606935e-01 -8.53045106e-01 7.97297359e-01 7.84046650e-01
-1.65445554e+00 -5.10983542e-02 -3.77083451e-01 -4.43234801e-01
3.83504361e-01 3.42667967e-01 -7.54553676e-01 6.27247632e-01
5.75958073e-01 9.95104015e-01 -5.48821449e-01 9.36374605e-01
-3.26961458e-01 5.00137806e-01 -2.43576601e-01 -1.30750090e-01
4.62818056e-01 2.23250255e-01 2.04452768e-01 1.14941716e+00
2.67149150e-01 3.75540704e-01 4.74644989e-01 4.86610562e-01
2.30126381e-01 -1.46175846e-01 -3.39296162e-01 2.24756911e-01
4.20591272e-02 9.85540628e-01 -7.85190344e-01 -4.00500476e-01
-3.34525138e-01 1.27395880e+00 5.94502747e-01 3.72614741e-01
-9.55330729e-01 -5.27343974e-02 4.59089041e-01 1.70386687e-01
6.08616650e-01 -6.01549506e-01 -8.67954344e-02 -1.21168590e+00
3.01033612e-02 -5.20477593e-01 3.32003564e-01 -6.67446136e-01
-8.46578062e-01 7.84871638e-01 -3.05667043e-01 -1.39318466e+00
-6.53552949e-01 -4.98629153e-01 -6.00828171e-01 7.34601617e-01
-1.37768996e+00 -1.26204789e+00 -3.67171675e-01 5.40777564e-01
5.12120664e-01 1.33033335e-01 5.99746883e-01 2.60406643e-01
-4.84934986e-01 3.95899951e-01 -5.69607556e-01 4.06020135e-01
2.40883276e-01 -9.65245605e-01 1.03214264e+00 9.76567626e-01
1.19670585e-01 3.92166913e-01 5.06565213e-01 -8.88114572e-01
-1.29398203e+00 -1.37716031e+00 1.20480025e+00 -4.98315752e-01
4.43272263e-01 -2.10821927e-01 -8.00192773e-01 8.35581839e-01
-4.23744274e-03 3.40282261e-01 2.64316678e-01 -2.47812215e-02
-1.69981867e-01 1.05946362e-01 -6.50988340e-01 8.43096554e-01
1.32318628e+00 -3.70791495e-01 -4.40835029e-01 2.72664011e-01
7.43066728e-01 -6.21278346e-01 -6.65357888e-01 5.10397315e-01
5.44921637e-01 -1.01580191e+00 8.41902196e-01 -5.90388298e-01
5.96623123e-01 -4.04107898e-01 -1.07565232e-01 -9.07125413e-01
-6.03010595e-01 -7.58041978e-01 -3.62952828e-01 4.34456378e-01
3.38935971e-01 -1.31134540e-01 1.24395537e+00 5.96440196e-01
-1.91885233e-01 -1.00580680e+00 -9.40497220e-01 -6.14387631e-01
-1.85836151e-01 -8.01811278e-01 4.41992223e-01 6.04294121e-01
1.34743348e-01 3.37587506e-01 -9.35204029e-01 2.57090777e-01
2.89872587e-01 9.64793786e-02 9.53000188e-01 -7.85321891e-01
-4.77620900e-01 -1.56793550e-01 -6.85805023e-01 -1.90508103e+00
2.83264201e-02 -5.55421710e-01 3.07589352e-01 -1.49735463e+00
9.01782606e-03 -1.38610363e-01 -2.81125993e-01 5.18081605e-01
-3.76014858e-01 1.83690071e-01 3.00986856e-01 4.52754319e-01
-9.80134189e-01 7.78091371e-01 1.28341866e+00 1.14494480e-01
-2.40705788e-01 1.64546698e-01 -1.56248927e-01 7.96992540e-01
5.95949829e-01 -2.45196223e-01 -4.79740173e-01 -5.22593975e-01
-1.50785828e-02 4.19969708e-01 5.50301611e-01 -1.50590694e+00
5.55709958e-01 -1.88704997e-01 4.98566806e-01 -9.35153604e-01
5.11070907e-01 -6.66162431e-01 1.22033238e-01 7.64477193e-01
-2.44290516e-01 1.04682520e-01 -6.03066944e-02 9.30659413e-01
-9.24051180e-02 3.44920039e-01 4.86525059e-01 -6.72653392e-02
-1.31080854e+00 6.97325349e-01 -2.50746071e-01 -2.09390208e-01
1.01192856e+00 -2.57270187e-01 1.55879930e-01 -5.73809028e-01
-9.54065621e-01 4.97878373e-01 2.19617233e-01 4.26145405e-01
7.38712311e-01 -1.38071156e+00 -4.28321421e-01 1.13032758e-01
1.28773190e-02 -9.87185463e-02 3.28511536e-01 1.04338312e+00
-6.21946394e-01 5.93121529e-01 -1.78725183e-01 -7.04591155e-01
-8.96534622e-01 6.07869327e-01 3.85404855e-01 -5.00266016e-01
-8.29710782e-01 1.16789067e+00 2.58651078e-01 -1.22709043e-01
2.80718476e-01 -5.38121998e-01 -3.96819621e-01 -3.14455092e-01
4.83065963e-01 2.58661628e-01 -2.56809592e-01 -9.12075281e-01
-4.89505708e-01 5.56808591e-01 1.28714174e-01 -1.86215222e-01
1.23957801e+00 -2.08717570e-01 3.01624209e-01 2.87321508e-01
1.20696378e+00 -2.68799126e-01 -1.80362749e+00 -3.42975974e-01
1.59081087e-01 -2.41815239e-01 -8.90140086e-02 -4.73819405e-01
-1.22163641e+00 9.64929521e-01 3.88399065e-01 -2.04942331e-01
1.16505325e+00 -1.76169366e-01 1.21711016e+00 4.44435239e-01
6.43902540e-01 -1.00040185e+00 2.57856041e-01 6.83306754e-01
8.40328038e-01 -1.16290534e+00 -8.09472129e-02 -3.90054882e-01
-9.29005861e-01 1.06522536e+00 7.55015969e-01 -3.94885242e-01
6.87475681e-01 -2.24648356e-01 -2.17974335e-01 -1.58378854e-01
-9.16335404e-01 -4.51358438e-01 7.13821590e-01 3.34823936e-01
3.81569654e-01 9.72049534e-02 -2.09698379e-01 7.39815116e-01
-1.15904242e-01 2.77297407e-01 -3.50583754e-02 6.87488794e-01
-6.63995147e-01 -8.77495885e-01 3.66760939e-02 1.52898759e-01
-2.75379241e-01 4.47387062e-02 -1.40594706e-01 7.18840122e-01
1.79909557e-01 8.27469945e-01 1.17655262e-01 -1.05065620e+00
2.71645635e-01 -1.33297533e-01 1.30427256e-01 -4.11405325e-01
-5.36953926e-01 2.58926660e-01 1.09327301e-01 -1.14530325e+00
-4.82203543e-01 -5.54422259e-01 -1.59007239e+00 -4.87514585e-01
-2.31790543e-03 1.15712816e-02 2.43140273e-02 1.13649392e+00
5.08159935e-01 4.78703558e-01 4.50848490e-01 -1.12202883e+00
-2.07000658e-01 -8.92570436e-01 -1.91585571e-01 3.47585589e-01
2.67023653e-01 -6.46602511e-01 -1.08654518e-02 9.43040326e-02] | [7.2687788009643555, -0.19213685393333435] |
5a703cf4-690a-40e8-9e3a-43c28d9e6c89 | trust-an-accurate-and-end-to-end-table | 2208.14687 | null | https://arxiv.org/abs/2208.14687v1 | https://arxiv.org/pdf/2208.14687v1.pdf | TRUST: An Accurate and End-to-End Table structure Recognizer Using Splitting-based Transformers | Table structure recognition is a crucial part of document image analysis domain. Its difficulty lies in the need to parse the physical coordinates and logical indices of each cell at the same time. However, the existing methods are difficult to achieve both these goals, especially when the table splitting lines are blurred or tilted. In this paper, we propose an accurate and end-to-end transformer-based table structure recognition method, referred to as TRUST. Transformers are suitable for table structure recognition because of their global computations, perfect memory, and parallel computation. By introducing novel Transformer-based Query-based Splitting Module and Vertex-based Merging Module, the table structure recognition problem is decoupled into two joint optimization sub-tasks: multi-oriented table row/column splitting and table grid merging. The Query-based Splitting Module learns strong context information from long dependencies via Transformer networks, accurately predicts the multi-oriented table row/column separators, and obtains the basic grids of the table accordingly. The Vertex-based Merging Module is capable of aggregating local contextual information between adjacent basic grids, providing the ability to merge basic girds that belong to the same spanning cell accurately. We conduct experiments on several popular benchmarks including PubTabNet and SynthTable, our method achieves new state-of-the-art results. In particular, TRUST runs at 10 FPS on PubTabNet, surpassing the previous methods by a large margin. | ['Jingdong Wang', 'Jingtuo Liu', 'Kun Yao', 'Zhihui Wang', 'Haojie Li', 'Chengquan Zhang', 'Pengyuan Lv', 'Yuechen Yu', 'Zengyuan Guo'] | 2022-08-31 | null | null | null | null | ['table-recognition'] | ['computer-vision'] | [-3.03421486e-02 -4.44911838e-01 -3.29414248e-01 -2.08238527e-01
-1.10590410e+00 -6.87292635e-01 5.88795356e-02 5.29809356e-01
1.93330366e-02 5.16781926e-01 -2.72845477e-02 -4.59278315e-01
-5.62836155e-02 -1.09853053e+00 -8.93665016e-01 -8.55381846e-01
4.07122038e-02 7.05803990e-01 3.57456625e-01 -2.32795715e-01
4.03491557e-01 6.15738750e-01 -1.30230951e+00 7.35962331e-01
8.63690853e-01 1.52393126e+00 1.29906341e-01 6.23869896e-01
-4.28712338e-01 1.06345856e+00 -4.19720888e-01 -4.54459727e-01
1.37930483e-01 1.58719886e-02 -7.69357324e-01 2.67076045e-01
5.37219524e-01 -3.80185574e-01 -3.18344295e-01 9.47004497e-01
1.77667320e-01 -1.75830349e-01 1.21819831e-01 -1.10499918e+00
-4.09522593e-01 9.15250242e-01 -9.09202456e-01 1.26539439e-01
2.99322546e-01 -1.65448427e-01 1.34420538e+00 -8.03256094e-01
6.64990425e-01 1.20352232e+00 4.76232886e-01 -2.06297338e-01
-1.15091932e+00 -5.02258003e-01 4.34317112e-01 5.17687559e-01
-1.71245050e+00 -3.75014842e-01 6.32815003e-01 -1.08307458e-01
1.06907678e+00 4.62959081e-01 3.29958647e-01 2.42195264e-01
3.38315666e-01 9.70912099e-01 7.01133192e-01 -1.26579646e-02
4.40012477e-02 -2.26707961e-02 1.14932314e-01 9.86908197e-01
2.71854639e-01 -7.63672173e-01 -6.33147657e-01 8.33259672e-02
8.14385295e-01 1.24595404e-01 -1.31694168e-01 -7.13778079e-01
-1.44179761e+00 5.07625222e-01 6.23999000e-01 1.45032033e-01
-2.17511848e-01 -1.14522077e-01 5.19697547e-01 1.81906551e-01
6.49899393e-02 -3.05058584e-02 -5.08938730e-01 -1.78653345e-01
-7.67434001e-01 1.98658884e-01 9.03666973e-01 1.19220245e+00
9.29268420e-01 -3.52581769e-01 -3.85259122e-01 5.31054080e-01
2.31375601e-02 5.25878668e-01 -1.76709592e-01 -4.21456873e-01
1.25547314e+00 1.19451582e+00 -1.44537643e-01 -1.43109155e+00
-3.94089550e-01 -3.94760758e-01 -1.32301915e+00 -2.35613406e-01
4.09939080e-01 2.98402101e-01 -7.95652151e-01 1.21421218e+00
5.08082867e-01 -2.68106908e-01 -2.35217549e-02 5.47216475e-01
7.29103863e-01 9.22893047e-01 -4.17214900e-01 -3.33931506e-01
1.77855945e+00 -1.05802977e+00 -6.84989691e-01 -7.40600675e-02
9.73669767e-01 -8.22256804e-01 9.46411014e-01 5.03045619e-01
-1.20284402e+00 -4.60989416e-01 -1.18396795e+00 -4.94159490e-01
-5.32127917e-01 3.60383749e-01 6.25932455e-01 3.47533554e-01
-7.48170555e-01 1.95548907e-01 -8.00009370e-01 9.26945359e-02
5.48850954e-01 5.40169656e-01 -4.97179389e-01 -1.44083932e-01
-7.41544306e-01 3.89721125e-01 3.83995712e-01 3.54972839e-01
-2.97850013e-01 -4.31150496e-01 -9.78792965e-01 4.51043665e-01
8.84026766e-01 -3.92867565e-01 9.01711583e-01 -2.49386609e-01
-9.50251281e-01 7.64396846e-01 -5.38825214e-01 -3.22999239e-01
2.37667948e-01 9.54839066e-02 -2.71069050e-01 -5.54638095e-02
2.74738729e-01 2.94956088e-01 3.84340078e-01 -1.09839904e+00
-9.42855597e-01 -8.96595359e-01 -9.79851782e-02 4.37514931e-01
-1.60343513e-01 -2.95655042e-01 -1.26046252e+00 -5.56164801e-01
5.35662353e-01 -4.75941569e-01 6.77314699e-02 -1.06358774e-01
-9.29468274e-01 -1.23217732e-01 8.76844049e-01 -8.06665421e-01
1.65543246e+00 -2.11850142e+00 6.39430806e-02 3.20555359e-01
4.51954812e-01 -3.65513787e-02 3.60639662e-01 3.91072124e-01
1.47912949e-01 4.26531471e-02 -1.75777823e-01 -2.39813954e-01
5.10543659e-02 7.97249749e-02 -4.53900605e-01 2.70565689e-01
-2.65951324e-02 9.03928101e-01 -5.40439308e-01 -6.90862000e-01
8.65427181e-02 2.69847989e-01 -4.61799294e-01 -5.18570840e-02
-2.72721559e-01 -8.84183943e-02 -4.74286318e-01 9.52597201e-01
9.99169767e-01 -5.98922431e-01 4.97591257e-01 -6.60934567e-01
-1.15545973e-01 4.59416091e-01 -1.67811465e+00 1.60430419e+00
-2.16143385e-01 2.42595673e-01 1.86557740e-01 -1.02191150e+00
9.13258731e-01 -1.04975760e-01 3.86168271e-01 -1.02992511e+00
9.17906314e-02 1.41336262e-01 -2.98396260e-01 -4.23929803e-02
6.45441175e-01 5.47571957e-01 -3.28315049e-01 3.77200335e-01
-4.80316818e-01 1.01114340e-01 4.70958978e-01 4.56599772e-01
9.12128389e-01 -2.72697695e-02 2.70617485e-01 -1.30023181e-01
9.29033935e-01 8.59701335e-02 6.92630053e-01 4.66970086e-01
1.39427289e-01 5.30377984e-01 9.78050113e-01 -6.36349618e-01
-8.58017147e-01 -9.96867239e-01 1.42606497e-01 9.05401289e-01
5.25297821e-01 -8.30367744e-01 -8.23322237e-01 -7.03644156e-01
8.79527330e-02 2.41406992e-01 -6.04567409e-01 7.54428804e-02
-9.15358007e-01 -6.81561708e-01 1.78756639e-01 8.89578640e-01
7.72178233e-01 -8.02733719e-01 -2.47528672e-01 3.72726291e-01
-4.84832436e-01 -1.42817390e+00 -8.94005477e-01 5.00330091e-01
-7.71193445e-01 -1.26834810e+00 -1.56577796e-01 -1.01822233e+00
7.47984529e-01 4.34623480e-01 1.14567459e+00 2.93938905e-01
-6.06680997e-02 -4.00893956e-01 2.05751769e-02 1.28312901e-01
4.60796570e-03 4.91502315e-01 -3.63921702e-01 1.54239804e-01
2.10119516e-01 -3.14044267e-01 -3.96759748e-01 6.76810443e-01
-8.11861694e-01 4.26959604e-01 5.90155065e-01 7.36726165e-01
1.20639968e+00 4.43437934e-01 -9.46120024e-02 -1.03690720e+00
2.29466364e-01 1.45709971e-02 -1.00096297e+00 6.87258244e-01
-5.40150344e-01 3.37682128e-01 9.39941406e-01 4.24980260e-02
-8.09287727e-01 1.35381937e-01 -3.87635417e-02 -1.60721883e-01
1.70177057e-01 4.63282883e-01 -8.58546853e-01 1.42802507e-01
-3.70003991e-02 5.29201150e-01 -3.18601161e-01 -4.86801267e-01
1.71559513e-01 5.16497791e-01 7.53250599e-01 -6.13916218e-01
9.14838433e-01 5.81615388e-01 6.73790649e-02 -3.22626531e-01
-7.39770710e-01 -3.25416893e-01 -8.35575104e-01 3.10354859e-01
7.25914538e-01 -8.27502191e-01 -1.27262330e+00 7.19644427e-01
-9.46831286e-01 -8.62055197e-02 1.35851782e-02 -1.99642554e-01
-3.46648902e-01 4.57163900e-01 -7.67122984e-01 -3.68780911e-01
-4.67323512e-01 -1.62466419e+00 1.27415955e+00 1.45328119e-01
3.33824277e-01 -6.62126839e-01 -3.32747549e-01 6.54699743e-01
1.78346727e-02 5.15614785e-02 1.29914320e+00 -4.68865395e-01
-1.27545130e+00 -1.24889106e-01 -6.92794383e-01 -9.29654390e-02
1.91774294e-01 9.16952342e-02 -4.91629601e-01 -1.79116562e-01
-3.24492604e-01 -7.49659464e-02 8.00865173e-01 1.76779822e-01
1.32749677e+00 -3.00281912e-01 -5.20078480e-01 9.96378064e-01
1.44529617e+00 4.28207517e-01 7.82476604e-01 4.81090933e-01
1.19266665e+00 2.26935104e-01 6.97490096e-01 4.46851492e-01
9.11405325e-01 8.72536004e-01 3.82662416e-01 -4.82365564e-02
1.60083383e-01 -3.92050803e-01 8.31102133e-02 8.11232567e-01
4.51011240e-01 -6.17178202e-01 -8.31316233e-01 2.48590335e-01
-1.90165925e+00 -5.65325618e-01 -1.75501674e-01 2.29437399e+00
7.48223186e-01 5.91426969e-01 1.33408070e-01 3.78167599e-01
8.25621724e-01 3.41058016e-01 -5.98339796e-01 -3.57490480e-01
-3.00013006e-01 -1.38201311e-01 6.88535154e-01 3.88183028e-01
-1.23146415e+00 9.84902203e-01 5.17054749e+00 1.12973988e+00
-8.56331706e-01 -4.71267104e-01 1.25910580e+00 1.60274640e-01
-1.39090359e-01 -1.99011993e-02 -1.24023926e+00 2.31435478e-01
3.79441530e-01 1.83351487e-01 6.55090332e-01 6.38527870e-01
-2.35511616e-01 -1.95907354e-01 -1.13773441e+00 1.25581169e+00
1.07259579e-01 -1.61246777e+00 2.55357593e-01 1.72453582e-01
3.30348134e-01 -5.57541847e-01 3.99968773e-02 1.15572259e-01
-3.48849110e-02 -8.88354421e-01 7.21261621e-01 1.20035507e-01
1.00250399e+00 -1.03728974e+00 7.49861479e-01 1.01373740e-01
-1.91563654e+00 8.42558406e-03 -2.59043902e-01 1.48753032e-01
8.63287656e-04 4.67320889e-01 -4.66174275e-01 7.29758918e-01
8.72255862e-01 7.87688017e-01 -6.30083084e-01 6.89479530e-01
4.13031392e-02 1.67022631e-01 -4.25361335e-01 6.09509237e-02
1.99796915e-01 -4.29101914e-01 1.23513140e-01 8.13099205e-01
8.60224590e-02 2.63909753e-02 2.81070173e-01 6.74683928e-01
-4.27820057e-01 2.09876850e-01 -2.48141840e-01 2.10906751e-02
5.37636697e-01 1.39142776e+00 -1.33195281e+00 -5.05396307e-01
-4.39060330e-01 8.44806433e-01 4.68126982e-01 -1.13334954e-02
-8.92243326e-01 -5.69818437e-01 5.43056548e-01 -5.68246171e-02
8.65974665e-01 -1.40869781e-01 -9.08447683e-01 -1.27930236e+00
6.07147157e-01 -1.21336520e+00 6.59445941e-01 -6.48668885e-01
-7.22362339e-01 7.20818579e-01 -2.84583211e-01 -1.02612090e+00
1.55420005e-01 -6.76150203e-01 -3.21413428e-01 7.17326641e-01
-1.22890997e+00 -1.05681109e+00 -4.98277634e-01 7.41249442e-01
3.64065647e-01 9.62945167e-03 5.44921994e-01 3.97588819e-01
-1.01999390e+00 9.51750934e-01 2.26878509e-01 4.61198509e-01
5.15462756e-01 -1.32786262e+00 7.95958042e-01 9.19782817e-01
3.12305778e-01 3.85677755e-01 2.84067214e-01 -5.67224801e-01
-1.96241224e+00 -8.35264266e-01 9.17617559e-01 -3.50967467e-01
4.71559346e-01 -9.15668130e-01 -1.18117034e+00 6.35033429e-01
-1.51323259e-01 6.80984408e-02 1.35849565e-01 6.98690265e-02
-6.18375719e-01 -8.54590654e-01 -8.74696076e-01 5.98979712e-01
9.97953594e-01 -4.50545877e-01 -1.41165584e-01 4.24909443e-01
5.76123416e-01 -9.45500791e-01 -8.66899610e-01 2.54314274e-01
5.07900953e-01 -1.16522670e+00 1.07779038e+00 -9.81862247e-02
3.72454166e-01 -4.44024891e-01 -2.08728716e-01 -5.46715677e-01
-3.53870243e-01 -5.17135620e-01 -2.49153674e-01 1.33808279e+00
4.17907953e-01 -5.65424502e-01 1.20379686e+00 3.19456011e-01
5.29740900e-02 -1.14156950e+00 -7.85989463e-01 -4.67353165e-01
-3.37246209e-01 -3.05596977e-01 1.06365812e+00 6.96081221e-01
-1.18681304e-01 5.22590280e-01 -2.32934013e-01 3.70345414e-01
3.78894538e-01 7.42146075e-01 9.53205705e-01 -9.46515560e-01
-6.51989728e-02 -6.27136290e-01 -3.38448793e-01 -1.49596179e+00
-3.93870890e-01 -6.67428374e-01 -1.03405043e-01 -1.66265595e+00
3.48995805e-01 -4.79389668e-01 -2.43690118e-01 5.52509010e-01
-1.45043373e-01 8.94329026e-02 2.49290973e-01 1.73302129e-01
-8.73658061e-01 3.00455958e-01 1.37492657e+00 -4.11447167e-01
-2.34370619e-01 -2.57651091e-01 -9.77196038e-01 3.97220492e-01
5.15268803e-01 -3.84245753e-01 -2.89580882e-01 -4.96185243e-01
5.10084212e-01 3.77276510e-01 -6.73071519e-02 -9.86221313e-01
5.37681162e-01 -1.29872695e-01 5.24086773e-01 -1.34567571e+00
1.49387822e-01 -7.83461034e-01 -7.14031830e-02 2.44935378e-01
-3.45659889e-02 6.48994863e-01 3.38826478e-01 4.21409965e-01
-4.03620988e-01 4.60810661e-01 4.89098459e-01 8.25596452e-02
-5.68129539e-01 6.87736273e-01 -3.20262872e-02 1.27098590e-01
8.15556943e-01 -5.12772679e-01 -4.13395345e-01 -9.46991742e-02
-3.57393563e-01 4.72775638e-01 5.49357355e-01 3.12066048e-01
5.75584173e-01 -1.28021073e+00 -4.33213681e-01 5.79236567e-01
2.30589956e-01 5.39430678e-01 3.69589001e-01 9.19346750e-01
-8.84824097e-01 6.53323770e-01 6.47087693e-02 -7.15784729e-01
-1.44924223e+00 7.65002906e-01 2.24512547e-01 -9.45250273e-01
-5.71202457e-01 8.08357239e-01 4.88283604e-01 -2.65903950e-01
4.74603683e-01 -7.16536880e-01 -1.00252405e-01 -6.98957145e-02
6.29260838e-01 1.25491500e-01 5.06712794e-01 -6.11089230e-01
-5.72577357e-01 6.77311480e-01 -6.87142491e-01 4.47869629e-01
1.04669273e+00 -1.47106797e-01 -3.74244660e-01 2.43264809e-01
1.07024372e+00 1.03001393e-01 -1.03237188e+00 -3.61764729e-01
-5.42714559e-02 -4.72641647e-01 2.04820093e-02 -8.06259573e-01
-1.35587001e+00 8.51048231e-01 2.77425855e-01 2.85786092e-01
1.36355722e+00 -3.34480882e-01 1.20448899e+00 6.44205213e-01
3.65048915e-01 -1.01700866e+00 -1.79581493e-01 5.68289340e-01
5.48199534e-01 -1.06771827e+00 1.96830958e-01 -8.34065855e-01
-5.88695824e-01 1.14791179e+00 6.79966867e-01 2.30709612e-01
4.21523124e-01 6.36840403e-01 -1.25663370e-01 -2.56140709e-01
-8.91236901e-01 4.07530740e-02 4.13628668e-01 1.64217472e-01
2.41401702e-01 -6.25022035e-03 2.45821610e-01 5.37396252e-01
-9.45049450e-02 -2.93163627e-01 1.78954050e-01 1.10655272e+00
-3.10544580e-01 -1.13294327e+00 -5.59852302e-01 5.94460130e-01
-3.17447752e-01 -2.73036271e-01 -2.00061992e-01 8.22562397e-01
-1.36849344e-01 6.75914824e-01 2.53414810e-01 -5.51227510e-01
3.95744264e-01 -2.73206890e-01 3.54420424e-01 -3.30542296e-01
-5.83863616e-01 2.73171157e-01 -6.47946745e-02 -6.91475689e-01
2.30382413e-01 -6.00591123e-01 -1.51185131e+00 -7.54787922e-01
-1.75392061e-01 1.34845465e-01 2.38287851e-01 9.23633873e-01
6.40782356e-01 5.22895992e-01 5.80115557e-01 -2.44817063e-01
-2.48040348e-01 -5.10707557e-01 -6.99926794e-01 3.11505169e-01
3.26697290e-01 -4.84495133e-01 3.87943350e-02 -6.05599955e-04] | [11.711078643798828, 3.0591719150543213] |
fe6eb47d-9611-42cb-bd5e-fae44a20580f | spice-semantic-pseudo-labeling-for-image | 2103.09382 | null | https://arxiv.org/abs/2103.09382v3 | https://arxiv.org/pdf/2103.09382v3.pdf | SPICE: Semantic Pseudo-labeling for Image Clustering | The similarity among samples and the discrepancy between clusters are two crucial aspects of image clustering. However, current deep clustering methods suffer from the inaccurate estimation of either feature similarity or semantic discrepancy. In this paper, we present a Semantic Pseudo-labeling-based Image ClustEring (SPICE) framework, which divides the clustering network into a feature model for measuring the instance-level similarity and a clustering head for identifying the cluster-level discrepancy. We design two semantics-aware pseudo-labeling algorithms, prototype pseudo-labeling, and reliable pseudo-labeling, which enable accurate and reliable self-supervision over clustering. Without using any ground-truth label, we optimize the clustering network in three stages: 1) train the feature model through contrastive learning to measure the instance similarity, 2) train the clustering head with the prototype pseudo-labeling algorithm to identify cluster semantics, and 3) jointly train the feature model and clustering head with the reliable pseudo-labeling algorithm to improve the clustering performance. Extensive experimental results demonstrate that SPICE achieves significant improvements (~10%) over existing methods and establishes the new state-of-the-art clustering results on six image benchmark datasets in terms of three popular metrics. Importantly, SPICE significantly reduces the gap between unsupervised and fully-supervised classification; e.g., there is only a 2% (91.8% vs 93.8%) accuracy difference on CIFAR-10. Our code has been made publically available at https://github.com/niuchuangnn/SPICE. | ['Hongming Shan', 'Ge Wang', 'Chuang Niu'] | 2021-03-17 | spice-semantic-pseudo-labeling-for-image-1 | https://arxiv.org/abs/2103.09382 | https://arxiv.org/pdf/2103.09382 | null | ['image-clustering'] | ['computer-vision'] | [-1.62031367e-01 -1.36803463e-01 -1.19181097e-01 -8.27200294e-01
-1.11265087e+00 -2.92401731e-01 3.89330000e-01 6.35766014e-02
-4.67783600e-01 6.50714636e-02 -1.23744942e-01 1.77240059e-01
-1.64591651e-02 -4.04974610e-01 -6.02758169e-01 -1.04789937e+00
-9.95274186e-02 6.96582615e-01 2.08384886e-01 4.23880160e-01
9.07447562e-02 -1.48732796e-01 -1.55702055e+00 4.55710351e-01
1.14911342e+00 1.22943544e+00 4.09698248e-01 2.05117360e-01
-7.81164020e-02 7.29822159e-01 -4.49922770e-01 4.06034514e-02
7.71187991e-02 -5.56811810e-01 -8.19317639e-01 5.43810189e-01
1.36828721e-01 4.15863357e-02 -7.76849315e-02 1.31403553e+00
3.14464122e-01 -2.43380554e-02 6.26618028e-01 -1.49550414e+00
-6.25030816e-01 8.43610525e-01 -8.72130752e-01 -2.35214662e-02
-2.60242224e-01 1.40406772e-01 8.93287241e-01 -8.87634099e-01
2.66465843e-01 1.11454093e+00 5.31608641e-01 5.20474732e-01
-1.16546965e+00 -1.01357031e+00 1.63011461e-01 2.75586993e-01
-1.85749817e+00 -3.53086621e-01 7.50473320e-01 -5.72546840e-01
4.76804137e-01 -2.34065056e-02 2.83520073e-01 7.18037605e-01
-4.93995875e-01 8.27515066e-01 1.21352601e+00 -6.06346317e-02
4.57764566e-01 1.41175687e-01 3.69592071e-01 6.07923388e-01
1.56404916e-02 -1.87662229e-01 -1.30992621e-01 1.20944083e-01
4.02044892e-01 1.53486416e-01 -2.70851195e-01 -3.65189105e-01
-1.13661957e+00 7.87584722e-01 8.89219701e-01 3.67792606e-01
-1.52793065e-01 2.60718852e-01 3.76711428e-01 -1.27642304e-01
4.24973696e-01 2.02186614e-01 -3.48711550e-01 1.29397959e-01
-1.06721377e+00 -2.28410080e-01 3.90377671e-01 1.02672613e+00
9.36499596e-01 -3.43397707e-01 -7.92840347e-02 9.38896358e-01
5.03697336e-01 3.08724225e-01 6.08329058e-01 -1.18557942e+00
2.90502787e-01 7.99900949e-01 -1.82667479e-01 -1.04052460e+00
-4.28837597e-01 -6.73716843e-01 -1.01903999e+00 -1.79045036e-01
2.53038198e-01 5.58780693e-02 -9.53907609e-01 1.87327766e+00
4.28282112e-01 6.04359329e-01 -9.71049294e-02 1.06724298e+00
7.77262688e-01 6.64374232e-01 1.55054361e-01 -9.88118276e-02
1.26616859e+00 -1.28499484e+00 -4.76621687e-01 -1.16140559e-01
7.48535872e-01 -5.75027525e-01 1.25101829e+00 3.08736086e-01
-7.91692615e-01 -7.09081173e-01 -1.02602720e+00 1.18102267e-01
-2.60432720e-01 3.66808325e-01 3.80274981e-01 2.94460207e-01
-1.07415068e+00 4.70138311e-01 -1.09478581e+00 -1.62900999e-01
7.71809578e-01 5.72410166e-01 -1.43473387e-01 -2.41230994e-01
-8.55072796e-01 2.22282410e-01 6.58840537e-01 -4.05746847e-02
-9.20039237e-01 -6.38700068e-01 -7.33243883e-01 1.39394715e-01
3.30571592e-01 -3.37716967e-01 8.73408794e-01 -1.30367994e+00
-1.14283717e+00 1.10896242e+00 -1.52520970e-01 -2.66437143e-01
1.84054613e-01 1.46111786e-01 -3.70493233e-01 2.59856343e-01
4.27955419e-01 1.06992853e+00 5.29273093e-01 -1.67318785e+00
-7.69201458e-01 -4.19057399e-01 -5.65755725e-01 1.99506521e-01
-3.69901597e-01 -2.63267398e-01 -1.06321931e+00 -4.50963527e-01
3.24263185e-01 -8.80376935e-01 -1.95187688e-01 -2.27181286e-01
-5.88996768e-01 -4.12913948e-01 7.94151902e-01 -3.74057978e-01
1.15656626e+00 -2.35524964e+00 -9.10570696e-02 3.26483786e-01
4.08468038e-01 1.66546747e-01 -3.09942633e-01 -1.32178217e-01
-1.47033542e-01 1.14596859e-01 -5.86736262e-01 -6.59183204e-01
1.94136035e-02 3.39125805e-02 2.84415055e-02 6.26782656e-01
1.41185701e-01 8.32241058e-01 -9.99702275e-01 -4.97722030e-01
2.95439065e-01 5.48158586e-01 -4.10110414e-01 2.18148872e-01
-3.42875272e-02 5.96169412e-01 -2.43680507e-01 5.29802561e-01
7.68099248e-01 -7.51935005e-01 2.51612723e-01 -3.44209611e-01
1.26049653e-01 2.12616116e-01 -1.06534755e+00 1.82949853e+00
-1.80629969e-01 3.69168937e-01 -1.96997095e-02 -1.18280232e+00
7.51531065e-01 3.43929939e-02 6.32161260e-01 -7.27537990e-01
2.85237789e-01 9.65368897e-02 -8.90613422e-02 -2.13219076e-01
2.89375056e-02 6.90326095e-02 -1.80003077e-01 5.18779576e-01
8.55721682e-02 3.46405506e-01 9.57294330e-02 3.85690629e-01
8.82663429e-01 -2.82816052e-01 -2.19071761e-01 -5.52838027e-01
4.59907144e-01 3.11079086e-03 9.16326284e-01 3.82932752e-01
-3.54040086e-01 8.69598806e-01 3.65887910e-01 -8.59533548e-02
-7.54798293e-01 -1.11355340e+00 -1.17855050e-01 8.60697210e-01
5.80547988e-01 -4.84949738e-01 -1.15916133e+00 -8.23007226e-01
-1.14630274e-01 5.64546108e-01 -6.19985223e-01 -3.78104180e-01
-1.71662062e-01 -9.22280133e-01 4.68921185e-01 6.02926493e-01
8.53426993e-01 -7.69930780e-01 -1.39563754e-01 -6.85174316e-02
-3.46361607e-01 -1.15780962e+00 -8.15866888e-01 1.36365846e-01
-6.39902234e-01 -1.15177190e+00 -1.59729660e-01 -1.16290927e+00
1.00531590e+00 6.53648198e-01 1.08709013e+00 5.06943524e-01
-1.34679228e-01 1.04001410e-01 -3.85638356e-01 1.52398571e-01
-1.49873435e-01 7.35202953e-02 -5.88057637e-02 9.77818817e-02
6.93668008e-01 -4.87396181e-01 -8.91193986e-01 5.27052581e-01
-7.94927537e-01 1.10279597e-01 5.38175046e-01 7.25721836e-01
9.53245044e-01 4.57219988e-01 7.37953067e-01 -1.05930018e+00
3.07956021e-02 -6.39874816e-01 -5.62238812e-01 1.63691461e-01
-7.92702198e-01 -4.60277162e-02 5.95083475e-01 -2.64725238e-01
-8.04020822e-01 3.90165567e-01 -4.06433791e-02 -6.38779938e-01
-3.55399609e-01 3.27980936e-01 -5.21162987e-01 2.82218277e-01
4.09854412e-01 1.47178307e-01 -2.67851539e-02 -3.86181027e-01
4.02090043e-01 9.10472572e-01 8.80226135e-01 -5.29572487e-01
7.58379221e-01 7.31639087e-01 -5.97666740e-01 -3.45018059e-01
-1.09625566e+00 -9.69560504e-01 -7.95289159e-01 -3.41414474e-02
1.05488503e+00 -1.32999444e+00 -6.67854607e-01 5.39291203e-01
-5.96175730e-01 -7.12145865e-01 -1.73603036e-02 3.85576934e-01
-4.22125250e-01 3.63081008e-01 -6.42798901e-01 -2.52390444e-01
-3.11914176e-01 -1.39109313e+00 1.12178206e+00 3.33672523e-01
1.29554734e-01 -9.84741509e-01 -3.52865666e-01 5.93371987e-01
8.88086930e-02 2.91102260e-01 6.49712801e-01 -6.11908376e-01
-4.97006744e-01 2.21491512e-02 -5.60389876e-01 4.92021084e-01
2.95959383e-01 -3.62121756e-03 -1.08941078e+00 -4.62805241e-01
-7.23697990e-02 -4.48133796e-01 1.10980892e+00 4.19685066e-01
1.67137384e+00 -1.09956875e-01 -5.04703403e-01 8.74938309e-01
1.52131474e+00 1.24460138e-01 4.51794565e-01 7.30153993e-02
9.02734697e-01 5.12310743e-01 6.63633347e-01 2.88168877e-01
8.42717290e-01 5.34523249e-01 4.18322206e-01 -1.22044548e-01
-3.58778059e-01 -2.69695431e-01 2.67260134e-01 1.12997437e+00
5.36448061e-01 -5.80682456e-02 -1.03185987e+00 6.55217528e-01
-2.07790756e+00 -6.63857102e-01 -1.93317726e-01 2.15943074e+00
9.81315196e-01 9.52232257e-02 1.67795449e-01 3.92255485e-02
1.18459737e+00 -1.74760252e-01 -7.08548307e-01 3.77726883e-01
4.36821580e-02 -1.29575565e-01 2.78957039e-01 5.24923861e-01
-1.46093524e+00 1.13521397e+00 4.51657867e+00 1.04568291e+00
-1.15133631e+00 3.14460456e-01 1.13038850e+00 2.85910089e-02
-6.89014941e-02 3.10886186e-02 -5.99572301e-01 8.91814530e-01
8.28311622e-01 1.56733006e-01 3.89218688e-01 8.51603687e-01
1.10241599e-01 -2.43371762e-02 -1.09134424e+00 1.21109557e+00
1.34533063e-01 -1.14854193e+00 -6.94839880e-02 1.86241344e-02
8.56797338e-01 3.09365898e-01 8.52923542e-02 2.47388452e-01
4.27425981e-01 -9.99298871e-01 7.20357597e-01 2.32396498e-01
7.72020280e-01 -9.86823082e-01 7.12551117e-01 3.54701608e-01
-1.27920616e+00 8.48420188e-02 -4.54129636e-01 2.43752360e-01
-1.87484026e-01 9.72803175e-01 -4.49387759e-01 3.27590257e-01
1.07186258e+00 1.06882048e+00 -6.86160088e-01 9.25279498e-01
-3.05896342e-01 8.85329783e-01 -2.81876624e-01 3.76942754e-01
3.11694086e-01 -4.40680772e-01 -5.82634248e-02 1.18567348e+00
3.83493230e-02 -1.82870198e-02 5.65956771e-01 1.10520101e+00
-3.20191205e-01 -2.72480100e-01 3.05259097e-02 2.20516056e-01
9.73288894e-01 1.56792915e+00 -1.24223697e+00 -5.31306744e-01
-1.52181923e-01 1.16369593e+00 4.74383026e-01 2.41457030e-01
-1.06188011e+00 -4.29284275e-01 6.14867568e-01 -1.62019283e-01
4.26467210e-01 1.61801185e-02 -4.06669945e-01 -1.06706202e+00
-1.55619919e-01 -6.57140493e-01 4.89014566e-01 -6.34275854e-01
-1.45800686e+00 6.53494596e-01 -1.92629889e-01 -1.07977223e+00
8.07220489e-02 -2.16172442e-01 -6.89401567e-01 4.64340597e-01
-1.33568358e+00 -1.08600330e+00 -6.88927174e-01 6.42116010e-01
4.47683096e-01 5.12639508e-02 5.08010864e-01 5.54132223e-01
-1.01743364e+00 8.25545967e-01 2.18616590e-01 5.06369889e-01
7.58901060e-01 -1.41526091e+00 1.17947690e-01 7.69352198e-01
6.79347664e-02 5.49122870e-01 2.49654457e-01 -3.66105109e-01
-8.73356164e-01 -1.75388360e+00 5.31784058e-01 -3.86653870e-01
4.44647998e-01 -6.80224717e-01 -1.03523076e+00 5.05349755e-01
4.56808880e-02 2.38641426e-01 8.93428981e-01 -1.94573656e-01
-6.08943760e-01 -3.54811877e-01 -1.10782111e+00 2.92785436e-01
9.55031455e-01 -4.70623016e-01 -2.31205285e-01 4.13914204e-01
9.09655869e-01 -1.11472934e-01 -1.01383948e+00 3.16006690e-01
1.18768156e-01 -1.00514245e+00 8.14416587e-01 -2.89870203e-02
2.47686878e-01 -8.44084203e-01 -2.44511873e-01 -1.24647164e+00
-4.82309937e-01 -9.13445875e-02 2.28509441e-01 1.60719573e+00
2.89313912e-01 -3.80114317e-01 7.30827689e-01 4.08718050e-01
-3.63772959e-01 -8.17964315e-01 -6.11011744e-01 -7.44407356e-01
1.82958439e-01 -4.08560604e-01 6.73891187e-01 1.41711795e+00
-9.75407213e-02 4.65020061e-01 1.23724498e-01 3.74533802e-01
9.24698651e-01 3.32232416e-01 5.07637203e-01 -1.03373837e+00
-1.02860108e-01 -6.09441400e-01 -3.46109390e-01 -9.61576402e-01
4.11048800e-01 -1.30119574e+00 3.73776227e-01 -1.38239658e+00
6.40265822e-01 -7.62674272e-01 -4.69342947e-01 6.30860269e-01
-3.84999633e-01 4.65618879e-01 -2.34394125e-03 5.76585770e-01
-1.33894062e+00 6.41157866e-01 8.04937780e-01 -2.09082976e-01
-4.64215949e-02 -2.29953498e-01 -9.68420982e-01 6.60042405e-01
1.02574182e+00 -7.03696311e-01 -5.50440192e-01 -4.50314194e-01
-2.11167485e-01 -5.61287284e-01 4.41081703e-01 -1.24000072e+00
4.32302535e-01 2.28370860e-01 4.16736990e-01 -5.63963234e-01
-1.27544180e-01 -6.60676718e-01 -5.16485125e-02 4.60420072e-01
-4.64723736e-01 -1.84799224e-01 -1.11086383e-01 6.75327718e-01
-3.90583038e-01 1.72337338e-01 1.15447104e+00 7.05365986e-02
-8.25976074e-01 4.56910789e-01 -2.13896893e-02 2.69234449e-01
1.09700263e+00 2.55145896e-02 -4.09811765e-01 -1.70355767e-01
-5.82155585e-01 8.10263813e-01 7.51444578e-01 2.70936489e-01
4.65217113e-01 -1.43620908e+00 -5.23927987e-01 7.17640147e-02
2.54739434e-01 4.48779285e-01 2.68906325e-01 1.05394113e+00
-3.46752495e-01 1.88375667e-01 2.65853733e-01 -1.09287786e+00
-1.01784408e+00 7.31659472e-01 4.12412882e-01 -3.01045030e-02
-4.17457968e-01 9.44549382e-01 5.66145957e-01 -7.26876795e-01
4.58936751e-01 -1.78119749e-01 8.11207294e-02 -2.31879205e-01
5.42113423e-01 2.04601884e-01 2.85584070e-02 -7.84829080e-01
-5.93077481e-01 6.65585518e-01 -1.76109120e-01 2.22466469e-01
1.16913784e+00 -3.02455842e-01 -3.22648555e-01 4.34053630e-01
1.63060319e+00 -5.94366729e-01 -1.55513096e+00 -3.29262763e-01
2.72185802e-01 -2.37622634e-01 2.05578879e-01 -8.33630860e-01
-1.64583004e+00 8.99355948e-01 8.13346982e-01 -1.20385729e-01
1.37603092e+00 4.24920291e-01 8.97367895e-01 9.00464654e-02
2.15536267e-01 -1.23771727e+00 2.26708919e-01 1.83263555e-01
3.77388835e-01 -1.52638090e+00 -1.37652129e-01 -4.73744661e-01
-8.71205628e-01 4.97039199e-01 8.56061578e-01 -3.13599020e-01
9.99437749e-01 1.68469965e-01 3.49928111e-01 -4.49301928e-01
-8.35906208e-01 -3.27178746e-01 3.04917216e-01 6.24298751e-01
4.59438801e-01 2.78481007e-01 5.85105307e-02 9.30092454e-01
-9.87316370e-02 -3.48182201e-01 7.96576962e-02 5.13007641e-01
-5.09216487e-01 -8.76029193e-01 -1.11245178e-01 4.04005796e-01
-1.82494611e-01 -2.98010632e-02 -4.41667110e-01 6.03249431e-01
3.36043715e-01 1.26399803e+00 4.49192047e-01 -8.29633474e-01
-1.03819869e-01 -2.62424231e-01 7.02957213e-02 -6.99896574e-01
-4.93475199e-01 2.98239052e-01 -4.49599445e-01 -7.50818431e-01
-7.03631461e-01 -5.28331637e-01 -1.92011940e+00 -1.97900683e-01
-3.84934276e-01 4.40287322e-01 4.86362129e-01 8.70329261e-01
6.44942820e-01 4.48307514e-01 1.01715982e+00 -6.67125702e-01
-2.48806998e-01 -8.21258962e-01 -6.14700139e-01 7.84622610e-01
-5.53875938e-02 -5.34904599e-01 -4.48119044e-01 1.94966123e-01] | [9.212972640991211, 3.3017261028289795] |
50e8a8b2-d076-40f2-bb7d-937b50128349 | chatgpt-for-zero-shot-dialogue-state-tracking | 2306.01386 | null | https://arxiv.org/abs/2306.01386v1 | https://arxiv.org/pdf/2306.01386v1.pdf | ChatGPT for Zero-shot Dialogue State Tracking: A Solution or an Opportunity? | Recent research on dialogue state tracking (DST) focuses on methods that allow few- and zero-shot transfer to new domains or schemas. However, performance gains heavily depend on aggressive data augmentation and fine-tuning of ever larger language model based architectures. In contrast, general purpose language models, trained on large amounts of diverse data, hold the promise of solving any kind of task without task-specific training. We present preliminary experimental results on the ChatGPT research preview, showing that ChatGPT achieves state-of-the-art performance in zero-shot DST. Despite our findings, we argue that properties inherent to general purpose models limit their ability to replace specialized systems. We further theorize that the in-context learning capabilities of such models will likely become powerful tools to support the development of dedicated and dynamic dialogue state trackers. | ['Milica Gašić', 'Carel van Niekerk', 'Hsien-Chin Lin', 'Christian Geishauser', 'Shutong Feng', 'Renato Vukovic', 'Benjamin Ruppik', 'Nurul Lubis', 'Michael Heck'] | 2023-06-02 | null | null | null | null | ['dialogue-state-tracking'] | ['natural-language-processing'] | [ 1.25561878e-02 5.10720789e-01 -4.43639576e-01 -4.58648682e-01
-5.71311593e-01 -5.77477872e-01 1.03977871e+00 -2.44495608e-02
-4.99297291e-01 7.67891049e-01 3.22623372e-01 -5.69150150e-01
4.28707629e-01 -2.46259883e-01 -3.92990261e-02 -7.29082618e-04
-1.92725405e-01 1.00555038e+00 5.33587039e-01 -9.08546329e-01
1.75920486e-01 1.42835364e-01 -1.16470301e+00 2.29809090e-01
6.43687546e-01 2.38852143e-01 1.29899725e-01 9.55373049e-01
-5.06930232e-01 8.32953393e-01 -8.02372634e-01 -1.72493517e-01
-7.28496388e-02 -5.67629814e-01 -1.15179992e+00 -7.21091703e-02
3.09956700e-01 -5.84778070e-01 -5.66027522e-01 5.59935093e-01
5.07781029e-01 5.03261745e-01 2.22108826e-01 -1.22496021e+00
-4.59508479e-01 6.92976832e-01 -1.04632370e-01 3.33951890e-01
6.16027296e-01 7.00852394e-01 7.69378483e-01 -4.61149156e-01
8.80042732e-01 1.37345183e+00 6.63814127e-01 1.34069467e+00
-1.39009678e+00 -3.13368022e-01 5.24144828e-01 -3.06307375e-01
-5.42293191e-01 -1.01870561e+00 5.31860888e-01 -1.90418020e-01
1.50913739e+00 2.33918712e-01 5.84710538e-01 1.39290917e+00
-6.83494061e-02 1.11308861e+00 1.03589308e+00 -4.42328036e-01
1.33836582e-01 4.23665822e-01 3.31403226e-01 7.27195919e-01
-1.24993138e-01 -2.23363683e-01 -7.60881305e-01 -2.66302347e-01
6.75132275e-01 -5.52013218e-01 -7.15516601e-03 -4.56132233e-01
-1.10211289e+00 8.28131616e-01 -1.63161382e-01 5.94781101e-01
-3.05565400e-03 -2.67506279e-02 8.27003241e-01 6.40237451e-01
5.67043364e-01 9.34042931e-01 -8.54492903e-01 -1.05545497e+00
-7.07544804e-01 3.29827368e-01 1.38981164e+00 1.05346429e+00
3.77475768e-01 2.75826961e-01 -3.88054639e-01 9.52032089e-01
-9.48434230e-03 6.23921789e-02 7.16646135e-01 -1.08862388e+00
5.85689723e-01 6.14187479e-01 1.60115182e-01 -1.97733477e-01
-5.67736328e-01 -6.09943159e-02 -2.17082739e-01 -8.42059478e-02
6.08738542e-01 -6.05130136e-01 -6.73642457e-01 1.87697101e+00
2.44347975e-01 -6.52626976e-02 2.85866916e-01 5.09347022e-01
6.05365217e-01 5.80945671e-01 3.82548869e-01 -1.80391803e-01
1.17496490e+00 -9.33401883e-01 -7.62805700e-01 -6.51954174e-01
1.17032111e+00 -5.16110718e-01 1.47825217e+00 2.00694442e-01
-1.26038539e+00 -3.20041627e-01 -7.83952951e-01 -1.30189076e-01
-3.95642817e-01 -4.04083699e-01 9.50212240e-01 8.08974802e-01
-1.14387560e+00 4.47098583e-01 -1.05261421e+00 -1.04137206e+00
4.55758534e-03 2.41123572e-01 -1.28556773e-01 2.18122184e-01
-1.31096077e+00 1.43135417e+00 3.03928673e-01 -2.26501137e-01
-8.59759212e-01 -4.29628074e-01 -8.52964520e-01 2.62694880e-02
5.72968781e-01 -4.55139339e-01 2.09014916e+00 -7.97966123e-01
-2.07744336e+00 8.82352054e-01 -1.38103485e-01 -5.33318520e-01
3.61903697e-01 -2.20957667e-01 -3.23326141e-01 -7.21405074e-02
-1.40782401e-01 5.02985656e-01 3.71257603e-01 -8.78601134e-01
-6.25229299e-01 -3.60721797e-02 3.71436775e-01 3.88488322e-01
-7.49384761e-01 2.97919482e-01 -3.19880337e-01 -6.71822652e-02
-3.38366002e-01 -9.00489509e-01 -2.58909434e-01 -1.24687470e-01
-1.19602039e-01 -4.50517893e-01 1.17798758e+00 -6.07697248e-01
1.16259384e+00 -1.77895617e+00 4.99131810e-03 -5.61892152e-01
5.26218303e-02 9.02535617e-01 -3.61689657e-01 8.38634551e-01
3.80315244e-01 1.79608667e-03 2.00751498e-01 -3.41347456e-01
2.27872968e-01 2.42858037e-01 1.15323499e-01 1.43213589e-02
2.64200747e-01 9.47219074e-01 -1.08177018e+00 -3.41591269e-01
3.81182224e-01 -4.73222062e-02 -6.34287894e-01 3.82460654e-01
-7.28073239e-01 3.83168548e-01 -4.66357142e-01 7.98196420e-02
9.80916917e-02 -3.95284057e-01 6.09568000e-01 3.29785138e-01
-2.42988005e-01 9.00051534e-01 -6.88511908e-01 2.02968240e+00
-5.27009845e-01 6.94302380e-01 4.32271689e-01 -7.88810074e-01
8.37342918e-01 4.94261801e-01 2.35088795e-01 -4.73654449e-01
1.56907246e-01 -2.00270295e-01 3.04732621e-01 -5.36668897e-01
9.59122956e-01 -1.04631878e-01 -3.12472194e-01 8.78540397e-01
4.27651882e-01 -1.05198957e-01 2.95768410e-01 5.59212506e-01
1.07255960e+00 1.33941427e-01 2.31440589e-01 -1.31448850e-01
1.94944486e-01 4.25449669e-01 3.29598725e-01 9.21614647e-01
-6.24205589e-01 -3.69338281e-02 6.29211605e-01 -2.41099313e-01
-1.12558460e+00 -4.59707290e-01 2.12902203e-01 1.96020901e+00
-1.77617744e-01 -4.64644045e-01 -8.81149709e-01 -7.99387991e-01
-1.59323793e-02 1.13211107e+00 -3.18624943e-01 -3.68628621e-01
-6.38662338e-01 -6.38671637e-01 8.09725225e-01 4.93331015e-01
5.61664402e-01 -1.04694164e+00 -3.92601311e-01 5.79328477e-01
-3.86068551e-03 -1.09531569e+00 -4.28757250e-01 1.98778033e-01
-9.88709450e-01 -7.89694428e-01 -5.94712317e-01 -7.25621164e-01
6.43718541e-02 3.75920624e-01 1.20611572e+00 1.09774381e-01
2.22728457e-02 7.43953705e-01 -3.27107459e-01 -1.32844001e-01
-1.21017754e+00 6.67155504e-01 1.35196924e-01 -6.15379393e-01
5.68804204e-01 -4.18262690e-01 -2.08988339e-01 2.60709494e-01
-3.93210649e-01 1.92835674e-01 2.11567953e-01 9.84442651e-01
-5.55092812e-01 -6.96842909e-01 9.45901513e-01 -1.15439451e+00
1.26090395e+00 -3.88902366e-01 -4.60226685e-01 3.58755797e-01
-6.96864486e-01 2.58987218e-01 2.95417964e-01 -5.98584294e-01
-1.53234792e+00 -3.08924913e-01 -1.31329134e-01 2.73651164e-02
-1.71196938e-01 4.11128342e-01 1.38427913e-01 3.46779404e-03
8.62648427e-01 3.40441406e-01 2.95311779e-01 -5.15723705e-01
6.46043658e-01 7.94406295e-01 4.38733906e-01 -8.76036704e-01
4.31978494e-01 -7.43632913e-02 -6.65920973e-01 -1.07436991e+00
-5.45855582e-01 -4.30974871e-01 -5.75558305e-01 1.54813973e-03
6.85802102e-01 -9.35103238e-01 -7.17894018e-01 3.97236973e-01
-8.79996955e-01 -1.02727687e+00 -3.30829740e-01 2.34720185e-02
-6.43029630e-01 5.15714347e-01 -1.04714453e+00 -1.06943786e+00
-4.36169356e-01 -7.75657713e-01 6.71329916e-01 4.33329672e-01
-7.86453962e-01 -1.40591943e+00 3.89413446e-01 5.22574604e-01
7.40336359e-01 -4.84473288e-01 8.53156209e-01 -1.43365848e+00
-3.09778303e-01 -2.50130266e-01 1.13563716e-01 2.26432219e-01
4.05287519e-02 6.48643225e-02 -1.11001396e+00 -5.17886579e-01
-1.43605322e-01 -8.22877824e-01 2.32090637e-01 -1.07185632e-01
4.40764755e-01 -2.21397981e-01 -3.38139355e-01 5.65520003e-02
8.00745130e-01 2.49132887e-01 6.72465116e-02 2.90191263e-01
4.07548785e-01 4.92562354e-01 7.04196870e-01 3.69515896e-01
7.29836881e-01 7.97616065e-01 -7.75972679e-02 4.80537154e-02
-2.26493344e-01 -4.18581545e-01 5.31027019e-01 7.76208878e-01
1.81646600e-01 -3.20695490e-01 -1.17750323e+00 5.41932821e-01
-1.95328200e+00 -9.91533160e-01 1.62802964e-01 1.94455183e+00
1.11122465e+00 4.60366458e-01 4.11225349e-01 -6.39980853e-01
6.75047696e-01 4.28886980e-01 -7.30696440e-01 -8.73190701e-01
2.47754812e-01 1.96305104e-02 6.69097761e-03 4.84570950e-01
-8.03560197e-01 1.46210158e+00 7.72690392e+00 3.11431497e-01
-9.32170331e-01 1.60177380e-01 2.73261845e-01 -3.45052592e-02
3.61089408e-03 7.92981535e-02 -9.71033037e-01 2.75027275e-01
1.36295509e+00 -4.13979471e-01 4.18289661e-01 9.39425468e-01
1.80757523e-01 -2.71713346e-01 -1.15976954e+00 3.93845320e-01
-9.84467492e-02 -1.25594950e+00 -2.74348289e-01 3.86419334e-02
6.29314959e-01 1.90757260e-01 -1.62081838e-01 9.92001355e-01
1.04019964e+00 -5.46970367e-01 5.81263788e-02 1.57991618e-01
6.75369680e-01 -3.09383810e-01 3.72281879e-01 7.55732477e-01
-6.22849762e-01 3.43643012e-03 -1.97086394e-01 -3.95388901e-01
1.96779132e-01 -3.38310778e-01 -1.51336241e+00 6.84696734e-02
6.05754368e-02 4.50004935e-01 -4.08187598e-01 7.25341558e-01
-6.19027205e-02 6.97714031e-01 -2.06181511e-01 -3.22838962e-01
2.95408159e-01 1.96389556e-01 5.20285904e-01 1.36847973e+00
-1.97919607e-01 4.72631603e-02 5.64831913e-01 5.29524148e-01
-6.71904348e-03 -1.78566128e-01 -7.15648293e-01 -4.20132101e-01
6.80686355e-01 1.20187128e+00 -3.39819580e-01 -8.30394804e-01
-6.56221151e-01 8.19311678e-01 5.06430149e-01 2.06859082e-01
-3.55705082e-01 -6.54956102e-02 1.00963068e+00 -7.66194016e-02
-2.69603869e-03 -4.23941821e-01 -2.22179428e-01 -1.51013768e+00
-4.36349511e-01 -1.19807553e+00 2.84429282e-01 -5.48757076e-01
-1.22610688e+00 4.47240293e-01 4.99832593e-02 -6.21732116e-01
-7.33428717e-01 -3.70542437e-01 -9.11276460e-01 8.18580329e-01
-1.02461016e+00 -1.06618106e+00 4.27987007e-03 5.39862156e-01
1.10140777e+00 -4.32683051e-01 1.19335186e+00 -2.14412972e-01
-8.57218623e-01 7.07556367e-01 1.87060386e-01 2.91500568e-01
9.92825866e-01 -1.27211082e+00 9.83454287e-01 5.41306198e-01
-1.83513418e-01 7.89711416e-01 1.03200924e+00 -7.62163103e-01
-1.46782756e+00 -5.93830109e-01 7.99781680e-01 -8.52605999e-01
9.67717588e-01 -6.08019292e-01 -1.16895592e+00 1.03886878e+00
6.16454840e-01 -4.31084037e-01 5.02048492e-01 9.43814278e-01
-2.24120542e-01 3.89302433e-01 -1.03128481e+00 7.29898393e-01
1.05030560e+00 -8.01508665e-01 -9.69325840e-01 4.33208346e-01
6.62816703e-01 -4.28589582e-01 -9.78649139e-01 -7.18518049e-02
4.51972663e-01 -7.60847211e-01 6.09344184e-01 -1.15590334e+00
3.86869051e-02 4.18401569e-01 3.65522921e-01 -1.31825137e+00
-1.73120305e-01 -1.13549089e+00 -3.20023239e-01 1.18893886e+00
3.68631333e-01 -6.68728709e-01 9.94667113e-01 1.17703032e+00
-3.84106994e-01 -2.24626929e-01 -6.42880380e-01 -8.24200749e-01
3.51192713e-01 -1.28659427e-01 2.73036659e-01 1.19684410e+00
8.50281358e-01 1.00310922e+00 -4.84860033e-01 -3.79893929e-01
1.41272873e-01 -2.31171295e-01 1.21692109e+00 -1.21469915e+00
-3.64676923e-01 -3.70781630e-01 1.52322429e-03 -1.28596127e+00
1.54781923e-01 -4.19878244e-01 2.66461313e-01 -1.34832501e+00
5.40707000e-02 -5.27922332e-01 -3.99739891e-02 7.14554071e-01
-3.22501510e-01 -4.43383873e-01 4.74123895e-01 -1.50130782e-03
-1.04338825e+00 6.62693322e-01 1.29735804e+00 1.04943840e-02
-5.24597406e-01 1.31819338e-01 -6.70009792e-01 6.85348153e-01
9.18576062e-01 -1.78970829e-01 -7.12306499e-01 -2.79911071e-01
-2.65188366e-01 3.99369806e-01 -1.33349925e-01 -1.02849340e+00
3.51103336e-01 -3.90873879e-01 -1.40015081e-01 -1.68231845e-01
6.91709578e-01 -1.86733410e-01 -4.66776073e-01 4.01148826e-01
-7.08240688e-01 -1.83004498e-01 5.21963418e-01 2.81873375e-01
8.66668448e-02 -3.22816133e-01 7.21866071e-01 -4.61392373e-01
-9.50439095e-01 1.92332789e-02 -9.62462068e-01 4.22901005e-01
8.78517389e-01 -2.86642253e-01 -6.49445176e-01 -7.29748309e-01
-8.48586082e-01 4.26817864e-01 6.87612057e-01 6.84338331e-01
1.30486354e-01 -7.44635761e-01 -5.09167254e-01 2.98943277e-02
2.20519796e-01 -3.25818658e-01 7.73625374e-02 4.89091247e-01
-1.49596795e-01 6.82239175e-01 -3.23695868e-01 -3.92571568e-01
-1.32582176e+00 4.10789460e-01 3.76813561e-01 -5.57293653e-01
-7.03574121e-01 7.06241965e-01 5.48064932e-02 -7.58972704e-01
2.20557541e-01 3.94490212e-02 -1.27018914e-01 -5.76979555e-02
4.17301208e-01 1.50453046e-01 -1.13143206e-01 -1.17008433e-01
-8.50884318e-02 -4.17313248e-01 -6.58955574e-01 -4.10346895e-01
1.24973381e+00 -2.70015568e-01 3.44217390e-01 7.65451193e-01
6.12301230e-01 -3.51417571e-01 -1.53358579e+00 -4.72454369e-01
3.21727723e-01 -1.95478410e-01 -2.21588179e-01 -1.18580079e+00
-3.54681909e-01 9.53031778e-01 3.72547537e-01 5.01951396e-01
4.95789409e-01 -1.03494167e-01 8.90441298e-01 8.10147941e-01
4.45911974e-01 -1.35169518e+00 3.31132501e-01 1.07854319e+00
4.08131242e-01 -1.27009809e+00 -2.08742246e-01 6.77041411e-02
-9.65166569e-01 9.58902538e-01 1.14530027e+00 1.33805722e-01
5.17991371e-02 1.55309796e-01 2.74376810e-01 -2.14452803e-01
-1.54197812e+00 -2.02756509e-01 -3.25406134e-01 8.08179379e-01
7.23902762e-01 -2.17190742e-01 -7.71628916e-02 2.23245367e-01
5.33601679e-02 2.21612528e-02 6.74760401e-01 1.32147050e+00
-8.26145828e-01 -1.36135828e+00 1.09487446e-02 3.75118583e-01
-6.47691041e-02 -1.20069258e-01 -5.35952210e-01 1.08581233e+00
-7.20893621e-01 9.82532680e-01 1.48036415e-02 -2.71199554e-01
4.89960074e-01 8.45820546e-01 4.08280849e-01 -1.13537931e+00
-1.13185346e+00 -1.28312483e-01 8.94505620e-01 -3.67507517e-01
-2.22510267e-02 -7.16595113e-01 -9.50259089e-01 -5.48776984e-01
-3.14077735e-01 3.07130247e-01 6.08075678e-01 8.60970318e-01
4.51415569e-01 3.74503970e-01 3.11961025e-01 -6.42301619e-01
-8.96759748e-01 -1.43298340e+00 -2.60993779e-01 1.66582644e-01
2.29480967e-01 -4.66956079e-01 7.05068484e-02 -1.39137641e-01] | [12.817146301269531, 7.8970489501953125] |
5cb7b1e9-55ef-45c6-a2ac-b8b2606ab433 | bert-embeddings-can-track-context-in | 2104.06529 | null | https://arxiv.org/abs/2104.06529v1 | https://arxiv.org/pdf/2104.06529v1.pdf | BERT Embeddings Can Track Context in Conversational Search | The use of conversational assistants to search for information is becoming increasingly more popular among the general public, pushing the research towards more advanced and sophisticated techniques. In the last few years, in particular, the interest in conversational search is increasing, not only because of the generalization of conversational assistants but also because conversational search is a step forward in allowing a more natural interaction with the system. In this work, the focus is on exploring the context present of the conversation via the historical utterances and respective embeddings with the aim of developing a conversational search system that helps people search for information in a natural way. In particular, this system must be able to understand the context where the question is posed, tracking the current state of the conversation and detecting mentions to previous questions and answers. We achieve this by using a context-tracking component based on neural query-rewriting models. Another crucial aspect of the system is to provide the most relevant answers given the question and the conversational history. To achieve this objective, we used a Transformer-based re-ranking method and expanded this architecture to use the conversational context. The results obtained with the system developed showed the advantages of using the context present in the natural language utterances and in the neural embeddings generated throughout the conversation. | ['Joao Magalhaes', 'David Semedo', 'Rafael Ferreira'] | 2021-04-13 | null | null | null | null | ['conversational-search'] | ['natural-language-processing'] | [-1.15454279e-01 8.68894756e-02 2.79105425e-01 -4.89589602e-01
-2.28147417e-01 -5.13220489e-01 1.17353666e+00 5.86133897e-01
-8.19139361e-01 4.23724622e-01 8.06024492e-01 -8.08070004e-02
-1.97929874e-01 -9.56459224e-01 1.27923399e-01 -2.28817701e-01
9.32128206e-02 7.76306093e-01 4.50472444e-01 -8.61901581e-01
4.06408906e-01 5.60292602e-01 -1.85312688e+00 4.24657971e-01
6.53190792e-01 5.01325905e-01 6.86229825e-01 8.20077181e-01
-5.86320460e-01 6.82790697e-01 -5.97545981e-01 -3.98976535e-01
-3.94281715e-01 -6.61126316e-01 -1.28631771e+00 -3.12379122e-01
9.56492051e-02 -1.96624160e-01 -1.75734580e-01 7.88801134e-01
3.66395175e-01 6.70987487e-01 2.79557973e-01 -8.54810297e-01
-3.64465803e-01 6.07954919e-01 5.94203413e-01 3.32502723e-01
8.92731190e-01 -1.30620211e-01 1.09725690e+00 -5.71659744e-01
8.72410953e-01 1.28595078e+00 1.72086135e-01 9.02046978e-01
-9.60824609e-01 -1.57687385e-02 2.62747049e-01 7.71030962e-01
-1.08883214e+00 -1.95640743e-01 7.68709302e-01 -4.21952248e-01
1.14861798e+00 6.47303462e-01 7.85528123e-01 1.09833872e+00
-1.22263208e-01 6.23544812e-01 7.61160672e-01 -9.36719298e-01
3.23300391e-01 8.90219450e-01 5.60040712e-01 4.08389837e-01
-4.73027170e-01 -7.08918869e-02 -3.96426201e-01 -2.61424273e-01
1.38417825e-01 1.33423090e-01 -3.73065233e-01 -1.20001473e-01
-8.17614436e-01 9.65215385e-01 4.04059857e-01 1.26733732e+00
-4.08227563e-01 -2.06984028e-01 5.04593492e-01 4.64864761e-01
2.84678727e-01 8.41880262e-01 -2.59929270e-01 -7.16540575e-01
-5.86814225e-01 5.28842509e-01 1.32942760e+00 3.22550446e-01
6.13519967e-01 -7.60411143e-01 -2.35813737e-01 1.00312340e+00
2.80020654e-01 8.99738297e-02 8.32058787e-01 -7.77027190e-01
2.32241690e-01 1.13629866e+00 8.52095038e-02 -1.19755375e+00
-3.57809961e-01 -1.34782448e-01 -2.72043258e-01 -6.17725030e-02
4.08206344e-01 1.39471769e-01 -3.06721061e-01 1.78701782e+00
3.80905807e-01 -2.68261999e-01 2.10250467e-01 6.31468832e-01
6.51907384e-01 7.24887490e-01 -9.77992043e-02 -1.45368427e-01
1.62413752e+00 -6.79942846e-01 -8.47211540e-01 -9.38932598e-02
6.89100862e-01 -7.88536489e-01 1.03491879e+00 2.45343335e-02
-6.04541242e-01 -5.37604213e-01 -6.99758708e-01 -1.89680740e-01
-8.99684370e-01 -1.38258278e-01 2.71246880e-01 4.36666071e-01
-1.07879961e+00 5.43592095e-01 -4.79511827e-01 -1.07511961e+00
-6.03049517e-01 8.05128366e-02 -2.03400016e-01 -6.60225600e-02
-1.65385783e+00 1.33295155e+00 1.60747260e-01 2.29726285e-02
-1.70235157e-01 -2.38173470e-01 -7.26741731e-01 3.76738518e-01
2.08790943e-01 -5.24356246e-01 1.44120514e+00 -6.54566050e-01
-1.50515473e+00 7.58626640e-01 -4.68162090e-01 -4.97268260e-01
3.43156278e-01 -8.09728056e-02 -5.12518942e-01 1.06034406e-01
-1.07019722e-01 2.66197532e-01 3.13147008e-01 -8.31853569e-01
-8.48179519e-01 -5.31388700e-01 4.75051165e-01 2.86825567e-01
-5.33487320e-01 9.61002931e-02 -5.40950418e-01 -1.42644092e-01
-2.81770557e-01 -1.10959470e+00 -3.81759815e-02 -3.54348809e-01
6.96919486e-02 -8.95968020e-01 8.37614000e-01 -7.37341225e-01
1.53708112e+00 -2.13341689e+00 2.67270356e-01 1.03308819e-01
-4.92276400e-02 5.00985205e-01 8.65678415e-02 1.09896433e+00
2.73212373e-01 -8.88405517e-02 6.53004870e-02 -5.29664814e-01
6.28039166e-02 3.49321246e-01 -3.43735754e-01 -2.48399869e-01
-1.36650354e-01 6.12645566e-01 -1.08210397e+00 -4.46533710e-01
3.45995575e-01 5.38322508e-01 -4.27390635e-01 5.69308758e-01
-3.82361114e-01 3.80886227e-01 -5.67964256e-01 -7.95561075e-02
-1.45220742e-01 1.08447716e-01 1.45822763e-01 4.95745204e-02
-3.83278489e-01 5.46787024e-01 -9.49108064e-01 1.51854587e+00
-8.71385038e-01 9.22842085e-01 -5.60363792e-02 -7.15403020e-01
9.75783288e-01 5.78253686e-01 1.58422187e-01 -9.44757342e-01
5.31915464e-02 1.62492350e-01 -1.24824703e-01 -9.12105680e-01
6.91605687e-01 1.27999723e-01 -1.02722168e-01 7.25489736e-01
-1.85089469e-01 -6.29651994e-02 4.03315157e-01 3.54189664e-01
9.40048873e-01 -1.46893770e-01 2.28624120e-01 -9.93416831e-03
1.11134779e+00 2.43168026e-02 -1.22811921e-01 6.00380421e-01
-3.03540062e-02 1.37712494e-01 2.73793429e-01 -4.63653207e-01
-5.98200381e-01 -5.11525393e-01 1.15074769e-01 1.23247027e+00
-1.66076362e-01 -5.10459721e-01 -7.85913229e-01 -5.07513583e-01
-1.70304418e-01 1.02152348e+00 -6.08720243e-01 -2.46126190e-01
-6.72864795e-01 -8.98099318e-03 1.52309209e-01 7.37920031e-02
4.94317681e-01 -1.43941522e+00 -7.98462391e-01 4.11019534e-01
-5.42825460e-01 -8.56628299e-01 -3.24610651e-01 -4.33486253e-02
-6.63980424e-01 -9.63167071e-01 -6.22996747e-01 -8.20320249e-01
1.46947354e-01 1.40736587e-02 9.90888774e-01 2.29797900e-01
-2.94681221e-01 8.88459921e-01 -6.33419633e-01 -3.10622722e-01
-7.92873204e-01 3.33701730e-01 -2.20087036e-01 6.65439814e-02
6.32012963e-01 -5.20524025e-01 -3.53772432e-01 1.44566670e-01
-9.16934371e-01 -3.54170322e-01 9.64066386e-02 6.45401597e-01
-2.72926450e-01 -1.21355176e-01 4.84352887e-01 -6.55901551e-01
1.37684977e+00 -3.09682965e-01 -3.90344769e-01 3.77161831e-01
-6.41792476e-01 3.22541326e-01 4.82509106e-01 -3.95731479e-01
-1.15313041e+00 -2.20041290e-01 -4.92183924e-01 1.87608317e-01
-3.11596155e-01 4.52795208e-01 -9.54885036e-03 3.73362273e-01
8.16871583e-01 2.39796996e-01 1.75875381e-01 -6.41706824e-01
4.16162252e-01 1.02473485e+00 3.66308093e-01 -3.32836300e-01
3.06665689e-01 5.29371984e-02 -3.86749059e-01 -1.23173773e+00
-3.98808807e-01 -9.10545766e-01 -3.71308327e-01 -5.29632926e-01
7.79045045e-01 -1.93933323e-01 -1.02346277e+00 1.70797303e-01
-1.45965195e+00 -5.37209213e-02 -2.17910677e-01 4.11806405e-01
-1.75119206e-01 5.24825811e-01 -2.70879507e-01 -1.09313869e+00
-4.13063347e-01 -1.14870119e+00 6.64430022e-01 3.08485925e-01
-7.31526554e-01 -1.16612542e+00 5.60153246e-01 3.03761274e-01
8.43090415e-01 -1.39233381e-01 1.20486045e+00 -1.15355957e+00
-3.85291874e-01 -4.25834715e-01 4.82070670e-02 3.44847828e-01
1.64544508e-01 -2.39437178e-01 -9.88467038e-01 1.60477579e-01
2.66362220e-01 1.17421135e-01 5.34962595e-01 -2.73631483e-01
4.34955925e-01 -2.95975119e-01 -4.22356188e-01 -3.44877541e-01
1.05166852e+00 3.95141631e-01 4.77579743e-01 4.15922374e-01
5.78199811e-02 1.17010927e+00 4.17663306e-01 1.90129772e-01
6.37276411e-01 1.24205971e+00 2.31298700e-01 4.69784170e-01
-1.47146389e-01 -2.58486718e-01 1.88492015e-01 7.07838893e-01
2.18953863e-01 -1.83608428e-01 -7.99043059e-01 7.27504313e-01
-2.02440548e+00 -1.18498456e+00 -3.93904112e-02 2.17136645e+00
8.30555081e-01 -7.87547156e-02 1.29552381e-02 -1.45893535e-02
5.86010635e-01 5.25626726e-02 -1.11900538e-01 -8.72840345e-01
4.82235819e-01 1.51105011e-02 -3.96544009e-01 9.96497571e-01
-6.59576595e-01 6.57584965e-01 5.34448624e+00 2.34424278e-01
-1.10513401e+00 -5.89570254e-02 9.67094954e-03 2.39409164e-01
-2.48629212e-01 -2.31788028e-02 -8.60692203e-01 4.83509481e-01
1.02171946e+00 -3.15449923e-01 6.41058028e-01 8.74362469e-01
1.95610344e-01 -2.98107177e-01 -1.36974931e+00 8.25088084e-01
3.09018880e-01 -1.16315150e+00 -1.05017737e-01 -7.45274127e-02
-1.96026769e-02 -1.81132495e-01 -4.60347176e-01 5.55497348e-01
-3.35905664e-02 -5.67685127e-01 2.79696643e-01 8.42550218e-01
-4.56544710e-03 -5.45742154e-01 8.72735202e-01 7.62578726e-01
-9.91588295e-01 -3.28911662e-01 6.74623996e-03 -2.52229244e-01
2.29941383e-01 2.61677086e-01 -1.43701494e+00 2.12563440e-01
5.90607643e-01 1.79746538e-01 -5.26763082e-01 9.03476238e-01
-2.78521210e-01 1.41068310e-01 -3.61827046e-01 -9.77081001e-01
2.25881115e-01 -2.21979737e-01 7.26133823e-01 1.40217650e+00
1.51448175e-01 -1.14594800e-02 -4.31490876e-02 7.65182972e-01
2.09647208e-01 3.58452916e-01 -7.45609820e-01 -1.48700267e-01
4.79795486e-01 1.09634638e+00 -3.65385175e-01 -2.41172731e-01
-1.56761885e-01 1.06749618e+00 3.20305884e-01 9.20787230e-02
-2.42049009e-01 -5.62684357e-01 7.04076469e-01 7.40611032e-02
4.90358621e-02 -2.36806318e-01 5.22415876e-01 -8.59181404e-01
2.68395126e-01 -8.40545774e-01 3.94831032e-01 -6.98208928e-01
-1.09421039e+00 1.03148425e+00 2.91294008e-01 -5.76943994e-01
-1.05276811e+00 -3.58120739e-01 -7.11810589e-01 1.18064499e+00
-1.20262408e+00 -7.91024268e-01 -3.75834644e-01 4.17334497e-01
6.92751169e-01 1.07669104e-02 1.30195856e+00 3.03973079e-01
-1.29313916e-01 1.58827588e-01 -1.84579104e-01 6.43254071e-02
6.16628945e-01 -1.16070056e+00 2.69026577e-01 6.28451645e-01
2.99237818e-01 9.96770620e-01 1.05292177e+00 -3.45228046e-01
-1.16907978e+00 -3.90722066e-01 1.87224734e+00 -5.80722094e-01
6.88904226e-01 -4.43606526e-01 -1.04237819e+00 1.52419478e-01
5.21363676e-01 -8.67416084e-01 4.97599453e-01 4.10561502e-01
-9.81108919e-02 -2.55364925e-01 -1.02045286e+00 6.64151192e-01
6.30017877e-01 -1.07185805e+00 -1.31519246e+00 2.45557174e-01
8.36312592e-01 -4.68940400e-02 -4.31434035e-01 -3.91250998e-02
4.75172281e-01 -8.24182689e-01 6.84254348e-01 -3.66069824e-01
-1.31318733e-01 -2.37710867e-02 -1.93018764e-02 -1.36639285e+00
4.23626006e-02 -5.49528003e-01 1.04328617e-01 1.47945547e+00
4.74675447e-01 -7.54425585e-01 5.98549306e-01 1.03364313e+00
1.51151150e-01 -4.52425212e-01 -1.04897428e+00 -1.97664186e-01
-4.04435068e-01 -3.40196192e-01 4.67940390e-01 6.22075498e-01
5.72966933e-01 6.55512333e-01 -8.79860073e-02 -1.74044102e-01
-2.41648197e-01 1.53611705e-01 6.63764954e-01 -1.51653099e+00
-2.00367630e-01 -5.16103923e-01 -3.40066552e-01 -1.03381300e+00
6.38796911e-02 -7.30833530e-01 5.67773767e-02 -1.70055175e+00
-1.87828749e-01 -8.55171755e-02 1.38451770e-01 1.82146892e-01
1.63615420e-02 -4.60625261e-01 3.04169953e-01 1.54523486e-02
-3.70478272e-01 3.97308469e-01 7.69490302e-01 -3.10199149e-02
-4.92626399e-01 3.18523735e-01 -4.56717342e-01 3.77560347e-01
6.07937217e-01 -2.88252473e-01 -4.32777792e-01 -1.65692374e-01
6.06342375e-01 -6.11602589e-02 1.97690204e-01 -1.00496233e+00
6.67379558e-01 2.66040057e-01 -3.03815603e-01 -4.90897119e-01
7.08901644e-01 -1.16165543e+00 -8.60051736e-02 5.33629775e-01
-7.51986742e-01 2.96134084e-01 8.21265727e-02 4.77101892e-01
-4.31372315e-01 -7.00413883e-01 2.98874408e-01 -3.11551630e-01
-8.47041488e-01 -2.94752419e-01 -6.89257920e-01 -2.05185413e-01
7.28212833e-01 -3.96157764e-02 -6.08943515e-02 -6.55794144e-01
-9.88712251e-01 3.19153190e-01 2.33074322e-01 9.13783789e-01
4.83350009e-01 -8.77844036e-01 -3.04293513e-01 7.21116289e-02
3.83280426e-01 -4.98840928e-01 4.48745675e-02 3.95673245e-01
-1.62774935e-01 7.28548765e-01 1.72385022e-01 -3.10795814e-01
-1.72127473e+00 4.15286332e-01 4.62594956e-01 -4.23156917e-01
-4.57432181e-01 6.01131022e-01 -4.55911905e-01 -4.73307192e-01
7.09700644e-01 -4.18273151e-01 -9.63268220e-01 5.59541106e-01
8.99613738e-01 3.33289325e-01 1.95539013e-01 -3.83741319e-01
-2.61561662e-01 2.36303881e-01 -1.21307939e-01 -3.73229682e-01
1.28204095e+00 -2.87402570e-01 -3.73307765e-01 6.84237421e-01
1.30773008e+00 8.43082219e-02 -3.09755951e-01 -4.33481425e-01
3.61523479e-01 -3.11119735e-01 -1.20335907e-01 -9.09827173e-01
-2.38781244e-01 8.55446160e-01 7.08889961e-01 9.39669788e-01
7.38730192e-01 2.25486875e-01 7.42476404e-01 8.46333683e-01
4.18966025e-01 -1.01534653e+00 4.50437516e-02 9.06702399e-01
1.02094901e+00 -9.34041321e-01 -3.37126046e-01 1.07960291e-02
-5.50772667e-01 1.39756382e+00 2.19864890e-01 2.99728453e-01
5.90902150e-01 -3.80300939e-01 2.02420861e-01 -3.69546413e-01
-7.27659822e-01 -4.31767583e-01 2.32248485e-01 4.71970707e-01
5.11239409e-01 -1.78756624e-01 -5.31333089e-01 1.68409750e-01
-2.72136480e-01 -1.25111332e-02 3.51523161e-01 8.39619577e-01
-5.97831249e-01 -1.61534393e+00 -2.27626741e-01 3.68576981e-02
-8.15474689e-02 4.68681976e-02 -8.50569725e-01 5.90238392e-01
-6.52698129e-02 1.38792467e+00 -5.67311682e-02 -2.47329697e-01
7.63091803e-01 6.57073379e-01 5.98659553e-02 -8.21999729e-01
-1.07131743e+00 -5.68468451e-01 4.43652302e-01 -5.09682000e-01
-3.49634528e-01 -6.67327166e-01 -8.97884071e-01 -6.89810142e-02
-4.18370396e-01 7.97096074e-01 1.11501527e+00 1.12091565e+00
3.51670027e-01 2.90187448e-01 6.42203033e-01 -5.22237659e-01
-6.01372182e-01 -1.07613420e+00 5.64159974e-02 5.34618855e-01
2.99950182e-01 -3.67936611e-01 -3.72619420e-01 -2.74797022e-01] | [12.162328720092773, 7.840281009674072] |
0c1e430f-4ffd-4d8e-86a0-43957a9d3501 | estimating-and-assessing-differential | 2212.10653 | null | https://arxiv.org/abs/2212.10653v2 | https://arxiv.org/pdf/2212.10653v2.pdf | Estimating and Assessing Differential Equation Models with Time-Course Data | Ordinary differential equation (ODE) models are widely used to describe chemical or biological processes. This article considers the estimation and assessment of such models on the basis of time-course data. Due to experimental limitations, time-course data are often noisy and some components of the system may not be observed. Furthermore, the computational demands of numerical integration have hindered the widespread adoption of time-course analysis using ODEs. To address these challenges, we explore the efficacy of the recently developed MAGI (MAnifold-constrained Gaussian process Inference) method for ODE inference. First, via a range of examples we show that MAGI is capable of inferring the parameters and system trajectories, including unobserved components, with appropriate uncertainty quantification. Second, we illustrate how MAGI can be used to assess and select different ODE models with time-course data based on MAGI's efficient computation of model predictions. Overall, we believe MAGI is a useful method for the analysis of time-course data in the context of ODE models, which bypasses the need for any numerical integration. | ['S. C. Kou', 'Shihao Yang', 'Samuel W. K. Wong'] | 2022-12-20 | null | null | null | null | ['numerical-integration'] | ['miscellaneous'] | [ 2.49725163e-01 -4.77070242e-01 2.46846855e-01 2.08333895e-01
-6.18547857e-01 -8.96429420e-01 6.65676177e-01 4.57630247e-01
-2.08772734e-01 9.23174500e-01 -4.49168712e-01 -4.60117489e-01
-5.61013579e-01 -1.80799335e-01 -3.95498544e-01 -1.09604442e+00
-2.72266537e-01 5.18378317e-01 2.15681177e-02 1.58617973e-01
3.11047196e-01 9.45247531e-01 -1.17859495e+00 -6.05633318e-01
7.68899620e-01 7.19137907e-01 -8.52792785e-02 8.15523684e-01
1.46522880e-01 4.04597431e-01 -3.64804268e-01 1.68187208e-02
-1.86481133e-01 -4.55216646e-01 -2.44296923e-01 -2.86130477e-02
-2.81409144e-01 6.43555000e-02 -2.91092813e-01 8.11178029e-01
6.20640934e-01 6.20951593e-01 1.15714335e+00 -1.39678776e+00
-4.89291660e-02 -2.44263485e-01 2.22671423e-02 1.37170359e-01
1.92423940e-01 5.44222832e-01 4.10965413e-01 -9.70008731e-01
5.10245979e-01 1.18172562e+00 7.79662967e-01 1.32675499e-01
-1.88963962e+00 -2.79967308e-01 5.52017502e-02 -2.50390410e-01
-1.71817493e+00 -5.64592540e-01 4.37075078e-01 -8.94551635e-01
7.19918787e-01 2.45902345e-01 4.61874217e-01 8.98278832e-01
6.45004511e-01 3.65000188e-01 1.14935994e+00 -9.05807987e-02
8.01829934e-01 -1.79677978e-01 1.68558136e-02 4.84135926e-01
2.15788916e-01 4.66139913e-01 -2.06377044e-01 -7.79339969e-01
7.68100381e-01 -4.15801406e-02 -1.20779447e-01 -5.94886780e-01
-9.78139281e-01 6.75776303e-01 -3.61317635e-01 7.21561015e-02
-4.04287905e-01 1.95847064e-01 1.88920841e-01 5.00224046e-02
6.74291372e-01 6.49291277e-01 -5.44626534e-01 -2.29754552e-01
-9.19906318e-01 5.94399810e-01 9.69042599e-01 8.51198733e-01
4.58337098e-01 -3.59034655e-03 -1.14077203e-01 3.96062613e-01
5.98280430e-01 4.56089497e-01 1.39932381e-02 -1.26805985e+00
-2.48918355e-01 3.08672816e-01 5.97005129e-01 -4.93763566e-01
-4.78137106e-01 -3.79289202e-02 -6.09543085e-01 4.49942797e-01
1.00880933e+00 -3.40360194e-01 -8.72582197e-01 1.54087973e+00
5.45494258e-01 2.28627041e-01 9.46204290e-02 4.69507366e-01
3.48850414e-02 7.13162601e-01 4.91181046e-01 -8.20549846e-01
1.05220795e+00 -1.74630657e-01 -8.06905031e-01 2.84266204e-01
6.13469899e-01 -6.76181436e-01 3.96382630e-01 3.92785579e-01
-1.09245932e+00 -1.26549184e-01 -6.48205519e-01 2.21985772e-01
-4.41485494e-01 -3.66183817e-02 5.36570013e-01 4.27410275e-01
-7.04782903e-01 1.17487586e+00 -1.33318365e+00 -4.42428738e-01
1.35409325e-01 3.46632719e-01 -1.27649635e-01 1.40226573e-01
-8.57805014e-01 9.84467566e-01 -8.20580497e-03 3.12946439e-01
-1.23321652e+00 -1.13545537e+00 -8.69302630e-01 3.23568285e-02
1.26086429e-01 -4.99618888e-01 1.31314361e+00 5.53321838e-02
-1.75033641e+00 7.85410404e-02 -4.50047135e-01 -1.02124885e-01
6.98285341e-01 3.19355130e-01 -1.49709165e-01 1.43669650e-01
-9.47665796e-02 2.60208726e-01 6.53494596e-01 -8.72583747e-01
5.72671667e-02 -3.28647196e-01 -5.36811233e-01 -3.57238427e-02
4.86812234e-01 6.33853069e-03 -2.76356824e-02 -2.16309205e-01
1.52528837e-01 -1.13193309e+00 -7.41297007e-01 4.03769255e-01
-3.38900805e-01 1.38609838e-02 6.24087811e-01 -4.45340246e-01
1.09174609e+00 -2.03355837e+00 8.62084180e-02 2.70188570e-01
1.88186973e-01 8.79497975e-02 1.05674401e-01 9.64751303e-01
-1.38879791e-01 4.84802723e-02 -7.32917845e-01 -2.65318185e-01
1.13798201e-01 7.37988427e-02 -1.93404019e-01 8.51509631e-01
3.50855678e-01 6.57015920e-01 -9.47813869e-01 -3.20906192e-01
5.07823348e-01 6.45167649e-01 -1.41679883e-01 1.56948954e-01
-2.98617870e-01 1.07148790e+00 -4.22850043e-01 5.26957929e-01
4.92723972e-01 -2.94806272e-01 2.67419666e-01 -2.68306416e-02
-5.70032895e-01 -2.17930585e-01 -1.28476954e+00 1.19701457e+00
-1.00179970e-01 4.83749777e-01 9.25000384e-02 -9.11303461e-01
6.27315879e-01 5.83233237e-01 9.00942028e-01 -2.87646595e-02
2.88475275e-01 3.69855106e-01 1.42806143e-01 -3.87960136e-01
-1.01676635e-01 -4.52892154e-01 1.11661386e-02 3.72221678e-01
1.20726958e-01 -6.24186158e-01 4.06098872e-01 -5.58072887e-02
9.10707176e-01 4.17973459e-01 4.99747396e-01 -6.32745802e-01
4.99921679e-01 1.60519898e-01 2.14830965e-01 5.57941675e-01
-5.28203189e-01 2.93508738e-01 7.08427966e-01 8.15632194e-02
-1.06502903e+00 -1.04699457e+00 -6.06707513e-01 4.19251949e-01
1.28050242e-02 -7.07557201e-02 -6.15813255e-01 2.10610196e-01
2.44956493e-01 6.49413824e-01 -5.04468799e-01 -6.91350475e-02
-1.09717138e-01 -1.20461190e+00 3.88249576e-01 3.72444332e-01
-3.20300937e-01 -3.82762820e-01 -1.48921505e-01 7.13474452e-01
2.53380060e-01 -9.14348304e-01 -2.05919042e-01 3.81514013e-01
-1.03769517e+00 -1.39630580e+00 -8.16513717e-01 -1.87612161e-01
7.74736285e-01 -1.01250127e-01 4.48343277e-01 -4.15759683e-01
-5.02567530e-01 6.60544872e-01 3.27783763e-01 -5.00391483e-01
-7.23339379e-01 -4.59283054e-01 2.58491069e-01 -2.34196275e-01
1.40529156e-01 -6.87241733e-01 -5.61283231e-01 3.95165890e-01
-8.13768864e-01 -5.25752485e-01 4.73584011e-02 7.00272977e-01
8.84936631e-01 2.09720299e-01 4.57608759e-01 -6.14205241e-01
9.52363968e-01 -6.59935713e-01 -1.19342792e+00 1.73393369e-01
-8.42332602e-01 2.55570799e-01 4.08281118e-01 -7.66663611e-01
-1.03300393e+00 2.68459439e-01 2.03487664e-01 -3.44623089e-01
-1.16104707e-01 8.12648773e-01 4.48522530e-02 -2.26734728e-01
5.41732728e-01 1.34410679e-01 6.17731869e-01 -4.00381505e-01
7.28746951e-02 1.96300790e-01 1.62104174e-01 -8.60893369e-01
3.01863164e-01 4.98095036e-01 8.14599037e-01 -1.00626314e+00
-2.00721920e-01 -3.99086237e-01 -5.87013185e-01 -3.04676592e-01
6.48765683e-01 -8.57443690e-01 -1.15825307e+00 5.98637283e-01
-7.90291011e-01 -7.37440944e-01 -8.89321119e-02 8.03633988e-01
-8.75780582e-01 5.32103002e-01 -7.56570995e-01 -1.43817091e+00
1.64648369e-01 -1.36225116e+00 1.02254677e+00 7.09946500e-03
-4.80814099e-01 -1.45599711e+00 2.70322800e-01 -2.91096300e-01
2.37761542e-01 5.38393617e-01 7.68188357e-01 -4.68527257e-01
-5.39169312e-01 -5.71595550e-01 2.55369604e-01 -1.90821931e-01
1.21573552e-01 5.98079145e-01 -9.93186653e-01 -2.88115144e-01
2.16770098e-01 2.20766708e-01 3.33833158e-01 9.49313402e-01
9.26377594e-01 6.14560395e-02 -5.98160803e-01 2.32062712e-01
1.34626043e+00 4.58750278e-01 3.69216204e-01 -2.06114709e-01
3.98117930e-01 6.44302726e-01 5.54525197e-01 6.35911763e-01
-1.07605517e-01 4.51012194e-01 9.39412266e-02 2.94981867e-01
4.76835519e-01 -2.16722965e-01 2.94995010e-01 4.75266397e-01
8.91170353e-02 -3.03971708e-01 -6.32735848e-01 3.09222013e-01
-1.80837560e+00 -6.81333780e-01 -5.57898581e-01 2.48524046e+00
6.90296829e-01 -2.02867582e-01 2.04401970e-01 -9.38417315e-02
6.56606734e-01 -4.30031002e-01 -9.53995168e-01 -1.94100633e-01
1.89375743e-01 7.32726306e-02 6.31691217e-01 8.01056266e-01
-9.07214046e-01 3.53304744e-01 7.77498674e+00 6.66441739e-01
-8.56534123e-01 -1.81676164e-01 5.77241123e-01 2.17608333e-01
7.42731318e-02 3.04650426e-01 -8.25359762e-01 6.00441515e-01
1.54234540e+00 -7.26414382e-01 3.39240581e-01 4.17840451e-01
9.45514500e-01 -5.08153319e-01 -1.21329439e+00 5.91122806e-01
-5.95383883e-01 -9.32129622e-01 -5.17276764e-01 4.04624581e-01
6.68178618e-01 -3.81880522e-01 -3.55312787e-02 2.15602592e-01
3.61462623e-01 -9.84781146e-01 3.25580627e-01 8.92074943e-01
6.15813911e-01 -4.72977310e-01 4.36658293e-01 4.37428087e-01
-1.02218473e+00 1.82213813e-01 -3.63571286e-01 -1.91477552e-01
5.27089834e-01 7.78788388e-01 -7.29334116e-01 3.80746990e-01
-1.55149177e-01 5.07189572e-01 2.10832637e-02 1.39016771e+00
1.54679507e-01 5.83299458e-01 -6.82893157e-01 -1.65115073e-02
3.30385566e-02 -7.51978159e-01 6.24186277e-01 9.11531150e-01
7.52264380e-01 1.69589177e-01 6.87898695e-02 1.18977571e+00
6.64650083e-01 -2.09599867e-01 -6.29288673e-01 -5.94969928e-01
5.87943614e-01 8.80056381e-01 -7.56550312e-01 -2.55047023e-01
-1.66839451e-01 5.85045636e-01 -2.38217250e-01 7.56482482e-01
-7.75413036e-01 -1.70772016e-01 1.06215560e+00 2.56680176e-02
1.23595551e-01 -6.06446028e-01 3.87180746e-02 -9.29419219e-01
-4.36788648e-01 -4.85720962e-01 2.55442202e-01 -8.26512516e-01
-1.17351890e+00 -1.05519064e-01 5.06467402e-01 -1.19688845e+00
-4.49894696e-01 -1.07479477e+00 -5.79366446e-01 1.00222051e+00
-1.09585619e+00 -3.51549506e-01 3.51006925e-01 1.73566848e-01
2.32707396e-01 3.38012457e-01 8.78637612e-01 1.07900016e-01
-8.84982586e-01 -1.09212741e-01 8.41949582e-01 -5.95832765e-01
5.22321284e-01 -1.29375207e+00 2.34177470e-01 7.51965582e-01
-4.66770291e-01 8.62366855e-01 1.30066681e+00 -6.81423545e-01
-1.44199753e+00 -9.49323118e-01 5.48044801e-01 -7.60076463e-01
1.05948222e+00 -8.36776569e-02 -9.79565978e-01 4.42216665e-01
-4.54190373e-01 -6.67755231e-02 7.72819757e-01 -3.29616398e-01
2.59625167e-01 3.28960776e-01 -1.20123744e+00 7.45588183e-01
5.48762918e-01 -5.23499727e-01 -1.75729953e-02 4.95672643e-01
2.07277343e-01 -3.63873243e-01 -1.43601012e+00 2.32870638e-01
5.13894618e-01 -3.86117369e-01 9.79109049e-01 -7.17860878e-01
-2.09244892e-01 -6.55891478e-01 4.15626690e-02 -1.35039592e+00
-1.95913196e-01 -9.26070690e-01 -5.08972764e-01 9.07193303e-01
3.52960020e-01 -7.36474931e-01 2.88505673e-01 1.41688013e+00
1.12157129e-01 -4.76271242e-01 -9.13920045e-01 -1.13802278e+00
3.05413365e-01 -3.88002872e-01 2.82836080e-01 5.89618981e-01
3.38418603e-01 9.91839333e-04 -8.74370709e-02 3.74465942e-01
7.68681467e-01 -1.17491968e-01 4.74008411e-01 -1.37047160e+00
-1.79259434e-01 -2.53793538e-01 -4.09261972e-01 -7.80459523e-01
5.60759101e-04 -3.66008162e-01 3.47821355e-01 -1.08758867e+00
-9.00396854e-02 -1.55274004e-01 -1.29845411e-01 -2.75933355e-01
-2.94597387e-01 -2.26755634e-01 -1.15468897e-01 2.39582375e-01
-2.12609068e-01 4.79035735e-01 1.24543929e+00 5.99522404e-02
-3.10013115e-01 2.37123340e-01 -3.66750360e-01 4.10449058e-01
7.49485970e-01 -5.13957143e-01 -5.83032310e-01 4.45947647e-01
-5.00154048e-02 5.07427216e-01 4.80579972e-01 -8.94599378e-01
2.45483994e-01 -3.87701988e-01 4.01636630e-01 -4.41710442e-01
5.12848854e-01 -7.33608902e-01 7.99006879e-01 5.64544618e-01
-3.35335612e-01 -1.36269957e-01 4.52788770e-01 1.03826988e+00
-1.63685083e-02 -3.97216558e-01 7.22005427e-01 -1.27020568e-01
-2.80135244e-01 5.54654002e-01 -1.30593109e+00 -2.19617099e-01
9.05350804e-01 -1.17351420e-01 -3.29071954e-02 -4.92555439e-01
-1.12128448e+00 1.87928945e-01 4.22644943e-01 -4.34739202e-01
1.54474497e-01 -1.04084265e+00 -2.53751010e-01 -2.51121856e-02
1.36479795e-01 -4.93221022e-02 4.35619473e-01 1.21644163e+00
-3.70498687e-01 5.66997647e-01 2.56569237e-01 -6.11712694e-01
-9.11129713e-01 8.21244299e-01 5.76666713e-01 -1.18511274e-01
-1.42120495e-01 2.60519952e-01 4.20664757e-01 -3.40142697e-01
-2.50106364e-01 -4.78894919e-01 2.00218290e-01 -1.10203564e-01
3.77465487e-01 7.15917289e-01 -1.95224449e-01 -3.71667862e-01
-1.94315344e-01 2.89982229e-01 4.17262256e-01 -3.26165706e-01
1.00964224e+00 -4.18962330e-01 -2.84135621e-02 8.26554120e-01
1.06242347e+00 -3.35281819e-01 -1.96850300e+00 2.29428157e-01
1.14415750e-01 -3.40494793e-03 -4.04134132e-02 -4.49439347e-01
-2.74894059e-01 8.92881215e-01 4.38036650e-01 2.49683321e-01
5.53488970e-01 -4.68630254e-01 1.36031240e-01 4.16839957e-01
3.70716937e-02 -9.54383612e-01 -3.68793577e-01 3.08949232e-01
5.86958230e-01 -8.83359015e-01 1.85667574e-01 -4.90156084e-01
-2.24538475e-01 1.14498484e+00 9.65766534e-02 6.12032949e-04
8.70672047e-01 2.41326168e-01 -9.23501253e-02 -2.77148277e-01
-8.41478825e-01 -1.62466884e-01 -3.75593081e-02 6.99902415e-01
4.43082631e-01 -6.66070431e-02 -3.73993725e-01 7.36773685e-02
6.50566936e-01 2.10685655e-01 6.86044574e-01 1.02691090e+00
-3.38046402e-01 -1.07178223e+00 -4.79437143e-01 2.55840540e-01
-3.22495729e-01 1.18020982e-01 -5.24918064e-02 9.07769024e-01
-4.92002368e-01 1.14322770e+00 -2.09019616e-01 4.15229321e-01
1.48387000e-01 5.80247402e-01 4.98127639e-01 -3.29871595e-01
2.02881902e-01 4.18214560e-01 -2.50606984e-02 -4.43866938e-01
-3.46365571e-01 -9.50668633e-01 -1.03074753e+00 -3.77248108e-01
-4.94794071e-01 3.14899832e-01 7.64187217e-01 1.11914825e+00
4.76264119e-01 3.72207850e-01 4.72848326e-01 -8.90196383e-01
-7.11377382e-01 -7.67208755e-01 -8.55462074e-01 -1.15213571e-02
2.54778683e-01 -9.97636378e-01 -6.68538451e-01 1.91407800e-01] | [6.422399997711182, 4.063368320465088] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.