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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
e688f15c-f7d8-4992-ad52-ccfc8ab04c8b | a-deterministic-algorithm-for-bridging | 1811.05721 | null | http://arxiv.org/abs/1811.05721v1 | http://arxiv.org/pdf/1811.05721v1.pdf | A Deterministic Algorithm for Bridging Anaphora Resolution | Previous work on bridging anaphora resolution (Poesio et al., 2004; Hou et
al., 2013b) use syntactic preposition patterns to calculate word relatedness.
However, such patterns only consider NPs' head nouns and hence do not fully
capture the semantics of NPs. Recently, Hou (2018) created word embeddings
(embeddings_PP) to capture associative similarity (ie, relatedness) between
nouns by exploring the syntactic structure of noun phrases. But embeddings_PP
only contains word representations for nouns. In this paper, we create new word
vectors by combining embeddings_PP with GloVe. This new word embeddings
(embeddings_bridging) are a more general lexical knowledge resource for
bridging and allow us to represent the meaning of an NP beyond its head easily.
We therefore develop a deterministic approach for bridging anaphora resolution,
which represents the semantics of an NP based on its head noun and
modifications. We show that this simple approach achieves the competitive
results compared to the best system in Hou et al.(2013b) which explores Markov
Logic Networks to model the problem. Additionally, we further improve the
results for bridging anaphora resolution reported in Hou (2018) by combining
our simple deterministic approach with Hou et al.(2013b)'s best system MLN II. | ['Yufang Hou'] | 2018-11-14 | a-deterministic-algorithm-for-bridging-1 | https://aclanthology.org/D18-1219 | https://aclanthology.org/D18-1219.pdf | emnlp-2018-10 | ['bridging-anaphora-resolution'] | ['natural-language-processing'] | [-3.02764863e-01 4.90342915e-01 -4.83291209e-01 -1.76276043e-01
-6.41929686e-01 -5.41209579e-01 5.62319994e-01 2.87361324e-01
-6.52646124e-01 7.38697529e-01 9.32036281e-01 -1.63908437e-01
-2.90928692e-01 -1.22249389e+00 -5.43219805e-01 -3.96060854e-01
-5.45906797e-02 9.08517540e-01 4.28098202e-01 -6.37936652e-01
9.54860672e-02 6.33726835e-01 -1.33672607e+00 4.52389836e-01
4.84269559e-01 3.75380665e-01 2.77924955e-01 2.33574137e-02
-4.05843019e-01 6.15729034e-01 -4.73290741e-01 -6.09021664e-01
-5.55590354e-03 -5.19490801e-02 -1.21669292e+00 -6.55679345e-01
1.05285339e-01 5.29226735e-02 -8.49759400e-01 1.09224772e+00
3.98689032e-01 3.21845651e-01 6.84421778e-01 -1.02600574e+00
-1.15004802e+00 1.44522023e+00 -3.39499235e-01 5.33874452e-01
5.07820129e-01 -3.01834792e-01 1.66539359e+00 -1.15167856e+00
1.01518989e+00 1.58638239e+00 8.28858495e-01 7.24107742e-01
-1.11510408e+00 -5.40419459e-01 8.98843929e-02 1.09601140e+00
-1.52159119e+00 -4.52319086e-02 7.46240258e-01 -6.42036228e-03
1.57597864e+00 9.87118781e-02 4.90732908e-01 1.38087201e+00
4.06622402e-02 7.36105561e-01 7.78552294e-01 -5.70498109e-01
1.68732494e-01 -3.49055111e-01 7.45453298e-01 8.61330777e-02
3.54801297e-01 9.84816998e-02 -5.38307548e-01 -3.91815186e-01
3.09706211e-01 -2.15509117e-01 -3.14443976e-01 -3.12066257e-01
-9.96670842e-01 1.26025426e+00 6.73472226e-01 5.67912698e-01
-5.46375394e-01 1.45478159e-01 4.69553351e-01 6.03323355e-02
-1.68921173e-01 5.09812772e-01 -3.85470212e-01 -2.01363415e-01
-2.08509237e-01 7.39814520e-01 9.37974334e-01 9.66409206e-01
5.72627544e-01 -4.12559122e-01 -2.06552580e-01 9.05776381e-01
3.44429195e-01 2.11470783e-01 5.54597199e-01 -1.20939922e+00
6.05241477e-01 4.76350129e-01 1.89985052e-01 -9.16855574e-01
-5.85556984e-01 1.71704292e-01 -2.59646654e-01 -3.99560928e-01
3.96865368e-01 1.90764591e-01 -1.89759403e-01 2.26698732e+00
2.38650441e-01 6.92851469e-02 7.35727727e-01 1.06753564e+00
1.16523170e+00 7.90862560e-01 3.05252135e-01 -3.52973044e-01
1.88659966e+00 -8.47066939e-01 -1.17442274e+00 -2.91955054e-01
6.20658636e-01 -4.60340261e-01 1.09828889e+00 4.88331914e-02
-1.08916569e+00 -1.50392473e-01 -1.04724646e+00 -3.29845250e-01
-5.16019821e-01 -6.11759722e-01 9.83135521e-01 1.74670234e-01
-7.31186271e-01 6.91588283e-01 -7.33692765e-01 -7.45020926e-01
3.08912158e-01 4.55582649e-01 -5.48799515e-01 -1.16495095e-01
-2.25918436e+00 1.58359790e+00 6.71050251e-01 -1.63705796e-01
-2.48171031e-01 -4.34246331e-01 -1.17197382e+00 3.17821413e-01
4.45250452e-01 -5.71913600e-01 1.03314734e+00 1.91050209e-02
-1.06616604e+00 7.60448158e-01 -1.52198374e-01 -6.83895886e-01
-3.14111710e-01 -2.82096148e-01 -3.11767966e-01 6.27759174e-02
3.14008385e-01 8.88622522e-01 1.06916577e-01 -1.11240530e+00
-4.07158554e-01 -4.17441905e-01 3.10033321e-01 3.32103699e-01
-5.33663392e-01 3.22231501e-01 -1.29418701e-01 -6.18065119e-01
3.02323699e-01 -7.43941188e-01 3.95528600e-02 -4.80899513e-01
-2.19434321e-01 -1.03575349e+00 5.35362720e-01 -3.76555175e-01
1.39045143e+00 -2.14260435e+00 6.84432089e-01 -2.35139728e-01
2.65990138e-01 1.54279411e-01 -3.80368531e-01 7.34486222e-01
-3.52749109e-01 1.44013941e-01 -5.36682665e-01 -5.98874763e-02
4.07460093e-01 8.19359660e-01 -6.48594797e-01 3.74427527e-01
3.21502149e-01 1.19304240e+00 -8.25707018e-01 -4.29035425e-01
-8.70326683e-02 3.98991823e-01 -7.52506554e-01 6.33708686e-02
-1.02202073e-02 -4.22719061e-01 -5.75149395e-02 4.34746087e-01
5.30847549e-01 3.72411400e-01 6.73651040e-01 -6.40278339e-01
-6.44646212e-02 6.46683931e-01 -1.29402149e+00 1.56960940e+00
-2.88180709e-01 4.67243075e-01 -2.35593006e-01 -1.00553405e+00
9.15497303e-01 5.39058089e-01 1.52738288e-01 -4.17106748e-01
2.77718663e-01 1.59509808e-01 3.21970195e-01 -3.57412159e-01
5.05881965e-01 -4.93879825e-01 -6.81623816e-01 7.08245516e-01
1.86640155e-02 -9.94034111e-02 2.86621749e-01 1.86920077e-01
1.36290073e+00 1.33427739e-01 5.19242287e-01 -4.24080670e-01
4.53223497e-01 1.01525038e-01 1.01506388e+00 6.99273705e-01
-4.09059405e-01 4.91505444e-01 7.27999210e-01 -4.55914199e-01
-7.90949821e-01 -1.51645875e+00 -5.05403161e-01 1.03243279e+00
3.85488629e-01 -7.17715323e-01 -6.87832832e-01 -5.34680486e-01
1.80030212e-01 9.52295244e-01 -8.25297117e-01 -3.20066690e-01
-1.15649939e+00 -9.13040102e-01 8.45306277e-01 1.06748343e+00
1.64902434e-02 -1.63503265e+00 -5.32508016e-01 4.40080822e-01
-4.79224116e-01 -1.08989704e+00 -1.72800198e-01 5.03498971e-01
-5.16571045e-01 -1.10754597e+00 9.01051462e-02 -9.69602525e-01
-2.78701968e-02 -2.12099731e-01 8.59754026e-01 -3.12052697e-01
-2.27703720e-01 -1.40051618e-02 -4.27499980e-01 -2.12332308e-01
-2.13844106e-01 -1.41278133e-01 5.40624976e-01 -5.97430468e-01
1.10138679e+00 -8.56552601e-01 -9.41225793e-03 -7.59089068e-02
-9.54373360e-01 -5.76177061e-01 1.92561209e-01 1.03426993e+00
6.20611787e-01 -1.67045116e-01 6.72157109e-01 -8.02197516e-01
9.22802150e-01 -7.11168408e-01 -1.81329638e-01 9.07876045e-02
-4.33018267e-01 1.81539506e-01 4.47143614e-01 -6.63877010e-01
-8.60851049e-01 -5.10675907e-01 -9.04875994e-02 -3.85834098e-01
7.09082261e-02 5.65648377e-01 -5.27975559e-01 4.96294796e-01
6.07086837e-01 -2.83948034e-01 -4.27304208e-01 -6.06476724e-01
8.35544825e-01 7.23861337e-01 7.68860579e-01 -9.39399123e-01
6.27840638e-01 2.84457207e-01 -2.50381649e-01 -2.27099061e-01
-1.04365575e+00 -3.24912935e-01 -7.39124060e-01 4.82461095e-01
1.05530310e+00 -6.09688878e-01 -6.68353558e-01 -1.67002827e-01
-1.78005290e+00 1.30387843e-01 -4.91357744e-01 5.86741805e-01
-7.38533497e-01 4.97330248e-01 -1.23368907e+00 -2.69358933e-01
-1.33580700e-01 -9.34983492e-01 6.06288552e-01 2.35804599e-02
-9.48293090e-01 -9.10803199e-01 2.34390676e-01 1.77156106e-01
2.94767231e-01 -9.50312316e-02 1.48399007e+00 -1.10887253e+00
-4.93190661e-02 1.30507141e-01 -5.42971909e-01 -1.19618371e-01
8.47833604e-02 -5.09471774e-01 -8.57099295e-01 -6.12820592e-03
2.78470010e-01 -1.67042539e-02 8.17797780e-01 1.13698594e-01
5.22701144e-01 -2.63676614e-01 -4.80176359e-01 3.03141713e-01
1.21664786e+00 1.12636298e-01 8.26690912e-01 8.10931921e-01
6.48368716e-01 7.54280090e-01 7.60079861e-01 1.21350519e-01
6.23839736e-01 9.76407528e-01 3.39927852e-01 5.72951198e-01
-2.40437627e-01 -7.36274794e-02 4.64532912e-01 9.45042133e-01
-4.91340943e-02 -2.68274695e-01 -9.62712765e-01 8.09379518e-01
-2.03890920e+00 -1.12104726e+00 -2.92247593e-01 1.74037528e+00
1.16302705e+00 1.05380472e-02 -3.07492375e-01 1.85059905e-01
9.40661788e-01 7.84886628e-02 -2.47134380e-02 -8.52876425e-01
-4.83391225e-01 7.59246349e-01 2.46271044e-01 8.38903904e-01
-1.03767061e+00 1.54070830e+00 5.49150944e+00 6.94055676e-01
-3.34540308e-01 5.11242449e-01 -4.91245419e-01 7.24897161e-02
-5.95961630e-01 3.05958629e-01 -9.13587809e-01 1.60492152e-01
9.83794630e-01 -3.00851077e-01 3.47111344e-01 3.92251909e-01
-4.24832493e-01 2.82072186e-01 -1.24913311e+00 7.40754485e-01
2.48400062e-01 -1.28388548e+00 1.94855571e-01 -1.16299696e-01
2.71489561e-01 -1.26770794e-01 -4.61544633e-01 6.95129275e-01
2.34049961e-01 -1.04909277e+00 4.63040173e-01 2.48404950e-01
2.30947196e-01 -8.11951637e-01 1.17494380e+00 5.00246976e-03
-1.08368111e+00 -3.42816204e-01 -7.37320602e-01 -2.79189169e-01
5.24025679e-01 1.10424682e-01 -4.59334791e-01 3.69342506e-01
7.06068754e-01 6.60952270e-01 -1.94431290e-01 3.03815067e-01
-8.87493014e-01 3.39838058e-01 -2.53008813e-01 -4.79125306e-02
2.50251442e-01 9.96216238e-02 8.85142267e-01 1.27206683e+00
-9.08548087e-02 4.87028927e-01 6.43272102e-02 1.25014555e+00
-6.98218793e-02 1.43558562e-01 -5.25576115e-01 1.19159378e-01
1.16548812e+00 9.68369663e-01 -4.16058332e-01 -2.27851063e-01
-4.03777540e-01 7.43983209e-01 8.08750212e-01 5.03596105e-03
-8.78286839e-01 -5.55629730e-01 9.16493177e-01 -1.68207824e-01
3.61303031e-01 -3.34144831e-02 -1.90971136e-01 -8.04302096e-01
-7.26327151e-02 -6.47927880e-01 7.75174618e-01 -6.86685383e-01
-1.73717856e+00 5.47162533e-01 4.02082086e-01 -7.50981867e-01
-1.26739204e-01 -7.75298715e-01 -4.75187778e-01 9.49971557e-01
-1.24985147e+00 -9.87117290e-01 2.45328650e-01 3.68491471e-01
4.06235635e-01 3.09511013e-02 1.37097561e+00 1.05080046e-01
-4.18136954e-01 4.37033236e-01 -3.43230397e-01 1.91655695e-01
8.33298326e-01 -1.21896625e+00 3.88224185e-01 5.03188848e-01
2.39924431e-01 1.00085866e+00 6.70973480e-01 -7.16632485e-01
-1.47081983e+00 -7.16034770e-01 1.71147823e+00 -7.52433002e-01
1.20488775e+00 -2.93655276e-01 -1.36123419e+00 9.14341211e-01
5.07771373e-01 -5.70900999e-02 7.93173313e-01 4.22041059e-01
-6.64473891e-01 2.76754439e-01 -1.16020095e+00 7.47704685e-01
1.25607860e+00 -5.83204269e-01 -1.91301155e+00 1.22674093e-01
1.15941644e+00 -2.79852569e-01 -1.19374776e+00 7.52285957e-01
6.26432970e-02 -5.91948569e-01 1.10586929e+00 -1.01245093e+00
4.38377917e-01 -2.49768600e-01 -7.32660294e-01 -9.39250588e-01
-5.57890594e-01 -1.28735125e-01 -4.84288454e-01 1.17626381e+00
2.65269220e-01 -5.56526959e-01 2.48853862e-01 3.91664714e-01
-1.68268397e-01 -7.57953286e-01 -1.24498475e+00 -8.44001055e-01
5.80609441e-01 -6.16239667e-01 8.15841377e-01 1.27698445e+00
8.16247940e-01 5.70082784e-01 2.68208027e-01 3.33434820e-01
4.70364630e-01 4.34151769e-01 -1.13331703e-02 -1.37365448e+00
-2.22069714e-02 -6.04489148e-01 -5.58424473e-01 -5.90046942e-01
9.79473352e-01 -1.38904476e+00 -4.64731529e-02 -1.59007835e+00
4.56561357e-01 -2.96659261e-01 -4.44139063e-01 7.55359232e-01
-1.76468030e-01 3.06947857e-01 2.94219136e-01 2.58386075e-01
-2.82858908e-01 5.31624794e-01 7.93107033e-01 -1.69565022e-01
-1.01246990e-01 -8.80195618e-01 -8.00795555e-01 6.97753787e-01
7.79042959e-01 -6.97848499e-01 3.10058668e-02 -4.62761790e-01
2.06874013e-01 -5.45284245e-03 2.64695406e-01 -4.24798429e-01
3.85546386e-01 -2.24330634e-01 -1.99924558e-01 -2.86628276e-01
4.73801553e-01 -5.32223582e-01 -1.31093934e-01 3.03791851e-01
-2.94980526e-01 2.73478568e-01 3.19805592e-01 3.93951148e-01
-2.51892924e-01 -6.89681172e-01 3.30871284e-01 5.36267348e-02
-8.40181887e-01 -1.41605258e-01 -5.32123566e-01 2.49366090e-01
8.46457958e-01 -1.49719324e-02 -4.05416667e-01 -4.42914665e-02
-1.10661089e+00 1.30883828e-01 2.17791736e-01 6.70853615e-01
7.20488191e-01 -1.75938058e+00 -6.97358251e-01 -2.17893675e-01
7.09179193e-02 -1.23645492e-01 -6.72648102e-02 7.41270483e-01
-2.17048109e-01 3.18464100e-01 -2.46277899e-01 -1.23758707e-02
-1.00844288e+00 9.02637720e-01 -3.64587009e-02 -2.16439575e-01
-8.85672331e-01 7.47640491e-01 9.46646631e-02 -6.37905240e-01
1.26568198e-01 -9.18960720e-02 -6.75561368e-01 4.73983794e-01
4.60932374e-01 2.30902791e-01 -1.86123848e-01 -7.22073674e-01
-7.23301351e-01 3.64466041e-01 -1.77328765e-01 -1.54027849e-01
1.56299448e+00 4.27323356e-02 -5.67353308e-01 4.68305707e-01
9.79742289e-01 2.17234939e-01 -2.57485330e-01 -3.00083518e-01
5.72548151e-01 -8.31383020e-02 -1.89648718e-01 -4.03102398e-01
-6.56636238e-01 8.23585033e-01 2.67851472e-01 -8.00071061e-02
5.48423350e-01 6.48680389e-01 9.51007605e-01 7.37553239e-01
4.15755481e-01 -9.62410212e-01 -2.90598750e-01 1.00717115e+00
9.27292168e-01 -5.94624996e-01 -1.28136516e-01 -6.69451356e-01
-5.07334709e-01 1.22885311e+00 7.70842254e-01 -2.59709716e-01
3.49915862e-01 5.40095210e-01 1.07204998e-02 -5.14050424e-01
-7.65497684e-01 -3.80218565e-01 -1.79576233e-01 6.82823420e-01
3.67921263e-01 2.23230213e-01 -1.02665353e+00 1.49334526e+00
-4.12253022e-01 -5.59041440e-01 7.28812039e-01 7.89564013e-01
-4.50377673e-01 -1.44426870e+00 -4.73230720e-01 -2.76973471e-03
-2.67206788e-01 -5.17715573e-01 -4.62801367e-01 1.00117135e+00
1.54111207e-01 8.10480595e-01 3.62388015e-01 -1.87740088e-01
6.31771684e-01 3.64826262e-01 6.50004268e-01 -8.90517473e-01
-4.34018075e-01 -4.08778399e-01 2.33215570e-01 -5.62590539e-01
-4.44654971e-01 -8.08252573e-01 -1.93897521e+00 -1.62598535e-01
-3.26628864e-01 4.28140104e-01 9.43695232e-02 1.04906869e+00
3.25219110e-02 3.81887317e-01 2.12474003e-01 -4.94182706e-01
-6.99165165e-01 -1.06221783e+00 -5.52957892e-01 6.55896902e-01
-4.59636986e-01 -1.08414364e+00 -5.46812534e-01 -3.88471007e-01] | [9.676392555236816, 9.249931335449219] |
f8f8b45c-f6ec-4da8-b798-ff75530b3fe1 | dict-tts-learning-to-pronounce-with-prior | 2206.02147 | null | https://arxiv.org/abs/2206.02147v2 | https://arxiv.org/pdf/2206.02147v2.pdf | Dict-TTS: Learning to Pronounce with Prior Dictionary Knowledge for Text-to-Speech | Polyphone disambiguation aims to capture accurate pronunciation knowledge from natural text sequences for reliable Text-to-speech (TTS) systems. However, previous approaches require substantial annotated training data and additional efforts from language experts, making it difficult to extend high-quality neural TTS systems to out-of-domain daily conversations and countless languages worldwide. This paper tackles the polyphone disambiguation problem from a concise and novel perspective: we propose Dict-TTS, a semantic-aware generative text-to-speech model with an online website dictionary (the existing prior information in the natural language). Specifically, we design a semantics-to-pronunciation attention (S2PA) module to match the semantic patterns between the input text sequence and the prior semantics in the dictionary and obtain the corresponding pronunciations; The S2PA module can be easily trained with the end-to-end TTS model without any annotated phoneme labels. Experimental results in three languages show that our model outperforms several strong baseline models in terms of pronunciation accuracy and improves the prosody modeling of TTS systems. Further extensive analyses demonstrate that each design in Dict-TTS is effective. The code is available at \url{https://github.com/Zain-Jiang/Dict-TTS}. | ['Zhenhui Ye', 'Jinglin Liu', 'Yi Ren', 'Qian Yang', 'Zhou Zhao', 'Su Zhe', 'Ziyue Jiang'] | 2022-06-05 | null | null | null | null | ['polyphone-disambiguation'] | ['natural-language-processing'] | [ 1.45951351e-02 -1.34414315e-01 -2.65455961e-01 -4.24241215e-01
-1.26464510e+00 -6.18099749e-01 3.66212130e-01 -3.24182838e-01
-2.59137243e-01 4.66797173e-01 5.09611130e-01 -5.89561582e-01
4.98290837e-01 -2.86651164e-01 -5.93386054e-01 -4.22341466e-01
6.57544434e-01 5.96521139e-01 -9.06313211e-03 -3.96506131e-01
-2.18895078e-01 -2.48684078e-01 -1.19731176e+00 2.85788953e-01
8.11705053e-01 1.05896890e+00 8.72400641e-01 8.08685064e-01
-3.52010190e-01 5.83458662e-01 -7.64856458e-01 -3.76343787e-01
-1.03239520e-02 -6.63571179e-01 -5.76001167e-01 -2.51462281e-01
1.55833215e-01 -8.46616030e-02 -2.76461869e-01 1.10965645e+00
7.54015923e-01 1.48546621e-01 3.51450086e-01 -9.36064422e-01
-6.58900201e-01 9.37312782e-01 1.43870503e-01 3.51484507e-01
4.05677050e-01 2.78919578e-01 1.24741066e+00 -1.08494282e+00
2.17857629e-01 1.34193313e+00 4.97428358e-01 7.63291836e-01
-8.44749033e-01 -7.16068089e-01 1.19091369e-01 1.35810450e-01
-1.46178150e+00 -9.31975305e-01 6.97600305e-01 -2.92178810e-01
1.12269390e+00 4.30695266e-01 4.36887681e-01 1.65145123e+00
-2.99074292e-01 1.04015958e+00 7.20898747e-01 -5.64810991e-01
1.60579577e-01 1.34413555e-01 3.48963402e-02 4.01488572e-01
-3.79383653e-01 3.75281237e-02 -9.76444364e-01 9.90725607e-02
4.86342698e-01 -3.77685785e-01 -3.02979678e-01 4.45068121e-01
-1.22164714e+00 6.46424592e-01 -2.70722896e-01 5.71894526e-01
-2.71809369e-01 1.32691070e-01 5.33909380e-01 2.11403459e-01
4.40974861e-01 2.00739101e-01 -7.36292303e-01 -6.44761741e-01
-9.09059286e-01 -1.24792732e-01 8.04533482e-01 1.20796502e+00
3.42562258e-01 5.82341850e-01 -1.99968562e-01 1.19424927e+00
3.47901970e-01 9.86310661e-01 8.99645448e-01 -5.96189141e-01
6.76524639e-01 -6.48746863e-02 -1.09476686e-01 -5.02218723e-01
1.69814155e-01 -5.21709383e-01 -6.08024895e-01 -7.36928165e-01
2.60981023e-01 -3.26587528e-01 -7.62388289e-01 1.88196206e+00
2.32126430e-01 3.46286476e-01 2.36786559e-01 7.89427400e-01
7.78043628e-01 1.00949824e+00 1.11261860e-01 -2.13986486e-01
1.51394057e+00 -1.02821100e+00 -1.13132131e+00 -5.60619652e-01
3.76581401e-01 -1.07056534e+00 1.71004450e+00 1.76738635e-01
-9.60579574e-01 -8.17848742e-01 -9.03979599e-01 -8.11122283e-02
-3.61528993e-01 5.04180789e-01 3.53202485e-02 7.59381294e-01
-9.42107141e-01 1.02635354e-01 -7.39260852e-01 -2.40437791e-01
-1.59557447e-01 -6.57797456e-02 3.10853928e-01 1.92130506e-01
-1.74701977e+00 6.11192048e-01 5.64953268e-01 -6.43806383e-02
-9.34627056e-01 -6.52473509e-01 -9.61155057e-01 -1.77669227e-02
5.03238201e-01 -6.67564213e-01 1.82620382e+00 -9.98472214e-01
-1.91591990e+00 6.92893207e-01 -5.69292247e-01 -4.75221515e-01
1.69311553e-01 -4.08007115e-01 -7.90700376e-01 -1.49595603e-01
1.45504370e-01 4.42129195e-01 9.19394374e-01 -8.10579181e-01
-6.96188569e-01 1.51047828e-02 -5.11594117e-01 5.49238145e-01
-4.38244611e-01 2.11615458e-01 -7.02691138e-01 -1.13165200e+00
-2.23014861e-01 -9.24755752e-01 1.43469930e-01 -7.74354398e-01
-6.64044738e-01 -4.29655522e-01 6.56036556e-01 -1.06969988e+00
1.62246013e+00 -2.28012419e+00 -9.35341492e-02 -1.23450115e-01
-3.92830491e-01 5.45905769e-01 -1.08154807e-02 4.67472494e-01
1.75281391e-01 -7.21870959e-02 -1.54916376e-01 -6.96734726e-01
3.88163120e-01 2.17572927e-01 -6.46069407e-01 -4.85215932e-02
2.72867223e-03 9.62474108e-01 -1.00303936e+00 -3.51044863e-01
3.97872418e-01 4.58722681e-01 -3.14213544e-01 5.13401389e-01
-4.64016795e-01 6.71023071e-01 -2.30734542e-01 5.63244581e-01
1.30426781e-02 7.36060217e-02 2.80199349e-01 8.76183584e-02
-6.02632463e-02 1.23180544e+00 -8.58228445e-01 1.82413220e+00
-9.31447208e-01 5.12428224e-01 -4.62244041e-02 -7.43748903e-01
8.61184120e-01 8.59934926e-01 3.18287224e-01 -7.22253919e-01
1.72240570e-01 6.44632518e-01 -1.40870452e-01 -3.74857962e-01
6.12300634e-01 -2.80702770e-01 -3.95479649e-01 5.05402870e-02
2.66821772e-01 -2.97048718e-01 -2.64282208e-02 1.26316279e-01
5.48301160e-01 -2.72640079e-01 2.81078398e-01 -1.21759757e-01
4.53042835e-01 -1.57249004e-01 5.55885196e-01 5.93642771e-01
-4.21607047e-02 6.05402231e-01 -1.65071145e-01 1.02174737e-01
-8.87142599e-01 -1.16153908e+00 1.31630078e-01 1.36513269e+00
-1.38072312e-01 -7.32803106e-01 -1.06333876e+00 -5.65067947e-01
-2.47730240e-01 1.24244237e+00 1.98843367e-02 -1.11453205e-01
-5.23924112e-01 -3.41728956e-01 9.29398298e-01 4.67922300e-01
3.13045114e-01 -1.10869861e+00 2.84175277e-01 4.29730654e-01
-9.19168413e-01 -1.42342436e+00 -1.30493665e+00 7.42272735e-02
-4.18654114e-01 -5.80186248e-01 -7.21802771e-01 -9.31237757e-01
-2.16827765e-02 1.46005452e-01 9.13828194e-01 -3.81445616e-01
1.51031241e-01 2.67550021e-01 -4.85592902e-01 -4.79347199e-01
-9.19032395e-01 2.74758786e-01 3.40084702e-01 1.00076556e-01
6.83446169e-01 -4.99541998e-01 -3.30488771e-01 3.11303318e-01
-5.73726714e-01 4.40250300e-02 4.02515948e-01 6.91661000e-01
7.52917469e-01 -5.88403568e-02 8.12503815e-01 -6.52009964e-01
6.45528257e-01 -6.08321190e-01 -4.73146766e-01 7.86674470e-02
-4.82012302e-01 -1.25753790e-01 8.80148053e-01 -4.95373279e-01
-1.03275287e+00 1.47334663e-02 -8.26504529e-01 -5.71190000e-01
-1.82947680e-01 5.80024362e-01 -7.09131479e-01 7.23951161e-01
3.19765151e-01 7.50032246e-01 -3.23619634e-01 -9.34850872e-01
4.84712809e-01 1.22322810e+00 8.72254610e-01 -5.05159974e-01
5.39892495e-01 -2.81447563e-02 -8.41592312e-01 -1.18039238e+00
-1.11150789e+00 -8.50188255e-01 -2.68289655e-01 -4.96973209e-02
7.95528412e-01 -1.08718610e+00 -4.80791360e-01 5.96331477e-01
-1.32193053e+00 -3.08991611e-01 -2.30235249e-01 7.30798185e-01
-6.65246189e-01 4.00772482e-01 -6.52092576e-01 -8.28138828e-01
-4.56979036e-01 -9.50687170e-01 1.25189650e+00 -3.30487862e-02
-5.02565801e-01 -1.06339788e+00 1.20424211e-01 6.22977316e-01
5.68650365e-01 -5.26897490e-01 6.76022649e-01 -8.36718440e-01
-3.02833706e-01 1.78908303e-01 3.33404154e-01 8.18342745e-01
3.63206625e-01 -3.37966204e-01 -1.24839270e+00 -8.79321247e-02
1.89152762e-01 -7.05926046e-02 6.12142146e-01 4.48960871e-01
1.12251890e+00 -7.69187748e-01 9.66408290e-03 4.53265518e-01
9.83402133e-01 4.82865542e-01 5.00406444e-01 -1.64660662e-01
8.35565507e-01 2.93365419e-01 5.37126601e-01 4.57108766e-01
4.55156446e-01 8.70688736e-01 -7.65261352e-02 3.82703431e-02
-6.94989443e-01 -9.18442905e-01 8.40207040e-01 1.92116940e+00
5.24968624e-01 -6.36726379e-01 -8.22241187e-01 8.34756732e-01
-1.60771799e+00 -6.61342084e-01 1.65182337e-01 2.16909289e+00
1.36154222e+00 8.63623843e-02 9.94689986e-02 4.65864800e-02
9.68119442e-01 2.81826586e-01 -5.11823952e-01 -3.28846693e-01
-1.54896617e-01 2.37112135e-01 1.98977560e-01 7.76955545e-01
-8.92294943e-01 1.46492136e+00 5.18832588e+00 1.59473085e+00
-1.17376995e+00 4.27081704e-01 4.52776223e-01 -7.04510063e-02
-5.07016063e-01 2.99767521e-03 -1.23090363e+00 8.31819832e-01
1.40255821e+00 -2.23612309e-01 6.95225477e-01 8.02548349e-01
4.23764378e-01 3.43152672e-01 -1.07314563e+00 1.27721405e+00
6.29927814e-02 -1.09585893e+00 1.46071389e-01 -3.99028897e-01
4.29769605e-01 2.50813812e-01 1.77143455e-01 4.00692493e-01
4.02129292e-01 -8.71497989e-01 1.10783064e+00 2.73082167e-01
1.17530847e+00 -5.75323105e-01 4.25581992e-01 4.30355698e-01
-1.49992752e+00 3.76584500e-01 5.92627116e-02 2.73364723e-01
4.74136591e-01 5.71074784e-01 -1.39695561e+00 4.26604211e-01
4.08554763e-01 5.63252449e-01 1.22530535e-02 5.52792966e-01
-4.54725355e-01 1.39269388e+00 -3.87170702e-01 -3.38684320e-01
2.69564152e-01 5.94681874e-02 8.58529747e-01 1.61236989e+00
4.73981738e-01 -6.73811808e-02 2.10701272e-01 6.84845209e-01
-2.02746630e-01 3.03115070e-01 -3.24454695e-01 -5.79891980e-01
1.00036955e+00 6.55874014e-01 -3.50047469e-01 -4.82631296e-01
-3.19015831e-01 1.20763981e+00 -2.77168415e-02 4.57644343e-01
-8.10962200e-01 -4.63139921e-01 9.04087901e-01 6.97622672e-02
3.36408973e-01 -4.04023737e-01 -1.86577905e-03 -1.11618459e+00
9.91337374e-02 -1.14041770e+00 3.67644399e-01 -7.41239846e-01
-1.41220236e+00 7.55689144e-01 -2.78710783e-01 -1.00050998e+00
-6.12017810e-01 -3.66170347e-01 -3.03969413e-01 1.09427059e+00
-1.55793774e+00 -9.98944044e-01 2.30538905e-01 6.42845392e-01
1.23053980e+00 -3.48838270e-01 8.80760014e-01 5.50321579e-01
-5.28123438e-01 8.31888855e-01 2.41522729e-01 3.36448431e-01
6.08989477e-01 -1.16225457e+00 9.77852046e-01 9.50765193e-01
6.00343823e-01 5.66344321e-01 6.44726336e-01 -6.78577125e-01
-1.33152223e+00 -1.27054429e+00 1.43613493e+00 -4.61946338e-01
7.15580344e-01 -6.95452571e-01 -7.87962735e-01 6.80746138e-01
4.11020011e-01 -4.57118809e-01 7.23592818e-01 4.30026762e-02
-2.47481242e-01 -5.47771640e-02 -4.74254221e-01 7.33393013e-01
1.09261167e+00 -1.10776043e+00 -6.33385420e-01 2.50339448e-01
1.32083857e+00 -4.83149886e-01 -3.59091401e-01 -1.58197209e-02
2.25706398e-01 -4.22194302e-01 7.92522550e-01 -3.97050858e-01
-2.25094274e-01 -3.35081935e-01 -6.42328441e-01 -1.47453105e+00
1.06694385e-01 -1.28547180e+00 -5.48104085e-02 1.39550686e+00
5.30513823e-01 -6.67065918e-01 2.18698323e-01 5.82296252e-02
-6.50299072e-01 -3.87199044e-01 -1.09587049e+00 -1.20289218e+00
-1.40851229e-01 -9.57728207e-01 6.75905466e-01 9.06754911e-01
1.43878981e-01 7.68287420e-01 -5.07235110e-01 3.36731225e-01
2.92343616e-01 -6.40894324e-02 5.14296770e-01 -6.19871259e-01
-6.64475858e-01 -3.54035556e-01 1.18680038e-01 -1.73970389e+00
2.74659008e-01 -1.09844697e+00 3.88211519e-01 -1.15729928e+00
-3.31156969e-01 -4.97868478e-01 -1.53795764e-01 5.97707689e-01
-2.47996733e-01 -9.11320001e-02 1.89867035e-01 7.35854656e-02
-5.52616000e-01 1.08296120e+00 9.43757534e-01 4.23707105e-02
-1.97795242e-01 1.99023098e-01 -6.10182881e-01 6.66237831e-01
9.73899961e-01 -6.15466654e-01 -4.55668122e-01 -5.05838275e-01
-2.52847075e-01 2.46524259e-01 9.39093679e-02 -8.70188892e-01
2.35913038e-01 -2.06573114e-01 -3.81705135e-01 -6.13780260e-01
5.92691660e-01 -4.30413395e-01 5.68225160e-02 2.49982342e-01
-2.97787279e-01 -1.76106095e-01 2.38996938e-01 4.59016293e-01
-4.28615123e-01 -1.53107777e-01 6.81190848e-01 -1.06281392e-01
-6.59670234e-01 2.55695909e-01 -6.81281328e-01 6.30244792e-01
4.02536184e-01 7.39109814e-02 -3.15910392e-03 -6.47848248e-01
-6.11647606e-01 2.38388199e-02 -4.06855978e-02 8.27904344e-01
5.17743766e-01 -1.38981628e+00 -7.18047559e-01 2.89626032e-01
1.44667298e-01 -1.86125502e-01 1.58159569e-01 3.70063692e-01
-6.45021489e-03 8.31765592e-01 4.81124043e-01 -4.34145123e-01
-1.13653457e+00 3.30035657e-01 3.55849534e-01 1.12788022e-01
-3.44713032e-01 9.85198081e-01 2.50523388e-01 -5.22002697e-01
5.63911796e-01 -5.40636063e-01 2.40685821e-01 -2.57780135e-01
3.70917410e-01 2.01363131e-01 1.57923639e-01 -8.04445803e-01
-4.38978583e-01 2.15740398e-01 6.21067099e-02 -5.91889977e-01
8.90801787e-01 -5.82226872e-01 3.98316711e-01 7.70802557e-01
1.25811458e+00 3.90974373e-01 -1.07751477e+00 -5.43994784e-01
8.72209761e-03 -2.97327459e-01 7.12941140e-02 -9.12154317e-01
-8.63455117e-01 1.01367474e+00 9.01607648e-02 3.20994551e-03
9.04572070e-01 1.30557284e-01 1.73088717e+00 3.12038034e-01
1.79498225e-01 -1.35459185e+00 1.77695528e-01 9.61025238e-01
8.49886537e-01 -1.03501451e+00 -1.02128112e+00 -3.52031708e-01
-1.00896227e+00 7.22152948e-01 3.45186621e-01 4.41182524e-01
7.92204440e-01 3.19483370e-01 4.02591735e-01 1.39992908e-01
-7.18003809e-01 -3.15832168e-01 3.91355067e-01 4.33836371e-01
4.30936873e-01 3.65571737e-01 -5.10557927e-03 1.10699558e+00
-7.86902487e-01 -3.87750655e-01 9.63744000e-02 4.44951594e-01
-5.08779347e-01 -1.28177452e+00 -1.21289209e-01 -1.37035355e-01
-3.83971065e-01 -7.59056032e-01 -5.66123903e-01 1.61930352e-01
-3.85885350e-02 1.21768427e+00 -1.43797949e-01 -4.00267869e-01
3.46873701e-01 5.37719250e-01 3.88521925e-02 -9.04708683e-01
-4.27632034e-01 4.97421712e-01 3.53423059e-01 -3.27267259e-01
1.38938069e-01 -6.52587414e-01 -1.37222433e+00 -2.35498667e-01
-3.53488177e-01 4.00697380e-01 7.92195082e-01 8.57908428e-01
4.74589914e-01 5.69650590e-01 7.75742769e-01 -3.62271875e-01
-6.31669939e-01 -1.07353544e+00 -4.68423396e-01 2.24426851e-01
3.94556284e-01 -2.40706220e-01 -3.16383600e-01 3.61333311e-01] | [14.635180473327637, 6.824821472167969] |
c0ad4733-fd42-49fe-abd6-ffa054870b44 | xbnet-an-extremely-boosted-neural-network | 2106.05239 | null | https://arxiv.org/abs/2106.05239v3 | https://arxiv.org/pdf/2106.05239v3.pdf | XBNet : An Extremely Boosted Neural Network | Neural networks have proved to be very robust at processing unstructured data like images, text, videos, and audio. However, it has been observed that their performance is not up to the mark in tabular data; hence tree-based models are preferred in such scenarios. A popular model for tabular data is boosted trees, a highly efficacious and extensively used machine learning method, and it also provides good interpretability compared to neural networks. In this paper, we describe a novel architecture XBNet, which tries to combine tree-based models with that of neural networks to create a robust architecture trained by using a novel optimization technique, Boosted Gradient Descent for Tabular Data which increases its interpretability and performance. | ['Tushar Sarkar'] | 2021-06-09 | null | null | null | null | ['breast-cancer-detection', 'breast-cancer-detection', 'diabetes-prediction', 'classification'] | ['knowledge-base', 'medical', 'medical', 'methodology'] | [-3.67543548e-02 1.76569402e-01 -2.99302310e-01 -4.85409141e-01
-2.53252953e-01 -1.79555267e-01 6.18125439e-01 4.10104275e-01
-2.20338345e-01 9.60126340e-01 9.96770859e-02 -5.24702907e-01
-3.78350616e-01 -8.90764177e-01 -4.71576720e-01 -5.09342730e-01
-1.13248840e-01 7.22118795e-01 -3.55715603e-02 -4.03603941e-01
2.75101990e-01 6.53390110e-01 -1.79023409e+00 6.16590023e-01
8.16115677e-01 1.35072505e+00 -1.87900096e-01 2.98968971e-01
-6.11506343e-01 1.36912692e+00 -6.10499561e-01 -7.01543093e-01
4.25164998e-02 6.89469948e-02 -7.03523755e-01 -1.36073500e-01
3.72380078e-01 -7.48231634e-02 6.88464344e-02 6.75206363e-01
1.87986508e-01 -4.38218080e-02 8.14670920e-01 -1.17167139e+00
-4.25442308e-01 6.15990818e-01 -3.34540606e-01 -1.19312860e-01
2.12691128e-01 -4.35354501e-01 1.02637208e+00 -9.67909396e-01
5.11950910e-01 1.47410297e+00 8.96658301e-01 3.18485081e-01
-8.78719985e-01 -3.79187137e-01 4.93600480e-02 6.61017716e-01
-9.68314230e-01 -1.21264048e-02 8.53642166e-01 -4.97295827e-01
6.46995604e-01 4.51220840e-01 6.52004302e-01 9.04931366e-01
5.18639326e-01 9.23429132e-01 1.20942235e+00 -5.08705378e-01
2.56920427e-01 3.36414963e-01 1.18162341e-01 7.88260996e-01
2.35635772e-01 -1.52800865e-02 -5.84144235e-01 9.10077840e-02
4.40357387e-01 3.06628831e-02 -8.33081901e-02 -4.09954429e-01
-8.43456268e-01 1.02788174e+00 6.39433980e-01 5.38710654e-01
-5.31000733e-01 -1.40597686e-01 7.88993359e-01 1.34397477e-01
4.83773082e-01 4.09802735e-01 -5.23733735e-01 -1.04152396e-01
-8.66424918e-01 1.75661504e-01 6.27114236e-01 5.30963719e-01
4.84257042e-01 4.42271411e-01 4.75155562e-02 8.75335217e-01
3.21744263e-01 1.39921755e-01 5.96145511e-01 -4.48551476e-01
3.90615910e-01 1.19435835e+00 -3.46620351e-01 -1.39955962e+00
-8.14805508e-01 -5.05851209e-01 -1.31538010e+00 5.23840189e-01
2.74120688e-01 8.86389539e-02 -1.13450515e+00 1.26709747e+00
2.18429551e-01 -7.27509022e-01 -1.09716296e-01 6.81591451e-01
1.11894381e+00 9.92479026e-01 -2.21446902e-02 -1.50714338e-01
1.16551757e+00 -1.04107475e+00 -1.10195649e+00 -1.18374728e-01
4.11990464e-01 -5.07316530e-01 9.01692212e-01 9.94339287e-01
-1.02491724e+00 -6.23829305e-01 -1.06470859e+00 -3.48548628e-02
-9.11775410e-01 4.03033793e-02 6.31314278e-01 8.22903633e-01
-8.05869639e-01 5.97743571e-01 -3.78872156e-01 -2.40585908e-01
3.51374805e-01 5.29897153e-01 -4.31724280e-01 1.85134709e-01
-1.18988848e+00 1.21200609e+00 9.91779864e-01 4.42447037e-01
-4.52126414e-01 4.94649224e-02 -7.82983661e-01 8.91125277e-02
5.12452245e-01 -4.61733520e-01 1.03192258e+00 -9.49713409e-01
-1.52533257e+00 4.76580858e-01 -8.72035697e-02 -8.78285348e-01
4.57756639e-01 -1.45405978e-01 -3.22135329e-01 -3.55351195e-02
-2.93919533e-01 5.81886053e-01 1.14657784e+00 -1.13633096e+00
-3.86683792e-01 -6.06490791e-01 4.55410220e-02 8.96532983e-02
-6.30798340e-01 -1.22407943e-01 -2.21502587e-01 -6.53354228e-01
3.28669786e-01 -6.57854915e-01 -7.76064768e-02 -1.48358971e-01
-1.98802859e-01 -4.27730739e-01 1.11762154e+00 -8.75320494e-01
1.43496466e+00 -1.70183957e+00 -1.90160479e-02 3.29855025e-01
2.93392539e-01 4.63414878e-01 3.19881558e-01 4.05415863e-01
-2.17271686e-01 1.91450715e-01 -2.63987035e-01 -9.70194787e-02
-1.52459472e-01 5.11382818e-01 -9.09488201e-02 -8.73592198e-02
1.88324407e-01 7.26977527e-01 -4.43987370e-01 -6.20908618e-01
4.78142947e-01 3.60053957e-01 -2.52781153e-01 1.44919261e-01
-3.21486980e-01 1.89172655e-01 -3.60747069e-01 7.58268178e-01
5.85723162e-01 -1.89822838e-01 2.18531176e-01 -2.61925697e-01
-1.59567371e-01 8.33530128e-02 -1.09714961e+00 1.12899137e+00
-4.90925491e-01 7.43330538e-01 6.18431671e-03 -1.22093964e+00
1.39458156e+00 5.28185129e-01 3.91302735e-01 -7.54051328e-01
3.25617701e-01 1.83826387e-01 -2.36560613e-01 -5.63120246e-01
5.29827058e-01 -1.09887883e-01 6.71479851e-02 -5.40564731e-02
-1.46541581e-01 -3.65718961e-01 3.20189536e-01 -4.15199576e-03
5.69940388e-01 1.88698381e-01 4.97482628e-01 -6.58761039e-02
9.09743190e-01 2.70222634e-01 2.64033347e-01 5.44419110e-01
2.79938698e-01 5.62798023e-01 5.42212665e-01 -1.24238372e+00
-1.24597192e+00 -5.42912364e-01 -2.33285457e-01 1.09923863e+00
-3.46517742e-01 -5.22836149e-01 -8.61755490e-01 -8.84691238e-01
-1.94914445e-01 5.45041144e-01 -6.70863032e-01 1.86124057e-01
-5.19232988e-01 -8.83848369e-01 2.34873548e-01 5.11963964e-01
9.44460690e-01 -1.22154737e+00 -5.70690632e-01 3.42359543e-01
-1.23295151e-01 -8.20236623e-01 4.51649278e-01 7.84476757e-01
-1.32617486e+00 -8.59701931e-01 -3.55772465e-01 -4.14098829e-01
4.44197267e-01 -3.05008471e-01 1.23621416e+00 -5.24020679e-02
3.88617888e-02 -3.94551247e-01 -5.30536890e-01 -6.30475283e-01
-5.79388678e-01 2.65524298e-01 -1.22438841e-01 5.84614575e-02
-5.47892302e-02 -2.22849056e-01 -1.05370365e-01 2.36087933e-01
-1.04335952e+00 3.20027173e-01 7.14753270e-01 1.15181100e+00
2.46367767e-01 4.42212194e-01 5.46708107e-01 -9.49608862e-01
8.00363839e-01 -2.48187035e-01 -6.05412066e-01 1.52733296e-01
-9.28838134e-01 3.65166664e-01 1.08685875e+00 1.05454862e-01
-9.95220065e-01 3.54948118e-02 -5.37804067e-01 -1.11171156e-01
-1.98096290e-01 9.68076885e-01 3.74476053e-03 -1.59704536e-01
8.45633149e-01 -6.85592592e-02 3.82279605e-02 -6.85991883e-01
1.34310082e-01 9.01647925e-01 5.16336858e-01 -3.41954082e-01
7.40899086e-01 3.68496895e-01 2.27435857e-01 -6.51656568e-01
-8.36085498e-01 -1.98705152e-01 -7.60166943e-01 -3.76360357e-01
7.82867074e-01 -3.48041892e-01 -7.65720725e-01 3.85068625e-01
-1.18383884e+00 5.33410050e-02 1.19985595e-01 2.63621986e-01
-3.79252464e-01 2.59091586e-01 -2.63997257e-01 -9.60313380e-01
-6.93617702e-01 -7.96959519e-01 6.07367158e-01 1.74656853e-01
-2.16982454e-01 -1.11839342e+00 -3.09963614e-01 6.58621550e-01
6.09716058e-01 5.10653079e-01 1.26672530e+00 -7.35097051e-01
-3.61342609e-01 -4.12138164e-01 -3.05953830e-01 4.46466416e-01
-1.65842455e-02 2.37773329e-01 -8.82973373e-01 5.37770614e-03
-6.99020410e-03 -4.70711648e-01 7.07474053e-01 3.72571945e-01
1.63221133e+00 -5.10444582e-01 3.34349344e-03 4.26837623e-01
1.23120487e+00 5.43167353e-01 7.09580839e-01 8.00476372e-01
5.80427527e-01 6.89693272e-01 7.70521343e-01 3.10465723e-01
2.06158116e-01 6.98545933e-01 9.85958695e-01 -4.04260963e-01
3.24653625e-01 -1.31790206e-01 -3.25736776e-02 7.20133722e-01
-2.03003630e-01 -3.54550958e-01 -1.12913477e+00 2.48438984e-01
-2.07543063e+00 -8.22441578e-01 -4.78799403e-01 2.03402853e+00
4.13166940e-01 6.73352301e-01 1.05743997e-01 8.04024458e-01
5.17297089e-01 4.44682837e-02 -2.55818278e-01 -1.13224649e+00
-1.27269581e-01 2.23621234e-01 3.38337839e-01 1.04629241e-01
-1.21149230e+00 7.64771283e-01 7.15584278e+00 1.05330718e+00
-1.29929459e+00 -6.70080334e-02 8.00010800e-01 2.75402635e-01
-8.19391280e-04 -4.42838490e-01 -4.97400254e-01 2.52647281e-01
1.04494989e+00 2.23224714e-01 1.86983630e-01 1.04404199e+00
3.43918681e-01 -7.08190501e-02 -7.89779544e-01 9.62717533e-01
6.18278943e-02 -1.48982298e+00 4.57275718e-01 -3.20404798e-01
4.97641236e-01 -3.81287307e-01 1.83185339e-01 3.89762551e-01
-4.04215530e-02 -1.50288177e+00 7.11565018e-01 1.34471253e-01
3.09713751e-01 -7.78764725e-01 1.23225927e+00 3.82442772e-01
-9.83982563e-01 -2.89475888e-01 -2.21839711e-01 -4.06205177e-01
-7.49026760e-02 5.79459310e-01 -1.16050518e+00 7.99349666e-01
1.06761992e+00 5.02628028e-01 -7.38338768e-01 1.04208136e+00
-1.81522846e-01 5.96648157e-01 -2.55563706e-01 -4.65128064e-01
6.44635022e-01 -1.17210761e-01 2.20812231e-01 1.10029399e+00
1.65287122e-01 -3.82351756e-01 -2.31027916e-01 5.81659257e-01
1.43312320e-01 4.59407985e-01 -8.41844261e-01 1.08507529e-01
1.23911174e-02 1.31438136e+00 -6.84352219e-01 -4.32244331e-01
-4.17067975e-01 3.59447658e-01 1.42134309e-01 -5.91538288e-02
-6.46334767e-01 -1.18888132e-01 -1.21013828e-01 2.34830633e-01
1.08863801e-01 -9.95008349e-02 -8.17262828e-01 -7.27402329e-01
8.33070949e-02 -1.15084445e+00 6.27147138e-01 -1.09103382e+00
-8.68590057e-01 1.11871362e+00 2.36298189e-01 -1.13848269e+00
-6.69813335e-01 -1.15510666e+00 -1.96936026e-01 5.63653827e-01
-1.20938647e+00 -1.23652768e+00 -5.53373516e-01 6.30048156e-01
5.01666069e-01 -3.41374636e-01 8.34680498e-01 1.31963536e-01
-5.47416568e-01 1.25739798e-01 3.41310233e-01 -4.71541807e-02
3.00360024e-01 -1.40492857e+00 1.66859373e-01 3.97474438e-01
1.91715941e-01 4.56717104e-01 7.90813208e-01 -3.68382692e-01
-1.19317853e+00 -8.39398682e-01 7.95993924e-01 -1.18331216e-01
3.76333565e-01 -2.94657767e-01 -1.06780660e+00 4.22693044e-01
3.38348001e-01 -3.61905873e-01 2.79674977e-01 2.47512370e-01
-2.27557525e-01 -4.92636412e-01 -1.07771742e+00 5.08510053e-01
3.58896345e-01 -9.02717113e-02 -7.48389721e-01 1.95039049e-01
3.06726605e-01 -3.33878428e-01 -8.18323016e-01 7.97246218e-01
6.85982466e-01 -1.27621198e+00 6.80744290e-01 -2.93690473e-01
5.88782132e-01 -1.78410739e-01 1.15654022e-01 -1.15076578e+00
-4.02501673e-02 -4.03749377e-01 -1.84476227e-01 9.61324394e-01
5.92987657e-01 -5.38958371e-01 1.18208456e+00 4.08111364e-01
2.54820250e-02 -1.01655531e+00 -9.67255890e-01 -6.22773170e-01
-1.36829652e-02 -3.05593848e-01 4.72391248e-01 8.00711989e-01
-6.74606934e-02 5.31141281e-01 -6.81966305e-01 -4.40088630e-01
3.77904832e-01 -1.28549263e-01 8.05031598e-01 -1.68257391e+00
1.84103087e-01 -4.73978430e-01 -5.23212135e-01 -6.24809980e-01
-8.11022967e-02 -5.45172274e-01 -2.51286834e-01 -1.74989033e+00
-6.12318516e-02 -4.65876490e-01 -1.99703231e-01 5.87681830e-01
7.46854544e-02 4.03818220e-01 1.61542609e-01 9.83641576e-03
-2.61905700e-01 4.39008504e-01 1.05573392e+00 -4.13339674e-01
-9.98871699e-02 4.25696634e-02 -4.40608531e-01 9.12830710e-01
8.85984600e-01 -4.80777234e-01 -2.45290428e-01 -2.00050220e-01
5.13603687e-01 1.04636893e-01 6.23947568e-02 -1.20024848e+00
8.29572603e-02 8.76934081e-02 4.87803221e-01 -1.07630956e+00
4.76907909e-01 -1.22249579e+00 3.11276793e-01 5.49705505e-01
-2.46208191e-01 3.06153744e-01 4.71810073e-01 7.27456659e-02
-7.50625193e-01 -5.46931326e-01 6.21582329e-01 -1.33078635e-01
-6.79408491e-01 8.65923092e-02 -4.65110630e-01 -4.90639746e-01
7.88434446e-01 -4.86881912e-01 -1.75225630e-01 -6.89673901e-01
-6.52949810e-01 3.79094593e-02 7.13665932e-02 5.19366622e-01
6.05030715e-01 -1.13745499e+00 -6.91374600e-01 8.57868195e-02
-7.63510168e-02 2.25497447e-02 -4.86797914e-02 5.90549529e-01
-9.13587153e-01 6.99903548e-01 -5.03884077e-01 -6.02194548e-01
-1.52063334e+00 7.49858916e-01 2.28544042e-01 -6.06860638e-01
-3.72831970e-01 4.32469517e-01 -1.46225214e-01 -6.29374564e-01
3.31876099e-01 -3.29620719e-01 -7.32790530e-01 1.90171510e-01
1.43901333e-01 3.77472043e-01 6.06267571e-01 -3.73060435e-01
-2.55666286e-01 5.97633123e-01 -2.69281641e-02 5.63307060e-03
1.29806507e+00 2.98434645e-01 -3.94282013e-01 5.64158976e-01
9.74956155e-01 -2.35270470e-01 -7.19871044e-01 4.28047441e-02
2.78014451e-01 -2.82265812e-01 1.97742149e-01 -1.00963652e+00
-8.89489949e-01 1.02677774e+00 4.87523109e-01 7.31918216e-01
1.38850284e+00 -4.97197866e-01 4.92695779e-01 6.61114991e-01
1.78651899e-01 -1.12904394e+00 7.26052597e-02 6.12220168e-01
1.09290469e+00 -1.31747115e+00 1.62912235e-01 -3.53989929e-01
-3.90616506e-01 1.68302763e+00 4.55319583e-01 2.84010500e-01
3.42380822e-01 3.86764824e-01 1.30145237e-01 -1.22215673e-01
-9.76951540e-01 2.76436880e-02 3.52460414e-01 6.39120519e-01
4.97750878e-01 -1.44779220e-01 -4.78351623e-01 1.87083259e-01
-3.87283951e-01 1.28546104e-01 1.43358007e-01 6.97309852e-01
-4.80702430e-01 -1.10052049e+00 -9.85083103e-01 5.46376348e-01
-7.84044504e-01 -1.47086188e-01 -5.73964894e-01 1.02903414e+00
1.01485007e-01 1.00808358e+00 -2.08104134e-01 -6.36830926e-01
1.43240660e-01 1.45099819e-01 1.28152043e-01 -9.98989716e-02
-8.71151209e-01 -3.17528605e-01 4.12507713e-01 -3.07453662e-01
-4.09458041e-01 -3.08264554e-01 -9.17065799e-01 -5.44450641e-01
-2.91977465e-01 3.83657336e-01 1.24276876e+00 8.78874004e-01
-1.03435047e-01 6.52845144e-01 5.86947978e-01 -6.77330256e-01
-2.35219896e-01 -1.04724562e+00 -1.57148853e-01 3.00122023e-01
1.84929192e-01 -7.58228660e-01 -2.23419026e-01 -2.15888191e-02] | [8.623873710632324, 4.0223565101623535] |
9680281a-2c92-4f16-baa8-b77239d6d686 | taspm-targeted-sequential-pattern-mining | 2202.13202 | null | https://arxiv.org/abs/2202.13202v1 | https://arxiv.org/pdf/2202.13202v1.pdf | TaSPM: Targeted Sequential Pattern Mining | Sequential pattern mining (SPM) is an important technique of pattern mining, which has many applications in reality. Although many efficient sequential pattern mining algorithms have been proposed, there are few studies can focus on target sequences. Targeted querying sequential patterns can not only reduce the number of sequences generated by SPM, but also improve the efficiency of users in performing pattern analysis. The current algorithms available on targeted sequence querying are based on specific scenarios and cannot be generalized to other applications. In this paper, we formulate the problem of targeted sequential pattern mining and propose a generic framework namely TaSPM, based on the fast CM-SPAM algorithm. What's more, to improve the efficiency of TaSPM on large-scale datasets and multiple-items-based sequence datasets, we propose several pruning strategies to reduce meaningless operations in mining processes. Totally four pruning strategies are designed in TaSPM, and hence it can terminate unnecessary pattern extensions quickly and achieve better performance. Finally, we conduct extensive experiments on different datasets to compare the existing SPM algorithms with TaSPM. Experiments show that the novel targeted mining algorithm TaSPM can achieve faster running time and less memory consumption. | ['Philip S. Yu', 'Wensheng Gan', 'Gengsen Huang'] | 2022-02-26 | null | null | null | null | ['sequential-pattern-mining'] | ['natural-language-processing'] | [ 6.59720421e-01 -6.48385584e-01 -3.25757980e-01 -1.62037119e-01
1.93880334e-01 -9.54509377e-02 -3.64024937e-02 9.86814424e-02
-3.85736674e-01 5.28309941e-01 -1.97012946e-01 -3.92280579e-01
-4.20870185e-01 -1.28478444e+00 -1.04432337e-01 -6.91971838e-01
-7.57257640e-02 4.28389072e-01 1.14338005e+00 -4.08467948e-02
5.27059138e-01 4.06910688e-01 -1.69614840e+00 6.53969467e-01
9.62529361e-01 8.90326440e-01 6.57868326e-01 1.12752110e-01
-7.25180328e-01 5.24490118e-01 -7.23292708e-01 -5.10054938e-02
2.82718420e-01 -6.54038370e-01 -6.53007984e-01 1.38023749e-01
-5.78720689e-01 1.48149654e-01 7.68141076e-02 1.12401223e+00
7.98732564e-02 1.50718555e-01 4.28340957e-02 -1.26477456e+00
-2.99822360e-01 9.05148685e-01 -1.07097042e+00 5.37178755e-01
3.56655926e-01 -9.11422744e-02 8.07633877e-01 -6.45873606e-01
4.52282369e-01 1.29511786e+00 5.85890234e-01 1.49180353e-01
-8.11259329e-01 -1.03244841e+00 4.10545141e-01 7.52859354e-01
-1.44863486e+00 2.00510457e-01 4.91486639e-01 5.48044108e-02
7.79883385e-01 6.91486955e-01 9.06123340e-01 5.21696568e-01
-1.56969745e-02 9.71228898e-01 9.85381782e-01 -4.93959814e-01
3.98707278e-02 8.18144754e-02 5.36333084e-01 4.42144632e-01
5.83312511e-01 -2.03139056e-02 -5.67503810e-01 -5.65644801e-01
2.90198922e-01 4.36570704e-01 -4.05016154e-01 -1.31049260e-01
-1.06968880e+00 7.29134262e-01 -2.87842780e-01 5.60653210e-01
-1.63099080e-01 -5.28390586e-01 5.58762729e-01 5.86158872e-01
3.33882719e-02 -1.07488886e-01 -6.56266093e-01 -2.35340267e-01
-6.83286786e-01 5.47309816e-01 8.04030240e-01 9.68934894e-01
8.38960707e-01 -2.80472159e-01 8.74135196e-02 9.04089630e-01
1.78570151e-01 4.29743975e-01 7.66651988e-01 -3.25554192e-01
3.09431642e-01 1.25208223e+00 -1.94053713e-03 -1.59952617e+00
-5.27317286e-01 -1.87463135e-01 -6.99208021e-01 -4.56991613e-01
-2.86643486e-02 1.07632577e-01 -4.58320171e-01 1.12856352e+00
5.39124787e-01 2.01661438e-01 -3.40827227e-01 6.52133346e-01
3.86140943e-01 9.88433480e-01 -3.13341767e-02 -1.01347339e+00
1.41010463e+00 -7.40012169e-01 -5.93557358e-01 -5.19001409e-02
4.61071998e-01 -9.44002390e-01 1.09242582e+00 8.00518692e-01
-6.59800410e-01 -5.75756729e-01 -8.98582935e-01 7.01056778e-01
-2.17779011e-01 -1.62456468e-01 6.16155386e-01 8.01597536e-01
-3.00433606e-01 5.04195809e-01 -7.82055855e-01 -5.45440793e-01
2.88585991e-01 3.32706928e-01 9.29084606e-03 -2.34023854e-03
-1.13919568e+00 5.40382028e-01 1.11452913e+00 -2.01771453e-01
-2.44342148e-01 -5.65143049e-01 -3.29502732e-01 4.92166243e-02
7.84301579e-01 -3.21877897e-01 1.17736757e+00 -7.52877772e-01
-9.12759125e-01 4.70109105e-01 -6.30573750e-01 -6.40636027e-01
2.00375125e-01 6.12756424e-03 -9.64641750e-01 8.49804282e-02
1.90557778e-01 -1.56878874e-01 4.44229871e-01 -7.31426477e-01
-1.34118974e+00 -3.01011801e-01 -5.36512375e-01 -5.69059663e-02
-7.69418836e-01 4.09234822e-01 -5.51411986e-01 -7.28046954e-01
2.42192864e-01 -6.66795731e-01 -7.30462849e-01 -4.48677123e-01
-1.14979617e-01 -4.75463122e-01 1.30265450e+00 -2.68862724e-01
1.88586879e+00 -2.21351337e+00 -2.28223190e-01 7.28305757e-01
-2.86868900e-01 3.78945440e-01 7.15519488e-02 6.69413269e-01
1.39851406e-01 6.33471087e-02 -4.64864373e-01 4.48947012e-01
-4.14844811e-01 5.90868890e-01 -4.60299224e-01 5.43538742e-02
-2.88079649e-01 4.37109590e-01 -6.70378447e-01 -7.18010724e-01
1.41645342e-01 -3.99649411e-01 -4.00326639e-01 2.69279275e-02
-3.85119736e-01 6.66168854e-02 -6.79846108e-01 6.13552928e-01
1.07120967e+00 -3.81676435e-01 5.18205941e-01 1.80960700e-01
-4.19582009e-01 -2.64017936e-02 -1.58134854e+00 1.14730120e+00
3.02897915e-02 -9.96333808e-02 -2.25695625e-01 -1.24270129e+00
1.33305895e+00 -8.19236711e-02 5.53961098e-01 -7.80813694e-01
-8.57872665e-02 3.33542913e-01 1.85303882e-01 -7.27838457e-01
2.63412058e-01 -1.11867249e-01 1.31204009e-01 7.42983103e-01
-6.35118365e-01 6.60760224e-01 7.29489148e-01 1.95797607e-02
1.14940596e+00 -3.17599446e-01 6.56056702e-01 -4.95660782e-01
1.02186036e+00 5.73667109e-01 1.20907128e+00 7.04917729e-01
2.01337963e-01 -2.47909818e-02 2.41353095e-01 -6.95020139e-01
-6.92868590e-01 -6.19983852e-01 -4.01737690e-02 8.47972393e-01
4.63878155e-01 -8.07309926e-01 -2.60825515e-01 -6.12918079e-01
1.74023941e-01 4.44067657e-01 -2.62836665e-01 1.24932239e-02
-9.62628424e-01 -1.26655936e+00 1.19641751e-01 1.87984884e-01
7.77757823e-01 -1.38327587e+00 -6.51188493e-01 5.74795067e-01
-1.50074959e-01 -8.52068663e-01 -4.16067153e-01 -1.88561499e-01
-9.75840569e-01 -1.52937162e+00 -5.12121506e-02 -9.39463735e-01
3.97152275e-01 8.75481546e-01 7.31128395e-01 4.44774747e-01
-3.62138510e-01 -3.33038360e-01 -9.71117973e-01 -8.73556793e-01
-3.66160125e-01 2.05995873e-01 1.02187559e-01 1.48552507e-01
1.06675279e+00 -7.95344174e-01 -4.84863043e-01 8.47781122e-01
-8.90701234e-01 4.94650416e-02 7.03377068e-01 5.93778074e-01
6.53977811e-01 7.16758668e-01 6.46257460e-01 -1.21086454e+00
6.06004298e-01 -5.36824167e-01 -5.70086658e-01 1.92441165e-01
-8.64776790e-01 -3.95122647e-01 8.29071641e-01 -4.53688711e-01
-1.27653432e+00 -1.21206328e-01 -3.73308152e-01 -1.33694783e-01
-2.03832462e-02 5.42896509e-01 -3.71710896e-01 3.76011170e-02
5.04855514e-01 9.43648696e-01 2.53468782e-01 -8.07634950e-01
-3.09297800e-01 8.15168500e-01 -3.19990286e-05 -3.21978182e-01
8.10490966e-01 5.84784806e-01 -3.96789014e-02 -7.55184352e-01
-6.15121841e-01 -7.85254240e-01 -2.86000639e-01 1.83820188e-01
2.07296073e-01 -3.33443016e-01 -9.02041137e-01 3.76886785e-01
-7.52287149e-01 2.17976689e-01 1.48852825e-01 1.73104569e-01
-2.33289003e-01 1.09991384e+00 -4.05749351e-01 -7.49079764e-01
-6.17865086e-01 -6.75301731e-01 2.30177864e-01 2.55341142e-01
-3.73188585e-01 -4.76048738e-01 1.30126193e-01 -4.87387069e-02
2.19330266e-01 -6.02296926e-02 1.06946087e+00 -8.60314369e-01
-5.71967304e-01 7.83776119e-02 -7.63671994e-02 -4.21637222e-02
4.11041290e-01 -4.17490788e-02 -1.85431525e-01 -1.46348909e-01
2.04139531e-01 3.22842568e-01 6.57100439e-01 -4.49064411e-02
1.67807615e+00 -4.76839304e-01 -9.95308936e-01 3.16308826e-01
1.22356105e+00 9.98716652e-01 5.23224533e-01 8.05446804e-01
1.86591804e-01 6.87952995e-01 1.49446464e+00 7.35747218e-01
-9.79709551e-02 5.56760430e-01 4.35617901e-02 3.79457921e-01
5.06227612e-01 -3.33031803e-01 7.29895607e-02 8.69838357e-01
2.15702951e-01 5.75336181e-02 -6.35110259e-01 5.54047465e-01
-2.20861888e+00 -1.30040240e+00 -6.68827474e-01 1.95441628e+00
8.41572762e-01 3.10850561e-01 4.94884700e-01 7.74166822e-01
8.30004871e-01 -1.48320735e-01 -3.08519602e-01 -4.53137785e-01
-2.40608186e-01 2.22753927e-01 1.34295449e-01 -8.84785950e-02
-9.41614270e-01 6.30007565e-01 6.15120029e+00 1.30244851e+00
-7.51363456e-01 1.27452403e-01 2.62548506e-01 -4.09836359e-02
-2.51900494e-01 1.47388712e-01 -1.09093785e+00 9.41185296e-01
6.38663590e-01 -5.44772387e-01 -1.11041695e-01 1.12664688e+00
4.52094525e-01 1.49932979e-02 -5.65673947e-01 1.08047211e+00
-2.95054108e-01 -1.46707141e+00 4.66824800e-01 -3.20360027e-02
4.19146895e-01 -3.79516751e-01 -4.94962245e-01 8.67479518e-02
-1.76410936e-02 -4.47171450e-01 2.00514734e-01 1.88937753e-01
1.48391991e-03 -1.11790788e+00 6.39181912e-01 8.37974846e-01
-1.40606356e+00 -5.24954855e-01 -8.29803169e-01 -1.73548967e-01
2.99794257e-01 1.07318175e+00 -5.50389290e-01 9.24973428e-01
1.12411225e+00 7.23421216e-01 -2.08474413e-01 1.31327569e+00
1.15857296e-01 7.28155732e-01 -5.14453769e-01 -5.92333674e-01
4.56124470e-02 -4.93447512e-01 5.02765715e-01 1.27631533e+00
3.69027764e-01 3.28219622e-01 6.40659392e-01 3.66151273e-01
5.27607679e-01 8.23244214e-01 -3.38522792e-01 1.96595982e-01
8.35268199e-01 1.05717313e+00 -7.80919135e-01 -3.17309171e-01
-6.94044471e-01 5.78309000e-01 -2.94450391e-02 -2.25197881e-01
-6.68010950e-01 -5.84970474e-01 4.56343621e-01 3.75986248e-01
4.99763489e-01 -2.44077072e-01 -3.77862185e-01 -8.58227968e-01
2.39491910e-01 -1.27071166e+00 1.09447694e+00 -9.54772905e-02
-1.19474173e+00 2.27084756e-01 3.93913873e-02 -1.26060259e+00
2.77741879e-01 -4.17926729e-01 -8.59072387e-01 5.79607844e-01
-1.06529629e+00 -6.99901223e-01 -2.76417971e-01 6.74555063e-01
5.96995473e-01 -2.19604492e-01 5.46491921e-01 3.74755532e-01
-7.08702743e-01 5.04203141e-01 -3.51787778e-03 -2.19928101e-01
2.65892088e-01 -5.59263527e-01 3.10148239e-01 1.11157465e+00
9.06578079e-02 9.98534679e-01 6.36607170e-01 -8.00675571e-01
-1.28165174e+00 -1.04444277e+00 6.67894900e-01 1.03503458e-01
5.55548847e-01 -2.58908495e-02 -1.18452406e+00 4.55308735e-01
-7.85512477e-02 -5.92844963e-01 1.02819550e+00 1.81054577e-01
1.67189594e-02 -3.00986022e-01 -1.16685367e+00 6.56771481e-01
1.50722909e+00 3.98509085e-01 -7.39216685e-01 2.47771740e-01
6.90979838e-01 8.62786621e-02 -7.07306743e-01 7.10153937e-01
5.58286250e-01 -1.06343114e+00 7.14107156e-01 -4.24259931e-01
-6.39876872e-02 -8.23897123e-01 1.39695868e-01 -6.05722964e-01
-6.10548556e-01 -5.20462334e-01 -2.48887792e-01 1.03918839e+00
3.03018451e-01 -8.55008006e-01 1.13017261e+00 -1.83147550e-01
4.06204015e-02 -1.08288002e+00 -6.62084222e-01 -1.00188112e+00
-5.20738363e-01 -4.99246359e-01 1.08027053e+00 7.94847310e-01
4.09000278e-01 7.56906532e-03 -4.75942403e-01 -3.59141491e-02
5.67963123e-01 9.18254673e-01 9.91058409e-01 -1.26661968e+00
-6.34903669e-01 -4.39406365e-01 -3.10670286e-01 -1.26667500e+00
-3.50780398e-01 -6.20898843e-01 -4.23463255e-01 -1.27841055e+00
5.20086527e-01 -6.46781683e-01 -2.68460393e-01 5.17447710e-01
-3.75024468e-01 -6.16591014e-02 -2.05450997e-01 6.23914659e-01
-4.69254881e-01 2.46934161e-01 1.09026301e+00 -3.68836038e-02
-3.25868219e-01 2.81697184e-01 -6.96920395e-01 6.12898827e-01
9.52716231e-01 -5.04852772e-01 -5.71703315e-01 -1.82224941e-02
1.43594548e-01 -2.62885660e-01 -3.88713956e-01 -7.52839625e-01
3.11842859e-01 -7.64862478e-01 3.02165538e-01 -1.09046900e+00
-3.36161137e-01 -8.70524108e-01 5.73890746e-01 1.21402752e+00
2.51698256e-01 7.19610155e-02 1.35844648e-01 6.96877658e-01
-2.24343836e-01 -4.65761870e-01 7.14937866e-01 -4.23358440e-01
-1.10304546e+00 5.14861345e-01 -6.61887944e-01 -3.81233573e-01
1.44679534e+00 -6.34114623e-01 -2.85733324e-02 3.26311558e-01
-5.61618805e-01 5.98971367e-01 3.99393022e-01 5.60869575e-01
7.59026766e-01 -1.21979034e+00 -4.25081819e-01 3.23331594e-01
2.95452178e-01 -2.35469583e-02 4.51110303e-01 7.35473514e-01
-7.15291619e-01 4.74511355e-01 -1.80713505e-01 -5.67502677e-01
-1.76584721e+00 1.03429639e+00 -3.90677862e-02 -3.15346301e-01
-9.26263809e-01 6.27616167e-01 5.18826805e-02 -3.35950881e-01
-9.56338421e-02 8.04446638e-02 -3.88009429e-01 -1.80985525e-01
1.07586229e+00 4.89460140e-01 -1.69153512e-02 9.06346664e-02
-4.47910666e-01 5.79172730e-01 -5.67165375e-01 4.07247007e-01
1.04548168e+00 -1.05746333e-02 -6.61363244e-01 2.14069858e-01
7.00038671e-01 2.22132921e-01 -3.02567661e-01 -3.62450391e-01
5.63654840e-01 -9.29733574e-01 -8.53480399e-01 -3.45738709e-01
-8.24919283e-01 3.58121455e-01 3.70944798e-01 4.67764258e-01
1.40315807e+00 -3.31357241e-01 1.25568485e+00 4.61691171e-01
8.60575378e-01 -1.03680980e+00 -1.06006265e-01 2.84903318e-01
5.27341247e-01 -6.35291278e-01 8.93017501e-02 -1.08778107e+00
-3.30691427e-01 1.04916823e+00 9.01689708e-01 -3.33019048e-02
7.14034975e-01 3.64840060e-01 -2.01372162e-01 4.47085835e-02
-8.12620044e-01 -9.03643966e-02 -3.82941902e-01 4.30242062e-01
6.16283789e-02 -1.27272502e-01 -1.22092271e+00 8.65552008e-01
-2.78835893e-01 4.17559631e-02 2.68130302e-01 1.25400424e+00
-1.17658746e+00 -1.73988473e+00 -4.22363281e-01 9.36194956e-01
-5.10500789e-01 2.21876651e-01 -2.81014532e-01 7.10693359e-01
3.46267849e-01 9.87365425e-01 4.74782251e-02 -6.39715970e-01
4.96063828e-01 -3.79926935e-02 3.07188839e-01 -6.53328419e-01
-3.52456033e-01 -1.38056412e-01 1.33236706e-01 -3.36765140e-01
-4.28684622e-01 -8.78738999e-01 -1.27061164e+00 -6.04458690e-01
-3.45317096e-01 7.35427797e-01 -1.28466457e-01 8.40460002e-01
4.40946311e-01 1.67992730e-02 7.53243983e-01 2.13826045e-01
-3.50200266e-01 -9.13131773e-01 -6.47971511e-01 3.91414344e-01
-5.50500631e-01 -6.25622869e-01 1.15067244e-01 -3.06386054e-01] | [8.31763744354248, 6.297606945037842] |
01dbde5e-1703-424c-b74b-f391ff7f55b1 | pandagpt-one-model-to-instruction-follow-them | 2305.16355 | null | https://arxiv.org/abs/2305.16355v1 | https://arxiv.org/pdf/2305.16355v1.pdf | PandaGPT: One Model To Instruction-Follow Them All | We present PandaGPT, an approach to emPower large lANguage moDels with visual and Auditory instruction-following capabilities. Our pilot experiments show that PandaGPT can perform complex tasks such as detailed image description generation, writing stories inspired by videos, and answering questions about audios. More interestingly, PandaGPT can take multimodal inputs simultaneously and compose their semantics naturally. For example, PandaGPT can connect how objects look in an image/video and how they sound in an audio. To do so, PandaGPT combines the multimodal encoders from ImageBind and the large language models from Vicuna. Notably, only aligned image-text pairs are required for the training of PandaGPT. Thanks to the strong capability of ImageBind in embedding data from different modalities into the same space, PandaGPT displays emergent, i.e. zero-shot, cross-modal behaviors for data other than image and text (e.g., video, audio, depth, thermal, and IMU). We hope that PandaGPT serves as an initial step toward building AGI that can perceive and understand inputs in different modalities holistically, as we humans do. Our project page is at https://panda-gpt.github.io/. | ['Deng Cai', 'Yan Wang', 'Jialu Xu', 'Huayang Li', 'Tian Lan', 'Yixuan Su'] | 2023-05-25 | null | null | null | null | ['instruction-following'] | ['natural-language-processing'] | [ 1.45850834e-02 1.37594774e-01 9.80392322e-02 -1.14825383e-01
-7.90980697e-01 -7.21081734e-01 7.58097410e-01 1.38245314e-01
3.29230167e-03 2.38137692e-01 8.26826692e-01 9.89171863e-02
3.31733733e-01 -7.30757356e-01 -1.08865499e+00 -4.13303524e-01
2.35480934e-01 1.97983056e-01 -1.19987629e-01 -9.87142697e-02
7.58861676e-02 5.83239272e-02 -1.92784595e+00 9.43902075e-01
1.06887646e-01 9.36888814e-01 5.90992689e-01 1.16661906e+00
-9.73032191e-02 1.29103804e+00 -2.64466345e-01 -3.64584208e-01
-2.47131795e-01 -3.36644590e-01 -9.91151094e-01 8.51192623e-02
7.84979880e-01 -8.27312350e-01 -8.47635031e-01 5.96635520e-01
5.48104882e-01 1.92306802e-01 5.21290541e-01 -1.51895368e+00
-1.08923864e+00 9.77041841e-01 -2.61550158e-01 -1.68981105e-01
6.72069669e-01 6.85306966e-01 1.16596889e+00 -1.02971673e+00
6.51713550e-01 1.55195343e+00 2.73885727e-01 8.00108135e-01
-1.25618887e+00 -4.54686046e-01 1.19955093e-01 3.30908060e-01
-1.19699764e+00 -6.00712061e-01 4.67542291e-01 -4.79306996e-01
1.06827545e+00 3.29378366e-01 9.08517599e-01 1.47531080e+00
-6.49546832e-02 1.14890802e+00 7.10073531e-01 -1.26894206e-01
-1.63540542e-01 -3.02153714e-02 -9.28124711e-02 7.52416372e-01
-4.06210154e-01 -8.19875896e-02 -1.44582486e+00 2.48472586e-01
8.93421292e-01 3.58841047e-02 -2.34636888e-01 -4.83522229e-02
-1.70304000e+00 4.60346639e-01 3.26959133e-01 8.02198946e-02
-2.67910600e-01 8.48338246e-01 2.94525921e-01 2.98363060e-01
-1.13572486e-01 3.96005869e-01 -1.19492561e-01 -6.11684442e-01
-5.88688016e-01 1.90246046e-01 5.11217356e-01 1.19214559e+00
7.10066855e-01 1.16215907e-01 -2.48934090e-01 5.97276926e-01
4.11672622e-01 9.10087705e-01 5.00079930e-01 -1.42854238e+00
4.08009499e-01 2.66681075e-01 -1.58132762e-01 -8.68216872e-01
-2.17114240e-01 2.34345660e-01 -4.17582929e-01 -1.04763910e-01
2.69411504e-01 -8.14167112e-02 -7.36592054e-01 1.92918348e+00
7.58492202e-02 4.43073139e-02 1.88269183e-01 9.35121775e-01
1.52466905e+00 1.20710933e+00 2.26256132e-01 2.07180142e-01
1.57302785e+00 -1.07526350e+00 -6.09414220e-01 -2.78950065e-01
5.10374904e-01 -6.82825625e-01 1.51978743e+00 4.82368678e-01
-1.45291841e+00 -7.85626709e-01 -7.25865841e-01 -5.89575589e-01
-1.95808873e-01 2.70071104e-02 3.29658747e-01 -1.07019380e-01
-1.16119349e+00 2.13710219e-01 -8.40586901e-01 -7.59384036e-01
1.95982724e-01 -7.45832622e-02 -6.87189758e-01 -1.19392008e-01
-9.00723815e-01 4.91974413e-01 4.31599468e-01 -1.46917611e-01
-1.38404953e+00 -5.53556800e-01 -1.00111556e+00 5.47574684e-02
2.68619895e-01 -9.03127670e-01 1.55470371e+00 -1.04082608e+00
-1.40593612e+00 8.44418764e-01 -1.70516044e-01 -3.14452171e-01
2.14627206e-01 -5.58749855e-01 -1.70115113e-01 6.55857086e-01
-9.42542106e-02 1.43897986e+00 9.30964410e-01 -1.27554727e+00
-2.41083980e-01 -2.13852406e-01 3.26262057e-01 5.41281819e-01
-4.93834853e-01 3.60373184e-02 -5.88607490e-01 -4.07780141e-01
-1.05926663e-01 -6.53796613e-01 2.62046218e-01 2.27811992e-01
-5.51413417e-01 3.78563553e-02 8.30741942e-01 -6.30503297e-01
9.04577494e-01 -2.42523861e+00 4.75812793e-01 -2.36464888e-01
4.25891846e-01 -2.35302716e-01 -4.09477890e-01 9.61238801e-01
1.76946133e-01 9.97799188e-02 2.42008433e-01 -4.38894272e-01
2.74148881e-01 2.68579245e-01 -5.30204117e-01 -1.74471363e-02
6.06577471e-02 1.25119567e+00 -9.45207775e-01 -6.08304501e-01
2.80254662e-01 6.04568958e-01 -6.90949440e-01 4.20303583e-01
-4.92047966e-01 4.73806560e-01 -1.59593433e-01 5.41444540e-01
1.25991106e-01 -5.51133394e-01 -2.23828554e-01 -5.38469791e-01
-1.09869771e-01 2.31211826e-01 -8.07241797e-01 2.03127241e+00
-5.31393528e-01 1.05343568e+00 -3.09164748e-02 -6.69596553e-01
4.47371304e-01 5.57277143e-01 2.98800528e-01 -8.07640851e-01
7.34465122e-02 -1.88384175e-01 -2.74920076e-01 -8.65974069e-01
7.45847344e-01 -1.47444636e-01 -1.49432540e-01 8.51555586e-01
5.20738125e-01 -4.81142282e-01 3.27418111e-02 7.34948516e-01
8.39253426e-01 2.54574805e-01 4.80927974e-02 3.84458542e-01
2.43138205e-02 -9.56261717e-03 -2.80332983e-01 9.31064725e-01
-6.04382483e-03 9.21453834e-01 3.49310666e-01 -1.23040952e-01
-9.74577725e-01 -1.50888062e+00 2.78729796e-01 1.44790840e+00
1.43442333e-01 -8.02090585e-01 -6.75844908e-01 -1.06618077e-01
-1.05294205e-01 7.30781734e-01 -4.82302189e-01 -1.32943660e-01
-2.37069875e-01 2.25921944e-02 7.73243606e-01 6.05781078e-01
3.95439178e-01 -1.43134212e+00 -6.36272192e-01 3.33742127e-02
-7.37087607e-01 -1.32421756e+00 -6.01354718e-01 -3.14497612e-02
-5.98896980e-01 -6.35635972e-01 -5.63993156e-01 -7.60725200e-01
3.63037258e-01 4.84217107e-01 1.09996963e+00 -1.82859659e-01
-2.47324541e-01 1.31527996e+00 -4.26201284e-01 -3.34765196e-01
-5.14230967e-01 -3.50587249e-01 -4.82960269e-02 8.39981809e-03
-9.21958983e-02 -6.21564269e-01 -5.84234953e-01 8.91803950e-02
-1.00324976e+00 9.72772598e-01 4.06504810e-01 5.21003962e-01
6.55147433e-01 -5.27476907e-01 2.91118026e-01 -3.60450506e-01
3.86739910e-01 -6.47319794e-01 6.10144623e-02 1.76673055e-01
2.95804262e-01 -1.73036575e-01 3.93257171e-01 -8.37658226e-01
-9.14418876e-01 -7.53693581e-02 -3.25456150e-02 -6.13424182e-01
-4.21858758e-01 6.39847040e-01 -8.90703276e-02 4.03500587e-01
6.64316058e-01 5.65324783e-01 1.22200727e-01 -2.17129722e-01
9.59069669e-01 6.62509322e-01 8.67826223e-01 -6.73320651e-01
5.62294602e-01 5.55838525e-01 -4.22369599e-01 -1.12811279e+00
-6.24367356e-01 -2.25784093e-01 -3.35324466e-01 -6.52747989e-01
1.16311955e+00 -1.23450065e+00 -1.06286418e+00 6.01270795e-01
-1.20584989e+00 -6.78917110e-01 -5.15698791e-01 3.44449729e-01
-9.98060107e-01 1.57689452e-01 -9.18469250e-01 -5.03463328e-01
-1.63536638e-01 -9.28261340e-01 1.24132729e+00 3.47820848e-01
-5.76566160e-01 -8.09336007e-01 -1.52408406e-01 7.20701575e-01
2.53911138e-01 1.76761858e-03 7.55028725e-01 -3.92159522e-01
-8.53695512e-01 6.28517866e-02 -4.17075634e-01 2.84979850e-01
3.57454941e-02 1.98836908e-01 -1.21136069e+00 -5.23258969e-02
-2.57223248e-01 -1.08052361e+00 6.41544819e-01 2.38207459e-01
1.22686088e+00 -6.06058419e-01 5.28228581e-02 4.22789663e-01
1.12896252e+00 -7.87041485e-02 4.28922266e-01 -1.12392500e-01
1.11788452e+00 6.64919198e-01 3.84700775e-01 5.55603623e-01
7.72354007e-01 5.35394907e-01 6.18283510e-01 1.22085825e-01
-5.27712584e-01 -6.94158316e-01 9.90387976e-01 1.13090265e+00
1.60311759e-01 -5.14476597e-01 -9.55300927e-01 5.50260663e-01
-1.84023356e+00 -1.20882583e+00 -4.05500457e-02 1.75200033e+00
8.80698740e-01 -2.50760883e-01 -2.07003094e-02 -3.72054517e-01
2.59327918e-01 2.73777604e-01 -5.31733155e-01 -4.98193383e-01
-4.31357563e-01 -1.48887545e-01 -1.22634150e-01 3.96171808e-01
-7.64794350e-01 9.72479880e-01 6.08454609e+00 6.40701175e-01
-1.26342022e+00 1.32192984e-01 3.09179574e-01 -5.01214445e-01
-6.29276216e-01 -1.50369376e-01 -3.15616012e-01 1.73977628e-01
1.04543662e+00 -1.91426665e-01 6.80321097e-01 3.09329361e-01
2.78165132e-01 -1.35976836e-01 -1.41493154e+00 1.25512946e+00
3.47771436e-01 -1.46228027e+00 5.90947390e-01 -2.30939239e-01
5.17668664e-01 2.84591258e-01 2.97433704e-01 1.36927396e-01
3.02339792e-01 -1.21360183e+00 1.36592484e+00 6.12026870e-01
9.58258510e-01 -2.24693745e-01 1.24135897e-01 3.12930405e-01
-1.17530465e+00 -1.82769895e-02 4.74371389e-02 -1.51411995e-01
3.86531293e-01 -1.65861979e-01 -6.05526328e-01 2.55264431e-01
7.33598173e-01 9.76050138e-01 -4.08805549e-01 6.28861725e-01
-2.02443749e-01 6.73114061e-01 -2.59034812e-01 1.62922479e-02
1.80729106e-01 3.03984702e-01 6.08770669e-01 1.17697966e+00
5.32391310e-01 2.73348570e-01 4.07527983e-02 1.01652384e+00
-1.04257837e-01 1.86734423e-02 -9.49129820e-01 -6.40411317e-01
4.02895451e-01 1.07499647e+00 -2.88443804e-01 -4.42524403e-01
-6.21616721e-01 1.06855249e+00 1.11770004e-01 6.50508165e-01
-8.94248366e-01 2.33159605e-02 5.94917119e-01 1.80400327e-01
2.02804878e-01 -3.77073556e-01 1.24922935e-02 -1.17601371e+00
-2.70187438e-01 -9.62699533e-01 3.08870494e-01 -1.95141399e+00
-1.09828627e+00 4.73731101e-01 1.62425965e-01 -1.28025484e+00
-2.69759446e-01 -5.45063019e-01 -5.58301687e-01 4.69927371e-01
-7.40602493e-01 -1.56379485e+00 -6.29760742e-01 8.95183086e-01
6.82136118e-01 1.50001019e-01 8.01829338e-01 1.83856934e-02
-1.77657127e-01 3.00496638e-01 -3.94985199e-01 6.60900101e-02
8.15518022e-01 -1.07040977e+00 2.33714372e-01 5.15564561e-01
6.65014148e-01 4.52696711e-01 7.71832466e-01 -4.40565705e-01
-1.73060918e+00 -7.60593534e-01 4.80306268e-01 -5.59416175e-01
9.33214307e-01 -3.11284900e-01 -7.35257804e-01 9.39112067e-01
7.26620913e-01 -2.45253235e-01 8.21806312e-01 -1.93499804e-01
-6.24450743e-01 1.07811853e-01 -6.00835323e-01 9.24833417e-01
1.01232719e+00 -1.18228269e+00 -5.66999197e-01 2.87409246e-01
1.14915252e+00 -4.86844599e-01 -7.42009282e-01 -1.13198906e-01
7.52180636e-01 -9.53672528e-01 1.01905119e+00 -7.00268745e-01
1.12231231e+00 -1.80240005e-01 -5.09060919e-01 -1.14762425e+00
-3.24402004e-04 -7.48573959e-01 -2.51405627e-01 1.34377074e+00
1.79799914e-01 -3.39410186e-01 3.57902169e-01 3.26482117e-01
-2.51310617e-01 -3.02075416e-01 -6.83670104e-01 -4.70579386e-01
-1.05682112e-01 -9.33148861e-01 3.97381127e-01 7.15700507e-01
3.65028232e-01 4.79544580e-01 -5.58594465e-01 1.09863102e-01
3.65301520e-01 7.43639246e-02 1.05334032e+00 -6.83317184e-01
-5.89935184e-01 -4.15848911e-01 -2.77838677e-01 -1.22169471e+00
-8.21707025e-02 -9.54334855e-01 9.48825777e-02 -1.70469892e+00
5.05271792e-01 1.11744486e-01 -7.49926120e-02 8.15978289e-01
2.19828919e-01 3.89660001e-01 6.83591366e-01 2.32626036e-01
-7.55086720e-01 5.92551708e-01 1.67144632e+00 -3.33386093e-01
3.63414660e-02 -6.72868907e-01 -7.96766520e-01 6.43456817e-01
7.18260527e-01 -5.95441312e-02 -5.74188411e-01 -8.24212670e-01
4.47017431e-01 4.33104217e-01 8.32252026e-01 -9.84380841e-01
1.93675891e-01 -2.32738912e-01 1.81941688e-01 -5.02714813e-01
7.18222320e-01 -5.44289112e-01 2.89305955e-01 1.13116302e-01
-6.10452414e-01 1.75170749e-01 5.02002954e-01 2.97826231e-01
-5.19955635e-01 1.28866926e-01 2.70364583e-01 -1.94497824e-01
-9.20370400e-01 1.99816763e-01 -6.84407651e-01 1.04371734e-01
5.93639791e-01 -3.32248151e-01 -7.92858660e-01 -1.06357956e+00
-9.87129092e-01 3.25502217e-01 4.75432038e-01 5.96870065e-01
9.16621327e-01 -1.43008256e+00 -6.41959071e-01 1.35058075e-01
4.77855057e-01 -4.88364510e-03 7.41119623e-01 8.73672128e-01
-4.57068652e-01 1.97059244e-01 -3.55854869e-01 -8.40267062e-01
-1.44240773e+00 2.93819010e-01 1.14862382e-01 3.65546793e-01
-6.80424392e-01 1.03173470e+00 8.17536473e-01 -2.93072164e-01
1.99511766e-01 -2.90803552e-01 1.95950661e-02 1.62455261e-01
7.63472199e-01 -5.59175052e-02 -5.60855508e-01 -7.92278886e-01
2.12907698e-02 5.24279833e-01 1.04164973e-01 -6.48169696e-01
1.19293857e+00 -4.97789234e-01 -5.87230176e-02 1.04628611e+00
1.17347383e+00 1.22288801e-02 -1.43248904e+00 -3.33447546e-01
-7.13007927e-01 -1.71309501e-01 -2.71715462e-01 -8.52016091e-01
-8.58105540e-01 1.37975740e+00 2.88242579e-01 1.68271661e-01
1.16813731e+00 6.08225107e-01 8.23885322e-01 4.41098154e-01
1.91601336e-01 -9.04395580e-01 8.97857070e-01 6.51879549e-01
1.30771542e+00 -1.13091969e+00 -3.85538280e-01 7.20579475e-02
-1.06761003e+00 1.14524961e+00 7.24847913e-01 4.08557445e-01
2.26225823e-01 1.78148732e-01 1.25612304e-01 -2.47910067e-01
-1.31193995e+00 -1.71528667e-01 3.01769733e-01 5.33849478e-01
4.09332126e-01 1.99528158e-01 5.73056579e-01 6.22528732e-01
-5.09599864e-01 -1.65312469e-01 7.54148901e-01 7.42120266e-01
-3.75916123e-01 -5.71664691e-01 -2.34193474e-01 1.49856895e-01
-2.02305652e-02 -3.28664720e-01 -4.39743191e-01 6.51783943e-01
-1.10940747e-01 8.74119639e-01 2.95946300e-01 -5.97746432e-01
1.42937392e-01 1.35809004e-01 6.91982150e-01 -6.52866185e-01
-4.56910312e-01 5.31586520e-02 1.63719401e-01 -8.75295341e-01
-3.76686871e-01 -7.41384983e-01 -1.68991756e+00 -3.73268306e-01
3.17830533e-01 -1.65141672e-01 4.92186010e-01 6.83955312e-01
3.41909707e-01 4.68613416e-01 2.16271520e-01 -9.45889771e-01
5.11704087e-02 -8.95383954e-01 -2.86611438e-01 5.29524088e-01
5.82011580e-01 -1.89756185e-01 -1.67313084e-01 5.07543564e-01] | [10.606285095214844, 1.2927532196044922] |
a3d71a48-776a-41c6-985c-b8efba9633da | a-spatio-temporal-identity-verification | 2111.00228 | null | https://arxiv.org/abs/2111.00228v2 | https://arxiv.org/pdf/2111.00228v2.pdf | whu-nercms at trecvid2021:instance search task | We will make a brief introduction of the experimental methods and results of the WHU-NERCMS in the TRECVID2021 in the paper. This year we participate in the automatic and interactive tasks of Instance Search (INS). For the automatic task, the retrieval target is divided into two parts, person retrieval, and action retrieval. We adopt a two-stage method including face detection and face recognition for person retrieval and two kinds of action detection methods consisting of three frame-based human-object interaction detection methods and two video-based general action detection methods for action retrieval. After that, the person retrieval results and action retrieval results are fused to initialize the result ranking lists. In addition, we make attempts to use complementary methods to further improve search performance. For interactive tasks, we test two different interaction strategies on the fusion results. We submit 4 runs for automatic and interactive tasks respectively. The introduction of each run is shown in Table 1. The official evaluations show that the proposed strategies rank 1st in both automatic and interactive tracks. | ['Jun Chen', 'Dongshu Xu', 'Shishi Wen', 'Ji Huang', 'Yue Zhang', 'Ankang Lu', 'Yanrui Niu', 'Zhongyuan Wang', 'Baojin Huang', 'Chao Liang', 'Jingyao Yang'] | 2021-10-30 | null | null | null | null | ['person-retrieval', 'instance-search'] | ['computer-vision', 'computer-vision'] | [ 2.26885289e-01 -3.51565987e-01 -1.54038474e-01 -5.48569918e-01
-1.14488411e+00 -6.00929558e-01 9.07025576e-01 -2.42963076e-01
-7.79439569e-01 5.35406649e-01 3.11621159e-01 4.27469164e-01
3.31465062e-03 -3.23216021e-01 -1.66108593e-01 -8.03064525e-01
1.83564410e-01 6.08867347e-01 5.55706620e-01 6.64302409e-02
3.48485976e-01 5.33030212e-01 -2.17527866e+00 6.60240412e-01
4.55601573e-01 1.03927255e+00 -1.69625916e-02 8.14343274e-01
-1.48061126e-01 5.52967370e-01 -6.52359366e-01 -3.60523045e-01
1.81115866e-01 -4.47246999e-01 -8.90909314e-01 7.59082362e-02
3.25610727e-01 -4.76668596e-01 -2.59900033e-01 9.49997365e-01
1.17553890e+00 6.25274539e-01 8.93424630e-01 -1.23234820e+00
-1.38499826e-01 1.35686204e-01 -5.98038316e-01 3.76408905e-01
1.13409674e+00 -2.35344656e-02 5.98595321e-01 -1.13622057e+00
7.36987770e-01 1.77725470e+00 6.15792610e-02 8.48151088e-01
-6.27248704e-01 -7.65364170e-01 -9.94989723e-02 5.73468745e-01
-1.76338410e+00 -7.98491836e-01 4.29127514e-01 -4.47223783e-01
8.43062103e-01 5.55282831e-01 4.28076476e-01 8.89081657e-01
-1.28414586e-01 1.10687339e+00 8.17545414e-01 -8.05489182e-01
-2.43849980e-04 2.64712483e-01 6.59763455e-01 5.13045967e-01
-7.90981501e-02 7.74413794e-02 -5.28917193e-01 -3.47654223e-01
4.13821161e-01 -5.18482290e-02 -1.05055563e-01 1.42109469e-01
-1.13263273e+00 5.30521154e-01 -4.81283478e-02 5.30837119e-01
-3.55366528e-01 5.18929437e-02 5.30200779e-01 4.49150465e-02
3.33349168e-01 -2.00397342e-01 -1.68328767e-03 8.82761553e-02
-9.72250819e-01 6.12624586e-01 6.01821005e-01 8.43554318e-01
3.70872706e-01 -5.52465558e-01 -9.81690764e-01 1.03353107e+00
4.13436711e-01 7.22750187e-01 4.88595724e-01 -7.53454685e-01
2.79416800e-01 4.81219143e-01 6.12888932e-01 -8.99720132e-01
-1.62946030e-01 4.77229476e-01 -2.51158804e-01 -1.65053964e-01
1.83962435e-01 2.77637281e-02 -1.15086174e+00 1.32499492e+00
6.19325161e-01 -2.12972853e-02 -8.29809830e-02 1.03379607e+00
1.54393542e+00 7.98507571e-01 4.71075118e-01 -5.53863943e-01
1.87392676e+00 -1.15207827e+00 -1.24624586e+00 4.52251554e-01
6.35085344e-01 -1.05772710e+00 6.47965908e-01 2.24169627e-01
-1.16174674e+00 -9.37516809e-01 -7.43874550e-01 6.51763156e-02
-6.30733371e-01 7.66765833e-01 2.30143905e-01 6.89930856e-01
-1.01360083e+00 6.64643422e-02 -4.78739262e-01 -7.74903715e-01
-1.38390690e-01 4.51157242e-01 -3.82183909e-01 3.60506494e-03
-1.52745390e+00 7.54402816e-01 6.35184646e-01 3.82636756e-01
-9.48196530e-01 3.13735276e-01 -4.91971880e-01 -1.29128069e-01
4.54610974e-01 -3.62131357e-01 1.25177240e+00 -8.82239223e-01
-1.27194154e+00 1.38816583e+00 -3.57767463e-01 -8.79771169e-03
5.97016156e-01 -4.64539945e-01 -6.88082278e-01 4.34839517e-01
1.12073772e-01 8.01826835e-01 9.53553975e-01 -9.30530488e-01
-9.26678598e-01 -5.76397300e-01 -1.47043332e-01 6.99388266e-01
-1.90526471e-01 1.09000731e+00 -1.42764223e+00 -4.21888888e-01
-1.85963571e-01 -7.93176949e-01 2.41806641e-01 -2.89477855e-01
-3.13075155e-01 -9.97742236e-01 8.04061890e-01 -7.69124568e-01
1.53772092e+00 -2.15641165e+00 7.50035793e-03 1.23531416e-01
-2.10108548e-01 3.39966983e-01 -1.87590122e-01 3.67172748e-01
-8.98300707e-02 6.11671992e-02 6.24727070e-01 -4.59459782e-01
-8.22068825e-02 -2.23213524e-01 4.53527942e-02 4.28246140e-01
-3.61423165e-01 7.60018349e-01 -6.27368391e-01 -1.16327178e+00
1.04420595e-01 3.51833969e-01 1.81815788e-01 3.90088946e-01
1.21640367e-02 1.05796963e-01 -1.00180352e+00 9.27444518e-01
6.44969821e-01 2.37750754e-01 -1.24313921e-01 -3.23564798e-01
-2.27019370e-01 -2.01025844e-01 -1.53392887e+00 1.53353202e+00
2.71746904e-01 2.28577137e-01 9.54424888e-02 -6.06284380e-01
4.71367240e-01 7.07605422e-01 4.15977985e-01 -7.52527714e-01
3.51751208e-01 -2.11667493e-02 -4.02057707e-01 -8.95706475e-01
5.36693811e-01 4.58099484e-01 3.02991699e-02 3.73402774e-01
5.18660098e-02 3.91552508e-01 6.00209832e-01 1.91527188e-01
7.41496086e-01 5.26305974e-01 2.70209789e-01 -1.97601497e-01
1.09425104e+00 -2.74532270e-02 3.28920633e-01 9.72281039e-01
-4.17912811e-01 4.52734470e-01 -3.09391748e-02 -5.70084155e-01
-5.12314498e-01 -6.64688528e-01 -1.11103013e-01 1.43381953e+00
3.45305085e-01 -4.35213447e-01 -1.01961052e+00 -8.87273073e-01
-3.43214959e-01 3.24958861e-01 -6.41999602e-01 1.34894505e-01
-4.66072798e-01 -6.96104944e-01 5.78145146e-01 2.60651976e-01
8.40510249e-01 -1.57623529e+00 -4.57902551e-01 -1.78830341e-01
-4.63378459e-01 -9.30068910e-01 -6.99616492e-01 -4.22884554e-01
-2.89572626e-01 -1.18072784e+00 -1.09925699e+00 -8.91056895e-01
4.18986589e-01 3.05983037e-01 7.35594034e-01 3.98443460e-01
-5.60021460e-01 8.52373242e-01 -6.83690310e-01 -5.33912063e-01
5.89807564e-03 -1.14626668e-01 1.90492764e-01 1.56678796e-01
8.79347563e-01 4.33548778e-01 -7.19429314e-01 6.19183123e-01
-6.66192651e-01 -3.09912920e-01 4.04826134e-01 4.96219665e-01
4.58977520e-01 -3.81018519e-02 -3.92539718e-04 -4.09112036e-01
6.70444906e-01 3.65812331e-03 -6.05801105e-01 9.47058558e-01
-3.38710964e-01 -6.05149306e-02 -4.23651159e-01 -4.20891523e-01
-1.32808805e+00 4.16663170e-01 7.31358379e-02 -1.80732384e-01
-4.78165388e-01 1.32891938e-01 -1.72869176e-01 -1.00482889e-01
7.34152079e-01 2.55995154e-01 -3.05130392e-01 -5.49342990e-01
1.31553691e-02 9.61858392e-01 2.50773907e-01 -4.35098857e-01
5.21311820e-01 3.99337858e-02 -4.99367923e-01 -7.97902405e-01
-6.93032086e-01 -1.13166237e+00 -5.29579401e-01 -8.97439599e-01
1.50516105e+00 -1.00559866e+00 -1.02094960e+00 6.66295052e-01
-1.36260629e+00 2.89044231e-01 2.10458413e-01 5.61500311e-01
-2.02627614e-01 5.32925785e-01 -5.25041699e-01 -1.39893091e+00
-7.83792853e-01 -1.17796469e+00 1.47554672e+00 5.03039837e-01
2.14953646e-01 -3.72729361e-01 1.81044340e-01 4.61557835e-01
3.32201347e-02 -8.98135751e-02 2.09812760e-01 -1.01949656e+00
-3.27773720e-01 -6.40724838e-01 -2.03781262e-01 -5.41534163e-02
-1.86179563e-01 -7.33137652e-02 -1.06175148e+00 -1.42181411e-01
-2.14626864e-01 -4.18702185e-01 9.05968428e-01 2.79123694e-01
9.79500890e-01 1.16800070e-01 -8.88180614e-01 -1.65706053e-01
1.19183803e+00 7.66340852e-01 9.04963970e-01 1.49721637e-01
1.60564825e-01 6.85128331e-01 1.09543312e+00 1.56823501e-01
-1.39934897e-01 1.26152551e+00 -2.26756915e-01 9.27009210e-02
-3.59148569e-02 1.54499635e-01 4.79291141e-01 1.06769867e-01
-7.69535244e-01 -6.55436039e-01 -8.04413736e-01 2.19591200e-01
-2.09760952e+00 -1.44050026e+00 -3.23119462e-02 2.34913445e+00
6.12485886e-01 -1.69245765e-01 4.16457027e-01 -7.62587264e-02
1.16166866e+00 -2.01517288e-02 -9.09426883e-02 1.36826159e-02
2.01010197e-01 -1.19638601e-02 5.05194739e-02 3.01758498e-01
-1.57610130e+00 1.02134013e+00 6.72244310e+00 1.15010595e+00
-4.60427135e-01 2.18857870e-01 5.12934744e-01 -2.16785409e-02
4.42683220e-01 -2.38355502e-01 -1.52520001e+00 3.31980765e-01
7.25662351e-01 2.09789239e-02 1.67070612e-01 7.07049608e-01
2.00940594e-01 -6.50471151e-01 -9.53699887e-01 1.39923334e+00
4.53430176e-01 -7.53552258e-01 3.21409941e-01 -2.54445016e-01
2.64454484e-01 -5.85924506e-01 -4.13606346e-01 5.71185887e-01
-1.19653314e-01 -7.44885385e-01 5.88623166e-01 1.01747334e+00
6.81683362e-01 -6.85094059e-01 9.61197257e-01 3.15650553e-01
-1.57113433e+00 4.97830100e-03 -1.80698559e-01 4.43144917e-01
1.30738392e-01 -8.22520405e-02 -2.77874708e-01 7.99576402e-01
9.52032804e-01 2.02188164e-01 -4.74982858e-01 1.26299202e+00
-1.05025925e-01 1.77452520e-01 -3.03732306e-01 -2.65080899e-01
-1.91095948e-01 -5.61250634e-02 4.31903303e-01 1.28190923e+00
8.24447721e-02 5.47912598e-01 4.92720395e-01 2.51824051e-01
1.27373055e-01 5.12422740e-01 -3.25455666e-01 4.57641631e-02
2.96179742e-01 1.43356323e+00 -8.47434044e-01 -6.60567045e-01
-3.39048743e-01 1.19996631e+00 -1.99551880e-01 5.05229771e-01
-9.88879323e-01 -4.91386354e-01 -4.57897373e-02 1.76602248e-02
2.92405632e-04 2.12368697e-01 6.44385755e-01 -9.83174086e-01
-1.05313867e-01 -7.50021815e-01 8.57621849e-01 -8.90095592e-01
-8.03748310e-01 6.37342393e-01 6.31689727e-01 -9.99033928e-01
-3.89854431e-01 -3.49322498e-01 -4.30993438e-01 8.05454016e-01
-8.71956408e-01 -1.06750333e+00 -4.08244789e-01 9.11103010e-01
9.76080239e-01 -3.14567715e-01 8.54091227e-01 4.95279342e-01
-7.12560296e-01 6.49610221e-01 -2.19752908e-01 2.56248802e-01
1.08816588e+00 -6.75722480e-01 -1.83645979e-01 6.86508596e-01
3.53155620e-02 6.29388809e-01 5.94352067e-01 -8.93272460e-01
-1.14458370e+00 -5.68771243e-01 1.14711010e+00 -4.06137019e-01
1.17269121e-02 -3.35878581e-01 -3.29208165e-01 4.15665686e-01
3.31631333e-01 -1.37753308e-01 4.41295475e-01 -3.87158059e-02
1.19305439e-01 3.25843617e-02 -1.17012215e+00 3.54780793e-01
9.68295932e-01 -4.22010660e-01 -7.25258231e-01 7.91871309e-01
3.04753542e-01 -2.92129129e-01 -3.77869099e-01 4.69791532e-01
8.51885915e-01 -6.72557592e-01 1.04628956e+00 -5.18070996e-01
-2.46374384e-01 -4.71788108e-01 -3.74278352e-02 -2.84550995e-01
-1.87115312e-01 -3.29674989e-01 1.28075451e-01 1.26361871e+00
1.07438512e-01 -1.41021743e-01 4.61743295e-01 7.60754466e-01
4.31589186e-01 -2.23097950e-01 -6.54014349e-01 -6.03602231e-01
-7.74697900e-01 1.61325652e-03 6.26506703e-03 3.89885962e-01
-1.37174770e-01 4.12486732e-01 -5.16354382e-01 -4.89556268e-02
5.55477262e-01 -7.23675191e-02 6.99588299e-01 -1.26762497e+00
2.73865797e-02 -6.90225363e-02 -4.47108120e-01 -8.72000635e-01
-8.29153955e-02 -3.95179212e-01 2.42482334e-01 -1.44802570e+00
8.93590152e-01 2.46755078e-01 -3.84793937e-01 5.47229588e-01
-2.17567667e-01 3.79046619e-01 1.79276153e-01 4.75541681e-01
-1.21597672e+00 2.87556380e-01 8.01444948e-01 -5.55652976e-02
-2.67158359e-01 1.17060035e-01 1.59633636e-01 5.37526548e-01
3.72378916e-01 -3.71352553e-01 -2.86286950e-01 -1.36735290e-02
-1.74844891e-01 2.71093309e-01 2.78280288e-01 -1.15880692e+00
3.95934075e-01 3.86331417e-03 7.35902488e-01 -1.11477995e+00
4.75407660e-01 -5.46982408e-01 9.55321342e-02 3.85263205e-01
-3.50718468e-01 1.00669391e-01 7.80510679e-02 5.41078687e-01
-3.82128686e-01 -4.37086403e-01 4.87207502e-01 -2.57470191e-01
-9.90186274e-01 3.21147650e-01 -5.69948196e-01 -3.82639647e-01
1.35223031e+00 -3.54275078e-01 -3.84615585e-02 -4.70598370e-01
-1.32268631e+00 4.55414355e-01 -1.79272622e-01 4.90423918e-01
5.76224387e-01 -1.40949750e+00 -6.73453987e-01 -1.30597040e-01
1.70403793e-01 -8.34993899e-01 4.71155524e-01 8.13730061e-01
-3.38845611e-01 5.87021947e-01 -4.86419499e-02 -4.01705503e-01
-2.10168314e+00 7.99744070e-01 1.93166658e-01 -3.28469783e-01
-9.87365469e-02 8.51948619e-01 2.11929351e-01 3.24383676e-02
8.20512295e-01 4.91713464e-01 -1.05513442e+00 3.62776846e-01
1.10744703e+00 4.81119275e-01 -1.84040099e-01 -8.08183968e-01
-6.40454292e-01 6.69926584e-01 -3.84407014e-01 -4.76123542e-01
6.94707215e-01 -1.01907812e-01 -1.38920024e-01 1.43933088e-01
9.18708265e-01 -2.54717797e-01 -3.59870285e-01 2.63824146e-02
1.20314695e-01 -3.18558693e-01 -7.58226737e-02 -9.15011883e-01
-7.76114047e-01 4.56616729e-01 1.39494014e+00 1.99614674e-01
1.33374810e+00 2.23505422e-02 3.51243675e-01 5.48542440e-01
3.71145695e-01 -1.55472708e+00 -1.08803071e-01 3.31452042e-01
1.11556816e+00 -1.18911433e+00 2.34466076e-01 -5.69907725e-01
-4.28969473e-01 9.92529392e-01 6.87116206e-01 1.04026593e-01
7.14334130e-01 -9.52739865e-02 -9.65179056e-02 -3.52421910e-01
-6.08442247e-01 -5.88828027e-01 7.74358988e-01 1.88015833e-01
5.05871654e-01 -3.50028634e-01 -1.21085715e+00 4.28259701e-01
5.98866105e-01 2.02284217e-01 -3.55817348e-01 1.02611804e+00
-6.29571080e-01 -1.16482425e+00 -6.99041367e-01 2.21934021e-01
-6.76855266e-01 4.45230491e-02 -7.40332901e-01 8.81307781e-01
1.09362379e-01 1.08624387e+00 -1.06874086e-01 -4.19587195e-01
4.79672670e-01 5.19344151e-01 5.61516881e-01 -3.20871770e-01
-6.93245173e-01 4.59066272e-01 3.56017470e-01 -9.29520369e-01
-8.60587537e-01 -8.10927153e-01 -1.05548429e+00 8.29162747e-02
-5.61644852e-01 4.89262819e-01 5.62292814e-01 9.73309159e-01
1.05959233e-02 -1.36601433e-01 3.37538511e-01 -8.86961937e-01
-3.65963191e-01 -1.27482450e+00 -4.04327989e-01 3.93333554e-01
-2.74229318e-01 -8.85324657e-01 -2.98478216e-01 2.66075432e-01] | [14.640499114990234, 0.8229175209999084] |
37470063-1e87-4124-a3bd-65232b82261c | multi-scale-alignment-and-spatial-roi-module | 2207.01345 | null | https://arxiv.org/abs/2207.01345v1 | https://arxiv.org/pdf/2207.01345v1.pdf | Multi-scale alignment and Spatial ROI Module for COVID-19 Diagnosis | Coronavirus Disease 2019 (COVID-19) has spread globally and become a health crisis faced by humanity since first reported. Radiology imaging technologies such as computer tomography (CT) and chest X-ray imaging (CXR) are effective tools for diagnosing COVID-19. However, in CT and CXR images, the infected area occupies only a small part of the image. Some common deep learning methods that integrate large-scale receptive fields may cause the loss of image detail, resulting in the omission of the region of interest (ROI) in COVID-19 images and are therefore not suitable for further processing. To this end, we propose a deep spatial pyramid pooling (D-SPP) module to integrate contextual information over different resolutions, aiming to extract information under different scales of COVID-19 images effectively. Besides, we propose a COVID-19 infection detection (CID) module to draw attention to the lesion area and remove interference from irrelevant information. Extensive experiments on four CT and CXR datasets have shown that our method produces higher accuracy of detecting COVID-19 lesions in CT and CXR images. It can be used as a computer-aided diagnosis tool to help doctors effectively diagnose and screen for COVID-19. | ['Arcot Sowmya', 'Dadong Wang', 'Hongyan Xu'] | 2022-07-04 | null | null | null | null | ['covid-19-detection'] | ['medical'] | [ 2.49762729e-01 -5.75787961e-01 1.63975686e-01 -9.83992070e-02
-4.43650991e-01 -4.37374800e-01 1.58477142e-01 3.31618458e-01
-6.97564542e-01 5.15457153e-01 4.87108417e-02 -4.26390946e-01
1.85039788e-02 -8.14958453e-01 -4.19166863e-01 -6.79746211e-01
-1.00142799e-01 2.31725380e-01 4.41115588e-01 1.45974532e-01
-1.35181084e-01 8.78650784e-01 -9.01975393e-01 4.80444431e-01
8.02112222e-01 7.40869462e-01 8.37133706e-01 6.57154977e-01
1.17542446e-01 6.00949347e-01 -5.32088339e-01 1.10098265e-01
4.85878736e-02 -3.23051125e-01 -5.00598729e-01 -1.62161812e-01
-3.19728046e-03 -7.04943120e-01 -3.51199508e-01 9.14787114e-01
6.30976677e-01 -1.64853349e-01 8.80132675e-01 -6.94370568e-01
-4.11646008e-01 -1.70554787e-01 -9.28493202e-01 9.66622114e-01
-6.47896677e-02 2.77276456e-01 3.55206549e-01 -8.86469483e-01
6.59068942e-01 1.07474029e+00 7.58274019e-01 3.49879146e-01
-7.88830876e-01 -5.87763727e-01 -6.23829514e-02 3.35946739e-01
-1.34040153e+00 4.61272478e-01 4.02724117e-01 -6.42855704e-01
9.33314681e-01 3.24903607e-01 8.13469410e-01 8.40804517e-01
7.29732573e-01 6.40629649e-01 9.17208552e-01 6.65301159e-02
-1.14259265e-01 -1.00421928e-01 1.24426614e-02 5.61528146e-01
5.55172205e-01 -8.92299712e-02 3.61702919e-01 -3.07147563e-01
1.24936211e+00 9.35681462e-01 -6.43902183e-01 9.16698426e-02
-1.26797140e+00 9.18000758e-01 9.02468383e-01 6.02776408e-01
-6.74526513e-01 -1.61025122e-01 5.00489533e-01 -1.71275377e-01
4.31115963e-02 2.94633329e-01 -3.70205551e-01 5.26316583e-01
-5.01356721e-01 9.09698009e-02 -1.07193582e-01 3.95731211e-01
3.58373702e-01 -2.58422792e-01 -5.91401458e-01 8.72940242e-01
2.11951628e-01 8.97249877e-01 6.64525390e-01 -4.87766474e-01
4.63308960e-01 7.27085114e-01 8.27428997e-02 -1.05283844e+00
-8.61372113e-01 -3.85978252e-01 -1.33161855e+00 -3.49123552e-02
5.27472086e-02 -2.70221204e-01 -1.10913610e+00 1.33501065e+00
1.89981371e-01 3.27559590e-01 -3.02671343e-01 1.16212451e+00
9.52136815e-01 7.53642917e-01 2.83384353e-01 -2.96854585e-01
1.94485044e+00 -6.73279166e-01 -6.10132635e-01 -2.10888475e-01
5.19366980e-01 -7.02706993e-01 8.00336003e-01 9.05072093e-02
-6.89179063e-01 -3.58472168e-01 -9.96047914e-01 1.64027125e-01
-3.75011206e-01 2.32693985e-01 3.25222671e-01 2.80764788e-01
-7.43317723e-01 1.76376700e-01 -8.33896339e-01 -3.73689741e-01
6.34968221e-01 1.09549850e-01 -3.26141596e-01 -3.75814199e-01
-1.24701142e+00 9.70064461e-01 1.92526683e-01 2.62831241e-01
-7.88847566e-01 -6.05859160e-01 -6.09713852e-01 5.32165878e-02
1.90186277e-01 -8.76403868e-01 8.52423608e-01 -2.15720221e-01
-5.90480745e-01 8.78809929e-01 -2.22779319e-01 -1.36825159e-01
3.19277257e-01 -2.71121562e-02 -3.15305978e-01 6.98479950e-01
1.11321725e-01 6.36545837e-01 9.31649268e-01 -1.08619642e+00
-4.50305879e-01 -5.78939915e-01 -1.31540611e-01 7.63983727e-02
9.36443433e-02 4.81546998e-01 -4.25698638e-01 -9.47780311e-01
7.40000606e-02 -7.55573273e-01 -5.12326837e-01 -2.75419019e-02
-1.53830141e-01 -1.43980891e-01 9.63372588e-01 -9.74882245e-01
1.16199684e+00 -1.98394847e+00 -4.63223368e-01 1.13696851e-01
5.26079297e-01 8.36881280e-01 -1.06821775e-01 -1.24963615e-02
-6.27092421e-02 2.80253530e-01 -2.85941511e-01 1.67963400e-01
-6.34483814e-01 1.42762750e-01 7.97125045e-03 6.11083627e-01
4.59028751e-01 1.13409984e+00 -6.48754478e-01 -9.37371016e-01
4.84384120e-01 8.08771133e-01 -4.16505754e-01 3.11686724e-01
8.09658617e-02 7.41458952e-01 -7.76907742e-01 5.75672388e-01
1.11436534e+00 -7.82217562e-01 -1.55610591e-01 -1.42698944e-01
9.74648818e-02 -2.54779667e-01 -6.78763092e-01 1.05819476e+00
-2.92009026e-01 3.67280632e-01 2.76821494e-01 -8.24102521e-01
4.84551221e-01 5.25699794e-01 5.66892326e-01 -5.77449739e-01
4.25805390e-01 -5.44224866e-02 1.39541805e-01 -1.00371277e+00
-7.69760683e-02 -2.14021906e-01 2.57824391e-01 4.87807989e-01
-4.72872496e-01 3.36585604e-02 -1.10546842e-01 -1.03214473e-01
1.02758324e+00 -7.20227659e-01 4.80898529e-01 3.18227820e-02
7.05938339e-01 -1.22560628e-01 7.16560960e-01 8.04119647e-01
-4.68437165e-01 9.69122529e-01 1.04689695e-01 -6.22096181e-01
-9.62396383e-01 -1.26350117e+00 -4.16290820e-01 4.17482048e-01
2.30424196e-01 1.83333009e-01 -6.87242091e-01 -6.32066607e-01
-2.42327407e-01 1.51892126e-01 -6.29961491e-01 1.70837641e-01
-1.07508135e+00 -9.42004502e-01 3.81556690e-01 7.37856328e-01
5.43705344e-01 -1.45270932e+00 -1.12198734e+00 2.60087520e-01
-4.49943781e-01 -1.11529088e+00 -5.16952634e-01 -7.75464252e-02
-9.38989282e-01 -1.24928415e+00 -1.25315607e+00 -1.02507341e+00
8.35573792e-01 6.22052312e-01 7.84819245e-01 4.07872170e-01
-8.85914564e-01 -1.16944820e-01 -2.40537480e-01 -3.84376436e-01
-1.19662724e-01 -2.13582158e-01 -2.92766660e-01 -3.65395486e-01
4.17786598e-01 -2.23311141e-01 -1.09849858e+00 2.41309926e-01
-9.79724288e-01 -4.08895925e-04 7.22419322e-01 8.46152306e-01
6.64774776e-01 2.47379038e-02 4.31450516e-01 -7.41061151e-01
6.78392529e-01 -4.95528817e-01 -3.79637688e-01 1.89383283e-01
-8.92372206e-02 -6.52683139e-01 7.26099253e-01 -2.10856438e-01
-1.02179003e+00 -2.88745821e-01 -2.53063828e-01 -7.21701026e-01
-2.73793489e-01 2.04432532e-01 1.31665885e-01 2.55896032e-01
4.36549246e-01 3.01438481e-01 -1.78493485e-01 -4.47117418e-01
-3.23998630e-01 8.46873999e-01 4.40239459e-01 1.04906358e-01
4.50305700e-01 8.30895782e-01 -1.89915411e-02 -9.28825736e-01
-5.24521887e-01 -7.01031923e-01 -4.35696334e-01 3.07216105e-04
1.44334531e+00 -9.68082607e-01 -7.03679502e-01 4.52267438e-01
-1.48638773e+00 2.94241726e-01 1.80349275e-01 8.17839384e-01
-1.73452348e-02 4.45893139e-01 -9.58739638e-01 -4.02402610e-01
-7.87989736e-01 -1.52178979e+00 1.07132542e+00 4.67819452e-01
-1.24169178e-02 -7.32523859e-01 -1.85969006e-03 3.69691104e-01
4.55016166e-01 2.46696487e-01 1.20648170e+00 -4.43437129e-01
-7.50160575e-01 -2.24496961e-01 -8.28371167e-01 3.85079414e-01
2.02782348e-01 -2.75816053e-01 -9.14460063e-01 -4.57868576e-01
2.62043625e-01 1.21905312e-01 9.83402908e-01 8.95320356e-01
1.36123049e+00 2.04193015e-02 -6.92098916e-01 6.51960611e-01
1.52906632e+00 7.38224566e-01 4.43302870e-01 -6.29610866e-02
7.62513936e-01 3.81730646e-01 4.70034420e-01 3.92762810e-01
1.11376144e-01 3.68775696e-01 3.92393738e-01 -5.87851584e-01
4.88476306e-02 9.58958492e-02 -3.07325393e-01 5.86869776e-01
-2.47791111e-01 -2.58019656e-01 -9.46856260e-01 6.55312061e-01
-1.41951621e+00 -7.86563516e-01 -2.94723660e-01 1.73008454e+00
4.53233749e-01 -2.57064104e-01 -1.78777650e-01 -1.93177491e-01
1.00894606e+00 6.45963708e-03 -6.90111041e-01 -3.39458048e-01
-9.56712812e-02 2.77230054e-01 3.82333398e-01 1.29098639e-01
-1.27420986e+00 2.65598327e-01 6.29794407e+00 6.77447975e-01
-1.36159265e+00 3.71382684e-01 6.35553658e-01 1.74124002e-01
-2.60806791e-02 -5.97528636e-01 -5.22558451e-01 5.97327292e-01
8.62286165e-02 2.00646907e-01 5.12616113e-02 6.60656571e-01
4.74885941e-01 -9.83578563e-02 -5.49318612e-01 1.09798098e+00
-1.48468623e-02 -1.27319789e+00 1.49265766e-01 -1.09132394e-01
7.47599423e-01 2.97658622e-01 -3.88422348e-02 6.22514449e-02
-2.39291862e-01 -1.22547638e+00 -2.02416882e-01 3.99903446e-01
9.07370567e-01 -6.50422096e-01 1.17649674e+00 2.67516851e-01
-1.28993952e+00 1.93832833e-02 -5.88007689e-01 2.96626121e-01
1.63560614e-01 4.73668545e-01 -1.07246697e+00 2.59990960e-01
1.02179575e+00 3.41373771e-01 -3.64807725e-01 1.20120370e+00
-1.86312899e-01 3.06506693e-01 -2.91546524e-01 -5.73038682e-02
2.71701276e-01 -3.11372690e-02 6.16501153e-01 1.45733249e+00
2.42648453e-01 4.57034111e-01 8.89638215e-02 9.44890022e-01
4.64197733e-02 2.61341780e-01 -6.11847460e-01 2.52585918e-01
2.27556527e-01 1.10651720e+00 -1.04365265e+00 -5.57010114e-01
-5.37636995e-01 7.28586376e-01 -1.38287082e-01 5.17873406e-01
-9.13605452e-01 -4.12185669e-01 5.86204112e-01 3.09134305e-01
7.57282257e-01 2.56238461e-01 -3.15105736e-01 -1.11396766e+00
-9.48329046e-02 -6.97635889e-01 3.88288438e-01 -8.16667795e-01
-1.34975183e+00 6.56482160e-01 -3.64159122e-02 -1.30366719e+00
5.72486967e-02 -5.87795198e-01 -9.55311358e-01 9.94920731e-01
-1.52129257e+00 -7.84059703e-01 -5.27296126e-01 6.13667011e-01
3.31309378e-01 9.78587642e-02 7.52621353e-01 3.29105318e-01
-4.40495998e-01 2.60123789e-01 1.20280296e-01 2.63246506e-01
4.33784425e-01 -8.81989717e-01 1.13101900e-01 7.76271284e-01
-5.47138333e-01 8.64833415e-01 1.78325757e-01 -8.14308047e-01
-7.65457630e-01 -1.29311800e+00 6.59643769e-01 -1.79604635e-01
3.74040082e-02 -2.98203658e-02 -1.00364876e+00 4.65656370e-01
2.02782005e-01 3.29285145e-01 4.98813897e-01 -7.28129447e-01
-5.77561446e-02 1.36813611e-01 -1.46182191e+00 4.09243345e-01
6.22832417e-01 -4.73306149e-01 -8.78927410e-01 2.49381661e-01
7.85524905e-01 -1.83231100e-01 -7.20643282e-01 8.78773212e-01
3.40588748e-01 -9.31443453e-01 1.22055387e+00 -3.70063454e-01
2.92228609e-01 -4.26706403e-01 1.56315640e-01 -9.80944276e-01
-4.14548874e-01 -6.84872419e-02 3.52339268e-01 4.43904191e-01
-1.12318076e-01 -6.97273314e-01 5.75249612e-01 -1.35838270e-01
1.39323667e-01 -9.57087040e-01 -6.74828053e-01 -3.15416873e-01
3.32733653e-02 -3.77142012e-01 5.59058309e-01 8.46117139e-01
-4.00757849e-01 8.45076293e-02 -4.84058000e-02 2.39549413e-01
3.97440672e-01 3.23912531e-01 2.49966726e-01 -1.06963015e+00
-1.73805565e-01 -3.72858316e-01 -3.10649842e-01 -8.03061068e-01
-5.98113060e-01 -7.10809946e-01 -1.70927107e-01 -1.95617735e+00
5.93738317e-01 -3.10578853e-01 -5.40897131e-01 2.59860367e-01
-4.78399813e-01 5.39404988e-01 2.36801848e-01 2.88380980e-01
-2.57146150e-01 1.14661470e-01 1.98502254e+00 -1.36682883e-01
-4.28865217e-02 2.07063463e-02 -3.23755890e-01 1.05030847e+00
8.18761766e-01 -4.86216456e-01 -2.53756642e-01 -2.30740577e-01
-1.13804169e-01 2.28099018e-01 4.43974197e-01 -7.21370041e-01
-7.06325248e-02 -5.03231548e-02 8.92690718e-01 -1.17057514e+00
2.37201273e-01 -8.71770620e-01 -9.18375924e-02 9.51754987e-01
2.11808890e-01 1.88598961e-01 2.47279495e-01 3.95757586e-01
-2.28098407e-01 -1.18781343e-01 1.01470053e+00 -3.32888365e-01
-3.96862060e-01 4.71941024e-01 -9.64278877e-01 1.77733049e-01
1.28953397e+00 -1.33014783e-01 -3.56519938e-01 2.87691299e-02
-5.08821905e-01 2.29480520e-01 2.73502439e-01 2.11806759e-01
9.82082784e-01 -9.41627920e-01 -7.80062556e-01 5.02268493e-01
1.11343704e-01 2.76016980e-01 8.33716512e-01 1.19409907e+00
-1.13839543e+00 7.99480677e-01 -2.31199518e-01 -8.47860754e-01
-1.40138328e+00 8.14133167e-01 3.89565736e-01 -6.69973195e-01
-9.14601684e-01 7.54583955e-01 9.19985592e-01 -1.23608917e-01
1.07786492e-01 -7.02739239e-01 -5.01279116e-01 -6.49422705e-02
8.82010400e-01 3.49239260e-02 -7.44077237e-03 -7.22718596e-01
-6.55446410e-01 8.32934499e-01 -3.43438655e-01 5.75963080e-01
1.19709492e+00 -1.07261397e-01 -1.34582341e-01 -1.75492808e-01
1.32682872e+00 -2.57292300e-01 -9.32446122e-01 -9.59571227e-02
-5.58670521e-01 -4.83350694e-01 -1.09250389e-01 -6.46122515e-01
-1.03901994e+00 1.22866738e+00 9.01589215e-01 -1.95713546e-02
1.21828043e+00 1.68659955e-01 1.04387021e+00 1.54525533e-01
1.72668308e-01 -6.90469563e-01 1.31399278e-02 2.51190364e-01
7.78908074e-01 -1.19093621e+00 -4.42008786e-02 -3.96837443e-01
-6.10247433e-01 8.70620072e-01 5.00026464e-01 -3.15250903e-01
8.23486328e-01 3.67866755e-01 2.54522979e-01 -3.66599023e-01
-4.31299895e-01 -5.14608882e-02 1.15164116e-01 8.14700663e-01
1.55875608e-01 1.94883972e-01 -2.87490249e-01 6.44434988e-01
4.30037707e-01 -2.26409873e-03 2.91730225e-01 1.01904523e+00
-5.17148197e-01 -5.30801356e-01 -7.32134581e-01 7.93429434e-01
-7.39182293e-01 -1.86851695e-01 1.39997313e-02 7.50489712e-01
6.45575523e-01 8.37047279e-01 1.51151538e-01 -1.15021244e-01
1.93021357e-01 -3.10077012e-01 3.92535031e-01 -5.29601455e-01
-5.96402287e-01 2.84798592e-01 -5.30824006e-01 -3.49826634e-01
-2.52001464e-01 -3.37343305e-01 -1.58378565e+00 -1.47270272e-02
-1.58336565e-01 -2.41273463e-01 4.37894583e-01 8.73050630e-01
9.38503295e-02 8.13248992e-01 4.76690769e-01 -4.26247656e-01
-1.14604585e-01 -9.54642832e-01 -6.86490655e-01 3.08938384e-01
6.27685547e-01 -5.26565254e-01 -1.66285172e-01 3.33838984e-02] | [15.54178237915039, -1.7393099069595337] |
9a28b92f-f300-45b7-91ce-17dbd904f2b6 | interpolating-item-and-user-fairness-in | 2306.10050 | null | https://arxiv.org/abs/2306.10050v1 | https://arxiv.org/pdf/2306.10050v1.pdf | Interpolating Item and User Fairness in Recommendation Systems | Online platforms employ recommendation systems to enhance customer engagement and drive revenue. However, in a multi-sided platform where the platform interacts with diverse stakeholders such as sellers (items) and customers (users), each with their own desired outcomes, finding an appropriate middle ground becomes a complex operational challenge. In this work, we investigate the ``price of fairness'', which captures the platform's potential compromises when balancing the interests of different stakeholders. Motivated by this, we propose a fair recommendation framework where the platform maximizes its revenue while interpolating between item and user fairness constraints. We further examine the fair recommendation problem in a more realistic yet challenging online setting, where the platform lacks knowledge of user preferences and can only observe binary purchase decisions. To address this, we design a low-regret online optimization algorithm that preserves the platform's revenue while achieving fairness for both items and users. Finally, we demonstrate the effectiveness of our framework and proposed method via a case study on MovieLens data. | ['Djallel Bouneffouf', 'Negin Golrezaei', 'Jason Cheuk Nam Liang', 'Qinyi Chen'] | 2023-06-12 | null | null | null | null | ['fairness', 'fairness'] | ['computer-vision', 'miscellaneous'] | [-2.74393737e-01 4.55874130e-02 -5.10721743e-01 -3.77220899e-01
-6.09472811e-01 -8.32425952e-01 6.35395423e-02 3.23545545e-01
-5.24173677e-01 5.23148358e-01 1.38627157e-01 -2.17195183e-01
-4.09702986e-01 -7.83117235e-01 -4.47193623e-01 -4.75904018e-01
-1.05287895e-01 1.73817545e-01 -4.43742454e-01 -3.51764530e-01
4.51752990e-01 -7.84929434e-04 -1.28470659e+00 1.36982009e-01
1.08477545e+00 1.43248320e+00 4.70922999e-02 2.42136613e-01
4.06921394e-02 7.40791559e-01 2.95953210e-02 -1.13669384e+00
1.11064863e+00 -2.61186380e-02 -3.06079417e-01 5.89533225e-02
1.61633119e-01 -7.17467368e-01 1.49873853e-01 1.15032530e+00
3.27715397e-01 2.30452657e-01 2.90779412e-01 -1.73202419e+00
-8.89550209e-01 7.49956667e-01 -6.77949727e-01 3.74235841e-03
5.19099273e-02 -1.63883284e-01 1.72212958e+00 -5.81222773e-01
3.95643353e-01 6.42422974e-01 3.03839743e-01 3.00989747e-01
-1.20711100e+00 -5.11189461e-01 5.05252838e-01 -1.52610928e-01
-1.00630832e+00 -4.15587395e-01 4.90437567e-01 -4.90189403e-01
2.03444332e-01 5.49712539e-01 4.76089567e-01 3.61104995e-01
-4.11070511e-02 8.28545749e-01 8.69412720e-01 -1.47212595e-01
4.06575531e-01 4.21410441e-01 1.91811964e-01 -2.61938155e-01
5.31690717e-01 5.05710132e-02 -5.05687058e-01 -4.92389649e-01
3.98274571e-01 4.04753894e-01 -1.96280256e-01 -3.50699335e-01
-5.11290729e-01 1.02490413e+00 3.07486624e-01 -2.96898305e-01
-9.02476847e-01 -1.42187744e-01 8.36991966e-02 5.96337438e-01
4.08636332e-01 4.01373237e-01 -2.30174720e-01 -9.29451827e-03
-9.60565627e-01 4.30596292e-01 9.79089439e-01 1.13912845e+00
3.03730488e-01 -3.03707778e-01 -2.26181433e-01 4.62486237e-01
5.66358924e-01 5.98083809e-02 -6.19072951e-02 -1.07320130e+00
6.46873534e-01 4.59369987e-01 1.13965166e+00 -1.12499905e+00
-5.17479926e-02 -7.73890853e-01 -4.24673617e-01 1.19959727e-01
6.52007282e-01 -3.52032453e-01 2.45044291e-01 1.86553860e+00
4.90815639e-01 -1.39854625e-01 -1.93858683e-01 1.47761524e+00
1.26991183e-01 4.22966272e-01 3.95937897e-02 -6.56703234e-01
1.17882550e+00 -9.89232898e-01 -6.51013196e-01 4.39640284e-02
7.69122839e-02 -8.78340006e-01 7.77323663e-01 4.79925603e-01
-1.58023250e+00 1.63638130e-01 -6.88992918e-01 1.89187471e-02
1.41597271e-01 -1.26396134e-01 6.14792645e-01 6.18345678e-01
-7.34964788e-01 5.57583332e-01 -1.99931726e-01 -1.46413118e-01
4.09454942e-01 4.70544636e-01 2.11830780e-01 5.89737706e-02
-1.19377887e+00 6.04622126e-01 -3.50637674e-01 9.72966701e-02
-6.09696031e-01 -8.55902195e-01 -1.14069819e-01 3.43723267e-01
8.78118932e-01 -9.12924111e-01 1.41976869e+00 -1.30082464e+00
-1.44935834e+00 5.13935149e-01 3.92753690e-01 -3.10404688e-01
1.10336542e+00 -1.39546126e-01 -1.22657456e-01 -4.22803432e-01
1.49808735e-01 -8.52498934e-02 4.60898548e-01 -1.23922801e+00
-1.30126572e+00 -5.80366671e-01 6.86499894e-01 5.88752329e-01
-4.89013582e-01 2.68944263e-01 -1.39497980e-01 -2.30725318e-01
-3.85597974e-01 -7.39939690e-01 -4.58383352e-01 -8.24877154e-03
-2.59699941e-01 9.37448740e-02 1.04277052e-01 -5.26492476e-01
1.28072941e+00 -2.10909939e+00 -1.86515570e-01 4.51528579e-01
3.21027458e-01 -2.62323856e-01 -1.48595586e-01 5.58117270e-01
5.82036614e-01 2.86969692e-01 3.44015360e-01 -4.37259078e-01
6.36805177e-01 -1.16112597e-01 -1.38144851e-01 4.91178602e-01
-4.57792729e-01 6.85519218e-01 -6.92317307e-01 1.89378895e-02
-2.84353465e-01 -7.28075355e-02 -9.74479139e-01 3.45249295e-01
-3.19702514e-02 1.17602095e-01 -7.69129097e-01 6.79138362e-01
9.59580302e-01 -3.75870585e-01 3.93030137e-01 1.61624461e-01
-4.18178797e-01 2.05360875e-02 -1.40403247e+00 1.10023487e+00
-6.84919596e-01 -3.01307924e-02 8.11932564e-01 -5.44464171e-01
5.74039280e-01 -7.99756199e-02 4.70160306e-01 -7.14298666e-01
4.10766482e-01 2.05009177e-01 3.71667854e-02 -1.81013912e-01
1.01033580e+00 -4.50339973e-01 -1.16140120e-01 9.96348441e-01
-6.00915253e-01 7.09517896e-01 -1.77873448e-01 3.79621208e-01
6.10204279e-01 -1.43400431e-01 4.63708639e-01 -4.02037621e-01
1.01750098e-01 -1.77599609e-01 6.04789138e-01 8.93784881e-01
-4.93924588e-01 1.64662763e-01 6.47000134e-01 -1.65488675e-01
-9.53185856e-01 -9.01500583e-01 1.81183130e-01 1.51370597e+00
7.29934514e-01 6.03800528e-02 -3.05289239e-01 -4.20669019e-01
6.19086027e-01 8.03336561e-01 -5.02754807e-01 3.22387457e-01
1.37922764e-01 -6.44933999e-01 -2.54084021e-01 1.54090524e-01
1.23556338e-01 -2.10918710e-01 -5.43849349e-01 3.18343818e-01
-3.77743006e-01 -1.00316978e+00 -1.15740931e+00 -4.02731568e-01
-4.15587485e-01 -9.28161383e-01 -4.17454869e-01 -2.64244586e-01
6.54944777e-01 8.00208509e-01 1.06558323e+00 -5.56762591e-02
2.03938439e-01 2.70977348e-01 -3.33337814e-01 -2.66692162e-01
5.55545986e-02 -1.19614348e-01 9.13872942e-02 4.53931153e-01
1.52976573e-01 -4.51200575e-01 -1.13088441e+00 6.78737044e-01
-9.29509759e-01 -1.86778605e-02 8.47587883e-02 4.78761494e-01
4.74023938e-01 1.84859172e-01 1.24352396e+00 -9.86495495e-01
1.09169388e+00 -1.13433397e+00 -9.37788248e-01 2.93735385e-01
-8.63975286e-01 -4.95086133e-01 1.03540552e+00 -2.81943947e-01
-1.20897126e+00 -3.28010321e-01 4.38809425e-01 3.89669426e-02
6.56402886e-01 5.33354521e-01 -4.28272337e-01 -4.67387959e-02
1.40499622e-01 -2.76672542e-01 1.04331970e-01 -4.57386822e-01
6.82282090e-01 9.23746169e-01 1.70612782e-01 -6.65976882e-01
5.08442879e-01 4.53927070e-01 -3.05700362e-01 2.08356641e-02
-8.46348226e-01 -4.32857215e-01 8.07012171e-02 -2.82990336e-01
1.82917655e-01 -9.88202214e-01 -1.47538400e+00 -5.99608123e-02
-7.81417668e-01 1.31351843e-01 -3.11784565e-01 4.38908458e-01
-4.35592383e-01 1.83306187e-01 -4.30036455e-01 -1.25700092e+00
-4.82649177e-01 -8.92316639e-01 3.69448960e-01 5.94963551e-01
7.32640103e-02 -7.90709853e-01 -1.35477439e-01 7.11579502e-01
7.17570782e-01 8.25429484e-02 4.22825664e-01 -7.85672009e-01
-8.69921386e-01 -3.77897650e-01 -3.54337364e-01 1.09978154e-01
1.57585487e-01 -2.33268961e-01 -4.88380611e-01 -5.77557802e-01
1.09667800e-01 -7.68426508e-02 3.43957655e-02 3.44467461e-01
8.61547172e-01 -8.00543904e-01 7.23241419e-02 2.52817541e-01
1.64631498e+00 6.92961970e-03 1.17199779e-01 4.38267767e-01
1.44784719e-01 9.34991956e-01 1.12156069e+00 1.30878329e+00
9.26200569e-01 7.70447254e-01 7.70140529e-01 3.34062904e-01
7.74721324e-01 -4.19036269e-01 1.95008025e-01 4.60327893e-01
-2.47143224e-01 -4.47229207e-01 -1.83641836e-01 3.05176109e-01
-2.30732465e+00 -8.62380922e-01 1.06208086e-01 2.71967435e+00
4.85085130e-01 -2.05520838e-01 6.40682101e-01 -3.64351392e-01
1.06860614e+00 -2.48709649e-01 -8.45001400e-01 -6.68407917e-01
4.83697578e-02 -4.24813360e-01 8.14022422e-01 3.04329932e-01
-6.21842146e-01 4.58313376e-01 5.40731716e+00 6.05577469e-01
-8.44661474e-01 2.48800188e-01 9.88098145e-01 -8.98553133e-01
-8.59762371e-01 1.09424248e-01 -6.04467392e-01 7.55589128e-01
7.86497355e-01 -9.77344334e-01 9.93288994e-01 7.15328455e-01
7.49718368e-01 8.97970200e-02 -1.16607201e+00 6.22777104e-01
-2.35849217e-01 -1.31765819e+00 -5.06175041e-01 5.44286907e-01
1.02543461e+00 -3.22521806e-01 2.51647264e-01 1.60237715e-01
4.98459160e-01 -7.47495592e-01 1.01946509e+00 5.41751206e-01
2.95664936e-01 -9.94948387e-01 5.86062670e-01 5.59465408e-01
-8.99839520e-01 -4.90264803e-01 -2.51173675e-01 -4.49975938e-01
4.73629862e-01 5.37849784e-01 -1.88656345e-01 6.63045883e-01
1.85659632e-01 1.75175890e-01 2.39482611e-01 1.18237329e+00
2.16881588e-01 6.24759234e-02 -1.88041210e-01 -1.54534141e-02
1.56930223e-01 -8.34966540e-01 1.88752964e-01 5.95918834e-01
5.29724181e-01 3.74842286e-01 2.63276428e-01 9.64752495e-01
-6.15972042e-01 7.75303483e-01 -1.27603486e-01 -1.22719593e-02
7.86724448e-01 1.67231011e+00 -4.64817613e-01 1.95328906e-01
-5.68028748e-01 5.07848084e-01 2.86054611e-01 1.46719724e-01
-9.82892811e-01 1.28634229e-01 1.02533937e+00 3.27222019e-01
5.39607704e-02 2.91114271e-01 -5.36318421e-01 -1.13960922e+00
1.58644632e-01 -8.41491759e-01 4.86161709e-01 -4.06272978e-01
-1.79452384e+00 -2.36874521e-02 -5.65803826e-01 -1.27792263e+00
2.94830978e-01 -2.58929223e-01 -8.40263844e-01 8.92614663e-01
-1.95014942e+00 -1.01225519e+00 4.50378060e-02 4.28395569e-01
-3.99387181e-02 1.04411289e-01 2.02441677e-01 6.75633550e-01
-5.85456014e-01 8.29947591e-01 6.35352492e-01 -6.67080522e-01
5.17539203e-01 -1.01116788e+00 -9.51534510e-02 7.15786457e-01
-4.13592547e-01 6.89182222e-01 7.88115978e-01 -4.07431334e-01
-1.90554452e+00 -9.77909207e-01 8.27636182e-01 -3.00727785e-02
8.76687706e-01 -2.07496688e-01 -3.52635443e-01 4.85960782e-01
-4.05585878e-02 -7.64873624e-02 1.26659012e+00 2.54312158e-01
-3.85508053e-02 -3.43646049e-01 -1.72036219e+00 6.98269069e-01
1.06292486e+00 -2.26824239e-01 -6.67538568e-02 2.81520605e-01
5.28089702e-01 -4.79976013e-02 -1.01461923e+00 -1.50637105e-01
9.08542573e-01 -9.34998751e-01 6.48806751e-01 -9.19173658e-01
5.08827090e-01 -9.29660127e-02 -4.97011006e-01 -1.26757658e+00
-3.55154425e-01 -9.72740769e-01 1.22022055e-01 1.36114955e+00
4.14904147e-01 -6.90243423e-01 7.89789557e-01 1.67408490e+00
2.75185227e-01 -6.73401475e-01 -7.84499228e-01 -5.61685264e-01
1.36038333e-01 -2.15175763e-01 9.51425910e-01 7.99579740e-01
3.66936862e-01 -1.04747809e-01 -7.93731153e-01 2.39466205e-01
8.36394370e-01 7.87880599e-01 7.50979602e-01 -9.70495224e-01
-5.28182685e-01 -4.68759209e-01 3.41294318e-01 -9.72692609e-01
-1.02770701e-01 -8.81365776e-01 -1.24728896e-01 -1.18896842e+00
5.90517223e-01 -8.06497693e-01 -6.79392695e-01 1.90600306e-02
4.15469520e-03 -3.17220413e-03 7.89964736e-01 2.75961101e-01
-1.05419981e+00 3.25496852e-01 1.32508051e+00 3.22193742e-01
-2.02816010e-01 3.96187365e-01 -1.91953123e+00 3.39352369e-01
6.19379461e-01 -3.00511122e-01 -4.59643841e-01 -2.37507239e-01
7.62262642e-01 4.08326089e-01 1.02374768e-02 3.14977080e-01
3.48902971e-01 -8.59130859e-01 -2.50823289e-01 -1.30673394e-01
-1.50340004e-02 -1.03878117e+00 4.89858091e-01 7.06345364e-02
-5.82556605e-01 -1.06157564e-01 -5.16864002e-01 7.66941726e-01
6.56455830e-02 -2.62719303e-01 6.45641267e-01 2.01521302e-03
-1.16709493e-01 7.46969700e-01 -2.34255772e-02 6.63194954e-02
1.34819746e+00 4.67533171e-02 -6.08857691e-01 -8.22031438e-01
-5.64122200e-01 8.76364827e-01 8.92564237e-01 3.01523864e-01
1.78712621e-01 -1.30946469e+00 -6.39562190e-01 -3.08202863e-01
1.71438307e-01 -4.98295546e-01 5.08658886e-01 8.01213622e-01
-2.94779558e-02 6.85850158e-02 -2.76685923e-01 1.46827191e-01
-8.84863198e-01 4.91220146e-01 2.60373414e-01 -3.03065866e-01
2.27709547e-01 6.40331030e-01 2.55901158e-01 -1.83921620e-01
4.92238291e-02 1.20628521e-01 -1.02619678e-01 4.41973120e-01
5.46813428e-01 7.40998507e-01 -1.69864878e-01 -4.85281348e-01
-1.88937485e-01 -3.27475481e-02 -1.34783641e-01 5.85990846e-02
1.36108756e+00 -8.64831150e-01 2.67986786e-02 -9.07219723e-02
8.31422925e-01 4.11843479e-01 -1.38442183e+00 -2.60486275e-01
-5.49834594e-02 -1.24002314e+00 2.25990996e-01 -8.59913051e-01
-1.35990131e+00 1.58113092e-01 1.67125821e-01 7.89224148e-01
1.19289494e+00 -5.00943422e-01 8.89746308e-01 -3.60088080e-01
7.01394737e-01 -1.31087232e+00 -3.26144159e-01 -2.31651425e-01
5.16286731e-01 -1.41486931e+00 -1.11586139e-01 -2.42184117e-01
-9.06970263e-01 7.49829888e-01 4.00104702e-01 -2.78998137e-01
8.13482106e-01 -9.45221186e-02 -1.15453944e-01 2.35863447e-01
-9.08823192e-01 -1.07108936e-01 3.11569721e-02 -2.14556083e-02
3.85683835e-01 6.92468345e-01 -8.91622245e-01 1.62782216e+00
-1.21165670e-01 2.08333746e-01 1.00853467e+00 8.72491539e-01
-3.37768883e-01 -1.23010635e+00 -1.59534901e-01 7.88314462e-01
-9.40143824e-01 -1.25707611e-01 -1.07377969e-01 3.47313106e-01
1.83403283e-01 1.21202087e+00 1.82024352e-02 -5.45090809e-02
5.04405379e-01 -5.19270360e-01 1.26387402e-01 -3.53682667e-01
-8.98473620e-01 2.52418131e-01 1.14416935e-01 -4.21445936e-01
-2.38763005e-01 -7.22720861e-01 -8.20960999e-01 -8.98819804e-01
-5.79946697e-01 3.40891182e-01 7.67645180e-01 7.93218195e-01
5.94227493e-01 7.27895722e-02 1.48815310e+00 -3.10577989e-01
-1.27927387e+00 -1.67477980e-01 -1.30240679e+00 5.41603982e-01
6.78942725e-02 -4.51018840e-01 -5.01236796e-01 -4.33722913e-01] | [9.575563430786133, 5.579000473022461] |
3df00d9d-5cb1-46cc-aaac-9af428476833 | focused-proofreading-efficiently-extracting | 1409.1199 | null | http://arxiv.org/abs/1409.1199v1 | http://arxiv.org/pdf/1409.1199v1.pdf | Focused Proofreading: Efficiently Extracting Connectomes from Segmented EM Images | Identifying complex neural circuitry from electron microscopic (EM) images
may help unlock the mysteries of the brain. However, identifying this circuitry
requires time-consuming, manual tracing (proofreading) due to the size and
intricacy of these image datasets, thus limiting state-of-the-art analysis to
very small brain regions. Potential avenues to improve scalability include
automatic image segmentation and crowd sourcing, but current efforts have had
limited success. In this paper, we propose a new strategy, focused
proofreading, that works with automatic segmentation and aims to limit
proofreading to the regions of a dataset that are most impactful to the
resulting circuit. We then introduce a novel workflow, which exploits
biological information such as synapses, and apply it to a large dataset in the
fly optic lobe. With our techniques, we achieve significant tracing speedups of
3-5x without sacrificing the quality of the resulting circuit. Furthermore, our
methodology makes the task of proofreading much more accessible and hence
potentially enhances the effectiveness of crowd sourcing. | ['Stephen M. Plaza'] | 2014-09-03 | null | null | null | null | ['art-analysis'] | ['computer-vision'] | [ 3.04562569e-01 1.41967326e-01 4.72597867e-01 -2.88406350e-02
-5.77225983e-01 -7.96712458e-01 3.47500771e-01 2.50382721e-01
-8.29856336e-01 8.95483494e-01 -2.90126622e-01 -4.64832574e-01
8.79961401e-02 -4.22250122e-01 -7.35930145e-01 -4.90429938e-01
3.01847756e-01 4.79951352e-01 5.73716819e-01 6.90825060e-02
7.57648766e-01 4.92718697e-01 -1.52524817e+00 7.88556039e-02
7.85144627e-01 6.56625390e-01 5.86030960e-01 9.35843706e-01
-3.60327154e-01 6.16323948e-01 -7.86158264e-01 -5.20922959e-01
2.39465967e-01 -4.92072523e-01 -1.08067203e+00 -1.22344576e-01
4.84104246e-01 -3.10771674e-01 2.00978845e-01 1.09704351e+00
3.67578030e-01 -3.98035675e-01 1.73309803e-01 -1.05915999e+00
-2.37476245e-01 2.30323076e-01 -5.78762889e-01 5.78554094e-01
-1.52995680e-02 2.41960496e-01 8.00103605e-01 -3.75073642e-01
1.01429772e+00 7.21178770e-01 7.13145435e-01 6.93493843e-01
-1.58397424e+00 -6.25928104e-01 9.40864068e-03 -7.65222684e-02
-1.13062668e+00 -7.04486310e-01 4.09387857e-01 -6.70818865e-01
8.34670663e-01 -1.12734221e-01 7.32924521e-01 5.54856896e-01
1.75482497e-01 5.84559143e-01 1.25071204e+00 -9.78031084e-02
5.75165570e-01 -1.34516463e-01 2.79935007e-03 9.26732719e-01
1.96050361e-01 -5.00865161e-01 -4.92201805e-01 -2.45570838e-01
8.14139187e-01 -9.97055992e-02 -2.97405034e-01 -1.69343814e-01
-1.40940392e+00 4.61286455e-01 7.07213581e-02 2.98851639e-01
-1.24079928e-01 2.58553296e-01 4.01558608e-01 -2.71933191e-02
3.18871975e-01 6.10591710e-01 -2.24757895e-01 -2.18540162e-01
-1.38725793e+00 3.86222541e-01 7.50484467e-01 7.12451994e-01
8.92407656e-01 -4.14771259e-01 2.47415274e-01 6.30811214e-01
-1.03516374e-02 1.47907630e-01 1.93428114e-01 -1.39183140e+00
4.86876555e-02 8.88832986e-01 5.44877946e-02 -7.01192439e-01
-5.37570477e-01 -1.60331458e-01 -3.50509495e-01 7.20444620e-01
8.86831045e-01 -1.09962717e-01 -6.86150968e-01 1.54053640e+00
5.98655701e-01 -1.27719074e-01 -4.36439574e-01 8.71397853e-01
3.36738110e-01 1.40678391e-01 -1.24310434e-01 2.10722148e-01
1.26080072e+00 -7.73930430e-01 -2.66385734e-01 -9.72395465e-02
6.00884438e-01 -6.15363479e-01 1.00660050e+00 4.58686501e-01
-1.14326155e+00 4.28807847e-02 -1.13582850e+00 -2.31429398e-01
-4.59155440e-01 -5.17446129e-03 6.27655923e-01 5.49222469e-01
-1.34314287e+00 8.01210046e-01 -9.75251436e-01 -5.61153114e-01
1.18281627e+00 7.04346657e-01 -3.70619267e-01 7.87827820e-02
-1.76761925e-01 6.72227919e-01 2.50705015e-02 -4.82237965e-01
-9.99126315e-01 -1.00882053e+00 -2.55191505e-01 -7.60215428e-03
3.20570916e-01 -8.07715774e-01 1.14371777e+00 -6.67317688e-01
-1.22538447e+00 1.18594539e+00 -2.51902521e-01 -5.42102933e-01
5.74128747e-01 2.33860895e-01 3.99795681e-01 6.86456800e-01
2.58294940e-01 1.14668608e+00 6.41767800e-01 -1.24453104e+00
-6.76127136e-01 -5.42943537e-01 7.63647631e-02 -3.90735358e-01
-2.22769335e-01 2.79018849e-01 -4.43285465e-01 -3.70747834e-01
-2.56708294e-01 -8.67764175e-01 -2.02497542e-01 5.56873679e-01
-1.55578941e-01 1.65775836e-01 7.59254575e-01 -6.71850920e-01
6.22456253e-01 -2.13132143e+00 3.78604494e-02 -1.52590886e-01
9.97520387e-01 4.18678761e-01 -3.15858126e-02 2.05562532e-01
5.66206098e-01 2.35449418e-01 -4.94285583e-01 -5.31966507e-01
-1.84999511e-01 1.31553188e-01 5.54071413e-03 6.93856478e-01
1.98964968e-01 8.80293608e-01 -8.70148361e-01 -8.27757061e-01
5.93567900e-02 3.28147352e-01 -5.67897439e-01 5.11312578e-03
-3.51615727e-01 5.07346988e-01 -2.30085552e-01 7.89233208e-01
6.65802717e-01 -5.03296494e-01 1.34623215e-01 2.24360615e-01
-3.11953157e-01 2.81999353e-02 -6.51403248e-01 1.76471853e+00
-2.30228558e-01 9.20897126e-01 8.75429869e-01 -7.21505940e-01
5.43579400e-01 -1.88465282e-01 4.45071220e-01 -3.92945051e-01
3.88662934e-01 1.56095415e-01 5.68810701e-02 -2.30624720e-01
2.68117756e-01 -1.63182676e-01 4.93207276e-01 8.40554059e-01
1.94411278e-01 -8.14037174e-02 3.98315549e-01 4.28794205e-01
1.60607779e+00 2.74915069e-01 1.14187837e-01 -5.45615077e-01
1.99274495e-02 3.80020976e-01 4.98916477e-01 5.65497220e-01
-7.12574184e-01 7.26077437e-01 6.44512832e-01 -3.23125273e-01
-1.37257266e+00 -7.78921604e-01 1.03571722e-02 8.95439029e-01
1.55959174e-01 -2.95453787e-01 -1.34195077e+00 -6.57711506e-01
-1.50025161e-02 1.01339996e-01 -6.24997437e-01 2.70564646e-01
-3.12700480e-01 -6.74308062e-01 1.02032220e+00 3.64076495e-01
4.74424541e-01 -9.19634521e-01 -1.32636690e+00 1.84202462e-01
-6.01259731e-02 -1.14282107e+00 -1.04885980e-01 6.42502233e-02
-8.90250921e-01 -1.13348663e+00 -9.56209660e-01 -5.85524917e-01
9.89984989e-01 3.34907413e-01 8.65295529e-01 4.68150973e-01
-7.77531683e-01 1.32079929e-01 -4.30300497e-02 -4.54411447e-01
-2.25146219e-01 -3.49271223e-02 -2.04729185e-01 -3.21579993e-01
1.65684953e-01 -7.58521974e-01 -6.37901664e-01 2.63398230e-01
-8.57295334e-01 7.31416792e-02 4.91330296e-01 6.96626782e-01
6.26006484e-01 -1.70118555e-01 4.22051311e-01 -1.06536138e+00
5.55607080e-01 -2.83009499e-01 -9.02045727e-01 1.49313003e-01
-6.10642731e-01 4.45128158e-02 8.18358302e-01 -3.65807503e-01
-8.25541973e-01 6.25011027e-02 8.96569155e-03 -2.71852553e-01
-3.01530719e-01 -1.09063178e-01 1.84604496e-01 -6.30466461e-01
4.77186650e-01 -4.55405489e-02 2.19729632e-01 -4.11971331e-01
2.58089192e-02 2.34327212e-01 5.75664401e-01 -4.41296488e-01
3.35992247e-01 1.09152102e+00 1.06414147e-01 -7.97748029e-01
-3.75523686e-01 -5.08935809e-01 -6.22387409e-01 -2.51942456e-01
6.90866292e-01 -4.45466697e-01 -1.09128177e+00 3.14298362e-01
-1.12399125e+00 -4.80739325e-01 9.37722847e-02 7.71039501e-02
-4.96124119e-01 3.88549358e-01 -7.52748549e-01 -7.59511054e-01
-3.81148577e-01 -1.06122422e+00 1.18679380e+00 1.95006728e-01
-4.68328059e-01 -7.55935013e-01 1.61172926e-01 6.57771111e-01
3.26837271e-01 1.97715893e-01 1.03059006e+00 -3.36475611e-01
-9.08698559e-01 4.85987738e-02 -5.13830245e-01 -5.65524213e-02
-3.39354873e-01 1.36899307e-01 -1.12976849e+00 -2.23130342e-02
-1.85016274e-01 -4.28944260e-01 7.71188736e-01 2.25784212e-01
1.00704741e+00 7.61479735e-02 -5.72861254e-01 5.92745781e-01
1.45656919e+00 -7.06157386e-02 5.62857747e-01 4.24474299e-01
5.20492971e-01 9.10182297e-01 3.19091648e-01 3.48974347e-01
2.58007079e-01 3.15074146e-01 5.40413737e-01 -2.87891090e-01
-2.21968114e-01 9.96634364e-03 -4.49549034e-02 4.02409196e-01
5.68194920e-03 -4.79186922e-02 -1.06784892e+00 9.52533066e-01
-1.70520914e+00 -7.68394291e-01 -1.21839158e-01 1.94337809e+00
7.42558658e-01 2.35360619e-02 3.23170394e-01 1.64961159e-01
7.04197705e-01 -2.88104743e-01 -7.43351400e-01 -2.51192540e-01
2.76268497e-02 1.34744212e-01 5.44463575e-01 3.25210750e-01
-6.25796139e-01 9.56067383e-01 6.86882639e+00 5.85242808e-01
-9.60697651e-01 2.03657344e-01 6.18553519e-01 -4.67531055e-01
-9.54565555e-02 2.00530112e-01 -7.34343290e-01 5.17521143e-01
7.41402090e-01 -7.45926192e-03 9.02698159e-01 5.65063179e-01
1.44097239e-01 -6.69891715e-01 -1.06537938e+00 9.70956922e-01
-9.51954424e-02 -1.57824063e+00 -2.85278022e-01 5.02451718e-01
5.34190357e-01 3.78016800e-01 -2.42709696e-01 -3.20037067e-01
4.13796663e-01 -8.51669371e-01 7.24119484e-01 3.32560331e-01
5.54720402e-01 -6.80223942e-01 6.29500687e-01 5.38271129e-01
-6.08817935e-01 7.92994350e-02 -5.59873283e-01 -2.11341400e-02
1.35171100e-01 8.33643556e-01 -1.05903721e+00 -1.58459261e-01
8.30564857e-01 2.14408144e-01 -7.13155746e-01 1.06466794e+00
1.96422338e-01 5.40499628e-01 -4.40520376e-01 -2.77133912e-01
-1.63725913e-02 -9.54675896e-04 4.99924719e-01 1.17315781e+00
-8.91604349e-02 1.70433596e-02 -1.80946648e-01 1.47284019e+00
-4.19741333e-01 -7.23688444e-03 -6.21944010e-01 -4.69249636e-01
3.15236181e-01 1.64991391e+00 -1.55097520e+00 -2.72848427e-01
-1.17451228e-01 9.52783167e-01 7.85015345e-01 -3.71316411e-02
-4.95735496e-01 -5.66385210e-01 5.56505144e-01 2.10740596e-01
4.71877366e-01 -2.84061164e-01 -7.39693522e-01 -8.66998971e-01
2.92953290e-02 -7.34272480e-01 -1.06752507e-01 -6.39351726e-01
-8.64810169e-01 2.04357415e-01 -4.63823974e-01 -4.82648134e-01
2.02707425e-01 -6.44706011e-01 -5.53877234e-01 5.86269200e-01
-1.22548974e+00 -8.47384393e-01 -3.55485082e-01 2.27308795e-01
2.67944753e-01 1.89948380e-01 4.95212525e-01 3.07219863e-01
-4.44608778e-01 3.66403013e-01 1.27353340e-01 -3.33582535e-02
4.56494778e-01 -1.17361760e+00 4.76475984e-01 6.84031785e-01
7.59783387e-02 9.31657434e-01 5.09001136e-01 -6.07473552e-01
-1.22707975e+00 -8.41629028e-01 6.87671185e-01 -6.42969012e-01
8.00822914e-01 -6.99431062e-01 -6.68244600e-01 3.09285492e-01
1.06952339e-01 -9.60069895e-02 7.01996922e-01 -8.77828002e-02
-1.86531216e-01 2.10517183e-01 -1.53379464e+00 7.46285737e-01
1.04388964e+00 -5.93106568e-01 -2.11585850e-01 7.27797523e-02
3.99177939e-01 1.93352774e-01 -7.68447101e-01 -1.51226193e-01
6.92999899e-01 -1.03075445e+00 6.04071677e-01 -2.00763941e-01
6.60717845e-01 -3.89330983e-01 1.78074017e-01 -1.07692420e+00
1.15455063e-02 -7.53692627e-01 2.67056346e-01 1.44265938e+00
3.92332137e-01 -5.41762710e-01 9.81149554e-01 5.85092127e-01
8.93820729e-03 -7.73785770e-01 -8.12500894e-01 -5.69229245e-01
4.65061665e-02 -6.74031302e-02 5.36376715e-01 7.05115557e-01
3.11307728e-01 1.87968642e-01 1.60755485e-01 -1.82237655e-01
7.63830721e-01 1.46763682e-01 8.43306124e-01 -1.11240220e+00
-3.29834282e-01 -6.02054596e-01 -6.49940252e-01 -8.01229358e-01
1.40183851e-01 -8.06221187e-01 2.69767851e-01 -1.43615580e+00
4.83465463e-01 -2.44656265e-01 1.63149267e-01 5.15845716e-01
5.58801442e-02 6.53183758e-01 1.41859174e-01 3.14517766e-01
-9.48171139e-01 5.70966629e-03 1.11279404e+00 -6.10140758e-03
2.57501781e-01 -5.70098162e-01 -7.62611032e-01 8.33155811e-01
8.52975011e-01 -6.48839772e-01 -1.69112429e-01 -6.93206668e-01
8.87829140e-02 -2.56923914e-01 6.41695142e-01 -1.03743517e+00
5.11189580e-01 1.89804107e-01 4.73376542e-01 -4.69580323e-01
2.06231385e-01 -4.94222224e-01 -1.99873015e-01 3.25231224e-01
-2.68548101e-01 4.14780006e-02 2.40251511e-01 5.70700109e-01
2.00417325e-01 -3.88649046e-01 9.92613673e-01 -4.62028623e-01
-2.50387698e-01 -6.26316145e-02 -6.37196958e-01 2.97128499e-01
1.02427828e+00 -2.72523552e-01 -5.44335783e-01 -6.58046603e-02
-2.40094453e-01 1.34417653e-01 1.23409700e+00 -4.10897225e-01
4.61070180e-01 -5.26221752e-01 -3.02226305e-01 -5.23605272e-02
-3.22421379e-02 -6.02602325e-02 -8.06409568e-02 9.51065421e-01
-8.99155021e-01 2.88980633e-01 -5.06738901e-01 -6.23724341e-01
-1.44048429e+00 5.64087570e-01 7.69471601e-02 -1.32951215e-02
-6.85294271e-01 9.70040083e-01 1.78356588e-01 -2.10055724e-01
-5.24132885e-03 -2.03055054e-01 2.70428951e-03 -3.10606241e-01
7.77223408e-01 7.86870658e-01 6.37607574e-02 -3.32482040e-01
-3.90364289e-01 3.75350744e-01 -2.00784862e-01 -2.90136188e-01
1.33188081e+00 -1.93085909e-01 -4.70566511e-01 2.40007102e-01
1.00998819e+00 1.18563481e-01 -1.58304477e+00 3.72217417e-01
1.65191904e-01 -4.27464247e-01 1.63969487e-01 -7.65247047e-01
-9.49205875e-01 1.11517036e+00 1.95008457e-01 1.45610407e-01
9.64083612e-01 1.10318929e-01 1.11093843e+00 3.68875176e-01
6.99808538e-01 -1.17764521e+00 -2.70352006e-01 1.82461739e-01
1.23995565e-01 -1.06588757e+00 8.21511354e-03 -2.69610345e-01
-2.17543513e-01 8.03874850e-01 3.82022768e-01 -1.92640647e-01
1.65222257e-01 8.33920300e-01 3.60948332e-02 -4.99400675e-01
-8.24978054e-01 -1.15109824e-01 -4.05556411e-01 8.01568747e-01
1.09105013e-01 -2.39584833e-01 -2.22751632e-01 3.41531813e-01
1.16185872e-02 3.97676796e-01 6.38661861e-01 1.16691566e+00
-5.69380403e-01 -9.01529968e-01 -3.74537170e-01 5.92127264e-01
-7.74741292e-01 -1.08339131e-01 -1.12031829e+00 5.10480225e-01
1.83827266e-01 9.95467901e-01 -4.14025970e-02 -1.66982636e-01
-1.45881042e-01 -8.77961665e-02 7.42703259e-01 -5.59199631e-01
-6.58037126e-01 -4.58077975e-02 -1.11537330e-01 -7.23947763e-01
-2.80955225e-01 -5.77393174e-01 -1.65784132e+00 -6.84520245e-01
-3.60587448e-01 -1.09071761e-01 9.82505918e-01 1.05783713e+00
7.34199464e-01 3.15045655e-01 -4.93543893e-02 -8.78340244e-01
-1.47162735e-01 -4.98895586e-01 -4.99561518e-01 2.01781407e-01
2.84814537e-01 -6.04607522e-01 -2.53065735e-01 4.38898146e-01] | [14.320436477661133, -3.141075611114502] |
8cb8aafa-1281-409d-9c44-1d82c4b75d50 | adaptive-network-combination-for-single-image | 2204.01505 | null | https://arxiv.org/abs/2204.01505v1 | https://arxiv.org/pdf/2204.01505v1.pdf | Adaptive Network Combination for Single-Image Reflection Removal: A Domain Generalization Perspective | Recently, multiple synthetic and real-world datasets have been built to facilitate the training of deep single image reflection removal (SIRR) models. Meanwhile, diverse testing sets are also provided with different types of reflection and scenes. However, the non-negligible domain gaps between training and testing sets make it difficult to learn deep models generalizing well to testing images. The diversity of reflections and scenes further makes it a mission impossible to learn a single model being effective to all testing sets and real-world reflections. In this paper, we tackle these issues by learning SIRR models from a domain generalization perspective. Particularly, for each source set, a specific SIRR model is trained to serve as a domain expert of relevant reflection types. For a given reflection-contaminated image, we present a reflection type-aware weighting (RTAW) module to predict expert-wise weights. RTAW can then be incorporated with adaptive network combination (AdaNEC) for handling different reflection types and scenes, i.e., generalizing to unknown domains. Two representative AdaNEC methods, i.e., output fusion (OF) and network interpolation (NI), are provided by considering both adaptation levels and efficiency. For images from one source set, we train RTAW to only predict expert-wise weights of other domain experts for improving generalization ability, while the weights of all experts are predicted and employed during testing. An in-domain expert (IDE) loss is presented for training RTAW. Extensive experiments show the appealing performance gain of our AdaNEC on different state-of-the-art SIRR networks. Source code and pre-trained models will available at https://github.com/csmliu/AdaNEC. | ['Lei Zhang', 'WangMeng Zuo', 'Zifei Yan', 'Jianan Pan', 'Ming Liu'] | 2022-04-04 | null | null | null | null | ['reflection-removal'] | ['computer-vision'] | [ 6.04261577e-01 -1.59627512e-01 3.25397760e-01 -5.55705607e-01
-9.21776235e-01 -1.67637676e-01 4.47950780e-01 -3.52124035e-01
-2.46497840e-01 6.23187542e-01 -1.18403807e-01 -1.46389022e-01
-8.99410099e-02 -9.10688877e-01 -6.65044248e-01 -9.04378831e-01
2.40710601e-01 2.90794730e-01 2.98348367e-01 -3.49848896e-01
-8.68905112e-02 6.20203376e-01 -1.68165827e+00 6.81712329e-01
1.35388303e+00 1.16790390e+00 5.15487731e-01 3.55944484e-01
1.11746885e-01 8.66234243e-01 -7.60568559e-01 -2.11603537e-01
4.85849947e-01 -3.26927304e-01 -1.01511203e-01 5.96842356e-03
4.89058495e-01 -3.32707435e-01 -4.02234167e-01 8.42370033e-01
7.32076228e-01 4.81644124e-01 7.41618812e-01 -9.84896719e-01
-8.09405327e-01 3.39021921e-01 -5.70537806e-01 -1.69521809e-01
8.08632746e-02 1.64857656e-01 5.46663165e-01 -1.06869268e+00
2.50158519e-01 1.06631124e+00 7.21786797e-01 5.50688982e-01
-9.28667307e-01 -9.38881576e-01 3.05045038e-01 2.19528243e-01
-1.33233285e+00 -4.19937849e-01 1.07041574e+00 -2.99675882e-01
6.73178434e-01 2.05098927e-01 3.30472648e-01 1.36333299e+00
-2.07016438e-01 7.54162073e-01 1.20280635e+00 -2.96896994e-01
2.20792279e-01 4.86149251e-01 1.59483608e-02 2.47093618e-01
8.27520713e-02 2.87854642e-01 -4.60469723e-01 1.45228341e-01
7.47950554e-01 -8.78568888e-02 -5.37811816e-01 1.10923662e-03
-8.13286483e-01 5.13530016e-01 6.21236861e-01 -2.47076135e-02
-4.97999996e-01 -3.21135014e-01 1.05925061e-01 5.04028261e-01
6.58620238e-01 2.00955003e-01 -4.54468310e-01 6.43959939e-01
-7.13796794e-01 6.48628697e-02 5.36257505e-01 8.57693076e-01
9.53524828e-01 3.71873885e-01 -1.81994528e-01 1.46288347e+00
2.15753436e-01 6.36294305e-01 1.59636885e-01 -9.19017076e-01
5.06296992e-01 5.80858231e-01 1.55312121e-01 -8.08011651e-01
-2.99947411e-01 -8.52154136e-01 -1.18123031e+00 4.24845755e-01
1.73302755e-01 -1.07184552e-01 -1.14702308e+00 1.54952788e+00
2.48051181e-01 4.15847033e-01 3.50314647e-01 1.07247770e+00
1.16914904e+00 7.93020546e-01 -1.43718839e-01 3.61309685e-02
9.91309762e-01 -9.40713167e-01 -1.93566099e-01 -5.46468973e-01
1.70541853e-01 -8.23506892e-01 1.02138078e+00 7.57961929e-01
-8.99350762e-01 -9.97537315e-01 -1.00977015e+00 1.43172890e-01
-2.23447591e-01 4.46034998e-01 2.80871212e-01 4.78670239e-01
-7.65890956e-01 3.67700487e-01 -4.77562755e-01 -3.84134799e-03
4.12176400e-01 3.22252512e-02 -8.84371921e-02 -6.32230639e-01
-1.42757535e+00 9.61441636e-01 1.30303249e-01 3.72917116e-01
-1.15737104e+00 -1.00886667e+00 -7.45764136e-01 -2.66743213e-01
3.74025106e-01 -4.85777497e-01 9.79643464e-01 -1.28851283e+00
-1.27379715e+00 6.52274430e-01 1.33464247e-01 -2.12002277e-01
7.28800476e-01 -2.40286022e-01 -8.97110283e-01 6.17220774e-02
-1.08912267e-01 2.86503285e-01 1.17495501e+00 -1.83083463e+00
-5.22463679e-01 -1.07346155e-01 2.25127995e-01 3.62080157e-01
-2.87633985e-01 -7.69257993e-02 -2.67880738e-01 -6.56088710e-01
1.51336705e-02 -5.64133525e-01 -9.07471776e-02 2.76272655e-01
-2.15344742e-01 1.11933641e-01 6.72140360e-01 -8.27042520e-01
9.34013546e-01 -2.31358361e+00 -1.38830885e-01 3.19496185e-01
3.69487070e-02 5.53771615e-01 -6.06106877e-01 1.01778641e-01
-2.11560145e-01 -3.57222497e-01 -5.13790727e-01 -1.45624429e-01
-2.70590127e-01 1.10764273e-01 -3.04748148e-01 3.31210107e-01
4.40982163e-01 3.76339257e-01 -6.66475534e-01 -7.77309015e-02
4.40563291e-01 9.59324002e-01 -4.35570747e-01 3.86929959e-01
-2.05613151e-01 7.07236409e-01 -4.02119517e-01 6.95423067e-01
1.07597756e+00 -7.21987337e-02 -1.03029720e-01 -5.81735909e-01
-5.49267465e-03 9.83994976e-02 -1.49578929e+00 1.22425520e+00
-1.06374395e+00 4.53458816e-01 -2.78850622e-03 -1.03472507e+00
1.47043228e+00 2.52341539e-01 2.68931478e-01 -1.03118074e+00
-4.47954722e-02 3.70847553e-01 -2.39156187e-02 -3.64492357e-01
2.48714179e-01 -8.39851890e-03 2.38181800e-01 2.53410727e-01
-4.74598855e-02 5.49771413e-02 -1.59884006e-01 -2.69473344e-01
9.23438430e-01 2.66403615e-01 -1.89757366e-02 9.50842351e-02
7.25051939e-01 -1.14003077e-01 7.41560280e-01 8.22853804e-01
1.50781631e-01 1.01769030e+00 -3.54725093e-01 -5.56849122e-01
-9.74806845e-01 -1.34428287e+00 -3.32469493e-01 1.04924417e+00
2.46823892e-01 3.43781084e-01 -3.02647889e-01 -3.50781590e-01
-1.80528134e-01 9.14792001e-01 -3.66214573e-01 -2.63074279e-01
-7.05500901e-01 -8.77715468e-01 3.19769681e-01 5.12207091e-01
9.57497001e-01 -1.17922425e+00 -5.31309187e-01 1.92955956e-01
-3.31251949e-01 -1.14720428e+00 7.81945884e-02 1.41226828e-01
-6.24838591e-01 -9.71223235e-01 -8.48868608e-01 -5.55142522e-01
7.53470182e-01 5.18768072e-01 1.26262975e+00 1.10048607e-01
-3.30995411e-01 2.46754408e-01 -4.86487925e-01 -4.86108184e-01
-5.65782428e-01 -3.89188737e-01 -1.83382839e-01 2.09509417e-01
1.82557911e-01 -6.55577719e-01 -8.87421072e-01 7.05283284e-01
-8.86100471e-01 8.38098526e-02 7.68560946e-01 8.03475440e-01
6.00106418e-01 2.75461763e-01 6.34690166e-01 -9.30272520e-01
3.64110291e-01 -5.00799060e-01 -5.15055478e-01 4.88041073e-01
-2.99778044e-01 -3.04026812e-01 9.13661122e-01 -7.01084256e-01
-1.71763742e+00 -1.89739078e-01 -2.74340659e-01 -3.84081572e-01
-3.76164496e-01 3.47023964e-01 -3.17055762e-01 -4.98569980e-02
1.11271477e+00 2.58759558e-01 -3.13046604e-01 -4.34883326e-01
1.16688557e-01 8.20192218e-01 5.70175707e-01 -7.85424590e-01
9.66010392e-01 4.20754731e-01 -3.15435022e-01 -9.32670474e-01
-1.09957206e+00 -4.47585136e-01 -2.30887026e-01 -4.69009399e-01
4.52025920e-01 -1.40284598e+00 -5.88738099e-02 7.75287211e-01
-9.60541725e-01 -7.01498508e-01 -2.74576157e-01 3.67277741e-01
-3.62384647e-01 8.61018300e-02 -2.82900274e-01 -9.73939598e-01
-2.85477996e-01 -9.97630894e-01 8.71886134e-01 2.94320673e-01
2.33982354e-01 -8.60208571e-01 -1.14089549e-01 4.18910712e-01
5.62397182e-01 1.80851638e-01 6.84233367e-01 -4.03791547e-01
-5.50481260e-01 -1.98772997e-01 -5.59752703e-01 9.23253298e-01
3.24057527e-02 -1.79818138e-01 -1.45617938e+00 -2.80186951e-01
1.91939756e-01 -4.03521270e-01 1.06251764e+00 4.06780809e-01
1.37431383e+00 -2.30863672e-02 2.26067845e-02 7.92355120e-01
1.59493995e+00 2.53742635e-01 1.01761448e+00 4.59543109e-01
6.01468146e-01 6.65179789e-01 8.13672006e-01 4.46721226e-01
1.63975716e-01 5.07900715e-01 4.07392949e-01 -3.95079613e-01
-3.48131150e-01 2.80665066e-02 3.94728661e-01 4.24866080e-01
-5.18956304e-01 -5.27637362e-01 -6.81851625e-01 3.86663616e-01
-1.65866745e+00 -8.39800656e-01 -9.04566124e-02 2.34597206e+00
6.56402409e-01 6.63431734e-02 -1.65872753e-01 1.52331680e-01
6.58792436e-01 1.79845318e-01 -8.74956906e-01 5.77255199e-03
-3.69067222e-01 3.56251389e-01 2.99744576e-01 2.47952998e-01
-8.94188404e-01 5.27921319e-01 4.55193710e+00 8.63679290e-01
-1.12260818e+00 1.21214256e-01 6.00569487e-01 1.42515659e-01
-5.19220889e-01 -2.24506631e-01 -7.02440679e-01 2.06904918e-01
6.08378232e-01 3.32795024e-01 6.17125630e-01 6.64327025e-01
2.48071417e-01 5.88094965e-02 -8.08722794e-01 8.41592312e-01
1.16245918e-01 -9.87396479e-01 1.02843745e-02 -4.33646709e-01
9.03256476e-01 2.08060756e-01 1.52743310e-01 4.89411891e-01
4.02563691e-01 -8.27838838e-01 4.96028662e-01 6.48392320e-01
9.34023798e-01 -6.56705618e-01 6.84196830e-01 3.47624451e-01
-1.05512071e+00 -3.78295630e-01 -6.64925098e-01 2.55234152e-01
2.68205311e-02 7.95978487e-01 -3.94464403e-01 9.48368549e-01
8.43312562e-01 7.11526394e-01 -3.04582268e-01 1.00154114e+00
-4.44421381e-01 4.64463562e-01 -3.01507741e-01 5.39104581e-01
-6.47163689e-02 -2.92856872e-01 4.87575740e-01 1.00997877e+00
5.70968390e-01 1.05260141e-01 1.05109222e-01 8.87302816e-01
7.96791613e-02 -2.25331500e-01 -3.34574282e-01 6.14798069e-01
6.00122213e-01 1.28287268e+00 -1.75812453e-01 -8.28838125e-02
-6.10610485e-01 8.15029383e-01 1.59670085e-01 8.48269165e-01
-9.36316490e-01 -2.33165056e-01 7.11112618e-01 3.00143182e-01
1.31045669e-01 6.13127947e-02 -2.72734523e-01 -1.13742924e+00
1.46017492e-01 -1.00351560e+00 4.36923593e-01 -1.07388997e+00
-1.65184152e+00 8.77761364e-01 -4.34859842e-03 -1.70352221e+00
2.54981697e-01 -7.20142841e-01 -7.89605200e-01 1.05978453e+00
-2.13378406e+00 -1.15789473e+00 -9.47411895e-01 7.64751852e-01
5.81938088e-01 -3.80882144e-01 5.04367590e-01 5.53710580e-01
-4.78044838e-01 6.78791404e-01 1.46996096e-01 7.33308792e-02
8.57020319e-01 -7.95424223e-01 8.23387206e-02 9.08098042e-01
-7.72151649e-02 3.16355586e-01 6.29971147e-01 -2.89479226e-01
-1.02831292e+00 -1.59626842e+00 3.15409720e-01 -9.45200864e-03
3.92471224e-01 -2.70955354e-01 -1.29701328e+00 5.31864941e-01
-2.12479755e-01 6.41087219e-02 5.40406168e-01 -3.59774493e-02
-5.99321902e-01 -5.65465331e-01 -1.20503306e+00 6.61269724e-01
1.15312266e+00 -3.81696850e-01 -2.70818263e-01 3.18828613e-01
5.02946198e-01 -3.99192959e-01 -8.69479060e-01 8.12183738e-01
4.12466079e-01 -1.23285961e+00 1.36768520e+00 5.81179149e-02
6.71126306e-01 -4.80925709e-01 -3.21800679e-01 -1.42710209e+00
-2.57956654e-01 -2.69548241e-02 8.24865848e-02 1.06709027e+00
4.90414351e-01 -9.08650398e-01 4.09021735e-01 5.33512056e-01
-4.98455048e-01 -5.11413217e-01 -6.91964328e-01 -8.02047849e-01
-6.15246072e-02 -6.58903003e-01 4.74433750e-01 8.88985872e-01
-9.10206378e-01 6.78466856e-02 -4.83333230e-01 5.84893465e-01
8.49782467e-01 2.23363295e-01 7.76856244e-01 -1.18043590e+00
-5.10662436e-01 -1.69056386e-01 -4.12654914e-02 -9.99938726e-01
7.37395734e-02 -8.41238737e-01 1.98311210e-01 -1.67104685e+00
-3.29319499e-02 -8.34866405e-01 -4.92697150e-01 5.00343442e-01
-1.44011855e-01 4.00745392e-01 4.32984903e-02 1.60400748e-01
-2.50475228e-01 6.95522666e-01 1.37835002e+00 -2.34090343e-01
-2.55884528e-01 3.27296674e-01 -6.93992078e-01 7.34556079e-01
9.75044727e-01 -3.32284808e-01 -6.47696257e-01 -7.27455974e-01
2.00756937e-01 -3.30844591e-03 7.84638584e-01 -1.25296474e+00
9.18000862e-02 -1.14177257e-01 6.20194972e-01 -5.23692012e-01
5.35392940e-01 -9.41689730e-01 3.58783066e-01 1.17980698e-02
-2.45539680e-01 -6.78386748e-01 2.19403550e-01 3.61048639e-01
-2.54678994e-01 -1.63225546e-01 1.07062256e+00 -1.10270388e-01
-8.82661402e-01 3.45506728e-01 -6.53327629e-03 2.81043071e-02
6.75139725e-01 -4.99149501e-01 -5.20712137e-01 -2.96789855e-01
-5.74146688e-01 2.08493605e-01 2.63627261e-01 3.25376719e-01
9.31659341e-01 -1.05053914e+00 -1.13303638e+00 3.27606797e-01
3.97065878e-01 4.64332610e-01 8.53978217e-01 5.62186778e-01
-7.00387135e-02 -4.20542121e-01 -1.69484302e-01 -5.00664771e-01
-1.06468487e+00 3.33405912e-01 5.38031697e-01 -9.71145406e-02
-7.10045576e-01 7.00063825e-01 6.79751337e-01 -8.58046710e-01
5.51920235e-02 -6.28333837e-02 -2.09780335e-01 -1.50291249e-01
6.92223787e-01 3.99212211e-01 1.91190556e-01 -3.55596930e-01
-1.39113590e-01 6.62476301e-01 -1.95877608e-02 3.69069725e-01
1.55814314e+00 1.26763210e-01 1.28179342e-01 2.59521753e-01
9.29959059e-01 -2.77515322e-01 -1.61377060e+00 -6.52995110e-01
-5.60576916e-01 -4.61982101e-01 4.50551249e-02 -1.20297146e+00
-1.35131466e+00 9.37567711e-01 6.81131005e-01 -1.84913591e-01
1.75288677e+00 -2.10260436e-01 5.52431881e-01 4.09428477e-01
3.39716285e-01 -1.01711237e+00 1.24472953e-01 4.16736096e-01
1.16413379e+00 -1.41529894e+00 -5.74541502e-02 -4.29787308e-01
-8.30990672e-01 1.05810273e+00 8.31549764e-01 -2.44021282e-01
5.62278152e-01 3.75238806e-01 3.06792378e-01 9.89149362e-02
-5.15539229e-01 -1.17649175e-01 4.17731702e-01 1.02711582e+00
2.68697351e-01 -1.45592978e-02 2.50221968e-01 5.23247063e-01
8.57949629e-02 4.16827686e-02 3.13503832e-01 5.73782861e-01
-4.92912143e-01 -8.48568618e-01 -4.58431721e-01 5.14016390e-01
5.67800440e-02 -1.00202776e-01 1.62643231e-02 7.02472985e-01
1.48948237e-01 1.00463724e+00 -1.19320191e-01 -4.20359403e-01
5.42211115e-01 -4.64252383e-01 3.80011916e-01 -5.20252407e-01
-4.74803835e-01 -3.33113619e-03 1.07106335e-01 -3.04860771e-01
-5.38659990e-01 -3.85517180e-01 -1.00791764e+00 -1.15977034e-01
-2.69462854e-01 -2.29941800e-01 4.81889546e-01 6.30312622e-01
1.15753017e-01 8.70512307e-01 9.02627110e-01 -8.82038891e-01
-6.33144498e-01 -8.70719612e-01 -5.27755260e-01 3.85254294e-01
2.83831298e-01 -6.88692212e-01 -5.52218735e-01 -1.21686421e-01] | [10.921873092651367, -2.681542158126831] |
f7bb2320-5f89-4c09-9122-7233e9142c86 | set-generation-networks-for-end-to-end | null | null | https://aclanthology.org/2021.emnlp-main.760 | https://aclanthology.org/2021.emnlp-main.760.pdf | Set Generation Networks for End-to-End Knowledge Base Population | The task of knowledge base population (KBP) aims to discover facts about entities from texts and expand a knowledge base with these facts. Previous studies shape end-to-end KBP as a machine translation task, which is required to convert unordered fact into a sequence according to a pre-specified order. However, the facts stated in a sentence are unordered in essence. In this paper, we formulate end-to-end KBP as a direct set generation problem, avoiding considering the order of multiple facts. To solve the set generation problem, we propose networks featured by transformers with non-autoregressive parallel decoding. Unlike previous approaches that use an autoregressive decoder to generate facts one by one, the proposed networks can directly output the final set of facts in one shot. Furthermore, to train the networks, we also design a set-based loss that forces unique predictions via bipartite matching. Compared with cross-entropy loss that highly penalizes small shifts in fact order, the proposed bipartite matching loss is invariant to any permutation of predictions. Benefiting from getting rid of the burden of predicting the order of multiple facts, our proposed networks achieve state-of-the-art (SoTA) performance on two benchmark datasets. | ['Wei Bi', 'Jun Zhao', 'Kang Liu', 'Yubo Chen', 'Chenhao Wang', 'Dianbo Sui'] | null | null | null | null | emnlp-2021-11 | ['knowledge-base-population'] | ['natural-language-processing'] | [ 3.57535332e-01 7.41927803e-01 -8.83132890e-02 -4.77144748e-01
-7.83715427e-01 -5.98857999e-01 4.53220606e-01 1.59136146e-01
-3.51040393e-01 1.28146863e+00 4.57050025e-01 -3.42345625e-01
-8.84660855e-02 -1.32049143e+00 -1.35232818e+00 -3.45956743e-01
2.73333758e-01 8.32633138e-01 4.25681099e-02 -5.49661934e-01
-1.32466238e-02 -4.67325188e-02 -1.25595212e+00 6.33034289e-01
1.12938595e+00 7.31074214e-01 8.05031583e-02 3.93188059e-01
-3.58335853e-01 9.20099437e-01 -3.36875111e-01 -1.44838154e+00
2.74287999e-01 -6.75554335e-01 -1.03842628e+00 -3.45183223e-01
4.49644208e-01 -1.15009122e-01 -4.84986156e-01 1.14115465e+00
5.17055035e-01 1.45087704e-01 7.83302784e-01 -1.01049829e+00
-1.08822834e+00 1.25296640e+00 -1.94748670e-01 2.40279198e-01
5.57784677e-01 -1.82447627e-01 1.46953130e+00 -9.03693795e-01
6.93198323e-01 9.82082129e-01 5.66506922e-01 4.66555923e-01
-1.00285327e+00 -5.94424069e-01 1.44711211e-01 5.54990768e-01
-1.28273213e+00 -4.37792927e-01 7.82468319e-01 -2.08205566e-01
1.21152759e+00 3.09049100e-01 5.91632485e-01 9.04154301e-01
2.14566112e-01 6.74154580e-01 5.94120085e-01 -4.32342112e-01
9.75575447e-02 1.45140037e-01 4.78308164e-02 7.26752102e-01
4.77636248e-01 -2.07069218e-01 -7.31941223e-01 1.71436125e-03
3.26446593e-01 -2.91528940e-01 -3.37630659e-01 -1.11187652e-01
-1.24995899e+00 6.97511971e-01 4.58822280e-01 1.88845862e-02
-4.99929547e-01 -6.29535392e-02 3.46897781e-01 4.98636186e-01
3.94746959e-01 5.05869687e-01 -7.57410824e-01 6.35053590e-02
-8.22641194e-01 3.00550491e-01 1.09664881e+00 1.14121377e+00
7.25691140e-01 -2.68980622e-01 -5.24924219e-01 5.56370854e-01
9.27346274e-02 5.00126719e-01 3.43138725e-01 -3.32705796e-01
1.02330494e+00 6.19236052e-01 -2.97552757e-02 -1.24208045e+00
-3.42273712e-01 -6.75874412e-01 -1.07272303e+00 -6.59958899e-01
1.80611923e-01 -2.76133388e-01 -7.19642878e-01 1.93842542e+00
2.84276128e-01 3.14162493e-01 4.61284399e-01 7.67683327e-01
9.87162173e-01 8.06206882e-01 -1.25556365e-01 -2.61609435e-01
1.31737757e+00 -9.27320719e-01 -5.92048764e-01 -1.77845120e-01
5.48553348e-01 -3.77106279e-01 8.34082067e-01 4.47385013e-02
-1.23151124e+00 -1.96129933e-01 -8.81606460e-01 -4.84648436e-01
-4.45192397e-01 -4.80797589e-02 4.34526145e-01 4.86352354e-01
-1.00633562e+00 4.57561672e-01 -2.71822214e-01 -6.25711828e-02
3.18169385e-01 3.62117201e-01 -3.55447173e-01 -8.70441645e-03
-1.77130795e+00 1.07749832e+00 9.62196648e-01 2.31376797e-01
-3.41664761e-01 -9.56036448e-01 -8.57370794e-01 5.95879197e-01
7.27531552e-01 -1.44302273e+00 9.69794154e-01 -6.76928639e-01
-1.62366664e+00 6.01079762e-01 -1.82947472e-01 -7.47679472e-01
4.32187915e-01 -6.98411688e-02 -4.09310758e-01 -7.27265403e-02
1.77852973e-01 5.38426459e-01 4.41002011e-01 -9.68567133e-01
-6.88945532e-01 -9.23639685e-02 3.77923459e-01 4.51076627e-01
-3.19817871e-01 -3.72520000e-01 -3.54826033e-01 -4.68730003e-01
-6.94180951e-02 -6.62664592e-01 -6.94593927e-03 -5.49901068e-01
-9.21258688e-01 4.98201549e-02 1.69966310e-01 -8.80461931e-01
1.12075269e+00 -1.67122662e+00 3.86215836e-01 2.96655118e-01
1.82447553e-01 1.29837498e-01 -7.13155940e-02 5.62537313e-01
5.99506721e-02 4.70416546e-02 -3.88670892e-01 -4.82659042e-02
2.35720918e-01 8.53087530e-02 -6.32671237e-01 -5.48898950e-02
3.17858130e-01 1.13630533e+00 -1.12036562e+00 -5.33567250e-01
-2.13600487e-01 1.85170859e-01 -9.41589177e-01 -8.04218203e-02
-4.48737085e-01 1.40917391e-01 -2.52488792e-01 1.75224036e-01
6.00593746e-01 -3.72778177e-01 3.72587949e-01 -3.74098331e-01
4.69053388e-01 6.87290788e-01 -8.21703613e-01 1.56358194e+00
-5.61267495e-01 1.58789188e-01 -7.29174137e-01 -9.94930804e-01
9.31232333e-01 3.74145150e-01 9.67407227e-02 -5.10765135e-01
6.72313478e-03 1.42095819e-01 4.27919887e-02 -3.27320635e-01
7.76303232e-01 -4.87885863e-01 -2.13456273e-01 1.39364824e-01
3.87951821e-01 1.53447059e-03 3.43883932e-01 3.54638696e-01
1.07554460e+00 -6.63788617e-02 4.99796540e-01 4.73955721e-02
5.48563421e-01 1.11445300e-02 6.77187324e-01 7.50119150e-01
4.76570874e-01 2.59842664e-01 7.34480202e-01 -3.57997924e-01
-1.12433207e+00 -1.08302224e+00 2.37149641e-01 7.22545028e-01
9.00962427e-02 -2.37549439e-01 -3.69288236e-01 -1.03397214e+00
-3.81297688e-03 1.18311727e+00 -4.68231142e-01 -3.48337114e-01
-3.78882796e-01 -5.68047702e-01 6.44703686e-01 3.31882358e-01
6.44491374e-01 -1.04398978e+00 -1.39640406e-01 4.44887251e-01
-7.77146995e-01 -1.14902198e+00 -7.66949117e-01 -2.18390767e-02
-4.57162261e-01 -9.73298728e-01 -5.34791470e-01 -8.99959803e-01
7.90969491e-01 -1.35493934e-01 1.29793227e+00 -2.05702513e-01
2.61586130e-01 -8.66053104e-02 -4.21336442e-01 -3.60313475e-01
-2.89445013e-01 4.02580976e-01 -5.59291840e-02 1.09701492e-01
3.70347381e-01 -7.24257529e-01 -4.13819969e-01 -6.33707494e-02
-8.12223375e-01 5.27199924e-01 8.03200781e-01 1.05162442e+00
5.95860183e-01 2.62518227e-01 8.58587921e-01 -1.15753055e+00
6.87328458e-01 -7.21884549e-01 -3.10975581e-01 6.48158014e-01
-3.41107041e-01 3.54615599e-01 1.11098182e+00 -5.10734431e-02
-1.43599081e+00 -8.10951293e-02 -1.38445869e-01 -1.63012475e-01
2.75633723e-01 9.77575958e-01 -2.36198962e-01 2.91563243e-01
5.31858444e-01 6.60718977e-01 -4.27510530e-01 -1.82467501e-03
7.68582225e-01 3.75142008e-01 5.46188354e-01 -5.70244968e-01
6.15031898e-01 1.63906083e-01 6.40235990e-02 -1.84218302e-01
-1.38104403e+00 -1.60526305e-01 -5.62770605e-01 -8.29939097e-02
5.68371058e-01 -8.79424632e-01 -8.73428881e-01 7.62223601e-02
-1.41611969e+00 1.43461362e-01 -3.58539283e-01 3.62484872e-01
-6.49535716e-01 2.47462392e-01 -4.29942161e-01 -6.03560030e-01
-7.38221049e-01 -5.49228072e-01 7.04034984e-01 5.85967675e-02
-2.14358456e-02 -8.45874667e-01 5.28845079e-02 3.59952986e-01
1.16449170e-01 1.09280191e-01 1.19377530e+00 -8.84982586e-01
-5.73600411e-01 -1.88857988e-01 -2.00736493e-01 1.95002839e-01
-2.38281991e-02 -3.90798390e-01 -5.72173953e-01 7.36999959e-02
-2.59135216e-01 -2.89406836e-01 9.90040898e-01 9.08402950e-02
8.04687738e-01 -9.06350791e-01 -2.55869955e-01 3.69488329e-01
1.46059430e+00 1.48205742e-01 7.85043955e-01 1.49982452e-01
7.39962101e-01 6.40791476e-01 3.11665505e-01 3.08093518e-01
8.96812558e-01 4.94719446e-01 1.57743067e-01 3.42747927e-01
3.94164212e-03 -7.96698451e-01 2.87250668e-01 1.00518858e+00
7.54326060e-02 -6.13251269e-01 -8.55098486e-01 7.58722842e-01
-1.93748820e+00 -1.36646438e+00 2.92188190e-02 2.04849505e+00
1.29700804e+00 2.94426888e-01 -1.54899716e-01 -1.54281706e-01
8.17847788e-01 -7.46128336e-03 -4.73867327e-01 -2.87107766e-01
-1.72799602e-01 9.32748616e-02 6.03643477e-01 5.83482802e-01
-7.10680664e-01 1.12778425e+00 5.32564735e+00 9.49241042e-01
-8.66083562e-01 6.34411126e-02 5.70697784e-01 -2.75706500e-01
-8.23203683e-01 9.93483216e-02 -8.76751244e-01 6.76414073e-01
6.76672637e-01 -5.45157671e-01 5.25099397e-01 5.32157660e-01
-1.32017285e-01 2.27396399e-01 -1.10034966e+00 5.98607838e-01
1.77416101e-01 -1.55037689e+00 5.79540491e-01 -3.44268918e-01
8.69712234e-01 -2.88868189e-01 -1.43359274e-01 6.93837702e-01
5.49054921e-01 -7.93889940e-01 8.48777592e-01 8.84609759e-01
7.51105428e-01 -9.54231501e-01 8.97486389e-01 4.25186574e-01
-1.00444281e+00 -7.52048418e-02 -5.59926271e-01 -2.24032894e-01
5.66096544e-01 1.07997060e+00 -1.10599291e+00 1.10565245e+00
2.90806681e-01 4.86915529e-01 -4.26793806e-02 9.23245609e-01
-4.77815062e-01 5.25726855e-01 -2.66643554e-01 -3.07098210e-01
5.10890260e-02 -1.06856667e-01 5.62553227e-01 1.15097868e+00
4.89092797e-01 2.31330901e-01 -7.40807205e-02 9.00190651e-01
-7.36152172e-01 1.57010853e-01 -6.52845144e-01 8.72814953e-02
6.12158477e-01 8.74155223e-01 -4.65960890e-01 -5.00620306e-01
-2.52280682e-01 1.05894375e+00 8.47848892e-01 1.57527089e-01
-8.81542683e-01 -6.10221148e-01 1.94767863e-01 -2.94098288e-01
5.50065875e-01 3.34645808e-01 -4.45696950e-01 -1.42237043e+00
3.98050815e-01 -5.94121158e-01 2.93123424e-01 -7.25890040e-01
-1.46896803e+00 6.73870921e-01 -1.20901719e-01 -9.65991259e-01
-1.92997262e-01 -1.66761979e-01 -5.23977816e-01 8.01432073e-01
-1.69662011e+00 -1.12593746e+00 1.69240549e-01 4.47545201e-01
1.84069291e-01 -2.20314916e-02 6.21391714e-01 4.76982355e-01
-4.00512099e-01 7.78815627e-01 2.74097528e-02 2.46835127e-01
5.55084348e-01 -1.31612790e+00 5.20400703e-01 9.67674136e-01
9.59991440e-02 6.22783542e-01 7.78049588e-01 -9.57342923e-01
-1.05129611e+00 -1.29841304e+00 1.53563011e+00 -3.28240782e-01
5.46441436e-01 -3.54824632e-01 -7.11788476e-01 9.81110334e-01
2.21104339e-01 -2.27836862e-01 5.64851224e-01 3.65747213e-01
-3.56225997e-01 -1.78637460e-01 -1.12839437e+00 7.22488046e-01
1.36341023e+00 -3.64603907e-01 -9.53267157e-01 3.70076865e-01
1.03718531e+00 -7.05780029e-01 -7.35515475e-01 2.15749383e-01
4.93498176e-01 -6.17439866e-01 7.55760193e-01 -9.48107123e-01
1.14245820e+00 -2.39726469e-01 4.55724448e-02 -1.67725444e+00
-5.41517258e-01 -6.25318468e-01 -1.92084283e-01 1.25970554e+00
1.10865653e+00 -6.90046787e-01 8.58901739e-01 4.54052657e-01
-2.51478523e-01 -8.30359340e-01 -1.07139587e+00 -7.72494137e-01
3.19991522e-02 -1.40607253e-01 9.99363780e-01 1.03165007e+00
4.66116458e-01 7.58877456e-01 -7.13760614e-01 2.95022100e-01
3.88017148e-01 2.22255215e-01 4.61563796e-01 -8.69703710e-01
-4.83090401e-01 -2.21189618e-01 -2.49890506e-01 -1.08972287e+00
2.95898318e-01 -1.33287835e+00 2.17040315e-01 -1.76251054e+00
4.06061441e-01 -3.26083571e-01 -1.63183123e-01 6.55509472e-01
-5.66110015e-01 -1.54693484e-01 2.75489658e-01 -1.23943619e-01
-5.82249761e-01 9.38624799e-01 1.25196123e+00 -2.37652376e-01
-2.98462026e-02 -1.58005744e-01 -9.88195121e-01 3.82383049e-01
7.52113581e-01 -6.91416919e-01 -6.06647789e-01 -5.44440091e-01
1.00650418e+00 4.83230382e-01 2.77814537e-01 -6.90885544e-01
5.78642190e-01 -6.58158511e-02 6.12049438e-02 -5.62228441e-01
3.41749310e-01 -8.87088895e-01 2.37351999e-01 1.85702428e-01
-5.78301251e-01 -1.28240630e-01 5.21276072e-02 7.23351896e-01
-2.61566460e-01 -2.67943174e-01 2.39762962e-01 -1.29082695e-01
-4.53742027e-01 2.05345958e-01 8.03717133e-03 3.65389168e-01
9.87817109e-01 3.11458278e-02 -5.82531571e-01 -4.55701649e-01
-6.17322505e-01 2.72738904e-01 -1.30101040e-01 1.82964757e-01
5.80141842e-01 -1.30580950e+00 -9.72683966e-01 -8.87427479e-02
5.84345348e-02 2.34512210e-01 5.64150393e-01 7.45670974e-01
-3.00716549e-01 6.13967538e-01 -2.05746740e-01 1.18778154e-01
-8.21052253e-01 6.24301195e-01 2.32405022e-01 -8.12184572e-01
-4.35517877e-01 1.04125881e+00 -4.66459803e-03 -7.59268403e-01
-2.56808370e-01 -2.54857332e-01 -1.69232950e-01 -1.49428071e-02
3.20249796e-01 1.33709297e-01 2.24196211e-01 -2.78842956e-01
-2.02390850e-01 5.13466746e-02 -4.50635910e-01 7.03840256e-02
1.33240497e+00 5.64016812e-02 -1.66430935e-01 8.94766003e-02
9.73139584e-01 1.18646272e-01 -7.52920687e-01 -4.76140290e-01
-1.49687514e-01 -1.98751733e-01 -2.06574351e-01 -1.10023284e+00
-9.51459825e-01 3.46960068e-01 -3.13213915e-01 5.94678111e-02
1.08832598e+00 -3.92245091e-02 1.08234775e+00 7.05264270e-01
4.63556528e-01 -8.06863129e-01 -2.50388533e-01 9.36547935e-01
8.86217654e-01 -9.48630095e-01 -2.59768337e-01 -6.97336733e-01
-8.94280553e-01 9.08380091e-01 5.40987194e-01 3.73577438e-02
4.18981642e-01 3.95155810e-02 -4.80158269e-01 -1.05237395e-01
-9.00515139e-01 -2.08608404e-01 3.84494990e-01 3.78209084e-01
1.32475510e-01 9.13322568e-02 -5.58200836e-01 9.11447704e-01
-8.19286585e-01 3.63776803e-01 5.93672156e-01 5.01824975e-01
-3.98785353e-01 -8.08285296e-01 1.00569753e-02 9.53575611e-01
-4.00112480e-01 -7.17308760e-01 -2.29966074e-01 3.25367868e-01
9.10938904e-02 8.39373767e-01 -1.09538801e-01 -5.48927546e-01
4.62628186e-01 2.36889645e-01 3.41058373e-01 -6.66788936e-01
-4.95859772e-01 -8.58155549e-01 4.79824185e-01 4.26386856e-02
-2.10192934e-01 -4.46110934e-01 -1.12684143e+00 -5.78443348e-01
-4.52352524e-01 2.62820125e-01 2.25782722e-01 1.21202552e+00
6.89794123e-01 7.28564858e-01 3.66223127e-01 -1.04577973e-01
-5.38429022e-01 -8.82821202e-01 -5.35666823e-01 4.44746614e-01
6.43015653e-02 -5.21193087e-01 -5.62796406e-02 1.68017507e-01] | [9.814414978027344, 8.489052772521973] |
d93833ec-75d0-42c1-a1f6-746474b19f38 | blind-source-extraction-based-on-multi | 2005.07976 | null | https://arxiv.org/abs/2005.07976v2 | https://arxiv.org/pdf/2005.07976v2.pdf | Target Speech Extraction Based on Blind Source Separation and X-vector-based Speaker Selection Trained with Data Augmentation | Extracting the desired speech from a mixture is a meaningful and challenging task. The end-to-end DNN-based methods, though attractive, face the problem of generalization. In this paper, we explore a sequential approach for target speech extraction by combining blind source separation (BSS) with the x-vector based speaker recognition (SR) module. Two promising BSS methods based on source independence assumption, independent low-rank matrix analysis (ILRMA) and multi-channel variational autoencoder (MVAE), are utilized and compared. ILRMA employs nonnegative matrix factorization (NMF) to capture spectral structures of source signals and MVAE utilizes the strong modeling power of deep neural networks (DNN). However, the investigation of MVAE has been limited to the training with very few speakers and the speech signals of test speakers are usually included. We extend the training of MVAE using clean speech signals of 500 speakers to evaluate its generalization to unseen speakers. To improve the correct extraction rate, two data augmentation strategies are implemented to train the SR module. The performance of the proposed cascaded approach is investigated with test data constructed with real room impulse responses under varied environments. | ['Jing Lu', 'Kai Chen', 'Lele Liao', 'Zhaoyi Gu'] | 2020-05-16 | null | null | null | null | ['speech-extraction'] | ['speech'] | [ 1.17223851e-01 -3.54451627e-01 2.65283823e-01 -2.29348525e-01
-8.76286149e-01 -4.82672930e-01 5.48268139e-01 -6.41044915e-01
-2.41406396e-01 7.24856436e-01 5.54969311e-01 -3.13113570e-01
-1.87786028e-01 -3.07311974e-02 -3.94972086e-01 -1.01535165e+00
2.03854933e-01 1.14788890e-01 -4.03897107e-01 -1.56999603e-01
-9.22301263e-02 5.40298462e-01 -1.73338437e+00 1.42426014e-01
9.71212804e-01 7.65089571e-01 3.09406996e-01 8.10209095e-01
-2.66912907e-01 5.82914829e-01 -8.40554893e-01 -6.82059005e-02
1.62789613e-01 -5.90396762e-01 -7.14346468e-02 2.40671396e-01
6.52197050e-04 -1.59518421e-01 -2.43697569e-01 1.09125209e+00
8.80545497e-01 5.36576450e-01 6.21150911e-01 -1.24845159e+00
-5.70932686e-01 8.16862702e-01 -3.69069010e-01 5.25094390e-01
3.88942361e-02 -2.06739873e-01 5.73584974e-01 -1.35289371e+00
-3.95436808e-02 1.20301986e+00 3.68000388e-01 7.03157067e-01
-9.58765090e-01 -8.27430487e-01 -1.62632321e-03 3.60343874e-01
-1.43091977e+00 -1.14474642e+00 1.31673646e+00 -3.75974894e-01
7.90359199e-01 4.44578826e-01 2.39716828e-01 1.49088502e+00
-4.60405022e-01 8.67903233e-01 1.11026800e+00 -4.64413881e-01
4.66220438e-01 3.91524315e-01 2.54900217e-01 8.98305327e-02
7.31212134e-03 7.45031834e-02 -7.46722400e-01 -1.77151754e-01
6.85030639e-01 -8.55563357e-02 -5.29793799e-01 1.93211231e-02
-1.17396820e+00 6.64802849e-01 -2.40604933e-02 6.54657543e-01
-6.68807983e-01 -4.25629228e-01 2.73095965e-01 2.42319271e-01
4.14877385e-01 -1.16242498e-01 -3.58230293e-01 -3.05294003e-02
-1.29138160e+00 -2.15497375e-01 8.28280210e-01 7.32285202e-01
2.44317025e-01 1.11528313e+00 -3.03924177e-02 1.22564018e+00
7.31823564e-01 8.37849259e-01 1.07145107e+00 -5.43643475e-01
5.68490803e-01 1.32822558e-01 3.42046120e-03 -7.48701036e-01
-8.78178552e-02 -1.08876133e+00 -1.10213411e+00 2.13998392e-01
6.65929094e-02 -4.72577184e-01 -1.01958227e+00 1.68373108e+00
3.53450418e-01 6.27997279e-01 7.70778239e-01 1.16101027e+00
1.10157871e+00 1.06338620e+00 -1.55190945e-01 -7.25286543e-01
9.20394897e-01 -9.28424358e-01 -1.05211389e+00 -2.51050532e-01
-1.24312989e-01 -9.39706922e-01 6.86858475e-01 7.01235116e-01
-9.60067511e-01 -8.72287929e-01 -9.54762995e-01 4.12680924e-01
-1.23871043e-01 5.83045483e-01 1.79535076e-01 9.90462065e-01
-9.62957919e-01 1.42558217e-01 -8.67628455e-01 -3.78839411e-02
1.11535348e-01 3.11000407e-01 -4.28846478e-01 8.25238600e-02
-1.17566454e+00 5.20014584e-01 1.36879152e-02 6.97921932e-01
-1.24813998e+00 -2.01270223e-01 -8.02619040e-01 2.06185326e-01
8.36984515e-02 -4.21123564e-01 9.26576674e-01 -1.29871452e+00
-1.85943019e+00 2.45299097e-02 -4.89391655e-01 -3.51412922e-01
1.53854564e-01 -2.22674429e-01 -9.46692228e-01 1.24929465e-01
-1.60373330e-01 -1.25585182e-03 1.51005316e+00 -1.30492008e+00
-1.30634144e-01 -4.63694185e-01 -8.09292972e-01 3.66554111e-01
-5.92218220e-01 2.30126575e-01 5.58400759e-03 -9.38816369e-01
4.49903637e-01 -6.71541512e-01 -7.48344362e-02 -7.20020771e-01
-5.11151373e-01 4.10325974e-02 8.02330196e-01 -1.20078754e+00
1.00904894e+00 -2.57503486e+00 4.12702739e-01 4.06965107e-01
-1.80031717e-01 4.93841082e-01 -2.08499774e-01 2.57260323e-01
-4.11800742e-01 -3.84937137e-01 -2.48709276e-01 -6.98356748e-01
-6.60627037e-02 8.70668665e-02 -4.70442772e-01 5.17759502e-01
7.66381845e-02 3.29025328e-01 -4.23843563e-01 -3.05764526e-01
3.62604469e-01 9.35846388e-01 -2.32703462e-01 4.48193192e-01
3.15595716e-01 7.01749861e-01 -1.27415031e-01 6.50783479e-01
7.52180696e-01 2.05648720e-01 -1.60200164e-01 -3.85077327e-01
-1.52924702e-01 3.69789377e-02 -1.73728240e+00 1.59171247e+00
-2.78980196e-01 5.46923995e-01 8.19176793e-01 -1.00298202e+00
8.89669895e-01 8.27109158e-01 2.50690430e-01 -7.00196177e-02
1.80107340e-01 3.22783232e-01 2.58496940e-01 -5.92498899e-01
1.61886975e-01 -3.04819405e-01 5.37984133e-01 1.15508296e-01
3.74090701e-01 3.96594048e-01 -1.60903752e-01 5.29886447e-02
7.61262417e-01 -1.00797668e-01 8.48371387e-02 -4.40723710e-02
7.76009738e-01 -5.34657836e-01 5.81941307e-01 4.36109006e-01
-1.85288727e-01 4.79232639e-01 -2.50480086e-01 2.40411565e-01
-7.41461694e-01 -1.26076245e+00 -1.27886847e-01 9.98943806e-01
-3.07106853e-01 7.48672336e-02 -5.73246956e-01 -1.93875298e-01
-3.31970155e-01 9.43579555e-01 -2.24260092e-01 9.09041390e-02
-3.90190691e-01 -9.59751844e-01 4.68218565e-01 4.29134220e-01
3.06088239e-01 -8.88228595e-01 5.70543036e-02 2.83920109e-01
-2.27606997e-01 -1.08938062e+00 -2.08823726e-01 2.20594555e-01
-6.99631453e-01 -3.58787984e-01 -1.19156146e+00 -7.04525709e-01
5.52442133e-01 4.31998461e-01 4.16635126e-01 -5.98900139e-01
2.70049661e-01 3.58575642e-01 -5.10911882e-01 -4.18713450e-01
-6.15660906e-01 -2.48541892e-01 7.84993172e-01 6.32448733e-01
1.91257462e-01 -1.00577211e+00 -3.66965741e-01 2.64988273e-01
-8.07656348e-01 -2.25795180e-01 9.06161368e-01 7.17895329e-01
2.75272489e-01 3.30099165e-01 9.61602569e-01 -2.66599059e-01
6.76960647e-01 -5.47313154e-01 -3.40327978e-01 3.04055531e-02
-2.36064464e-01 -8.38504359e-02 7.20008075e-01 -7.33850420e-01
-1.50113380e+00 1.02955811e-01 -5.07085919e-01 -9.81608808e-01
-3.94977540e-01 5.32526314e-01 -6.55183792e-01 1.13303430e-01
6.82817757e-01 7.75905967e-01 -7.68112615e-02 -8.34012210e-01
3.61175507e-01 1.08756387e+00 6.17155313e-01 -6.08522408e-02
9.48169053e-01 1.34062931e-01 -4.70184386e-01 -1.37972534e+00
-4.58491743e-01 -6.81060851e-01 -3.35861206e-01 -2.93220788e-01
6.32701337e-01 -1.26886833e+00 -3.65315288e-01 5.53447008e-01
-1.12725210e+00 2.00754985e-01 -3.07962000e-02 1.17682707e+00
-1.04705729e-01 4.34636861e-01 -5.03177345e-01 -1.40754640e+00
-4.74851966e-01 -1.19108462e+00 8.14682543e-01 1.59920380e-01
1.70183226e-01 -5.59025288e-01 -2.81781461e-02 6.31537557e-01
4.90411460e-01 -3.21965218e-01 3.67755026e-01 -1.15869343e+00
-1.54386997e-01 -2.37043649e-02 2.72137612e-01 9.88882840e-01
3.35023642e-01 -1.64202645e-01 -1.51831663e+00 -3.72795761e-01
7.54572272e-01 1.04216881e-01 7.58775353e-01 5.90982020e-01
7.12546766e-01 -4.78781402e-01 -4.64525260e-02 3.32820654e-01
1.17236257e+00 3.28687489e-01 4.60007221e-01 -1.61366507e-01
9.31707203e-01 5.39420068e-01 4.60481308e-02 3.57491076e-01
-4.13079709e-02 2.59126097e-01 1.22766979e-01 -1.52062336e-02
-8.64181519e-02 1.35304049e-01 7.85821140e-01 1.65802777e+00
-7.17695877e-02 -5.11496842e-01 -6.16517663e-01 5.18256783e-01
-1.43661809e+00 -1.06206667e+00 -1.89966157e-01 2.09189820e+00
5.67976773e-01 2.37442832e-02 1.06060095e-01 7.96608806e-01
9.26201344e-01 2.94025578e-02 -4.82741326e-01 -4.83472012e-02
-3.36318821e-01 9.52421427e-02 1.23207092e-01 5.59751451e-01
-9.21471953e-01 4.67507392e-01 5.18930435e+00 9.80925381e-01
-1.23398113e+00 3.34779501e-01 3.06610972e-01 -1.10082917e-01
-2.44231448e-01 -3.58086914e-01 -6.50088012e-01 4.95560497e-01
1.20093811e+00 3.76331598e-01 6.50454938e-01 7.69927382e-01
4.84294236e-01 2.96714067e-01 -7.21311986e-01 1.40282369e+00
3.46380293e-01 -7.11941183e-01 -8.11799169e-02 -2.00828001e-01
5.22605360e-01 1.22842677e-01 2.28602901e-01 2.06506461e-01
-3.87415588e-02 -8.29549134e-01 7.25040853e-01 6.26813531e-01
3.84651244e-01 -8.13154042e-01 6.43790483e-01 5.81238925e-01
-9.47461903e-01 -2.78466940e-01 -2.49905020e-01 2.01926947e-01
2.46207133e-01 9.01128829e-01 -7.79372156e-01 5.89414418e-01
4.78746951e-01 3.70167106e-01 -2.44509652e-01 8.42371881e-01
-1.11731388e-01 1.15711093e+00 -4.47600335e-01 1.06859639e-01
-3.88880260e-02 -3.88141483e-01 1.17097521e+00 1.14503312e+00
5.40162325e-01 1.34770438e-01 -3.27269346e-01 8.67359042e-01
1.07710689e-01 2.54049242e-01 -3.33753943e-01 -3.11954200e-01
3.35561782e-01 1.40285063e+00 -5.25868833e-01 -1.90825775e-01
-3.76377821e-01 1.09086835e+00 -1.37186259e-01 8.21208656e-01
-6.75793886e-01 -1.44380048e-01 4.64754701e-01 -2.36002907e-01
5.62132061e-01 -4.94193077e-01 2.96152337e-03 -1.36524153e+00
1.24297455e-01 -1.12647629e+00 -1.25350416e-01 -7.95146585e-01
-1.37374949e+00 9.36897933e-01 -2.17390895e-01 -1.44893026e+00
-3.54064018e-01 -3.90221357e-01 -6.46013618e-01 1.22560823e+00
-1.36739028e+00 -9.36655879e-01 -9.24381316e-02 8.56699586e-01
8.04571867e-01 -7.00868964e-01 7.20739901e-01 6.20678008e-01
-9.76632297e-01 4.96897787e-01 6.20340168e-01 1.15318902e-01
2.98673183e-01 -8.68499517e-01 2.88426094e-02 1.22774780e+00
4.69587147e-01 6.95100904e-01 9.28983390e-01 -4.74950761e-01
-1.46248198e+00 -9.91385698e-01 5.93642056e-01 -1.40932232e-01
3.76672983e-01 -4.34886634e-01 -9.68258619e-01 4.81021911e-01
3.20992947e-01 -6.16549253e-02 1.11362946e+00 -1.88886374e-01
-4.22343910e-02 -1.67090580e-01 -9.05000091e-01 2.66868025e-01
6.17301822e-01 -5.46827555e-01 -8.04758668e-01 1.59901124e-03
5.21002769e-01 -4.43876199e-02 -5.31602979e-01 2.20015064e-01
2.13048503e-01 -8.58762205e-01 1.01371574e+00 -4.37140316e-01
-3.59198428e-03 -6.31796300e-01 -5.13007522e-01 -1.58561683e+00
-2.17181951e-01 -5.68242908e-01 -3.20709884e-01 1.78166294e+00
4.73672211e-01 -5.47830880e-01 5.65925062e-01 2.94668376e-01
-2.56084085e-01 -9.10258591e-02 -1.06015229e+00 -7.11118340e-01
-4.11679566e-01 -5.99760652e-01 3.37188989e-01 9.60659921e-01
-2.54618853e-01 6.16711915e-01 -6.58926070e-01 6.97489738e-01
7.75515258e-01 -1.45464286e-01 4.80567932e-01 -1.10498548e+00
-5.88394403e-01 -5.05123697e-02 -2.23848224e-01 -8.57449293e-01
3.74823719e-01 -7.10056603e-01 4.58018146e-02 -1.31254721e+00
-4.16720033e-01 7.35897869e-02 -4.16794777e-01 -2.58331914e-02
-1.56633571e-01 -1.57113478e-01 9.58556905e-02 3.01596850e-01
-1.12757064e-01 9.71407294e-01 7.23540127e-01 -1.23793185e-01
-4.83235985e-01 4.77692664e-01 -5.05223572e-01 6.24671102e-01
6.02504194e-01 -4.01851237e-01 -7.24622250e-01 -2.98761189e-01
-1.48680389e-01 4.25776690e-01 2.80403703e-01 -1.21386778e+00
3.43399882e-01 2.32604623e-01 7.12991118e-01 -6.52388692e-01
7.35733032e-01 -8.92358720e-01 3.88397515e-01 1.08016118e-01
-2.11635768e-01 -2.82935828e-01 2.37574980e-01 7.00339556e-01
-4.74587739e-01 -2.37209439e-01 5.68593144e-01 -1.46732014e-02
-4.76968318e-01 -3.91851552e-02 -6.15083158e-01 -4.55806881e-01
5.26206315e-01 -2.69104421e-01 1.94947720e-01 -5.23836076e-01
-1.03866768e+00 -2.89583981e-01 -3.44571024e-01 2.74610311e-01
9.54441428e-01 -1.38830125e+00 -9.57519352e-01 5.97402275e-01
-3.16860735e-01 -1.07642293e-01 6.40162468e-01 7.61062324e-01
3.20491455e-02 2.24046916e-01 2.17241924e-02 -3.56496871e-01
-1.23919690e+00 7.62585938e-01 2.18274400e-01 3.74896616e-01
-1.64196014e-01 1.07597697e+00 5.22867963e-02 -3.49443376e-01
3.73685688e-01 -9.26552266e-02 -7.33381331e-01 1.95100352e-01
5.52046239e-01 6.09562993e-01 1.12888992e-01 -1.11244440e+00
-3.63709390e-01 2.46235743e-01 2.09914461e-01 -5.54729521e-01
1.46712244e+00 -3.33703250e-01 -1.52532160e-01 5.53959966e-01
1.22049594e+00 3.88587475e-01 -8.53333533e-01 -2.95579106e-01
-2.78024286e-01 -1.74996674e-01 4.83735919e-01 -6.01279914e-01
-1.08395755e+00 9.99508142e-01 9.03061032e-01 1.78980961e-01
1.24236572e+00 -2.69477338e-01 6.16682291e-01 2.04441011e-01
-2.92635448e-02 -8.57633471e-01 -1.44979745e-01 1.09925121e-01
9.84835804e-01 -1.17945886e+00 -3.78426522e-01 -2.46358186e-01
-6.32486701e-01 9.94883299e-01 3.32208574e-01 1.44161612e-01
8.87593389e-01 1.87661633e-01 2.55979657e-01 1.65234178e-01
-3.86177897e-01 -1.58594564e-01 4.83349323e-01 4.42815900e-01
2.16237873e-01 -1.91738065e-02 1.32446587e-01 1.05905020e+00
-2.34028429e-01 -2.06077486e-01 3.59082580e-01 6.83555067e-01
-4.04478312e-01 -8.01895499e-01 -1.01866961e+00 2.98054934e-01
-5.74134767e-01 -3.17113757e-01 -1.01108201e-01 3.81676666e-02
-1.81263298e-01 1.40555549e+00 -3.32099438e-01 -4.18428361e-01
8.36747587e-02 3.63585591e-01 1.28116578e-01 -4.32446122e-01
-3.22018981e-01 7.31473923e-01 -1.41174495e-01 8.63348041e-03
-6.94067478e-01 -7.61246383e-01 -9.69370246e-01 1.64075121e-01
-5.97312868e-01 4.69691008e-01 1.08548629e+00 1.05506861e+00
1.44297957e-01 6.01707160e-01 8.94403577e-01 -8.95660639e-01
-6.91474974e-01 -1.29526746e+00 -9.07234669e-01 1.12928711e-01
8.33415926e-01 -4.01035279e-01 -9.42151070e-01 1.60672158e-01] | [15.068680763244629, 5.760958194732666] |
7640a3c9-c052-46c5-a9a8-73e551ff658e | non-deep-networks-1 | 2110.07641 | null | https://arxiv.org/abs/2110.07641v1 | https://arxiv.org/pdf/2110.07641v1.pdf | Non-deep Networks | Depth is the hallmark of deep neural networks. But more depth means more sequential computation and higher latency. This begs the question -- is it possible to build high-performing "non-deep" neural networks? We show that it is. To do so, we use parallel subnetworks instead of stacking one layer after another. This helps effectively reduce depth while maintaining high performance. By utilizing parallel substructures, we show, for the first time, that a network with a depth of just 12 can achieve top-1 accuracy over 80% on ImageNet, 96% on CIFAR10, and 81% on CIFAR100. We also show that a network with a low-depth (12) backbone can achieve an AP of 48% on MS-COCO. We analyze the scaling rules for our design and show how to increase performance without changing the network's depth. Finally, we provide a proof of concept for how non-deep networks could be used to build low-latency recognition systems. Code is available at https://github.com/imankgoyal/NonDeepNetworks. | ['Vladlen Koltun', 'Jia Deng', 'Alexey Bochkovskiy', 'Ankit Goyal'] | 2021-10-14 | non-deep-networks | https://openreview.net/forum?id=Xg47v73CDaj | https://openreview.net/pdf?id=Xg47v73CDaj | null | ['real-time-object-detection'] | ['computer-vision'] | [-1.61480442e-01 1.32277906e-01 1.58765689e-01 -5.47589362e-01
-3.77306074e-01 -4.53486234e-01 2.96698641e-02 -2.54327387e-01
-8.63762379e-01 3.72391999e-01 -1.63230732e-01 -7.43080974e-01
1.31889313e-01 -7.54171491e-01 -9.26554263e-01 -5.16656339e-01
-2.09677383e-01 2.72826433e-01 6.98539674e-01 -6.81604594e-02
5.27411923e-02 7.85445452e-01 -1.38454962e+00 6.65519297e-01
3.31621468e-02 1.04447246e+00 1.32156909e-01 9.42397535e-01
1.13885678e-01 8.47139001e-01 -7.63018906e-01 -5.01667857e-01
5.48162162e-01 2.30127620e-03 -1.05502605e+00 -4.63441670e-01
8.25503826e-01 -5.81905305e-01 -5.72009385e-01 8.96230519e-01
5.02155423e-01 -2.29706168e-01 1.63629279e-01 -1.12082171e+00
-1.42004848e-01 9.25137520e-01 -4.01238680e-01 4.99838650e-01
-3.47842216e-01 1.25024408e-01 1.02639401e+00 -6.65953457e-01
3.17701131e-01 1.13407516e+00 9.05073464e-01 7.22677469e-01
-1.13313329e+00 -1.11669064e+00 2.69377232e-01 -3.93218175e-02
-1.36287606e+00 -8.48784328e-01 2.52179593e-01 -1.65435180e-01
1.42782521e+00 5.72279803e-02 6.67372465e-01 9.12169337e-01
1.55001700e-01 5.86938024e-01 9.01272833e-01 -2.03627810e-01
8.03825557e-02 -2.47357056e-01 6.39025986e-01 9.16935384e-01
5.67746341e-01 -2.14319870e-01 -4.42757457e-01 1.49018198e-01
9.16714728e-01 1.49489343e-02 -1.69765234e-01 2.73000240e-01
-1.04973102e+00 7.03603446e-01 7.05475390e-01 5.12068450e-01
-2.53856666e-02 7.42768407e-01 5.27942657e-01 6.84248269e-01
1.56078801e-01 6.57209873e-01 -6.50850356e-01 -1.62649423e-01
-1.00022876e+00 3.00861318e-02 1.22558463e+00 8.85225892e-01
7.89850771e-01 2.69178659e-01 5.40392876e-01 7.05022395e-01
-1.54150929e-03 2.98088670e-01 2.91716188e-01 -1.24117541e+00
4.47623610e-01 3.16209853e-01 -4.80347544e-01 -5.81305385e-01
-7.32679427e-01 -5.30156791e-01 -9.24358070e-01 5.21775246e-01
7.44349062e-01 -5.77891767e-01 -9.50196743e-01 1.76158237e+00
-2.21135616e-01 3.46412733e-02 -3.83060686e-02 8.32876205e-01
6.31167114e-01 6.97505534e-01 -2.19422117e-01 3.55311781e-01
1.47096848e+00 -1.14566839e+00 -1.13365436e-02 -4.61611658e-01
9.34452891e-01 -5.83845377e-01 8.22284281e-01 5.85258842e-01
-1.16184926e+00 -4.52437013e-01 -1.31896150e+00 -1.76689908e-01
-2.15026855e-01 -7.30672404e-02 9.02417243e-01 7.58875191e-01
-1.62072754e+00 8.06895375e-01 -1.11908507e+00 -2.87411302e-01
4.59061205e-01 7.73387611e-01 -3.90366733e-01 9.44748074e-02
-8.47599566e-01 6.93669915e-01 4.22839493e-01 -7.57256895e-02
-8.91246498e-01 -6.52912259e-01 -2.22463176e-01 2.15408146e-01
-8.21654424e-02 -5.69765151e-01 1.40768886e+00 -1.10714412e+00
-1.35431993e+00 8.02676022e-01 -2.02692941e-01 -8.92277598e-01
2.98214138e-01 -2.00527921e-01 -2.46537492e-01 2.56201476e-01
-4.09889013e-01 1.13591015e+00 2.21122950e-01 -7.53597498e-01
-7.27893591e-01 -3.69530648e-01 3.51044983e-01 -9.90699306e-02
-5.52530229e-01 5.56611605e-02 -5.93539834e-01 -2.91337401e-01
3.23327452e-01 -1.17698014e+00 -2.77496457e-01 1.99507236e-01
-3.04877192e-01 -9.14588571e-02 5.59893608e-01 -4.18276250e-01
8.37302148e-01 -2.26916409e+00 -4.26850200e-01 3.94666255e-01
7.03771830e-01 3.52780014e-01 -3.78004313e-01 9.99019220e-02
-2.98904516e-02 2.75326908e-01 1.42802864e-01 -2.83470362e-01
-2.44082272e-01 1.38043523e-01 -1.24013036e-01 4.05225724e-01
4.87402119e-02 6.69661522e-01 -4.48692083e-01 4.39892858e-02
-3.24334979e-01 6.06356263e-01 -8.79604697e-01 -2.24949479e-01
-4.90932576e-02 -2.24939466e-01 -2.37765089e-02 3.70361656e-01
6.11758530e-01 -5.08704126e-01 3.23336124e-01 -3.41537148e-01
-2.83060223e-01 5.77161849e-01 -8.87814403e-01 1.46059215e+00
-2.12558553e-01 1.12060881e+00 3.68727118e-01 -8.27971697e-01
8.16787541e-01 1.51543498e-01 2.37364937e-02 -8.82842362e-01
3.45441908e-01 4.18607801e-01 5.31750202e-01 1.10745721e-01
1.81636408e-01 9.83285755e-02 3.01883757e-01 3.82857352e-01
3.69729884e-02 3.15076143e-01 1.77299842e-01 7.97368437e-02
1.47825348e+00 -7.28179276e-01 -4.82884496e-01 -4.56529647e-01
-8.72820541e-02 -8.30083713e-02 4.51133966e-01 9.02842104e-01
-2.62856394e-01 5.05696177e-01 8.13057959e-01 -5.77772200e-01
-1.36067772e+00 -9.54860270e-01 -4.81323972e-02 1.24210811e+00
-3.07951838e-01 -4.13214594e-01 -9.88175392e-01 -1.41330913e-01
-1.28293455e-01 1.51745170e-01 -2.79111952e-01 6.67762011e-02
-6.76698744e-01 -8.08422744e-01 1.18856704e+00 7.39037931e-01
8.87884140e-01 -6.39246881e-01 -5.21511376e-01 1.44856632e-01
1.94557801e-01 -1.17770147e+00 -7.53049403e-02 5.90384305e-01
-1.29559982e+00 -8.50916743e-01 -7.26166129e-01 -1.08514273e+00
5.69078147e-01 4.02698219e-01 9.67624187e-01 3.21808726e-01
7.57074961e-03 -2.34320730e-01 -1.35953166e-02 -1.92527875e-01
-1.53241649e-01 6.88548207e-01 1.23055287e-01 -5.82979023e-01
3.71767223e-01 -9.08359587e-01 -8.37382972e-01 2.57036000e-01
-5.33188999e-01 2.94915050e-01 7.78391540e-01 5.27082682e-01
2.55292118e-01 -8.22636709e-02 4.53224301e-01 -9.75047410e-01
2.04629257e-01 -2.47599199e-01 -6.81726694e-01 -1.30287766e-01
-5.82016408e-01 1.00819757e-02 6.32749736e-01 -4.69478428e-01
-3.86759013e-01 1.35173559e-01 -6.64834142e-01 -3.71537715e-01
2.47794781e-02 6.30468279e-02 1.38703406e-01 -2.91303575e-01
7.14299083e-01 -8.63636136e-02 1.78318918e-01 -5.88648915e-01
1.22677200e-01 4.13403243e-01 3.55460793e-01 -4.18579221e-01
3.03393900e-01 5.01266837e-01 -1.32045776e-01 -7.55756795e-01
-6.27487063e-01 -2.12033838e-01 -2.69687474e-01 1.54678389e-01
6.64933801e-01 -1.22750938e+00 -1.11931205e+00 7.07278132e-01
-1.33029306e+00 -9.39081013e-01 1.28808692e-01 3.29606771e-01
-1.69769190e-02 -1.60830960e-01 -1.33985174e+00 -4.10676241e-01
-6.29481733e-01 -1.10092425e+00 4.44296449e-01 1.60839230e-01
-1.05214693e-01 -7.32485056e-01 -3.17899555e-01 2.16380909e-01
6.01396263e-01 -2.50120848e-01 7.20740318e-01 -7.17282712e-01
-8.30744743e-01 -1.51527403e-02 -7.50445545e-01 6.82621419e-01
-3.62915456e-01 1.30885122e-02 -1.23066473e+00 -5.13774753e-01
-2.09270820e-01 -2.67780870e-01 1.07920623e+00 2.06912324e-01
1.26775277e+00 -5.40448427e-01 -2.42102385e-01 9.84004140e-01
1.54635501e+00 2.67616183e-01 6.95482254e-01 3.93153757e-01
8.05848122e-01 3.60152721e-01 -4.66835260e-01 2.31571719e-01
2.65956551e-01 3.24026287e-01 2.96109349e-01 -1.48015663e-01
-3.36574912e-01 1.15865029e-01 3.63187432e-01 1.19018555e+00
-4.82718879e-03 -2.73968816e-01 -1.23188686e+00 4.21164989e-01
-1.39812386e+00 -7.70520329e-01 -2.53201902e-01 1.81350279e+00
8.29723179e-01 7.36303985e-01 1.45373628e-01 1.46266982e-01
5.30291140e-01 1.04203999e-01 -4.52027559e-01 -6.44123435e-01
-1.71103880e-01 4.35447961e-01 9.98886526e-01 6.70613706e-01
-7.85735548e-01 1.03385782e+00 6.95839119e+00 6.25288785e-01
-1.32234740e+00 2.99730539e-01 9.57199633e-01 -4.24887091e-01
-1.11329909e-02 -7.72975758e-02 -1.25810575e+00 2.93189496e-01
1.48262858e+00 2.83103883e-01 3.02278131e-01 9.60112810e-01
-1.30153134e-01 3.08486074e-02 -1.19707584e+00 8.51013601e-01
-2.76272506e-01 -1.58473456e+00 8.79906863e-02 1.95384502e-01
6.38135672e-01 8.08134437e-01 -2.05780849e-01 1.54453680e-01
2.68763572e-01 -9.84785378e-01 7.01901257e-01 1.95199817e-01
8.23141694e-01 -8.25811446e-01 7.34226048e-01 2.49761060e-01
-1.03916788e+00 -3.44835371e-02 -5.90930045e-01 -4.46605861e-01
-2.82121629e-01 7.69605041e-01 -7.35025704e-01 -2.82551646e-01
7.47579932e-01 3.09680372e-01 -5.83897948e-01 1.07978737e+00
9.41805243e-02 7.32561767e-01 -7.55832970e-01 -3.18599135e-01
3.62951249e-01 3.29350442e-01 5.97288180e-03 1.65647912e+00
2.14186907e-01 7.37195238e-02 -2.36953273e-01 5.22024333e-01
-5.91392994e-01 -3.56651902e-01 -5.59112012e-01 -2.87113455e-03
7.27780163e-01 1.03803957e+00 -8.87240708e-01 -4.86537904e-01
-3.31743389e-01 9.92202818e-01 5.10047257e-01 2.40388587e-01
-6.16984487e-01 -6.63177073e-01 9.46383715e-01 1.82279631e-01
9.02202949e-02 -5.12885571e-01 -5.27712643e-01 -1.08434522e+00
-9.99480709e-02 -1.00857437e+00 -1.77402161e-02 -5.15411556e-01
-9.20654297e-01 8.17398727e-01 -5.68576336e-01 -5.54551184e-01
1.79217145e-01 -1.03972352e+00 -4.76719379e-01 6.89376235e-01
-1.18968892e+00 -5.89416087e-01 -2.80960888e-01 4.69502300e-01
3.65497977e-01 -6.51699901e-02 7.27115393e-01 7.36860514e-01
-6.59363210e-01 1.02354705e+00 -3.23463255e-03 7.37879097e-01
3.42880309e-01 -8.51040542e-01 9.29626703e-01 8.67117047e-01
9.54039842e-02 8.09958875e-01 4.31616873e-01 -1.77501500e-01
-1.51644051e+00 -8.94484878e-01 8.40024889e-01 -1.50586218e-01
6.99435294e-01 -6.28933787e-01 -9.25733209e-01 8.23717892e-01
2.77903508e-02 -1.26916289e-01 4.84688044e-01 2.92680681e-01
-8.52370441e-01 -5.90449572e-01 -7.82604873e-01 8.20947528e-01
1.31344247e+00 -5.22255778e-01 -2.17529327e-01 3.07987630e-01
1.00930858e+00 -3.54483902e-01 -7.98659623e-01 2.36581922e-01
7.94222951e-01 -1.13253546e+00 9.02401745e-01 -2.80895352e-01
3.52615863e-01 4.16017771e-02 -1.59643605e-01 -8.93741906e-01
-4.39318568e-01 -4.33154553e-01 -6.11842610e-02 7.49105752e-01
7.12541401e-01 -1.10617101e+00 1.32534182e+00 4.93529677e-01
-2.53213316e-01 -8.74064982e-01 -8.68920982e-01 -1.09258306e+00
3.77359331e-01 -6.35312438e-01 4.01894987e-01 6.58254743e-01
-2.11280718e-01 3.77370805e-01 9.26533937e-02 3.60086381e-01
4.60712641e-01 -4.28762853e-01 6.70077264e-01 -1.05035532e+00
-4.15102810e-01 -8.64464283e-01 -5.58329105e-01 -1.49992144e+00
-5.44553511e-02 -8.30233157e-01 -1.16162382e-01 -1.22369766e+00
7.66935945e-02 -6.78234339e-01 -3.18666130e-01 8.14154208e-01
6.02545321e-01 5.86484671e-01 4.35477227e-01 2.72215098e-01
-4.16270107e-01 -2.72671998e-01 9.92339432e-01 1.65508598e-01
4.73862700e-02 -3.47670048e-01 -8.61874402e-01 8.69708359e-01
1.19413495e+00 -4.70948428e-01 -3.56516421e-01 -1.06678760e+00
2.76112705e-01 -1.30927384e-01 3.01498741e-01 -1.72249544e+00
6.43060923e-01 4.05294478e-01 3.20213795e-01 -3.09896857e-01
5.13686299e-01 -5.52979887e-01 2.44082566e-02 8.70265245e-01
-3.23146909e-01 4.54934835e-01 4.42156285e-01 2.87091117e-02
-2.34598853e-02 -2.02317804e-01 1.01702619e+00 -1.04414970e-01
-6.67012274e-01 3.05181891e-01 -4.55086142e-01 -3.04584727e-02
3.97171140e-01 -1.36686146e-01 -8.58951449e-01 -3.55269723e-02
-5.48126638e-01 1.05421707e-01 3.46307099e-01 -1.82921458e-02
5.20847082e-01 -8.33068967e-01 -5.37825823e-01 2.32458234e-01
-3.43646199e-01 -3.39777991e-02 5.45291305e-02 7.44338870e-01
-1.20328498e+00 5.37434995e-01 -2.77136385e-01 -4.86484408e-01
-1.25478458e+00 -1.14260033e-01 7.86495566e-01 -1.92157447e-03
-7.50917315e-01 1.45505130e+00 3.27756196e-01 -7.29330480e-02
5.40516555e-01 -4.08522785e-01 4.03938562e-01 -7.04801381e-02
7.41054475e-01 2.54388541e-01 2.57011503e-01 8.62509906e-02
-5.36820292e-01 3.15751791e-01 -4.76756960e-01 -3.89678180e-01
1.41564333e+00 1.83482155e-01 -2.02908859e-01 3.25641334e-01
1.64639151e+00 -3.99877220e-01 -1.34239876e+00 -6.87130839e-02
-1.91667844e-02 7.99369514e-02 2.63053983e-01 -6.04589224e-01
-1.52375293e+00 1.10517049e+00 6.04344487e-01 8.39521512e-02
1.15513408e+00 -1.59029588e-01 1.05249012e+00 1.08615363e+00
4.66550469e-01 -9.04940426e-01 1.03842013e-01 1.10597527e+00
4.79002506e-01 -8.53583932e-01 -3.04185480e-01 -1.77797228e-01
-7.25728124e-02 1.22769403e+00 7.87396550e-01 -5.27266324e-01
8.52211237e-01 8.97845745e-01 -2.40291599e-02 -2.46873975e-01
-1.10163260e+00 -4.75154445e-02 -2.64194548e-01 1.74905896e-01
6.11038744e-01 5.33585511e-02 1.29340857e-01 3.72480452e-01
-5.22720098e-01 -4.44365578e-05 5.44404685e-01 9.28157568e-01
-8.39510977e-01 -8.34454477e-01 -2.78044455e-02 5.69213629e-01
-7.42764056e-01 -5.18586397e-01 -1.94544449e-01 7.64113605e-01
4.92371470e-02 7.67489731e-01 4.90948677e-01 -7.46225715e-01
1.83329150e-01 1.34519145e-01 4.67780530e-01 -5.92966855e-01
-7.06738293e-01 -1.05117865e-01 4.19140399e-01 -5.79623282e-01
3.80386636e-02 -1.37451157e-01 -1.34002721e+00 -1.33304584e+00
-1.35821134e-01 -1.60993293e-01 8.85666013e-01 6.03939176e-01
4.26823586e-01 4.46361721e-01 3.96890849e-01 -7.44111538e-01
-2.40273163e-01 -6.23714149e-01 -3.55471045e-01 -2.93684155e-01
3.84856194e-01 2.54489854e-02 -5.27415872e-01 -2.22577497e-01] | [8.618408203125, 2.951463460922241] |
5b291ff6-ca7a-4183-8eed-d59e9a877032 | unveiling-the-two-faced-truth-disentangling | 2306.03002 | null | https://arxiv.org/abs/2306.03002v1 | https://arxiv.org/pdf/2306.03002v1.pdf | Unveiling the Two-Faced Truth: Disentangling Morphed Identities for Face Morphing Detection | Morphing attacks keep threatening biometric systems, especially face recognition systems. Over time they have become simpler to perform and more realistic, as such, the usage of deep learning systems to detect these attacks has grown. At the same time, there is a constant concern regarding the lack of interpretability of deep learning models. Balancing performance and interpretability has been a difficult task for scientists. However, by leveraging domain information and proving some constraints, we have been able to develop IDistill, an interpretable method with state-of-the-art performance that provides information on both the identity separation on morph samples and their contribution to the final prediction. The domain information is learnt by an autoencoder and distilled to a classifier system in order to teach it to separate identity information. When compared to other methods in the literature it outperforms them in three out of five databases and is competitive in the remaining. | ['Jaime S. Cardoso', 'Ana F. Sequeira', 'Naser Damer', 'Tiago Gonçalves', 'Pedro C. Neto', 'Eduarda Caldeira'] | 2023-06-05 | null | null | null | null | ['face-recognition'] | ['computer-vision'] | [ 2.77286232e-01 2.76400298e-01 1.34643286e-01 -5.49460709e-01
-2.28109106e-01 -7.06246793e-01 7.39829004e-01 1.35946080e-01
-2.37271711e-01 6.06781602e-01 -2.02806950e-01 -3.89970124e-01
-2.85823405e-01 -5.60213804e-01 -4.25010502e-01 -7.73890257e-01
7.76092783e-02 8.44575167e-01 -1.67906746e-01 -3.02327216e-01
3.15756828e-01 9.15813386e-01 -1.57308507e+00 5.60463190e-01
4.72215682e-01 1.17174840e+00 -6.10294402e-01 4.40510511e-01
1.03877634e-01 6.51979208e-01 -6.47447884e-01 -1.19167006e+00
4.45462793e-01 -4.13669795e-01 -9.91365552e-01 -1.29371583e-01
8.05675149e-01 -3.84535819e-01 -9.57152396e-02 1.11061370e+00
6.36216164e-01 -5.80657780e-01 6.91233873e-01 -1.23015273e+00
-7.42818713e-01 3.71612787e-01 -4.48959380e-01 -1.01386413e-01
2.75497228e-01 1.13054737e-01 9.41869199e-01 -6.62566662e-01
5.98472655e-01 1.31851578e+00 7.20308602e-01 8.46239686e-01
-1.43946934e+00 -7.06152856e-01 -2.56995440e-01 3.67721468e-01
-1.10379875e+00 -6.29544258e-01 6.48333609e-01 -4.39040184e-01
6.32835865e-01 3.74643445e-01 5.56703269e-01 1.43016434e+00
-5.98841645e-02 5.87773561e-01 1.31539202e+00 -5.71569443e-01
1.36083692e-01 5.50881624e-01 7.05449358e-02 6.93224788e-01
3.98359179e-01 2.00826272e-01 -4.34384614e-01 -9.27213207e-02
4.05876189e-01 -2.29730502e-01 -5.68725243e-02 -2.82004505e-01
-4.11218584e-01 7.73531079e-01 1.15437128e-01 3.89354408e-01
-2.39384070e-01 -2.73452938e-01 3.01711738e-01 3.55545580e-01
3.61327022e-01 5.88132262e-01 -4.53457832e-01 -5.78789338e-02
-1.05572999e+00 1.93064511e-01 1.17209864e+00 2.28807226e-01
5.34366131e-01 1.53188542e-01 4.25183564e-01 6.50765657e-01
5.84944785e-02 2.82819062e-01 2.05672041e-01 -6.08632743e-01
4.03542258e-03 9.49754477e-01 -3.21864873e-01 -1.10664630e+00
-3.50561023e-01 -3.66305172e-01 -8.43530118e-01 6.67800903e-01
7.61479378e-01 6.31624088e-02 -1.01894331e+00 1.55732822e+00
1.06446154e-01 8.64665136e-02 1.70391854e-02 8.55103910e-01
4.43271011e-01 2.37317532e-01 1.70958769e-02 1.97117612e-01
1.47513318e+00 -1.84496790e-01 -5.51713824e-01 -9.64980945e-02
1.92524374e-01 -9.42448556e-01 6.25514209e-01 8.97050560e-01
-7.38985837e-01 -4.89884675e-01 -1.30456376e+00 1.67655256e-02
-7.25592315e-01 3.03013604e-02 7.87880599e-01 1.41491163e+00
-8.85600805e-01 8.93951595e-01 -5.52705288e-01 -4.73408043e-01
7.01071024e-01 9.45339561e-01 -7.96432674e-01 1.43964931e-01
-1.04009426e+00 1.10622716e+00 2.96065986e-01 1.54929847e-01
-5.55282712e-01 -5.19024849e-01 -4.67590034e-01 7.48767052e-03
-2.48903758e-04 -4.00555551e-01 7.32794702e-01 -1.42916012e+00
-1.52945244e+00 1.28073561e+00 -1.70019586e-02 -6.83075130e-01
5.60720503e-01 -3.23328115e-02 -6.17182493e-01 3.76158580e-02
-4.25852716e-01 5.54317176e-01 1.01298106e+00 -1.23518085e+00
-4.98797625e-01 -6.85702145e-01 1.17891036e-01 -3.37606966e-01
-6.44437432e-01 9.93359461e-02 -1.33206591e-01 -3.57963353e-01
-8.31092671e-02 -1.08080149e+00 2.69589603e-01 -6.51171058e-02
-4.82138067e-01 -3.84213775e-02 1.07004368e+00 -9.58230257e-01
8.46958399e-01 -2.07361245e+00 1.59620687e-01 3.33177686e-01
3.10182542e-01 7.91551709e-01 4.10393104e-02 2.95637995e-01
-5.11094451e-01 2.64599055e-01 -1.90271646e-01 -4.41915393e-01
-6.50238171e-02 2.74210632e-01 -2.68561214e-01 5.05452693e-01
6.74400389e-01 4.71932620e-01 -5.05431235e-01 -2.67817855e-01
2.92670757e-01 6.27775311e-01 -5.23750722e-01 1.84211642e-01
1.43340766e-01 4.14170980e-01 -8.00337642e-02 7.36212254e-01
8.57943773e-01 1.63355216e-01 3.11289161e-01 -2.41180882e-01
2.35340223e-01 1.92329120e-02 -1.28498173e+00 9.77548718e-01
-2.60208309e-01 9.81811881e-01 -1.02991462e-01 -1.15183103e+00
9.97442544e-01 3.27072978e-01 4.50716704e-01 -5.21205544e-01
2.42957488e-01 3.75403255e-01 3.80878150e-01 -4.47445810e-01
2.33423173e-01 -4.70449775e-01 1.88731283e-01 3.78130972e-01
2.89317071e-01 1.89059168e-01 -1.77261397e-01 -2.99291879e-01
7.43791521e-01 5.61500117e-02 1.62538037e-01 -2.89672971e-01
8.63296688e-01 -2.54190117e-01 3.92973930e-01 4.20886785e-01
-2.54922092e-01 6.46674037e-01 6.05200171e-01 -1.01207054e+00
-1.06635797e+00 -9.41656530e-01 -2.50116855e-01 6.67010307e-01
-7.32289925e-02 -8.36596787e-02 -1.19852495e+00 -6.71116948e-01
1.20217636e-01 5.63646495e-01 -8.77208650e-01 -3.61132532e-01
-4.81846809e-01 -7.41379440e-01 9.69997287e-01 2.77130514e-01
4.28868473e-01 -7.75445938e-01 -3.25906307e-01 2.60693170e-02
1.35145709e-01 -1.20074511e+00 1.59168959e-01 3.05030376e-01
-5.69209158e-01 -1.25239825e+00 -3.17186832e-01 -4.21614558e-01
5.74027240e-01 -3.74880433e-01 1.05720806e+00 1.83536515e-01
-5.77337325e-01 3.95610034e-02 -4.39505689e-02 -9.98954356e-01
-6.74166679e-01 2.35245764e-01 1.86562419e-01 5.91520965e-01
7.66913772e-01 -4.74630803e-01 -4.49559391e-01 1.75605536e-01
-9.92645621e-01 -2.83030629e-01 6.21038795e-01 8.90796185e-01
-2.19566077e-02 7.18147382e-02 3.31766933e-01 -9.79765832e-01
5.40853620e-01 -1.33513227e-01 -3.97379905e-01 2.43177429e-01
-6.94415629e-01 2.64310032e-01 6.45329714e-01 -4.07154799e-01
-9.28573132e-01 1.67476997e-01 -3.21186900e-01 -1.23485692e-01
-4.46868062e-01 3.23823914e-02 -5.16139507e-01 -4.86511946e-01
8.36369455e-01 2.83006411e-02 2.96601683e-01 -4.51395065e-01
2.41488311e-02 8.68858576e-01 4.05004323e-01 -3.56805742e-01
9.96106505e-01 5.83375216e-01 2.21301526e-01 -1.03076637e+00
-5.37210047e-01 -8.34673345e-02 -7.99351275e-01 -2.54459381e-01
8.17537725e-01 -4.14519727e-01 -9.84890044e-01 7.72357583e-01
-1.12631500e+00 2.56489813e-01 -7.50704631e-02 8.83060470e-02
-2.03451246e-01 4.29912686e-01 -3.97355050e-01 -8.72552395e-01
-1.46232024e-01 -1.33573365e+00 8.70416641e-01 2.28608683e-01
-4.58059788e-01 -1.11875677e+00 -2.10970327e-01 6.73441112e-01
3.51981610e-01 4.49254960e-01 9.64590907e-01 -1.11853945e+00
-3.31428200e-01 -5.82511961e-01 -1.91544935e-01 7.23913252e-01
2.58501351e-01 1.80515513e-01 -1.66947281e+00 -1.67382956e-01
2.15907171e-01 -2.36056700e-01 5.95394790e-01 -3.39453630e-02
1.17603230e+00 -4.41785991e-01 -7.21754804e-02 6.11356437e-01
1.25287819e+00 1.71384722e-01 8.47853720e-01 3.12073261e-01
4.01286751e-01 9.96531904e-01 1.14978917e-01 5.05328923e-02
-2.67756581e-02 8.51692379e-01 5.60635090e-01 -2.88512170e-01
-9.92860049e-02 -7.31669813e-02 1.91033423e-01 1.12599611e-01
-3.44000846e-01 1.43750785e-02 -9.84600544e-01 1.98303595e-01
-1.49753582e+00 -1.12914574e+00 1.06716909e-01 2.13730121e+00
7.81084001e-01 2.85333872e-01 1.54677063e-01 6.18197680e-01
5.56829154e-01 -1.48204699e-01 -4.67379510e-01 -8.38832557e-01
-1.89945385e-01 4.72232193e-01 3.95181715e-01 4.26849604e-01
-1.09693992e+00 9.51210439e-01 5.93574524e+00 7.08609104e-01
-1.40073955e+00 -3.37358087e-01 9.92744505e-01 4.93897408e-01
2.36428692e-03 -1.02543518e-01 -7.92037070e-01 3.41466099e-01
1.04126108e+00 3.29501003e-01 4.68288302e-01 8.24489534e-01
-1.73739344e-01 6.92864060e-02 -1.37327909e+00 1.02343822e+00
3.66035700e-01 -1.09773743e+00 2.15075761e-01 4.34248924e-01
3.00117522e-01 -6.46991789e-01 3.47214341e-01 -2.32034884e-02
6.55327439e-02 -1.49420094e+00 4.95292038e-01 2.82588631e-01
4.32766497e-01 -8.50264490e-01 1.00849879e+00 1.75690219e-01
-5.37913442e-01 -9.89971906e-02 -2.96506405e-01 -1.05976187e-01
-3.91733766e-01 4.10730720e-01 -1.19049430e+00 5.59054554e-01
4.71979588e-01 1.62293166e-01 -7.17520595e-01 8.11438262e-01
-2.08549798e-01 6.07038915e-01 -2.35990688e-01 5.53274155e-02
6.74795732e-02 -1.19926631e-01 3.60077441e-01 1.17141843e+00
-8.50118622e-02 -6.98879287e-02 -3.58098388e-01 6.04871988e-01
-8.60102922e-02 -1.48241207e-01 -4.85962659e-01 -2.18542576e-01
1.14654668e-01 1.28829217e+00 -5.79749703e-01 -4.74084429e-02
-8.03276449e-02 1.04738939e+00 1.46385282e-01 -8.92991498e-02
-7.70770490e-01 -2.67086238e-01 9.71165419e-01 1.51485145e-01
4.68269512e-02 4.55198064e-02 -6.83475256e-01 -9.67316687e-01
-2.84200348e-02 -1.45416391e+00 4.05001342e-01 -1.98021978e-01
-1.24829602e+00 8.83520186e-01 -1.89082339e-01 -1.01104891e+00
-3.75929445e-01 -1.26446831e+00 -2.49464482e-01 1.06911564e+00
-1.38307822e+00 -1.37432778e+00 -2.54755020e-01 3.38409513e-01
1.38184786e-01 -5.78080237e-01 1.28435528e+00 3.83348316e-01
-5.38166285e-01 9.15179729e-01 -1.46935180e-01 4.51554865e-01
7.82905638e-01 -1.23508584e+00 4.29025084e-01 7.86244214e-01
5.22984922e-01 7.31919587e-01 7.86578894e-01 -1.79027647e-01
-1.37360787e+00 -4.74358112e-01 8.68144453e-01 -6.93292618e-01
5.24579167e-01 -5.90573728e-01 -9.34451759e-01 3.63804668e-01
2.44990423e-01 -2.17135295e-01 1.12758875e+00 2.06772193e-01
-8.74655902e-01 -4.44774657e-01 -1.47142208e+00 4.15889978e-01
6.68625653e-01 -6.95160985e-01 -5.10603487e-01 3.89847234e-02
-8.30748007e-02 -2.79214472e-01 -7.68934369e-01 3.41181993e-01
1.00678039e+00 -1.38843548e+00 9.71012175e-01 -9.61877286e-01
3.65159631e-01 -1.35708705e-01 5.36242675e-04 -1.06222093e+00
-5.83693236e-02 -5.05969226e-01 -6.20631427e-02 1.41909373e+00
4.82301831e-01 -8.28829587e-01 1.20725203e+00 9.39698339e-01
5.09346783e-01 -5.76805174e-01 -8.88927579e-01 -7.94839382e-01
7.53876716e-02 -1.41229287e-01 6.53629959e-01 1.00798774e+00
-2.76252508e-01 2.63809264e-01 -5.03697813e-01 3.33730817e-01
8.07275474e-01 1.62137132e-02 7.84021974e-01 -1.65377510e+00
-2.97744513e-01 -6.37408137e-01 -1.03337896e+00 -3.67811263e-01
2.52082855e-01 -6.28337622e-01 -5.48108101e-01 -9.23460007e-01
1.15330731e-04 -1.12015337e-01 -1.88855723e-01 7.12856770e-01
-4.16330919e-02 6.46393180e-01 2.22073525e-01 1.02655608e-02
2.20345974e-01 -5.16982377e-02 7.99475431e-01 -2.74270386e-01
2.19201982e-01 1.59754351e-01 -8.75577748e-01 8.53330255e-01
9.40563560e-01 -4.28685784e-01 -1.92238107e-01 -2.28422746e-01
-1.29918205e-02 -3.19753408e-01 2.92128563e-01 -1.16873753e+00
5.05261943e-02 1.10574536e-01 8.55721354e-01 -9.85468999e-02
5.87240756e-01 -1.10611367e+00 2.27749914e-01 6.99977160e-01
-1.29272878e-01 -2.48762950e-01 3.59777778e-01 1.19367167e-01
-2.82944649e-01 -2.03073859e-01 1.06853306e+00 8.12646672e-02
-5.50063610e-01 2.42654383e-01 -7.14300498e-02 -1.71522900e-01
1.01007473e+00 -5.81144035e-01 -9.34233516e-02 -4.33637947e-01
-7.23088682e-01 -2.78815240e-01 4.59264308e-01 5.92292130e-01
4.88798320e-01 -9.83693540e-01 -7.58665562e-01 4.54721212e-01
-5.05997464e-02 -6.29129052e-01 9.68547463e-02 4.84790683e-01
-7.59542286e-01 3.74920517e-01 -6.61132634e-01 -5.31062841e-01
-1.76374364e+00 3.50160450e-01 5.88959157e-01 -1.36586934e-01
3.98075804e-02 8.12224686e-01 -6.86068088e-02 -2.88993508e-01
3.52443933e-01 2.93500721e-01 -2.08647951e-01 2.89820880e-01
6.85303152e-01 1.68231130e-01 3.90549123e-01 -6.50050819e-01
-5.50369620e-01 4.00814772e-01 -2.26108894e-01 1.21569037e-01
1.54108942e+00 4.71412241e-01 -2.84615993e-01 1.17370440e-02
1.21152508e+00 1.04061186e-01 -9.32216942e-01 2.79899120e-01
4.22564745e-01 -6.10696435e-01 -3.14787388e-01 -1.00672662e+00
-1.07739866e+00 1.22788846e+00 8.55587482e-01 5.23629189e-01
1.22497761e+00 -4.32183951e-01 4.13429946e-01 2.92059630e-01
2.02244312e-01 -8.96851003e-01 -3.81074324e-02 3.11728328e-01
7.19818532e-01 -1.32944107e+00 2.08966583e-01 -4.83258337e-01
-4.83120799e-01 1.56111300e+00 3.81069154e-01 7.14938343e-03
5.27169108e-01 3.58964443e-01 2.65861899e-01 -6.08710721e-02
-2.22968027e-01 8.60092510e-03 4.53199774e-01 8.93685758e-01
4.66989547e-01 -1.22074701e-01 -2.81025581e-02 4.21173513e-01
-3.53058606e-01 -2.14928120e-01 3.36016268e-01 3.99874479e-01
-3.02035594e-03 -1.79625940e+00 -3.94958884e-01 1.70110643e-01
-8.57180059e-01 8.63846317e-02 -1.16078413e+00 9.93800640e-01
4.54937518e-01 8.50813150e-01 -2.78944820e-01 -6.29292607e-01
2.75786251e-01 3.03532481e-01 6.24153078e-01 -2.13608369e-01
-9.64486837e-01 -5.84573984e-01 2.35970676e-01 -3.13870072e-01
-2.49320194e-01 -6.74826622e-01 -6.90282285e-01 -6.72500134e-01
-2.69956052e-01 2.92489342e-02 8.84098530e-01 9.73218143e-01
3.17081898e-01 4.60949428e-02 6.40456915e-01 -7.80829132e-01
-7.46973753e-01 -7.03473449e-01 -3.32213044e-01 7.03801095e-01
2.57109970e-01 -4.82387036e-01 -1.87517509e-01 1.41770437e-01] | [13.011161804199219, 1.0850123167037964] |
a3fb0083-1c05-489a-ab80-ccec1485d10c | decision-trees-for-decision-making-under-the | 2003.00360 | null | https://arxiv.org/abs/2003.00360v2 | https://arxiv.org/pdf/2003.00360v2.pdf | Decision Trees for Decision-Making under the Predict-then-Optimize Framework | We consider the use of decision trees for decision-making problems under the predict-then-optimize framework. That is, we would like to first use a decision tree to predict unknown input parameters of an optimization problem, and then make decisions by solving the optimization problem using the predicted parameters. A natural loss function in this framework is to measure the suboptimality of the decisions induced by the predicted input parameters, as opposed to measuring loss using input parameter prediction error. This natural loss function is known in the literature as the Smart Predict-then-Optimize (SPO) loss, and we propose a tractable methodology called SPO Trees (SPOTs) for training decision trees under this loss. SPOTs benefit from the interpretability of decision trees, providing an interpretable segmentation of contextual features into groups with distinct optimal solutions to the optimization problem of interest. We conduct several numerical experiments on synthetic and real data including the prediction of travel times for shortest path problems and predicting click probabilities for news article recommendation. We demonstrate on these datasets that SPOTs simultaneously provide higher quality decisions and significantly lower model complexity than other machine learning approaches (e.g., CART) trained to minimize prediction error. | ['Jason Cheuk Nam Liang', 'Ryan McNellis', 'Adam N. Elmachtoub'] | 2020-02-29 | null | https://proceedings.icml.cc/static/paper_files/icml/2020/6150-Paper.pdf | https://proceedings.icml.cc/static/paper_files/icml/2020/6150-Paper.pdf | icml-2020-1 | ['parameter-prediction'] | ['miscellaneous'] | [ 3.37868690e-01 4.11885262e-01 -5.04608572e-01 -8.77402306e-01
-7.40508258e-01 -3.88459116e-01 2.00050190e-01 2.96085894e-01
-4.11756575e-01 6.62664413e-01 -3.96149121e-02 -8.09332371e-01
-5.74255228e-01 -9.53815162e-01 -6.18361592e-01 -6.64477587e-01
-9.87933129e-02 7.68888533e-01 -7.27450475e-02 1.42156882e-02
4.71804410e-01 2.51357406e-01 -1.40725935e+00 2.82610804e-01
1.20219290e+00 1.44919920e+00 1.97936848e-01 6.27092242e-01
8.42667371e-02 4.89443690e-01 -9.55192670e-02 -7.39600062e-01
4.16468143e-01 -4.91409525e-02 -9.85891283e-01 1.35136619e-01
8.21084157e-02 -7.19159022e-02 2.97651350e-01 8.27917874e-01
1.56014010e-01 3.33930939e-01 9.37236726e-01 -1.30349898e+00
-1.53289959e-01 7.75028169e-01 -3.25148523e-01 7.62464330e-02
1.22922771e-01 -1.63067773e-01 1.77701986e+00 -7.10552752e-01
3.86392832e-01 1.32417941e+00 7.10781097e-01 1.39371797e-01
-1.60836446e+00 -2.61516839e-01 3.54433537e-01 3.87739182e-01
-1.07436454e+00 -2.49968961e-01 3.61987054e-01 -6.42839015e-01
5.63945889e-01 3.90100449e-01 1.54256791e-01 6.98689103e-01
4.98458117e-01 1.01308107e+00 9.48582292e-01 -4.39332902e-01
5.17014623e-01 5.30324161e-01 4.42408234e-01 6.86000466e-01
9.51859802e-02 3.46555382e-01 -1.98790014e-01 -2.24967092e-01
1.20295621e-01 3.29525083e-01 -6.64995611e-02 -4.22883332e-01
-8.19432259e-01 1.22423577e+00 4.96904641e-01 -3.44121784e-01
-3.53459865e-01 -2.17731595e-01 9.22557265e-02 6.90077722e-01
5.37561834e-01 5.61747611e-01 -7.95128226e-01 1.16541214e-01
-6.34896398e-01 4.49374110e-01 1.17196560e+00 7.04646230e-01
8.75135422e-01 -3.92474324e-01 -2.67101079e-01 7.75348485e-01
3.61495316e-01 2.59164512e-01 2.13752404e-01 -9.23807144e-01
8.04461241e-01 3.38330954e-01 4.84208465e-01 -1.06737649e+00
-5.56668878e-01 -6.78800106e-01 -5.31784773e-01 2.97655314e-01
6.27432048e-01 -3.85051668e-01 -7.36289918e-01 1.48684478e+00
3.79746675e-01 -4.31354582e-01 -1.30632684e-01 8.72302175e-01
-2.64124483e-01 7.48346567e-01 2.17454080e-02 -2.98278689e-01
8.65597188e-01 -9.69217002e-01 -3.78779322e-01 -2.22762331e-01
1.04757273e+00 -4.51058656e-01 1.10872400e+00 6.49382591e-01
-7.63495564e-01 -3.33602428e-01 -9.74992752e-01 2.61269331e-01
-2.39331424e-01 7.80989081e-02 6.35340571e-01 5.86284578e-01
-6.32277966e-01 1.19304550e+00 -7.59241402e-01 -1.82418734e-01
2.55394697e-01 4.23590183e-01 2.45389864e-01 -1.55728400e-01
-9.71877813e-01 9.91038442e-01 1.68702513e-01 3.99355032e-02
-5.63264012e-01 -5.85513175e-01 -5.33791482e-01 1.50651559e-01
7.06022739e-01 -6.57963037e-01 1.49893236e+00 -1.07826698e+00
-1.34150600e+00 5.53065538e-01 -1.42650247e-01 -9.29477036e-01
7.41990745e-01 -4.87676948e-01 -3.26296329e-01 -4.86660987e-01
2.64136791e-01 3.20979923e-01 8.93534601e-01 -9.91190612e-01
-1.20731449e+00 -4.78629053e-01 1.71780959e-01 3.79300751e-02
-7.03745848e-03 -2.73773611e-01 -1.30162397e-02 -4.70783591e-01
1.27675878e-02 -1.09440660e+00 -7.87756860e-01 -1.58381350e-02
-8.41873288e-01 -2.85766304e-01 4.23732400e-01 -7.07186699e-01
1.51102662e+00 -1.84498608e+00 -2.51730029e-02 6.10327363e-01
1.48352385e-01 -1.28817409e-01 5.96908629e-02 3.41323793e-01
1.44055784e-01 3.07788968e-01 -3.36066306e-01 -3.15079540e-01
2.23316014e-01 3.76330584e-01 -4.38603342e-01 3.42338353e-01
-8.80003255e-03 5.05387485e-01 -5.08705020e-01 -3.23439896e-01
-1.93396416e-02 -2.23684937e-01 -8.55222106e-01 1.75639525e-01
-5.53449154e-01 1.48328185e-01 -9.10217881e-01 5.30018806e-01
3.38411659e-01 -3.84605676e-01 2.59148598e-01 2.28045672e-01
-1.14222109e-01 6.03900075e-01 -1.13250411e+00 9.52888131e-01
-7.81396091e-01 3.87897760e-01 -1.11806847e-01 -1.50149179e+00
8.96071732e-01 -2.82815903e-01 2.55633414e-01 -6.02512300e-01
7.76364878e-02 1.15642183e-01 -1.58013716e-01 -3.97374272e-01
2.65891910e-01 -2.23093480e-02 -2.20598668e-01 4.41240430e-01
-3.60143632e-01 2.45471150e-01 -7.65282959e-02 -1.01179078e-01
9.87643301e-01 -9.36165601e-02 5.33133507e-01 -2.49585167e-01
2.62207627e-01 9.14035216e-02 8.50298882e-01 1.01821816e+00
2.03475162e-01 1.82418257e-01 6.67888224e-01 -6.85091197e-01
-9.38854516e-01 -8.24287832e-01 -4.28344697e-01 1.26746905e+00
-1.43588841e-01 -2.96650469e-01 -4.81134325e-01 -9.89628971e-01
4.97902721e-01 1.18857670e+00 -5.66575766e-01 1.19140245e-01
-2.53824562e-01 -7.59555101e-01 -2.59857118e-01 2.94413447e-01
2.10727051e-01 -5.30686617e-01 -4.19024974e-01 4.04995531e-01
-1.32752806e-01 -7.70859778e-01 -3.88296843e-01 5.20556092e-01
-1.03146839e+00 -1.04102027e+00 -3.17518830e-01 -5.66499829e-01
6.11662209e-01 -1.25836164e-01 1.26661897e+00 -9.00637880e-02
-7.54554123e-02 -5.30886613e-02 -4.42764670e-01 -3.78440768e-01
-2.16383725e-01 1.27790973e-01 1.39844194e-01 4.01398242e-01
1.69920072e-01 -5.08036137e-01 -6.07193947e-01 8.96152139e-01
-3.69839966e-01 1.28135651e-01 7.24303782e-01 1.01991415e+00
9.60063696e-01 2.64290214e-01 4.67768431e-01 -1.36705303e+00
5.76979637e-01 -7.74767876e-01 -8.35338652e-01 5.00206113e-01
-1.34033358e+00 5.66064179e-01 8.66025031e-01 -1.09272517e-01
-8.26449454e-01 9.00911018e-02 4.30950001e-02 -4.30107899e-02
2.51479503e-02 6.54193163e-01 -1.49511665e-01 5.28988950e-02
5.12096107e-01 -1.43413037e-01 -3.87389541e-01 -6.68257594e-01
2.75664806e-01 8.14374387e-01 8.07764307e-02 -6.36798978e-01
4.40433592e-01 2.54656672e-01 1.51946977e-01 -3.58820587e-01
-1.34431195e+00 -2.74980724e-01 -4.65674669e-01 -7.15214573e-03
6.82538271e-01 -2.87089467e-01 -7.75759757e-01 -3.19248624e-02
-8.13464403e-01 -2.56941557e-01 -3.52992594e-01 3.63476455e-01
-7.99518883e-01 -1.08695244e-02 -1.36789188e-01 -1.09934604e+00
-2.16780365e-01 -9.71654177e-01 7.25291967e-01 9.01414752e-02
-2.53902882e-01 -1.04798186e+00 -1.16206028e-01 4.52247530e-01
5.21528497e-02 3.37913752e-01 1.27159715e+00 -9.90375459e-01
-5.21757901e-01 -4.05630678e-01 -7.46543333e-02 3.84840071e-01
-1.36475548e-01 -1.00130588e-01 -6.58593535e-01 -1.16709776e-01
8.99356902e-02 4.56072763e-03 9.01732624e-01 6.44748926e-01
1.40424168e+00 -1.03542590e+00 -6.37007058e-01 6.38655782e-01
1.43934560e+00 3.50835025e-01 1.95056170e-01 4.23708290e-01
2.69293010e-01 9.45822299e-01 1.05881548e+00 6.87480390e-01
5.87217510e-01 8.75038803e-01 2.83877254e-01 1.45192146e-01
5.99457443e-01 -6.07363760e-01 7.34373704e-02 3.54554445e-01
3.12639058e-01 -4.38917816e-01 -1.00404036e+00 2.87086874e-01
-2.13621330e+00 -4.98553783e-01 1.61961973e-01 2.58043742e+00
6.42723560e-01 4.24915224e-01 3.33894998e-01 2.23244935e-01
6.71479583e-01 -2.75003135e-01 -9.04478490e-01 -7.17013955e-01
2.21233949e-01 -6.21530600e-02 8.15203249e-01 6.70027912e-01
-1.17528558e+00 5.40172815e-01 6.07979679e+00 8.16411912e-01
-8.61772001e-01 -2.49049380e-01 1.19659710e+00 5.41260652e-02
-2.19275549e-01 1.84642836e-01 -1.13918555e+00 5.81222653e-01
1.05738473e+00 -3.43214154e-01 4.69760716e-01 1.14714015e+00
5.31518161e-01 -8.33268687e-02 -1.48624551e+00 6.91839695e-01
-5.47504365e-01 -1.26952350e+00 -3.96784060e-02 1.92750007e-01
7.74590433e-01 -1.53101385e-01 2.25791767e-01 1.03844732e-01
9.24085319e-01 -9.79899108e-01 6.03097141e-01 7.61647344e-01
4.21643168e-01 -8.98201704e-01 6.53170884e-01 5.73073745e-01
-7.37778544e-01 -6.21850193e-01 -4.78119224e-01 -2.20893994e-01
8.43397230e-02 1.08313370e+00 -1.10442114e+00 2.85913378e-01
6.26386642e-01 9.39927220e-01 -1.47295088e-01 1.30871964e+00
-1.07597403e-01 8.03558111e-01 -5.75153053e-01 -4.26977664e-01
3.67119938e-01 -4.50442731e-01 5.50247550e-01 9.97560441e-01
2.01579988e-01 -3.44912969e-02 3.91372442e-01 7.28517354e-01
1.55871078e-01 2.45072693e-01 -3.08969468e-01 5.12737259e-02
2.82224953e-01 7.58661866e-01 -4.99123573e-01 -2.09384924e-03
-2.94021487e-01 5.77349663e-01 3.66325021e-01 4.50522572e-01
-4.63507473e-01 -4.40296322e-01 7.73052812e-01 1.45624325e-01
4.40955579e-01 4.30772230e-02 -8.90946925e-01 -8.62139523e-01
2.03603312e-01 -6.14465535e-01 6.64370537e-01 -2.67842233e-01
-1.40817881e+00 3.58484119e-01 -2.19822526e-01 -1.38960862e+00
-3.31589699e-01 -6.07270360e-01 -7.18284190e-01 6.73701048e-01
-1.48338127e+00 -4.36401933e-01 2.59272426e-01 3.14967901e-01
6.13707304e-01 1.84302196e-01 5.36152780e-01 -8.52869600e-02
-7.03084707e-01 7.19755292e-01 7.00381875e-01 -2.77580380e-01
2.67902613e-01 -1.44631970e+00 3.80959511e-01 4.26109493e-01
7.12152943e-02 1.09352067e-01 8.77116680e-01 -2.47799292e-01
-1.08647728e+00 -1.08483374e+00 1.21630454e+00 -3.88605058e-01
6.91771805e-01 -1.48903653e-01 -4.74658996e-01 6.84229374e-01
-5.68347573e-01 -4.33780611e-01 6.73788607e-01 4.76312011e-01
-2.62753069e-02 -3.88814926e-01 -1.26909959e+00 4.80596900e-01
8.51842642e-01 -2.70629562e-02 -3.60022396e-01 6.45970225e-01
8.02799165e-01 3.16987187e-02 -8.11559498e-01 1.50309488e-01
6.76957190e-01 -7.98274159e-01 7.86195755e-01 -1.18036282e+00
4.81534988e-01 7.15179965e-02 -3.45991582e-01 -1.52416539e+00
-3.45717520e-01 -8.34591031e-01 1.14235461e-01 8.41144502e-01
1.20041108e+00 -8.52041364e-01 8.34237754e-01 1.05857193e+00
1.76818401e-01 -1.29356647e+00 -7.98538089e-01 -6.17974877e-01
1.02860965e-01 -6.41630411e-01 5.10913312e-01 5.02663612e-01
-9.43568200e-02 2.95239300e-01 -4.82699901e-01 1.41719922e-01
8.11783671e-01 5.59638321e-01 5.92082083e-01 -1.38184440e+00
-6.16282165e-01 -2.61465609e-01 -2.66771048e-01 -1.45766473e+00
2.68712584e-02 -8.51844668e-01 7.92503357e-02 -1.38375497e+00
-2.84782827e-01 -1.07219505e+00 -2.98530340e-01 3.39457333e-01
-1.21326938e-01 -6.57639802e-01 2.29821846e-01 2.96158731e-01
-7.59617150e-01 4.68497306e-01 1.12685645e+00 5.39973285e-03
-3.73729080e-01 1.16406691e+00 -8.53112936e-01 8.51071417e-01
6.34014785e-01 -7.47895896e-01 -2.37169296e-01 -2.08413720e-01
4.70102012e-01 4.44983780e-01 6.37844531e-03 -7.07972765e-01
1.88298017e-01 -5.39942682e-01 2.01062486e-01 -4.50132698e-01
-3.06305178e-02 -8.95296991e-01 -7.96356276e-02 3.95643681e-01
-9.21521604e-01 -3.67800631e-02 -4.48314995e-01 8.51423681e-01
-5.71000613e-02 -2.49785587e-01 6.91035271e-01 1.96204215e-01
-5.56523323e-01 2.96521991e-01 -2.87056267e-01 1.38873637e-01
8.76880586e-01 -7.52629712e-02 -2.49055959e-03 -5.92508197e-01
-1.15933669e+00 7.38989294e-01 1.15806438e-01 1.06408305e-01
4.51767206e-01 -1.01955521e+00 -7.36381412e-01 6.51696920e-02
3.22216004e-02 -1.56463459e-01 -2.32701316e-01 8.04854333e-01
-2.06728086e-01 5.91551125e-01 2.41086230e-01 -3.60091925e-01
-9.04964209e-01 4.54831630e-01 4.63658392e-01 -8.33777606e-01
-4.19521779e-01 8.41276109e-01 2.36711815e-01 -5.89287937e-01
4.14973080e-01 -4.85564291e-01 -1.42790005e-01 1.36480927e-02
3.77493382e-01 6.25333607e-01 9.88540426e-02 1.07156869e-03
-2.24921092e-01 3.70357871e-01 -2.16264084e-01 2.49151830e-02
1.36035395e+00 -3.59845787e-01 4.46064889e-01 3.73738796e-01
1.12757456e+00 -3.48619699e-01 -1.31375337e+00 -6.75581694e-01
4.95156825e-01 -5.28852046e-01 1.46697640e-01 -1.19354939e+00
-8.93338025e-01 7.87108421e-01 4.08268332e-01 5.04608154e-01
1.06927085e+00 -2.30493635e-01 7.96981454e-01 7.88791955e-01
4.06654924e-01 -1.39015305e+00 -5.18577576e-01 2.91988462e-01
7.47858346e-01 -1.22955155e+00 -1.72721833e-01 -3.26001108e-01
-9.63551283e-01 1.15443766e+00 1.53536707e-01 -6.21892996e-02
1.04487109e+00 7.26290047e-02 -1.76913142e-01 2.25464240e-01
-1.17217040e+00 -1.23226397e-01 2.77500302e-01 3.70025277e-01
-9.06384960e-02 6.24528110e-01 -2.72349149e-01 8.99914205e-01
-4.82600361e-01 -2.98895948e-02 6.90945163e-02 6.44853115e-01
-6.52994156e-01 -9.79163468e-01 -9.19313207e-02 1.12551177e+00
-4.11935598e-01 -7.36158267e-02 -3.03705841e-01 3.04672629e-01
-1.63231492e-01 1.21327198e+00 -5.32559678e-02 -8.02751005e-01
4.04491603e-01 3.47077884e-02 -2.27461681e-01 -4.03670579e-01
-2.38264009e-01 -3.55212927e-01 5.57831347e-01 -7.01043963e-01
8.64018053e-02 -8.35431278e-01 -8.28864813e-01 -4.68684465e-01
-4.44903284e-01 4.18548614e-01 8.63633573e-01 1.20734394e+00
3.59198630e-01 1.12916470e-01 1.32610250e+00 -4.63829219e-01
-1.20163870e+00 -5.79229116e-01 -7.93681502e-01 2.29613036e-01
4.28248078e-01 -6.75802350e-01 -7.24119067e-01 -2.66277701e-01] | [8.091716766357422, 4.159327507019043] |
31a7595e-4b37-4d20-9603-d4de5ff59043 | fair-information-spread-on-social-networks | 2305.08791 | null | https://arxiv.org/abs/2305.08791v1 | https://arxiv.org/pdf/2305.08791v1.pdf | Fair Information Spread on Social Networks with Community Structure | Information spread through social networks is ubiquitous. Influence maximiza- tion (IM) algorithms aim to identify individuals who will generate the greatest spread through the social network if provided with information, and have been largely devel- oped with marketing in mind. In social networks with community structure, which are very common, IM algorithms focused solely on maximizing spread may yield signifi- cant disparities in information coverage between communities, which is problematic in settings such as public health messaging. While some IM algorithms aim to remedy disparity in information coverage using node attributes, none use the empirical com- munity structure within the network itself, which may be beneficial since communities directly affect the spread of information. Further, the use of empirical network struc- ture allows us to leverage community detection techniques, making it possible to run fair-aware algorithms when there are no relevant node attributes available, or when node attributes do not accurately capture network community structure. In contrast to other fair IM algorithms, this work relies on fitting a model to the social network which is then used to determine a seed allocation strategy for optimal fair information spread. We develop an algorithm to determine optimal seed allocations for expected fair coverage, defined through maximum entropy, provide some theoretical guarantees under appropriate conditions, and demonstrate its empirical accuracy on both simu- lated and real networks. Because this algorithm relies on a fitted network model and not on the network directly, it is well-suited for partially observed and noisy social networks. | ['Ji Zhu', 'Elizaveta Levina', 'Octavio Mesner'] | 2023-05-15 | null | null | null | null | ['community-detection', 'marketing'] | ['graphs', 'miscellaneous'] | [ 3.06977004e-01 4.25922453e-01 -6.82159662e-01 1.24100782e-01
-2.78884619e-02 -7.47950017e-01 4.50718105e-01 6.72551334e-01
-2.89041311e-01 7.50005424e-01 2.11346760e-01 -4.05267239e-01
-4.43163902e-01 -1.38765693e+00 -2.23776221e-01 -2.24950463e-01
-6.24041378e-01 6.99114382e-01 1.87775925e-01 -3.30313236e-01
2.30103120e-01 1.58285290e-01 -8.93527091e-01 -3.29650670e-01
9.00564790e-01 2.94572383e-01 -1.80521101e-01 6.94617569e-01
-2.51214772e-01 7.64713764e-01 -6.99036896e-01 -6.55924559e-01
3.74203831e-01 -5.63329637e-01 -6.97349668e-01 5.20334095e-02
-2.92323440e-01 -2.41768092e-01 -1.53733879e-01 1.06434739e+00
3.70514423e-01 -1.62809268e-01 8.69129121e-01 -1.50251710e+00
-7.75469393e-02 1.26409209e+00 -8.19451272e-01 4.20668036e-01
6.01246178e-01 -1.90637484e-01 1.44516027e+00 2.60162558e-02
8.45910847e-01 1.31235981e+00 9.01520729e-01 1.36390209e-01
-1.65298033e+00 -7.06920743e-01 3.24489176e-02 -3.55185270e-01
-1.33257282e+00 -1.36967525e-01 6.82971656e-01 -3.95568281e-01
3.38354707e-01 5.16339779e-01 1.01237988e+00 7.25211203e-01
6.46565259e-02 6.17451787e-01 9.50886548e-01 -2.04422623e-01
3.22816372e-01 1.31757051e-01 -1.20522939e-01 2.73013592e-01
7.46569872e-01 -8.32379162e-02 -4.14531767e-01 -8.73404503e-01
6.80095673e-01 1.29032820e-01 -3.98550391e-01 -1.32735714e-01
-1.11055684e+00 1.25867009e+00 7.60693848e-01 2.09345475e-01
-5.92785835e-01 1.14323452e-01 4.47017178e-02 5.52612543e-01
8.57307732e-01 6.08449697e-01 5.49802743e-02 -3.36542204e-02
-1.31625462e+00 2.52945006e-01 1.49800420e+00 5.96094429e-01
9.24643397e-01 -5.46641648e-01 4.07771654e-02 5.24687111e-01
4.41758633e-01 5.73370695e-01 -2.45780766e-01 -1.01123297e+00
3.84944797e-01 7.84686506e-01 -8.71329382e-02 -1.55794585e+00
-3.98827374e-01 -4.50915515e-01 -9.80745494e-01 -1.33880466e-01
5.39654911e-01 -7.07470238e-01 -3.01779121e-01 1.74181736e+00
4.12166625e-01 2.11373061e-01 -3.16413552e-01 6.17260098e-01
3.94234926e-01 3.58673245e-01 -2.29926139e-01 -5.52399814e-01
8.18244934e-01 -2.93657064e-01 -6.17035568e-01 -1.14667155e-01
4.94881272e-01 -7.55363226e-01 3.88913929e-01 -2.87437830e-02
-1.05365276e+00 4.03727531e-01 -8.63844395e-01 7.46095717e-01
-1.10425547e-01 -1.07327378e+00 6.42306566e-01 1.01628113e+00
-1.27880442e+00 7.61413217e-01 -5.22922695e-01 -7.03094780e-01
7.36524761e-01 3.50213647e-01 1.69438958e-01 -1.23993628e-01
-1.15256655e+00 4.54296112e-01 -4.78122868e-02 -1.88167408e-01
-5.00484884e-01 -7.13799179e-01 -4.48521525e-01 8.40487182e-02
6.94596350e-01 -7.52154708e-01 8.20661604e-01 -1.10695589e+00
-7.60949731e-01 4.07192677e-01 -8.59257728e-02 -4.69645143e-01
8.21870267e-01 6.16119087e-01 -1.13930888e-01 3.97406787e-01
3.39505017e-01 4.24871922e-01 5.23107648e-01 -1.20544267e+00
-3.58185202e-01 -6.05679117e-03 4.17160779e-01 2.83325374e-01
-5.18568099e-01 2.10953519e-01 -1.33852929e-01 -3.85627985e-01
3.71399261e-02 -8.45289707e-01 -8.20540071e-01 -4.35489677e-02
-7.47212231e-01 9.59132314e-02 6.63224161e-01 -3.02756548e-01
1.37052011e+00 -1.46270180e+00 -1.34226725e-01 1.15813649e+00
8.74955833e-01 -2.12986127e-01 -1.13077953e-01 8.14171255e-01
4.98407304e-01 8.88794363e-01 -1.57859176e-01 -8.64456668e-02
-2.34558836e-01 4.14772592e-02 2.48462081e-01 5.51869571e-01
-8.67293030e-03 6.18837953e-01 -1.07969403e+00 -5.78229368e-01
-2.19252720e-01 2.34333366e-01 -6.99192286e-01 -1.37980878e-01
-2.39453241e-01 7.42371380e-02 -7.01530814e-01 5.19723654e-01
7.21349418e-01 -7.25966752e-01 5.93790352e-01 4.78424072e-01
2.43784532e-01 3.10602151e-02 -1.29492497e+00 7.20138431e-01
-1.13298893e-01 6.89073265e-01 6.00794971e-01 -7.80642450e-01
7.07775295e-01 2.08923399e-01 9.34189081e-01 4.71854210e-02
2.94343412e-01 4.43568379e-02 2.33998716e-01 -6.25368729e-02
4.26518202e-01 -1.50445282e-01 2.33826756e-01 1.34577692e+00
-5.11185884e-01 8.69502947e-02 4.10497516e-01 9.91176307e-01
1.56706989e+00 -9.47935522e-01 3.37951839e-01 -4.63475436e-01
-7.82558024e-02 6.85063154e-02 2.59544194e-01 1.08670044e+00
-2.97021717e-01 3.12983036e-01 1.00188053e+00 1.97665676e-01
-1.11478484e+00 -1.00139880e+00 -3.17435414e-02 7.75781035e-01
5.28820634e-01 -4.21381205e-01 -9.20464277e-01 -5.38208246e-01
1.57232508e-01 1.20147407e-01 -5.81128955e-01 2.66325772e-01
-7.12951506e-03 -1.05997050e+00 4.41143870e-01 -1.85792118e-01
3.37124646e-01 -7.75312722e-01 -4.68784422e-02 5.16547620e-01
-5.16817093e-01 -7.95906246e-01 -7.45940089e-01 -2.67287970e-01
-8.07829976e-01 -1.56159806e+00 -7.17637181e-01 -2.11138815e-01
8.82323563e-01 4.11228269e-01 1.26520395e+00 7.09056973e-01
-2.05211103e-01 6.22275889e-01 -3.50845575e-01 -3.95401269e-01
-7.19858408e-01 4.84572589e-01 -1.21709868e-01 1.54920295e-01
4.03779447e-01 -7.89237618e-01 -7.62702584e-01 4.29889888e-01
-8.75059068e-01 -2.31016174e-01 3.58450353e-01 3.00336510e-01
-1.61163762e-01 2.75167316e-01 9.45452213e-01 -1.11762249e+00
1.38488877e+00 -1.15329087e+00 -3.00239235e-01 3.78696322e-02
-8.85925949e-01 -2.19022095e-01 1.17597520e-01 -4.10964489e-01
-6.88649595e-01 -5.96482873e-01 2.35987261e-01 2.25737274e-01
5.22020876e-01 9.05457377e-01 3.01052302e-01 -1.02822237e-01
9.51901019e-01 -3.95006895e-01 5.78498900e-01 -5.23791909e-02
2.34319627e-01 8.33278894e-01 -4.66238290e-01 -2.80994058e-01
6.98059082e-01 5.82960725e-01 1.19368769e-01 -8.20739746e-01
-4.14335221e-01 -5.45236886e-01 -3.16160858e-01 -4.46651399e-01
2.23574445e-01 -7.19747186e-01 -9.57508206e-01 1.95553020e-01
-8.48534346e-01 -1.96916506e-01 -1.75552174e-01 2.60507613e-01
-1.07294746e-01 4.44588572e-01 -7.34848619e-01 -1.13696074e+00
-2.83971637e-01 -7.04352915e-01 2.35327542e-01 1.29553989e-01
-5.07754266e-01 -1.54309905e+00 -3.89081589e-03 1.81361556e-01
7.24618852e-01 4.88060087e-01 4.17103797e-01 -9.24199581e-01
-8.94321740e-01 -3.40038896e-01 -4.65322107e-01 -1.12901285e-01
3.00059348e-01 2.14412540e-01 -4.34464246e-01 -4.60846782e-01
-5.56415260e-01 -5.57406479e-03 7.23520696e-01 8.06233585e-01
4.05025303e-01 -6.08975410e-01 -8.20438683e-01 -6.36603460e-02
1.28924727e+00 -4.73215431e-01 4.36005473e-01 -7.00223669e-02
3.16035897e-01 9.06659007e-01 1.59540966e-01 7.79039860e-01
5.96572816e-01 3.29399675e-01 5.34116864e-01 -1.71132341e-01
1.11230962e-01 -4.76557255e-01 2.06052035e-01 4.97839451e-01
-1.12108424e-01 -5.11651456e-01 -7.69572914e-01 6.96999609e-01
-1.70061886e+00 -1.20119977e+00 -2.70720869e-01 2.30265594e+00
1.04189622e+00 3.23219895e-01 6.38138175e-01 1.34566709e-01
1.17638338e+00 2.06589073e-01 -4.54144865e-01 -3.83493081e-02
-1.92133188e-01 -1.17754720e-01 9.71359730e-01 9.26825345e-01
-6.51747167e-01 3.79368216e-01 7.23568726e+00 7.43608117e-01
-6.50427461e-01 6.28983183e-03 8.49071205e-01 -2.84828335e-01
-7.83842444e-01 2.15584472e-01 -5.19070625e-01 5.00793457e-01
9.23544765e-01 -8.12316239e-01 7.28606105e-01 3.12619448e-01
6.09226525e-01 -3.01141113e-01 -8.43415260e-01 4.10879135e-01
-3.00428301e-01 -1.44666219e+00 -3.75542372e-01 6.74763918e-01
1.02739561e+00 2.31188253e-01 -3.61167759e-01 -3.09416205e-01
9.62471783e-01 -1.01277041e+00 1.35784969e-01 3.20024610e-01
5.87929845e-01 -7.55618155e-01 6.69190347e-01 5.60843885e-01
-1.07358849e+00 -1.73561845e-03 -4.02727276e-01 -2.26666957e-01
4.35790092e-01 1.23894191e+00 -1.19782448e+00 1.54718652e-01
3.90948236e-01 9.13120985e-01 -2.50147730e-01 1.31793809e+00
1.26790673e-01 9.65370178e-01 -8.73858690e-01 -5.17116010e-01
-1.20424735e-03 -1.39704987e-01 1.04760396e+00 9.93820548e-01
1.53399080e-01 -1.56759232e-01 3.87827247e-01 1.03878689e+00
-3.95768046e-01 3.57040048e-01 -7.58962095e-01 -3.70474458e-01
9.47946250e-01 1.20266545e+00 -1.14637434e+00 -1.39394328e-01
-1.25810727e-01 3.18828106e-01 6.91857710e-02 5.61951637e-01
-3.06113422e-01 -2.69240767e-01 6.57478869e-01 8.45439672e-01
-1.99104976e-02 1.13644749e-01 -3.54033262e-01 -9.20320928e-01
-3.57173443e-01 -7.98056006e-01 4.05782729e-01 -1.48115158e-01
-1.53466332e+00 2.61932194e-01 9.52665284e-02 -8.10173333e-01
-2.66304195e-01 7.04070032e-02 -9.01657879e-01 7.84668326e-01
-1.21445155e+00 -6.94498837e-01 -3.26535851e-02 3.50264668e-01
-1.46123707e-01 6.84449077e-02 2.67300248e-01 1.51305109e-01
-3.33851755e-01 5.35148680e-01 1.94579005e-01 1.08544298e-01
3.07800174e-01 -1.10999584e+00 2.26606801e-01 7.09511459e-01
-5.61069406e-04 6.42597914e-01 9.21691239e-01 -1.10964179e+00
-9.19391870e-01 -8.47688019e-01 9.03135717e-01 -2.45559305e-01
8.43147635e-01 -3.11770290e-01 -5.90052366e-01 4.52134013e-01
2.07080424e-01 -2.90655762e-01 7.60743260e-01 5.24228454e-01
-8.82714428e-03 1.03404053e-01 -1.56335366e+00 8.13403606e-01
1.18680048e+00 -2.65653104e-01 3.03219467e-01 5.08606672e-01
6.93179488e-01 1.29069135e-01 -9.55322504e-01 1.76038984e-02
2.97336727e-01 -8.80079985e-01 7.66334891e-01 -2.61649966e-01
3.41419429e-01 1.29742967e-03 2.60712564e-01 -1.47787106e+00
-1.43382013e-01 -1.09776807e+00 1.81349993e-01 1.18929529e+00
8.52718890e-01 -9.52570975e-01 1.16976857e+00 6.21065974e-01
8.92151952e-01 -6.03397906e-01 -5.69414914e-01 -5.00310719e-01
-1.54698402e-01 -1.41595006e-01 7.24285424e-01 1.14335287e+00
2.46134967e-01 3.73814732e-01 -2.35033691e-01 -1.35634705e-01
1.01838732e+00 -1.14513554e-01 7.04397917e-01 -1.75677824e+00
-2.33489573e-01 -7.17496574e-01 -1.58389166e-01 -9.02930617e-01
-4.39245924e-02 -8.23645294e-01 -1.96435079e-01 -1.60747099e+00
5.81129968e-01 -1.07578850e+00 3.14219564e-01 2.20581263e-01
-2.55555600e-01 2.95963854e-01 9.54055488e-02 4.28109348e-01
-5.93251944e-01 5.62506728e-02 1.24744833e+00 -2.00497031e-01
-5.87684393e-01 5.07967591e-01 -1.26228452e+00 4.43613738e-01
8.21363509e-01 -7.31200933e-01 -6.63648188e-01 4.16935295e-01
8.95713568e-01 7.26565048e-02 1.89716950e-01 -4.13437665e-01
4.00887042e-01 -3.80537689e-01 -2.52077300e-02 -2.72216469e-01
4.06300835e-02 -7.14150608e-01 4.14326876e-01 5.29153466e-01
-5.30668914e-01 -3.10296118e-01 -3.83825153e-01 9.64643955e-01
5.62061816e-02 -4.31903869e-01 4.06388134e-01 -1.93455756e-01
2.27591336e-01 4.26330298e-01 -4.81633544e-01 3.55280995e-01
9.99869108e-01 -3.63508493e-01 -4.97511625e-01 -1.26688027e+00
-5.23998559e-01 5.47480464e-01 7.21346617e-01 -7.98883885e-02
3.44903916e-01 -1.01081681e+00 -1.18631732e+00 -1.64730087e-01
-1.54552519e-01 -2.04804108e-01 -7.60839805e-02 9.96774137e-01
-5.64192116e-01 -2.11557612e-01 1.68801308e-01 -5.23239732e-01
-1.34419382e+00 2.73912579e-01 3.23668242e-01 -4.09450084e-01
-1.32320687e-01 5.46247005e-01 -2.17149809e-01 -3.15907896e-01
-2.59135365e-02 1.92490280e-01 -1.05008215e-01 3.26317608e-01
5.29856265e-01 5.12665510e-01 -3.82902533e-01 -4.80461210e-01
-1.30214602e-01 3.58389341e-03 -1.77455191e-02 -2.55207628e-01
1.22346485e+00 -4.81392503e-01 -4.94160712e-01 9.40989107e-02
9.85811293e-01 2.31514111e-01 -1.02270699e+00 -4.41759408e-01
6.11564666e-02 -7.51019597e-01 1.24816604e-01 -4.46952909e-01
-1.22352910e+00 3.32120240e-01 -2.20752463e-01 1.24088442e+00
7.86040843e-01 1.46580487e-01 4.19974416e-01 7.40514249e-02
3.88231307e-01 -8.71539474e-01 4.51412909e-02 -5.70577458e-02
4.33715045e-01 -1.12898338e+00 4.45891798e-01 -7.75524020e-01
-3.44815433e-01 7.22237945e-01 8.67189318e-02 -3.03030349e-02
1.25912035e+00 3.20401758e-01 -2.74202317e-01 -3.49997342e-01
-7.52039671e-01 -2.88311362e-01 -3.61406989e-02 6.50122106e-01
3.04387003e-01 2.15172634e-01 -5.32787919e-01 1.92323789e-01
-1.62171319e-01 -1.56531319e-01 1.09448552e+00 6.96675897e-01
-8.17347109e-01 -1.25035560e+00 -4.72076416e-01 1.15343463e+00
-7.35016167e-01 -2.54149973e-01 -8.39088023e-01 5.80132782e-01
-3.41681153e-01 1.49018252e+00 1.81955546e-02 -2.12343216e-01
-1.20109834e-01 -6.13451004e-01 1.21728525e-01 -5.40322602e-01
-7.31675804e-01 3.00095212e-02 4.43392098e-01 -1.77874357e-01
-6.37827039e-01 -7.12849855e-01 -9.03017640e-01 -1.38416004e+00
-8.28241527e-01 2.65455693e-01 4.39250350e-01 7.87256181e-01
3.52636188e-01 3.06039527e-02 9.82425332e-01 -4.85718817e-01
-2.95736551e-01 -8.52422416e-01 -8.76068592e-01 2.92347968e-01
3.23452264e-01 -3.61909032e-01 -7.90144861e-01 -2.91656524e-01] | [6.871538162231445, 5.338607311248779] |
18af1368-70ef-4d51-87bd-93b6880f1fb9 | face-recognition-in-the-age-of-clip-billion | 2301.07315 | null | https://arxiv.org/abs/2301.07315v1 | https://arxiv.org/pdf/2301.07315v1.pdf | Face Recognition in the age of CLIP & Billion image datasets | CLIP (Contrastive Language-Image Pre-training) models developed by OpenAI have achieved outstanding results on various image recognition and retrieval tasks, displaying strong zero-shot performance. This means that they are able to perform effectively on tasks for which they have not been explicitly trained. Inspired by the success of OpenAI CLIP, a new publicly available dataset called LAION-5B was collected which resulted in the development of open ViT-H/14, ViT-G/14 models that outperform the OpenAI L/14 model. The LAION-5B dataset also released an approximate nearest neighbor index, with a web interface for search & subset creation. In this paper, we evaluate the performance of various CLIP models as zero-shot face recognizers. Our findings show that CLIP models perform well on face recognition tasks, but increasing the size of the CLIP model does not necessarily lead to improved accuracy. Additionally, we investigate the robustness of CLIP models against data poisoning attacks by testing their performance on poisoned data. Through this analysis, we aim to understand the potential consequences and misuse of search engines built using CLIP models, which could potentially function as unintentional face recognition engines. | ['Shrey Jain', 'Aaditya Bhat'] | 2023-01-18 | null | null | null | null | ['data-poisoning'] | ['adversarial'] | [ 1.56688914e-02 -2.59570777e-01 -3.06265652e-01 1.97290182e-02
-7.48501062e-01 -5.11941195e-01 7.54802227e-01 -1.09911695e-01
-3.98331195e-01 4.69492346e-01 9.67207178e-02 -1.94338083e-01
-1.79473832e-01 -6.82899296e-01 -7.06309140e-01 -5.27871966e-01
-1.13204323e-01 3.67002964e-01 8.64269491e-03 -3.55897874e-01
3.07548583e-01 7.09875345e-01 -1.70332241e+00 5.78828394e-01
3.62521410e-01 9.72388566e-01 -2.16086894e-01 3.87178540e-01
1.18055992e-01 9.60625350e-01 -8.88798058e-01 -1.03356779e+00
2.31261194e-01 -1.14422612e-01 -6.86471462e-01 -4.39284742e-01
6.94229424e-01 -6.83235288e-01 -5.71434855e-01 9.06865239e-01
8.16326797e-01 -3.06850851e-01 6.03374183e-01 -1.51136386e+00
-9.99909997e-01 6.40673757e-01 -2.11248830e-01 3.26035112e-01
6.22885644e-01 4.86981601e-01 5.54650187e-01 -8.15721214e-01
6.88316107e-01 1.22073638e+00 8.84480178e-01 7.85102129e-01
-1.02973616e+00 -1.27006412e+00 -6.40993536e-01 5.80241382e-01
-1.82742202e+00 -8.74828756e-01 3.98007959e-01 -2.00336382e-01
1.08251476e+00 4.76800114e-01 1.58223778e-01 1.62673032e+00
9.34797525e-02 6.33811474e-01 8.57396126e-01 -2.91845858e-01
1.47557065e-01 3.10753971e-01 6.62174523e-02 6.54539466e-01
1.36658385e-01 1.86920941e-01 -9.51249063e-01 -7.60201633e-01
-1.25756012e-02 -4.98589575e-02 -3.11222583e-01 1.39981300e-01
-3.17826480e-01 1.00596035e+00 3.65710735e-01 3.47331583e-01
-3.53620082e-01 -3.02519530e-01 5.43214023e-01 2.05281377e-01
3.81669313e-01 3.69167089e-01 -1.45053333e-02 2.33738795e-01
-1.04454887e+00 1.59598321e-01 9.45644975e-01 7.18381107e-01
4.33507472e-01 -1.65158305e-02 -5.20781100e-01 9.33388472e-01
9.59016457e-02 7.52428651e-01 5.10612190e-01 -6.55537128e-01
1.05188169e-01 3.61493587e-01 -7.95043856e-02 -1.01920533e+00
-6.17650673e-02 -1.85728237e-01 -4.96127605e-01 3.67196207e-03
1.54190063e-01 3.77881020e-01 -8.86970341e-01 1.52783513e+00
-7.61062130e-02 2.83521801e-01 8.55394527e-02 3.71783316e-01
8.17703724e-01 6.78559482e-01 5.30514896e-01 -1.75706282e-01
1.38665462e+00 -5.02701700e-01 -6.97824597e-01 2.28601187e-01
5.99314272e-01 -6.50863826e-01 1.04141903e+00 5.81834733e-01
-8.07502568e-01 -2.76262373e-01 -1.06483781e+00 1.78123280e-01
-8.72183204e-01 -2.26804599e-01 4.64489698e-01 1.28663266e+00
-1.00489593e+00 4.50973213e-01 -2.48956963e-01 -4.98815596e-01
8.52917135e-01 4.22415882e-01 -5.26799738e-01 -3.86392713e-01
-1.27055311e+00 9.05055940e-01 7.67118856e-02 -3.51048857e-01
-1.46057498e+00 -7.40981758e-01 -5.48714697e-01 1.27415463e-01
2.87682384e-01 -4.80908379e-02 8.60760868e-01 -4.98313218e-01
-9.31910157e-01 9.89838779e-01 2.05694772e-02 -8.45655680e-01
1.48834363e-01 -9.80517939e-02 -7.19210327e-01 5.36977828e-01
1.31632134e-01 7.97173917e-01 9.79227841e-01 -9.47674274e-01
-1.49823576e-01 -6.75196290e-01 -5.23725599e-02 -3.68955493e-01
-1.05027246e+00 6.15857720e-01 -5.24830222e-01 -5.92150450e-01
-8.08140576e-01 -7.10131645e-01 3.57521802e-01 7.32793063e-02
-3.22617948e-01 -4.17521149e-01 1.04214180e+00 -7.34082997e-01
1.40452147e+00 -2.23164368e+00 -5.08653045e-01 1.28517389e-01
-1.39988109e-01 9.91109610e-01 -5.07707834e-01 6.13189042e-01
4.84880842e-02 5.67456841e-01 -1.82926297e-01 -1.90853164e-01
-1.50217727e-01 1.38517141e-01 -6.70543253e-01 5.20548105e-01
1.03065155e-01 7.60069251e-01 -5.24129689e-01 -4.11658287e-01
-2.66546644e-02 6.59840286e-01 -6.42673433e-01 1.01485573e-01
-2.73371428e-01 -8.93465579e-02 -2.86728710e-01 1.11768186e+00
7.37955332e-01 1.42368093e-01 -2.97201991e-01 -5.70703410e-02
2.90998328e-03 -2.40627155e-01 -3.16813529e-01 1.21152472e+00
-1.09475806e-01 5.44107378e-01 -5.50682396e-02 -4.78440285e-01
7.13061810e-01 5.55623710e-01 4.48345751e-01 -1.11297917e+00
1.44943595e-01 1.06110359e-02 -3.49319965e-01 -7.35811710e-01
2.54010648e-01 2.18430564e-01 5.58560975e-02 4.68673974e-01
2.18648404e-01 2.57051706e-01 2.13344589e-01 2.58023143e-01
1.34491336e+00 -3.16147804e-01 -1.22960523e-01 -1.93833545e-01
7.14961946e-01 7.00006932e-02 2.16522455e-01 1.05556989e+00
-3.49796116e-01 4.75091964e-01 2.93995768e-01 -5.00753462e-01
-8.22314382e-01 -8.72619092e-01 -6.22302473e-01 1.17945015e+00
-2.10085452e-01 -8.00736964e-01 -1.24428380e+00 -6.39387667e-01
-6.59550875e-02 8.21753919e-01 -7.13680804e-01 -5.40086806e-01
-5.37100136e-02 -7.21010923e-01 1.62359691e+00 3.39815542e-02
7.65271902e-01 -1.39964402e+00 -6.19543910e-01 -1.20105319e-01
-8.30995739e-02 -9.19580281e-01 -4.44959819e-01 -1.59804136e-01
-1.99221969e-01 -1.46459758e+00 -6.80736601e-01 -5.27254522e-01
2.65754998e-01 1.65267587e-01 7.34060466e-01 3.63830179e-01
-8.18764329e-01 6.62487924e-01 -5.53572297e-01 -5.30314922e-01
-4.34182882e-01 -7.44987950e-02 3.34953666e-01 2.07282268e-02
8.36489081e-01 -2.09149688e-01 -2.86110073e-01 5.34645379e-01
-1.27892566e+00 -6.40108585e-01 2.99587667e-01 6.07139349e-01
4.07094620e-02 -2.76040763e-01 6.96307540e-01 -6.96822107e-01
8.04736555e-01 -7.28232443e-01 -4.53192890e-01 7.47545958e-01
-5.98887920e-01 -7.43907839e-02 3.04382801e-01 -6.60986960e-01
-7.51767814e-01 -1.00225620e-01 -4.90820736e-01 -8.19405198e-01
-8.97526070e-02 3.15380365e-01 -1.54255375e-01 -4.37877715e-01
1.07099664e+00 3.87896240e-01 6.80922791e-02 -7.09315062e-01
9.79800969e-02 1.18151546e+00 6.33590519e-01 -3.22585076e-01
5.96997976e-01 4.03049976e-01 -3.47595960e-01 -1.30351126e+00
-5.59175551e-01 -2.63061553e-01 -1.60300419e-01 -2.19001442e-01
8.43249202e-01 -7.50351667e-01 -7.42111325e-01 7.92211115e-01
-1.09911847e+00 -8.86511728e-02 1.82544917e-01 -1.71620309e-01
-1.85618237e-01 3.96827668e-01 -6.21482491e-01 -8.95143926e-01
-7.40387976e-01 -1.14431238e+00 8.05326998e-01 7.69116655e-02
-6.56183586e-02 -4.57233220e-01 5.24956696e-02 5.78047752e-01
5.30361116e-01 -6.60191402e-02 9.28586245e-01 -1.07314813e+00
-1.52128384e-01 -2.76594877e-01 -1.42740369e-01 4.57168251e-01
-2.32934043e-01 -1.35529578e-01 -1.25713420e+00 -3.40130508e-01
1.41661376e-01 -8.08895051e-01 7.92016387e-01 -1.27059564e-01
1.24678719e+00 -7.85471201e-01 -2.99416453e-01 7.03545809e-01
1.23812127e+00 3.05123299e-01 1.05938220e+00 1.93135381e-01
2.33371466e-01 5.08665860e-01 1.56841010e-01 4.64117616e-01
-2.56025583e-01 7.28924334e-01 4.03101981e-01 5.16877174e-01
-2.81397879e-01 -4.74543393e-01 6.50163651e-01 1.27450868e-01
3.62514287e-01 -3.60304415e-01 -8.88770819e-01 3.63301724e-01
-1.20005977e+00 -1.16517985e+00 8.88875499e-02 2.14703751e+00
8.53055954e-01 -1.75842255e-01 5.06405011e-02 -3.56971920e-02
5.43641448e-01 -6.06815778e-02 -6.09098554e-01 -4.06091869e-01
-2.83730030e-01 5.95203698e-01 4.01320130e-01 1.44530684e-01
-9.80050385e-01 1.11320662e+00 7.35083008e+00 1.26831985e+00
-1.03701258e+00 3.61701280e-01 6.45941794e-01 -1.92482024e-01
8.89479592e-02 -3.72467935e-01 -1.00979042e+00 6.47721052e-01
1.38858891e+00 -1.77967221e-01 6.60980880e-01 8.67918670e-01
-3.07977825e-01 -3.63992713e-02 -9.44082141e-01 1.08439028e+00
8.60922515e-01 -1.37144613e+00 4.91547197e-01 3.20992649e-01
3.55988503e-01 1.86024398e-01 2.99612164e-01 4.72746134e-01
-6.73163608e-02 -1.24231005e+00 5.33667028e-01 3.87077332e-01
9.59542990e-01 -8.46209168e-01 4.47529584e-01 3.31760257e-01
-5.74270964e-01 -4.70602810e-01 -5.85486472e-01 3.23636264e-01
-2.63785094e-01 -1.01564288e-01 -7.50964344e-01 1.37345130e-02
9.03225899e-01 2.90884733e-01 -1.21670091e+00 1.12281907e+00
3.56338918e-02 6.88057125e-01 -4.25090343e-01 3.07431519e-02
-1.58436652e-02 2.34846011e-01 5.25058687e-01 1.03785884e+00
2.00845256e-01 9.69888344e-02 -1.56219810e-01 8.81380379e-01
-4.18026328e-01 1.23062447e-01 -1.17572927e+00 -2.68325239e-01
5.92682004e-01 9.90865409e-01 -1.83513522e-01 -2.15297222e-01
-3.67616802e-01 9.09633517e-01 1.59656152e-01 1.74455673e-01
-8.99404645e-01 -2.83339500e-01 5.88113129e-01 1.95311561e-01
4.83102202e-02 2.74578184e-01 1.50046289e-01 -8.02566707e-01
-3.12887847e-01 -1.26516974e+00 7.04651117e-01 -9.91306961e-01
-1.25838315e+00 9.16155696e-01 2.20004663e-01 -6.27445877e-01
-7.62638599e-02 -6.11574471e-01 -3.58165890e-01 5.85726738e-01
-9.08062100e-01 -1.31001306e+00 1.09799348e-01 9.40886438e-01
5.64774275e-01 -6.91786766e-01 1.18027639e+00 4.44173753e-01
-7.75410831e-01 1.11825359e+00 5.08209094e-02 1.89625174e-01
6.60266042e-01 -1.74505383e-01 -2.48808283e-02 6.57616854e-01
4.15463418e-01 8.07179451e-01 4.61743116e-01 -8.00280094e-01
-1.47467113e+00 -1.01163220e+00 7.61563838e-01 -6.42143130e-01
4.34629023e-01 -3.70865554e-01 -9.17335391e-01 3.72153759e-01
3.41916054e-01 -2.74563164e-01 8.53783607e-01 -2.97469974e-01
-8.10788095e-01 -1.93514049e-01 -1.63644886e+00 4.46311921e-01
9.14836049e-01 -9.44115996e-01 -4.40637648e-01 5.60900092e-01
4.81094003e-01 4.93617386e-01 -5.95162511e-01 3.80973876e-01
4.69296515e-01 -9.36177909e-01 1.16033292e+00 -7.12808967e-01
3.02217752e-01 2.57254750e-01 -3.33206415e-01 -5.40991306e-01
-8.83103441e-03 -4.24366444e-01 -2.28317738e-01 1.41195309e+00
3.06419492e-01 -4.25845832e-01 4.54900324e-01 7.51391053e-01
2.89460510e-01 -4.01433975e-01 -9.77895617e-01 -9.24897194e-01
-9.06825811e-02 -4.85069752e-01 5.59127927e-01 7.57170975e-01
-4.51318510e-02 -4.78885062e-02 -6.41551018e-01 3.40481028e-02
8.11167419e-01 -5.94819367e-01 5.83347082e-01 -1.00661159e+00
4.82081883e-02 -2.87109375e-01 -3.67466807e-01 -1.97262615e-01
4.13016200e-01 -9.67234850e-01 -4.51455623e-01 -6.42483234e-01
5.11315703e-01 -2.37095281e-01 -1.76387116e-01 1.07903540e+00
2.28542820e-01 9.91142929e-01 2.83406407e-01 5.27715623e-01
-5.01678526e-01 3.92461032e-01 3.57188612e-01 -4.01468396e-01
4.56494465e-02 -2.30162248e-01 -6.88455582e-01 4.38669056e-01
7.62287378e-01 -6.38693452e-01 -2.38422036e-01 -2.66283393e-01
-9.37343016e-02 -1.88848600e-01 5.14181852e-01 -1.45004523e+00
4.52523082e-01 1.67137280e-01 3.21553558e-01 -3.37843359e-01
5.47462702e-01 -6.96544230e-01 2.68676251e-01 6.21009767e-01
-3.25242430e-01 -1.66914970e-01 4.38986421e-01 2.74159759e-01
7.60518387e-02 -6.77719593e-01 6.91765904e-01 -1.07095778e-01
-6.11974418e-01 4.53452945e-01 -5.16823351e-01 -2.04539433e-01
1.13783824e+00 -5.30082658e-02 -5.75088501e-01 -2.68638223e-01
-4.82190430e-01 -3.01957093e-02 4.62821215e-01 7.27064073e-01
7.28061199e-01 -1.10451281e+00 -6.26714826e-01 5.15318692e-01
3.87717903e-01 -1.05441034e+00 9.28141996e-02 3.73262107e-01
-5.86900771e-01 7.05178678e-01 -3.78274977e-01 -2.79182553e-01
-1.48988211e+00 1.05656648e+00 1.19779609e-01 5.15521318e-03
-3.69027704e-01 6.14683092e-01 -1.57095194e-01 -6.97038546e-02
6.82168424e-01 7.04435468e-01 -2.20579281e-01 6.45970702e-02
1.23365998e+00 3.27789366e-01 1.19673714e-01 -8.14811110e-01
-4.09351796e-01 9.79064927e-02 -3.54367048e-01 -1.36478424e-01
1.39480960e+00 3.53548080e-01 -2.49049217e-01 -2.56801158e-01
1.30080032e+00 -9.15730894e-02 -6.98413253e-01 1.47051334e-01
9.26227644e-02 -4.11280185e-01 -2.64731189e-03 -1.10228622e+00
-9.47027385e-01 7.72020817e-01 1.00621855e+00 5.81015833e-02
1.06597316e+00 -9.08043608e-03 7.51945257e-01 5.61896920e-01
6.58864558e-01 -7.89146900e-01 1.24832526e-01 3.92463833e-01
1.06260836e+00 -9.50828493e-01 -1.76125988e-01 -4.05307338e-02
-4.50199634e-01 7.54284680e-01 5.92397690e-01 3.17942172e-01
6.32136106e-01 1.88077420e-01 -1.78619713e-01 -3.11656654e-01
-6.74384356e-01 -1.17018662e-01 -4.65570204e-02 1.01514018e+00
-6.07723594e-02 -2.97882408e-01 -2.15655461e-01 7.40720749e-01
-6.80849478e-02 1.46537095e-01 1.80525839e-01 8.35856497e-01
-3.13952297e-01 -1.19449747e+00 -6.64230168e-01 4.20774758e-01
-7.31607378e-01 -2.30031937e-01 -8.68169367e-01 5.32910466e-01
1.61414653e-01 1.18287599e+00 -2.27472961e-01 -6.46557927e-01
8.58020559e-02 6.12388134e-01 2.38271609e-01 -2.79155999e-01
-9.04637218e-01 -4.50271666e-01 -4.98426631e-02 -5.98060071e-01
1.06726609e-01 -4.83696222e-01 -4.97540653e-01 -5.70444942e-01
-2.47412652e-01 3.25183809e-01 7.50961900e-01 5.93915224e-01
5.48529625e-01 -1.40262559e-01 4.78905648e-01 -5.46589255e-01
-8.28611732e-01 -1.00824833e+00 -4.15974945e-01 5.04116058e-01
-4.21652645e-02 -4.55973059e-01 -3.06507349e-01 9.66549944e-03] | [12.770979881286621, 1.0708577632904053] |
18edeaa6-9e6f-4ccf-bb8d-2108610c9cdc | frustratingly-easy-transferability-estimation | 2106.09362 | null | https://arxiv.org/abs/2106.09362v4 | https://arxiv.org/pdf/2106.09362v4.pdf | Frustratingly Easy Transferability Estimation | Transferability estimation has been an essential tool in selecting a pre-trained model and the layers in it for transfer learning, to transfer, so as to maximize the performance on a target task and prevent negative transfer. Existing estimation algorithms either require intensive training on target tasks or have difficulties in evaluating the transferability between layers. To this end, we propose a simple, efficient, and effective transferability measure named TransRate. Through a single pass over examples of a target task, TransRate measures the transferability as the mutual information between features of target examples extracted by a pre-trained model and their labels. We overcome the challenge of efficient mutual information estimation by resorting to coding rate that serves as an effective alternative to entropy. From the perspective of feature representation, the resulting TransRate evaluates both completeness (whether features contain sufficient information of a target task) and compactness (whether features of each class are compact enough for good generalization) of pre-trained features. Theoretically, we have analyzed the close connection of TransRate to the performance after transfer learning. Despite its extraordinary simplicity in 10 lines of codes, TransRate performs remarkably well in extensive evaluations on 32 pre-trained models and 16 downstream tasks. | ['Junzhou Huang', 'Qiang Yang', 'Yu Rong', 'Ying WEI', 'Long-Kai Huang'] | 2021-06-17 | null | null | null | null | ['mutual-information-estimation'] | ['methodology'] | [ 4.19349343e-01 2.73524314e-01 -2.13548124e-01 -4.06116188e-01
-8.29965115e-01 -3.58861893e-01 7.15569675e-01 2.15839684e-01
-6.82506740e-01 8.82178664e-01 -3.63042504e-02 -1.22562088e-01
-1.67261094e-01 -7.05494404e-01 -8.35760772e-01 -5.90489268e-01
-3.22937757e-01 2.90078253e-01 1.09763175e-01 7.22034425e-02
1.47738442e-01 4.13941056e-01 -1.45407724e+00 3.04068923e-01
1.08629954e+00 1.26853800e+00 3.78591508e-01 4.92497534e-01
2.94282913e-01 5.73571205e-01 -5.36742151e-01 -6.05296612e-01
1.40024394e-01 -6.70533299e-01 -1.06630182e+00 -1.90702617e-01
2.13892668e-01 -7.07786754e-02 -2.66083479e-01 9.33363378e-01
3.11494648e-01 1.26878768e-01 1.41813552e+00 -1.20455909e+00
-4.11397964e-01 5.44186652e-01 -1.06161654e-01 3.21212053e-01
-7.54452646e-02 -1.26630709e-01 1.11996663e+00 -1.09445477e+00
3.89884204e-01 8.43657434e-01 6.39924228e-01 4.96237785e-01
-1.29829800e+00 -7.22619116e-01 -3.54438186e-01 1.16765276e-01
-1.28725052e+00 -6.09894156e-01 3.57793003e-01 -5.64752162e-01
8.69821727e-01 1.42557919e-01 4.29315984e-01 6.93183362e-01
2.14724585e-01 8.17374349e-01 9.82307196e-01 -5.52780867e-01
1.87332883e-01 7.13077962e-01 -9.83077958e-02 6.10874057e-01
2.74895057e-02 1.64853513e-01 -5.58834910e-01 8.71906355e-02
6.35364532e-01 -3.03978622e-01 -5.00778794e-01 -4.75889295e-01
-1.25931549e+00 8.46884072e-01 8.65885079e-01 5.17185807e-01
-3.93857360e-02 -1.42425701e-01 3.76797408e-01 6.54583037e-01
6.61700726e-01 7.21669316e-01 -5.37117362e-01 -4.31575365e-02
-8.29517186e-01 -2.51191944e-01 7.05319881e-01 9.13890362e-01
1.08589768e+00 -3.62158090e-01 -4.07566041e-01 7.39538491e-01
-1.48160150e-02 2.98937708e-01 7.81526804e-01 -6.16284609e-01
7.71991014e-01 6.38386488e-01 -3.86116326e-01 -6.62972212e-01
-1.85729474e-01 -8.56431901e-01 -1.05390286e+00 1.04572602e-01
5.29332459e-01 -1.29505903e-01 -4.18202549e-01 1.99860096e+00
-7.95646384e-02 -1.44100115e-01 3.11428875e-01 4.00888711e-01
4.10840154e-01 7.71514893e-01 2.64838133e-02 -3.26354653e-01
9.99675333e-01 -8.22152495e-01 -2.66578466e-01 -2.68298179e-01
1.23460782e+00 -7.02039719e-01 1.03909791e+00 1.31929129e-01
-1.06810904e+00 -5.95391035e-01 -1.09810805e+00 1.27546445e-01
-4.10545498e-01 5.19386113e-01 3.77007455e-01 5.94392657e-01
-1.03751540e+00 1.06575644e+00 -3.45057189e-01 -1.09834060e-01
6.60638511e-01 4.91480112e-01 -5.54996073e-01 7.03892931e-02
-1.18874753e+00 1.26302922e+00 7.51198113e-01 -3.82130831e-01
-8.00525725e-01 -8.03013742e-01 -7.38507271e-01 6.87864542e-01
-4.55170050e-02 -4.86357629e-01 1.17267275e+00 -1.10993350e+00
-1.35527062e+00 9.12719309e-01 1.42511249e-01 -5.79695463e-01
6.54775321e-01 -1.75829366e-01 2.30229292e-02 -3.53853963e-02
-9.34100449e-02 7.46552289e-01 9.69216168e-01 -8.34364116e-01
-6.87069893e-01 -2.48611003e-01 -1.37287825e-01 4.33597326e-01
-1.03240418e+00 -2.35456392e-01 -7.66398907e-02 -4.08168256e-01
-1.11600786e-01 -9.15847003e-01 1.65479586e-01 9.93973315e-02
-2.65686363e-01 -2.65290976e-01 4.53923047e-01 -4.32432413e-01
1.01857781e+00 -2.29761505e+00 2.51213610e-01 1.70087442e-01
1.77718118e-01 5.57125211e-01 -1.30842149e-01 1.50342897e-01
-2.25928456e-01 1.20910622e-01 -3.58827502e-01 -7.63782337e-02
-1.10959433e-01 -3.06532890e-01 -2.53834277e-01 4.85939980e-01
4.62382406e-01 7.89569139e-01 -9.00258064e-01 -5.21910071e-01
1.02060340e-01 4.85903442e-01 -5.85683703e-01 4.71977085e-01
2.36466780e-01 3.44009340e-01 -2.51020789e-01 -3.78962047e-02
4.62198436e-01 -3.22606713e-01 -2.69240420e-02 -1.43217683e-01
2.33449206e-01 4.30797458e-01 -6.90832794e-01 1.35886395e+00
-9.17587578e-01 8.93630266e-01 -4.50217664e-01 -1.19668877e+00
1.10646033e+00 2.60897756e-01 2.59335726e-01 -6.36490703e-01
1.21442169e-01 4.59014177e-01 2.16137812e-01 -1.04151599e-01
3.02363962e-01 -2.67732829e-01 -2.10793428e-02 6.11030102e-01
5.89205563e-01 -1.23869084e-01 1.26920626e-01 2.44064964e-02
9.96243894e-01 -4.59545180e-02 5.83608568e-01 -3.92934948e-01
5.71231067e-01 -5.07696509e-01 7.40502924e-02 5.02507985e-01
-1.65565774e-01 4.55681503e-01 5.43461919e-01 -6.86325952e-02
-1.05067396e+00 -1.05085242e+00 -3.35659117e-01 1.31102097e+00
-2.34998822e-01 -3.68525028e-01 -7.00214803e-01 -1.04588568e+00
-5.73684797e-02 7.23796427e-01 -8.19853544e-01 -7.88487911e-01
-2.38804683e-01 -5.27396321e-01 4.99654919e-01 5.26737571e-01
5.80278218e-01 -1.05561447e+00 -6.86018169e-01 -1.41920432e-01
-2.69379228e-01 -8.40186059e-01 -4.38051850e-01 5.89926898e-01
-1.12819040e+00 -8.10978711e-01 -1.02763498e+00 -8.72991562e-01
8.05672109e-01 6.53311536e-02 1.12836587e+00 -1.43698543e-01
-9.88029093e-02 -8.69789571e-02 -2.24261582e-01 -3.07330370e-01
-5.24762273e-01 5.50303161e-01 -9.16672051e-02 -1.09682269e-02
1.45109937e-01 -4.39459562e-01 -4.35967028e-01 4.17445123e-01
-6.68204606e-01 5.92214018e-02 9.95840311e-01 9.49370325e-01
1.71912104e-01 -1.96632892e-01 7.00868547e-01 -7.43199587e-01
6.66218817e-01 -5.02704740e-01 -3.13558221e-01 4.81562138e-01
-6.22717857e-01 2.98845828e-01 7.21428573e-01 -3.09416831e-01
-9.22135532e-01 6.89112768e-02 1.16671510e-01 -2.07315490e-01
2.18147874e-01 4.95174021e-01 -9.19051096e-02 1.07098006e-01
9.01623905e-01 3.56519550e-01 1.19926639e-01 -1.88898668e-01
3.26346874e-01 7.71226943e-01 1.69992596e-01 -3.50878656e-01
6.17708921e-01 -4.38580401e-02 1.59700948e-03 -5.94930172e-01
-1.07060134e+00 -4.63489175e-01 -1.02455306e+00 -1.25914648e-01
6.01578176e-01 -7.45382130e-01 -6.39060199e-01 2.77032673e-01
-1.04409266e+00 -4.37422782e-01 -4.33202118e-01 7.27346241e-01
-8.56564343e-01 8.49976018e-02 -1.57197222e-01 -7.77088642e-01
-2.90029883e-01 -9.30748224e-01 7.35298753e-01 1.00350576e-02
-2.21707880e-01 -1.30356514e+00 -7.31819794e-02 -6.57625869e-02
4.38813448e-01 -2.73196429e-01 1.07486904e+00 -1.02357507e+00
-2.14716151e-01 -2.84734488e-01 -4.61100161e-01 7.85561681e-01
2.92342424e-01 -2.66644627e-01 -1.23945808e+00 -4.82640088e-01
4.35232520e-02 -7.60404825e-01 1.28839362e+00 2.89527088e-01
1.16356289e+00 -3.61364514e-01 -3.49467009e-01 5.47284126e-01
1.23331904e+00 -3.78089361e-02 6.77454472e-01 2.26765350e-01
2.70369470e-01 7.33652115e-01 6.51125550e-01 2.88842499e-01
-1.51036054e-01 6.88754439e-01 1.22161463e-01 -1.75912119e-02
-1.69125706e-01 -2.37995461e-01 6.17019355e-01 9.35652852e-01
-6.80879056e-02 -1.48039997e-01 -8.38154733e-01 4.58930641e-01
-1.55338323e+00 -7.40855873e-01 2.71250069e-01 2.62539887e+00
1.15966773e+00 3.03636551e-01 -5.05976714e-02 2.36195207e-01
7.90332377e-01 -2.72804052e-01 -3.99359733e-01 -2.95528591e-01
1.14821851e-01 6.25264645e-02 2.97055811e-01 3.63191307e-01
-1.18120456e+00 6.68989003e-01 6.14137030e+00 1.17531598e+00
-9.36420560e-01 8.98244679e-02 9.26114917e-01 1.18360817e-01
-7.56633356e-02 -2.29900345e-01 -8.91950727e-01 4.03170735e-01
1.15875602e+00 -5.32374799e-01 1.74935758e-01 8.72249961e-01
-4.11191612e-01 2.50220876e-02 -1.71948087e+00 7.64220953e-01
-3.56472246e-02 -1.03025508e+00 3.10188867e-02 1.76584050e-01
7.20709145e-01 -1.25866488e-01 4.89792258e-01 6.12578988e-01
-9.19168890e-02 -1.11779809e+00 4.96558666e-01 2.59931505e-01
1.02186954e+00 -7.61170685e-01 7.95772493e-01 5.22801518e-01
-8.62163603e-01 -1.08303040e-01 -7.27737606e-01 1.53370784e-03
-3.55678678e-01 6.49142981e-01 -1.37479663e+00 3.44302833e-01
2.89950460e-01 7.99345136e-01 -6.32415831e-01 1.11699390e+00
-1.96213543e-01 4.71926123e-01 -1.54054031e-01 -1.22468196e-01
1.21252194e-01 1.26923574e-02 1.80934325e-01 1.36450148e+00
5.28886437e-01 -4.30163741e-01 -2.47171551e-01 7.65271664e-01
-3.90025526e-01 4.17782336e-01 -8.79604876e-01 6.38538897e-02
3.04172546e-01 1.17747569e+00 -5.23006856e-01 -3.26887310e-01
-2.06032708e-01 8.83157074e-01 6.81717634e-01 6.53921962e-02
-7.08052993e-01 -5.95925152e-01 2.84503460e-01 -2.54883431e-02
3.08775902e-01 1.33974537e-01 -3.58379990e-01 -1.04404759e+00
4.13751602e-02 -5.07121563e-01 3.11340630e-01 -4.60035831e-01
-1.14390635e+00 8.08146954e-01 -9.46157947e-02 -1.72018349e+00
-4.17347908e-01 -6.15695894e-01 -4.75836426e-01 9.40487266e-01
-1.53118885e+00 -7.81186759e-01 -9.24841985e-02 6.25439286e-01
4.05667156e-01 -4.31189567e-01 9.29729640e-01 2.08817512e-01
-3.71733397e-01 1.01964533e+00 3.21241975e-01 7.68192112e-02
7.52663314e-01 -1.23514593e+00 7.81969279e-02 4.58364218e-01
2.10459962e-01 3.57932121e-01 3.92440379e-01 -2.92058945e-01
-6.79722607e-01 -1.22781408e+00 1.09864461e+00 -3.35351378e-01
5.00284791e-01 -2.79421359e-01 -9.68658984e-01 6.04319036e-01
-4.58932333e-02 -4.55674566e-02 7.36850739e-01 2.91062474e-01
-4.00468558e-01 -1.40643924e-01 -8.81771326e-01 3.27522069e-01
8.38446975e-01 -6.08123600e-01 -6.38525724e-01 3.72784883e-01
5.04165471e-01 6.35237172e-02 -9.79596257e-01 3.73303533e-01
5.31364620e-01 -1.16586959e+00 7.55623877e-01 -5.87126434e-01
8.23228478e-01 3.98016840e-01 -9.40443352e-02 -1.62313175e+00
-3.41359496e-01 -2.67746806e-01 8.88872817e-02 1.23892808e+00
9.07585859e-01 -6.11886263e-01 5.86978734e-01 4.31806207e-01
-1.47145838e-01 -8.20637107e-01 -1.00133383e+00 -1.04157925e+00
3.69262636e-01 -8.49159136e-02 1.53609455e-01 8.75346184e-01
4.67415512e-01 6.59902275e-01 -2.46962324e-01 -4.37557429e-01
4.39306617e-01 -9.08169076e-02 5.21791101e-01 -1.49336278e+00
-1.64473504e-01 -5.48307836e-01 -4.53452885e-01 -9.65089619e-01
3.91643882e-01 -1.39535105e+00 1.69041753e-01 -1.06198311e+00
5.26589155e-01 -7.61122525e-01 -5.93944192e-01 4.81815219e-01
-4.30792496e-02 2.44193152e-01 3.00701559e-01 5.69586992e-01
-4.96500194e-01 8.04241359e-01 1.38261020e+00 -7.91479051e-02
-2.31548846e-01 3.29588681e-01 -5.32168806e-01 6.49787188e-01
9.01356459e-01 -6.37340069e-01 -5.36912978e-01 -1.72677815e-01
2.63839185e-01 -9.73488763e-03 1.34273231e-01 -1.15546727e+00
-1.47256413e-02 2.00011507e-01 5.64511240e-01 -9.79554504e-02
4.59084094e-01 -7.08198011e-01 -4.55863297e-01 8.22883904e-01
-8.89439166e-01 -3.96010906e-01 8.12404081e-02 3.81113410e-01
-3.48003626e-01 -7.24325478e-01 9.15055096e-01 5.61935082e-02
-4.61656868e-01 1.99597985e-01 -1.61759108e-01 1.35793328e-01
1.17399895e+00 -7.32189417e-02 -1.90771535e-01 -3.60262722e-01
-8.36364508e-01 -1.51692539e-01 3.24381560e-01 2.19361201e-01
6.62808955e-01 -1.36730933e+00 -8.16006482e-01 3.58546734e-01
1.97668731e-01 -3.24373454e-01 -1.38286695e-01 8.63494217e-01
-1.88164175e-01 6.11125767e-01 -5.85866094e-01 -5.97034574e-01
-1.07835972e+00 4.50627208e-01 3.18280399e-01 -5.87904930e-01
-2.47980550e-01 9.84669268e-01 6.51013494e-01 -3.09275806e-01
2.05589041e-01 -2.73515046e-01 -2.71794647e-01 3.00631732e-01
4.81042266e-01 3.65989536e-01 3.52557480e-01 -5.29650629e-01
-1.84269801e-01 3.95773917e-01 -1.54665679e-01 -1.19656213e-01
1.25342691e+00 9.04554799e-02 1.46666646e-01 4.88738120e-01
1.79422116e+00 -4.29227561e-01 -1.39772511e+00 -4.03690606e-01
-2.50809155e-02 -3.40051234e-01 -7.33041838e-02 -4.79585409e-01
-1.00079632e+00 1.30274701e+00 5.43780744e-01 3.26064110e-01
1.08047962e+00 1.01029895e-01 3.69811296e-01 5.54251313e-01
1.58635527e-01 -8.64636302e-01 3.86476576e-01 6.72026753e-01
1.02534175e+00 -1.19574225e+00 -9.42352936e-02 -5.56358635e-01
-6.91788077e-01 1.05944097e+00 5.20247459e-01 -8.44294801e-02
7.64622033e-01 -1.21488445e-01 -4.06527072e-01 2.42648020e-01
-8.61077070e-01 -1.14673458e-01 6.85064673e-01 5.79183519e-01
6.30267262e-01 -3.32138240e-02 -2.84665804e-02 2.42630944e-01
-3.49419385e-01 -1.45132225e-02 2.69357681e-01 5.58429360e-01
-6.49940729e-01 -7.65165806e-01 1.15955099e-01 7.07132399e-01
-3.43067944e-01 -1.89454570e-01 -3.23068827e-01 7.49534965e-01
-1.33845329e-01 5.80244064e-01 2.35426158e-01 -5.72334349e-01
3.96856442e-02 1.72011867e-01 6.11419082e-01 -7.83518553e-01
-5.96367121e-01 -3.59625697e-01 -1.18601426e-01 -1.77577168e-01
-2.73123801e-01 -5.11005998e-01 -7.33239353e-01 6.81438595e-02
-6.47385836e-01 2.91565746e-01 6.57143056e-01 1.06154990e+00
1.66902393e-01 3.87851089e-01 8.79137337e-01 -8.78851473e-01
-9.82639611e-01 -1.29502237e+00 -6.37495100e-01 4.23273325e-01
2.73932755e-01 -6.86624527e-01 -4.37190413e-01 1.31279334e-01] | [9.631421089172363, 3.02311635017395] |
393bde1c-2df8-4435-9a87-7656e058fd72 | generating-3d-bio-printable-patches-using | 2203.03814 | null | https://arxiv.org/abs/2203.03814v1 | https://arxiv.org/pdf/2203.03814v1.pdf | Generating 3D Bio-Printable Patches Using Wound Segmentation and Reconstruction to Treat Diabetic Foot Ulcers | We introduce AiD Regen, a novel system that generates 3D wound models combining 2D semantic segmentation with 3D reconstruction so that they can be printed via 3D bio-printers during the surgery to treat diabetic foot ulcers (DFUs). AiD Regen seamlessly binds the full pipeline, which includes RGB-D image capturing, semantic segmentation, boundary-guided point-cloud processing, 3D model reconstruction, and 3D printable G-code generation, into a single system that can be used out of the box. We developed a multi-stage data preprocessing method to handle small and unbalanced DFU image datasets. AiD Regen's human-in-the-loop machine learning interface enables clinicians to not only create 3D regenerative patches with just a few touch interactions but also customize and confirm wound boundaries. As evidenced by our experiments, our model outperforms prior wound segmentation models and our reconstruction algorithm is capable of generating 3D wound models with compelling accuracy. We further conducted a case study on a real DFU patient and demonstrated the effectiveness of AiD Regen in treating DFU wounds. | ['Taebin Lim', 'Seungyeob Han', 'Hyewon Son', 'Seunghwan Lee', 'Han Joo Chae'] | 2022-03-08 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Chae_Generating_3D_Bio-Printable_Patches_Using_Wound_Segmentation_and_Reconstruction_To_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Chae_Generating_3D_Bio-Printable_Patches_Using_Wound_Segmentation_and_Reconstruction_To_CVPR_2022_paper.pdf | cvpr-2022-1 | ['2d-semantic-segmentation'] | ['computer-vision'] | [ 5.44590592e-01 3.15377772e-01 7.48632252e-02 4.24020253e-02
-8.33278298e-01 -4.74653065e-01 -1.08591788e-01 3.36333185e-01
6.89740255e-02 4.70994711e-01 2.44421795e-01 -7.43050337e-01
1.09879172e-03 -9.38757002e-01 -5.67325771e-01 3.58692044e-03
2.99479216e-02 8.57866406e-01 4.23922896e-01 -3.19491953e-01
3.86762947e-01 6.62215829e-01 -1.30079615e+00 5.64041376e-01
1.08212483e+00 1.07998252e+00 2.66657919e-01 8.67382050e-01
-4.31002975e-01 1.05707303e-01 -2.96057433e-01 -6.15181513e-02
4.74704713e-01 -4.75516409e-01 -4.40434188e-01 2.26642981e-01
1.11286324e-02 -6.02800429e-01 2.66435534e-01 1.59399644e-01
9.33538496e-01 -6.37282610e-01 6.09268427e-01 -7.02768683e-01
-3.63875538e-01 -2.36860260e-01 -5.15048206e-01 -5.91666698e-01
6.58983111e-01 3.39863271e-01 3.53443902e-03 -8.04075480e-01
1.12075329e+00 9.90276039e-01 9.94241774e-01 8.30842614e-01
-1.52614284e+00 -3.34878236e-01 -4.52259779e-01 -5.00359058e-01
-9.46615577e-01 -1.51023939e-01 5.19909322e-01 -6.41440392e-01
7.19570816e-01 5.28648555e-01 1.65209818e+00 8.33832622e-01
6.61806166e-01 6.68149829e-01 1.55924141e+00 -6.52642071e-01
5.62666714e-01 -3.56255233e-01 -1.52921915e-01 8.91966105e-01
1.03313297e-01 2.50841647e-01 -3.75665963e-01 -2.48793408e-01
1.83458328e+00 -1.12032019e-01 -3.33341360e-01 -4.11069334e-01
-1.02212906e+00 3.99974883e-01 3.04924607e-01 -5.60294807e-01
-5.74873865e-01 3.56505036e-01 1.11630738e-01 -2.46353745e-01
6.02076888e-01 3.18049401e-01 -5.53812742e-01 -1.74857780e-01
-5.58524430e-01 3.33808154e-01 8.09630811e-01 1.25182426e+00
2.15204358e-01 -5.98068655e-01 -1.31357655e-01 8.39379549e-01
1.02817185e-01 3.09234381e-01 6.78747073e-02 -1.12386763e+00
-1.81069538e-01 8.55091333e-01 1.62273034e-01 -3.36400092e-01
-4.93717164e-01 2.52208374e-02 -4.01468486e-01 1.00890958e+00
2.98379153e-01 -3.32852930e-01 -1.68382311e+00 6.60678685e-01
6.08018875e-01 -2.55659789e-01 -2.98839957e-01 1.06733477e+00
7.72709370e-01 1.17350362e-01 9.38707814e-02 9.62484851e-02
1.28373992e+00 -4.85661507e-01 -5.57219088e-01 -2.18273923e-01
8.09380054e-01 -8.60539138e-01 1.41343486e+00 6.19497240e-01
-1.37376595e+00 2.17635676e-01 -7.56797194e-01 -2.86083758e-01
-2.96110325e-02 -2.36841559e-01 7.26972401e-01 3.20121735e-01
-1.19672918e+00 6.43399298e-01 -9.94252264e-01 -6.92363679e-01
1.01902139e+00 2.69811481e-01 -4.97822791e-01 -4.02333587e-01
-4.17909026e-01 1.09793699e+00 -1.08350672e-01 -1.61735207e-01
-1.89269692e-01 -1.06151628e+00 -5.99486589e-01 -7.42524147e-01
-4.96832319e-02 -1.60037994e+00 1.32167327e+00 -3.85109335e-01
-1.59971726e+00 1.39769757e+00 -4.59203497e-02 4.35578078e-02
1.05265427e+00 -4.38743383e-01 8.09313431e-02 2.72100836e-01
1.19811408e-01 6.62841082e-01 2.95667678e-01 -1.68124878e+00
-5.46145916e-01 -5.20870864e-01 -3.89357328e-01 5.05353585e-02
5.19474506e-01 -1.24826908e-01 -6.31196976e-01 -6.09817863e-01
6.24602497e-01 -6.79638863e-01 -6.74014270e-01 1.13356149e+00
-4.87540960e-01 4.75413293e-01 5.86086631e-01 -9.96645987e-01
9.73108709e-01 -1.68094611e+00 -2.16189653e-01 4.29056048e-01
-1.07999794e-01 9.18391868e-02 2.22527355e-01 5.26281297e-01
6.66413037e-03 4.79465514e-01 -4.60575640e-01 -2.80695319e-01
-2.60532230e-01 3.86674285e-01 3.01066995e-01 2.72649061e-03
6.77112341e-02 1.00578678e+00 -7.55400956e-01 -8.06529582e-01
3.57944518e-01 5.36915302e-01 -6.99563384e-01 6.50661439e-02
-2.25302696e-01 6.28648341e-01 -2.85028845e-01 1.23480237e+00
6.23840034e-01 -6.38444647e-02 -5.80169111e-02 -2.23485872e-01
-1.65447637e-01 -2.19965160e-01 -8.95295560e-01 2.34580636e+00
-5.86059034e-01 4.96013230e-03 4.35699463e-01 -1.20689742e-01
8.50647807e-01 1.62043080e-01 6.95661545e-01 -7.79093981e-01
1.75692454e-01 6.21930897e-01 -7.68314719e-01 -8.66970122e-01
-9.38536227e-02 -5.23576140e-01 2.42384538e-01 3.84751201e-01
-4.87138331e-01 -7.86729991e-01 -1.90668225e-01 -8.77222940e-02
1.01502287e+00 8.47256601e-01 6.62957653e-02 -3.38507921e-01
-6.56938791e-01 7.80388236e-01 1.34092674e-01 2.45767504e-01
1.86137050e-01 1.42466009e+00 4.07008201e-01 -4.77021217e-01
-1.10993683e+00 -1.30494499e+00 -4.12169367e-01 1.63690478e-01
3.89515847e-01 -1.10665031e-01 -9.09320831e-01 -3.30553591e-01
5.19522667e-01 6.80921555e-01 -8.18071485e-01 2.07530022e-01
-3.85732472e-01 -4.15901631e-01 -5.89652397e-02 5.94755113e-01
1.05040021e-01 -6.93011284e-01 -1.05345309e+00 5.09559929e-01
1.26888335e-01 -4.46762294e-01 -2.96331912e-01 2.61662304e-01
-1.02232635e+00 -1.22129142e+00 -1.00594616e+00 -1.09083414e+00
9.94536161e-01 -5.28116003e-02 7.77484775e-01 2.44835764e-01
-1.14352727e+00 3.12073171e-01 -5.90634704e-01 -9.05463934e-01
-4.28403646e-01 -5.36158025e-01 -2.71786451e-01 -7.30488539e-01
-1.40380979e-01 -9.08411562e-01 -1.29092360e+00 2.86292464e-01
-9.12020683e-01 9.84497190e-01 4.46268439e-01 4.85497892e-01
1.19510067e+00 -4.90549028e-01 3.95614684e-01 -9.70149755e-01
7.86657870e-01 -1.26463026e-01 -8.68171826e-02 5.59988432e-02
-6.07100129e-01 -5.18653274e-01 1.10528320e-01 -1.70589045e-01
-6.60544753e-01 1.05277546e-01 -2.73492724e-01 -3.13270725e-02
-2.77366012e-01 5.25195479e-01 2.56821454e-01 -9.62523073e-02
1.00080037e+00 -1.24915138e-01 7.78422058e-01 -7.06307948e-01
4.51077402e-01 8.45124006e-01 6.74788892e-01 -4.24030751e-01
2.84629345e-01 4.80620682e-01 -2.66174793e-01 -3.93271327e-01
-2.18281269e-01 -2.13014111e-01 -8.40412557e-01 -5.84168851e-01
6.80877388e-01 -5.71427464e-01 -3.65491748e-01 8.35071027e-01
-1.09338140e+00 -1.10290384e+00 -5.94194829e-01 6.29205778e-02
-8.69840026e-01 -2.15129972e-01 -8.24488461e-01 -7.00737357e-01
-6.11985326e-01 -8.73910010e-01 1.31272113e+00 1.58514917e-01
-8.02016735e-01 -6.88500404e-01 -1.30596220e-01 5.06927431e-01
5.43300271e-01 1.24628580e+00 1.27657926e+00 4.58126336e-01
-3.34359527e-01 -5.45726061e-01 -1.34873614e-01 1.39616281e-01
4.37643707e-01 1.85571849e-01 -6.11841738e-01 2.92831004e-01
-5.80818653e-01 3.37188020e-02 3.56428802e-01 7.76174247e-01
9.14025784e-01 -1.99314103e-01 -7.26739526e-01 6.31210804e-01
1.63050795e+00 2.28891924e-01 9.93171215e-01 3.62442911e-01
4.24441546e-01 5.77651918e-01 7.42129803e-01 3.97824496e-01
2.62637556e-01 2.92320818e-01 3.65550756e-01 -9.07331645e-01
-6.22084796e-01 -4.50575054e-01 -3.74967217e-01 2.71462083e-01
-4.09459531e-01 1.35814339e-01 -1.16041338e+00 4.67863798e-01
-1.51238871e+00 -1.94069669e-01 -4.80180919e-01 2.00098252e+00
1.05142200e+00 3.33942696e-02 6.22701906e-02 4.92762513e-02
3.58362556e-01 -8.80352378e-01 -6.19145811e-01 -4.82232124e-01
2.79703587e-01 1.04230416e+00 7.73178339e-01 3.85572463e-01
-5.10332942e-01 8.76442134e-01 6.65166378e+00 2.34314173e-01
-1.18807971e+00 -1.82398081e-01 5.98810196e-01 -2.15991437e-01
-6.28796875e-01 -9.23651829e-02 1.23853445e-01 2.95525640e-01
2.23027602e-01 -4.75099981e-02 6.87670261e-02 6.88546777e-01
6.08496308e-01 -7.19332874e-01 -8.60247552e-01 9.45396364e-01
-4.31230664e-01 -1.96769905e+00 2.18845129e-01 2.01521307e-01
4.66251880e-01 -3.60416502e-01 -4.18050617e-01 -5.18399000e-01
4.15463448e-01 -1.07139027e+00 9.82210338e-01 9.55466032e-01
1.56656241e+00 -6.81153834e-01 5.22904158e-01 3.35179031e-01
-3.91966671e-01 1.59661964e-01 3.37237984e-01 -1.59050170e-02
7.24453211e-01 7.52934575e-01 -1.17795527e+00 3.84683311e-01
7.96723783e-01 2.11559176e-01 -1.68467417e-01 1.47218287e+00
1.55369103e-01 1.06505357e-01 -4.73844290e-01 2.45055735e-01
-4.19788301e-01 -3.86080235e-01 1.31341636e-01 9.62094665e-01
5.77099025e-01 7.12840736e-01 -3.90033200e-02 7.81080484e-01
5.64797044e-01 2.85313189e-01 -3.54154557e-01 1.88253060e-01
2.85566181e-01 7.71965265e-01 -9.32452679e-01 5.14239557e-02
1.82720095e-01 1.39945233e+00 -2.31798127e-01 1.63764045e-01
-3.31721455e-01 -3.21109712e-01 6.26737535e-01 1.03143811e+00
-1.73954353e-01 -4.66368884e-01 -1.22021019e+00 -2.82963157e-01
4.35073487e-02 -5.65546453e-01 -1.15475357e-01 -1.41375148e+00
-1.18312705e+00 1.01199351e-01 -4.41238970e-01 -1.11687231e+00
2.90021986e-01 -7.81338811e-01 -4.52690005e-01 1.12121654e+00
-1.25493610e+00 -1.56264520e+00 -7.22065151e-01 4.10565376e-01
3.36392134e-01 6.84060872e-01 1.52311194e+00 2.24040751e-03
-2.74390340e-01 3.83072160e-02 -2.45818645e-01 -1.49937853e-01
6.08971298e-01 -1.10815954e+00 5.07230997e-01 2.77465761e-01
-9.67762828e-01 3.01945239e-01 4.95276868e-01 -1.35233879e+00
-1.58908021e+00 -9.93537962e-01 2.17262819e-01 -5.50984204e-01
5.79184145e-02 -3.08859617e-01 -4.64935333e-01 3.26349020e-01
-2.16882691e-01 -5.54521829e-02 9.71680999e-01 -4.61132437e-01
2.13920280e-01 3.18253756e-01 -1.77755892e+00 8.48785520e-01
1.64918458e+00 -1.39995664e-01 -5.25575131e-02 5.27703702e-01
5.07659495e-01 -1.14403212e+00 -1.18689847e+00 5.84093809e-01
1.04586160e+00 -7.78631806e-01 7.96311080e-01 -4.19056505e-01
9.59766209e-01 -2.03079700e-01 9.54846963e-02 -1.29571199e+00
-1.77317306e-01 -7.42789209e-01 1.29230529e-01 5.63494146e-01
5.57753623e-01 -6.99066699e-01 8.69798362e-01 1.15430164e+00
-7.03933358e-01 -1.27433968e+00 -7.12901771e-01 -1.83065966e-01
-2.61000842e-02 -4.74717855e-01 3.73143226e-01 6.01217449e-01
4.65148121e-01 -4.97962505e-01 3.82003427e-01 1.08636115e-02
6.34700298e-01 1.49402753e-01 6.76108181e-01 -1.36334538e+00
1.13497242e-01 -2.54721314e-01 -3.49186659e-01 -7.15132058e-01
-8.69679511e-01 -8.51936877e-01 8.96192715e-02 -2.79940009e+00
-2.91054457e-01 -1.17339814e+00 5.14731884e-01 7.36613512e-01
1.86511986e-02 2.45444417e-01 -2.21593916e-01 2.17708424e-01
5.61916113e-01 -1.15256444e-01 1.91081989e+00 3.31267327e-01
-6.08945966e-01 -7.26435184e-02 -1.01419544e+00 3.95018131e-01
6.71251535e-01 7.88718238e-02 -2.86575377e-01 -4.60356772e-01
-1.73698604e-01 1.98784932e-01 5.14081240e-01 -1.00846612e+00
-5.29160425e-02 -2.65772909e-01 8.84168029e-01 -4.47146088e-01
1.49738148e-01 -8.07413042e-01 5.91949880e-01 5.88234425e-01
1.94817185e-01 -5.13907075e-01 5.32599747e-01 1.73265293e-01
4.67985392e-01 5.42302072e-01 7.72944748e-01 -5.60100734e-01
-4.53996837e-01 3.54717165e-01 -3.62989426e-01 -3.53132337e-01
1.46832418e+00 -1.20224941e+00 -3.12781483e-01 -3.68030858e-03
-1.13530242e+00 2.68446684e-01 1.21621156e+00 1.36274070e-01
9.81723428e-01 -1.23225474e+00 -7.05406904e-01 3.76382142e-01
9.30648521e-02 9.29668188e-01 4.22971338e-01 6.70413017e-01
-1.69840038e+00 -1.48777440e-01 -4.52230960e-01 -6.52745068e-01
-9.55042720e-01 1.45689443e-01 4.84426409e-01 3.60569030e-01
-1.25433755e+00 8.79946291e-01 -3.32300693e-01 -4.22544688e-01
-1.26072466e-01 -5.87487578e-01 5.09553134e-01 -2.39012420e-01
3.13713312e-01 2.05028832e-01 3.34456772e-01 1.34918228e-01
-2.67264307e-01 9.31331277e-01 3.27429503e-01 -2.12301344e-01
1.58299363e+00 -9.01158061e-03 1.03650324e-01 6.03937507e-02
8.52534056e-01 -3.20864946e-01 -1.55268753e+00 4.01309580e-01
-4.64628398e-01 -5.92235565e-01 1.26843750e-01 -1.61501586e+00
-7.01924086e-01 4.56958264e-01 8.07306826e-01 -1.93845674e-01
1.16895783e+00 -5.39924903e-03 1.40032601e+00 -7.76813507e-01
1.02709496e+00 -1.31760740e+00 -1.43385202e-01 -2.37064451e-01
1.36400354e+00 -5.93220592e-01 -2.66449489e-02 -1.03568256e+00
-5.05245090e-01 1.19983351e+00 1.65900737e-01 -1.99920833e-01
7.59386301e-01 7.70504534e-01 6.98996842e-01 -4.44138974e-01
-1.44365430e-01 4.84280623e-02 -1.90933853e-01 1.12316048e+00
1.83494359e-01 2.23832726e-01 -5.36311209e-01 3.99918586e-01
-3.05455327e-01 8.39731574e-01 3.66455257e-01 1.66730964e+00
-2.88241118e-01 -1.10534608e+00 -2.43838862e-01 7.37958193e-01
4.45890985e-02 5.44837750e-02 -4.27266866e-01 8.42110574e-01
2.08337590e-01 5.19206762e-01 3.85317132e-02 -3.09178561e-01
1.00828946e+00 1.67407304e-01 6.38937771e-01 -8.39798450e-01
-4.47815299e-01 2.13735774e-01 3.49510759e-01 -8.42056334e-01
1.65779099e-01 -3.83887291e-01 -1.55850267e+00 -8.27372149e-02
1.15769142e-02 -6.95568204e-01 1.16733706e+00 5.18893361e-01
8.30218434e-01 5.26432633e-01 3.26746285e-01 -1.02353168e+00
2.50042528e-01 -5.46414554e-01 -8.71970534e-01 2.27299213e-01
1.13647342e-01 -3.45490724e-01 1.04009621e-01 4.84467655e-01] | [12.898773193359375, -2.813842296600342] |
88fd2c82-f7a2-4817-b7cc-2281dbcb05c1 | sketchbetween-video-to-video-synthesis-for | 2209.00185 | null | https://arxiv.org/abs/2209.00185v1 | https://arxiv.org/pdf/2209.00185v1.pdf | SketchBetween: Video-to-Video Synthesis for Sprite Animation via Sketches | 2D animation is a common factor in game development, used for characters, effects and background art. It involves work that takes both skill and time, but parts of which are repetitive and tedious. Automated animation approaches exist, but are designed without animators in mind. The focus is heavily on real-life video, which follows strict laws of how objects move, and does not account for the stylistic movement often present in 2D animation. We propose a problem formulation that more closely adheres to the standard workflow of animation. We also demonstrate a model, SketchBetween, which learns to map between keyframes and sketched in-betweens to rendered sprite animations. We demonstrate that our problem formulation provides the required information for the task and that our model outperforms an existing method. | ['Matthew Guzdial', 'Dagmar Lukka Loftsdóttir'] | 2022-09-01 | null | null | null | null | ['video-to-video-synthesis'] | ['computer-vision'] | [ 5.44775352e-02 1.10590737e-02 5.04120179e-02 2.47190714e-01
-1.47530232e-02 -6.91325963e-01 8.70866418e-01 -3.11786830e-01
-3.87302190e-02 3.13812464e-01 1.68768261e-02 -3.22881460e-01
1.10654451e-01 -7.55644441e-01 -6.94283247e-01 -1.94553420e-01
-2.85787612e-01 4.91790295e-01 4.99104202e-01 -5.26089668e-01
1.65141299e-01 7.04691529e-01 -1.42181528e+00 1.43546611e-01
2.58151621e-01 3.44485104e-01 1.31775647e-01 1.29383636e+00
-3.81917238e-01 1.24260640e+00 -7.03447938e-01 -5.03252625e-01
3.27034444e-01 -7.02019095e-01 -7.78588116e-01 1.65722296e-01
6.69607818e-01 -5.10231853e-01 -3.81562948e-01 6.84951603e-01
3.42328280e-01 1.92100421e-01 7.35238552e-01 -1.47712457e+00
-4.23737228e-01 3.59279454e-01 -6.23247385e-01 3.88777517e-02
6.64932907e-01 3.89782280e-01 8.14376831e-01 -5.18485427e-01
8.72904420e-01 1.31896722e+00 8.68113875e-01 9.50960338e-01
-1.15302253e+00 -6.25399470e-01 1.44429848e-01 -2.81392902e-01
-1.09977460e+00 -1.26890942e-01 9.13526654e-01 -8.29574227e-01
7.42147207e-01 4.01177496e-01 1.45589709e+00 1.04867697e+00
5.22542819e-02 8.91762912e-01 8.60111415e-01 -5.54217160e-01
7.75302649e-02 -3.30602616e-01 -2.81181216e-01 8.27932000e-01
-2.35499829e-01 8.60185772e-02 -5.01255214e-01 -6.48830906e-02
1.66432798e+00 -2.17744067e-01 -1.27441734e-01 -5.73137105e-01
-1.09776008e+00 5.38385749e-01 -2.16368645e-01 3.31006318e-01
-2.05457598e-01 8.58389854e-01 2.12869123e-01 2.01606303e-01
3.06743085e-01 5.37294567e-01 -1.60557479e-01 -7.26200461e-01
-1.19039249e+00 8.73880208e-01 9.99735355e-01 1.08320594e+00
5.81499219e-01 2.40269423e-01 1.74402669e-01 4.20442969e-01
9.53787342e-02 3.24125469e-01 1.53566986e-01 -1.51687860e+00
-8.30543488e-02 5.29170156e-01 2.45817930e-01 -1.19669950e+00
-2.97930896e-01 1.69833973e-01 -5.91771841e-01 8.72062802e-01
6.30308568e-01 3.45462561e-02 -6.07251644e-01 1.64515758e+00
3.94985616e-01 8.44043732e-01 -5.59066594e-01 7.59700954e-01
9.26002502e-01 5.22431791e-01 8.88388306e-02 -9.73385759e-03
1.19337726e+00 -9.97836471e-01 -7.75941551e-01 -8.54916349e-02
4.44585413e-01 -8.22514832e-01 1.57156551e+00 4.05480295e-01
-1.65768361e+00 -4.47427362e-01 -9.42169785e-01 -6.07557818e-02
-1.20006897e-01 -3.70367140e-01 8.45023751e-01 6.62460089e-01
-1.16244996e+00 1.08895433e+00 -8.82127762e-01 -1.40337676e-01
3.43462825e-01 2.71205187e-01 -3.19339842e-01 4.38183784e-01
-8.11371148e-01 1.04476023e+00 -1.84852369e-02 -2.82722384e-01
-5.80925584e-01 -1.05776179e+00 -7.64884174e-01 -1.13130331e-01
4.56714869e-01 -8.35890949e-01 1.79154325e+00 -9.90043342e-01
-2.07621527e+00 1.00796223e+00 2.81716645e-01 -5.13475053e-02
9.04098868e-01 -4.54296112e-01 8.42985138e-02 1.14177071e-01
-2.94541895e-01 7.47778654e-01 9.40818310e-01 -1.29793000e+00
-4.73747790e-01 2.32880637e-01 4.47022051e-01 1.97168618e-01
4.03591543e-02 2.21110821e-01 -7.48100281e-01 -9.07089829e-01
-4.27991629e-01 -1.09490085e+00 -3.06844771e-01 2.51659453e-01
-1.78778604e-01 -2.55766064e-01 1.05232632e+00 -5.43486118e-01
1.63295555e+00 -1.72992015e+00 4.22157705e-01 1.49066718e-02
4.67568308e-01 1.46800220e-01 1.25577986e-01 6.28050685e-01
-3.48669291e-02 2.72863299e-01 -3.38300355e-02 -3.29703629e-01
3.19242537e-01 1.36419386e-01 -1.89390004e-01 1.04855224e-01
1.09943435e-01 9.48104501e-01 -1.10443735e+00 -5.34917235e-01
2.62027830e-01 4.38820183e-01 -6.24371231e-01 4.08900321e-01
-4.70457256e-01 4.95958090e-01 -2.86524236e-01 4.28433806e-01
3.38262677e-01 -2.84954250e-01 2.49096900e-01 2.75102798e-02
-3.29775095e-01 8.93511400e-02 -1.32524657e+00 1.95359302e+00
-3.43201637e-01 7.66033947e-01 1.39585853e-01 -5.41955352e-01
5.04113078e-01 3.39690745e-01 5.44683039e-01 -3.26455206e-01
9.66784824e-03 4.20618802e-02 -3.79480794e-02 -6.67270064e-01
5.36587059e-01 -1.47906691e-01 -8.56230706e-02 9.66616988e-01
-2.89984554e-01 -9.71123695e-01 2.40602821e-01 4.25991923e-01
1.20021486e+00 8.41861725e-01 5.72411299e-01 -1.76846273e-02
8.46737623e-02 1.74445868e-01 1.15070716e-01 6.87829971e-01
1.24820814e-01 5.01597524e-01 8.24765384e-01 -7.31086433e-01
-1.35706711e+00 -8.29127133e-01 5.73131084e-01 1.04383445e+00
1.50890857e-01 -8.14932704e-01 -9.00983036e-01 -6.97739899e-01
-1.48789853e-01 3.64256799e-01 -7.50642359e-01 3.08158159e-01
-1.18611073e+00 1.96043104e-01 6.09531999e-01 5.56025803e-01
-6.70343190e-02 -1.29698491e+00 -1.04152918e+00 1.67107955e-01
3.85605007e-01 -9.88647819e-01 -7.84772992e-01 -4.27022517e-01
-5.98382413e-01 -1.27364647e+00 -7.90378690e-01 -6.36517107e-01
4.02263492e-01 2.34353393e-01 1.71609271e+00 6.34571731e-01
-4.07337993e-01 9.54689622e-01 -1.91157669e-01 -5.02840877e-01
-7.83230782e-01 -1.94246173e-01 -1.27628893e-01 -5.08699775e-01
-5.14885262e-02 -9.72075105e-01 -4.86997515e-01 1.66604877e-01
-8.60427260e-01 7.67487288e-01 2.35368535e-01 5.97989023e-01
3.21337938e-01 -2.27548793e-01 -2.08939184e-02 -1.00346231e+00
7.85502255e-01 -3.71754877e-02 -5.89546800e-01 -5.64247966e-02
-8.10611248e-02 -2.12056190e-01 5.75070500e-01 -9.24262524e-01
-6.67908669e-01 1.34140834e-01 -8.48692656e-02 -8.59973729e-01
-2.29866266e-01 1.39263332e-01 4.00176756e-02 -3.90121728e-01
4.78304058e-01 -1.16296433e-01 1.81355402e-01 -4.41480160e-01
5.01318872e-01 4.12844196e-02 6.17444873e-01 -8.11986566e-01
1.05667043e+00 4.39426839e-01 2.71457583e-01 -9.38891411e-01
-2.63547003e-01 3.54838893e-02 -8.21922123e-01 -7.14957833e-01
6.41874135e-01 -5.72266459e-01 -1.09709835e+00 3.87236029e-01
-1.32478487e+00 -1.03105628e+00 -5.10883689e-01 2.27793887e-01
-9.68996406e-01 3.79872948e-01 -7.75704384e-01 -7.62598872e-01
5.31592481e-02 -1.04459524e+00 1.03029060e+00 2.41101027e-01
-7.97655463e-01 -1.22306764e+00 4.41836119e-01 -1.30306065e-01
2.76037186e-01 6.88572049e-01 8.78587723e-01 -5.03913276e-02
-6.41986191e-01 -5.91430143e-02 2.75182813e-01 -7.65260383e-02
2.00058192e-01 6.00700617e-01 -5.55396736e-01 1.14347480e-01
-4.90843862e-01 -2.96116352e-01 1.59960926e-01 3.18568289e-01
1.05396020e+00 -3.78415793e-01 -1.40288040e-01 4.41744268e-01
9.48094487e-01 1.95788756e-01 5.67314744e-01 3.62711012e-01
1.03942621e+00 7.50987828e-01 3.49537343e-01 4.53417450e-01
4.91474211e-01 9.64532077e-01 3.29562873e-01 -1.91835091e-01
-2.64471531e-01 -5.00331759e-01 1.24455854e-01 9.00912285e-01
-6.34696007e-01 3.66168246e-02 -1.01695156e+00 3.52602094e-01
-1.95648289e+00 -1.30585265e+00 -1.88814342e-01 1.79669952e+00
9.69300091e-01 1.77258208e-01 6.97446942e-01 2.37180561e-01
3.86347860e-01 9.50766057e-02 -2.81185567e-01 -6.23114467e-01
3.57630789e-01 6.41500771e-01 -2.32823920e-02 5.46318114e-01
-9.83425140e-01 1.24291348e+00 7.52915478e+00 7.38364339e-01
-9.67634022e-01 -2.76980132e-01 8.12005438e-03 -4.90534529e-02
-4.61358696e-01 3.10759302e-02 -2.02749684e-01 3.37619603e-01
4.31494415e-01 -3.17554682e-01 6.14831805e-01 7.33333409e-01
4.33483839e-01 7.12465793e-02 -1.46234751e+00 1.01281047e+00
-8.63289088e-02 -1.67215276e+00 2.34480985e-02 -5.42957187e-02
6.12500846e-01 -9.71169949e-01 1.17672846e-01 1.50146291e-01
7.67497718e-01 -1.27635729e+00 9.59622860e-01 3.46223056e-01
8.23975921e-01 -6.06930077e-01 -9.43120494e-02 4.02280420e-01
-1.32381701e+00 5.45103729e-01 1.59557432e-01 -4.42787737e-01
2.03593522e-01 -2.51534224e-01 -7.44365096e-01 2.03821193e-02
5.03335357e-01 7.37411857e-01 -1.92850351e-01 9.00231183e-01
-4.81472015e-01 4.94341850e-01 -1.43660277e-01 -8.90754312e-02
2.46198922e-01 -3.18104535e-01 5.35122871e-01 1.45286369e+00
3.46149057e-01 3.76534522e-01 2.51814842e-01 7.27294564e-01
8.02397579e-02 9.83584896e-02 -9.40534949e-01 -3.01516116e-01
4.46330607e-01 1.05008578e+00 -8.56326878e-01 -2.91248083e-01
-4.99954462e-01 9.80515778e-01 1.10982880e-01 1.21728733e-01
-8.91927898e-01 -7.99828619e-02 9.10312474e-01 5.88911653e-01
2.89409310e-01 -6.23027503e-01 -1.38055056e-01 -9.23171818e-01
-3.33861768e-01 -1.13412178e+00 -5.10165840e-03 -8.62876534e-01
-1.02594876e+00 2.60727972e-01 5.87678775e-02 -1.42503023e+00
-6.39893651e-01 -5.77847362e-01 -9.30428863e-01 5.82412064e-01
-1.14141738e+00 -1.44440627e+00 -4.97638136e-01 5.47455788e-01
7.58965433e-01 1.09067462e-01 9.13493514e-01 2.14430287e-01
-3.88145447e-01 3.97156060e-01 -4.77525800e-01 1.64827898e-01
5.60409844e-01 -1.42161322e+00 9.17107105e-01 5.23895919e-01
2.01996848e-01 4.66111958e-01 1.06049371e+00 -6.74831152e-01
-1.58052933e+00 -4.83206332e-01 5.29554069e-01 -6.44796014e-01
9.51603293e-01 -2.90610403e-01 -8.68340313e-01 7.02011704e-01
3.53633106e-01 -1.75448313e-01 5.61118305e-01 -1.88093752e-01
-1.64977938e-01 6.02613449e-01 -6.41761899e-01 1.06349921e+00
1.35246682e+00 -4.55398381e-01 -5.15359938e-01 1.53963968e-01
6.03695333e-01 -1.07326412e+00 -6.72953546e-01 -1.72692567e-01
1.02490997e+00 -9.48978245e-01 1.24234581e+00 -1.05995095e+00
8.06139648e-01 -2.85458475e-01 3.21785599e-01 -1.22054255e+00
-3.82407337e-01 -1.29320478e+00 -5.87990165e-01 9.35882092e-01
-8.66841152e-02 2.52497047e-01 1.12325919e+00 7.47900784e-01
-5.42745814e-02 -5.69299579e-01 -4.49508131e-01 -8.79243076e-01
2.09899157e-01 -5.54183125e-01 6.45348191e-01 1.21803761e+00
-4.30408418e-02 1.89215511e-01 -6.20779037e-01 -7.34328181e-02
5.24228275e-01 1.56374305e-01 1.45879912e+00 -1.38655984e+00
-5.04873931e-01 -9.39079642e-01 -4.03963089e-01 -1.09608817e+00
3.64862204e-01 -4.42466080e-01 -2.20321462e-01 -1.38998389e+00
-8.48618373e-02 -3.74704719e-01 4.94309664e-01 3.43185157e-01
-2.85446256e-01 2.46724993e-01 5.83180547e-01 2.89377004e-01
-4.61967081e-01 7.71690980e-02 1.72989655e+00 -4.41600457e-02
-4.90875900e-01 7.85270408e-02 -3.68203074e-01 1.20641541e+00
5.09173751e-01 -3.53164077e-01 -5.21004558e-01 -3.81829023e-01
3.84043068e-01 1.84026882e-01 4.63115990e-01 -8.12555969e-01
2.55685270e-01 -5.61251998e-01 2.59282533e-02 -2.44255111e-01
4.73942846e-01 -8.10721040e-01 5.37934422e-01 3.97957385e-01
-3.60700846e-01 4.79285628e-01 1.80396587e-01 2.66805708e-01
9.53841731e-02 -2.32550234e-01 6.30385578e-01 -4.55945313e-01
-6.35320485e-01 3.72744232e-01 -3.66455615e-01 2.01797575e-01
1.06013775e+00 -5.61136901e-01 -4.29951400e-03 -7.18786955e-01
-7.97213554e-01 -7.57470280e-02 8.97410989e-01 2.53871500e-01
5.04667997e-01 -1.60334659e+00 -6.33222818e-01 -2.23738983e-01
-4.55424249e-01 1.38721289e-02 9.73066688e-02 4.41495717e-01
-1.23879623e+00 -2.28103116e-01 -4.37528133e-01 -4.16904449e-01
-1.47738671e+00 5.17203987e-01 1.28590971e-01 -2.97386587e-01
-1.01443231e+00 6.96453035e-01 2.33509585e-01 -7.11053014e-02
2.11539596e-01 -2.37936974e-01 -2.96227992e-01 -9.60973278e-02
6.54005349e-01 4.91476923e-01 -3.80338311e-01 -3.51916939e-01
-7.50519559e-02 8.44994187e-01 1.67565644e-01 -2.64569789e-01
1.38566041e+00 1.99571744e-01 1.32542580e-01 6.41131520e-01
5.25549054e-01 2.18063980e-01 -1.62752223e+00 1.04674004e-01
-3.83389592e-02 -7.53118873e-01 -2.67701745e-01 -2.23784089e-01
-8.22366536e-01 1.01078260e+00 -8.04364681e-05 5.96226633e-01
7.61031330e-01 -1.70340210e-01 8.71979713e-01 1.86176926e-01
3.40513527e-01 -1.00571811e+00 5.91897368e-01 5.39234042e-01
8.79113376e-01 -8.15473735e-01 1.29919380e-01 -4.71867651e-01
-6.88314140e-01 1.28707945e+00 8.06683898e-01 -3.77296358e-01
4.73306268e-01 9.38869953e-01 1.55467857e-02 -3.06166202e-01
-7.86138535e-01 -3.15817185e-02 7.72606879e-02 7.98235595e-01
4.81202215e-01 -1.21276416e-01 -2.35203430e-01 6.13762736e-01
-3.69832128e-01 1.66368365e-01 6.36699796e-01 1.27518368e+00
-2.01331541e-01 -1.20731592e+00 -2.85406202e-01 -5.89977019e-02
-3.56768370e-01 7.07605854e-02 -5.86935639e-01 1.28642488e+00
6.88547343e-02 3.41444254e-01 2.19943672e-01 -5.02251685e-01
5.49744785e-01 -3.31915915e-04 1.04197860e+00 -5.81155598e-01
-9.10176873e-01 1.26646727e-01 -2.05003154e-02 -5.67768395e-01
-5.17675757e-01 -5.19359231e-01 -1.14548314e+00 -7.95969605e-01
5.03175333e-02 7.04277540e-03 4.18351889e-01 6.91451311e-01
7.87735656e-02 5.92103541e-01 4.39730167e-01 -1.55035901e+00
-9.05907433e-03 -5.08547306e-01 -4.97940063e-01 5.00861526e-01
1.74238652e-01 -7.43313611e-01 -1.17390677e-01 4.89974022e-01] | [11.021347045898438, -0.5605491995811462] |
7a82e064-5ce0-4c6e-9fec-56bf9d57947d | learning-semantic-representations-for-1 | null | null | https://aclanthology.org/D15-1164 | https://aclanthology.org/D15-1164.pdf | Learning Semantic Representations for Nonterminals in Hierarchical Phrase-Based Translation | null | ['Xing Wang', 'Deyi Xiong', 'Min Zhang'] | 2015-09-01 | null | null | null | emnlp-2015-9 | ['learning-semantic-representations'] | ['methodology'] | [-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] | [-7.266462802886963, 3.744016170501709] |
2ba86980-6b6d-474b-a70e-110eae160dc2 | vararray-meets-t-sot-advancing-the-state-of | 2209.04974 | null | https://arxiv.org/abs/2209.04974v2 | https://arxiv.org/pdf/2209.04974v2.pdf | VarArray Meets t-SOT: Advancing the State of the Art of Streaming Distant Conversational Speech Recognition | This paper presents a novel streaming automatic speech recognition (ASR) framework for multi-talker overlapping speech captured by a distant microphone array with an arbitrary geometry. Our framework, named t-SOT-VA, capitalizes on independently developed two recent technologies; array-geometry-agnostic continuous speech separation, or VarArray, and streaming multi-talker ASR based on token-level serialized output training (t-SOT). To combine the best of both technologies, we newly design a t-SOT-based ASR model that generates a serialized multi-talker transcription based on two separated speech signals from VarArray. We also propose a pre-training scheme for such an ASR model where we simulate VarArray's output signals based on monaural single-talker ASR training data. Conversation transcription experiments using the AMI meeting corpus show that the system based on the proposed framework significantly outperforms conventional ones. Our system achieves the state-of-the-art word error rates of 13.7% and 15.5% for the AMI development and evaluation sets, respectively, in the multiple-distant-microphone setting while retaining the streaming inference capability. | ['Takuya Yoshioka', 'Jinyu Li', 'Zhuo Chen', 'Xiaofei Wang', 'Jian Wu', 'Naoyuki Kanda'] | 2022-09-12 | null | null | null | null | ['speech-separation'] | ['speech'] | [ 3.37931007e-01 -2.51264274e-01 6.69643521e-01 -4.35042948e-01
-1.84404624e+00 -5.89423954e-01 1.76829129e-01 -2.83309937e-01
-6.48256838e-02 2.18478039e-01 4.12778765e-01 -5.47523677e-01
1.96101721e-02 -7.77659267e-02 -6.10872149e-01 -1.00434899e+00
9.30522084e-02 2.13679776e-01 4.83165830e-02 -1.94284841e-01
-2.24436477e-01 1.76842809e-01 -1.76660514e+00 6.20833397e-01
6.40084267e-01 1.13889992e+00 4.98891860e-01 1.38873374e+00
-3.18355784e-02 5.34600556e-01 -1.11513889e+00 -1.64291725e-01
2.65796781e-01 -4.85372275e-01 -9.49393660e-02 -9.85645801e-02
3.44767928e-01 -6.32563010e-02 -3.00858229e-01 5.34543514e-01
1.33340693e+00 1.27611369e-01 3.57325882e-01 -5.82898378e-01
-1.39507368e-01 1.01272154e+00 -1.86057568e-01 4.96231765e-01
4.75458652e-01 -6.60259351e-02 9.51585114e-01 -1.23727238e+00
-2.94183135e-01 1.13070130e+00 7.00872421e-01 5.93551159e-01
-9.67551827e-01 -7.61858523e-01 -7.50040933e-02 -4.24649343e-02
-1.83883679e+00 -1.19314551e+00 9.18226004e-01 -1.90037459e-01
1.29509270e+00 8.77685905e-01 3.54071528e-01 1.09509611e+00
-2.36180052e-01 8.67784858e-01 8.31623375e-01 -4.57678497e-01
4.43828970e-01 -1.86293826e-01 1.41803250e-01 1.67974129e-01
-4.36185986e-01 -1.38847753e-01 -8.85296226e-01 -2.88632154e-01
3.85072798e-01 -4.95015651e-01 -5.32503724e-01 5.85177243e-01
-1.11923134e+00 3.93089652e-01 -3.67531955e-01 3.81169498e-01
-3.75669241e-01 -6.49020225e-02 3.85496080e-01 3.97386342e-01
6.45825088e-01 -5.39233685e-02 -3.73858482e-01 -3.29055101e-01
-1.24717319e+00 -2.72253565e-02 8.74804080e-01 1.14831746e+00
8.86883959e-02 7.79740214e-01 -2.34606132e-01 1.47098649e+00
3.86637509e-01 1.05974054e+00 8.13140392e-01 -6.34042084e-01
7.79422343e-01 -4.04076278e-01 -3.44390720e-02 -3.62289429e-01
-2.24298257e-02 -8.75560939e-01 -8.54901254e-01 -2.83939838e-01
-1.56872913e-01 -4.78789091e-01 -8.03056240e-01 1.51409173e+00
3.22986990e-01 6.56170726e-01 6.44662321e-01 8.95599365e-01
9.84798968e-01 1.10733807e+00 -6.37125731e-01 -7.34830737e-01
1.17853963e+00 -1.13982022e+00 -1.15571761e+00 -1.44699767e-01
3.14633518e-01 -1.02569032e+00 8.59376907e-01 8.37268233e-01
-1.45469344e+00 -7.01859474e-01 -9.49897587e-01 3.87679696e-01
2.33015388e-01 2.52916515e-01 -7.18677565e-02 1.21170187e+00
-1.19992483e+00 -1.44268647e-01 -6.37353778e-01 1.08227752e-01
-1.65439025e-01 9.64324623e-02 3.24890320e-03 2.27291137e-01
-1.06280935e+00 1.21413112e-01 -2.70649642e-01 4.85680938e-01
-1.03548133e+00 -5.96764386e-01 -8.01769853e-01 1.67840004e-01
1.87068760e-01 -6.20227993e-01 1.71398568e+00 -8.42400372e-01
-2.22290993e+00 4.85154063e-01 -8.63878429e-01 -4.81533319e-01
2.29955912e-01 -3.53908062e-01 -9.92846966e-01 1.55791119e-01
-2.39212811e-01 -3.41440856e-01 1.05059958e+00 -1.31157863e+00
-3.74729961e-01 -3.25619578e-01 -6.81460798e-01 4.58965898e-01
-1.98786572e-01 3.55134755e-01 -9.52774584e-02 -9.69583690e-01
3.42100471e-01 -5.92297912e-01 -9.72977653e-02 -9.09590185e-01
-4.86297756e-01 7.48672243e-03 5.84515095e-01 -9.56604183e-01
1.54010987e+00 -2.47046471e+00 2.50528064e-02 1.19570486e-01
-2.60935336e-01 5.99481583e-01 -9.00536180e-02 4.59581196e-01
-2.23258719e-01 -2.75781423e-01 -2.76955247e-01 -8.60903144e-01
-5.99984676e-02 -1.04513682e-01 -6.06100559e-01 1.82808429e-01
-1.26481146e-01 2.02587903e-01 -5.68154752e-01 -2.30459988e-01
1.91048414e-01 5.35929441e-01 -4.35874373e-01 7.12572455e-01
2.27452338e-01 4.20543879e-01 -8.11912417e-02 5.55027485e-01
7.05477178e-01 5.00963569e-01 -5.50734671e-03 -9.14562028e-03
-2.63249576e-01 6.74429238e-01 -1.30739939e+00 1.74335372e+00
-8.03095758e-01 3.85176271e-01 7.44091213e-01 -1.12402320e+00
1.35184395e+00 1.09997177e+00 1.20757967e-01 -2.03620240e-01
3.10896505e-02 6.86467767e-01 3.99673693e-02 -7.54347324e-01
4.29139078e-01 -3.93323660e-01 1.75015815e-02 5.28724566e-02
8.51608440e-02 -2.81620383e-01 -5.42829990e-01 -1.60056666e-01
1.27403450e+00 -5.36528051e-01 8.85363817e-02 1.24047443e-01
7.86827862e-01 -7.56090462e-01 5.51627159e-01 8.99887681e-01
-1.15908328e-02 1.14890480e+00 -4.04115379e-01 4.19259250e-01
-7.90275514e-01 -1.49048579e+00 1.30183473e-01 9.62489128e-01
-4.05810833e-01 -2.12409049e-01 -8.37284982e-01 -9.28611774e-03
-3.80999804e-01 8.95137310e-01 1.27881557e-01 3.85046929e-01
-7.62838542e-01 -4.82888728e-01 1.14217794e+00 4.08298731e-01
2.09314018e-01 -6.05908155e-01 9.32033733e-02 4.35287863e-01
-4.79989648e-01 -1.30390644e+00 -6.23046994e-01 2.56431073e-01
-3.97962868e-01 -1.21648386e-01 -1.17408109e+00 -1.00885808e+00
4.22865823e-02 5.64660668e-01 7.03414202e-01 -4.78783816e-01
3.99428084e-02 6.29501164e-01 -6.30498946e-01 -4.12098527e-01
-7.03462780e-01 -1.28608420e-01 4.28623855e-01 7.19707668e-01
1.07590174e-02 -1.00608170e+00 -4.54474330e-01 4.47896332e-01
-7.06618190e-01 -2.97648013e-01 4.89509284e-01 6.82776988e-01
4.96231526e-01 6.93645105e-02 1.24178386e+00 -2.22364068e-01
7.73274839e-01 -5.29928744e-01 -2.35594004e-01 2.06972554e-01
2.58356594e-02 -3.23633999e-01 8.93169463e-01 -6.68549836e-01
-1.43051302e+00 -2.49723177e-02 -1.00441861e+00 -6.77161098e-01
-2.00446874e-01 3.24094474e-01 -6.22225821e-01 4.05251890e-01
5.13072908e-01 7.91850209e-01 -3.35649610e-01 -6.46626353e-01
3.46656173e-01 1.90470493e+00 8.37215543e-01 -3.85917515e-01
4.91445899e-01 1.77460909e-01 -7.18359172e-01 -1.39243889e+00
-2.66395539e-01 -1.04952621e+00 -1.74675867e-01 -7.93533698e-02
6.87513828e-01 -1.25925958e+00 -6.76841080e-01 8.96987677e-01
-1.31194794e+00 -2.89070666e-01 -8.96247551e-02 8.90884161e-01
-6.11657560e-01 3.73105049e-01 -6.47980154e-01 -1.63683116e+00
-7.92772830e-01 -9.98327613e-01 1.31452477e+00 -1.91396758e-01
-1.27709121e-01 -5.82836032e-01 1.43937126e-01 7.57795990e-01
5.23036718e-01 -4.60745603e-01 2.63419986e-01 -9.51206505e-01
-2.61483252e-01 -2.54222732e-02 5.88795125e-01 8.96382987e-01
2.11758733e-01 -3.78368974e-01 -1.53047109e+00 -3.04550618e-01
6.25063241e-01 8.96084607e-02 4.96250778e-01 5.94583094e-01
9.33323920e-01 -5.33461452e-01 -1.34117439e-01 4.95265782e-01
9.98328030e-01 5.79546094e-01 5.99851012e-01 -3.48842829e-01
4.64291960e-01 3.36328268e-01 4.92601901e-01 6.44285381e-01
1.79582924e-01 8.18638921e-01 -1.15948722e-01 1.05664752e-01
-2.65183866e-01 -3.86635587e-02 7.96181202e-01 2.10404086e+00
2.91060597e-01 -7.07763910e-01 -6.68530226e-01 8.07835281e-01
-1.47229660e+00 -9.60265338e-01 -2.47283533e-01 2.33454561e+00
8.12116325e-01 -1.16675593e-01 1.11804552e-01 7.23775446e-01
9.22234893e-01 2.06604540e-01 -3.64707828e-01 -6.27589703e-01
-4.54020679e-01 5.48164845e-01 2.37606987e-01 7.46002793e-01
-4.52350527e-01 4.18171018e-01 5.72595072e+00 1.13571751e+00
-1.11866879e+00 4.94156629e-01 3.25484633e-01 -3.65786910e-01
-3.08768511e-01 -4.78084713e-01 -8.90162289e-01 3.41049582e-01
1.60079336e+00 -1.36114731e-01 5.36962211e-01 5.16758859e-01
7.08685517e-01 2.91074753e-01 -1.02481651e+00 1.40005934e+00
4.96574104e-01 -8.66510689e-01 -1.39268963e-02 -4.81547505e-01
3.01929891e-01 -9.52081382e-02 2.37837538e-01 1.45073995e-01
-1.69118032e-01 -8.38775635e-01 9.99306560e-01 1.97663337e-01
9.58217442e-01 -5.27005732e-01 6.53399646e-01 5.37561476e-01
-1.49027646e+00 -1.83350340e-01 1.02558434e-01 -3.20102721e-02
5.09091139e-01 8.30165148e-01 -1.19391143e+00 9.21768606e-01
5.78339279e-01 8.25856179e-02 1.65100947e-01 1.00974500e+00
1.31940059e-02 1.46594167e+00 -4.46273178e-01 -5.07071242e-03
-1.76358119e-01 2.07210332e-01 1.13628483e+00 1.61418653e+00
8.20198774e-01 3.80876750e-01 -1.51945800e-01 3.54416877e-01
2.09445670e-01 1.16972618e-01 -4.29481596e-01 3.48617017e-01
1.02231622e+00 9.39404845e-01 -2.08061188e-01 -4.50475782e-01
-1.54882878e-01 9.39972281e-01 -4.13587391e-01 6.19387269e-01
-8.17388296e-01 -7.28957772e-01 6.48977339e-01 3.95361334e-02
7.34301865e-01 -3.53219002e-01 -2.21089169e-01 -1.01318455e+00
3.09614509e-01 -1.09308708e+00 -1.01540595e-01 -1.01159942e+00
-1.19075334e+00 1.01561809e+00 -3.22244346e-01 -1.23060501e+00
-4.45291340e-01 -2.88134754e-01 -6.19954824e-01 1.12033248e+00
-1.22708952e+00 -9.02864933e-01 1.25942007e-01 8.30776393e-01
1.13132942e+00 -4.38861966e-01 1.00813544e+00 7.71578252e-01
-4.60497588e-01 1.00884402e+00 3.91038239e-01 -2.32605692e-02
4.92303282e-01 -9.69706357e-01 5.52748680e-01 8.21681321e-01
5.27445860e-02 5.34465194e-01 8.55489910e-01 -1.76668406e-01
-1.46400225e+00 -1.08502603e+00 8.78141522e-01 -2.80835778e-01
4.43391412e-01 -9.56369877e-01 -8.95161867e-01 6.32422745e-01
1.04535975e-01 -2.29862422e-01 9.88922894e-01 -8.62599164e-02
-2.12664470e-01 -5.83093405e-01 -8.92059684e-01 3.43946010e-01
9.30462301e-01 -7.58363664e-01 -9.36528981e-01 1.35444209e-01
1.19330084e+00 -4.64022517e-01 -7.81403422e-01 2.93111533e-01
3.85951132e-01 -8.45380843e-01 9.83887374e-01 1.47737384e-01
-2.06986040e-01 -3.97698313e-01 -6.67284906e-01 -1.43250787e+00
1.37503907e-01 -1.42870128e+00 7.62484372e-02 1.73097980e+00
5.02500236e-01 -8.18707645e-01 4.30242002e-01 -1.81927413e-01
-8.64806652e-01 -3.83099645e-01 -1.39526904e+00 -1.09680438e+00
-2.50822604e-01 -8.55887055e-01 5.88640928e-01 4.75270718e-01
-1.68415569e-02 6.70465350e-01 -3.72030765e-01 7.67813265e-01
5.07909358e-01 -2.58857608e-01 7.62317955e-01 -6.29544914e-01
-7.69674778e-01 -3.84146348e-02 -1.84690505e-01 -1.72599900e+00
6.06654249e-02 -6.56346083e-01 5.24805665e-01 -1.17423916e+00
-4.77517784e-01 -3.92809391e-01 -2.18711942e-01 -1.74185112e-01
7.60219572e-03 -1.78820446e-01 -4.39989306e-02 -1.21762969e-01
-3.22964102e-01 6.96668088e-01 6.24672413e-01 -1.04548112e-01
-4.63361830e-01 6.52035475e-01 -3.41184825e-01 4.93783951e-01
6.34863198e-01 -3.12305808e-01 -5.51810324e-01 -4.18502390e-01
-4.42468911e-01 7.72236288e-01 -2.77054477e-02 -1.31229413e+00
3.37882131e-01 3.76660794e-01 -3.97831529e-01 -7.27231026e-01
1.00372851e+00 -6.04750395e-01 3.26002836e-01 -7.82166272e-02
-3.87303889e-01 -1.42117098e-01 2.49527216e-01 5.06318808e-01
-4.51869577e-01 6.87167123e-02 4.01538581e-01 1.51240379e-01
4.32643108e-02 -2.51610667e-01 -8.03259730e-01 -1.55206937e-02
5.27446926e-01 -1.69787750e-01 -3.11362883e-03 -6.84521794e-01
-8.54193747e-01 -1.90383032e-01 -4.63625610e-01 1.92745104e-01
8.93176377e-01 -1.13903856e+00 -1.05480325e+00 3.96721423e-01
-9.11232233e-02 1.75746724e-01 6.17509961e-01 6.73007607e-01
3.44488658e-02 5.46096683e-01 5.72908998e-01 -8.39680254e-01
-1.73579144e+00 2.71681160e-01 3.53589505e-01 8.85445774e-02
-4.59799856e-01 1.14403534e+00 2.88478822e-01 -4.15674299e-01
4.39850479e-01 -4.83105242e-01 1.30210370e-01 -4.39815730e-01
5.51982582e-01 4.41663653e-01 5.69977582e-01 -7.02178657e-01
-2.37854376e-01 3.81356448e-01 3.41270655e-01 -8.46903324e-01
1.09628224e+00 -5.30479610e-01 1.99147031e-01 1.20173013e+00
1.19804549e+00 6.78515434e-01 -7.00516701e-01 -2.67187327e-01
-6.00472152e-01 -3.11148107e-01 -5.75495698e-02 -7.84158111e-01
-6.23882771e-01 1.01825356e+00 5.33043742e-01 4.97310340e-01
1.38205123e+00 -1.17339343e-01 1.24110353e+00 2.75378644e-01
2.85595030e-01 -7.38628924e-01 2.27263123e-02 4.49941128e-01
9.15564537e-01 -5.78541636e-01 -8.58764410e-01 -4.34212387e-01
-4.59946871e-01 8.67748022e-01 1.31268933e-01 2.42888451e-01
7.04955578e-01 7.13131726e-01 5.51891863e-01 3.98318768e-01
-8.33665133e-01 1.78107284e-02 -6.06846884e-02 5.75878978e-01
4.82861102e-01 2.01085046e-01 8.80212411e-02 1.07276750e+00
-7.47912645e-01 -2.27663100e-01 4.17638808e-01 5.35575092e-01
-5.49379885e-01 -7.73217916e-01 -1.03290546e+00 1.42573640e-01
-6.71141207e-01 -4.78232026e-01 -1.06133565e-01 -1.70097381e-01
-3.45008433e-01 1.73187113e+00 6.38646558e-02 -6.65343404e-01
5.28380573e-01 2.38676623e-01 2.02131554e-01 -5.39997637e-01
-8.12197089e-01 6.87141001e-01 3.78096074e-01 9.83324926e-03
-3.42934072e-01 -6.91626370e-01 -1.12974370e+00 -1.81674380e-02
-5.73767722e-01 5.73640704e-01 9.38603818e-01 8.20254385e-01
4.42657232e-01 1.00309086e+00 1.09000051e+00 -8.52156997e-01
-8.28814626e-01 -1.28733099e+00 -8.29545498e-01 -2.96770055e-02
8.32468271e-01 -6.42257184e-02 -7.35291660e-01 1.26652300e-01] | [14.77794075012207, 6.095434665679932] |
ee1a1016-9e32-444b-934a-f8d4d0327dd7 | continuous-3d-label-stereo-matching-using | 1603.08328 | null | http://arxiv.org/abs/1603.08328v3 | http://arxiv.org/pdf/1603.08328v3.pdf | Continuous 3D Label Stereo Matching using Local Expansion Moves | We present an accurate stereo matching method using local expansion moves
based on graph cuts. This new move-making scheme is used to efficiently infer
per-pixel 3D plane labels on a pairwise Markov random field (MRF) that
effectively combines recently proposed slanted patch matching and curvature
regularization terms. The local expansion moves are presented as many
alpha-expansions defined for small grid regions. The local expansion moves
extend traditional expansion moves by two ways: localization and spatial
propagation. By localization, we use different candidate alpha-labels according
to the locations of local alpha-expansions. By spatial propagation, we design
our local alpha-expansions to propagate currently assigned labels for nearby
regions. With this localization and spatial propagation, our method can
efficiently infer MRF models with a continuous label space using randomized
search. Our method has several advantages over previous approaches that are
based on fusion moves or belief propagation; it produces submodular moves
deriving a subproblem optimality; it helps find good, smooth, piecewise linear
disparity maps; it is suitable for parallelization; it can use cost-volume
filtering techniques for accelerating the matching cost computations. Even
using a simple pairwise MRF, our method is shown to have best performance in
the Middlebury stereo benchmark V2 and V3. | ['Yasuyuki Matsushita', 'Tatsunori Taniai', 'Takeshi Naemura', 'Yoichi Sato'] | 2016-03-28 | null | null | null | null | ['patch-matching'] | ['computer-vision'] | [ 2.65405476e-01 6.66501746e-02 -4.10899609e-01 -3.98169249e-01
-1.31188595e+00 -3.48919779e-01 3.18553418e-01 2.41010740e-01
-3.68052274e-01 6.05113983e-01 1.40727043e-01 -1.43361121e-01
-1.05147801e-01 -9.10331070e-01 -1.00337148e+00 -5.58932483e-01
-1.12922154e-01 8.08684647e-01 9.06897902e-01 -1.89412415e-01
8.28918517e-01 7.36382484e-01 -1.69453871e+00 3.90201509e-01
7.99807250e-01 9.23267961e-01 1.53243840e-01 7.01478124e-01
-1.00467563e-01 7.38637030e-01 1.14371538e-01 -2.87635237e-01
7.41508424e-01 -1.10740647e-01 -1.10120785e+00 9.15282220e-02
1.07228220e+00 -2.39635751e-01 -3.35898716e-03 1.01275814e+00
3.75335693e-01 4.05209303e-01 4.41340655e-01 -1.03940427e+00
-1.67962179e-01 3.00350845e-01 -1.23190916e+00 2.83958353e-02
6.80835962e-01 1.10188417e-01 9.63707805e-01 -8.59060466e-01
1.11145771e+00 1.46835935e+00 1.10757148e+00 3.61318946e-01
-1.67578924e+00 -4.79882985e-01 2.33452141e-01 2.13763699e-01
-1.58246899e+00 -1.97059229e-01 6.53238952e-01 -6.08826280e-01
1.16848159e+00 4.64579850e-01 7.89444387e-01 1.23271318e-02
4.33297932e-01 7.60620654e-01 1.12595701e+00 -5.31987071e-01
2.97189146e-01 -3.43241543e-01 2.67994463e-01 1.12816679e+00
-2.94881076e-01 3.12193125e-01 -6.52848661e-01 -6.26430213e-01
1.11500084e+00 -2.42494598e-01 -5.47873855e-01 -8.43912780e-01
-1.33888078e+00 8.93216252e-01 6.12302125e-01 -6.49873912e-02
-3.90107125e-01 5.30672431e-01 6.07666373e-02 2.11297926e-02
4.66050863e-01 2.18111143e-01 -4.30958211e-01 2.47261543e-02
-1.28838575e+00 5.03524244e-01 5.60616553e-01 1.04407883e+00
1.52003598e+00 -6.70267761e-01 -1.63930863e-01 6.90031111e-01
3.38301688e-01 5.26626647e-01 -7.47215301e-02 -1.76598883e+00
5.08799016e-01 5.66374421e-01 1.36102185e-01 -1.32626426e+00
-4.55340952e-01 1.79714009e-01 -4.82605815e-01 7.01658964e-01
2.73905605e-01 1.91935852e-01 -1.26270163e+00 1.53376913e+00
7.23783672e-01 3.26022267e-01 -3.05321693e-01 8.56001079e-01
7.01357782e-01 6.94874644e-01 -3.67713541e-01 -1.58102334e-01
9.55878377e-01 -1.17576194e+00 -3.46274048e-01 1.58232033e-01
7.21882463e-01 -9.58901107e-01 5.47725618e-01 4.65312898e-01
-1.50127792e+00 -5.14633596e-01 -7.19389141e-01 -4.67748523e-01
-5.14320210e-02 -5.00902832e-01 7.59467125e-01 4.53372240e-01
-1.71560013e+00 6.76248133e-01 -8.74002934e-01 2.18160450e-02
3.80654126e-01 7.47361243e-01 -2.41718292e-01 -3.24989349e-01
-4.76184189e-01 6.88754559e-01 3.36627066e-02 -3.54282744e-02
-3.72927785e-01 -8.26415300e-01 -1.02966297e+00 -3.16194504e-01
1.49496198e-01 -1.03552866e+00 9.45588231e-01 -6.71120822e-01
-1.28757370e+00 1.03543651e+00 -8.11317205e-01 -2.96203315e-01
5.03064752e-01 2.85937309e-01 3.02534431e-01 3.29750389e-01
3.53882760e-01 1.58336794e+00 4.90677565e-01 -1.34227586e+00
-1.03426850e+00 -4.06242013e-01 -1.74505766e-02 4.84878570e-01
5.96213758e-01 -3.39136362e-01 -8.11189473e-01 -4.24407870e-01
8.41870546e-01 -9.43356156e-01 -7.02542841e-01 1.83328733e-01
-2.98733890e-01 -1.94043204e-01 4.32240039e-01 -5.20954251e-01
9.85946774e-01 -1.97006500e+00 3.99096906e-01 8.41573179e-01
4.05940801e-01 -4.00998265e-01 8.75637457e-02 3.03662680e-02
2.11713225e-01 -1.03134669e-01 -5.31049371e-01 -5.38344324e-01
-1.64823979e-01 1.06199786e-01 6.85987920e-02 6.47447407e-01
-3.81313920e-01 8.93328846e-01 -9.03870404e-01 -8.57283533e-01
3.88978988e-01 5.91354549e-01 -9.40526366e-01 -3.93099338e-01
-2.04501510e-01 2.60811478e-01 -2.05662131e-01 7.47730970e-01
1.16862535e+00 -7.81802312e-02 -4.05016840e-01 -2.11270764e-01
-3.45739812e-01 -1.42152354e-01 -1.54263806e+00 2.06149840e+00
-1.57796144e-01 6.96517706e-01 1.52422369e-01 -5.79877377e-01
8.93193960e-01 -1.87549815e-01 6.67073190e-01 -4.54247326e-01
-1.20279804e-01 4.47264016e-01 -8.38573456e-01 -4.10756506e-02
6.79664791e-01 -5.16702197e-02 2.79199243e-01 1.96673766e-01
-3.63846719e-01 -5.67342520e-01 3.88086103e-02 2.14888304e-01
1.19799054e+00 2.63786882e-01 -5.62450364e-02 -4.53110725e-01
6.56363904e-01 5.29691041e-01 9.45977926e-01 6.49896920e-01
-7.48729855e-02 1.00721097e+00 7.69503862e-02 -7.24006712e-01
-7.20862031e-01 -9.88258779e-01 -4.75464076e-01 7.57677019e-01
8.95877302e-01 -3.52723986e-01 -6.18802369e-01 -5.18944383e-01
2.22768992e-01 3.24712515e-01 -6.08393013e-01 5.71878016e-01
-8.02575707e-01 -3.92640799e-01 7.34413834e-03 5.86863995e-01
5.80603957e-01 -7.45984197e-01 -6.50653780e-01 3.26744199e-01
-3.66703540e-01 -7.14677036e-01 -6.77483320e-01 -1.69799086e-02
-1.14352572e+00 -1.11566341e+00 -9.94269371e-01 -1.13530827e+00
9.86960530e-01 5.16092241e-01 1.13162959e+00 5.37289344e-02
-2.08704978e-01 5.23480356e-01 -4.05476876e-02 2.92169929e-01
-7.99549744e-02 -3.08126748e-01 -4.80917871e-01 -1.71767101e-01
4.53912199e-01 -4.28853214e-01 -8.13685060e-01 6.57478452e-01
-3.89061660e-01 4.04909045e-01 1.18972532e-01 6.56639814e-01
1.41502655e+00 -6.23815954e-02 -4.43448067e-01 -7.76690841e-01
1.30687147e-01 -6.12450205e-02 -1.15602577e+00 2.96436429e-01
-4.14965332e-01 2.80966610e-01 -9.86481309e-02 -5.65961860e-02
-1.03865230e+00 4.33676034e-01 -5.57609648e-02 -3.53838027e-01
-3.12485676e-02 8.57415795e-02 3.95910859e-01 -8.49830449e-01
5.40788531e-01 -2.04098538e-01 -2.18399718e-01 -2.81924367e-01
4.67159510e-01 1.46998495e-01 5.92312813e-01 -6.73018098e-01
4.05456126e-01 9.96544242e-01 4.07931268e-01 -6.13344431e-01
-5.98044768e-02 -9.94284630e-01 -5.23446023e-01 -3.68775487e-01
1.10027301e+00 -6.76416993e-01 -7.35419452e-01 5.06408870e-01
-1.35186934e+00 -3.75457138e-01 -3.34587008e-01 4.04908180e-01
-9.10289049e-01 6.11479580e-01 -7.54936993e-01 -6.17157638e-01
-8.75495076e-02 -1.52965021e+00 1.56576359e+00 8.47311765e-02
-2.22501680e-01 -1.11481667e+00 5.16943753e-01 4.51001346e-01
1.76994845e-01 4.31980371e-01 7.13452578e-01 3.39055836e-01
-1.30121291e+00 8.96934941e-02 -2.64964342e-01 -1.99716404e-01
-2.93626755e-01 3.26698497e-02 -7.36662090e-01 -2.99989074e-01
-3.20511520e-01 -1.82761736e-02 1.10319245e+00 1.09792376e+00
8.85750115e-01 2.29575858e-01 -6.80839002e-01 1.02633512e+00
1.72533333e+00 1.42968059e-01 8.33743572e-01 4.52665538e-01
8.04796576e-01 6.68325126e-01 7.50495076e-01 1.38851866e-01
7.10681915e-01 8.27800751e-01 3.44898671e-01 -3.81436050e-01
-2.91799873e-01 -2.09575638e-01 3.62857953e-02 5.25331914e-01
-3.59655708e-01 1.87196344e-01 -1.02130592e+00 5.92104912e-01
-2.13745260e+00 -7.34459281e-01 -6.28868937e-01 2.31970429e+00
5.45569658e-01 -5.18621840e-02 6.32700101e-02 7.76800513e-03
9.07374501e-01 -1.93600301e-02 -2.23011971e-01 -4.22742635e-01
-1.34246126e-01 2.94953436e-01 1.13156343e+00 1.36873591e+00
-9.12025332e-01 9.62738872e-01 6.48906422e+00 9.24653053e-01
-8.07494104e-01 1.52694598e-01 6.05476677e-01 1.35283023e-01
-7.64848173e-01 3.28097850e-01 -1.11449993e+00 8.66317749e-02
7.04453290e-02 2.98250169e-01 3.17467928e-01 5.65745354e-01
6.98755458e-02 -8.44576716e-01 -1.06017506e+00 1.26754165e+00
-2.10223626e-03 -1.76847517e+00 -5.40026017e-02 2.26578057e-01
1.29942346e+00 2.07920581e-01 -1.26392633e-01 -2.18515083e-01
4.64984655e-01 -7.39869237e-01 6.77914739e-01 4.17076200e-01
5.83119094e-01 -7.69083202e-01 5.70120454e-01 2.87200689e-01
-1.42231345e+00 2.39659771e-01 -5.69620550e-01 6.18635453e-02
4.95688647e-01 6.62174106e-01 -4.60617870e-01 3.53669763e-01
6.80853605e-01 4.98492479e-01 -1.97373882e-01 1.36147571e+00
1.24228252e-02 -1.14117123e-01 -7.13704288e-01 1.90828055e-01
3.25744390e-01 -3.95250916e-01 5.05201221e-01 1.07901216e+00
1.68338820e-01 2.18810633e-01 3.04041117e-01 5.94924390e-01
2.65547007e-01 2.34284386e-01 -3.91922921e-01 1.02740800e+00
2.78075695e-01 5.74840069e-01 -1.06268728e+00 -3.45433891e-01
-1.89812064e-01 8.84838402e-01 2.47076824e-01 4.76499259e-01
-6.68056786e-01 -2.63415992e-01 3.71914029e-01 3.38775128e-01
4.44962442e-01 -1.51186541e-01 -5.77693105e-01 -1.04102695e+00
1.11709476e-01 -4.80439454e-01 3.94031227e-01 -7.58713663e-01
-9.02445257e-01 4.67109561e-01 2.42160633e-01 -1.09103477e+00
-2.26759806e-01 -3.18353117e-01 -1.90164998e-01 9.30295050e-01
-1.65369773e+00 -9.56565201e-01 -2.84922451e-01 8.54534686e-01
6.59053862e-01 3.50452214e-01 5.89276254e-01 1.70531124e-01
1.42194837e-01 2.86389530e-01 -1.29052669e-01 -3.13137591e-01
4.03802931e-01 -1.19232833e+00 3.78843516e-01 8.82003367e-01
8.78360793e-02 2.37264708e-01 3.74000788e-01 -8.22553754e-01
-1.00060594e+00 -6.51690960e-01 1.14361274e+00 -3.13928187e-01
9.52837393e-02 9.20986943e-03 -8.60265195e-01 7.30914712e-01
6.15428276e-02 9.57903862e-02 1.91826075e-01 5.85995615e-02
-7.49189705e-02 -1.39101341e-01 -1.56094420e+00 4.15065855e-01
1.22459781e+00 -2.47916430e-01 -4.69998538e-01 5.17534971e-01
3.96306813e-01 -9.53466237e-01 -6.66708171e-01 6.55525982e-01
4.13600594e-01 -1.34440517e+00 1.07940412e+00 1.99722379e-01
1.62246957e-01 -6.02532506e-01 -2.46006966e-01 -9.79745924e-01
-6.11603975e-01 -8.73855948e-01 2.58394420e-01 9.26525831e-01
4.67113286e-01 -6.58912718e-01 1.17176020e+00 8.33155394e-01
-1.02541491e-01 -8.77411664e-01 -1.20279694e+00 -5.39132535e-01
-1.31181002e-01 -5.98878264e-01 5.48298240e-01 8.31449568e-01
-1.86578613e-02 -2.46748403e-01 4.55118380e-02 2.62225270e-01
1.16045237e+00 3.22236985e-01 8.04123163e-01 -9.66414750e-01
-4.13848221e-01 -6.15585685e-01 -8.54191065e-01 -1.76167011e+00
-3.33758034e-02 -8.20236742e-01 3.03965360e-01 -1.68803048e+00
1.77121222e-01 -9.77051914e-01 2.78214714e-03 2.85014480e-01
1.30741680e-02 4.34088677e-01 -7.42692873e-02 3.10969293e-01
-4.51870292e-01 1.10015079e-01 1.27522016e+00 -2.01720014e-01
-5.18007815e-01 -6.89759403e-02 -6.63940310e-02 8.43436897e-01
4.40403670e-01 -4.58742052e-01 -3.30214471e-01 -5.15905499e-01
3.06780070e-01 2.54867971e-01 2.68065363e-01 -9.78732228e-01
5.76343060e-01 -2.22094193e-01 -1.17453961e-02 -1.14691091e+00
5.23051798e-01 -6.57054365e-01 4.22196299e-01 3.06475818e-01
-3.00925255e-01 1.51658773e-01 -3.24129825e-03 5.28511345e-01
-2.83156723e-01 -2.07389981e-01 9.92899001e-01 -2.60152012e-01
-1.21702564e+00 3.73083532e-01 -5.44801205e-02 3.06725688e-02
1.21173573e+00 -7.94038951e-01 1.55834049e-01 -3.52782488e-01
-8.35572720e-01 5.16364038e-01 9.66695011e-01 -1.23465136e-01
9.06074166e-01 -1.35493565e+00 -6.37922406e-01 1.90412030e-01
-3.46602313e-02 5.82923055e-01 1.97826475e-01 9.59447980e-01
-1.23490500e+00 2.06991225e-01 9.34165157e-03 -1.21889782e+00
-1.47191024e+00 4.74227458e-01 3.57820839e-01 -1.27532244e-01
-8.14639688e-01 1.42977500e+00 2.77983040e-01 -2.86452979e-01
3.52981627e-01 -5.58334529e-01 1.59043938e-01 -3.24493468e-01
2.89834499e-01 7.62098014e-01 1.12003675e-02 -7.29642153e-01
-4.30254519e-01 1.43284869e+00 2.43938416e-01 -3.77943844e-01
1.05911553e+00 -2.94760853e-01 -2.29346618e-01 1.30119145e-01
1.27604520e+00 2.25997970e-01 -1.33961999e+00 -1.19329214e-01
-3.10711056e-01 -1.11119020e+00 4.92732614e-01 -3.66932482e-01
-1.24587131e+00 6.19185030e-01 5.13914943e-01 -1.92885160e-01
1.04568481e+00 2.21995547e-01 9.37123835e-01 -6.46918416e-02
9.34201479e-01 -1.09286499e+00 -5.09112477e-01 2.84738868e-01
5.93825281e-01 -1.15108252e+00 1.67609602e-01 -1.06822062e+00
-2.35665679e-01 1.07419813e+00 3.34158629e-01 -1.82972267e-01
7.24811912e-01 3.69360030e-01 2.96907052e-02 -3.99123728e-01
-3.12214166e-01 -1.87380463e-01 4.00422513e-01 6.42994165e-01
5.46013117e-02 4.09835912e-02 -4.88647640e-01 -6.15671813e-01
-3.67787592e-02 2.23188028e-02 2.52025187e-01 9.52890873e-01
-5.56541622e-01 -1.09428835e+00 -7.31542110e-01 1.81086525e-01
3.49237956e-02 -1.40584767e-01 2.73237620e-02 7.36889720e-01
4.78138179e-01 7.88278759e-01 2.71563947e-01 -1.82311550e-01
3.56169730e-01 -3.47861737e-01 8.22207987e-01 -3.36548448e-01
-3.54052603e-01 2.33485267e-01 4.32596616e-02 -1.08044004e+00
-8.63824487e-01 -8.90088797e-01 -1.46555650e+00 -5.61759412e-01
-4.00598049e-01 7.55562484e-02 5.78741074e-01 7.72183836e-01
3.27511817e-01 -2.38856912e-01 5.84990263e-01 -1.24913526e+00
7.38081336e-02 -9.98924747e-02 -4.12582338e-01 2.04700455e-01
2.78241277e-01 -7.51538396e-01 -4.61888045e-01 2.77027953e-02] | [8.937685012817383, -2.4124374389648438] |
c41952f6-63ec-4d71-a4d5-f1c31b8b4619 | incremental-predictive-process-monitoring-how | 1804.03967 | null | http://arxiv.org/abs/1804.03967v1 | http://arxiv.org/pdf/1804.03967v1.pdf | Incremental Predictive Process Monitoring: How to Deal with the Variability of Real Environments | A characteristic of existing predictive process monitoring techniques is to
first construct a predictive model based on past process executions, and then
use it to predict the future of new ongoing cases, without the possibility of
updating it with new cases when they complete their execution. This can make
predictive process monitoring too rigid to deal with the variability of
processes working in real environments that continuously evolve and/or exhibit
new variant behaviors over time. As a solution to this problem, we propose the
use of algorithms that allow the incremental construction of the predictive
model. These incremental learning algorithms update the model whenever new
cases become available so that the predictive model evolves over time to fit
the current circumstances. The algorithms have been implemented using different
case encoding strategies and evaluated on a number of real and synthetic
datasets. The results provide a first evidence of the potential of incremental
learning strategies for predicting process monitoring in real environments, and
of the impact of different case encoding strategies in this setting. | ['Fabrizio Maria Maggi', 'Chiara Ghidini', 'Chiara Di Francescomarino', 'Williams Rizzi', 'Cosimo Damiano Persia'] | 2018-04-11 | null | null | null | null | ['predictive-process-monitoring'] | ['time-series'] | [ 6.08367741e-01 4.94043194e-02 -2.10646316e-02 -9.88419577e-02
-1.10324249e-01 -3.24342608e-01 8.40969622e-01 5.85731864e-01
-2.07120746e-01 6.51716948e-01 -4.98769321e-02 -2.83574134e-01
-4.38013405e-01 -9.44980383e-01 -1.14108182e-01 -4.72690314e-01
-6.66356325e-01 8.95198405e-01 7.36055255e-01 2.12897927e-01
3.33380967e-01 7.56998777e-01 -1.74711668e+00 2.55767852e-01
3.85360241e-01 6.28608644e-01 -7.55586252e-02 1.04273582e+00
-3.22731793e-01 8.73428762e-01 -8.94945741e-01 4.99290824e-02
3.44959110e-01 -2.65566260e-01 -5.33636451e-01 5.60419202e-01
-3.67280841e-01 -1.24324210e-01 -1.28661439e-01 2.90663421e-01
4.33327183e-02 9.83445272e-02 2.22876325e-01 -1.49379659e+00
2.09346846e-01 5.50036967e-01 -3.40447932e-01 4.63737845e-01
5.80962837e-01 6.54365182e-01 3.17810744e-01 1.94385350e-01
6.08156085e-01 1.07326531e+00 8.11985850e-01 4.31614608e-01
-1.39624810e+00 -1.88870236e-01 4.28775877e-01 8.50428641e-02
-9.54088151e-01 -3.77065361e-01 4.40563232e-01 -4.51432407e-01
1.20822930e+00 2.84340829e-01 7.77035892e-01 9.30515528e-01
6.21170044e-01 4.15339053e-01 9.84175563e-01 -4.51100081e-01
6.38619483e-01 -3.45923305e-02 4.84943911e-02 2.55850762e-01
5.15513241e-01 2.45949030e-01 -3.71465355e-01 -7.18871772e-01
6.85400665e-01 4.14078891e-01 -9.62287784e-02 -3.35649908e-01
-1.08956742e+00 1.59607917e-01 -4.74274427e-01 5.08954644e-01
-6.52816474e-01 1.08234592e-01 5.55402637e-01 5.90602994e-01
2.74011880e-01 6.61311567e-01 -7.04265058e-01 -6.51366532e-01
-9.22015667e-01 2.94327766e-01 1.49399507e+00 6.54218793e-01
7.03543663e-01 2.02196185e-02 -3.86617899e-01 3.24112296e-01
1.66277200e-01 -3.60751040e-02 9.64046717e-01 -9.44070816e-01
8.83098841e-02 8.58453453e-01 3.53676260e-01 -3.91815335e-01
-1.82696179e-01 -2.79273629e-01 -3.85448515e-01 2.08904639e-01
4.12909538e-01 -7.34955370e-02 -7.50560164e-01 1.27955949e+00
2.09158480e-01 5.24458945e-01 -1.59180969e-01 -1.03537329e-01
-4.99585271e-01 6.49852514e-01 1.33851662e-01 -7.72144198e-01
8.07661593e-01 -6.68450177e-01 -6.67269230e-01 -1.62470102e-01
3.85658175e-01 -3.90260667e-01 7.13510275e-01 5.88084638e-01
-8.93184900e-01 -6.75712526e-01 -8.93685222e-01 1.17260265e+00
4.45122495e-02 -5.00223815e-01 7.64056444e-01 8.77855122e-01
-9.69217420e-01 8.72901320e-01 -1.57233810e+00 -5.12580931e-01
-2.42983270e-02 2.95232773e-01 1.57763973e-01 6.44592345e-02
-6.29877746e-01 7.13969648e-01 6.38018370e-01 -8.09646472e-02
-9.59411681e-01 -5.10732412e-01 -3.36798996e-01 2.10877806e-01
6.18817151e-01 -5.04915297e-01 1.57940376e+00 -9.55286741e-01
-1.45521355e+00 5.80019616e-02 -2.97683328e-01 -5.56534529e-01
8.53419840e-01 -8.34639817e-02 -7.55182445e-01 -3.90429050e-01
-3.36284876e-01 -3.39451551e-01 9.34057891e-01 -1.37851608e+00
-8.13221633e-01 -3.00288856e-01 -2.36475855e-01 -1.90933108e-01
-4.11517397e-02 -1.35754466e-01 -2.97589034e-01 -1.05091542e-01
4.62213084e-02 -9.98472810e-01 -6.40321136e-01 -5.96942425e-01
4.68563428e-03 -4.28746492e-02 9.81555462e-01 -2.78052479e-01
1.38415873e+00 -1.95355701e+00 -1.13425955e-01 4.51919228e-01
-5.58361225e-02 1.72018573e-01 9.55517739e-02 8.81316543e-01
5.75125180e-02 9.11670849e-02 -2.34224528e-01 -2.79176295e-01
-2.29629248e-01 6.84678972e-01 -2.32838437e-01 -7.21504390e-02
1.94038182e-01 2.81506389e-01 -8.02110255e-01 -3.08550328e-01
1.72548383e-01 -9.01488704e-04 -6.76969215e-02 5.61133265e-01
-6.03145599e-01 4.66987252e-01 -4.22352344e-01 4.43377852e-01
2.30932757e-01 -2.06372917e-01 6.18532181e-01 6.60335422e-01
-2.86428183e-01 1.23911120e-01 -1.46491551e+00 8.09261084e-01
-5.56549907e-01 1.80051148e-01 -1.35745302e-01 -6.90218866e-01
9.07309771e-01 5.11580408e-01 7.50779271e-01 -3.95915300e-01
-2.39905864e-01 -1.23235457e-01 1.67415157e-01 -6.01325691e-01
2.48730540e-01 -2.09098250e-01 2.04510793e-01 1.01831222e+00
-2.94575214e-01 1.32299366e-03 5.41186273e-01 -3.08850706e-01
1.80594182e+00 1.76185951e-01 7.20829129e-01 3.47757667e-01
6.52920902e-01 -1.59202486e-01 7.87112892e-01 9.08817232e-01
-1.93968832e-01 -4.43296619e-02 8.15037668e-01 -9.30497289e-01
-9.89507198e-01 -9.96178687e-01 2.17050284e-01 1.02612460e+00
-3.36051494e-01 -4.67768610e-01 -1.85570478e-01 -5.22765815e-01
2.18003429e-02 8.80743206e-01 -5.89812934e-01 -1.86932981e-01
-7.06585050e-01 -9.49510694e-01 1.53126612e-01 3.64887595e-01
2.46558249e-01 -1.22913229e+00 -1.42053390e+00 9.19305563e-01
5.25210440e-01 -7.68375337e-01 3.53772819e-01 5.14143765e-01
-1.26560366e+00 -1.21642256e+00 2.51796752e-01 -7.28412196e-02
3.13417375e-01 -3.04633170e-01 1.15353692e+00 1.76666364e-01
-1.94525957e-01 7.33205378e-01 -2.32699901e-01 -5.99294245e-01
-1.18211889e+00 8.03286731e-02 -1.56910151e-01 1.23033775e-02
1.82751983e-01 -6.35396838e-01 1.29811512e-02 2.73956209e-01
-1.17790282e+00 -2.53465980e-01 5.25178134e-01 6.03263021e-01
6.41627014e-01 6.28188670e-01 4.89596635e-01 -1.14156878e+00
8.85350704e-01 -5.05027533e-01 -7.45710552e-01 6.12908840e-01
-1.10157001e+00 2.40702257e-01 6.68358862e-01 -7.12948978e-01
-1.36046350e+00 1.23786397e-01 2.18841314e-01 -2.56732643e-01
-4.88213271e-01 5.15959740e-01 -9.27650928e-02 5.47274172e-01
3.27829421e-01 4.32983249e-01 1.32250428e-01 -2.81439334e-01
-1.49718806e-01 2.71551341e-01 2.08994970e-01 -6.03508174e-01
6.19594038e-01 3.88227075e-01 1.18218794e-01 -4.63040203e-01
-1.75336301e-01 -5.22082508e-01 -6.64802074e-01 -3.21080744e-01
1.15172952e-01 -2.76475102e-01 -5.14549971e-01 6.36658013e-01
-6.33381486e-01 -6.05912983e-01 -6.26503050e-01 9.14950203e-03
-5.56458116e-01 2.63640165e-01 -6.73041582e-01 -1.11058056e+00
-7.26618767e-02 -8.66231501e-01 7.34325290e-01 1.25717118e-01
-6.27151966e-01 -1.26722395e+00 6.07859612e-01 -2.14966148e-01
6.30301952e-01 4.33777601e-01 9.49240148e-01 -1.08767664e+00
-6.01026535e-01 -7.39361763e-01 4.97830480e-01 -5.21660671e-02
5.15801191e-01 3.80353540e-01 -8.76762033e-01 -5.72302401e-01
3.52465123e-01 5.71073472e-01 3.54006916e-01 -2.93361604e-01
1.19483685e+00 -4.03848320e-01 -6.59482658e-01 1.44664600e-01
1.40039372e+00 6.20116472e-01 5.53871870e-01 6.68737292e-01
3.55756164e-01 4.61839199e-01 4.51040596e-01 8.56060922e-01
-3.51380929e-02 5.21860003e-01 3.18300366e-01 5.74080050e-01
4.17463094e-01 -1.73367709e-02 4.24449742e-01 4.02424484e-01
-3.26321304e-01 -5.11160269e-02 -1.31621981e+00 5.01300275e-01
-1.86629975e+00 -1.41607666e+00 -4.12469637e-03 2.61350918e+00
5.44123411e-01 6.15978181e-01 1.63265333e-01 4.22857672e-01
4.59088236e-01 -4.07478064e-02 -6.42561972e-01 -7.03274667e-01
4.77635294e-01 2.36874610e-01 4.81097579e-01 1.30537182e-01
-9.43276167e-01 2.87546664e-01 7.02310181e+00 -7.11174076e-03
-1.23895264e+00 2.91734445e-03 5.28213441e-01 -8.69769678e-02
1.76697932e-02 2.40112707e-01 -6.22616827e-01 5.01304150e-01
1.57269895e+00 -5.55076122e-01 2.65825838e-01 8.53993475e-01
3.65766257e-01 -4.09418941e-01 -1.47167826e+00 5.46307921e-01
-1.65835798e-01 -1.14208627e+00 -1.29489779e-01 1.60250619e-01
6.84900522e-01 -7.35274097e-03 -3.38178813e-01 4.27734375e-01
5.67817211e-01 -8.09552550e-01 3.96693170e-01 8.06773126e-01
2.22835952e-04 -6.31354451e-01 9.35134470e-01 6.47623539e-01
-1.04032254e+00 -6.33284152e-01 1.23719446e-01 -5.31141698e-01
1.54703975e-01 5.41837871e-01 -1.50703084e+00 4.21777993e-01
4.95990127e-01 4.08078849e-01 -7.36100495e-01 1.24973321e+00
6.47362471e-02 8.68541121e-01 -3.31939608e-01 1.35332793e-01
-1.80793554e-01 1.61169872e-01 6.26010656e-01 1.03455687e+00
4.16067690e-01 -5.74119508e-01 4.06654149e-01 5.61792374e-01
6.16548121e-01 -1.99339092e-01 -5.42086899e-01 1.04194142e-01
5.48557937e-01 8.46339762e-01 -8.64344120e-01 -4.46165472e-01
-3.24750811e-01 6.54260278e-01 6.22907951e-02 3.81111801e-01
-5.23064554e-01 3.36600512e-01 5.36228538e-01 5.67583382e-01
2.56589830e-01 -4.48198318e-01 -2.82788128e-01 -7.50192881e-01
8.34753066e-02 -8.69545758e-01 6.35640860e-01 -2.67567962e-01
-1.39608335e+00 4.15870875e-01 6.68350384e-02 -1.11960173e+00
-7.53049076e-01 -1.72476664e-01 -9.65641856e-01 6.42389834e-01
-1.03025699e+00 -7.36419678e-01 -2.85558492e-01 2.37723917e-01
5.41497529e-01 -1.05649635e-01 8.14493775e-01 -4.74431902e-01
-4.46373761e-01 -7.44646117e-02 2.24081844e-01 -5.18515050e-01
5.66703975e-01 -1.23692906e+00 4.15087909e-01 7.63482094e-01
-4.88180779e-02 6.18476331e-01 9.53048706e-01 -9.68339086e-01
-1.43806005e+00 -1.19605684e+00 4.81273383e-01 -6.83905602e-01
8.06022465e-01 -5.70844337e-02 -1.56886303e+00 9.10657823e-01
-1.85186006e-02 -1.84871942e-01 7.38017261e-01 2.15388238e-01
-1.17646180e-01 -1.96464226e-01 -1.15606678e+00 2.48175040e-01
7.57318974e-01 -1.93926647e-01 -5.76474845e-01 -1.29454911e-01
2.77362198e-01 -1.55551732e-01 -1.08476257e+00 2.50804067e-01
4.00152594e-01 -1.11166239e+00 3.99807364e-01 -4.16800648e-01
-1.19979471e-01 -3.26751024e-01 2.51299560e-01 -1.26799643e+00
-3.38493139e-01 -8.95000696e-01 -5.36104321e-01 1.06226146e+00
3.72098714e-01 -8.38870108e-01 5.75743318e-01 9.51052010e-01
2.08483562e-01 -5.77719986e-01 -1.07332838e+00 -8.64019692e-01
-4.80926841e-01 -5.28394341e-01 1.06254554e+00 8.41409206e-01
6.74190298e-02 -1.45655587e-01 -3.94560732e-02 1.44105405e-01
3.36941779e-01 1.21596456e-01 9.94838297e-01 -1.42146027e+00
-9.03470278e-01 -3.52978051e-01 -6.24162018e-01 -1.79061830e-01
-4.69794646e-02 -2.08221316e-01 -7.24228844e-02 -1.37995398e+00
1.62904799e-01 -5.21141231e-01 -4.31417912e-01 5.06367624e-01
-3.96794051e-01 -7.79428601e-01 1.25175327e-01 5.96739292e-01
-6.20162070e-01 1.25444561e-01 7.71640360e-01 1.02669254e-01
-6.93696916e-01 5.94824791e-01 -1.10014461e-01 5.68754494e-01
8.81614745e-01 -4.73835021e-01 -5.74823916e-01 1.45397842e-01
1.87356547e-01 1.90856710e-01 1.69608861e-01 -1.49416661e+00
2.34380707e-01 -4.92617428e-01 4.48241979e-01 -2.69359767e-01
-8.35491344e-02 -1.13499904e+00 9.27497685e-01 7.83200979e-01
-3.79993051e-01 4.47777092e-01 2.56788909e-01 1.00896740e+00
-1.38113067e-01 -1.76586583e-01 3.61791015e-01 -2.37582281e-01
-9.30427611e-01 2.42860004e-01 -8.75606775e-01 -4.02669013e-01
1.53337550e+00 -6.14050329e-01 2.08770786e-03 -2.94901520e-01
-9.45272088e-01 1.01463705e-01 9.03436661e-01 4.09063011e-01
1.98353797e-01 -6.10690534e-01 -3.11842322e-01 3.72373462e-01
2.93254405e-01 -1.33422464e-01 -9.94798765e-02 5.82144320e-01
-3.70738238e-01 4.09694649e-02 -3.82475227e-01 -5.08689761e-01
-1.17016637e+00 9.14389729e-01 5.77809393e-01 -9.44352567e-01
-7.22352862e-01 -3.26895081e-02 -5.43232083e-01 -9.15161371e-02
-8.89887661e-02 -3.16264242e-01 -1.28445655e-01 -1.73287660e-01
6.67335808e-01 3.51432055e-01 2.49741033e-01 1.72909461e-02
-5.29253334e-02 -1.30966455e-01 1.74323302e-02 -1.72897547e-01
1.49069357e+00 -1.07707486e-01 -3.39429259e-01 1.16387284e+00
1.79751843e-01 9.57914740e-02 -1.56509185e+00 -2.12106138e-01
7.93602109e-01 -5.34445763e-01 -4.92917717e-01 -7.49184430e-01
-7.08351672e-01 3.55246305e-01 5.85540652e-01 6.39042974e-01
1.12431371e+00 -1.28885418e-01 2.77824104e-01 4.31303710e-01
6.32945895e-01 -9.85279202e-01 -8.10640529e-02 4.81756032e-01
4.42855060e-01 -6.14863276e-01 1.20988287e-01 -2.37141430e-01
-4.86853123e-01 1.32530046e+00 6.69691026e-01 2.30168924e-01
5.75632691e-01 6.28802538e-01 -1.22858793e-01 5.94307892e-02
-1.40994382e+00 2.20351398e-01 -5.46642184e-01 9.60580289e-01
3.97364318e-01 3.54506791e-01 6.12209458e-03 2.43112534e-01
9.59849879e-02 3.75073552e-01 8.17711830e-01 1.59892750e+00
-3.10586333e-01 -1.48593771e+00 -5.91015756e-01 8.33134472e-01
-2.79826015e-01 4.95897084e-01 -3.16098183e-01 8.05229485e-01
1.88897029e-01 7.96060681e-01 3.41294974e-01 -1.21747881e-01
5.47372222e-01 7.41359115e-01 4.37953025e-01 -7.61886537e-01
-8.33018720e-01 -7.22667649e-02 2.85091877e-01 -4.60236073e-01
-1.77834809e-01 -1.19722545e+00 -1.02828729e+00 -7.77649209e-02
1.65343642e-01 -3.32159139e-02 4.76273894e-01 9.42333519e-01
2.69027740e-01 8.89371514e-01 4.99891013e-01 -5.61437070e-01
-7.89359987e-01 -9.19081390e-01 -2.72449553e-01 6.66181862e-01
-6.09880965e-03 -4.62944329e-01 -1.40434608e-01 3.90315413e-01] | [8.587271690368652, 6.007868766784668] |
e2ffbc50-8f82-4872-8723-03a0987d834d | dualnet-locate-then-detect-effective-payload | 2010.12171 | null | https://arxiv.org/abs/2010.12171v1 | https://arxiv.org/pdf/2010.12171v1.pdf | DualNet: Locate Then Detect Effective Payload with Deep Attention Network | Network intrusion detection (NID) is an essential defense strategy that is used to discover the trace of suspicious user behaviour in large-scale cyberspace, and machine learning (ML), due to its capability of automation and intelligence, has been gradually adopted as a mainstream hunting method in recent years. However, traditional ML based network intrusion detection systems (NIDSs) are not effective to recognize unknown threats and their high detection rate often comes with the cost of high false alarms, which leads to the problem of alarm fatigue. To address the above problems, in this paper, we propose a novel neural network based detection system, DualNet, which is constructed with a general feature extraction stage and a crucial feature learning stage. DualNet can rapidly reuse the spatial-temporal features in accordance with their importance to facilitate the entire learning process and simultaneously mitigate several optimization problems occurred in deep learning (DL). We evaluate the DualNet on two benchmark cyber attack datasets, NSL-KDD and UNSW-NB15. Our experiment shows that DualNet outperforms classical ML based NIDSs and is more effective than existing DL methods for NID in terms of accuracy, detection rate and false alarm rate. | ['Hui Guo', 'Peilun Wu', 'Shiyi Yang'] | 2020-10-23 | null | null | null | null | ['deep-attention', 'deep-attention'] | ['computer-vision', 'natural-language-processing'] | [ 4.28836159e-02 -7.92763233e-01 -5.99101111e-02 -6.83208406e-02
1.72552496e-01 -3.02284092e-01 6.46566570e-01 2.20383629e-01
-6.64344370e-01 5.20902097e-01 -4.80405927e-01 -5.71799040e-01
-4.29859757e-01 -9.66275811e-01 8.39301385e-03 -4.88789201e-01
-1.91589281e-01 2.64648646e-01 7.27082193e-01 -2.41001874e-01
3.69177729e-01 1.01438391e+00 -1.01105928e+00 -6.28029602e-03
6.24393642e-01 1.27119327e+00 -3.41788918e-01 2.95103699e-01
-2.48534381e-01 8.43772352e-01 -8.84316385e-01 -1.17084511e-01
3.70623171e-01 -2.82433689e-01 -4.69993651e-01 -4.60414439e-01
-3.11072439e-01 -3.34947437e-01 -7.20171273e-01 1.16105664e+00
4.17782813e-01 2.33249620e-01 2.89849550e-01 -1.45797300e+00
-8.33927933e-03 1.68149337e-01 -8.87431502e-01 8.00970793e-01
-3.49916443e-02 3.07034522e-01 5.66530108e-01 -5.26716650e-01
1.82985678e-01 1.29652715e+00 4.98734832e-01 4.35477495e-01
-6.76198542e-01 -1.22795725e+00 2.30291113e-01 5.55671453e-01
-1.21348333e+00 -3.54211628e-02 8.74579549e-01 -1.32356599e-01
7.64662385e-01 1.30583510e-01 3.97670656e-01 1.22937739e+00
2.89781332e-01 6.44301176e-01 8.36241424e-01 -1.35977820e-01
2.12107286e-01 -5.04828580e-02 5.37960410e-01 7.37700820e-01
5.41538537e-01 4.44754899e-01 -8.52367058e-02 -1.87953964e-01
6.71190798e-01 5.84724963e-01 -8.47585499e-02 6.72578663e-02
-7.73813546e-01 9.84650671e-01 7.43405700e-01 5.43765545e-01
-3.50703448e-01 -3.33297461e-01 7.75538146e-01 4.47046965e-01
1.40923768e-01 5.27179599e-01 -4.49975818e-01 -1.97939217e-01
-4.23258871e-01 5.89201003e-02 8.36018324e-01 8.53686705e-02
3.89228433e-01 5.14810801e-01 1.45064801e-01 6.66116774e-01
1.16856031e-01 3.06087434e-01 5.65608144e-01 -6.64058030e-02
1.82477534e-01 1.19322419e+00 -3.34676057e-01 -1.46842110e+00
-7.86261201e-01 -7.21028030e-01 -1.25961018e+00 2.31590331e-01
7.47783929e-02 -4.22057509e-01 -8.60734582e-01 1.39891183e+00
3.75355422e-01 6.85082138e-01 -8.05873871e-02 6.31424129e-01
5.51178753e-01 7.22493947e-01 2.03141481e-01 -3.33020747e-01
1.11601019e+00 -4.65460807e-01 -5.68996966e-01 -3.04837048e-01
4.60686445e-01 -3.97460729e-01 5.92665315e-01 5.25986910e-01
-2.18264461e-01 -4.19141471e-01 -1.20225346e+00 8.70072961e-01
-7.63625979e-01 -2.48904541e-01 8.82697344e-01 7.63212264e-01
-2.28700981e-01 4.82809544e-01 -6.91383302e-01 -2.52676904e-01
6.47951961e-01 3.83639783e-01 -1.59686714e-01 -4.73363996e-02
-1.51005113e+00 7.45157301e-01 8.15017819e-01 1.67440772e-01
-9.34711754e-01 -4.29501176e-01 -4.88395572e-01 1.11325473e-01
7.78194249e-01 1.09572709e-01 8.42227161e-01 -5.64491987e-01
-1.15416014e+00 1.36408612e-01 4.20268744e-01 -6.12620533e-01
1.96233258e-01 -7.64786974e-02 -1.14019716e+00 2.67957952e-02
-2.75574207e-01 -6.17436022e-02 9.06060874e-01 -8.21632981e-01
-8.09557617e-01 -4.96823460e-01 7.52963871e-02 -1.22950427e-01
-6.90519452e-01 1.69807360e-01 -2.88463570e-02 -5.12916088e-01
1.44773483e-01 -5.74191451e-01 -3.43344569e-01 -5.12922406e-01
-5.03522754e-01 -3.28232914e-01 1.58068979e+00 -3.24491262e-01
1.49966168e+00 -1.97378302e+00 -2.86366314e-01 6.92483544e-01
3.51043403e-01 1.24008572e+00 -9.31908414e-02 1.87868059e-01
-2.04769447e-01 -6.36272207e-02 -2.76605099e-01 2.57827014e-01
-3.15852910e-01 1.33786023e-01 -4.42021191e-01 2.53793746e-01
1.83548853e-01 4.98528898e-01 -6.93252087e-01 -1.97222307e-01
3.82092446e-01 1.73207656e-01 -2.81901717e-01 3.43916416e-01
-8.97296593e-02 3.95632863e-01 -6.82060242e-01 8.57320309e-01
7.08912492e-01 -1.22585870e-01 -2.90999152e-02 -1.51730329e-01
1.09310290e-02 -6.25793934e-02 -1.15750408e+00 8.79904807e-01
-1.39571711e-01 4.29635406e-01 -7.63793364e-02 -1.31622207e+00
1.14464617e+00 2.12707371e-01 4.11549121e-01 -9.83796656e-01
4.60109383e-01 2.58264691e-01 5.24660587e-01 -4.23959225e-01
-2.45531470e-01 3.58562209e-02 -9.41345543e-02 5.52963316e-01
-1.98302239e-01 4.51668143e-01 7.98360705e-02 1.49659127e-01
1.51637185e+00 -7.34471262e-01 3.94764453e-01 7.97967911e-02
8.97583783e-01 -9.16908681e-02 8.58553767e-01 7.45896280e-01
-3.56903404e-01 -3.73626262e-01 5.93855381e-01 -9.28409100e-01
-4.34505194e-01 -9.12864625e-01 -1.51697800e-01 5.15458882e-01
2.00682834e-01 -1.96312577e-01 -3.49905699e-01 -1.18713582e+00
-9.73979309e-02 5.57984591e-01 -3.80630821e-01 -7.05864370e-01
-7.02985287e-01 -1.07405221e+00 8.42331171e-01 2.99923867e-01
1.15111136e+00 -1.42106366e+00 -5.36449313e-01 3.53321850e-01
3.42645198e-01 -1.13933003e+00 5.75437546e-02 2.30736002e-01
-6.82692051e-01 -1.57957053e+00 4.95266244e-02 -4.38361138e-01
4.81053233e-01 3.02019924e-01 4.18925613e-01 3.86768401e-01
-8.01368594e-01 -2.01835573e-01 -5.39652526e-01 -5.33761561e-01
-2.76245892e-01 2.01161042e-01 4.81054515e-01 1.87601656e-01
7.99217105e-01 -6.89986467e-01 -3.05412769e-01 3.08897227e-01
-1.22483397e+00 -5.09716690e-01 9.44449484e-01 7.95934558e-01
-6.94893897e-02 7.32239008e-01 8.67323279e-01 -9.00670946e-01
9.84474599e-01 -6.07897997e-01 -8.19413900e-01 9.72339809e-02
-5.44247508e-01 -2.31137127e-01 1.05338335e+00 -4.94722784e-01
-7.86596537e-01 -4.83939677e-01 -2.71463305e-01 -3.70438159e-01
-3.24799180e-01 6.03634834e-01 -2.56255656e-01 -4.22874182e-01
6.23451352e-01 2.95696110e-01 5.80213331e-02 -5.11018574e-01
-3.77247661e-01 6.03561640e-01 1.83794767e-01 -1.99919581e-01
1.14203656e+00 3.81270535e-02 3.69986266e-01 -8.78036559e-01
-8.77120495e-01 -4.43020254e-01 -4.38528299e-01 -5.23060486e-02
4.94132936e-01 -2.97069907e-01 -1.08532166e+00 6.35209322e-01
-1.11115861e+00 2.75741071e-01 2.83834726e-01 4.67348248e-01
3.23744446e-01 4.09883410e-01 -7.41932452e-01 -8.44496846e-01
-4.49388862e-01 -9.90534186e-01 8.55888575e-02 4.94387448e-01
1.47382632e-01 -8.87547851e-01 7.45566655e-03 -3.32933158e-01
6.68823898e-01 4.33145642e-01 1.01365376e+00 -1.34359384e+00
-3.93951535e-01 -7.43181109e-01 -6.79956377e-01 4.78480607e-01
3.00702244e-01 -5.85938506e-02 -6.07647419e-01 -4.82450247e-01
2.48501897e-01 -2.57344395e-01 7.66457558e-01 1.22196786e-02
1.20237029e+00 -3.49077255e-01 -4.72904801e-01 5.21818876e-01
1.48126245e+00 1.01296699e+00 4.57780510e-01 6.09903514e-01
6.32845640e-01 2.85987526e-01 4.66521919e-01 5.80605388e-01
-3.28864336e-01 3.64909291e-01 7.33596802e-01 -1.41437694e-01
4.93505180e-01 6.89381436e-02 2.10523859e-01 5.75002909e-01
2.15927169e-01 -1.38191208e-01 -1.04843342e+00 -1.15726568e-01
-1.74611700e+00 -1.05391729e+00 1.71826094e-01 1.99125099e+00
2.97253072e-01 7.42411673e-01 2.08666682e-01 4.97968048e-01
6.36565626e-01 3.03356111e-01 -6.82968915e-01 -3.56322914e-01
6.00275807e-02 3.05224299e-01 2.78524369e-01 2.73019411e-02
-1.30389404e+00 7.54433334e-01 5.35426283e+00 9.18502927e-01
-1.24830985e+00 -7.91010112e-02 5.80858827e-01 1.14260748e-01
5.01416743e-01 -1.71170995e-01 -8.10745656e-01 4.55764562e-01
9.67721045e-01 1.12919152e-01 3.44106704e-01 8.60933661e-01
6.05739653e-02 2.21330762e-01 -5.28460860e-01 8.93630445e-01
-8.16371962e-02 -1.21550047e+00 2.48702168e-01 -4.52364981e-02
1.84456483e-01 -1.40798107e-01 1.37391193e-02 6.61063075e-01
2.70743877e-01 -9.72370148e-01 -1.93582535e-01 1.12227440e-01
3.18315119e-01 -1.27046537e+00 1.10413599e+00 5.22313297e-01
-1.22918880e+00 -6.56600118e-01 -3.20354462e-01 -1.12681888e-01
5.23227677e-02 6.11330450e-01 -7.32212126e-01 6.58148646e-01
5.99615037e-01 7.24115014e-01 -5.26724517e-01 1.21678674e+00
-1.34860918e-01 6.87455416e-01 -2.80790418e-01 -3.15212846e-01
6.98025942e-01 -4.96909022e-02 7.36373365e-01 9.34404552e-01
-1.59392767e-02 1.63740098e-01 6.37083411e-01 4.93234962e-01
8.76239166e-02 -7.61913881e-02 -6.25628471e-01 -2.76786029e-01
6.57662868e-01 1.33643579e+00 -8.85230243e-01 -7.26020485e-02
-2.19439954e-01 4.62695032e-01 6.69446513e-02 8.70592669e-02
-8.87516975e-01 -9.87549245e-01 7.03242660e-01 -8.88659135e-02
-1.61698908e-02 -2.35102400e-01 -5.15889488e-02 -1.02218723e+00
-3.17776054e-01 -1.11878574e+00 6.99192107e-01 1.21613473e-01
-1.22821128e+00 9.64732647e-01 -1.33649766e-01 -1.28490329e+00
6.81679472e-02 -9.18726563e-01 -1.02829933e+00 5.34062684e-01
-1.33310461e+00 -7.88989246e-01 -3.29541087e-01 8.43502343e-01
5.43820500e-01 -6.62335396e-01 7.21102297e-01 5.37741423e-01
-1.41744864e+00 5.29675841e-01 -2.32410222e-01 5.87123215e-01
3.20251405e-01 -6.41706824e-01 3.25590551e-01 1.14091599e+00
-1.75683379e-01 6.43348575e-01 2.82545447e-01 -7.43878007e-01
-1.06844163e+00 -9.86178100e-01 1.27309322e-01 4.63071465e-02
8.08924377e-01 -1.66160882e-01 -1.00940025e+00 4.77810651e-01
-2.98542351e-01 -6.30107196e-03 5.93533993e-01 7.38815889e-02
-4.00075614e-01 -4.04695600e-01 -1.27118063e+00 6.17073536e-01
6.40207350e-01 -1.96703404e-01 -3.67886871e-01 2.06444532e-01
9.03082490e-01 4.44421731e-02 -4.40786064e-01 7.24004686e-01
1.69211656e-01 -9.48927283e-01 9.69726086e-01 -8.07908416e-01
-2.63190538e-01 -3.32660377e-01 1.38947636e-01 -9.70122635e-01
-4.42423373e-01 -3.35333496e-01 -2.76895106e-01 1.02624273e+00
9.99384597e-02 -1.05838323e+00 7.25465298e-01 6.03417791e-02
2.11758688e-02 -8.58169317e-01 -1.00678205e+00 -8.82638335e-01
-5.17595351e-01 -4.66129035e-01 6.77769363e-01 1.10078442e+00
-1.49813324e-01 3.04086357e-01 -3.88136059e-01 2.53644109e-01
6.96768403e-01 -1.51385665e-01 5.04722893e-01 -1.70780969e+00
-1.08672641e-01 -6.45417571e-01 -6.45391941e-01 -5.30280352e-01
7.89847076e-02 -4.81622279e-01 -4.29416448e-01 -9.41772997e-01
-2.27909863e-01 -5.70409119e-01 -7.83026874e-01 6.52544081e-01
1.78597942e-02 1.41605781e-02 -1.28777087e-01 2.01068431e-01
-6.33693516e-01 3.49479735e-01 8.39024961e-01 -1.21017009e-01
-2.76983410e-01 3.30845684e-01 -4.12277132e-01 9.52822804e-01
9.61795866e-01 -4.54547048e-01 -3.69908541e-01 1.17013752e-01
-2.51696110e-01 -5.65976612e-02 2.21223131e-01 -1.32116175e+00
3.67882013e-01 -3.47075254e-01 6.49461925e-01 -5.03851831e-01
2.03062966e-01 -9.07915473e-01 -3.00104141e-01 1.14360666e+00
-1.80537663e-02 3.22925627e-01 4.48919743e-01 6.63167119e-01
-3.32842529e-01 -9.77053866e-02 9.85248864e-01 -3.58354636e-02
-1.02151263e+00 8.13391328e-01 -3.35817397e-01 -2.16931030e-01
1.20342374e+00 -6.27605990e-02 -4.42903876e-01 8.50448199e-03
-2.06951991e-01 2.17339188e-01 -1.98074758e-01 6.34792805e-01
1.02069521e+00 -1.20063877e+00 -4.12178099e-01 6.32067859e-01
-4.20959368e-02 -2.81291753e-01 2.85052150e-01 8.11683118e-01
-5.71422637e-01 5.43056250e-01 -3.93914312e-01 -2.29224041e-01
-1.05949259e+00 6.26291513e-01 3.21625143e-01 -7.16404140e-01
-6.40655637e-01 7.41574049e-01 -1.41384825e-01 -3.65554601e-01
5.28719604e-01 4.23527807e-01 -5.03194392e-01 -1.39353126e-01
8.36047292e-01 4.06925529e-01 6.46474287e-02 -2.78869420e-01
-6.20313346e-01 1.23229660e-01 -6.22174144e-01 4.52442169e-01
1.14915729e+00 4.92048681e-01 -2.35608175e-01 -4.80738934e-03
9.82799768e-01 -4.66027915e-01 -7.15711176e-01 -3.87219846e-01
4.89634961e-01 -4.71902043e-01 4.20174971e-02 -8.54817510e-01
-1.15744281e+00 8.22588921e-01 6.66476429e-01 4.74794686e-01
1.31434572e+00 -6.59267426e-01 1.16324580e+00 7.21751213e-01
3.53791475e-01 -7.61612713e-01 4.47168797e-01 8.52699816e-01
2.67995566e-01 -1.22201455e+00 -9.86380950e-02 -1.12677872e-01
-3.64952207e-01 1.33298445e+00 1.14770067e+00 -2.95368642e-01
9.28277552e-01 2.45167777e-01 -1.70492396e-01 -3.56076896e-01
-5.15664041e-01 1.47606000e-01 7.08909631e-02 4.49122995e-01
-1.23303220e-01 -1.99273899e-01 -1.57600939e-01 5.10563970e-01
3.44622403e-01 -4.80355918e-01 1.88916162e-01 1.12700331e+00
-7.73343146e-01 -1.05170619e+00 -2.53035426e-01 7.10773647e-01
-4.67149466e-01 1.01763926e-01 -2.80951619e-01 9.38591301e-01
1.44567475e-01 9.86977816e-01 -5.12122698e-02 -9.52098012e-01
3.81974369e-01 -2.06910819e-01 -1.63927376e-01 -4.54793423e-01
-5.45326710e-01 -4.25066411e-01 -2.15883315e-01 -6.78379238e-01
9.67011526e-02 -1.22813895e-01 -1.23673081e+00 -5.30889571e-01
-4.01754737e-01 3.84251416e-01 5.15589893e-01 1.19444585e+00
1.78222716e-01 8.64377856e-01 9.66034949e-01 -3.38464379e-01
-7.21229672e-01 -8.91792536e-01 -6.34555101e-01 2.46679559e-01
1.39847800e-01 -9.13417220e-01 -5.20713091e-01 -9.70401943e-01] | [5.239187717437744, 7.1727190017700195] |
0de56492-fc1f-4185-bd9c-3437029323c0 | bjtu-wechat-s-systems-for-the-wmt22-chat | 2211.15009 | null | https://arxiv.org/abs/2211.15009v1 | https://arxiv.org/pdf/2211.15009v1.pdf | BJTU-WeChat's Systems for the WMT22 Chat Translation Task | This paper introduces the joint submission of the Beijing Jiaotong University and WeChat AI to the WMT'22 chat translation task for English-German. Based on the Transformer, we apply several effective variants. In our experiments, we utilize the pre-training-then-fine-tuning paradigm. In the first pre-training stage, we employ data filtering and synthetic data generation (i.e., back-translation, forward-translation, and knowledge distillation). In the second fine-tuning stage, we investigate speaker-aware in-domain data generation, speaker adaptation, prompt-based context modeling, target denoising fine-tuning, and boosted self-COMET-based model ensemble. Our systems achieve 0.810 and 0.946 COMET scores. The COMET scores of English-German and German-English are the highest among all submissions. | ['Jie zhou', 'Yufeng Chen', 'Jinan Xu', 'Fandong Meng', 'Yunlong Liang'] | 2022-11-28 | null | null | null | null | ['synthetic-data-generation', 'synthetic-data-generation'] | ['medical', 'miscellaneous'] | [ 1.05193086e-01 -1.79793268e-01 5.91411367e-02 -4.52163517e-01
-1.43807530e+00 -6.22463942e-01 7.73820221e-01 -3.34822029e-01
-6.23838603e-01 1.21333385e+00 5.45194447e-01 -4.58102763e-01
2.49632820e-01 -4.91228044e-01 -6.18839383e-01 -3.04866821e-01
5.76135397e-01 8.36095273e-01 -2.50118300e-02 -6.01293981e-01
-6.46592816e-03 -1.49831787e-01 -9.11903799e-01 6.92461789e-01
1.35537398e+00 6.46681070e-01 3.46904784e-01 8.82413924e-01
-3.15037489e-01 3.58173579e-01 -9.11733091e-01 -7.08067834e-01
-9.13377404e-02 -7.44575322e-01 -1.02128363e+00 -2.62927145e-01
-2.02417914e-02 -1.56080976e-01 -4.56464887e-02 7.87164688e-01
9.75507200e-01 1.91967309e-01 4.92470413e-01 -9.40810502e-01
-8.88064563e-01 1.15993631e+00 -2.10583314e-01 1.53165713e-01
5.37181497e-01 1.78799659e-01 5.69579720e-01 -1.28014874e+00
6.22372150e-01 1.26638389e+00 4.41666543e-01 1.11818397e+00
-1.07922697e+00 -7.49376237e-01 7.74572939e-02 3.47503573e-01
-1.30413687e+00 -7.13932335e-01 6.89007580e-01 -3.18039060e-01
1.03582168e+00 3.50787431e-01 2.46776953e-01 1.61155283e+00
-7.44269565e-02 7.72005856e-01 1.20916283e+00 -8.29542160e-01
1.51700526e-01 3.85787636e-01 -2.70827949e-01 1.03836238e-01
-3.86036426e-01 1.49288058e-01 -7.85653830e-01 -6.50982931e-02
5.02187312e-01 -5.66356361e-01 -3.31182368e-02 6.26342654e-01
-1.66857529e+00 6.16802990e-01 4.37637307e-02 2.71443427e-01
-4.73737359e-01 -1.67736635e-01 3.91295135e-01 6.52129591e-01
5.83022177e-01 4.54980701e-01 -8.05313587e-01 -5.37620485e-01
-1.07716370e+00 4.04916704e-01 8.81042123e-01 1.36775231e+00
3.68582845e-01 7.60177076e-02 -7.24153280e-01 1.28495657e+00
1.27537176e-01 6.44939840e-01 9.73382831e-01 -6.91651762e-01
1.01563323e+00 7.56001920e-02 2.08567694e-01 -5.63939176e-02
-1.41089723e-01 -2.91690648e-01 -8.60821307e-01 -5.24624109e-01
1.56849951e-01 -7.71452487e-01 -1.02718747e+00 1.89203286e+00
3.20214391e-01 1.31164193e-01 5.42624116e-01 7.88631916e-01
9.94389117e-01 8.61015975e-01 2.28853494e-01 -3.86551917e-01
1.29643345e+00 -1.58438182e+00 -9.45424497e-01 -2.14572236e-01
5.17879725e-01 -1.30988657e+00 1.32217729e+00 5.50433248e-02
-1.26422071e+00 -8.38553786e-01 -7.10490227e-01 -2.98192352e-02
-2.92998374e-01 3.85139465e-01 4.83665951e-02 4.77652282e-01
-8.75259578e-01 4.37548876e-01 -6.33379698e-01 -4.74355012e-01
-9.96932387e-02 5.20505682e-02 -1.70814171e-01 -1.14032282e-02
-1.71858001e+00 1.09231174e+00 3.64156842e-01 -3.48932445e-02
-8.77623439e-01 -7.89022088e-01 -3.04721534e-01 -2.33069584e-01
1.35631263e-01 -9.64055598e-01 1.72041142e+00 -8.67729247e-01
-2.26311350e+00 7.02706158e-01 -3.37040633e-01 -4.81412202e-01
7.79454172e-01 -3.82371306e-01 -9.01586354e-01 -2.98659891e-01
2.16578264e-02 5.79395413e-01 7.00854599e-01 -9.56421375e-01
-8.29379082e-01 5.09935357e-02 -2.51974076e-01 3.99164796e-01
-2.53204733e-01 3.94563943e-01 -3.97168756e-01 -8.55279446e-01
-2.94843286e-01 -1.06119382e+00 -2.20438406e-01 -9.38752830e-01
-5.88670433e-01 -2.31541440e-01 5.85578561e-01 -1.13773358e+00
1.59146655e+00 -2.11592960e+00 1.57961547e-01 -1.27101332e-01
-3.26519907e-01 4.65029538e-01 -4.88465607e-01 6.51549459e-01
1.53938070e-01 1.42816275e-01 -1.12785257e-01 -5.77362239e-01
9.82176214e-02 1.94151953e-01 -3.39856565e-01 -3.90684724e-01
4.97139663e-01 1.00861466e+00 -8.67953956e-01 -4.09631163e-01
-3.70326713e-02 2.30069339e-01 -5.51266432e-01 5.80971897e-01
-2.66464740e-01 7.00738490e-01 -3.42506438e-01 2.80113846e-01
5.37525475e-01 3.15238327e-01 8.63348544e-02 1.42521918e-01
-2.51457721e-01 8.39037120e-01 -9.47004318e-01 1.90600145e+00
-7.58904576e-01 4.07184362e-01 -1.91336963e-02 -4.21719193e-01
1.04563522e+00 6.62040174e-01 -1.04668878e-01 -6.80683613e-01
-1.20624498e-01 6.86897039e-01 8.83520767e-02 -5.37877917e-01
9.08790171e-01 -2.33283550e-01 -2.85149157e-01 4.02726054e-01
3.43477190e-01 -2.67928541e-01 2.73099780e-01 5.97293042e-02
7.13385284e-01 3.60990286e-01 3.45768258e-02 -1.59984782e-01
4.90072221e-01 1.27818257e-01 5.71842551e-01 5.44418752e-01
-6.76427782e-02 8.11049342e-01 5.31141795e-02 3.86394821e-02
-1.04510832e+00 -9.69783425e-01 2.34157443e-01 1.44981253e+00
-1.78254426e-01 -4.29383427e-01 -1.35052979e+00 -6.51894510e-01
-4.43206757e-01 1.06678474e+00 -4.02173221e-01 -3.70708019e-01
-8.50104213e-01 -6.00659966e-01 8.42384338e-01 4.23724443e-01
7.85500646e-01 -1.06825268e+00 2.56760597e-01 6.22130275e-01
-9.73908842e-01 -1.16701066e+00 -9.88741219e-01 1.12096176e-01
-6.84816658e-01 -1.39708370e-01 -1.06239629e+00 -9.57250059e-01
8.35757181e-02 -1.58538952e-01 1.17978930e+00 -3.93602908e-01
3.58302742e-01 -1.77707583e-01 -5.11832595e-01 -4.52015936e-01
-9.67303276e-01 7.26773202e-01 2.54342616e-01 -5.89367934e-02
4.84769285e-01 -5.41534424e-01 -3.30800742e-01 4.97479498e-01
-3.35940659e-01 2.29781315e-01 8.49910080e-01 1.14095902e+00
5.79930663e-01 -6.08234882e-01 1.01902354e+00 -8.36436987e-01
1.17397034e+00 -4.02319729e-01 -1.89084709e-01 6.45346820e-01
-6.05622828e-01 4.40118276e-02 7.16039240e-01 -7.32090950e-01
-1.49922538e+00 -1.78013682e-01 -4.48822439e-01 -2.53666550e-01
-2.91398734e-01 5.95085084e-01 -3.59810799e-01 3.85172069e-01
8.85855198e-01 4.28907454e-01 -3.52889597e-01 -6.44059420e-01
5.93779325e-01 1.30783701e+00 6.30082250e-01 -8.42303038e-01
6.59915566e-01 -3.95990431e-01 -8.56448650e-01 -6.61345363e-01
-6.56180263e-01 -1.61824584e-01 -6.41745627e-01 -6.66897884e-03
8.58705223e-01 -1.15360665e+00 -1.82079375e-01 5.23637772e-01
-1.44363189e+00 -4.84741807e-01 -4.37156171e-01 6.03193164e-01
-5.77989101e-01 1.29724130e-01 -6.75157428e-01 -6.23103142e-01
-7.41769493e-01 -1.07910228e+00 9.90217924e-01 3.18424314e-01
-4.50745374e-01 -8.34963441e-01 3.17633569e-01 6.00873232e-01
7.84246743e-01 -1.49891421e-01 5.92336535e-01 -9.04670775e-01
-1.80802181e-01 4.75882292e-02 1.87227905e-01 6.24843657e-01
-1.03235424e-01 -6.15926012e-02 -1.05042887e+00 -9.44403037e-02
-2.99006760e-01 -3.91258240e-01 5.71752369e-01 6.52366504e-02
5.63137293e-01 -4.33713883e-01 -3.26373428e-02 4.14807737e-01
7.85306334e-01 2.49222890e-01 5.84913671e-01 2.09331632e-01
3.93923610e-01 3.33609819e-01 6.72271132e-01 2.88379550e-01
7.50921726e-01 8.73690426e-01 -4.54529971e-01 1.53287202e-01
-4.63378489e-01 -5.42662740e-01 7.28603005e-01 1.47062922e+00
-2.90446132e-01 -2.56902486e-01 -8.16478372e-01 6.26621306e-01
-1.90476871e+00 -8.72414112e-01 -7.28979111e-02 1.99666727e+00
1.42758679e+00 9.61518735e-02 3.30411255e-01 -2.53459185e-01
8.92569423e-01 -2.55397201e-01 -2.64867395e-01 -7.55863905e-01
-2.48845652e-01 3.15280676e-01 5.03109880e-02 7.34995604e-01
-8.18538725e-01 1.47181869e+00 5.81090784e+00 1.06990838e+00
-1.08755982e+00 5.14424860e-01 4.08563644e-01 -1.20898403e-01
-3.21393490e-01 -1.70822963e-02 -9.74522531e-01 5.95613837e-01
1.48385251e+00 -3.67620438e-01 7.63509035e-01 6.75298989e-01
2.07293689e-01 3.04930538e-01 -9.30066645e-01 7.83668458e-01
-1.21002741e-01 -1.21860015e+00 1.15116455e-01 -2.71513939e-01
1.09031391e+00 2.48327613e-01 -2.20140710e-01 8.09796691e-01
4.41168219e-01 -7.72045612e-01 9.49088693e-01 4.65998143e-01
9.30280089e-01 -7.09446669e-01 6.89694166e-01 5.51522195e-01
-8.29077661e-01 6.39654696e-02 -1.70746222e-01 1.33826524e-01
4.17024463e-01 6.05787337e-01 -1.13766837e+00 8.51589859e-01
3.60635519e-01 1.07755870e-01 -1.61511436e-01 7.27228999e-01
-5.33330083e-01 1.13196433e+00 -3.13830376e-01 -2.73492992e-01
3.84782106e-02 8.57021008e-03 5.54215431e-01 1.71460426e+00
4.42231804e-01 -4.96465303e-02 1.47692412e-01 6.91199481e-01
-3.48435342e-01 1.05923675e-01 8.87569562e-02 -1.22347837e-02
7.86315858e-01 1.07364416e+00 1.88410729e-01 -5.77973008e-01
-6.43320903e-02 1.30538750e+00 2.75217831e-01 6.07167184e-01
-8.05547714e-01 -6.11960590e-01 7.22966373e-01 -2.61223644e-01
7.54704252e-02 -1.77865714e-01 -4.01202440e-01 -1.30643129e+00
2.88439691e-02 -1.03382909e+00 1.22928612e-01 -6.65896595e-01
-1.29458702e+00 1.01679873e+00 -5.74735701e-02 -1.08643389e+00
-7.09810376e-01 -1.45139784e-01 -7.06438482e-01 1.38096321e+00
-1.49712729e+00 -1.34259355e+00 8.40225294e-02 7.49515772e-01
8.40101123e-01 -3.77557933e-01 1.08927929e+00 5.61401427e-01
-5.98240316e-01 1.08803630e+00 3.56772691e-01 2.27994874e-01
1.03224766e+00 -1.11304462e+00 8.36428106e-01 7.70950496e-01
6.11264110e-02 5.75597286e-01 5.83780468e-01 -5.87588787e-01
-9.93508399e-01 -1.28600836e+00 1.57551622e+00 -5.61469138e-01
4.99776155e-01 -5.28831244e-01 -7.04517841e-01 5.35819173e-01
5.72643518e-01 -4.68231529e-01 5.92084289e-01 2.21993253e-01
-1.41703680e-01 -1.59516588e-01 -9.37131524e-01 7.17542529e-01
8.96984041e-01 -5.31477690e-01 -7.77956963e-01 3.21621537e-01
1.05351460e+00 -7.27965474e-01 -9.69381869e-01 1.20185085e-01
4.60525632e-01 -2.59237051e-01 7.35584378e-01 -8.82185459e-01
4.93092835e-01 -8.55232626e-02 -3.13948780e-01 -1.88199592e+00
-3.54329050e-01 -1.23261619e+00 3.10526416e-02 1.72497833e+00
1.03938854e+00 -4.03773874e-01 2.98123091e-01 3.36578667e-01
-4.81276929e-01 -4.35024947e-01 -1.18069303e+00 -8.84524763e-01
4.54400152e-01 -3.38315815e-01 9.34819460e-01 1.03765404e+00
1.72251612e-01 9.06270444e-01 -6.35330975e-01 -1.57129958e-01
3.15206766e-01 1.88187452e-03 1.00074399e+00 -7.17354894e-01
-4.21935856e-01 -3.32137346e-01 3.59289199e-01 -1.01569760e+00
-1.02884740e-01 -7.84189105e-01 2.01351672e-01 -1.53782368e+00
-4.88225184e-02 -2.18981653e-01 -1.83174923e-01 4.06777769e-01
-5.65733254e-01 8.23795944e-02 2.82780379e-01 2.21505806e-01
-3.77274901e-01 6.03759706e-01 1.43621731e+00 7.87495747e-02
-4.94666845e-01 2.46836111e-01 -7.43050694e-01 2.22227097e-01
1.05096030e+00 -4.24335361e-01 -1.28766075e-01 -7.04443455e-01
-2.83180118e-01 5.87164648e-02 -2.10447647e-02 -8.91373932e-01
1.43474922e-01 -3.75196010e-01 9.91360620e-02 -2.96319038e-01
3.17049354e-01 -3.27791214e-01 1.71532452e-01 2.43279755e-01
-6.01527393e-01 2.48307083e-02 2.64571905e-01 1.70690402e-01
-4.04686719e-01 5.30774817e-02 7.93543816e-01 -1.08850628e-01
-4.20798540e-01 -5.96339889e-02 -4.58528131e-01 4.19610083e-01
6.62757635e-01 7.36487657e-02 -3.83040279e-01 -3.99322599e-01
-8.54139209e-01 2.79355228e-01 -2.20036075e-01 7.62702525e-01
4.10178304e-01 -1.63986111e+00 -1.38701463e+00 2.66918063e-01
1.44441605e-01 -2.09189922e-01 4.51279134e-01 8.73957276e-01
-7.17842951e-02 3.16455305e-01 -5.02308132e-03 -3.68009120e-01
-1.08421040e+00 1.12250723e-01 1.20825335e-01 -6.25499249e-01
-3.32787521e-02 1.03367209e+00 -2.55886525e-01 -9.56181347e-01
1.03420906e-01 -2.30109453e-01 6.38002679e-02 -1.66805387e-01
5.68913877e-01 3.54602993e-01 3.12401503e-01 -3.67559791e-01
-3.78193706e-01 1.39629304e-01 -3.06687355e-01 -7.82754421e-01
9.63258922e-01 -2.25912467e-01 1.52869463e-01 3.80885035e-01
7.84561217e-01 1.46143824e-01 -8.32482994e-01 -5.03588140e-01
9.82888639e-02 -4.76820171e-02 -2.25544527e-01 -1.57108784e+00
-4.38436061e-01 1.10658634e+00 2.94243366e-01 -6.52550310e-02
1.13556826e+00 -2.08172008e-01 1.14954257e+00 5.11748731e-01
2.23006159e-01 -1.60705924e+00 -2.42209092e-01 1.14474893e+00
1.13400221e+00 -1.05399930e+00 -6.40960455e-01 -1.67984188e-01
-1.14234269e+00 9.11166251e-01 7.83996642e-01 3.51157337e-01
3.43781203e-01 2.39945009e-01 4.97700781e-01 7.07705736e-01
-1.25097036e+00 -4.33474891e-02 2.41346970e-01 6.90651178e-01
6.70828283e-01 3.90784472e-01 -4.27268237e-01 1.08026087e+00
-6.26464367e-01 3.35489959e-01 2.52198607e-01 6.09144628e-01
-3.26692969e-01 -1.29187107e+00 -3.80089492e-01 -1.51264695e-02
-3.81676227e-01 -4.18183178e-01 -6.84031963e-01 5.12778819e-01
-6.04865737e-02 1.23265290e+00 -3.29876184e-01 -8.02295208e-01
8.62150073e-01 5.19069791e-01 2.43290529e-01 -5.24217069e-01
-1.16112864e+00 3.71757925e-01 6.07004464e-01 -1.44966707e-01
-2.11556986e-01 -5.32744586e-01 -9.47551191e-01 -3.73029143e-01
-5.36772370e-01 6.32935047e-01 8.09144974e-01 9.18814719e-01
7.22495198e-01 6.81458712e-01 6.67610884e-01 -5.31595826e-01
-7.48119831e-01 -1.68668973e+00 -1.24342062e-01 2.06349239e-01
1.25266716e-01 -1.38772845e-01 -7.55155459e-02 2.69425184e-01] | [14.343335151672363, 7.293102264404297] |
26c614ab-45b8-4a21-922c-89d59160ddf7 | machine-learning-approach-of-automatic | null | null | https://ieeexplore.ieee.org/abstract/document/8822896 | https://ieeexplore.ieee.org/abstract/document/8822896 | Machine learning approach of automatic identification and counting of blood cells | A complete blood cell count is an important test in medical diagnosis to evaluate overall health condition. Traditionally blood cells are counted manually using haemocytometer along with other laboratory equipment’s and chemical compounds, which is a time-consuming and tedious task. In this work, the authors present a machine learning approach for automatic identification and counting of three types of blood cells using ‘you only look once’ (YOLO) object detection and classification algorithm. YOLO framework has been trained with a modified configuration BCCD Dataset of blood smear images to automatically identify and count red blood cells, white blood cells, and platelets. Moreover, this study with other convolutional neural network architectures considering architecture complexity, reported accuracy, and running time with this framework and compare the accuracy of the models for blood cells detection. They also tested the trained model on smear images from a different dataset and found that the learned models are generalized. Overall the computer-aided system of detection and counting enables us to count blood cells from smear images in less than a second, which is useful for practical applications. | ['Mohammad Tariqul Islam', 'Mohammad Mahmudul Alam'] | 2019-09-05 | null | null | null | healthcare-technology-letters-iet-2019-9 | ['blood-cell-count', 'cbc-test', 'blood-cell-detection'] | ['computer-vision', 'computer-vision', 'medical'] | [-2.85869271e-01 -1.78321660e-01 1.93232462e-01 5.08937202e-02
-8.55425596e-02 -3.92728746e-01 2.84559518e-01 8.78212452e-01
-8.65863204e-01 8.19182217e-01 -3.32588017e-01 -1.84972033e-01
3.14700246e-01 -1.03054380e+00 1.45440295e-01 -8.71471882e-01
1.77409887e-01 1.09892368e+00 6.29656240e-02 3.05198640e-01
2.99428433e-01 1.07193303e+00 -1.31144178e+00 2.60573119e-01
3.85563940e-01 9.46861386e-01 -3.49558204e-01 1.37377214e+00
-3.77447575e-01 1.31165314e+00 -9.07640278e-01 -2.85987526e-01
-1.88612640e-02 -3.28845859e-01 -5.03736198e-01 2.01613471e-01
3.77129853e-01 -5.91476202e-01 3.15398663e-01 4.90846962e-01
7.55431116e-01 -4.05627400e-01 1.10924244e+00 -6.64837718e-01
-5.61085820e-01 3.27123664e-02 -4.46102649e-01 5.94690323e-01
8.31112191e-02 2.62035400e-01 3.10613401e-02 -7.38624752e-01
1.97576895e-01 1.09284806e+00 8.82276177e-01 5.14209628e-01
-8.77034843e-01 -4.50047761e-01 -1.01681542e+00 1.90276541e-02
-1.30567098e+00 -2.01517805e-01 -7.44179934e-02 -7.06473887e-01
1.28497660e+00 3.35913837e-01 7.27438569e-01 1.55529439e-01
4.29041147e-01 6.19734943e-01 9.61221397e-01 -7.68918514e-01
3.51397872e-01 6.95511520e-01 4.54578668e-01 9.53320086e-01
1.07702601e+00 -2.01636180e-01 -5.01036979e-02 8.97992253e-02
8.40088487e-01 1.50580943e-01 4.02086794e-01 3.37350249e-01
-7.57303298e-01 7.71279216e-01 5.76774441e-02 1.01423764e+00
-1.73518479e-01 1.87713746e-02 6.26886785e-01 -1.01815008e-01
1.71534434e-01 5.96974134e-01 -2.37861246e-01 1.54896244e-01
-1.03205895e+00 -3.31466973e-01 1.00080168e+00 4.32574451e-01
4.83026773e-01 4.87071946e-02 -5.02187431e-01 7.65948892e-01
-1.99809819e-01 5.04605770e-01 5.55538297e-01 -5.16925633e-01
-4.01870549e-01 1.01978791e+00 1.37279540e-01 -1.01737881e+00
-7.95514464e-01 -2.12268844e-01 -1.06988227e+00 2.32517481e-01
1.04546595e+00 -1.24586582e-01 -8.92790437e-01 6.44452095e-01
4.13225114e-01 4.68095206e-02 -6.04192987e-02 7.30902076e-01
1.14184761e+00 2.63245314e-01 2.85364032e-01 3.62390815e-03
1.62366569e+00 -8.96121323e-01 -6.83936954e-01 3.86516809e-01
1.03156495e+00 -6.90496624e-01 4.63786811e-01 5.77662408e-01
-9.43800926e-01 -6.52154148e-01 -1.16271305e+00 -3.80735517e-01
-8.67466331e-01 8.04300070e-01 8.69390190e-01 1.29038882e+00
-9.51361001e-01 4.58142728e-01 -8.20078194e-01 -6.64617002e-01
7.00705945e-01 5.64596236e-01 -3.93188894e-01 2.38344803e-01
-4.68890339e-01 9.26242888e-01 2.78317541e-01 1.44181430e-01
-3.54113162e-01 -5.53876877e-01 -6.62499726e-01 4.17631269e-02
-1.80500194e-01 -5.47073603e-01 9.45123374e-01 -2.19936848e-01
-1.32436025e+00 1.30018497e+00 -1.44639090e-01 -6.37461066e-01
5.54516017e-01 -1.68903414e-02 -9.08935890e-02 5.75468957e-01
1.18933134e-02 3.97978097e-01 8.38972926e-02 -4.85004514e-01
-9.51604009e-01 -3.74129057e-01 -4.48446572e-01 -4.32250559e-01
4.37439792e-03 -1.38388529e-01 -2.82283157e-01 -1.45859897e-01
-3.31519246e-01 -1.85683057e-01 9.25235301e-02 6.59270957e-02
-1.30314246e-01 -5.39963543e-01 3.41919273e-01 -6.37454271e-01
8.89079571e-01 -1.68964767e+00 -7.04732716e-01 7.08405599e-02
3.71493340e-01 6.80997491e-01 3.14814955e-01 1.02323629e-01
3.35404456e-01 1.86024040e-01 9.84280109e-02 -3.26665461e-01
-1.87784657e-01 -4.03320044e-02 3.04572821e-01 5.49713969e-01
4.18682963e-01 7.93604553e-01 -7.02492893e-01 -9.91084278e-01
8.37496161e-01 9.12898958e-01 -2.07074299e-01 1.14058949e-01
3.49967062e-01 4.68614131e-01 -1.09177576e-02 1.04880059e+00
6.20408952e-01 -5.05670488e-01 -1.14512235e-01 -3.75687301e-01
3.91426384e-02 -4.74352419e-01 -8.75184834e-01 6.32685840e-01
-4.20929193e-01 7.25765586e-01 -2.41415545e-01 -8.92618299e-01
1.03324747e+00 3.38611901e-01 3.05725843e-01 -7.20451176e-01
8.64743829e-01 3.55509222e-01 -5.64789735e-02 -9.03714478e-01
-1.05130719e-02 -2.92441040e-01 4.56227541e-01 4.71037030e-01
1.60413489e-01 -4.40121181e-02 7.62646377e-01 -3.55870388e-02
8.63844633e-01 -3.76305670e-01 8.18637490e-01 -3.32613587e-01
1.10545790e+00 1.03594154e-01 8.66806135e-02 7.98770547e-01
-5.27317047e-01 4.55534399e-01 3.37886125e-01 -1.17219746e+00
-8.62419903e-01 -7.02106595e-01 -5.28576970e-01 7.09234893e-01
-3.57418329e-01 4.66657430e-01 -1.06253779e+00 -1.94536597e-01
2.21303239e-01 3.70921493e-01 -1.00009429e+00 4.37971830e-01
-7.01704204e-01 -1.01475215e+00 8.47658515e-01 5.36329627e-01
7.10240602e-01 -1.23196912e+00 -1.35070777e+00 3.46222699e-01
4.05199796e-01 -8.65846634e-01 4.49163318e-02 -1.59647991e-03
-6.56177342e-01 -1.73118103e+00 -7.84794927e-01 -6.92641854e-01
6.43210471e-01 -2.53103733e-01 1.01125979e+00 7.99822867e-01
-1.26501155e+00 6.76995590e-02 -2.30851397e-01 -7.88169444e-01
-3.93631697e-01 -2.01817870e-01 -1.89870596e-01 1.25970459e-02
1.39996815e+00 1.52003959e-01 -8.41600716e-01 -1.88493535e-01
-7.38811195e-01 -1.97486594e-01 6.07667506e-01 8.01129937e-01
4.24560487e-01 -1.61746681e-01 6.00587249e-01 -1.29501641e+00
3.00670862e-01 -1.41069978e-01 -5.11920154e-01 3.43829513e-01
-4.97803152e-01 -2.22290501e-01 7.17744768e-01 -1.26552861e-02
-6.70696735e-01 -8.25791806e-02 -7.53647536e-02 1.56976372e-01
-5.90945363e-01 1.79284401e-02 5.33237755e-01 -2.79211532e-02
5.98121762e-01 2.42950618e-01 8.25362802e-02 -3.81997466e-01
-3.56524467e-01 1.00060141e+00 7.23219335e-01 -5.82089834e-02
-9.85462368e-02 1.00422359e+00 4.72094744e-01 -1.07828426e+00
-7.32735097e-01 -8.43653262e-01 -8.85913253e-01 -1.43251032e-01
9.56007063e-01 -6.32244825e-01 -1.38159156e+00 9.07338917e-01
-1.01453054e+00 -1.30813092e-01 2.44414508e-02 6.88102067e-01
-8.08815062e-02 4.75485861e-01 -1.13940883e+00 -1.21678889e+00
-9.48733032e-01 -7.58486867e-01 9.81650770e-01 7.08127141e-01
-3.28665465e-01 -1.35110140e+00 4.65646833e-02 2.04203501e-01
6.91982388e-01 6.70992017e-01 8.77865493e-01 -7.11743116e-01
5.90974875e-02 -7.81281948e-01 -6.39125168e-01 1.28982320e-01
5.00091255e-01 2.54740298e-01 -1.03958476e+00 -2.60581225e-01
3.28028083e-01 -1.91414490e-01 9.57787395e-01 7.56960988e-01
9.23789799e-01 -1.93062007e-01 -4.56844270e-01 6.25723779e-01
1.70601356e+00 5.23098826e-01 6.59413278e-01 2.62377739e-01
4.44134772e-01 5.85605145e-01 3.38615835e-01 4.69963640e-01
2.71763164e-03 -1.62621975e-01 -1.43549755e-01 -4.23907131e-01
-2.62115359e-01 4.68767077e-01 -5.27531922e-01 5.44661462e-01
-1.87657788e-01 -1.87467441e-01 -9.75256622e-01 7.65568197e-01
-1.04397166e+00 -8.22806656e-01 -2.16341347e-01 1.78790510e+00
7.39605963e-01 -2.55613178e-01 3.33449751e-01 4.56477702e-01
8.78502488e-01 -9.45904732e-01 -4.35515255e-01 -5.30928552e-01
-3.58466879e-02 1.00334716e+00 5.29877543e-01 5.06188333e-01
-1.00120103e+00 5.75076401e-01 6.69043922e+00 6.22377872e-01
-1.52725029e+00 -9.81955826e-02 1.15944207e+00 -1.38500780e-01
7.79462099e-01 -7.04278350e-01 -9.52454329e-01 6.43761456e-01
6.77835941e-01 1.91749185e-01 2.03513764e-02 4.28297013e-01
-1.57018472e-02 -9.74422574e-01 -1.31099904e+00 1.21012282e+00
3.32424819e-01 -1.47233903e+00 4.91848402e-02 -1.29759997e-01
2.15199038e-01 -5.29310048e-01 -3.14993709e-01 2.57113785e-01
3.40114050e-02 -1.45304704e+00 8.41077194e-02 6.00385189e-01
9.91006911e-01 -6.70395672e-01 1.55859077e+00 8.73422697e-02
-9.51265216e-01 9.05999392e-02 -6.96120143e-01 -3.48589331e-01
-2.01033220e-01 6.47806585e-01 -1.45920432e+00 -2.57640481e-01
7.06350386e-01 3.38330418e-01 -9.29763913e-01 1.28092325e+00
4.35827762e-01 3.93061817e-01 -3.95774603e-01 -6.10911787e-01
1.50304111e-02 -1.93559498e-01 -2.92456567e-01 1.74031174e+00
4.13969368e-01 5.95940240e-02 -4.35670435e-01 7.14542389e-01
2.36126363e-01 2.22636119e-01 -2.52557784e-01 1.08439006e-01
4.61433947e-01 1.43857431e+00 -1.56270862e+00 -7.99363375e-01
-1.40728593e-01 3.88300657e-01 1.94666296e-01 -1.27607554e-01
-5.02974749e-01 -5.81567764e-01 -3.37629735e-01 8.39845613e-02
9.99141857e-02 3.39586347e-01 -7.71613657e-01 -6.21594906e-01
-5.61850727e-01 -1.97660521e-01 7.28457451e-01 -4.37029511e-01
-1.22318614e+00 3.04583400e-01 -3.04130793e-01 -8.63610327e-01
-5.41257709e-02 -1.25208867e+00 -5.54306865e-01 7.78134584e-01
-1.35441363e+00 -7.79333353e-01 -6.18464887e-01 1.21411644e-01
2.06926301e-01 -5.39402544e-01 9.24721241e-01 2.69679278e-01
-7.89449513e-01 6.34270489e-01 8.88990834e-02 6.92225039e-01
5.31101048e-01 -1.42110789e+00 -2.15639889e-01 4.16418582e-01
-3.07677001e-01 6.39966965e-01 4.82654601e-01 -4.40525830e-01
-8.75269294e-01 -8.61448765e-01 9.94866908e-01 -3.63263488e-01
-4.33362424e-02 4.65584062e-02 -6.68840408e-01 2.05056310e-01
3.66543978e-01 1.93348862e-02 1.21505940e+00 -5.17971039e-01
-2.30561513e-02 -8.31403583e-02 -1.77445507e+00 1.91131443e-01
-6.76717386e-02 -1.25529915e-01 -2.91420937e-01 4.68196839e-01
-2.47351035e-01 -4.28376019e-01 -7.39166141e-01 5.08838519e-02
8.36294234e-01 -1.38661420e+00 8.24745417e-01 -2.87845612e-01
2.37468064e-01 -2.83439964e-01 2.69519359e-01 -4.86585796e-01
-9.58235487e-02 1.22669777e-02 -3.00319325e-02 9.95681822e-01
-7.74825290e-02 -7.94553041e-01 9.47024822e-01 5.55528760e-01
3.37056309e-01 -7.50728965e-01 -9.54802692e-01 -5.11211097e-01
2.82828540e-01 3.29767644e-01 5.63402057e-01 7.19009399e-01
1.35054216e-01 2.86282092e-01 1.99225657e-02 -3.54199201e-01
6.52909875e-01 -1.33604720e-01 7.86474645e-01 -1.23871958e+00
-2.69037575e-01 -5.52114904e-01 -6.03177488e-01 -1.61535278e-01
-6.52309954e-01 -5.33190012e-01 -5.14166772e-01 -1.45044887e+00
3.79107744e-01 -2.25156218e-01 -1.59546465e-01 3.54512542e-01
-1.39604628e-01 7.00164676e-01 -1.39732748e-01 1.72999278e-01
-2.86147594e-01 -5.88695347e-01 1.00787830e+00 -1.33415565e-01
9.43623390e-03 -3.71928960e-01 -4.24397081e-01 4.31800842e-01
1.06800926e+00 -4.69450772e-01 2.58816868e-01 -3.48209798e-01
9.33195502e-02 -1.56650424e-01 3.05103570e-01 -1.27866900e+00
2.57450610e-01 2.19256148e-01 1.08433259e+00 -8.87349963e-01
1.07279301e-01 -4.52956855e-01 3.98333594e-02 1.28916395e+00
5.58066480e-02 -3.81038159e-01 4.21652704e-01 1.04691245e-01
-1.27983898e-01 -6.12142920e-01 1.26619983e+00 -5.37256181e-01
-1.70322791e-01 -2.00956210e-01 -6.81928039e-01 -6.55482352e-01
1.29694784e+00 -9.27390933e-01 -5.95307589e-01 1.10197715e-01
-5.64460099e-01 1.97106838e-01 3.20049584e-01 -4.22060043e-01
3.55990440e-01 -8.51227164e-01 -8.84225309e-01 2.19863445e-01
-1.19297147e-01 1.67685136e-01 3.07784021e-01 1.03683054e+00
-1.88014281e+00 9.00925457e-01 -4.56560820e-01 -8.55324209e-01
-1.29953277e+00 6.13093555e-01 6.24517977e-01 -4.81930375e-01
-3.54992837e-01 9.26892698e-01 -6.92189410e-02 -4.65353057e-02
4.59471121e-02 -4.51980054e-01 -8.45029593e-01 1.43012106e-01
9.61817980e-01 1.02325320e+00 1.72741756e-01 -3.87840301e-01
-2.91318029e-01 6.49101913e-01 -1.66453794e-01 4.00946885e-01
1.19096243e+00 6.80100471e-02 -3.23866874e-01 4.57640678e-01
1.05441117e+00 -7.10487366e-02 -5.64523101e-01 1.16938218e-01
-1.80012491e-02 -3.58939976e-01 -2.17236131e-01 -7.94213533e-01
-1.03322458e+00 1.24555564e+00 9.81035531e-01 4.01538461e-01
8.96748662e-01 -2.68195093e-01 5.73111176e-01 4.54230368e-01
1.18096538e-01 -1.24976909e+00 -5.61761390e-03 1.98660910e-01
3.38588148e-01 -1.26515508e+00 3.39971215e-01 -4.25043136e-01
-1.72930107e-01 1.61814797e+00 6.15163147e-01 -3.76769722e-01
5.82257032e-01 4.35662627e-01 3.37504894e-01 -4.14510906e-01
-4.08319563e-01 -1.38526514e-01 -1.52567118e-01 8.85417461e-01
9.83997166e-01 6.05131686e-02 -6.08166933e-01 3.96800667e-01
-6.28804117e-02 5.67155004e-01 6.98342681e-01 7.41655946e-01
-5.36660135e-01 -4.75532770e-01 -6.04023159e-01 8.19760025e-01
-1.03977370e+00 1.23166025e-01 -5.76881826e-01 9.56391335e-01
3.04423511e-01 1.21378756e+00 4.93514955e-01 2.05107644e-01
-1.08276330e-01 -4.98623215e-02 6.45591617e-01 -6.73628986e-01
-7.97922909e-01 1.21808276e-02 -2.56100625e-01 1.06241502e-01
-4.87621903e-01 -3.57688367e-01 -1.24440789e+00 -4.07766879e-01
-3.55026454e-01 5.39908931e-03 3.42136592e-01 9.71864581e-01
6.95580617e-03 3.52929980e-01 5.90759404e-02 -5.12918711e-01
-5.68965776e-03 -1.45806539e+00 -8.04430187e-01 3.83086145e-01
4.60309595e-01 -4.74933952e-01 -3.31664175e-01 3.47312003e-01] | [14.858492851257324, -3.1367743015289307] |
7fe292be-509a-4c4e-8c04-855827d2a8da | template-free-articulated-neural-point-clouds | 2305.19065 | null | https://arxiv.org/abs/2305.19065v1 | https://arxiv.org/pdf/2305.19065v1.pdf | Template-free Articulated Neural Point Clouds for Reposable View Synthesis | Dynamic Neural Radiance Fields (NeRFs) achieve remarkable visual quality when synthesizing novel views of time-evolving 3D scenes. However, the common reliance on backward deformation fields makes reanimation of the captured object poses challenging. Moreover, the state of the art dynamic models are often limited by low visual fidelity, long reconstruction time or specificity to narrow application domains. In this paper, we present a novel method utilizing a point-based representation and Linear Blend Skinning (LBS) to jointly learn a Dynamic NeRF and an associated skeletal model from even sparse multi-view video. Our forward-warping approach achieves state-of-the-art visual fidelity when synthesizing novel views and poses while significantly reducing the necessary learning time when compared to existing work. We demonstrate the versatility of our representation on a variety of articulated objects from common datasets and obtain reposable 3D reconstructions without the need of object-specific skeletal templates. Code will be made available at https://github.com/lukasuz/Articulated-Point-NeRF. | ['Petr Kellnhofer', 'Elmar Eisemann', 'Lukas Uzolas'] | 2023-05-30 | null | null | null | null | ['specificity'] | ['natural-language-processing'] | [ 2.38598168e-01 -2.01795414e-01 7.91849494e-02 -1.18795894e-01
-8.74446809e-01 -7.99642086e-01 6.01040244e-01 -6.82001770e-01
2.08002254e-01 4.19806600e-01 9.66393948e-02 5.68358563e-02
-2.45621398e-01 -5.39473236e-01 -9.96949077e-01 -7.33340621e-01
2.69292325e-01 3.54050994e-01 1.88237712e-01 -9.28254873e-02
-2.20507346e-02 9.89213228e-01 -1.74304461e+00 7.70044774e-02
5.22727370e-01 8.42968345e-01 4.02190030e-01 7.40163922e-01
2.10402340e-01 5.80977142e-01 -1.60498530e-01 -2.27928445e-01
6.73787296e-01 -3.06446612e-01 -4.65671510e-01 4.05389875e-01
9.46804464e-01 -6.09196961e-01 -7.98278928e-01 5.59391141e-01
5.92339635e-01 3.23772669e-01 3.98854762e-01 -1.08227289e+00
-6.91391230e-01 -4.50941771e-02 -5.33981204e-01 -2.27410384e-02
5.54317474e-01 3.01438361e-01 7.00322866e-01 -1.18244994e+00
1.19096494e+00 1.06214309e+00 7.47899354e-01 7.56735623e-01
-1.37481320e+00 -5.33455610e-01 1.46918952e-01 1.75366640e-01
-1.30058861e+00 -6.61473572e-01 1.14792836e+00 -4.70328063e-01
9.45344865e-01 3.02163869e-01 8.86712551e-01 1.34826815e+00
3.06552440e-01 5.25037050e-01 1.01207042e+00 -1.07136197e-01
-5.59547218e-03 -5.95900595e-01 -5.48842072e-01 9.20156300e-01
-6.91500902e-02 3.60714614e-01 -7.40822077e-01 -7.77464956e-02
1.27665913e+00 3.52585256e-01 -4.61398810e-01 -9.35522258e-01
-1.40006447e+00 3.18783253e-01 4.15054142e-01 6.44245446e-02
-5.19353926e-01 6.26780272e-01 -6.47859797e-02 1.03182487e-01
4.27740782e-01 9.51739028e-02 -3.10903370e-01 -1.40715733e-01
-8.44239712e-01 4.55808610e-01 3.98187518e-01 8.67291987e-01
6.72182441e-01 5.69544733e-01 3.16168040e-01 6.64796472e-01
2.52677381e-01 7.62500942e-01 2.03329369e-01 -1.49314666e+00
1.96273804e-01 3.01977456e-01 6.93536252e-02 -8.34889770e-01
-2.37955227e-01 -4.51273680e-01 -6.95321918e-01 4.98864740e-01
1.63464442e-01 1.36471018e-01 -1.21788001e+00 1.63243937e+00
6.71705127e-01 3.36593002e-01 -1.22840039e-01 9.72322941e-01
7.90112555e-01 7.26900876e-01 -3.46303135e-01 -4.27742936e-02
8.79655242e-01 -8.24122071e-01 -5.07064104e-01 -2.01484129e-01
-6.43892353e-03 -8.01581025e-01 9.61211860e-01 4.12570715e-01
-1.47624493e+00 -3.93793464e-01 -9.08934653e-01 -3.28205645e-01
1.49838567e-01 -2.50843614e-01 5.37214518e-01 2.29126722e-01
-9.48436916e-01 6.52916133e-01 -1.24522543e+00 -2.19402716e-01
3.34367663e-01 2.85562038e-01 -6.93848908e-01 -3.21337581e-01
-5.78947127e-01 8.18321943e-01 -1.64943188e-01 1.26917481e-01
-1.37400806e+00 -1.11887360e+00 -8.60825598e-01 -2.68622637e-01
3.75434011e-01 -1.22716367e+00 1.25885212e+00 -7.78179705e-01
-1.71835172e+00 7.95790553e-01 4.02553827e-02 2.78644487e-02
7.56273746e-01 -1.86190963e-01 -1.45979553e-01 3.05403143e-01
-1.88528344e-01 5.80265820e-01 9.34335053e-01 -1.67541003e+00
-1.32096201e-01 -3.62083793e-01 1.22034237e-01 3.63874406e-01
9.48390886e-02 -3.06949884e-01 -5.18229008e-01 -9.15133715e-01
4.57167804e-01 -1.09733665e+00 -1.66026667e-01 7.13728130e-01
-5.98607659e-02 3.43155056e-01 1.09732628e+00 -8.63704860e-01
6.99406087e-01 -2.00326467e+00 7.94550717e-01 -1.15043603e-01
9.19762924e-02 -6.08023778e-02 -2.55083203e-01 4.45394218e-01
-9.65091810e-02 -2.59879589e-01 -3.73539090e-01 -2.87602007e-01
-1.97229728e-01 3.53212476e-01 -1.93227097e-01 6.85899079e-01
1.98648814e-02 9.14859176e-01 -8.48994076e-01 -2.06248343e-01
4.88115489e-01 1.01665938e+00 -6.10608935e-01 2.55485982e-01
-3.12744409e-01 8.66269886e-01 -3.39545012e-01 9.65983450e-01
5.55088758e-01 -2.88484961e-01 6.31073639e-02 -5.83897471e-01
-1.27042696e-01 -1.15253609e-02 -1.18486834e+00 2.43575788e+00
-6.53789580e-01 3.75047892e-01 1.99241847e-01 -6.50913954e-01
7.75101900e-01 3.25809568e-01 8.83689046e-01 -4.56092268e-01
1.01318583e-02 2.23331481e-01 -3.53781939e-01 -4.56107587e-01
3.25814664e-01 -4.46858495e-01 1.20623350e-01 3.52182209e-01
1.90970957e-01 -4.90397573e-01 -2.23370120e-01 -1.20417036e-01
1.02378786e+00 9.77172077e-01 1.71120256e-01 1.11004703e-01
2.80521214e-01 -1.64252833e-01 5.08532643e-01 1.71835288e-01
2.68395215e-01 1.11745775e+00 -1.97642103e-01 -6.10698938e-01
-1.33172154e+00 -1.60877478e+00 -8.67443532e-02 5.73488951e-01
1.34766832e-01 -2.44642332e-01 -2.74390668e-01 -2.61073411e-01
1.17486380e-01 4.91485447e-01 -4.71456021e-01 -4.41276617e-02
-8.38390350e-01 -1.79567665e-01 2.54623979e-01 4.88391936e-01
3.01019579e-01 -7.00092852e-01 -9.31569815e-01 1.48231030e-01
-2.85742819e-01 -1.13747573e+00 -3.62578839e-01 -1.24654524e-01
-1.20928144e+00 -9.45958555e-01 -9.92473125e-01 -5.61313272e-01
6.42676175e-01 4.55955893e-01 9.50904369e-01 7.94746075e-03
-4.16774690e-01 8.98992836e-01 5.30267972e-03 9.29150507e-02
-3.58447224e-01 -2.94444472e-01 1.00131616e-01 -1.48990929e-01
-5.03808737e-01 -1.03317153e+00 -7.74417222e-01 4.14368451e-01
-1.03517771e+00 4.17535633e-01 2.72223204e-01 8.26475918e-01
8.90328705e-01 -3.65879536e-01 7.95580670e-02 -4.25764292e-01
-1.79073766e-01 -1.52057767e-01 -5.44687867e-01 2.21751943e-01
-2.66352504e-01 -1.15021011e-02 4.11065519e-01 -5.95643520e-01
-1.27040625e+00 3.50371510e-01 -9.75644365e-02 -1.08201325e+00
1.10473163e-01 2.03348294e-01 1.21657411e-02 -3.09107780e-01
6.25368476e-01 2.58756250e-01 1.08894989e-01 -6.35704935e-01
5.04621089e-01 9.76656377e-02 7.49958754e-01 -6.56086862e-01
1.11940229e+00 8.11645150e-01 2.08540320e-01 -6.94701135e-01
-6.56956971e-01 -3.52480635e-02 -7.22004771e-01 -6.54781520e-01
7.00839996e-01 -1.05516422e+00 -4.65442330e-01 5.86020529e-01
-9.21874166e-01 -5.05088091e-01 -4.82323229e-01 4.47953880e-01
-1.04827249e+00 4.41920549e-01 -5.07250607e-01 -4.44782019e-01
-2.80224085e-01 -1.03133285e+00 1.23651445e+00 5.06736711e-03
1.38231814e-02 -7.82768428e-01 1.60465270e-01 7.55738199e-01
2.58944452e-01 8.54008973e-01 5.69849074e-01 5.44397950e-01
-1.08562756e+00 1.48399370e-02 3.32702756e-01 1.17414713e-01
2.17254698e-01 2.05834776e-01 -8.64288211e-01 -5.01484990e-01
-5.96268922e-02 -2.89712727e-01 5.58183193e-01 3.85238379e-01
1.03630328e+00 -1.72349066e-01 -7.38967955e-02 1.07723987e+00
1.63355339e+00 1.30258635e-01 4.84768748e-01 3.08263659e-01
9.63449240e-01 4.50494379e-01 3.02185327e-01 6.46772027e-01
2.31087148e-01 1.00673091e+00 6.73821449e-01 2.65353411e-01
-5.66533864e-01 -3.56359035e-01 4.89653111e-01 8.89183342e-01
-6.19780660e-01 -1.78772882e-01 -9.86186206e-01 5.21458566e-01
-1.71727979e+00 -1.08463037e+00 2.74026483e-01 1.93408608e+00
6.30659163e-01 -3.12175393e-01 -1.02473721e-01 8.99947700e-05
4.30052787e-01 4.04397279e-01 -8.05218756e-01 3.90233472e-02
-1.10745735e-01 1.05543375e-01 2.40136459e-01 5.27383924e-01
-5.91366172e-01 6.84472442e-01 5.39632893e+00 4.09850061e-01
-1.24308503e+00 2.31661007e-01 1.23028010e-01 -6.15876436e-01
-6.76519990e-01 -2.17393842e-02 -3.85299921e-01 -1.90188140e-02
7.00716376e-01 -2.44142979e-01 6.24381483e-01 6.47471130e-01
7.06418827e-02 1.79583505e-01 -9.21842217e-01 1.19076252e+00
2.85107315e-01 -1.58922696e+00 1.89269707e-01 2.77775135e-02
9.40619588e-01 8.63500535e-02 1.39736012e-01 -3.27872485e-01
1.84344977e-01 -7.39508450e-01 1.16982114e+00 9.96168137e-01
1.08100605e+00 -6.32457137e-01 -2.35755965e-02 1.05296850e-01
-1.13580215e+00 -4.33519706e-02 -1.60826653e-01 1.87031686e-01
5.96579492e-01 2.76055783e-01 -2.80153871e-01 7.93274760e-01
8.75873804e-01 1.02075100e+00 -3.61758381e-01 7.56621778e-01
-5.82497604e-02 1.71423391e-01 -4.25720483e-01 6.14023030e-01
2.27655191e-02 -1.76756218e-01 8.43655705e-01 6.16331339e-01
7.00759768e-01 2.70375699e-01 -7.93462470e-02 7.66825855e-01
-5.60070015e-02 -2.41262659e-01 -7.21410036e-01 7.50816092e-02
1.54543474e-01 1.19377267e+00 -5.52711248e-01 -2.28527766e-02
-4.10978258e-01 1.03979504e+00 1.50404811e-01 3.68988365e-01
-8.85894835e-01 3.04502428e-01 6.67401731e-01 4.23326015e-01
5.01341701e-01 -5.67935884e-01 -9.05641913e-02 -1.40338421e+00
2.49497414e-01 -7.77515769e-01 7.62779713e-02 -1.31336236e+00
-1.03729498e+00 4.60711181e-01 2.15836972e-01 -1.49563205e+00
-3.48698169e-01 -3.85903001e-01 -8.48445296e-02 4.62071508e-01
-1.21445847e+00 -1.46641040e+00 -6.71655536e-01 7.99953759e-01
7.79719710e-01 2.50774380e-02 6.75374508e-01 2.91165322e-01
-1.99999556e-01 1.59680098e-01 1.30551666e-01 -3.84301037e-01
5.57382047e-01 -8.48464906e-01 4.06066567e-01 9.53961074e-01
2.13714570e-01 2.90386438e-01 7.07609117e-01 -4.89769995e-01
-2.00835538e+00 -9.16314304e-01 2.16879740e-01 -5.93015730e-01
3.90208572e-01 -3.17029685e-01 -8.89048755e-01 8.12398970e-01
2.61079296e-02 3.35211724e-01 3.84669542e-01 -4.33625966e-01
-4.37373370e-01 -2.13815823e-01 -1.18645740e+00 7.05673695e-01
1.53396702e+00 -5.93178630e-01 -4.79621559e-01 1.75971687e-01
5.13736546e-01 -9.32858467e-01 -1.18848515e+00 4.97781634e-01
1.00747347e+00 -9.85024035e-01 1.39536214e+00 -3.10052693e-01
6.26057684e-01 -6.13064826e-01 -3.80473346e-01 -1.14709175e+00
-4.05535579e-01 -7.19093561e-01 -4.72499162e-01 9.02156115e-01
-1.08860768e-01 -3.19565654e-01 7.48269856e-01 5.86879253e-01
-3.80251288e-01 -7.84698963e-01 -1.07193470e+00 -8.51144135e-01
-1.59787297e-01 -2.98998713e-01 4.15948182e-01 9.19954240e-01
-6.32453322e-01 -4.82642688e-02 -5.96179724e-01 2.37666503e-01
8.24699581e-01 3.94718528e-01 8.80964696e-01 -9.08738315e-01
-4.53598291e-01 -6.93094917e-03 -5.31494081e-01 -8.99185061e-01
1.06327161e-01 -9.53121603e-01 -6.62119165e-02 -1.56942606e+00
-7.28274286e-02 -2.54895866e-01 2.11904701e-02 5.29785454e-01
3.30369592e-01 5.65596998e-01 4.25276011e-01 3.62025589e-01
-6.91549927e-02 8.80613148e-01 1.63268292e+00 -1.04672350e-02
-9.60303247e-02 -2.78785467e-01 -2.36096993e-01 6.50865853e-01
6.74454927e-01 -4.76744115e-01 -5.26819050e-01 -8.55378330e-01
2.75028676e-01 4.10320014e-01 7.34899402e-01 -1.03012764e+00
1.42402142e-01 -3.58932137e-01 5.32325804e-01 -6.20804310e-01
8.73843789e-01 -1.03029132e+00 1.29094660e+00 5.04472375e-01
4.49007004e-02 4.25082237e-01 2.45750219e-01 7.44952559e-01
1.54604614e-01 6.69375882e-02 8.70631933e-01 -3.20965916e-01
-8.43388438e-01 6.37070775e-01 3.64852361e-02 -1.77458838e-01
9.86476302e-01 -4.55034047e-01 -1.70911744e-01 -2.75306493e-01
-7.63389826e-01 -2.39325494e-01 1.10320985e+00 5.79766631e-01
9.77534533e-01 -1.43699872e+00 -6.01285636e-01 3.43496501e-01
1.24839045e-01 3.08406413e-01 7.11425126e-01 6.08710885e-01
-9.66917217e-01 -6.16173111e-02 -6.01026535e-01 -8.40393543e-01
-1.31089842e+00 4.84021127e-01 5.69475949e-01 1.39851436e-01
-1.28208673e+00 6.10970914e-01 2.41131335e-01 -4.63659585e-01
-1.35700434e-01 -1.20133631e-01 4.97688174e-01 -2.60680139e-01
1.13078885e-01 5.27772009e-01 -2.72941147e-03 -8.41124356e-01
-3.34246248e-01 1.15998733e+00 2.91397363e-01 -2.67994732e-01
1.78383768e+00 -1.12699308e-01 2.27906853e-01 3.99361402e-01
1.08018005e+00 -6.50577247e-02 -2.02440357e+00 -2.61945128e-01
-6.23795211e-01 -8.32715631e-01 1.48489773e-01 -6.50338709e-01
-1.44136274e+00 6.49364531e-01 5.92734277e-01 -5.70264637e-01
1.18279767e+00 8.80662277e-02 8.76183271e-01 3.26452434e-01
6.49414599e-01 -6.89305604e-01 2.39932746e-01 1.72764301e-01
1.38977528e+00 -7.91208863e-01 4.02416110e-01 -3.42553139e-01
-4.31364000e-01 1.22511983e+00 4.46715981e-01 -2.62231559e-01
6.03559792e-01 3.37015182e-01 4.27720249e-02 -3.17400932e-01
-7.84814000e-01 2.18784809e-01 2.75203556e-01 7.34159708e-01
9.27999094e-02 -3.22963148e-01 1.30042538e-01 -1.75098851e-01
-2.26509094e-01 -2.25213449e-02 6.15544617e-01 1.16501188e+00
3.26977409e-02 -9.92339313e-01 -2.67708659e-01 2.99737733e-02
-3.45249146e-01 2.68053234e-01 -4.90064695e-02 7.12164521e-01
-1.26638055e-01 3.85527521e-01 -2.70171799e-02 -2.58921444e-01
5.31786501e-01 -1.47975003e-02 1.22705221e+00 -4.36902940e-01
-4.02960569e-01 1.66540220e-01 -3.04007065e-02 -9.10319388e-01
-9.74622190e-01 -9.02928770e-01 -1.04837418e+00 -3.97002131e-01
-5.02511710e-02 -5.51579952e-01 7.06512868e-01 4.80856150e-01
5.32286525e-01 3.58382463e-01 5.24506450e-01 -1.36704481e+00
-3.99239212e-01 -3.74625295e-01 -3.89252871e-01 5.04322171e-01
4.77555245e-01 -9.54352498e-01 -2.88998663e-01 4.91112292e-01] | [8.894805908203125, -2.819120168685913] |
8fb27434-a903-456d-8110-b3b1d442a644 | research-on-dynamic-target-detection-and | 1912.01992 | null | https://arxiv.org/abs/1912.01992v1 | https://arxiv.org/pdf/1912.01992v1.pdf | Research on dynamic target detection and tracking system of hexapod robot | Dynamic target detection and target tracking are hot issues in the field of image. In order to explore its application value in the field of mobile robot, a dynamic target detection and tracking system is designed based on hexapod robot. Firstly, the dynamic target detection method is introduced with region merging and adaptive external point filtering based on motion compensation method. This method achieves the accurate compensation of the moving background through symmetric matching and adaptive external point filtering, and achieves complete detection of non-rigid objects by region merging. Secondly, the application of target tracking algorithm based on KCF in hexapod robot platform is studied, and the Angle tracking of moving target is realized by adaptive adjustment of tracking speed. The last, the architecture of robot monitoring system is designed, which consists of operator, processor, hexapod robot and vision sensor, and the moving object detection and tracking algorithm proposed in this paper is applied to the system. The experimental results show that the improved algorithm can effectively detect and track the moving target when applied to the system of the mobile hexapod robot. | ['Dexin Wang'] | 2019-12-04 | null | null | null | null | ['moving-object-detection'] | ['computer-vision'] | [-2.56055653e-01 -3.44280362e-01 -1.77050438e-02 2.42387906e-01
2.62253672e-01 -3.27315927e-01 1.73615739e-01 -4.15039569e-01
-6.55255973e-01 1.46433199e-02 -4.61300582e-01 -2.43256483e-02
-6.06554188e-02 -6.83892608e-01 1.87109003e-03 -1.09064770e+00
2.48949468e-01 4.94102091e-01 1.06664097e+00 -1.35662109e-01
3.97706449e-01 9.28169012e-01 -1.47844243e+00 -3.96132737e-01
4.78467584e-01 9.45287585e-01 1.01275587e+00 5.70148289e-01
2.03044057e-01 3.18063080e-01 -4.38010216e-01 4.47016805e-01
5.43936372e-01 2.74997987e-02 -8.43606964e-02 4.09101218e-01
-1.05844222e-01 -2.99782395e-01 -5.18247128e-01 1.38698649e+00
7.97195017e-01 1.61671519e-01 7.74184406e-01 -1.43678796e+00
-3.98345232e-01 2.92583436e-01 -6.97740674e-01 4.80098516e-01
-2.76286602e-01 2.47987136e-01 -5.41928783e-02 -1.14252126e+00
5.59827149e-01 1.48062658e+00 7.48011649e-01 3.70161057e-01
-7.23034859e-01 -5.08609653e-01 -2.07891032e-01 7.75764704e-01
-1.59202206e+00 -4.30691950e-02 3.95699799e-01 -6.32293582e-01
5.61930478e-01 -6.78241551e-02 6.54275298e-01 8.48049298e-02
8.30069542e-01 4.74610120e-01 5.92893064e-01 -4.93166476e-01
-1.17353804e-01 9.39099491e-02 5.07420301e-01 5.29679835e-01
7.25837827e-01 4.32352483e-01 3.51938188e-01 3.15879673e-01
7.61530638e-01 4.41465825e-01 -6.07633650e-01 -7.85198748e-01
-1.48487484e+00 6.71082377e-01 4.24080282e-01 2.80826151e-01
-3.74575973e-01 -1.85501620e-01 3.50593507e-01 1.32233322e-01
-2.80472577e-01 -1.32458270e-01 -2.99489975e-01 1.35622486e-01
-1.83586121e-01 1.29342976e-03 6.36726499e-01 1.32355571e+00
2.59163111e-01 2.59648502e-01 -1.06989391e-01 5.68480968e-01
7.22761333e-01 1.46604753e+00 7.06205249e-01 -9.44501698e-01
6.21254966e-02 6.56156719e-01 5.90258896e-01 -1.38362074e+00
-8.99275780e-01 -2.02856492e-03 -5.25436819e-01 1.05879724e+00
2.26684853e-01 -1.96819574e-01 -1.01288986e+00 1.01142883e+00
9.92076755e-01 -2.96903461e-01 3.06375325e-01 1.20203674e+00
7.83120573e-01 1.08246517e+00 -3.05862308e-01 -6.95984006e-01
1.50997102e+00 -1.01168180e+00 -1.04287493e+00 -5.88822961e-02
2.52325475e-01 -1.04044366e+00 4.93610293e-01 3.20807695e-01
-5.66453815e-01 -8.74856174e-01 -9.80652511e-01 2.17588544e-01
-3.24480623e-01 4.63751554e-01 -4.25537042e-02 3.26367587e-01
-4.80551243e-01 -1.34593800e-01 -8.55472207e-01 -6.89509153e-01
-3.60001892e-01 5.17373502e-01 -1.11839511e-01 -2.07219005e-01
-7.28136241e-01 1.52913606e+00 8.63758564e-01 2.97762543e-01
-7.03882992e-01 -2.75825471e-01 -6.30583704e-01 -2.27942273e-01
3.02210987e-01 -4.72752094e-01 1.22642171e+00 -6.08190000e-01
-1.51745880e+00 6.47652090e-01 1.43919781e-01 -9.41031873e-02
4.37884271e-01 -1.15480796e-01 -6.81032658e-01 1.85148895e-01
2.58207023e-01 3.00937146e-01 1.03279972e+00 -8.93436074e-01
-1.40719736e+00 -7.14735985e-01 -7.37404823e-01 7.26077914e-01
6.98678270e-02 1.69081017e-01 -3.17279488e-01 6.53652996e-02
4.05681878e-01 -9.29406226e-01 -1.84546903e-01 1.30105361e-01
7.40157142e-02 -1.43566370e-01 2.10908699e+00 -4.17711854e-01
8.38714719e-01 -2.44390917e+00 2.80421264e-02 1.26045078e-01
-1.61660075e-01 3.46969604e-01 1.56912968e-01 9.72574353e-02
3.80662262e-01 -1.08617818e+00 2.77017295e-01 3.91948849e-01
-2.55512148e-01 -8.02389532e-03 -1.34051189e-01 7.90195823e-01
-5.13509750e-01 4.33582366e-01 -5.42370677e-01 -6.78998470e-01
4.73742783e-01 3.68878603e-01 2.93153703e-01 -2.65482720e-02
8.70411098e-02 1.02346554e-01 -4.36842144e-01 8.33439052e-01
1.21404397e+00 2.80380249e-01 -2.51610994e-01 -3.47499698e-01
-9.13541615e-01 -6.53915882e-01 -1.68577945e+00 6.07496619e-01
9.80032831e-02 4.99512762e-01 5.70353210e-01 -5.66066265e-01
1.35109949e+00 1.12240296e-02 4.93881583e-01 -5.95732152e-01
5.76078951e-01 4.52639759e-01 1.37362465e-01 -8.10431421e-01
6.41086042e-01 9.95820984e-02 2.65661925e-01 -1.11747026e-01
-6.02188945e-01 -1.75866470e-01 -9.73702520e-02 -3.02933812e-01
8.83906960e-01 -2.92056240e-02 4.15785789e-01 -4.15382892e-01
8.39156151e-01 1.14565969e+00 7.73824036e-01 4.45014715e-01
-3.14443827e-01 8.38883594e-02 -5.96363783e-01 -4.32485819e-01
-1.06177628e+00 -9.42048311e-01 -2.70131618e-01 5.68566740e-01
1.32676435e+00 6.36007369e-01 -3.99330467e-01 -1.40433442e-02
3.60805631e-01 1.43034652e-01 -1.80786163e-01 -3.23653847e-01
-7.62522161e-01 -7.16182292e-01 -1.15182385e-01 2.16719314e-01
9.63126898e-01 -1.05534971e+00 -1.22207952e+00 5.29842317e-01
6.95222318e-02 -8.24737966e-01 -5.64162314e-01 -1.78862214e-01
-7.63943672e-01 -1.26214659e+00 -7.11270332e-01 -1.46004057e+00
5.99067748e-01 1.33534753e+00 1.58114940e-01 -4.13323268e-02
-4.23225164e-01 7.36674786e-01 -3.31227899e-01 -7.17174411e-01
-2.77277172e-01 -5.37135005e-01 4.59421515e-01 -8.25430080e-02
6.44962370e-01 2.77509630e-01 -5.46262503e-01 1.03514230e+00
-3.45451891e-01 -3.02541018e-01 7.28327334e-01 7.93122828e-01
5.52859247e-01 5.39645791e-01 -7.83629250e-04 2.25807875e-02
4.42386083e-02 -1.61793418e-02 -1.48402488e+00 -6.52030781e-02
-7.15494573e-01 -8.69903445e-01 2.99589008e-01 -8.75612080e-01
-1.07587647e+00 3.17950189e-01 5.91778457e-01 -6.85104609e-01
1.22187197e-01 1.36534557e-01 -5.09939551e-01 -4.30117279e-01
6.05102956e-01 3.71028632e-01 8.42300773e-01 -2.67588317e-01
1.60567343e-01 1.06245995e+00 9.31583464e-01 3.50778967e-01
8.62400711e-01 6.86346829e-01 4.03562456e-01 -9.46946144e-01
1.45034000e-01 -8.63546193e-01 -5.95157981e-01 -5.29699862e-01
1.06234908e+00 -1.11330175e+00 -9.33059752e-01 1.00840807e+00
-1.36307752e+00 -2.94824839e-01 -8.25473592e-02 1.15993631e+00
-5.50517917e-01 4.96648073e-01 -5.51922202e-01 -6.61858380e-01
-4.51448262e-01 -9.82168615e-01 7.08374619e-01 6.17746711e-01
6.24654412e-01 -6.70585454e-01 4.08212142e-03 -2.39821300e-01
3.24999720e-01 -5.40297292e-02 1.84765026e-01 -1.65712729e-01
-8.57958138e-01 -3.95188808e-01 -1.98917776e-01 -5.35017438e-02
1.71429098e-01 1.75221100e-01 -2.28356138e-01 -5.65245152e-01
5.26809990e-01 5.98052859e-01 4.26057398e-01 8.86695921e-01
3.77262086e-02 -1.26051024e-01 -1.02897274e+00 4.58092749e-01
1.66063857e+00 1.03991199e+00 7.24751413e-01 1.03204370e+00
5.53888798e-01 5.50106406e-01 1.76620543e+00 1.70806170e-01
9.05002505e-02 9.30858076e-01 6.28620684e-01 8.87151212e-02
1.00582261e-02 3.70627493e-01 7.20963001e-01 4.28518027e-01
1.39340773e-01 -2.14220062e-02 -6.82578444e-01 3.84418458e-01
-2.03418303e+00 -1.27101719e+00 -1.09657431e+00 1.85257375e+00
-1.48282025e-03 -1.21665344e-01 2.65884101e-01 1.07655097e-02
1.57489455e+00 -5.50272644e-01 -5.07165790e-01 1.41899645e-01
1.07827790e-01 -9.10341740e-01 8.93655181e-01 7.17021406e-01
-1.12734246e+00 7.09615350e-01 5.61917830e+00 7.11932421e-01
-1.50098634e+00 1.54924229e-01 -4.68637735e-01 5.33098936e-01
7.11774707e-01 -2.24835426e-01 -1.58542800e+00 7.41340756e-01
1.00882493e-01 -3.33291471e-01 2.69314218e-02 1.31282663e+00
4.18763608e-01 -1.53216049e-01 -3.34204435e-01 1.21320474e+00
1.37940705e-01 -6.77555263e-01 -2.29563147e-01 4.16512564e-02
2.88216978e-01 -2.08339785e-04 -2.05401052e-02 1.67513534e-03
2.53818426e-02 -1.52475700e-01 8.84576917e-01 4.38160062e-01
2.74442405e-01 -4.48282778e-01 8.25252712e-01 6.33314252e-01
-1.36862111e+00 -7.27536976e-01 -1.07235980e+00 -8.46238509e-02
3.94007504e-01 4.27156538e-01 -1.21821070e+00 2.91220069e-01
8.07330072e-01 5.77913821e-01 -1.90136939e-01 1.59265840e+00
1.83262244e-01 -2.11951345e-01 -6.00421429e-01 -3.97766680e-01
-6.11964874e-02 -5.49241364e-01 1.14559519e+00 1.04991591e+00
8.17511022e-01 4.28035527e-01 3.29431087e-01 4.17140573e-01
9.71073329e-01 1.15401171e-01 -7.36491740e-01 5.80656528e-01
7.19063759e-01 1.55414271e+00 -6.98920727e-01 -3.95834029e-01
-3.15015197e-01 4.75060105e-01 -5.00240445e-01 3.71493697e-02
-9.64771986e-01 -9.89879906e-01 8.75404850e-02 1.97170421e-01
7.34757245e-01 -5.07894337e-01 -2.08947137e-01 -7.10959375e-01
-2.64529675e-01 -3.22661847e-01 5.50220966e-01 -1.10913301e+00
-6.86851203e-01 1.72920376e-01 -5.00470996e-02 -1.80827487e+00
2.29012087e-01 -1.21644318e+00 -9.29400802e-01 6.62958145e-01
-9.03046250e-01 -1.15435207e+00 -7.88846731e-01 7.78108239e-01
5.48276365e-01 -5.01431882e-01 1.43556505e-01 2.14412764e-01
-4.27828431e-01 -9.89843234e-02 4.82371867e-01 -2.01794326e-01
5.89565396e-01 -6.46180511e-01 -4.76922989e-01 1.24141121e+00
-7.82057762e-01 2.06294313e-01 9.18018460e-01 -9.44416106e-01
-1.61013019e+00 -1.06366980e+00 3.51707160e-01 -3.10000628e-01
4.61319268e-01 -2.23942921e-02 -6.43092811e-01 7.17870891e-01
-5.28588481e-02 -1.39315829e-01 -2.80007899e-01 -8.80934358e-01
4.41665351e-01 -3.81156027e-01 -1.43484116e+00 5.15681326e-01
6.46379709e-01 6.01352394e-01 -8.51165891e-01 3.28659564e-01
1.04839933e+00 -7.50609994e-01 -6.86257660e-01 7.24792063e-01
4.02351558e-01 -2.99606115e-01 9.40495610e-01 8.79948363e-02
-6.40188456e-01 -1.43295527e+00 -7.24588707e-02 -9.15969551e-01
-9.16155696e-01 -3.43132854e-01 1.50788978e-01 1.10956442e+00
-3.42954062e-02 -8.11631024e-01 6.17132425e-01 1.42476141e-01
-6.88658237e-01 -6.25264421e-02 -1.04933405e+00 -1.08612263e+00
-5.81952572e-01 5.33452690e-01 4.08141166e-02 6.64621115e-01
8.36751685e-02 4.46097255e-01 -1.77968219e-01 8.69626999e-01
7.77246773e-01 1.43400803e-01 1.14133906e+00 -1.22533584e+00
1.01342522e-01 -7.78379962e-02 -8.24744999e-01 -1.06642199e+00
-4.33383346e-01 -4.46010858e-01 4.22307551e-01 -1.59671664e+00
3.09249479e-02 -9.43238363e-02 3.95335048e-01 -4.59733866e-02
8.30300301e-02 1.00353464e-01 1.82889774e-01 6.93278491e-01
-4.64937449e-01 3.02248091e-01 1.23131049e+00 -3.79941165e-01
-3.39189887e-01 3.33390206e-01 8.24444368e-02 5.49897611e-01
5.93700588e-01 -3.51478457e-01 -8.23454186e-02 -2.41872013e-01
-6.43588483e-01 5.76027576e-03 4.33308065e-01 -1.14177251e+00
8.03355992e-01 -2.16433212e-01 5.31179488e-01 -1.10063839e+00
3.55083346e-01 -1.55490685e+00 5.02662182e-01 1.55800045e+00
8.10729384e-01 3.32009733e-01 1.33301452e-01 5.64016044e-01
-1.25270143e-01 -4.20612603e-01 1.43181968e+00 4.50811908e-03
-1.44319093e+00 2.15959802e-01 -7.33935297e-01 -5.94387531e-01
1.81061947e+00 -5.48190176e-01 -7.99543619e-01 1.45803438e-02
-7.58268237e-01 5.16026676e-01 5.74118674e-01 2.12458819e-01
8.32061172e-01 -1.47456026e+00 -4.03793693e-01 8.22573379e-02
-4.02804315e-02 8.38154275e-03 1.98142841e-01 1.21305883e+00
-1.09553993e+00 2.63600767e-01 -3.86083335e-01 -9.99879718e-01
-1.71332777e+00 1.19691384e+00 6.37444735e-01 3.65177721e-01
-9.11143124e-01 2.03354806e-01 2.40971353e-02 -2.70415843e-01
2.25695789e-01 -5.07965535e-02 -4.87889022e-01 -2.35151723e-01
7.31162190e-01 8.29251528e-01 -4.10775691e-01 -8.02883506e-01
-4.62623060e-01 1.26844478e+00 9.39694420e-02 -6.35778680e-02
8.33442092e-01 -6.23698771e-01 -2.52828360e-01 1.53121144e-01
5.53983331e-01 -9.35229957e-02 -1.16770840e+00 -6.82784244e-03
-1.69180781e-01 -6.01952434e-01 -3.30439955e-01 -5.46519160e-01
-6.41221583e-01 4.87149954e-01 1.24067223e+00 1.82709351e-01
9.23218906e-01 -3.07878315e-01 4.66424406e-01 4.59327996e-01
7.15777755e-01 -1.20294023e+00 -3.30744386e-02 8.89551222e-01
6.31244183e-01 -1.03701854e+00 -1.11123519e-02 -7.54345179e-01
-5.00234246e-01 1.24780536e+00 1.24675333e+00 -3.32564950e-01
5.39883733e-01 4.53098327e-01 4.84306008e-01 -5.82419187e-02
-2.58936405e-01 -1.41702771e-01 -2.60243803e-01 1.10240352e+00
-7.88714647e-01 -2.06783935e-01 -6.18500948e-01 1.38763428e-01
2.93881655e-01 -8.29520375e-02 7.81464636e-01 1.01837814e+00
-1.64686286e+00 -4.22757655e-01 -1.34195614e+00 -7.47040585e-02
-1.40808865e-01 5.33176780e-01 2.58451283e-01 1.16667414e+00
1.09899707e-01 9.23541546e-01 3.87735873e-01 -4.10509676e-01
7.86539853e-01 -5.04704475e-01 2.62061328e-01 -3.21788669e-01
1.79830775e-01 1.75573245e-01 -4.41470981e-01 -1.56090587e-01
-2.39123031e-01 -7.48098791e-01 -1.74070263e+00 -2.86229681e-02
-6.19074583e-01 4.73187089e-01 7.60857880e-01 3.22545499e-01
2.21351415e-01 5.55614457e-02 5.88220060e-01 -1.10384822e+00
-7.02507436e-01 -1.03311658e+00 -9.67806101e-01 -5.29808663e-02
1.31765381e-01 -1.13634980e+00 -5.29625952e-01 3.67634394e-03] | [6.915713787078857, -1.9407333135604858] |
ecb96aa6-15b7-4365-bbb5-6851ffa9fc45 | zero-resource-cross-domain-named-entity | 2002.05923 | null | https://arxiv.org/abs/2002.05923v2 | https://arxiv.org/pdf/2002.05923v2.pdf | Zero-Resource Cross-Domain Named Entity Recognition | Existing models for cross-domain named entity recognition (NER) rely on numerous unlabeled corpus or labeled NER training data in target domains. However, collecting data for low-resource target domains is not only expensive but also time-consuming. Hence, we propose a cross-domain NER model that does not use any external resources. We first introduce a Multi-Task Learning (MTL) by adding a new objective function to detect whether tokens are named entities or not. We then introduce a framework called Mixture of Entity Experts (MoEE) to improve the robustness for zero-resource domain adaptation. Finally, experimental results show that our model outperforms strong unsupervised cross-domain sequence labeling models, and the performance of our model is close to that of the state-of-the-art model which leverages extensive resources. | ['Genta Indra Winata', 'Pascale Fung', 'Zihan Liu'] | 2020-02-14 | zero-resource-cross-domain-named-entity-1 | https://aclanthology.org/2020.repl4nlp-1.1 | https://aclanthology.org/2020.repl4nlp-1.1.pdf | ws-2020-7 | ['cross-domain-named-entity-recognition'] | ['natural-language-processing'] | [-9.51099768e-02 -2.46423498e-01 -3.76872480e-01 -3.83741170e-01
-1.17629921e+00 -9.44208086e-01 6.05990708e-01 -5.64953722e-02
-8.96664679e-01 1.09797585e+00 4.75679012e-03 -2.11248174e-01
4.54754651e-01 -6.28913403e-01 -6.37825966e-01 -1.71906546e-01
3.58168781e-01 7.20487356e-01 5.06470084e-01 -8.06938484e-02
-2.10248515e-01 2.76180387e-01 -8.56539667e-01 1.72366798e-01
1.26701164e+00 5.40668726e-01 2.17059553e-01 1.93802044e-01
-6.15833044e-01 6.40404344e-01 -6.64440393e-01 -7.26603568e-01
1.70931190e-01 -1.81623593e-01 -1.07635057e+00 -2.92081833e-01
1.08078778e-01 -1.57851592e-01 -8.14003497e-02 1.07306075e+00
6.85046732e-01 3.67622793e-01 7.93991446e-01 -1.06604719e+00
-1.08056295e+00 6.60739005e-01 -4.45739359e-01 8.02429318e-02
3.12868990e-02 -1.80159047e-01 8.10015738e-01 -1.01228833e+00
7.76434302e-01 8.84035289e-01 7.72091150e-01 8.71867239e-01
-1.06277120e+00 -9.75651681e-01 2.81837136e-01 6.56792149e-03
-1.35109878e+00 -5.25233924e-01 5.03987670e-01 -2.71068960e-01
1.03335392e+00 -4.59730268e-01 -2.63542831e-01 1.52099514e+00
-3.85247648e-01 9.76777136e-01 1.17752326e+00 -5.53200960e-01
2.55515814e-01 1.90695748e-01 6.86649084e-02 4.44406837e-01
3.14581305e-01 -1.63215399e-01 -1.94129169e-01 -2.12027594e-01
7.02101052e-01 -2.07600981e-01 2.98993923e-02 -3.78994703e-01
-1.17683852e+00 6.97043180e-01 -1.15243539e-01 6.27544641e-01
-3.51011604e-01 -4.44954842e-01 5.18904746e-01 2.20779821e-01
6.32961988e-01 4.00952101e-01 -1.22237003e+00 -1.41166046e-01
-8.44635487e-01 -1.46239087e-01 9.86120880e-01 1.25035119e+00
7.33510137e-01 -3.62219326e-02 7.97320455e-02 1.26490462e+00
1.65752292e-01 6.08362436e-01 6.30926192e-01 -4.58487391e-01
7.86638141e-01 6.15041375e-01 4.18011844e-01 -2.97737539e-01
-3.45383793e-01 -7.10980967e-02 -6.71895385e-01 -2.18471900e-01
4.77295488e-01 -7.11289823e-01 -1.00694561e+00 1.87221348e+00
4.59266186e-01 4.79199260e-01 4.40110952e-01 4.89731491e-01
8.66389811e-01 4.81085092e-01 8.73926163e-01 1.31509885e-01
1.52577591e+00 -1.16340947e+00 -7.05572665e-01 -5.57792902e-01
8.31315935e-01 -6.55682027e-01 9.25041080e-01 -1.34781733e-01
-6.30827129e-01 -5.72589517e-01 -7.42564201e-01 -1.24080233e-01
-7.96047211e-01 4.55671757e-01 4.18460041e-01 7.22877800e-01
-5.48849165e-01 3.37900877e-01 -7.93376803e-01 -4.33526486e-01
3.26847136e-01 1.19345054e-01 -6.83278084e-01 -2.35822976e-01
-1.62824404e+00 1.21931732e+00 8.24601829e-01 -4.79569972e-01
-8.26588988e-01 -8.51926744e-01 -1.03163838e+00 4.03485261e-02
4.04215962e-01 -4.43358243e-01 1.36399126e+00 -7.72250414e-01
-1.61278975e+00 9.70871985e-01 -1.53347790e-01 -1.59414694e-01
2.79380620e-01 -4.81477737e-01 -1.05435550e+00 -1.22658640e-01
5.39461970e-01 4.93219346e-01 3.61048520e-01 -1.26505637e+00
-8.38107824e-01 -9.80216563e-02 1.06629888e-02 -1.54674407e-02
-5.78025937e-01 3.49869102e-01 -4.12838280e-01 -6.66340172e-01
-5.96394360e-01 -7.67419159e-01 -4.45366055e-01 -7.05010653e-01
-3.55491698e-01 -5.30347407e-01 5.41861236e-01 -7.44990468e-01
1.22226000e+00 -2.05140400e+00 -3.68305266e-01 -2.45902732e-01
2.86670141e-02 6.97363853e-01 -4.85899568e-01 4.06421870e-01
-1.61740199e-01 4.41207141e-01 -2.02388763e-01 -3.43131632e-01
1.69612810e-01 1.67360246e-01 -1.40320316e-01 5.69104031e-02
5.37648141e-01 7.37069368e-01 -1.11760235e+00 -6.86452448e-01
1.25787817e-02 3.07422012e-01 -2.06587762e-01 3.60024750e-01
-5.99594824e-02 5.41892111e-01 -7.32420564e-01 5.49690962e-01
8.94641697e-01 -6.10482037e-01 5.53485632e-01 -1.07461989e-01
-2.56870389e-01 5.15975058e-01 -1.26019454e+00 1.81348860e+00
-6.44406259e-01 -6.82189092e-02 -1.83144689e-01 -8.96324456e-01
9.91091192e-01 6.57694876e-01 4.40617710e-01 -5.76908231e-01
-2.05117520e-02 4.09661651e-01 -4.28459674e-01 -2.95299232e-01
4.35257703e-01 -3.03552955e-01 -5.67271829e-01 4.01552022e-01
6.46727622e-01 4.84173119e-01 2.92606950e-01 -9.32585895e-02
1.19333601e+00 4.02410388e-01 6.83538914e-01 5.80070429e-02
4.76557970e-01 2.64021277e-01 1.18687689e+00 5.68237901e-01
-5.34903109e-01 1.54532358e-01 5.04589081e-02 -7.75750875e-02
-1.14244986e+00 -1.04488552e+00 -1.16015829e-01 1.55958843e+00
1.25435039e-01 -1.03728473e-01 -6.70747340e-01 -1.38908172e+00
-2.37458631e-01 7.78419137e-01 -3.28697443e-01 2.46257782e-01
-6.64945364e-01 -6.91280425e-01 1.07111382e+00 8.35065782e-01
6.72954082e-01 -1.08542764e+00 3.56055312e-02 3.43076050e-01
-5.51935732e-01 -1.60717618e+00 -6.69269085e-01 3.59365463e-01
-7.02257097e-01 -8.22768867e-01 -9.20224726e-01 -1.11759961e+00
4.61082399e-01 3.26258061e-03 1.42356825e+00 -6.79548860e-01
2.87105858e-01 2.52038777e-01 -6.41220629e-01 -3.35000902e-01
-4.04715598e-01 5.79165757e-01 2.16772348e-01 -3.25754166e-01
1.00356305e+00 -5.69290459e-01 -1.35754630e-01 4.77887332e-01
-9.06912386e-01 -3.45851302e-01 8.69854152e-01 9.81159449e-01
5.64048767e-01 -1.01115420e-01 1.05639911e+00 -1.18171310e+00
4.28752631e-01 -8.14406872e-01 -5.22659481e-01 7.36948192e-01
-3.27999383e-01 2.24984035e-01 9.44531500e-01 -7.69647896e-01
-1.71896493e+00 3.26925635e-01 -9.49508697e-02 -3.85785341e-01
-6.72721982e-01 4.21722233e-01 -5.99498093e-01 1.75814942e-01
6.27698064e-01 8.68171602e-02 -8.95738959e-01 -9.53550100e-01
7.00069189e-01 7.85466313e-01 5.37279189e-01 -8.87307644e-01
8.43636751e-01 1.74380556e-01 -6.14379764e-01 -5.32437027e-01
-1.22096622e+00 -7.46644676e-01 -1.11719370e+00 3.09287965e-01
9.41769540e-01 -1.30418420e+00 -2.10604399e-01 2.90659547e-01
-1.42584586e+00 -3.56943548e-01 3.45857963e-02 6.27091289e-01
-2.57601410e-01 4.71664995e-01 -7.67748117e-01 -5.84446609e-01
-2.57999063e-01 -6.17222309e-01 8.02188814e-01 3.44879240e-01
1.78139806e-02 -1.40207350e+00 5.47119796e-01 3.67484540e-01
1.52715579e-01 1.53025137e-02 8.23487163e-01 -1.62024713e+00
-7.81662464e-02 1.03053255e-02 -3.24754059e-01 2.78866023e-01
1.96926415e-01 -4.84207302e-01 -9.39896584e-01 -9.74519551e-03
-3.85309517e-01 -6.42741501e-01 5.14789462e-01 -8.35826173e-02
5.98314941e-01 -5.49488068e-02 -6.64484084e-01 3.66957366e-01
1.50357282e+00 9.19407010e-02 3.63741219e-01 6.39582932e-01
6.96364403e-01 5.29363155e-01 7.59810507e-01 3.92472804e-01
7.96260178e-01 3.65971297e-01 -2.44910404e-01 -4.27142888e-01
1.57404412e-02 -5.09223640e-01 3.98413062e-01 9.02378082e-01
7.38676488e-02 -4.51210618e-01 -1.19018579e+00 1.11142790e+00
-1.76367998e+00 -8.82769167e-01 3.25653404e-02 1.88016820e+00
1.20623255e+00 -1.87514931e-01 3.93914469e-02 -7.00616002e-01
1.10208595e+00 -8.52258429e-02 -8.33903134e-01 1.56646911e-02
-1.98260725e-01 6.01175427e-01 6.63275421e-01 -2.69273333e-02
-1.65795481e+00 1.35917485e+00 6.48622465e+00 9.82291400e-01
-7.14314044e-01 5.77969730e-01 1.14147574e-01 4.51946527e-01
-1.85190365e-02 -7.47537315e-02 -1.21422899e+00 4.90676790e-01
1.23830426e+00 -2.05585778e-01 1.48557387e-02 1.21595800e+00
-3.22476774e-01 5.07065177e-01 -8.35365534e-01 6.32402956e-01
-7.76298121e-02 -8.31504226e-01 -1.11527309e-01 -3.46533954e-02
9.10016596e-01 4.66092348e-01 -3.60719770e-01 8.78812075e-01
1.44237983e+00 -5.65252542e-01 2.55138874e-01 1.58422366e-01
9.31041121e-01 -6.02101386e-01 7.40109384e-01 5.36283731e-01
-1.32031095e+00 2.26266354e-01 -5.14281750e-01 5.71268082e-01
4.91265327e-01 5.49873769e-01 -7.17414141e-01 8.64111245e-01
5.67467093e-01 4.76366937e-01 -2.90498525e-01 8.87888312e-01
-4.98339891e-01 6.26429558e-01 -1.78760648e-01 2.32203096e-01
1.22418083e-01 4.44890521e-02 1.45210102e-01 1.63644469e+00
3.32093686e-01 9.14609507e-02 5.82534671e-01 5.94360471e-01
-6.28404140e-01 3.50796103e-01 -6.08643889e-01 -2.96964973e-01
8.92383397e-01 1.36186528e+00 -4.10166740e-01 -5.56942403e-01
-9.57147896e-01 1.35819030e+00 7.54246533e-01 3.89081657e-01
-8.66839707e-01 -6.24119580e-01 7.99509823e-01 -5.54564238e-01
7.26321042e-01 -2.07816288e-01 -1.32296667e-01 -1.70415378e+00
-9.20379609e-02 -7.26689160e-01 7.96106040e-01 -2.77549744e-01
-2.10931396e+00 6.67162001e-01 -3.13012540e-01 -1.38905478e+00
-1.51678771e-01 -6.37301922e-01 -3.72860909e-01 8.80048215e-01
-2.16410208e+00 -1.45460713e+00 1.84837639e-01 6.96524978e-01
3.82293373e-01 -2.12097585e-01 1.17114782e+00 7.79768944e-01
-6.48124933e-01 8.94047558e-01 4.39910620e-01 9.36069489e-01
1.40638995e+00 -1.25186169e+00 7.21436441e-01 9.67605293e-01
-2.40073390e-02 7.66155422e-01 1.36188418e-01 -8.17578852e-01
-7.44738102e-01 -1.37111974e+00 1.35080183e+00 -7.69846618e-01
8.69786739e-01 -3.41817588e-01 -1.04035175e+00 1.01721632e+00
2.10143998e-01 1.70460820e-01 1.41641915e+00 4.61186320e-01
-8.07724655e-01 4.02587712e-01 -1.21254706e+00 3.59433383e-01
1.03030550e+00 -5.26054442e-01 -1.09490240e+00 2.02379093e-01
7.70732403e-01 -1.31272718e-01 -1.28971887e+00 3.61781150e-01
2.30761498e-01 -3.84260952e-01 1.00379169e+00 -1.12983990e+00
1.80569574e-01 -3.30476463e-01 -9.45254490e-02 -1.45001638e+00
-4.40648645e-01 -1.67643890e-01 8.78221095e-02 1.85061061e+00
7.80752182e-01 -5.76388776e-01 4.63968396e-01 7.10111558e-01
-8.07178095e-02 1.89794540e-01 -7.53136456e-01 -1.33211005e+00
4.28930283e-01 -1.78381354e-01 6.56977654e-01 1.77613986e+00
1.20939828e-01 8.18633735e-01 -4.18796480e-01 5.50839365e-01
5.85422516e-01 -2.66947541e-02 6.00550950e-01 -1.42814445e+00
-7.50139654e-02 -6.19031005e-02 1.87544852e-01 -1.20790291e+00
7.36334264e-01 -8.57489288e-01 2.03539744e-01 -1.40664446e+00
2.59227902e-01 -7.19834745e-01 -7.50608206e-01 8.91981244e-01
-5.20259380e-01 -1.26232266e-01 1.00967735e-01 3.20105225e-01
-1.13927150e+00 6.03506207e-01 7.96264946e-01 1.60625488e-01
-2.26052478e-01 -3.07999164e-01 -7.28159308e-01 7.51372159e-01
6.94860995e-01 -9.29140449e-01 4.61579785e-02 -6.05889142e-01
-1.65613815e-01 -5.36507443e-02 -2.60887057e-01 -8.16816807e-01
3.07187319e-01 -3.78424168e-01 4.07892346e-01 -3.55884194e-01
1.58002768e-02 -7.77178764e-01 -2.44282827e-01 -1.33552641e-01
-2.19058260e-01 -7.64777213e-02 2.35385969e-01 6.12493038e-01
-2.39191413e-01 -3.76740724e-01 8.82912278e-01 -2.49799386e-01
-1.14844990e+00 3.89779210e-01 -2.54737794e-01 5.77767491e-01
9.44710314e-01 2.27678075e-01 -3.61345619e-01 1.74764976e-01
-5.89023590e-01 4.07082260e-01 4.45501596e-01 5.30265510e-01
-4.38361615e-02 -1.52286518e+00 -7.30325162e-01 -3.47158164e-01
5.08216739e-01 -2.43414894e-01 3.25494558e-01 2.61793911e-01
1.47393748e-01 4.64252114e-01 -2.74284303e-01 -7.77754709e-02
-9.39335287e-01 7.21212864e-01 3.05140726e-02 -9.92659569e-01
-1.80520505e-01 6.26733184e-01 2.65385360e-01 -1.39368284e+00
-2.35022679e-01 2.61028439e-01 -4.66138512e-01 7.92616699e-03
5.20174265e-01 3.43880355e-01 -1.40997887e-01 -5.22204280e-01
-5.44493854e-01 3.42935354e-01 -2.21651137e-01 -1.56520799e-01
1.45993674e+00 -1.97211564e-01 2.27418542e-01 3.90130609e-01
9.14914548e-01 8.81084278e-02 -1.13264644e+00 -7.67567098e-01
6.36819422e-01 -1.17102198e-01 -2.95413256e-01 -1.19400489e+00
-5.25352776e-01 7.55065620e-01 3.56346846e-01 -1.89207375e-01
9.96311963e-01 6.26724493e-03 1.13559794e+00 5.98244011e-01
5.66320479e-01 -1.43916821e+00 -7.87281543e-02 1.02221310e+00
4.14933823e-02 -1.56431079e+00 -3.96474928e-01 -3.72832507e-01
-1.05987179e+00 8.30705881e-01 9.59399939e-01 1.14055000e-01
6.73657358e-01 3.59474927e-01 4.65955108e-01 1.55097529e-01
-6.05839670e-01 -6.72240913e-01 1.16476573e-01 9.86020029e-01
7.00109005e-01 3.98622751e-02 -2.41366476e-01 1.12373817e+00
3.42577308e-01 1.85071275e-01 7.90600628e-02 8.87745857e-01
-3.50379646e-01 -1.74044633e+00 -9.49143171e-02 -3.79252322e-02
-8.11063588e-01 -4.42453384e-01 -4.01868194e-01 8.72095585e-01
2.41540357e-01 1.06827331e+00 -2.43631124e-01 -1.28091842e-01
5.95884621e-01 6.44532204e-01 8.11143070e-02 -8.55631649e-01
-5.58940589e-01 -1.59781784e-01 2.85031706e-01 -1.66180879e-01
-6.56281948e-01 -4.91015285e-01 -1.26821005e+00 -2.81630736e-02
-4.64399844e-01 3.11127305e-01 5.29103696e-01 1.10864162e+00
7.57474422e-01 2.13455066e-01 5.12105525e-01 -2.99199939e-01
-4.60053146e-01 -1.07274783e+00 -5.47624826e-01 5.32477736e-01
-3.48978229e-02 -7.53047347e-01 2.53207907e-02 3.62793684e-01] | [9.801044464111328, 9.534828186035156] |
4980fcfd-c926-4152-ad6d-6aeba51229cf | leveraged-weighted-loss-for-partial-label-1 | 2106.05731 | null | https://arxiv.org/abs/2106.05731v1 | https://arxiv.org/pdf/2106.05731v1.pdf | Leveraged Weighted Loss for Partial Label Learning | As an important branch of weakly supervised learning, partial label learning deals with data where each instance is assigned with a set of candidate labels, whereas only one of them is true. Despite many methodology studies on learning from partial labels, there still lacks theoretical understandings of their risk consistent properties under relatively weak assumptions, especially on the link between theoretical results and the empirical choice of parameters. In this paper, we propose a family of loss functions named \textit{Leveraged Weighted} (LW) loss, which for the first time introduces the leverage parameter $\beta$ to consider the trade-off between losses on partial labels and non-partial ones. From the theoretical side, we derive a generalized result of risk consistency for the LW loss in learning from partial labels, based on which we provide guidance to the choice of the leverage parameter $\beta$. In experiments, we verify the theoretical guidance, and show the high effectiveness of our proposed LW loss on both benchmark and real datasets compared with other state-of-the-art partial label learning algorithms. | ['Zhouchen Lin', 'Yisen Wang', 'Jiabin Liu', 'Hanyuan Hang', 'Jingyi Cui', 'Hongwei Wen'] | 2021-06-10 | leveraged-weighted-loss-for-partial-label | https://openreview.net/forum?id=DHkGKg2fJay | https://openreview.net/pdf?id=DHkGKg2fJay | null | ['partial-label-learning'] | ['methodology'] | [ 1.77919045e-01 4.85030621e-01 -7.06622005e-01 -7.64198780e-01
-1.00431848e+00 -5.19693136e-01 3.67176890e-01 5.09729028e-01
-4.12599832e-01 8.36256087e-01 -2.68034071e-01 -2.64081627e-01
-4.58312690e-01 -5.36814392e-01 -7.12689102e-01 -9.12387729e-01
-4.26242650e-02 3.97958666e-01 1.27182424e-01 4.72305894e-01
3.04550473e-02 1.40900567e-01 -1.38005555e+00 3.53187695e-02
7.49943674e-01 1.33075595e+00 -3.85689735e-01 -7.95825571e-02
-8.01032335e-02 8.43392611e-01 -1.75902978e-01 -8.19952428e-01
4.49457318e-01 -4.09367383e-01 -9.12818134e-01 -8.23074877e-02
4.53343719e-01 -2.02695772e-01 1.52024493e-01 1.21733797e+00
3.56919110e-01 -7.30657056e-02 1.03312576e+00 -1.62221813e+00
-2.56243110e-01 1.02015018e+00 -7.21238494e-01 -5.49529083e-02
-5.67896068e-02 -1.17837489e-01 1.60454488e+00 -7.05626011e-01
3.83726358e-01 1.02146876e+00 8.79566312e-01 4.66386557e-01
-1.22749925e+00 -9.44195509e-01 2.52092987e-01 -9.43143070e-02
-1.24331760e+00 -6.40745014e-02 6.19212091e-01 -5.75289845e-01
1.79287583e-01 1.63848419e-02 -6.33191988e-02 7.84166515e-01
-1.93387568e-01 9.16492462e-01 1.44214129e+00 -5.75006843e-01
2.84331769e-01 7.46006191e-01 8.18389237e-01 8.77762556e-01
3.99695635e-01 3.00584614e-01 -4.07380283e-01 -5.04272401e-01
2.14564234e-01 -2.80057341e-01 -2.64562666e-01 -7.77982473e-01
-7.35604167e-01 9.17679787e-01 1.81997135e-01 7.50420475e-03
1.20710008e-01 1.07738711e-01 4.42064643e-01 3.46008837e-01
8.21100354e-01 5.04459068e-02 -6.44110858e-01 1.56468570e-01
-8.40539634e-01 -4.38275635e-02 8.76120150e-01 8.81721556e-01
7.56368995e-01 -5.15335381e-01 -2.13119760e-01 9.08677638e-01
2.95312375e-01 2.23425299e-01 -4.77712080e-02 -7.82191336e-01
5.59122026e-01 4.45306361e-01 2.45010480e-01 -4.59959745e-01
-6.04592025e-01 -7.72704303e-01 -6.06984079e-01 -1.93774719e-02
8.18803251e-01 -1.71024382e-01 -3.69220585e-01 2.17660475e+00
3.70233804e-01 2.61069775e-01 -3.46351296e-01 6.44974768e-01
3.92292857e-01 3.16556334e-01 3.11884910e-01 -5.19527614e-01
1.04537809e+00 -8.30490589e-01 -5.22549152e-01 6.67629614e-02
1.03554213e+00 -4.08211142e-01 1.11796331e+00 2.96685666e-01
-9.43454266e-01 -2.79777467e-01 -9.38003659e-01 5.89272194e-02
-5.49756512e-02 5.33886105e-02 6.83336973e-01 8.05563331e-01
-6.16836429e-01 7.41217315e-01 -5.56942642e-01 -1.62905410e-01
4.33717906e-01 3.21178317e-01 -1.85674101e-01 -4.29628557e-03
-1.20388889e+00 7.10089862e-01 2.61716694e-01 -1.80326715e-01
-5.86882174e-01 -1.02079153e+00 -5.84495962e-01 1.48963809e-01
7.02326238e-01 -4.50217307e-01 1.23626328e+00 -8.07766616e-01
-1.06319594e+00 1.20763707e+00 1.11236624e-01 -4.93690014e-01
8.91704202e-01 -1.84592634e-01 4.66530211e-02 -9.31714624e-02
1.76530257e-01 3.83728951e-01 5.31673431e-01 -1.42434537e+00
-9.47728276e-01 -4.00088668e-01 1.24623887e-01 -3.94047201e-02
-4.37652647e-01 -1.55603826e-01 8.81911293e-02 -5.13480902e-01
-3.45517918e-02 -8.64078701e-01 1.01827666e-01 2.09066764e-01
-5.00607967e-01 -7.00867534e-01 5.02499342e-01 -1.76678002e-01
1.20272565e+00 -2.28867126e+00 -2.42501289e-01 2.80930966e-01
-9.09532681e-02 9.98631492e-02 1.23953134e-01 2.06218913e-01
-2.20089570e-01 2.93390334e-01 -4.39350754e-01 -5.61741233e-01
3.58322531e-01 3.11764106e-02 -5.30473590e-01 6.66468680e-01
-8.75798464e-02 6.28263533e-01 -9.40622091e-01 -5.08651376e-01
-1.35935351e-01 1.59020364e-01 -3.06019247e-01 2.22151101e-01
-1.67952523e-01 2.58164294e-02 -6.01608992e-01 5.49230516e-01
7.12796807e-01 -4.68621582e-01 1.11789897e-01 -1.91718593e-01
2.80117512e-01 2.06885591e-01 -1.21823990e+00 1.16836727e+00
-3.80748093e-01 -9.59045365e-02 1.73873395e-01 -1.29334915e+00
7.45706320e-01 2.99496710e-01 4.66312706e-01 -2.04389825e-01
2.13408902e-01 4.11490709e-01 -4.72988218e-01 -2.70554096e-01
-1.07482374e-01 -7.19512343e-01 -3.34126167e-02 9.47248459e-01
7.50758424e-02 1.52956754e-01 3.61221582e-02 5.16993515e-02
9.13112521e-01 1.57899305e-01 3.71608615e-01 -4.97825027e-01
4.79419202e-01 -4.09225523e-01 6.46785080e-01 8.60979557e-01
-4.38423544e-01 5.03133416e-01 9.62015212e-01 -3.09405774e-01
-7.38237143e-01 -9.63039398e-01 -7.21395731e-01 1.19652474e+00
1.92408308e-01 -7.75100663e-02 -6.94286346e-01 -1.41026533e+00
3.20415884e-01 8.37761819e-01 -7.67227054e-01 -3.09844315e-01
-2.33109593e-01 -1.07750523e+00 6.39009833e-01 3.82524908e-01
3.79529625e-01 -6.95167840e-01 -2.66376972e-01 -1.69855103e-01
1.11490563e-02 -9.94461000e-01 -4.13445085e-01 6.52815938e-01
-8.17512214e-01 -1.31058776e+00 -3.98413211e-01 -8.19245934e-01
7.30068743e-01 -5.46415001e-02 1.19890440e+00 5.13752252e-02
1.19527400e-01 3.09530854e-01 -2.68958449e-01 -3.73226285e-01
-3.28599662e-01 2.98242718e-01 -1.05178999e-02 1.30026132e-01
3.62707585e-01 -5.18385053e-01 -2.82581925e-01 5.09776890e-01
-8.37564886e-01 -2.13806689e-01 4.64848489e-01 8.90597939e-01
7.43333817e-01 2.62269706e-01 8.48162532e-01 -1.63340211e+00
3.71247053e-01 -6.65525675e-01 -6.51759207e-01 5.86891830e-01
-1.04453135e+00 3.64113539e-01 4.26515758e-01 -3.28520268e-01
-1.04998457e+00 -9.18862224e-02 3.28565426e-02 1.38169050e-03
5.06770723e-02 4.18521434e-01 -2.83848047e-01 9.47833955e-02
3.72145385e-01 -2.67694294e-01 -1.47119313e-01 -6.79062247e-01
3.31631213e-01 5.89323401e-01 9.45185870e-02 -9.31397796e-01
7.39361703e-01 5.90585828e-01 3.21042925e-01 -2.88462549e-01
-1.77801013e+00 -3.81967455e-01 -6.54963911e-01 8.48631188e-02
3.72441471e-01 -5.82379818e-01 -6.51229262e-01 3.49801362e-01
-5.35449445e-01 -4.00213152e-01 -5.20133555e-01 3.68698359e-01
-6.06144667e-01 3.98822576e-01 -6.78867698e-01 -9.29626822e-01
-1.18434712e-01 -9.00992453e-01 7.70954251e-01 -2.22494096e-01
9.92181599e-02 -1.18635869e+00 3.00084017e-02 4.27819252e-01
-9.95005574e-03 2.00582728e-01 1.37402058e+00 -9.70571160e-01
-1.95783198e-01 -3.01211625e-01 -4.50468212e-01 6.81209087e-01
-2.16142423e-02 -3.17953408e-01 -1.05622244e+00 -4.57128882e-01
2.50620425e-01 -7.26214170e-01 1.14884770e+00 4.84901577e-01
1.10304940e+00 -2.46168390e-01 -2.92153418e-01 4.30435330e-01
1.43782091e+00 -2.49560893e-01 1.20930433e-01 1.35749519e-01
4.20066655e-01 9.00389075e-01 8.63921523e-01 5.48131049e-01
2.12959319e-01 5.88150799e-01 4.52029616e-01 6.51533157e-02
2.01410487e-01 -3.37664187e-01 2.23472804e-01 4.51333553e-01
3.92238408e-01 -8.43873098e-02 -5.25546312e-01 2.69363821e-01
-1.73761296e+00 -5.84651768e-01 1.25014737e-01 2.58607745e+00
1.05460119e+00 4.68435079e-01 2.39923611e-01 2.62272477e-01
9.27862108e-01 9.05441493e-02 -6.61282599e-01 -1.24902420e-01
-1.18001863e-01 5.96141219e-02 6.03507876e-01 6.20230496e-01
-1.33703780e+00 5.91459751e-01 5.73321056e+00 1.19967306e+00
-8.78408790e-01 2.88190126e-01 9.71914887e-01 -1.41586140e-01
-2.39346176e-01 1.86334923e-01 -1.07687330e+00 4.59060311e-01
7.22138047e-01 -7.03568235e-02 3.41427363e-02 1.03922522e+00
-1.89931661e-01 2.63126940e-02 -1.52911747e+00 4.55069900e-01
-7.39211962e-02 -6.78468823e-01 -2.31220365e-01 2.40475580e-01
8.75030220e-01 -2.61243433e-01 2.13339537e-01 4.27678883e-01
3.51422817e-01 -7.67130554e-01 8.68067920e-01 2.02062219e-01
7.57102549e-01 -9.62856650e-01 9.83483136e-01 6.69806540e-01
-8.89967084e-01 -3.00126731e-01 -3.77984732e-01 -3.66943516e-02
6.68384209e-02 1.16941214e+00 -5.11561692e-01 4.27120298e-01
2.13274956e-01 5.33342600e-01 -4.59929734e-01 8.04233909e-01
-3.35380435e-01 9.81603086e-01 -4.14595187e-01 2.66126722e-01
1.41589075e-01 -2.11432576e-02 1.38271004e-01 1.08884335e+00
1.56301573e-01 1.12396628e-02 2.97517926e-01 7.17318535e-01
-4.31992352e-01 2.93745100e-01 -4.28666204e-01 3.55516016e-01
5.03445864e-01 1.20029271e+00 -7.63189375e-01 1.24296183e-02
-4.70691949e-01 3.14389139e-01 6.63159430e-01 2.11722448e-01
-8.01802099e-01 -2.69783679e-02 2.26723045e-01 2.55248904e-01
6.62005544e-02 3.38117450e-01 -4.87919778e-01 -9.88914192e-01
2.53554225e-01 -3.71924400e-01 9.27339554e-01 -1.64088935e-01
-1.83394861e+00 7.40369260e-02 2.98468530e-01 -1.07717812e+00
9.54645872e-02 -5.82824409e-01 -2.43600681e-01 5.38463295e-01
-1.61417639e+00 -1.08020985e+00 2.37047032e-01 3.06515038e-01
4.59887460e-02 1.65940627e-01 5.70343077e-01 3.09555471e-01
-5.43565273e-01 1.00580382e+00 2.68356025e-01 1.07996119e-03
8.10942769e-01 -1.29785144e+00 -3.09293687e-01 4.31363463e-01
1.36229753e-01 3.79570127e-01 6.77048385e-01 -2.63588578e-01
-6.05948627e-01 -1.02802050e+00 8.86429250e-01 -4.15412784e-01
5.61850011e-01 -3.05351019e-01 -9.31660414e-01 6.69368267e-01
-3.08106124e-01 3.86291653e-01 9.31741774e-01 2.88452983e-01
-7.85578489e-01 -2.62633055e-01 -1.50537026e+00 1.66310742e-01
9.10616636e-01 -3.57713968e-01 -3.38432372e-01 3.96132708e-01
7.24585950e-01 8.14023018e-02 -8.12775075e-01 8.75742018e-01
6.39687479e-01 -1.10778868e+00 8.05041909e-01 -6.17381334e-01
3.16295207e-01 3.41776647e-02 -1.75010175e-01 -1.02637792e+00
-1.11149818e-01 -1.70542866e-01 -1.12246767e-01 1.59839487e+00
6.39007330e-01 -6.72203481e-01 8.88855219e-01 8.94458950e-01
1.98417589e-01 -1.15267360e+00 -1.06533051e+00 -9.61095095e-01
4.72906113e-01 -5.51250637e-01 3.08048368e-01 1.03637540e+00
1.35181710e-01 2.18964458e-01 -4.63464379e-01 4.10520807e-02
9.86254811e-01 1.61473155e-01 2.15157092e-01 -1.48637116e+00
-4.36398774e-01 -4.31865036e-01 -1.06962651e-01 -8.95905852e-01
6.25298738e-01 -1.06283939e+00 -9.21042189e-02 -1.10190535e+00
6.63463831e-01 -9.92146969e-01 -8.96088064e-01 6.49271369e-01
-3.18946004e-01 1.80575818e-01 4.88916412e-02 1.72526151e-01
-7.50893891e-01 4.29126501e-01 9.47603166e-01 -6.75134063e-02
5.33273369e-02 4.25200552e-01 -7.95781076e-01 8.81035268e-01
5.97535491e-01 -8.63470793e-01 -4.95822459e-01 2.89126206e-02
4.68994498e-01 -1.11477725e-01 1.38649955e-01 -6.16957128e-01
-2.49414742e-02 -1.92454249e-01 -3.72150838e-01 -3.37538302e-01
-9.54647511e-02 -8.20279121e-01 -2.25947201e-01 4.22213942e-01
-1.08274329e+00 -6.97739840e-01 -3.34977865e-01 7.80530334e-01
7.62041658e-02 -5.95842183e-01 9.90914881e-01 5.63737825e-02
-4.35709916e-02 3.23959589e-01 3.88031304e-01 6.33369327e-01
1.07842457e+00 2.46429726e-01 -1.68382674e-01 -2.81736314e-01
-6.32699013e-01 3.41733396e-01 2.44877204e-01 1.15493871e-01
4.51788828e-02 -1.24288034e+00 -7.28706300e-01 1.06963841e-02
1.38068378e-01 -7.22060204e-02 9.09695178e-02 8.23543847e-01
7.24425018e-02 4.41944599e-01 3.50526392e-01 -4.08393115e-01
-1.03926969e+00 8.00866365e-01 1.75931424e-01 -6.07327938e-01
-3.64576459e-01 9.48298872e-01 5.96949041e-01 -5.13567090e-01
7.99967587e-01 -1.81280822e-01 6.93713576e-02 2.25313708e-01
2.49101236e-01 5.49591124e-01 6.95908591e-02 -4.20649141e-01
-1.69815049e-01 5.54581881e-01 -9.73681957e-02 -1.22219026e-02
1.09584975e+00 -2.28668094e-01 -1.01176471e-01 6.94721639e-01
1.32203948e+00 -5.94680607e-02 -1.45617819e+00 -6.99867547e-01
3.57594669e-01 -4.20477420e-01 -1.25484869e-01 -9.18390393e-01
-1.19399989e+00 9.13196981e-01 5.70636988e-01 2.37438142e-01
8.78789186e-01 2.49456957e-01 5.96045136e-01 2.86183357e-01
6.66711807e-01 -1.06759799e+00 -2.57473290e-02 2.13236094e-01
3.26985806e-01 -1.25701833e+00 6.90757781e-02 -5.95042109e-01
-5.86849749e-01 7.12670267e-01 4.37019825e-01 1.01371065e-01
1.08926225e+00 1.47243023e-01 -2.82399543e-03 1.69184897e-02
-7.82352507e-01 3.28469388e-02 1.73578098e-01 2.68304944e-01
4.12891150e-01 1.15491450e-01 -7.20891953e-01 9.66154337e-01
4.83683031e-03 -8.87911767e-02 1.35404214e-01 6.56979978e-01
-4.50373769e-01 -1.31506896e+00 5.87977320e-02 5.66246986e-01
-8.53566468e-01 1.23703934e-01 -2.07571939e-01 7.81599939e-01
3.66856664e-01 8.12668920e-01 -4.20994639e-01 -1.02935126e-02
2.06904843e-01 3.80395472e-01 4.00122136e-01 -6.43399954e-01
-3.42660666e-01 -1.62945166e-01 -1.43584935e-02 -1.46694854e-01
-5.30241549e-01 -6.66042984e-01 -1.14294553e+00 1.50131220e-02
-6.77794099e-01 4.38276559e-01 4.07331735e-01 1.10634327e+00
-1.12117901e-01 -3.08965594e-02 8.27758908e-01 -3.22982848e-01
-1.32864797e+00 -8.35349023e-01 -1.23417783e+00 7.88603365e-01
2.94987410e-01 -9.53128695e-01 -9.32927251e-01 -2.30475783e-01] | [9.16380500793457, 4.167608737945557] |
41e07ea9-7d59-4b1f-9754-f63230ff8493 | relation-dependent-contrastive-learning-with | 2211.12266 | null | https://arxiv.org/abs/2211.12266v1 | https://arxiv.org/pdf/2211.12266v1.pdf | Relation-dependent Contrastive Learning with Cluster Sampling for Inductive Relation Prediction | Relation prediction is a task designed for knowledge graph completion which aims to predict missing relationships between entities. Recent subgraph-based models for inductive relation prediction have received increasing attention, which can predict relation for unseen entities based on the extracted subgraph surrounding the candidate triplet. However, they are not completely inductive because of their disability of predicting unseen relations. Moreover, they fail to pay sufficient attention to the role of relation as they only depend on the model to learn parameterized relation embedding, which leads to inaccurate prediction on long-tail relations. In this paper, we introduce Relation-dependent Contrastive Learning (ReCoLe) for inductive relation prediction, which adapts contrastive learning with a novel sampling method based on clustering algorithm to enhance the role of relation and improve the generalization ability to unseen relations. Instead of directly learning embedding for relations, ReCoLe allocates a pre-trained GNN-based encoder to each relation to strengthen the influence of relation. The GNN-based encoder is optimized by contrastive learning, which ensures satisfactory performance on long-tail relations. In addition, the cluster sampling method equips ReCoLe with the ability to handle both unseen relations and entities. Experimental results suggest that ReCoLe outperforms state-of-the-art methods on commonly used inductive datasets. | ['Haifeng Hu', 'Sijie Mai', 'Jianfeng Wu'] | 2022-11-22 | null | null | null | null | ['inductive-relation-prediction'] | ['graphs'] | [ 1.01910323e-01 7.55197585e-01 -5.01368105e-01 -3.91639441e-01
-2.80559957e-01 -2.39395916e-01 5.50604582e-01 3.59247267e-01
-1.68226898e-01 7.51746356e-01 2.26574495e-01 -2.52834946e-01
-4.50933278e-01 -1.20609415e+00 -7.69084156e-01 -4.95628983e-01
-2.78013319e-01 7.52883434e-01 2.81809598e-01 -4.19702321e-01
-4.59160358e-01 3.24054301e-01 -1.36363626e+00 3.46042007e-01
1.10318828e+00 7.92137563e-01 1.06666565e-01 1.02533408e-01
-1.01885200e-01 1.18444526e+00 -2.05986381e-01 -7.75749147e-01
-6.86461329e-02 -4.38157201e-01 -1.16206288e+00 -2.71200359e-01
-6.06018901e-02 3.03649977e-02 -6.84944451e-01 7.19828486e-01
3.23788106e-01 1.88975304e-01 7.42838800e-01 -1.31324470e+00
-8.21264207e-01 1.05437255e+00 -3.25980455e-01 1.85709700e-01
4.91219342e-01 -5.98765433e-01 1.61980987e+00 -9.43417072e-01
7.62569070e-01 1.06604433e+00 6.01465940e-01 4.14952427e-01
-1.21950877e+00 -6.36069179e-01 2.67735898e-01 6.62431896e-01
-1.61629474e+00 -3.17018688e-01 1.00377917e+00 -2.13843912e-01
1.23893929e+00 3.05041313e-01 5.58102608e-01 9.21550035e-01
-1.47456571e-01 7.27567434e-01 2.66579360e-01 -5.41640043e-01
-8.77925977e-02 1.16596729e-01 1.59254760e-01 5.81452310e-01
3.07633996e-01 -1.80243365e-02 -5.08594513e-01 2.69170720e-02
3.84036243e-01 -1.80358112e-01 -4.37865108e-01 -5.84927797e-01
-1.06987584e+00 6.35949492e-01 1.09286249e+00 3.96388203e-01
-2.88655013e-01 -2.19106510e-01 1.98813647e-01 1.85124263e-01
6.20142102e-01 5.30647516e-01 -7.02030778e-01 4.09252882e-01
-3.41655165e-01 -1.19059250e-01 9.34550226e-01 1.26500070e+00
1.08312976e+00 -3.83481085e-01 -1.16185889e-01 7.44793355e-01
2.43051797e-01 -8.77806023e-02 1.09152578e-01 -4.02837962e-01
7.47212052e-01 1.34099245e+00 -3.13554734e-01 -1.08893418e+00
-6.52761519e-01 -7.22199261e-01 -1.14049292e+00 -5.78772426e-01
2.59023625e-02 -1.13880523e-02 -6.52051389e-01 1.65436411e+00
4.11613137e-01 4.42846626e-01 3.26064587e-01 5.63148320e-01
1.32230878e+00 5.92599928e-01 2.68770158e-01 -3.39441925e-01
1.04640675e+00 -1.05182803e+00 -7.32216001e-01 -2.24242508e-01
1.25035763e+00 -2.58832961e-01 7.50365973e-01 -1.69652164e-01
-5.95663905e-01 -5.66151440e-01 -8.29448104e-01 -2.32421324e-01
-6.47514462e-01 1.47734970e-01 1.04098201e+00 1.72378391e-01
-7.97465861e-01 4.87071484e-01 -6.81955457e-01 -4.39665318e-01
4.23501700e-01 6.10724390e-01 -5.79286933e-01 -6.02681115e-02
-1.76726007e+00 1.08333004e+00 1.06816924e+00 3.84511083e-01
-1.64480269e-01 -6.45681441e-01 -1.26778758e+00 4.43525672e-01
6.35870099e-01 -7.41434395e-01 5.11767089e-01 -5.89623034e-01
-9.67336595e-01 6.63698673e-01 -3.58802646e-01 -5.32051086e-01
-8.31337571e-02 -7.40478337e-02 -7.72939861e-01 6.75268248e-02
1.34355456e-01 5.73328018e-01 2.59614825e-01 -1.47421372e+00
-5.23060501e-01 -2.13412687e-01 2.46273741e-01 3.78106564e-01
-5.69912493e-01 -4.05379534e-01 -5.01830339e-01 -4.29895878e-01
3.68882298e-01 -7.72870958e-01 -5.10502905e-02 -6.25856519e-01
-6.97931111e-01 -6.89541817e-01 7.37161875e-01 -3.19629997e-01
1.54555571e+00 -2.04221964e+00 1.37679487e-01 3.87918144e-01
4.84925032e-01 3.98785442e-01 -1.96532547e-01 6.25243723e-01
-4.26488876e-01 -3.73627711e-03 -6.36456627e-03 -6.03282899e-02
-1.48715064e-01 5.08501709e-01 -1.40170857e-01 9.69158560e-02
5.20510554e-01 1.37189364e+00 -1.21493506e+00 -7.28505731e-01
6.00643978e-02 4.75628585e-01 -4.15310681e-01 2.62172699e-01
-1.58144623e-01 2.77930826e-01 -4.94824022e-01 3.89178395e-01
4.90464032e-01 -7.16619849e-01 7.28551984e-01 -4.96626884e-01
5.10634899e-01 5.73943496e-01 -9.26267266e-01 1.36399782e+00
-4.05041665e-01 1.58832788e-01 -5.70471466e-01 -1.23743463e+00
1.11326849e+00 4.18571413e-01 5.71148396e-01 -4.41804647e-01
-5.78017198e-02 1.51337907e-01 2.26635575e-01 -5.29068053e-01
3.35181087e-01 2.66310573e-02 1.56461775e-01 -4.57218476e-02
3.50999206e-01 3.19056481e-01 3.20294321e-01 5.83767295e-01
1.27863061e+00 3.01454723e-01 3.98667514e-01 8.50119591e-02
7.98365116e-01 -1.44232526e-01 8.05600286e-01 2.19564393e-01
3.58256370e-01 1.84303150e-01 5.50252616e-01 -4.47618723e-01
-4.72981602e-01 -1.01086318e+00 -4.43806797e-02 1.06204391e+00
5.23393273e-01 -8.82325470e-01 -1.70046642e-01 -1.24407029e+00
-1.41538545e-01 7.42597163e-01 -7.05688596e-01 -5.66173255e-01
-5.55999875e-01 -8.49425137e-01 1.53446451e-01 5.06482065e-01
4.64497387e-01 -1.14251697e+00 3.36311370e-01 2.54273236e-01
-5.59516966e-01 -1.34010124e+00 -8.25221166e-02 4.84797925e-01
-6.71240866e-01 -1.33745456e+00 -6.51056366e-03 -1.27928078e+00
9.49820995e-01 -1.25072161e-02 1.31065679e+00 2.19451055e-01
3.77095312e-01 5.09306677e-02 -7.01747298e-01 3.19186077e-02
-2.47105613e-01 7.31261253e-01 -1.80186570e-01 1.22415356e-01
7.80900836e-01 -7.09812760e-01 -2.44808748e-01 2.52920300e-01
-5.79834938e-01 1.42434850e-01 5.88126242e-01 1.02756727e+00
4.68209177e-01 4.03194398e-01 7.79430807e-01 -1.35159099e+00
3.29225272e-01 -5.65560877e-01 -7.66900033e-02 6.85402870e-01
-7.93039143e-01 2.80442655e-01 6.90352023e-01 -3.64518523e-01
-1.12046945e+00 2.11796448e-01 -4.57634078e-03 -1.47942737e-01
8.36791098e-02 7.95281053e-01 -4.71584827e-01 1.05376676e-01
5.80515563e-01 3.30373943e-02 -5.58101296e-01 -2.61717796e-01
3.76169175e-01 3.20676565e-01 3.73474628e-01 -4.38799798e-01
9.03494596e-01 1.41189337e-01 2.60663331e-01 -4.00004655e-01
-1.26091743e+00 -4.54896271e-01 -8.61953855e-01 6.32522106e-02
7.46919990e-01 -9.97801304e-01 -7.74207115e-01 -6.30423948e-02
-1.17409098e+00 -8.95388871e-02 -3.47067863e-01 6.53082311e-01
-2.00762525e-01 2.49024555e-01 -7.82822549e-01 -6.27378464e-01
-2.75847197e-01 -6.82431400e-01 7.64512599e-01 2.39404272e-02
-1.99384585e-01 -1.19679296e+00 -1.26830101e-01 3.23289126e-01
-5.34570813e-02 2.10656822e-01 1.28229380e+00 -1.03214049e+00
-7.03600824e-01 -1.97542563e-01 -3.82579595e-01 8.51798803e-02
3.04209977e-01 -3.55899453e-01 -8.37911069e-01 -1.03412624e-02
-5.99659383e-01 -4.08462107e-01 1.02884090e+00 -4.57295850e-02
8.01402509e-01 -2.91261435e-01 -9.33666170e-01 5.04569173e-01
1.12167311e+00 3.17998566e-02 6.83628559e-01 2.74636120e-01
1.23957777e+00 7.51287282e-01 7.21951604e-01 4.52396162e-02
9.65404928e-01 6.14881814e-01 4.25835580e-01 -2.49707505e-01
-1.53433114e-01 -6.24199867e-01 -9.37095210e-02 8.53480220e-01
-3.07293445e-01 -3.45302641e-01 -8.77512157e-01 5.29003263e-01
-2.11726403e+00 -7.78533101e-01 -4.76674289e-01 1.96756041e+00
1.09736300e+00 2.28096932e-01 -2.17125803e-01 1.89104676e-01
8.03332865e-01 -3.67513373e-02 -2.96333432e-01 -9.21748485e-03
-2.64089346e-01 1.40580639e-01 6.08702414e-02 5.69281340e-01
-1.12971163e+00 1.31798077e+00 4.59514761e+00 9.32728231e-01
-5.95557213e-01 -6.09846152e-02 4.39358234e-01 5.04989922e-01
-5.08761644e-01 2.98177689e-01 -8.68056476e-01 1.12956107e-01
6.60824299e-01 4.07445198e-03 1.24313742e-01 6.88307464e-01
-3.28418106e-01 5.23002408e-02 -1.32763255e+00 6.77099705e-01
-1.11556247e-01 -1.07095397e+00 1.25580817e-01 -3.57886180e-02
7.26318717e-01 -2.61844993e-01 -4.06111151e-01 7.26866663e-01
1.87994093e-01 -9.67576325e-01 4.61955890e-02 4.55979645e-01
5.67700088e-01 -8.95230532e-01 1.08576274e+00 3.82439554e-01
-1.50336707e+00 1.17585495e-01 -3.74032050e-01 -5.90852425e-02
3.08354795e-02 8.33265364e-01 -1.11904979e+00 1.06239879e+00
4.85569835e-01 1.00117373e+00 -7.14509130e-01 6.44938052e-01
-7.48735249e-01 5.60531080e-01 -2.83720255e-01 2.89946813e-02
-1.33470550e-01 -7.32923597e-02 1.69745117e-01 9.62100267e-01
1.15110494e-01 2.59663731e-01 1.78274795e-01 5.94326138e-01
-3.26027542e-01 2.24123955e-01 -5.76878309e-01 2.14198783e-01
5.91741621e-01 1.26420641e+00 -4.88257796e-01 -2.13439777e-01
-5.46124279e-01 7.85582602e-01 9.85355318e-01 5.11203289e-01
-6.68474615e-01 -3.92709762e-01 9.18175504e-02 1.68190420e-01
4.53047782e-01 1.13417290e-01 -1.33607030e-01 -1.26237321e+00
1.20350763e-01 -2.97343105e-01 6.30799234e-01 -4.46957707e-01
-1.32007647e+00 8.91662002e-01 -1.02333650e-01 -1.15994608e+00
-3.22777063e-01 -2.10411444e-01 -3.31094176e-01 6.60664976e-01
-1.63056815e+00 -1.63494110e+00 -1.93437889e-01 5.76784372e-01
-1.45752236e-01 -1.09905347e-01 9.64173317e-01 4.19345200e-01
-7.16267765e-01 6.95565820e-01 -4.56871480e-01 3.36946249e-01
6.98508561e-01 -1.33447003e+00 1.60730541e-01 6.10779643e-01
3.00059766e-01 6.32675827e-01 4.00455028e-01 -7.85959482e-01
-9.96717393e-01 -1.38267577e+00 1.67078638e+00 -4.07729685e-01
6.37069881e-01 -4.44801390e-01 -1.07626057e+00 9.83247995e-01
-9.27826855e-03 5.80782413e-01 9.54162419e-01 8.90124142e-01
-4.55478936e-01 -4.56623822e-01 -7.97204256e-01 5.35349905e-01
1.40691280e+00 -4.90292549e-01 -6.62112653e-01 3.39439958e-01
8.15228283e-01 -2.88386494e-01 -1.23590755e+00 9.72068787e-01
1.25747934e-01 -7.11719215e-01 1.03494024e+00 -6.07475758e-01
5.66138148e-01 -3.31150413e-01 1.80715397e-01 -1.18881559e+00
-5.93309760e-01 -3.13234270e-01 -8.73837471e-01 1.60635984e+00
9.21584308e-01 -6.25421166e-01 9.73610044e-01 4.51842844e-01
6.53654411e-02 -1.18366289e+00 -5.27891755e-01 -7.16591597e-01
-3.51170689e-01 -1.70182019e-01 6.80226505e-01 1.33221805e+00
4.42335874e-01 1.24117577e+00 -3.55261445e-01 4.73433435e-01
3.78657669e-01 4.29536819e-01 5.91256738e-01 -1.40572083e+00
-2.96466470e-01 6.50695637e-02 -5.39098561e-01 -1.01877415e+00
6.26676381e-01 -1.28486955e+00 -1.82330295e-01 -1.75301969e+00
2.64146775e-01 -7.65649974e-01 -3.52469683e-01 7.19398797e-01
-6.96064055e-01 2.01706082e-01 -9.52879414e-02 1.13205709e-01
-9.05141532e-01 8.89681339e-01 1.30939221e+00 -1.85811937e-01
-3.36098462e-01 2.09265292e-01 -7.63598144e-01 5.50050855e-01
6.72013700e-01 -3.97989899e-01 -8.07073772e-01 -1.89649910e-01
6.02985263e-01 -6.16661943e-02 2.76490152e-01 -7.19227970e-01
4.34683114e-01 2.10551262e-01 5.03368914e-01 -6.22975409e-01
3.39352071e-01 -9.92741704e-01 1.27325073e-01 1.60779312e-01
-4.30841237e-01 -4.99192953e-01 -2.10828736e-01 6.61778927e-01
-4.73490000e-01 -8.67497176e-02 2.18100578e-01 1.87824279e-01
-8.12358439e-01 5.03844619e-01 2.43256778e-01 2.47618124e-01
9.75236773e-01 3.13432283e-05 -2.85894871e-01 -3.13901693e-01
-1.06154358e+00 5.18039346e-01 -1.49536217e-02 3.45280439e-01
4.97343779e-01 -1.53407824e+00 -6.38527811e-01 -2.17612181e-02
4.52989310e-01 4.87931907e-01 8.45436975e-02 7.39902258e-01
-4.80327243e-03 3.15694362e-01 4.37021136e-01 -2.43871585e-01
-1.05605423e+00 9.97784674e-01 9.62876529e-02 -8.35139096e-01
-5.54716527e-01 1.06785071e+00 2.83459574e-01 -7.15615332e-01
1.30483255e-01 -1.04592308e-01 -5.08251905e-01 8.33678171e-02
7.54311979e-02 1.12230428e-01 2.55531162e-01 -6.33709252e-01
-6.35924637e-01 3.41261327e-01 -2.75952339e-01 6.07446551e-01
1.27877915e+00 -7.62568042e-02 -2.84664035e-01 2.49463275e-01
1.17255330e+00 -9.29822102e-02 -8.37594688e-01 -5.66521883e-01
3.51929665e-01 -1.33287096e-02 -1.76234454e-01 -6.47028208e-01
-1.05670130e+00 4.80489045e-01 -2.15538159e-01 3.73225838e-01
1.16669738e+00 4.68429804e-01 7.64286637e-01 5.93138814e-01
3.99863362e-01 -7.98777223e-01 -1.90550476e-01 7.12015092e-01
6.18414164e-01 -1.31268919e+00 1.91032767e-01 -1.23374116e+00
-5.32917500e-01 9.57980514e-01 9.17479455e-01 1.19640246e-01
8.68549645e-01 1.44006982e-01 -3.16681862e-01 -2.60170788e-01
-8.11192155e-01 -6.94352090e-01 6.43615365e-01 7.47565866e-01
6.26695216e-01 8.32923576e-02 -3.42339188e-01 6.32922947e-01
-2.30758548e-01 -2.88675874e-01 -6.37868941e-02 5.06713986e-01
-8.68992284e-02 -1.33065689e+00 2.04106137e-01 6.63930893e-01
-7.64269903e-02 -4.01383042e-01 -4.50256318e-01 7.54381597e-01
4.65827852e-01 1.07117462e+00 -1.40666768e-01 -7.20526695e-01
3.31743330e-01 -1.10597443e-02 4.16459680e-01 -8.24926615e-01
-4.24525231e-01 -3.24506998e-01 4.72639382e-01 -3.34858924e-01
-5.69816649e-01 -3.46347362e-01 -1.41234481e+00 -1.03704832e-01
-8.94262850e-01 5.08451343e-01 -1.64707989e-01 1.21313000e+00
5.26565373e-01 7.08796024e-01 5.79235137e-01 -2.84000784e-01
1.37304794e-02 -9.91004109e-01 -6.77166700e-01 5.21717370e-01
8.15616250e-02 -7.32368410e-01 -1.33008614e-01 -1.84962705e-01] | [8.954375267028809, 8.119715690612793] |
86933b98-e14d-423e-8665-bf4c15afe133 | explainable-inference-on-sequential-data-via | null | null | https://www.ijcai.org/Proceedings/2020/278 | https://www.ijcai.org/Proceedings/2020/0278.pdf | Explainable Inference on Sequential Data via Memory-Tracking | In this paper we present a novel mechanism to
get explanations that allow to better understand
network predictions when dealing with sequential
data. Specifically, we adopt memory-based networks — Differential Neural Computers — to exploit their capability of storing data in memory and
reusing it for inference. By tracking both the memory access at prediction time, and the information
stored by the network at each step of the input
sequence, we can retrieve the most relevant input
steps associated to each prediction. We validate
our approach (1) on a modified T-maze, which is a
non-Markovian discrete control task evaluating an
algorithm’s ability to correlate events far apart in
history, and (2) on the Story Cloze Test, which is
a commonsense reasoning framework for evaluating story understanding that requires a system to
choose the correct ending to a four-sentence story.
Our results show that we are able to explain agent’s
decisions in (1) and to reconstruct the most relevant
sentences used by the network to select the story
ending in (2). Additionally, we show not only that
by removing those sentences the network prediction changes, but also that the same are sufficient to
reproduce the inference. | ['Daniele Nardi', 'Roberto Capobianco', 'Biagio La Rosa'] | 2020-07-11 | null | null | null | null | ['cloze-test'] | ['natural-language-processing'] | [ 5.78942895e-01 5.20113349e-01 -1.24179069e-02 -1.24907844e-01
8.35814849e-02 -5.25831580e-01 6.25943184e-01 4.63258594e-01
-4.16112483e-01 9.64743376e-01 1.81173742e-01 -4.12840813e-01
-3.74878764e-01 -1.18551803e+00 -7.10028708e-01 -6.28104210e-01
-1.62765607e-01 7.05118001e-01 4.18456346e-01 -3.51304710e-01
7.67767787e-01 6.64200127e-01 -1.67748463e+00 5.80233693e-01
5.75662792e-01 5.79883993e-01 4.08996522e-01 9.60450411e-01
1.26669958e-01 1.35349739e+00 -4.64633167e-01 -3.23689312e-01
-9.50885713e-02 -1.01054215e+00 -1.30961370e+00 -5.99412918e-01
-1.63691938e-01 -3.48486394e-01 -4.64038730e-01 9.92141843e-01
8.66509676e-02 3.73265803e-01 8.32204461e-01 -1.12168562e+00
-3.56285006e-01 1.18280792e+00 2.99352586e-01 3.47612977e-01
6.48186207e-01 3.00838768e-01 7.99623907e-01 -3.11347932e-01
1.06432700e+00 1.10059845e+00 6.37293100e-01 8.74008775e-01
-1.41396511e+00 -2.03976631e-01 5.37065752e-02 7.19223499e-01
-1.16620409e+00 -5.50641418e-01 7.32404292e-01 -4.63321626e-01
1.27035487e+00 5.62589645e-01 1.08199835e+00 1.22237313e+00
5.03038406e-01 5.37735105e-01 9.71597433e-01 -5.44741213e-01
5.86312115e-01 -1.59273088e-01 3.55092138e-01 8.20460320e-01
2.82409228e-02 6.33773983e-01 -8.62100184e-01 -3.47228311e-02
7.27810860e-01 -6.91618025e-02 -2.89249718e-01 -6.99595511e-02
-1.21890318e+00 6.23142481e-01 1.57467559e-01 7.86571205e-01
-5.71772099e-01 2.62342572e-01 2.66702414e-01 4.59086329e-01
-2.19704881e-01 6.92915916e-01 -3.20621580e-01 -5.82281761e-02
-8.69003773e-01 5.04427433e-01 1.01254737e+00 3.92559648e-01
4.92302030e-01 -1.95015058e-01 -4.35442328e-01 4.94065136e-02
1.50249049e-01 -3.22881248e-03 7.48431802e-01 -1.22715819e+00
1.50323257e-01 4.28087026e-01 -1.32558635e-02 -1.19056940e+00
-6.82299912e-01 -2.12665424e-01 -7.48361647e-01 3.06414068e-01
7.51520276e-01 1.47558868e-01 -4.00431216e-01 2.32332230e+00
-2.50711769e-01 3.58088195e-01 2.67869473e-01 6.67024672e-01
3.06848407e-01 5.53641856e-01 4.27934676e-02 -5.47942877e-01
1.01881433e+00 -5.87884784e-01 -4.93124723e-01 -2.27115244e-01
8.70844603e-01 1.94113031e-01 8.16668570e-01 4.32205200e-01
-1.35945261e+00 -4.02136207e-01 -1.19114232e+00 1.12278342e-01
-4.13305044e-01 -4.97180969e-01 5.80201805e-01 2.61808038e-01
-1.23109567e+00 1.45356083e+00 -8.18362594e-01 -5.04596055e-01
9.68129113e-02 3.27742487e-01 -1.03175312e-01 1.50513157e-01
-1.51970589e+00 1.37824953e+00 7.99817622e-01 6.93556294e-02
-1.14102077e+00 -3.43281537e-01 -5.52086294e-01 4.59821641e-01
3.02768946e-01 -1.00198185e+00 9.90081131e-01 -9.38971102e-01
-1.52078116e+00 8.86510074e-01 -3.94068092e-01 -8.70390117e-01
4.10972297e-01 4.40459877e-01 -2.91348875e-01 1.79082856e-01
-1.73091158e-01 6.34698927e-01 3.96880507e-01 -9.22118425e-01
-2.01924428e-01 -5.59157789e-01 -7.31937587e-02 1.03093814e-02
2.70643085e-01 -2.73389190e-01 2.63176560e-01 -4.85118687e-01
2.49372676e-01 -8.19090843e-01 -2.07644999e-01 -1.94160685e-01
-6.82729363e-01 -6.71442747e-02 1.40226871e-01 -7.55567849e-01
1.13419986e+00 -1.83697498e+00 5.97530127e-01 3.42535377e-01
3.53552729e-01 -5.54709285e-02 -1.14896983e-01 5.00673532e-01
-3.29482108e-01 3.52185607e-01 -4.69048113e-01 -2.20667347e-01
4.59443294e-02 2.68763959e-01 -4.34198141e-01 1.30079240e-01
-5.42628579e-02 9.11449790e-01 -8.06761205e-01 -3.78219336e-01
-7.04216808e-02 -3.65287438e-02 -4.27818090e-01 1.78375989e-01
-5.36983430e-01 5.94145417e-01 -2.96602219e-01 -3.47899087e-02
1.43488407e-01 -5.39541095e-02 5.20424128e-01 3.11271220e-01
-6.80239424e-02 5.09851158e-01 -8.79121840e-01 1.56025541e+00
-9.74691659e-02 8.63548815e-01 -6.44722879e-01 -1.09654701e+00
7.83209085e-01 4.12439555e-01 -3.32892418e-01 -7.96528101e-01
2.91979998e-01 1.27091939e-02 3.34711939e-01 -6.39121175e-01
3.09277356e-01 -2.42128983e-01 -3.50967608e-02 9.55173910e-01
-1.10163495e-01 1.51552409e-01 3.59758884e-01 1.53717741e-01
1.44806540e+00 6.52908906e-02 5.40044367e-01 -4.74164523e-02
5.55164218e-01 -4.15666960e-03 4.64909285e-01 1.27891326e+00
-6.45714551e-02 2.70818084e-01 1.05368054e+00 -5.84509313e-01
-1.10945356e+00 -1.05034852e+00 4.87076104e-01 9.76114094e-01
1.52367121e-03 -5.00365496e-02 -9.76457059e-01 -2.16792494e-01
-4.53763723e-01 1.81355679e+00 -9.93113518e-01 -6.95340633e-01
-6.85941160e-01 -2.62558162e-01 6.67120636e-01 3.48796040e-01
3.14376444e-01 -1.67500544e+00 -1.14041770e+00 1.01586998e-01
-3.47072452e-01 -3.35832447e-01 5.69181226e-04 5.95886528e-01
-1.14760089e+00 -1.18058956e+00 -8.75610393e-03 -6.19118273e-01
5.32513559e-01 -3.45629036e-01 9.92800653e-01 4.44498122e-01
5.55893332e-02 1.16011398e-02 -4.80906032e-02 3.26361060e-02
-8.82782638e-01 6.61732303e-03 -9.72220749e-02 -3.55020732e-01
1.15626432e-01 -1.02453661e+00 -1.52427927e-01 -5.29433563e-02
-9.33634043e-01 4.60154295e-01 3.37710321e-01 7.20549345e-01
4.13119256e-01 1.67725638e-01 4.28282738e-01 -8.78247976e-01
7.61785626e-01 -5.24250805e-01 -4.85003024e-01 4.69331890e-01
-7.26647973e-01 5.71274698e-01 8.71947944e-01 -3.07235599e-01
-9.79606569e-01 -7.29878694e-02 5.68746664e-02 1.82875052e-01
-4.45580363e-01 5.12690306e-01 -4.23248559e-02 4.76131618e-01
7.26977468e-01 6.50440633e-01 1.66980624e-02 -1.63039416e-01
1.70974240e-01 -8.77463594e-02 6.15780175e-01 -3.92045230e-01
4.03064460e-01 1.01108691e-02 1.96738601e-01 -1.61623821e-01
-5.19902706e-01 4.67999667e-01 -8.77400398e-01 -2.64377505e-01
1.04264367e+00 -1.09782182e-01 -1.30064940e+00 3.33736330e-01
-1.67967141e+00 -5.44197679e-01 -3.95316452e-01 3.59682828e-01
-1.16898584e+00 -2.55475454e-02 -8.09578001e-01 -8.82692456e-01
9.17778313e-02 -8.99769485e-01 3.16597551e-01 1.39570341e-01
-7.76034355e-01 -1.04923987e+00 3.64833355e-01 -9.16678384e-02
1.18754007e-01 1.65101066e-01 1.71135581e+00 -1.07922232e+00
-5.67936540e-01 8.63858834e-02 2.58831412e-01 -3.64290982e-01
-5.69294214e-01 -1.72063500e-01 -9.00116026e-01 2.39346802e-01
3.12747449e-01 -6.36126334e-03 1.07351387e+00 1.09146766e-01
1.24366713e+00 -4.29142952e-01 -4.47463661e-01 1.63708389e-01
1.14154649e+00 5.33970892e-01 1.11787486e+00 4.47528660e-01
2.93607339e-02 7.11059451e-01 2.21075207e-01 2.56104738e-01
1.54424593e-01 6.00231886e-01 5.15645027e-01 6.78310573e-01
1.08644702e-01 -6.68324172e-01 4.94617879e-01 4.59077626e-01
-1.78079322e-01 -3.72596413e-01 -8.26055348e-01 3.47185165e-01
-2.09427929e+00 -1.54001570e+00 -2.03388035e-02 2.27750087e+00
9.05549586e-01 3.82980973e-01 -1.84682786e-01 4.02956843e-01
6.77409530e-01 -5.26924878e-02 -7.92965651e-01 -6.25898957e-01
-2.74411976e-01 2.46987671e-01 2.52077952e-02 8.22811067e-01
-3.17663610e-01 8.21716964e-01 6.85108566e+00 6.31198704e-01
-7.11657107e-01 -6.23119995e-02 6.18304908e-01 -1.78447608e-02
-4.05438900e-01 6.57141134e-02 -2.51363337e-01 2.80086964e-01
1.47283626e+00 -3.27491343e-01 9.58852887e-01 4.16216195e-01
2.42086291e-01 -5.79789579e-01 -1.80482388e+00 5.48251927e-01
6.55598752e-03 -1.73880005e+00 2.69676447e-01 -2.37836272e-01
1.39329135e-01 -6.57180309e-01 -5.54246195e-02 1.36440873e-01
2.14396045e-01 -1.37497771e+00 1.06696880e+00 1.25847709e+00
8.35928172e-02 -6.85518622e-01 5.91785252e-01 8.76811385e-01
-5.82738519e-01 -1.10547282e-01 -2.75993913e-01 -6.43628299e-01
-3.90514247e-02 2.67514914e-01 -9.35286105e-01 3.55729610e-02
8.25194493e-02 3.69309664e-01 -4.92869467e-01 7.85269737e-01
-6.81173980e-01 5.15156806e-01 -9.05954931e-03 -4.16949093e-01
-2.16672376e-01 3.39845940e-02 7.93224633e-01 7.50234902e-01
4.69941795e-01 3.46080512e-01 -4.90124285e-01 1.46982539e+00
1.23672366e-01 -4.59134728e-01 -8.23372066e-01 -1.47396564e-01
6.07617915e-01 5.16372621e-01 -9.96682465e-01 -4.46267664e-01
4.92681503e-01 1.29340672e+00 4.60044622e-01 3.20717931e-01
-7.47718334e-01 -2.05760241e-01 3.03787678e-01 -1.10698916e-01
2.03952864e-01 -1.44353405e-01 -5.61710298e-01 -7.49374151e-01
-3.44019800e-01 -5.61318636e-01 2.20122933e-01 -1.29671741e+00
-7.24068701e-01 6.65570080e-01 4.62906696e-02 -5.81207037e-01
-7.57604182e-01 -4.66104537e-01 -8.16659033e-01 9.02373254e-01
-8.23325217e-01 -3.87838423e-01 1.53325662e-01 5.31186759e-01
2.33797535e-01 -6.83733001e-02 1.26652360e+00 -2.89322406e-01
-6.18236959e-01 3.16729605e-01 -2.11243361e-01 -3.75676826e-02
-1.13861397e-01 -1.03098989e+00 3.09761524e-01 7.27433264e-01
1.58959314e-01 8.09480131e-01 1.12892365e+00 -7.08262622e-01
-1.03275251e+00 -7.10961759e-01 1.41821873e+00 -3.90119910e-01
4.64309573e-01 -1.76969662e-01 -1.02393150e+00 9.91752982e-01
1.16190299e-01 -6.64050579e-01 6.39457822e-01 -9.40842628e-02
-3.00396085e-01 2.97339439e-01 -1.33633912e+00 1.02301109e+00
1.31715512e+00 -6.09010637e-01 -1.21642327e+00 1.90203920e-01
8.26713502e-01 -1.33981124e-01 -4.00039732e-01 -2.87374824e-01
4.78709757e-01 -1.50176299e+00 6.48352027e-01 -1.05506742e+00
7.96876669e-01 -2.05116764e-01 -8.19661245e-02 -1.34040236e+00
-5.09499669e-01 -4.38226253e-01 -2.89836675e-01 7.33311892e-01
6.32914901e-01 -7.55185127e-01 7.53504038e-01 7.24969506e-01
1.57298133e-01 -5.38748205e-01 -9.64801192e-01 -5.42585969e-01
-1.17374383e-01 -6.05627596e-01 6.83634579e-01 6.99332118e-01
3.67925644e-01 4.42307234e-01 -2.38747805e-01 2.33324561e-02
4.00090188e-01 2.12830693e-01 2.61877640e-03 -1.37932169e+00
-5.17687023e-01 -7.56506920e-01 -2.32776657e-01 -7.74344027e-01
4.75353748e-01 -1.06793559e+00 9.44956541e-02 -1.28286481e+00
4.91066307e-01 -2.07097158e-02 -9.47274417e-02 4.19817924e-01
3.09831262e-01 -2.11769670e-01 3.81942600e-01 1.97572067e-01
-5.37267923e-01 3.40870768e-01 9.44951117e-01 -6.51805801e-03
-2.06827447e-01 -1.59354746e-01 -5.59302688e-01 8.77266347e-01
8.16352725e-01 -8.58663976e-01 -2.47384131e-01 -1.41949624e-01
7.62875021e-01 8.06350410e-01 7.49202728e-01 -9.98623073e-01
5.37364781e-01 -3.10130179e-01 6.57019436e-01 -4.78358477e-01
2.81474799e-01 -6.26369536e-01 6.33757770e-01 1.11761749e+00
-1.13829136e+00 9.88046378e-02 1.19900756e-01 6.12709582e-01
1.62147671e-01 -9.87141967e-01 4.33179379e-01 -3.66112322e-01
-7.49240875e-01 -1.93451837e-01 -9.35169339e-01 -2.22481474e-01
9.78211880e-01 -2.42458031e-01 -3.37611705e-01 -5.49169421e-01
-1.42180419e+00 -5.38315400e-02 4.02851731e-01 -6.40305206e-02
7.12672710e-01 -1.11700916e+00 -3.18749487e-01 7.34457374e-02
-2.74780333e-01 -4.96861398e-01 1.97947145e-01 8.06996942e-01
-5.23724198e-01 4.99413580e-01 -5.72415173e-01 -1.82838216e-02
-1.05110884e+00 7.92597711e-01 5.96701384e-01 -4.80400592e-01
-3.32046777e-01 6.66863859e-01 -1.97538286e-01 -1.94355235e-01
7.79485330e-02 -4.53150719e-01 -4.58333373e-01 8.51177126e-02
7.93814540e-01 4.28454965e-01 -1.12922743e-01 -2.00339764e-01
-2.10038885e-01 1.16427772e-01 1.43333018e-01 -4.22025323e-01
1.46188319e+00 -7.85248801e-02 -5.84239721e-01 9.56658542e-01
6.90386593e-01 -3.75728041e-01 -8.53865445e-01 6.97102100e-02
2.45012090e-01 -8.41340646e-02 -2.69692421e-01 -1.00404096e+00
-6.27411664e-01 8.47347021e-01 3.27198893e-01 5.77080846e-01
9.76300895e-01 -1.16682708e-01 5.19062459e-01 8.53128672e-01
5.24716556e-01 -8.57604265e-01 -5.31868860e-02 7.71437645e-01
9.23276961e-01 -4.65800911e-01 -3.82155985e-01 -5.40234558e-02
-3.98441732e-01 1.54528320e+00 2.46778965e-01 -9.33684036e-02
1.75567269e-01 -3.27709243e-02 -6.48739040e-01 -2.19180912e-01
-1.15580094e+00 8.35756436e-02 -1.73325911e-01 6.95216656e-01
4.03049625e-02 -2.01108288e-02 -2.19087541e-01 7.83995867e-01
-3.68234783e-01 1.98502332e-01 7.94753730e-01 6.52265847e-01
-6.89145505e-01 -8.61098230e-01 -3.99116665e-01 4.03730720e-01
1.17482357e-01 -2.48986423e-01 -7.66222000e-01 5.80180347e-01
1.64201632e-02 7.98408091e-01 1.65626556e-01 -5.94212830e-01
1.54704675e-01 6.25039935e-01 8.11710417e-01 -3.74972075e-01
-6.13740981e-01 -7.59867430e-01 1.77732974e-01 -8.35538089e-01
-2.00341284e-01 -7.40350962e-01 -1.72133565e+00 -7.71990001e-01
2.17558101e-01 1.87517852e-01 1.82278052e-01 1.33281744e+00
2.51604110e-01 8.59365404e-01 1.54989526e-01 -8.22996557e-01
-3.99805605e-01 -7.10000396e-01 -6.61712587e-01 2.03815088e-01
1.69483855e-01 -5.63112617e-01 -3.58648390e-01 1.11388080e-01] | [8.251848220825195, 3.23345947265625] |
19b0b821-79d2-499c-bae2-b17570782046 | optimizing-industrial-hvac-systems-with | 2209.08112 | null | https://arxiv.org/abs/2209.08112v1 | https://arxiv.org/pdf/2209.08112v1.pdf | Optimizing Industrial HVAC Systems with Hierarchical Reinforcement Learning | Reinforcement learning (RL) techniques have been developed to optimize industrial cooling systems, offering substantial energy savings compared to traditional heuristic policies. A major challenge in industrial control involves learning behaviors that are feasible in the real world due to machinery constraints. For example, certain actions can only be executed every few hours while other actions can be taken more frequently. Without extensive reward engineering and experimentation, an RL agent may not learn realistic operation of machinery. To address this, we use hierarchical reinforcement learning with multiple agents that control subsets of actions according to their operation time scales. Our hierarchical approach achieves energy savings over existing baselines while maintaining constraints such as operating chillers within safe bounds in a simulated HVAC control environment. | ['Jerry Luo', 'Cosmin Paduraru', 'Yuri Chervonyi', 'Octavian Voicu', 'Praneet Dutta', 'William Wong'] | 2022-09-16 | null | null | null | null | ['hierarchical-reinforcement-learning'] | ['methodology'] | [ 2.42154986e-01 2.94457793e-01 -4.85549539e-01 -7.61359259e-02
-3.86308163e-01 -7.49065638e-01 3.63146245e-01 3.48848939e-01
-2.64405906e-01 1.24781525e+00 -2.99139142e-01 -4.50168461e-01
-4.85220194e-01 -8.91312540e-01 -5.34790397e-01 -7.72781491e-01
-3.10741633e-01 5.13474464e-01 -1.44901901e-01 -3.11203718e-01
2.03181565e-01 5.46134353e-01 -1.39982140e+00 -7.14959651e-02
7.07573533e-01 7.78868020e-01 3.29248250e-01 8.81659269e-01
4.76861894e-01 9.62321162e-01 -8.37854743e-01 6.42468214e-01
5.02207935e-01 -5.56033492e-01 -7.21068382e-01 3.30472589e-01
-4.31678444e-01 -5.99602044e-01 -1.87559053e-01 8.09704363e-01
1.94320515e-01 7.30107188e-01 3.86533797e-01 -1.58664739e+00
-3.92205976e-02 7.09077775e-01 -4.30582464e-01 6.73645586e-02
2.50961632e-01 6.65169060e-01 8.77645314e-01 4.76304889e-01
1.70495108e-01 1.00258017e+00 -5.41361002e-03 5.11226714e-01
-1.23298240e+00 -5.21388710e-01 3.96638811e-01 2.54075944e-01
-9.42271590e-01 5.36609069e-02 4.96427834e-01 2.22506940e-01
1.55161285e+00 2.96286136e-01 1.10353637e+00 6.50654018e-01
7.79086471e-01 6.67721987e-01 1.19950557e+00 -4.82038885e-01
7.28465319e-01 -2.68366486e-01 -5.23249924e-01 4.15561855e-01
3.35875332e-01 6.34241641e-01 3.04842126e-02 -7.13836625e-02
6.17556095e-01 2.15299744e-02 4.97926995e-02 -4.16447729e-01
-1.00132334e+00 9.46439624e-01 1.29118517e-01 -9.46080778e-03
-6.92075133e-01 5.51270962e-01 6.04005098e-01 4.91885245e-01
-1.86289713e-01 1.34089792e+00 -8.23388040e-01 -3.62857670e-01
-5.09306848e-01 6.20017946e-01 8.41784120e-01 9.43750918e-01
5.56246698e-01 5.49305260e-01 -2.64825314e-01 3.95227373e-01
-1.04472779e-01 3.27663571e-01 2.07843676e-01 -1.61129713e+00
2.25523844e-01 2.30316132e-01 6.49362564e-01 -1.26623303e-01
-4.35202032e-01 5.30663133e-02 -3.91107261e-01 6.64516032e-01
8.13436806e-02 -7.10075676e-01 -8.14814508e-01 1.23147035e+00
2.40535498e-01 -4.05737348e-02 1.74034059e-01 6.79989755e-01
-5.07960856e-01 8.48623037e-01 1.98372185e-01 -7.41106033e-01
9.78652179e-01 -7.91859925e-01 -9.22045350e-01 -9.11082774e-02
4.04873729e-01 -4.41277951e-01 6.84461653e-01 7.08029330e-01
-1.21310604e+00 -2.64308244e-01 -1.35960090e+00 7.77820647e-01
-3.46782923e-01 -2.44320974e-01 8.77438664e-01 7.93440521e-01
-7.09145844e-01 1.03624988e+00 -1.17683935e+00 -1.39007801e-02
1.33327037e-01 6.91243827e-01 2.92247355e-01 3.47298048e-02
-1.05159891e+00 1.24466288e+00 6.30320430e-01 1.11807883e-01
-1.44277310e+00 -5.06555974e-01 -9.11946595e-01 1.39462709e-01
1.02711558e+00 -2.50729710e-01 1.99888337e+00 -5.19690096e-01
-2.01837754e+00 -3.56274933e-01 5.10590792e-01 -6.69575155e-01
3.86221588e-01 -3.42853844e-01 -3.35157424e-01 1.56615332e-01
-3.85491163e-01 3.95154208e-01 7.88440704e-01 -1.34426475e+00
-1.02536166e+00 1.20835565e-01 6.12028837e-01 6.35057926e-01
-1.04526937e-01 -3.54121655e-01 3.18541467e-01 -7.19092637e-02
-6.07289910e-01 -1.30795312e+00 -9.27965879e-01 -8.25865388e-01
-2.64725268e-01 -3.87274981e-01 9.57409382e-01 -2.18756363e-01
8.31828475e-01 -1.64964139e+00 1.86686605e-01 3.11011016e-01
-4.02195185e-01 -6.96934536e-02 2.20208708e-02 5.20971715e-01
8.97921324e-02 -5.42653948e-02 1.44229159e-01 3.45908850e-01
3.23683947e-01 6.84240460e-01 -3.99128906e-02 3.50680262e-01
6.02320163e-03 5.20093799e-01 -1.21771479e+00 -2.61068434e-01
7.32206404e-01 -2.44163617e-01 -4.05170411e-01 4.42845762e-01
-7.14000106e-01 3.25835913e-01 -7.69236565e-01 7.47678638e-01
1.14910752e-01 1.20202191e-01 7.84841478e-01 2.08499834e-01
-2.64611036e-01 7.80314058e-02 -1.11901677e+00 1.30362177e+00
-8.72156441e-01 1.74493030e-01 3.54402125e-01 -1.21379793e+00
3.95766467e-01 4.68606055e-01 1.12767768e+00 -7.75910079e-01
3.38048451e-02 -2.54406542e-01 2.09978461e-01 -3.75536323e-01
6.34751499e-01 -1.22373566e-01 -1.60954192e-01 5.77169776e-01
-3.23644876e-01 -1.09814167e+00 6.15960658e-01 -2.17400789e-01
1.53336346e+00 3.61792922e-01 3.51486892e-01 -3.33466053e-01
-1.04611190e-02 4.67188418e-01 8.07257354e-01 7.14024723e-01
-3.32518131e-01 -2.73908347e-01 5.81708848e-01 -3.50508600e-01
-1.00188863e+00 -8.86145890e-01 2.71402627e-01 1.11316717e+00
2.57583499e-01 -1.99775398e-01 -5.86915672e-01 -7.97541261e-01
2.09553093e-01 1.25119007e+00 -2.51851439e-01 -3.61600488e-01
-7.67928898e-01 -8.25917840e-01 -9.76632312e-02 5.30986249e-01
3.09508383e-01 -1.13139307e+00 -1.52632976e+00 4.90320086e-01
4.08549160e-01 -1.00206172e+00 -3.96400392e-01 9.41738069e-01
-7.18876064e-01 -1.28818750e+00 -7.88002983e-02 -2.58551866e-01
7.28875697e-01 9.64623690e-02 1.22026145e+00 -6.32266328e-02
-7.20953643e-01 5.03223479e-01 -2.55639791e-01 -7.33302534e-01
-4.78257149e-01 1.08283972e-02 1.58548310e-01 -9.49979722e-01
-4.88229357e-02 -1.07360877e-01 -7.82161057e-01 2.20555007e-01
-7.25265384e-01 -2.69435376e-01 7.61449158e-01 9.63689268e-01
5.56914747e-01 1.15084863e+00 6.72487855e-01 -7.87595630e-01
9.10528183e-01 -2.15412974e-01 -1.06030679e+00 4.62690324e-01
-1.08118129e+00 1.52970627e-01 1.03108621e+00 -4.18500870e-01
-1.20659530e+00 3.66409034e-01 4.90713358e-01 -1.64014995e-01
-2.16981336e-01 5.36065213e-02 -2.04511303e-02 1.41603157e-01
1.64483950e-01 -9.74293128e-02 -9.94270220e-02 2.46323735e-01
9.88249779e-02 1.75685555e-01 -1.50745560e-03 -1.03808594e+00
7.95814216e-01 -3.18080157e-01 2.99258918e-01 -5.64947665e-01
-4.94820893e-01 -4.09138724e-02 -2.74909765e-01 -3.72804075e-01
6.61669254e-01 -6.75510943e-01 -1.33916235e+00 -1.62925124e-01
-4.30542767e-01 -1.00110686e+00 -6.64836287e-01 5.89520276e-01
-9.75814700e-01 1.46221509e-02 -6.99312389e-01 -1.14779639e+00
-4.90811244e-02 -1.17279577e+00 5.31700075e-01 5.69413364e-01
-3.98513049e-01 -6.43443048e-01 -4.50828560e-02 9.21664089e-02
3.11411530e-01 6.05618834e-01 1.10206962e+00 -2.21611653e-02
-6.58619225e-01 2.78177410e-02 4.92860287e-01 3.59895706e-01
4.85087901e-01 6.53815195e-02 -4.13131654e-01 -8.15492570e-01
-2.46210784e-01 -8.47536325e-01 1.69186011e-01 5.46427608e-01
1.46093130e+00 -3.55325967e-01 -4.15744454e-01 -1.09313801e-01
1.36779070e+00 1.09603477e+00 2.59775221e-01 5.61486304e-01
1.55376196e-01 3.85947019e-01 1.42553604e+00 9.41682398e-01
3.46290134e-02 4.70878333e-01 7.52762735e-01 -2.12115869e-02
6.52906775e-01 -3.99669074e-02 3.48426342e-01 2.05818281e-01
-2.98358560e-01 -2.62580812e-01 -4.78328615e-01 3.07988435e-01
-1.78839719e+00 -9.05704558e-01 5.18441796e-01 2.28206444e+00
8.11310112e-01 5.91597319e-01 7.98153281e-02 1.16930023e-01
3.44045699e-01 4.92732190e-02 -1.20374465e+00 -9.86655593e-01
5.63777089e-01 3.17636758e-01 1.12734807e+00 3.15692514e-01
-9.25907016e-01 4.49031830e-01 7.48258543e+00 3.46669853e-01
-6.44916415e-01 -3.57112736e-01 5.25294304e-01 -6.82544529e-01
5.51429801e-02 1.77292898e-02 -4.41440046e-01 2.66461581e-01
1.34026873e+00 -4.40142751e-01 1.14693749e+00 9.41451371e-01
7.62319267e-01 -5.78177810e-01 -1.29019701e+00 4.90378618e-01
-6.06519938e-01 -9.65453863e-01 -5.64324379e-01 6.51240274e-02
1.04209375e+00 -3.53266031e-01 -1.81600258e-01 8.42427492e-01
1.26281524e+00 -1.17603767e+00 2.85698146e-01 1.02717556e-01
3.16662639e-01 -1.64011431e+00 5.92440605e-01 2.51110226e-01
-1.17338490e+00 -6.79765105e-01 -3.16314429e-01 -2.03635067e-01
9.63890702e-02 3.01449060e-01 -1.11263573e+00 5.92200339e-01
6.12352729e-01 1.31872743e-01 -1.12987570e-02 6.81981146e-01
-3.11830282e-01 5.07881522e-01 -3.03317666e-01 -4.45170343e-01
5.07774532e-01 -3.15444410e-01 -8.87604710e-03 6.07051253e-01
2.70947486e-01 1.85682982e-01 7.78611839e-01 6.37384057e-01
2.28444099e-01 -4.93909240e-01 -6.68076515e-01 -2.77512640e-01
6.84644401e-01 1.32643092e+00 -9.61755216e-01 -1.20058939e-01
-1.10421538e-01 8.58405054e-01 -9.98222753e-02 3.50790679e-01
-1.02059650e+00 -5.85440099e-01 8.83801699e-01 -2.72258312e-01
1.07864864e-01 -5.79398096e-01 3.32603268e-02 -2.86142588e-01
-4.54383045e-01 -9.71670568e-01 3.18681479e-01 -4.84522551e-01
-7.49501586e-01 -4.18301374e-02 2.96659648e-01 -9.83546913e-01
-7.26899922e-01 -6.43273115e-01 -5.53534687e-01 3.83304507e-01
-1.53754187e+00 -4.07770395e-01 2.00509608e-01 3.35662574e-01
9.76500571e-01 -1.40467361e-01 8.60036075e-01 -3.62096667e-01
-7.16549456e-01 -2.35044714e-02 3.01926076e-01 -6.51274979e-01
3.43777329e-01 -1.88145483e+00 -1.33139538e-02 4.39100266e-01
-4.92464542e-01 2.55638510e-01 1.16229725e+00 -5.15229583e-01
-2.04097533e+00 -9.45047379e-01 -2.40212977e-01 -4.25253948e-03
5.19414544e-01 -5.47070871e-04 -3.75646055e-01 5.09141326e-01
8.49289000e-01 -3.25130433e-01 3.02732915e-01 7.60474876e-02
6.44550860e-01 -6.52163103e-02 -1.23093724e+00 6.42722785e-01
4.63495344e-01 -1.54575557e-01 -2.79578388e-01 7.44616151e-01
5.92396915e-01 -6.63140774e-01 -1.16927147e+00 4.67337430e-01
2.07750365e-01 -4.39895630e-01 5.96711934e-01 -6.00320399e-01
1.38136834e-01 -2.91178435e-01 2.35940054e-01 -2.03666449e+00
-2.36422896e-01 -1.00155544e+00 -3.01578015e-01 7.44429529e-01
2.51817912e-01 -2.89940596e-01 8.92677963e-01 9.82108057e-01
-2.39345089e-01 -8.58438075e-01 -8.42848063e-01 -1.15420902e+00
-5.75976563e-04 -3.39695364e-02 6.75532341e-01 8.20710123e-01
4.40760404e-01 -4.45470760e-05 -3.37967396e-01 3.43142182e-01
6.19661987e-01 2.70988077e-01 2.70525485e-01 -5.27762115e-01
-5.56699395e-01 -2.16593564e-01 2.44162545e-01 -3.98949951e-01
3.58629525e-01 -1.07530296e-01 7.02424824e-01 -1.61246240e+00
-2.09217668e-01 -3.42526168e-01 -7.42573380e-01 7.46519148e-01
1.22047670e-01 -4.59374070e-01 2.31639802e-01 -4.58929181e-01
-7.12668538e-01 6.82408452e-01 1.44808352e+00 -4.13408905e-01
-5.97659707e-01 1.02620363e-01 -3.03968877e-01 3.53576511e-01
1.29666853e+00 -2.77273238e-01 -8.68176937e-01 5.65270893e-02
-6.80160290e-03 6.26367033e-01 -2.64936149e-01 -8.29292655e-01
-4.29099947e-02 -1.08581603e+00 4.12116826e-01 -3.59840274e-01
2.37509489e-01 -1.10739744e+00 1.52842239e-01 1.02782595e+00
-5.32209992e-01 5.23469329e-01 3.28466326e-01 7.28903472e-01
7.16273934e-02 -3.31000626e-01 7.50522435e-01 -4.74622130e-01
-7.79376268e-01 -3.74854058e-02 -1.05109572e+00 -3.72214526e-01
1.73665690e+00 1.28058583e-01 -1.46525398e-01 -2.18029082e-01
-4.48484063e-01 1.01099718e+00 2.88616955e-01 5.14946759e-01
3.34724903e-01 -9.76883888e-01 -1.99851811e-01 -2.78703958e-01
-1.74588352e-01 2.23545786e-02 4.06579161e-03 2.80573756e-01
-4.97412711e-01 5.01980543e-01 -5.80286622e-01 -1.41471639e-01
-1.05488241e+00 8.07270169e-01 4.79690015e-01 -6.48272932e-01
-6.02567554e-01 1.40589923e-01 -2.74031013e-01 -9.81133506e-02
5.94771244e-02 -6.13179445e-01 1.32741019e-01 -2.26184398e-01
1.96934029e-01 7.10767508e-01 4.11627293e-02 4.22280967e-01
-1.21157274e-01 -1.82025786e-02 -1.14503197e-01 -6.33892566e-02
1.26815498e+00 1.01109125e-01 3.67463589e-01 2.18489677e-01
4.04074997e-01 -5.07009387e-01 -1.79833984e+00 5.55231988e-01
1.97384685e-01 -4.05409575e-01 3.17353427e-01 -1.02440619e+00
-9.50630665e-01 1.19852148e-01 6.74303114e-01 5.90822935e-01
1.27807665e+00 -6.42729402e-01 6.16143346e-01 8.02633166e-01
7.19986796e-01 -2.16088176e+00 1.65145084e-01 2.67833412e-01
4.50511754e-01 -9.97135043e-01 4.85948712e-01 -3.28294672e-02
-8.93404961e-01 1.15156519e+00 9.92020309e-01 -2.12663800e-01
1.65397778e-01 6.21715844e-01 -2.10912019e-01 1.19759459e-02
-1.19154954e+00 -5.31734154e-02 -5.96266866e-01 6.34183943e-01
2.01458737e-01 6.06735766e-01 -1.59933507e-01 -3.33598256e-01
2.12881237e-01 -1.11590862e-01 7.12271929e-01 1.56804311e+00
-4.53592896e-01 -1.23920798e+00 -4.74045068e-01 5.69334209e-01
-3.71109933e-01 5.87823868e-01 7.06750453e-02 1.05503082e+00
-3.82225886e-02 1.29751348e+00 -7.75218010e-02 7.48379761e-03
5.16932130e-01 -1.17400132e-01 7.38987625e-01 -8.00196409e-01
-7.16498137e-01 1.77642882e-01 3.80370170e-01 -7.27360368e-01
-2.28006780e-01 -7.55578279e-01 -1.82044196e+00 -7.79304802e-02
-2.71632701e-01 4.13386106e-01 6.25069380e-01 9.45571899e-01
-8.30462277e-02 1.44791043e+00 1.12181532e+00 -8.56568336e-01
-1.01546800e+00 -5.82375586e-01 -9.40658689e-01 -3.12974118e-02
2.68922031e-01 -7.58036017e-01 -8.73083845e-02 -2.09326491e-01] | [4.688526630401611, 2.161872386932373] |
e85ee8a0-98fb-4108-a593-bde5bf573286 | retinal-oct-disease-classification-with | 1904.00790 | null | http://arxiv.org/abs/1904.00790v1 | http://arxiv.org/pdf/1904.00790v1.pdf | Retinal OCT disease classification with variational autoencoder regularization | According to the World Health Organization, 285 million people worldwide live
with visual impairment. The most commonly used imaging technique for diagnosis
in ophthalmology is optical coherence tomography (OCT). However, analysis of
retinal OCT requires trained ophthalmologists and time, making a comprehensive
early diagnosis unlikely. A recent study established a diagnostic tool based on
convolutional neural networks (CNN), which was trained on a large database of
retinal OCT images. The performance of the tool in classifying retinal
conditions was on par to that of trained medical experts. However, the training
of these networks is based on an enormous amount of labeled data, which is
expensive and difficult to obtain. Therefore, this paper describes a method
based on variational autoencoder regularization that improves classification
performance when using a limited amount of labeled data. This work uses a
two-path CNN model combining a classification network with an autoencoder (AE)
for regularization. The key idea behind this is to prevent overfitting when
using a limited training dataset size with small number of patients. Results
show superior classification performance compared to a pre-trained and fully
fine-tuned baseline ResNet-34. Clustering of the latent space in relation to
the disease class is distinct. Neural networks for disease classification on
OCTs can benefit from regularization using variational autoencoders when
trained with limited amount of patient data. Especially in the medical imaging
domain, data annotated by experts is expensive to obtain. | ['Lüder A. Kahrs', 'Max-Heinrich Laves', 'Tobias Ortmaier', 'Sontje Ihler'] | 2019-03-23 | null | null | null | null | ['retinal-oct-disease-classification'] | ['computer-vision'] | [-2.90723771e-01 4.39408869e-02 1.04155652e-01 -1.65405840e-01
-2.41911218e-01 -1.07563384e-01 -6.81173801e-02 -4.40771170e-02
-6.96229458e-01 9.05884266e-01 2.02035442e-01 -2.90795743e-01
-7.16826618e-02 -5.86865425e-01 -4.42360103e-01 -6.95098579e-01
3.77688408e-01 4.99958515e-01 1.72330618e-01 2.02573866e-01
6.20103218e-02 4.77691203e-01 -2.02827048e+00 3.71539682e-01
1.30887783e+00 9.94635105e-01 4.60002869e-01 5.55473328e-01
5.47074247e-03 6.71253622e-01 -3.27917933e-01 8.94384161e-02
3.89631867e-01 -4.65504378e-01 -6.91079438e-01 3.65327060e-01
4.56951648e-01 -4.18796152e-01 7.71545023e-02 9.35315549e-01
7.37163305e-01 -3.43318880e-02 7.71361768e-01 -4.08503532e-01
-6.59564555e-01 -1.59999311e-01 -1.62441805e-01 2.90589780e-01
-2.74661034e-01 3.39490533e-01 3.46150070e-01 -5.52335203e-01
5.65297723e-01 6.76961839e-01 6.97227418e-01 7.28987217e-01
-1.25133479e+00 -3.68896216e-01 -3.75098765e-01 1.80503547e-01
-1.16871309e+00 -1.03521302e-01 2.91646242e-01 -9.90053654e-01
1.09089661e+00 -4.64233905e-02 1.25990415e+00 7.97273457e-01
1.34890497e-01 -3.89884524e-02 1.48650575e+00 -4.10603613e-01
1.27523080e-01 5.17204881e-01 1.46571681e-01 6.79892957e-01
2.41905808e-01 3.24316919e-01 2.13771760e-01 4.27373089e-02
9.69823122e-01 -4.04190384e-02 -3.80366057e-01 -6.97816610e-02
-5.06857216e-01 1.04705679e+00 5.20964146e-01 3.21744919e-01
-6.07198298e-01 -2.84661174e-01 3.90583903e-01 1.76842034e-01
7.58501470e-01 5.38268268e-01 -3.66988599e-01 6.97896034e-02
-1.18506503e+00 -4.17346388e-01 4.20318872e-01 4.62210290e-02
5.89866638e-01 1.98928371e-01 -1.17048107e-01 1.07211733e+00
4.74404454e-01 2.05189303e-01 8.01776946e-01 -8.47835064e-01
-3.16163093e-01 1.04702520e+00 -1.29357696e-01 -4.89562720e-01
-3.80313009e-01 -7.97481358e-01 -1.02725053e+00 9.24556434e-01
4.43897873e-01 -4.49393451e-01 -1.30829048e+00 1.10256815e+00
8.81604850e-02 4.10238177e-01 5.72178662e-02 1.17075229e+00
8.41131926e-01 3.23963702e-01 -4.69637550e-02 -3.57646465e-01
1.28013504e+00 -7.89037943e-01 -6.30186141e-01 6.60760850e-02
5.54517090e-01 -8.16738844e-01 8.72138679e-01 5.60128450e-01
-9.36279953e-01 -6.85031414e-01 -7.44251072e-01 -3.02351862e-01
-3.54352355e-01 7.67700732e-01 6.32533550e-01 5.71080625e-01
-1.34950602e+00 5.51735520e-01 -9.65992570e-01 -6.44127786e-01
7.14104176e-01 5.87376595e-01 -4.07542616e-01 1.27484977e-01
-6.82563484e-01 1.08784366e+00 3.22630972e-01 2.11122662e-01
-5.19239426e-01 -5.91868460e-01 -5.39900661e-01 -1.64211020e-01
-1.49978086e-01 -1.14091110e+00 6.32518530e-01 -1.21525729e+00
-1.63362730e+00 1.06958568e+00 -3.18827808e-01 -7.77579188e-01
5.07297814e-01 -2.82178223e-01 -2.84583211e-01 3.61723274e-01
-1.52996466e-01 7.12691963e-01 1.09849906e+00 -8.51898313e-01
-5.08060575e-01 -3.00521225e-01 -2.22348630e-01 -2.75818914e-01
-3.51696461e-01 -1.54630719e-02 -2.22049817e-01 -2.06127480e-01
-9.98337120e-02 -1.00211132e+00 -2.81946898e-01 -6.61874115e-02
-3.28549385e-01 -3.47948462e-01 5.95612347e-01 -8.10115159e-01
1.10719204e+00 -1.93880391e+00 -1.10111646e-01 3.12340725e-02
6.59565151e-01 8.79063666e-01 1.68531880e-01 -4.61757220e-02
-5.46938002e-01 1.91907644e-01 2.39802562e-02 -1.04005836e-01
-7.39486217e-01 1.48347706e-01 1.31165877e-01 4.54811811e-01
4.42938924e-01 6.25360310e-01 -4.81883645e-01 -3.62139344e-01
5.46380639e-01 7.66490877e-01 -9.72489655e-01 2.77002305e-01
-1.39191955e-01 9.06017363e-01 -9.45667773e-02 4.10810918e-01
5.60673416e-01 -8.16154480e-01 -1.41383037e-01 -4.36983198e-01
-2.95303941e-01 -1.65286839e-01 -8.49901676e-01 1.26852691e+00
-3.75805348e-01 1.04554439e+00 -3.69921595e-01 -9.39860582e-01
5.66877782e-01 4.54806447e-01 4.61010307e-01 -1.08956493e-01
4.01144564e-01 3.01975250e-01 3.14474136e-01 -1.11483586e+00
-2.24342763e-01 -2.38508478e-01 1.12332487e+00 -1.89081486e-02
6.93632066e-02 2.80262232e-01 1.40226468e-01 -2.76704937e-01
8.19207847e-01 7.11451396e-02 2.75332421e-01 1.24464624e-01
5.27486026e-01 1.09893680e-01 6.79875612e-01 2.88596004e-01
-6.02084137e-02 7.39223003e-01 5.16746104e-01 -7.86275864e-01
-1.16051006e+00 -5.34247160e-01 -7.47257650e-01 2.95829158e-02
-4.60604310e-01 -2.81494737e-01 -7.89234400e-01 -2.85424173e-01
-2.02548757e-01 1.67192161e-01 -7.88473189e-01 8.22537616e-02
-5.25770560e-02 -9.86576855e-01 1.95284799e-01 3.57671767e-01
6.61198735e-01 -7.91439831e-01 -6.95599318e-01 1.05841279e-01
1.54708475e-01 -9.35973465e-01 2.75462627e-01 -3.49907368e-01
-1.21694875e+00 -1.34795344e+00 -9.88032639e-01 -8.05573702e-01
9.76927817e-01 -2.31313661e-01 7.78880358e-01 8.01726133e-02
-7.73263097e-01 1.98705807e-01 -1.16943166e-01 -5.99403560e-01
-3.25422704e-01 -2.69378871e-01 8.78902227e-02 3.60386580e-01
6.30033195e-01 -6.12301409e-01 -9.81592119e-01 -6.28356412e-02
-5.56611955e-01 -5.52147515e-02 1.04590321e+00 1.11220980e+00
8.83844137e-01 1.47018760e-01 4.10059653e-02 -8.48204792e-01
5.40397584e-01 -2.69656956e-01 -7.70325065e-01 -2.24791397e-03
-8.39801013e-01 -1.39593467e-01 2.51289219e-01 -5.01464427e-01
-8.72564137e-01 -1.14254408e-01 -8.42178166e-02 -8.98057282e-01
-3.56171250e-01 6.86264396e-01 6.02777541e-01 -3.21768165e-01
9.35324252e-01 -1.63476691e-01 5.54415107e-01 -5.51648736e-01
-2.54065365e-01 8.32165062e-01 4.04544622e-02 1.46709457e-01
3.02455097e-01 5.07405102e-01 1.77420974e-01 -1.07423484e+00
-8.81847978e-01 -5.38273394e-01 -4.81250703e-01 -2.43558481e-01
1.29367101e+00 -9.73671257e-01 -8.83020461e-01 5.10809958e-01
-1.00670493e+00 -1.97022080e-01 -3.86660874e-01 1.24675405e+00
-1.62648171e-01 1.64169937e-01 -4.66524720e-01 -6.21962607e-01
-2.03297243e-01 -1.31435716e+00 4.72775280e-01 3.65355313e-01
1.76504478e-01 -1.08311129e+00 2.79826820e-01 6.17372394e-01
3.34779799e-01 2.11374611e-01 9.56541002e-01 -3.19578499e-01
-6.91338480e-01 -2.67108083e-01 -5.41122913e-01 1.29443657e+00
7.56683648e-02 4.13031369e-01 -1.11351275e+00 -1.81068361e-01
1.94974929e-01 -3.75100017e-01 1.17194164e+00 1.24278998e+00
1.29138637e+00 2.88378391e-02 -1.29626185e-01 8.68745506e-01
1.64371240e+00 1.47066694e-02 8.63314509e-01 7.43005872e-02
6.11746073e-01 6.59538448e-01 -2.45548915e-02 1.10507086e-01
7.48908147e-02 1.93529308e-01 4.43115741e-01 -4.27409053e-01
-3.91038954e-01 2.39653125e-01 -2.06617683e-01 7.99685717e-01
-1.01807904e+00 1.58576131e-01 -1.01727676e+00 5.17413557e-01
-1.57772684e+00 -6.98431313e-01 -3.79927725e-01 2.22911477e+00
6.20889366e-01 -1.60584196e-01 -4.51884829e-02 -1.19609416e-01
6.95909917e-01 -5.96171558e-01 -5.12880027e-01 -1.93037510e-01
-1.56385258e-01 5.04888892e-01 4.87061620e-01 3.98160011e-01
-9.25079465e-01 8.88001919e-01 6.08269978e+00 5.65604806e-01
-1.66672564e+00 2.53634363e-01 4.01140898e-01 -2.22940356e-01
3.79979700e-01 4.97790612e-02 -5.24941981e-01 5.48455477e-01
1.11943638e+00 4.19696808e-01 1.85398981e-01 4.75381285e-01
7.04065919e-01 -2.60571986e-01 -6.78711832e-01 1.18570697e+00
-1.13036921e-02 -1.57710063e+00 1.61005616e-01 3.63291949e-01
7.70984173e-01 3.10506731e-01 3.27973425e-01 -9.30383876e-02
-2.38981470e-01 -1.39253879e+00 -4.23881084e-01 1.10977638e+00
1.00414753e+00 -4.15012687e-01 1.06487477e+00 7.12971091e-02
-5.38961589e-01 -1.75109372e-01 -7.19665647e-01 -1.40855387e-01
-1.36402339e-01 7.81371057e-01 -9.31074917e-01 1.99973524e-01
9.20748591e-01 1.02232254e+00 -5.99169195e-01 1.49038363e+00
-2.33106464e-01 8.52505803e-01 -1.45836577e-01 1.80622935e-01
1.65044546e-01 -5.90865195e-01 6.60582602e-01 6.07847273e-01
4.40529048e-01 2.04716817e-01 -2.41979547e-02 9.63300467e-01
3.07849586e-01 3.75367582e-01 -5.20061016e-01 -1.76933035e-01
-1.36820376e-01 1.08225989e+00 -3.49060178e-01 -1.64885491e-01
-5.60861707e-01 6.14718318e-01 2.10762560e-01 5.28837025e-01
-3.03176731e-01 -2.90213246e-02 4.50887442e-01 5.24603426e-01
8.59958455e-02 2.50816315e-01 -2.39867300e-01 -1.10834038e+00
-1.06985182e-01 -5.72186589e-01 1.44294277e-01 -1.12025666e+00
-1.30819297e+00 9.73591983e-01 -4.63391274e-01 -1.63878214e+00
-1.74340084e-01 -9.97198224e-01 -3.16129178e-01 1.28762186e+00
-1.64141083e+00 -1.22307503e+00 -5.55958927e-01 7.47359455e-01
1.63716659e-01 -7.34655380e-01 9.70062912e-01 3.01815480e-01
-7.91495383e-01 1.57945767e-01 4.72983979e-02 2.79328853e-01
8.62121522e-01 -1.18605661e+00 -7.29349852e-01 6.96212709e-01
-3.95930320e-01 6.88912511e-01 4.28033024e-01 -6.70184553e-01
-3.29351306e-01 -1.16703022e+00 7.94883847e-01 -2.10399047e-01
4.46303993e-01 4.75237757e-01 -9.63799000e-01 6.35242939e-01
4.15403634e-01 3.79256546e-01 1.09354472e+00 1.89870819e-01
1.42812088e-01 -1.55705839e-01 -1.06386840e+00 3.86096358e-01
5.33577561e-01 -5.43650925e-01 -7.10367143e-01 6.36151433e-01
2.04334855e-01 -3.73164356e-01 -1.04997981e+00 4.90342319e-01
3.83146018e-01 -1.23978209e+00 6.41657352e-01 -8.73102367e-01
6.65052652e-01 -1.98986173e-01 3.77730340e-01 -1.23914862e+00
-2.29976311e-01 -4.03941512e-01 2.54628599e-01 5.88360071e-01
3.40766013e-01 -1.08200014e+00 8.80500793e-01 5.10919511e-01
-1.00844711e-01 -8.57556224e-01 -6.82716906e-01 -4.95506525e-01
1.41699594e-02 -1.71027139e-01 -1.24839716e-01 8.37522805e-01
-6.29832447e-01 2.94003814e-01 -2.81580567e-01 2.99085945e-01
5.26127398e-01 -1.90723568e-01 4.66674924e-01 -1.88269353e+00
-2.03748897e-01 -2.44176567e-01 -9.32675183e-01 -2.37975687e-01
1.50195539e-01 -8.90542269e-01 -7.06108510e-01 -1.72026455e+00
3.10443051e-04 -2.58052915e-01 -3.18698555e-01 3.83903563e-01
1.16822511e-01 4.23559219e-01 -3.18491161e-01 4.92004663e-01
2.76439905e-01 2.99406081e-01 1.48913884e+00 4.04680967e-02
-6.06933296e-01 2.41616324e-01 -3.95173579e-01 1.05455542e+00
8.55528235e-01 -3.73384058e-01 -6.06117249e-01 -3.77286255e-01
3.06667447e-01 -1.35464594e-01 7.74301112e-01 -1.22985744e+00
4.20159936e-01 2.16988713e-01 5.43431938e-01 -4.48897272e-01
3.22796643e-01 -6.62644565e-01 1.41546369e-01 5.27730525e-01
9.85979661e-02 -6.37829125e-01 3.53449881e-01 5.08849561e-01
-6.48235440e-01 -2.17070699e-01 1.13060606e+00 -3.64321738e-01
-5.08718967e-01 3.65088314e-01 -4.63995785e-01 -4.21693735e-02
9.83711004e-01 -5.70688188e-01 -2.29747131e-01 -7.28955939e-02
-1.32060289e+00 -1.21883705e-01 3.25066447e-01 -2.14878365e-01
6.92823946e-01 -7.69549072e-01 -8.66296530e-01 4.04420108e-01
2.37399265e-02 8.55198428e-02 5.69734156e-01 1.31526029e+00
-9.87005055e-01 5.99994600e-01 -6.22048020e-01 -8.21528614e-01
-1.43376040e+00 3.47632736e-01 8.17705035e-01 -4.63163778e-02
-8.14252019e-01 1.00953019e+00 -8.62499774e-02 7.09171295e-02
4.17346768e-02 -5.08525074e-01 -8.19477797e-01 3.21564257e-01
4.13953036e-01 3.56408864e-01 1.41454414e-01 -2.34215021e-01
1.11560889e-01 9.45822716e-01 -1.91377878e-01 1.11092843e-01
1.51685488e+00 1.03173047e-01 -6.62960887e-01 2.80340999e-01
1.02268374e+00 -3.45553271e-02 -1.02894104e+00 -6.78519085e-02
-5.16945720e-01 -4.09241110e-01 7.65161335e-01 -8.61105621e-01
-1.27273011e+00 1.15435708e+00 1.23306417e+00 3.98885971e-03
1.22057068e+00 -3.48162830e-01 3.78890067e-01 2.50308812e-01
-5.44664301e-02 -1.11857486e+00 -9.44071189e-02 9.16175023e-02
8.36725533e-01 -1.56215680e+00 -2.26918280e-01 -3.91630590e-01
-5.98852873e-01 1.18301535e+00 4.99043316e-01 -4.83478487e-01
1.16402757e+00 -2.86805719e-01 4.95389193e-01 -3.09340298e-01
-5.34503758e-01 -5.84745884e-01 7.22073615e-01 7.09100783e-01
3.84111285e-01 -1.00354113e-01 -3.43442500e-01 4.17242885e-01
-3.99360359e-02 4.92772937e-01 6.20944619e-01 4.86198932e-01
-3.43049556e-01 -1.14216471e+00 -1.46019801e-01 9.79397297e-01
-7.22322643e-01 -3.11148375e-01 -7.51885399e-02 6.96571112e-01
7.25357890e-01 8.73775184e-01 2.44725242e-01 -3.44690904e-02
-8.97622555e-02 1.44208714e-01 4.21355069e-01 -9.96413052e-01
-3.77486259e-01 4.16537523e-01 -2.29432359e-02 -5.88480830e-01
-8.40005875e-01 -4.50875282e-01 -8.11363339e-01 1.07803121e-01
-2.35820323e-01 -2.61995066e-02 6.20015442e-01 8.77208352e-01
5.21130562e-01 8.21519613e-01 2.29368970e-01 -4.57814276e-01
-1.74894467e-01 -1.20542264e+00 -8.57324481e-01 5.17832562e-02
6.33750856e-01 -7.51589060e-01 -5.28738141e-01 3.55354339e-01] | [15.812399864196777, -3.9625892639160156] |
42ddb20d-2921-45f7-a4ac-704327796544 | what-matters-in-unsupervised-optical-flow | 2006.04902 | null | https://arxiv.org/abs/2006.04902v2 | https://arxiv.org/pdf/2006.04902v2.pdf | What Matters in Unsupervised Optical Flow | We systematically compare and analyze a set of key components in unsupervised optical flow to identify which photometric loss, occlusion handling, and smoothness regularization is most effective. Alongside this investigation we construct a number of novel improvements to unsupervised flow models, such as cost volume normalization, stopping the gradient at the occlusion mask, encouraging smoothness before upsampling the flow field, and continual self-supervision with image resizing. By combining the results of our investigation with our improved model components, we are able to present a new unsupervised flow technique that significantly outperforms the previous unsupervised state-of-the-art and performs on par with supervised FlowNet2 on the KITTI 2015 dataset, while also being significantly simpler than related approaches. | ['Anelia Angelova', 'Ariel Gordon', 'Rico Jonschkowski', 'Jonathan T. Barron', 'Austin Stone', 'Kurt Konolige'] | 2020-06-08 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/3651_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123470545.pdf | eccv-2020-8 | ['occlusion-handling'] | ['computer-vision'] | [-8.00337922e-03 -2.12233469e-01 -2.44596407e-01 -2.20714018e-01
1.38012007e-01 -4.62585121e-01 8.51341069e-01 -2.74904985e-02
-5.18879533e-01 1.01544404e+00 6.30024016e-01 -4.21008877e-02
-3.23567212e-01 -5.06026447e-01 -4.05497253e-01 -4.30743277e-01
-2.50529617e-01 3.70510489e-01 3.06037575e-01 3.50885987e-02
7.01674104e-01 7.08155990e-01 -1.61124659e+00 -2.50592411e-01
9.55587983e-01 9.14778054e-01 -3.60935867e-01 8.46988857e-01
-1.90639928e-01 1.47839522e+00 -3.19756359e-01 -2.96218127e-01
5.31975329e-01 -5.05686402e-01 -1.32708514e+00 3.37133378e-01
1.35979235e+00 -6.05988979e-01 -5.59837759e-01 6.30941033e-01
4.22292948e-01 7.89632261e-01 6.11528933e-01 -1.04571664e+00
-4.07230169e-01 1.07818969e-01 -2.71261275e-01 7.46890306e-01
2.73044527e-01 4.92280662e-01 9.30699289e-01 -8.06730747e-01
1.18671477e+00 1.28980935e+00 5.87888837e-01 3.99311066e-01
-1.59087253e+00 -2.28833050e-01 -3.88319790e-02 2.17321739e-01
-8.77398014e-01 -8.03021371e-01 7.67157555e-01 -7.51164615e-01
1.29606104e+00 -1.09959766e-01 6.95963085e-01 7.09321976e-01
4.15093042e-02 5.50338566e-01 9.69139636e-01 -4.02325958e-01
1.17279790e-01 -2.63258014e-02 1.78056136e-01 9.67441618e-01
6.21187165e-02 3.47515255e-01 -4.62655514e-01 1.13130383e-01
9.85256553e-01 -3.18069458e-01 -3.68427962e-01 -6.39249146e-01
-1.14079714e+00 7.79573798e-01 6.25332892e-01 3.17723125e-01
-2.05328420e-01 3.95871431e-01 5.73185384e-01 1.19217686e-01
7.33553588e-01 5.83393753e-01 -2.77898908e-01 -3.27297419e-01
-1.23575759e+00 3.37902755e-01 8.51803660e-01 7.36295402e-01
1.46496427e+00 3.37094218e-01 -2.77780324e-01 4.87175941e-01
2.10841790e-01 2.58515719e-02 2.50903755e-01 -1.72821665e+00
4.31709170e-01 3.89521211e-01 -4.07630112e-03 -1.08906162e+00
-3.16279233e-01 -3.75326395e-01 -4.93124515e-01 5.03877759e-01
6.99545920e-01 -1.76701590e-01 -8.96670818e-01 1.36078548e+00
2.01851368e-01 6.58559084e-01 -2.04636380e-01 9.20348883e-01
6.24342561e-01 3.96302968e-01 1.57632157e-01 -2.65276849e-01
7.07382977e-01 -1.41782677e+00 -8.89447093e-01 -4.55459617e-02
8.32199931e-01 -9.68492329e-01 7.18783915e-01 1.98485598e-01
-1.12467229e+00 -8.89601707e-01 -9.57500160e-01 -4.17951554e-01
-3.51088136e-01 -4.86280844e-02 1.24565279e+00 6.92044020e-01
-1.35788107e+00 1.18862808e+00 -8.79220545e-01 -5.18018842e-01
7.13808775e-01 2.17782542e-01 -5.55911362e-01 9.58095640e-02
-7.04242945e-01 9.53738451e-01 2.61384428e-01 3.25050647e-03
-5.93133628e-01 -1.15929437e+00 -1.08947909e+00 -1.98029637e-01
1.54625639e-01 -9.99541581e-01 8.04330230e-01 -9.36135948e-01
-1.82971621e+00 7.99046159e-01 -3.93699914e-01 -7.43669450e-01
8.43999267e-01 -3.49222839e-01 -1.31148277e-02 5.14377773e-01
6.82072863e-02 1.18455565e+00 8.91088903e-01 -9.63259757e-01
-5.58850408e-01 9.91407335e-02 5.66232624e-03 5.60490526e-02
-1.39462784e-01 -1.97507627e-03 -1.62527829e-01 -6.30794168e-01
-3.78868431e-01 -6.66600883e-01 -3.53089720e-01 4.53605503e-02
-1.68968201e-01 -1.09452888e-01 9.41437066e-01 -5.36813617e-01
1.10083067e+00 -1.93577445e+00 3.37055355e-01 3.43577960e-03
4.54484969e-01 3.29815775e-01 5.08774929e-02 1.00239456e-01
-1.99261323e-01 1.06426857e-01 -5.49202502e-01 -9.00709033e-01
-3.05536121e-01 3.90196294e-01 -1.64776161e-01 7.96241462e-01
4.30896282e-01 9.09584343e-01 -1.04951608e+00 -7.55041659e-01
1.00910068e+00 5.69073319e-01 -9.05233920e-01 1.19640917e-01
8.10587853e-02 9.97907698e-01 2.26640701e-02 4.07247573e-01
6.57522917e-01 -1.31301507e-01 -4.85081047e-01 -8.69398490e-02
-3.38161409e-01 5.56882441e-01 -1.45341039e+00 1.96798110e+00
-2.18147904e-01 1.06082439e+00 -1.69284344e-01 -9.22258079e-01
6.40837133e-01 1.11581355e-01 9.59236681e-01 -3.52368534e-01
1.89496100e-01 -7.37372264e-02 -8.64265710e-02 -7.32011616e-01
5.05779624e-01 -3.93983312e-02 8.71098042e-01 3.15472245e-01
7.88156569e-01 -1.79395825e-01 8.63813281e-01 3.66472661e-01
9.51995194e-01 6.62031770e-01 -5.58762699e-02 -7.67445624e-01
1.02277303e+00 -2.81842127e-02 4.75711882e-01 6.45259023e-01
-9.05769467e-01 7.95715988e-01 3.84148657e-01 -7.40166962e-01
-7.93701589e-01 -8.99808228e-01 -2.39013031e-01 7.63869345e-01
-2.11200882e-02 -6.23237312e-01 -5.95618963e-01 -8.70563149e-01
8.43588486e-02 1.83178872e-01 -8.34103227e-01 3.33917528e-01
-9.15590525e-01 -5.41378200e-01 4.14526135e-01 4.26473171e-01
6.13519371e-01 -1.22573340e+00 -4.66987729e-01 3.19900662e-01
-1.54150620e-01 -1.36386442e+00 -4.72978622e-01 -3.20510045e-02
-1.11273026e+00 -1.27102268e+00 -4.61571604e-01 -6.13788962e-01
6.03653610e-01 -6.45376295e-02 1.33413410e+00 2.65425622e-01
-5.53348720e-01 3.76013637e-01 -1.20792285e-01 1.82609055e-02
-1.54029667e-01 2.80106187e-01 -2.46155828e-01 1.87045515e-01
-8.55907872e-02 -6.23619854e-01 -8.65723491e-01 -1.24572270e-01
-8.81518424e-01 -3.98570746e-01 -8.48974138e-02 5.34760952e-01
1.41426072e-01 -2.51077920e-01 1.84505060e-01 -9.57922041e-01
2.54747540e-01 -6.84743524e-02 -4.59226251e-01 -2.26130575e-01
-4.94524688e-01 3.06224138e-01 6.11903250e-01 -1.42267257e-01
-1.21501386e+00 1.47809729e-01 -1.90138564e-01 -6.57447815e-01
-3.16983938e-01 -1.64963380e-01 3.43026072e-01 -6.70638978e-01
7.75768518e-01 -2.51655996e-01 2.21434683e-01 -2.55130380e-01
6.50905311e-01 -1.03646062e-01 5.92231154e-01 -4.56741333e-01
7.83648551e-01 1.01808619e+00 4.25237834e-01 -9.93105412e-01
-6.80080295e-01 -8.22157502e-01 -1.11005199e+00 -4.18112218e-01
9.72938776e-01 -5.03818333e-01 -5.59934258e-01 5.03141940e-01
-9.76936340e-01 -3.34351003e-01 -7.40351140e-01 7.73736596e-01
-6.59777701e-01 7.12880850e-01 -8.13939571e-01 -5.59285343e-01
-9.07975957e-02 -1.21279061e+00 6.52682900e-01 2.64955521e-01
-2.13528812e-01 -1.59998763e+00 5.09638548e-01 3.79561901e-01
7.88939774e-01 2.56246060e-01 2.65307724e-01 -9.69042704e-02
-5.52798688e-01 2.42924973e-01 -2.53569812e-01 6.50420189e-01
3.00137758e-01 5.49925506e-01 -1.09916008e+00 -9.50858667e-02
-1.18992120e-01 -2.61116982e-01 1.37656569e+00 6.29929304e-01
8.03360760e-01 9.28795040e-02 1.87324807e-02 1.23942947e+00
1.42536569e+00 -3.48993063e-01 8.98935318e-01 3.91088039e-01
9.82479692e-01 8.97290409e-01 1.55837700e-01 1.41185045e-01
2.12593004e-01 2.84708083e-01 4.61048245e-01 -1.78078309e-01
-5.76770723e-01 -3.90357524e-02 3.93832251e-02 4.88058805e-01
-5.90431690e-01 1.00999601e-01 -3.33190501e-01 6.36730015e-01
-1.76693773e+00 -9.47550595e-01 -2.84241736e-01 1.94053113e+00
4.03365761e-01 1.66399926e-01 3.12829435e-01 2.01442912e-01
3.90879035e-01 6.35944843e-01 -3.53169404e-02 -4.39297169e-01
-1.13612808e-01 7.46493101e-01 7.85677671e-01 1.21311128e+00
-1.53402114e+00 1.33592176e+00 7.68220949e+00 2.63420165e-01
-9.79873002e-01 -1.11676119e-01 4.19917166e-01 1.20270150e-02
3.74391885e-03 4.74196315e-01 -6.54485524e-01 2.49258369e-01
5.59541762e-01 3.16424996e-01 5.82071364e-01 4.47990596e-01
3.36897612e-01 -1.95615500e-01 -1.01649296e+00 7.55735397e-01
2.13997498e-01 -1.65363860e+00 8.76470353e-04 -6.36579394e-02
8.85760248e-01 1.35937348e-01 -3.43331516e-01 -6.28738105e-02
2.01395839e-01 -8.66769135e-01 3.41977507e-01 5.90606272e-01
2.41226390e-01 -3.56688291e-01 5.36954939e-01 -3.21177006e-01
-1.15973139e+00 1.00573294e-01 -1.46384850e-01 -5.06862819e-01
4.68777508e-01 6.09034479e-01 -2.44365722e-01 6.65369093e-01
6.48590505e-01 1.48374104e+00 -7.23671377e-01 1.23658442e+00
-3.95712048e-01 5.66426158e-01 -3.23826969e-01 5.76355338e-01
4.32375431e-01 -5.64294517e-01 7.52722323e-01 1.45502090e+00
-2.80887038e-01 -3.10909510e-01 1.29416853e-01 9.13834929e-01
4.40394729e-02 3.87000918e-01 -4.83188897e-01 4.84420091e-01
-1.32319763e-01 1.26270103e+00 -9.20632362e-01 -6.83040559e-01
-4.74881291e-01 7.58082330e-01 5.21818221e-01 4.98391390e-01
-5.00966489e-01 -2.95159847e-01 8.49295616e-01 9.17667150e-03
3.83532554e-01 -4.50514078e-01 -3.11737299e-01 -1.41464257e+00
-8.10774788e-02 -9.49184671e-02 4.05005753e-01 -5.65973759e-01
-1.11457527e+00 2.92795151e-01 -1.70443226e-02 -1.04831672e+00
-3.06661844e-01 -6.64170325e-01 -7.87495911e-01 7.50962138e-01
-1.93555248e+00 -6.94085479e-01 -3.56775463e-01 6.81539893e-01
3.64470154e-01 1.20659463e-01 4.30621952e-01 5.13835311e-01
-6.46041632e-01 1.72212198e-01 -3.45674187e-01 2.04944581e-01
9.79034185e-01 -1.45893550e+00 4.14334565e-01 1.24838555e+00
2.60253340e-01 5.49968541e-01 6.47869170e-01 -5.07002950e-01
-9.23213065e-01 -8.26518476e-01 9.42123175e-01 -5.29168010e-01
8.67900848e-01 -1.79394364e-01 -9.04489458e-01 8.63163173e-01
4.95998770e-01 5.07849693e-01 1.52195498e-01 -1.90535992e-01
-5.52224927e-02 -5.39128520e-02 -1.20291042e+00 3.34456682e-01
1.27605319e+00 -3.08642149e-01 -6.54191256e-01 1.95023552e-01
3.62574697e-01 -3.77818853e-01 -8.25890303e-01 4.31925863e-01
3.12632829e-01 -1.41752100e+00 1.12646651e+00 -5.12140453e-01
5.11724234e-01 -2.72871286e-01 3.97043258e-01 -1.23215759e+00
-4.51720536e-01 -1.15343082e+00 -2.41419673e-01 1.22230792e+00
2.17270423e-02 -8.51073444e-01 9.30701911e-01 4.28781927e-01
-9.65537056e-02 -2.09913194e-01 -8.44942927e-01 -5.96540332e-01
6.91559091e-02 -4.43356633e-01 1.48005188e-02 9.56634820e-01
-1.64315641e-01 1.55995518e-01 -5.27419508e-01 -3.67781609e-01
8.23386073e-01 -3.36357027e-01 9.91664052e-01 -1.15870988e+00
4.53500077e-02 -7.99919724e-01 -8.39900613e-01 -1.22666657e+00
4.19486344e-01 -8.36982191e-01 -1.81254596e-01 -1.46981478e+00
-4.58538204e-01 -1.88192189e-01 -1.72483250e-01 2.97479779e-01
-1.83480144e-01 5.63608706e-01 2.05525652e-01 3.22967112e-01
-3.98833156e-01 4.47001845e-01 1.53432393e+00 4.82472293e-02
-5.46523213e-01 -2.62981564e-01 -3.27713847e-01 7.57682204e-01
5.42635858e-01 -2.43516549e-01 -3.37117583e-01 -4.04034883e-01
-2.10408837e-01 -3.50311160e-01 3.39072526e-01 -1.11943364e+00
2.70769566e-01 -3.00695444e-03 3.53122860e-01 -2.76350617e-01
-5.17988205e-02 -6.08347595e-01 -4.04130250e-01 4.34263319e-01
-3.10992062e-01 -3.91040109e-02 2.43946820e-01 2.14294612e-01
-1.84135303e-01 -2.32186928e-01 8.26231778e-01 -2.48759523e-01
-8.82547915e-01 3.68672639e-01 -3.46124828e-01 4.39551890e-01
7.41755724e-01 -2.67126501e-01 -3.47439766e-01 -8.62617195e-02
-8.67110372e-01 1.30395144e-01 3.57070059e-01 5.05036116e-01
1.09763883e-01 -9.97949719e-01 -7.31942058e-01 5.94206810e-01
-1.66041732e-01 -9.62682143e-02 2.07127124e-01 9.92606878e-01
-1.21658969e+00 4.99027044e-01 -4.61653531e-01 -7.98681140e-01
-7.14814305e-01 4.50466305e-01 5.88846982e-01 -3.32288563e-01
-9.80200529e-01 7.08019614e-01 -8.17522109e-02 -3.16157967e-01
3.60964358e-01 -4.50057656e-01 -3.14524382e-01 7.92009085e-02
3.70322853e-01 8.42739344e-01 8.36157054e-02 -9.12001073e-01
-4.11408752e-01 9.17065382e-01 2.22393990e-01 -2.03459710e-03
1.16129017e+00 -2.69532114e-01 -3.04831356e-01 2.33111173e-01
1.28995228e+00 5.33480644e-02 -1.68025517e+00 6.52593933e-03
-2.02329248e-01 -8.69269490e-01 1.95124701e-01 -3.43458146e-01
-1.49939525e+00 7.77752161e-01 4.15461004e-01 2.18485221e-01
8.73989880e-01 -3.55162084e-01 4.71023977e-01 1.10403121e-01
-9.96031687e-02 -1.03530777e+00 -1.16678707e-01 6.35880709e-01
4.41735834e-01 -1.37676966e+00 3.37452203e-01 -6.38164639e-01
-2.69184500e-01 1.14078188e+00 5.33601463e-01 -9.42821681e-01
1.02043819e+00 2.41858795e-01 1.06143281e-01 -1.63288623e-01
-4.71121788e-01 -5.37969470e-01 4.50425535e-01 5.87897718e-01
4.67613965e-01 -4.75912720e-01 -3.72348011e-01 -7.16640830e-01
-3.12634587e-01 1.57797024e-01 4.74018484e-01 9.90730822e-01
-1.62702128e-01 -1.08363402e+00 -7.21684247e-02 3.14617306e-01
-6.68996811e-01 -1.29080087e-01 -1.85863018e-01 9.39689934e-01
1.34640709e-01 8.95269632e-01 3.09093803e-01 9.32505950e-02
3.72066349e-01 8.94825757e-02 7.55363166e-01 -3.76152098e-01
-7.72922218e-01 5.87760545e-02 6.26090122e-03 -9.01699364e-01
-1.21631849e+00 -6.21892452e-01 -1.03694034e+00 -3.69063139e-01
-2.44168267e-01 -1.81109309e-01 3.56380075e-01 1.39670515e+00
2.23547995e-01 4.91983116e-01 5.51443517e-01 -1.32525480e+00
2.82899499e-01 -1.10200214e+00 -2.73409218e-01 8.35523129e-01
6.61268950e-01 -8.37675691e-01 -1.00000024e+00 4.85233366e-01] | [8.793191909790039, -1.788657546043396] |
4fa3bb2f-3d4c-4c45-ae11-890aeff34fd0 | generating-medically-accurate-summaries-of | 2305.05982 | null | https://arxiv.org/abs/2305.05982v1 | https://arxiv.org/pdf/2305.05982v1.pdf | Generating medically-accurate summaries of patient-provider dialogue: A multi-stage approach using large language models | A medical provider's summary of a patient visit serves several critical purposes, including clinical decision-making, facilitating hand-offs between providers, and as a reference for the patient. An effective summary is required to be coherent and accurately capture all the medically relevant information in the dialogue, despite the complexity of patient-generated language. Even minor inaccuracies in visit summaries (for example, summarizing "patient does not have a fever" when a fever is present) can be detrimental to the outcome of care for the patient. This paper tackles the problem of medical conversation summarization by discretizing the task into several smaller dialogue-understanding tasks that are sequentially built upon. First, we identify medical entities and their affirmations within the conversation to serve as building blocks. We study dynamically constructing few-shot prompts for tasks by conditioning on relevant patient information and use GPT-3 as the backbone for our experiments. We also develop GPT-derived summarization metrics to measure performance against reference summaries quantitatively. Both our human evaluation study and metrics for medical correctness show that summaries generated using this approach are clinically accurate and outperform the baseline approach of summarizing the dialog in a zero-shot, single-prompt setting. | ['Anitha Kannan', 'Elliot Schumacher', 'Varun Nair'] | 2023-05-10 | null | null | null | null | ['dialogue-understanding'] | ['natural-language-processing'] | [ 6.60277069e-01 6.59416497e-01 -2.35745177e-01 -6.76999152e-01
-1.47664475e+00 -5.14227688e-01 4.98319536e-01 1.18392026e+00
-1.98505044e-01 9.18992162e-01 1.24148309e+00 -2.87153989e-01
-2.22191930e-01 -3.04893047e-01 -2.50872541e-02 -3.43858689e-01
3.26896384e-02 9.37679350e-01 -1.13364495e-01 -2.22740695e-01
2.14012057e-01 1.79577425e-01 -7.45468020e-01 6.73142731e-01
1.03411651e+00 5.58182776e-01 8.79133772e-03 1.21742344e+00
-9.66008902e-02 1.06127548e+00 -1.14114666e+00 -4.10691440e-01
-3.23171824e-01 -1.01864874e+00 -1.23392808e+00 2.14996383e-01
1.71820506e-01 -7.11323142e-01 -5.31576127e-02 5.55390179e-01
7.73292482e-01 3.19719791e-01 6.08305097e-01 -7.90454686e-01
-2.25826446e-02 7.56676555e-01 2.60247849e-02 2.36606285e-01
9.85983431e-01 2.20639795e-01 9.91192222e-01 -1.10674739e-01
6.98535740e-01 1.21823275e+00 6.41698062e-01 8.99070919e-01
-1.06431770e+00 -8.39029402e-02 -7.71263316e-02 -4.25696641e-01
-5.52326083e-01 -7.91063309e-01 9.10578296e-02 -3.24465215e-01
9.82628167e-01 8.94871175e-01 4.40830469e-01 1.01320124e+00
3.31370443e-01 5.09568453e-01 3.27410609e-01 1.47735532e-02
2.44045824e-01 9.40008089e-02 6.72337115e-01 5.79039037e-01
2.71378607e-01 -5.74875116e-01 -3.96172673e-01 -9.05735731e-01
3.48903507e-01 1.41574636e-01 -6.63888812e-01 4.81905073e-01
-1.25491452e+00 9.49522913e-01 3.39891426e-02 2.39929512e-01
-6.88638866e-01 -2.26212338e-01 7.75968432e-01 1.36640653e-01
3.58622342e-01 7.89349914e-01 -1.67585716e-01 -4.19703573e-01
-9.51943159e-01 4.80009347e-01 1.53660870e+00 9.00752902e-01
-7.68768862e-02 -4.75878388e-01 -1.11520302e+00 7.72308230e-01
-2.05703557e-01 2.79823065e-01 5.02444983e-01 -1.21525860e+00
4.30234879e-01 5.02985597e-01 6.08275235e-01 -8.05317402e-01
-6.22762740e-01 -1.24364749e-01 -8.27500761e-01 -7.29593694e-01
2.38564923e-01 -5.43332815e-01 -6.05105400e-01 1.52731168e+00
3.18775117e-01 -2.38273114e-01 4.40513164e-01 5.45487583e-01
1.53013587e+00 7.28426397e-01 3.52747202e-01 -9.37816739e-01
1.71702886e+00 -1.02165627e+00 -1.22596836e+00 -1.33155942e-01
1.03202510e+00 -9.93154883e-01 6.27167642e-01 -5.03903553e-02
-1.33105552e+00 -1.69709489e-01 -5.83952129e-01 -1.94612429e-01
2.36377627e-01 -7.71332756e-02 2.56632119e-01 1.77454248e-01
-9.03431058e-01 7.71773815e-01 -6.79290414e-01 -6.79868102e-01
1.43284068e-01 1.33196086e-01 -2.24387586e-01 7.58489892e-02
-1.06474543e+00 1.01986539e+00 1.92000598e-01 -2.46296659e-01
-5.23958206e-01 -8.37676883e-01 -1.08968556e+00 3.59458238e-01
3.55022818e-01 -1.28280687e+00 2.17325735e+00 -2.39822641e-01
-1.11029875e+00 8.81780624e-01 -5.56725502e-01 -5.98316908e-01
5.79285443e-01 -1.91338181e-01 -1.37924567e-01 4.68275338e-01
2.22678915e-01 5.33641160e-01 2.21979171e-01 -8.82599533e-01
-7.75146425e-01 -3.07966601e-02 1.24765188e-01 5.10573804e-01
2.05848575e-01 1.89986452e-01 -3.07342976e-01 -4.12483037e-01
-2.51658410e-01 -8.06008339e-01 -5.57140470e-01 -3.46350968e-01
-9.98167634e-01 -2.66325384e-01 3.83917451e-01 -9.02706265e-01
1.59513450e+00 -1.93518794e+00 -2.87100732e-01 -2.68419117e-01
5.57112277e-01 3.15160096e-01 4.07239124e-02 1.00329542e+00
7.50586167e-02 1.84707835e-01 -5.74523807e-01 -4.92258132e-01
-2.57869184e-01 2.64714986e-01 -3.49320292e-01 4.11913060e-02
1.74342290e-01 1.04974973e+00 -1.18799162e+00 -8.04009318e-01
1.25529811e-01 3.21212649e-01 -4.04457271e-01 6.17408454e-01
-2.65128702e-01 8.33675027e-01 -5.93811154e-01 2.36183897e-01
-5.78216985e-02 -7.63258219e-01 1.18271552e-01 -1.41659379e-01
2.13707134e-01 8.51415992e-01 -3.87990803e-01 1.47597134e+00
-2.77914435e-01 3.17702115e-01 9.19979289e-02 -5.12856424e-01
4.35002714e-01 7.39429176e-01 6.69281960e-01 -1.85785443e-01
2.51353741e-01 -3.52455750e-02 3.83718908e-02 -7.30414152e-01
6.88374579e-01 -2.63636768e-01 -3.64841878e-01 7.42695212e-01
-4.50120151e-01 -4.44786966e-01 3.04094076e-01 7.94139266e-01
1.42660654e+00 -6.91563010e-01 9.63982463e-01 -3.19743603e-02
3.81817788e-01 3.07355464e-01 2.10541710e-01 1.04772723e+00
-1.82410255e-01 7.22050548e-01 7.24277854e-01 -5.29473126e-02
-7.48604953e-01 -5.86246848e-01 -7.15962518e-03 6.44204378e-01
4.02444676e-02 -7.79459596e-01 -8.06853473e-01 -6.68849826e-01
-1.35750636e-01 1.22641516e+00 -4.74924445e-01 -1.71219379e-01
-4.92517203e-01 -6.20177448e-01 4.99236882e-01 4.21047747e-01
1.24362633e-01 -1.03513765e+00 -1.13322330e+00 5.61368763e-01
-8.60805690e-01 -1.11554205e+00 -9.80275512e-01 -7.66240209e-02
-8.20923507e-01 -1.19271731e+00 -8.55287731e-01 -6.23111904e-01
5.09862959e-01 1.55873448e-01 1.20672798e+00 1.86948106e-01
-4.32014793e-01 8.67390871e-01 -2.20932946e-01 -3.84063125e-01
-1.12394261e+00 -8.69072899e-02 -3.89337957e-01 -4.74445969e-01
8.94895718e-02 -1.25041619e-01 -7.53401697e-01 -9.99637321e-02
-7.91420162e-01 3.36093068e-01 1.43099099e-01 9.28310513e-01
3.80406737e-01 -6.11794949e-01 6.69485271e-01 -1.36853516e+00
1.56110775e+00 -4.28340733e-01 3.05587262e-01 4.16731089e-01
-1.59945369e-01 8.14987123e-02 4.00919437e-01 -3.87379341e-02
-9.37033892e-01 -1.41794935e-01 -3.70147258e-01 2.89703697e-01
-9.69181210e-02 3.62927765e-01 4.46581334e-01 6.65084660e-01
6.45733654e-01 1.72473654e-01 1.37828156e-01 -3.29934776e-01
1.85852930e-01 8.76183987e-01 7.79814005e-01 -2.31960759e-01
1.60509404e-02 1.48434535e-01 -3.40998948e-01 -8.57514799e-01
-1.27169681e+00 -8.73425961e-01 -1.37294427e-01 5.96258193e-02
9.30977523e-01 -5.49691916e-01 -1.01884401e+00 -1.64877817e-01
-1.58181036e+00 3.67921032e-02 -6.28246427e-01 2.94922978e-01
-5.52287757e-01 5.62538445e-01 -7.30316758e-01 -8.66010785e-01
-9.50866938e-01 -1.05841744e+00 1.30445266e+00 2.49255568e-01
-1.13303292e+00 -1.04201007e+00 3.11279148e-01 2.40564257e-01
2.67861038e-01 5.40274024e-01 1.03290939e+00 -1.40720797e+00
2.09521037e-02 -3.36318642e-01 -1.38457194e-01 -9.73627344e-03
7.70439327e-01 -1.55988038e-01 -7.51524389e-01 -3.03877965e-02
3.48723292e-01 -2.11915240e-01 6.56330049e-01 5.05630314e-01
8.51046741e-01 -9.64728355e-01 -5.52591443e-01 -4.80280966e-02
7.70563722e-01 4.14943337e-01 1.33682594e-01 -6.09740973e-01
2.92067438e-01 9.78800774e-01 6.78979814e-01 8.01546097e-01
4.71323580e-01 3.10328603e-01 -7.64377117e-02 -1.65050223e-01
2.67268177e-02 -8.72382522e-02 -1.42101645e-01 8.98651898e-01
2.77595103e-01 -5.91295123e-01 -8.36166084e-01 4.76835310e-01
-2.01620531e+00 -8.69763613e-01 -4.24194373e-02 2.05662274e+00
1.17786467e+00 -1.05317168e-01 9.03713927e-02 -3.06543767e-01
6.43696308e-01 1.52490884e-01 -4.73898262e-01 -5.82325637e-01
3.70503247e-01 -6.29397184e-02 9.38808694e-02 8.86547506e-01
-8.89751315e-01 4.26755935e-01 6.66443443e+00 1.65321082e-01
-7.95250833e-01 -4.84450720e-02 9.09577668e-01 -4.26025093e-02
-2.35078067e-01 -4.70091522e-01 -7.85304070e-01 3.71419936e-01
1.21605265e+00 -7.09594727e-01 -1.76190227e-01 5.13396204e-01
6.26175404e-01 -3.82570505e-01 -1.82427776e+00 8.66773367e-01
1.48022696e-01 -1.43195915e+00 1.16535470e-01 -3.40828449e-01
5.09250164e-01 -3.66188884e-01 -4.14630383e-01 1.07071869e-01
5.06963015e-01 -9.81015742e-01 -8.64369869e-02 5.23054898e-01
7.77374148e-01 -2.61757940e-01 8.61185372e-01 4.40185368e-01
-7.35579431e-01 3.05341989e-01 7.49173090e-02 2.44208679e-01
6.88972175e-01 3.48135144e-01 -1.70689476e+00 5.73931277e-01
9.57748145e-02 5.07428527e-01 -6.02069385e-02 1.00597954e+00
8.65294635e-02 3.19491744e-01 -2.13497326e-01 -6.67925403e-02
3.91705424e-01 1.34040356e-01 7.99225628e-01 1.61072791e+00
9.36754495e-02 9.70898032e-01 4.12355900e-01 3.95009369e-01
-2.60703623e-01 3.46344352e-01 -6.91215813e-01 -4.10242409e-01
4.41226393e-01 1.08136737e+00 -6.05883598e-01 -1.03216577e+00
-5.75507581e-02 9.91427958e-01 -1.46221504e-01 4.69809100e-02
-5.98283887e-01 -3.08712095e-01 4.90153015e-01 -5.06147854e-02
-2.00482354e-01 3.27206433e-01 -1.61756352e-01 -8.64312410e-01
-1.61969870e-01 -1.06745696e+00 6.36223137e-01 -5.83358407e-01
-1.13036001e+00 7.86698759e-01 -1.88443027e-02 -1.08331037e+00
-9.33633804e-01 1.70189351e-01 -7.34465957e-01 9.44820285e-01
-1.16516900e+00 -4.80731666e-01 -4.89680022e-01 1.95733562e-01
9.56028879e-01 4.41686124e-01 1.34233189e+00 -2.47645658e-02
-3.92592818e-01 3.00165832e-01 -2.16004193e-01 4.99327108e-03
1.02638948e+00 -1.15007699e+00 4.65748698e-01 3.34962249e-01
-4.71962452e-01 8.90757024e-01 1.06597698e+00 -9.55939293e-01
-9.26228464e-01 -1.01626468e+00 1.44970882e+00 -6.52120531e-01
3.08443040e-01 2.27913976e-01 -1.10677540e+00 4.80484337e-01
3.51038277e-01 -6.30764723e-01 9.75042164e-01 -2.22616851e-01
1.71356171e-01 2.64734566e-01 -1.21595478e+00 6.65009916e-01
7.67624021e-01 -3.34215343e-01 -1.05009794e+00 9.00499463e-01
1.22039866e+00 -7.75286615e-01 -8.98568451e-01 2.88100511e-01
4.14928555e-01 -6.57216847e-01 6.86968982e-01 -1.08563435e+00
6.46512270e-01 3.31231803e-01 2.50787348e-01 -1.31456614e+00
1.22262008e-01 -1.06689990e+00 3.86462063e-02 9.48914349e-01
4.52464074e-01 -4.86767560e-01 3.68937105e-01 1.32130122e+00
-2.59353071e-01 -7.75230706e-01 -5.39317727e-01 -4.29627486e-02
-5.31003535e-01 1.47025406e-01 2.68184781e-01 8.42628479e-01
4.98474509e-01 7.79714048e-01 -3.38799477e-01 -8.83100256e-02
2.52281815e-01 1.56433702e-01 5.36807358e-01 -1.13704133e+00
-2.88671792e-01 -3.50478172e-01 2.93092161e-01 -1.19092834e+00
-3.77325654e-01 -6.16283953e-01 4.70512480e-01 -2.23702836e+00
4.86330211e-01 -1.35118514e-01 2.49151289e-01 5.38470328e-01
-4.84010458e-01 -4.08151329e-01 4.63927574e-02 2.82174259e-01
-8.16858232e-01 1.33019358e-01 1.32503664e+00 -2.10684344e-01
-6.06997430e-01 3.48797917e-01 -1.10182345e+00 4.97744769e-01
5.56170940e-01 -4.40316081e-01 -4.50960100e-01 2.37599462e-02
-2.73768127e-01 1.09753323e+00 -8.20346400e-02 -4.09585088e-01
4.32314783e-01 -1.38434932e-01 -1.59151226e-01 -6.01353824e-01
3.97073358e-01 -5.34627497e-01 2.38721464e-02 8.30788851e-01
-1.01246655e+00 1.07780501e-01 1.75515145e-01 5.65850139e-01
-1.93943620e-01 -2.65263885e-01 6.07482016e-01 -4.14691269e-01
3.19988206e-02 5.71894534e-02 -3.48051161e-01 5.39280176e-01
7.61978805e-01 1.12883989e-02 -5.72164774e-01 -9.28171813e-01
-9.81489778e-01 7.54353821e-01 2.06710815e-01 1.30900577e-01
5.52128971e-01 -5.85897624e-01 -1.13541234e+00 -3.99013668e-01
1.45743772e-01 1.51651055e-01 4.30138916e-01 9.76758718e-01
-5.96114337e-01 7.73465872e-01 2.52212197e-01 -5.43150425e-01
-1.87219739e+00 2.61570543e-01 1.66482091e-01 -7.27863431e-01
-9.13280964e-01 6.89608335e-01 4.48645353e-01 -9.59001184e-02
7.16630697e-01 -7.51068890e-01 -3.40356857e-01 3.17061186e-01
1.01964700e+00 1.56030953e-01 5.49696200e-02 -2.63576716e-01
-2.54430473e-01 1.13208611e-02 -4.54747796e-01 -1.65398300e-01
1.18972147e+00 -3.33732814e-01 -1.04965948e-01 5.86436212e-01
8.62318993e-01 -1.08894352e-02 -5.33112466e-01 -2.60579199e-01
-2.23270804e-02 -3.25660557e-02 -4.75655615e-01 -1.12346065e+00
-1.20022818e-01 7.18631029e-01 -1.15353823e-01 5.48639357e-01
9.54849303e-01 1.49058923e-01 1.02238953e+00 5.96240222e-01
-1.95093229e-01 -5.43700397e-01 3.49517092e-02 2.19726980e-01
1.07594728e+00 -1.26207006e+00 3.75382341e-02 -5.55125713e-01
-1.20339584e+00 9.91565108e-01 1.75254971e-01 5.38138747e-01
1.83019787e-01 1.21295258e-01 1.70676291e-01 -4.49346334e-01
-1.09817684e+00 9.34506506e-02 2.50210822e-01 2.03295738e-01
6.74322486e-01 1.44057617e-01 -5.54146409e-01 5.94109416e-01
-2.30967745e-01 -1.94301873e-01 7.16498673e-01 1.06736994e+00
-5.57153761e-01 -7.59503365e-01 -7.46708736e-02 8.29094410e-01
-6.48056805e-01 -3.23250800e-01 -7.51356781e-01 3.71036857e-01
-6.37977600e-01 1.24272215e+00 -5.57001233e-02 -1.28094675e-02
4.96107191e-01 2.89345294e-01 2.27485821e-01 -1.29380190e+00
-1.14599991e+00 1.19364560e-01 9.29048777e-01 -4.92567331e-01
-4.31891263e-01 -4.79270995e-01 -1.44244254e+00 -2.41657794e-01
1.27853837e-03 6.67706788e-01 2.84800321e-01 9.37465370e-01
4.59364831e-01 8.60892236e-01 4.16445255e-01 -2.65004128e-01
-7.94134796e-01 -1.00766110e+00 -2.83055492e-02 4.97759998e-01
9.10015881e-01 3.29969786e-02 -1.03616923e-01 1.53728858e-01] | [12.260376930236816, 8.83348274230957] |
90e01783-b7ed-4d0d-94d2-325bd70f0fb6 | deep-video-harmonization-with-color-mapping | 2205.00687 | null | https://arxiv.org/abs/2205.00687v1 | https://arxiv.org/pdf/2205.00687v1.pdf | Deep Video Harmonization with Color Mapping Consistency | Video harmonization aims to adjust the foreground of a composite video to make it compatible with the background. So far, video harmonization has only received limited attention and there is no public dataset for video harmonization. In this work, we construct a new video harmonization dataset HYouTube by adjusting the foreground of real videos to create synthetic composite videos. Moreover, we consider the temporal consistency in video harmonization task. Unlike previous works which establish the spatial correspondence, we design a novel framework based on the assumption of color mapping consistency, which leverages the color mapping of neighboring frames to refine the current frame. Extensive experiments on our HYouTube dataset prove the effectiveness of our proposed framework. Our dataset and code are available at https://github.com/bcmi/Video-Harmonization-Dataset-HYouTube. | ['Liqing Zhang', 'Wenyan Cong', 'Li Niu', 'Shengyuan Huang', 'Xinyuan Lu'] | 2022-05-02 | null | null | null | null | ['video-harmonization'] | ['computer-vision'] | [-1.25330716e-01 -3.90125126e-01 -1.19017377e-01 3.77273038e-02
-4.52706546e-01 -5.30912519e-01 3.74540120e-01 -3.14428687e-01
-1.70510367e-01 7.42186964e-01 2.64193267e-01 -4.73166853e-02
1.45943940e-01 -6.70369208e-01 -7.84360468e-01 -7.20907569e-01
4.26028848e-01 -3.20899844e-01 4.59612221e-01 -1.17924541e-01
1.21213473e-01 -2.57177632e-02 -1.33191550e+00 1.15534745e-01
8.78353775e-01 8.22877586e-01 2.15748504e-01 5.92375636e-01
2.95443058e-01 5.48817635e-01 -5.61909556e-01 -2.79941529e-01
7.05757976e-01 -8.07606697e-01 -4.76350218e-01 2.43130088e-01
4.51082647e-01 -2.42890656e-01 -6.69999540e-01 1.08388734e+00
7.00195730e-01 3.17452163e-01 -5.42894043e-02 -1.61309600e+00
-7.09725440e-01 5.59529841e-01 -7.92699635e-01 3.63851666e-01
3.86097014e-01 1.77185640e-01 8.70400667e-01 -7.10889399e-01
9.03009951e-01 8.89534175e-01 4.20447141e-01 3.49569470e-01
-1.09421015e+00 -1.00794852e+00 1.74177513e-01 6.75640643e-01
-1.70110881e+00 -4.12610680e-01 9.72621799e-01 -2.10707545e-01
1.49618536e-01 4.33988005e-01 1.16900253e+00 8.38675082e-01
-2.81934291e-02 6.76477849e-01 9.63993549e-01 -3.12511086e-01
6.32657334e-02 -3.65080208e-01 -3.98286343e-01 4.55408067e-01
1.29765704e-01 -1.48460686e-01 -6.83256269e-01 3.28232199e-02
8.23993981e-01 -8.02571997e-02 -7.15529978e-01 -5.34892142e-01
-1.41516209e+00 5.37102938e-01 1.59484938e-01 3.34864765e-01
-1.94638982e-01 1.22120656e-01 3.52470160e-01 1.70896173e-01
2.90129453e-01 1.48924366e-01 7.45373368e-02 -1.43915340e-01
-9.52355087e-01 3.24742258e-01 4.25121248e-01 1.09464812e+00
7.80350208e-01 7.07677603e-02 -1.63641050e-01 7.87303865e-01
-2.44685784e-02 2.76800841e-01 4.30288196e-01 -1.37579882e+00
3.33955824e-01 3.45117509e-01 3.51127870e-02 -1.21459568e+00
-4.77531925e-02 -1.31287739e-01 -7.06726372e-01 -6.88330159e-02
1.91050738e-01 -2.04611614e-01 -3.00488055e-01 1.78754830e+00
6.03903830e-01 8.92906308e-01 -1.62429810e-01 1.10350859e+00
8.49186540e-01 7.07386136e-01 -2.14093059e-01 -4.34081763e-01
9.92437005e-01 -1.13168097e+00 -8.29710841e-01 5.36759257e-01
1.40920296e-01 -1.15517223e+00 9.91083384e-01 2.81079650e-01
-1.36198306e+00 -6.20715439e-01 -1.15296841e+00 3.79788466e-02
9.71637964e-02 -1.03953853e-01 1.79818898e-01 4.91947234e-01
-1.12948489e+00 4.86534685e-02 -7.29887962e-01 -3.10324311e-01
1.42814681e-01 8.62259045e-02 -3.37445378e-01 4.77927588e-02
-1.15442038e+00 3.05840701e-01 6.10343218e-01 6.75102472e-02
-6.94373965e-01 -8.27887118e-01 -6.78053737e-01 -2.90865988e-01
6.23164296e-01 -7.89718449e-01 1.18426895e+00 -1.20688498e+00
-1.50439477e+00 5.77803731e-01 -1.36090413e-01 -2.44663447e-01
7.30550766e-01 -8.16571265e-02 -4.31348026e-01 3.03779453e-01
5.34671210e-02 6.79123759e-01 8.08974504e-01 -1.44931638e+00
-8.17419767e-01 2.26168782e-02 2.45758235e-01 2.30920628e-01
-3.91648322e-01 -9.49271843e-02 -1.33482158e+00 -1.08153808e+00
-4.41222303e-02 -1.14349329e+00 2.91214436e-02 -1.36390954e-01
-6.25667274e-01 2.11438030e-01 9.55932260e-01 -7.12975562e-01
1.74782908e+00 -2.28323412e+00 3.29033852e-01 2.60446072e-01
2.77181119e-01 9.51420292e-02 -2.28111044e-01 3.40494156e-01
-1.75732523e-01 2.97414605e-02 -9.62236077e-02 -1.57805383e-01
-7.40209520e-02 1.52659854e-02 -1.30763352e-01 6.18554533e-01
-1.51491508e-01 7.64510572e-01 -8.57494414e-01 -7.02547610e-01
2.44608358e-01 5.68462491e-01 -9.09142733e-01 2.15208977e-01
2.86242794e-02 6.23056114e-01 -3.59573901e-01 6.44081652e-01
9.56035376e-01 -7.98336491e-02 4.47779924e-01 -6.45108402e-01
-3.16883504e-01 -3.82346034e-01 -1.31387722e+00 1.75613260e+00
-1.25942200e-01 6.75559282e-01 5.38016707e-02 -7.17621386e-01
5.75907648e-01 2.53757536e-01 9.88512218e-01 -8.05484653e-01
2.14242443e-01 3.76223288e-02 -9.66935679e-02 -2.85779357e-01
6.84242368e-01 2.84982443e-01 9.38438103e-02 3.67265910e-01
-3.77038032e-01 -1.34243086e-01 5.79766035e-01 1.94582462e-01
8.09976578e-01 3.62222463e-01 2.49519080e-01 -1.00824118e-01
5.87179065e-01 3.57369706e-02 9.80857015e-01 3.68322343e-01
-3.89720976e-01 1.04471791e+00 4.11575228e-01 -3.03679556e-01
-1.18467152e+00 -1.17834198e+00 -7.33540766e-03 7.99167395e-01
8.31698954e-01 -7.03631699e-01 -1.06847370e+00 -1.11170039e-01
-2.49046862e-01 2.77856439e-01 -5.75251758e-01 -2.25383081e-02
-8.38961601e-01 -6.19286418e-01 4.40568715e-01 1.27488166e-01
9.15909827e-01 -7.10613489e-01 -5.68412006e-01 4.74156775e-02
-5.88699043e-01 -1.20703936e+00 -1.03728497e+00 -7.58630037e-01
-3.37537199e-01 -1.44016290e+00 -7.64357269e-01 -8.87308359e-01
6.24523818e-01 7.45964646e-01 9.75056350e-01 3.26514512e-01
-2.66567051e-01 4.26431268e-01 -5.57159841e-01 1.38681959e-02
-2.90108085e-01 -8.42014551e-02 4.73152241e-03 1.13797627e-01
-3.57269906e-02 -5.43463767e-01 -9.33969200e-01 6.89091563e-01
-1.16238546e+00 6.02113843e-01 9.07427296e-02 5.79167247e-01
8.82385552e-01 3.52028847e-01 1.89564958e-01 -6.06432915e-01
5.54342985e-01 -4.31451946e-01 -6.47703707e-01 3.11609149e-01
-1.92756042e-01 -5.72831035e-01 5.26280463e-01 -4.60372120e-01
-1.00065017e+00 -1.29631817e-01 2.79496461e-01 -8.20497870e-01
2.51846075e-01 3.21238965e-01 -2.38474637e-01 -1.48022063e-02
8.64454731e-02 1.44051373e-01 -1.56273320e-01 -1.48187563e-01
4.25505966e-01 2.93132544e-01 9.11631942e-01 -5.46603262e-01
9.76498842e-01 6.83602035e-01 -3.07227135e-01 -5.57065487e-01
-2.99010873e-01 -2.68359423e-01 -4.59971428e-01 -4.76069868e-01
1.04674399e+00 -1.03577113e+00 -6.00263953e-01 5.45788109e-01
-8.71253312e-01 -4.52863365e-01 -9.70312431e-02 4.44111675e-01
-5.60904980e-01 5.54428041e-01 -4.65911955e-01 -2.18015194e-01
-2.00391471e-01 -1.08818185e+00 8.11973453e-01 3.16712618e-01
-9.00476649e-02 -7.08691239e-01 3.35888624e-01 3.63982528e-01
1.36278540e-01 4.56395507e-01 4.81259197e-01 1.58707947e-01
-7.11878538e-01 2.39426956e-01 -7.60381669e-02 6.73182905e-02
3.88041496e-01 6.11416280e-01 -4.80256706e-01 -3.51355910e-01
-2.26041660e-01 2.19251379e-01 6.04476571e-01 5.33736169e-01
1.25241506e+00 -1.75594285e-01 1.03705171e-02 8.00682545e-01
1.30830050e+00 3.77735823e-01 9.16784823e-01 6.48352325e-01
8.28787386e-01 3.79017591e-01 8.75952482e-01 8.46960545e-01
5.17317772e-01 1.19119692e+00 1.85257792e-01 -2.37182796e-01
-3.64669681e-01 -2.04211652e-01 2.60998011e-01 1.02027655e+00
-2.56975472e-01 -4.83593136e-01 -6.84108078e-01 4.09943879e-01
-2.07974410e+00 -1.21514285e+00 1.45047039e-01 2.05285907e+00
6.97781205e-01 -2.15329736e-01 3.42761934e-01 3.30068208e-02
1.16752422e+00 1.89045921e-01 -3.70840847e-01 2.15618923e-01
-4.43122149e-01 -1.25212625e-01 3.40335548e-01 5.01596749e-01
-1.10973454e+00 8.33114088e-01 5.46670151e+00 9.11572635e-01
-1.24721515e+00 3.25644791e-01 6.87638044e-01 -5.79977036e-01
-4.55828279e-01 -1.22469719e-02 -2.29838252e-01 8.61059189e-01
3.06815386e-01 -5.92855215e-01 5.29368222e-01 3.45272660e-01
7.57053137e-01 -5.18501066e-02 -6.40113711e-01 1.24311054e+00
1.80009142e-01 -1.50689268e+00 6.17450196e-03 -2.57227093e-01
1.08610761e+00 -3.35875928e-01 1.65249392e-01 -1.51676059e-01
6.26395717e-02 -5.01128137e-01 1.01130188e+00 3.86964053e-01
7.25501478e-01 -7.55201280e-01 3.97801340e-01 -4.11230683e-01
-1.60783923e+00 1.59687147e-01 -1.39673308e-01 3.03050131e-01
3.76590550e-01 3.05181623e-01 -3.38509023e-01 7.13890910e-01
9.39333379e-01 8.82293880e-01 -6.66763365e-01 1.24473941e+00
-1.18478641e-01 4.25376624e-01 -7.11898580e-02 5.75229049e-01
-7.81738982e-02 -6.12563133e-01 5.26984334e-01 9.20418859e-01
6.03336811e-01 1.60120934e-01 3.17096651e-01 5.63361645e-01
-1.08163349e-01 3.74393433e-01 -3.38617980e-01 2.98280753e-02
7.62021959e-01 1.12785876e+00 -7.67645776e-01 -1.93264887e-01
-5.73626995e-01 1.07461655e+00 -1.30610228e-01 5.56339502e-01
-1.36573553e+00 3.18362415e-02 7.51146495e-01 3.05411220e-02
3.60021561e-01 -2.28887379e-01 -6.47728890e-02 -1.39434564e+00
1.32703483e-01 -1.07201600e+00 4.53332961e-01 -8.33844185e-01
-8.62401068e-01 5.10296166e-01 8.80829990e-02 -1.65287983e+00
-9.56352428e-02 -3.84393870e-03 -7.79096007e-01 2.90134013e-01
-1.19235051e+00 -1.15542364e+00 -8.58925700e-01 9.49759364e-01
5.49222171e-01 -8.01684633e-02 1.91643998e-01 5.95465124e-01
-9.68623519e-01 5.69697380e-01 2.00323984e-01 4.76908498e-02
9.74665761e-01 -6.57695949e-01 1.11145802e-01 1.13611376e+00
4.32317555e-02 5.12644172e-01 8.43912840e-01 -6.68167770e-01
-1.45480835e+00 -9.49750662e-01 2.37008780e-01 -5.78689054e-02
6.67440593e-01 -1.85348615e-01 -8.29538584e-01 5.13613105e-01
5.41770458e-01 4.98770103e-02 6.96913719e-01 -5.53421259e-01
-2.19550639e-01 -2.02352688e-01 -7.95061290e-01 9.46098864e-01
1.24784780e+00 -3.64500970e-01 -4.98277992e-02 1.22478269e-01
7.51032829e-01 -5.58018923e-01 -9.45917130e-01 3.62392962e-01
7.46596634e-01 -1.22030675e+00 9.19153929e-01 -4.90257107e-02
4.83558476e-01 -9.35884476e-01 -3.02072406e-01 -1.12157369e+00
-2.26875618e-01 -7.99474359e-01 1.13335922e-01 1.29443669e+00
1.89444777e-02 -3.86929542e-01 6.89991951e-01 3.71642649e-01
-1.42554075e-01 -5.95187664e-01 -8.37930143e-01 -6.87169254e-01
-3.58746022e-01 -2.58450180e-01 6.83876574e-01 1.08578515e+00
-1.49873346e-01 -1.92003652e-01 -6.08044267e-01 6.72976896e-02
4.45390373e-01 4.23502028e-01 1.14565587e+00 -4.88217115e-01
-4.36469406e-01 -5.03013432e-01 -4.40893561e-01 -7.08874822e-01
2.59719074e-01 -4.63018119e-01 -2.03942910e-01 -1.31670761e+00
4.09966946e-01 -3.62584561e-01 -3.47941577e-01 2.55666018e-01
-3.20239663e-01 8.76402795e-01 5.41667819e-01 3.93639565e-01
-8.64098787e-01 7.05290377e-01 1.39987159e+00 -8.54284391e-02
-2.78305024e-01 -5.18051684e-01 -6.37010217e-01 5.27155757e-01
9.45650876e-01 -8.12881887e-02 -5.46668530e-01 -5.27517855e-01
-6.63383082e-02 -4.85864580e-02 2.52759308e-01 -9.50979948e-01
2.27629155e-01 -3.29089582e-01 2.07618326e-01 -4.01520282e-01
2.35641479e-01 -7.78699994e-01 9.63217676e-01 4.03852582e-01
-3.89408283e-02 4.78137195e-01 1.95862025e-01 4.41112012e-01
-4.87156332e-01 1.54013619e-01 8.69594634e-01 1.92808196e-01
-1.07055593e+00 5.06475449e-01 -1.55764699e-01 1.71797529e-01
1.40735579e+00 -3.28705013e-01 -3.73684704e-01 -5.26880503e-01
-4.27623808e-01 2.46561483e-01 1.01155508e+00 5.96712589e-01
6.10691547e-01 -1.58763659e+00 -7.73562789e-01 1.20013878e-01
2.24669129e-02 -2.77737111e-01 5.24274290e-01 1.04285049e+00
-9.24574792e-01 -2.52834260e-02 -5.27986705e-01 -3.70381296e-01
-1.58348656e+00 6.05920672e-01 1.99105114e-01 3.21609855e-01
-6.02339804e-01 3.94445032e-01 5.70456028e-01 9.58632827e-02
-1.96300987e-02 -4.09227272e-04 -7.53906593e-02 -1.28296306e-02
5.21500289e-01 5.15291393e-01 -2.98022360e-01 -1.03456676e+00
-2.64111876e-01 7.15332150e-01 2.44368628e-01 -1.05889514e-01
1.01467037e+00 -5.52667141e-01 -2.38753065e-01 3.75830457e-02
1.11619151e+00 3.96977425e-01 -1.30166352e+00 -3.43621001e-02
-3.48786265e-01 -9.66704071e-01 -1.64869696e-01 -2.62722913e-02
-1.48748851e+00 1.60941467e-01 5.64508617e-01 9.62878540e-02
1.52589655e+00 -2.70914704e-01 1.03550828e+00 -1.10300735e-01
2.01559961e-01 -1.15955997e+00 2.70248592e-01 1.29156545e-01
8.38413835e-01 -9.49719250e-01 3.94366384e-02 -6.73489094e-01
-5.67330241e-01 8.08315516e-01 7.16525912e-01 -1.23711146e-01
4.62533742e-01 1.99228913e-01 1.90118670e-01 1.38093859e-01
-4.89543974e-01 -4.85079549e-02 4.36256975e-02 3.15247178e-01
4.59656388e-01 -6.44509643e-02 -6.40312195e-01 2.26736590e-01
-2.84081370e-01 4.71979827e-02 6.63989007e-01 8.03491056e-01
-1.11662783e-01 -1.26414502e+00 -5.39506316e-01 -7.71259889e-02
-3.44166607e-01 -4.01695818e-02 -2.21769542e-01 7.39479423e-01
3.09605628e-01 1.02117944e+00 1.59821361e-01 -4.56656277e-01
2.26908222e-01 -5.40091813e-01 4.70824957e-01 -1.88543618e-01
-3.29777718e-01 4.35512125e-01 -1.27279535e-01 -7.49758601e-01
-8.94459605e-01 -5.68360329e-01 -1.23230219e+00 -7.80393898e-01
1.54437497e-01 1.91670090e-01 1.42969728e-01 4.00908381e-01
3.89813155e-01 4.81396407e-01 7.75801241e-01 -9.97718990e-01
3.06902438e-01 -3.51766378e-01 -6.17535233e-01 7.08913982e-01
6.95533305e-02 -6.67069733e-01 -3.82767394e-02 5.14856339e-01] | [11.190349578857422, -1.1719237565994263] |
d5045fe7-80e2-4cd2-8689-7af9a8e80308 | dynamic-named-entity-recognition | 2302.10314 | null | https://arxiv.org/abs/2302.10314v1 | https://arxiv.org/pdf/2302.10314v1.pdf | Dynamic Named Entity Recognition | Named Entity Recognition (NER) is a challenging and widely studied task that involves detecting and typing entities in text. So far,NER still approaches entity typing as a task of classification into universal classes (e.g. date, person, or location). Recent advances innatural language processing focus on architectures of increasing complexity that may lead to overfitting and memorization, and thus, underuse of context. Our work targets situations where the type of entities depends on the context and cannot be solved solely by memorization. We hence introduce a new task: Dynamic Named Entity Recognition (DNER), providing a framework to better evaluate the ability of algorithms to extract entities by exploiting the context. The DNER benchmark is based on two datasets, DNER-RotoWire and DNER-IMDb. We evaluate baseline models and present experiments reflecting issues and research axes related to this novel task. | ['Aurélien Baelde', 'Siwar Jendoubi', 'Vincent Guigue', 'Laure Soulier', 'Tristan Luiggi'] | 2023-02-16 | null | null | null | null | ['memorization', 'entity-typing'] | ['natural-language-processing', 'natural-language-processing'] | [-1.32262230e-01 -8.36181045e-02 -1.00898571e-01 -4.25677299e-01
-5.68935275e-01 -8.34260523e-01 9.22127903e-01 6.89125836e-01
-1.29656661e+00 1.04355335e+00 3.54065835e-01 -3.43575656e-01
2.29121417e-01 -9.00089145e-01 -5.84896743e-01 -1.80358380e-01
4.52449769e-02 5.62267303e-01 2.67095000e-01 -9.87389311e-02
9.32303593e-02 7.40711927e-01 -1.40056634e+00 2.91249424e-01
8.06248128e-01 6.98565602e-01 -9.99359861e-02 6.56005204e-01
-6.12045228e-01 6.88020587e-01 -1.11056316e+00 -4.89716291e-01
-1.07412264e-01 -2.10944843e-02 -1.25662053e+00 -4.68601257e-01
4.04837728e-01 7.94097856e-02 -1.18193202e-01 7.12820172e-01
5.61330378e-01 4.18650836e-01 5.93156815e-01 -8.53038311e-01
-6.73581541e-01 9.79245961e-01 -2.26531290e-02 5.15911937e-01
4.60815966e-01 -3.62567127e-01 1.03624582e+00 -1.00903928e+00
1.10111976e+00 8.15471768e-01 8.59312177e-01 7.18606174e-01
-1.11512411e+00 -5.17540634e-01 2.45907590e-01 2.61604726e-01
-1.60241735e+00 -7.44301081e-01 -1.57555602e-02 -3.59928370e-01
1.37471545e+00 4.02037382e-01 -8.30783248e-02 1.01217401e+00
-4.29621823e-02 7.71797955e-01 8.80820572e-01 -5.58844864e-01
4.60486740e-01 2.30226547e-01 6.97742164e-01 2.47369081e-01
7.26734996e-01 -2.16658801e-01 -5.25197864e-01 -1.37134701e-01
3.13506097e-01 -2.17526034e-01 -1.84567556e-01 8.11461825e-03
-1.29442322e+00 4.22735214e-01 2.22959951e-01 8.56173038e-01
-5.66597760e-01 -2.38281548e-01 5.11255264e-01 1.99288592e-01
2.47806445e-01 9.56372738e-01 -9.30337667e-01 -2.27298826e-01
-9.95674908e-01 1.07463382e-01 1.29842639e+00 8.97005677e-01
7.27080941e-01 -1.58209622e-01 -3.19263935e-01 7.44742572e-01
-1.06498718e-01 4.84117717e-01 7.10727751e-01 -1.03485435e-01
5.05780637e-01 7.07852185e-01 4.96831179e-01 -8.20921600e-01
-7.86021173e-01 -1.00074314e-01 -6.04699254e-01 -2.30383649e-01
4.10440892e-01 -5.99384308e-01 -1.14262891e+00 1.67469203e+00
3.34344029e-01 1.66710407e-01 3.43629479e-01 3.91594946e-01
1.15467763e+00 5.44228971e-01 8.08528006e-01 2.57229488e-02
1.54632032e+00 -6.52268708e-01 -8.21393907e-01 -4.62244958e-01
1.06347907e+00 -5.32817245e-01 4.03935760e-01 -4.57738861e-02
-8.63735795e-01 -2.60957569e-01 -7.23719835e-01 -2.00488389e-01
-1.24355626e+00 2.38560021e-01 5.79666555e-01 8.85695040e-01
-8.74012470e-01 5.70182502e-01 -8.54940176e-01 -6.12938881e-01
-2.57886834e-02 3.13330591e-01 -7.35784531e-01 1.50596321e-01
-1.40198088e+00 1.16326690e+00 9.38498735e-01 1.01033859e-01
-1.89125076e-01 -6.47815943e-01 -1.11418974e+00 1.99977621e-01
3.82771313e-01 -5.24682999e-01 1.19915199e+00 -6.98327303e-01
-9.15863514e-01 1.04410100e+00 -4.35285121e-01 -7.06300557e-01
1.68507248e-01 -5.28177917e-01 -1.01985037e+00 -1.98136538e-01
1.61744013e-01 2.62117654e-01 2.68666208e-01 -9.14474726e-01
-8.36071193e-01 -3.52529287e-01 1.19503863e-01 -3.09961904e-02
-4.34859812e-01 4.15338635e-01 -1.60853773e-01 -6.88784957e-01
-9.95658934e-02 -8.02958131e-01 -2.43278235e-01 -7.32494235e-01
-6.99344337e-01 -4.80834275e-01 3.50794077e-01 -7.49942660e-01
1.68950772e+00 -1.97089756e+00 -8.22395757e-02 2.28296407e-02
2.33112350e-01 5.52994251e-01 8.59904066e-02 5.00308990e-01
-3.55980814e-01 6.30749166e-01 1.96941197e-02 -3.40934515e-01
5.43341860e-02 8.97144247e-03 -3.00228655e-01 7.99096599e-02
2.23557279e-01 8.72993112e-01 -8.23867381e-01 -4.40793693e-01
-2.62786627e-01 4.05601472e-01 -1.97795048e-01 2.31942952e-01
-7.26246682e-04 -1.27602518e-02 -3.45546186e-01 5.66395521e-01
4.88414884e-01 -2.68315166e-01 2.48643190e-01 -8.35879222e-02
-4.48337197e-01 6.44726276e-01 -1.57177806e+00 1.21775699e+00
-6.50793850e-01 6.84822083e-01 -1.43083677e-01 -4.57345098e-01
7.22483099e-01 4.68009919e-01 1.58179864e-01 -4.11986262e-01
-5.98157384e-02 2.35947713e-01 -3.20432246e-01 -4.19369370e-01
1.10309494e+00 1.32469833e-01 -3.97479832e-01 3.95317793e-01
3.07876050e-01 6.72750413e-01 5.49877405e-01 -1.43682072e-02
1.33488476e+00 -1.27756536e-01 6.81076109e-01 -1.40501812e-01
4.23379093e-01 1.90301955e-01 6.52246594e-01 1.03311336e+00
-7.14591444e-02 1.31372060e-03 3.43567110e-03 -4.03587639e-01
-8.24822724e-01 -9.60615575e-01 -3.09571624e-01 1.51529503e+00
-1.16830207e-01 -5.21994293e-01 -5.29700696e-01 -8.91732037e-01
-7.53124654e-02 9.96237457e-01 -6.82516217e-01 5.51677048e-02
-8.39661419e-01 -7.86403954e-01 9.28451180e-01 8.63405704e-01
4.16289657e-01 -1.27238691e+00 -5.32123446e-01 4.25146461e-01
-2.58729547e-01 -1.30881035e+00 -3.71999741e-01 3.06508750e-01
-7.09798992e-01 -7.82039762e-01 -7.26606190e-01 -8.35551083e-01
5.33395588e-01 -6.95936605e-02 1.63018703e+00 -3.64586003e-02
-1.93092003e-01 5.06967306e-01 -4.07361418e-01 -6.42376542e-01
-2.71687418e-01 7.41396487e-01 2.25661639e-02 -1.92351952e-01
7.73061931e-01 -2.39694610e-01 -2.79867053e-01 3.53591919e-01
-9.28388119e-01 -3.52844208e-01 6.58038735e-01 7.57044852e-01
3.14803928e-01 -1.30340800e-01 5.70286870e-01 -1.36308300e+00
4.83907282e-01 -7.23370135e-01 -2.85151184e-01 7.35357821e-01
-5.69036186e-01 3.32524836e-01 5.16217768e-01 -4.48278427e-01
-1.46442497e+00 -3.05771809e-02 -2.26964667e-01 2.34412238e-01
-6.56847954e-01 7.50136495e-01 -3.14094216e-01 2.25390568e-01
7.98669577e-01 1.49712920e-01 -1.08913577e+00 -7.03605831e-01
4.62302625e-01 7.42258668e-01 5.47344983e-01 -6.85848653e-01
4.50604886e-01 2.84291416e-01 -4.55836475e-01 -9.77989674e-01
-9.08667982e-01 -8.06707919e-01 -7.84098923e-01 1.99949235e-01
8.83191168e-01 -9.27665651e-01 -3.86358976e-01 3.84308457e-01
-1.33922935e+00 -1.39630279e-02 -3.16667259e-01 4.71245229e-01
1.41176209e-01 1.41419202e-01 -8.21596622e-01 -7.91875660e-01
-5.39603233e-01 -3.92235190e-01 8.40716124e-01 5.46984434e-01
-4.32721376e-01 -1.21013939e+00 2.57322311e-01 -2.88417041e-02
4.88490671e-01 1.60379350e-01 7.85173714e-01 -1.51560712e+00
-2.08729610e-01 -4.27532464e-01 -4.27914783e-02 -1.06082559e-01
-1.78120568e-01 -1.57099023e-01 -9.68675196e-01 -6.26669079e-02
-3.40022832e-01 -1.67172298e-01 8.25791001e-01 -2.97687709e-01
6.72337592e-01 -3.56783777e-01 -6.19462013e-01 4.41196918e-01
1.46210074e+00 2.61240870e-01 6.90190911e-01 7.71407127e-01
6.71084344e-01 5.27704835e-01 3.89522731e-01 4.04526234e-01
6.35492802e-01 4.96481359e-01 -2.24441782e-01 -8.73940215e-02
1.65277570e-01 -2.53560841e-01 1.26618862e-01 5.22945523e-01
-2.52628326e-01 -5.20459950e-01 -1.35724437e+00 7.67521203e-01
-1.55167818e+00 -1.08573699e+00 -1.24944439e-02 2.14877582e+00
9.77956712e-01 3.06101870e-02 -1.57198191e-01 -2.78634906e-01
9.92685139e-01 2.72617340e-02 -4.11517113e-01 -2.78489769e-01
-3.92824292e-01 4.84865010e-01 6.91840231e-01 7.88423195e-02
-1.41379225e+00 1.02904677e+00 6.40418291e+00 6.49855912e-01
-1.13304222e+00 -7.06876591e-02 3.36839616e-01 4.53622729e-01
6.63231462e-02 -1.01589933e-01 -1.55885625e+00 2.80764908e-01
1.33751738e+00 -5.08670628e-01 5.78520857e-02 9.25496042e-01
-2.84897745e-01 7.01621873e-03 -1.22869921e+00 7.24369109e-01
1.34318352e-01 -1.14812768e+00 9.08180997e-02 -2.85622686e-01
4.50409949e-01 2.20133871e-01 -4.36481625e-01 7.35866666e-01
5.41069388e-01 -9.37714100e-01 6.05412066e-01 6.47662699e-01
5.73626339e-01 -4.83612776e-01 9.33909535e-01 2.63909250e-01
-1.26102245e+00 1.97940115e-02 -2.34426975e-01 2.90209442e-01
2.10995376e-01 3.49563062e-01 -9.20758247e-01 6.09629512e-01
8.11972797e-01 4.03784394e-01 -7.79935718e-01 1.33026206e+00
-2.88695961e-01 6.22255504e-01 -4.14013535e-01 -1.38396189e-01
-1.48765415e-01 3.78241092e-01 4.88897264e-01 1.74261773e+00
2.29697824e-01 6.77536055e-02 6.36267290e-02 4.22168612e-01
-4.74388927e-01 5.17711282e-01 -3.96679193e-01 -2.69499838e-01
7.05664933e-01 1.25295818e+00 -7.86032259e-01 -5.42101800e-01
-3.38603169e-01 9.23723698e-01 6.20490909e-01 3.15157205e-01
-5.30640662e-01 -9.06110406e-01 4.53149021e-01 -7.42984191e-02
5.08741081e-01 -4.71839875e-01 -3.63683343e-01 -1.53734505e+00
-1.82252601e-02 -5.12988865e-01 8.92443895e-01 -4.22966629e-01
-1.22674656e+00 9.34061706e-01 -1.24470599e-01 -7.20698714e-01
-3.20413381e-01 -6.67872787e-01 -5.34479380e-01 7.32732594e-01
-1.53163075e+00 -8.73909473e-01 -2.50301391e-01 4.25191462e-01
2.51110911e-01 1.54017538e-01 1.23618233e+00 6.48444176e-01
-8.50840867e-01 8.50397170e-01 2.38459051e-01 8.11486542e-01
1.02420747e+00 -1.46552896e+00 6.01033866e-01 9.71991956e-01
3.17376077e-01 1.06774342e+00 6.37561679e-01 -7.16074407e-01
-8.99117291e-01 -1.02481949e+00 1.81280065e+00 -9.27143574e-01
7.89602935e-01 -4.37118560e-01 -1.03946781e+00 8.53396058e-01
1.10921197e-01 -6.48129880e-02 1.11352122e+00 5.72691023e-01
-6.16964698e-01 3.63561094e-01 -1.17330289e+00 5.37727475e-01
1.05088758e+00 -5.30245006e-01 -1.01208603e+00 3.34044732e-02
6.52435660e-01 -3.48914623e-01 -1.17790234e+00 1.99479863e-01
2.26209730e-01 -5.35488307e-01 7.81944633e-01 -1.13389802e+00
3.17069376e-03 -9.62103754e-02 1.83059275e-01 -1.26478708e+00
-5.41861244e-02 -4.35159266e-01 -2.41147548e-01 1.72731864e+00
8.81888032e-01 -6.42084956e-01 6.48519695e-01 1.16251969e+00
1.64490700e-01 -2.63146967e-01 -8.53468418e-01 -8.53440344e-01
1.80907071e-01 -2.04013720e-01 8.59461844e-01 1.39731932e+00
1.28532335e-01 6.00935817e-01 -2.07879879e-02 2.22126976e-01
1.03068464e-01 -4.50756922e-02 4.68411356e-01 -1.32017958e+00
3.35053802e-02 -2.24270478e-01 -6.09265208e-01 -8.80939603e-01
2.63063341e-01 -7.39125967e-01 1.19148813e-01 -1.47668850e+00
-7.33504519e-02 -6.17346227e-01 -4.95929092e-01 8.01213026e-01
-3.77041221e-01 -1.34132430e-01 1.24161988e-01 3.04407150e-01
-1.05098176e+00 1.16377190e-01 2.77163357e-01 4.38469090e-02
-3.09497535e-01 -1.31101564e-01 -4.30576473e-01 5.72818398e-01
6.61094785e-01 -6.57336414e-01 3.13625485e-01 -5.64397156e-01
4.73965466e-01 -6.78356737e-02 -1.62509494e-02 -9.83898997e-01
6.26323283e-01 -4.67584059e-02 4.87257153e-01 -4.16636050e-01
-8.90316442e-02 -7.10754633e-01 -2.09452435e-02 3.88991646e-02
-4.65509146e-01 4.22287405e-01 2.76316136e-01 4.52398777e-01
-2.76492953e-01 -4.22524124e-01 3.46637189e-01 -1.14144593e-01
-1.26105034e+00 1.06662348e-01 -4.70776230e-01 4.71218646e-01
9.41536248e-01 9.09001157e-02 -4.37705278e-01 1.42241821e-01
-1.06730115e+00 1.58219904e-01 3.42597544e-01 5.33902943e-01
2.25399181e-01 -1.10504782e+00 -7.04917252e-01 -1.89893454e-01
3.35488468e-01 -2.95579433e-01 2.48070866e-01 2.63198733e-01
-3.96543711e-01 5.67899406e-01 -2.27119271e-02 -1.09003060e-01
-1.30140793e+00 5.36220551e-01 3.83785218e-01 -7.36832738e-01
-3.05599988e-01 7.08246648e-01 -2.02197731e-01 -5.72431207e-01
1.81259826e-01 -1.59372136e-01 -8.67123425e-01 3.84227574e-01
8.04972112e-01 4.91130352e-01 4.92137164e-01 -7.82392800e-01
-4.67602640e-01 -5.05547673e-02 -3.42388928e-01 9.81677994e-02
1.26778054e+00 -6.10685833e-02 -1.47302896e-01 5.31510949e-01
8.74977231e-01 2.36060470e-01 -3.80568475e-01 -4.77744162e-01
1.08472729e+00 -4.73548509e-02 -3.05112183e-01 -9.48094308e-01
-5.34936786e-01 4.71011847e-01 4.81251091e-01 2.83920616e-01
8.02749991e-01 -1.41247138e-01 7.68254578e-01 1.13430393e+00
4.21718717e-01 -1.29865384e+00 -6.23563886e-01 1.07469130e+00
3.62633049e-01 -1.17462778e+00 -2.05933928e-01 -2.44610101e-01
-4.05620933e-01 1.12991929e+00 6.88837171e-01 2.45684922e-01
8.63903046e-01 2.58607149e-01 1.91232339e-01 -1.73986219e-02
-5.72590351e-01 -4.49176908e-01 4.49128836e-01 4.31310952e-01
8.38967681e-01 5.56697622e-02 -4.76600736e-01 9.34597611e-01
-1.58122316e-01 -8.70365128e-02 3.97693962e-01 1.41822839e+00
-3.60423416e-01 -1.11993313e+00 -2.81906545e-01 4.11324233e-01
-1.02273953e+00 -4.18988615e-01 -4.99266982e-01 8.12353551e-01
1.41278028e-01 8.99804413e-01 8.53787288e-02 -2.67259795e-02
6.98431075e-01 3.74958694e-01 -4.32234108e-02 -9.70778346e-01
-1.01575565e+00 -6.93291306e-01 5.83912373e-01 -2.55977750e-01
-3.29040557e-01 -6.00555062e-01 -1.37450778e+00 -1.84930101e-01
-6.05086446e-01 6.71020567e-01 7.71353900e-01 9.70034778e-01
6.40052378e-01 1.98066711e-01 1.66013762e-01 -3.45357865e-01
-3.09168369e-01 -7.86632955e-01 -4.51697588e-01 6.40513957e-01
5.55953234e-02 -4.75292981e-01 -2.29272321e-01 6.29169419e-02] | [9.62482738494873, 9.487841606140137] |
a76f4ca3-8255-4747-a17a-64fd6c3224cf | extractive-summarization-as-text-matching | 2004.08795 | null | https://arxiv.org/abs/2004.08795v1 | https://arxiv.org/pdf/2004.08795v1.pdf | Extractive Summarization as Text Matching | This paper creates a paradigm shift with regard to the way we build neural extractive summarization systems. Instead of following the commonly used framework of extracting sentences individually and modeling the relationship between sentences, we formulate the extractive summarization task as a semantic text matching problem, in which a source document and candidate summaries will be (extracted from the original text) matched in a semantic space. Notably, this paradigm shift to semantic matching framework is well-grounded in our comprehensive analysis of the inherent gap between sentence-level and summary-level extractors based on the property of the dataset. Besides, even instantiating the framework with a simple form of a matching model, we have driven the state-of-the-art extractive result on CNN/DailyMail to a new level (44.41 in ROUGE-1). Experiments on the other five datasets also show the effectiveness of the matching framework. We believe the power of this matching-based summarization framework has not been fully exploited. To encourage more instantiations in the future, we have released our codes, processed dataset, as well as generated summaries in https://github.com/maszhongming/MatchSum. | ['Xuanjing Huang', 'PengFei Liu', 'Yiran Chen', 'Xipeng Qiu', 'Danqing Wang', 'Ming Zhong'] | 2020-04-19 | extractive-summarization-as-text-matching-1 | https://aclanthology.org/2020.acl-main.552 | https://aclanthology.org/2020.acl-main.552.pdf | acl-2020-6 | ['extractive-document-summarization'] | ['natural-language-processing'] | [ 5.12216508e-01 5.00254154e-01 -1.65940911e-01 -4.17513371e-01
-8.80266905e-01 -5.11596739e-01 7.98502445e-01 4.47381616e-01
-3.95789534e-01 6.44298851e-01 1.23596036e+00 1.21857770e-01
-1.38091102e-01 -8.40952814e-01 -6.53676808e-01 -1.93366036e-01
2.61541396e-01 1.76317573e-01 -1.08696438e-01 -4.86903191e-01
6.78514481e-01 7.17834756e-02 -1.55174470e+00 5.69077790e-01
1.10095406e+00 7.13840842e-01 1.64687410e-01 6.79616511e-01
-3.50700051e-01 6.04365885e-01 -8.41729641e-01 -7.70301998e-01
9.10365302e-03 -7.34792113e-01 -1.13044906e+00 -6.68552071e-02
7.62769938e-01 -2.53514677e-01 -6.08154178e-01 1.14074934e+00
5.56221604e-01 2.26930320e-01 6.32702589e-01 -9.17833447e-01
-9.02141631e-01 1.22638977e+00 -1.49500787e-01 8.15709010e-02
4.34093922e-01 -1.32950008e-01 1.48763812e+00 -7.36656010e-01
8.79592419e-01 1.07277334e+00 5.29850066e-01 7.14691579e-01
-7.79349566e-01 -3.61854792e-01 1.59054026e-01 1.22581527e-01
-9.02158260e-01 -7.34124422e-01 8.53687584e-01 -7.66362296e-03
1.25119579e+00 6.77443087e-01 6.14090741e-01 1.26371801e+00
1.13680139e-01 1.18099165e+00 5.05465984e-01 -2.92694271e-01
9.31168571e-02 -1.76871434e-01 6.04075134e-01 5.07194161e-01
5.74444592e-01 -3.76934081e-01 -8.52079153e-01 9.68820080e-02
-1.84672810e-02 -1.15779117e-01 -3.62298459e-01 5.29525541e-02
-1.29441321e+00 8.50713134e-01 3.94347906e-01 5.55910110e-01
-4.51115280e-01 1.89902544e-01 8.52112830e-01 1.75039366e-01
7.47141242e-01 8.75613630e-01 -1.61140949e-01 -1.39751405e-01
-1.52605295e+00 5.39263427e-01 9.65439439e-01 1.14973712e+00
4.93698239e-01 1.44372080e-02 -7.10537553e-01 6.76719308e-01
-4.43954356e-02 3.67312610e-01 8.12732577e-01 -8.86232734e-01
8.24845314e-01 6.24760032e-01 -2.20933050e-01 -1.05919886e+00
-3.13240051e-01 -6.63411438e-01 -8.02798569e-01 -5.50885677e-01
-9.46099460e-02 -1.07425608e-01 -6.35143161e-01 1.57545662e+00
-2.27687076e-01 4.87701967e-02 4.54883754e-01 6.71870291e-01
1.34121644e+00 6.02520525e-01 -2.40536615e-01 -1.60928026e-01
1.36531031e+00 -1.20243573e+00 -8.77088428e-01 -3.81493598e-01
6.91950142e-01 -6.38987362e-01 1.02091122e+00 1.37928665e-01
-1.32065344e+00 -3.89500260e-01 -1.36239696e+00 -4.79542702e-01
-4.87205386e-01 1.60198942e-01 5.41669905e-01 2.34765127e-01
-1.15675890e+00 1.00908887e+00 -6.32449090e-01 -8.32364023e-01
4.03105974e-01 -1.26760527e-01 -1.81205124e-01 2.17392057e-01
-1.22065675e+00 8.62397611e-01 8.15573573e-01 -1.09217159e-01
-3.67438078e-01 -7.13085949e-01 -8.30753446e-01 3.79885763e-01
2.72484213e-01 -1.35798168e+00 1.42315841e+00 -7.43228734e-01
-1.28999698e+00 8.26515377e-01 -3.83227646e-01 -7.93850124e-01
3.87857735e-01 -4.50462639e-01 -3.33263904e-01 3.28674942e-01
3.48031014e-01 6.27306104e-01 3.97987694e-01 -1.09180629e+00
-6.80697381e-01 -4.54858750e-01 -1.25542328e-01 3.52245927e-01
-6.70572162e-01 1.83445826e-01 -2.24877000e-01 -8.69530320e-01
-1.03198506e-01 -5.07737160e-01 -8.51646438e-03 -7.61055231e-01
-9.50277269e-01 -3.68509084e-01 3.93558294e-01 -9.84458745e-01
1.76460600e+00 -1.84794545e+00 3.27444792e-01 -3.09801370e-01
2.63486087e-01 1.70666575e-01 -3.81262779e-01 1.01802647e+00
1.06208347e-01 3.57063830e-01 -5.87151885e-01 -7.75644720e-01
4.02654767e-01 -2.51833677e-01 -9.11215723e-01 -3.00868880e-02
1.65746897e-01 1.24705207e+00 -1.17259133e+00 -5.59960842e-01
-2.38433018e-01 1.35829583e-01 -3.93736959e-01 4.03200984e-02
-1.50936857e-01 -7.46810287e-02 -5.70128202e-01 3.97135198e-01
3.71871293e-01 5.71049154e-02 -1.38049930e-01 -3.40622097e-01
-1.01531491e-01 8.31584096e-01 -6.44055903e-01 2.31616545e+00
-2.16427326e-01 6.52543902e-01 -3.29178363e-01 -9.58256304e-01
9.18068647e-01 2.04105929e-01 4.18355435e-01 -6.16183043e-01
-5.91081791e-02 3.08235675e-01 -3.58831793e-01 -5.29280365e-01
1.55836308e+00 -1.19314175e-02 -4.43602920e-01 4.80340332e-01
3.97814214e-01 -4.48258072e-01 7.69821227e-01 6.42159283e-01
1.15162039e+00 7.02465475e-02 3.33396018e-01 -2.77569473e-01
2.04472870e-01 2.96998322e-01 2.49781489e-01 1.03053033e+00
1.21639542e-01 9.24469769e-01 4.33306426e-01 1.55172367e-02
-1.08713758e+00 -8.69420886e-01 -1.38491457e-02 7.30659902e-01
-1.15702339e-01 -9.58637655e-01 -1.07126582e+00 -7.79042900e-01
-6.46343380e-02 1.22835743e+00 -5.42665124e-01 -3.78051192e-01
-6.91507757e-01 -6.00315154e-01 9.55152094e-01 4.66679484e-01
5.97587824e-01 -1.18639183e+00 -6.43452048e-01 2.05248356e-01
-6.16364777e-01 -1.00216019e+00 -5.84174693e-01 -1.50648415e-01
-1.04234457e+00 -6.78260446e-01 -5.24691820e-01 -5.67505181e-01
2.90874064e-01 2.87077039e-01 1.29359710e+00 7.37746656e-02
1.83672726e-01 4.32815105e-01 -6.05287373e-01 -5.90634406e-01
-6.17195129e-01 7.60118365e-01 -2.43981138e-01 -2.16026157e-01
2.95713097e-01 -6.06010854e-01 -6.92768276e-01 -6.23230815e-01
-1.23853242e+00 3.64346713e-01 5.34510732e-01 4.76949841e-01
1.41627222e-01 -2.01401174e-01 1.07685399e+00 -8.72028768e-01
1.34111845e+00 -5.25570750e-01 2.35834554e-01 3.00598085e-01
-6.13873184e-01 1.25734031e-01 7.65341401e-01 1.75069511e-01
-1.04102898e+00 -3.66320282e-01 -1.69104740e-01 1.41284406e-01
-4.92124148e-02 8.57697010e-01 -7.77897378e-03 8.60514998e-01
5.64957082e-01 5.47633111e-01 -1.63948774e-01 -4.71605331e-01
6.89055741e-01 9.64520514e-01 9.08073366e-01 -3.72332305e-01
6.70220971e-01 4.95538920e-01 -3.03749055e-01 -7.31614292e-01
-1.15684998e+00 -4.49829757e-01 -5.15918434e-01 3.58087234e-02
8.98105979e-01 -7.46279359e-01 -1.84299752e-01 2.98771262e-01
-1.48283958e+00 1.21528067e-01 -7.31601417e-01 1.68721572e-01
-6.14375114e-01 7.62795925e-01 -5.43493390e-01 -3.33828628e-01
-1.12301207e+00 -7.59126902e-01 1.23407984e+00 2.24406064e-01
-7.31519282e-01 -1.04837608e+00 9.82081965e-02 5.15021443e-01
5.25414467e-01 2.54141867e-01 8.07965875e-01 -1.12590325e+00
-2.36945286e-01 -2.90820628e-01 -1.71703905e-01 4.66767341e-01
2.05195069e-01 1.00306176e-01 -1.04629707e+00 -1.48737729e-01
1.66236445e-01 -9.69709307e-02 1.45228410e+00 2.33197272e-01
9.14472699e-01 -6.17055178e-01 -8.90282691e-02 3.05247664e-01
1.21528971e+00 -2.03104109e-01 5.93319535e-01 4.17850614e-01
5.80811918e-01 7.77223349e-01 4.93341357e-01 4.37077194e-01
5.79131305e-01 4.81800437e-01 3.42359424e-01 1.34323552e-01
-5.54348528e-01 -5.93789876e-01 5.19248366e-01 1.34184432e+00
1.84381127e-01 -4.96447206e-01 -5.50480485e-01 6.13100290e-01
-2.05990458e+00 -1.30681515e+00 -6.62195981e-02 1.89357686e+00
1.00808597e+00 2.18799524e-02 3.60668972e-02 -1.26778677e-01
5.70567787e-01 5.51821530e-01 -3.47170889e-01 -6.19737983e-01
-3.95337701e-01 2.22845346e-01 3.01983386e-01 4.22156453e-01
-8.30295622e-01 1.04993260e+00 5.59073544e+00 1.02695787e+00
-1.00456297e+00 -1.29121870e-01 3.25752079e-01 -2.17530057e-01
-7.25508392e-01 1.10330358e-01 -8.53625059e-01 5.71393967e-01
9.99884605e-01 -7.55062282e-01 3.15433800e-01 4.28954393e-01
3.20143551e-01 -3.83875333e-03 -1.35820150e+00 6.58090115e-01
5.48155069e-01 -1.69640803e+00 5.37649751e-01 -2.57389128e-01
6.65556729e-01 1.10544696e-01 -2.26576507e-01 4.89594072e-01
1.07875161e-01 -9.18809295e-01 9.57910955e-01 7.05783069e-01
3.96326303e-01 -5.36500871e-01 6.95477664e-01 3.39725524e-01
-8.63923252e-01 3.63150425e-02 -3.47622514e-01 -4.44530621e-02
2.83007711e-01 6.77803040e-01 -7.75348365e-01 1.20702827e+00
4.41489846e-01 1.10269082e+00 -8.48346472e-01 8.83029759e-01
-3.62375379e-01 5.46202779e-01 9.46419016e-02 -2.11097151e-01
3.04719895e-01 -2.49663115e-01 9.10196960e-01 1.69209170e+00
4.89108056e-01 -2.95688421e-01 -2.75591910e-01 1.02448976e+00
-4.55958307e-01 2.07454860e-01 -6.11193299e-01 -2.37113252e-01
5.54355204e-01 1.31240833e+00 -5.65317869e-01 -7.16244698e-01
-1.86229423e-01 1.12595546e+00 2.85082906e-01 3.66928101e-01
-6.27088666e-01 -7.46673942e-01 3.48499388e-01 -1.56418383e-01
1.37409449e-01 -3.88319939e-02 -6.61051095e-01 -1.50063026e+00
3.76083881e-01 -8.57717276e-01 3.08113158e-01 -8.17624092e-01
-1.13425899e+00 7.13783205e-01 2.00123265e-01 -1.08852851e+00
-2.99046159e-01 -1.60051920e-02 -9.32409644e-01 6.22673094e-01
-1.49268663e+00 -1.11443305e+00 -2.61375189e-01 6.58077970e-02
9.83071089e-01 -9.14738402e-02 8.31467152e-01 5.93703659e-03
-6.05999649e-01 3.86451721e-01 1.15212584e-02 1.31755650e-01
7.41285145e-01 -1.31602085e+00 7.80568302e-01 1.21933079e+00
2.51229495e-01 7.40072429e-01 1.01443064e+00 -6.91019475e-01
-1.28327978e+00 -1.15674782e+00 1.43139625e+00 -5.43404698e-01
8.28160167e-01 -2.43724167e-01 -8.13925147e-01 6.63402081e-01
9.55120981e-01 -8.39438498e-01 5.58102906e-01 -1.89925153e-02
-1.52032599e-01 -2.67450768e-03 -8.25811565e-01 9.03601587e-01
1.35317469e+00 -4.21097040e-01 -1.14489520e+00 3.56879294e-01
1.33772969e+00 -2.33055204e-01 -7.68466473e-01 4.40778673e-01
2.84193546e-01 -9.39432681e-01 7.24183381e-01 -8.36380601e-01
1.10761130e+00 3.78679782e-02 -1.73835099e-01 -1.60701203e+00
-8.50081220e-02 -5.68091631e-01 -2.34089836e-01 1.81439734e+00
3.66507083e-01 -5.75364172e-01 5.91911972e-01 3.82232666e-01
-7.96571076e-01 -8.55280101e-01 -7.57210493e-01 -7.95864880e-01
3.84433955e-01 -3.29226017e-01 8.05137694e-01 7.98299849e-01
3.33688408e-01 7.22055972e-01 1.77323620e-03 -2.81095088e-01
3.99809629e-01 3.54387671e-01 7.06873596e-01 -9.66361761e-01
-4.08201069e-02 -1.03312671e+00 6.82047233e-02 -1.01479423e+00
5.33501983e-01 -1.50186765e+00 -1.51768491e-01 -2.20068741e+00
5.44649601e-01 1.64949119e-01 -1.56175941e-01 2.19868898e-01
-3.30660611e-01 -7.40995705e-02 4.27386582e-01 2.65765667e-01
-7.65632451e-01 8.85866880e-01 1.06863785e+00 -2.40909204e-01
-5.07761426e-02 -9.45415944e-02 -1.43478453e+00 4.88032937e-01
1.01246929e+00 -4.68001068e-01 -3.20079446e-01 -6.72857821e-01
3.75764251e-01 -4.09217812e-02 2.82834888e-01 -8.93606365e-01
3.96114439e-01 9.34406519e-02 -1.82122320e-01 -7.32405484e-01
1.11145534e-01 -2.87297785e-01 -2.22311318e-01 3.50630522e-01
-8.84724855e-01 1.04969874e-01 7.54786283e-02 2.80464321e-01
-4.74192083e-01 -6.57951176e-01 1.72479615e-01 -1.51021123e-01
-5.96970558e-01 -3.00232638e-02 -1.52127117e-01 3.20257753e-01
4.23158199e-01 -3.45028847e-01 -7.91770875e-01 -3.92630011e-01
-3.24080855e-01 2.28849694e-01 3.76574159e-01 5.94561994e-01
6.65137649e-01 -1.12977839e+00 -1.21107793e+00 -2.46395007e-01
2.36040622e-01 -9.58686471e-02 2.70884752e-01 7.24519193e-01
-2.19381079e-01 7.44393289e-01 -7.03303739e-02 -4.75279503e-02
-1.05415392e+00 3.77381206e-01 -7.88340792e-02 -4.87570077e-01
-7.29185998e-01 4.28660780e-01 -8.48577544e-02 -3.77049744e-01
5.00437133e-02 -5.40088236e-01 -2.48152450e-01 4.28114355e-01
4.67938036e-01 6.73468471e-01 3.68336141e-01 -4.58103895e-01
-4.97043207e-02 1.84133232e-01 -1.33726895e-01 -1.87343121e-01
1.46176755e+00 -2.43285805e-01 -4.17768240e-01 2.77280390e-01
1.22506595e+00 2.76942998e-01 -7.75529325e-01 -2.58366197e-01
3.30397606e-01 1.80861019e-02 -1.96883976e-01 -7.54949272e-01
-8.13304961e-01 6.81117177e-01 -2.44069085e-01 4.56136286e-01
1.13716269e+00 6.66532442e-02 1.26974058e+00 4.67178166e-01
-1.88304901e-01 -1.22317576e+00 1.25933900e-01 7.08521485e-01
1.22322249e+00 -8.86178792e-01 1.51604235e-01 -2.06756309e-01
-6.09490335e-01 1.20579898e+00 2.66605526e-01 -2.82792002e-01
-4.33770707e-03 -1.67014990e-02 -3.58338028e-01 -3.51517558e-01
-7.88190544e-01 -1.90628216e-01 4.02949691e-01 3.38983834e-01
6.52307332e-01 4.10064124e-02 -6.44664288e-01 9.55150485e-01
-8.45637381e-01 -7.56799355e-02 9.09502625e-01 7.51531482e-01
-6.77171946e-01 -9.56055641e-01 1.06647022e-01 6.04159355e-01
-5.48481941e-01 -4.52715486e-01 -7.72708237e-01 6.54117823e-01
-4.41202670e-01 1.07578909e+00 -5.15934732e-03 -3.41714919e-01
5.71270406e-01 1.70300037e-01 2.33835101e-01 -9.19295847e-01
-8.73129129e-01 -3.95293564e-01 4.78905618e-01 -3.42954338e-01
-3.62524182e-01 -6.79462254e-01 -1.42882621e+00 -3.05310279e-01
3.98227759e-02 2.84883201e-01 6.95537388e-01 8.56850743e-01
6.85544491e-01 7.41683185e-01 3.66171032e-01 -7.65114069e-01
-9.62586939e-01 -1.21074343e+00 -4.05338734e-01 5.79867899e-01
1.53071702e-01 9.95445475e-02 -2.86751539e-01 1.63569972e-02] | [12.358423233032227, 9.398187637329102] |
ae8281c9-1ef7-43e6-8f9a-6c14f2934698 | hierarchical-multiresolution-feature-and | 2306.02143 | null | https://arxiv.org/abs/2306.02143v1 | https://arxiv.org/pdf/2306.02143v1.pdf | Hierarchical Multiresolution Feature- and Prior-based Graphs for Classification | To incorporate spatial (neighborhood) and bidirectional hierarchical relationships as well as features and priors of the samples into their classification, we formulated the classification problem on three variants of multiresolution neighborhood graphs and the graph of a hierarchical conditional random field. Each of these graphs was weighted and undirected and could thus incorporate the spatial or hierarchical relationships in all directions. In addition, each variant of the proposed neighborhood graphs was composed of a spatial feature-based subgraph and an aspatial prior-based subgraph. It expanded on a random walker graph by using novel mechanisms to derive the edge weights of its spatial feature-based subgraph. These mechanisms included implicit and explicit edge detection to enhance detection of weak boundaries between different classes in spatial domain. The implicit edge detection relied on the outlier detection capability of the Tukey's function and the classification reliabilities of the samples estimated by a hierarchical random forest classifier. Similar mechanism was used to derive the edge weights and thus the energy function of the hierarchical conditional random field. This way, the classification problem boiled down to a system of linear equations and a minimization of the energy function which could be done via fast and efficient techniques. | ['Faezeh Fallah'] | 2023-06-03 | null | null | null | null | ['edge-detection', 'outlier-detection'] | ['computer-vision', 'methodology'] | [ 2.07196310e-01 -2.95970100e-03 2.28804667e-02 -5.08125007e-01
-3.59071761e-01 -4.18667495e-01 7.90372849e-01 2.93388724e-01
-2.55026042e-01 9.45087314e-01 -3.07786446e-02 -2.70600498e-01
-6.95199192e-01 -1.30315495e+00 -5.15877068e-01 -9.56434190e-01
-3.67463440e-01 3.34447920e-01 8.24928403e-01 1.00089014e-01
4.83536988e-01 8.00362051e-01 -1.97353745e+00 3.06485087e-01
8.94453347e-01 8.74135435e-01 2.30095327e-01 7.37697840e-01
7.46173263e-02 5.20671666e-01 -2.64688730e-01 2.44339600e-01
5.31197786e-02 -2.20440716e-01 -3.91406029e-01 1.28268421e-01
4.62752163e-01 1.48584247e-01 -2.12868266e-02 1.01566851e+00
3.14197898e-01 6.09674275e-01 1.08101010e+00 -1.42411876e+00
-7.75485814e-01 2.23700136e-01 -8.29287350e-01 9.72391367e-02
4.01856691e-01 -3.59480798e-01 9.51849103e-01 -1.05443788e+00
6.03599668e-01 1.08992076e+00 7.07184434e-01 -3.16096514e-01
-1.46784258e+00 -3.92506748e-01 3.19045573e-01 5.29487193e-01
-1.68052447e+00 -1.29703224e-01 7.96387851e-01 -6.82604909e-01
6.83734357e-01 5.13054729e-01 9.26188290e-01 6.63210034e-01
2.67876565e-01 1.83791980e-01 1.41610146e+00 -5.92154324e-01
4.10355359e-01 1.10171333e-01 4.94782031e-01 8.41373682e-01
3.93355727e-01 8.91783759e-02 -4.68520701e-01 -5.68390191e-01
5.64783335e-01 -1.80909246e-01 -2.38411590e-01 -7.33876526e-01
-8.31836879e-01 8.58971953e-01 6.15020573e-01 2.84077436e-01
-3.09717923e-01 -1.98402092e-01 -7.50180706e-02 -2.49291345e-01
4.66996521e-01 -1.00349881e-01 -1.47437379e-01 6.81401968e-01
-1.07230163e+00 4.46183532e-02 7.04210162e-01 7.58436561e-01
1.20722270e+00 -1.98646367e-01 2.48162970e-02 4.88345712e-01
5.36117256e-01 3.35370302e-01 1.16546884e-01 -7.76619196e-01
1.53132334e-01 7.90541053e-01 1.39259189e-01 -1.54321957e+00
-5.42393863e-01 -5.61935067e-01 -9.04716551e-01 4.72775519e-01
5.35347700e-01 2.16756314e-01 -9.35279727e-01 1.64031851e+00
7.83804715e-01 2.05991074e-01 -6.17166579e-01 6.26301229e-01
4.78441536e-01 4.71505195e-01 -1.33007228e-01 -3.54874879e-01
1.25605941e+00 -5.32536805e-01 -5.89652419e-01 1.68966979e-01
3.22953939e-01 -4.98343974e-01 6.70550525e-01 2.13624641e-01
-6.98044538e-01 -7.71454155e-01 -1.32551980e+00 2.04802558e-01
-7.34659493e-01 1.41008748e-02 4.67590094e-01 6.00576103e-01
-1.27889872e+00 5.81434906e-01 -6.62351847e-01 -3.54967058e-01
-1.09277397e-01 2.77697682e-01 -4.70504224e-01 8.90808478e-02
-1.04651976e+00 8.09166968e-01 3.20739150e-01 4.78992283e-01
-3.97034198e-01 -2.39867344e-01 -9.19573724e-01 -1.09734677e-01
3.75712886e-02 -6.11910999e-01 1.09443232e-01 -8.71870875e-01
-9.80693996e-01 6.76503599e-01 -3.18606883e-01 1.12542927e-01
5.12433171e-01 2.61478066e-01 -5.44535339e-01 1.36469705e-02
5.75052798e-01 1.91987380e-01 8.02115142e-01 -1.51232505e+00
-7.94541299e-01 -6.95243180e-01 -3.02289248e-01 2.02694893e-01
1.43390641e-01 -4.04503137e-01 -8.95561874e-02 -6.42319083e-01
7.51185060e-01 -7.75021732e-01 -1.56072512e-01 -8.03374127e-02
-3.60178441e-01 -8.10814798e-02 8.44599783e-01 -9.25552726e-01
1.36295402e+00 -2.25861883e+00 7.42209926e-02 9.71273124e-01
3.72906834e-01 -4.77781415e-01 1.08177103e-01 3.85055453e-01
-1.49707884e-01 1.77102879e-01 -4.38806742e-01 4.48780507e-01
-2.70095557e-01 5.21625616e-02 1.63068891e-01 9.75940883e-01
-5.39570823e-02 1.34333476e-01 -1.05973840e+00 -6.36670172e-01
2.80132800e-01 6.21460080e-01 -4.64568645e-01 1.83391361e-03
1.55277133e-01 2.02389657e-01 -4.64475155e-01 3.59446794e-01
1.11457169e+00 1.59229338e-02 2.85422504e-01 -3.50224674e-01
-3.97664845e-01 -8.17228109e-02 -1.73375440e+00 1.20885372e+00
-1.81231797e-01 3.58096004e-01 3.22557420e-01 -1.04298866e+00
1.20046854e+00 -8.09242856e-03 5.86556971e-01 -3.06201339e-01
-3.37576389e-01 6.26640767e-02 -1.97673574e-01 -1.78435996e-01
5.87900281e-01 -9.38432664e-02 2.19360456e-01 1.66310593e-01
-1.13226160e-01 6.03963546e-02 -3.49046253e-02 3.84684233e-03
9.85197484e-01 4.94856179e-01 5.37611842e-01 -7.30193555e-01
6.50869608e-01 -6.92967176e-02 4.46828306e-01 8.74597371e-01
5.00594415e-02 4.60391939e-01 2.56588645e-02 -3.19644958e-01
-6.73652053e-01 -1.53347850e+00 -5.31215847e-01 1.06106341e+00
3.44347954e-01 -3.19833934e-01 -3.95757586e-01 -4.98619258e-01
2.47844197e-02 7.24957347e-01 -8.49043250e-01 2.15877201e-02
-3.23426574e-01 -1.03473246e+00 -6.96191341e-02 2.96998203e-01
5.14032364e-01 -6.01169765e-01 -2.20539317e-01 5.57245500e-02
-9.95691791e-02 -5.52761614e-01 -1.98997736e-01 6.08691156e-01
-7.30679393e-01 -1.26861215e+00 -1.86900020e-01 -7.96154976e-01
7.76755393e-01 2.11886555e-01 7.94932365e-01 -1.17980205e-01
-1.67618692e-01 2.33318582e-01 -3.81339848e-01 7.20873028e-02
1.50569575e-02 -1.23774052e-01 5.52074909e-02 3.36714625e-01
1.36019394e-01 -9.57362771e-01 -4.91369665e-01 6.84758723e-01
-5.76995194e-01 -1.73373997e-01 3.13279539e-01 9.09034610e-01
6.89588368e-01 4.77153212e-01 4.34547991e-01 -7.38443375e-01
3.38004291e-01 -6.76406920e-01 -6.06032789e-01 2.64983773e-01
-4.21061307e-01 1.33367941e-01 4.71027881e-01 -2.53383458e-01
-1.13361430e+00 2.97485560e-01 1.92685530e-01 1.95924729e-01
-1.17921121e-01 4.91787404e-01 -5.16109288e-01 -1.61220819e-01
8.11724424e-01 1.64305538e-01 -3.74803811e-01 -2.51117319e-01
6.00736797e-01 8.29617381e-01 3.43671948e-01 -7.67267108e-01
6.50349498e-01 5.98924398e-01 4.21096951e-01 -1.16811228e+00
-4.46657658e-01 -4.47818786e-01 -1.05797660e+00 -5.10173023e-01
1.06468070e+00 -6.06101096e-01 -3.98804754e-01 2.32520640e-01
-9.50180709e-01 1.19900413e-01 -1.88902631e-01 5.95313013e-01
-3.94719303e-01 4.60444957e-01 -1.25607044e-01 -9.92056370e-01
3.05656195e-01 -7.97162354e-01 8.06810558e-01 1.20660532e-02
-9.53641459e-02 -1.07202387e+00 9.07726213e-02 2.96268370e-02
1.32632032e-01 5.71986914e-01 1.02394795e+00 -4.24935013e-01
-5.58389544e-01 -3.71407390e-01 -3.74029756e-01 3.53925116e-02
2.74633974e-01 3.28994811e-01 -8.17408800e-01 -2.31961861e-01
2.17873320e-01 3.63095194e-01 7.25470901e-01 5.03798366e-01
8.65847945e-01 -2.82030702e-02 -5.95717251e-01 5.63866019e-01
1.63281190e+00 3.61880869e-01 3.65065306e-01 2.98805326e-01
6.87127829e-01 6.52621925e-01 4.38237846e-01 4.52335715e-01
2.39915088e-01 7.66940713e-01 3.22819114e-01 -2.52089407e-02
-7.95542002e-02 -2.57722527e-01 1.09746255e-01 7.19762743e-01
-4.84345019e-01 -1.53134502e-02 -8.39340031e-01 4.95396078e-01
-1.97857153e+00 -1.10070884e+00 -8.00422072e-01 2.44804311e+00
3.10339749e-01 -3.49411853e-02 -3.53524350e-02 3.40735167e-01
1.45533812e+00 1.10600650e-01 -4.02423948e-01 -2.88002908e-01
-2.21913606e-01 -5.85903227e-02 4.85632747e-01 7.92063713e-01
-1.03921342e+00 2.91290760e-01 6.86732769e+00 8.17021012e-01
-4.85070378e-01 -5.16877435e-02 3.97748441e-01 4.96306777e-01
-3.89368266e-01 2.32201979e-01 -8.82349610e-01 3.39558005e-01
5.07501125e-01 2.72998028e-02 3.43317837e-01 7.72530973e-01
2.44919986e-01 -5.75412571e-01 -7.81872153e-01 4.12566811e-01
-1.95709795e-01 -8.11671078e-01 9.09512118e-02 2.32375264e-01
6.35455251e-01 -1.99386463e-01 -3.68803263e-01 -1.00830197e-01
5.85304856e-01 -7.64760554e-01 6.21141613e-01 7.83316433e-01
5.11155784e-01 -4.83671695e-01 6.06277108e-01 4.37974423e-01
-1.85356951e+00 5.26888333e-02 -2.65923053e-01 -4.10917908e-01
-7.67856687e-02 9.06312943e-01 -3.10388178e-01 8.80408227e-01
7.84911156e-01 5.26248991e-01 -6.43042505e-01 8.27905416e-01
-2.36423790e-01 2.96517253e-01 -5.71070015e-01 1.32692680e-01
1.56158963e-02 -7.30541527e-01 5.37332177e-01 1.18886256e+00
4.16718155e-01 -1.59213305e-01 3.49169493e-01 5.60049355e-01
5.22287965e-01 2.32282892e-01 -6.73985004e-01 6.36343122e-01
6.39259636e-01 1.27210343e+00 -1.09077418e+00 -2.38620713e-01
-5.55563211e-01 5.91131091e-01 4.93444979e-01 5.29022157e-01
-8.96883130e-01 -4.01574790e-01 1.65233567e-01 6.63651347e-01
3.54735970e-01 -2.95137346e-01 -4.04573470e-01 -1.01098728e+00
1.22972831e-01 -3.19251984e-01 6.28878176e-01 -8.37168515e-01
-1.54892731e+00 5.26389360e-01 2.94499934e-01 -1.06527817e+00
1.12695344e-01 -5.51850975e-01 -4.54253644e-01 1.09736931e+00
-1.01470435e+00 -1.15815532e+00 -4.84767616e-01 8.42438102e-01
-1.23450376e-01 4.30720225e-02 7.27481306e-01 6.41882047e-02
-2.33498260e-01 9.65432376e-02 3.33415031e-01 -1.40418977e-01
6.07159734e-01 -1.28852010e+00 -3.97000521e-01 9.70883548e-01
-1.23309821e-01 5.13537347e-01 5.74736893e-01 -9.78487253e-01
-9.02941406e-01 -1.02192092e+00 7.73322165e-01 -1.78391457e-01
9.29447889e-01 -2.77918577e-01 -8.18155050e-01 7.25284517e-01
-1.01569995e-01 1.72461510e-01 5.94404697e-01 1.71977550e-01
-1.06823839e-01 -8.35314617e-02 -1.28867662e+00 3.47633660e-01
1.24274814e+00 -4.29620266e-01 -4.74039346e-01 1.82348102e-01
2.49587327e-01 4.83965874e-03 -9.55431163e-01 5.28228939e-01
6.00298345e-01 -1.17003822e+00 9.51731265e-01 -2.87452489e-01
3.19578014e-02 -8.44536066e-01 -5.14808536e-01 -1.09947491e+00
-1.02765834e+00 1.33538842e-01 2.14247882e-01 1.30508602e+00
4.24393415e-01 -8.48820090e-01 7.08262324e-01 2.33632937e-01
-3.54270972e-02 -5.56989074e-01 -1.10907543e+00 -5.93859136e-01
-2.90443182e-01 -2.12356776e-01 2.34569162e-01 9.39088166e-01
-1.08883590e-01 2.34532997e-01 -5.86044043e-02 5.73973238e-01
9.71535861e-01 2.57298231e-01 3.25015634e-01 -1.45642757e+00
-2.65878946e-01 -2.17411622e-01 -7.76006877e-01 -6.24355912e-01
8.24250281e-02 -1.09494138e+00 -1.03249475e-02 -1.70491171e+00
7.52974004e-02 -4.93616313e-01 -2.30054304e-01 1.49358794e-01
-1.90339401e-01 6.55733943e-02 -1.95959806e-01 1.16718933e-01
-1.62050962e-01 3.80343527e-01 9.61469412e-01 -5.49530238e-02
-3.53328884e-01 1.37371063e-01 -2.63325363e-01 8.74919176e-01
5.49771428e-01 -5.24756014e-01 -4.13676143e-01 7.88406879e-02
3.10359061e-01 1.91802278e-01 3.37320179e-01 -1.02793622e+00
4.29719001e-01 -1.42188519e-01 6.33461416e-01 -7.88426578e-01
3.31293344e-01 -1.00847030e+00 6.91722453e-01 1.64427370e-01
1.28392456e-02 8.37366804e-02 -1.86920598e-01 8.72004569e-01
-2.66399801e-01 -1.66899368e-01 7.97544062e-01 -1.76151425e-01
-6.44575059e-01 -1.67926520e-01 -6.08182728e-01 -2.75072992e-01
1.27636075e+00 -6.37737751e-01 -1.60826460e-01 -1.85454577e-01
-1.26771998e+00 -1.35378353e-02 5.38062453e-01 -1.17286794e-01
6.14883244e-01 -1.48299301e+00 -6.21930540e-01 3.15367848e-01
-8.80121812e-02 -6.67942092e-02 5.29182740e-02 7.94258833e-01
-4.40102875e-01 -7.95523524e-02 -4.85541940e-01 -7.14063048e-01
-1.12381983e+00 6.74739957e-01 3.37450236e-01 -1.84652045e-01
-2.99463421e-01 4.77184385e-01 1.81502804e-01 -5.17171383e-01
-4.90815975e-02 -1.92047790e-01 -4.03102934e-01 4.29427058e-01
1.12151511e-01 8.72347951e-01 4.10573184e-02 -9.95816946e-01
-7.19900846e-01 9.14559662e-01 5.36671042e-01 -1.13525473e-01
1.27590024e+00 -3.05539787e-01 -4.61791933e-01 5.41663647e-01
1.00214481e+00 3.03530216e-01 -8.56868267e-01 6.34629205e-02
1.36532217e-01 -4.46599334e-01 1.57182366e-01 -5.06040812e-01
-7.01889098e-01 3.70310247e-01 4.62109715e-01 4.78116035e-01
1.23538280e+00 -1.91228837e-01 -2.23965108e-01 -4.68867645e-02
5.54181695e-01 -9.53737557e-01 -4.02558655e-01 3.34841341e-01
7.36884117e-01 -1.01659107e+00 4.21677500e-01 -8.71311665e-01
-1.23011403e-01 1.23388004e+00 3.87505561e-01 -2.71244228e-01
1.25920928e+00 2.83050120e-01 -4.11478609e-01 -1.36472493e-01
-2.47532070e-01 -2.52184272e-01 4.86288935e-01 8.79626274e-01
3.50428969e-01 3.05961043e-01 -6.28308833e-01 4.18755293e-01
-1.34645745e-01 -7.32505694e-02 2.39345163e-01 7.98738539e-01
-6.35066211e-01 -7.16770709e-01 -6.32390082e-01 5.01687944e-01
2.39473492e-01 5.74792549e-02 -1.10500813e-01 8.40724707e-01
6.22161925e-01 1.06434608e+00 1.66825786e-01 -6.38747334e-01
2.74369121e-01 4.14565653e-02 3.31714869e-01 -4.34486806e-01
-1.87372789e-01 1.46744223e-02 1.69768512e-01 -4.31077629e-01
-6.08855128e-01 -7.43142545e-01 -1.05117679e+00 -1.53919369e-01
-6.22392297e-01 3.36653858e-01 7.51476645e-01 7.72511959e-01
1.91356353e-02 3.14061642e-01 6.47541285e-01 -1.14195967e+00
-2.76136864e-02 -8.59170258e-01 -1.04442906e+00 1.50772884e-01
6.00579008e-03 -8.84204626e-01 -6.97523117e-01 -1.05441272e-01] | [7.621306896209717, 4.567877769470215] |
45532aee-7baa-4496-bd3a-ee2116441b0b | learning-with-noisy-labels-by-efficient-1 | 2111.14932 | null | https://arxiv.org/abs/2111.14932v2 | https://arxiv.org/pdf/2111.14932v2.pdf | Learning with Noisy Labels by Efficient Transition Matrix Estimation to Combat Label Miscorrection | Recent studies on learning with noisy labels have shown remarkable performance by exploiting a small clean dataset. In particular, model agnostic meta-learning-based label correction methods further improve performance by correcting noisy labels on the fly. However, there is no safeguard on the label miscorrection, resulting in unavoidable performance degradation. Moreover, every training step requires at least three back-propagations, significantly slowing down the training speed. To mitigate these issues, we propose a robust and efficient method that learns a label transition matrix on the fly. Employing the transition matrix makes the classifier skeptical about all the corrected samples, which alleviates the miscorrection issue. We also introduce a two-head architecture to efficiently estimate the label transition matrix every iteration within a single back-propagation, so that the estimated matrix closely follows the shifting noise distribution induced by label correction. Extensive experiments demonstrate that our approach shows the best performance in training efficiency while having comparable or better accuracy than existing methods. | ['Buru Chang', 'Joonyoung Yi', 'Kwanghee Choi', 'Seong Min Kye'] | 2021-11-29 | learning-with-noisy-labels-by-efficient | https://openreview.net/forum?id=g1D7SfQKbg | https://openreview.net/pdf?id=g1D7SfQKbg | null | ['learning-with-noisy-labels', 'learning-with-noisy-labels'] | ['computer-vision', 'natural-language-processing'] | [ 3.66090208e-01 1.33031741e-01 -1.90217271e-01 -6.60835564e-01
-1.22465897e+00 -4.39585328e-01 4.09829140e-01 5.71635365e-01
-7.27808833e-01 6.16728723e-01 -5.37395850e-02 -9.76965427e-02
7.85795227e-02 -4.62045729e-01 -7.36540794e-01 -9.30111289e-01
4.91762727e-01 2.30045035e-01 9.83980075e-02 1.05398804e-01
2.59469450e-01 -2.20771786e-02 -1.47683060e+00 2.55404234e-01
7.80760884e-01 8.75741899e-01 3.50841135e-02 4.65690643e-01
-6.83250055e-02 1.15189254e+00 -5.37557602e-01 -6.31646574e-01
1.19237699e-01 -4.35992777e-01 -8.89153361e-01 1.30890042e-01
6.09338999e-01 -5.22858724e-02 -4.36073132e-02 1.50484753e+00
3.53304029e-01 1.77799210e-01 5.53050578e-01 -9.42864299e-01
-3.74263525e-01 9.45615172e-01 -7.30661333e-01 -1.33733511e-01
-1.40972614e-01 -9.73656587e-03 1.02516174e+00 -8.40900540e-01
3.99272829e-01 1.01680183e+00 1.14564228e+00 7.33284056e-01
-1.38085437e+00 -8.43728721e-01 5.01284659e-01 1.15314312e-01
-1.46006858e+00 -3.22912723e-01 5.53460896e-01 -2.59749383e-01
4.00154382e-01 3.97605151e-01 2.28466555e-01 9.89127398e-01
6.43958920e-04 7.66289592e-01 1.32456791e+00 -5.79122305e-01
3.46444696e-01 5.07040858e-01 4.99361455e-01 8.54772747e-01
3.16260457e-01 2.50519104e-02 -4.70268428e-01 -3.38537544e-01
1.38703451e-01 3.74911949e-02 -1.68423429e-01 -2.78964311e-01
-9.89892483e-01 7.43411362e-01 4.96140540e-01 1.90657541e-01
-1.44137487e-01 2.93679684e-01 5.41543245e-01 3.16159755e-01
6.64669752e-01 2.98847497e-01 -5.63290834e-01 8.45992193e-02
-1.11202610e+00 -4.23883647e-02 5.22536159e-01 7.46672332e-01
1.07420361e+00 -2.68016070e-01 -2.48631641e-01 9.55945492e-01
2.73862094e-01 3.37298542e-01 3.26392472e-01 -1.04929876e+00
3.44525248e-01 4.60084260e-01 1.88485831e-01 -9.48351085e-01
-6.01420045e-01 -9.77394462e-01 -9.57300782e-01 1.16530336e-01
5.96229613e-01 -1.08640932e-01 -1.03919375e+00 2.06272960e+00
5.57639718e-01 3.52740407e-01 -2.88108170e-01 6.50384486e-01
3.51399809e-01 3.24564517e-01 3.34878922e-01 -3.15355450e-01
1.37849510e+00 -1.31056643e+00 -6.92260444e-01 -3.83719265e-01
1.16281116e+00 -7.16496527e-01 9.71015871e-01 5.61058164e-01
-6.21688485e-01 -4.15036052e-01 -9.73530352e-01 -4.53382060e-02
-1.27696320e-01 2.08118424e-01 6.34638250e-01 9.23590243e-01
-7.72298574e-01 7.62381256e-01 -8.26895297e-01 -1.55737147e-01
4.70763087e-01 3.60513300e-01 -2.34283134e-01 -2.30504304e-01
-8.37877989e-01 7.65610576e-01 4.59204733e-01 1.99308798e-01
-7.37356842e-01 -7.47681022e-01 -7.09923923e-01 1.11691073e-01
4.90009665e-01 -4.03930426e-01 1.61533844e+00 -9.53900516e-01
-1.37243402e+00 7.59815216e-01 -3.41760933e-01 -3.61283481e-01
8.01203370e-01 -7.70064443e-02 -3.53571087e-01 -3.04214299e-01
1.34815857e-01 5.79235554e-01 9.07911599e-01 -1.59513366e+00
-9.53107953e-01 -1.82321951e-01 -1.65026173e-01 -1.61977205e-02
-5.50134003e-01 -2.28320181e-01 -4.88784641e-01 -6.48000062e-01
4.34085757e-01 -1.19646299e+00 -3.57001692e-01 -2.67996937e-01
-4.48029101e-01 -1.31195813e-01 4.00277466e-01 -3.26171070e-01
1.47701907e+00 -2.16513252e+00 -4.99099106e-01 3.15732569e-01
4.35962886e-01 3.55535835e-01 -1.84933767e-01 1.06906772e-01
-6.20418787e-02 8.17419365e-02 -3.86897683e-01 -6.54554665e-01
-1.06840588e-01 1.68729678e-01 -2.12224171e-01 5.24957657e-01
-1.39401436e-01 6.90023363e-01 -1.25762594e+00 -3.72488648e-01
-1.77012414e-01 3.58808577e-01 -4.99647528e-01 -3.92040126e-02
-1.19663283e-01 5.88713467e-01 -1.40223145e-01 2.71741927e-01
8.66179645e-01 -5.94194651e-01 4.21914339e-01 -1.94297209e-01
2.38044277e-01 3.19208801e-01 -1.31985378e+00 1.51747513e+00
-6.51138365e-01 3.66929263e-01 2.63912473e-02 -8.61088216e-01
6.80669248e-01 2.54168719e-01 3.31492692e-01 -4.93851662e-01
3.77466112e-01 3.25541914e-01 -3.46976876e-01 -2.63582855e-01
3.83683383e-01 -2.13017374e-01 6.47232309e-02 8.68766725e-01
-2.37189122e-02 3.88649344e-01 1.62033364e-01 2.41103277e-01
1.01524031e+00 -1.60338148e-01 -1.96540859e-02 -1.67749710e-02
4.20273632e-01 -1.06193796e-01 8.38353872e-01 1.22102749e+00
-2.13518232e-01 5.10568917e-01 2.29266793e-01 -4.26210731e-01
-7.54243791e-01 -6.01123810e-01 -1.53216133e-02 1.46358907e+00
1.12927414e-01 -6.12289488e-01 -1.03717613e+00 -1.15560317e+00
-6.42424077e-02 6.97410941e-01 -7.63215721e-01 -3.62985522e-01
-4.27141488e-01 -1.12798142e+00 5.10441482e-01 4.16536540e-01
3.65159720e-01 -5.91140509e-01 -4.90073301e-02 2.71840334e-01
-4.53406155e-01 -8.44412088e-01 -5.48853517e-01 5.14551163e-01
-9.93315041e-01 -9.05148685e-01 -4.10779655e-01 -7.77498364e-01
1.11994481e+00 4.98918086e-01 8.55045199e-01 2.81607658e-01
1.65571645e-01 -1.89378709e-01 -2.20037967e-01 -1.79153636e-01
-6.96373045e-01 1.81287006e-01 3.86355817e-02 1.81963146e-01
5.15199900e-01 -3.20211470e-01 -4.09720689e-01 2.69433260e-01
-7.15149105e-01 1.80290426e-05 3.84745240e-01 1.08523786e+00
6.40070796e-01 3.16914409e-01 5.71842015e-01 -1.60456407e+00
3.96992952e-01 -4.45797503e-01 -4.92584974e-01 3.56939316e-01
-1.19824719e+00 3.50940496e-01 6.12409592e-01 -4.40516055e-01
-1.22392714e+00 3.62064719e-01 -1.50726184e-01 -5.47474734e-02
-4.36536446e-02 2.95229971e-01 1.73298672e-01 -1.86559945e-01
1.01005280e+00 4.67591546e-02 -3.00013661e-01 -7.45289147e-01
4.18034732e-01 7.37982273e-01 5.69421828e-01 -3.72602433e-01
7.58781552e-01 5.80112576e-01 -1.09154969e-01 -1.52873978e-01
-1.69405496e+00 -6.39346957e-01 -5.54512918e-01 -1.03210799e-01
2.54070461e-01 -1.11248744e+00 -7.90936053e-01 5.82503200e-01
-8.74799371e-01 -1.85243130e-01 -1.27034098e-01 4.75641787e-01
-1.72846884e-01 3.44122201e-01 -9.07671928e-01 -7.37284064e-01
-2.73714572e-01 -9.97274935e-01 7.35908210e-01 1.33134261e-01
-2.14998394e-01 -9.22338784e-01 4.15754365e-03 4.95360583e-01
2.77749568e-01 -8.98639262e-02 8.96979988e-01 -6.44868433e-01
-4.67561603e-01 -4.57655609e-01 -2.90447414e-01 4.81704175e-01
8.23634788e-02 -2.97334969e-01 -1.47721064e+00 -5.07766545e-01
1.64981303e-03 -3.09472650e-01 1.10352802e+00 5.87979928e-02
1.14240813e+00 -1.44273922e-01 -3.46208841e-01 4.42460835e-01
1.26069856e+00 -2.55293936e-01 9.18379202e-02 2.91560978e-01
8.96488547e-01 3.33034426e-01 7.53478229e-01 3.66472155e-01
4.84728515e-01 3.80681187e-01 2.88057983e-01 -1.77147999e-01
-2.56573230e-01 -3.65109116e-01 1.02040470e-01 1.04837275e+00
3.98984700e-01 -1.08123653e-01 -8.00641537e-01 4.12068546e-01
-2.05022168e+00 -6.49813712e-01 -4.58656728e-01 2.39801002e+00
1.20924938e+00 2.65982896e-01 -2.48738676e-01 3.25844556e-01
7.75533736e-01 -2.88393651e-03 -5.16967952e-01 -1.96512058e-01
1.32399082e-01 -3.09427958e-02 7.43552029e-01 7.24763513e-01
-1.20885086e+00 9.41212893e-01 6.65182066e+00 9.66831088e-01
-1.11936438e+00 4.85160947e-01 6.51799321e-01 -1.24163432e-02
-2.88994759e-01 1.14846915e-01 -9.23128307e-01 4.33219135e-01
1.00415933e+00 1.88594237e-01 3.23859364e-01 9.29135561e-01
6.66418672e-02 -3.61415416e-01 -1.00329530e+00 8.90160501e-01
1.12615190e-02 -1.10851669e+00 -1.90838978e-01 -2.10937545e-01
1.10912776e+00 2.20128745e-01 -5.05534187e-02 3.82693529e-01
6.02651656e-01 -5.31837165e-01 8.00150752e-01 3.93384039e-01
6.49425089e-01 -8.92016709e-01 8.58236194e-01 6.43889546e-01
-8.65869343e-01 -2.43612796e-01 -3.98590624e-01 4.75781523e-02
-1.87357351e-01 1.17457306e+00 -9.32592869e-01 2.31509119e-01
4.35361207e-01 4.01611239e-01 -8.97360802e-01 1.01971388e+00
-5.96985996e-01 9.29464817e-01 -2.73817033e-01 1.97894722e-01
1.65757820e-01 1.23170696e-01 9.35901329e-02 1.31939006e+00
1.11980341e-01 -1.56687379e-01 1.57289371e-01 4.77653682e-01
-5.49817026e-01 1.08921736e-01 -1.15658037e-01 3.07640463e-01
5.84125817e-01 1.36667860e+00 -9.35352623e-01 -5.34432530e-01
-3.14374149e-01 1.04369998e+00 6.13750339e-01 1.97710440e-01
-7.96286881e-01 -2.87333876e-01 3.35036457e-01 -1.55809835e-01
1.11015648e-01 1.70429096e-01 -6.22919977e-01 -1.11377096e+00
-1.21203050e-01 -7.81001508e-01 3.64811629e-01 -3.07437599e-01
-1.15798521e+00 4.07464415e-01 -5.33756793e-01 -1.03660882e+00
-2.08261237e-02 -1.21347100e-01 -1.95993692e-01 6.05675340e-01
-1.59152782e+00 -8.23189735e-01 -3.98204446e-01 2.11841986e-01
1.86383188e-01 2.77668506e-01 8.70813191e-01 6.22408688e-01
-8.16079617e-01 1.15247858e+00 5.07199347e-01 -4.45982218e-02
1.20217705e+00 -1.28235400e+00 3.72226477e-01 8.49041760e-01
3.10789675e-01 5.93792677e-01 6.61425292e-01 -5.83734453e-01
-7.22153842e-01 -1.34946465e+00 1.17185283e+00 -5.46400607e-01
4.58772868e-01 -4.37825829e-01 -1.21579909e+00 6.81501508e-01
-2.67171472e-01 1.90330505e-01 7.93373287e-01 4.26699668e-01
-8.74391556e-01 -1.86412334e-01 -1.08994591e+00 4.38641995e-01
8.71098995e-01 -6.91355884e-01 -1.51863232e-01 6.19715154e-01
4.26653713e-01 -4.25225616e-01 -6.03677034e-01 1.85252368e-01
3.84799331e-01 -8.31232607e-01 5.00322580e-01 -3.50347370e-01
1.38886541e-01 -4.50526088e-01 1.75588727e-01 -1.27261531e+00
-6.01345599e-01 -5.66450834e-01 -1.29103303e-01 1.34185326e+00
6.60298169e-01 -5.53723335e-01 9.10191417e-01 7.34420002e-01
1.14935741e-01 -5.98792076e-01 -6.23260200e-01 -6.38211310e-01
-1.55279040e-01 -6.53272450e-01 5.33986986e-01 1.20970988e+00
-1.33802891e-01 3.14065427e-01 -6.37739301e-01 2.84549803e-01
8.70018065e-01 -2.38305051e-02 5.96950352e-01 -1.27516270e+00
-1.99955046e-01 -1.21792899e-02 7.00402036e-02 -9.07863736e-01
2.41280109e-01 -1.15803969e+00 4.56006080e-01 -1.11941910e+00
5.66778898e-01 -8.44008386e-01 -8.01271081e-01 8.16651165e-01
-6.77672029e-01 7.39759862e-01 4.74959724e-02 3.43118280e-01
-9.64616358e-01 1.21512108e-01 9.08431172e-01 -7.43079409e-02
2.58894265e-02 4.41387929e-02 -9.86501217e-01 8.91204059e-01
7.94125736e-01 -1.27052176e+00 -3.00384611e-01 -5.59492767e-01
4.58437800e-01 -4.57545877e-01 9.17746872e-02 -9.74359453e-01
5.37204802e-01 2.99801938e-02 9.28807110e-02 -2.15250552e-01
2.52614822e-02 -7.78713405e-01 1.10503107e-01 3.56362522e-01
-7.36816585e-01 -4.38822180e-01 -1.14148326e-01 9.49545324e-01
-1.15635887e-01 -5.76445103e-01 1.20081937e+00 5.11906750e-04
-2.99832463e-01 1.83313608e-01 -3.18006188e-01 3.02873645e-02
6.76036060e-01 1.57692030e-01 -2.79873013e-01 -1.53223708e-01
-7.61117399e-01 1.80071637e-01 5.03390491e-01 1.56297028e-01
-9.21049565e-02 -1.26559734e+00 -4.92675483e-01 1.92625165e-01
2.93447554e-01 -8.69387388e-02 3.76428217e-01 8.87401938e-01
-1.69041604e-01 1.24713294e-01 3.14209998e-01 -5.26887715e-01
-1.24591768e+00 5.20961881e-01 2.71316737e-01 -4.62525100e-01
-4.55112517e-01 1.30561352e+00 -8.83120596e-02 -6.17720306e-01
4.43884760e-01 -9.59701911e-02 3.69670726e-02 2.71384656e-01
8.01518500e-01 4.49163377e-01 5.25623441e-01 -4.57985401e-01
-1.60356507e-01 5.18263280e-01 -6.12067580e-01 1.62936151e-01
1.14422512e+00 -4.12289143e-01 -2.61741392e-02 4.63542193e-01
1.21931171e+00 1.15605466e-01 -1.42071021e+00 -7.42217302e-01
3.74400914e-01 -4.71753657e-01 4.07877117e-01 -1.06325912e+00
-1.12272656e+00 6.63355410e-01 6.05398953e-01 1.45713136e-01
1.09881794e+00 -4.21548665e-01 9.07421649e-01 5.95842957e-01
4.13526356e-01 -1.37720418e+00 -1.91430226e-01 5.82463562e-01
5.02777211e-02 -1.54181373e+00 2.54182424e-02 -5.71957290e-01
-5.96639454e-01 7.90978968e-01 4.09457833e-01 1.25604004e-01
6.00478709e-01 1.69830889e-01 5.20555437e-01 4.01133765e-03
-8.73188794e-01 1.02012306e-01 -9.66709033e-02 3.62733275e-01
3.30740541e-01 7.80496523e-02 -2.26444378e-01 4.63280022e-01
8.37972313e-02 -8.99242833e-02 5.08009791e-01 8.65765095e-01
-5.95155597e-01 -1.25542164e+00 -4.18599278e-01 4.46862072e-01
-6.17185652e-01 -1.15698218e-01 -1.25525817e-01 2.63199270e-01
1.94515824e-01 1.18842161e+00 -2.53438115e-01 -4.22942728e-01
2.33523265e-01 3.52068156e-01 3.01260322e-01 -7.07086325e-01
-8.84008348e-01 -1.30239234e-03 -8.19552839e-02 -4.80896503e-01
-3.70904297e-01 -3.19611102e-01 -1.35797691e+00 -5.22679806e-01
-8.41644585e-01 3.79858106e-01 7.39034295e-01 9.80359375e-01
4.61228490e-01 4.39258248e-01 8.78648102e-01 -4.60173607e-01
-8.92004013e-01 -9.69561875e-01 -6.13495231e-01 5.33219278e-01
4.03044015e-01 -5.47742128e-01 -4.63267475e-01 6.94068298e-02] | [9.360489845275879, 3.9904983043670654] |
49540e4f-5f02-4664-9a94-344602047dad | reasoning-circuits-few-shot-multihop-question | 2211.08466 | null | https://arxiv.org/abs/2211.08466v1 | https://arxiv.org/pdf/2211.08466v1.pdf | Reasoning Circuits: Few-shot Multihop Question Generation with Structured Rationales | Multi-hop Question Generation is the task of generating questions which require the reader to reason over and combine information spread across multiple passages using several reasoning steps. Chain-of-thought rationale generation has been shown to improve performance on multi-step reasoning tasks and make model predictions more interpretable. However, few-shot performance gains from including rationales have been largely observed only in +100B language models, and otherwise require large scale manual rationale annotation. In this work, we introduce a new framework for applying chain-of-thought inspired structured rationale generation to multi-hop question generation under a very low supervision regime (8- to 128-shot). We propose to annotate a small number of examples following our proposed multi-step rationale schema, treating each reasoning step as a separate task to be performed by a generative language model. We show that our framework leads to improved control over the difficulty of the generated questions and better performance compared to baselines trained without rationales, both on automatic evaluation metrics and in human evaluation. Importantly, we show that this is achievable with a modest model size. | ['Anna Rumshisky', 'Saurabh Kulshreshtha'] | 2022-11-15 | null | null | null | null | ['question-generation'] | ['natural-language-processing'] | [ 4.44365561e-01 9.13861334e-01 1.51348159e-01 -4.33424860e-01
-1.75792098e+00 -6.75279915e-01 9.97122586e-01 5.02346754e-01
-2.08935514e-01 9.92364347e-01 6.56202018e-01 -3.90352428e-01
-7.27886185e-02 -6.62895501e-01 -7.68498778e-01 -9.40395519e-02
4.97695237e-01 9.02761996e-01 2.75397569e-01 -4.70984310e-01
3.72493714e-01 -1.61733806e-01 -1.33984232e+00 8.68122280e-01
9.28134441e-01 3.91736090e-01 -2.48880293e-02 1.18468297e+00
-2.48932198e-01 1.38991892e+00 -8.59839797e-01 -9.71839607e-01
-4.59050089e-02 -8.87845278e-01 -1.36957097e+00 3.92002203e-02
6.71362042e-01 -2.87442833e-01 1.21295668e-01 4.70452219e-01
4.99871969e-01 3.23930949e-01 7.61322558e-01 -9.40597951e-01
-8.92276406e-01 1.04238403e+00 -1.45356148e-01 1.89320385e-01
7.16201663e-01 1.28055587e-01 1.55405617e+00 -9.07173157e-01
7.56800771e-01 1.20439756e+00 6.42841816e-01 8.37682247e-01
-1.29628599e+00 -9.51257870e-02 2.11242259e-01 1.93264216e-01
-8.43767643e-01 -3.80117118e-01 6.05315030e-01 -3.08481991e-01
1.13545763e+00 3.97777796e-01 4.56139505e-01 1.16886365e+00
-2.29139831e-02 8.88494372e-01 8.64396036e-01 -6.78975999e-01
3.91225904e-01 6.97998330e-02 2.19717309e-01 7.96870828e-01
3.12945068e-01 -6.75804973e-01 -6.16512239e-01 -3.28456104e-01
2.91907996e-01 -3.65721464e-01 -1.37534469e-01 1.11251429e-01
-1.01891911e+00 1.28514659e+00 2.66260684e-01 1.43153548e-01
-4.82134491e-01 3.78827393e-01 1.16104610e-01 2.55952567e-01
4.20407951e-01 9.40278947e-01 -3.26263487e-01 -2.94126123e-01
-1.11690116e+00 7.86911547e-01 1.08716702e+00 9.65650260e-01
4.78923589e-01 -3.05144221e-01 -8.81551206e-01 6.91622913e-01
3.19791198e-01 3.19659680e-01 4.89894897e-01 -1.26840520e+00
8.40582550e-01 4.88797128e-01 4.77152884e-01 -5.92358410e-01
-2.55428910e-01 -2.50871420e-01 -4.98916000e-01 -1.00379311e-01
5.01923740e-01 -3.22200507e-01 -8.30913901e-01 1.79712188e+00
2.79185027e-01 -3.89244765e-01 2.20233500e-01 4.97108817e-01
8.22212219e-01 5.95788419e-01 1.48473218e-01 -1.32157817e-01
1.59160340e+00 -1.39355505e+00 -6.08273685e-01 -3.76298934e-01
7.84687102e-01 -8.20840597e-01 1.45312905e+00 3.31190079e-01
-1.47145724e+00 -3.23373348e-01 -6.78275526e-01 -5.38997293e-01
3.18309367e-02 8.68442729e-02 5.79402506e-01 5.46171129e-01
-1.18900728e+00 3.94511580e-01 -3.75085741e-01 -3.79925817e-01
4.36590225e-01 -1.10092647e-01 1.27421513e-01 -1.53432533e-01
-1.31919539e+00 8.87848914e-01 1.35725364e-01 -2.47407839e-01
-7.96620190e-01 -8.49812567e-01 -7.20722973e-01 1.81024805e-01
5.50693393e-01 -1.35566294e+00 1.90339470e+00 -3.29925269e-01
-1.44690990e+00 5.88035703e-01 -3.99483085e-01 -7.02627540e-01
6.92016661e-01 -6.92043006e-01 1.88859314e-01 2.96754688e-01
4.68666553e-01 8.83769631e-01 5.90597034e-01 -1.12332630e+00
-3.40610713e-01 2.27866948e-01 5.23346007e-01 3.61170113e-01
-1.51013220e-02 -4.13245633e-02 -2.85877794e-01 -7.33245552e-01
-2.27723777e-01 -9.38672900e-01 -4.47564423e-01 -2.65048355e-01
-4.37257528e-01 -5.06317019e-01 2.81879008e-01 -8.52142453e-01
1.12182093e+00 -1.39155853e+00 4.11040336e-03 -3.79521251e-01
1.04285568e-01 1.28960446e-01 -2.18327135e-01 7.37246931e-01
1.94851443e-01 4.56595093e-01 -3.57198477e-01 -5.09419918e-01
2.40176827e-01 3.08077782e-02 -6.23595893e-01 -4.83977705e-01
4.45898086e-01 1.15138495e+00 -1.00837779e+00 -5.60230136e-01
-2.27005363e-01 1.94967374e-01 -9.51431096e-01 5.79232931e-01
-8.74927998e-01 1.28893964e-02 -5.39154708e-01 1.13786109e-01
1.37745412e-02 -8.18808734e-01 1.40842854e-03 7.15155154e-02
4.68335807e-01 8.30450177e-01 -7.92753041e-01 1.95768225e+00
-7.15127230e-01 5.12122214e-01 -5.16641796e-01 -3.22243750e-01
6.11546159e-01 5.95632076e-01 -1.13815784e-01 -3.62475246e-01
-1.90534517e-01 7.66650913e-03 -1.23728350e-01 -6.33123338e-01
7.03152478e-01 -3.81053686e-01 -2.68777013e-01 1.07358813e+00
1.49886347e-02 -6.44208014e-01 7.37011731e-01 8.05380046e-01
1.38753164e+00 3.21718641e-02 3.64272654e-01 1.25693113e-01
4.89156783e-01 3.70712638e-01 -4.39476855e-02 1.24800718e+00
2.39669383e-01 7.34099388e-01 5.36150515e-01 -2.04098299e-01
-1.05245745e+00 -9.40390408e-01 5.41812956e-01 1.13244116e+00
-1.89728871e-01 -6.05788827e-01 -8.10024083e-01 -7.89002299e-01
-3.09028774e-01 1.48325503e+00 -6.25135124e-01 -8.66373032e-02
-7.17162967e-01 -8.28714550e-01 7.45393932e-01 6.27189577e-01
2.99519241e-01 -1.18845224e+00 -7.32591927e-01 3.81804734e-01
-5.94487071e-01 -1.12451649e+00 -4.52194870e-01 -9.79335746e-04
-8.37205231e-01 -9.49522197e-01 -6.22763634e-01 -4.19295579e-01
5.40902138e-01 1.22054197e-01 1.60312557e+00 3.13035309e-01
-2.03634664e-01 5.30455232e-01 -5.20421624e-01 -3.19317222e-01
-7.21303344e-01 2.21734866e-01 -5.90531230e-01 -3.54195893e-01
-1.35694966e-01 -4.76514578e-01 -5.39664209e-01 -2.55324215e-01
-9.31345701e-01 5.90711117e-01 8.11405778e-01 9.05817330e-01
2.77342826e-01 -4.38832730e-01 7.12331951e-01 -1.26323986e+00
1.17661774e+00 -4.26358312e-01 5.44725955e-02 5.84892988e-01
-6.92882657e-01 5.67899525e-01 6.65591776e-01 -9.51939896e-02
-1.56699145e+00 -4.43580240e-01 -2.87727445e-01 2.11147651e-01
-1.74262583e-01 3.57068092e-01 1.96081594e-01 4.97122467e-01
8.56573403e-01 1.03071211e-02 -3.08815151e-01 -3.36246073e-01
9.82507467e-01 2.29813352e-01 2.75697559e-01 -8.01921666e-01
9.29758012e-01 2.05489337e-01 -2.07724884e-01 -3.81925821e-01
-1.53789508e+00 -2.92332947e-01 -3.45674962e-01 -6.30029216e-02
8.54754269e-01 -9.23383474e-01 -2.97933489e-01 -1.15329497e-01
-1.59739542e+00 -4.79908973e-01 -6.06430471e-01 6.54064305e-03
-6.71044111e-01 2.67482728e-01 -7.93338597e-01 -8.58771980e-01
-7.86147237e-01 -7.02229679e-01 1.10329843e+00 3.80676329e-01
-7.73954034e-01 -1.06491232e+00 2.46034727e-01 1.11458230e+00
4.70250070e-01 1.33959763e-02 1.21907628e+00 -1.01675570e+00
-7.64498889e-01 -1.33072659e-01 -7.08429068e-02 7.53485188e-02
-5.51565550e-03 -1.58995122e-01 -9.03673947e-01 2.00624123e-01
8.50610733e-02 -8.63324642e-01 8.65345418e-01 -1.91853493e-01
8.44685197e-01 -6.45344675e-01 9.67593119e-02 -1.52246237e-01
1.16258752e+00 -2.83503652e-01 3.83130848e-01 1.15616381e-01
6.05379283e-01 5.93643606e-01 5.17409444e-01 3.35758030e-01
8.24274421e-01 5.25124431e-01 1.18062347e-01 9.92824882e-02
-4.02569532e-01 -4.42176819e-01 2.21916959e-01 7.58039713e-01
3.40771042e-02 -6.10558331e-01 -8.78165722e-01 7.79930949e-01
-1.85244179e+00 -1.25075817e+00 -4.54086930e-01 1.72941351e+00
1.19554114e+00 2.27693737e-01 -4.39987052e-03 2.70399135e-02
4.17670012e-01 1.70471892e-01 -2.93353200e-01 -5.19820809e-01
-2.65076719e-02 4.14648026e-01 -1.80584505e-01 7.89061546e-01
-5.62020540e-01 8.53498995e-01 6.80639935e+00 6.14999652e-01
-4.34652030e-01 2.82664686e-01 6.29622042e-01 -3.21375757e-01
-9.88615572e-01 3.40882808e-01 -7.79522002e-01 1.12260886e-01
1.06595016e+00 -3.54231775e-01 1.84986994e-01 6.51895165e-01
1.30342051e-01 -1.20350800e-01 -1.17881131e+00 4.37202543e-01
4.99198198e-01 -1.44392765e+00 5.67111433e-01 -3.58235240e-01
1.01862907e+00 -3.55734676e-01 -2.57412791e-01 5.00019252e-01
7.37401783e-01 -9.22392964e-01 8.25427532e-01 3.84708881e-01
1.98875532e-01 -3.63390267e-01 5.67190409e-01 5.52185714e-01
-8.31196189e-01 -6.10144064e-02 -1.56370565e-01 -1.78707257e-01
6.45816207e-01 6.20155454e-01 -1.41592503e+00 6.25239313e-01
1.24143377e-01 7.79790506e-02 -8.99278462e-01 7.06183791e-01
-7.99811900e-01 7.95283437e-01 -4.96648019e-03 -3.64624709e-01
3.22306871e-01 3.37693393e-01 2.25189239e-01 1.17213714e+00
2.97899216e-01 2.23960847e-01 -8.84142444e-02 9.36899960e-01
-4.03102934e-01 1.31315798e-01 -2.26249188e-01 -1.72222331e-01
2.85764962e-01 1.23656142e+00 -5.77547491e-01 -7.26099491e-01
-2.65528232e-01 1.04883468e+00 6.71755552e-01 1.19828738e-01
-9.98907208e-01 -2.99073786e-01 8.20560455e-02 2.36944482e-01
3.08028758e-01 4.95300367e-02 -4.13116515e-01 -1.26799202e+00
1.28331408e-01 -9.24171448e-01 5.45419872e-01 -1.21065784e+00
-1.21945298e+00 6.26203597e-01 4.10707779e-02 -6.58179522e-01
-8.63750339e-01 -2.60722250e-01 -8.86411190e-01 8.13642323e-01
-1.50998235e+00 -1.05145109e+00 -2.98030227e-01 1.30691633e-01
1.07321668e+00 3.10399085e-01 8.91106546e-01 -2.23776177e-01
-2.41294786e-01 4.39860106e-01 -4.03092712e-01 -5.82148135e-02
6.83957040e-01 -1.62774491e+00 8.28014731e-01 9.55865085e-01
4.70075428e-01 6.78834975e-01 9.83903766e-01 -6.93660259e-01
-1.01307440e+00 -1.08003712e+00 1.45038319e+00 -9.39475656e-01
6.97208881e-01 -2.41440266e-01 -9.69695270e-01 6.66707397e-01
6.26081169e-01 -5.20455718e-01 9.70011890e-01 1.87157705e-01
-4.58429992e-01 2.88871288e-01 -8.99972200e-01 7.51947939e-01
9.44725871e-01 -5.92540443e-01 -1.37563944e+00 7.00844228e-01
1.01494992e+00 -2.66576856e-01 -5.26417911e-01 -4.17913869e-02
2.78956473e-01 -9.37362850e-01 8.36605072e-01 -8.58455122e-01
9.84077156e-01 -9.89635810e-02 6.34702668e-02 -1.11854708e+00
-2.03082174e-01 -9.25770581e-01 -2.74181873e-01 1.18018031e+00
7.68040895e-01 -1.72957465e-01 7.06708431e-01 9.26966429e-01
-1.61232024e-01 -7.51786411e-01 -6.05868936e-01 -4.53454554e-01
6.53600320e-02 -2.21537575e-01 4.62845743e-01 4.00388151e-01
2.19475515e-02 9.75489736e-01 -3.31898421e-01 -1.30278572e-01
6.18295610e-01 2.67667323e-01 8.02563846e-01 -1.01400852e+00
-6.26872241e-01 -3.27212393e-01 4.65820789e-01 -1.12650037e+00
-3.16534340e-02 -8.35651398e-01 4.20051843e-01 -2.22206593e+00
3.79447371e-01 -7.74244294e-02 1.11595891e-01 4.59312677e-01
-8.78243744e-01 1.09487951e-01 3.27493370e-01 1.59787878e-01
-9.56842601e-01 5.33349037e-01 1.21068311e+00 -2.07891762e-02
1.08377360e-01 -8.78791660e-02 -1.20517552e+00 6.26116097e-01
6.15384459e-01 -6.34330690e-01 -7.40875125e-01 -6.17444694e-01
7.11954594e-01 4.08146322e-01 4.25545871e-01 -1.03154612e+00
3.28796595e-01 -2.07794636e-01 7.20966458e-02 -3.80009085e-01
4.48773414e-01 -2.65758652e-02 2.48678979e-02 3.65761667e-01
-1.04308176e+00 1.76743835e-01 2.99987961e-02 6.50036812e-01
4.26784391e-03 -5.57142198e-01 4.11086828e-01 -4.25288320e-01
-1.53225541e-01 -2.57195324e-01 -3.63506377e-01 7.29034305e-01
7.52222180e-01 2.03259751e-01 -4.78309780e-01 -6.92482352e-01
-5.31883061e-01 1.46560803e-01 2.49283507e-01 2.70420521e-01
3.79430383e-01 -1.11734760e+00 -9.85801578e-01 -4.42747623e-01
1.09044671e-01 1.41439825e-01 2.42356971e-01 4.50481474e-01
-3.58804971e-01 7.29327738e-01 1.86057001e-01 -1.73818246e-01
-1.10463643e+00 1.93214759e-01 -4.55774441e-02 -1.05587363e+00
-5.17408609e-01 1.17934453e+00 -3.46885957e-02 -2.11136311e-01
-1.26504824e-01 -3.52285832e-01 -2.17133403e-01 2.55566686e-01
5.69642365e-01 1.96005687e-01 1.08904958e-01 -8.86736251e-03
-3.11331693e-02 3.40028107e-01 -4.39819962e-01 -4.74592805e-01
1.16542852e+00 -1.17723085e-01 1.37314901e-01 4.72928137e-01
6.90314412e-01 1.38658494e-01 -1.12000871e+00 -5.56586776e-03
1.68092653e-01 -1.67466462e-01 -5.51820636e-01 -1.25467837e+00
-1.85536295e-01 9.01286602e-01 -2.93886393e-01 4.13075417e-01
7.38449216e-01 2.49542475e-01 1.06767690e+00 8.61504972e-01
1.71614796e-01 -8.41233253e-01 7.31063783e-01 4.90966529e-01
1.10183573e+00 -1.17089272e+00 -5.13510182e-02 -2.92816788e-01
-9.52630222e-01 8.48414660e-01 5.86519182e-01 1.30085066e-01
-1.23926789e-01 -1.17505662e-01 1.60632849e-01 -3.41841191e-01
-1.59005678e+00 -4.96343598e-02 3.35079670e-01 2.12921068e-01
6.59248054e-01 -1.91094846e-01 -2.96248287e-01 6.68215930e-01
-4.57532287e-01 8.10751691e-02 9.59609687e-01 9.52896178e-01
-6.10365629e-01 -1.21391070e+00 -1.62198797e-01 5.17715096e-01
-4.02260095e-01 -4.96236593e-01 -5.55913091e-01 5.82176626e-01
-3.88367891e-01 1.24626243e+00 -3.00170004e-01 1.45989716e-01
2.84085453e-01 6.75840676e-01 4.61193711e-01 -1.08239734e+00
-9.50109124e-01 -4.68071818e-01 5.80601096e-01 -2.99902141e-01
-3.06975067e-01 -5.67679167e-01 -1.34709275e+00 -1.39368353e-02
-1.85241833e-01 5.95030785e-01 3.27971727e-01 1.28643763e+00
5.53269148e-01 6.25976861e-01 2.17187524e-01 -1.94506869e-01
-9.97592986e-01 -1.07261896e+00 8.68288800e-02 6.27922177e-01
8.63183662e-02 -1.52786359e-01 -4.42013711e-01 4.03284520e-01] | [11.355720520019531, 8.182151794433594] |
13cd1500-0a00-4c0a-8a7d-db588707135d | reconvat-a-semi-supervised-automatic-music | 2107.04954 | null | https://arxiv.org/abs/2107.04954v2 | https://arxiv.org/pdf/2107.04954v2.pdf | ReconVAT: A Semi-Supervised Automatic Music Transcription Framework for Low-Resource Real-World Data | Most of the current supervised automatic music transcription (AMT) models lack the ability to generalize. This means that they have trouble transcribing real-world music recordings from diverse musical genres that are not presented in the labelled training data. In this paper, we propose a semi-supervised framework, ReconVAT, which solves this issue by leveraging the huge amount of available unlabelled music recordings. The proposed ReconVAT uses reconstruction loss and virtual adversarial training. When combined with existing U-net models for AMT, ReconVAT achieves competitive results on common benchmark datasets such as MAPS and MusicNet. For example, in the few-shot setting for the string part version of MusicNet, ReconVAT achieves F1-scores of 61.0% and 41.6% for the note-wise and note-with-offset-wise metrics respectively, which translates into an improvement of 22.2% and 62.5% compared to the supervised baseline model. Our proposed framework also demonstrates the potential of continual learning on new data, which could be useful in real-world applications whereby new data is constantly available. | ['Li Su', 'Dorien Herremans', 'Kin Wai Cheuk'] | 2021-07-11 | null | null | null | null | ['music-transcription'] | ['music'] | [ 6.86093807e-01 -9.77015197e-02 2.37580277e-02 -1.19148307e-01
-1.18086743e+00 -9.98317838e-01 2.46195406e-01 -1.76940158e-01
-2.34894797e-01 7.75613129e-01 1.62100911e-01 2.32179929e-02
-3.29445675e-02 -4.72635210e-01 -8.92130435e-01 -5.66952765e-01
6.09330609e-02 4.88310128e-01 -1.42737314e-01 -2.16007471e-01
1.12663150e-01 -3.40075903e-02 -1.45793080e+00 4.75925356e-01
8.55771780e-01 9.46367502e-01 2.02501938e-03 6.69547021e-01
5.77954762e-02 8.01153362e-01 -8.41421366e-01 -4.99337971e-01
4.89420027e-01 -6.62705243e-01 -6.51936293e-01 -2.22654790e-02
6.78249598e-01 -3.87771125e-03 -1.37414530e-01 9.39765453e-01
9.36344802e-01 3.93588513e-01 5.57898283e-01 -1.15052724e+00
-5.08829474e-01 1.23867297e+00 -6.10852957e-01 -1.64698675e-01
2.76581228e-01 5.82773127e-02 1.14257216e+00 -7.87730753e-01
5.19171536e-01 9.56764221e-01 9.38306510e-01 6.51308298e-01
-1.32520378e+00 -1.14113605e+00 -1.39929667e-01 1.85274541e-01
-1.38668656e+00 -5.37026882e-01 9.12070096e-01 -2.64380783e-01
8.36119175e-01 5.47221065e-01 3.67831409e-01 1.32220972e+00
-3.34950686e-01 9.92580533e-01 1.06916320e+00 -3.90941292e-01
1.22826636e-01 -2.16394007e-01 -5.08890152e-01 8.75514820e-02
-3.90469313e-01 -6.33673519e-02 -9.82542992e-01 1.61009487e-02
7.84164071e-01 -2.75311947e-01 -2.60746330e-01 5.40354401e-02
-1.36381054e+00 3.05122256e-01 2.80672818e-01 2.15650320e-01
-1.53109476e-01 6.08218424e-02 9.00519729e-01 5.13788044e-01
3.84888798e-01 6.50084257e-01 -4.44600254e-01 -6.31530643e-01
-1.31839705e+00 5.95359430e-02 5.01426935e-01 1.08102429e+00
2.60466367e-01 5.73754609e-01 -1.36153132e-01 1.22546577e+00
-1.77968130e-01 3.84387165e-01 9.37032878e-01 -1.03700960e+00
6.72982097e-01 3.34240854e-01 -2.00396404e-01 -7.49661803e-01
2.08940897e-02 -1.08251476e+00 -1.17278457e+00 -4.76503633e-02
1.69262424e-01 2.26914342e-02 -9.15240467e-01 1.90504420e+00
-8.26270208e-02 9.98799622e-01 2.56886154e-01 8.01913202e-01
7.93986917e-01 6.96347952e-01 -2.18437955e-01 -2.91208208e-01
7.02274740e-01 -1.30406988e+00 -7.34450519e-01 -1.25147954e-01
2.29846984e-01 -1.07542932e+00 1.53253388e+00 8.55188847e-01
-1.09579647e+00 -9.78333712e-01 -1.22122109e+00 1.66897386e-01
5.35793304e-02 1.20489284e-01 2.98401296e-01 5.22981703e-01
-7.40682721e-01 7.49657512e-01 -6.23679638e-01 -1.06586091e-01
4.38602537e-01 3.27365160e-01 -2.65282840e-01 -6.74064830e-02
-1.05306041e+00 1.60797954e-01 5.17439783e-01 -1.31575949e-02
-1.15117371e+00 -6.88125074e-01 -4.28600788e-01 2.97601670e-02
3.97048742e-01 -4.87166554e-01 1.46236145e+00 -1.08841753e+00
-1.63014495e+00 5.33685684e-01 1.90016910e-01 -6.45000339e-01
7.85341144e-01 -4.39459115e-01 -5.01718879e-01 -1.55542493e-01
3.55862230e-02 5.42062402e-01 7.49319911e-01 -1.14736581e+00
-4.64082897e-01 -1.66458264e-01 -2.47424364e-01 4.06124473e-01
-4.48617697e-01 -2.86125720e-01 -4.01752442e-01 -1.29664445e+00
-1.95470899e-02 -1.11296499e+00 -2.09565330e-02 -4.04218793e-01
-6.19991541e-01 9.50537995e-02 7.39393771e-01 -7.06257164e-01
1.31208313e+00 -2.24714994e+00 3.19347411e-01 -6.26387894e-02
-3.73401523e-01 4.66106772e-01 -3.43362570e-01 5.80484450e-01
-2.11613163e-01 1.64734498e-02 -4.10760552e-01 -3.85320187e-01
1.67256258e-02 2.25865886e-01 -5.98826826e-01 -5.32351024e-02
-1.36818394e-01 8.16976488e-01 -9.67776358e-01 -9.42877755e-02
2.53227446e-02 3.39639902e-01 -6.08057320e-01 1.96679890e-01
-3.31862003e-01 8.66054595e-01 1.42416775e-01 6.55857563e-01
3.47897470e-01 2.33661920e-01 8.44260976e-02 1.40949935e-01
2.88901716e-01 4.01763350e-01 -1.25841737e+00 2.49746656e+00
-5.25283933e-01 4.26637828e-01 -3.09785515e-01 -8.86424601e-01
1.08923090e+00 6.57636821e-01 3.24395955e-01 -4.41592991e-01
5.82468696e-02 4.40174401e-01 7.59851858e-02 -1.02642298e-01
4.49862063e-01 -4.74479944e-01 -2.93431759e-01 5.27937293e-01
4.28542763e-01 3.28932181e-02 8.44565257e-02 5.69657469e-03
1.24501514e+00 4.81666297e-01 1.37271985e-01 2.42021963e-01
3.95466954e-01 -2.88790703e-01 8.73101950e-01 7.59774804e-01
3.06222104e-02 8.92177284e-01 5.50600067e-02 -1.51074886e-01
-9.89929557e-01 -1.07573104e+00 1.24695651e-01 1.20400655e+00
-3.08483213e-01 -5.74111819e-01 -7.07190037e-01 -5.95353901e-01
-1.59569547e-01 7.44178116e-01 -3.59606594e-01 -3.05026442e-01
-4.88332212e-01 -4.12925333e-01 1.03397858e+00 7.01842129e-01
6.70736849e-01 -1.29920828e+00 -7.33816177e-02 4.31508631e-01
-4.39228684e-01 -9.56322253e-01 -6.33276522e-01 1.64435312e-01
-1.08382511e+00 -6.23011053e-01 -7.50824213e-01 -8.37363839e-01
1.66900799e-01 -5.27883768e-02 1.08151281e+00 -3.79355103e-01
4.75391261e-02 -3.69213521e-02 -6.08884811e-01 -5.36379933e-01
-4.87456799e-01 4.73191559e-01 4.32672828e-01 3.49557906e-01
4.35786620e-02 -1.25100899e+00 -3.06925327e-01 3.97966862e-01
-9.14148748e-01 8.58861730e-02 5.98652542e-01 1.00760031e+00
1.02004993e+00 1.55638292e-01 1.26948917e+00 -1.00445569e+00
5.78309774e-01 -3.35035712e-01 -7.56683126e-02 1.42908931e-01
-4.62378770e-01 -3.59566897e-01 9.09416556e-01 -8.26602042e-01
-8.36161077e-01 1.11373156e-01 -2.50773668e-01 -8.60260129e-01
-1.05828591e-01 4.61735874e-01 -3.32223564e-01 2.58737117e-01
8.89580131e-01 3.50279331e-01 -3.70167434e-01 -8.12263548e-01
2.55125970e-01 1.00407243e+00 1.11077797e+00 -6.49812400e-01
8.56125653e-01 5.30255288e-02 -2.80496687e-01 -4.09321934e-01
-1.02377176e+00 -4.58711386e-01 -5.90808749e-01 -2.09952340e-01
4.12826627e-01 -9.84821081e-01 -4.16459084e-01 5.06140649e-01
-6.82423353e-01 -3.97912741e-01 -7.33434916e-01 4.37008739e-01
-1.00690281e+00 2.43895739e-01 -6.13037705e-01 -8.45405161e-01
-6.79633081e-01 -8.31391156e-01 9.03707564e-01 -5.98481148e-02
-4.51449990e-01 -6.93191171e-01 1.85515389e-01 5.70968449e-01
2.97946572e-01 3.16255778e-01 7.17488408e-01 -8.05678904e-01
-4.16126162e-01 -2.81298518e-01 3.34851146e-01 8.97364557e-01
2.43320510e-01 -4.57675636e-01 -1.32061076e+00 -4.16707247e-01
-1.26249820e-01 -6.82368577e-01 7.56674170e-01 -4.31458279e-02
1.32086635e+00 -2.84149230e-01 2.84473926e-01 8.08313906e-01
1.17552352e+00 2.78300017e-01 6.12291396e-01 2.66106457e-01
8.09502006e-01 4.90807667e-02 6.71829343e-01 3.24528903e-01
-2.32150003e-01 7.43386388e-01 3.14744443e-01 2.18376130e-01
-3.44256580e-01 -8.88709903e-01 4.70199704e-01 1.59344435e+00
-2.14612350e-01 -2.58348733e-01 -8.32156003e-01 7.03310847e-01
-2.01744103e+00 -1.01985645e+00 2.63556510e-01 2.31626463e+00
1.14555490e+00 1.94869429e-01 1.83821425e-01 8.44643593e-01
6.38342023e-01 -7.40112960e-02 -8.65566373e-01 -3.38255644e-01
-2.60737568e-01 5.69715023e-01 2.31980756e-01 -4.32754941e-02
-1.04488695e+00 9.53311265e-01 6.02167273e+00 1.27424359e+00
-1.00205576e+00 2.57235646e-01 3.50836575e-01 -2.90730357e-01
-2.51624845e-02 -1.33382782e-01 -3.07375908e-01 4.20063734e-01
1.12375700e+00 -8.93929005e-02 7.70885706e-01 6.17644727e-01
3.71998083e-03 5.54688632e-01 -1.01723301e+00 1.19181442e+00
2.47889802e-01 -1.06279886e+00 3.09633404e-01 -1.18369743e-01
1.13919199e+00 -6.81629702e-02 2.75528997e-01 5.90269268e-01
8.66430029e-02 -1.16450882e+00 8.05937529e-01 3.21696818e-01
1.32449341e+00 -1.09364772e+00 7.80539930e-01 6.32656038e-01
-1.01395833e+00 1.76043455e-02 -2.23774850e-01 -2.32127383e-01
2.45883018e-01 2.87875950e-01 -1.13529348e+00 8.31106007e-01
5.60084820e-01 1.05720854e+00 -4.40110475e-01 1.19459963e+00
-3.30337763e-01 1.35765755e+00 -2.48411238e-01 4.29329634e-01
1.12538196e-01 6.77381177e-03 7.06874549e-01 1.01338649e+00
4.96179044e-01 -3.53636265e-01 1.41656518e-01 4.83512223e-01
-5.35127580e-01 3.92104894e-01 -4.02018398e-01 -2.20703915e-01
5.04331708e-01 8.98757160e-01 -4.41411018e-01 -2.21147552e-01
-5.57213873e-02 1.21991289e+00 1.59069330e-01 2.82193184e-01
-6.76146686e-01 -4.57390517e-01 4.07717556e-01 -3.37297767e-02
2.83945948e-01 1.00008860e-01 -3.26413184e-01 -1.21774185e+00
7.15792701e-02 -1.37803984e+00 3.53377730e-01 -9.04282033e-01
-1.50105083e+00 8.15606534e-01 -4.16649759e-01 -1.75593364e+00
-3.59519750e-01 -2.51720399e-01 -5.34670115e-01 6.61508858e-01
-9.79091227e-01 -1.37855673e+00 2.06400454e-02 4.76669908e-01
8.42766821e-01 -7.47081697e-01 1.19366884e+00 5.15692294e-01
-3.70700657e-01 1.05441380e+00 2.71095842e-01 2.72609234e-01
1.12832320e+00 -1.27786708e+00 6.50025845e-01 6.75993919e-01
9.68183815e-01 4.29070622e-01 5.42774737e-01 -4.31487352e-01
-1.02904439e+00 -1.34715426e+00 4.98949856e-01 -5.51047504e-01
5.10664940e-01 -4.99528766e-01 -9.75976825e-01 7.74038076e-01
1.90603346e-01 -2.42135659e-01 1.14436412e+00 1.82899877e-01
-6.45812869e-01 -1.93555489e-01 -9.35937345e-01 5.30042827e-01
1.34806228e+00 -6.25118494e-01 -6.04964018e-01 7.44960234e-02
7.94287562e-01 -4.91504461e-01 -9.22285616e-01 6.42513096e-01
7.12981462e-01 -7.51513124e-01 8.62493515e-01 -6.86974823e-01
3.97719741e-01 -3.13849628e-01 -4.34590876e-01 -1.50458634e+00
-5.77485375e-02 -9.84183729e-01 -1.76476076e-01 1.62249303e+00
4.85169709e-01 -2.30978861e-01 7.04727709e-01 -2.93271989e-01
-2.92235792e-01 -5.75179100e-01 -1.03542054e+00 -1.07663095e+00
6.40306622e-02 -8.42080414e-01 6.59766555e-01 1.28419983e+00
-1.95212185e-01 6.74745023e-01 -8.79322588e-01 -2.53794938e-01
6.53218687e-01 8.43831301e-02 9.80582714e-01 -1.09630549e+00
-7.66019225e-01 -1.46130964e-01 -6.36967957e-01 -8.71095657e-01
1.58680141e-01 -1.34281385e+00 -7.68752322e-02 -1.22712398e+00
1.55799523e-01 -4.91543114e-01 -8.27552974e-01 6.60665214e-01
3.39300074e-02 8.44815552e-01 3.39535475e-01 3.43160748e-01
-4.95670676e-01 5.91482699e-01 1.26794004e+00 -3.39391440e-01
-3.73824954e-01 2.14589521e-01 -6.69070959e-01 6.63932204e-01
1.04809248e+00 -5.30003548e-01 -6.38885498e-01 -4.54453051e-01
1.48450419e-01 -3.41334939e-02 -8.53874162e-03 -1.42726362e+00
1.33881658e-01 1.37785271e-01 2.02286690e-01 -5.77420712e-01
5.26342034e-01 -5.58415771e-01 4.06224936e-01 -1.08789422e-01
-5.93919516e-01 -1.55492574e-01 4.36569631e-01 8.33525002e-01
-5.00319719e-01 -1.45346269e-01 4.18678224e-01 -1.56626105e-02
-4.17438000e-01 6.88866302e-02 1.24635845e-01 2.25817308e-01
6.71018124e-01 -1.14830084e-01 -4.10034228e-03 -7.08574712e-01
-8.70650649e-01 -9.54287350e-02 3.21920455e-01 6.91919863e-01
3.82462889e-01 -1.70997572e+00 -9.41800535e-01 2.13938281e-01
2.63554722e-01 1.50957316e-01 4.19446766e-01 5.76388299e-01
-1.90476909e-01 6.52221143e-02 -2.92872667e-01 -5.11620402e-01
-1.25547135e+00 3.76801133e-01 5.00352345e-02 -1.80591255e-01
-5.43011785e-01 9.38934684e-01 -1.52419716e-01 -7.56607056e-01
5.08531213e-01 -1.73030660e-01 1.01376310e-01 -1.28140211e-01
4.56566721e-01 2.96934187e-01 2.06242397e-01 -7.71960557e-01
2.40778625e-02 4.26710576e-01 1.73129719e-02 -3.41201097e-01
1.41583145e+00 2.58644879e-01 1.63900167e-01 1.07742202e+00
8.82121384e-01 2.58809865e-01 -1.15036595e+00 -4.00287807e-01
-1.21632554e-01 -4.25591767e-01 -2.42141828e-01 -1.31866801e+00
-1.01722944e+00 1.09879839e+00 5.47360778e-01 -1.58128828e-01
1.43433940e+00 -2.92723089e-01 1.17109644e+00 3.54410708e-01
5.78822553e-01 -9.77589369e-01 9.47360024e-02 3.88391912e-01
9.78349626e-01 -7.90506303e-01 -3.21486503e-01 -9.67445970e-02
-7.74943769e-01 5.81963778e-01 4.75477993e-01 -1.32835642e-01
1.77029222e-01 2.35722780e-01 3.22293162e-01 4.24212933e-01
-8.98440659e-01 2.36713737e-02 4.17284161e-01 5.41673541e-01
4.21773821e-01 1.22432165e-01 1.11525701e-02 1.12234473e+00
-7.18316138e-01 5.60907312e-02 3.83270919e-01 7.58446872e-01
-1.44346386e-01 -1.35218894e+00 -2.84905672e-01 3.78298163e-01
-6.88812971e-01 -1.68547425e-02 -5.96546233e-01 3.17089170e-01
2.70562470e-01 9.02734995e-01 -2.23621503e-01 -8.72938991e-01
4.63854373e-01 4.44071740e-01 4.08560336e-01 -7.97895074e-01
-7.18930244e-01 4.83217508e-01 5.47909439e-02 -8.45444053e-02
-5.98960042e-01 -5.29535592e-01 -1.06256306e+00 8.99355188e-02
-2.84337938e-01 1.59397617e-01 5.17469227e-01 5.68337321e-01
3.16450566e-01 8.30551028e-01 4.95215178e-01 -8.67132902e-01
-5.44979990e-01 -1.44348598e+00 -6.24494612e-01 5.80918968e-01
1.21234506e-01 -4.77364451e-01 -2.49954253e-01 3.16092044e-01] | [15.721281051635742, 5.339897155761719] |
e688a58a-8422-4e45-ba26-097f6e6a7cea | muller-multilayer-laplacian-resizer-for | 2304.02859 | null | https://arxiv.org/abs/2304.02859v1 | https://arxiv.org/pdf/2304.02859v1.pdf | MULLER: Multilayer Laplacian Resizer for Vision | Image resizing operation is a fundamental preprocessing module in modern computer vision. Throughout the deep learning revolution, researchers have overlooked the potential of alternative resizing methods beyond the commonly used resizers that are readily available, such as nearest-neighbors, bilinear, and bicubic. The key question of our interest is whether the front-end resizer affects the performance of deep vision models? In this paper, we present an extremely lightweight multilayer Laplacian resizer with only a handful of trainable parameters, dubbed MULLER resizer. MULLER has a bandpass nature in that it learns to boost details in certain frequency subbands that benefit the downstream recognition models. We show that MULLER can be easily plugged into various training pipelines, and it effectively boosts the performance of the underlying vision task with little to no extra cost. Specifically, we select a state-of-the-art vision Transformer, MaxViT, as the baseline, and show that, if trained with MULLER, MaxViT gains up to 0.6% top-1 accuracy, and meanwhile enjoys 36% inference cost saving to achieve similar top-1 accuracy on ImageNet-1k, as compared to the standard training scheme. Notably, MULLER's performance also scales with model size and training data size such as ImageNet-21k and JFT, and it is widely applicable to multiple vision tasks, including image classification, object detection and segmentation, as well as image quality assessment. | ['Hossein Talebi', 'Peyman Milanfar', 'Zhengzhong Tu'] | 2023-04-06 | null | null | null | null | ['image-quality-assessment'] | ['computer-vision'] | [ 5.15756644e-02 9.87617648e-04 -1.68001339e-01 -1.93786383e-01
-7.63090611e-01 -3.78053427e-01 3.57135594e-01 -2.83604026e-01
-7.15359509e-01 4.79939789e-01 1.88649818e-02 -3.82876664e-01
2.99900830e-01 -5.72700500e-01 -1.06893158e+00 -7.38307834e-01
5.09114802e-01 -1.61838070e-01 3.68947238e-01 -9.92139578e-02
-2.66386978e-02 4.34147179e-01 -1.24734032e+00 9.95814055e-03
9.57385600e-01 1.27384210e+00 2.31196150e-01 6.54384077e-01
2.49550417e-01 9.09886301e-01 -4.86852497e-01 -6.02762103e-01
3.48212481e-01 -5.36717102e-03 -7.03081787e-01 -1.62539169e-01
8.56951118e-01 -4.50905919e-01 -8.22291970e-01 1.23917258e+00
5.77339232e-01 6.07233588e-03 5.21821797e-01 -8.29652965e-01
-1.22260630e+00 5.78845024e-01 -8.46714675e-01 4.23783779e-01
-4.12661850e-01 3.79679739e-01 8.07145119e-01 -1.15048051e+00
1.62616715e-01 1.12006640e+00 9.76042747e-01 5.98881364e-01
-1.24691880e+00 -7.43290007e-01 1.99294701e-01 4.84015048e-01
-1.28443444e+00 -4.27245915e-01 5.01949608e-01 -2.90708810e-01
7.59636283e-01 8.40107650e-02 5.73706746e-01 9.76297975e-01
3.34527314e-01 9.20314312e-01 1.25952184e+00 -2.70586759e-01
3.68420668e-02 -1.42984018e-01 3.32938194e-01 7.46452332e-01
5.57984635e-02 -1.18930086e-01 -3.84869128e-01 2.92594343e-01
8.26545179e-01 4.55746911e-02 -4.60472226e-01 3.27486426e-01
-1.06650972e+00 6.48567557e-01 1.08487070e+00 9.34817304e-04
-2.13974044e-01 4.51198280e-01 3.12935859e-01 3.60823497e-02
6.79219723e-01 2.30122671e-01 -3.66565287e-01 2.82240123e-01
-8.82816017e-01 -1.10909395e-01 3.58074129e-01 7.89233625e-01
8.28879893e-01 3.11759263e-01 -3.79662305e-01 1.04471397e+00
8.61778185e-02 5.07185817e-01 6.18101299e-01 -8.57227027e-01
2.03732207e-01 5.28764009e-01 -2.95541614e-01 -6.21918023e-01
-2.64291465e-01 -7.77973056e-01 -1.25129473e+00 1.47596151e-01
2.05365509e-01 -1.40263841e-01 -1.42886937e+00 1.71614838e+00
1.92230329e-01 4.63795125e-01 -3.21375467e-02 9.29164231e-01
1.10428643e+00 5.37134767e-01 3.16894390e-02 1.24438263e-01
1.40191031e+00 -1.39513803e+00 -3.08543235e-01 -4.83137310e-01
8.33947062e-02 -8.70310426e-01 1.28049076e+00 1.06679469e-01
-1.14598775e+00 -8.82083118e-01 -1.14333105e+00 -5.60117185e-01
-2.00004593e-01 2.42264733e-01 7.01381981e-01 4.85558838e-01
-1.44818342e+00 5.49776077e-01 -8.59691501e-01 -1.14605404e-01
6.93269908e-01 2.66622543e-01 -1.99809000e-01 -2.21450850e-01
-9.77913380e-01 1.01619744e+00 8.14343616e-02 1.80492312e-01
-1.00764227e+00 -1.01026416e+00 -7.10409343e-01 1.06251650e-01
1.31728500e-01 -8.82253468e-01 1.32652342e+00 -9.56454039e-01
-1.64215255e+00 9.40281391e-01 -2.60536790e-01 -7.72361457e-01
5.95751762e-01 -3.02307963e-01 -1.69192418e-01 6.91720657e-03
-4.35906090e-02 9.17703986e-01 1.21801627e+00 -8.60755324e-01
-6.02005005e-01 -2.96236724e-01 1.97029307e-01 9.96410549e-02
-5.19343853e-01 -1.19000226e-01 -8.05145621e-01 -8.79567206e-01
-4.14499938e-02 -1.03426337e+00 -3.54419500e-01 1.28728911e-01
-5.05641401e-01 -1.54270604e-01 6.24850988e-01 -6.80227935e-01
7.15563595e-01 -2.24605823e+00 -1.14528589e-01 -1.97966516e-01
3.98066640e-01 6.24709964e-01 -1.77664012e-01 -1.52909309e-01
-6.07639104e-02 7.78012574e-02 -3.10962677e-01 -3.09181035e-01
-2.15810567e-01 -1.31364064e-02 -4.37413871e-01 5.96184969e-01
1.57200918e-01 1.31896758e+00 -5.98701060e-01 -1.41423613e-01
3.95744026e-01 7.84260809e-01 -5.25860667e-01 5.08931372e-03
2.73034666e-02 2.96958208e-01 -2.36277863e-01 4.88670379e-01
8.67956400e-01 -4.85676259e-01 -3.64185840e-01 -5.95305860e-01
-2.77881436e-02 9.05835405e-02 -5.27570069e-01 1.41904140e+00
-5.03666878e-01 9.97523904e-01 2.83323936e-02 -9.62104142e-01
6.06934369e-01 -7.64056370e-02 1.86076298e-01 -7.11670518e-01
9.34099033e-02 -7.62720406e-02 -1.24793880e-01 -2.02277154e-01
3.58903140e-01 2.29525000e-01 3.36564571e-01 -9.63579305e-03
1.46124616e-01 1.93159301e-02 -1.68681547e-01 -8.07125643e-02
9.72968936e-01 -2.24630207e-01 2.08668917e-01 -3.55763465e-01
3.50403756e-01 -9.08056349e-02 4.34893787e-01 7.62413144e-01
-2.15420559e-01 8.00871253e-01 1.06657252e-01 -2.97989398e-01
-9.00403976e-01 -1.13767493e+00 -4.55570161e-01 1.27497149e+00
2.69829690e-01 -3.69194224e-02 -1.00521219e+00 -3.82067233e-01
-6.42551333e-02 3.68793994e-01 -5.88387787e-01 -3.75809163e-01
-5.21341383e-01 -1.09383023e+00 8.38851511e-01 6.24291480e-01
1.24324930e+00 -7.98392057e-01 -3.58783841e-01 1.77414380e-02
-5.28327860e-02 -1.38410282e+00 -9.03165638e-01 1.25590384e-01
-8.21048796e-01 -1.01627123e+00 -1.01539648e+00 -9.98950660e-01
6.65467978e-01 5.81715703e-01 9.79460120e-01 -1.52159974e-01
-3.80934894e-01 2.56955624e-01 -1.60804197e-01 -4.02384430e-01
-1.20510750e-01 1.61215946e-01 -1.57035679e-01 1.37051165e-01
-7.86329526e-03 -5.00285149e-01 -1.03740096e+00 1.82572141e-01
-8.15097153e-01 2.13178009e-01 8.24679732e-01 9.59362030e-01
6.68966770e-01 -1.03912480e-01 5.45748949e-01 -7.27310002e-01
4.20421094e-01 -1.21626556e-01 -5.44287682e-01 2.29534373e-01
-6.45245075e-01 -1.07803978e-01 8.13026428e-01 -5.93686461e-01
-9.49965954e-01 -8.04928876e-03 -4.25566316e-01 -6.28835678e-01
1.71069950e-01 4.39438283e-01 -1.80055539e-03 -6.11108243e-01
7.26583838e-01 3.64373058e-01 -1.99348971e-01 -4.95044798e-01
6.28065169e-01 5.48152506e-01 9.91535723e-01 -1.16795018e-01
8.46754372e-01 5.25104761e-01 -1.07481346e-01 -1.03393173e+00
-1.06817663e+00 -1.71069354e-01 -1.49672598e-01 -4.48856987e-02
9.35383260e-01 -1.28780389e+00 -8.61212194e-01 9.09954071e-01
-9.22032475e-01 -5.37238538e-01 -1.99571237e-01 3.14898968e-01
-1.67316929e-01 2.15384632e-01 -1.05927086e+00 -3.14825684e-01
-7.32655525e-01 -1.36891985e+00 8.34219873e-01 6.10786140e-01
4.40928638e-01 -7.93142557e-01 -5.11115313e-01 5.73339283e-01
6.34518206e-01 -7.37202689e-02 8.64096820e-01 -1.61393672e-01
-6.09088659e-01 -5.64015247e-02 -7.21086442e-01 8.45404088e-01
-2.84734480e-02 -8.46687984e-03 -1.47476768e+00 -4.74447578e-01
6.88602924e-02 -3.94182742e-01 1.43214417e+00 6.71370983e-01
1.68077660e+00 -3.59490693e-01 4.15695421e-02 1.19439161e+00
1.34219432e+00 -9.31332558e-02 7.61779785e-01 1.47178486e-01
1.03708375e+00 1.17807567e-01 -3.68058085e-02 -5.24990894e-02
6.01313055e-01 4.75591987e-01 5.24105847e-01 -5.57257175e-01
-5.92065156e-01 -2.11489961e-01 3.51516664e-01 8.62981200e-01
-7.21846595e-02 5.92467785e-02 -7.05762267e-01 3.65440756e-01
-1.65456533e+00 -6.71764791e-01 7.62969777e-02 2.07180905e+00
1.05185091e+00 9.85115990e-02 -3.36045295e-01 -1.81123599e-01
7.34164476e-01 2.76697725e-01 -1.09241223e+00 -3.42810601e-02
-1.87344268e-01 4.41488117e-01 1.01226032e+00 3.81372929e-01
-1.24875712e+00 1.24293804e+00 6.77965593e+00 1.13142276e+00
-1.51640487e+00 2.29936749e-01 1.08004224e+00 1.24250539e-02
6.73619956e-02 -2.26934418e-01 -7.64187336e-01 3.87347490e-01
7.14559138e-01 -7.76083618e-02 6.25962615e-01 9.36988175e-01
9.64126289e-02 1.08628385e-01 -9.33655381e-01 1.11260343e+00
-6.98713819e-03 -1.54029131e+00 6.66118041e-02 -1.05053067e-01
8.75212610e-01 5.56543946e-01 4.23787653e-01 4.45661694e-01
2.47657388e-01 -1.16831183e+00 7.29113817e-01 2.29218185e-01
9.89653170e-01 -6.18307590e-01 5.32284796e-01 2.27323413e-01
-1.10154545e+00 4.96477783e-02 -6.25146985e-01 7.51900524e-02
-2.62512118e-01 8.26973498e-01 -5.76558471e-01 2.05253005e-01
9.49774325e-01 8.05761218e-01 -7.65447199e-01 1.16069758e+00
-2.98607469e-01 9.56759632e-01 -1.74089238e-01 3.52002650e-01
2.20563740e-01 -3.97196822e-02 2.54425555e-01 1.16006482e+00
3.13584469e-02 2.84382113e-04 6.76050186e-02 7.08382070e-01
-6.38623953e-01 -2.17150733e-01 -1.59434900e-01 3.26982081e-01
4.48205203e-01 1.26219559e+00 -5.67221820e-01 -3.78386289e-01
-4.11064327e-01 1.05889392e+00 5.01493990e-01 6.10794425e-01
-1.02492416e+00 -2.60093361e-01 8.22439015e-01 -1.54986650e-01
2.53254026e-01 -1.35985836e-01 -4.08426285e-01 -1.27388477e+00
1.38025850e-01 -8.75481606e-01 6.28196299e-02 -7.63648450e-01
-1.40157282e+00 6.80849850e-01 -3.26934248e-01 -8.68982911e-01
2.86945999e-01 -7.75113463e-01 -6.40797555e-01 9.06961381e-01
-1.80814672e+00 -1.30879998e+00 -6.05783939e-01 7.06307709e-01
6.50841415e-01 -2.30865832e-03 4.28711236e-01 4.06484395e-01
-9.10378754e-01 9.66062486e-01 1.90094590e-01 3.38117570e-01
7.75226176e-01 -1.12718868e+00 8.34065259e-01 1.08867657e+00
1.07044600e-01 6.19064152e-01 4.52834606e-01 -2.08302245e-01
-1.46597755e+00 -1.59333622e+00 3.27876508e-01 -8.04877356e-02
6.64133966e-01 -1.80163875e-01 -1.04562640e+00 6.77829206e-01
2.50578254e-01 4.79983747e-01 1.63644314e-01 -2.26312965e-01
-6.21483266e-01 -4.79433030e-01 -9.74325657e-01 8.17466676e-01
1.05844247e+00 -8.07632327e-01 -3.60130548e-01 4.78366226e-01
1.00558317e+00 -6.86422288e-01 -5.93357325e-01 6.04572415e-01
4.35606927e-01 -9.16471422e-01 1.33673441e+00 -4.03466821e-01
4.88885671e-01 -1.43940046e-01 -4.72335704e-02 -1.35855448e+00
-5.16289353e-01 -5.41522920e-01 5.03121503e-02 1.12512910e+00
3.56175750e-01 -9.75954413e-01 6.52980745e-01 4.50497061e-01
-3.40737402e-01 -9.22982156e-01 -9.50896084e-01 -6.16794288e-01
2.12610886e-01 -4.69312012e-01 3.01261991e-01 6.65452003e-01
-6.07898951e-01 3.62952679e-01 -3.08318138e-01 2.37173706e-01
7.22793758e-01 -6.80821687e-02 5.54457605e-01 -8.92748654e-01
-2.76424319e-01 -6.33273125e-01 -3.76383901e-01 -1.49750781e+00
3.25164527e-01 -9.70336080e-01 1.35192685e-02 -1.46716702e+00
2.77502537e-01 -3.57838392e-01 -4.63663369e-01 5.86785972e-01
-5.72347403e-01 8.42163265e-01 2.08720759e-01 3.05962205e-01
-2.66236782e-01 6.68816686e-01 1.51493835e+00 -4.73475546e-01
-2.37623323e-02 5.69750294e-02 -1.00408161e+00 1.01322711e+00
6.58568203e-01 -4.18525301e-02 -3.18783402e-01 -8.66139770e-01
-2.05200285e-01 -2.85478920e-01 6.93982720e-01 -1.06894183e+00
3.94415677e-01 1.91559121e-02 4.85147119e-01 -1.76682696e-01
3.21036547e-01 -5.11543155e-01 -8.32266808e-02 3.04993868e-01
-4.17646468e-01 -5.89741133e-02 2.93946981e-01 4.80329037e-01
-2.03088179e-01 5.84813207e-02 1.29163146e+00 5.11788130e-02
-9.05106187e-01 6.06615543e-01 -1.09391525e-01 2.14480937e-01
8.63601208e-01 -1.90374383e-04 -6.89948320e-01 -1.81131691e-01
-3.58120084e-01 4.56482880e-02 3.12880337e-01 2.88918704e-01
6.06198370e-01 -1.08764064e+00 -7.53722310e-01 9.84759182e-02
-1.01923339e-01 1.91674739e-01 3.89269471e-01 8.28831017e-01
-5.41454136e-01 1.97124571e-01 -4.74633425e-02 -6.99688613e-01
-9.60970700e-01 3.68382782e-01 5.31590879e-01 -4.90172300e-04
-9.86398339e-01 1.08226812e+00 6.20862663e-01 4.76484075e-02
4.35667753e-01 -6.57189548e-01 -1.22035667e-01 -1.14543892e-01
7.02978730e-01 3.58518273e-01 2.63759524e-01 -5.40858865e-01
-3.35523158e-01 7.37888992e-01 -2.15370789e-01 2.29957119e-01
1.11640358e+00 -3.99558432e-02 -1.35342553e-01 1.06238477e-01
1.22766221e+00 -3.26250941e-01 -1.73403621e+00 -4.28341120e-01
-3.58682662e-01 -2.43941709e-01 5.19519567e-01 -6.02185071e-01
-1.53239655e+00 8.53635550e-01 7.23325372e-01 2.50895899e-02
1.40468466e+00 -7.42778927e-02 9.47663903e-01 3.99912328e-01
3.41997206e-01 -6.72301710e-01 -4.20180894e-02 7.02015162e-01
8.63984823e-01 -1.38547564e+00 -1.14387177e-01 -2.71598369e-01
-5.45840025e-01 7.51085639e-01 6.87723219e-01 -3.50025147e-01
6.71769261e-01 2.86952972e-01 1.46013200e-01 3.07930022e-01
-6.74770534e-01 -3.48881930e-01 5.70829213e-01 4.53337073e-01
3.26296538e-01 3.88943292e-02 1.38752311e-01 3.85095149e-01
-3.22479546e-01 2.65335478e-03 3.63116741e-01 2.80677736e-01
-4.01867032e-01 -4.37073529e-01 -2.11007908e-01 6.52221084e-01
-7.51042545e-01 -4.25670147e-01 -1.59194022e-02 4.90640908e-01
1.78029805e-01 8.88396680e-01 -4.12313687e-03 -4.89637762e-01
2.63141751e-01 -3.34644854e-01 3.60254556e-01 -3.89620215e-01
-5.93255639e-01 -1.57464847e-01 -4.53316748e-01 -5.86560845e-01
-3.15818876e-01 -5.17309383e-02 -1.08004022e+00 -4.95029181e-01
-2.96054095e-01 -2.97761202e-01 6.06004298e-01 1.00326884e+00
3.40218455e-01 7.42046177e-01 5.03759921e-01 -1.16543221e+00
-6.95514143e-01 -9.55602348e-01 -2.03158498e-01 8.26669410e-02
6.92423165e-01 -3.87272120e-01 -2.71379590e-01 2.59068221e-01] | [10.94921588897705, -1.7900110483169556] |
9c3d9247-c796-4990-8558-2c3012f69fc9 | ingram-inductive-knowledge-graph-embedding | 2305.19987 | null | https://arxiv.org/abs/2305.19987v2 | https://arxiv.org/pdf/2305.19987v2.pdf | InGram: Inductive Knowledge Graph Embedding via Relation Graphs | Inductive knowledge graph completion has been considered as the task of predicting missing triplets between new entities that are not observed during training. While most inductive knowledge graph completion methods assume that all entities can be new, they do not allow new relations to appear at inference time. This restriction prohibits the existing methods from appropriately handling real-world knowledge graphs where new entities accompany new relations. In this paper, we propose an INductive knowledge GRAph eMbedding method, InGram, that can generate embeddings of new relations as well as new entities at inference time. Given a knowledge graph, we define a relation graph as a weighted graph consisting of relations and the affinity weights between them. Based on the relation graph and the original knowledge graph, InGram learns how to aggregate neighboring embeddings to generate relation and entity embeddings using an attention mechanism. Experimental results show that InGram outperforms 14 different state-of-the-art methods on varied inductive learning scenarios. | ['Joyce Jiyoung Whang', 'Chanyoung Chung', 'Jaejun Lee'] | 2023-05-31 | null | null | null | null | ['graph-embedding', 'knowledge-graph-embedding', 'inductive-knowledge-graph-completion', 'knowledge-graph-completion', 'knowledge-graphs', 'entity-embeddings'] | ['graphs', 'graphs', 'knowledge-base', 'knowledge-base', 'knowledge-base', 'methodology'] | [ 3.42092067e-02 8.87694836e-01 -5.64652085e-01 -3.62585336e-01
3.84784043e-02 -5.18883467e-01 5.58772326e-01 6.97951138e-01
-2.62797028e-01 9.31690812e-01 2.75158048e-01 -4.36114043e-01
-4.02278453e-01 -1.49851584e+00 -9.56985116e-01 -2.96601504e-01
-5.04154027e-01 9.19603348e-01 9.66456831e-02 -1.18643187e-01
-2.11809263e-01 3.34231585e-01 -1.29530942e+00 1.45642161e-01
1.06655216e+00 3.76688898e-01 -2.02377871e-01 3.90063167e-01
-3.50368679e-01 9.06119585e-01 -3.10016543e-01 -9.19688582e-01
-4.62528244e-02 -1.53427795e-02 -1.14056015e+00 -3.51699919e-01
1.41423136e-01 -2.23396942e-01 -8.92751455e-01 8.36485326e-01
2.72015184e-01 6.45684302e-01 7.95137465e-01 -1.50304186e+00
-1.68052292e+00 1.05262220e+00 -1.82792187e-01 2.41115615e-01
6.96524441e-01 -4.95862156e-01 1.53761566e+00 -1.18201697e+00
1.01962078e+00 1.13133645e+00 5.67840517e-01 4.09671605e-01
-1.18840837e+00 -5.84679306e-01 4.77431387e-01 7.74723828e-01
-1.51623142e+00 3.68838757e-02 7.15040863e-01 -2.73450613e-01
1.41826463e+00 -1.00854307e-01 7.57204473e-01 7.67251551e-01
-1.86500758e-01 5.58577418e-01 3.02748084e-01 -6.63468361e-01
1.54229596e-01 1.45669326e-01 5.23715913e-01 7.34006226e-01
8.26804101e-01 -1.47070006e-01 -5.08609593e-01 -1.57481924e-01
4.65038002e-01 3.38633098e-02 -1.93286240e-01 -6.54840946e-01
-1.19602370e+00 8.34652483e-01 8.63448381e-01 2.17112914e-01
-3.46393079e-01 7.20830113e-02 1.47812769e-01 3.03907573e-01
5.67213416e-01 5.80188811e-01 -4.87849355e-01 4.67430562e-01
-7.41055161e-02 1.07088767e-01 9.72791672e-01 1.32146823e+00
1.29933596e+00 -2.91718304e-01 4.85043135e-03 5.88354528e-01
2.60244191e-01 2.46886700e-01 1.85951769e-01 -4.08363044e-01
6.26771748e-01 1.28225660e+00 2.89187254e-03 -1.21069098e+00
-2.77050376e-01 -3.02554816e-01 -5.86536705e-01 -4.16333199e-01
-1.71213248e-03 -1.79205775e-01 -8.53300869e-01 1.51671445e+00
5.94698250e-01 7.42426336e-01 4.64627504e-01 3.91720831e-01
1.27059007e+00 5.35535991e-01 8.88425857e-02 -4.62036394e-02
1.04852891e+00 -1.00625384e+00 -8.93365502e-01 -3.23195457e-01
8.63837719e-01 -1.61251515e-01 6.09901249e-01 -2.05069587e-01
-6.78304434e-01 -4.26954329e-01 -9.14629400e-01 -3.06940049e-01
-1.23707700e+00 -2.66297221e-01 1.15068543e+00 2.25512922e-01
-9.29201007e-01 4.64061618e-01 -6.41285777e-01 -5.34744501e-01
3.28575432e-01 4.79106724e-01 -6.93750024e-01 -3.47654760e-01
-1.71565676e+00 1.17944741e+00 1.02939558e+00 3.21037203e-01
-5.39653599e-01 -7.92635620e-01 -1.34407473e+00 2.46083170e-01
6.47374809e-01 -9.48585272e-01 6.44109607e-01 -3.91909331e-01
-6.90598845e-01 7.09957182e-01 -2.93034673e-01 -4.40791339e-01
-4.20399338e-01 -3.09167445e-01 -9.69113827e-01 -2.20569789e-01
1.51006121e-03 4.62021321e-01 3.96312833e-01 -1.50180089e+00
-5.82825124e-01 -2.88727373e-01 5.71583152e-01 4.26459193e-01
-4.98078257e-01 -6.14893734e-01 -5.39183915e-01 -3.08319390e-01
8.26042518e-02 -6.90200090e-01 -3.96502018e-03 -4.43672627e-01
-7.55023003e-01 -7.06676781e-01 8.87797773e-01 -3.98370266e-01
1.43446338e+00 -1.83723009e+00 3.76996160e-01 4.47158277e-01
5.75644016e-01 4.16211933e-01 -2.69452363e-01 7.66028464e-01
-5.12884319e-01 1.55738816e-01 1.06471740e-01 3.10780550e-03
1.95587322e-01 7.26748705e-01 -4.46089536e-01 7.27362279e-03
2.58829385e-01 1.29039752e+00 -1.43992734e+00 -3.83583605e-01
3.94180454e-02 5.81521332e-01 -5.21716952e-01 3.13446999e-01
-3.47511053e-01 -7.67431185e-02 -5.11563480e-01 4.62659538e-01
5.97141325e-01 -4.97001231e-01 6.76223397e-01 -3.82062435e-01
4.89477754e-01 2.32188970e-01 -1.16383350e+00 1.58596635e+00
-4.16833282e-01 4.84713405e-01 -7.54549444e-01 -1.15639699e+00
1.06069422e+00 2.31105790e-01 1.96555182e-01 -2.50155032e-01
-1.21938519e-01 -1.07495531e-01 -5.12525924e-02 -6.43494546e-01
7.36690879e-01 8.62607919e-03 2.07790822e-01 3.48980635e-01
5.70888519e-01 1.45899624e-01 4.13329512e-01 7.32738376e-01
1.15637481e+00 1.88655198e-01 3.82356137e-01 2.38554642e-01
3.17513853e-01 -1.30751982e-01 6.49126470e-01 5.08253574e-01
2.54130036e-01 -1.64740067e-02 5.25178730e-01 -6.14552319e-01
-7.12923765e-01 -1.62476504e+00 2.69795299e-01 1.00518394e+00
3.03291589e-01 -8.57113719e-01 -1.60985291e-01 -1.15262437e+00
3.38257343e-01 7.50289083e-01 -9.03965294e-01 -4.77114230e-01
-4.22493488e-01 -4.98564035e-01 2.43868962e-01 6.35918379e-01
1.52736098e-01 -1.13663793e+00 3.13838392e-01 1.73785359e-01
-2.89064169e-01 -1.34730327e+00 -2.28353083e-01 4.19721194e-02
-7.38015056e-01 -1.47113371e+00 -8.53446648e-02 -1.36978543e+00
1.20339549e+00 -8.03010166e-02 1.33657408e+00 2.47515202e-01
-1.37791350e-01 5.19600630e-01 -6.79662406e-01 -2.05005556e-01
-9.56161842e-02 2.22549349e-01 1.49164200e-01 2.56444365e-02
8.82049561e-01 -8.10758829e-01 -2.55259782e-01 -1.77122340e-01
-9.32804823e-01 -8.25488567e-02 3.88565511e-01 9.18468654e-01
6.56623185e-01 3.66624802e-01 8.64757240e-01 -1.47551572e+00
7.32656837e-01 -8.08414340e-01 -1.58310056e-01 7.62976885e-01
-7.74221778e-01 3.36213619e-01 5.54990530e-01 -4.51811790e-01
-1.08061051e+00 -5.55743463e-02 2.76015729e-01 -2.85891950e-01
1.17579684e-01 1.08342910e+00 -1.41496465e-01 6.09015487e-02
6.63127184e-01 -7.17251822e-02 -6.20561361e-01 -1.99419290e-01
1.15402246e+00 2.31660470e-01 4.35061634e-01 -5.58724344e-01
1.15272820e+00 3.07467520e-01 -7.75838345e-02 -5.83643079e-01
-1.06896532e+00 -5.06683528e-01 -1.11962378e+00 6.49946705e-02
7.86087215e-01 -9.29287672e-01 -6.33791208e-01 -1.59441203e-01
-1.24393308e+00 -8.87696669e-02 -6.95917785e-01 6.16583943e-01
1.04258563e-02 3.76591563e-01 -3.29429656e-01 -6.22672677e-01
-1.66795149e-01 -3.61974567e-01 4.86136079e-01 2.84177154e-01
-1.80500805e-01 -1.60444295e+00 3.26435715e-01 9.33646336e-02
6.09937273e-02 4.12420869e-01 1.29205227e+00 -9.77456033e-01
-7.89637506e-01 -3.18793237e-01 -3.29148352e-01 1.05036683e-01
4.83147681e-01 -6.28128350e-02 -6.60398364e-01 -4.98450138e-02
-9.72553313e-01 -3.53113472e-01 8.85237992e-01 -2.22956374e-01
5.84189653e-01 -4.68407363e-01 -9.01324034e-01 4.38431889e-01
1.33841395e+00 8.05797577e-02 8.22064281e-01 7.43385255e-02
1.19822443e+00 3.98600638e-01 4.37481880e-01 2.49785379e-01
9.01846945e-01 2.98498094e-01 3.11682463e-01 2.97385845e-02
-1.65298179e-01 -8.14580917e-01 1.16789620e-02 1.06988311e+00
-2.69740313e-01 -3.21758926e-01 -8.52978528e-01 9.72726285e-01
-1.99365544e+00 -7.45464146e-01 -2.34658420e-01 2.06098032e+00
1.03623211e+00 -2.05962034e-03 -4.46757495e-01 -1.83738545e-01
8.08216929e-01 2.28163581e-02 -5.35386264e-01 -3.68003309e-01
-1.09147325e-01 5.63922405e-01 1.76599011e-01 9.06118155e-01
-8.84595811e-01 1.26527977e+00 5.63191080e+00 2.90069312e-01
-3.20088267e-01 2.35199668e-02 -8.74256417e-02 3.58772188e-01
-7.70854473e-01 4.01409686e-01 -8.11083078e-01 -1.61830187e-01
4.49086994e-01 -4.58099604e-01 4.06367302e-01 6.83384418e-01
-6.17619216e-01 1.23315684e-01 -1.45225561e+00 7.89143443e-01
1.86013132e-01 -1.28956914e+00 4.79087502e-01 -2.09748507e-01
1.02505898e+00 -2.37076387e-01 -2.29610443e-01 7.71919966e-01
8.58033299e-01 -1.05449522e+00 -2.57676810e-01 7.84798920e-01
7.32091844e-01 -6.65092409e-01 8.90868008e-01 -6.25810325e-02
-1.56123304e+00 1.52948707e-01 -4.03738052e-01 -2.51608968e-01
1.84953958e-01 7.48248696e-01 -1.26522708e+00 1.20230365e+00
3.79038155e-01 1.04607749e+00 -6.61495268e-01 7.58161306e-01
-1.05409324e+00 4.35232222e-01 -1.39044836e-01 -2.29800344e-02
-1.45082116e-01 -1.28275827e-01 1.92505762e-01 9.97902513e-01
1.30290583e-01 1.82524920e-01 -5.99755254e-03 9.14139688e-01
-5.25401592e-01 4.20611911e-02 -1.00705981e+00 -2.04191193e-01
8.39232802e-01 1.14888549e+00 -3.13145578e-01 -5.26384175e-01
-6.00730002e-01 1.03781617e+00 9.97763574e-01 9.09344494e-01
-6.70589328e-01 -7.02831566e-01 5.91193497e-01 -1.40186474e-01
4.76069778e-01 -1.63704306e-01 1.72350183e-01 -1.32792509e+00
1.85160190e-01 -3.36373299e-02 7.96104848e-01 -8.72671545e-01
-1.44587481e+00 4.87068862e-01 2.91969702e-02 -6.36638999e-01
-4.32767905e-02 -4.27952260e-01 -6.31145000e-01 6.48396671e-01
-1.72448826e+00 -1.31289423e+00 -3.01273346e-01 6.16737247e-01
-1.79698855e-01 -1.78524274e-02 1.24372709e+00 2.88108706e-01
-5.63805997e-01 6.36793256e-01 -8.64594281e-02 5.00969291e-01
5.81441283e-01 -1.43014169e+00 6.35078788e-01 6.93566918e-01
5.52950323e-01 9.20456946e-01 3.79063517e-01 -9.11919534e-01
-1.36111724e+00 -1.37487996e+00 1.56033194e+00 -6.21171176e-01
8.87460530e-01 -4.80872601e-01 -1.18462384e+00 1.45458376e+00
2.88518578e-01 5.94573975e-01 8.44767928e-01 7.89711177e-01
-7.22940087e-01 -1.37927204e-01 -7.38676965e-01 6.47032976e-01
1.40597057e+00 -7.06440330e-01 -1.12437475e+00 4.88106310e-01
1.04345953e+00 -2.49575302e-01 -1.24902856e+00 5.26717961e-01
1.94137990e-01 -3.29584420e-01 1.15848196e+00 -1.15196371e+00
2.66146392e-01 -4.25484359e-01 1.13489673e-01 -1.69883502e+00
-5.31445146e-01 -3.74243528e-01 -1.00608110e+00 1.34587646e+00
9.39755082e-01 -8.12131464e-01 8.29401433e-01 6.31811976e-01
1.41261280e-01 -7.23185539e-01 -6.88183606e-01 -9.36434865e-01
-1.24250405e-01 -8.40379521e-02 7.76660740e-01 1.66182649e+00
7.65440881e-01 9.43405330e-01 -1.12275757e-01 5.84154427e-01
5.39051533e-01 2.10454389e-01 7.38664746e-01 -1.30927837e+00
-6.44878373e-02 1.90102622e-01 -8.08681846e-01 -6.92403972e-01
6.22894704e-01 -1.30107343e+00 -3.89480412e-01 -2.31857538e+00
2.13615537e-01 -5.78313887e-01 -4.43161011e-01 8.13968360e-01
-6.33730590e-01 -1.77631333e-01 -2.77029157e-01 -4.02653068e-01
-7.98480630e-01 7.50512421e-01 1.10707128e+00 -3.66205692e-01
-2.26158023e-01 -4.42771345e-01 -6.91515267e-01 4.90850210e-01
6.97440922e-01 -3.35591882e-01 -8.82115185e-01 -4.23102915e-01
8.03044796e-01 -3.28016698e-01 1.88072056e-01 -5.66567481e-01
5.43247819e-01 6.59829602e-02 1.81096852e-01 -3.69227529e-01
2.41878286e-01 -8.81949306e-01 2.69158423e-01 5.74281160e-03
-2.19940901e-01 -8.10289606e-02 -7.46871578e-03 9.27914202e-01
-3.96095246e-01 -1.00587897e-01 -2.85706725e-02 1.43520892e-01
-1.17104113e+00 5.49313486e-01 1.69273466e-01 3.67961526e-01
1.16826296e+00 -5.52438647e-02 -6.27054453e-01 -1.55445114e-01
-1.25927210e+00 4.81986463e-01 7.03328177e-02 6.80788696e-01
9.84080255e-01 -1.70272267e+00 -5.95663071e-01 -4.91748303e-02
5.75164497e-01 4.49166775e-01 1.45421371e-01 3.75846624e-01
-1.94348365e-01 2.24461421e-01 1.77533522e-01 1.63817808e-01
-1.06785178e+00 1.10262859e+00 1.50442138e-01 -5.08816659e-01
-6.86723113e-01 9.73083317e-01 1.49118140e-01 -8.79889131e-01
2.76476324e-01 -4.47692633e-01 -4.64476138e-01 1.03124723e-01
2.68113732e-01 2.12553993e-01 -1.24622248e-02 -3.94291610e-01
-4.01677668e-01 3.98258418e-01 -3.29150289e-01 4.30671275e-01
1.30237067e+00 6.45942762e-02 -3.04876864e-01 4.43838954e-01
1.15952075e+00 -3.39759104e-02 -6.40052855e-01 -6.97277904e-01
1.76874608e-01 -3.94471288e-01 -4.30167168e-01 -6.21726453e-01
-9.77710843e-01 5.39404273e-01 -1.20735846e-01 2.32489228e-01
7.89735913e-01 3.30788404e-01 6.92777872e-01 7.86480010e-01
2.56266594e-01 -1.00213695e+00 -1.57566909e-02 6.27907038e-01
6.85341716e-01 -1.18067575e+00 3.12389731e-01 -9.41882968e-01
-4.87752020e-01 9.52476084e-01 8.19906354e-01 -6.45651370e-02
9.22512114e-01 -1.11774899e-01 -3.99880469e-01 -5.16085386e-01
-8.36063862e-01 -5.61910033e-01 4.24400836e-01 1.18589664e+00
2.36997053e-01 3.26346666e-01 -6.95330799e-02 4.75709349e-01
-2.19818912e-02 4.35375609e-02 4.35135305e-01 7.96536863e-01
-2.07477570e-01 -1.27817178e+00 1.95617780e-01 5.99395812e-01
8.52233544e-02 -3.90248299e-01 -6.27252817e-01 9.12820458e-01
3.72252643e-01 9.50818479e-01 1.22260200e-02 -5.95202684e-01
5.35796106e-01 2.02386260e-01 5.13140500e-01 -1.06729722e+00
-5.37719503e-02 -1.01033103e+00 5.05226314e-01 -1.54950306e-01
-3.78830910e-01 -1.99010015e-01 -1.64562583e+00 -2.19334394e-01
-7.84422278e-01 5.16475320e-01 2.51696426e-02 9.63854134e-01
5.45113325e-01 8.25945139e-01 2.79780060e-01 -1.22558780e-01
2.51511604e-01 -8.22024047e-01 -7.76593864e-01 6.40700698e-01
9.74789932e-02 -9.13613439e-01 -1.24892041e-01 2.73364127e-01] | [8.862540245056152, 8.0051851272583] |
1760168d-2d13-41d5-a819-d97447ad3e2b | align-perturb-and-decouple-toward-better | 2305.18714 | null | https://arxiv.org/abs/2305.18714v1 | https://arxiv.org/pdf/2305.18714v1.pdf | Align, Perturb and Decouple: Toward Better Leverage of Difference Information for RSI Change Detection | Change detection is a widely adopted technique in remote sense imagery (RSI) analysis in the discovery of long-term geomorphic evolution. To highlight the areas of semantic changes, previous effort mostly pays attention to learning representative feature descriptors of a single image, while the difference information is either modeled with simple difference operations or implicitly embedded via feature interactions. Nevertheless, such difference modeling can be noisy since it suffers from non-semantic changes and lacks explicit guidance from image content or context. In this paper, we revisit the importance of feature difference for change detection in RSI, and propose a series of operations to fully exploit the difference information: Alignment, Perturbation and Decoupling (APD). Firstly, alignment leverages contextual similarity to compensate for the non-semantic difference in feature space. Next, a difference module trained with semantic-wise perturbation is adopted to learn more generalized change estimators, which reversely bootstraps feature extraction and prediction. Finally, a decoupled dual-decoder structure is designed to predict semantic changes in both content-aware and content-agnostic manners. Extensive experiments are conducted on benchmarks of LEVIR-CD, WHU-CD and DSIFN-CD, demonstrating our proposed operations bring significant improvement and achieve competitive results under similar comparative conditions. Code is available at https://github.com/wangsp1999/CD-Research/tree/main/openAPD | ['Wenbing Zhu', 'Chengjie Wang', 'Yabiao Wang', 'Mingmin Chi', 'Ming Xie', 'Yuxi Li', 'Supeng Wang'] | 2023-05-30 | null | null | null | null | ['change-detection'] | ['computer-vision'] | [ 4.54528153e-01 -3.60253662e-01 -4.44722641e-03 -4.92233664e-01
-7.68578529e-01 -5.11913419e-01 7.97164500e-01 5.88076711e-02
-2.19297320e-01 4.18381870e-01 5.65511942e-01 3.42363268e-02
-3.57847139e-02 -8.11338902e-01 -5.97656369e-01 -8.73098195e-01
-2.67898768e-01 -1.90114558e-01 4.33325917e-01 -4.30871755e-01
3.52666914e-01 2.97462881e-01 -1.57557714e+00 2.29961774e-03
8.90882492e-01 1.01334536e+00 4.32982236e-01 4.67503309e-01
3.22274379e-02 5.72326779e-01 -3.45879905e-02 -6.24029152e-03
4.94490862e-01 -4.13707018e-01 -7.64309943e-01 1.37681484e-01
3.01739872e-01 -2.81373739e-01 -4.99738336e-01 1.35103607e+00
7.77202666e-01 4.64823842e-02 5.46849012e-01 -8.65386665e-01
-5.54802299e-01 3.77773017e-01 -1.00067770e+00 5.00546753e-01
1.56085044e-01 2.55831659e-01 1.13638163e+00 -1.11007273e+00
6.65450454e-01 1.15502632e+00 7.26254344e-01 -8.35807398e-02
-9.61476803e-01 -5.50726295e-01 4.44570065e-01 5.38108170e-01
-1.45135152e+00 -6.50152087e-01 9.47063327e-01 -5.04236460e-01
7.38524199e-01 3.62934411e-01 5.42209506e-01 6.59100831e-01
8.83765072e-02 9.38869536e-01 1.05516088e+00 -1.99670300e-01
1.04942784e-01 -1.81221902e-01 -2.90817022e-02 3.57157558e-01
1.71217304e-02 6.84835538e-02 -5.17138183e-01 1.43838182e-01
4.04040426e-01 1.11205719e-01 -6.76387727e-01 -1.01220161e-01
-1.14572394e+00 6.85349464e-01 7.00102389e-01 2.82787025e-01
-4.22414631e-01 2.37457994e-02 4.19278741e-01 4.94419605e-01
9.06287074e-01 2.10289329e-01 -5.63444614e-01 -1.72050998e-01
-9.74358141e-01 2.74144262e-01 3.90675217e-01 7.89927363e-01
1.28141451e+00 -7.33915046e-02 4.08842489e-02 9.99457598e-01
2.54888356e-01 6.96321547e-01 6.29483700e-01 -5.33490300e-01
3.33956987e-01 4.08154219e-01 -9.94355306e-02 -1.34300578e+00
-4.40763712e-01 -5.95796525e-01 -7.43185878e-01 -3.55743468e-02
-1.50828630e-01 1.08835258e-01 -7.78151214e-01 1.72721839e+00
5.76988637e-01 3.74675184e-01 -1.47377744e-01 9.07150626e-01
6.88128114e-01 5.82576811e-01 -1.55001640e-01 6.79448154e-03
1.28486347e+00 -8.37200582e-01 -3.48907173e-01 -5.21281660e-01
5.60130656e-01 -6.14417493e-01 1.00986910e+00 -1.01692334e-01
-6.60217345e-01 -3.56148064e-01 -1.03635418e+00 4.82673477e-03
-4.00981277e-01 -8.85612443e-02 2.98487216e-01 3.91899571e-02
-1.03380120e+00 5.71681142e-01 -7.50147581e-01 -5.83612084e-01
4.43227381e-01 -1.80778541e-02 -2.92404890e-01 -4.41817380e-02
-1.34682870e+00 6.92957819e-01 3.94675940e-01 3.05605024e-01
-7.09412158e-01 -9.17284071e-01 -8.59619260e-01 -1.81508884e-01
4.47552562e-01 -3.39134842e-01 9.43520725e-01 -1.05950141e+00
-1.37362528e+00 9.62426603e-01 -1.65869713e-01 -3.86582017e-01
6.12533867e-01 -1.54953703e-01 -5.43235123e-01 1.09208927e-01
2.43147507e-01 6.96986318e-01 9.53261077e-01 -1.31986320e+00
-9.79939938e-01 -4.64235187e-01 7.19438940e-02 5.17542005e-01
-2.18626559e-01 -5.57060465e-02 -7.36361325e-01 -8.50307941e-01
4.73031074e-01 -1.04657733e+00 -9.93621722e-02 1.15781978e-01
-2.28119105e-01 3.66386563e-01 8.11978579e-01 -9.63128746e-01
1.43633711e+00 -2.34439254e+00 9.17210206e-02 1.48096159e-01
7.79362842e-02 6.81351200e-02 -3.29341054e-01 5.52543819e-01
-8.25556964e-02 5.44859469e-02 -7.90170789e-01 -5.39896786e-02
-2.37139374e-01 1.59973472e-01 -4.16096568e-01 6.24645472e-01
3.01154196e-01 8.55661750e-01 -1.01424205e+00 -2.57568121e-01
1.57049999e-01 1.49224356e-01 -6.47055387e-01 -7.44978935e-02
-7.28402287e-02 6.28219485e-01 -3.62537116e-01 8.71706009e-01
1.05049396e+00 -5.83598949e-02 6.71700388e-02 -4.61652547e-01
-3.58822703e-01 2.15886489e-01 -1.25167120e+00 1.76713920e+00
-4.14131671e-01 5.59911430e-01 -1.58010766e-01 -1.11461377e+00
7.02549458e-01 -3.14414650e-01 4.44661170e-01 -8.52844834e-01
-2.21932188e-01 2.50394911e-01 -1.58928558e-01 -5.83759606e-01
6.83590233e-01 -1.49301416e-03 -1.06539302e-01 1.68579742e-01
-2.73593038e-01 -1.14871949e-01 -6.59182742e-02 4.95598800e-02
1.08151948e+00 2.69809037e-01 5.64872980e-01 -4.56204861e-01
6.65038824e-01 -2.44751815e-02 8.25459003e-01 5.95121682e-01
-3.32490474e-01 7.59898245e-01 1.33989230e-01 -2.91320771e-01
-7.09359884e-01 -9.27236199e-01 -2.68318534e-01 1.01128411e+00
6.83076978e-01 -2.89674461e-01 -1.82041347e-01 -6.17800951e-01
1.23254798e-01 5.86290002e-01 -7.03800976e-01 -3.30259442e-01
-5.50743103e-01 -1.09243953e+00 3.48268718e-01 2.70194441e-01
9.21388745e-01 -8.65536630e-01 -5.64116478e-01 2.31292218e-01
-3.24644476e-01 -8.18985164e-01 -4.46249366e-01 -2.25759647e-03
-8.42684865e-01 -8.51726174e-01 -3.92589003e-01 -5.15791118e-01
5.12591898e-01 8.19971740e-01 7.60430634e-01 -1.21922925e-01
-3.65388215e-01 3.08848590e-01 -6.51794732e-01 -4.10386138e-02
-3.75827067e-02 1.29720524e-01 -2.31957585e-01 1.35992616e-01
2.58925468e-01 -9.27508235e-01 -1.05357194e+00 2.84823328e-01
-1.14056635e+00 1.74514920e-01 7.03371406e-01 9.28287983e-01
5.64418793e-01 9.43896696e-02 3.76188725e-01 -8.44510257e-01
3.23099941e-02 -7.94826984e-01 -2.85388410e-01 1.86056837e-01
-7.42980242e-01 7.21033290e-02 1.87146381e-01 -1.09348193e-01
-1.29825282e+00 -1.19316183e-01 -1.04363091e-01 -4.08656374e-02
-1.22970454e-02 9.57572997e-01 -3.82878184e-01 -1.98767963e-03
4.76846427e-01 6.39625669e-01 -1.65603518e-01 -5.59998870e-01
4.05462086e-01 7.63031304e-01 6.67150140e-01 -3.11637372e-01
1.07334757e+00 8.51350129e-01 -4.49346960e-01 -8.86916518e-01
-7.19844639e-01 -7.18827784e-01 -5.97875953e-01 -2.13157460e-01
3.73195678e-01 -1.33038831e+00 2.21922293e-01 9.64765012e-01
-6.48798764e-01 -3.78200948e-01 -2.32262194e-01 2.10326165e-01
-2.14767218e-01 5.95812142e-01 -4.03843641e-01 -2.82800376e-01
-4.65971917e-01 -9.82592285e-01 1.13846219e+00 2.44086027e-01
1.47387072e-01 -1.00164557e+00 2.88873017e-01 1.22032855e-02
6.55532300e-01 2.63751507e-01 7.73997068e-01 -3.52933347e-01
-4.80754733e-01 -1.32361293e-01 -3.90435070e-01 3.05754691e-01
6.43674612e-01 -8.88348520e-02 -1.02176952e+00 -5.16009986e-01
-8.99284407e-02 5.76631241e-02 1.14154899e+00 1.69472784e-01
1.14541554e+00 -3.19358408e-01 -2.86220551e-01 1.01714718e+00
1.72413087e+00 -1.37406273e-03 7.60335147e-01 7.10575819e-01
7.10648537e-01 4.23251212e-01 8.02366972e-01 9.00003731e-01
7.74911106e-01 7.11135089e-01 5.87708592e-01 -1.30349817e-02
-3.51841152e-01 -8.02291632e-02 5.41445792e-01 9.41382647e-01
8.88200626e-02 -2.59864014e-02 -9.03033674e-01 7.40117669e-01
-1.76726413e+00 -9.68308866e-01 4.22521774e-03 2.16793561e+00
1.03526485e+00 -1.21335544e-01 -2.87876189e-01 5.19675855e-03
7.39826918e-01 7.59740591e-01 -8.52437973e-01 3.05093616e-01
-4.38685209e-01 -6.54513091e-02 7.94570625e-01 5.16719341e-01
-1.26295018e+00 1.13585961e+00 4.90565681e+00 8.61021578e-01
-1.49418962e+00 3.50538105e-01 2.66774029e-01 3.87749895e-02
-4.85835761e-01 3.28993738e-01 -5.94922900e-01 6.19693637e-01
5.00794291e-01 -1.85859188e-01 2.46991485e-01 5.85134983e-01
3.91565770e-01 -3.45504880e-01 -6.05473578e-01 9.14072931e-01
3.98454443e-02 -1.12242973e+00 -5.93222259e-03 -9.70768481e-02
7.40770221e-01 5.14008105e-01 -9.63091291e-03 1.70360953e-01
5.31620309e-02 -3.29637289e-01 9.72635329e-01 6.13988221e-01
8.03569078e-01 -4.65274006e-01 5.62268555e-01 -3.45372595e-02
-1.44105458e+00 -5.48664555e-02 -3.27907592e-01 -1.06357999e-01
-5.20624220e-02 7.73152590e-01 -6.04997694e-01 1.03124070e+00
9.15972054e-01 1.44612610e+00 -8.99855554e-01 9.26084697e-01
-2.51215518e-01 7.83962846e-01 -2.56550103e-01 5.56606889e-01
2.34131157e-01 -2.62192845e-01 8.21343958e-01 1.32420516e+00
4.31278318e-01 2.98339650e-02 6.00297339e-02 4.15242940e-01
9.22924280e-02 9.81716365e-02 -3.89126003e-01 3.65257800e-01
6.83958292e-01 1.28340328e+00 -5.39677083e-01 -2.13393196e-01
-6.00428164e-01 1.34569919e+00 8.98922235e-02 2.12938339e-01
-8.48124385e-01 -2.80697674e-01 1.03237772e+00 1.25825003e-01
5.97221971e-01 -1.04642063e-01 -1.35449052e-01 -1.54868925e+00
1.67032942e-01 -7.95432925e-01 3.78125370e-01 -6.62576437e-01
-1.20189440e+00 3.41082573e-01 5.48775028e-03 -1.67775381e+00
1.45869944e-02 -1.73529908e-01 -7.49088943e-01 7.19520926e-01
-2.18592119e+00 -1.29731345e+00 -7.62557924e-01 3.56967837e-01
6.50315464e-01 1.68582231e-01 3.47721010e-01 3.98433656e-01
-7.61833489e-01 5.04185140e-01 5.60542464e-01 -1.20201357e-01
9.14506793e-01 -1.13263690e+00 5.30443192e-01 1.31104743e+00
-2.14121982e-01 2.57482797e-01 7.26368189e-01 -6.74451172e-01
-1.35946059e+00 -1.44868433e+00 5.00264764e-01 5.45168445e-02
8.24355781e-01 -1.79162214e-03 -1.15401804e+00 4.81843650e-01
-9.77576450e-02 1.42349213e-01 4.69836026e-01 -3.35127443e-01
-3.96456331e-01 -3.80406231e-01 -8.76478314e-01 5.80909550e-01
1.37637579e+00 -6.52970314e-01 -4.19327557e-01 8.29773992e-02
5.70043802e-01 -3.49561185e-01 -9.16733265e-01 5.33659995e-01
4.49458241e-01 -9.12723064e-01 8.78027081e-01 -7.53082186e-02
4.67599601e-01 -7.00508058e-01 -5.49394190e-01 -1.35840142e+00
-4.52691287e-01 -4.22161788e-01 2.26990446e-01 1.42608392e+00
1.28024548e-01 -8.02565277e-01 2.31369510e-01 1.65211082e-01
-3.27619940e-01 -6.17118478e-01 -7.65906870e-01 -7.29029059e-01
-3.81837296e-03 -4.84732687e-01 7.78718054e-01 1.09931242e+00
-3.26804876e-01 2.63566561e-02 -5.49140930e-01 5.26670337e-01
3.34734261e-01 4.42606270e-01 7.32758522e-01 -1.02926683e+00
-2.46238202e-01 -4.86878783e-01 -5.78490973e-01 -1.16297448e+00
-3.28426994e-02 -1.05208910e+00 3.10935259e-01 -1.33892906e+00
2.22545907e-01 -4.17312950e-01 -3.02576602e-01 5.45587778e-01
-5.22977591e-01 2.51365930e-01 2.25944426e-02 7.18695164e-01
-3.32483023e-01 8.75688076e-01 8.69104028e-01 -2.26492137e-01
-1.88961178e-01 -2.69481152e-01 -7.70784497e-01 7.04979122e-01
7.34819651e-01 -4.78715509e-01 -4.05063391e-01 -6.22644842e-01
2.92892307e-01 -3.98089260e-01 4.58604157e-01 -1.11665690e+00
1.73845708e-01 -1.74565241e-01 9.93174911e-02 -5.90023994e-01
-1.35985747e-01 -5.69206178e-01 2.41124406e-01 4.51728851e-01
-1.06454365e-01 -1.83184803e-01 -2.30914317e-02 6.91288173e-01
-5.34363091e-01 -1.51809603e-01 8.90996516e-01 -1.22886389e-01
-1.48960519e+00 5.40230691e-01 -1.02456659e-01 1.28316119e-01
7.68832386e-01 -2.74492323e-01 -2.56538630e-01 -2.48631731e-01
-2.94768631e-01 3.03460509e-01 7.63690412e-01 5.15760005e-01
5.42602301e-01 -1.18565369e+00 -8.97872984e-01 1.26988381e-01
5.82470357e-01 7.93592110e-02 6.87938929e-01 1.06011915e+00
-4.57200289e-01 -1.61931857e-01 -7.33616063e-03 -6.20111465e-01
-9.50022101e-01 2.00656071e-01 3.87930751e-01 -9.60559249e-02
-8.66871476e-01 7.99633324e-01 4.45355892e-01 -3.77488971e-01
-3.16904753e-01 -2.22434431e-01 -1.71973091e-02 4.89421725e-01
5.35821378e-01 1.81863666e-01 2.35284701e-01 -8.40381086e-01
-5.29041529e-01 8.07263911e-01 -1.26956463e-01 -2.03462914e-02
1.66378415e+00 -8.57316136e-01 -1.24233663e-01 3.72888356e-01
1.34070110e+00 -4.98404168e-02 -1.52663302e+00 -9.54401553e-01
1.88935101e-01 -7.49310076e-01 3.23224008e-01 -6.39380276e-01
-1.27301002e+00 6.79227233e-01 8.57627273e-01 -2.81573027e-01
1.60349059e+00 -3.47028747e-02 7.70492136e-01 2.62962908e-01
2.95129448e-01 -1.03934383e+00 -9.28391516e-02 5.42368948e-01
1.09729302e+00 -1.47279811e+00 1.87027842e-01 -2.28063598e-01
-7.57672191e-01 9.23574805e-01 4.16362524e-01 -2.45191273e-03
8.70435655e-01 6.97103050e-03 -1.52568780e-02 -2.24594355e-01
-5.51312089e-01 -5.54522932e-01 1.28789082e-01 3.13831955e-01
6.37085885e-02 2.79320553e-02 -2.27155700e-01 2.71132141e-01
7.99056236e-03 -1.79110304e-01 4.22320902e-01 1.20445275e+00
-5.90335131e-01 -7.47490942e-01 -1.12896226e-01 5.10920644e-01
-1.02324247e-01 -3.99876058e-01 2.93064658e-02 6.40696824e-01
8.31231996e-02 6.54286981e-01 1.15002528e-01 -5.90266645e-01
2.06020579e-01 -1.41342789e-01 9.58177354e-03 -4.05110687e-01
-2.43133813e-01 2.21484527e-02 -7.56752789e-02 -6.41330600e-01
-6.11414075e-01 -1.13501275e+00 -1.05082822e+00 -1.56292319e-01
-2.74816573e-01 -2.14020178e-01 5.24625719e-01 8.50724101e-01
6.85759902e-01 2.67090261e-01 1.08223999e+00 -9.14551198e-01
-4.68846947e-01 -1.14379334e+00 -6.48877025e-01 3.71060967e-01
3.98497730e-01 -7.15417027e-01 -6.25270426e-01 2.01222852e-01] | [9.683207511901855, -1.2590593099594116] |
643971c0-ac9d-44f6-849d-546ef492cef4 | a-unified-view-of-deep-learning-for-reaction | 2306.15890 | null | https://arxiv.org/abs/2306.15890v1 | https://arxiv.org/pdf/2306.15890v1.pdf | A Unified View of Deep Learning for Reaction and Retrosynthesis Prediction: Current Status and Future Challenges | Reaction and retrosynthesis prediction are fundamental tasks in computational chemistry that have recently garnered attention from both the machine learning and drug discovery communities. Various deep learning approaches have been proposed to tackle these problems, and some have achieved initial success. In this survey, we conduct a comprehensive investigation of advanced deep learning-based models for reaction and retrosynthesis prediction. We summarize the design mechanisms, strengths, and weaknesses of state-of-the-art approaches. Then, we discuss the limitations of current solutions and open challenges in the problem itself. Finally, we present promising directions to facilitate future research. To our knowledge, this paper is the first comprehensive and systematic survey that seeks to provide a unified understanding of reaction and retrosynthesis prediction. | ['Irwin King', 'Yang Yu', 'Peilin Zhao', 'Ziqiao Meng'] | 2023-06-28 | null | null | null | null | ['drug-discovery', 'retrosynthesis'] | ['medical', 'medical'] | [ 2.33620003e-01 -1.72337964e-01 -7.32656837e-01 -5.00797778e-02
-4.28964227e-01 -7.32759953e-01 8.52214873e-01 5.43078363e-01
-3.02414298e-01 8.68668199e-01 9.93523002e-02 -6.72480881e-01
4.35351208e-03 -6.43324375e-01 -3.94891858e-01 -1.00030005e+00
-3.66075486e-02 2.97492146e-01 -6.77294433e-02 -1.67309374e-01
5.57258010e-01 8.36837888e-01 -1.16160655e+00 5.53071916e-01
5.28213203e-01 9.34881330e-01 9.00924206e-02 5.18822074e-01
-1.29721791e-01 1.17210400e+00 -4.22371209e-01 -5.60981095e-01
3.77870984e-02 -3.66957009e-01 -1.00192332e+00 -4.74308878e-01
-7.65380636e-02 -2.80505847e-02 -5.32188416e-01 6.23295248e-01
9.15690541e-01 2.30349690e-01 9.69884157e-01 -8.42466414e-01
-5.82918346e-01 5.86709678e-01 -3.02410632e-01 3.83723885e-01
2.43353710e-01 1.62900046e-01 1.25340629e+00 -1.10509980e+00
6.18328273e-01 8.02975416e-01 4.58797425e-01 7.07870603e-01
-9.64430392e-01 -8.79568279e-01 2.84612685e-01 4.24104810e-01
-1.30756938e+00 -5.40666461e-01 6.98859274e-01 -6.28392935e-01
1.56785107e+00 6.21806085e-02 6.11654282e-01 1.11513126e+00
5.43118775e-01 8.76429319e-01 6.47220254e-01 -1.72831446e-01
3.44254166e-01 -2.10794955e-01 -3.44223857e-01 6.57008290e-01
9.57332626e-02 1.99239925e-01 -4.74265158e-01 3.15562002e-02
5.57110369e-01 -6.20982908e-02 2.80312523e-02 -3.47689629e-01
-1.08028865e+00 1.28946030e+00 3.77842486e-01 5.11975110e-01
-3.71732414e-01 1.91843614e-01 5.87944567e-01 -2.40620419e-01
5.09009004e-01 9.40345287e-01 -9.73042190e-01 1.20279426e-02
-8.18447590e-01 4.43350762e-01 7.96929061e-01 6.19052410e-01
2.90139347e-01 1.60707772e-01 1.53538942e-01 6.68135703e-01
1.83679163e-01 -4.56339687e-01 2.26277024e-01 -6.63689256e-01
5.07719517e-02 1.02649644e-01 5.35967089e-02 -6.24619782e-01
-6.71441615e-01 -3.62680942e-01 -7.90157080e-01 7.87089244e-02
4.23867196e-01 -1.06857635e-01 -6.96891546e-01 1.26924515e+00
2.69734502e-01 -2.21425742e-01 1.53583318e-01 5.23028553e-01
1.18768382e+00 1.02720106e+00 6.52130961e-01 -5.43779135e-01
1.05210185e+00 -1.30116355e+00 -7.49524295e-01 2.10578710e-01
6.96480870e-01 -1.24689722e+00 2.28840575e-01 7.99343586e-01
-1.24301255e+00 -2.37764791e-01 -1.08463383e+00 -4.91846919e-01
-8.45938027e-01 -5.84270768e-02 1.23708415e+00 8.14928770e-01
-4.13859010e-01 1.00976956e+00 -8.88849676e-01 -2.83409834e-01
4.96144831e-01 5.79385519e-01 -2.50710454e-02 1.87096685e-01
-1.09309638e+00 1.13320029e+00 6.80921614e-01 -8.74876156e-02
-1.23880756e+00 -9.47504163e-01 -3.92768234e-01 -6.27645627e-02
5.85703969e-01 -5.41460574e-01 1.83457208e+00 -3.96943033e-01
-1.64367032e+00 7.79757679e-01 -6.53981790e-02 -2.35029265e-01
1.53453678e-01 2.56612241e-01 -3.33528131e-01 -2.21565709e-01
-2.88586408e-01 7.59076416e-01 9.16819274e-02 -1.03057921e+00
-6.90671682e-01 -2.40566358e-01 -2.78276578e-02 4.74548601e-02
-2.92249285e-02 4.29690987e-01 -4.96431142e-02 -7.66650140e-01
-1.59180820e-01 -7.11637139e-01 -6.19580448e-01 -1.30376652e-01
-4.74905103e-01 -6.60417199e-01 2.46827215e-01 -1.63039818e-01
1.25328600e+00 -1.32200742e+00 4.22648579e-01 -2.45417491e-01
2.76297629e-01 4.68351662e-01 -1.23364534e-02 1.10200930e+00
-7.11329460e-01 4.63063151e-01 1.00393325e-01 -1.02974363e-01
-1.93568572e-01 -1.61062747e-01 -2.03303307e-01 5.44448018e-01
-1.38601556e-03 9.06896591e-01 -1.02650809e+00 4.91322428e-02
3.23925465e-01 3.23772460e-01 -3.40510994e-01 -5.32174446e-02
-7.49507487e-01 5.92583179e-01 -6.56794012e-01 9.65615809e-01
2.14389041e-01 -2.62125015e-01 4.36947048e-01 -4.73954588e-01
-9.10015881e-01 9.32581246e-01 -7.50823140e-01 1.66793740e+00
-2.11565420e-01 2.51962394e-01 -3.26252490e-01 -1.44513440e+00
8.31640661e-01 6.53862059e-01 8.32812786e-01 -5.05474746e-01
3.65563393e-01 3.89608771e-01 3.41399640e-01 -2.45255083e-01
4.27996218e-01 -7.03403711e-01 1.76283240e-01 3.37584555e-01
1.32029831e-01 -1.49202302e-01 4.71398115e-01 -1.21087976e-01
5.98656356e-01 1.88197643e-01 1.00303066e+00 1.41906431e-02
8.63294542e-01 6.87013194e-02 4.46099252e-01 4.89067167e-01
-2.20728785e-01 1.85361072e-01 3.55714679e-01 -1.15845454e+00
-1.26739299e+00 -5.10136724e-01 -7.73879886e-02 1.10585594e+00
-2.58027822e-01 -9.37738955e-01 -4.53299999e-01 -6.37704670e-01
-4.56101507e-01 3.00944179e-01 -4.55348015e-01 -8.11343547e-03
-4.49967742e-01 -1.10512245e+00 5.74249804e-01 6.15501702e-01
1.10653624e-01 -1.02341008e+00 1.39440298e-01 5.53960145e-01
2.61270218e-02 -9.04814005e-01 -4.96341474e-02 5.50683558e-01
-7.33260274e-01 -1.37403190e+00 -4.73174959e-01 -9.87914443e-01
-4.52806652e-02 2.62726337e-01 7.54998863e-01 -2.14968443e-01
-6.52721882e-01 -1.41901389e-01 -1.15535058e-01 -8.14466894e-01
-5.86903691e-01 3.14446867e-01 -5.00417873e-02 -5.88238537e-01
3.53271157e-01 -3.35351348e-01 -6.66404724e-01 2.42848005e-02
-6.28115594e-01 -8.78207088e-02 4.10892725e-01 5.73565304e-01
7.37718046e-01 2.16789424e-01 5.21954775e-01 -5.82032800e-01
5.00998139e-01 -6.11515224e-01 -6.65412843e-01 1.53600752e-01
-8.09181571e-01 5.13191596e-02 7.76790440e-01 -1.81855753e-01
-1.00270700e+00 2.30483696e-01 -6.55077696e-01 1.72760829e-01
-1.97807342e-01 8.42215657e-01 -1.31327212e-01 -2.85945922e-01
5.33407331e-01 1.41994566e-01 -3.81898433e-01 -7.22226441e-01
5.06491184e-01 3.14761519e-01 -1.21737070e-01 -7.45026648e-01
3.43397141e-01 2.09036365e-01 4.69289392e-01 -8.78823221e-01
-1.01231182e+00 -5.11639774e-01 -4.49918956e-01 1.34307444e-01
1.16099179e+00 -7.94460833e-01 -1.11904442e+00 4.47786331e-01
-1.34321570e+00 -2.55888999e-01 1.70084074e-01 3.63000929e-01
-6.38379693e-01 5.54647326e-01 -9.77743566e-01 -5.97069681e-01
-5.80267131e-01 -1.47906530e+00 6.93263173e-01 1.36961550e-01
-3.19670826e-01 -1.25882435e+00 2.63930649e-01 3.86169583e-01
-6.95958659e-02 3.08832437e-01 1.02278471e+00 -7.86741316e-01
-6.71201408e-01 -2.41929963e-02 1.06856324e-01 -8.30201358e-02
4.49881196e-01 2.62570590e-01 -9.62058961e-01 -1.19762138e-01
-2.72504389e-01 -5.81514657e-01 1.12288940e+00 6.35796845e-01
1.51428723e+00 -8.27422068e-02 -6.87681019e-01 4.94113266e-01
1.21239531e+00 9.02002156e-01 4.54131037e-01 3.31167728e-01
6.46556377e-01 5.71127117e-01 6.31995857e-01 5.73812604e-01
1.63015038e-01 4.12630707e-01 7.06396401e-01 -2.53195226e-01
1.11103237e-01 -2.80083477e-01 6.99437857e-02 4.39985096e-01
-6.07819796e-01 -7.09266722e-01 -9.00196433e-01 4.90279645e-02
-1.62630439e+00 -9.88635123e-01 -1.38697982e-01 1.90657175e+00
8.94149065e-01 -3.77948023e-02 3.87330174e-01 3.58035743e-01
4.68200326e-01 3.86808306e-01 -6.00293100e-01 -6.61137044e-01
8.31387341e-02 7.42363751e-01 4.28690076e-01 3.99459988e-01
-1.36244023e+00 1.50866973e+00 7.71999979e+00 9.29301858e-01
-1.37254882e+00 -2.20103160e-01 8.96758735e-01 6.67041019e-02
-1.03572860e-01 7.29204193e-02 -8.14508319e-01 -9.11030099e-02
8.95885766e-01 -2.79384285e-01 3.64257514e-01 7.07477152e-01
4.09154713e-01 3.59877348e-01 -1.52905524e+00 7.77646303e-01
-2.92846829e-01 -2.25764155e+00 1.03324249e-01 1.86655268e-01
6.55761719e-01 -3.48220058e-02 2.86300749e-01 8.83775279e-02
2.20410764e-01 -1.27280378e+00 5.25661111e-01 -8.02165568e-02
5.49949408e-01 -8.71196508e-01 2.40741059e-01 6.54685199e-02
-1.36000049e+00 -2.60316759e-01 -2.70587772e-01 -4.18467164e-01
-4.80701327e-02 4.94340032e-01 -8.31936598e-01 7.50287533e-01
2.95115620e-01 1.21336687e+00 3.13274637e-02 9.20727789e-01
-5.01748860e-01 3.37984979e-01 1.62831262e-01 -3.64914030e-01
6.59002185e-01 -3.06354225e-01 -2.11821292e-02 1.21059024e+00
1.88332751e-01 4.05145347e-01 5.08606791e-01 8.86462748e-01
-2.55153537e-01 2.93804556e-01 -3.13331544e-01 -7.96899915e-01
3.94933373e-01 1.11446786e+00 -7.60522902e-01 -3.57010573e-01
-8.25147331e-01 4.25745219e-01 9.94408131e-02 2.23636732e-01
-8.64688456e-01 -3.42371799e-02 1.12154806e+00 -2.03942254e-01
2.09780395e-01 -3.21073353e-01 -3.47498775e-01 -1.04412806e+00
-7.39298403e-01 -8.47989798e-01 6.56600952e-01 -3.71599197e-01
-9.86344337e-01 -1.31999344e-01 -2.80384600e-01 -6.84537947e-01
3.37736517e-01 -1.01665652e+00 -5.80484092e-01 5.32265604e-01
-1.47871375e+00 -8.55491996e-01 4.31668460e-01 1.63323462e-01
9.54281747e-01 -3.13800305e-01 1.10962915e+00 2.89525151e-01
-1.04668570e+00 1.89304173e-01 4.51579928e-01 -3.68167758e-01
7.20151484e-01 -1.03632963e+00 5.59610307e-01 2.74618864e-01
-1.24662489e-01 7.29515433e-01 8.09646785e-01 -4.32736754e-01
-1.48628891e+00 -1.07846916e+00 1.10046530e+00 -1.53699607e-01
8.35141838e-01 -2.70551622e-01 -4.56696182e-01 6.74491465e-01
1.71409532e-01 -3.09427619e-01 1.32706499e+00 5.67555837e-02
-3.00622135e-01 5.72622903e-02 -9.55931365e-01 7.61389673e-01
7.71176696e-01 -3.16469580e-01 -9.48778167e-02 6.32679820e-01
4.87410873e-01 -2.58476347e-01 -1.20947206e+00 2.63047785e-01
6.42633617e-01 -7.90445447e-01 1.38532794e+00 -8.97391081e-01
6.98721051e-01 -5.25613204e-02 2.26014167e-01 -9.98237848e-01
-7.31986761e-01 -7.40299463e-01 -1.21822879e-01 7.18593180e-01
5.69529653e-01 -1.63369238e-01 1.00054908e+00 3.48860502e-01
-6.64329112e-01 -1.12307477e+00 -4.79970187e-01 -5.97438216e-01
7.87304282e-01 -2.06579462e-01 3.66309434e-01 1.07030034e+00
3.70737791e-01 7.40967512e-01 -4.37457800e-01 -2.81852424e-01
2.63510108e-01 1.69716939e-01 3.45197111e-01 -1.27913260e+00
-2.59370971e-02 -7.49586344e-01 7.10376874e-02 -1.05598938e+00
2.09700316e-01 -1.04998362e+00 -3.77638012e-01 -1.66921365e+00
4.09817457e-01 1.37432128e-01 -4.55257177e-01 5.41286111e-01
8.30669180e-02 1.04910582e-01 -1.30599648e-01 -1.14013284e-01
-5.51952481e-01 6.41864359e-01 1.37712979e+00 -4.79693145e-01
-1.81209058e-01 -2.69994766e-01 -9.26109493e-01 6.73934042e-01
1.37917888e+00 -2.90470809e-01 -3.12000513e-01 -5.02426587e-02
4.77764666e-01 -1.85106546e-02 -4.09415305e-01 -4.44245517e-01
1.30256340e-01 -6.82119191e-01 5.58265924e-01 -8.49687874e-01
3.56347203e-01 -2.80759633e-01 1.35117490e-02 7.47104049e-01
-4.72237110e-01 -3.08926165e-01 2.05780134e-01 4.16276723e-01
3.14333066e-02 -8.03340450e-02 9.38395679e-01 -7.39688396e-01
-5.58131516e-01 8.89201462e-01 -1.25808847e+00 -5.51161766e-01
1.07935178e+00 3.36144003e-03 -1.56855494e-01 -3.97631899e-03
-8.49182308e-01 1.81814730e-02 1.01342954e-01 3.83521140e-01
4.52834517e-01 -9.40001845e-01 -1.76478118e-01 -5.07850587e-01
1.49474427e-01 -8.23910013e-02 -1.54822528e-01 6.46592200e-01
-5.39844096e-01 1.12832522e+00 9.10046976e-03 1.37522966e-01
-1.32444572e+00 1.29714441e+00 5.49751043e-01 -2.84383565e-01
9.55737010e-02 7.97169268e-01 5.05798221e-01 -1.85802311e-01
4.88505870e-01 -3.79938573e-01 -3.64345342e-01 4.43698876e-02
4.28942531e-01 5.31345487e-01 3.00500631e-01 -2.75446266e-01
-4.65867430e-01 3.28946292e-01 -5.43954790e-01 5.49238443e-01
1.34897077e+00 3.36250663e-01 -1.22600831e-01 4.84824777e-02
1.02939034e+00 -4.50795799e-01 -7.82096803e-01 -2.01469474e-02
8.87845457e-02 2.26106703e-01 2.88311154e-01 -9.59792972e-01
-8.51935506e-01 1.17646921e+00 2.34729834e-02 -7.12115467e-02
7.84364998e-01 6.72904029e-02 8.75265837e-01 6.69298947e-01
-3.70957553e-02 -9.74370301e-01 4.65042025e-01 6.54852748e-01
6.75661564e-01 -1.19069111e+00 3.44906032e-01 -5.92731476e-01
-1.82193801e-01 1.63166821e+00 5.20147681e-01 4.28120553e-01
5.84130526e-01 -9.04024229e-04 -5.10147393e-01 -2.38705248e-01
-7.85138786e-01 -2.07349822e-01 -8.62322971e-02 4.43463892e-01
1.44757521e+00 -1.68456078e-01 -8.20694983e-01 5.47718287e-01
5.62600754e-02 -2.84216031e-02 3.79858285e-01 1.06738961e+00
-7.42704570e-01 -1.97159576e+00 -2.19159216e-01 -2.01600432e-01
-7.84070790e-01 -3.30555677e-01 -1.12795091e+00 5.24495304e-01
2.80759484e-01 1.10209835e+00 -4.24276501e-01 -2.73661673e-01
4.81497496e-02 1.39247537e-01 1.01662076e+00 -5.66190362e-01
-4.83336627e-01 4.04629052e-01 4.40095335e-01 -3.89853358e-01
-4.83198941e-01 -3.99811476e-01 -1.34306037e+00 -4.97845501e-01
-1.72719941e-01 2.99140006e-01 7.18019783e-01 8.58063042e-01
1.29196256e-01 5.73363662e-01 6.61394477e-01 -9.01582479e-01
-3.21148783e-01 -6.31203771e-01 -3.98438841e-01 -3.84343356e-01
2.72969544e-01 -6.34671092e-01 1.66244775e-01 6.35016933e-02] | [4.576940059661865, 6.065827369689941] |
51a86fec-3217-4bc6-b423-aa6f63b94fb6 | study-of-lexical-aspect-in-the-french-medical | null | null | https://aclanthology.org/W19-1907 | https://aclanthology.org/W19-1907.pdf | Study of lexical aspect in the French medical language. Development of a lexical resource | This paper details the development of a linguistic resource designed to improve temporal information extraction systems and to integrate aspectual values. After a brief review of recent works in temporal information extraction for the medical area, we discuss the linguistic notion of aspect and how it got a place in the NLP field. Then, we present our clinical data and describe the five-step approach adopted in this study. Finally, we represent the linguistic resource itself and explain how we elaborated it and which properties were selected for the creation of the tables. | ["C{\\'e}drick Fairon", 'Agathe Pierson'] | 2019-06-01 | null | null | null | ws-2019-6 | ['temporal-information-extraction'] | ['natural-language-processing'] | [ 6.41147420e-02 5.81541538e-01 -9.51820016e-01 -5.79414546e-01
-4.00125772e-01 -3.37168992e-01 6.67298257e-01 1.00432599e+00
-6.65530920e-01 1.18993425e+00 7.03268886e-01 -4.23550844e-01
-6.78117573e-01 -9.53094184e-01 -2.84207892e-03 -4.19428468e-01
-3.75161439e-01 7.09254801e-01 2.15315521e-01 -1.66377053e-01
-1.57943685e-02 5.18098891e-01 -1.28954399e+00 8.01609576e-01
3.66574466e-01 7.88749397e-01 -3.87661934e-01 2.95264393e-01
-5.81960678e-01 1.06222546e+00 -7.90446699e-01 -4.97724146e-01
-3.11023027e-01 -5.88009477e-01 -1.20779049e+00 -1.09811746e-01
-5.08591771e-01 1.15241349e-01 -1.76721569e-02 5.60268342e-01
4.12439048e-01 1.56665459e-01 5.70837677e-01 -9.30028260e-01
-3.45206887e-01 1.15997589e+00 -7.39161521e-02 5.13886750e-01
8.62322569e-01 -5.01171589e-01 7.09483922e-01 -3.77065748e-01
1.46906173e+00 8.77790749e-01 5.57084322e-01 5.43311656e-01
-8.04684460e-01 -6.99030533e-02 1.12729289e-01 2.20009163e-01
-1.25214672e+00 -4.26407218e-01 6.12383723e-01 -4.05110389e-01
1.52677178e+00 3.83664876e-01 1.32980692e+00 7.53042102e-01
5.06331205e-01 6.68340921e-01 1.17501652e+00 -8.10087144e-01
1.31594092e-01 3.09407175e-01 4.39670086e-01 6.04807973e-01
4.54250008e-01 7.06491396e-02 -7.35743940e-01 -3.56874555e-01
4.82895106e-01 -5.33040285e-01 1.95384368e-01 1.49362519e-01
-1.07131755e+00 5.16788244e-01 -6.77521303e-02 1.21313202e+00
-4.97818857e-01 -2.32828006e-01 6.82828367e-01 1.21741757e-01
6.14249527e-01 1.60757124e-01 -7.98255622e-01 -4.42688391e-02
-9.16871190e-01 -1.91777963e-02 1.03216302e+00 8.38662744e-01
-1.50399983e-01 -4.47113991e-01 -1.71964541e-01 6.23451591e-01
6.00779206e-02 1.09953545e-01 3.88843566e-01 -5.55767655e-01
2.20077530e-01 7.23086834e-01 -6.62287921e-02 -8.78141105e-01
-9.24625099e-01 -1.01228967e-01 -1.63891166e-01 -3.94436777e-01
2.87255138e-01 -6.91316724e-02 -8.06343317e-01 1.31186533e+00
2.10886806e-01 -6.08703911e-01 2.31551066e-01 2.72061825e-01
1.24454689e+00 7.65690431e-02 5.17230392e-01 -1.10707104e+00
2.00862837e+00 -5.18120766e-01 -1.61687779e+00 2.16021910e-01
6.95927799e-01 -7.79721916e-01 2.35656440e-01 2.32012480e-01
-1.02654231e+00 1.55723348e-01 -8.64385724e-01 -1.05333216e-01
-1.00771308e+00 6.27395324e-03 1.10382986e+00 4.17828441e-01
-8.47956777e-01 4.92910624e-01 -9.82270181e-01 -6.45083785e-01
1.08900100e-01 2.03580499e-01 -3.94884884e-01 5.49643576e-01
-1.56650841e+00 1.31309319e+00 8.37020993e-01 -6.11139573e-02
1.01944715e-01 -5.13920724e-01 -9.21008587e-01 -5.64810097e-01
6.24278724e-01 -9.27398980e-01 9.64728236e-01 -1.80598140e-01
-9.81509686e-01 1.51190937e+00 -4.72334713e-01 -6.15794599e-01
2.33331695e-01 1.15727492e-01 -1.12390614e+00 3.35348427e-01
2.54623026e-01 2.78903604e-01 -5.32597713e-02 -7.06758201e-01
-8.23803246e-01 -3.08117867e-01 4.72762249e-02 -4.94823828e-02
4.89021204e-02 3.91812533e-01 -7.70418704e-01 -9.92815375e-01
2.39874553e-02 -4.82009530e-01 -5.12947619e-01 -3.30155253e-01
-2.09913850e-01 -4.71653372e-01 -6.00169599e-02 -6.63021803e-01
2.18991899e+00 -1.95201123e+00 -5.09482957e-02 4.08213913e-01
1.94934294e-01 -1.73045948e-01 6.29093468e-01 9.37769413e-01
-4.28281665e-01 4.56629276e-01 -3.49858463e-01 7.90887997e-02
-3.50240558e-01 5.31324923e-01 2.02711523e-02 3.18348408e-01
-7.93356970e-02 1.06483388e+00 -1.10518599e+00 -1.24905860e+00
1.56800300e-01 4.36739326e-01 -4.57382686e-02 -4.22675282e-01
-1.57218963e-01 1.84641123e-01 -6.63928390e-01 8.96650016e-01
-2.30847728e-02 1.37950122e-01 4.82211024e-01 -2.23482952e-01
-5.55738032e-01 8.33711624e-01 -7.33062983e-01 1.45365596e+00
-2.97502458e-01 3.82644713e-01 -5.36646426e-01 -3.13758492e-01
7.80771375e-01 8.47155035e-01 1.09715462e+00 -7.26470530e-01
3.48499984e-01 4.77322459e-01 -1.60929441e-01 -7.76530266e-01
4.66648132e-01 -6.40395343e-01 -3.03104043e-01 4.26887870e-01
-4.75588106e-02 -8.36574584e-02 8.94158363e-01 -9.83721763e-03
6.96385860e-01 2.40567416e-01 1.58379745e+00 -2.51528263e-01
1.03327453e+00 5.80035448e-01 6.76420927e-01 1.23258576e-01
-2.51814842e-01 -2.65493598e-02 9.35808480e-01 -8.73201370e-01
-6.12519205e-01 -5.26616871e-01 -8.58856082e-01 4.44612592e-01
-3.74051571e-01 -1.03312099e+00 -2.80778915e-01 -7.75469124e-01
-4.13193792e-01 8.99274647e-01 -1.09046185e+00 2.55138010e-01
-5.54809511e-01 -1.09708118e+00 4.28111404e-01 2.78955013e-01
-1.36485055e-01 -1.34841704e+00 -8.38591337e-01 4.70446706e-01
-4.88788486e-01 -1.03166306e+00 5.28362133e-02 4.38411653e-01
-9.53906536e-01 -1.20462453e+00 -1.54683217e-01 -5.85393548e-01
5.13068855e-01 -7.08903551e-01 1.25129557e+00 -3.91647518e-02
-3.59147966e-01 3.05157542e-01 -6.17136478e-01 -8.32046747e-01
-3.52121621e-01 -5.88698406e-03 -1.24067917e-01 -7.48530507e-01
1.09494901e+00 -2.87510276e-01 -1.87297240e-01 -2.64167309e-01
-1.03124940e+00 -2.64665812e-01 2.99494803e-01 4.19165790e-01
8.23504269e-01 2.42696851e-01 1.63467556e-01 -1.31596458e+00
8.78103495e-01 -4.36150819e-01 -2.86913514e-01 3.86579454e-01
-9.37065542e-01 3.41997921e-01 -1.69689775e-01 6.90536723e-02
-9.32125866e-01 -1.30281774e-02 -4.37682152e-01 5.70293188e-01
-5.11050820e-02 9.63689744e-01 2.12204307e-01 3.24655652e-01
5.88639677e-01 -1.73452899e-01 -1.64242119e-01 -4.86009151e-01
3.82607877e-01 2.22571716e-01 2.45578796e-01 -3.93242836e-01
-4.77580819e-03 5.71852922e-01 2.78455973e-01 -5.30525982e-01
-8.02774191e-01 -6.99784517e-01 -1.10055923e+00 -2.26444840e-01
1.02189422e+00 -4.70683724e-01 -4.44382668e-01 5.95778972e-03
-1.10308707e+00 1.58459008e-01 -8.49720836e-01 6.60061955e-01
-6.97943568e-01 7.04890043e-02 -4.76704359e-01 -8.30501854e-01
-3.99199426e-01 -9.14830983e-01 6.89938188e-01 -2.45464295e-01
-7.42703557e-01 -1.48572755e+00 4.08655763e-01 -4.65773754e-02
-5.71839996e-02 7.67864525e-01 1.05862105e+00 -6.62039816e-01
5.73315285e-02 -9.34742987e-02 3.39101166e-01 -6.88636005e-01
6.67933881e-01 2.35157106e-02 -6.11797810e-01 5.18286347e-01
3.64850640e-01 4.25138921e-01 6.29925370e-01 4.39185649e-01
5.24914861e-01 -4.53798145e-01 -1.00445712e+00 3.17873746e-01
1.39860535e+00 7.67954826e-01 3.89994681e-01 7.20789909e-01
-9.27936733e-02 9.83029008e-01 1.19920397e+00 3.64914984e-01
3.33141446e-01 7.85833597e-01 -2.37033233e-01 3.96325402e-02
-1.38818517e-01 3.62769291e-02 -1.69834048e-01 9.70101058e-01
-3.71569604e-01 -2.27815993e-02 -1.07992530e+00 8.39364290e-01
-1.66176176e+00 -5.44265151e-01 -2.09702671e-01 1.72180033e+00
1.38489938e+00 2.72648066e-01 2.80525923e-01 3.09496343e-01
1.63663730e-01 7.58777708e-02 1.30696505e-01 -7.44111419e-01
-2.97621816e-01 2.15619475e-01 5.06257594e-01 7.58067250e-01
-1.12183344e+00 9.36847746e-01 7.90528059e+00 4.34549540e-01
-6.23108208e-01 1.37714580e-01 2.89417297e-01 1.43627018e-01
-3.78462225e-01 -7.30923712e-02 -7.67561555e-01 1.29456408e-02
1.35034108e+00 -6.04370177e-01 -2.99835742e-01 4.74578261e-01
4.94422883e-01 -4.12090153e-01 -9.04524207e-01 4.83978868e-01
-1.76560115e-02 -1.33280897e+00 2.67070830e-01 6.61796611e-03
2.25177363e-01 -4.06902820e-01 -4.19865936e-01 -3.83628577e-01
-3.62383500e-02 -7.28889108e-01 5.44051349e-01 8.27310860e-01
7.66107380e-01 -5.70613146e-01 1.10961211e+00 -1.37062535e-01
-1.44603038e+00 2.42577747e-01 8.43857974e-02 1.12609938e-01
7.56108165e-01 8.60797286e-01 -9.75290418e-01 1.17605734e+00
7.18388498e-01 6.95502222e-01 -5.01022041e-01 9.55830455e-01
-5.78643918e-01 -1.93421505e-02 -2.69802988e-01 -5.90055510e-02
1.04916006e-01 -1.91740781e-01 7.73831546e-01 1.44690931e+00
-8.56366679e-02 4.16122049e-01 -1.75009593e-01 2.56542832e-01
4.78627235e-01 7.82242060e-01 -8.35500419e-01 -3.01373363e-01
2.54593790e-01 7.49218762e-01 -1.20687568e+00 -5.39769113e-01
-4.45708513e-01 4.12323684e-01 -1.01639159e-01 -5.65591343e-02
-6.84968948e-01 -4.02395695e-01 3.23450685e-01 1.40675604e-01
-6.01904988e-02 -6.74446300e-02 -5.77931345e-01 -9.23121333e-01
9.39449221e-02 -7.15907633e-01 1.20912874e+00 -4.08104599e-01
-7.50299156e-01 1.19485199e+00 7.84606814e-01 -1.18426692e+00
-7.14213073e-01 -5.14822960e-01 2.30073214e-01 5.74740291e-01
-9.52039361e-01 -1.01235008e+00 4.74957347e-01 3.40443552e-01
1.60205573e-01 2.08216365e-02 1.09449577e+00 5.17511010e-01
-1.14164881e-01 1.76591352e-01 -6.54064238e-01 4.61305566e-02
7.01322794e-01 -1.11895621e+00 1.22373834e-01 7.60273814e-01
1.63379626e-03 1.05530798e+00 8.89084160e-01 -1.17743671e+00
-4.78246331e-01 -6.43371224e-01 2.24293089e+00 -5.68673432e-01
9.64626610e-01 9.41463262e-02 -5.27166069e-01 7.92347908e-01
3.83142322e-01 -4.61710453e-01 1.04776406e+00 3.04117166e-02
1.72534510e-02 6.53099865e-02 -1.33254743e+00 6.16129696e-01
1.08585930e+00 -4.42923158e-01 -1.27559614e+00 3.93310905e-01
9.32601213e-01 -4.15440261e-01 -1.45513773e+00 6.07281685e-01
5.04639983e-01 -5.48376441e-01 9.66174543e-01 -6.95356905e-01
-6.39312267e-02 -3.76659900e-01 2.72810221e-01 -7.64585376e-01
-1.53267786e-01 -5.53088427e-01 -1.00488402e-01 1.13673031e+00
9.74731207e-01 -6.14748716e-01 3.08110714e-01 5.44681966e-01
1.35817071e-02 -9.09570992e-01 -9.99485433e-01 -3.40037078e-01
-8.73267502e-02 -6.86741829e-01 5.91441691e-01 9.50780213e-01
8.73053432e-01 1.38053060e-01 -4.92577218e-02 -4.41229254e-01
3.34624738e-01 6.18106462e-02 -3.68582368e-01 -9.00038362e-01
1.02693193e-01 -4.58154142e-01 -3.92021537e-01 -4.94328365e-02
-1.41279772e-01 -9.10107195e-01 -3.20794523e-01 -1.97351611e+00
3.40084061e-02 -1.54041216e-01 -2.09568784e-01 8.79365325e-01
4.08717170e-02 6.40377998e-02 -1.97661698e-01 4.18487974e-02
-2.18122721e-01 -1.99306846e-01 9.03095841e-01 -7.06675425e-02
-6.43932819e-01 5.35790659e-02 -7.63120234e-01 6.61820889e-01
8.11419547e-01 -8.92313719e-01 -3.32183391e-01 -7.83415046e-03
7.88175881e-01 7.63582662e-02 -5.39821029e-01 -5.80293298e-01
3.63461375e-01 -3.94621909e-01 1.20527856e-01 -1.18492723e+00
1.04089625e-01 -1.21419370e+00 5.04405081e-01 9.19262528e-01
-2.04024747e-01 2.12904543e-01 5.12538135e-01 4.33422364e-02
-4.92070556e-01 -5.02989829e-01 4.42923844e-01 -2.63317555e-01
-7.12019920e-01 -6.17475137e-02 -6.66714787e-01 -1.89998299e-01
1.15956843e+00 1.37921110e-01 -1.00540832e-01 1.06795095e-01
-1.43029833e+00 3.31464171e-01 1.08652167e-01 3.58072847e-01
5.33533812e-01 -1.21143484e+00 -3.20376247e-01 -3.27419162e-01
3.20803761e-01 -4.57457155e-01 -4.91159968e-02 1.28594065e+00
-7.54006803e-01 1.12064326e+00 -2.12186381e-01 -7.25115538e-02
-1.53500354e+00 1.18994761e+00 2.51060635e-01 -7.55483866e-01
-8.01459312e-01 2.77407557e-01 -3.61027032e-01 2.44677365e-02
1.59661576e-01 -6.61802828e-01 -1.10845101e+00 8.07898283e-01
5.52919209e-01 1.14578672e-01 1.86735675e-01 -6.67405844e-01
-1.06333935e+00 5.72070003e-01 3.24824423e-01 -4.90866303e-01
1.51692188e+00 -3.36322814e-01 -9.63467836e-01 1.11860967e+00
7.61837780e-01 3.96590322e-01 6.28586635e-02 -1.48701876e-01
7.66740978e-01 2.33503982e-01 -2.80764878e-01 -1.05266893e+00
-8.47976983e-01 1.25687182e-01 1.29549190e-01 3.79002899e-01
1.32923186e+00 4.58957523e-01 3.54511440e-01 1.94463909e-01
5.36326945e-01 -1.05203772e+00 -7.68601060e-01 4.56007749e-01
9.30495203e-01 -5.99095941e-01 6.60817623e-01 -9.97344434e-01
-6.33163750e-01 1.27562726e+00 -1.79545090e-01 5.35249054e-01
1.15070951e+00 6.76746786e-01 2.39504784e-01 -7.21854270e-01
-9.64644492e-01 -5.15680075e-01 6.02854609e-01 5.45941949e-01
1.01738775e+00 5.25185093e-02 -1.63533235e+00 7.08124161e-01
-5.29083669e-01 5.22660673e-01 3.42303038e-01 1.32696128e+00
4.38295715e-02 -1.71165335e+00 -2.61020899e-01 3.38727385e-01
-1.09043944e+00 -2.39339724e-01 -8.20873797e-01 1.18579280e+00
4.13869441e-01 8.31547678e-01 -2.13809133e-01 -6.22880943e-02
5.91571391e-01 3.58229637e-01 5.32941699e-01 -5.50538123e-01
-8.22353244e-01 4.16579545e-01 9.33767617e-01 -7.32501805e-01
-1.24658442e+00 -9.95104551e-01 -1.51945388e+00 7.19604343e-02
1.81970358e-01 6.24277472e-01 5.08642733e-01 9.55265164e-01
-2.84974664e-01 8.67818475e-01 -1.75730884e-01 3.07968169e-01
5.94301581e-01 -7.60386467e-01 -4.05964285e-01 -1.71352744e-01
3.41552675e-01 -6.31673634e-01 -1.73633069e-01 3.88085216e-01] | [8.580159187316895, 8.991235733032227] |
9a299693-c7ff-42ac-a249-cfd93c32d7fc | fastano-fast-anomaly-detection-via-spatio | 2106.08613 | null | https://arxiv.org/abs/2106.08613v4 | https://arxiv.org/pdf/2106.08613v4.pdf | FastAno: Fast Anomaly Detection via Spatio-temporal Patch Transformation | Video anomaly detection has gained significant attention due to the increasing requirements of automatic monitoring for surveillance videos. Especially, the prediction based approach is one of the most studied methods to detect anomalies by predicting frames that include abnormal events in the test set after learning with the normal frames of the training set. However, a lot of prediction networks are computationally expensive owing to the use of pre-trained optical flow networks, or fail to detect abnormal situations because of their strong generative ability to predict even the anomalies. To address these shortcomings, we propose spatial rotation transformation (SRT) and temporal mixing transformation (TMT) to generate irregular patch cuboids within normal frame cuboids in order to enhance the learning of normal features. Additionally, the proposed patch transformation is used only during the training phase, allowing our model to detect abnormal frames at fast speed during inference. Our model is evaluated on three anomaly detection benchmarks, achieving competitive accuracy and surpassing all the previous works in terms of speed. | ['Sangyoun Lee', 'Minhyeok Lee', 'MyeongAh Cho', 'Chaewon Park'] | 2021-06-16 | null | null | null | null | ['anomaly-detection-in-surveillance-videos', 'anomaly-detection-in-surveillance-videos'] | ['computer-vision', 'methodology'] | [ 3.28721374e-01 -2.78247416e-01 -8.92768241e-03 -6.31394088e-02
-3.02034076e-02 -2.04972923e-01 6.33262813e-01 2.48133376e-01
-2.46376902e-01 4.66887325e-01 -4.46460181e-04 -2.25280434e-01
1.27586082e-01 -7.15829611e-01 -7.06080794e-01 -7.00392485e-01
-4.42830652e-01 5.67329228e-02 5.55304110e-01 7.57689327e-02
1.49195135e-01 5.80464363e-01 -1.76798081e+00 3.15972269e-01
1.13026965e+00 1.16995108e+00 -2.78159708e-01 6.37126923e-01
-1.55459729e-03 1.21381700e+00 -4.74802703e-01 -2.45484903e-01
2.89223105e-01 -5.94938993e-01 -3.47320199e-01 2.75678009e-01
5.70673406e-01 -7.28469610e-01 -4.76940572e-01 9.76513743e-01
1.68895349e-01 3.40357840e-01 3.99252355e-01 -1.40924656e+00
-5.52786477e-02 4.97478060e-02 -6.62748694e-01 7.40875900e-01
5.51269948e-01 1.65100291e-01 8.03513527e-01 -8.13883960e-01
4.64619368e-01 1.05153060e+00 4.47147191e-01 6.32394195e-01
-9.82178986e-01 -5.34278750e-01 3.67855996e-01 8.26184094e-01
-9.24163163e-01 -4.13634241e-01 8.83988917e-01 -4.28495586e-01
8.45026076e-01 2.60301977e-01 9.94914114e-01 1.15955889e+00
1.42543420e-01 8.92808795e-01 4.77596551e-01 -6.25688136e-02
1.36359677e-01 -4.21653241e-01 -1.30836487e-01 8.62197340e-01
2.29306936e-01 3.25712748e-02 -3.51354092e-01 -3.33861142e-01
6.53314471e-01 4.67972159e-01 -5.45331299e-01 -2.66044796e-01
-1.14320207e+00 6.80469453e-01 2.46559128e-01 1.35916963e-01
-6.28264546e-01 -1.13516197e-01 6.30193651e-01 2.62738883e-01
6.16637528e-01 3.31111878e-01 -2.15547904e-01 -4.11381662e-01
-8.34966660e-01 1.77002266e-01 4.48464721e-01 4.09665316e-01
5.01980603e-01 3.43815267e-01 -2.78768092e-01 5.61070502e-01
1.20625675e-01 2.62836874e-01 5.14418304e-01 -9.02529657e-01
4.38864201e-01 8.56358051e-01 5.87968864e-02 -1.47422397e+00
-3.43575418e-01 -3.01488817e-01 -1.18035662e+00 -9.80986282e-03
7.02825546e-01 -1.70698896e-01 -8.19068015e-01 1.50246084e+00
3.25099587e-01 1.17319489e+00 -2.17213463e-02 7.89807737e-01
4.73778903e-01 9.32621658e-01 2.23554876e-02 -5.14029860e-01
9.93167818e-01 -8.88872504e-01 -5.67172289e-01 -8.48838836e-02
7.84116924e-01 -6.78043306e-01 6.53891563e-01 5.26026428e-01
-8.51254284e-01 -5.89379668e-01 -7.02321708e-01 4.82434124e-01
4.63536233e-02 5.41697778e-02 4.07602698e-01 3.61127794e-01
-6.72363818e-01 5.19348204e-01 -1.29945588e+00 -3.36801827e-01
6.91560388e-01 4.07939218e-02 -3.83866608e-01 -3.07967663e-01
-8.89037311e-01 6.39660358e-01 3.01573545e-01 1.98928356e-01
-8.41255724e-01 -6.85604632e-01 -9.79414344e-01 1.14054695e-01
3.86107534e-01 -3.87379348e-01 6.76826119e-01 -1.07189655e+00
-1.22129393e+00 3.82287681e-01 -3.64001453e-01 -7.44718432e-01
7.26665020e-01 -4.17225927e-01 -6.38471603e-01 4.94353622e-01
-1.50102243e-01 4.54678774e-01 1.10317492e+00 -6.77817881e-01
-9.10858333e-01 -1.56846702e-01 3.88654061e-02 9.48167313e-03
-4.62389708e-01 -2.05703571e-01 -2.55910575e-01 -9.17944908e-01
1.48573592e-01 -8.91578734e-01 -2.19547197e-01 -1.88086256e-02
-3.51189315e-01 -3.21082592e-01 1.33095706e+00 -7.33093262e-01
1.32318974e+00 -2.30565596e+00 -1.07213780e-02 3.12686563e-01
2.61121988e-01 5.74190497e-01 -1.99441612e-02 4.88005765e-02
-1.22773208e-01 -2.52500981e-01 -2.38376483e-01 -1.29604787e-01
-5.41942894e-01 3.09436619e-01 -4.79653656e-01 4.67005581e-01
5.79085767e-01 4.27012920e-01 -1.04914820e+00 -4.73418951e-01
5.79495668e-01 3.59532952e-01 -8.59818518e-01 4.65596139e-01
-2.67353565e-01 1.01968586e+00 -4.38741744e-01 4.85736579e-01
5.72436512e-01 -1.24333076e-01 -1.94885999e-01 -4.32187580e-02
1.55019104e-01 2.58970894e-02 -1.11107838e+00 1.24501336e+00
-4.64103222e-02 5.70084691e-01 -4.71154004e-01 -1.33170569e+00
7.31792271e-01 6.07648253e-01 9.40697134e-01 -3.63080263e-01
-4.56740931e-02 1.15219615e-01 2.69316018e-01 -7.75096416e-01
1.65384244e-02 3.65039766e-01 6.18832290e-01 2.90719837e-01
-1.34916350e-01 4.99149352e-01 6.02884412e-01 1.54731013e-02
1.35356164e+00 8.45180675e-02 1.86811090e-01 2.64723450e-01
1.02428043e+00 -1.25993624e-01 1.06532681e+00 6.20863318e-01
-3.09615612e-01 4.69572008e-01 7.84020662e-01 -1.17572391e+00
-8.78424227e-01 -8.39124560e-01 4.67706285e-02 7.38898158e-01
1.51927218e-01 -4.55552161e-01 -6.16957664e-01 -1.03190351e+00
-3.28877032e-01 5.43459713e-01 -5.40330231e-01 -3.22372675e-01
-9.03417349e-01 -9.43723679e-01 3.72574180e-01 4.37198073e-01
7.45502353e-01 -1.22041047e+00 -9.90094244e-01 2.47737750e-01
-4.20859635e-01 -1.35291791e+00 -2.45819390e-01 -4.87337857e-01
-9.66891170e-01 -1.55958498e+00 -5.68534315e-01 -5.25790930e-01
1.02737415e+00 1.15302503e-01 8.25676858e-01 5.17287910e-01
-3.47098023e-01 1.59184739e-01 -5.62094748e-01 -2.63988137e-01
-5.10338366e-01 -2.22203493e-01 7.52725676e-02 5.58237851e-01
5.14020324e-01 -5.77594161e-01 -7.55611777e-01 4.62253928e-01
-9.98311222e-01 3.80563992e-03 3.65121305e-01 7.97912419e-01
4.16898251e-01 1.45372778e-01 3.77611756e-01 -7.00264096e-01
3.67205329e-02 -4.98954505e-01 -6.68888807e-01 1.07031055e-02
-2.20661297e-01 -2.03856483e-01 9.08801496e-01 -4.75257218e-01
-8.33212197e-01 -9.62627679e-02 -9.29048508e-02 -7.14935780e-01
-4.09759760e-01 1.87996730e-01 2.19116285e-01 7.57878423e-02
4.21524674e-01 3.87718946e-01 1.61052361e-01 -9.48520303e-02
-3.05027455e-01 -4.18719426e-02 4.37390208e-01 -2.62138456e-01
8.65094483e-01 6.00229323e-01 2.54547924e-01 -8.80288780e-01
-8.59469712e-01 -2.85053939e-01 -5.64663351e-01 -3.73386055e-01
8.70909393e-01 -5.80722213e-01 -4.89740461e-01 6.38137281e-01
-1.17143273e+00 1.73278928e-01 -1.82862371e-01 6.85181201e-01
-2.69784153e-01 7.25751996e-01 -5.59864163e-01 -7.24180043e-01
-2.58971512e-01 -1.31286061e+00 6.41336977e-01 2.01140657e-01
-2.70073721e-03 -8.06610763e-01 1.65554695e-02 4.24879789e-02
4.04888302e-01 6.72651410e-01 9.28418934e-01 -7.22298563e-01
-7.03712165e-01 -4.28424567e-01 -1.61646128e-01 4.61278349e-01
3.53805423e-01 2.86979675e-01 -7.55014360e-01 -4.24658269e-01
-1.34065107e-01 2.30894610e-01 8.55381429e-01 6.12881839e-01
1.44942987e+00 -3.85789603e-01 -3.91549140e-01 6.51695788e-01
9.52678680e-01 3.38509589e-01 6.33904457e-01 2.96690553e-01
7.30450690e-01 4.29439276e-01 5.73055923e-01 6.96727574e-01
8.32271203e-02 5.46227515e-01 6.53912485e-01 1.67258307e-02
1.79490864e-01 1.14670983e-02 4.04736161e-01 4.38964516e-01
-5.12044251e-01 -1.78086787e-01 -7.84574509e-01 4.17953014e-01
-1.97503459e+00 -1.30028152e+00 -1.43218741e-01 2.28218389e+00
2.45573267e-01 3.10407400e-01 2.25958094e-01 3.88298392e-01
7.20448196e-01 3.57363135e-01 -4.37037379e-01 -8.79429653e-02
3.98109816e-02 -7.92520344e-02 -3.93035896e-02 1.28974736e-01
-1.39751232e+00 6.20790005e-01 5.36365128e+00 4.88233060e-01
-1.14045429e+00 -2.88338721e-01 7.73297131e-01 7.36151263e-02
2.68833578e-01 -2.40256056e-01 -4.26107317e-01 7.18156576e-01
6.00340605e-01 2.08128244e-01 2.79114604e-01 7.15470135e-01
4.39569324e-01 3.72523535e-03 -1.04220438e+00 9.29240227e-01
1.08359940e-01 -1.09630787e+00 2.98371732e-01 -2.46168405e-01
6.50962412e-01 -7.60052800e-02 -1.42616794e-01 9.90312845e-02
-2.72187293e-01 -7.38589585e-01 1.43634275e-01 4.53358591e-01
2.53681600e-01 -7.67839611e-01 1.05047691e+00 2.77529776e-01
-1.13965464e+00 -3.30937713e-01 -4.00046706e-01 -9.73501578e-02
1.32482857e-01 6.42135799e-01 -7.57728696e-01 3.64196390e-01
8.36092114e-01 1.18689537e+00 -6.40359998e-01 1.29233742e+00
-1.29295275e-01 8.76562178e-01 -4.57077026e-01 4.32215750e-01
2.81230479e-01 -2.62651294e-01 9.21009302e-01 8.62847805e-01
5.62482953e-01 -1.90131322e-01 4.80599493e-01 3.78594726e-01
2.38898247e-01 1.47916839e-01 -6.74001753e-01 2.69592792e-01
1.15428686e-01 1.08120120e+00 -6.01322532e-01 -4.46202040e-01
-7.27568150e-01 8.24883580e-01 -1.35957539e-01 2.35257864e-01
-9.58317518e-01 -1.30344182e-01 7.05916524e-01 2.08990484e-01
2.28073224e-01 -2.73541734e-03 2.10910439e-01 -1.47665799e+00
2.55132735e-01 -1.00899005e+00 5.85255980e-01 -2.44418919e-01
-1.07635283e+00 7.19671011e-01 -3.55761759e-02 -1.77228928e+00
-6.05923712e-01 -4.35588419e-01 -9.12325680e-01 2.28871152e-01
-1.59589660e+00 -8.22458923e-01 -7.41100311e-01 8.12819600e-01
7.09414363e-01 -3.08124691e-01 7.11752534e-01 5.22583544e-01
-8.00538540e-01 3.38824242e-01 -2.48944998e-01 4.74540740e-01
5.22513509e-01 -1.07303214e+00 3.33737880e-01 1.39419591e+00
1.42069459e-01 6.81150854e-02 7.07117021e-01 -5.87032378e-01
-8.98686111e-01 -1.30588651e+00 5.15313864e-01 -2.99089909e-01
4.33975488e-01 1.67778328e-01 -1.31827545e+00 7.04666913e-01
-2.04427429e-02 5.58058262e-01 5.24193466e-01 -4.27416354e-01
-2.22281858e-01 -1.68659016e-01 -9.56503212e-01 5.78866899e-01
9.39010203e-01 5.35344742e-02 -4.30705577e-01 2.10222512e-01
1.09777257e-01 -4.87472504e-01 -5.44342101e-01 7.78649688e-01
4.03688014e-01 -1.18243074e+00 8.48076284e-01 -5.23721576e-01
4.99722749e-01 -5.23711979e-01 1.49743587e-01 -1.21977103e+00
-7.40536451e-02 -5.20665824e-01 -5.98730326e-01 9.49368596e-01
7.95989558e-02 -8.58384371e-01 8.69478822e-01 2.72318751e-01
1.01828454e-02 -7.90318191e-01 -9.38659132e-01 -4.95079458e-01
-7.39647627e-01 -4.34701592e-01 5.04505336e-01 9.03894365e-01
-2.77195305e-01 -1.71852589e-01 -5.84474504e-01 4.49133843e-01
5.85704446e-01 1.03470139e-01 7.93583930e-01 -1.41065800e+00
-1.89559236e-01 -3.85377377e-01 -9.47270453e-01 -9.03328836e-01
1.07431278e-01 -5.82317233e-01 -2.00227395e-01 -1.04811537e+00
-2.63491217e-02 -1.40593275e-01 -4.07856673e-01 4.14863855e-01
-4.28605616e-01 4.06072229e-01 4.61325832e-02 2.76907235e-01
-6.35831177e-01 5.54847598e-01 1.19320118e+00 -5.86487986e-02
-3.04480463e-01 3.96930784e-01 4.58380766e-02 1.15792406e+00
6.86261237e-01 -2.76879132e-01 -3.84385109e-01 -3.35517555e-01
-9.96513665e-02 6.08702637e-02 6.12867057e-01 -1.48250699e+00
-5.32655567e-02 -1.63433850e-01 6.13654971e-01 -6.58875704e-01
1.05970390e-01 -9.76211965e-01 -1.52788326e-01 7.22803533e-01
-2.05024138e-01 4.84569132e-01 8.08221474e-02 7.89142072e-01
-4.32473600e-01 2.07974073e-02 8.22182119e-01 9.98233333e-02
-8.73653769e-01 7.07872570e-01 -5.62831223e-01 1.17722534e-01
1.19390035e+00 -2.62660086e-01 -1.49396256e-01 -4.95333403e-01
-5.92500985e-01 1.93042114e-01 3.05871069e-01 6.15431190e-01
7.34837592e-01 -1.14247143e+00 -7.25311995e-01 5.40344000e-01
5.03438264e-02 2.27486551e-01 3.71506751e-01 1.19727623e+00
-9.09119368e-01 1.93679437e-01 -3.04867506e-01 -9.22336698e-01
-1.09171152e+00 4.89876181e-01 3.74966115e-01 -4.30612773e-01
-1.09601665e+00 3.97268116e-01 3.19588065e-01 1.94004193e-01
2.41994828e-01 -2.61175245e-01 -6.15665913e-01 -2.69171804e-01
7.97702968e-01 5.17164946e-01 -6.21782057e-02 -6.48139775e-01
-3.07315350e-01 5.52787364e-01 -1.17836319e-01 6.73452377e-01
1.18791234e+00 1.76524669e-01 -1.18831716e-01 3.29689421e-02
8.77464950e-01 -1.80199683e-01 -1.45009208e+00 -2.67484725e-01
1.67082950e-01 -7.62280583e-01 -1.85775876e-01 -9.10033956e-02
-1.39931560e+00 8.50809753e-01 6.66147768e-01 3.77172023e-01
1.37309837e+00 -4.30291682e-01 1.02708602e+00 4.81861472e-01
2.68727280e-02 -7.14351058e-01 1.95628434e-01 3.27752650e-01
5.57888150e-01 -1.23310268e+00 -1.15278929e-01 -4.62491393e-01
-5.42321324e-01 1.33860254e+00 9.15777266e-01 -4.83487695e-01
6.42731786e-01 -1.29315063e-01 -3.25626023e-02 4.26310413e-02
-7.13952780e-01 1.18687421e-01 5.04059196e-01 5.25663555e-01
2.99218744e-01 -3.42764288e-01 -1.08565994e-01 -1.36555701e-01
1.75025076e-01 -1.86681196e-01 3.60987097e-01 7.73561120e-01
-3.21632296e-01 -8.65665078e-01 -3.35256189e-01 8.71827364e-01
-7.68161893e-01 1.66804686e-01 3.07532281e-01 5.23404777e-01
1.80179089e-01 8.88559997e-01 4.08271134e-01 -1.34913951e-01
2.83141434e-01 1.17433265e-01 1.49314955e-01 -2.71949261e-01
-1.32189244e-01 2.85741519e-02 -3.40635091e-01 -8.82885754e-01
-5.88023901e-01 -6.80754364e-01 -1.05125475e+00 -1.31694123e-01
-1.81563571e-01 2.51484960e-01 -3.70576866e-02 1.20855713e+00
3.71752709e-01 4.62304950e-01 7.90900826e-01 -7.16342390e-01
-4.07150760e-02 -9.01961446e-01 -1.95243597e-01 8.52778375e-01
7.26836681e-01 -8.08785439e-01 -4.06459153e-01 1.50251880e-01] | [7.873534679412842, 1.5441735982894897] |
164e10c0-0cbb-463d-84a3-6388a30368e3 | stapi-an-automatic-scraper-for-extracting | null | null | https://aclanthology.org/2022.lrec-1.371 | https://aclanthology.org/2022.lrec-1.371.pdf | STAPI: An Automatic Scraper for Extracting Iterative Title-Text Structure from Web Documents | Formal documents often are organized into sections of text, each with a title, and extracting this structure remains an under-explored aspect of natural language processing. This iterative title-text structure is valuable data for building models for headline generation and section title generation, but there is no corpus that contains web documents annotated with titles and prose texts. Therefore, we propose the first title-text dataset on web documents that incorporates a wide variety of domains to facilitate downstream training. We also introduce STAPI (Section Title And Prose text Identifier), a two-step system for labeling section titles and prose text in HTML documents. To filter out unrelated content like document footers, its first step involves a filter that reads HTML documents and proposes a set of textual candidates. In the second step, a typographic classifier takes the candidates from the filter and categorizes each one into one of the three pre-defined classes (title, prose text, and miscellany). We show that STAPI significantly outperforms two baseline models in terms of title-text identification. We release our dataset along with a web application to facilitate supervised and semi-supervised training in this domain. | ['Prasenjit Mitra', 'Shomir Wilson', 'Nan Zhang'] | null | null | null | null | lrec-2022-6 | ['headline-generation'] | ['natural-language-processing'] | [ 6.93955004e-01 1.90496817e-01 -6.94237292e-01 -2.07116857e-01
-1.03367651e+00 -1.09693646e+00 9.37926352e-01 5.24874032e-01
-1.75011531e-01 7.12055326e-01 5.58760583e-01 -6.39963269e-01
-1.95715055e-01 -6.00575864e-01 -6.10111177e-01 -1.46617919e-01
3.97119433e-01 7.76139021e-01 4.74438012e-01 -5.03489515e-03
8.02136064e-01 -5.99923059e-02 -1.60541284e+00 8.67783368e-01
1.23944247e+00 6.51508749e-01 4.30412233e-01 5.95675588e-01
-1.01566708e+00 5.26898146e-01 -7.58258522e-01 -3.55451047e-01
-7.54771531e-02 -6.19279981e-01 -1.13680279e+00 2.05733150e-01
6.29994035e-01 -1.19838551e-01 1.06857933e-01 9.25149977e-01
8.68223235e-02 -2.85034686e-01 9.65516508e-01 -1.13876855e+00
-3.04534078e-01 1.43480468e+00 -4.40545917e-01 -5.44170775e-02
7.76902795e-01 -3.59477013e-01 1.50576341e+00 -9.88528728e-01
1.08074450e+00 1.16871381e+00 6.56341314e-01 4.83166993e-01
-1.25564420e+00 -3.52132082e-01 -4.98251170e-02 -5.06010175e-01
-1.04439139e+00 -3.06882381e-01 5.39968431e-01 -7.65368164e-01
9.01616991e-01 3.41844141e-01 2.44241759e-01 1.44368243e+00
2.65864804e-02 1.11612332e+00 9.25809562e-01 -8.64765644e-01
1.18469208e-01 5.82239032e-01 6.36600971e-01 3.11329365e-01
3.57854933e-01 -4.41074461e-01 -5.35060465e-01 -3.15338612e-01
8.57398063e-02 -3.95684689e-01 -9.77586061e-02 -7.26600736e-02
-1.28609753e+00 7.72074163e-01 -3.92876476e-01 3.46653968e-01
-9.67519581e-02 -4.43291456e-01 5.50094426e-01 -2.25812830e-02
1.96309641e-01 7.43684530e-01 -4.18410599e-01 -1.62779018e-01
-1.31235504e+00 5.21596372e-01 1.23954284e+00 1.39863944e+00
5.02270460e-01 -5.93129992e-01 -2.20592111e-01 1.27316737e+00
1.76848561e-01 3.52863818e-01 6.77136540e-01 -5.80542803e-01
1.06631672e+00 9.02342200e-01 2.77103812e-01 -5.12252152e-01
-3.02366734e-01 -1.36151895e-01 -3.28932613e-01 -3.15134913e-01
4.97136027e-01 -2.72249002e-02 -6.61069572e-01 1.05406284e+00
-4.59784176e-04 -1.08511662e+00 -4.98545617e-02 2.32259780e-01
7.97594547e-01 8.82719576e-01 -7.64683411e-02 -2.61691451e-01
1.58107567e+00 -8.27582479e-01 -7.46755838e-01 -3.18894982e-01
7.78684556e-01 -1.24943674e+00 1.22863805e+00 3.92138273e-01
-9.83065546e-01 -4.04898375e-01 -8.16900074e-01 -1.60497352e-01
-7.78413475e-01 5.68985224e-01 1.81570858e-01 5.42265117e-01
-5.07539630e-01 6.14185274e-01 -3.39844257e-01 -7.08916724e-01
2.81985134e-01 -1.82869926e-01 -1.06652733e-02 1.55629262e-01
-1.18196952e+00 3.52053285e-01 6.48493230e-01 -6.50802195e-01
-2.18157157e-01 -6.58312380e-01 -6.42708600e-01 1.98311564e-02
6.31856382e-01 -3.84913683e-01 1.39972568e+00 -8.31647992e-01
-1.09602439e+00 9.90749657e-01 -2.61567861e-01 -2.00960100e-01
4.59590375e-01 -2.68883049e-01 -4.67411131e-01 1.00557841e-01
4.72319871e-01 4.71072763e-01 9.61388707e-01 -1.27360368e+00
-1.08080745e+00 -2.86481351e-01 -1.78285986e-01 1.88654196e-02
-5.54386675e-01 4.16160703e-01 -5.62018573e-01 -9.70778227e-01
9.68027115e-02 -8.52575839e-01 4.31590050e-01 -4.85463500e-01
-1.05447781e+00 -7.28744507e-01 5.66481590e-01 -8.46228659e-01
1.97189927e+00 -1.79529130e+00 -3.28143865e-01 4.07601029e-01
1.14831343e-01 8.30244645e-03 -1.43243801e-02 9.70273495e-01
1.05543859e-01 8.50071609e-01 -5.68450913e-02 -6.11635633e-02
3.23938012e-01 -1.61735341e-01 -6.21421874e-01 -3.14436346e-01
-1.91904251e-02 6.06259108e-01 -8.46014738e-01 -7.76223421e-01
-3.95684421e-01 -2.02764779e-01 -4.81712341e-01 1.79001957e-01
-6.24578953e-01 -2.17821553e-01 -4.56401527e-01 6.96646690e-01
3.82804066e-01 -3.08626324e-01 2.33780712e-01 -2.93886345e-02
-5.57628810e-01 1.09929895e+00 -1.30663502e+00 1.15554416e+00
-1.74730733e-01 6.31843448e-01 -1.67579129e-01 -4.80803072e-01
7.32522249e-01 1.19335853e-01 3.82722199e-01 -3.97567153e-01
2.39441656e-02 5.63392341e-01 -1.39321268e-01 -5.74242175e-01
8.93942416e-01 5.45205593e-01 -4.79295015e-01 8.43192279e-01
5.30939251e-02 -6.67475685e-02 1.16359818e+00 6.96212649e-01
1.07944107e+00 2.47264832e-01 2.94410735e-01 -2.50913084e-01
6.10299528e-01 4.01542276e-01 2.26396531e-01 1.09159601e+00
3.10080647e-01 5.45031905e-01 8.91838670e-01 3.74066792e-02
-1.37755370e+00 -8.59381080e-01 -2.74829209e-01 1.37308586e+00
-2.13010162e-01 -9.82282877e-01 -9.69801188e-01 -1.00828290e+00
1.85699761e-01 8.94692361e-01 -4.01524186e-01 3.65460157e-01
-6.86025560e-01 -3.64965409e-01 5.44224083e-01 4.94451851e-01
-3.60785909e-02 -1.14840436e+00 -1.22050174e-01 2.17518732e-01
-4.78432059e-01 -8.36128354e-01 -8.37854147e-01 4.25706327e-01
-4.46136624e-01 -1.06554019e+00 -6.57623470e-01 -1.03323960e+00
6.65262640e-01 1.34703875e-01 1.04713857e+00 -2.64074057e-02
-1.30963773e-01 8.94935504e-02 -8.26694667e-01 -4.28007662e-01
-1.14496577e+00 6.62035525e-01 -2.12767154e-01 -4.07520086e-01
4.85619694e-01 4.77802083e-02 2.98041012e-03 2.59582847e-01
-9.86671984e-01 1.16365552e-01 4.62896347e-01 9.22493041e-01
2.36617059e-01 4.07532081e-02 3.13362807e-01 -1.43079019e+00
8.75795186e-01 -3.46160501e-01 -6.50945127e-01 4.72741842e-01
-7.78349519e-01 2.07146987e-01 7.77243137e-01 -3.79836977e-01
-1.20784259e+00 4.59711626e-02 -1.22228265e-01 4.48185623e-01
-3.12160730e-01 9.14921880e-01 -2.12936521e-01 6.50020540e-01
7.34113634e-01 3.07648927e-01 -1.80886090e-01 -9.27073717e-01
3.01186711e-01 1.22287095e+00 4.70574111e-01 -5.27939737e-01
1.11089575e+00 -1.66212231e-01 -6.34352446e-01 -1.03039742e+00
-9.22073424e-01 -7.83385694e-01 -8.86577547e-01 -1.55798480e-01
5.30099928e-01 -4.87423211e-01 -2.02142686e-01 3.77606452e-01
-1.22092426e+00 -1.63498670e-01 -1.84290931e-01 1.39857054e-01
-2.01667339e-01 6.30809665e-01 -6.05907321e-01 -6.04206383e-01
-3.95889074e-01 -7.40702033e-01 1.10292697e+00 1.14549913e-01
-1.02917385e+00 -6.93512082e-01 -9.62868426e-03 5.00155091e-01
-3.21621560e-02 -2.37550884e-01 1.36358345e+00 -1.25433171e+00
-2.32473746e-01 -2.61609703e-01 -1.86507255e-01 2.12124527e-01
8.92763138e-02 4.10965830e-01 -4.59258229e-01 5.68928123e-02
-4.67923969e-01 -3.82438421e-01 8.86817694e-01 -3.25768925e-02
1.07666564e+00 -7.31355965e-01 -6.48642659e-01 4.46422368e-01
1.11549926e+00 4.16875601e-01 4.36309546e-01 8.43204916e-01
5.09218574e-01 9.40829515e-01 6.57156646e-01 5.15532851e-01
2.19347104e-01 3.38205367e-01 -1.49376675e-01 2.08284199e-01
-7.20559135e-02 -6.29926145e-01 4.84343976e-01 7.32865453e-01
5.17352223e-01 -8.01889837e-01 -1.06561744e+00 5.97562909e-01
-1.43822622e+00 -1.17026889e+00 -3.58891070e-01 1.91018653e+00
1.27251995e+00 5.75592995e-01 3.35087657e-01 4.90186393e-01
8.06697965e-01 1.68947168e-02 -1.55985624e-01 -3.28518927e-01
-1.02909990e-01 8.08049366e-02 4.77362067e-01 2.90182263e-01
-1.33014238e+00 1.02571511e+00 6.00174952e+00 8.15832019e-01
-6.52323961e-01 -5.14719486e-01 9.01227072e-02 4.16871786e-01
-4.85738307e-01 2.97529638e-01 -1.65833426e+00 8.19894075e-01
8.43875647e-01 -2.70942390e-01 2.24305794e-01 1.06634939e+00
2.43374124e-01 -2.33787209e-01 -1.26466084e+00 3.46515775e-01
1.13239497e-01 -1.43247712e+00 3.79863799e-01 8.99130777e-02
7.15705037e-01 -4.01747227e-01 -3.02341521e-01 3.64617705e-01
3.46277654e-01 -5.10586739e-01 1.06990850e+00 1.41923293e-01
7.64493883e-01 -3.68731797e-01 5.52915156e-01 5.17994642e-01
-8.87040138e-01 -1.24846073e-03 -5.28886206e-02 4.86589372e-01
-2.43342891e-02 4.98562664e-01 -1.06559718e+00 2.85585761e-01
4.83801693e-01 6.79444611e-01 -8.13104451e-01 8.77583623e-01
-2.08881035e-01 8.97332191e-01 -1.33817881e-01 -5.58081806e-01
2.44422331e-01 -1.54041901e-01 6.29433990e-01 1.58751845e+00
3.14447254e-01 -2.40065679e-01 3.35521042e-01 9.76130366e-01
-2.20560595e-01 4.16844130e-01 -3.50298017e-01 -6.39794827e-01
8.70369196e-01 1.10513771e+00 -1.07135451e+00 -5.54933429e-01
-4.70541507e-01 7.94787645e-01 -5.71440421e-02 2.60531068e-01
-4.70682949e-01 -1.32449210e+00 9.51057673e-02 5.69456577e-01
4.26794082e-01 9.80750192e-03 -4.65437591e-01 -1.15732431e+00
2.61240155e-01 -1.18738532e+00 5.08706093e-01 -7.37032294e-01
-1.26558185e+00 4.04221863e-01 -3.77705395e-02 -1.19789755e+00
-4.43795621e-01 -6.83567643e-01 -4.54254478e-01 8.26792836e-01
-1.10052776e+00 -9.37041521e-01 -2.15133056e-01 1.86674565e-01
9.31186974e-01 -1.73612982e-01 4.97296035e-01 1.08279198e-01
-7.27511525e-01 6.22400939e-01 6.51271045e-01 6.98491395e-01
1.01522005e+00 -1.50137591e+00 5.96645534e-01 7.79417932e-01
3.11289094e-02 1.13262236e+00 7.28055000e-01 -1.13962817e+00
-1.20231926e+00 -1.03063500e+00 1.65695715e+00 -5.98657370e-01
1.05176497e+00 -7.18498647e-01 -1.00118828e+00 6.07546031e-01
3.81377131e-01 -1.00431514e+00 7.23751366e-01 1.36478424e-01
-4.19485182e-01 1.67901114e-01 -5.38194239e-01 6.43215835e-01
8.60136151e-01 -3.27635288e-01 -1.13387251e+00 5.83233595e-01
6.85812712e-01 -2.76838720e-01 -6.74069703e-01 -2.44478479e-01
5.59449255e-01 -5.05548894e-01 5.36549687e-01 -4.12138373e-01
8.31151545e-01 -7.73854926e-02 3.01360101e-01 -1.07294321e+00
-1.53535813e-01 -9.46942866e-01 1.36524007e-01 2.02048135e+00
8.78415227e-01 -3.78801286e-01 6.71544313e-01 3.22141290e-01
-1.23445414e-01 -2.83257037e-01 -1.57734647e-01 -9.85219061e-01
1.00167945e-01 -2.91259080e-01 4.89214063e-01 7.65500844e-01
6.88557506e-01 7.95092702e-01 4.26916257e-02 -5.53530037e-01
4.71643597e-01 2.74099141e-01 8.39975595e-01 -1.43545389e+00
1.18668973e-01 -8.72782469e-01 4.57689166e-01 -1.27813411e+00
5.72884120e-02 -1.26120341e+00 3.63142371e-01 -1.74686885e+00
3.14109802e-01 -4.35477883e-01 4.42334473e-01 3.81262958e-01
-1.11355409e-01 -4.95854914e-01 -2.01815471e-01 6.69197857e-01
-4.59304839e-01 3.05767953e-02 6.29192173e-01 -1.60814151e-01
-5.31525433e-01 2.01726884e-01 -6.86208904e-01 8.34145784e-01
6.66348219e-01 -6.52850270e-01 -1.29551262e-01 2.93144006e-02
3.31141800e-01 -1.73241884e-01 -2.12926596e-01 -8.62044871e-01
1.84867904e-01 -3.63432944e-01 3.63884747e-01 -1.30009151e+00
-4.03109521e-01 -3.55945170e-01 -3.68782490e-01 1.53890669e-01
-1.04912102e+00 -3.75385918e-02 -1.05209708e-01 1.76101491e-01
9.16757993e-03 -8.74188721e-01 5.91796637e-01 -2.74239570e-01
-3.30371201e-01 -1.00943275e-01 -9.71040368e-01 5.86184442e-01
5.18766284e-01 -2.89596766e-01 -5.76272488e-01 1.04641885e-01
-2.95007497e-01 1.21220976e-01 5.65177441e-01 5.69835842e-01
2.97725350e-01 -9.44031417e-01 -5.22608399e-01 2.27542669e-01
4.03623104e-01 -4.60566878e-01 -4.11739558e-01 4.29775685e-01
-5.77792346e-01 8.42386842e-01 2.71405131e-02 -3.57198864e-01
-1.26912475e+00 5.01836538e-01 -4.39435273e-01 -5.36653996e-01
-4.76699084e-01 5.06570160e-01 -1.81504264e-01 -3.87488931e-01
5.79272389e-01 -5.16230702e-01 -5.86083114e-01 5.71875811e-01
6.48713350e-01 4.04841691e-01 1.15762003e-01 -5.31012952e-01
-1.41362831e-01 1.40589863e-01 -4.71931666e-01 -3.46908867e-01
1.14223278e+00 -2.60764658e-01 -3.47576529e-01 5.34892380e-01
1.28110611e+00 3.83667886e-01 -5.86968899e-01 -3.48494768e-01
8.87831151e-01 -6.86300620e-02 -2.60116607e-01 -9.52871621e-01
-3.04445401e-02 5.79509437e-01 -1.90273046e-01 6.54317677e-01
6.37764871e-01 2.10161179e-01 9.90304828e-01 4.06967163e-01
-7.56376833e-02 -1.66641748e+00 2.12311476e-01 1.00844109e+00
7.62337387e-01 -1.00361681e+00 3.63581441e-02 -5.90315521e-01
-5.40472329e-01 1.40140676e+00 6.73956096e-01 4.98254299e-01
3.77304345e-01 3.47041577e-01 -1.19465120e-01 -1.41248822e-01
-6.77833140e-01 -2.00137705e-01 5.24732351e-01 4.53850865e-01
7.97189951e-01 -3.23426127e-01 -8.93048942e-01 7.35330462e-01
-6.67685390e-01 -3.45295072e-01 7.53105938e-01 1.11171901e+00
-7.09227264e-01 -1.30522549e+00 -5.43930233e-01 9.21380877e-01
-7.23035157e-01 -3.31827074e-01 -8.92194152e-01 8.30223262e-01
-1.17603928e-01 8.31012428e-01 5.94155006e-02 -2.16679916e-01
2.09661126e-01 5.34856796e-01 1.10222530e-02 -1.01863658e+00
-8.31331670e-01 4.85064209e-01 4.32739973e-01 -7.81294256e-02
-8.64756852e-03 -1.04181087e+00 -1.14985538e+00 -8.67011175e-02
-4.01216358e-01 2.94932216e-01 7.21773684e-01 7.87512422e-01
1.11261383e-01 2.65844077e-01 5.53926706e-01 -4.79710728e-01
-6.92090034e-01 -9.88337874e-01 -3.86266023e-01 6.57665193e-01
5.79816289e-02 -3.67505789e-01 -5.09167969e-01 6.12990797e-01] | [11.304261207580566, 8.663585662841797] |
820529b0-3541-4567-8da1-0f2223bd6275 | amharic-text-clustering-using-encyclopedic | 2105.00809 | null | https://arxiv.org/abs/2105.00809v2 | https://arxiv.org/pdf/2105.00809v2.pdf | Amharic Text Clustering Using Encyclopedic Knowledge with Neural Word Embedding | In this digital era, almost in every discipline people are using automated systems that generate information represented in document format in different natural languages. As a result, there is a growing interest towards better solutions for finding, organizing and analyzing these documents. In this paper, we propose a system that clusters Amharic text documents using Encyclopedic Knowledge (EK) with neural word embedding. EK enables the representation of related concepts and neural word embedding allows us to handle the contexts of the relatedness. During the clustering process, all the text documents pass through preprocessing stages. Enriched text document features are extracted from each document by mapping with EK and word embedding model. TF-IDF weighted vector of enriched feature was generated. Finally, text documents are clustered using popular spherical K-means algorithm. The proposed system is tested with Amharic text corpus and Amharic Wikipedia data. Test results show that the use of EK with word embedding for document clustering improves the average accuracy over the use of only EK. Furthermore, changing the size of the class has a significant effect on accuracy. | ['Yeregal Assabie', 'Dessalew Yohannes'] | 2021-03-31 | null | null | null | null | ['text-clustering'] | ['natural-language-processing'] | [-4.0925592e-01 -1.5475189e-02 7.5328231e-02 -1.1590752e-01
-1.0483571e-01 -5.6034452e-01 8.1049562e-01 7.8108662e-01
-7.9917794e-01 3.3971110e-01 7.3323363e-01 7.4942581e-02
-3.7570456e-01 -1.0317972e+00 1.4203332e-02 -6.2472707e-01
2.3404171e-01 4.5415682e-01 9.3941696e-02 -3.1632152e-01
8.5100514e-01 2.9502755e-01 -1.7193881e+00 3.3605620e-01
8.4322304e-01 3.2980317e-01 5.4352587e-01 7.1266586e-01
-9.4617420e-01 3.7736702e-01 -5.9625900e-01 -1.3575189e-01
9.1595072e-03 -1.8354648e-01 -8.9821434e-01 -2.4129671e-01
-1.3588996e-01 2.1775848e-01 -3.1833827e-01 8.1739801e-01
4.4167715e-01 5.2307171e-01 9.5618629e-01 -9.4822770e-01
-9.7415394e-01 9.6786904e-01 -3.2994798e-01 4.7939522e-03
1.6052468e-01 -7.1321839e-01 9.5305026e-01 -1.1831213e+00
6.9858712e-01 1.2321627e+00 3.9699608e-01 1.8762681e-01
-7.4458873e-01 -2.4635127e-01 -1.6514228e-01 4.5568135e-01
-1.5610890e+00 2.2573805e-01 6.0678500e-01 -5.5430597e-01
1.1339707e+00 1.7092595e-01 6.7017537e-01 3.2740152e-01
2.3282769e-01 3.4591767e-01 4.9915576e-01 -8.8204956e-01
1.3757949e-01 6.8452525e-01 8.5545319e-01 4.5518488e-01
5.5109656e-01 -7.6889426e-01 -1.5405840e-01 -1.0619828e-01
-2.0357147e-02 1.1117527e-01 -6.5962777e-02 -2.6020110e-01
-1.0013125e+00 9.1410321e-01 1.3106029e-01 8.3832240e-01
-3.2882896e-01 -7.7674016e-02 7.2000688e-01 1.6566277e-02
3.0803719e-01 6.0135370e-01 -9.5496491e-02 -5.6504078e-02
-8.5178852e-01 8.1431046e-02 4.8870593e-01 7.9218918e-01
7.6634330e-01 -2.3420997e-01 -1.8681115e-02 1.1056324e+00
5.2139276e-01 2.8419670e-01 1.3550800e+00 -3.7242672e-01
2.0275190e-01 1.2532773e+00 1.9238902e-02 -1.6204835e+00
-4.8127648e-01 -5.8214560e-02 -5.0630724e-01 -1.9748524e-01
-1.4839645e-01 -6.0295098e-02 -7.3433131e-01 1.1538187e+00
4.2029765e-01 -2.5337678e-01 5.5671465e-01 5.1069409e-01
9.0060270e-01 1.2463229e+00 -8.5793212e-02 -1.1236392e-02
1.6318618e+00 -7.9189086e-01 -9.4226682e-01 3.4258375e-01
8.8436407e-01 -1.0892410e+00 8.6315882e-01 1.3843934e-01
-5.6493706e-01 -5.9000182e-01 -9.8286295e-01 -2.7352539e-01
-1.2399173e+00 3.4207469e-01 3.0411121e-01 5.9973031e-01
-9.2024451e-01 3.8891125e-01 -5.2060437e-01 -8.5883790e-01
-8.9129947e-02 3.3066803e-01 -6.6482145e-01 -2.0697365e-02
-1.1564618e+00 8.5861778e-01 1.1367149e+00 -3.1959382e-01
-6.5769158e-02 -3.3116004e-01 -7.5689077e-01 2.4505749e-01
-9.8564468e-02 -2.0798261e-01 4.0118366e-01 -7.3325098e-01
-9.4514811e-01 4.9205378e-01 -1.2140813e-01 -3.2268617e-01
-2.4792026e-01 -2.1091397e-01 -6.4509177e-01 1.9828339e-01
1.0137726e-01 5.2884871e-01 4.3767175e-01 -1.0383588e+00
-6.9877601e-01 -4.5445052e-01 -3.6010265e-01 4.4082299e-01
-1.2386795e+00 2.2808323e-02 -5.0556737e-01 -5.6036198e-01
1.2457383e-01 -8.9574641e-01 8.5556865e-02 -5.0107247e-01
-1.4760014e-01 -6.9711304e-01 1.0898269e+00 -7.4775702e-01
1.5550535e+00 -2.3548894e+00 1.3061620e-01 5.0770992e-01
2.0999444e-01 1.6872002e-01 6.4835899e-02 9.8936659e-01
5.0308585e-02 3.1205025e-01 2.5535442e-02 1.6585515e-01
6.8831734e-02 8.1301808e-02 -1.2892736e-01 4.3822363e-02
-3.8556191e-01 3.9909077e-01 -7.3229313e-01 -8.3651966e-01
1.4971079e-01 6.1285472e-01 -5.1563257e-01 -2.6976924e-02
1.5461311e-01 -4.3588588e-01 -4.6317452e-01 1.0495013e-01
4.0039116e-01 1.9781858e-02 9.9911846e-02 -3.2976702e-01
-3.5648036e-01 -3.2424450e-01 -1.3411421e+00 1.2010734e+00
-5.3498906e-01 9.7345370e-01 -5.1592356e-01 -9.9335676e-01
1.1363755e+00 4.0306950e-01 3.9413077e-01 -1.5389444e-01
2.9918262e-01 5.9533719e-02 1.3225307e-01 -7.4993181e-01
1.2758349e+00 2.7910337e-01 4.1563820e-02 7.2518319e-01
1.3118741e-01 9.5685758e-02 5.1278114e-01 6.8053293e-01
8.2514888e-01 -3.8704941e-01 3.2392898e-01 -4.5075724e-01
7.6313704e-01 2.0258440e-01 -4.7172777e-02 1.9292322e-01
2.6543000e-01 2.6455852e-01 4.0152043e-02 -2.6579422e-01
-1.1089696e+00 -6.6556072e-01 -1.0230266e-01 1.0367264e+00
5.7367440e-02 -9.0788931e-01 -7.0347261e-01 -3.0766177e-01
4.4003394e-02 1.0080968e+00 -7.0037943e-01 -3.9023474e-01
-3.0990744e-01 -6.3253146e-01 4.6587360e-01 3.6305407e-01
2.4345112e-01 -1.0576180e+00 -3.9461184e-01 2.8803948e-01
-1.4092937e-01 -6.0318363e-01 -3.5466558e-01 1.1720575e-03
-7.5863969e-01 -9.0765792e-01 -6.3369787e-01 -1.1379472e+00
9.7064173e-01 4.3731838e-01 4.7073051e-01 -2.7383503e-02
-5.6598902e-01 6.3698596e-01 -8.6149192e-01 -3.8932788e-01
-3.0472827e-01 2.8105450e-01 2.5062969e-01 -1.2869124e-01
9.2596573e-01 -4.3320496e-02 -3.5470697e-01 -1.0063092e-01
-1.1251374e+00 -9.2149884e-02 2.1914716e-01 5.3678495e-01
2.2675423e-01 4.9410614e-01 6.2658775e-01 -6.1337537e-01
1.2474530e+00 -7.3321974e-01 -1.5983318e-01 3.6675847e-01
-9.2910767e-01 4.0282252e-01 8.5418117e-01 -1.6882555e-01
-1.0419585e+00 -3.0278030e-01 3.0495158e-01 1.3763380e-01
8.8061467e-03 7.4479586e-01 1.5514864e-01 4.8068479e-01
7.0000041e-01 4.6553686e-01 -1.5431871e-01 -3.7963146e-01
6.5835571e-01 1.4461471e+00 1.8715344e-01 -4.1558632e-01
5.0691181e-01 2.6974478e-01 -4.2096090e-01 -1.1288894e+00
-5.1591370e-02 -7.8726608e-01 -6.0303271e-01 -3.9316925e-01
1.1327593e+00 -5.8908129e-01 -2.0366223e-01 -1.0994484e-01
-1.1050205e+00 5.7926005e-01 -4.4339623e-02 7.7355844e-01
1.7835006e-02 5.1727676e-01 -2.3027758e-01 -6.5052193e-01
-5.3155702e-01 -8.0992538e-01 4.2686519e-01 2.5038284e-01
-4.8184267e-01 -1.0517526e+00 4.7256789e-01 2.3794584e-01
1.9702230e-01 -1.5959142e-01 1.3835570e+00 -9.5733231e-01
4.8180358e-03 -4.6444353e-01 -9.1301374e-02 4.2177844e-01
4.4758010e-01 3.3184928e-01 -5.2426714e-01 -8.6900048e-02
-5.0901037e-01 1.6078584e-01 8.1703889e-01 2.2343303e-01
9.2315966e-01 -2.4922773e-01 -3.8603112e-01 1.2849872e-01
1.6037458e+00 4.4621560e-01 5.1151705e-01 6.9523966e-01
7.3318070e-01 8.2372564e-01 4.6221629e-01 5.1243728e-01
4.3435067e-01 3.1383061e-01 -1.0343732e-01 4.6384764e-01
8.2391426e-02 1.2213184e-02 3.3494207e-01 1.4278920e+00
-5.5201214e-02 -3.7064096e-01 -1.1557320e+00 9.1875613e-01
-1.8182814e+00 -1.0199672e+00 -2.2865862e-01 1.8240024e+00
6.8075889e-01 -3.0059645e-01 -2.6397493e-01 4.4048265e-01
9.9859172e-01 -2.1993411e-01 -4.9167354e-02 -7.3632699e-01
-1.5361142e-02 -6.4825505e-02 3.1282720e-01 5.4324120e-01
-6.8047082e-01 1.1164403e+00 4.9486690e+00 6.7268395e-01
-8.5852504e-01 -1.4462802e-01 3.0158618e-03 -1.0549550e-02
-4.1156709e-01 -8.4779784e-02 -8.0011070e-01 5.6975329e-01
8.9522821e-01 -9.4929314e-01 1.8946655e-01 7.6516014e-01
1.5289019e-01 -1.2761648e-01 -6.0930192e-01 9.5885259e-01
4.5635656e-01 -1.2779748e+00 4.8131528e-01 -1.1890115e-01
6.8112057e-01 -2.2299252e-01 -9.9610351e-02 2.2921750e-01
3.2494152e-01 -8.2904524e-01 2.5897646e-01 5.2212983e-01
2.1096812e-01 -1.2281002e+00 1.0839033e+00 1.5186061e-01
-1.0826869e+00 -6.7883193e-02 -7.6767480e-01 2.4779886e-01
-1.1767037e-01 5.4192203e-01 -8.6313379e-01 4.5679140e-01
7.2277808e-01 4.8923901e-01 -7.4596834e-01 7.7165729e-01
3.3277044e-01 3.4693566e-01 -2.2094333e-01 -5.9877306e-01
4.0505996e-01 -3.3252323e-01 3.4327656e-01 1.5569092e+00
6.8930018e-01 9.7729668e-02 -3.5803431e-01 3.0460751e-01
-7.6004617e-02 9.9949622e-01 -8.1370914e-01 -5.0813919e-01
7.0625198e-01 1.3578556e+00 -1.1506685e+00 -5.0882810e-01
-2.0397171e-01 8.6514276e-01 7.1276367e-02 4.5730218e-02
-4.8170018e-01 -1.1288487e+00 2.5351340e-01 -3.1262573e-02
1.9313216e-01 -3.0359262e-01 -4.6076220e-02 -8.2970822e-01
-1.9349544e-01 -4.6654436e-01 2.7614224e-01 -8.0946141e-01
-9.7668320e-01 7.5217259e-01 1.0352389e-01 -8.9050066e-01
-1.5929508e-01 -5.5515969e-01 -5.0053251e-01 6.4904851e-01
-7.6649910e-01 -7.1765238e-01 -3.6675367e-01 5.5201942e-01
5.3057235e-01 -7.4237388e-01 9.9594003e-01 2.0760550e-01
-4.0109891e-01 1.3518803e-01 8.4057689e-01 2.6183259e-01
9.2287070e-01 -1.2510538e+00 -1.8375124e-01 5.9936202e-01
3.1893343e-01 1.0922779e+00 6.0728842e-01 -6.8412715e-01
-1.0150499e+00 -1.0624236e+00 1.2586840e+00 -1.3802318e-01
5.1508433e-01 -2.1597829e-01 -8.6942875e-01 4.3690953e-01
6.9259655e-01 -5.4780889e-01 1.3510811e+00 1.4757797e-03
-9.7012304e-02 -3.9518528e-02 -1.0906616e+00 7.7526814e-01
2.3687845e-01 -3.7025601e-01 -1.0681105e+00 2.3908348e-01
7.0459640e-01 4.3438983e-01 -8.3589846e-01 -3.8109520e-01
4.6229830e-01 -4.2553243e-01 8.0017626e-01 -5.2391762e-01
4.4887796e-01 -5.4721874e-01 -3.2837126e-01 -1.6132735e+00
-4.3281966e-01 1.2739407e-01 1.2671997e-01 1.3427457e+00
3.5470924e-01 -3.5350397e-01 4.6738321e-01 3.6181834e-01
-2.2389574e-02 -3.2619849e-01 -4.8001578e-01 -4.7202450e-01
9.9283993e-02 -1.6576423e-01 6.1425442e-01 1.1991643e+00
5.9996170e-01 4.0398598e-01 2.6097579e-02 -4.6676993e-02
3.1680652e-01 -1.9723225e-01 5.3488350e-01 -1.4741343e+00
2.9536885e-01 -3.4409779e-01 -7.2773445e-01 -2.5525254e-01
1.8730724e-01 -1.3864810e+00 -2.4554823e-01 -2.0106483e+00
2.6912907e-01 -1.3705558e-01 -2.9471099e-01 2.4564056e-01
-2.9293975e-02 -9.6105244e-03 1.8073834e-01 2.5490108e-01
-4.1850597e-01 6.1680961e-01 5.7795519e-01 -7.1090102e-02
-3.3178854e-01 -7.6702470e-01 -6.1498898e-01 6.7998010e-01
1.1075155e+00 -5.7146245e-01 -6.4570796e-01 -4.8980898e-01
4.8554185e-01 -6.0269475e-01 -2.0190921e-01 -1.0616031e+00
5.3973395e-01 -3.0848792e-02 4.4010088e-01 -7.4932736e-01
2.3390243e-02 -9.5589513e-01 -1.1936442e-01 2.4620891e-01
-4.4699970e-01 5.3801608e-01 2.4699114e-01 3.5042214e-01
-3.5188627e-01 -8.0430478e-01 6.3979435e-01 5.0672814e-02
-7.7667344e-01 -2.0169945e-01 -1.0187536e+00 -1.8182018e-01
1.1795105e+00 -4.2311573e-01 -5.4274671e-02 -1.5813027e-01
-5.4669887e-01 2.3577264e-01 2.7869210e-01 6.8720686e-01
7.6829594e-01 -1.4570490e+00 -5.9657949e-01 8.1706949e-02
3.5663468e-01 -3.0990756e-01 1.2882185e-01 1.1977160e-01
-8.3690327e-01 7.4683893e-01 -2.3098272e-01 -7.5157963e-02
-1.6206472e+00 4.2461053e-01 -2.3335734e-01 6.0215347e-02
-5.5932647e-01 3.8558173e-01 -2.6118356e-01 -6.1376393e-01
7.2537432e-03 -5.8574561e-02 -1.0157546e+00 5.4112136e-01
5.6034297e-01 5.9417558e-01 -9.8179698e-02 -8.9084220e-01
-1.9969051e-01 7.4645156e-01 -3.7222520e-01 -3.0231041e-01
1.3834764e+00 -1.5430468e-01 -4.8837891e-01 6.0727721e-01
1.3461375e+00 2.1303168e-01 -3.2093149e-02 6.2586293e-03
3.0446437e-01 -1.9247429e-01 3.0051169e-01 -5.1830482e-01
-5.5802143e-01 7.4706173e-01 8.1847078e-01 2.1974143e-01
6.9264340e-01 -1.2238796e-01 5.3442991e-01 9.2183858e-01
-9.2387080e-02 -1.7905211e+00 -1.1342318e-01 6.6531485e-01
5.8866280e-01 -9.8794729e-01 4.4457458e-02 9.6696742e-02
-6.5572280e-01 1.4101195e+00 4.9434778e-01 -9.9688098e-02
9.4845569e-01 -1.6545122e-02 1.1997914e-01 -2.3031467e-01
-6.1749923e-01 -5.4643542e-02 2.5315860e-01 4.7186762e-01
7.9316300e-01 -6.2936045e-02 -8.8915318e-01 6.6627014e-01
-5.3212166e-01 -3.1670177e-01 7.0972133e-01 8.1303775e-01
-9.1803288e-01 -1.1196864e+00 -5.6659263e-01 5.2889639e-01
-3.8243586e-01 -1.2657185e-01 -6.4267796e-01 8.3684826e-01
3.2936129e-01 1.0137823e+00 3.9586353e-01 -3.0148596e-01
-1.7301807e-01 4.1498682e-01 -2.6158616e-03 -7.5076652e-01
-4.5771819e-01 -3.0695412e-01 -2.1738431e-01 1.0777961e-01
-3.3842304e-01 -3.7350091e-01 -1.8640854e+00 -2.9442480e-01
-5.1627690e-01 9.3625349e-01 1.0486996e+00 6.2829155e-01
4.5193931e-01 4.6211150e-01 5.0066000e-01 -3.0207974e-01
5.0443970e-02 -1.1642358e+00 -4.9859902e-01 4.4499919e-01
-2.0929456e-01 -4.8186880e-01 -2.8985077e-01 2.6882699e-01] | [10.314980506896973, 8.580270767211914] |
970b3835-fe6f-49e1-bed8-e05c7db81596 | language-models-are-causal-knowledge | 2304.03754 | null | https://arxiv.org/abs/2304.03754v1 | https://arxiv.org/pdf/2304.03754v1.pdf | Language Models are Causal Knowledge Extractors for Zero-shot Video Question Answering | Causal Video Question Answering (CVidQA) queries not only association or temporal relations but also causal relations in a video. Existing question synthesis methods pre-trained question generation (QG) systems on reading comprehension datasets with text descriptions as inputs. However, QG models only learn to ask association questions (e.g., ``what is someone doing...'') and result in inferior performance due to the poor transfer of association knowledge to CVidQA, which focuses on causal questions like ``why is someone doing ...''. Observing this, we proposed to exploit causal knowledge to generate question-answer pairs, and proposed a novel framework, Causal Knowledge Extraction from Language Models (CaKE-LM), leveraging causal commonsense knowledge from language models to tackle CVidQA. To extract knowledge from LMs, CaKE-LM generates causal questions containing two events with one triggering another (e.g., ``score a goal'' triggers ``soccer player kicking ball'') by prompting LM with the action (soccer player kicking ball) to retrieve the intention (to score a goal). CaKE-LM significantly outperforms conventional methods by 4% to 6% of zero-shot CVidQA accuracy on NExT-QA and Causal-VidQA datasets. We also conduct comprehensive analyses and provide key findings for future research. | ['Shih-Fu Chang', 'Winston H. Hsu', 'Xudong Lin', 'Yulei Niu', 'Hung-Ting Su'] | 2023-04-07 | null | null | null | null | ['video-question-answering', 'reading-comprehension', 'question-generation'] | ['computer-vision', 'natural-language-processing', 'natural-language-processing'] | [ 2.23620296e-01 4.20705140e-01 -1.79571077e-01 -3.19270432e-01
-1.13528228e+00 -5.81450462e-01 6.60349369e-01 9.76136550e-02
8.17802548e-02 1.03598082e+00 7.99521625e-01 -5.90788007e-01
-4.36029255e-01 -1.01232326e+00 -1.10475159e+00 -2.16504291e-01
7.54612163e-02 2.11594835e-01 4.17629659e-01 -3.62779796e-01
1.39357537e-01 -4.20965880e-01 -1.65920246e+00 8.96743476e-01
8.40661466e-01 8.66926670e-01 2.24336028e-01 1.13399088e+00
-2.66027629e-01 2.30515742e+00 -8.68470430e-01 -7.93816149e-01
-3.76598656e-01 -1.26325786e+00 -1.48572814e+00 -3.77299815e-01
6.28970742e-01 -6.67962313e-01 -6.38495564e-01 6.75819337e-01
6.91415370e-02 1.94646537e-01 7.22722411e-01 -1.71020591e+00
-9.03361261e-01 8.36425126e-01 5.63982725e-02 6.47127390e-01
1.32505631e+00 3.53251815e-01 1.54466462e+00 -4.45609719e-01
6.57526553e-01 1.45283127e+00 2.06912801e-01 7.82882810e-01
-7.92503476e-01 -6.22581244e-01 2.99351275e-01 1.20055008e+00
-1.02561677e+00 -9.74942595e-02 6.44164383e-01 -4.80157435e-01
9.97420609e-01 5.53664029e-01 6.25071764e-01 1.44846082e+00
3.05612147e-01 1.08326232e+00 8.58458638e-01 -2.79876173e-01
1.66554108e-01 -3.80260289e-01 1.61709800e-01 7.20701814e-01
-2.70926148e-01 -8.76465514e-02 -1.20604801e+00 -1.27539188e-01
6.35486424e-01 -6.67207539e-01 -2.39643738e-01 2.69405663e-01
-1.42570817e+00 1.06741714e+00 1.06571965e-01 -1.66765392e-01
-5.82675874e-01 6.69457018e-01 1.61537841e-01 3.34654570e-01
-2.83174068e-01 4.95589465e-01 -3.80973399e-01 -4.84556615e-01
-3.15916836e-01 8.83529305e-01 9.56023633e-01 1.19580400e+00
2.65738815e-01 -2.28703216e-01 -7.45585620e-01 3.52581292e-01
3.97864342e-01 6.04116917e-01 -3.74812480e-05 -1.49716997e+00
6.12709284e-01 4.61132526e-01 1.10724531e-01 -1.17709970e+00
-1.86574832e-01 2.28123784e-01 -3.00480276e-01 -5.94457626e-01
5.89077830e-01 -3.10182154e-01 -4.47591603e-01 1.78020692e+00
2.55933076e-01 3.81075740e-01 1.59476161e-01 1.18056989e+00
1.46189797e+00 9.17102873e-01 4.91632819e-01 -3.56819689e-01
1.98210514e+00 -6.44684613e-01 -1.17933953e+00 -1.09575167e-01
5.02841890e-01 -4.69514281e-01 1.44130075e+00 5.70740938e-01
-1.06642497e+00 -4.35100347e-01 -5.96981108e-01 -2.21745148e-01
1.90197065e-01 -2.48506814e-01 6.74600184e-01 2.17013717e-01
-7.77079880e-01 -5.49068712e-02 -2.62125909e-01 -2.41792306e-01
3.73577595e-01 -9.58778784e-02 -3.20805833e-02 -3.24244678e-01
-2.00107121e+00 6.18678689e-01 3.38143259e-01 -2.42519587e-01
-1.56150687e+00 -9.23260093e-01 -7.90403843e-01 -7.58736804e-02
1.15276003e+00 -1.06512833e+00 1.81811905e+00 -7.85351813e-01
-1.45554423e+00 4.32400405e-01 -6.04793608e-01 -5.37779391e-01
1.19386218e-01 -7.15175152e-01 -5.60971200e-01 7.68864810e-01
5.33961654e-01 5.98781943e-01 7.13391483e-01 -1.10818207e+00
-8.51086676e-01 1.21268714e-02 7.17553079e-01 1.27837807e-01
1.82431802e-01 3.80683124e-01 -2.54587620e-01 -5.39816380e-01
-2.37409696e-01 -3.59733343e-01 2.32789963e-01 -5.31355083e-01
-4.31612015e-01 -8.45589638e-01 6.40489221e-01 -8.96515608e-01
1.69056976e+00 -1.57687759e+00 1.02849528e-01 -3.34180921e-01
3.03219557e-01 -7.65442923e-02 -1.52582034e-01 6.70156240e-01
-1.52581677e-01 1.63145542e-01 6.93902746e-02 5.56181669e-01
-1.64895412e-03 3.62852603e-01 -7.46425748e-01 -2.52921224e-01
7.15546191e-01 1.21090138e+00 -1.56424260e+00 -9.17110145e-01
-1.95256379e-02 4.45806049e-03 -7.76776195e-01 6.38691843e-01
-9.91658509e-01 4.56616133e-01 -6.06003582e-01 4.77729052e-01
-8.08876604e-02 -3.96964431e-01 -8.12824443e-02 -2.56050587e-01
1.99270487e-01 8.36834311e-01 -8.37532818e-01 1.46307933e+00
-7.02220500e-02 7.23092735e-01 -5.52915454e-01 -7.63488889e-01
5.61026692e-01 7.05983043e-01 2.50327319e-01 -8.45882297e-01
-1.57303326e-02 -3.30309600e-01 5.61880656e-02 -1.36472821e+00
1.14944138e-01 -3.56529474e-01 -2.31389225e-01 3.51401746e-01
2.43276298e-01 -2.20022187e-01 4.40304965e-01 1.00478959e+00
1.58852744e+00 3.19935530e-01 2.58845478e-01 2.22014293e-01
7.99721420e-01 6.02299690e-01 3.89447242e-01 9.32118833e-01
-2.18287691e-01 4.09656435e-01 1.26775956e+00 -5.32281622e-02
-5.38121939e-01 -1.32882500e+00 5.14476299e-01 1.24938822e+00
2.08227724e-01 -7.33707845e-01 -7.97602773e-01 -9.09465075e-01
-3.33368927e-01 1.71604407e+00 -6.69080436e-01 -2.71932304e-01
-7.40457535e-01 -3.27456385e-01 8.80322874e-01 5.31082809e-01
6.35356903e-01 -1.25280058e+00 -5.25100470e-01 2.80120373e-01
-1.21336210e+00 -1.17575443e+00 -2.34875068e-01 -4.29756910e-01
-4.84717935e-01 -1.48523211e+00 -5.65819405e-02 -3.19092512e-01
5.66772483e-02 4.29369546e-02 1.66069996e+00 2.39006476e-03
-3.87259945e-02 8.90896142e-01 -8.11905205e-01 -4.59163517e-01
-4.70677763e-01 -4.46480751e-01 -1.92637593e-01 -9.79678780e-02
6.98079169e-01 -2.42305361e-02 -6.36959851e-01 4.78365362e-01
-7.44014919e-01 2.41423860e-01 2.74200648e-01 5.50506055e-01
3.26322049e-01 3.99619602e-02 9.27170455e-01 -5.62495470e-01
9.13049817e-01 -9.76270974e-01 -8.93588439e-02 3.53988856e-01
-1.83409899e-01 -7.38574713e-02 5.06340563e-01 -3.22797000e-01
-1.54007745e+00 -6.39349759e-01 -2.88989335e-01 -1.94474086e-01
-3.62523705e-01 5.96162617e-01 -3.27746928e-01 1.01769376e+00
8.16337407e-01 2.65824497e-01 -2.71369874e-01 2.35580117e-01
6.77971959e-01 1.32835194e-01 9.33624506e-01 -8.45106602e-01
5.68077147e-01 3.12851459e-01 -1.14096977e-01 -3.19130987e-01
-1.21984351e+00 -4.46942598e-01 -1.70787126e-01 -8.69254529e-01
1.50149727e+00 -8.68855357e-01 -1.32157207e+00 5.66502288e-02
-1.36352086e+00 -3.96220207e-01 -1.57364711e-01 4.47680473e-01
-8.80696476e-01 1.30278230e-01 -6.00786567e-01 -9.97235715e-01
8.22975636e-02 -5.50490141e-01 8.55450213e-01 1.50257841e-01
-5.77793419e-01 -7.61237621e-01 -5.58388680e-02 1.04048216e+00
-2.00164184e-01 7.90808201e-02 1.26358306e+00 -4.33036089e-01
-8.13135505e-01 1.62673101e-01 -3.33581626e-01 9.90534723e-02
9.77715477e-02 -1.40212709e-03 -6.15979671e-01 5.71220636e-01
-3.71738896e-02 -5.73880136e-01 2.83448786e-01 2.29528800e-01
9.90658700e-01 -6.99249148e-01 -8.74680001e-03 -3.41294467e-01
1.01617932e+00 4.44698274e-01 8.80435228e-01 4.43804897e-02
6.61702335e-01 9.36191738e-01 1.05427539e+00 3.09018463e-01
1.09805977e+00 4.94944870e-01 6.16210639e-01 5.46629012e-01
-4.17638093e-01 -8.57204914e-01 7.34331667e-01 6.84159100e-01
-3.02966565e-01 -4.25455213e-01 -9.95001256e-01 7.69777298e-01
-1.91286242e+00 -1.31296635e+00 -1.08528221e+00 1.48251534e+00
1.12209356e+00 -1.99879125e-01 4.93748039e-02 2.20818464e-02
2.62790620e-01 -5.41435331e-02 -3.80301833e-01 -2.11689264e-01
-1.36095926e-01 1.22367203e-01 -1.23206578e-01 5.62921405e-01
-6.93953037e-01 1.07555282e+00 5.15257692e+00 7.92083740e-01
-3.15544814e-01 2.40571469e-01 3.46446574e-01 -4.15002517e-02
-6.26232743e-01 3.22261542e-01 -6.84058070e-01 3.24831039e-01
1.04878831e+00 -7.61540532e-02 2.16129765e-01 3.32949609e-01
5.72273552e-01 -4.93942738e-01 -1.15197241e+00 7.59308517e-01
3.27328414e-01 -1.42198181e+00 5.39367855e-01 -6.17967546e-01
4.39928979e-01 -8.10844123e-01 -4.69211996e-01 5.95838428e-01
7.01433420e-01 -1.11392474e+00 8.88779640e-01 9.16423321e-01
4.81029034e-01 -5.09441078e-01 5.80106497e-01 4.98769045e-01
-8.50209475e-01 -2.44859621e-01 -6.55035069e-03 -4.43920940e-01
5.55387914e-01 2.63607383e-01 -9.09233868e-01 8.35193455e-01
1.00659013e+00 3.48623872e-01 -3.77730787e-01 4.37457383e-01
-1.08230138e+00 1.42217159e+00 1.62804291e-01 -4.11134899e-01
1.88588157e-01 1.20201364e-01 6.66686594e-01 8.74439061e-01
1.40949422e-02 1.00748718e+00 -4.66909587e-01 1.09544551e+00
7.74604231e-02 5.92385679e-02 -3.47630292e-01 -1.70925576e-02
2.61349201e-01 6.50330901e-01 -1.10248297e-01 -6.45370483e-01
-4.94888097e-01 8.91505957e-01 9.90685727e-03 5.03340304e-01
-1.24193394e+00 -1.42279670e-01 4.17802364e-01 2.88238049e-01
-1.17181540e-02 3.61532867e-02 7.48360083e-02 -9.59211349e-01
-1.69461697e-01 -1.13501358e+00 8.61979365e-01 -1.52661598e+00
-1.35082531e+00 1.24542557e-01 5.44146597e-01 -7.36211777e-01
-6.23496473e-01 -2.86314160e-01 -5.36107481e-01 6.16837502e-01
-1.21140039e+00 -8.70942891e-01 -5.08278966e-01 8.87886763e-01
8.09012592e-01 1.67512774e-01 4.55677986e-01 5.62587120e-02
-1.79123893e-01 1.63076028e-01 -1.15654647e+00 1.73243880e-01
7.45002985e-01 -1.28856742e+00 -4.76195328e-02 7.71177888e-01
3.74601446e-02 5.11823058e-01 8.94848347e-01 -9.57078099e-01
-1.58581877e+00 -1.01578343e+00 1.32273316e+00 -1.01037729e+00
8.46534669e-01 -4.15148586e-02 -1.14381707e+00 7.26005137e-01
5.09825349e-01 -6.41268075e-01 8.52333724e-01 -1.85250938e-01
-3.31666529e-01 8.21774155e-02 -6.85275972e-01 8.84314656e-01
1.16381192e+00 -5.30740261e-01 -1.27542651e+00 3.76314163e-01
1.27650535e+00 -9.39300731e-02 -7.44417727e-01 2.89185673e-01
2.21641481e-01 -9.81806338e-01 8.90348196e-01 -1.18087554e+00
1.26433420e+00 -5.93336821e-01 -1.93182915e-01 -7.97457099e-01
-9.82441828e-02 -5.39801359e-01 -5.53825617e-01 1.21656227e+00
3.68352860e-01 -5.26068322e-02 4.77909565e-01 7.39829361e-01
-3.88453640e-02 -4.16767299e-01 -8.56500447e-01 -4.58273768e-01
-3.19481567e-02 -1.05763137e+00 4.56829280e-01 8.37532341e-01
2.19234124e-01 8.43838573e-01 -3.94500673e-01 3.83224189e-01
3.76812637e-01 -3.16954345e-01 8.13172877e-01 -6.60465539e-01
-3.61292154e-01 -1.16452657e-01 8.08675122e-03 -1.34071970e+00
1.48139745e-01 -4.82568234e-01 1.77251071e-01 -1.84615934e+00
1.93722203e-01 -1.43813081e-02 1.73998252e-01 4.39450592e-01
-7.86597550e-01 -1.29815072e-01 1.58670291e-01 -2.12406330e-02
-8.52216125e-01 4.94775265e-01 1.57522106e+00 -1.22502394e-01
1.54636363e-02 -2.24227324e-01 -6.76789403e-01 5.59069216e-01
5.75000465e-01 -3.45519423e-01 -9.34898913e-01 -3.18531632e-01
1.00534225e+00 9.85539377e-01 9.73860621e-01 -5.92243552e-01
3.42421085e-01 -4.49885458e-01 -1.03462361e-01 -6.00639343e-01
3.58390957e-02 -2.92102784e-01 7.74235139e-03 2.57328570e-01
-6.94535136e-01 -3.67174260e-02 1.48612738e-01 7.01709569e-01
-5.13689578e-01 -3.86342704e-01 -6.39820769e-02 -1.32732317e-01
-9.06072617e-01 -1.58815190e-01 -8.77873540e-01 4.39389020e-01
1.05942309e+00 -2.26579942e-02 -5.38272560e-01 -1.15830004e+00
-6.30998075e-01 6.66615605e-01 -5.48037291e-01 7.55321026e-01
9.35745001e-01 -1.29885185e+00 -1.02345562e+00 -5.75684130e-01
2.85246193e-01 9.51153263e-02 5.31859577e-01 8.84537339e-01
-3.38886142e-01 6.62063122e-01 2.43873522e-01 -4.28374469e-01
-1.06112659e+00 5.83935261e-01 2.55878597e-01 6.09994028e-03
-3.33623767e-01 1.17419255e+00 3.46026838e-01 1.76618807e-03
4.54644226e-02 -3.52899969e-01 -6.15065455e-01 1.59999669e-01
6.06840193e-01 3.04984778e-01 -3.64447057e-01 -2.82135397e-01
-3.38865250e-01 3.16166244e-02 2.51493722e-01 -2.88418204e-01
6.02842927e-01 -1.99934199e-01 -3.39126438e-01 6.26434445e-01
6.91172421e-01 -1.19757771e-01 -1.04047513e+00 -9.44119617e-02
3.38573009e-02 -4.14583445e-01 -3.51550400e-01 -1.25353360e+00
-4.38883454e-01 6.37384355e-01 -1.11549541e-01 2.56837666e-01
1.10193443e+00 6.40565217e-01 7.44488120e-01 1.46349683e-01
3.39535683e-01 -7.50832260e-01 7.16468811e-01 4.66984481e-01
1.33326042e+00 -1.03849411e+00 -1.93672329e-01 -6.04159236e-01
-9.97348487e-01 9.35093939e-01 9.83834386e-01 1.98193997e-01
2.17219397e-01 -2.20927402e-01 1.74853310e-01 -6.55005872e-01
-1.46247053e+00 -2.27764398e-01 3.24532390e-01 5.02676189e-01
3.69267851e-01 1.75218686e-01 -3.68646443e-01 9.90332603e-01
-4.54237878e-01 1.00624554e-01 6.79868758e-01 5.89358032e-01
-1.48011193e-01 -7.12395668e-01 -6.21572435e-01 6.46504879e-01
-3.19903761e-01 -2.70575345e-01 -6.58389151e-01 7.34117866e-01
2.24924698e-01 1.75899529e+00 5.67726344e-02 -4.49042588e-01
5.21775007e-01 2.04890102e-01 4.51884627e-01 -4.60422993e-01
-6.06255412e-01 -3.27780068e-01 5.00451684e-01 -9.46882665e-01
-6.44472480e-01 -9.57768083e-01 -1.55579054e+00 -2.29288533e-01
-1.20090351e-01 1.73626989e-01 2.24251412e-02 1.20160615e+00
-1.00095067e-02 8.97482991e-01 1.93687484e-01 4.46177930e-01
-1.73540488e-01 -8.84010851e-01 8.14858824e-02 5.51503122e-01
1.59844622e-01 -6.27459526e-01 -2.83356369e-01 6.78802073e-01] | [10.425924301147461, 1.0854419469833374] |
deff59f2-77c7-4c55-b5cc-c27dec2315b9 | fault-tolerant-fpga-implementation-on | 2304.08165 | null | https://arxiv.org/abs/2304.08165v1 | https://arxiv.org/pdf/2304.08165v1.pdf | Fault Tolerant FPGA Implementation on Redundancy Techniques and ECG Denoising | As more the communications and signal process we use in the today life the more we intend to develop more reliable devices which gives fewer errors due to transient fault, So we use a technique called 5-modular redundancy to generate fewer errors. 5-Modular redundancy is an approach to increasing the reliability of hardware systems constructed from component devices that are subject to failure. In digital signal processing and many other important daily life subjects FIR digital filters are used and they alike performing multiplications, complex computations and selecting desired frequency for different applications. FIR filters comprise of multipliers, adders, delay units. FIR filters has been chosen because it offers more steadiness and simple execution because of its limited length and no feedback within the circuit. Fir filter is most important for the signal processing and also used in daily life basis. The noise reduction can be performed by denoising of the signals of all the implementations .The proposed FPGA methods use Xilinx Vivado EDA. | ['Sakthivel SM', 'Sathvik Reddy O'] | 2023-04-17 | null | null | null | null | ['ecg-denoising'] | ['medical'] | [ 1.42132670e-01 -2.41796702e-01 3.73729795e-01 -1.54023930e-01
4.40910101e-01 -5.14592409e-01 4.98659685e-02 4.76992399e-01
-2.15190932e-01 8.04391444e-01 -1.47286728e-01 -2.29507312e-01
-1.57752231e-01 -7.65187919e-01 -2.95558184e-01 -2.50438720e-01
-2.05532476e-01 -2.22108752e-01 6.18542373e-01 -5.32262087e-01
3.81393790e-01 7.01047719e-01 -1.61163294e+00 1.14387915e-01
5.73408365e-01 8.04014027e-01 1.75592661e-01 9.56802070e-01
4.42044288e-01 6.90847635e-01 -8.62640202e-01 1.62696153e-01
5.37021160e-01 -4.95735407e-01 -4.48942065e-01 -1.55891776e-01
-2.52112508e-01 -3.73347193e-01 -1.72748148e-01 1.00153375e+00
3.28377873e-01 8.27424824e-02 6.68592155e-01 -7.86538422e-01
-8.89020041e-02 7.78004825e-01 -5.21529853e-01 5.83209753e-01
3.01572680e-01 6.66785892e-03 1.62934706e-01 -4.71588403e-01
3.83325815e-02 1.28232825e+00 7.62372077e-01 5.52171357e-02
-9.25215960e-01 -5.50147653e-01 -5.52969158e-01 1.91771701e-01
-1.31197822e+00 -3.28546971e-01 4.64102924e-01 -2.00467020e-01
1.23546052e+00 4.34348285e-01 5.04205465e-01 2.54206479e-01
1.13445997e+00 -1.39830157e-01 1.12233460e+00 -6.21318519e-01
1.01996325e-01 2.27347702e-01 4.94286925e-01 5.76377690e-01
9.04495776e-01 -7.31663853e-02 -2.48719618e-01 8.60720426e-02
6.79671824e-01 2.95763701e-01 -2.91022390e-01 9.83804047e-01
-8.11411798e-01 3.59584749e-01 7.58210197e-02 7.87674129e-01
-4.95276988e-01 8.22668672e-01 5.09388328e-01 6.04654372e-01
-3.00558388e-01 5.98716214e-02 -5.52776814e-01 -1.57422081e-01
-7.86813915e-01 1.08769491e-01 4.52072382e-01 8.42068493e-01
3.26671660e-01 6.78080082e-01 7.16384828e-01 1.74161896e-01
4.33841348e-01 4.20019090e-01 9.17970300e-01 -2.38674268e-01
-1.82087794e-01 7.57125318e-01 -1.35447428e-01 -9.98212218e-01
-6.65541410e-01 -6.48101270e-01 -8.07982087e-01 6.87097788e-01
-1.17477275e-01 -2.45639175e-01 -1.03654385e+00 1.06008029e+00
4.47037220e-02 6.98754191e-02 1.72432899e-01 4.82431412e-01
5.92404902e-01 1.16340125e+00 -1.48026437e-01 -1.59917384e-01
1.70914388e+00 -1.69485927e-01 -9.53130662e-01 1.28463969e-01
6.10035248e-02 -1.55701959e+00 3.84201288e-01 9.38045025e-01
-9.26959574e-01 -1.00802672e+00 -1.93032563e+00 2.19360963e-01
-2.48617306e-01 4.48669851e-01 5.00572622e-01 8.86053562e-01
-7.46310294e-01 1.02056384e+00 -8.41366410e-01 -2.06631660e-01
-9.82441753e-02 7.05955863e-01 -3.00386488e-01 5.14008045e-01
-9.64530826e-01 1.00117147e+00 4.30259347e-01 1.59452423e-01
-5.83324552e-01 -5.34462750e-01 -2.26397768e-01 -7.94133451e-03
-1.59099177e-01 -4.38362896e-01 9.32779133e-01 -1.31977963e+00
-1.33057308e+00 2.15028509e-01 2.51109183e-01 -9.22507882e-01
2.77029544e-01 1.96906514e-02 -7.47915089e-01 8.15275311e-02
-5.12801945e-01 2.36924753e-01 1.02354240e+00 -2.15605140e-01
-7.76836753e-01 -1.50480002e-01 -2.22029462e-01 -3.10788542e-01
3.23010460e-02 -1.62404731e-01 8.94102156e-01 -7.97666192e-01
3.77332807e-01 -7.53944933e-01 -3.48392129e-01 -1.94860503e-01
-5.57147451e-02 -7.10574687e-02 1.22962439e+00 -6.32587790e-01
1.13298333e+00 -2.08557105e+00 -1.33216351e-01 3.69353086e-01
-2.11326152e-01 4.16005403e-01 6.85117304e-01 6.23658359e-01
-1.39297470e-01 -1.04960811e-03 2.32977927e-01 4.48486179e-01
-6.24606669e-01 -1.47242740e-01 5.38499020e-02 6.28358662e-01
3.58034730e-01 -3.32540900e-01 -2.35524416e-01 -2.30913192e-01
1.40445530e-01 7.05761135e-01 -2.34335676e-01 -3.11763406e-01
1.85426444e-01 1.90237418e-01 -2.95221895e-01 3.96981955e-01
6.65343821e-01 3.90953660e-01 4.03310843e-02 -7.94924557e-01
-3.27700347e-01 8.46043602e-02 -1.72020996e+00 1.15759039e+00
-4.64116663e-01 1.01815367e+00 -5.21681495e-02 -8.83029163e-01
1.32573044e+00 7.75968909e-01 1.48969665e-01 -3.41952384e-01
7.36003160e-01 5.46011865e-01 4.01092619e-01 -4.05540735e-01
4.62425351e-01 -4.95721411e-04 3.45612943e-01 1.24119952e-01
-3.04658469e-02 -1.26188084e-01 3.75798047e-01 1.41074330e-01
1.22418499e+00 -2.96787284e-02 6.40511513e-01 -7.26915121e-01
9.24145877e-01 -3.13106142e-02 7.54474521e-01 -5.54569289e-02
-2.05161527e-01 -8.20853487e-02 1.19326459e-02 -3.71791095e-01
-1.30266047e+00 -7.75453925e-01 -3.98125380e-01 -1.16855681e-01
-7.65831023e-02 -1.50448814e-01 -1.08313464e-01 3.23154926e-01
-1.14447840e-01 1.83409438e-01 4.27446552e-02 -1.95629105e-01
-7.64440656e-01 -1.35081023e-01 4.69317883e-01 1.46153256e-01
7.46637523e-01 -6.17974520e-01 -1.41159594e+00 7.19807804e-01
9.15829599e-01 -7.67869413e-01 1.79209679e-01 2.72852629e-01
-1.56639075e+00 -8.12214434e-01 -8.84451270e-02 -9.06163633e-01
6.44245923e-01 3.14943224e-01 6.61362946e-01 7.41955817e-01
-1.02981782e+00 1.13372751e-01 -6.06750846e-01 -7.60598004e-01
-7.05741167e-01 -4.49885130e-01 1.43972784e-01 -5.71043491e-01
1.62165046e-01 -9.67065096e-01 -9.11519051e-01 -1.04909644e-01
-8.49259138e-01 -3.19209516e-01 5.96379340e-01 6.18052840e-01
-5.41467071e-02 9.67223823e-01 1.02921367e+00 -6.72167003e-01
5.64288020e-01 -2.70288020e-01 -9.11305487e-01 -1.24039002e-01
-4.30636972e-01 8.84616598e-02 1.22444320e+00 -4.73696254e-02
-8.29971373e-01 2.31232479e-01 -5.49010001e-02 3.32121819e-01
2.89712697e-01 1.20074980e-01 2.31899217e-01 -3.23919654e-01
9.51537907e-01 -3.31605047e-01 9.17942375e-02 -4.14848328e-01
-1.57095000e-01 7.59050548e-01 3.90928805e-01 -4.83807512e-02
5.48545301e-01 1.48605555e-01 7.54546463e-01 -1.33777773e+00
5.58162391e-01 -1.13625988e-01 -4.74595055e-02 -4.16010827e-01
4.44425702e-01 -8.02867413e-01 -9.92482901e-01 2.14831278e-01
-1.39459181e+00 6.98393166e-01 3.74948978e-01 7.40888178e-01
1.28749207e-01 2.44729042e-01 -6.26816213e-01 -1.32848513e+00
-9.62810934e-01 -1.30071795e+00 1.82802439e-01 8.63408804e-01
-3.84780914e-01 -5.11453688e-01 -5.69563687e-01 -6.06707394e-01
4.03817624e-01 4.64943379e-01 9.57509577e-01 5.69147095e-02
-7.15651274e-01 -4.57486361e-01 4.35918450e-01 6.10952973e-01
5.30423701e-01 6.57579720e-01 -6.65444732e-01 -1.25250816e-01
3.64217371e-01 2.71650136e-01 5.59380889e-01 1.47397697e-01
4.97622371e-01 -1.57555550e-01 -1.99674249e-01 2.70897467e-02
2.21357751e+00 9.37557697e-01 7.71006763e-01 -8.98205563e-02
-4.72244099e-02 3.65404159e-01 7.12835252e-01 5.91048956e-01
-4.26769793e-01 8.51638019e-02 2.27738976e-01 2.00643405e-01
-1.53179407e-01 2.06183583e-01 3.96008492e-01 8.24824154e-01
-1.93974793e-01 -1.84948668e-01 -3.74608845e-01 3.00509512e-01
-1.22568011e+00 -9.13886309e-01 -8.47541034e-01 2.25594163e+00
7.11695611e-01 5.04120111e-01 -1.31991386e-01 1.33950591e+00
6.16617382e-01 -5.62983334e-01 8.36017728e-02 -8.31288993e-01
2.23428547e-01 1.07267284e+00 9.92801905e-01 4.73284215e-01
-6.04532361e-01 2.91913807e-01 4.98752928e+00 5.32352746e-01
-1.42300367e+00 -9.19932127e-03 3.94081324e-01 1.37200072e-01
1.79245938e-02 2.41117731e-01 -7.18085229e-01 7.01444507e-01
1.01104951e+00 -6.49269968e-02 4.14149202e-02 6.89979851e-01
6.67726338e-01 -8.10584068e-01 -9.68475044e-01 9.44856524e-01
-5.97762942e-01 -1.12580824e+00 -4.35004920e-01 -4.66048837e-01
6.66724622e-01 -8.48203719e-01 -2.01537117e-01 -4.45556641e-01
-2.90604413e-01 -8.68101895e-01 6.57863259e-01 5.59666932e-01
2.17401594e-01 -1.15169239e+00 8.75536084e-01 -3.17308605e-02
-1.34772861e+00 -2.40525126e-01 -6.41590714e-01 -3.59076053e-01
1.51210561e-01 1.02834678e+00 -9.55233037e-01 3.14576566e-01
5.52134216e-01 2.08140090e-01 -2.34836116e-01 1.12250137e+00
1.26997009e-01 6.32437170e-01 -4.90249783e-01 -3.93589973e-01
-1.46163881e-01 -2.52652675e-01 5.94780326e-01 9.70123053e-01
5.36365807e-01 2.57515937e-01 -3.41879427e-01 2.24825427e-01
1.93097591e-01 1.57999814e-01 -5.77408314e-01 3.75266373e-02
8.48293543e-01 1.12655985e+00 -1.01253331e+00 -2.61951149e-01
-8.03289339e-02 6.45800173e-01 -5.79804242e-01 -4.25218940e-01
-6.09287977e-01 -1.15440190e+00 5.33792257e-01 3.54751736e-01
-2.43946910e-01 -8.04183900e-01 -3.40884328e-01 -2.07063466e-01
-2.05122009e-01 -9.54971075e-01 -6.07339963e-02 -2.55299956e-01
-4.75380480e-01 6.69774354e-01 -2.93030798e-01 -1.27597046e+00
-6.43365979e-02 -6.72320068e-01 -5.77947080e-01 1.01947355e+00
-8.42486918e-01 -4.56618994e-01 -2.91729927e-01 2.81666279e-01
9.51741874e-01 -2.58367896e-01 5.84605038e-01 6.80533946e-01
-3.48898172e-01 1.19486205e-01 2.31404398e-02 -4.02896285e-01
7.39934504e-01 -7.29688764e-01 -7.53044114e-02 1.29885602e+00
-3.13506186e-01 7.98375189e-01 1.20731556e+00 -8.59084845e-01
-1.69815242e+00 -3.56603056e-01 4.38101500e-01 6.00876093e-01
2.09653184e-01 3.83482836e-02 -3.93628567e-01 3.99755418e-01
5.80756128e-01 -2.85214722e-01 3.68045300e-01 -7.19322383e-01
4.58509326e-01 -6.55856073e-01 -1.57998669e+00 3.90301913e-01
2.03767791e-01 1.36298582e-01 -2.45573089e-01 2.01137766e-01
4.43654716e-01 -2.19045326e-01 -7.25188971e-01 -2.16091797e-01
5.89850307e-01 -1.24777436e+00 5.90135992e-01 -2.83369441e-02
1.88263148e-01 -8.18131328e-01 -1.49595618e-01 -8.91301394e-01
-3.00832033e-01 -7.76313722e-01 7.04620898e-01 1.33249450e+00
3.84711564e-01 -7.22010374e-01 4.58708584e-01 -2.78012501e-03
-2.47575805e-01 -2.38857821e-01 -7.84174979e-01 -4.53631133e-01
-7.04216957e-01 2.05992967e-01 4.81719643e-01 4.53877777e-01
-1.13403998e-01 2.84751743e-01 -3.93320113e-01 3.42624635e-01
5.41921556e-01 -6.95750952e-01 2.08633438e-01 -1.53873074e+00
-6.51565731e-01 1.24534011e-01 -1.18513453e+00 -4.87995774e-01
-7.87049294e-01 -3.25201571e-01 -1.61530808e-01 -1.11217701e+00
-7.92826712e-01 -5.63561678e-01 2.04827026e-01 -4.76983190e-02
3.05991769e-01 2.37518877e-01 -1.15980908e-01 -1.14083908e-01
4.73054886e-01 -2.31181473e-01 6.95476532e-01 1.06360763e-01
-1.23953439e-01 1.05893359e-01 -1.09361419e-02 6.33259356e-01
8.58503222e-01 -4.74832833e-01 -8.17667484e-01 -7.98199251e-02
5.41768551e-01 1.26487792e-01 -1.35768875e-01 -1.63573933e+00
2.64218271e-01 1.90427288e-01 9.02690291e-01 -7.11050212e-01
1.67210117e-01 -1.55317521e+00 8.10880244e-01 1.35260499e+00
5.06366313e-01 7.72718132e-01 9.05788392e-02 1.75883785e-01
-4.47152257e-02 -9.73994493e-01 1.05518389e+00 -2.84598947e-01
-5.82370520e-01 -4.36570674e-01 -8.45736444e-01 -9.28566933e-01
1.27466512e+00 -6.22404158e-01 -3.01285952e-01 -1.32811129e-01
-5.20659208e-01 -1.78609297e-01 4.05825824e-02 -9.54190642e-02
5.58085144e-01 -8.79209757e-01 -5.44332683e-01 1.20479099e-01
-7.50680923e-01 -2.59528339e-01 3.46272737e-01 6.31604612e-01
-1.63495982e+00 4.65166718e-01 -9.67796683e-01 -2.83519089e-01
-1.67983603e+00 8.58920440e-02 3.42718124e-01 2.13797107e-01
-6.30028784e-01 7.58787096e-01 -9.81346011e-01 9.68773901e-01
-1.52387261e-01 -6.12839282e-01 -3.43861490e-01 -2.30464667e-01
6.12260580e-01 1.05869591e+00 4.66243923e-01 -4.75625023e-02
-3.67994249e-01 5.91563880e-01 5.56427129e-02 -2.17223451e-01
1.25650263e+00 1.05536222e-01 -5.14910161e-01 2.68454194e-01
8.57043862e-01 2.39131644e-01 -5.29832423e-01 5.55766463e-01
1.17071746e-02 -5.43264747e-01 2.12307319e-01 -5.57258248e-01
-7.77308166e-01 5.43774009e-01 1.14607847e+00 3.83282304e-01
1.38876009e+00 -1.25822997e+00 5.70569694e-01 2.33485907e-01
6.98482990e-01 -1.19119442e+00 -1.18694827e-01 1.76834717e-01
4.84829307e-01 -6.75651968e-01 6.39422238e-01 -5.13642251e-01
8.26606061e-04 1.93191743e+00 4.15690213e-01 -8.12493145e-01
9.48908985e-01 8.03673625e-01 -2.76710659e-01 1.25346452e-01
-7.14972138e-01 3.54443759e-01 -2.62572080e-01 6.93584681e-01
1.07711029e+00 -6.09277049e-03 -1.44619024e+00 5.43309152e-01
-2.91888148e-01 1.33271620e-01 1.21835828e+00 1.22997105e+00
-9.38276112e-01 -1.50995600e+00 -7.89816380e-01 6.52668536e-01
-9.13732052e-01 2.35072076e-01 1.06884435e-01 6.81673408e-01
4.56199855e-01 1.37554574e+00 1.63938671e-01 -4.93135452e-01
1.83042854e-01 -2.43267864e-01 8.37045848e-01 -1.68335527e-01
-1.35575795e+00 -2.92819999e-02 1.86332345e-01 -6.42326772e-02
-1.01760074e-01 -2.79970109e-01 -1.80744350e+00 -8.09842229e-01
-4.19702560e-01 3.03932548e-01 1.39627981e+00 5.91729045e-01
1.77573025e-01 1.08769727e+00 5.97623289e-01 -1.68769896e-01
-5.04947662e-01 -9.45561647e-01 -6.05429113e-01 -3.74301642e-01
2.32692823e-01 -4.55074668e-01 -1.36677511e-02 1.27790138e-01] | [7.922689437866211, 2.6324403285980225] |
cc2e8045-a95a-4b54-bf6f-1d7fa73ebc1d | speaker-clustering-in-textual-dialogue-with | null | null | https://openreview.net/forum?id=s-5K23UWff7 | https://openreview.net/pdf?id=s-5K23UWff7 | Speaker Clustering in Textual Dialogue with Utterance Correlation and Cross-corpus Dialogue Act Supervision | We propose a textual dialogue speaker clustering model, which groups the utterances of a multi-party dialogue without speaker annotations, so that the real speakers are identical inside each cluster. We find that, even without knowing the speakers, the interactions between utterances are still implied in the text. Such interactions suggest the correlations of the speakers. In this work, we model the semantic content of an utterance with a pre-trained language model, and the correlations of speakers with an utterance-level pairwise matrix. The semantic content representation can be further enhanced by additional cross-corpus supervised dialogue act modeling. The speaker labels are finally generated by spectral clustering. Experiment shows that our model outperforms the sequence classification baseline, and benefits from the set-specific dialogue act classification auxiliary task. We also discuss the detail of correlation modeling and step-wise training process. | ['Anonymous'] | 2022-01-16 | null | null | null | acl-arr-january-2022-1 | ['cross-corpus', 'dialogue-act-classification'] | ['computer-vision', 'natural-language-processing'] | [ 1.12971686e-01 6.62800252e-01 4.12881672e-02 -1.03501248e+00
-7.67450452e-01 -6.81216776e-01 9.92171526e-01 2.95324158e-02
-1.78418476e-02 3.96809578e-01 9.59682524e-01 1.00416675e-01
1.80697501e-01 -1.76403508e-01 -1.27268657e-01 -7.35337853e-01
-1.33425087e-01 1.11270213e+00 -1.11355633e-01 -5.34997344e-01
1.11911543e-01 -3.69320601e-01 -1.11145878e+00 7.78328240e-01
7.53813565e-01 6.08624518e-01 1.71949089e-01 8.57694507e-01
-6.91719234e-01 1.30081248e+00 -9.38749731e-01 -3.80501330e-01
7.18137994e-03 -9.72347379e-01 -1.26394939e+00 7.85350502e-01
-5.71715795e-02 -5.95256314e-03 -1.46855503e-01 8.93840492e-01
1.18544243e-01 4.58637387e-01 8.20168376e-01 -1.38158560e+00
-1.98621839e-01 1.34506404e+00 -3.41834366e-01 -4.81797546e-01
6.19339824e-01 -2.28394449e-01 1.34811032e+00 -6.93664670e-01
3.84938896e-01 1.82635117e+00 3.74460965e-01 7.97406137e-01
-1.25683427e+00 -4.84377950e-01 3.53913426e-01 1.93925723e-01
-1.11924517e+00 -6.11505449e-01 1.01991141e+00 -4.88749117e-01
8.35098624e-01 4.44914490e-01 2.96281010e-01 1.09778368e+00
-5.02461195e-01 9.39153671e-01 8.22831392e-01 -6.27364218e-01
2.24683642e-01 4.40652817e-01 5.14941871e-01 5.19778252e-01
-1.00265121e+00 -6.90455437e-01 -7.46167541e-01 -4.37669963e-01
2.13664785e-01 -2.12485805e-01 -3.22915204e-02 -2.82264024e-01
-9.98400331e-01 1.16411293e+00 -1.51524916e-01 5.05481303e-01
-2.57690586e-02 -3.06696981e-01 5.80519438e-01 4.30417717e-01
6.17970943e-01 1.76045388e-01 -6.65713966e-01 -1.15766183e-01
-6.90633357e-01 -5.41659109e-02 1.35232711e+00 1.18801975e+00
9.12427783e-01 -4.38774943e-01 6.93551153e-02 1.47970617e+00
7.20157683e-01 7.69925490e-02 6.00197911e-01 -1.25527346e+00
5.17309725e-01 7.00490296e-01 -1.36609256e-01 -7.43200123e-01
-4.25778300e-01 3.14060122e-01 -7.57852733e-01 -5.43598831e-01
4.78291601e-01 -4.68599528e-01 -2.57430673e-01 1.82012355e+00
3.86483431e-01 1.93873435e-01 4.92296964e-01 5.43981969e-01
6.99824333e-01 7.38177538e-01 7.21380934e-02 -6.18837476e-01
1.42779398e+00 -1.35141051e+00 -1.19968259e+00 -6.14857748e-02
9.05535042e-01 -6.80160642e-01 8.68669271e-01 1.28851563e-01
-9.34690833e-01 -5.01228094e-01 -7.47497261e-01 2.34584082e-02
-3.09658572e-02 1.13760186e-02 5.66948771e-01 6.69417620e-01
-9.80625927e-01 1.42918661e-01 -6.60394490e-01 -3.57416302e-01
-2.36160353e-01 3.28516722e-01 -1.51686996e-01 4.52565193e-01
-1.37330127e+00 7.23289371e-01 2.36875460e-01 -1.43979385e-01
-6.16083443e-01 -3.22696686e-01 -9.25035059e-01 -2.26753931e-02
2.86096185e-01 -2.31855944e-01 1.64483786e+00 -1.11567461e+00
-2.03352070e+00 7.63431072e-01 -6.21471763e-01 -2.70494133e-01
3.26936066e-01 6.48935810e-02 -2.50641108e-01 1.10538898e-03
2.56814845e-02 5.95353782e-01 5.63617766e-01 -1.55863273e+00
-7.80699372e-01 -3.63442063e-01 -6.64862692e-02 7.34764457e-01
-3.59228373e-01 3.76065254e-01 -7.07237959e-01 -1.04660824e-01
9.55009535e-02 -1.23284411e+00 -3.18628758e-01 -7.56255984e-01
-7.23413467e-01 -7.40297496e-01 8.02506804e-01 -6.97697639e-01
1.10718715e+00 -2.25010347e+00 2.53040582e-01 3.30996752e-01
1.14564829e-01 -3.80025715e-01 -7.59381726e-02 7.34302402e-01
-1.10225461e-01 -2.15143226e-02 -3.62061411e-01 -9.86067176e-01
1.58415809e-01 4.05052513e-01 -2.58755863e-01 2.86305070e-01
-6.99806362e-02 4.54191327e-01 -7.84518719e-01 -6.69563949e-01
1.46089211e-01 1.68079257e-01 -4.32229340e-01 5.81520796e-01
-4.20770288e-01 8.65296125e-01 -3.01709741e-01 -1.08175278e-01
4.14007336e-01 -1.39282584e-01 9.59499657e-01 -1.45870377e-03
1.05853230e-01 8.44354451e-01 -1.18940914e+00 1.82380581e+00
-4.16961491e-01 5.35811126e-01 4.09453988e-01 -1.06450033e+00
1.01832402e+00 7.31902957e-01 8.10879469e-01 -1.33858830e-01
-1.08294405e-01 -2.03195870e-01 4.24219757e-01 -3.79972517e-01
5.57538390e-01 -1.68618545e-01 -4.23090100e-01 8.42528999e-01
2.89305001e-01 -6.89661726e-02 8.80969390e-02 7.02819288e-01
7.26151228e-01 -4.01710510e-01 2.52232015e-01 -1.74328357e-01
7.86676586e-01 -1.94922805e-01 3.62076849e-01 4.25476342e-01
-1.51226118e-01 2.58329779e-01 6.72224760e-01 4.48187329e-02
-7.61384249e-01 -7.06795633e-01 -1.43117145e-01 1.65845025e+00
2.45314222e-02 -8.34137082e-01 -9.61326301e-01 -7.81639397e-01
-4.12519097e-01 8.85406196e-01 -5.33505678e-01 3.44355881e-01
-5.80385029e-01 -6.46840811e-01 5.58078945e-01 1.31856024e-01
3.97101611e-01 -7.24251151e-01 3.98521066e-01 1.67404592e-01
-7.83361018e-01 -1.21081102e+00 -7.51504898e-01 2.67071515e-01
-4.80692625e-01 -9.42188859e-01 -8.63534659e-02 -1.06794643e+00
5.85602641e-01 8.29612240e-02 8.84811878e-01 3.01414225e-02
3.29862982e-01 5.38822353e-01 -5.20738125e-01 -1.75642014e-01
-1.07327652e+00 -1.78916797e-01 2.47109979e-01 2.55501896e-01
6.72341287e-01 -3.33834738e-01 -8.29967931e-02 5.79566658e-01
-3.46388817e-01 9.07624811e-02 -1.06008269e-01 9.88712072e-01
-2.09965676e-01 1.87781364e-01 5.52364409e-01 -1.02585483e+00
7.97288716e-01 -4.25258577e-01 1.01357773e-01 2.65604913e-01
-3.64156328e-02 2.13913605e-01 5.00941336e-01 -3.97865206e-01
-1.55040658e+00 3.42585474e-01 -3.79443057e-02 8.46082941e-02
-4.61436272e-01 3.16483587e-01 -5.69382846e-01 6.90164387e-01
4.69527006e-01 9.99361929e-03 2.63383389e-01 -4.16802019e-01
7.32660890e-01 1.11309600e+00 2.34289840e-01 -6.58347309e-01
4.49735343e-01 2.46894240e-01 -6.39592648e-01 -1.22003973e+00
-8.84326160e-01 -9.88236785e-01 -1.04561591e+00 -3.30218703e-01
9.77173030e-01 -1.07101309e+00 -9.56946135e-01 1.17397167e-01
-1.51738882e+00 -4.58790421e-01 1.64653528e-02 4.94129777e-01
-5.32880962e-01 7.12040544e-01 -9.48209107e-01 -1.34791338e+00
1.86686844e-01 -1.06504345e+00 9.08748031e-01 -1.55692592e-01
-8.82160783e-01 -1.46217847e+00 1.99327976e-01 9.11693990e-01
-2.06833959e-01 -5.00488400e-01 7.76332736e-01 -1.51545012e+00
-1.10295422e-01 2.55018979e-01 3.43737602e-01 4.49122131e-01
5.42595565e-01 -4.35251296e-02 -1.22937357e+00 1.78125843e-01
5.68169355e-01 -2.50049353e-01 4.89530116e-01 2.61364490e-01
6.62405193e-01 -6.58048153e-01 -3.75502974e-01 -1.05625749e-01
4.51438755e-01 5.65330327e-01 2.41056576e-01 -2.74519712e-01
7.93309391e-01 1.61297226e+00 3.61031771e-01 3.77155244e-01
7.95144498e-01 6.28863454e-01 -1.10356562e-01 -9.77826416e-02
3.34628880e-01 -1.93929836e-01 6.95999980e-01 1.53089583e+00
1.78503290e-01 -2.65843391e-01 -8.44630480e-01 3.56210351e-01
-2.18357992e+00 -1.26204598e+00 -4.71385181e-01 1.78869557e+00
1.20090985e+00 -2.87824512e-01 4.47079062e-01 -7.17810690e-02
9.21246707e-01 1.65762424e-01 -1.53253078e-01 -3.42489272e-01
-1.93546876e-01 -3.64376217e-01 1.40113831e-02 1.15203393e+00
-1.09613168e+00 1.01107121e+00 6.65548229e+00 5.17680407e-01
-3.40976566e-01 2.23331913e-01 7.41534293e-01 1.36155114e-01
-1.87975109e-01 2.31958866e-01 -8.02094221e-01 4.12558675e-01
1.21923220e+00 -1.31472915e-01 3.55919510e-01 6.77851915e-01
3.23122948e-01 3.32123542e-04 -1.59384763e+00 7.56269038e-01
4.28071946e-01 -8.60554218e-01 -2.26149410e-01 7.18785524e-02
6.90844655e-01 -2.51628965e-01 -3.37963879e-01 4.47796881e-01
9.52694416e-01 -8.45307291e-01 3.11264485e-01 1.31851614e-01
8.20549147e-04 -6.01229846e-01 5.81165910e-01 7.59906590e-01
-1.05240846e+00 5.47127202e-02 -9.29495916e-02 -8.32702667e-02
3.08483869e-01 1.39213696e-01 -1.42269170e+00 5.04146695e-01
3.88487488e-01 6.47634685e-01 -1.62602320e-01 8.71857977e-04
-9.97160822e-02 8.35862100e-01 -2.11001277e-01 -1.62824407e-01
1.69758826e-01 -7.06433475e-01 2.86504030e-01 1.62639642e+00
-1.56297088e-01 2.71915615e-01 5.92320919e-01 6.67112887e-01
-6.23476878e-02 3.42734188e-01 -3.38308275e-01 9.72956717e-02
5.07422090e-01 1.26086152e+00 -4.10474747e-01 -6.14471912e-01
-6.66699529e-01 1.17915916e+00 2.14991421e-01 3.20032239e-01
-5.66370606e-01 2.49338984e-01 7.90791988e-01 -3.71329397e-01
-1.29898712e-01 -1.78559259e-01 -2.50024796e-01 -1.06197453e+00
-2.37879232e-01 -9.71003354e-01 2.94634074e-01 -1.63051531e-01
-1.71864259e+00 3.21604818e-01 -7.68036023e-02 -9.96770442e-01
-7.64706790e-01 -3.61560464e-01 -7.08893180e-01 8.33801448e-01
-6.78049803e-01 -8.61890495e-01 1.81642532e-01 7.81277835e-01
1.07671154e+00 -4.22160953e-01 1.19952095e+00 -1.02662347e-01
-5.23285985e-01 4.14884627e-01 2.83628609e-02 6.60480380e-01
9.90229785e-01 -1.48859143e+00 8.86357352e-02 1.62677646e-01
2.51564622e-01 9.32942152e-01 8.11314702e-01 -5.92518330e-01
-9.33757007e-01 -8.31272423e-01 1.15772247e+00 -6.32993042e-01
9.69714344e-01 -8.59560132e-01 -1.05473709e+00 9.01599288e-01
1.01314533e+00 -9.26960707e-01 1.53952503e+00 7.55624652e-01
-3.56748223e-01 3.44074845e-01 -7.82437444e-01 6.43044829e-01
7.93074250e-01 -9.24743295e-01 -9.70963180e-01 7.23660052e-01
8.06256413e-01 -5.93644567e-02 -9.16304648e-01 -2.71144331e-01
1.60608515e-01 -8.11111748e-01 7.28563726e-01 -5.81565559e-01
3.13550472e-01 -1.35265097e-01 -3.12837362e-01 -1.58934391e+00
-5.59221916e-02 -8.48298490e-01 2.27586329e-01 1.65102947e+00
8.14964771e-01 -2.09172979e-01 6.83722138e-01 1.20120656e+00
-2.37843528e-01 2.04190999e-01 -7.97271609e-01 -6.02148592e-01
1.01394631e-01 -4.89299595e-01 3.52081716e-01 1.54021728e+00
1.23254728e+00 1.28475046e+00 -6.53260350e-01 1.76475868e-01
5.16032398e-01 -1.36675954e-01 9.14634824e-01 -1.26083004e+00
-5.73917449e-01 -2.71936893e-01 -6.85712993e-02 -1.54845452e+00
7.69557893e-01 -7.98011303e-01 5.95633984e-01 -1.22344828e+00
3.99084628e-01 -3.24289471e-01 3.65199775e-01 3.05426002e-01
-4.36544806e-01 -5.63320637e-01 8.54767933e-02 4.67101812e-01
-8.48602116e-01 8.09248984e-01 7.93580115e-01 -2.63739288e-01
-5.91283619e-01 7.01450780e-02 -6.07533574e-01 8.16900790e-01
9.19856131e-01 -3.23479295e-01 -5.62161982e-01 3.25300507e-02
-4.37430054e-01 4.16105062e-01 -2.59297669e-01 -3.60418320e-01
6.23397291e-01 -1.56243786e-01 -1.02067098e-01 -5.78396857e-01
7.79266715e-01 -7.51061499e-01 -2.09085029e-02 9.88192484e-02
-1.12170696e+00 -4.57177043e-01 -2.24957228e-01 5.61962605e-01
-5.48295796e-01 -3.08969378e-01 4.77384090e-01 -1.55225575e-01
-2.68496156e-01 -1.73448786e-01 -8.26018393e-01 -8.46902505e-02
8.66966903e-01 1.28573239e-01 -1.35036722e-01 -1.02330351e+00
-1.02042997e+00 5.82631052e-01 5.87411784e-02 3.37100148e-01
9.67388749e-02 -1.11786127e+00 -8.11840773e-01 -1.16002418e-01
7.24931061e-02 -2.10618321e-02 3.97512883e-01 5.94224155e-01
3.26120615e-01 3.94805729e-01 3.59165251e-01 -7.92322636e-01
-1.74461567e+00 3.78066003e-01 2.14092299e-01 -2.20504940e-01
-2.29686141e-01 7.12555647e-01 5.87724209e-01 -1.07353795e+00
5.01255751e-01 1.01789981e-01 -5.10228992e-01 3.67432177e-01
5.15164912e-01 7.83559382e-02 -4.45478141e-01 -1.00815487e+00
-2.25642741e-01 1.68091491e-01 -2.73309469e-01 -6.60050750e-01
1.12491262e+00 -6.49753928e-01 -3.19439083e-01 1.08309901e+00
1.26559997e+00 2.84495324e-01 -1.14088190e+00 -5.76662183e-01
4.18662459e-01 7.15227202e-02 -5.76907933e-01 -4.74525005e-01
-5.68307459e-01 7.17912912e-01 -5.02462573e-02 6.40688479e-01
5.42546868e-01 4.93476421e-01 2.41035968e-01 5.69511235e-01
-2.22313348e-02 -1.43062150e+00 4.25225228e-01 9.57903385e-01
8.23803067e-01 -1.47819996e+00 -3.90075356e-01 -9.35352802e-01
-1.44169879e+00 9.45908725e-01 5.72285116e-01 2.94018835e-01
7.41063297e-01 2.99129903e-01 5.11518121e-01 -1.39272943e-01
-1.06978309e+00 -1.34958744e-01 1.96007621e-02 5.99802434e-01
1.00966632e+00 3.56468946e-01 -4.32236232e-02 6.22243702e-01
-6.66489840e-01 -9.54749763e-01 4.38625753e-01 4.15100187e-01
-4.49593842e-01 -1.45169997e+00 -4.00585681e-01 8.73116180e-02
-5.06833307e-02 -1.96795702e-01 -1.05721104e+00 3.07583004e-01
-2.39278972e-01 1.77253187e+00 3.54856342e-01 -5.52223861e-01
7.56461844e-02 7.48516560e-01 -1.07412830e-01 -8.94933879e-01
-9.06686604e-01 5.34689307e-01 6.51944637e-01 -1.50514126e-01
-8.48715603e-01 -9.98815715e-01 -1.57493317e+00 -1.54774010e-01
-4.33660358e-01 7.91524291e-01 7.35746384e-01 1.12548351e+00
-5.42446859e-02 3.78280938e-01 1.17964590e+00 -4.97754931e-01
-5.05793810e-01 -1.50561929e+00 -6.34911954e-01 7.41198599e-01
1.11627743e-01 -1.30340725e-01 -6.39395833e-01 5.52836239e-01] | [12.675891876220703, 7.739925861358643] |
a57216ac-fe99-4324-917e-51d7fc836712 | painterly-image-harmonization-in-dual-domains | 2212.08846 | null | https://arxiv.org/abs/2212.08846v4 | https://arxiv.org/pdf/2212.08846v4.pdf | Painterly Image Harmonization in Dual Domains | Image harmonization aims to produce visually harmonious composite images by adjusting the foreground appearance to be compatible with the background. When the composite image has photographic foreground and painterly background, the task is called painterly image harmonization. There are only few works on this task, which are either time-consuming or weak in generating well-harmonized results. In this work, we propose a novel painterly harmonization network consisting of a dual-domain generator and a dual-domain discriminator, which harmonizes the composite image in both spatial domain and frequency domain. The dual-domain generator performs harmonization by using AdaIN modules in the spatial domain and our proposed ResFFT modules in the frequency domain. The dual-domain discriminator attempts to distinguish the inharmonious patches based on the spatial feature and frequency feature of each patch, which can enhance the ability of generator in an adversarial manner. Extensive experiments on the benchmark dataset show the effectiveness of our method. Our code and model are available at https://github.com/bcmi/PHDNet-Painterly-Image-Harmonization. | ['Li Niu', 'Yan Hong', 'Junyan Cao'] | 2022-12-17 | null | null | null | null | ['image-harmonization'] | ['computer-vision'] | [ 2.33476818e-01 -2.33144000e-01 9.53015238e-02 8.24377015e-02
-5.61995804e-01 -5.52959681e-01 4.37954575e-01 -4.90584135e-01
-1.58661306e-02 6.45530462e-01 4.55090255e-02 5.00101708e-02
2.42077604e-01 -1.01838267e+00 -7.58665025e-01 -9.59375560e-01
4.41260725e-01 -1.33492783e-01 1.70164630e-01 -4.40595657e-01
1.35933489e-01 2.05147937e-01 -1.36738110e+00 3.33704025e-01
1.16302907e+00 8.62455547e-01 1.52076364e-01 8.36237669e-01
1.41390786e-01 5.02418280e-01 -9.06872451e-01 -4.50202733e-01
6.83532178e-01 -1.05940354e+00 -4.35377181e-01 -5.31797670e-02
5.00067592e-01 -3.02096363e-02 -2.75697708e-01 1.39711452e+00
8.13452959e-01 7.87058324e-02 5.30244470e-01 -1.41898298e+00
-1.08165574e+00 3.76826018e-01 -1.11432552e+00 3.64961661e-02
1.68789342e-01 3.61097932e-01 5.73579311e-01 -5.52386045e-01
5.00071406e-01 1.42224216e+00 5.15150845e-01 5.90425551e-01
-1.22645617e+00 -1.03236675e+00 -1.52446687e-01 2.51235545e-01
-1.41316092e+00 -8.65203887e-02 1.29368198e+00 -2.49325007e-01
5.82639202e-02 5.68676531e-01 8.28354776e-01 9.20916677e-01
4.62268293e-01 5.64003766e-01 1.36552083e+00 -4.42042977e-01
-5.30602261e-02 -1.93119457e-03 -6.53382540e-01 4.52900171e-01
3.10624726e-02 3.09868246e-01 -3.11563402e-01 4.03286740e-02
1.04673386e+00 -1.55567989e-01 -4.16228652e-01 -2.22347125e-01
-9.79622662e-01 7.40674615e-01 7.94434011e-01 4.55677986e-01
-2.25358665e-01 1.07651643e-01 1.21886201e-01 2.03125596e-01
5.09794176e-01 5.73713243e-01 2.61418641e-01 2.94841230e-01
-9.03365433e-01 3.58836442e-01 4.44872737e-01 6.33172750e-01
7.59225130e-01 2.13854164e-01 -3.82963061e-01 1.05788076e+00
-9.63765457e-02 6.58229053e-01 5.03293335e-01 -9.30610240e-01
1.98893771e-01 5.36483228e-01 -1.94347892e-02 -1.28792393e+00
-1.00500107e-01 -3.81880492e-01 -1.27729011e+00 7.88235068e-01
6.02012612e-02 -2.98054934e-01 -8.12420666e-01 1.87245703e+00
5.49875438e-01 4.09056574e-01 -4.41315733e-02 1.23048162e+00
9.39726114e-01 9.39762115e-01 -2.83981836e-03 -5.61803691e-02
1.34150064e+00 -1.18069255e+00 -7.78943300e-01 1.58301368e-02
-3.44537407e-01 -1.42227089e+00 1.18044758e+00 1.44138709e-01
-1.39644516e+00 -1.04871023e+00 -1.21923411e+00 -6.97769523e-02
-1.64019912e-01 -2.57090051e-02 2.16825083e-02 5.66087902e-01
-9.95513201e-01 2.61505485e-01 -9.07615274e-02 2.42641047e-02
3.39852959e-01 -5.20024914e-03 -3.22914094e-01 3.67268682e-01
-1.31404376e+00 7.54289150e-01 3.77943069e-01 -5.48250601e-02
-5.36949337e-01 -7.66437352e-01 -7.84095824e-01 -1.51189610e-01
-1.40461409e-02 -9.28762436e-01 9.40523922e-01 -1.49465227e+00
-1.59845102e+00 9.92783010e-01 2.10045815e-01 -2.40328446e-01
7.54868984e-01 2.62811482e-01 -5.37336469e-01 1.92834124e-01
1.76270232e-01 8.77563477e-01 1.20792258e+00 -1.68407619e+00
-6.83380902e-01 -6.33706003e-02 -2.57820487e-01 3.80809575e-01
-1.43475473e-01 -3.41205060e-01 -5.74431121e-01 -1.37032020e+00
-3.57056409e-01 -8.91385496e-01 -7.13198408e-02 8.47811699e-02
-6.93363070e-01 2.52317280e-01 1.10401654e+00 -8.34948778e-01
1.23734260e+00 -2.38486576e+00 1.49630919e-01 2.31432617e-01
1.88408092e-01 3.82333845e-01 -3.88460606e-01 2.35511482e-01
-3.34480554e-01 9.97169092e-02 -3.74785662e-01 2.70212647e-02
-1.49572968e-01 -9.96625349e-02 -4.15114909e-01 4.16430295e-01
3.25544715e-01 9.81404006e-01 -7.41668582e-01 -6.29540563e-01
3.33156705e-01 5.53913891e-01 -4.62853730e-01 4.91748869e-01
2.34277099e-02 6.35996640e-01 -1.24316134e-01 5.92605174e-01
1.12152314e+00 1.84501588e-01 -1.18164010e-01 -5.29647708e-01
-9.95637104e-02 -5.08506179e-01 -1.14824569e+00 1.43499839e+00
-4.78995353e-01 5.53566158e-01 1.18863486e-01 -7.32828140e-01
1.18666029e+00 9.75520164e-03 3.00758898e-01 -1.17516541e+00
2.46796042e-01 1.68946177e-01 -3.39415967e-02 -2.90002823e-01
5.05957007e-01 -3.53111327e-01 -1.21671051e-01 3.30097467e-01
-3.02955866e-01 -5.66161335e-01 5.83838597e-02 -1.43165320e-01
6.03427887e-01 3.23497504e-02 2.37775683e-01 -2.46395648e-01
6.17348433e-01 -4.66561951e-02 5.45853913e-01 4.35994565e-01
-1.22079834e-01 1.02680981e+00 3.09457719e-01 -4.13161337e-01
-1.19437170e+00 -1.26925051e+00 2.50438619e-02 8.65833402e-01
8.87946010e-01 4.17702049e-02 -9.79047239e-01 -2.22872689e-01
-1.13358200e-01 4.02005076e-01 -7.75585592e-01 -5.38110137e-01
-6.25966549e-01 -5.00964940e-01 5.61264634e-01 1.70791462e-01
1.18337417e+00 -1.17103970e+00 -5.74493170e-01 -6.32027611e-02
-3.11907679e-01 -6.79460287e-01 -9.95839834e-01 -3.46452832e-01
-2.48398677e-01 -1.05072498e+00 -1.19192922e+00 -1.10341239e+00
6.87476277e-01 2.27029651e-01 1.01327372e+00 8.78111422e-02
-6.36537611e-01 -1.39341867e-02 -2.23602235e-01 -4.69712019e-01
-4.96965468e-01 -2.71742523e-01 -4.28372413e-01 2.65434712e-01
-2.96101809e-01 -6.55435085e-01 -1.08715105e+00 5.60706079e-01
-1.30997014e+00 4.66749579e-01 5.84769309e-01 9.96991754e-01
6.43730879e-01 3.21286440e-01 3.24529439e-01 -6.71660721e-01
9.41224396e-01 -2.44613677e-01 -4.24013793e-01 2.80462027e-01
-1.58229157e-01 -2.57786572e-01 7.43328989e-01 -7.17985630e-01
-1.18606973e+00 -1.05408490e-01 4.86899912e-02 -5.85789263e-01
1.80287585e-01 -5.67531288e-02 -3.16572636e-01 -3.17930728e-01
7.90778458e-01 3.13116699e-01 1.96072385e-01 -1.42486900e-01
5.95923543e-01 3.70849729e-01 1.12518990e+00 -5.02292275e-01
1.19345391e+00 5.18527627e-01 -4.44436222e-02 -3.35347921e-01
-3.12807918e-01 7.51786083e-02 -9.60059464e-02 -4.89346534e-01
9.71745431e-01 -8.16907823e-01 -5.87844789e-01 7.63698637e-01
-1.02693152e+00 -5.41658819e-01 -4.89045173e-01 -3.25753652e-02
-4.79249030e-01 2.72737890e-01 -4.85658318e-01 -4.42119330e-01
-6.07583344e-01 -1.00457096e+00 1.05433381e+00 8.48694742e-01
-1.16175458e-01 -7.11265147e-01 3.19766313e-01 1.63523629e-01
5.33225477e-01 7.98445880e-01 8.22317481e-01 2.19757184e-01
-3.05107504e-01 1.51207112e-02 -2.10281029e-01 4.73479360e-01
3.06914002e-01 2.61162877e-01 -8.53509665e-01 -1.41137555e-01
4.48219217e-02 -1.48371205e-01 7.76468098e-01 4.51599479e-01
1.14037037e+00 -5.39806187e-01 5.33495210e-02 8.72654974e-01
1.49845433e+00 5.36833405e-01 9.83536959e-01 4.62255895e-01
5.81844568e-01 5.69578230e-01 6.59357131e-01 3.77362043e-01
6.11695535e-02 9.18969452e-01 3.44759762e-01 -9.34245169e-01
-6.93596125e-01 -3.85448843e-01 8.05961117e-02 3.22843969e-01
-5.91670536e-03 -2.83715427e-01 -5.95276654e-01 3.88278067e-01
-1.73454630e+00 -1.21920097e+00 2.06486806e-01 1.74083626e+00
1.18191838e+00 -2.43227765e-01 2.37681285e-01 2.72661094e-02
1.26774776e+00 2.51971662e-01 -6.10608995e-01 -4.90490496e-01
-5.43823600e-01 3.73060614e-01 2.85682559e-01 5.49449503e-01
-1.08480799e+00 7.28167057e-01 5.47046232e+00 1.29226017e+00
-1.38931870e+00 3.83990742e-02 1.11119974e+00 -7.45765269e-02
-5.48500001e-01 -2.48520195e-01 -1.41431674e-01 8.93106639e-01
-1.66975224e-04 -4.60586905e-01 4.70681012e-01 6.44547582e-01
1.28701359e-01 -1.14086382e-01 -3.84009570e-01 1.26470494e+00
1.12248912e-01 -1.41787434e+00 1.81983054e-01 -4.14404243e-01
1.02923477e+00 -7.17938542e-01 6.46872938e-01 -5.89543879e-02
2.33685210e-01 -1.07879877e+00 9.66665983e-01 4.61868316e-01
1.04784167e+00 -9.41853642e-01 4.09178376e-01 -3.09206415e-02
-1.39490378e+00 8.85259360e-02 -1.40397206e-01 2.67679334e-01
8.12966079e-02 4.19401050e-01 -2.08597839e-01 6.10048771e-01
8.46894205e-01 5.17689764e-01 -5.10547161e-01 1.01931489e+00
-3.35677832e-01 1.43575206e-01 1.06149226e-01 3.33155215e-01
3.71220289e-03 -5.31046331e-01 6.40671968e-01 1.21515942e+00
3.21467429e-01 7.97408372e-02 5.61280586e-02 1.11558521e+00
-7.60606900e-02 8.33353177e-02 -5.07345021e-01 2.93506861e-01
5.02172589e-01 1.40100193e+00 -5.66723526e-01 -2.03068405e-01
9.34686735e-02 1.20337284e+00 -1.21541999e-01 3.27484637e-01
-1.22398400e+00 -7.22908795e-01 5.10636806e-01 1.25339394e-02
-2.97186896e-02 1.63819969e-01 -6.54154181e-01 -8.20444822e-01
1.00300133e-01 -1.16129899e+00 4.19713467e-01 -1.09279549e+00
-1.47436345e+00 7.80775785e-01 -1.95367217e-01 -1.48161900e+00
4.20466773e-02 -1.48846731e-01 -1.13398826e+00 1.09353805e+00
-1.26215017e+00 -1.38912857e+00 -8.08030546e-01 9.71743941e-01
3.01674515e-01 -1.51373714e-01 4.99953836e-01 3.57506335e-01
-5.55482388e-01 6.90243661e-01 -1.04370266e-01 4.42999862e-02
9.80618894e-01 -1.05099142e+00 2.28296533e-01 9.54086185e-01
-7.58264959e-02 1.91984802e-01 8.19656372e-01 -5.18996060e-01
-9.60756361e-01 -1.13486505e+00 2.36065403e-01 1.95282742e-01
2.77373552e-01 -2.41415694e-01 -8.15947354e-01 -7.71272779e-02
7.76408315e-01 -1.80112153e-01 4.68201309e-01 -6.99679852e-01
-4.54066962e-01 -4.33840930e-01 -1.29207206e+00 9.13620114e-01
7.64440000e-01 -3.87676358e-01 -3.69012445e-01 7.49187320e-02
6.27911985e-01 -5.90289176e-01 -7.06391454e-01 4.96986389e-01
5.56109786e-01 -1.14850771e+00 1.10778725e+00 8.09050575e-02
7.96797633e-01 -7.05802023e-01 1.58367053e-01 -1.42045224e+00
-5.39591730e-01 -8.15780520e-01 5.25662184e-01 1.33901036e+00
9.04586241e-02 -6.01141095e-01 3.74057323e-01 1.00181349e-01
4.95559685e-02 -6.58914387e-01 -8.05588424e-01 -7.82749593e-01
1.64351761e-02 2.55724281e-01 7.36284137e-01 8.72941315e-01
-3.27711642e-01 1.71594575e-01 -6.39524817e-01 -1.30743869e-02
4.39668775e-01 5.50576806e-01 1.00703669e+00 -4.72625405e-01
-4.57773834e-01 -6.93767309e-01 -2.76613414e-01 -4.01707560e-01
-8.17474909e-03 -5.54797411e-01 7.98993781e-02 -1.26484132e+00
3.20312530e-01 -3.39193285e-01 -1.28967434e-01 5.01991808e-01
-5.35497248e-01 8.81502509e-01 4.47433531e-01 2.25582615e-01
-2.03564823e-01 6.88311994e-01 1.67526352e+00 -4.12226498e-01
-2.87365735e-01 -3.68414789e-01 -9.55051959e-01 4.76887733e-01
1.13386798e+00 -1.30779713e-01 -4.31410521e-01 -3.78174484e-01
-2.61203021e-01 -1.56904429e-01 6.58343017e-01 -1.08312178e+00
1.04892269e-01 -2.81494856e-01 7.02014208e-01 -2.80003995e-01
3.22806448e-01 -6.19000673e-01 8.14466417e-01 7.89950132e-01
-1.83883250e-01 1.47102654e-01 3.83304983e-01 2.96324372e-01
-5.80015838e-01 3.25728178e-01 1.47620106e+00 2.81082280e-02
-6.64975524e-01 2.94710040e-01 1.41465038e-01 2.22355574e-01
1.46525729e+00 -2.77503550e-01 -6.17052794e-01 -6.17505014e-01
-1.50858030e-01 -9.05822217e-02 7.92642713e-01 5.56690454e-01
6.41679585e-01 -1.72079420e+00 -9.18750644e-01 3.36088777e-01
-1.79960579e-01 -1.97844669e-01 6.91869378e-01 4.69494522e-01
-8.54419768e-01 -4.11187291e-01 -9.47748661e-01 -3.89366925e-01
-1.36697567e+00 6.03737473e-01 5.03463209e-01 -8.91733468e-02
-4.38599885e-01 8.71121228e-01 7.55065024e-01 -1.01365056e-02
-4.34102602e-02 3.04067343e-01 -2.55941804e-02 -1.14402622e-01
4.87274766e-01 3.48848224e-01 -3.96033317e-01 -7.16216922e-01
-1.61868364e-01 1.01595771e+00 3.03262055e-01 -2.90109754e-01
9.69134390e-01 1.03940070e-02 -3.61849040e-01 -5.28360195e-02
1.12787890e+00 2.24515766e-01 -1.25665510e+00 7.90050253e-02
-7.30635166e-01 -7.49755383e-01 -3.32551956e-01 -7.26028204e-01
-1.48996472e+00 6.06018782e-01 8.03632736e-01 2.79190600e-01
1.90145612e+00 -2.61011660e-01 8.93885732e-01 -6.08926177e-01
-1.29754826e-01 -9.82982755e-01 4.36802626e-01 -7.63749108e-02
1.42349696e+00 -8.21107149e-01 -1.12706333e-01 -4.91426915e-01
-7.84960628e-01 8.45105231e-01 8.45018566e-01 -5.34188092e-01
5.04007936e-01 3.08793366e-01 4.37458932e-01 2.90906392e-02
-2.38697514e-01 7.48571241e-03 4.90395963e-01 7.22771227e-01
1.06022544e-01 1.89732864e-01 -5.37188172e-01 4.80982095e-01
-5.05168557e-01 -4.13954198e-01 1.35295197e-01 5.38588226e-01
-2.89475352e-01 -1.22293925e+00 -8.01538825e-01 -1.13088578e-01
-2.60451317e-01 -4.07926291e-02 -5.52251041e-01 8.19855511e-01
6.56790674e-01 1.02498209e+00 1.17823921e-01 -5.79892755e-01
4.05888587e-01 -5.67537248e-01 3.46402854e-01 -6.73927814e-02
-6.97129607e-01 2.62655854e-01 -3.79405051e-01 -5.33854842e-01
-3.81314397e-01 9.17011872e-02 -9.23592150e-01 -5.42840540e-01
1.43511742e-01 2.94323219e-03 3.24093729e-01 1.34649992e-01
3.06172967e-01 5.79387784e-01 1.06785738e+00 -9.44673955e-01
-9.30731073e-02 -5.92944264e-01 -6.59194589e-01 8.54216039e-01
1.89121187e-01 -4.44373310e-01 -5.07938452e-02 3.06902647e-01] | [11.243170738220215, -1.1750848293304443] |
25e930e6-bad8-4ea3-a3af-3a3291241af3 | autoqnn-an-end-to-end-framework-for | 2304.03782 | null | https://arxiv.org/abs/2304.03782v1 | https://arxiv.org/pdf/2304.03782v1.pdf | AutoQNN: An End-to-End Framework for Automatically Quantizing Neural Networks | Exploring the expected quantizing scheme with suitable mixed-precision policy is the key point to compress deep neural networks (DNNs) in high efficiency and accuracy. This exploration implies heavy workloads for domain experts, and an automatic compression method is needed. However, the huge search space of the automatic method introduces plenty of computing budgets that make the automatic process challenging to be applied in real scenarios. In this paper, we propose an end-to-end framework named AutoQNN, for automatically quantizing different layers utilizing different schemes and bitwidths without any human labor. AutoQNN can seek desirable quantizing schemes and mixed-precision policies for mainstream DNN models efficiently by involving three techniques: quantizing scheme search (QSS), quantizing precision learning (QPL), and quantized architecture generation (QAG). QSS introduces five quantizing schemes and defines three new schemes as a candidate set for scheme search, and then uses the differentiable neural architecture search (DNAS) algorithm to seek the layer- or model-desired scheme from the set. QPL is the first method to learn mixed-precision policies by reparameterizing the bitwidths of quantizing schemes, to the best of our knowledge. QPL optimizes both classification loss and precision loss of DNNs efficiently and obtains the relatively optimal mixed-precision model within limited model size and memory footprint. QAG is designed to convert arbitrary architectures into corresponding quantized ones without manual intervention, to facilitate end-to-end neural network quantization. We have implemented AutoQNN and integrated it into Keras. Extensive experiments demonstrate that AutoQNN can consistently outperform state-of-the-art quantization. | ['Tao Li', 'Chenkun Du', 'Deng Qian', 'Surong Dai', 'Ye Lu', 'Cheng Gong'] | 2023-04-07 | null | null | null | null | ['architecture-search'] | ['methodology'] | [ 1.56949788e-01 -2.74511546e-01 -6.87541068e-01 -4.56109852e-01
-9.03365552e-01 -3.56004387e-01 3.25867385e-01 -1.65939242e-01
-7.18529999e-01 6.85375571e-01 -2.36798838e-01 -6.37113929e-01
-1.58979788e-01 -8.71964157e-01 -9.83646631e-01 -6.93888664e-01
1.74912736e-01 5.27323604e-01 2.78555214e-01 -9.31401923e-02
2.34729111e-01 5.91551721e-01 -1.76088488e+00 3.01323295e-01
7.81135499e-01 1.59722698e+00 6.24175191e-01 6.44288480e-01
-4.49658960e-01 7.57043958e-01 -6.24817789e-01 -5.99899709e-01
5.91094136e-01 -8.79152268e-02 -7.08697140e-01 -3.33218306e-01
9.66479242e-01 -8.74783576e-01 -6.18025422e-01 1.37250006e+00
4.78357434e-01 -1.01587705e-01 5.22171855e-01 -1.16463876e+00
-5.81088305e-01 9.95432138e-01 -2.15309054e-01 4.67568934e-01
-6.21852219e-01 4.15255845e-01 1.05378294e+00 -4.69926506e-01
1.97958007e-01 1.42492914e+00 4.69689667e-01 8.54842663e-01
-1.31094790e+00 -1.26397598e+00 3.33209112e-02 2.01342389e-01
-1.46199989e+00 -7.50791013e-01 6.05337739e-01 -1.63094655e-01
1.17591417e+00 -1.11951940e-02 7.04997063e-01 7.70954728e-01
-7.08334241e-03 1.02249658e+00 7.04123616e-01 -2.39417911e-01
7.25888968e-01 -1.14394084e-01 2.18413305e-02 8.98645282e-01
2.14466706e-01 3.95378977e-01 -6.29689157e-01 -2.20836997e-02
9.79592741e-01 -8.64296257e-02 -1.71157524e-01 -1.66818097e-01
-7.70023525e-01 7.92141140e-01 5.58732808e-01 -7.22357929e-02
-2.51579374e-01 7.19337165e-01 5.34935832e-01 4.65013862e-01
-7.80916959e-02 5.60802400e-01 -7.15866625e-01 -4.50856000e-01
-1.39567268e+00 4.59777206e-01 4.94364142e-01 1.28487837e+00
1.00576758e+00 3.41100991e-01 -1.86425298e-01 8.15699995e-01
1.03857979e-01 4.35535759e-01 7.49679089e-01 -1.42056310e+00
7.86183715e-01 4.59723532e-01 -3.50893915e-01 -5.64945638e-01
-1.27098253e-02 -3.07964444e-01 -1.23467171e+00 -2.37548035e-02
1.80921391e-01 -2.36125290e-02 -9.40748155e-01 1.74505258e+00
4.39602956e-02 8.20794627e-02 3.94002460e-02 7.88311720e-01
4.95212078e-01 7.82461584e-01 1.11708283e-01 -1.52263746e-01
1.23250318e+00 -9.50377345e-01 -5.95389485e-01 -3.97373848e-02
8.50374639e-01 -2.74570614e-01 1.22976053e+00 7.13169098e-01
-1.40100861e+00 -6.03676975e-01 -1.35575187e+00 -3.24796766e-01
-1.61840320e-01 1.49207473e-01 6.09017253e-01 7.45092988e-01
-1.37541986e+00 9.28654253e-01 -1.12940526e+00 3.38646114e-01
9.32542861e-01 7.22239316e-01 2.78103292e-01 7.46117681e-02
-1.31141579e+00 5.06660759e-01 9.83364463e-01 -2.25140065e-01
-9.71018612e-01 -9.79789257e-01 -7.26680279e-01 4.12480295e-01
2.57355362e-01 -5.50155282e-01 1.72428489e+00 -8.63263845e-01
-1.70287454e+00 4.06585693e-01 1.41035393e-02 -1.04904723e+00
1.50727451e-01 5.97103685e-02 -1.38337970e-01 2.83190250e-01
-4.49118972e-01 1.24362159e+00 8.97617757e-01 -8.06569099e-01
-8.40295732e-01 -1.94747537e-01 -1.81104448e-02 -1.75985880e-02
-8.37868214e-01 -3.34903747e-01 -5.43467581e-01 -6.15778208e-01
4.59115021e-02 -5.09234071e-01 -6.57338053e-02 3.65161091e-01
-3.76701690e-02 -5.05243361e-01 5.59994876e-01 -3.36836785e-01
1.58446980e+00 -2.06104422e+00 -5.32624498e-02 8.81504342e-02
3.53567868e-01 6.97500646e-01 -2.48161316e-01 -1.75981864e-01
4.47888255e-01 3.07516783e-01 1.41718276e-02 -6.20814145e-01
3.57562512e-01 5.76465070e-01 -4.97501403e-01 6.00919835e-02
2.89849907e-01 8.58959734e-01 -9.36717033e-01 -7.31035769e-01
5.48932888e-02 3.12187582e-01 -1.06657720e+00 1.94312349e-01
-6.67815864e-01 -3.45844358e-01 -2.34015554e-01 8.06646645e-01
6.54490948e-01 -3.34811360e-01 1.11541115e-01 -5.75941384e-01
4.08225805e-02 6.83882475e-01 -9.96961057e-01 1.77756941e+00
-5.19367814e-01 4.81455058e-01 -6.37567136e-03 -9.60274160e-01
1.03911710e+00 6.74368069e-02 1.72313169e-01 -1.02383280e+00
2.02871203e-01 4.89103973e-01 -8.93392041e-02 -2.27358919e-02
6.63985193e-01 -1.07564516e-02 7.24971369e-02 1.58871099e-01
4.82794821e-01 -1.89434420e-02 3.80481154e-01 2.12739743e-02
8.69151652e-01 -2.81518370e-01 1.28435433e-01 -1.80948466e-01
2.04131678e-01 -2.85779953e-01 6.07282579e-01 7.52941251e-01
-4.79795605e-01 1.35007024e-01 5.09154260e-01 -4.69959468e-01
-1.28532529e+00 -9.76035655e-01 -2.32724562e-01 1.38403964e+00
-7.80854970e-02 -4.66237456e-01 -8.45613897e-01 -4.18208659e-01
-1.30849034e-01 5.50363958e-01 -1.81333125e-01 -3.34484428e-01
-6.07525647e-01 -2.46059120e-01 9.68547940e-01 6.17392778e-01
9.90605831e-01 -1.06729329e+00 -8.48527193e-01 1.96996436e-01
7.70568997e-02 -1.02417350e+00 -8.09260070e-01 5.99564433e-01
-1.20358658e+00 -4.09781247e-01 -5.04539728e-01 -8.39902639e-01
1.67642146e-01 -9.82441977e-02 1.15683281e+00 3.11646685e-02
8.47140253e-02 -4.01155472e-01 2.89130248e-02 -1.38094753e-01
-4.63551164e-01 6.00667000e-01 2.52223730e-01 -4.80378628e-01
3.05829912e-01 -6.46330237e-01 -6.64012730e-01 3.74822542e-02
-1.11597252e+00 8.90645832e-02 8.37732255e-01 8.91365349e-01
1.08030009e+00 4.35429007e-01 4.57765311e-01 -5.78679562e-01
5.43432176e-01 -5.07920608e-02 -1.16330266e+00 2.31075451e-01
-9.66102183e-01 7.23556519e-01 9.79581058e-01 -6.35585487e-01
-4.23342049e-01 4.49927002e-02 -2.79567331e-01 -1.16524923e+00
-1.93767752e-02 2.86273807e-01 -3.55520904e-01 -5.64510487e-02
7.59302735e-01 4.19239014e-01 -2.41152234e-02 -4.97589290e-01
4.61587906e-01 7.20686197e-01 7.35778928e-01 -7.25614667e-01
3.81795436e-01 -7.05857426e-02 -9.69984382e-02 -5.28344989e-01
-9.10019398e-01 -2.42956802e-01 -3.69124860e-01 3.15108955e-01
5.03628433e-01 -9.18183982e-01 -8.68034244e-01 3.53998572e-01
-1.14294672e+00 -7.64799416e-01 -2.89291650e-01 9.66245681e-02
-5.95836341e-01 4.62401882e-02 -7.47488499e-01 -6.77078068e-01
-8.48726869e-01 -1.68889463e+00 1.05618668e+00 2.37224653e-01
1.98982328e-01 -6.43933296e-01 -3.30746680e-01 1.35114744e-01
6.08327806e-01 -4.63996202e-01 1.04607666e+00 -3.64334643e-01
-8.74704659e-01 4.25829828e-01 -5.59269845e-01 6.39525115e-01
-1.63520068e-01 -2.00583324e-01 -8.79669189e-01 -3.94085914e-01
-1.99783280e-01 -8.26491833e-01 1.04705966e+00 3.05169970e-01
1.96881211e+00 -8.24419916e-01 6.82465360e-02 1.33619285e+00
1.50278580e+00 3.68395686e-01 5.74242949e-01 4.76714194e-01
6.17584050e-01 -2.60378659e-01 2.75350451e-01 4.31586772e-01
2.77364880e-01 5.65571547e-01 5.69767535e-01 3.85388941e-01
-1.61767796e-01 -4.14826214e-01 3.89164567e-01 8.73708248e-01
4.72893745e-01 -6.75445199e-02 -7.54878402e-01 3.73617738e-01
-1.43709981e+00 -8.21496189e-01 8.18316758e-01 1.95578361e+00
1.55711865e+00 5.44554055e-01 3.03672310e-02 3.43907237e-01
3.51111144e-01 7.36630931e-02 -9.70524907e-01 -7.27331460e-01
2.33112916e-01 4.91977125e-01 1.01998210e+00 2.61391491e-01
-1.03723717e+00 1.08938277e+00 5.40194750e+00 1.47150278e+00
-1.33674157e+00 -1.36961624e-01 8.66024733e-01 -3.15022677e-01
-2.11513773e-01 -3.90542984e-01 -1.54864216e+00 5.32697082e-01
1.43854356e+00 1.85770020e-01 1.08414435e+00 1.10445178e+00
-1.88889518e-01 4.95496273e-01 -1.28812325e+00 1.32182336e+00
-4.09108281e-01 -1.86965346e+00 5.87498128e-01 -1.83756964e-03
7.36476481e-01 2.31485665e-01 2.93002933e-01 6.61709547e-01
2.17384741e-01 -9.90986288e-01 1.00461721e+00 2.54667103e-01
1.39008129e+00 -1.06442916e+00 4.28333580e-01 4.42822754e-01
-1.18366539e+00 -3.35885048e-01 -8.70188892e-01 2.69054085e-01
-1.36845112e-01 2.18621954e-01 -7.91180193e-01 -2.41209328e-01
7.32456505e-01 4.13520485e-01 -4.50524420e-01 7.93504536e-01
1.52150899e-01 7.58230507e-01 -3.69326860e-01 -3.68873000e-01
5.51335275e-01 5.07963412e-02 -1.35832593e-01 1.04270470e+00
2.88731098e-01 1.79367140e-01 1.58150829e-02 1.04749966e+00
-5.23111761e-01 -2.28922352e-01 -1.79682262e-02 -3.70303690e-01
1.13662863e+00 9.18482661e-01 -3.95439088e-01 -4.31721240e-01
-6.64266795e-02 6.77436590e-01 7.08239496e-01 1.67905122e-01
-8.14375937e-01 -5.15192389e-01 7.20978141e-01 7.73427933e-02
6.67515516e-01 -1.17185831e-01 -4.17343050e-01 -8.53850484e-01
-1.96856841e-01 -1.18497133e+00 2.34663337e-01 -4.96940851e-01
-1.01093340e+00 6.52163804e-01 7.72447418e-03 -1.06096613e+00
-9.67788473e-02 -7.70550728e-01 -6.41947761e-02 7.80571938e-01
-1.82565236e+00 -6.27122998e-01 -1.76051557e-02 6.15229070e-01
6.60120428e-01 -5.05330682e-01 5.65376997e-01 4.27622050e-01
-5.35842240e-01 1.45080757e+00 1.53402865e-01 -3.42276949e-03
3.11529487e-01 -1.17383397e+00 4.90975380e-01 4.29194152e-01
5.83765842e-02 7.27504551e-01 3.89801919e-01 -2.12096125e-01
-1.49303651e+00 -1.54164326e+00 6.86000884e-01 3.47476870e-01
7.48672187e-01 -2.59460270e-01 -1.00406969e+00 2.32833087e-01
-1.66514695e-01 3.67048196e-02 4.54304755e-01 -2.93510616e-01
-4.33145374e-01 -6.18962288e-01 -9.98410940e-01 6.95155382e-01
1.02966094e+00 -4.86254543e-01 -1.35841161e-01 2.48962447e-01
1.29960597e+00 -5.83555758e-01 -9.13657308e-01 2.47313634e-01
5.21027982e-01 -8.48501146e-01 9.83601332e-01 -3.52897972e-01
5.83570361e-01 -6.11305237e-02 -6.26551151e-01 -8.95252466e-01
-1.78913742e-01 -6.08133614e-01 -7.82464027e-01 1.02679300e+00
3.12144607e-01 -3.66375148e-01 1.04677141e+00 4.77906615e-01
-2.79282242e-01 -1.24508977e+00 -1.02264714e+00 -8.97869766e-01
2.93422431e-01 -5.39602578e-01 1.18803048e+00 4.26775157e-01
-4.37627226e-01 1.18564472e-01 -7.05130175e-02 1.05222993e-01
7.04945803e-01 -4.50206064e-02 4.30914164e-01 -9.03992057e-01
-6.43945932e-01 -9.82596517e-01 -3.98015320e-01 -1.69124901e+00
1.96591750e-01 -9.25193310e-01 2.62943748e-02 -9.59378302e-01
-1.71703473e-01 -9.22106206e-01 -4.31796193e-01 5.57853878e-01
2.24079087e-01 -5.66007756e-03 2.48424560e-01 3.81046265e-01
-7.30148852e-01 7.94240475e-01 1.31008458e+00 -4.09309119e-01
-3.22606146e-01 -8.46274197e-02 -7.81255841e-01 4.80126411e-01
9.62702274e-01 -2.33757049e-01 -7.86794245e-01 -7.56741822e-01
1.95638984e-01 1.62364274e-01 1.89041883e-01 -1.28648400e+00
3.47500950e-01 -3.12635779e-01 2.46808141e-01 -7.90114880e-01
2.86975175e-01 -6.33232415e-01 -3.07142943e-01 4.52383012e-01
-7.41782486e-01 8.09037238e-02 3.70240062e-01 2.53642052e-01
1.80131402e-02 -3.74248445e-01 1.01058364e+00 1.41384755e-03
-7.27324247e-01 7.77848482e-01 5.24292402e-02 1.48955509e-01
4.00349647e-01 -1.70622826e-01 -3.98017824e-01 2.68364344e-02
-3.05393696e-01 2.99462289e-01 2.92896032e-01 -5.56523129e-02
7.88859427e-01 -1.51415718e+00 -2.08049119e-01 4.55261171e-01
-1.73877627e-01 5.19856930e-01 2.06550080e-02 2.25910544e-01
-8.18885386e-01 6.20130897e-01 -8.42079669e-02 -6.79571867e-01
-8.32770646e-01 6.57641411e-01 5.78700066e-01 -5.33026457e-01
-3.12544882e-01 1.15432501e+00 -1.05383225e-01 -2.54322469e-01
8.51198614e-01 -9.86938000e-01 1.41314641e-01 -3.55595022e-01
7.55641937e-01 3.04327667e-01 2.44668052e-01 -1.34235352e-01
6.99995235e-02 1.27870485e-01 -4.10789728e-01 -2.21583378e-02
1.02881801e+00 1.07637756e-01 1.56695113e-01 1.46328747e-01
1.55068195e+00 -9.91244078e-01 -1.64903092e+00 -4.94519085e-01
-2.13957563e-01 -9.36277509e-02 6.10034466e-01 -5.43036997e-01
-1.45288682e+00 1.12618995e+00 8.18031251e-01 -1.73078887e-02
1.49673951e+00 -2.82614619e-01 1.34209692e+00 8.60754967e-01
4.89693463e-01 -1.04899466e+00 2.88547873e-02 7.08830059e-01
6.15070939e-01 -8.68865430e-01 -1.06168807e-01 1.37833908e-01
-3.45705211e-01 1.06285334e+00 9.81341302e-01 -9.01794210e-02
7.31503725e-01 4.73428488e-01 -1.53976306e-01 6.81250989e-02
-1.21913791e+00 2.02758878e-01 2.01586887e-01 4.51181054e-01
3.58001031e-02 2.08882540e-02 9.18312222e-02 6.56892538e-01
-7.65010238e-01 1.12274587e-01 3.06406580e-02 7.45357037e-01
-7.35010207e-01 -9.03517663e-01 -4.60819453e-02 7.81796992e-01
-2.18096271e-01 -3.82818997e-01 3.26901704e-01 2.97503322e-01
3.94030064e-01 4.33987588e-01 3.48505139e-01 -5.83130062e-01
9.86387804e-02 -1.03717417e-01 5.39520621e-01 -2.58479834e-01
-4.05247629e-01 -2.77839694e-02 -3.78815114e-01 -6.79159343e-01
2.11154446e-02 -2.77939975e-01 -1.36849487e+00 -5.12549281e-01
-3.86714876e-01 -1.63917884e-01 5.58552086e-01 9.01366472e-01
3.21718007e-01 5.34904242e-01 4.85302001e-01 -8.61949742e-01
-1.29281414e+00 -7.13909268e-01 -4.99012828e-01 -6.18599989e-02
5.58259487e-01 -5.05787671e-01 -9.10718739e-02 1.01617083e-01] | [8.637969970703125, 3.0399389266967773] |
fa953f7a-5003-4ce5-a90a-cdf627b29ec2 | persistent-anti-muslim-bias-in-large-language | 2101.05783 | null | https://arxiv.org/abs/2101.05783v2 | https://arxiv.org/pdf/2101.05783v2.pdf | Persistent Anti-Muslim Bias in Large Language Models | It has been observed that large-scale language models capture undesirable societal biases, e.g. relating to race and gender; yet religious bias has been relatively unexplored. We demonstrate that GPT-3, a state-of-the-art contextual language model, captures persistent Muslim-violence bias. We probe GPT-3 in various ways, including prompt completion, analogical reasoning, and story generation, to understand this anti-Muslim bias, demonstrating that it appears consistently and creatively in different uses of the model and that it is severe even compared to biases about other religious groups. For instance, "Muslim" is analogized to "terrorist" in 23% of test cases, while "Jewish" is mapped to "money" in 5% of test cases. We quantify the positive distraction needed to overcome this bias with adversarial text prompts, and find that use of the most positive 6 adjectives reduces violent completions for "Muslims" from 66% to 20%, but which is still higher than for other religious groups. | ['James Zou', 'Maheen Farooqi', 'Abubakar Abid'] | 2021-01-14 | null | null | null | null | ['adversarial-text'] | ['adversarial'] | [-1.81965027e-02 6.63898230e-01 -3.78488809e-01 -2.69343048e-01
-5.69970787e-01 -8.09649348e-01 9.36422110e-01 5.01363099e-01
-3.01422119e-01 8.99235666e-01 1.22989690e+00 -3.66187304e-01
1.02825716e-01 -9.63729382e-01 -5.18036962e-01 -2.39182547e-01
3.97545904e-01 6.05389118e-01 -4.59326893e-01 -8.59310389e-01
8.73883486e-01 2.28890195e-01 -9.54678297e-01 5.13141036e-01
1.04831910e+00 5.28176762e-02 -7.79201150e-01 2.64628768e-01
-1.46331429e-01 1.35373425e+00 -9.34015751e-01 -9.37812686e-01
-2.01529697e-01 -4.16127652e-01 -9.95981991e-01 -4.21268523e-01
8.80587697e-01 -5.36881685e-01 -8.42369422e-02 9.49734151e-01
8.69033873e-01 -7.02434452e-03 9.66991365e-01 -8.20807874e-01
-1.57955611e+00 1.27463055e+00 -5.97596705e-01 1.26602277e-01
1.00350201e+00 1.97756112e-01 8.43803585e-01 -9.09681022e-01
8.57874155e-01 1.98948944e+00 7.21713543e-01 6.52043283e-01
-1.77980828e+00 -1.08061159e+00 4.97879647e-02 -2.66930729e-01
-1.36500096e+00 -2.96375722e-01 8.58781278e-01 -7.97353208e-01
7.24643469e-01 7.20866203e-01 1.04568315e+00 1.80725408e+00
4.72767085e-01 1.09701842e-01 1.58736646e+00 -3.06980535e-02
2.02696845e-01 4.52858716e-01 1.85813233e-01 4.98650700e-01
4.63286310e-01 1.35933414e-01 -8.62134218e-01 -5.16669989e-01
6.18706167e-01 -4.09712762e-01 2.27404409e-03 4.62229103e-01
-1.16335869e+00 1.42895734e+00 4.40589190e-01 2.40086511e-01
-1.55983463e-01 9.93885100e-02 7.45100752e-02 2.00265363e-01
8.06822717e-01 9.41156566e-01 7.03145862e-02 -1.80287242e-01
-8.70487630e-01 7.21646309e-01 8.27126801e-01 4.88501370e-01
5.42078197e-01 3.14016432e-01 -4.80124325e-01 1.06903028e+00
-5.49153276e-02 7.20909774e-01 1.46307290e-01 -1.04327333e+00
3.01510006e-01 6.14192247e-01 2.53644019e-01 -1.58717334e+00
-5.67749619e-01 -5.30941248e-01 -6.91384435e-01 2.41366331e-03
4.09977734e-01 -4.27643061e-01 -4.07910049e-01 2.01868749e+00
8.39427710e-02 -3.90776873e-01 -1.14223033e-01 9.67530429e-01
9.98272538e-01 4.41663235e-01 5.86791873e-01 2.24262234e-02
1.33590794e+00 -1.44458622e-01 -4.53189343e-01 -8.06860209e-01
7.63745904e-01 -9.23505485e-01 1.46853554e+00 1.64609998e-02
-1.32158482e+00 -9.31815356e-02 -6.88938200e-01 -2.67312020e-01
-3.40841949e-01 -5.69317162e-01 4.99170661e-01 9.96783316e-01
-1.06408179e+00 3.90151680e-01 1.95053935e-01 -5.70157290e-01
3.31186026e-01 -1.13331474e-01 -2.25637168e-01 -7.53564164e-02
-1.46314001e+00 1.23957920e+00 1.57568231e-02 -3.36829185e-01
-8.69792700e-01 -1.05490196e+00 -8.98934186e-01 -1.62581190e-01
2.13997155e-01 -8.27675045e-01 7.90923953e-01 -1.04056847e+00
-9.53349710e-01 1.14926100e+00 -1.39986081e-02 -1.46384016e-01
4.73867148e-01 -3.26415032e-01 -5.61783552e-01 -6.21735938e-02
2.65417904e-01 9.80595171e-01 6.18828595e-01 -1.24339485e+00
1.02418572e-01 -3.22230011e-01 4.83867615e-01 2.73751110e-01
-6.26271784e-01 6.91547573e-01 8.39117408e-01 -1.19195771e+00
-1.22957610e-01 -8.72456610e-01 -4.89485264e-02 -4.35523123e-01
-4.42769110e-01 -1.04006380e-01 2.53390193e-01 -8.19409013e-01
1.41261601e+00 -1.81239939e+00 5.13714626e-02 2.19389021e-01
3.05253714e-01 -3.41689169e-01 1.79037768e-02 6.91224992e-01
-2.47697622e-01 8.29417527e-01 4.58246358e-02 2.36554131e-01
1.92824110e-01 1.34271294e-01 -6.20076120e-01 4.66614753e-01
4.10886467e-01 8.68950188e-01 -9.46166933e-01 -6.78812027e-01
-1.55411050e-01 4.61974889e-01 -1.16182625e+00 -4.42296684e-01
1.85746014e-01 2.75999546e-01 2.69785285e-01 8.21691692e-01
6.06775165e-01 1.50302425e-01 1.36643931e-01 1.73329487e-01
-2.69201905e-01 5.67853451e-01 -7.19527960e-01 1.09453833e+00
-6.68344125e-02 7.17571378e-01 -1.09225631e-01 -1.81980893e-01
9.15036678e-01 1.36917487e-01 -3.52746904e-01 -7.65018821e-01
8.98555741e-02 1.27292261e-01 1.82343096e-01 -3.67173105e-01
9.59859192e-01 -8.31444860e-01 -5.49796402e-01 4.30712610e-01
-4.48581368e-01 -6.28374875e-01 7.89337978e-03 6.11313164e-01
6.30483091e-01 -3.88626903e-01 2.43409157e-01 -9.10955667e-01
1.11777127e-01 -1.95503719e-02 5.62012553e-01 8.98774624e-01
6.28931448e-02 3.37506890e-01 8.16264033e-01 -2.85804003e-01
-7.98315167e-01 -1.07154655e+00 -1.21882468e-01 1.46465921e+00
-1.79562569e-01 -2.68008560e-01 -6.97295547e-01 -4.05669540e-01
2.04253659e-01 1.75824225e+00 -1.04022038e+00 -4.98536408e-01
-4.36460912e-01 -6.12098873e-01 7.39017367e-01 3.86619270e-01
-5.17103681e-03 -8.90654624e-01 -5.03401577e-01 -4.71147709e-02
-4.12240624e-01 -6.48536742e-01 -3.13399196e-01 -3.58617097e-01
-5.86608410e-01 -6.76419020e-01 -4.74535197e-01 -3.24625790e-01
6.28648043e-01 -1.28096700e-01 1.44468737e+00 2.49692947e-01
1.06581107e-01 2.98431247e-01 -2.25456312e-01 -5.94628155e-01
-7.17039585e-01 -2.05344930e-01 -7.64514580e-02 -4.82052326e-01
3.62672687e-01 -4.58956212e-01 -4.41896707e-01 1.25444923e-02
-9.18337345e-01 1.03241868e-01 3.37931693e-01 5.53941190e-01
-3.70470703e-01 -6.56216800e-01 5.80979645e-01 -1.18924129e+00
9.71229315e-01 -7.86985457e-01 1.17258735e-01 -2.35056058e-01
-4.54435945e-01 -3.30811620e-01 2.87681192e-01 -8.01429987e-01
-8.68934751e-01 -1.14250541e+00 -1.30969509e-01 3.16814810e-01
-1.50552792e-02 5.63448370e-01 4.12532508e-01 -4.61378247e-02
1.21046937e+00 -1.38154492e-01 -2.64333755e-01 -4.24057469e-02
6.64694160e-02 4.04242218e-01 2.53643274e-01 -9.68714774e-01
9.43134964e-01 4.45962846e-01 -2.28070989e-01 -8.45195234e-01
-8.70254993e-01 3.12669784e-01 1.39297694e-01 -3.29509437e-01
8.26766312e-01 -1.07841778e+00 -6.99548602e-01 7.18469992e-02
-1.15535891e+00 -4.74899888e-01 -2.94352829e-01 4.56942677e-01
-1.74888641e-01 3.65813114e-02 -9.52493727e-01 -9.89626348e-01
-2.66145676e-01 -7.50022590e-01 7.99615204e-01 -6.89397231e-02
-1.32937634e+00 -1.16175222e+00 1.82068527e-01 5.28225005e-01
4.57968295e-01 5.24582922e-01 1.60969436e+00 -8.20877790e-01
2.57307649e-01 1.14880726e-01 -1.56746581e-01 -1.67757794e-01
-1.91874936e-01 9.19214040e-02 -6.20430827e-01 -1.31995663e-01
6.39802366e-02 -5.72741389e-01 3.31463009e-01 -4.08144966e-02
5.38488865e-01 -8.61422956e-01 8.31175521e-02 9.72209275e-02
1.20290613e+00 -6.10476024e-02 7.54926682e-01 3.19849342e-01
5.45659363e-01 8.94324720e-01 4.66503352e-01 5.71429849e-01
6.56577289e-01 4.29594785e-01 1.91567898e-01 -6.87602535e-02
1.10866584e-01 -5.31284809e-01 7.52815545e-01 4.93472517e-01
1.27979726e-01 1.80211544e-01 -1.20361543e+00 7.49697149e-01
-1.14996052e+00 -1.24751770e+00 -3.45393687e-01 1.87026393e+00
1.11623466e+00 1.74417228e-01 -1.05841104e-02 -4.59149294e-02
5.20596743e-01 4.88608420e-01 -9.78357568e-02 -9.89363968e-01
-4.61206466e-01 1.66057467e-01 1.25746667e-01 9.61030185e-01
-4.04691696e-01 1.05037749e+00 7.71281576e+00 5.90372384e-01
-9.94817674e-01 1.92820188e-02 8.80424559e-01 -4.83568847e-01
-1.21873510e+00 1.98045477e-01 -3.49850118e-01 5.15159190e-01
7.62586415e-01 -3.67427945e-01 5.48206270e-01 4.80661273e-01
3.26729000e-01 -1.96524709e-01 -1.17149079e+00 7.15809643e-01
4.68394935e-01 -1.25644422e+00 5.97634971e-01 -9.40421503e-03
8.97508800e-01 -5.64691305e-01 3.62351447e-01 5.59170187e-01
5.63596070e-01 -1.53046429e+00 1.43211412e+00 3.46101195e-01
7.73328066e-01 -9.10912633e-01 5.43971717e-01 3.73018652e-01
-2.32241392e-01 -2.29138076e-01 -3.66766006e-01 -8.13001931e-01
-4.90070246e-02 5.08085012e-01 -7.40099430e-01 -4.78343517e-02
6.62744701e-01 1.70600355e-01 -8.07996929e-01 -6.43714443e-02
-3.71014267e-01 7.18255997e-01 7.85024315e-02 -4.74521630e-02
2.73195654e-01 -2.26463884e-01 6.48739398e-01 1.51068723e+00
3.71896833e-01 3.48414570e-01 -6.40599877e-02 1.00489616e+00
-2.61475116e-01 3.10336024e-01 -9.44553554e-01 -1.98812470e-01
5.87525547e-01 9.59588766e-01 -3.66342723e-01 -4.79750454e-01
-1.25604182e-01 5.57756722e-01 3.71977776e-01 4.65374887e-01
-9.87451732e-01 -4.35555689e-02 7.48272121e-01 5.75978100e-01
-5.68149149e-01 6.92975521e-02 -8.76706064e-01 -8.88403952e-01
-5.92297852e-01 -1.37579858e+00 2.55239606e-01 -1.17152381e+00
-1.19181609e+00 -1.54955447e-01 1.49564818e-01 -3.85903746e-01
-2.92139083e-01 -2.45186955e-01 -5.69656491e-01 8.82949114e-01
-7.85801113e-01 -1.02652431e+00 -1.14053838e-01 5.41603684e-01
2.82663226e-01 1.94112547e-02 8.15651834e-01 -3.35092992e-02
-3.28330100e-01 7.31605351e-01 -6.96462154e-01 2.28821665e-01
1.17017925e+00 -1.05888629e+00 2.26255774e-01 6.32488847e-01
-3.40050817e-01 1.02807164e+00 1.22571421e+00 -1.06234562e+00
-1.00710464e+00 -7.97515154e-01 1.37762368e+00 -8.28585446e-01
9.27304387e-01 -3.05158079e-01 -8.29369962e-01 9.61006045e-01
5.99632323e-01 -8.61735582e-01 1.23093605e+00 4.18714166e-01
-9.45646465e-01 3.79940122e-01 -1.46963751e+00 1.43805027e+00
1.27200758e+00 -7.79876113e-01 -9.92353439e-01 4.17793423e-01
7.95486510e-01 -2.33990371e-01 -8.77888381e-01 -6.69116229e-02
4.37533915e-01 -1.12654912e+00 1.01616609e+00 -8.37753654e-01
1.11548698e+00 1.70865342e-01 -1.48289457e-01 -1.61029470e+00
-8.02171886e-01 -7.00592637e-01 3.16681653e-01 1.38691354e+00
4.35580671e-01 -1.01523888e+00 3.52436662e-01 6.62821770e-01
-8.83734673e-02 -3.54734123e-01 -8.52258563e-01 -2.00050145e-01
8.80975008e-01 -2.43152991e-01 5.21724761e-01 1.80598629e+00
4.14371818e-01 5.65673172e-01 -4.23979789e-01 -1.10237017e-01
4.30974692e-01 -2.10947260e-01 5.70640981e-01 -1.25673997e+00
-9.85597670e-02 -6.62899971e-01 4.12539952e-02 -1.90959305e-01
3.23597342e-01 -9.85529721e-01 -7.19422996e-01 -1.39334655e+00
5.62959611e-01 -4.38355990e-02 3.74822527e-01 3.15157205e-01
-3.68104964e-01 5.84971189e-01 3.17959487e-01 -2.29887739e-01
1.15755729e-01 2.98483104e-01 1.12205374e+00 -1.80878341e-01
7.41271898e-02 -1.02442491e+00 -1.86717713e+00 1.01955485e+00
7.86420941e-01 -6.34687841e-01 -5.52691743e-02 -5.88792145e-01
1.04428387e+00 -3.67691405e-02 7.35590398e-01 -6.46089256e-01
-1.86743751e-01 -7.97841907e-01 7.56470442e-01 -1.60090491e-01
4.08125401e-01 -4.71566916e-01 2.73912519e-01 6.16818249e-01
-6.44959509e-01 3.13366920e-01 2.57275373e-01 2.28438620e-02
1.44815251e-01 -6.91389218e-02 6.32819593e-01 -9.92394015e-02
7.33541995e-02 -5.53172350e-01 -6.99784517e-01 6.38356030e-01
8.32980871e-01 -1.67916939e-01 -9.02739763e-01 -7.58721352e-01
-2.47029185e-01 -6.10021539e-02 7.96106577e-01 2.81624168e-01
5.74300766e-01 -1.49586833e+00 -1.26909494e+00 -1.59786373e-01
1.90862507e-01 -7.04496324e-01 2.71188736e-01 9.21260178e-01
-4.61154342e-01 8.76916423e-02 -2.22259343e-01 3.69324796e-02
-1.07822669e+00 5.64497352e-01 -8.74718800e-02 5.14844321e-02
-1.01100765e-01 7.71316886e-01 3.23140144e-01 -2.91173428e-01
-3.55049461e-01 -1.16427131e-01 -1.91833884e-01 4.95379150e-01
4.18131232e-01 7.23130643e-01 -5.31882167e-01 -7.73875535e-01
-3.90787750e-01 5.35591185e-01 -6.75956383e-02 -3.50434899e-01
9.26997483e-01 1.98645502e-01 -5.70592105e-01 6.56515956e-01
6.67600572e-01 8.60484838e-01 -3.48546714e-01 3.01626831e-01
-1.20322838e-01 -6.67593598e-01 -3.11256170e-01 -1.09078920e+00
-2.86068738e-01 6.79103255e-01 -2.35336889e-02 2.87840277e-01
5.37762284e-01 -6.71584010e-02 2.52662092e-01 6.99509755e-02
3.46820861e-01 -1.41551638e+00 1.50029242e-01 7.44300902e-01
1.51580930e+00 -8.34170759e-01 1.74099371e-01 -3.57103080e-01
-8.63429606e-01 5.39279282e-01 6.43981099e-01 -2.46449366e-01
-2.29970645e-02 1.18604833e-02 1.21005498e-01 -2.48266757e-01
-7.74774909e-01 3.79444420e-01 -4.10220958e-02 5.79411268e-01
9.92374718e-01 4.47477430e-01 -7.92403042e-01 5.62720597e-01
-9.73129928e-01 -4.51633960e-01 9.24182773e-01 6.34909391e-01
-2.31566921e-01 -4.18095618e-01 -1.00107121e+00 2.30079502e-01
-7.45800793e-01 -6.45842493e-01 -1.17984509e+00 8.44957650e-01
5.67680262e-02 1.15678644e+00 2.79335260e-01 -4.25850809e-01
1.64582193e-01 2.69251525e-01 3.74018818e-01 -6.66192293e-01
-1.38512981e+00 -1.30288124e-01 7.09344745e-01 -3.60797733e-01
-7.50109479e-02 -4.51006114e-01 -9.38909829e-01 -1.16863465e+00
3.22314501e-01 -6.72617555e-02 1.41317695e-01 6.81975484e-01
1.23690844e-01 2.98798710e-01 2.05645412e-01 -6.76881254e-01
-5.78934491e-01 -1.12174177e+00 -3.80569398e-01 6.39181733e-01
5.51220104e-02 -6.08032286e-01 -5.06677270e-01 -4.77918178e-01] | [9.2589750289917, 10.202986717224121] |
6e4aeb9a-88cd-4285-b901-c4dd68c6b90f | 3d-human-pose-estimation-via-intuitive | 2303.18246 | null | https://arxiv.org/abs/2303.18246v2 | https://arxiv.org/pdf/2303.18246v2.pdf | 3D Human Pose Estimation via Intuitive Physics | Estimating 3D humans from images often produces implausible bodies that lean, float, or penetrate the floor. Such methods ignore the fact that bodies are typically supported by the scene. A physics engine can be used to enforce physical plausibility, but these are not differentiable, rely on unrealistic proxy bodies, and are difficult to integrate into existing optimization and learning frameworks. In contrast, we exploit novel intuitive-physics (IP) terms that can be inferred from a 3D SMPL body interacting with the scene. Inspired by biomechanics, we infer the pressure heatmap on the body, the Center of Pressure (CoP) from the heatmap, and the SMPL body's Center of Mass (CoM). With these, we develop IPMAN, to estimate a 3D body from a color image in a "stable" configuration by encouraging plausible floor contact and overlapping CoP and CoM. Our IP terms are intuitive, easy to implement, fast to compute, differentiable, and can be integrated into existing optimization and regression methods. We evaluate IPMAN on standard datasets and MoYo, a new dataset with synchronized multi-view images, ground-truth 3D bodies with complex poses, body-floor contact, CoM and pressure. IPMAN produces more plausible results than the state of the art, improving accuracy for static poses, while not hurting dynamic ones. Code and data are available for research at https://ipman.is.tue.mpg.de. | ['Dimitrios Tzionas', 'Michael J. Black', 'Omid Taheri', 'Chun-Hao P. Huang', 'Lea Müller', 'Shashank Tripathi'] | 2023-03-31 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Tripathi_3D_Human_Pose_Estimation_via_Intuitive_Physics_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Tripathi_3D_Human_Pose_Estimation_via_Intuitive_Physics_CVPR_2023_paper.pdf | cvpr-2023-1 | ['3d-human-pose-estimation'] | ['computer-vision'] | [-2.99081534e-01 2.52172202e-01 8.50124285e-02 1.72742717e-02
-1.45486176e-01 -4.58192617e-01 5.10496795e-01 -8.41734372e-03
-1.32239982e-01 7.97411561e-01 1.01893701e-01 1.41549706e-01
-1.59371980e-02 -5.70695877e-01 -1.14556110e+00 -3.27155441e-01
-2.49891356e-01 7.93047845e-01 2.53832400e-01 -3.33702207e-01
-3.66354920e-02 3.86541307e-01 -1.52869892e+00 -2.30536371e-01
8.01778555e-01 6.92613423e-01 -6.40550107e-02 7.98285484e-01
5.68117857e-01 3.04426581e-01 -3.60146582e-01 -6.71314657e-01
3.55253994e-01 -2.72557557e-01 -5.35172045e-01 8.18095654e-02
9.37077522e-01 -4.29939240e-01 -9.37584862e-02 6.01873875e-01
3.53009164e-01 2.90831715e-01 6.68487191e-01 -1.32493019e+00
-5.46328723e-01 8.39567035e-02 -7.65928924e-01 -4.06067431e-01
8.77113044e-01 3.44362915e-01 8.64469588e-01 -7.63121545e-01
8.30397785e-01 1.52242506e+00 1.13304305e+00 6.71801746e-01
-1.49395895e+00 -3.22824776e-01 3.14341933e-01 -3.06406289e-01
-1.35471725e+00 -1.93097040e-01 7.59119034e-01 -7.07852900e-01
6.48684621e-01 5.30408859e-01 1.49178600e+00 1.18773270e+00
6.52391136e-01 5.40155232e-01 1.00318241e+00 -1.88427195e-01
9.59616303e-02 -1.08087569e-01 -2.67696977e-01 1.10499406e+00
5.06413877e-01 1.66700423e-01 -6.59920931e-01 -3.15630794e-01
1.07330441e+00 -9.61254612e-02 -2.86371797e-01 -9.50483441e-01
-1.12142372e+00 4.51512665e-01 5.28788388e-01 -4.43075895e-01
-2.53265083e-01 6.17028713e-01 -1.00016437e-01 -3.96343946e-01
4.28730726e-01 4.04508114e-01 -5.58882415e-01 -1.82682753e-01
-6.15680397e-01 9.43539500e-01 1.04991543e+00 8.18785489e-01
6.59582138e-01 -1.02149121e-01 2.45416164e-01 4.98054385e-01
5.56009650e-01 8.17124069e-01 -1.20101318e-01 -1.39390588e+00
3.41917604e-01 6.27066135e-01 4.23378646e-01 -1.17687237e+00
-6.83733702e-01 -2.13862032e-01 -5.88184297e-01 5.23171604e-01
5.74827373e-01 -3.06366920e-01 -8.53750169e-01 1.61603105e+00
8.14515471e-01 9.37393233e-02 -4.47218478e-01 1.36228418e+00
5.13021529e-01 3.66964906e-01 -1.98243052e-01 3.10984790e-01
1.21108246e+00 -8.24637771e-01 -3.18681180e-01 -5.23659408e-01
1.86785892e-01 -4.24061269e-01 1.26046813e+00 4.95757341e-01
-1.43512475e+00 -4.47120070e-01 -9.21238840e-01 -1.92116186e-01
-1.36869505e-01 -1.04694545e-01 7.88233221e-01 4.51334298e-01
-8.13949764e-01 9.52125549e-01 -1.22323680e+00 -3.53761941e-01
-5.25644086e-02 3.25077772e-01 -3.82613420e-01 4.05040324e-01
-9.40699637e-01 1.22726643e+00 3.09800021e-02 2.70643294e-01
-8.12900126e-01 -9.59916651e-01 -1.07129157e+00 -5.11618197e-01
3.71210247e-01 -1.44761050e+00 1.04904485e+00 -7.42956877e-01
-1.65579402e+00 9.03437257e-01 1.92499235e-01 -2.73548752e-01
1.11407983e+00 -9.48986471e-01 1.80859283e-01 1.34897590e-01
-1.25894666e-01 8.18000019e-01 8.89590263e-01 -1.52291811e+00
2.84422040e-01 -2.52472132e-01 1.04732014e-01 4.42346185e-01
3.74401540e-01 -5.12417912e-01 -5.02060711e-01 -4.31991994e-01
2.08008751e-01 -1.27625895e+00 -3.68562847e-01 6.93289757e-01
-7.83505261e-01 3.35975647e-01 3.77721995e-01 -8.06161523e-01
7.18510866e-01 -1.79781926e+00 6.67713284e-01 2.53934264e-01
9.80954841e-02 -4.01811719e-01 4.75027829e-01 2.22454056e-01
3.09358060e-01 1.74592182e-01 -5.02373695e-01 -5.05047619e-01
2.37245947e-01 3.14479262e-01 1.00547764e-02 8.29909325e-01
-2.52505601e-03 8.02215934e-01 -8.33488047e-01 -7.51433313e-01
4.94782001e-01 6.39584661e-01 -9.07899559e-01 1.51489571e-01
-3.18240166e-01 6.57049775e-01 -2.08268926e-01 8.27443540e-01
4.78016466e-01 -2.16821015e-01 1.24038838e-01 -3.09654355e-01
-3.43279094e-02 5.49486279e-02 -1.50458598e+00 1.77580786e+00
-2.74513811e-01 7.08057806e-02 3.04970026e-01 -3.72174472e-01
6.73172534e-01 -4.56477748e-03 5.50182521e-01 7.17139244e-02
9.08506215e-02 -3.61137965e-04 -1.78902015e-01 -3.25794935e-01
4.54270929e-01 -3.24135721e-01 -1.53292045e-01 2.76573710e-02
-1.40126079e-01 -8.27383757e-01 -6.81106523e-02 -3.74303311e-02
6.36456132e-01 1.04922605e+00 2.51477957e-01 -3.91454101e-01
1.06389135e-01 4.81131040e-02 3.97536457e-01 2.76184678e-01
1.29739940e-01 9.28877175e-01 2.18190119e-01 -4.27139550e-01
-1.04369628e+00 -1.69621062e+00 -2.84924209e-01 7.64650047e-01
6.22536361e-01 -4.55774516e-01 -8.36015224e-01 -7.57733360e-02
6.40407622e-01 3.58386815e-01 -6.50182962e-01 -9.85437185e-02
-6.22716665e-01 -4.95862782e-01 3.69683921e-01 4.95174140e-01
2.30187863e-01 -5.40911555e-01 -1.02189040e+00 7.00976104e-02
-3.92170846e-01 -9.98111308e-01 -3.47972155e-01 -2.47510493e-01
-1.03021240e+00 -1.25176060e+00 -6.34983718e-01 2.71457247e-02
7.38189697e-01 -3.67424905e-01 1.42985415e+00 2.15776727e-01
-5.14350653e-01 6.65627420e-01 9.38074365e-02 -3.60686928e-01
-1.14176340e-01 -3.85673791e-01 5.22450089e-01 -4.95219409e-01
-5.53877592e-01 -5.45839727e-01 -7.79146433e-01 5.30692577e-01
-3.25933278e-01 5.54686069e-01 -6.43687770e-02 5.19395590e-01
8.47607613e-01 -3.08850944e-01 -3.14969331e-01 -2.95601279e-01
-1.08251860e-02 -2.63123930e-01 -6.08057320e-01 -1.11066632e-01
-5.30423261e-02 -1.45007461e-01 4.17446256e-01 -5.53995371e-01
-9.86603916e-01 1.40007600e-01 -2.29326244e-02 -5.20000875e-01
-1.07504293e-01 1.13094188e-01 -1.46620311e-02 3.00317234e-03
6.89623773e-01 -2.91055113e-01 -3.98551226e-02 -6.77494526e-01
4.98129040e-01 -2.57399023e-01 7.07929611e-01 -1.29368830e+00
9.56312716e-01 8.56080532e-01 2.82322109e-01 -7.85926819e-01
-6.46429479e-01 1.66789978e-03 -9.25579786e-01 -5.17841280e-01
8.11615825e-01 -7.86748528e-01 -1.16512001e+00 3.07265103e-01
-9.21464920e-01 -6.36879325e-01 -3.16451460e-01 5.58892488e-01
-8.22794914e-01 4.30011809e-01 -6.58347011e-01 -9.18124318e-01
-2.39468113e-01 -6.51458442e-01 1.46308506e+00 2.06898823e-01
-7.95939147e-01 -9.57296014e-01 1.48952603e-01 5.05429447e-01
-1.36473492e-01 9.86839294e-01 4.98709112e-01 4.13179934e-01
-5.23898542e-01 -1.04451559e-01 3.28908265e-01 -3.16420868e-02
-1.19642289e-02 6.28071785e-01 -7.07855761e-01 -2.83165485e-01
-4.58625674e-01 -4.01157260e-01 3.06804895e-01 6.90186322e-01
9.89540100e-01 -4.88972843e-01 -4.10407156e-01 8.96336913e-01
1.25964963e+00 -6.76488638e-01 3.01246881e-01 3.36372465e-01
8.73504758e-01 9.27545369e-01 5.31514287e-01 7.18591928e-01
5.58015049e-01 7.84931958e-01 7.58682132e-01 -1.12476550e-01
-3.66654471e-02 -7.56192088e-01 5.57239771e-01 5.03698349e-01
-7.02940345e-01 1.61454260e-01 -1.10970843e+00 2.40887031e-01
-1.72569633e+00 -5.64331830e-01 -4.85037357e-01 2.24526858e+00
8.53897333e-01 3.74377847e-01 3.03896427e-01 -2.40920082e-01
2.89248884e-01 -6.44695163e-02 -6.63856149e-01 -2.27982387e-01
9.00263637e-02 3.45028378e-02 5.33961356e-01 7.79991746e-01
-8.99567485e-01 7.61731923e-01 6.29167843e+00 8.54394361e-02
-9.85971332e-01 -9.71272588e-03 2.68169224e-01 -4.95415837e-01
-2.68547952e-01 7.91815817e-02 -6.59214973e-01 3.34131062e-01
4.77372140e-01 3.07880826e-02 3.07735264e-01 1.01247990e+00
2.72787124e-01 -3.92128199e-01 -1.38515520e+00 8.18328321e-01
-3.59247588e-02 -9.72744346e-01 -2.39924893e-01 -2.25505931e-03
5.88335276e-01 -1.94253415e-01 -3.95263471e-02 -5.08782379e-02
4.29362535e-01 -1.02238190e+00 1.39877534e+00 8.81573498e-01
6.28575563e-01 -2.06237704e-01 1.62576303e-01 5.71995735e-01
-1.24655926e+00 3.29843163e-01 -2.94483930e-01 -4.13203776e-01
4.58747894e-01 5.59779882e-01 -2.61642396e-01 5.78122258e-01
1.07138586e+00 5.66081405e-01 -5.11227727e-01 8.48569989e-01
-4.61340755e-01 1.73498124e-01 -8.63533378e-01 8.00614282e-02
-1.69603586e-01 -4.67856884e-01 7.66324997e-01 9.33576405e-01
1.20085314e-01 -3.69772911e-02 3.99008304e-01 1.29137123e+00
1.73934981e-01 -2.91848909e-02 -5.82974315e-01 4.79024231e-01
9.79684070e-02 1.15183139e+00 -6.20050430e-01 -2.06091017e-01
1.76533967e-01 8.82150114e-01 2.01153502e-01 2.76761323e-01
-1.20973969e+00 2.13497102e-01 8.47314358e-01 5.46678424e-01
-3.18464816e-01 -5.07730365e-01 -3.46005172e-01 -1.38047552e+00
3.03301126e-01 -5.44924438e-01 7.34640807e-02 -1.16428256e+00
-1.18860221e+00 -3.43607850e-02 6.18734479e-01 -1.09240186e+00
-2.25946791e-02 -7.85927296e-01 -3.98499519e-01 6.06171787e-01
-9.37565446e-01 -1.26551175e+00 -4.52531606e-01 4.89379585e-01
1.94680005e-01 9.36844766e-01 5.87667823e-01 -6.08961172e-02
-4.20753896e-01 1.88924149e-01 -5.26887476e-01 -2.39794016e-01
6.59554303e-01 -1.52989137e+00 4.44141716e-01 5.42414308e-01
-1.15666442e-01 7.76256323e-01 1.04407752e+00 -1.00659144e+00
-1.78460407e+00 -6.98405385e-01 1.80600837e-01 -1.04945254e+00
5.14439046e-01 -6.66854858e-01 -7.71146119e-01 8.59425187e-01
-1.84680164e-01 8.55693594e-02 4.18568760e-01 2.79637307e-01
-2.05952913e-01 7.18906969e-02 -1.06233239e+00 9.82054353e-01
1.31255889e+00 -9.82736349e-02 -8.55325580e-01 2.80662566e-01
3.24104398e-01 -1.16985941e+00 -9.66902316e-01 4.24862057e-01
1.11539042e+00 -9.01808500e-01 1.55314195e+00 -4.06747073e-01
4.74872351e-01 -3.72165203e-01 -8.99097184e-04 -1.23110640e+00
-9.11335871e-02 -7.00662792e-01 -5.39653242e-01 6.79542124e-01
1.77949175e-01 -4.58229601e-01 8.19317579e-01 1.01361620e+00
-1.41826048e-01 -8.36714447e-01 -1.01483357e+00 -8.57353091e-01
1.61604598e-01 -4.25155103e-01 3.60072374e-01 7.06583858e-01
-4.15473357e-02 -6.55059963e-02 -6.97111726e-01 4.02184576e-01
9.66457427e-01 1.67605653e-01 1.28075123e+00 -1.30227804e+00
-7.15694189e-01 -2.46175230e-01 -1.41581252e-01 -1.02276921e+00
-4.91501018e-02 -4.44955677e-01 2.09526166e-01 -1.63652384e+00
-5.66059090e-02 -4.65975851e-01 3.00021678e-01 6.31898046e-01
-1.18044838e-01 2.35586002e-01 3.56430769e-01 9.63648558e-02
-1.79127350e-01 5.24837255e-01 1.53266490e+00 1.37537926e-01
-2.41907611e-01 -2.49877498e-01 -2.51831472e-01 1.50209582e+00
5.73817611e-01 -3.21259797e-01 -8.92410353e-02 -2.90026456e-01
4.29487765e-01 4.25781384e-02 9.57615852e-01 -1.09479487e+00
-1.52309999e-01 -3.34635764e-01 8.85746598e-01 -4.06997114e-01
9.25975204e-01 -8.04176211e-01 7.73221076e-01 7.89707601e-01
1.85595974e-01 -3.83965783e-02 3.78747076e-01 4.12924230e-01
5.31826258e-01 1.51990756e-01 7.59184361e-01 -4.53894049e-01
-1.84373379e-01 2.03007400e-01 -8.17443710e-03 8.29115137e-02
7.66152978e-01 -5.36154985e-01 -1.37562275e-01 -2.12174818e-01
-8.11645806e-01 3.34730804e-01 1.13330185e+00 3.87940675e-01
6.49786413e-01 -1.09311068e+00 -4.78160650e-01 5.50211146e-02
-1.65904656e-01 3.45560730e-01 1.22689977e-01 1.02493215e+00
-9.91051078e-01 -2.56092578e-01 -2.15934575e-01 -9.56029296e-01
-1.10067928e+00 2.13384628e-01 7.78479457e-01 9.95510966e-02
-9.02556479e-01 7.41724610e-01 2.73587167e-01 -7.67983735e-01
-3.44846286e-02 -7.23324835e-01 4.01141465e-01 -3.75241697e-01
-1.29695758e-01 5.53334713e-01 -4.66852844e-01 -6.60447836e-01
-5.14768898e-01 1.21125579e+00 6.92091703e-01 -2.02645034e-01
1.08725357e+00 -1.07601024e-01 -5.76665923e-02 6.13718092e-01
5.54001927e-01 3.75767171e-01 -1.72123957e+00 3.71146202e-01
-3.74837011e-01 -5.67066193e-01 -5.42103171e-01 -6.82657897e-01
-6.89442694e-01 7.45806634e-01 2.40827829e-01 -1.35973558e-01
5.75257123e-01 1.54831469e-01 6.00612164e-01 1.81055382e-01
5.39900243e-01 -1.08510530e+00 1.69428393e-01 1.77050099e-01
1.34058368e+00 -9.97616351e-01 5.81477523e-01 -7.44119167e-01
-4.42527026e-01 9.00667608e-01 1.05278790e+00 -3.87426019e-01
6.24128461e-01 3.64783615e-01 -1.54936776e-01 -2.13666156e-01
-3.88336957e-01 9.11252871e-02 6.42742455e-01 5.34451962e-01
3.68025661e-01 2.39028931e-01 -2.02682465e-02 3.76213104e-01
-7.80403078e-01 -2.32020527e-01 2.80903429e-01 1.06333995e+00
-2.58895904e-01 -6.72111511e-01 -7.55221248e-01 1.71451226e-01
-3.39492619e-01 3.06466192e-01 -4.78444278e-01 1.05427432e+00
4.11755055e-01 4.46681857e-01 4.80254516e-02 -1.79503649e-01
6.17884994e-01 -1.74792334e-01 8.78777981e-01 -5.54878652e-01
-3.32536131e-01 1.33213714e-01 2.50639945e-01 -8.34902585e-01
-3.59835535e-01 -6.15100741e-01 -1.40840602e+00 -5.40557683e-01
-3.28181565e-01 -1.55877277e-01 5.26628613e-01 6.99801207e-01
7.87113309e-02 2.80268133e-01 -1.14198081e-01 -1.59098840e+00
-2.43359417e-01 -6.40322983e-01 -4.16059524e-01 6.93463504e-01
3.11094701e-01 -1.18815458e+00 -3.45403433e-01 3.14866453e-01] | [7.116305828094482, -1.1527976989746094] |
3f4737df-4a09-415d-9f10-0b32940fead4 | single-stage-multi-human-parsing-via-point | 2304.11356 | null | https://arxiv.org/abs/2304.11356v1 | https://arxiv.org/pdf/2304.11356v1.pdf | Single-stage Multi-human Parsing via Point Sets and Center-based Offsets | This work studies the multi-human parsing problem. Existing methods, either following top-down or bottom-up two-stage paradigms, usually involve expensive computational costs. We instead present a high-performance Single-stage Multi-human Parsing (SMP) deep architecture that decouples the multi-human parsing problem into two fine-grained sub-problems, i.e., locating the human body and parts. SMP leverages the point features in the barycenter positions to obtain their segmentation and then generates a series of offsets from the barycenter of the human body to the barycenters of parts, thus performing human body and parts matching without the grouping process. Within the SMP architecture, we propose a Refined Feature Retain module to extract the global feature of instances through generated mask attention and a Mask of Interest Reclassify module as a trainable plug-in module to refine the classification results with the predicted segmentation. Extensive experiments on the MHPv2.0 dataset demonstrate the best effectiveness and efficiency of the proposed method, surpassing the state-of-the-art method by 2.1% in AP50p, 1.0% in APvolp, and 1.2% in PCP50. In particular, the proposed method requires fewer training epochs and a less complex model architecture. We will release our source codes, pretrained models, and online demos to facilitate further studies. | ['Jian Zhao', 'Junliang Xing', 'Lei Jin', 'Jiaming Chu'] | 2023-04-22 | null | null | null | null | ['multi-human-parsing', 'human-parsing'] | ['computer-vision', 'computer-vision'] | [ 2.70325512e-01 5.38619995e-01 -1.51664698e-02 -5.48489869e-01
-1.02794659e+00 -4.70169693e-01 1.83598578e-01 4.69591767e-02
-3.86255771e-01 3.53776187e-01 -1.56674981e-01 7.17531983e-03
2.33845055e-01 -7.09215641e-01 -9.55542386e-01 -5.41631997e-01
1.81124732e-01 8.04541230e-01 6.40755415e-01 -6.48827553e-02
-2.55724844e-02 1.75323054e-01 -1.44007587e+00 3.86919200e-01
8.23024750e-01 1.07915318e+00 1.96889430e-01 6.70204997e-01
5.10144942e-02 1.84401691e-01 -6.00142598e-01 -6.90059006e-01
4.52950597e-01 -2.24616930e-01 -9.30563092e-01 8.21897984e-02
4.83504117e-01 -3.12465370e-01 9.60729271e-02 8.40848625e-01
5.56427419e-01 4.72254157e-02 3.87815744e-01 -1.08729649e+00
-4.63258982e-01 5.87284088e-01 -1.05240297e+00 -1.89190537e-01
1.64832085e-01 -7.68795609e-03 1.03113008e+00 -8.10060382e-01
5.04942238e-01 1.32813466e+00 7.78287530e-01 7.84317434e-01
-1.08218443e+00 -6.41686499e-01 2.54482895e-01 -1.44092962e-01
-1.18536365e+00 -1.90135628e-01 6.60811603e-01 -4.37911451e-01
8.02624702e-01 1.01647891e-01 5.56959569e-01 7.32271552e-01
1.07184172e-01 1.02736795e+00 7.35835671e-01 -2.43855864e-01
-1.26032252e-02 -3.59992266e-01 3.39078724e-01 1.12799108e+00
2.72987872e-01 -1.90527707e-01 -2.95379490e-01 1.98749895e-03
7.78040230e-01 -8.16984549e-02 1.90818176e-01 -3.31116259e-01
-9.26137447e-01 8.79542887e-01 5.08541822e-01 9.52341873e-03
-4.32637215e-01 3.72362547e-02 3.52467149e-01 -4.22429264e-01
3.76402795e-01 1.82998627e-01 -8.38769317e-01 2.31930181e-01
-1.11562026e+00 4.29365933e-01 5.33335626e-01 1.17880356e+00
7.11348414e-01 -4.58799303e-01 -2.69239634e-01 9.81058180e-01
2.53240943e-01 1.14670612e-01 3.89430970e-01 -9.58479106e-01
8.12644482e-01 7.97670066e-01 1.67975966e-02 -8.35141599e-01
-7.78831542e-01 -4.32672888e-01 -6.45625710e-01 -1.01156287e-01
3.90508980e-01 -3.02873611e-01 -1.28086388e+00 1.58262980e+00
9.45238233e-01 1.38584778e-01 -1.11295961e-01 9.76875365e-01
9.87716794e-01 5.99892378e-01 3.69360209e-01 2.74587631e-01
1.87940860e+00 -1.61682737e+00 -2.09396303e-01 -5.93176186e-01
5.16606271e-01 -5.71399033e-01 8.70573759e-01 1.71153545e-01
-1.30264127e+00 -8.66907835e-01 -9.45616186e-01 -4.63252991e-01
-1.41453624e-01 5.52979171e-01 6.77522123e-01 5.06591141e-01
-7.48044848e-01 8.56224775e-01 -1.14210260e+00 -1.54585138e-01
5.53585172e-01 4.75561172e-01 -3.98038954e-01 1.61359668e-01
-9.19776976e-01 5.42732716e-01 5.41507185e-01 3.79888773e-01
-6.25121534e-01 -7.28895485e-01 -1.15257025e+00 2.07796499e-01
4.62106347e-01 -8.46237957e-01 1.33097637e+00 -6.50796056e-01
-1.49347174e+00 1.11453557e+00 -9.92065370e-02 -3.43430191e-01
5.82079470e-01 -5.77416718e-01 1.36935376e-02 2.70708591e-01
4.67789233e-01 1.04663384e+00 7.30141461e-01 -1.12919259e+00
-9.86112356e-01 -4.75768656e-01 -7.92866051e-02 7.81107247e-02
2.69548655e-01 4.54033948e-02 -1.18457794e+00 -6.26989603e-01
4.91524547e-01 -9.82254684e-01 -3.99651855e-01 -2.14491680e-01
-8.35387290e-01 -3.46112818e-01 4.08628553e-01 -1.02511656e+00
1.13311422e+00 -2.01754951e+00 2.43393645e-01 -3.96432206e-02
1.78416207e-01 1.74524754e-01 -2.20013469e-01 1.24931904e-02
5.57612963e-02 1.17251486e-01 -4.83628482e-01 -7.47520506e-01
7.30804820e-03 9.01632607e-02 2.34228328e-01 3.96292239e-01
4.73965943e-01 9.29916382e-01 -6.78809643e-01 -8.26395392e-01
1.69440001e-01 2.87860394e-01 -7.17928350e-01 2.28369758e-01
-6.85541704e-02 4.21499431e-01 -4.55361009e-01 9.15057421e-01
8.06308270e-01 -3.08683455e-01 1.89927980e-01 -2.27294922e-01
-7.21888542e-02 2.52517253e-01 -1.10193527e+00 1.80166245e+00
-2.17749193e-01 -7.02625290e-02 3.34204018e-01 -8.58845711e-01
7.68772483e-01 1.51207866e-02 5.48122227e-01 -4.79844689e-01
1.11818597e-01 7.89518803e-02 -1.95420589e-02 -3.08045775e-01
5.07814109e-01 9.22980383e-02 -5.15330493e-01 -8.59308168e-02
1.93685442e-01 1.79698706e-01 3.77514929e-01 4.05438915e-02
9.30160403e-01 4.17088926e-01 3.19783032e-01 -2.85372913e-01
4.42749053e-01 -9.46956780e-03 1.15623868e+00 6.33690894e-01
-2.66173929e-01 9.81383204e-01 6.57407522e-01 -4.28851187e-01
-8.49087596e-01 -8.62596512e-01 2.91304626e-02 1.33949029e+00
2.16242090e-01 -4.73511368e-01 -1.21280551e+00 -9.39900100e-01
-3.05304280e-03 3.87035012e-01 -7.76513934e-01 2.04839453e-01
-1.13130045e+00 -8.74403656e-01 3.73799354e-01 8.68502080e-01
5.76162457e-01 -1.22913182e+00 -8.46124828e-01 2.73851961e-01
-1.80749446e-01 -1.29361713e+00 -5.05929828e-01 1.76955417e-01
-7.64318585e-01 -1.20685339e+00 -7.51384377e-01 -1.11079895e+00
7.41850734e-01 -2.48304993e-01 1.07019496e+00 3.60968113e-01
-3.15757215e-01 -1.21333919e-01 -4.55763161e-01 -6.20824173e-02
-9.91488770e-02 3.97495419e-01 -3.89264226e-01 -2.14673609e-01
1.63958389e-02 -1.98038414e-01 -7.45489061e-01 3.66077244e-01
-4.41441745e-01 5.01130819e-01 7.38413751e-01 8.53773952e-01
9.32803154e-01 -1.60224095e-01 2.78202444e-01 -1.19163632e+00
-5.44325635e-02 -4.63817894e-01 -5.99083722e-01 2.54963875e-01
-2.56405443e-01 -1.48781366e-03 5.74088931e-01 -2.23343715e-01
-9.51897025e-01 6.15014672e-01 -4.51313317e-01 -1.93689093e-01
-2.70455807e-01 1.34698883e-01 -5.13572335e-01 2.06845596e-01
1.33335516e-01 -1.03507474e-01 -4.28070962e-01 -7.58359909e-01
4.31500852e-01 3.63716722e-01 8.98955226e-01 -7.78509200e-01
6.79660678e-01 2.21595615e-01 -3.73045243e-02 -2.26343438e-01
-1.02428639e+00 -4.22748387e-01 -1.06698728e+00 9.41983312e-02
1.25212753e+00 -8.70046258e-01 -5.25621712e-01 4.70614344e-01
-1.19588780e+00 -3.50186765e-01 4.58133481e-02 -3.20115872e-02
-4.23752576e-01 3.74897361e-01 -9.72337008e-01 -4.95461673e-01
-7.13676572e-01 -1.25790679e+00 1.74191058e+00 4.20780808e-01
-9.22812000e-02 -3.47361803e-01 -2.56152481e-01 8.19056690e-01
-2.13164195e-01 4.44722056e-01 9.61386740e-01 -8.12762380e-01
-3.10246915e-01 -1.98011070e-01 -2.12403998e-01 5.15384451e-02
-2.02723429e-01 -3.68165821e-02 -8.21188211e-01 -9.54205617e-02
-4.52690601e-01 -1.43597409e-01 8.58256280e-01 4.90701526e-01
1.42012167e+00 -4.94248904e-02 -6.52108669e-01 8.05193782e-01
1.09078276e+00 6.67612329e-02 3.41952294e-01 3.57097000e-01
9.56770658e-01 8.24642181e-01 8.36026192e-01 3.53489429e-01
7.01080024e-01 6.96115494e-01 4.33180213e-01 -2.97184050e-01
-2.16116995e-01 -4.32966322e-01 9.67443548e-03 5.51894546e-01
1.51416259e-02 -1.66187197e-01 -8.91506314e-01 5.96469343e-01
-1.92754996e+00 -5.44108689e-01 -1.00869641e-01 1.91319263e+00
6.57068551e-01 2.53984988e-01 2.97162026e-01 -1.36216134e-01
1.01754200e+00 1.61870465e-01 -5.95175564e-01 -2.92315066e-01
3.06109607e-01 4.19139057e-01 2.85126001e-01 3.84832263e-01
-1.56077218e+00 1.34284830e+00 5.59139538e+00 5.75571895e-01
-6.41402245e-01 7.02355355e-02 7.70114064e-01 -8.26624855e-02
2.55881578e-01 -1.05751917e-01 -1.17538798e+00 5.00651360e-01
6.13210559e-01 5.11288702e-01 1.83132276e-01 1.02591383e+00
1.09579964e-02 8.97524208e-02 -1.14518261e+00 7.68496752e-01
-1.28995568e-01 -9.44445431e-01 -6.03166744e-02 -2.33676538e-01
2.98892111e-01 -9.45097357e-02 -3.29508901e-01 6.46788418e-01
1.55425072e-01 -7.99830079e-01 1.11620677e+00 7.30573609e-02
5.47462523e-01 -7.04520643e-01 6.72010005e-01 3.43131810e-01
-1.58632421e+00 -1.38819307e-01 -3.74040455e-01 1.02563389e-01
4.32913721e-01 4.28937584e-01 -4.57469165e-01 6.48616850e-01
9.48983669e-01 3.89195383e-01 -6.92934990e-01 8.43397260e-01
-5.47100782e-01 4.86943871e-01 -3.34632099e-01 2.64553249e-01
1.30171135e-01 -2.73305103e-02 1.65037885e-01 1.22431695e+00
1.66013509e-01 2.62084365e-01 3.33256513e-01 6.65381610e-01
-2.18868181e-01 1.83880925e-01 2.52844274e-01 3.52545559e-01
4.77816641e-01 1.58555758e+00 -1.04920387e+00 -5.07884979e-01
-4.56706911e-01 1.14593983e+00 5.93450069e-01 3.11396345e-02
-1.15782285e+00 -4.26416457e-01 4.75039661e-01 7.89642483e-02
7.34180748e-01 1.12899207e-01 -4.22246218e-01 -1.00951529e+00
2.41146654e-01 -8.31282735e-01 7.45816946e-01 -5.30774713e-01
-1.06095862e+00 7.56903470e-01 -7.96751585e-03 -7.52294064e-01
-2.58854836e-01 -6.03234231e-01 -5.27399838e-01 7.99636245e-01
-1.18120134e+00 -1.58378637e+00 -2.69645065e-01 2.31155396e-01
7.60612845e-01 3.79000306e-01 5.88918209e-01 2.96292752e-01
-1.04124177e+00 7.97234178e-01 -5.31534493e-01 4.68625247e-01
4.10265982e-01 -1.29586947e+00 9.49185252e-01 1.05537963e+00
-1.58735961e-01 5.08284628e-01 4.33835953e-01 -7.73751378e-01
-9.64538574e-01 -1.28397071e+00 8.37714672e-01 -3.17443550e-01
2.94693410e-01 -6.58970892e-01 -7.22859681e-01 8.91081691e-01
-1.04550079e-01 2.17844680e-01 3.64779234e-01 2.02454180e-01
-2.12416530e-01 7.33863339e-02 -1.28252685e+00 4.58968371e-01
1.20326376e+00 1.20154709e-01 -6.54998481e-01 1.60128653e-01
9.29403901e-01 -8.56609225e-01 -9.04988408e-01 6.64078534e-01
6.80105567e-01 -7.94332862e-01 1.10349667e+00 -5.18241048e-01
5.52378058e-01 -2.63251662e-01 -4.28251065e-02 -8.73591840e-01
-5.06871998e-01 -3.90821874e-01 -1.36310846e-01 1.31890106e+00
6.48610592e-01 -3.69999647e-01 1.08931851e+00 7.90949523e-01
-5.30996621e-01 -1.10976911e+00 -8.54475856e-01 -3.82657081e-01
2.00786188e-01 -1.89043418e-01 6.18613362e-01 5.31859457e-01
-3.55804741e-01 2.99633473e-01 -3.34432721e-01 5.95416486e-01
6.31717443e-01 4.39757615e-01 7.59389043e-01 -1.15970159e+00
-5.79844773e-01 -2.57444739e-01 -1.95593312e-01 -1.23625040e+00
2.18674839e-01 -7.33654022e-01 4.38167065e-01 -1.58354962e+00
3.24984372e-01 -3.81042600e-01 -1.40904754e-01 9.12829638e-01
-6.77990496e-01 9.96807292e-02 3.68396252e-01 1.37607930e-02
-6.57200336e-01 2.49660790e-01 1.14938736e+00 4.24415544e-02
-3.30048293e-01 1.50462866e-01 -7.41193652e-01 1.03110945e+00
8.92617404e-01 -6.60663009e-01 -6.37343079e-02 -6.89934373e-01
-2.81199217e-01 3.80399227e-02 3.25298190e-01 -1.10454392e+00
1.26674175e-01 6.28784075e-02 6.14355326e-01 -8.84995580e-01
4.30179417e-01 -5.52434683e-01 -8.21326077e-02 5.35456419e-01
-1.47488981e-01 2.11048976e-01 2.62430966e-01 3.15899014e-01
8.00147131e-02 -2.54359096e-01 8.63614440e-01 -2.99146801e-01
-6.45672977e-01 3.62125814e-01 5.34808934e-02 2.46833395e-02
1.02642262e+00 -2.10370999e-02 -2.11589918e-01 3.24786156e-01
-8.96943212e-01 5.28477490e-01 2.76474595e-01 4.34919924e-01
3.69767994e-01 -8.68487835e-01 -5.80581248e-01 2.29706213e-01
-1.46902099e-01 5.39727986e-01 3.90058756e-01 6.40434504e-01
-5.97256422e-01 2.62582511e-01 -5.61502092e-02 -5.87938011e-01
-1.42520463e+00 4.63942945e-01 2.92896777e-01 -4.82470423e-01
-9.59159791e-01 1.11330593e+00 4.24467176e-01 -6.84364557e-01
7.78229311e-02 -5.13666213e-01 -9.55109000e-02 -1.05438642e-01
3.71393919e-01 3.84547383e-01 -4.75380458e-02 -8.29419732e-01
-6.82721972e-01 7.38686562e-01 -9.26055536e-02 1.24981888e-01
1.35709250e+00 1.44600660e-01 -7.86407851e-03 -1.40825599e-01
9.61639643e-01 -6.43453822e-02 -1.42486191e+00 2.21688777e-01
4.42657061e-02 -1.02730848e-01 -3.49527627e-01 -9.81317997e-01
-1.34043229e+00 7.72194803e-01 3.71677279e-01 -8.88100415e-02
1.16475368e+00 3.76798660e-01 1.14184320e+00 -1.00888185e-01
3.56194913e-01 -9.15890276e-01 -3.01588207e-01 3.07641298e-01
7.04571486e-01 -1.09273493e+00 1.97919942e-02 -1.01293004e+00
-5.56262970e-01 8.24379742e-01 1.08242095e+00 -1.55036956e-01
3.64347577e-01 2.28319526e-01 -3.98884602e-02 -1.73843950e-01
-4.75565344e-01 -1.02434240e-01 3.56817096e-01 3.32384437e-01
4.06860232e-01 2.01267749e-01 -5.09716153e-01 1.41353130e+00
-4.02294248e-01 -3.89736265e-01 -5.35726286e-02 9.67908323e-01
-3.74619395e-01 -1.05091572e+00 -2.96697974e-01 3.48277658e-01
-6.84024811e-01 -4.75628451e-02 -2.79381186e-01 9.48011160e-01
5.53864121e-01 7.88855672e-01 7.33397901e-02 -3.59402090e-01
6.42371118e-01 9.68436524e-02 3.24174315e-01 -8.21690559e-01
-8.61864448e-01 1.83823675e-01 2.18303069e-01 -9.07481611e-01
-2.31152728e-01 -7.46032715e-01 -1.72146583e+00 8.70358422e-02
-4.48812306e-01 1.20533044e-02 5.43279350e-01 8.68975103e-01
6.64590180e-01 5.47506213e-01 1.30121127e-01 -1.13800967e+00
-3.03150475e-01 -9.37263250e-01 -3.42884570e-01 4.28683102e-01
-3.84295918e-02 -6.75825596e-01 -4.55225036e-02 2.20578145e-02] | [8.609846115112305, -0.0028932930435985327] |
199d5419-a28e-4dcd-b6f6-9b694975a123 | online-anomaly-detection-in-surveillance | 2010.07110 | null | https://arxiv.org/abs/2010.07110v1 | https://arxiv.org/pdf/2010.07110v1.pdf | Online Anomaly Detection in Surveillance Videos with Asymptotic Bounds on False Alarm Rate | Anomaly detection in surveillance videos is attracting an increasing amount of attention. Despite the competitive performance of recent methods, they lack theoretical performance analysis, particularly due to the complex deep neural network architectures used in decision making. Additionally, online decision making is an important but mostly neglected factor in this domain. Much of the existing methods that claim to be online, depend on batch or offline processing in practice. Motivated by these research gaps, we propose an online anomaly detection method in surveillance videos with asymptotic bounds on the false alarm rate, which in turn provides a clear procedure for selecting a proper decision threshold that satisfies the desired false alarm rate. Our proposed algorithm consists of a multi-objective deep learning module along with a statistical anomaly detection module, and its effectiveness is demonstrated on several publicly available data sets where we outperform the state-of-the-art algorithms. All codes are available at https://github.com/kevaldoshi17/Prediction-based-Video-Anomaly-Detection-. | ['Yasin Yilmaz', 'Keval Doshi'] | 2020-10-10 | null | null | null | null | ['anomaly-detection-in-surveillance-videos', 'anomaly-detection-in-surveillance-videos'] | ['computer-vision', 'methodology'] | [ 5.28151505e-02 -3.69426221e-01 -1.45301104e-01 -3.71351331e-01
-4.97422069e-01 -2.85171628e-01 2.17508703e-01 1.36582732e-01
-3.52080762e-01 2.38648191e-01 -2.98519224e-01 -4.00204778e-01
-1.31543025e-01 -5.49427509e-01 -7.35538840e-01 -8.66336882e-01
-4.90189910e-01 1.05006769e-01 3.73218626e-01 1.12244263e-01
1.84264749e-01 3.03551167e-01 -1.64791417e+00 1.06596239e-01
1.01873446e+00 1.60606647e+00 -3.34943086e-01 7.99162626e-01
1.45218283e-01 8.51031363e-01 -2.69786447e-01 -4.85467762e-01
7.01893747e-01 -4.61315274e-01 -4.28122371e-01 3.31982613e-01
3.17984939e-01 -9.22260582e-01 -5.01592398e-01 1.32319248e+00
3.55879605e-01 1.18459247e-01 4.64742124e-01 -1.52495289e+00
-3.27541083e-01 9.92631540e-02 -6.92055345e-01 8.56977940e-01
1.22848794e-01 3.60838860e-01 9.90361810e-01 -7.87722528e-01
-2.92959157e-02 8.52917731e-01 5.78767598e-01 4.70637202e-01
-6.36876285e-01 -6.63177013e-01 4.84925985e-01 7.31334329e-01
-1.19893229e+00 -3.83714914e-01 6.99693620e-01 -5.72477698e-01
6.33736908e-01 9.35254097e-02 8.56064677e-01 1.20179403e+00
3.71417552e-01 1.05511832e+00 5.47554612e-01 -1.10121638e-01
3.38586509e-01 -3.62475514e-01 -4.54162583e-02 8.87671113e-01
5.54168940e-01 1.94363043e-01 -2.89029926e-01 -3.09405148e-01
5.90557277e-01 5.43778360e-01 -2.07366318e-01 -2.77401954e-01
-7.90069401e-01 8.88516009e-01 1.57918647e-01 3.30685079e-02
-7.33585119e-01 5.09704240e-02 6.94911063e-01 6.41921461e-01
7.38080442e-01 2.21719697e-01 -4.79985058e-01 -2.63905883e-01
-8.45141768e-01 1.97888255e-01 6.54058099e-01 5.85714459e-01
2.33275369e-01 2.85009205e-01 -1.42870650e-01 5.52363813e-01
3.13647777e-01 4.52975988e-01 2.97633082e-01 -9.32107151e-01
1.55186236e-01 7.01851428e-01 2.75936157e-01 -1.24431574e+00
-2.29071259e-01 -3.30887645e-01 -9.92416739e-01 2.58500040e-01
5.37885606e-01 -4.51009154e-01 -6.11376822e-01 1.51746166e+00
5.49738705e-01 4.67281044e-01 -1.93354189e-01 1.10244656e+00
1.62574589e-01 5.95721483e-01 -1.32607091e-02 -3.87805760e-01
1.14462173e+00 -9.34228659e-01 -8.11344624e-01 -2.12885708e-01
5.28699577e-01 -5.21914423e-01 5.78162253e-01 7.18356192e-01
-7.46889591e-01 -2.14879602e-01 -9.34262633e-01 4.55177397e-01
-1.29561514e-01 2.36510932e-01 5.92791200e-01 5.42031169e-01
-9.64953899e-01 2.70026982e-01 -1.09794354e+00 -4.84915227e-01
6.28279209e-01 1.71956122e-01 -2.13911399e-01 -7.68384039e-02
-9.99573529e-01 5.84238827e-01 3.47352564e-01 3.90068650e-01
-1.22789931e+00 -4.55321312e-01 -7.86646068e-01 -1.15610600e-01
1.02363598e+00 -5.51073790e-01 1.54086185e+00 -1.38004363e+00
-1.34424531e+00 7.10451841e-01 -9.76473391e-02 -8.83650422e-01
9.55359221e-01 -5.83878100e-01 -5.08674085e-01 3.52390677e-01
-9.03718248e-02 1.99669629e-01 1.24098432e+00 -6.66903019e-01
-1.09106672e+00 -3.86840165e-01 1.75890356e-01 5.47198802e-02
-3.90598863e-01 2.44047314e-01 -3.28051656e-01 -8.26217890e-01
-1.52022362e-01 -7.37462640e-01 -4.78896707e-01 4.40378487e-01
-2.63322473e-01 -3.10325176e-01 8.80756974e-01 -9.16890919e-01
1.46387064e+00 -2.18371940e+00 -1.74853444e-01 9.04683843e-02
2.10640788e-01 4.84544665e-01 4.33054417e-02 1.98706269e-01
-9.08924118e-02 -1.53785467e-01 -3.76863897e-01 -1.19003542e-01
-1.49769172e-01 -2.50443723e-02 -1.67206839e-01 7.75576770e-01
6.08004093e-01 4.21903700e-01 -9.60507452e-01 -2.52942085e-01
2.93003350e-01 7.20858276e-02 -6.58185124e-01 4.74607497e-01
-1.79811478e-01 3.87594312e-01 -6.20229304e-01 1.02226663e+00
6.41279459e-01 -2.86763370e-01 -8.48632753e-02 1.78545550e-01
-6.47639781e-02 -1.62214935e-01 -1.02558494e+00 1.24055398e+00
2.21559569e-01 6.63937986e-01 3.16855371e-01 -1.53670835e+00
6.22714221e-01 4.42245245e-01 8.41393948e-01 -4.40757513e-01
4.11370367e-01 2.01430365e-01 2.26122975e-01 -8.00644159e-01
8.39663669e-02 3.15509528e-01 2.25504577e-01 2.06100285e-01
-1.90492213e-01 6.89873576e-01 3.99083108e-01 -1.02714345e-01
1.41081357e+00 -1.13440350e-01 4.36086923e-01 -1.42783120e-01
7.03868389e-01 5.74178845e-02 8.78301561e-01 8.82753313e-01
-8.44837248e-01 3.65185022e-01 6.45385563e-01 -8.16938043e-01
-9.59113300e-01 -7.32544184e-01 3.72980232e-03 1.11193752e+00
6.18146025e-02 5.08287810e-02 -7.75227666e-01 -7.90404975e-01
-1.74222179e-02 3.69546115e-01 -7.51490355e-01 -2.96405673e-01
-4.33830768e-01 -9.82079625e-01 3.24875563e-01 4.37728792e-01
5.14932990e-01 -9.66334224e-01 -9.09060299e-01 1.55878261e-01
-1.69857830e-01 -1.21017957e+00 -2.02177227e-01 -2.63273895e-01
-8.67443025e-01 -1.53367460e+00 -6.25027478e-01 -2.56520152e-01
7.89946735e-01 3.22970748e-01 9.16197956e-01 2.48396114e-01
-3.72312158e-01 6.37605846e-01 -6.14673436e-01 -7.62298048e-01
-3.58957112e-01 -2.08884120e-01 2.73735702e-01 4.85751569e-01
7.45161891e-01 -1.69489905e-01 -8.59704494e-01 3.89008015e-01
-1.04902697e+00 -2.47042194e-01 8.58786345e-01 7.69005120e-01
3.97790372e-01 -5.05333133e-02 5.96826851e-01 -4.27133054e-01
3.28786463e-01 -8.13406229e-01 -1.13471913e+00 7.90461898e-02
-3.76327127e-01 -2.71913797e-01 6.63464963e-01 -3.15338671e-01
-7.16043711e-01 -1.44439861e-01 -1.03303581e-01 -7.73438931e-01
-3.23579699e-01 3.28358084e-01 8.53368193e-02 7.40084574e-02
3.35501373e-01 2.12029234e-01 3.15896690e-01 -1.44875437e-01
-2.61325657e-01 2.99227029e-01 2.48350263e-01 -1.93003774e-01
5.86924076e-01 6.26190960e-01 -1.69279322e-01 -6.99996948e-01
-1.07462025e+00 -5.52132189e-01 -4.13208783e-01 -4.89224464e-01
8.77898812e-01 -8.92718077e-01 -6.10014737e-01 9.35523391e-01
-1.12973762e+00 -1.61516398e-01 2.21580341e-02 5.28291941e-01
-3.23884606e-01 5.16575396e-01 -5.50162077e-01 -1.15882611e+00
-4.24335331e-01 -9.73615289e-01 7.50633419e-01 1.45552382e-01
1.24674365e-01 -7.98852384e-01 1.92489300e-03 2.65435517e-01
5.09722292e-01 3.98188174e-01 3.33565027e-01 -9.63077247e-01
-6.71526015e-01 -6.12778187e-01 -3.50772411e-01 5.81376731e-01
1.27875939e-01 2.84666151e-01 -8.11956346e-01 -4.51673657e-01
7.25050196e-02 3.92338028e-04 9.01855290e-01 7.66347587e-01
1.63141489e+00 -6.25447035e-01 -9.68612880e-02 5.97967207e-01
1.23171067e+00 5.03956735e-01 3.48954767e-01 4.05644655e-01
5.09121418e-01 3.42353791e-01 7.86842644e-01 9.52805102e-01
1.77363917e-01 3.45395893e-01 8.48140895e-01 1.10690378e-01
4.90829766e-01 3.14569086e-01 6.25312090e-01 6.40164733e-01
-2.15975881e-01 -4.65398967e-01 -9.33760524e-01 5.87042689e-01
-2.24274778e+00 -1.15305495e+00 1.33400828e-01 2.38057661e+00
3.44182670e-01 1.31503344e-02 3.43580246e-01 1.55572683e-01
7.91761577e-01 1.67515025e-01 -9.17525768e-01 -2.65611231e-01
-3.81332636e-02 -4.44190204e-01 3.98769081e-01 1.13436624e-01
-1.69285381e+00 5.03841519e-01 5.79477406e+00 6.94057167e-01
-1.00355577e+00 -9.67707660e-04 1.01350749e+00 -2.24119753e-01
5.41197419e-01 -4.63395268e-01 -4.44856793e-01 7.69366145e-01
9.02974248e-01 -4.62558605e-02 2.02510774e-01 1.11399639e+00
4.80616838e-01 -1.42054468e-01 -1.00621688e+00 1.02320850e+00
9.57632363e-02 -9.33948576e-01 9.97799486e-02 -4.27196920e-02
6.51209056e-01 8.85786042e-02 -2.30335910e-02 1.52067468e-01
-7.47121423e-02 -6.93203032e-01 4.67335731e-01 5.30792236e-01
2.80695826e-01 -7.95312166e-01 1.14272141e+00 1.31743655e-01
-9.34030473e-01 -5.74656606e-01 -3.90659422e-01 -1.77993119e-01
-9.88822337e-03 8.31664205e-01 -4.18135136e-01 4.46803063e-01
1.05629134e+00 8.18983376e-01 -3.71090233e-01 1.41468978e+00
-9.48753953e-02 8.91201258e-01 -3.84896010e-01 -6.32305071e-02
4.01875466e-01 -2.51609683e-01 8.88499916e-01 1.06838071e+00
5.55643439e-01 8.03256258e-02 3.74368787e-01 4.36446100e-01
1.08915895e-01 2.26103067e-02 -7.64517903e-01 -2.09701121e-01
2.75242001e-01 1.12253881e+00 -4.52959299e-01 -2.11624369e-01
-6.68729067e-01 9.18694317e-01 -6.91136345e-02 2.34332040e-01
-1.05490232e+00 -1.71277061e-01 8.98558021e-01 2.48401780e-02
4.40083802e-01 -2.02071950e-01 -1.08278431e-02 -1.23176134e+00
5.31482637e-01 -1.15516376e+00 7.89295673e-01 -6.00991808e-02
-1.48471725e+00 3.27603042e-01 -4.37575057e-02 -1.62368679e+00
-4.00539398e-01 -8.65312636e-01 -8.91185939e-01 1.68182492e-01
-1.40945888e+00 -5.88447094e-01 -4.70448136e-01 4.87637401e-01
7.55162716e-01 -3.54786903e-01 5.45229495e-01 6.24513268e-01
-1.06575489e+00 5.79725027e-01 1.52451262e-01 4.52327728e-01
4.56284344e-01 -9.48966980e-01 2.68023133e-01 1.55232096e+00
-3.87548625e-01 -5.75423054e-02 7.53402412e-01 -3.77367705e-01
-1.45822513e+00 -1.08554256e+00 2.12034002e-01 -3.62474352e-01
9.59782839e-01 -1.03043556e-01 -1.10487843e+00 5.59924781e-01
2.28922609e-02 2.29387403e-01 6.52034760e-01 -2.12389037e-01
-6.53027371e-02 -1.87684476e-01 -1.14016283e+00 5.40140152e-01
1.01888609e+00 -2.48143170e-02 -3.03876072e-01 3.25118095e-01
4.11433786e-01 -2.99626648e-01 -2.75895983e-01 5.32872319e-01
5.31088233e-01 -1.11987627e+00 6.22108400e-01 -7.18888342e-01
4.62584794e-01 -2.70070165e-01 7.05589131e-02 -1.00810969e+00
-1.74778938e-01 -6.38012528e-01 -6.72809005e-01 6.73952937e-01
2.94543326e-01 -9.20768917e-01 6.72964096e-01 5.49905479e-01
-8.05180967e-02 -1.09685862e+00 -1.15311515e+00 -8.29047024e-01
-5.63151419e-01 -6.00605190e-01 4.91923541e-02 8.63576829e-01
-4.37859029e-01 -2.83495992e-01 -6.86847687e-01 5.60763717e-01
7.39658296e-01 -2.25844875e-01 5.54209888e-01 -1.22468328e+00
-2.08187863e-01 -6.06969774e-01 -7.88703382e-01 -9.01826918e-01
-1.53775409e-01 -1.61185950e-01 7.76162604e-03 -1.16893697e+00
4.35946472e-02 -6.58769310e-02 -7.00208306e-01 4.22787488e-01
-3.44685912e-01 -1.29384711e-01 -2.84569532e-01 9.20084119e-03
-1.05807436e+00 7.13012397e-01 6.48269951e-01 8.77345446e-03
5.20952381e-02 2.32255057e-01 -4.88261878e-01 1.11408520e+00
1.10015953e+00 -3.65129977e-01 -1.54458657e-01 -5.14993727e-01
1.98067069e-01 -1.98061287e-01 5.60908139e-01 -1.16430140e+00
2.33718693e-01 -3.14794421e-01 3.68717462e-01 -4.50709313e-01
1.52972713e-01 -9.86055434e-01 -3.73430789e-01 7.48280346e-01
-1.12313509e-01 5.24479091e-01 2.24411413e-01 1.00828099e+00
-2.73099065e-01 -7.96195567e-02 7.84626007e-01 1.32532388e-01
-1.00374806e+00 8.52221727e-01 -6.89978659e-01 2.12462824e-02
1.25233734e+00 -1.15142412e-01 -1.70316964e-01 -5.00403881e-01
-3.95607412e-01 5.31344593e-01 1.63869634e-01 6.16388440e-01
7.01121986e-01 -1.13615739e+00 -9.61263120e-01 2.41863474e-01
1.85775280e-01 -4.86785872e-03 4.20996815e-01 1.34373379e+00
-6.17636204e-01 1.71840623e-01 -2.60780156e-01 -6.62473381e-01
-1.06256473e+00 5.75028598e-01 4.19589967e-01 -1.11963421e-01
-5.81726968e-01 7.01630414e-01 1.35283768e-01 1.37043774e-01
4.52807933e-01 -3.44976574e-01 -1.32575966e-02 -8.52892324e-02
8.42671156e-01 5.99123895e-01 -6.93650767e-02 -2.55407155e-01
-4.64708924e-01 1.81655318e-01 -2.06856877e-01 3.49547893e-01
1.14534664e+00 5.14021888e-02 3.09831053e-02 3.31209570e-01
8.07174861e-01 -5.09630680e-01 -1.45370913e+00 -2.44271263e-01
8.44148472e-02 -6.80132151e-01 2.45188624e-02 -3.65222692e-01
-1.25289917e+00 7.04325795e-01 9.74831223e-01 5.63156307e-01
1.50618374e+00 -2.90688962e-01 7.51485467e-01 7.85685956e-01
1.24398194e-01 -1.20546818e+00 1.37106746e-01 5.67346573e-01
8.61526668e-01 -1.89586866e+00 -1.41628504e-01 -1.66290328e-01
-4.50491995e-01 1.01397336e+00 8.27160060e-01 -4.23261106e-01
7.79596269e-01 3.10994107e-02 1.09612316e-01 -1.20292321e-01
-8.63975823e-01 -4.28947695e-02 3.69190931e-01 2.61284739e-01
2.44480744e-01 -8.60676169e-02 -1.67611077e-01 4.35580224e-01
3.73541653e-01 -1.36582134e-02 2.49759376e-01 9.02177453e-01
-4.00481403e-01 -4.43068147e-01 -2.49583036e-01 7.85850883e-01
-9.80103135e-01 6.64323941e-02 -1.15775131e-01 5.09518147e-01
-8.06380734e-02 9.83194113e-01 2.61274040e-01 -1.99631840e-01
1.26253277e-01 -6.02702461e-02 1.46933272e-02 -9.69269350e-02
-1.23421572e-01 -2.48801887e-01 -1.17857344e-01 -8.89022529e-01
-4.47423428e-01 -7.56628275e-01 -6.69784546e-01 -3.94314796e-01
-2.55074143e-01 -1.28370464e-01 2.96875805e-01 1.11766851e+00
5.81237555e-01 3.71218026e-01 7.21164882e-01 -6.46747112e-01
-6.12035990e-01 -9.19173598e-01 -2.70946264e-01 2.63325989e-01
6.73651934e-01 -7.40536809e-01 -5.01937568e-01 -9.93293226e-02] | [7.866522789001465, 1.549054503440857] |
07ac3024-ae71-4e7e-b5d6-18f99010c205 | terrestrial-laser-interferometers | 2103.01740 | null | https://arxiv.org/abs/2103.01740v1 | https://arxiv.org/pdf/2103.01740v1.pdf | Terrestrial Laser Interferometers | Terrestrial laser interferometers for gravitational-wave detection made the landmark first detection of gravitational waves in 2015. We provide an overview of the history of how these laser interferometers prevailed as the most promising technology in the search for gravitational waves. We describe their working principles and their limitations, and provide examples of some of the most important technologies that enabled their construction. We introduce each of the four large-scale laser interferometer gravitational-wave detectors in operation around the world today and provide a brief outlook for the future of ground-based detectors. | ['Jo van den Brand', 'Hartmut Grote', 'Katherine L Dooley'] | 2021-03-02 | null | null | null | null | ['gravitational-wave-detection'] | ['miscellaneous'] | [-3.06891561e-01 -4.90613461e-01 1.98411047e-01 -5.63127883e-02
-2.73808062e-01 -6.52615190e-01 5.93294144e-01 -1.04408085e+00
-3.99873644e-01 6.30886137e-01 1.09016828e-01 -4.28715348e-01
-9.45912972e-02 -1.05647874e+00 1.03559403e-03 -6.70860231e-01
-5.27164102e-01 5.71473122e-01 6.35859489e-01 -6.49172366e-01
6.85985804e-01 6.74468040e-01 -1.33677161e+00 -5.04239798e-01
5.05560338e-01 5.77167690e-01 -5.24949282e-02 8.00507247e-01
3.94074053e-01 3.88284773e-02 -4.52505410e-01 1.11876559e-02
4.71130550e-01 -6.84264004e-01 -6.06544435e-01 -5.83910227e-01
1.40205055e-01 -4.79863077e-01 -9.21342850e-01 7.84929752e-01
8.27795208e-01 1.59173071e-01 5.19336641e-01 -6.83552146e-01
-3.68012607e-01 1.42119616e-01 -4.61238414e-01 9.00952756e-01
4.56492037e-01 3.83876771e-01 7.31428385e-01 -9.77923989e-01
5.06929636e-01 6.47807539e-01 6.20821476e-01 3.43124866e-01
-7.84413695e-01 -7.36889720e-01 -1.06068254e+00 3.03333670e-01
-1.41380858e+00 -8.79594028e-01 4.63737428e-01 -6.73636556e-01
1.37723494e+00 -7.51167610e-02 8.44081521e-01 1.14892915e-01
3.20835292e-01 -1.22100219e-01 1.01595938e+00 -8.90144289e-01
1.05630860e-01 -4.11663085e-01 -1.48533806e-01 4.90655780e-01
1.02230763e+00 7.65455306e-01 -9.73981798e-01 -2.53952652e-01
9.13072646e-01 -8.83841991e-01 -4.92934257e-01 4.72147837e-02
-7.83063412e-01 1.05751574e+00 3.58152688e-01 8.20426285e-01
4.87491628e-03 1.71124175e-01 -3.52769613e-01 3.30520391e-01
6.13677979e-01 9.92173374e-01 -6.67568296e-02 -2.62999743e-01
-9.88838553e-01 6.83075726e-01 3.38183165e-01 7.16414332e-01
8.79364312e-01 6.36139035e-01 5.48513412e-01 1.83154136e-01
5.42928040e-01 8.62012386e-01 1.30546853e-01 -3.45296025e-01
1.30416453e-01 -3.78024429e-02 4.22437549e-01 -3.54897588e-01
-6.56299829e-01 -4.44619775e-01 -8.91442597e-02 6.26433611e-01
1.79669291e-01 -7.59464979e-01 -8.24496150e-01 8.32581699e-01
-5.09989485e-02 6.28067017e-01 1.20427541e-01 8.48021626e-01
1.12797940e+00 2.29602888e-01 -4.47530001e-01 -8.88667181e-02
1.37085545e+00 -1.23606466e-01 -2.00243860e-01 -8.70367885e-01
6.32716179e-01 -1.01285696e+00 -4.28633690e-02 -4.40306425e-01
-9.59751368e-01 -4.32495400e-02 -1.27547657e+00 4.13836747e-01
2.18861520e-01 -3.66805255e-01 8.64061058e-01 1.24644399e+00
-1.07830966e+00 5.67334652e-01 -9.18020308e-01 -4.97553378e-01
-2.07082167e-01 1.10126734e-01 1.91635996e-01 4.71215218e-01
-1.25828040e+00 1.02402222e+00 3.53970528e-01 3.78078334e-02
-1.51505787e-02 -5.55153131e-01 -3.93401653e-01 -2.52927005e-01
-2.17378423e-01 -7.74963081e-01 1.33248734e+00 8.12402129e-01
-1.57900512e+00 1.25790977e+00 -1.20980866e-01 -6.00706935e-01
-1.30465522e-01 -1.44093648e-01 -9.72860515e-01 3.38871241e-01
1.66491896e-01 -2.45545477e-01 2.46233225e-01 -3.53606761e-01
-9.33497667e-01 -2.41441026e-01 -4.36554283e-01 -7.98849575e-03
5.69866240e-01 8.14636648e-01 1.86999336e-01 1.02337621e-01
8.72871876e-01 -1.04012597e+00 -1.05334051e-01 -7.63425887e-01
5.10190465e-02 -3.13050866e-01 5.55229187e-01 5.63377514e-02
8.42183053e-01 -1.85580218e+00 -5.34017920e-01 -1.04305707e-01
1.07392967e-01 2.64429331e-01 3.85146469e-01 1.03173542e+00
2.75341868e-02 2.13576600e-01 -1.10799618e-01 2.06035793e-01
-1.97707146e-01 -3.69571358e-01 -5.02842903e-01 9.34021294e-01
-2.07079127e-01 1.12464941e+00 -9.05112982e-01 2.76987702e-01
4.94732499e-01 -2.01844737e-01 -3.33241552e-01 8.14306438e-02
3.41539890e-01 9.89342570e-01 -5.81711411e-01 7.37520456e-01
1.18014276e+00 5.92649043e-01 -2.90534943e-01 2.27318272e-01
-1.19162142e+00 8.85875881e-01 -8.07271421e-01 1.34412897e+00
1.96643651e-01 9.44048285e-01 2.83506513e-01 -7.32674003e-01
1.12947917e+00 5.68236470e-01 3.03400099e-01 -6.21600091e-01
-1.47226602e-01 8.56878519e-01 2.35731229e-01 -7.22816229e-01
6.69309437e-01 -1.24559283e+00 -1.80269301e-01 3.65220040e-01
9.77453366e-02 -1.02788270e+00 2.10597858e-01 -1.16764322e-01
1.06210220e+00 1.61569007e-02 3.93015683e-01 -6.10850334e-01
2.54897892e-01 1.60890650e-02 2.93812156e-01 9.32629228e-01
-1.74204692e-01 5.45642138e-01 -3.66025954e-01 -7.74220347e-01
-6.67527616e-01 -8.42882931e-01 -5.47714114e-01 6.40830100e-01
3.40659291e-01 -3.98697317e-01 1.66631579e-01 4.42360908e-01
4.92340587e-02 5.00261366e-01 3.72157735e-03 1.23032264e-01
-5.31619132e-01 -1.49903035e+00 7.30291367e-01 2.96977349e-02
6.05055153e-01 -7.59757578e-01 -9.09598887e-01 1.78087294e-01
-3.21313560e-01 -9.45656300e-01 6.08232260e-01 -1.69879571e-01
-1.10951173e+00 -9.86490726e-01 -3.33058357e-01 -3.62551063e-01
-2.40427524e-01 6.87491477e-01 1.26953745e+00 2.14173242e-01
-4.49101776e-01 1.01446345e-01 -4.32980239e-01 -5.09841263e-01
-3.07100285e-02 -3.12529922e-01 6.26232326e-01 -7.18291104e-01
1.12040234e+00 -9.74160075e-01 -6.36481345e-01 5.58177792e-02
4.50680777e-03 -7.08884478e-01 3.71300966e-01 5.48874922e-02
-5.74276924e-01 2.37886515e-03 2.65508384e-01 -3.96974325e-01
2.22289205e-01 -3.74238342e-01 -1.19673312e+00 -5.03025830e-01
-4.63553995e-01 -1.05809905e-02 -2.70055324e-01 6.46372080e-01
-9.08656180e-01 -5.66717327e-01 -3.62411857e-01 7.22244382e-01
-3.51404212e-02 6.62842512e-01 6.50776625e-01 -1.17014146e+00
1.26857734e+00 1.52096406e-01 -5.64559877e-01 -6.48171127e-01
2.33159021e-01 4.99962330e-01 7.96794653e-01 -3.90816987e-01
1.59426856e+00 5.31675637e-01 6.23344779e-01 -1.30198097e+00
-3.79730076e-01 -1.04874134e+00 -8.33949029e-01 -2.68869311e-01
6.20763302e-01 -8.48676741e-01 -6.62891626e-01 7.22889066e-01
-1.04532015e+00 1.74160913e-01 -8.39930922e-02 1.35010982e+00
-2.52356827e-01 3.19964111e-01 -4.03266430e-01 -1.24363494e+00
-2.94233233e-01 -4.48780715e-01 6.61405623e-01 8.59823525e-01
2.04121023e-01 -8.41733694e-01 9.37339723e-01 -9.56478789e-02
9.24715102e-01 3.38120937e-01 1.16029002e-01 3.26248318e-01
-8.20352256e-01 -5.21916151e-01 -1.80942446e-01 -6.52200758e-01
7.82157630e-02 -6.86436519e-02 -9.51494575e-01 -4.03596431e-01
5.18346786e-01 -4.30531166e-02 1.25089896e+00 7.59210646e-01
-3.51567805e-01 4.59014535e-01 -7.72282362e-01 1.04312146e+00
1.82384121e+00 4.05881256e-01 8.60305905e-01 5.78127325e-01
1.72186643e-01 4.83357124e-02 2.67806053e-01 3.46291780e-01
-3.89157325e-01 7.15969980e-01 2.84288675e-01 2.37927005e-01
6.35638386e-02 2.66111910e-01 1.41999334e-01 5.82392752e-01
-1.20502901e+00 1.08707622e-01 -9.18473840e-01 4.75491822e-01
-1.11713278e+00 -1.35897505e+00 -9.38684583e-01 2.51418638e+00
-3.27513414e-03 9.70476866e-02 2.85177201e-01 -3.86971653e-01
5.38929164e-01 3.70050132e-01 2.02923659e-02 -5.64467385e-02
-2.21827868e-02 9.79361951e-01 1.15641630e+00 7.28327692e-01
-1.06969869e+00 9.84954834e-01 9.41192818e+00 -1.19631819e-01
-9.33752239e-01 1.31781310e-01 -8.31569552e-01 1.72754005e-01
-1.39288381e-01 5.21271884e-01 -9.61764574e-01 2.57026523e-01
1.00937629e+00 -7.48560607e-01 9.17613879e-02 3.42771471e-01
2.44497016e-01 -6.25394881e-01 -2.51961797e-01 7.17893958e-01
-4.13852930e-01 -1.36680818e+00 -6.88494265e-01 1.12736011e-02
6.68403864e-01 1.07712293e+00 -4.99331892e-01 8.51892531e-02
7.35162152e-03 -4.68655646e-01 3.68880987e-01 3.22147697e-01
1.00818753e+00 -4.75728214e-01 7.03774810e-01 1.15936652e-01
-1.55542421e+00 4.17409360e-01 -1.10542285e+00 -1.11817360e+00
8.69106174e-01 1.01128125e+00 -8.60316098e-01 9.52544689e-01
5.67147553e-01 3.40326846e-01 -1.86460659e-01 1.60100949e+00
-7.97126532e-01 9.41499949e-01 -7.38100529e-01 9.03476849e-02
3.77235264e-01 -7.76391923e-01 1.07878685e+00 1.00744009e+00
8.99499953e-01 5.61233521e-01 -1.67145759e-01 7.45734274e-01
2.18869552e-01 -5.29290795e-01 -1.00442982e+00 -1.23641476e-01
7.96541750e-01 9.56952035e-01 -4.33695942e-01 -4.12118360e-02
-7.84690201e-01 2.49671966e-01 -4.45816904e-01 1.56111062e-01
-2.93856144e-01 -8.99867773e-01 1.04624605e+00 4.32784468e-01
-1.48940548e-01 -1.01926219e+00 -2.82182068e-01 -1.20145583e+00
-2.53537327e-01 4.30504620e-01 3.27504337e-01 -6.54877484e-01
-7.07405806e-01 2.76331812e-01 9.01969150e-02 -1.60866797e+00
-4.90662605e-01 -7.47399032e-01 -1.28121018e+00 1.40629900e+00
-1.22705746e+00 -5.08784413e-01 -3.68071765e-01 -2.38537878e-01
-7.71075711e-02 -3.06455135e-01 8.76596689e-01 -1.22674089e-02
9.33169015e-03 -2.34319955e-01 3.00875872e-01 -9.16782618e-02
7.56077766e-01 -1.12705922e+00 1.48356545e+00 1.30583107e+00
7.35712377e-03 4.99635130e-01 1.37752879e+00 -1.07194185e+00
-1.36440539e+00 -2.26321235e-01 9.53713655e-01 -4.97288555e-01
1.04721701e+00 -1.87894210e-01 -4.38752353e-01 9.92036223e-01
2.16814205e-01 -1.48803607e-01 4.70816821e-01 4.50413615e-01
1.99772015e-01 3.14972609e-01 -9.71319914e-01 1.13333896e-01
1.37095189e+00 -6.79707110e-01 -9.96284902e-01 6.70426369e-01
2.64090031e-01 -4.84211415e-01 -3.52530301e-01 6.51173115e-01
9.74062562e-01 -1.32279921e+00 8.42615664e-01 -1.30097017e-01
-1.29949510e-01 -5.29685855e-01 4.84239250e-01 -1.22016251e+00
-1.17998886e+00 -1.15077877e+00 4.21299905e-01 4.97027338e-01
1.26159653e-01 -1.24806559e+00 1.13088524e+00 1.37104362e-01
-6.59822285e-01 2.91040182e-01 -1.33241427e+00 -1.17965424e+00
2.23764300e-01 -2.80298293e-01 2.05826223e-01 4.70842302e-01
5.89045167e-01 4.51629847e-01 -5.46458125e-01 5.69549680e-01
9.12650585e-01 4.56265330e-01 6.46059513e-01 -1.54069889e+00
-2.84092695e-01 -3.87517452e-01 -1.08247185e+00 -1.07448626e+00
-7.20949650e-01 -7.21911490e-01 2.16904357e-01 -1.60741556e+00
-2.88832366e-01 2.39830394e-03 3.02703232e-01 -1.95135489e-01
-7.07803071e-02 9.68210042e-01 -1.82654873e-01 5.44801354e-01
3.85382831e-01 1.29544169e-01 6.40466928e-01 3.68146002e-01
-2.48251572e-01 4.99603122e-01 -4.78586435e-01 6.17538333e-01
7.65065253e-01 -4.58360612e-01 2.79120170e-02 -4.32046384e-01
4.85405594e-01 -2.03916922e-01 -2.43801922e-01 -1.62788963e+00
4.98959869e-01 -3.22919965e-01 4.99300919e-02 -8.67087007e-01
1.36210024e-01 5.70380278e-02 3.37557793e-01 7.43346870e-01
1.15391278e+00 -4.98897225e-01 1.38972402e-01 6.74818307e-02
-2.27350757e-01 -7.82610536e-01 1.17374659e+00 -4.31379676e-01
-1.09922361e+00 1.64397702e-01 -8.12008083e-01 3.77265513e-02
6.64097905e-01 -1.66065082e-01 -8.06242287e-01 -2.11190879e-01
-4.40120041e-01 9.78000369e-03 5.51681876e-01 -1.35874510e-01
1.94390357e-01 -1.10077727e+00 -1.17467511e+00 3.50263894e-01
1.93391114e-01 -4.17835295e-01 1.86412081e-01 7.37612844e-01
-1.09797943e+00 1.00852704e+00 -4.20509756e-01 -3.53076339e-01
-8.40804756e-01 1.13683753e-01 5.85099578e-01 6.13340065e-02
-9.90205169e-01 1.30819368e+00 1.67630881e-01 1.56795047e-02
-1.08449519e+00 1.52363449e-01 -2.03154936e-01 -4.39125478e-01
9.26676273e-01 5.55014849e-01 3.54657441e-01 -6.30541742e-01
-8.35205376e-01 1.15119159e+00 5.69248796e-01 -5.30215323e-01
1.30552018e+00 -2.43821412e-01 -2.86499977e-01 3.18076611e-01
3.05288285e-01 4.67437625e-01 -3.52837533e-01 -2.95257300e-01
-3.75744589e-02 -7.25033820e-01 2.22754002e-01 -2.44007215e-01
-4.85827625e-01 6.12019837e-01 3.67200673e-01 4.34922695e-01
9.20739710e-01 3.48336190e-01 6.35744691e-01 4.35915053e-01
1.04168296e+00 -6.38405085e-01 -6.44511819e-01 1.01287031e+00
7.56815195e-01 -5.63232660e-01 4.40079093e-01 -6.69042468e-01
4.68067020e-01 1.40967381e+00 1.02719463e-01 -5.33895135e-01
7.52462685e-01 1.36626586e-01 -1.13152772e-01 -8.34100723e-01
-7.32907876e-02 -7.87977755e-01 1.42291561e-01 8.31165612e-01
7.31630445e-01 2.04465866e-01 -1.07221520e+00 -1.10800751e-01
-6.96027994e-01 -3.63542512e-02 1.02691829e+00 1.34002066e+00
-1.36702180e+00 -1.19847345e+00 -7.97873974e-01 3.48910213e-01
-7.80257657e-02 -1.47705600e-01 -3.03339541e-01 4.02904391e-01
-6.20126724e-02 1.24235773e+00 -1.91121437e-02 -3.89617234e-01
3.88439357e-01 5.47923326e-01 8.89243603e-01 -9.69316542e-01
9.41115469e-02 1.93310771e-02 3.93285602e-01 -2.79211611e-01
-7.80290246e-01 -4.23434645e-01 -9.40657020e-01 -5.35432994e-01
-7.48967052e-01 6.53218150e-01 6.63873672e-01 1.31839216e+00
-2.73857862e-02 -6.56076670e-02 6.80836797e-01 -1.48451924e+00
-1.38373896e-01 -1.41359043e+00 -1.45041943e+00 -6.13738298e-01
2.27591902e-01 -9.06913280e-01 -1.00990462e+00 -5.33555686e-01] | [7.554080009460449, 3.102977752685547] |
d5be0b99-9eab-488a-b60b-ab895ae984b6 | mot16-a-benchmark-for-multi-object-tracking | 1603.00831 | null | http://arxiv.org/abs/1603.00831v2 | http://arxiv.org/pdf/1603.00831v2.pdf | MOT16: A Benchmark for Multi-Object Tracking | Standardized benchmarks are crucial for the majority of computer vision
applications. Although leaderboards and ranking tables should not be
over-claimed, benchmarks often provide the most objective measure of
performance and are therefore important guides for reseach.
Recently, a new benchmark for Multiple Object Tracking, MOTChallenge, was
launched with the goal of collecting existing and new data and creating a
framework for the standardized evaluation of multiple object tracking methods.
The first release of the benchmark focuses on multiple people tracking, since
pedestrians are by far the most studied object in the tracking community. This
paper accompanies a new release of the MOTChallenge benchmark. Unlike the
initial release, all videos of MOT16 have been carefully annotated following a
consistent protocol. Moreover, it not only offers a significant increase in the
number of labeled boxes, but also provides multiple object classes beside
pedestrians and the level of visibility for every single object of interest. | ['Laura Leal-Taixe', 'Anton Milan', 'Stefan Roth', 'Konrad Schindler', 'Ian Reid'] | 2016-03-02 | null | null | null | null | ['multiple-people-tracking'] | ['computer-vision'] | [-2.32315406e-01 -4.33747768e-01 -2.55163938e-01 -1.18481293e-01
-5.75266659e-01 -6.76436543e-01 6.81540251e-01 3.95007968e-01
-7.76882112e-01 9.40478325e-01 -3.03309355e-02 1.48810849e-01
1.01421140e-02 -3.91588658e-01 -5.90812981e-01 -6.50067747e-01
-9.37974229e-02 6.66383386e-01 9.14799929e-01 4.54031825e-02
-1.15178436e-01 3.41945559e-01 -1.98266590e+00 8.20832103e-02
3.07485759e-01 8.39365780e-01 1.75044492e-01 7.82411039e-01
1.30725026e-01 6.68504536e-01 -8.91863227e-01 -7.01369047e-01
3.98585945e-01 -1.18936092e-01 -5.18944621e-01 -9.52315610e-03
1.31931269e+00 -4.29298729e-01 -1.62435427e-01 1.03953063e+00
6.36553347e-01 5.00519350e-02 2.58658975e-01 -1.74539053e+00
-1.81191623e-01 4.43082452e-01 -5.35406470e-01 6.09362066e-01
2.20489800e-01 1.73680454e-01 1.09987235e+00 -5.28018475e-01
6.95393503e-01 1.23643875e+00 8.89739513e-01 6.25731707e-01
-1.09433115e+00 -8.25284183e-01 1.75389364e-01 3.23023081e-01
-1.18458736e+00 -4.37352687e-01 1.11676559e-01 -6.28262043e-01
4.19491142e-01 6.99927211e-01 8.12868834e-01 1.08656383e+00
-4.67863716e-02 8.27766001e-01 9.73598957e-01 -1.70307934e-01
-8.44149590e-02 5.18789887e-01 4.67566729e-01 4.97746646e-01
9.24497426e-01 3.45186293e-01 -2.52707988e-01 -1.20344132e-01
4.05029058e-01 1.64834894e-02 -3.28682992e-03 -6.96594477e-01
-1.52531528e+00 5.24158359e-01 2.45837271e-01 1.84548602e-01
-2.31689308e-02 2.80430555e-01 5.59807837e-01 5.67351049e-03
1.19845964e-01 2.40639858e-02 -1.64123118e-01 -1.62253425e-01
-8.37933660e-01 7.68896937e-01 4.68652248e-01 1.00354326e+00
2.72714525e-01 -2.05058962e-01 -7.39649534e-01 4.11912531e-01
3.69942456e-01 8.40336084e-01 -6.87691793e-02 -9.96141195e-01
3.30255538e-01 6.49599075e-01 6.64854944e-01 -8.32154989e-01
-4.02370334e-01 -5.86470485e-01 -4.76240486e-01 5.73861063e-01
8.88952672e-01 -1.39238477e-01 -4.86073047e-01 1.63502598e+00
5.54865599e-01 1.99278906e-01 -3.92480463e-01 7.75149643e-01
1.08523965e+00 1.36611432e-01 2.93554962e-01 -7.87553042e-02
1.57574499e+00 -1.07928419e+00 -5.35745084e-01 -7.94709399e-02
3.64181072e-01 -6.74239457e-01 6.09108210e-01 2.00786442e-01
-9.19389307e-01 -7.92172849e-01 -9.32705879e-01 3.42804223e-01
-3.94238770e-01 -1.56958982e-01 3.65535289e-01 1.04339421e+00
-1.04654014e+00 3.08800459e-01 -5.66492915e-01 -6.32105291e-01
3.50918144e-01 2.05250308e-01 -2.62967646e-01 2.25345306e-02
-7.41043091e-01 1.16217721e+00 3.82986337e-01 -2.47509167e-01
-9.82514083e-01 -6.59057438e-01 -3.74183863e-01 -2.19138682e-01
5.34910738e-01 -4.82858121e-01 1.34555268e+00 -3.33295792e-01
-8.03920388e-01 9.74931002e-01 6.47454411e-02 -5.44048011e-01
1.00014567e+00 -2.52634376e-01 -5.42275667e-01 -1.35079771e-01
5.42977631e-01 8.18841040e-01 6.96035743e-01 -1.25792205e+00
-1.18453312e+00 -6.87086582e-02 2.23905772e-01 -3.21514048e-02
-1.88320667e-01 5.24687827e-01 -6.28068328e-01 -3.58378649e-01
-6.96957111e-01 -9.25148249e-01 -1.59881398e-01 2.15420485e-01
-2.75544941e-01 -2.57109791e-01 1.02611899e+00 -2.74843901e-01
1.07041728e+00 -2.06834126e+00 -2.41223481e-02 -3.26223522e-01
6.49214625e-01 3.67175728e-01 -2.73919106e-03 -5.15988618e-02
1.82963386e-01 -4.75169383e-02 3.87731016e-01 -5.68975747e-01
1.55640677e-01 7.41713941e-02 -3.41705494e-02 6.39150202e-01
-7.63881207e-02 7.22283304e-01 -1.14970183e+00 -7.92703152e-01
5.54267228e-01 2.31262207e-01 -2.09530666e-01 -6.65113032e-02
-2.02949151e-01 6.23433292e-01 -2.57214010e-01 6.72204912e-01
6.21776998e-01 -3.89058471e-01 -1.22452810e-01 -5.79379424e-02
-4.28254306e-01 -3.97033870e-01 -1.60180092e+00 9.63942111e-01
2.72761792e-01 8.62124681e-01 3.32987122e-02 -2.64893115e-01
5.30093372e-01 2.32008234e-01 9.05029953e-01 -5.60118139e-01
1.81105644e-01 -1.66417509e-01 6.50901697e-04 1.47379283e-02
8.93602073e-01 4.10159260e-01 -6.21048324e-02 1.23769999e-01
-1.36907354e-01 4.05306250e-01 9.83713150e-01 2.63164192e-01
1.09067929e+00 -1.87078677e-03 4.01869625e-01 -2.50741720e-01
4.75421727e-01 3.16742539e-01 6.61258101e-01 1.02031851e+00
-7.98075259e-01 4.39684272e-01 1.88617155e-01 -6.14349604e-01
-9.93681014e-01 -1.19262707e+00 -2.52806306e-01 1.34087706e+00
5.12033939e-01 -6.10562861e-01 -5.83465993e-01 -6.77825689e-01
2.32074156e-01 2.13229865e-01 -6.46370173e-01 4.22396898e-01
-3.92023623e-01 -7.45272934e-01 6.74360812e-01 5.33175349e-01
4.72167253e-01 -9.58586812e-01 -1.02173829e+00 5.00496179e-02
-2.39862919e-01 -1.43706727e+00 -4.54357952e-01 -1.32553160e-01
-5.84902704e-01 -1.56922424e+00 -8.73795748e-01 -4.03125137e-01
4.05833751e-01 6.51478052e-01 1.32752335e+00 4.20538217e-01
-4.56396580e-01 6.33978724e-01 -2.31399730e-01 -7.50149071e-01
-2.58359224e-01 -1.49877042e-01 1.94718286e-01 -1.51926339e-01
2.10979730e-01 1.87140018e-01 -4.10663575e-01 9.15697873e-01
-3.40972066e-01 -1.68406367e-01 3.73183161e-01 3.74671221e-01
5.26115358e-01 -1.32985339e-01 1.75422151e-02 -5.42763829e-01
1.80710718e-01 -1.39429346e-01 -1.19725370e+00 4.38689172e-01
-3.00691605e-01 -3.44006419e-01 3.89766470e-02 -3.91071826e-01
-7.36857057e-01 -3.66337039e-02 3.85048926e-01 -1.52323484e-01
-3.68287295e-01 -2.95040816e-01 7.95037076e-02 -3.07839602e-01
7.21444249e-01 -2.55640417e-01 -4.05142307e-01 -4.57685649e-01
2.43378311e-01 1.54840246e-01 5.83760679e-01 -3.83976609e-01
9.59655941e-01 5.63182712e-01 1.41156703e-01 -9.45369482e-01
-8.20557892e-01 -8.53359699e-01 -6.85697854e-01 -7.47714221e-01
9.59551632e-01 -9.08441246e-01 -1.14033628e+00 4.46100801e-01
-1.12532961e+00 -1.19440958e-01 -2.69017577e-01 4.13272232e-01
-5.59872314e-02 3.54953796e-01 -1.69971630e-01 -7.91777670e-01
-4.15205285e-02 -1.15575337e+00 1.04530585e+00 4.22292680e-01
-1.87543556e-01 -7.82334208e-01 1.89612716e-01 4.22354341e-01
4.49160606e-01 5.01891434e-01 1.39764488e-01 -5.70450187e-01
-8.82779300e-01 -4.32123452e-01 -4.32340354e-01 3.59085463e-02
-1.80740237e-01 2.09369406e-01 -8.80771697e-01 -5.27212679e-01
-6.19848132e-01 -1.08169086e-01 7.77481914e-01 5.36140203e-01
6.48600161e-01 2.79153228e-01 -7.51372933e-01 2.12167472e-01
1.23618436e+00 7.77340010e-02 3.40154141e-01 7.10410655e-01
5.41370034e-01 3.56791645e-01 6.11633003e-01 4.66333419e-01
5.30584872e-01 1.03858709e+00 6.87482297e-01 -3.35649550e-02
-3.79839629e-01 3.12465936e-01 4.89953309e-01 2.12820500e-01
-5.01231432e-01 -2.89346009e-01 -9.21711862e-01 3.03047568e-01
-1.82584071e+00 -1.36958611e+00 -6.69686317e-01 2.44838262e+00
1.88014567e-01 3.70354801e-01 8.28141928e-01 -4.63003218e-02
9.64865208e-01 6.83009103e-02 -3.02432209e-01 5.80393851e-01
-7.74143264e-02 -2.79105753e-01 7.12324560e-01 2.32229635e-01
-1.39518940e+00 6.16207182e-01 7.53574705e+00 5.82864106e-01
-5.78909218e-01 3.66205812e-01 -6.84241802e-02 -3.45569223e-01
5.88836730e-01 -2.30826080e-01 -1.61532354e+00 5.79919934e-01
8.84684980e-01 -2.48198584e-01 -6.62635639e-02 1.00233531e+00
-1.13206161e-02 -4.04291213e-01 -1.00356162e+00 9.45254564e-01
-1.75772727e-01 -1.20380843e+00 -3.34632188e-01 7.39566609e-02
4.74991769e-01 2.78987467e-01 -2.75463983e-02 4.08885568e-01
5.61742067e-01 -6.30106747e-01 1.19224417e+00 2.93570727e-01
4.46784735e-01 -3.75153095e-01 8.69549274e-01 8.89390521e-03
-1.63853872e+00 -2.63737589e-02 -4.17073935e-01 1.27389371e-01
2.92924196e-01 2.23051056e-01 -5.05347550e-01 5.02118945e-01
1.20410347e+00 6.83536649e-01 -1.11714685e+00 1.81762767e+00
1.37315452e-01 2.49225616e-01 -2.67419338e-01 -2.00671121e-01
3.41092274e-02 2.00244740e-01 8.06511819e-01 1.19995809e+00
-3.52917910e-02 -5.10625482e-01 7.09107876e-01 1.52720913e-01
1.71486869e-01 -1.24481253e-01 -4.36970800e-01 1.36869073e-01
3.88762921e-01 1.38759089e+00 -1.08509159e+00 -4.30221230e-01
-5.98798752e-01 3.62005502e-01 1.67577475e-01 1.79171632e-03
-1.29552984e+00 1.66764140e-01 8.19891989e-01 1.55076608e-01
4.26534802e-01 -1.50546297e-01 1.37551025e-01 -7.96945989e-01
-1.26236618e-01 -7.15278685e-01 5.80048263e-01 -5.47541440e-01
-1.03153872e+00 5.91827810e-01 4.11457807e-01 -1.48721349e+00
1.04517557e-01 -7.58440197e-01 -2.52829552e-01 3.79653931e-01
-1.21685231e+00 -1.17215550e+00 -8.72960329e-01 4.95059162e-01
3.01321268e-01 -2.58794963e-01 6.72400475e-01 7.13219762e-01
-7.42974937e-01 4.46131110e-01 2.13099673e-01 1.97158873e-01
8.71537626e-01 -1.19354463e+00 3.64185333e-01 7.93912470e-01
1.64102435e-01 2.93924719e-01 1.08622730e+00 -6.84073746e-01
-8.85293126e-01 -1.19203901e+00 3.39252979e-01 -1.22604537e+00
6.35973990e-01 -3.70978057e-01 -3.99945855e-01 7.91859388e-01
1.99063644e-01 1.99511766e-01 4.74419355e-01 1.61945790e-01
-1.60224929e-01 -2.44522050e-01 -9.15366769e-01 4.89002258e-01
1.02174985e+00 4.65316139e-02 -3.40499341e-01 4.89095539e-01
4.34754521e-01 -6.79885507e-01 -7.42671251e-01 1.75235733e-01
7.07068145e-01 -1.25002563e+00 1.12661994e+00 -3.74661118e-01
-2.54066586e-01 -6.29977047e-01 -2.23487601e-01 -8.92863691e-01
-4.42973852e-01 -3.40934724e-01 -1.98289156e-01 1.32290554e+00
2.64432102e-01 -3.11808646e-01 9.44657207e-01 4.08738524e-01
4.17496264e-03 -1.12158932e-01 -8.27707708e-01 -1.29768288e+00
-4.09355581e-01 -3.54895622e-01 4.72779483e-01 5.05309343e-01
-7.45588839e-01 1.08637348e-01 -6.52911365e-01 5.76310195e-02
1.37777412e+00 -1.76255658e-01 1.30717146e+00 -1.64508593e+00
-1.20018236e-01 -6.22962236e-01 -7.96296179e-01 -5.99706113e-01
-4.84743476e-01 -7.31894016e-01 2.23766584e-02 -1.50760710e+00
6.49563313e-01 -3.17468286e-01 -3.68528694e-01 2.98125535e-01
-2.28717536e-01 5.12508929e-01 4.69406307e-01 1.22186035e-01
-1.40308309e+00 3.66962962e-02 1.14695787e+00 -2.64284372e-01
2.39300087e-01 4.28434432e-01 -5.04119098e-01 5.83175898e-01
5.02968967e-01 -7.74181485e-01 4.85381857e-02 -3.50776136e-01
-3.69470306e-02 -4.54366922e-01 9.69319880e-01 -1.55010951e+00
3.88534963e-01 -1.67420015e-01 3.98011595e-01 -9.69824135e-01
5.00087857e-01 -9.55401719e-01 5.43215334e-01 6.75118923e-01
7.00947195e-02 3.87526542e-01 4.21055555e-01 6.53082967e-01
4.50122636e-03 -5.64743355e-02 9.88023520e-01 -2.09933788e-01
-1.23202598e+00 4.96067435e-01 -6.61543980e-02 2.00570643e-01
1.52790427e+00 -4.61622804e-01 -7.16779947e-01 6.48662522e-02
-5.03945827e-01 4.97561365e-01 5.13343751e-01 6.30327404e-01
7.80851319e-02 -1.36741316e+00 -7.78025210e-01 -1.89981669e-01
3.58430266e-01 -3.95978808e-01 5.73066100e-02 9.27423775e-01
-3.66014749e-01 4.21647847e-01 -6.18920386e-01 -1.00500870e+00
-1.98093259e+00 5.29729366e-01 1.72006965e-01 -4.21355695e-01
-7.39434063e-01 5.64998388e-01 -6.81331009e-02 -4.86141965e-02
6.89947844e-01 -1.41826153e-01 -3.61268461e-01 2.13774025e-01
1.00340819e+00 8.50438654e-01 -7.90451840e-02 -9.53897595e-01
-6.74681842e-01 4.53005999e-01 -1.09979816e-01 5.54794781e-02
1.14191282e+00 5.91169856e-03 6.87800497e-02 3.55702817e-01
4.47416544e-01 3.70167494e-02 -1.25445282e+00 -3.41799594e-02
2.64790148e-01 -7.09497869e-01 -2.30239138e-01 -6.85100496e-01
-8.45218122e-01 4.64063078e-01 9.38226223e-01 4.67114866e-01
4.94174838e-01 4.12010401e-02 3.67343277e-01 2.51319259e-01
7.10914433e-01 -7.90263176e-01 1.83983624e-01 4.04170036e-01
5.59880316e-01 -1.46445000e+00 2.07902342e-01 -1.71715304e-01
-5.19346535e-01 7.53905058e-01 9.69366610e-01 3.06368917e-01
3.36006612e-01 4.29554671e-01 1.75926998e-01 -1.94458261e-01
-4.27331775e-01 -4.92687017e-01 4.53482121e-01 8.46531451e-01
2.55050212e-01 4.41372674e-03 3.42296809e-02 1.47972226e-01
5.92804812e-02 -1.69913307e-01 2.26548925e-01 8.39798629e-01
-6.95509195e-01 -1.08938718e+00 -9.06786561e-01 3.92475098e-01
-3.45397234e-01 4.23684746e-01 -3.87815356e-01 1.17944753e+00
3.40243280e-01 1.03061998e+00 -1.26531780e-01 -3.08858812e-01
5.96742332e-01 -4.40492660e-01 6.72649145e-01 -4.74093646e-01
-7.46301234e-01 -4.20460880e-01 9.37673748e-02 -5.66160738e-01
-6.07231736e-01 -8.04602146e-01 -7.27362931e-01 -7.26380587e-01
-4.48288977e-01 1.48412183e-01 6.00202382e-01 7.68935919e-01
4.36621867e-02 6.35463774e-01 -6.58191293e-02 -1.03996265e+00
-2.69151539e-01 -8.23181331e-01 -4.20543641e-01 5.32858908e-01
3.10894400e-01 -1.27465379e+00 -3.35606784e-02 -8.83676484e-02] | [6.359598636627197, -2.009089469909668] |
9fad9e55-34d8-46fc-8987-3059317c3c79 | moc-gan-mixing-objects-and-captions-to | 2106.03128 | null | https://arxiv.org/abs/2106.03128v1 | https://arxiv.org/pdf/2106.03128v1.pdf | MOC-GAN: Mixing Objects and Captions to Generate Realistic Images | Generating images with conditional descriptions gains increasing interests in recent years. However, existing conditional inputs are suffering from either unstructured forms (captions) or limited information and expensive labeling (scene graphs). For a targeted scene, the core items, objects, are usually definite while their interactions are flexible and hard to clearly define. Thus, we introduce a more rational setting, generating a realistic image from the objects and captions. Under this setting, objects explicitly define the critical roles in the targeted images and captions implicitly describe their rich attributes and connections. Correspondingly, a MOC-GAN is proposed to mix the inputs of two modalities to generate realistic images. It firstly infers the implicit relations between object pairs from the captions to build a hidden-state scene graph. So a multi-layer representation containing objects, relations and captions is constructed, where the scene graph provides the structures of the scene and the caption provides the image-level guidance. Then a cascaded attentive generative network is designed to coarse-to-fine generate phrase patch by paying attention to the most relevant words in the caption. In addition, a phrase-wise DAMSM is proposed to better supervise the fine-grained phrase-patch consistency. On COCO dataset, our method outperforms the state-of-the-art methods on both Inception Score and FID while maintaining high visual quality. Extensive experiments demonstrate the unique features of our proposed method. | ['Yikang Li', 'Tao Ma'] | 2021-06-06 | null | null | null | null | ['implicit-relations'] | ['natural-language-processing'] | [ 4.14306402e-01 3.53267550e-01 -1.35639578e-01 -5.85954309e-01
-6.63526595e-01 -3.25007468e-01 8.26258898e-01 -3.94994944e-01
1.12745576e-01 7.57551491e-01 3.83955508e-01 2.09232062e-01
1.67851835e-01 -9.00161207e-01 -1.16319621e+00 -8.22738945e-01
5.53300977e-01 5.78938305e-01 7.42734596e-02 -1.81221649e-01
-1.35312423e-01 -1.39753804e-01 -1.35969877e+00 5.77413857e-01
1.01676679e+00 9.23418343e-01 8.42597902e-01 2.64688790e-01
-2.68176913e-01 9.37682450e-01 -3.25289339e-01 -6.86451852e-01
-7.56471232e-02 -8.30122530e-01 -6.26638353e-01 7.61917353e-01
4.03416336e-01 -3.26736659e-01 -3.68382275e-01 1.16648054e+00
1.72578931e-01 1.06276371e-01 6.81227803e-01 -1.26853442e+00
-1.37978709e+00 8.75706673e-01 -7.04852700e-01 -3.08793068e-01
3.45024109e-01 5.67310333e-01 1.16071796e+00 -9.26959872e-01
5.73220432e-01 1.45641112e+00 9.88136604e-03 7.47976184e-01
-1.37764776e+00 -6.01807594e-01 5.60814381e-01 1.28771529e-01
-1.37771201e+00 -3.65884751e-01 9.93323624e-01 -3.52747560e-01
4.07050520e-01 1.83091789e-01 6.52339458e-01 1.36591482e+00
-2.14206055e-01 9.01000738e-01 9.00135458e-01 -2.05278516e-01
4.09697220e-02 3.75770688e-01 -3.72571677e-01 6.83625579e-01
9.12251845e-02 -1.22656457e-01 -3.30553025e-01 2.93610185e-01
1.05653787e+00 1.80809811e-01 -3.02309155e-01 -2.32564881e-01
-1.34183753e+00 7.44909167e-01 7.55787373e-01 1.74003944e-01
-6.39617741e-01 3.31749320e-01 -8.98866430e-02 -3.52048635e-01
9.84293744e-02 3.55645955e-01 -1.32168368e-01 4.06233847e-01
-6.49337113e-01 2.46812224e-01 2.79043347e-01 1.51130521e+00
9.10678387e-01 9.86471623e-02 -7.51213074e-01 7.82922626e-01
6.45470679e-01 5.94388068e-01 -8.85051582e-03 -8.73785794e-01
8.17606091e-01 6.40893221e-01 6.47818297e-02 -1.15490198e+00
2.55359679e-01 -7.03899741e-01 -1.17323971e+00 -4.46675479e-01
-6.92724660e-02 3.13763767e-02 -1.27253056e+00 2.01206422e+00
2.31677562e-01 3.10598433e-01 1.47285417e-01 1.10314000e+00
9.85084057e-01 1.13261151e+00 5.41047454e-01 -1.78739637e-01
1.54875052e+00 -1.29849696e+00 -8.77910018e-01 -6.38402998e-01
4.73891236e-02 -5.66488981e-01 1.11772919e+00 -4.98834886e-02
-1.21699786e+00 -8.38769615e-01 -7.65305936e-01 -1.80975974e-01
-3.36651728e-02 1.70607999e-01 5.69701016e-01 1.20760240e-01
-8.23671460e-01 -1.15088969e-01 -4.12894696e-01 -8.98556411e-03
8.18411648e-01 7.24182948e-02 -1.16062820e-01 -4.70117092e-01
-1.26169896e+00 6.01798296e-01 7.24831283e-01 3.89519423e-01
-1.42514813e+00 -4.23938513e-01 -1.17843103e+00 2.34570131e-01
5.49830914e-01 -1.24477744e+00 9.59720373e-01 -1.11387944e+00
-1.18939197e+00 8.08890462e-01 -1.08644396e-01 -6.69611990e-02
3.36401522e-01 -2.02917922e-02 -2.59248793e-01 1.13420300e-01
3.36770386e-01 1.21712506e+00 9.55582619e-01 -1.92131352e+00
-6.20507896e-01 4.31297310e-02 3.92402023e-01 5.17293155e-01
-6.24492653e-02 -1.79547966e-01 -1.13183391e+00 -7.14214861e-01
3.31942784e-03 -6.89870298e-01 -3.40565175e-01 -2.28733063e-01
-8.39776337e-01 -1.13755213e-02 6.39088690e-01 -5.09150922e-01
9.89544570e-01 -2.18004537e+00 3.91476125e-01 -1.87225029e-01
2.36824751e-01 -7.21212849e-02 -3.45009565e-01 2.81709224e-01
-1.94562495e-01 1.55828908e-01 -5.56914151e-01 -6.33388102e-01
1.42361373e-01 5.34842432e-01 -5.45155168e-01 5.06922491e-02
5.69585741e-01 1.41829753e+00 -1.01856625e+00 -7.85665691e-01
2.42865101e-01 6.00366533e-01 -4.88597572e-01 6.74066067e-01
-7.30778694e-01 5.90901017e-01 -6.84817493e-01 4.42699134e-01
6.60330534e-01 -7.50607014e-01 -1.48383364e-01 -5.16575038e-01
3.25432390e-01 6.28245845e-02 -9.25122261e-01 1.81160736e+00
-5.23961008e-01 1.99120805e-01 -1.35923743e-01 -8.92339051e-01
8.73597682e-01 1.69939741e-01 8.38002469e-03 -5.91571867e-01
6.27096668e-02 -1.65867090e-01 -1.32209122e-01 -5.99683642e-01
2.44030058e-01 -3.73909682e-01 -2.96886742e-01 1.68992281e-01
1.53046027e-01 -3.89908731e-01 1.31233722e-01 5.95691860e-01
4.13047671e-01 4.47410524e-01 5.29461689e-02 1.32640656e-02
5.62866986e-01 -1.26376495e-01 5.63599408e-01 5.27523041e-01
2.63643026e-01 1.09738481e+00 4.84681547e-01 -1.08273372e-01
-1.02752137e+00 -1.10354674e+00 1.95290193e-01 9.32791054e-01
6.21987104e-01 -3.94344538e-01 -7.60915399e-01 -6.51942492e-01
-4.80730146e-01 9.42528129e-01 -9.19493377e-01 -2.46357352e-01
-3.87783855e-01 -5.90410233e-01 4.38667201e-02 5.35988629e-01
7.15148330e-01 -1.44168580e+00 -1.20582685e-01 3.54326852e-02
-4.54270869e-01 -1.29781437e+00 -7.80720651e-01 -1.79160371e-01
-4.42957014e-01 -6.89577222e-01 -7.38161504e-01 -1.17258477e+00
1.20970690e+00 1.11097142e-01 1.36045527e+00 3.54710333e-02
2.03303263e-01 5.87435104e-02 -3.85486037e-01 -1.42324507e-01
-3.64896983e-01 -2.30618820e-01 -2.88678288e-01 4.70973462e-01
-5.32259382e-02 -5.77344060e-01 -6.47440195e-01 2.72875249e-01
-1.10243046e+00 9.40434754e-01 1.13711190e+00 9.00554180e-01
8.52427304e-01 9.54685211e-02 4.79760885e-01 -1.25231433e+00
1.06711581e-01 -6.58855081e-01 -3.51514131e-01 4.01452631e-01
-2.13694990e-01 1.24323986e-01 7.99009204e-01 -4.18886483e-01
-1.41326571e+00 2.84703463e-01 1.19319730e-01 -6.34276390e-01
-2.86538720e-01 3.83651733e-01 -8.94237459e-01 5.91045737e-01
2.46459693e-01 7.34249532e-01 -4.26983893e-01 -1.36422768e-01
8.02214026e-01 2.84311920e-01 7.17314422e-01 -9.47843134e-01
1.00741124e+00 3.34325016e-01 -3.02628487e-01 -3.19361180e-01
-1.30999362e+00 8.47676471e-02 -4.50911939e-01 -1.33528367e-01
1.18441904e+00 -1.14946449e+00 -2.39325687e-01 3.80288422e-01
-1.49942875e+00 -2.05941081e-01 -3.43784451e-01 1.12070449e-01
-6.92594290e-01 -7.83705115e-02 -4.98378187e-01 -6.31625056e-01
-7.93514401e-02 -1.25486982e+00 1.43670869e+00 5.28014302e-01
2.97696680e-01 -7.84383297e-01 -3.81820202e-01 5.75516522e-01
1.22706629e-01 4.24579591e-01 9.46581066e-01 -4.00803983e-02
-1.13560593e+00 2.12989822e-01 -6.82746828e-01 3.02058995e-01
2.40847364e-01 -1.25040635e-01 -9.09782469e-01 1.82059202e-02
-4.93837148e-02 -3.82791966e-01 7.56884158e-01 3.11587721e-01
1.42408621e+00 -6.04862273e-01 -3.03393543e-01 5.90551496e-01
1.49762845e+00 1.39452487e-01 7.75771439e-01 -1.80707157e-01
1.22271705e+00 7.88238764e-01 4.67454433e-01 1.83493063e-01
5.52430987e-01 5.15981317e-01 6.71072006e-01 -3.68771702e-01
-3.30923766e-01 -8.76510978e-01 2.57666677e-01 8.12552929e-01
8.20303559e-02 -5.69751143e-01 -5.40348232e-01 5.95775485e-01
-1.97573411e+00 -9.40139055e-01 -1.38041764e-01 1.64659262e+00
1.04883754e+00 1.01367868e-01 -2.59446293e-01 -5.28172255e-01
1.09794104e+00 3.40491951e-01 -6.65632963e-01 2.85862982e-01
-3.64718318e-01 -3.08985025e-01 -8.98792669e-02 2.77174950e-01
-8.42959642e-01 1.13786721e+00 5.16092873e+00 9.53881860e-01
-8.09494674e-01 6.38883412e-02 1.18679202e+00 8.39044601e-02
-7.60465086e-01 1.61875233e-01 -7.18452811e-01 7.20455885e-01
3.20661604e-01 2.93809883e-02 3.35753858e-01 6.83558345e-01
1.47916853e-01 2.58349657e-01 -1.12165034e+00 1.06953204e+00
2.48891890e-01 -1.47314370e+00 7.66061425e-01 1.13470800e-01
1.03963685e+00 -4.59636182e-01 2.59462208e-01 4.49604750e-01
4.01064157e-01 -1.22592998e+00 1.07405293e+00 7.80848026e-01
1.11032915e+00 -5.96186578e-01 5.76373160e-01 3.94694626e-01
-1.27937663e+00 1.75829917e-01 -3.55746359e-01 3.82404774e-02
5.13769209e-01 4.28163767e-01 -3.99342149e-01 6.45081222e-01
5.35855114e-01 8.98243845e-01 -5.26691496e-01 5.78312159e-01
-6.84791684e-01 5.27994275e-01 3.01217549e-02 8.77670944e-02
5.22209048e-01 -3.75661165e-01 2.38850191e-01 1.00230169e+00
1.57522038e-01 4.62955713e-01 1.76878482e-01 1.59536064e+00
-2.52166063e-01 -1.38955832e-01 -4.77340579e-01 -5.46233952e-02
3.69886935e-01 1.46998000e+00 -5.78227043e-01 -5.04335821e-01
-4.19870645e-01 1.00454307e+00 3.43309075e-01 6.65090501e-01
-1.13661695e+00 -2.17288397e-02 2.77034193e-01 1.10688679e-01
4.60289061e-01 1.10799879e-01 -1.85423538e-01 -1.27367544e+00
1.02953114e-01 -8.72560143e-01 2.60125518e-01 -1.30983865e+00
-1.53560328e+00 7.51325071e-01 8.45189318e-02 -1.28824592e+00
-8.65233168e-02 -3.00354570e-01 -6.72030807e-01 1.07211554e+00
-1.29831076e+00 -1.63133407e+00 -4.97954905e-01 5.51123023e-01
7.75613844e-01 1.31531790e-01 4.96007711e-01 1.96308181e-01
-7.09988058e-01 3.39273393e-01 -4.80067402e-01 2.34594196e-01
3.47652137e-01 -1.15325344e+00 1.99597806e-01 9.96879220e-01
2.96554476e-01 6.91335201e-01 7.08213449e-01 -6.94366992e-01
-1.25840211e+00 -1.48085284e+00 4.74177092e-01 -5.19334435e-01
4.09588665e-01 -7.23701656e-01 -9.03609693e-01 6.81145310e-01
5.38231790e-01 7.96302482e-02 3.68463755e-01 -2.78374255e-01
-3.89170945e-01 -1.24610245e-01 -7.64320314e-01 7.33802795e-01
1.17486906e+00 -4.63037223e-01 -6.91469491e-01 4.57696527e-01
1.27881265e+00 -4.57270116e-01 -3.63678008e-01 2.15312228e-01
1.13899261e-01 -7.79855132e-01 8.30172122e-01 -6.16402805e-01
1.08252728e+00 -6.81842625e-01 -2.33795956e-01 -1.22060609e+00
-6.05150700e-01 -5.17092884e-01 -1.34321019e-01 1.68987346e+00
5.94666362e-01 -8.08174908e-02 6.34530246e-01 6.50280476e-01
-3.43023568e-01 -7.59705842e-01 -2.80470848e-01 -3.22247475e-01
-2.45909542e-01 -3.22975904e-01 7.43445456e-01 9.49779451e-01
-4.27612036e-01 1.06043530e+00 -6.25483751e-01 2.66139477e-01
6.55057788e-01 4.70025957e-01 6.08244240e-01 -6.92587256e-01
-5.78113556e-01 -2.63382524e-01 -9.72496942e-02 -1.42545414e+00
1.20636523e-01 -7.28435397e-01 3.31474841e-01 -1.92848110e+00
7.25504577e-01 -5.48787832e-01 -2.64565974e-01 5.80445528e-01
-6.11621320e-01 3.74366820e-01 2.69461751e-01 1.96144566e-01
-9.02343512e-01 9.92502749e-01 1.95728505e+00 -4.24148828e-01
1.57258108e-01 -3.53592664e-01 -1.16149652e+00 4.58401799e-01
3.50424767e-01 -3.68600041e-01 -7.81277776e-01 -8.21876943e-01
2.57523537e-01 1.50534093e-01 6.10735893e-01 -6.51409388e-01
1.58493310e-01 -4.84940737e-01 5.47161579e-01 -5.53993046e-01
4.92411643e-01 -7.68081367e-01 4.11679447e-01 5.98734617e-02
-4.83031660e-01 -2.67290622e-01 -1.28912047e-01 8.29897046e-01
-3.61177891e-01 2.04925351e-02 6.71769142e-01 -3.28255832e-01
-6.67597175e-01 8.10073376e-01 3.53745878e-01 1.24202043e-01
1.11520255e+00 -2.63019074e-02 -3.50204349e-01 -5.08684695e-01
-6.13239765e-01 4.95674342e-01 5.01245737e-01 5.68330526e-01
7.81976640e-01 -1.63576341e+00 -9.72690344e-01 1.43774927e-01
3.54115158e-01 7.77306378e-01 6.52913511e-01 4.49227393e-01
-1.99770480e-01 2.83728361e-01 -1.99241906e-01 -6.99566424e-01
-6.41385138e-01 7.87800014e-01 1.92111656e-01 -1.43854704e-03
-4.56313133e-01 1.13252127e+00 1.26001060e+00 -7.93454349e-02
-5.85574948e-04 -1.37402445e-01 -3.06572169e-01 -1.26967341e-01
4.63956118e-01 -4.67229903e-01 -6.63075149e-01 -9.88343775e-01
-9.13457274e-02 5.07447958e-01 -1.92979246e-01 -8.96552578e-02
1.30884194e+00 -3.22411180e-01 -2.38363460e-01 3.08778971e-01
9.98803318e-01 -2.03754067e-01 -1.69941962e+00 -2.12276533e-01
-6.81992412e-01 -4.31856394e-01 -1.07925788e-01 -8.13707173e-01
-1.26229692e+00 8.80745888e-01 2.22745091e-02 5.09538017e-02
1.32311797e+00 4.86395895e-01 7.69143283e-01 -1.10836908e-01
1.88006833e-01 -5.49084067e-01 3.61891061e-01 1.83810443e-01
1.09448266e+00 -1.22261548e+00 -2.27723479e-01 -7.79769361e-01
-9.83617246e-01 5.36590457e-01 1.16595495e+00 -1.17211200e-01
1.12580478e-01 -1.19407341e-01 -1.89528316e-01 -3.42731237e-01
-8.43108833e-01 -2.47131661e-01 4.65881705e-01 7.34422028e-01
1.19320258e-01 1.25203967e-01 9.66665596e-02 9.39604878e-01
-3.09594125e-02 -3.46783429e-01 2.74899930e-01 3.12761396e-01
-9.89279225e-02 -7.64596581e-01 -2.06766829e-01 3.54254454e-01
-7.28934333e-02 -4.40850049e-01 -3.33296508e-01 4.96106803e-01
5.01871228e-01 7.59773493e-01 5.81591055e-02 -3.43132377e-01
1.80745289e-01 -3.05137694e-01 3.44187677e-01 -9.61129248e-01
-1.11440040e-01 1.47954106e-01 -1.32446468e-01 -3.97545546e-01
-5.63701570e-01 -4.66440022e-01 -1.19653499e+00 2.14095652e-01
-3.09184998e-01 2.86531389e-01 2.15026125e-01 1.08127105e+00
3.16910207e-01 9.00061250e-01 6.71090961e-01 -7.58131444e-01
-2.58948244e-02 -9.76754248e-01 -4.22469646e-01 7.01108932e-01
1.82046250e-01 -5.91457725e-01 -2.19457820e-01 6.61372960e-01] | [10.870280265808105, 0.8913512825965881] |
db4065fa-c0b6-4f81-afbd-e82baa38d79d | icassp-2021-acoustic-echo-cancellation | 2009.04972 | null | https://arxiv.org/abs/2009.04972v2 | https://arxiv.org/pdf/2009.04972v2.pdf | ICASSP 2021 Acoustic Echo Cancellation Challenge: Datasets and Testing Framework | The ICASSP 2021 Acoustic Echo Cancellation Challenge is intended to stimulate research in the area of acoustic echo cancellation (AEC), which is an important part of speech enhancement and still a top issue in audio communication and conferencing systems. Many recent AEC studies report reasonable performance on synthetic datasets where the train and test samples come from the same underlying distribution. However, the AEC performance often degrades significantly on real recordings. Also, most of the conventional objective metrics such as echo return loss enhancement (ERLE) and perceptual evaluation of speech quality (PESQ) do not correlate well with subjective speech quality tests in the presence of background noise and reverberation found in realistic environments. In this challenge, we open source two large datasets to train AEC models under both single talk and double talk scenarios. These datasets consist of recordings from more than 2,500 real audio devices and human speakers in real environments, as well as a synthetic dataset. We open source an online subjective test framework based on ITU-T P.808 for researchers to quickly test their results. The winners of this challenge will be selected based on the average P.808 Mean Opinion Score (MOS) achieved across all different single talk and double talk scenarios. | ['Ross Cutler', 'Kusha Sridhar', 'Sebastian Braun', 'Hannes Gamper', 'Sriram Srinivasan', 'Robert Aichner', 'Ando Saabas', 'Tanel Parnamaa'] | 2020-09-10 | null | null | null | null | ['acoustic-echo-cancellation', 'acoustic-echo-cancellation'] | ['medical', 'speech'] | [ 2.84620430e-02 -5.52739501e-01 8.37351501e-01 -3.06871623e-01
-1.23212934e+00 -5.22115648e-01 2.69694507e-01 6.45450577e-02
-4.06667352e-01 4.38738197e-01 5.28306067e-01 -2.95783073e-01
5.94601082e-03 -1.53561220e-01 -4.45490241e-01 -6.19181216e-01
-4.24376309e-01 -1.57622620e-01 2.92798728e-01 -3.29961807e-01
-4.58198786e-03 8.45841989e-02 -1.78671455e+00 3.40108544e-01
9.16173100e-01 1.03782272e+00 2.51200765e-01 1.16955090e+00
2.62981743e-01 2.22212061e-01 -1.23952925e+00 -4.11221534e-01
1.18824609e-01 -4.50795740e-01 -6.56195059e-02 -3.37985665e-01
3.55806142e-01 -2.15449426e-02 -3.02363366e-01 1.23698890e+00
1.43103397e+00 2.17812166e-01 3.70979398e-01 -9.57199097e-01
5.74431801e-03 4.69172269e-01 -1.12492487e-01 3.12672436e-01
5.30735850e-01 1.96945935e-01 7.70073593e-01 -1.12489712e+00
-1.57508235e-02 1.06800854e+00 9.06453252e-01 3.74700159e-01
-8.25532377e-01 -9.89615560e-01 -3.40555161e-01 5.30291915e-01
-1.19475830e+00 -1.00970435e+00 7.07239330e-01 4.71748970e-03
7.90085971e-01 5.92627764e-01 5.70272684e-01 1.13744688e+00
-3.24161872e-02 6.93851829e-01 1.22840285e+00 -5.29870808e-01
4.47361946e-01 2.90654182e-01 -5.87563477e-02 -1.24357022e-01
-2.39617035e-01 5.42143583e-01 -8.77094626e-01 -1.08378530e-01
4.25483175e-02 -7.73223281e-01 -7.89250493e-01 2.90755242e-01
-1.03611326e+00 2.61878610e-01 -8.18540454e-02 4.40296024e-01
-2.92205065e-01 -1.10489435e-01 4.11696792e-01 7.75526226e-01
4.52122986e-01 3.27924073e-01 -3.62012178e-01 -5.42041481e-01
-9.12190020e-01 2.59202838e-01 1.14140677e+00 7.71799326e-01
1.63859189e-01 6.09742939e-01 -1.90132394e-01 1.39243364e+00
4.26266015e-01 1.06384301e+00 4.31868285e-01 -7.99933255e-01
6.65704727e-01 -4.82495040e-01 2.70428509e-01 -9.98513401e-01
-1.01005726e-01 -9.37004983e-01 -7.24044085e-01 1.30509064e-01
1.01730958e-01 -5.14779031e-01 -6.49058461e-01 1.62500322e+00
1.12787429e-02 4.75219220e-01 2.18629882e-01 9.11123514e-01
8.92321885e-01 9.04690802e-01 -2.57605851e-01 -2.97850907e-01
1.05281138e+00 -7.48345196e-01 -1.26345122e+00 -2.48834565e-01
2.41474837e-01 -1.28268492e+00 9.85539556e-01 8.56426120e-01
-1.20512235e+00 -6.98947191e-01 -1.25373149e+00 6.43426776e-01
-1.74768195e-01 -2.02242866e-01 7.22835660e-02 1.40244257e+00
-1.20118237e+00 2.00238481e-01 -3.78123969e-01 2.56274372e-01
-2.60944903e-01 1.34480342e-01 -2.52769738e-01 -1.08162098e-01
-1.43813181e+00 5.59348166e-01 -3.56121033e-01 3.65369827e-01
-1.03086257e+00 -9.05056417e-01 -6.20199680e-01 7.18139112e-02
2.41720267e-02 1.66644588e-01 1.54454243e+00 -6.34425819e-01
-1.66202390e+00 3.06517959e-01 -1.10923648e-01 -4.40543443e-01
4.38313246e-01 -5.02863705e-01 -1.39405119e+00 -1.56320319e-01
-2.90920526e-01 6.40938133e-02 8.40740561e-01 -1.28685963e+00
-6.73679471e-01 -5.73619753e-02 -6.42261982e-01 2.99612910e-01
-2.58097827e-01 2.37030730e-01 -4.34372783e-01 -7.84090459e-01
6.47736788e-02 -6.33816123e-01 1.04447588e-01 -5.07894754e-01
-2.87176579e-01 3.60791504e-01 6.95325613e-01 -1.08457279e+00
1.35388517e+00 -2.43332624e+00 -5.16379774e-01 3.22571665e-01
-3.52715403e-01 8.35917354e-01 -2.88864613e-01 3.51295739e-01
-4.65304255e-02 -1.40147835e-01 -1.80909857e-01 -4.60860163e-01
1.27506182e-01 -3.57432812e-01 -2.62327760e-01 3.16837907e-01
-1.62424490e-01 1.32813066e-01 -7.98218131e-01 -8.90719071e-02
2.73809403e-01 6.68366671e-01 -6.51159167e-01 4.72567230e-01
4.26886708e-01 4.83897060e-01 1.10962771e-01 3.38486135e-01
1.07609355e+00 7.30966568e-01 -1.52816370e-01 -9.83036309e-02
-1.53749883e-01 4.63467211e-01 -1.66395581e+00 1.34228730e+00
-8.17824006e-01 1.07255495e+00 6.87445402e-01 -5.77522337e-01
1.02430046e+00 9.10713375e-01 6.68249130e-02 -9.84885395e-01
-8.30415338e-02 7.57557750e-01 3.50038052e-01 -3.97880852e-01
3.93970877e-01 -1.38455346e-01 3.42466414e-01 1.48339450e-01
2.55016476e-01 -5.63802481e-01 -6.85864091e-02 6.77049253e-03
1.22358787e+00 -5.95708370e-01 -2.68423051e-01 -2.76880383e-01
7.44761586e-01 -9.32258427e-01 4.56444770e-01 6.93613052e-01
-5.42627692e-01 8.45977306e-01 -6.85957819e-02 3.56755793e-01
-8.27334166e-01 -1.28580296e+00 -2.33901978e-01 8.38869512e-01
-1.21700518e-01 -3.52066904e-01 -8.75756443e-01 8.08880851e-02
-4.71627325e-01 9.01085913e-01 2.00274847e-02 -1.21967070e-01
-4.38325793e-01 -5.00595808e-01 9.61714506e-01 -1.92305036e-02
5.70563316e-01 -1.06098318e+00 -8.45755488e-02 4.99983460e-01
-4.89268243e-01 -1.24355078e+00 -6.84993327e-01 1.60611942e-01
-2.71480560e-01 -5.81937551e-01 -1.14329696e+00 -8.03451777e-01
-1.95223335e-02 1.89187557e-01 1.04969895e+00 -2.31047124e-01
-1.16419666e-01 5.74932694e-01 -5.44542611e-01 -7.34600782e-01
-7.65955210e-01 -3.94279838e-01 3.97392273e-01 1.67114303e-01
8.32679719e-02 -5.85059941e-01 -8.20402324e-01 8.41732681e-01
-6.29578114e-01 -4.54898775e-01 2.45162293e-01 5.77263415e-01
1.56292930e-01 4.26194131e-01 8.46021354e-01 -1.53361678e-01
1.09752345e+00 -1.82156786e-01 -4.56016332e-01 1.74832225e-01
-3.45219821e-01 -4.21709985e-01 4.72027212e-01 -4.40046459e-01
-1.38093996e+00 -5.70331812e-01 -8.88514698e-01 -5.53596579e-02
-1.81592360e-01 2.88751364e-01 -5.05263269e-01 1.30549848e-01
6.29801214e-01 2.91260034e-01 -4.50645268e-01 -6.59093142e-01
-1.65339142e-01 1.29030025e+00 8.02913010e-01 -2.67435670e-01
6.67496622e-01 -9.69668105e-02 -3.69391382e-01 -1.28710639e+00
-3.26567352e-01 -6.47867680e-01 1.50527477e-01 -4.39633936e-01
3.08731437e-01 -8.95363271e-01 -5.09140193e-01 9.92784560e-01
-1.05519021e+00 -2.98560202e-01 5.98311499e-02 1.10384250e+00
-2.32396334e-01 2.45529518e-01 -4.57556695e-01 -1.35325599e+00
-4.88973916e-01 -1.32357955e+00 7.17663467e-01 1.94228187e-01
-3.77535112e-02 -7.31383741e-01 3.19190621e-01 1.10495865e-01
1.04913735e+00 -3.76835704e-01 3.82856190e-01 -4.50825721e-01
3.91951203e-02 -4.09888238e-01 2.91136503e-01 9.17821765e-01
1.91048205e-01 -2.62191087e-01 -1.45603013e+00 -4.02704209e-01
3.40616584e-01 -9.49866995e-02 4.88092035e-01 4.73972201e-01
7.88069308e-01 2.20123734e-02 1.32927060e-01 4.10438150e-01
1.01884055e+00 6.01263583e-01 1.03234446e+00 -2.40469828e-01
-4.55583669e-02 7.15919435e-01 4.83019114e-01 4.18483704e-01
-2.16075942e-01 7.78954566e-01 1.85109168e-01 -8.48162249e-02
-4.27035570e-01 -1.75513864e-01 6.16488695e-01 1.53635228e+00
2.42107004e-01 -6.88608408e-01 -8.08629990e-01 4.58319217e-01
-7.93279350e-01 -8.36987257e-01 -3.29747677e-01 2.49972868e+00
9.74001586e-01 5.97236790e-02 -2.36341760e-01 7.64472902e-01
7.76943982e-01 -1.87497959e-02 -1.25747070e-01 -6.11769736e-01
-4.35485125e-01 6.11201763e-01 3.53290170e-01 8.03804696e-01
-6.58578515e-01 3.52315485e-01 6.20573044e+00 1.03171802e+00
-1.30299473e+00 1.47721887e-01 5.52519798e-01 5.08862287e-02
-2.21018136e-01 -4.79854107e-01 -5.67758679e-01 3.91570508e-01
1.55333102e+00 -3.05512361e-02 4.55775529e-01 3.53850901e-01
4.93522346e-01 -8.93774256e-02 -7.82459915e-01 1.24163723e+00
8.91939774e-02 -6.54564679e-01 -6.11894548e-01 -6.10226691e-02
6.28584146e-01 1.26429111e-01 4.89316434e-01 5.45315266e-01
-7.81769231e-02 -9.85926151e-01 7.34856308e-01 2.88178027e-01
8.60654771e-01 -5.24588883e-01 8.82093072e-01 1.79359168e-01
-1.15798688e+00 1.31592646e-01 -1.68495819e-01 4.05513197e-01
3.49786550e-01 7.42657483e-01 -7.72165239e-01 4.19161886e-01
8.94747853e-01 1.07133254e-01 -1.43457845e-01 1.75576711e+00
-1.44941360e-01 1.26012373e+00 -4.14625198e-01 1.64581891e-02
-4.33258265e-02 1.79518163e-01 9.18329358e-01 1.53658521e+00
6.65352046e-01 9.54151526e-02 -7.02236950e-01 2.66485959e-01
-4.38172929e-02 1.99301749e-01 -2.54752368e-01 1.97114170e-01
6.45092189e-01 8.09361935e-01 -9.50958133e-02 2.71765776e-02
-2.84111768e-01 8.14593852e-01 -7.38208354e-01 6.78262532e-01
-8.40822458e-01 -8.15538824e-01 6.89123273e-01 -7.19477534e-02
1.44909665e-01 -1.47335812e-01 1.63602352e-01 -7.25083172e-01
1.16233706e-01 -1.40758932e+00 -3.04008514e-01 -7.58384764e-01
-1.05910790e+00 8.28897953e-01 -2.68553942e-01 -1.45626020e+00
-1.43877327e-01 -6.13689244e-01 -8.97372663e-01 1.27045202e+00
-1.38839376e+00 -7.81167373e-02 -1.43483832e-01 5.30789554e-01
7.44441390e-01 -2.81652361e-01 9.83337998e-01 1.06594312e+00
-2.29435667e-01 1.03180337e+00 4.84015286e-01 -9.87226740e-02
9.41635013e-01 -1.08910334e+00 4.48238075e-01 7.15507269e-01
1.04638137e-01 4.13527787e-01 1.29418707e+00 -1.93342924e-01
-1.10114157e+00 -8.00583124e-01 7.23558664e-01 1.60343815e-02
3.89844775e-01 -5.94391286e-01 -1.06848514e+00 -2.13281676e-01
4.79396939e-01 -1.37789369e-01 7.66453087e-01 -4.84713949e-02
-2.35449001e-02 -4.21022594e-01 -9.93844628e-01 4.12723452e-01
7.16939092e-01 -5.42057872e-01 -2.08118871e-01 -9.82428566e-02
5.90295374e-01 -3.52735430e-01 -6.51344121e-01 5.76186657e-01
5.01706660e-01 -1.18434501e+00 8.23904634e-01 2.32657697e-02
-7.25820959e-02 -2.51073748e-01 -6.24149203e-01 -1.81540012e+00
3.03158790e-01 -1.06156743e+00 1.33513898e-01 1.61854684e+00
8.22821379e-01 -6.71842217e-01 3.48382682e-01 1.05940394e-01
-4.48987305e-01 -2.19331667e-01 -9.41599250e-01 -9.95778441e-01
-7.49808624e-02 -1.11656666e+00 5.16712427e-01 3.40821236e-01
-1.69453755e-01 1.05204903e-01 -5.24120986e-01 3.52425039e-01
6.73298717e-01 -8.06990564e-01 6.90286696e-01 -8.43032062e-01
-5.46951473e-01 -2.34668851e-01 -3.57303530e-01 -1.08054340e+00
-2.83855885e-01 -3.08118761e-01 5.70954442e-01 -1.28173411e+00
-5.71628511e-01 -4.10519093e-01 -4.70331430e-01 -3.87192607e-01
-1.22411445e-01 1.40200689e-01 1.07591718e-01 -2.42021784e-01
-1.74682438e-01 8.53428364e-01 1.03473830e+00 -3.27292174e-01
-2.64096379e-01 7.01274395e-01 -3.57816219e-01 5.50965130e-01
7.45793402e-01 -5.10661483e-01 -4.47567493e-01 -4.39707637e-01
-8.64977241e-02 4.00597095e-01 8.44974741e-02 -1.55859923e+00
3.38701636e-01 4.09929633e-01 -8.91205072e-02 -5.94892085e-01
6.92500234e-01 -8.10218036e-01 4.93104123e-02 2.84449756e-01
-4.10310864e-01 -2.53823012e-01 4.51116353e-01 5.32170296e-01
-6.67023063e-01 -1.02859840e-01 7.84035027e-01 2.82985896e-01
-3.72531444e-01 -1.13343611e-01 -4.97215599e-01 1.49675459e-01
3.40444028e-01 -6.65524527e-02 -6.51012510e-02 -1.13260031e+00
-5.83448589e-01 -5.29883392e-02 -3.13004702e-01 4.84054893e-01
7.21533179e-01 -1.01178312e+00 -1.14141119e+00 3.07496578e-01
4.03767601e-02 -5.93155622e-01 5.52595973e-01 7.53927827e-01
-2.47966915e-01 4.70875055e-01 1.11748472e-01 -6.46754265e-01
-1.35471487e+00 -1.26691952e-01 7.11544573e-01 1.83401611e-02
-3.12052429e-01 9.88179386e-01 1.94248080e-01 -3.15146387e-01
7.30179310e-01 -6.66708425e-02 -1.45466730e-01 -4.35573846e-01
9.20638204e-01 5.13147831e-01 6.72108114e-01 -6.21126115e-01
-2.05104128e-01 2.34650761e-01 3.00864458e-01 -8.76023769e-01
9.66765642e-01 -2.62595654e-01 3.78964543e-01 3.83398443e-01
1.37277043e+00 6.49426162e-01 -8.44219863e-01 -2.22918361e-01
-2.04057768e-01 -5.71051240e-01 2.69497871e-01 -1.14088154e+00
-9.40507710e-01 1.11109793e+00 1.39277339e+00 1.88896909e-01
1.33083403e+00 -4.95500416e-01 7.66712546e-01 1.67145625e-01
3.78513485e-01 -1.18126345e+00 1.76600069e-01 6.11189187e-01
1.17025900e+00 -9.84971941e-01 -6.84273303e-01 -2.16695815e-01
-5.99135160e-01 7.02478886e-01 3.48576128e-01 2.95127392e-01
1.00350893e+00 6.60350084e-01 5.88655889e-01 3.05422574e-01
-5.71081519e-01 8.39182884e-02 2.47197285e-01 9.21450555e-01
7.85879493e-01 1.03920758e-01 -1.51849210e-01 6.31996214e-01
-6.14952981e-01 -4.67429608e-01 3.02927613e-01 5.43559432e-01
-5.11310697e-01 -1.11287367e+00 -6.81266487e-01 2.63354063e-01
-7.52426505e-01 -2.76159972e-01 -1.25736892e-01 2.06413940e-01
-1.50223598e-01 1.72078109e+00 -9.44802985e-02 -5.50135911e-01
7.99877226e-01 -7.32737929e-02 8.69379938e-02 -2.08542362e-01
-7.09474027e-01 3.38191479e-01 2.68342048e-01 -8.36028308e-02
-1.54752105e-01 -4.87282783e-01 -9.52131212e-01 -2.86540031e-01
-5.31893730e-01 5.83588123e-01 1.18682170e+00 5.50950587e-01
6.58713207e-02 8.42805564e-01 8.94771039e-01 -7.09901869e-01
-6.46574259e-01 -1.36969578e+00 -9.26669061e-01 3.71382445e-01
4.70140368e-01 -3.01525235e-01 -6.94872737e-01 -1.45992786e-01] | [14.986717224121094, 5.916009426116943] |
20b6888e-bc69-41d4-ae15-e6b45968d380 | bayesian-optimization-with-conformal-coverage | 2210.12496 | null | https://arxiv.org/abs/2210.12496v3 | https://arxiv.org/pdf/2210.12496v3.pdf | Bayesian Optimization with Conformal Prediction Sets | Bayesian optimization is a coherent, ubiquitous approach to decision-making under uncertainty, with applications including multi-arm bandits, active learning, and black-box optimization. Bayesian optimization selects decisions (i.e. objective function queries) with maximal expected utility with respect to the posterior distribution of a Bayesian model, which quantifies reducible, epistemic uncertainty about query outcomes. In practice, subjectively implausible outcomes can occur regularly for two reasons: 1) model misspecification and 2) covariate shift. Conformal prediction is an uncertainty quantification method with coverage guarantees even for misspecified models and a simple mechanism to correct for covariate shift. We propose conformal Bayesian optimization, which directs queries towards regions of search space where the model predictions have guaranteed validity, and investigate its behavior on a suite of black-box optimization tasks and tabular ranking tasks. In many cases we find that query coverage can be significantly improved without harming sample-efficiency. | ['Andrew Gordon Wilson', 'Wesley Maddox', 'Samuel Stanton'] | 2022-10-22 | null | null | null | null | ['decision-making-under-uncertainty', 'prediction-intervals', 'decision-making-under-uncertainty'] | ['medical', 'miscellaneous', 'reasoning'] | [ 1.60181180e-01 5.14821231e-01 -9.02751982e-01 -5.73749483e-01
-1.54537034e+00 -9.57157969e-01 3.92856389e-01 3.35678130e-01
-4.19698805e-01 1.11451566e+00 4.25163925e-01 -6.89331055e-01
-8.01929295e-01 -8.19969356e-01 -7.29123592e-01 -4.82179105e-01
2.31395513e-02 1.02763546e+00 -2.38920316e-01 4.55214798e-01
3.57194036e-01 7.77029693e-02 -1.15663016e+00 -6.89525306e-02
1.03724241e+00 1.21514952e+00 -1.66014418e-01 4.69314128e-01
8.64726007e-02 6.25514865e-01 -5.31062603e-01 -8.41234326e-01
1.91946894e-01 2.40994260e-01 -5.97978115e-01 -2.96825618e-01
4.09199577e-03 -4.10189778e-01 8.61391947e-02 1.22166336e+00
3.73321772e-01 2.68782318e-01 9.42824841e-01 -1.03100002e+00
-3.64432454e-01 1.08389771e+00 -5.42861879e-01 9.55480263e-02
4.77928430e-01 1.67555124e-01 1.49688113e+00 -5.91713607e-01
2.71898896e-01 1.74554431e+00 2.67553061e-01 8.66801515e-02
-1.66234457e+00 -5.67451596e-01 2.24252224e-01 -2.26045832e-01
-1.15491748e+00 -6.42425358e-01 3.57840329e-01 -5.87360084e-01
4.35575783e-01 7.90753663e-01 3.51092607e-01 9.71018076e-01
4.09982383e-01 8.80315661e-01 1.05515587e+00 -3.52112353e-01
8.65222275e-01 3.25663805e-01 1.69960946e-01 2.17134774e-01
7.62549937e-01 7.73068070e-01 -8.54786456e-01 -1.00113940e+00
2.88117617e-01 -6.48071915e-02 -2.49556854e-01 -5.17441273e-01
-9.76286411e-01 1.18912959e+00 -1.45214841e-01 -6.07788801e-01
-5.08716643e-01 3.76334190e-01 -1.19438231e-01 1.79732755e-01
7.94906914e-01 7.80258179e-01 -6.36938572e-01 -2.15842202e-01
-7.29346812e-01 5.92984498e-01 1.03332949e+00 8.71268988e-01
5.41170537e-01 -2.44554386e-01 -9.99560714e-01 6.28115475e-01
7.47527897e-01 7.84257352e-01 -4.07093912e-02 -1.14762425e+00
6.45458579e-01 9.42242444e-02 9.25881624e-01 -6.11238837e-01
-1.32221654e-01 -6.18086874e-01 -9.98942852e-02 1.53886750e-01
6.77892268e-01 -4.41527665e-01 -7.08554804e-01 1.73269224e+00
3.67591381e-01 -5.86112738e-01 -1.64215773e-01 8.16229701e-01
7.84015656e-02 5.28110862e-01 2.00830594e-01 -5.46648383e-01
1.18090069e+00 -4.11422253e-01 -7.84818470e-01 -3.83708268e-01
5.85419595e-01 -6.23927414e-01 9.21181262e-01 6.71031058e-01
-1.14544666e+00 3.61152411e-01 -8.46964121e-01 3.21103811e-01
8.32880139e-02 -4.66370285e-01 9.60335672e-01 1.11405540e+00
-5.64412475e-01 4.05231357e-01 -5.61025739e-01 4.22217995e-01
5.51817715e-01 1.82205766e-01 2.66130775e-01 -2.46220380e-02
-1.02575135e+00 8.64515603e-01 3.79451871e-01 -1.20349228e-01
-1.11468875e+00 -9.50382113e-01 -5.28044105e-01 5.69322109e-01
1.13647282e+00 -7.41207540e-01 1.60456073e+00 -6.48567855e-01
-1.38218033e+00 3.96517694e-01 -2.24721372e-01 -7.04496264e-01
7.00451434e-01 -3.73266131e-01 -4.08673324e-02 -4.90427464e-01
4.44562212e-02 2.59835035e-01 1.03010571e+00 -9.17346358e-01
-6.58191502e-01 -6.44199252e-01 7.34331906e-02 3.63262028e-01
1.82679594e-01 2.72806026e-02 1.18541438e-02 -6.78237200e-01
2.07897022e-01 -8.06077659e-01 -4.71383661e-01 -1.21821232e-01
-6.98471844e-01 -3.01338881e-01 -3.79626006e-02 -4.30807710e-01
1.46919239e+00 -1.74946916e+00 -2.27239430e-01 6.72384858e-01
3.98049206e-02 -6.39176309e-01 3.04619998e-01 1.05771460e-01
2.77449369e-01 3.91794235e-01 -1.90484062e-01 -7.87444636e-02
5.98959625e-01 -1.04719140e-01 -6.97579324e-01 5.94815493e-01
-1.72835797e-01 9.06694114e-01 -6.36113763e-01 -1.93702161e-01
-1.77921176e-01 -3.19066346e-01 -7.11245179e-01 9.90486518e-02
-7.94758916e-01 1.07691534e-01 -4.65010643e-01 9.06353295e-01
4.40071583e-01 -3.77170533e-01 1.11775912e-01 3.36506128e-01
1.35851726e-01 4.81890917e-01 -1.33510256e+00 1.02676630e+00
-2.68112004e-01 2.94220924e-01 -4.68865447e-02 -7.53892601e-01
4.36843157e-01 1.52078763e-01 1.41475320e-01 -2.91847020e-01
-1.53220780e-02 2.34804034e-01 -1.97993070e-01 -2.74099618e-01
3.48836184e-01 5.13928160e-02 -3.12413096e-01 6.78678334e-01
-2.69956559e-01 -1.73106745e-01 -2.80154318e-01 -9.18709207e-03
8.23186636e-01 -1.50104538e-01 5.92828631e-01 -6.32831395e-01
-4.01440799e-01 -6.42165914e-02 7.31046677e-01 1.68894076e+00
4.92799915e-02 3.12983394e-01 8.11981142e-01 -1.98562354e-01
-5.21449983e-01 -1.29973662e+00 -3.72626126e-01 1.37014925e+00
-4.48101610e-02 -1.17306136e-01 -2.56525874e-01 -5.56349456e-01
5.39212584e-01 1.33225179e+00 -6.50730789e-01 -1.15894228e-01
3.64922732e-01 -8.81914556e-01 1.42572140e-02 1.77358240e-01
-1.13137975e-01 -4.87174511e-01 -6.41852677e-01 2.81571358e-01
-5.95663004e-02 -3.49069774e-01 -6.27128363e-01 2.37623632e-01
-9.05772865e-01 -9.89627719e-01 -5.98818839e-01 4.38161016e-01
3.17284018e-01 -1.64105490e-01 1.33203554e+00 -5.38311958e-01
1.79288000e-01 4.69651520e-01 9.65507925e-02 -8.82060468e-01
-5.60932867e-02 -4.03059453e-01 -2.39152964e-02 -6.20937347e-02
3.80265206e-01 -3.62219326e-02 -7.53483176e-01 3.67817134e-01
-6.45115554e-01 -4.05383468e-01 2.74500310e-01 8.75095308e-01
6.82473898e-01 -2.33582914e-01 5.38418114e-01 -1.14150941e+00
9.16649640e-01 -7.00552940e-01 -1.48039520e+00 7.42151439e-01
-1.03225720e+00 4.61385429e-01 -4.90377486e-01 -4.69451070e-01
-1.38233840e+00 -1.53797790e-01 6.42429411e-01 4.83125485e-02
2.48504639e-01 1.05950427e+00 -2.11402535e-01 3.30399066e-01
9.27844584e-01 -4.21593428e-01 -3.15501183e-01 -4.95288074e-01
1.51867256e-01 4.86435026e-01 1.51532605e-01 -1.03267050e+00
3.51257414e-01 3.62596363e-01 1.23591192e-01 -1.80623308e-01
-1.26833189e+00 -9.08777714e-02 4.33972895e-01 -7.15846494e-02
3.78265172e-01 -8.11680257e-01 -9.47807074e-01 -2.14677304e-01
-9.24149513e-01 -2.58265644e-01 -3.64728659e-01 8.19946468e-01
-7.67651439e-01 -2.49547586e-02 1.49551198e-01 -1.64758396e+00
-2.34686628e-01 -1.24480522e+00 9.43209529e-01 1.92027360e-01
-6.31272972e-01 -7.66039729e-01 1.13042921e-01 4.76145566e-01
3.71256530e-01 -6.55264556e-02 1.07400632e+00 -9.69288051e-01
-9.16092873e-01 -1.83981672e-01 1.06547408e-01 -3.74832839e-01
-3.00372064e-01 -1.19192814e-02 -1.03174543e+00 -2.86574960e-01
4.47691269e-02 -2.56012648e-01 8.95796537e-01 1.08834636e+00
1.36831951e+00 -8.65999639e-01 -4.50704724e-01 3.45620364e-01
1.00189257e+00 4.28812861e-01 2.46583924e-01 2.10799947e-01
-4.36117768e-01 6.86299920e-01 9.33010042e-01 9.74212170e-01
1.21195182e-01 6.63008869e-01 4.95060652e-01 5.28665841e-01
7.86363959e-01 -3.91962558e-01 4.91157845e-02 -4.55819339e-01
4.40240562e-01 -4.95855719e-01 -9.41517770e-01 4.86396700e-01
-1.98359895e+00 -6.82202816e-01 4.71422791e-01 3.05430102e+00
1.04876816e+00 5.10231733e-01 1.01256430e-01 -3.29693913e-01
8.01554561e-01 -8.33123550e-02 -1.09101892e+00 -2.27840662e-01
-1.54075310e-01 -1.53923586e-01 8.49171340e-01 9.21105981e-01
-9.23372269e-01 3.64316046e-01 6.94650316e+00 1.00265217e+00
-4.89777565e-01 2.17663541e-01 1.11073864e+00 -6.99926674e-01
-9.21552479e-01 3.10827196e-01 -8.16107273e-01 5.57987750e-01
6.77032292e-01 -5.79590321e-01 5.68687856e-01 1.06271100e+00
1.44363731e-01 -5.55018723e-01 -1.52079678e+00 9.15520787e-01
-6.71958864e-01 -1.45887244e+00 -2.78514147e-01 3.19856316e-01
9.19776678e-01 -1.74509510e-01 4.29147750e-01 2.37891361e-01
1.04696596e+00 -1.18217707e+00 1.04480398e+00 9.51389849e-01
7.22358406e-01 -8.74551237e-01 6.14445090e-01 5.68037570e-01
-1.09589867e-01 -4.70115095e-01 -2.74813920e-01 9.27513614e-02
5.07966727e-02 9.67066824e-01 -7.99209893e-01 -1.61273733e-01
6.53178871e-01 -5.24916165e-02 -1.00354925e-01 1.45798886e+00
-1.08980604e-01 8.74156177e-01 -7.26975799e-01 -3.80475938e-01
-5.66312224e-02 -2.89995909e-01 8.29519272e-01 5.66390216e-01
4.77763861e-01 2.12830827e-02 -2.34764665e-01 1.23167598e+00
5.28106792e-03 -9.32977870e-02 -5.99020720e-01 -6.92664683e-02
1.12207592e+00 4.70976979e-01 -2.54302144e-01 -1.92275882e-01
-5.55813350e-02 3.16529535e-04 -6.64160848e-02 8.14865649e-01
-4.31788713e-01 1.58727169e-03 4.58071828e-01 -2.02543154e-01
5.99010065e-02 3.24933290e-01 -9.21446145e-01 -1.06891811e+00
-3.07238754e-02 -7.79890358e-01 8.92063200e-01 -7.48408914e-01
-1.21850944e+00 -2.93077290e-01 4.71316367e-01 -4.77012396e-01
-6.24676645e-01 -2.86020011e-01 -4.01214063e-01 1.07071674e+00
-1.00905645e+00 -1.41422093e-01 5.05967379e-01 1.47095174e-01
2.04277679e-01 -1.54006720e-01 6.32320046e-01 -3.96376610e-01
-3.42534721e-01 7.75846303e-01 5.32468796e-01 -4.87897098e-01
5.72655797e-01 -1.33940113e+00 3.38764228e-02 5.21394372e-01
5.71876243e-02 7.18711853e-01 1.09154153e+00 -7.65834510e-01
-1.32243156e+00 -6.09263837e-01 6.88714623e-01 -6.30477250e-01
5.88778019e-01 -1.89091578e-01 -4.81985271e-01 7.22265124e-01
-2.35482633e-01 -5.15128851e-01 7.67976940e-01 7.32733130e-01
-4.00842607e-01 -2.13229284e-01 -1.41593742e+00 7.41066873e-01
5.22577047e-01 -3.12242210e-01 -5.12067437e-01 5.33340514e-01
8.99727941e-01 -2.99425393e-01 -7.17924893e-01 5.07553160e-01
8.32695544e-01 -8.22375834e-01 8.02258790e-01 -9.51365888e-01
5.36739454e-02 2.47237399e-01 -5.74416816e-01 -1.22217858e+00
-2.35510468e-01 -1.27255166e+00 -4.50555027e-01 7.83398747e-01
1.01583302e+00 -9.98446643e-01 8.99246037e-01 1.37878871e+00
3.49904925e-01 -6.26729608e-01 -1.33958793e+00 -5.20216584e-01
5.56707792e-02 -7.31931329e-01 1.01543486e+00 4.49895442e-01
-1.29763102e-02 3.30894254e-02 -2.30461881e-01 1.08557992e-01
1.26043737e+00 4.06219959e-01 4.77154672e-01 -1.43264580e+00
-5.63321054e-01 -5.55133164e-01 2.94637144e-01 -1.13990545e+00
-1.01669371e-01 -2.71841884e-01 5.70909567e-02 -9.71676648e-01
3.77581358e-01 -4.13305998e-01 -2.84740597e-01 1.29955947e-01
-2.89821923e-01 -5.41595876e-01 -1.21492475e-01 3.30290981e-02
-6.81564510e-01 4.21570957e-01 9.38419223e-01 -2.67523348e-01
-2.75640935e-01 8.07311118e-01 -1.01679444e+00 6.44338548e-01
5.99500418e-01 -6.68881536e-01 -5.49749196e-01 -1.51938930e-01
1.13862455e+00 8.44113648e-01 2.76631713e-01 -5.48573658e-02
2.34895602e-01 -8.85694623e-01 2.25645855e-01 -8.97451222e-01
2.92073578e-01 -7.53079057e-01 4.47819829e-01 2.94512510e-01
-1.05357027e+00 -3.37122232e-01 -1.90973744e-01 1.09822834e+00
1.73173726e-01 -5.40819645e-01 4.56527442e-01 2.22810488e-02
7.51733482e-02 1.82853237e-01 -5.09532571e-01 3.40220243e-01
6.82110369e-01 1.16325207e-01 -4.73548204e-01 -8.94811034e-01
-8.70360851e-01 5.46884656e-01 7.76280686e-02 9.61294845e-02
4.14502591e-01 -1.09439075e+00 -7.76125669e-01 -1.23041756e-01
2.36960500e-01 1.08774506e-01 -7.28245005e-02 6.35698318e-01
-1.24228129e-03 8.75646353e-01 7.18643725e-01 -5.34027576e-01
-7.17805326e-01 3.22540194e-01 4.96122986e-01 -3.48514766e-01
2.89962798e-01 9.31069314e-01 3.77618402e-01 -2.72728115e-01
7.29115367e-01 -3.02261263e-01 1.70886606e-01 1.21782862e-01
3.48824859e-01 6.81906104e-01 -8.10713917e-02 3.49323541e-01
-1.90509275e-01 -1.32088244e-01 4.34574634e-02 -6.81166291e-01
8.57579589e-01 -2.39298686e-01 -6.55801892e-02 5.48701704e-01
5.39494455e-01 3.03450655e-02 -1.41028726e+00 -5.66144526e-01
5.59519410e-01 -8.09821665e-01 4.11801308e-01 -1.30218661e+00
-4.21515226e-01 4.46341455e-01 4.38590109e-01 4.52088207e-01
5.19306719e-01 1.97322503e-01 -1.17593080e-01 7.20557213e-01
5.79578876e-01 -1.17863321e+00 -3.71275395e-01 -2.52694376e-02
1.04262900e+00 -1.43022418e+00 2.31372476e-01 5.22755012e-02
-5.69842041e-01 5.10529101e-01 1.71962291e-01 3.34215254e-01
8.92888188e-01 -7.40072783e-03 -3.54671925e-01 -1.05302230e-01
-1.24835300e+00 2.82412797e-01 6.03009939e-01 4.38543737e-01
-1.27300862e-02 5.09864867e-01 -1.98646039e-01 1.14270556e+00
-2.31974304e-01 -2.93532252e-01 1.17193535e-01 4.90719289e-01
-6.53823435e-01 -5.03320694e-01 -8.84709835e-01 1.03274083e+00
-8.02846014e-01 -1.93434149e-01 -2.66832769e-01 3.57673794e-01
-7.31697321e-01 1.30391979e+00 3.98425311e-02 2.79230982e-01
3.90963741e-02 -1.80484876e-02 2.82166988e-01 -6.04148686e-01
-1.28749147e-01 3.86630893e-01 4.18274999e-01 -6.69019818e-01
2.18105838e-01 -8.78977060e-01 -3.39989543e-01 -1.54419705e-01
-7.33087480e-01 4.49055046e-01 5.13504207e-01 8.64118576e-01
3.49622756e-01 -4.64851372e-02 5.05350590e-01 -1.40209496e-01
-1.77080035e+00 -1.00184202e+00 -7.91937649e-01 -2.75579900e-01
4.63064224e-01 -8.62553418e-01 -6.12342954e-01 -6.52008533e-01] | [4.526650428771973, 3.24007511138916] |
9d235b16-61a7-448b-90a3-4590fb015036 | enhancing-the-protein-tertiary-structure | 2306.01824 | null | https://arxiv.org/abs/2306.01824v1 | https://arxiv.org/pdf/2306.01824v1.pdf | Enhancing the Protein Tertiary Structure Prediction by Multiple Sequence Alignment Generation | The field of protein folding research has been greatly advanced by deep learning methods, with AlphaFold2 (AF2) demonstrating exceptional performance and atomic-level precision. As co-evolution is integral to protein structure prediction, AF2's accuracy is significantly influenced by the depth of multiple sequence alignment (MSA), which requires extensive exploration of a large protein database for similar sequences. However, not all protein sequences possess abundant homologous families, and consequently, AF2's performance can degrade on such queries, at times failing to produce meaningful results. To address this, we introduce a novel generative language model, MSA-Augmenter, which leverages protein-specific attention mechanisms and large-scale MSAs to generate useful, novel protein sequences not currently found in databases. These sequences supplement shallow MSAs, enhancing the accuracy of structural property predictions. Our experiments on CASP14 demonstrate that MSA-Augmenter can generate de novo sequences that retain co-evolutionary information from inferior MSAs, thereby improving protein structure prediction quality on top of strong AF2. | ['Siqi Sun', 'Yu Li', 'Tao Shen', 'Jiayang Chen', 'Le Zhang'] | 2023-06-02 | null | null | null | null | ['multiple-sequence-alignment', 'protein-structure-prediction', 'protein-folding'] | ['medical', 'miscellaneous', 'natural-language-processing'] | [ 4.20085728e-01 4.15477343e-02 -3.18046324e-02 -4.64566380e-01
-1.00366735e+00 -7.61569738e-01 -2.49715019e-02 3.94442946e-01
-1.09660096e-01 1.31504774e+00 2.23266780e-01 -4.89287019e-01
3.42231810e-01 -5.69771826e-01 -1.13268864e+00 -7.37548709e-01
-4.02823910e-02 4.30186540e-01 8.28262642e-02 -3.63970608e-01
2.11738229e-01 3.34436297e-01 -1.18694794e+00 6.49694145e-01
1.31997573e+00 2.19705433e-01 5.07340372e-01 3.63441080e-01
-6.55495971e-02 3.03615779e-01 -5.39874375e-01 -3.23564738e-01
7.48184249e-02 -7.95064270e-01 -8.58307898e-01 -3.60395521e-01
2.15895876e-01 -2.60820501e-02 -4.74884100e-02 7.81733930e-01
4.90165323e-01 -1.14434518e-01 4.40871239e-01 -6.58538401e-01
-8.96304429e-01 2.34544903e-01 -3.72319967e-01 2.04048514e-01
2.90464342e-01 8.77321839e-01 1.31021190e+00 -8.94866824e-01
1.09467185e+00 9.64902878e-01 7.40639865e-01 5.03037393e-01
-1.84476721e+00 -4.99934435e-01 -1.01106964e-01 9.70557481e-02
-1.08663869e+00 -7.94267580e-02 3.77572596e-01 -3.95530641e-01
1.76689577e+00 -2.05051620e-02 7.76310086e-01 1.18264890e+00
4.92878526e-01 6.78873956e-01 6.70687735e-01 4.99960594e-02
2.27980405e-01 -5.88004947e-01 5.47433533e-02 8.04369211e-01
7.55985081e-02 -4.09458652e-02 -7.42095172e-01 -7.09601104e-01
3.72816086e-01 5.09699397e-02 -3.66993636e-01 -4.44101959e-01
-1.19457042e+00 9.05649185e-01 6.01307929e-01 2.86806822e-02
-7.54701316e-01 -1.70520753e-01 1.64283052e-01 3.97662282e-01
2.19183058e-01 1.04707968e+00 -7.70953178e-01 -2.67791599e-01
-9.09942091e-01 7.16112256e-01 5.53795636e-01 6.07584476e-01
8.79525423e-01 -2.20586687e-01 2.26945058e-01 7.11397886e-01
3.42948847e-02 3.97493780e-01 4.92304683e-01 -5.90563059e-01
1.47568256e-01 8.10899019e-01 1.30527958e-01 -5.43734729e-01
-3.18999797e-01 -4.42746848e-01 -4.00202841e-01 1.31256968e-01
2.55106628e-01 1.07342377e-01 -8.55800331e-01 2.08969927e+00
2.28410766e-01 -2.17581958e-01 1.83648914e-01 7.64784813e-01
3.50587040e-01 7.28633702e-01 3.40524077e-01 -2.49335155e-01
1.16054535e+00 -7.33382463e-01 -2.62280524e-01 -1.00752808e-01
6.40000165e-01 -6.20055795e-01 1.22730160e+00 3.85242999e-01
-1.02585089e+00 -3.36374789e-01 -1.26942170e+00 -2.29761511e-01
-1.02628246e-02 -2.90381759e-01 7.88652480e-01 2.46371016e-01
-6.99478090e-01 1.02551746e+00 -9.31725919e-01 -2.22073823e-01
7.36710191e-01 3.87270391e-01 -5.39706171e-01 -3.10432799e-02
-1.06971776e+00 1.12382281e+00 4.53533202e-01 -4.81530726e-01
-8.86768699e-01 -1.05802619e+00 -5.45388341e-01 1.80150703e-01
-8.23708475e-02 -8.84791493e-01 1.19360960e+00 -9.31558847e-01
-1.04359424e+00 6.15824580e-01 -5.76351464e-01 -7.19243050e-01
1.27621233e-01 -3.64738852e-01 -1.44030392e-01 7.48742744e-02
9.06172469e-02 8.27912807e-01 4.00103360e-01 -7.28322804e-01
-3.18187445e-01 -5.16111612e-01 -3.40091914e-01 2.47953922e-01
3.77579331e-01 -1.55043244e-01 3.67829770e-01 -5.92091739e-01
1.06254287e-01 -8.57356668e-01 -6.98114038e-01 -6.25749677e-02
-7.99404234e-02 -1.07436031e-01 4.94134277e-01 -8.37023437e-01
1.02699661e+00 -1.80252230e+00 4.18417096e-01 2.14734897e-01
3.76660287e-01 3.52192193e-01 -3.67613465e-01 6.94570899e-01
-1.93992764e-01 2.45429128e-02 -5.68045676e-01 4.29573059e-01
-4.87877369e-01 1.64856702e-01 -2.96790004e-01 3.44869435e-01
7.81724095e-01 1.35613060e+00 -9.88254428e-01 1.53252646e-01
-2.71667629e-01 4.96089667e-01 -8.58290017e-01 1.69537634e-01
-9.42526877e-01 4.93918210e-01 -3.33873391e-01 8.66020977e-01
4.97408330e-01 -8.16740632e-01 7.21164107e-01 -1.72847897e-01
6.24754615e-02 6.33780181e-01 -2.23448902e-01 1.75239813e+00
9.71422344e-02 1.17088027e-01 -4.13612753e-01 -7.68935263e-01
1.04940546e+00 -4.68479964e-04 6.23208225e-01 -5.93555927e-01
-4.21128005e-01 4.37853575e-01 5.29783010e-01 -2.19111934e-01
2.29455933e-01 -3.35129887e-01 2.82199442e-01 4.02258962e-01
3.42868567e-02 2.75587052e-01 6.56753033e-02 1.52201310e-01
1.50368249e+00 5.24468124e-01 4.03079540e-01 -2.31495306e-01
1.30072057e-01 5.85232794e-01 1.01317513e+00 3.56917679e-01
-1.96552295e-02 3.96509737e-01 5.72537839e-01 -6.37098789e-01
-1.64652872e+00 -1.07103193e+00 -1.26405910e-01 1.00870240e+00
5.30973263e-02 -6.51699543e-01 -7.66213536e-01 -7.84481943e-01
1.82153076e-01 4.88224924e-01 -3.12676758e-01 -5.49719334e-01
-7.46362031e-01 -1.24387503e+00 6.82874739e-01 3.34445387e-01
-7.53467949e-03 -1.33333194e+00 -6.86036587e-01 8.05034220e-01
-3.66459668e-01 -3.81624252e-01 -5.26036739e-01 5.73918104e-01
-9.83071387e-01 -1.32571399e+00 -7.55660892e-01 -8.26236188e-01
4.76855397e-01 1.22120343e-01 1.48491788e+00 5.96774481e-02
-5.30422986e-01 -5.65270066e-01 -2.39876270e-01 -2.20216051e-01
-8.42860162e-01 2.20980629e-01 2.14119270e-01 -5.41437685e-01
8.46152246e-01 -1.00583267e+00 -5.54072857e-01 2.00738236e-01
-8.39074373e-01 1.88708991e-01 9.54717398e-01 1.36644864e+00
8.55824947e-01 -4.93391484e-01 1.04369235e+00 -8.60652924e-01
6.19128346e-01 -5.38257658e-01 -5.00640213e-01 3.57114643e-01
-7.20949888e-01 5.25390744e-01 5.56860566e-01 -3.56408089e-01
-7.35334754e-01 3.10688078e-01 -4.15058702e-01 -4.20055650e-02
9.76075679e-02 6.57312810e-01 -2.52986103e-01 5.20449877e-02
9.69967484e-01 7.18858659e-01 5.58841407e-01 -6.51190162e-01
1.11753598e-01 2.19551742e-01 4.74208474e-01 -6.06778204e-01
2.63150007e-01 -2.16049850e-01 -6.30587637e-02 -6.05098367e-01
-5.85643291e-01 -2.12866083e-01 -4.73700017e-01 5.33518076e-01
6.83805943e-01 -8.87515068e-01 -9.28195179e-01 2.74507195e-01
-1.06120288e+00 -2.18153626e-01 -7.25002377e-04 2.67811120e-01
-6.39857531e-01 5.74459732e-01 -6.69492900e-01 -2.90873230e-01
-6.41046524e-01 -1.48669505e+00 9.33923960e-01 -1.06855221e-01
-6.12079203e-01 -3.16139221e-01 3.48025382e-01 5.17013013e-01
1.51558548e-01 2.85052389e-01 1.59307790e+00 -7.60163784e-01
-8.42113733e-01 3.91617686e-01 1.53687894e-01 2.32970715e-01
1.71559140e-01 -2.25189447e-01 -3.17864746e-01 -2.96381056e-01
-4.24611300e-01 -6.16247833e-01 9.97064173e-01 1.29005983e-01
7.99373507e-01 -3.49708766e-01 -4.52749163e-01 5.98462284e-01
1.27800465e+00 2.82620013e-01 5.69853187e-01 4.35493112e-01
3.83873969e-01 2.70003825e-01 6.81052268e-01 1.37666434e-01
-1.60276126e-02 6.01469994e-01 4.95763153e-01 -7.06299245e-02
1.69554025e-01 -4.84479249e-01 4.16920632e-01 3.50695133e-01
-2.35350914e-02 -2.07149893e-01 -9.34511781e-01 4.20690209e-01
-1.73565865e+00 -1.13347590e+00 7.26589561e-02 1.91808641e+00
1.42916739e+00 -5.06137423e-02 2.15906486e-01 -4.13902760e-01
4.68572140e-01 -1.95603400e-01 -1.46283162e+00 -1.76779553e-01
-5.99134684e-01 5.01538575e-01 3.19399565e-01 3.16204101e-01
-7.88016200e-01 9.90024745e-01 7.15225887e+00 4.68301028e-01
-1.00901461e+00 -2.85254478e-01 6.31198704e-01 -2.13087231e-01
-6.23086393e-01 1.19166315e-01 -7.52109826e-01 6.82976246e-01
9.87521350e-01 -2.59280175e-01 3.51544440e-01 9.73557115e-01
6.00428618e-02 2.44244814e-01 -1.06977332e+00 4.73194510e-01
-4.59876508e-01 -1.93865061e+00 2.67563045e-01 2.82188445e-01
6.02663934e-01 4.48445112e-01 1.87843516e-02 7.71068037e-02
6.38382196e-01 -1.27495384e+00 6.82830513e-02 2.95102924e-01
6.22495532e-01 -1.08673179e+00 5.89890718e-01 3.76223415e-01
-4.56533700e-01 1.81136087e-01 -5.72276950e-01 3.01776499e-01
2.51138777e-01 6.19485199e-01 -1.17164898e+00 2.68653959e-01
4.11242723e-01 7.47467458e-01 -4.62581068e-01 5.91730654e-01
3.66954878e-02 7.48770177e-01 -2.51130104e-01 -6.36644587e-02
2.18050942e-01 -3.88898283e-01 5.17919302e-01 9.19898391e-01
1.56378567e-01 2.34534264e-01 2.29050010e-01 1.12798285e+00
-3.63298327e-01 1.02992699e-01 -3.35591525e-01 -4.43656594e-01
4.71359760e-01 7.70512462e-01 -1.16183192e-01 -3.01974416e-01
-3.23601156e-01 1.16316366e+00 6.25028491e-01 1.12805136e-01
-7.77534127e-01 -1.86331481e-01 1.48558354e+00 2.33380735e-01
2.50409871e-01 -7.29534850e-02 8.64362866e-02 -1.07051945e+00
-2.87452433e-02 -1.70473003e+00 2.37673745e-01 -6.23657346e-01
-1.50456798e+00 5.72204232e-01 -7.01509595e-01 -7.52677977e-01
-1.90732315e-01 -4.37105179e-01 -1.07435912e-01 1.04179287e+00
-1.09581435e+00 -1.15341902e+00 3.40199262e-01 -1.73274316e-02
6.38813734e-01 -4.05235797e-01 9.79108751e-01 1.06694244e-01
-3.02136302e-01 7.23444581e-01 4.27035242e-01 -4.83797759e-01
8.03253889e-01 -1.11387837e+00 1.13102543e+00 5.24942398e-01
-2.14670636e-02 1.16239583e+00 8.99074852e-01 -1.19370711e+00
-1.28693473e+00 -1.17475939e+00 8.26882660e-01 -5.70768118e-01
3.74931037e-01 -2.85841614e-01 -1.57783806e+00 6.54612005e-01
-2.22555056e-01 -4.18566883e-01 1.09133399e+00 1.71830833e-01
-6.47793770e-01 4.66921061e-01 -1.21549428e+00 5.71524382e-01
1.34129250e+00 -4.96237576e-01 -6.54410481e-01 4.23465490e-01
1.02949405e+00 -4.85638715e-02 -9.15521502e-01 6.78831279e-01
5.69158435e-01 -9.00313377e-01 1.20539021e+00 -1.33283079e+00
6.95998788e-01 -4.32064891e-01 -2.00288430e-01 -1.16768348e+00
-7.03229308e-01 -5.50005496e-01 -3.48612577e-01 6.19098723e-01
8.00398111e-01 -5.11730194e-01 1.00425184e+00 3.04022133e-01
-3.23500276e-01 -1.05617857e+00 -5.92751563e-01 -6.71596706e-01
4.09229994e-01 2.80196786e-01 8.48250031e-01 8.15507948e-01
1.09171145e-01 6.61575675e-01 -3.91228557e-01 -1.16458073e-01
2.66931534e-01 4.47621614e-01 7.33488381e-01 -1.06870329e+00
-7.48744249e-01 -3.48464012e-01 -2.85555780e-01 -1.24226868e+00
-5.58302887e-02 -1.10613370e+00 4.93654050e-02 -1.16977310e+00
4.98123288e-01 -1.65244699e-01 -2.71584362e-01 6.26367688e-01
-6.86056852e-01 1.02772497e-01 -1.87137142e-01 4.34998870e-01
-4.36013609e-01 7.17386186e-01 1.08752918e+00 6.17792504e-03
-2.34829769e-01 -1.82084441e-01 -7.04632640e-01 3.77587020e-01
8.54390562e-01 -2.92615116e-01 -2.04585612e-01 8.39140117e-02
2.77205855e-01 -9.62401107e-02 1.13014996e-01 -7.55440176e-01
-3.66795450e-01 -3.60831499e-01 6.71619892e-01 -7.41819739e-01
2.19528630e-01 -2.09743559e-01 5.14206350e-01 8.60582709e-01
-4.93957400e-01 4.04703379e-01 1.78167596e-02 8.05996597e-01
-5.97119965e-02 1.90303415e-01 9.28638995e-01 -3.81820232e-01
-4.97729361e-01 3.89637113e-01 -5.51947951e-01 -1.06157206e-01
7.49164939e-01 -1.10346086e-01 -2.25079730e-01 7.76331499e-02
-7.20266402e-01 -5.21240197e-02 1.08630288e+00 2.60791600e-01
4.01606321e-01 -9.01778460e-01 -7.72920728e-01 4.41871107e-01
3.59836787e-01 -1.79610729e-01 6.55543879e-02 3.53686750e-01
-7.30925083e-01 3.55531096e-01 -3.89489055e-01 -5.77307761e-01
-1.52679360e+00 7.54584432e-01 2.31907025e-01 -2.91452676e-01
-7.13410735e-01 8.14149141e-01 2.59448618e-01 -6.02564752e-01
-2.98401684e-01 1.05401628e-01 2.87441522e-01 -5.67423463e-01
3.82729411e-01 -1.19965822e-01 4.92616221e-02 -4.00163203e-01
-3.14757198e-01 4.09866869e-02 -5.06404698e-01 4.11731184e-01
1.68940163e+00 2.40511447e-01 -2.14169458e-01 -6.07484393e-02
9.57735658e-01 -1.94891065e-01 -1.35684681e+00 -2.36103714e-01
3.66475731e-01 -3.01168233e-01 -6.57012880e-01 -1.25249743e+00
-4.29259926e-01 5.78378201e-01 4.15444434e-01 -6.87445462e-01
7.67072976e-01 1.92899201e-02 1.33818328e+00 8.74884486e-01
5.74997306e-01 -5.99810183e-01 1.25572979e-01 5.43233633e-01
5.99668205e-01 -1.26999009e+00 -2.22635418e-01 -7.61314481e-02
-6.49654090e-01 8.10620785e-01 7.75789857e-01 1.91518277e-01
-7.48248100e-02 1.51162893e-01 -7.12792352e-02 -2.21984863e-01
-1.07005620e+00 1.76919386e-01 8.62233639e-02 6.01646841e-01
6.05043411e-01 -9.95864198e-02 -4.32749271e-01 5.86770892e-01
-2.28869125e-01 -6.20124117e-02 6.61091655e-02 9.93258119e-01
-7.72550464e-01 -1.53972995e+00 1.58872977e-01 4.72689986e-01
-7.35263228e-01 -4.48896348e-01 -7.57572532e-01 2.23601565e-01
-1.07733987e-01 5.16218126e-01 -1.93033934e-01 -1.78499252e-01
6.93129972e-02 4.85291123e-01 5.07332385e-01 -5.29370964e-01
-7.12276638e-01 -1.68462887e-01 3.92773822e-02 -5.14384210e-01
1.29496336e-01 -6.99591696e-01 -1.53925705e+00 -3.88643920e-01
-2.40644172e-01 4.17486012e-01 4.19609755e-01 5.78631699e-01
1.09934556e+00 2.93311507e-01 3.17841172e-01 -3.01576436e-01
-8.54657173e-01 -7.84101725e-01 -1.72386423e-01 5.88237643e-01
3.41851205e-01 -4.03843939e-01 1.74633220e-01 2.59533226e-01] | [4.70175838470459, 5.616098403930664] |
3dba634d-aa8f-42c1-84f5-df70a8237286 | learning-discriminative-feature-with-crf-for | 2008.01270 | null | https://arxiv.org/abs/2008.01270v1 | https://arxiv.org/pdf/2008.01270v1.pdf | Learning Discriminative Feature with CRF for Unsupervised Video Object Segmentation | In this paper, we introduce a novel network, called discriminative feature network (DFNet), to address the unsupervised video object segmentation task. To capture the inherent correlation among video frames, we learn discriminative features (D-features) from the input images that reveal feature distribution from a global perspective. The D-features are then used to establish correspondence with all features of test image under conditional random field (CRF) formulation, which is leveraged to enforce consistency between pixels. The experiments verify that DFNet outperforms state-of-the-art methods by a large margin with a mean IoU score of 83.4% and ranks first on the DAVIS-2016 leaderboard while using much fewer parameters and achieving much more efficient performance in the inference phase. We further evaluate DFNet on the FBMS dataset and the video saliency dataset ViSal, reaching a new state-of-the-art. To further demonstrate the generalizability of our framework, DFNet is also applied to the image object co-segmentation task. We perform experiments on a challenging dataset PASCAL-VOC and observe the superiority of DFNet. The thorough experiments verify that DFNet is able to capture and mine the underlying relations of images and discover the common foreground objects. | ['Jiaxiang Shang', 'Long Quan', 'Lei Zhou', 'Haoan Feng', 'Mingmin Zhen', 'Tian Fang', 'Shiwei Li'] | 2020-08-04 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/5794_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123720443.pdf | eccv-2020-8 | ['unsupervised-video-object-segmentation'] | ['computer-vision'] | [ 2.94508666e-01 -2.58972943e-01 -5.32606244e-01 -4.54530269e-01
-5.53671181e-01 -2.98889726e-01 6.58276081e-01 -3.72746795e-01
-2.59187192e-01 5.54813981e-01 1.05768412e-01 1.32084772e-01
-7.87521526e-02 -3.94978166e-01 -9.88389611e-01 -6.74307108e-01
-1.50135338e-01 8.16126466e-02 7.78207362e-01 1.39946193e-01
2.38731071e-01 2.54042536e-01 -1.72493267e+00 3.00463319e-01
8.52601230e-01 1.29440284e+00 5.88069916e-01 3.33406955e-01
2.80745983e-01 9.97880936e-01 -2.78528482e-01 -9.02032331e-02
3.31181824e-01 -2.29597896e-01 -9.44728017e-01 3.77424747e-01
7.09155202e-01 -4.51282918e-01 -5.45569718e-01 1.18220639e+00
9.31913853e-02 2.48435348e-01 6.14569545e-01 -1.37661564e+00
-6.17371976e-01 4.41185147e-01 -6.59938455e-01 6.45712376e-01
2.05462247e-01 2.47367397e-01 1.24124682e+00 -9.84447420e-01
7.15831518e-01 1.33793402e+00 2.42089480e-01 3.36783439e-01
-1.02183735e+00 -6.36365116e-01 4.41945553e-01 5.39366424e-01
-1.32603788e+00 -2.69620299e-01 6.99653089e-01 -4.75899011e-01
6.34846389e-01 3.34315374e-02 5.90301752e-01 1.14289796e+00
-7.03699589e-02 1.50199091e+00 1.18676686e+00 3.86861176e-03
3.42392735e-03 -1.95217818e-01 2.27797970e-01 9.79987204e-01
1.78022459e-01 1.21215470e-01 -8.32481921e-01 1.91725835e-01
8.61458778e-01 2.61038572e-01 -3.30345243e-01 -3.06942791e-01
-1.26370859e+00 7.81999111e-01 6.27621830e-01 1.30522504e-01
-2.71159112e-01 2.40741402e-01 1.75393000e-01 -9.88709182e-02
2.81627476e-01 1.23646900e-01 -5.72444320e-01 -1.03350580e-01
-9.54801142e-01 2.69081563e-01 5.22685885e-01 1.10988331e+00
9.68564868e-01 -4.36630361e-02 -5.46822071e-01 5.90170622e-01
4.02312487e-01 6.05255008e-01 2.78143138e-01 -1.18453836e+00
1.34840056e-01 5.86791992e-01 -2.15090379e-01 -1.06774950e+00
-1.14916369e-01 -5.59831023e-01 -7.16266990e-01 -8.84920657e-02
2.50132293e-01 8.74793008e-02 -1.14046884e+00 1.70228839e+00
3.25060070e-01 7.46201396e-01 -1.81563303e-01 1.12957716e+00
9.57725644e-01 6.06786549e-01 1.04572669e-01 -4.30163145e-02
1.24398422e+00 -1.27207851e+00 -3.61504942e-01 -2.98575163e-01
1.62279665e-01 -5.70142746e-01 8.54342937e-01 2.79438376e-01
-7.20294654e-01 -8.87440860e-01 -8.09569538e-01 -1.04029045e-01
-2.30212752e-02 2.62881905e-01 9.11930084e-01 4.23577689e-02
-8.97154570e-01 6.27531946e-01 -9.47450161e-01 -2.52149761e-01
1.02564406e+00 3.36398333e-01 -2.42980167e-01 -1.61924586e-01
-9.18445289e-01 4.99999315e-01 3.10473770e-01 -1.14517435e-02
-1.50399137e+00 -7.86866188e-01 -8.95704746e-01 2.46774983e-02
6.77005112e-01 -6.39066815e-01 9.41511810e-01 -1.14011979e+00
-1.27964687e+00 8.78657103e-01 -3.24361175e-01 -4.73259449e-01
3.14183801e-01 -5.06745279e-01 -2.92615026e-01 4.56734627e-01
3.61410558e-01 1.06639040e+00 9.23174858e-01 -1.22534037e+00
-9.76491153e-01 -4.73190509e-02 2.08454773e-01 1.52057726e-02
-1.39725745e-01 9.34304073e-02 -8.65679920e-01 -6.88183665e-01
3.85333523e-02 -9.88458991e-01 -5.88067845e-02 5.70273809e-02
-5.95454216e-01 -4.75794703e-01 1.08011007e+00 -3.86528403e-01
1.09706318e+00 -2.20855546e+00 2.27266818e-01 8.87371376e-02
4.71895933e-01 3.37148041e-01 -1.64864644e-01 -2.73007363e-01
1.31218031e-01 -1.76318273e-01 -4.33270186e-01 -1.54832050e-01
-1.49970427e-01 3.30480963e-01 -3.38639706e-01 5.74577570e-01
6.37090802e-01 1.28053236e+00 -9.67423201e-01 -7.36179948e-01
1.08454436e-01 2.76796877e-01 -6.07436955e-01 3.12418133e-01
-3.74666482e-01 4.31673348e-01 -7.81887114e-01 9.59383667e-01
5.65672219e-01 -6.07973039e-01 2.42077056e-02 -3.93950313e-01
4.70007733e-02 2.24420443e-01 -9.27016258e-01 1.85077035e+00
3.25887144e-01 8.33552897e-01 -2.70071477e-01 -1.22870922e+00
8.27885389e-01 -2.56790519e-01 5.25586188e-01 -6.09868765e-01
1.83980361e-01 -6.25627190e-02 -3.20668961e-03 -6.43683851e-01
3.48662227e-01 4.56210643e-01 9.79560316e-02 1.87955618e-01
4.49126452e-01 3.15930575e-01 2.15504915e-01 4.54343677e-01
9.78156447e-01 3.22697222e-01 1.65275171e-01 -7.28170276e-01
4.97912556e-01 -2.67708838e-01 9.01199400e-01 8.30972910e-01
-4.60021406e-01 5.12433112e-01 5.65653086e-01 -3.30122977e-01
-5.18855929e-01 -1.10678828e+00 -2.10096881e-01 1.11706007e+00
6.87593997e-01 -4.87490177e-01 -6.30418777e-01 -1.09692192e+00
2.04026952e-01 3.23568434e-01 -8.85641456e-01 4.04691361e-02
-3.40756387e-01 -4.98194188e-01 2.63073713e-01 7.51835227e-01
8.58138561e-01 -1.01340950e+00 -7.05803096e-01 -2.18996078e-01
-2.60309458e-01 -1.52022350e+00 -4.47763562e-01 1.45227373e-01
-8.11861813e-01 -1.29048133e+00 -4.43295777e-01 -9.56660748e-01
6.26178741e-01 5.77950001e-01 1.18861043e+00 2.15455487e-01
-2.34482422e-01 4.01116937e-01 -3.45038474e-01 -1.89832851e-01
2.27074549e-01 1.36604104e-02 -9.46840048e-02 1.39225423e-01
5.05782306e-01 -4.32933360e-01 -7.48113275e-01 5.89376032e-01
-8.90831053e-01 2.14704648e-01 5.73380470e-01 8.04897785e-01
8.57945144e-01 -5.28965928e-02 4.04040098e-01 -8.08845758e-01
-2.36086160e-01 -6.36947453e-01 -5.66454172e-01 1.58287331e-01
-2.96017945e-01 -3.28696296e-02 2.07235903e-01 -3.85749042e-01
-7.57151723e-01 1.59219041e-01 2.72518814e-01 -7.30134428e-01
-2.92293489e-01 2.52876788e-01 -3.32174450e-01 -1.48861930e-02
1.59938514e-01 2.60369509e-01 -1.65089682e-01 -3.56733680e-01
2.94604421e-01 3.56668502e-01 8.73148739e-01 -5.85520804e-01
8.74912977e-01 6.22835100e-01 5.43385334e-02 -5.66801369e-01
-1.45418608e+00 -6.17337286e-01 -6.53906643e-01 -3.58453900e-01
1.01919496e+00 -1.22551847e+00 -5.31898618e-01 5.24727106e-01
-9.78521168e-01 -3.76770467e-01 -9.05652121e-02 4.41029549e-01
-6.18767440e-01 3.07341486e-01 -4.77178186e-01 -3.05592120e-01
8.42060298e-02 -1.33998990e+00 1.12095964e+00 5.35926282e-01
9.48286951e-02 -8.05074155e-01 -3.47965658e-01 3.67602795e-01
1.65455922e-01 2.55378366e-01 5.87515473e-01 -6.40811503e-01
-1.13315296e+00 3.91950786e-01 -6.12662256e-01 4.92433786e-01
1.61385849e-01 1.37873009e-01 -1.11703384e+00 -1.61471024e-01
-1.70406476e-01 -3.65961492e-01 1.36973345e+00 5.33690691e-01
1.58563769e+00 3.04992925e-02 -4.50203389e-01 8.24897230e-01
1.31190109e+00 -1.98022127e-01 5.17314255e-01 2.04736054e-01
9.60788667e-01 4.01012897e-01 8.64267170e-01 2.73168653e-01
5.91221929e-01 6.92021489e-01 7.03077793e-01 -1.64182797e-01
-3.67213726e-01 -3.36220384e-01 4.11160439e-01 6.89846158e-01
-1.08989999e-01 -5.89318946e-02 -5.86932361e-01 7.01917768e-01
-2.06717157e+00 -8.96925092e-01 -1.16736002e-01 1.65265608e+00
7.72145092e-01 2.52734721e-01 2.40914404e-01 -3.07194442e-01
8.67292643e-01 3.63251805e-01 -6.50978506e-01 3.91547650e-01
-3.98276150e-01 1.42073691e-01 4.91657466e-01 9.37132761e-02
-1.52167261e+00 1.34817612e+00 6.56950569e+00 1.01859212e+00
-1.04136038e+00 8.15897733e-02 7.97988117e-01 1.54756188e-01
-1.44403398e-01 1.20542750e-01 -8.39322865e-01 6.33668661e-01
4.43745792e-01 6.75395876e-02 3.33037466e-01 9.47760165e-01
-7.59343952e-02 -3.17962348e-01 -1.08726037e+00 1.05253172e+00
1.89062789e-01 -1.47532856e+00 -2.13259645e-02 -5.68143539e-02
1.10689604e+00 4.70150024e-01 1.96605325e-01 2.37315893e-01
2.69567579e-01 -1.04913187e+00 8.02765012e-01 6.60505235e-01
6.32834315e-01 -5.45871317e-01 6.10568881e-01 1.65774040e-02
-1.24815941e+00 -6.15981072e-02 -4.90012735e-01 -2.97102518e-02
9.57454070e-02 7.75950372e-01 -4.87616122e-01 4.95886683e-01
1.08261132e+00 1.50451958e+00 -8.62106800e-01 1.05144823e+00
-4.17332172e-01 9.06370640e-01 -2.46618971e-01 1.78622440e-01
4.34038162e-01 6.44365838e-03 4.83473241e-01 1.27350688e+00
-5.52318022e-02 3.07693556e-02 5.55600405e-01 1.11451960e+00
-2.80969799e-01 -2.84547806e-01 -3.41818213e-01 3.74033526e-02
2.75333256e-01 1.43429410e+00 -9.19830143e-01 -5.22632897e-01
-3.84956092e-01 8.29059422e-01 3.11554313e-01 4.24760014e-01
-1.14206576e+00 5.03167100e-02 8.18124056e-01 -2.18479514e-01
8.80197525e-01 -2.48986423e-01 1.78065896e-02 -1.19525826e+00
-3.35109457e-02 -8.91107738e-01 1.73748061e-01 -7.11654902e-01
-1.34743345e+00 4.41700101e-01 7.49090537e-02 -1.20314705e+00
3.11737508e-02 -6.89060688e-01 -5.04550755e-01 3.94425184e-01
-1.96031690e+00 -1.05100238e+00 -5.29554009e-01 8.65318418e-01
7.09394932e-01 -2.29383692e-01 3.05121243e-01 1.91952676e-01
-6.70188546e-01 4.19581532e-01 -7.44372755e-02 3.29355448e-01
5.90575457e-01 -1.14935815e+00 2.48555705e-01 9.38746274e-01
4.08675104e-01 4.22260463e-01 3.25751662e-01 -6.42125845e-01
-1.45659328e+00 -1.32460773e+00 3.99569511e-01 -4.79348183e-01
6.82279527e-01 -2.96788424e-01 -8.81004512e-01 5.79831898e-01
9.80990604e-02 5.21675229e-01 5.23557603e-01 -9.86835808e-02
-4.41126198e-01 -8.50481093e-02 -7.67135978e-01 2.13934481e-01
1.42897987e+00 -6.06662035e-01 -6.30349576e-01 1.90058261e-01
9.59927738e-01 -3.02023053e-01 -6.96352303e-01 6.41078472e-01
4.30955321e-01 -1.00491333e+00 9.16356206e-01 -6.61563396e-01
8.20267737e-01 -6.05324268e-01 -4.94820237e-01 -8.83432508e-01
-5.14692307e-01 -6.19358122e-01 -3.42774004e-01 1.34535253e+00
1.87726215e-01 -2.60204762e-01 5.73101640e-01 5.90138547e-02
-4.37239185e-02 -9.85584438e-01 -7.48227298e-01 -7.24342167e-01
-3.75853270e-01 -5.72183609e-01 4.68707025e-01 7.88451612e-01
-5.48906147e-01 2.17874408e-01 -3.37261587e-01 2.39573419e-01
7.06574857e-01 3.64382714e-01 7.54713356e-01 -1.18203771e+00
-3.68667185e-01 -3.91212672e-01 -7.21867561e-01 -1.60277283e+00
5.77310979e-01 -1.00626791e+00 2.00697064e-01 -1.19197106e+00
7.63901055e-01 -3.34952325e-01 -7.17791796e-01 4.53138053e-01
-5.52581012e-01 4.04693574e-01 3.58554065e-01 3.25188726e-01
-1.39265418e+00 7.28247702e-01 1.36232448e+00 -2.87354410e-01
1.76039606e-01 -1.53641880e-01 -8.05102885e-01 9.18549776e-01
4.74625677e-01 -4.38336045e-01 -2.50907511e-01 -4.94508028e-01
-2.50898123e-01 -4.30094332e-01 8.22357297e-01 -9.25003946e-01
1.46891490e-01 -2.81432599e-01 5.79770386e-01 -7.54904330e-01
-4.01108377e-02 -6.21571600e-01 -2.82529175e-01 1.41777992e-01
-2.93079227e-01 -1.60700381e-01 -1.56370215e-02 7.19315588e-01
-2.98122346e-01 5.65204173e-02 7.26719022e-01 3.35720778e-01
-1.24287868e+00 6.64117873e-01 -1.67127587e-02 3.25057417e-01
1.07793331e+00 -6.74462914e-02 -3.82572591e-01 -1.72001109e-01
-3.18879098e-01 4.43615407e-01 4.43403691e-01 5.42661369e-01
7.08955109e-01 -1.30450165e+00 -5.46166241e-01 2.83181727e-01
1.04707867e-01 2.12058768e-01 1.22625679e-01 9.85084236e-01
-2.60626853e-01 2.83898622e-01 -2.32603639e-01 -1.19315767e+00
-1.01718283e+00 4.02290940e-01 8.84372741e-02 -3.06716245e-02
-7.22860575e-01 1.01876330e+00 6.11243784e-01 5.48748188e-02
2.67577618e-01 -3.29549700e-01 -1.86411723e-01 -1.39557049e-01
4.03671473e-01 2.50891298e-01 -4.41800565e-01 -8.29483390e-01
-5.92727721e-01 6.12622023e-01 -1.10341169e-01 3.35929602e-01
1.46834385e+00 -1.24613695e-01 -9.40762088e-02 1.76023528e-01
1.29060781e+00 -2.19021380e-01 -2.03036809e+00 -5.77970088e-01
8.88701454e-02 -6.86362267e-01 2.09111631e-01 -5.64970911e-01
-1.62537742e+00 6.16164505e-01 4.86320496e-01 -1.30936638e-01
1.23309839e+00 4.12484765e-01 6.58776402e-01 2.45592758e-01
2.52083838e-01 -9.82761979e-01 3.10627073e-01 5.35368741e-01
5.87012649e-01 -1.42259061e+00 1.16213165e-01 -5.44116795e-01
-9.26320255e-01 1.03290236e+00 6.96210623e-01 -5.10636926e-01
7.48020291e-01 1.50813699e-01 -2.95925528e-01 -3.31591636e-01
-6.71168029e-01 -5.76887369e-01 6.59634948e-01 4.94034976e-01
3.67305726e-02 2.89775021e-02 4.82570045e-02 5.11553228e-01
4.61555161e-02 7.66012445e-02 1.46158129e-01 7.16455400e-01
-5.32416523e-01 -6.96655810e-01 1.14387728e-01 4.82664287e-01
-4.09427881e-01 -1.66819412e-02 -1.51373655e-01 7.22583473e-01
3.38946909e-01 9.00398612e-01 1.84765488e-01 -4.32240129e-01
-8.05909038e-02 -3.49222690e-01 4.47217077e-01 -4.18317825e-01
-1.94646537e-01 1.55061647e-01 -2.97656566e-01 -1.07853901e+00
-1.09661043e+00 -7.92543113e-01 -1.30815828e+00 -1.48343798e-02
-3.67062598e-01 -1.30121768e-01 3.68338525e-01 1.18006074e+00
3.87596369e-01 6.71468794e-01 7.70167589e-01 -9.94837999e-01
-2.93836117e-01 -8.30181599e-01 -5.84096193e-01 5.67335010e-01
3.72622401e-01 -1.06526887e+00 -2.28362188e-01 3.20213735e-01] | [9.331472396850586, -0.23569414019584656] |
8d27e6be-c87c-4289-ab74-ff446bdde96c | exploiting-diffusion-prior-for-real-world | 2305.07015 | null | https://arxiv.org/abs/2305.07015v2 | https://arxiv.org/pdf/2305.07015v2.pdf | Exploiting Diffusion Prior for Real-World Image Super-Resolution | We present a novel approach to leverage prior knowledge encapsulated in pre-trained text-to-image diffusion models for blind super-resolution (SR). Specifically, by employing our time-aware encoder, we can achieve promising restoration results without altering the pre-trained synthesis model, thereby preserving the generative prior and minimizing training cost. To remedy the loss of fidelity caused by the inherent stochasticity of diffusion models, we introduce a controllable feature wrapping module that allows users to balance quality and fidelity by simply adjusting a scalar value during the inference process. Moreover, we develop a progressive aggregation sampling strategy to overcome the fixed-size constraints of pre-trained diffusion models, enabling adaptation to resolutions of any size. A comprehensive evaluation of our method using both synthetic and real-world benchmarks demonstrates its superiority over current state-of-the-art approaches. | ['Chen Change Loy', 'Kelvin C. K. Chan', 'Shangchen Zhou', 'Zongsheng Yue', 'Jianyi Wang'] | 2023-05-11 | null | null | null | null | ['image-super-resolution'] | ['computer-vision'] | [ 3.77807409e-01 -9.89102721e-02 -9.86004993e-02 -4.16994318e-02
-9.23116922e-01 -4.45915073e-01 7.01384783e-01 -3.79232645e-01
-1.66039422e-01 6.89618886e-01 7.62609363e-01 1.15205854e-01
-1.59472123e-01 -7.82223284e-01 -6.49990022e-01 -5.66546917e-01
1.36967286e-01 1.05294669e-02 2.30671540e-01 -5.37926592e-02
2.20731333e-01 4.28900272e-01 -1.46938598e+00 2.49084994e-01
1.19346678e+00 8.04714739e-01 5.01404881e-01 6.74743652e-01
1.11740485e-01 1.06588793e+00 -5.96914828e-01 -3.16998124e-01
3.77654582e-01 -6.14319026e-01 -5.54227591e-01 4.88017619e-01
4.09218907e-01 -7.47911274e-01 -6.88011050e-01 1.07293534e+00
6.45535529e-01 9.60839838e-02 5.30202270e-01 -2.86062330e-01
-1.29556715e+00 3.31271529e-01 -5.86441159e-01 4.72130179e-01
4.22988653e-01 3.46563607e-01 9.47376668e-01 -8.79033446e-01
7.93541908e-01 1.26311123e+00 4.37242955e-01 6.71881139e-01
-1.63175845e+00 -4.80394274e-01 1.44498572e-01 1.16509445e-01
-1.38527751e+00 -1.01575911e+00 5.69351315e-01 -5.05500317e-01
7.59033024e-01 1.02742575e-01 4.80680555e-01 1.18182206e+00
5.23304427e-03 4.93049085e-01 1.23270988e+00 -3.27461988e-01
3.73670161e-01 2.51826458e-02 -4.97870892e-01 7.58053362e-01
7.55289719e-02 2.46110499e-01 -7.50668883e-01 -1.01284862e-01
1.40659928e+00 -2.23914519e-01 -6.43310845e-01 -3.35808963e-01
-1.17780972e+00 6.50624812e-01 3.17244500e-01 1.09659344e-01
-4.71405387e-01 1.02727644e-01 3.15670297e-02 2.24614844e-01
7.77572751e-01 3.23280662e-01 -1.05617151e-01 6.48079067e-02
-1.41148531e+00 1.18870297e-02 3.93226355e-01 7.40791738e-01
6.03690624e-01 2.28698015e-01 -8.26673865e-01 9.96497869e-01
1.86681822e-01 4.05597299e-01 4.47749943e-01 -1.40491557e+00
1.83787212e-01 1.39725834e-01 3.70077223e-01 -6.07570231e-01
1.53287172e-01 -7.26005733e-01 -8.70615959e-01 2.39587590e-01
1.56052023e-01 1.09642230e-01 -1.13106322e+00 1.72742081e+00
2.15663031e-01 2.52058595e-01 -1.63797125e-01 1.04678988e+00
4.09561485e-01 5.46420217e-01 -2.38053188e-01 -4.96431977e-01
1.14043987e+00 -1.21934116e+00 -8.57110202e-01 -1.52188435e-01
-5.01811579e-02 -7.41910934e-01 1.19915116e+00 2.42003649e-01
-1.42620075e+00 -3.16051543e-01 -1.09757805e+00 -1.45444080e-01
3.34105581e-01 2.19429135e-01 3.20582122e-01 5.18445790e-01
-1.25240731e+00 8.96232724e-01 -8.86027694e-01 -1.40211448e-01
6.76061511e-01 1.71753794e-01 -1.05373420e-01 -3.59991223e-01
-8.22846830e-01 7.27570236e-01 -1.87275946e-01 -3.27779949e-01
-1.30224144e+00 -8.97373259e-01 -4.51661646e-01 6.06074594e-02
2.58076757e-01 -1.16258073e+00 1.11045825e+00 -8.14534366e-01
-1.98805404e+00 4.96460497e-01 -4.16966051e-01 -3.91584009e-01
7.49197304e-01 -2.99241006e-01 -4.23765868e-01 3.41473520e-01
7.86540005e-03 2.88916647e-01 1.49590635e+00 -1.50709724e+00
-2.32206807e-01 -5.26186265e-02 2.67370224e-01 2.56323695e-01
-7.36928403e-01 9.85686630e-02 -7.30331242e-01 -1.04476023e+00
7.16586038e-02 -6.42339587e-01 -3.41508329e-01 3.08604211e-01
-3.10120523e-01 3.97673696e-01 5.67510843e-01 -9.65574801e-01
1.41919827e+00 -2.16236591e+00 6.16088569e-01 -1.39879823e-01
4.95571166e-01 1.05050139e-01 -2.80920476e-01 2.56399661e-01
4.37918901e-01 4.50651757e-02 -5.43121636e-01 -7.94067919e-01
-1.53808460e-01 2.36158073e-01 -3.20006460e-01 3.93225282e-01
1.43992797e-01 5.77685058e-01 -8.11177075e-01 -4.35716897e-01
1.83957055e-01 8.93240094e-01 -7.86401033e-01 4.13611650e-01
-3.19504559e-01 8.71206582e-01 -2.10093081e-01 5.25483966e-01
6.21910334e-01 -6.32931471e-01 1.77910969e-01 -1.90488607e-01
-2.45431084e-02 4.16018248e-01 -1.10087144e+00 1.95583963e+00
-8.87206733e-01 4.52970684e-01 2.83494204e-01 -4.16975588e-01
5.36575019e-01 3.24516147e-01 2.57190615e-01 -7.53578424e-01
-1.94438457e-01 8.04616287e-02 -4.33755606e-01 -1.68569311e-01
6.37948751e-01 -4.95691691e-03 6.36459708e-01 5.98978221e-01
8.55163932e-02 -4.12296019e-02 1.25350952e-01 2.17746705e-01
1.18182480e+00 1.34668425e-01 -8.40531886e-02 -1.94440648e-01
3.59406680e-01 -5.24712622e-01 4.82052416e-01 6.36750281e-01
1.08849127e-02 1.02210605e+00 5.68077564e-02 -9.69862491e-02
-1.39968598e+00 -1.31405437e+00 -1.37687594e-01 8.01411629e-01
-6.55220589e-03 -4.81177330e-01 -8.66879344e-01 -3.92810136e-01
-2.70056039e-01 4.33598399e-01 -6.16651416e-01 5.18242493e-02
-4.30801749e-01 -8.19950640e-01 2.23650098e-01 3.67172241e-01
5.57482004e-01 -4.70782131e-01 -2.62397707e-01 2.26333439e-01
-2.79614091e-01 -1.36374414e+00 -8.04044843e-01 -3.93183112e-01
-1.05511117e+00 -4.29385483e-01 -1.08749425e+00 -4.72255945e-01
8.09346735e-01 3.77509385e-01 1.03711152e+00 1.22879781e-01
-1.18929692e-01 3.12118798e-01 -1.62067592e-01 4.68745947e-01
-4.10555273e-01 -9.10197422e-02 -7.11375475e-02 1.80612266e-01
-3.54605049e-01 -9.41555858e-01 -1.10731483e+00 1.73029169e-01
-9.16230798e-01 1.06789134e-01 5.02050519e-01 8.87672484e-01
7.49845445e-01 1.81488961e-01 3.85908991e-01 -4.27521706e-01
7.77129829e-01 -2.63661742e-01 -6.56225502e-01 1.79939926e-01
-8.59259784e-01 3.26141864e-01 6.01364076e-01 -7.50719905e-01
-1.26376402e+00 -1.70936227e-01 2.34537065e-01 -6.47499204e-01
4.07585859e-01 1.96912006e-01 -5.90710379e-02 -3.55766445e-01
6.78504288e-01 4.57280576e-01 -9.72683057e-02 -7.26349771e-01
6.93105340e-01 5.32357395e-01 7.09490240e-01 -5.60303867e-01
9.44451571e-01 7.81889260e-01 -3.36742997e-01 -5.84865153e-01
-8.35787773e-01 4.21382114e-02 -3.70132089e-01 -1.34706154e-01
7.14293897e-01 -1.27300322e+00 -3.30534071e-01 3.95363718e-01
-1.09677052e+00 -5.10578215e-01 -5.77560484e-01 3.23203534e-01
-6.05180383e-01 4.86211300e-01 -9.60321903e-01 -6.50418222e-01
-4.06283438e-01 -1.17984700e+00 9.69087124e-01 1.56668499e-01
-4.72107194e-02 -9.58508015e-01 -1.51691819e-02 4.32047904e-01
9.66578722e-01 -1.61899328e-01 7.48326540e-01 2.30411842e-01
-1.10744870e+00 1.97815612e-01 -5.47075808e-01 4.59955603e-01
3.14771175e-01 -2.26552621e-01 -9.65105414e-01 -5.41829824e-01
1.55062824e-01 -4.77752909e-02 1.08264625e+00 3.69403034e-01
1.05772007e+00 -5.67913234e-01 -1.06623814e-01 8.47707987e-01
1.40898716e+00 -3.58823895e-01 7.24042296e-01 4.50270891e-01
6.10015392e-01 1.43356949e-01 2.36474741e-02 7.45507538e-01
5.51030695e-01 7.08502948e-01 1.93379536e-01 -1.76183172e-02
-8.67009699e-01 -4.84971493e-01 4.97551560e-01 8.42067361e-01
-3.74202490e-01 -1.71113148e-01 -3.62304002e-01 7.08840907e-01
-1.57834351e+00 -9.20014679e-01 4.17819500e-01 2.32994342e+00
1.21884441e+00 1.85122769e-02 -1.35221511e-01 4.82600853e-02
6.23458862e-01 5.48977435e-01 -6.44500911e-01 9.67220217e-02
-2.69569039e-01 2.57378787e-01 5.50594509e-01 7.49110818e-01
-7.74771035e-01 1.01046956e+00 7.25491333e+00 9.35148895e-01
-9.69004631e-01 4.36882466e-01 4.11542147e-01 -4.86335337e-01
-7.87529111e-01 7.70650133e-02 -3.73474717e-01 3.74417931e-01
7.84301579e-01 -3.15648079e-01 1.17435431e+00 2.96377987e-01
4.71230328e-01 1.88501120e-01 -8.20265353e-01 7.46373713e-01
-4.16356884e-02 -1.57414794e+00 2.68571019e-01 1.23878986e-01
1.15199232e+00 1.63507275e-02 4.15185869e-01 -3.17541778e-01
5.66547215e-01 -9.76340950e-01 8.00526202e-01 8.35196853e-01
1.14819765e+00 -4.48174596e-01 1.86678674e-02 1.04441471e-01
-9.98705864e-01 -2.22630605e-01 -3.73629183e-01 2.23972932e-01
4.34333742e-01 9.80535030e-01 -3.13605309e-01 5.56305408e-01
5.92603505e-01 7.14197993e-01 -3.21702003e-01 7.97350287e-01
-3.83594841e-01 5.73600113e-01 -7.61818290e-02 7.07476318e-01
-2.91708887e-01 -1.39858693e-01 8.17337692e-01 9.63731766e-01
5.80250740e-01 7.08033442e-02 -1.55463874e-01 1.19111443e+00
-2.73174644e-01 -1.66401312e-01 -1.88027427e-01 -2.33665071e-02
5.39846838e-01 9.14172471e-01 -2.90264606e-01 -3.15171003e-01
-5.33090055e-01 1.49998653e+00 2.92473346e-01 6.77817643e-01
-6.22237265e-01 1.74977392e-01 7.49183714e-01 4.21019077e-01
6.60331905e-01 -5.50031841e-01 -3.95246059e-01 -1.35421574e+00
1.45075068e-01 -9.43669617e-01 -2.24016495e-02 -7.24769294e-01
-1.19242287e+00 8.14807892e-01 -4.52839881e-01 -1.07418287e+00
-8.95533431e-03 5.32315038e-02 -2.33128116e-01 1.08816171e+00
-1.90798926e+00 -1.12030125e+00 -2.41170079e-01 7.19535589e-01
6.37256682e-01 -1.59514561e-01 6.73787773e-01 4.56729561e-01
-4.68554944e-01 6.79349959e-01 2.96648949e-01 -3.42712969e-01
7.78170586e-01 -9.30896580e-01 4.96339887e-01 1.22715318e+00
-6.07885085e-02 6.31164968e-01 7.91031063e-01 -5.92001021e-01
-1.40650427e+00 -1.07743597e+00 6.17616296e-01 -4.30333465e-01
7.07614422e-01 -1.22553356e-01 -9.34608340e-01 4.55063850e-01
1.92944005e-01 2.61208296e-01 2.79244423e-01 -2.62790501e-01
-6.67530775e-01 -4.53900136e-02 -1.24577129e+00 6.54958308e-01
1.44013321e+00 -9.35460806e-01 -1.27643093e-01 2.50905603e-01
9.58218753e-01 -5.70721090e-01 -1.07093334e+00 4.22397643e-01
4.32185143e-01 -1.03427780e+00 1.16127062e+00 -7.16607943e-02
8.39476824e-01 -3.73807549e-01 -2.47293413e-01 -1.31329429e+00
-6.84682965e-01 -8.85214388e-01 -8.57798994e-01 1.16751242e+00
2.45696530e-01 -4.03537869e-01 5.91338813e-01 6.25732720e-01
1.05861574e-01 -4.10426944e-01 -8.64608765e-01 -9.01316643e-01
-1.34022385e-01 -3.08429450e-03 6.64693236e-01 7.36324131e-01
-4.01565909e-01 1.17614180e-01 -7.89117575e-01 4.42649603e-01
9.90855277e-01 -2.92210039e-02 4.56185788e-01 -7.72979856e-01
-7.86129177e-01 -4.24795032e-01 7.60414917e-03 -1.48836029e+00
-1.93778560e-01 -6.83310330e-01 -1.46633804e-01 -1.70761490e+00
3.31758738e-01 -3.38120878e-01 -2.46570662e-01 1.24918289e-01
-2.20573694e-01 4.34985101e-01 1.91193849e-01 6.97730601e-01
-3.65789652e-01 8.21654916e-01 1.86310589e+00 -2.63791699e-02
-2.65922606e-01 -2.95626909e-01 -8.20892155e-01 3.83620143e-01
4.24669087e-01 -4.47239161e-01 -5.44412076e-01 -9.81729746e-01
1.04295403e-01 2.00613618e-01 4.07425731e-01 -9.58895504e-01
3.54345620e-01 -2.01566353e-01 2.15010032e-01 5.17174974e-02
4.80851203e-01 -5.46203136e-01 1.96674883e-01 1.01017058e-01
-3.97665292e-01 -2.86296785e-01 4.47621495e-02 7.92177796e-01
-4.42161262e-02 1.12858653e-01 1.01429379e+00 8.13719109e-02
-1.92194179e-01 5.04449069e-01 -1.89450771e-01 -2.61727143e-02
5.31396925e-01 1.23656936e-01 -4.72099155e-01 -5.62723637e-01
-6.97123945e-01 -1.17830567e-01 9.13197100e-01 3.74220103e-01
5.92884541e-01 -1.27205217e+00 -8.66303027e-01 1.87467694e-01
-5.80146722e-02 -1.57068431e-01 4.43699300e-01 6.45047963e-01
-3.15163881e-01 -1.57490999e-01 3.99659947e-02 -2.49141857e-01
-8.57427716e-01 6.80751383e-01 3.44826490e-01 -2.82938093e-01
-1.02898955e+00 7.88929224e-01 7.24731311e-02 8.16545039e-02
1.94649190e-01 6.38738424e-02 2.02027671e-02 -3.98371488e-01
8.24914813e-01 3.64662975e-01 -6.79620951e-02 -5.05568266e-01
-8.77963752e-02 5.12520790e-01 -2.18523830e-01 -4.93256330e-01
1.35930443e+00 -6.48149371e-01 -3.17845680e-02 4.12151255e-02
7.44001329e-01 3.75076830e-01 -1.86551106e+00 -6.55135930e-01
-3.77786458e-01 -7.85915554e-01 5.48035324e-01 -1.05403709e+00
-1.28655922e+00 6.64046466e-01 5.98947763e-01 -3.12051885e-02
1.29830730e+00 -1.72852784e-01 9.81699288e-01 -1.47099450e-01
3.11948061e-01 -1.00038362e+00 3.10565114e-01 1.39078870e-01
1.02815175e+00 -1.03080857e+00 1.31994858e-01 -3.97524595e-01
-6.42072439e-01 8.01653087e-01 1.32018566e-01 -6.61672875e-02
5.01979291e-01 3.84113550e-01 -3.96171734e-02 1.15371443e-01
-8.34299326e-01 -1.35132000e-01 2.69518167e-01 6.22782111e-01
2.96614379e-01 -1.02994680e-01 -3.16051364e-01 4.43850815e-01
1.09898020e-03 4.10829484e-01 5.84868133e-01 7.10113525e-01
-3.74046981e-01 -1.10413194e+00 -2.29961887e-01 2.07138464e-01
-5.92968166e-01 -4.86575067e-01 6.55369610e-02 2.76512764e-02
-2.25757480e-01 1.06267691e+00 -1.49851426e-01 -3.48327667e-01
8.84674117e-02 -3.92858028e-01 7.08029032e-01 -3.79729807e-01
-9.40906331e-02 2.79756635e-01 6.56178072e-02 -6.13759935e-01
-5.66984713e-01 -4.64606732e-01 -8.47125709e-01 -4.86402571e-01
-1.71102405e-01 -2.54121214e-01 4.45162624e-01 8.64799082e-01
7.41525710e-01 7.83965766e-01 8.23231518e-01 -8.70828688e-01
-6.40804946e-01 -8.69328141e-01 -4.72178578e-01 1.59879863e-01
7.09037244e-01 -5.67429662e-01 -4.33390319e-01 3.14781427e-01] | [11.311005592346191, -1.9554007053375244] |
68cc3944-e51c-4e24-8269-f563af8020c5 | anomaly-detection-in-video-via-self | 2011.07491 | null | https://arxiv.org/abs/2011.07491v3 | https://arxiv.org/pdf/2011.07491v3.pdf | Anomaly Detection in Video via Self-Supervised and Multi-Task Learning | Anomaly detection in video is a challenging computer vision problem. Due to the lack of anomalous events at training time, anomaly detection requires the design of learning methods without full supervision. In this paper, we approach anomalous event detection in video through self-supervised and multi-task learning at the object level. We first utilize a pre-trained detector to detect objects. Then, we train a 3D convolutional neural network to produce discriminative anomaly-specific information by jointly learning multiple proxy tasks: three self-supervised and one based on knowledge distillation. The self-supervised tasks are: (i) discrimination of forward/backward moving objects (arrow of time), (ii) discrimination of objects in consecutive/intermittent frames (motion irregularity) and (iii) reconstruction of object-specific appearance information. The knowledge distillation task takes into account both classification and detection information, generating large prediction discrepancies between teacher and student models when anomalies occur. To the best of our knowledge, we are the first to approach anomalous event detection in video as a multi-task learning problem, integrating multiple self-supervised and knowledge distillation proxy tasks in a single architecture. Our lightweight architecture outperforms the state-of-the-art methods on three benchmarks: Avenue, ShanghaiTech and UCSD Ped2. Additionally, we perform an ablation study demonstrating the importance of integrating self-supervised learning and normality-specific distillation in a multi-task learning setting. | ['Mubarak Shah', 'Marius Popescu', 'Fahad Shahbaz Khan', 'Radu Tudor Ionescu', 'Antonio Barbalau', 'Mariana-Iuliana Georgescu'] | 2020-11-15 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Georgescu_Anomaly_Detection_in_Video_via_Self-Supervised_and_Multi-Task_Learning_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Georgescu_Anomaly_Detection_in_Video_via_Self-Supervised_and_Multi-Task_Learning_CVPR_2021_paper.pdf | cvpr-2021-1 | ['anomaly-detection-in-surveillance-videos', 'abnormal-event-detection-in-video', 'anomaly-detection-in-surveillance-videos', 'abnormal-event-detection-in-video'] | ['computer-vision', 'computer-vision', 'methodology', 'methodology'] | [ 2.96369016e-01 -1.60539642e-01 1.61000028e-01 -2.72654533e-01
-8.66093934e-01 -2.68836975e-01 5.96386015e-01 4.96347845e-01
-5.36145151e-01 2.74473488e-01 -1.10959418e-01 -2.03123555e-01
2.52481282e-01 -3.91650945e-01 -1.06524980e+00 -6.03536069e-01
-3.51894110e-01 3.94225955e-01 7.93019712e-01 2.91307848e-02
-1.26186749e-02 4.16726887e-01 -1.82961726e+00 6.27224326e-01
7.87197590e-01 1.45598626e+00 -2.84826607e-01 9.75875974e-01
1.17409937e-01 1.10267210e+00 -6.74165905e-01 -7.77834281e-03
2.73148000e-01 -3.28572392e-01 -6.60136700e-01 3.03394556e-01
8.15676689e-01 -6.91530406e-01 -2.91015744e-01 7.88190484e-01
4.56043959e-01 5.34280129e-02 7.19788194e-01 -1.58346426e+00
-4.50129747e-01 1.53108999e-01 -6.07648253e-01 1.10945940e+00
3.58379692e-01 4.36178058e-01 8.13378930e-01 -8.89623523e-01
3.43643248e-01 9.40132082e-01 7.00461149e-01 3.63376051e-01
-1.06889439e+00 -5.94922483e-01 5.26624501e-01 7.33480513e-01
-1.00136852e+00 -2.90937245e-01 7.52767324e-01 -7.79700339e-01
1.18331695e+00 -5.78948185e-02 5.12248814e-01 1.72884822e+00
8.56121927e-02 1.18422759e+00 6.09105706e-01 -5.84934950e-02
3.18603575e-01 -1.89778790e-01 1.62383392e-02 6.72785819e-01
1.96522877e-01 -5.95783442e-03 -5.65967083e-01 -2.13083938e-01
4.91796970e-01 3.86671960e-01 7.46715516e-02 -3.10913682e-01
-1.15413845e+00 5.66457629e-01 -1.13613300e-01 1.72700837e-01
-4.69462633e-01 1.65667310e-01 9.69104528e-01 5.91351032e-01
7.02940226e-01 1.26823947e-01 -8.61439109e-01 -2.25630403e-01
-1.00082254e+00 2.36009136e-01 5.95988333e-01 7.24202752e-01
5.11731148e-01 5.14196754e-01 -4.18750882e-01 5.56558669e-01
2.19003499e-01 1.94221869e-01 7.49610186e-01 -6.97360754e-01
3.12863618e-01 5.96742272e-01 -1.82316378e-01 -6.91341221e-01
-5.47016621e-01 -4.14723516e-01 -6.66867971e-01 4.98922527e-01
6.17711782e-01 -2.95920074e-01 -1.07737005e+00 1.53614330e+00
3.34584832e-01 1.00509202e+00 9.74489823e-02 7.23946095e-01
8.81579995e-01 4.50649381e-01 2.10115969e-01 -1.37033999e-01
1.26204836e+00 -1.05977786e+00 -4.87826854e-01 -2.08856210e-01
9.27982748e-01 -5.26238501e-01 7.07677364e-01 6.48245215e-01
-9.73142028e-01 -6.65424109e-01 -9.12320256e-01 2.64526755e-01
-4.43479925e-01 9.00266916e-02 2.95026958e-01 1.43374547e-01
-6.54386818e-01 5.26354253e-01 -1.09792161e+00 -3.11234593e-01
8.58945429e-01 -3.98502201e-02 -3.90539855e-01 1.79011405e-01
-9.80372369e-01 8.42341244e-01 4.49730128e-01 -2.47179806e-01
-1.40117526e+00 -8.72015357e-01 -1.07323885e+00 -1.16278604e-01
6.77473843e-01 -5.23929894e-01 1.34793818e+00 -1.38784873e+00
-1.02423882e+00 1.08859444e+00 1.09417073e-01 -8.74527633e-01
8.36339056e-01 -6.62337601e-01 -6.52518690e-01 3.31547886e-01
1.82375684e-01 2.78637230e-01 1.37633502e+00 -8.83878112e-01
-1.08304048e+00 -1.82203501e-01 -1.52660742e-01 2.45613921e-02
-1.20550446e-01 9.19941217e-02 -3.25273365e-01 -1.03915393e+00
-1.31655946e-01 -6.55754864e-01 5.45782894e-02 1.96723700e-01
-3.78717363e-01 -4.81610954e-01 1.34377599e+00 -6.95155561e-01
1.11521220e+00 -2.42579865e+00 7.10735619e-02 -8.49851295e-02
2.81068146e-01 5.64386845e-01 -1.63603693e-01 -6.04172014e-02
-4.67166573e-01 -3.75222206e-01 -1.69574901e-01 -5.50471485e-01
-2.55825996e-01 2.80750960e-01 -3.84692281e-01 5.67043126e-01
6.95554495e-01 6.78982615e-01 -1.15191829e+00 -5.60369849e-01
3.26713711e-01 2.32836306e-01 -6.40827477e-01 3.78765225e-01
-3.83346379e-01 4.88800049e-01 -3.51721197e-01 1.09228730e+00
3.21311980e-01 -1.87281266e-01 -5.63428760e-01 -1.18318245e-01
1.87111929e-01 2.07898691e-02 -1.31280422e+00 1.73353577e+00
-1.21613085e-01 6.74312174e-01 -2.95990050e-01 -1.42316675e+00
5.32009482e-01 6.11227930e-01 8.96591425e-01 -6.19377196e-01
-1.88364416e-01 1.84257090e-01 -3.57270800e-02 -8.68574262e-01
1.13832593e-01 3.30609977e-01 9.57564935e-02 3.89935821e-01
5.46447217e-01 3.65287334e-01 2.75966078e-01 2.50702024e-01
1.60722065e+00 4.90566134e-01 2.34661281e-01 1.92555845e-01
4.79409307e-01 -2.04306856e-01 8.13154638e-01 9.36413586e-01
-5.66435337e-01 8.00370216e-01 6.75058424e-01 -9.12435293e-01
-9.19391274e-01 -1.34619963e+00 -8.41889754e-02 1.42537344e+00
-1.75566152e-01 -2.37975597e-01 -4.32352632e-01 -1.25583827e+00
3.39424938e-01 6.88660264e-01 -7.33826399e-01 -3.70614946e-01
-6.57644153e-01 -7.71291673e-01 8.52815628e-01 7.72950053e-01
3.72210681e-01 -1.28864181e+00 -1.02743995e+00 2.50742614e-01
-5.49663007e-02 -1.49393725e+00 -3.07118356e-01 4.34628874e-01
-7.84892023e-01 -1.44502592e+00 -3.77875686e-01 -4.76169556e-01
4.74614680e-01 -3.02780252e-02 1.26686144e+00 1.13162629e-01
-6.99728489e-01 8.47476959e-01 -4.09077078e-01 -6.81854486e-01
-2.92646766e-01 -3.58391851e-01 2.68978566e-01 3.48044127e-01
4.87727314e-01 -8.13910246e-01 -6.34694517e-01 2.88899034e-01
-9.99031961e-01 -2.34996721e-01 6.51530504e-01 7.04220772e-01
7.32869327e-01 -1.87211052e-01 5.19303799e-01 -7.17836678e-01
4.12244946e-02 -8.36684465e-01 -3.64902616e-01 -1.82437338e-02
-3.30980271e-01 -9.30474773e-02 5.44559479e-01 -7.02693224e-01
-9.03072357e-01 2.59113759e-01 -9.01409239e-02 -8.44765484e-01
-7.85620987e-01 -4.75200638e-02 1.40075549e-01 2.23919302e-01
8.92908156e-01 3.43205631e-01 -1.72527924e-01 -3.07214618e-01
1.18439473e-01 1.63469478e-01 7.97650456e-01 -4.69381809e-01
6.23011827e-01 6.00751281e-01 5.92372902e-02 -7.77727842e-01
-1.19717550e+00 -6.36271000e-01 -8.11309278e-01 -2.41280556e-01
1.18634963e+00 -1.26174998e+00 -3.70279878e-01 8.03669155e-01
-1.01392210e+00 -4.39721674e-01 -7.49080181e-01 4.03095216e-01
-7.90013909e-01 4.16618913e-01 -4.79165792e-01 -5.63073397e-01
-1.46150708e-01 -1.00517440e+00 1.29216564e+00 -1.54845461e-01
-1.42753139e-01 -9.92315471e-01 8.96728933e-02 2.42627814e-01
1.65577590e-01 7.57243276e-01 4.02536035e-01 -1.46095121e+00
-5.28501451e-01 -1.83187112e-01 -1.98064417e-01 4.95100975e-01
5.40818237e-02 -4.48543094e-02 -1.12891674e+00 -1.56253308e-01
-1.32544711e-01 -6.58608198e-01 1.32293296e+00 3.60736340e-01
1.59641349e+00 -1.12922579e-01 -1.47376955e-01 8.82427871e-01
9.22933459e-01 -8.18486791e-03 5.00382841e-01 6.00586414e-01
8.75273943e-01 8.97099450e-02 5.37107229e-01 7.09959269e-01
5.17031252e-01 4.55544740e-01 7.51810908e-01 -1.33891940e-01
-2.04039827e-01 9.53407660e-02 5.46387017e-01 4.01637889e-02
-1.05389625e-01 4.18585502e-02 -1.01302326e+00 7.01033890e-01
-2.22373033e+00 -1.28411567e+00 -1.79241598e-01 1.84613752e+00
6.52773380e-01 4.30084109e-01 3.84222955e-01 2.29808778e-01
3.64599615e-01 1.74365029e-01 -9.13600326e-01 -2.12870002e-01
-2.41430447e-01 -1.37880608e-01 2.40894586e-01 -6.02167211e-02
-1.85339940e+00 7.21020341e-01 5.32426405e+00 5.37450314e-01
-9.50557649e-01 2.21467912e-01 6.25063777e-01 -4.15841520e-01
4.49266434e-01 -3.69601697e-01 -7.85907447e-01 6.58074498e-01
8.32674563e-01 3.13754916e-01 -1.89477608e-01 9.03097868e-01
1.21845666e-03 -1.67065725e-01 -1.47921336e+00 1.04431236e+00
3.10472906e-01 -1.04102325e+00 8.06656573e-03 -3.20161223e-01
6.57859683e-01 3.32453102e-01 -9.12370756e-02 5.66465735e-01
7.33990669e-02 -7.10473537e-01 7.06821978e-01 6.01944625e-01
2.68694639e-01 -3.28315169e-01 6.34185493e-01 1.89000592e-01
-1.05995274e+00 -3.42811257e-01 1.85137302e-01 3.63531858e-02
1.94248736e-01 7.25212753e-01 -5.16101480e-01 3.97233993e-01
1.17439091e+00 1.10929632e+00 -8.16389501e-01 1.11672664e+00
-1.70152724e-01 8.33159804e-01 -3.88657391e-01 7.81219065e-01
2.78081685e-01 2.03645766e-01 9.70429301e-01 1.53851509e+00
2.79626012e-01 -1.37134731e-01 5.54030836e-01 5.13996720e-01
2.39176989e-01 -3.02890152e-01 -5.58467805e-01 3.58317673e-01
-8.27427208e-02 1.13093960e+00 -5.08706391e-01 -5.31412899e-01
-8.70167255e-01 1.31124568e+00 2.51583546e-01 3.65340412e-01
-9.75459039e-01 -2.30764155e-03 8.51343095e-01 6.41179308e-02
6.03554010e-01 -3.96308750e-02 -1.19042337e-01 -1.39824116e+00
1.70167089e-01 -7.75501013e-01 1.02987480e+00 -4.54871356e-01
-1.41866076e+00 3.34442347e-01 9.46651865e-03 -1.37906742e+00
-5.56193411e-01 -6.05232358e-01 -1.02929544e+00 2.89959639e-01
-1.75207853e+00 -1.14129007e+00 -4.56555873e-01 9.57426071e-01
9.42243278e-01 -5.64556360e-01 6.60174072e-01 4.91154730e-01
-8.60378981e-01 6.37887418e-01 -1.97242767e-01 4.17868197e-01
1.08384693e+00 -1.64741611e+00 3.21965188e-01 1.17460167e+00
1.28680006e-01 -3.55306119e-01 6.32662296e-01 -6.57125056e-01
-9.31825280e-01 -1.57246411e+00 3.21298331e-01 -6.62675500e-01
8.09697092e-01 -6.96871355e-02 -1.37449157e+00 9.77756679e-01
-9.25915316e-02 8.27233016e-01 7.14462161e-01 -7.85767660e-02
-6.36020362e-01 -1.55029795e-03 -1.00278163e+00 1.82661563e-01
1.14093232e+00 -4.03244257e-01 -6.96809530e-01 5.75980365e-01
5.29770911e-01 -7.09556878e-01 -6.33736551e-01 5.82921863e-01
2.78111845e-01 -1.09154499e+00 1.17598724e+00 -9.22953129e-01
4.59536999e-01 -2.86271781e-01 -2.60799509e-02 -1.26166964e+00
-5.37246875e-02 -4.41516221e-01 -1.09057105e+00 9.27317142e-01
2.83424944e-01 -4.13814247e-01 5.67787111e-01 1.28632262e-01
-5.50624788e-01 -8.03133547e-01 -9.31623816e-01 -9.08425152e-01
-2.34879330e-01 -7.34471321e-01 -1.67016333e-05 9.78730202e-01
-4.25526023e-01 1.69392359e-02 -4.49146777e-01 5.05699635e-01
6.83051169e-01 -1.89882234e-01 6.63242638e-01 -1.43599892e+00
-3.61095101e-01 -3.91801596e-01 -7.67044365e-01 -7.15765059e-01
1.87592849e-01 -6.66667223e-01 -1.23604544e-01 -1.15068185e+00
1.84748170e-03 -1.32058084e-03 -5.04099905e-01 7.88107336e-01
-3.22638005e-01 1.93812340e-01 -8.08383673e-02 9.87828448e-02
-1.35157287e+00 4.74890977e-01 6.44509494e-01 3.01704444e-02
-2.33575255e-01 2.53006101e-01 -2.88576305e-01 1.13988256e+00
7.05249310e-01 -3.92687470e-01 -1.32439300e-01 -4.11948949e-01
3.97266895e-02 -2.13116676e-01 8.79308760e-01 -1.28758037e+00
2.23121300e-01 1.04633808e-01 6.86433196e-01 -6.56100452e-01
1.12756439e-01 -8.95452440e-01 -5.67834795e-01 4.30987120e-01
-1.77853003e-01 2.19915122e-01 4.77936000e-01 8.82886767e-01
-2.90239155e-01 1.31473690e-01 6.80295825e-01 -5.25259748e-02
-1.26719785e+00 5.68647146e-01 -5.38944662e-01 3.96236479e-01
1.51678574e+00 -2.78078586e-01 -2.05446884e-01 -2.26306185e-01
-1.07914305e+00 5.12231350e-01 -7.56319538e-02 8.10783207e-01
7.81220675e-01 -1.28492510e+00 -8.79755378e-01 4.16314423e-01
4.30392921e-01 2.38120586e-01 3.60425711e-01 9.39701676e-01
-2.83892244e-01 -7.03751370e-02 -2.55120844e-01 -1.11625981e+00
-1.19630051e+00 6.05341136e-01 4.41012114e-01 -2.61088192e-01
-8.92252445e-01 6.13320947e-01 8.28543678e-02 -1.24543838e-01
5.71110427e-01 -3.44529957e-01 -1.99171215e-01 2.99688488e-01
6.21503532e-01 5.88170826e-01 1.63707092e-01 -4.95588779e-01
-3.90651792e-01 3.42647552e-01 -2.65850484e-01 3.66365403e-01
1.30493355e+00 1.32454336e-01 4.12785351e-01 5.77918172e-01
9.82226968e-01 -5.04671931e-01 -1.71338248e+00 -5.78869939e-01
1.67197719e-01 -2.41545290e-01 -6.01234958e-02 -6.61680639e-01
-1.04201639e+00 7.97678530e-01 7.84540534e-01 2.21214473e-01
1.32506287e+00 9.98784155e-02 7.78636158e-01 4.64450389e-01
-2.37730980e-01 -1.29646230e+00 7.77279973e-01 6.16346061e-01
6.96629822e-01 -1.80468369e+00 -1.25149041e-01 6.27221242e-02
-8.05360377e-01 1.04273415e+00 1.20957386e+00 -2.49373704e-01
7.99634397e-01 2.14510217e-01 3.39920223e-02 -2.51400948e-01
-9.37701106e-01 -4.38003778e-01 6.31883681e-01 6.31346703e-01
1.38561279e-01 -4.17653531e-01 3.72646153e-01 4.90370542e-01
3.75994831e-01 -8.96617398e-02 3.69334102e-01 1.09773064e+00
-4.37043458e-01 -5.36880493e-01 -2.30721623e-01 7.80654609e-01
-8.15469980e-01 1.82429641e-01 2.99862772e-02 5.46171725e-01
4.69852328e-01 6.99368238e-01 5.52240551e-01 -7.63766319e-02
5.02100646e-01 4.38992292e-01 1.72063649e-01 -5.88216603e-01
-5.11945188e-01 -1.76499724e-01 -3.38851288e-02 -1.05572820e+00
-2.81116515e-01 -9.42053616e-01 -1.16838443e+00 3.03454161e-01
2.37038787e-02 -3.44008654e-01 3.11799109e-01 1.28397703e+00
5.17553747e-01 1.03077209e+00 3.83285731e-01 -8.71321857e-01
-4.92982894e-01 -7.78425455e-01 -3.00803572e-01 8.54399681e-01
8.85565102e-01 -7.13632405e-01 -4.87405658e-01 2.42765233e-01] | [7.840611934661865, 1.6341747045516968] |
f25bc6fa-8be8-4784-ab86-78da2577491b | visual-re-ranking-with-natural-language | 1810.12738 | null | http://arxiv.org/abs/1810.12738v1 | http://arxiv.org/pdf/1810.12738v1.pdf | Visual Re-ranking with Natural Language Understanding for Text Spotting | Many scene text recognition approaches are based on purely visual information
and ignore the semantic relation between scene and text. In this paper, we
tackle this problem from natural language processing perspective to fill the
gap between language and vision. We propose a post-processing approach to
improve scene text recognition accuracy by using occurrence probabilities of
words (unigram language model), and the semantic correlation between scene and
text. For this, we initially rely on an off-the-shelf deep neural network,
already trained with a large amount of data, which provides a series of text
hypotheses per input image. These hypotheses are then re-ranked using word
frequencies and semantic relatedness with objects or scenes in the image. As a
result of this combination, the performance of the original network is boosted
with almost no additional cost. We validate our approach on ICDAR'17 dataset. | ['Lluís Padró', 'Francesc Moreno-Noguer', 'Ahmed Sabir'] | 2018-10-29 | null | null | null | null | ['text-spotting'] | ['computer-vision'] | [ 3.58764082e-01 -3.42826009e-01 2.29729518e-01 -4.18522567e-01
-2.88972497e-01 -4.11497504e-01 1.04853487e+00 5.93578517e-01
-9.31133628e-01 1.98771387e-01 5.01131296e-01 -1.54660657e-01
2.49266148e-01 -8.01491261e-01 -5.09787202e-01 -2.26059183e-01
6.35232329e-01 4.37042862e-01 4.05036032e-01 -9.94682834e-02
5.35259426e-01 2.92875379e-01 -1.73301244e+00 7.80716479e-01
5.50110638e-01 9.50979114e-01 6.01733923e-01 6.10834241e-01
-7.16667950e-01 1.01271474e+00 -4.28614378e-01 -3.77395838e-01
-4.14574295e-02 -4.37012076e-01 -7.84928679e-01 4.90421414e-01
5.66654146e-01 -2.15304598e-01 -5.71254790e-01 1.00457036e+00
2.60873556e-01 1.17429703e-01 6.19261563e-01 -6.93589985e-01
-5.19970953e-01 5.13269544e-01 -4.84264344e-01 1.56539977e-01
3.92647624e-01 -1.03764363e-01 1.07817519e+00 -1.32529581e+00
5.24190247e-01 1.19161880e+00 4.76338804e-01 1.82588473e-01
-1.04639602e+00 -1.32169127e-01 3.03075165e-01 3.33635956e-01
-1.52443635e+00 -4.75490630e-01 7.94412971e-01 -4.98132706e-01
1.20587802e+00 1.17163047e-01 5.95013440e-01 1.08980262e+00
4.23758589e-02 8.24209750e-01 7.29578197e-01 -7.10861981e-01
2.32118756e-01 3.48288804e-01 3.53829235e-01 5.92960179e-01
2.96388894e-01 -4.76676464e-01 -6.20651305e-01 2.44717598e-01
4.12314802e-01 2.30883241e-01 -1.29240021e-01 -4.04301345e-01
-1.21480322e+00 7.73780406e-01 4.19324100e-01 6.16373897e-01
-4.62862909e-01 6.94175512e-02 4.57020104e-01 -1.23570278e-01
4.75568444e-01 3.43335479e-01 -5.80345280e-02 1.78473353e-01
-1.14880180e+00 -9.42732617e-02 8.13881695e-01 5.35994768e-01
7.51682460e-01 -2.07252562e-01 -3.64901811e-01 1.13636410e+00
5.75426757e-01 4.60933566e-01 7.35954702e-01 -1.53412938e-01
4.53606486e-01 8.46358836e-01 -3.20385039e-01 -1.29867887e+00
-4.33161676e-01 -3.37105274e-01 -6.43396318e-01 -1.89846888e-01
4.49543111e-02 3.06589186e-01 -1.26244843e+00 1.29501522e+00
8.92572924e-02 -6.75395727e-02 9.11101848e-02 8.27558219e-01
8.47313702e-01 7.38082469e-01 1.74487740e-01 8.20981525e-03
1.54287910e+00 -1.07210541e+00 -5.98803818e-01 -6.19931698e-01
7.14913309e-01 -9.45281208e-01 1.25007498e+00 2.96486467e-01
-7.36519456e-01 -5.28160930e-01 -9.62158024e-01 -3.10183287e-01
-7.49114454e-01 4.33803529e-01 2.69684404e-01 5.21172583e-01
-1.07063210e+00 2.63298213e-01 -6.50297046e-01 -9.18937981e-01
4.63864237e-01 1.01347916e-01 -2.85797507e-01 -3.37831050e-01
-8.58355224e-01 9.71349001e-01 6.60616100e-01 -1.94965467e-01
-5.77335715e-01 -2.97830284e-01 -8.79265845e-01 1.69776291e-01
2.99843848e-01 -6.96810186e-01 1.05979192e+00 -1.34568131e+00
-1.21559060e+00 9.62508202e-01 -4.02300924e-01 -4.40740585e-01
3.79284292e-01 -2.18229011e-01 -2.56177008e-01 3.84876549e-01
1.32327542e-01 7.16263533e-01 9.04953957e-01 -1.30669296e+00
-6.70074105e-01 -3.97622168e-01 -1.16996542e-01 5.16968131e-01
-6.53978467e-01 1.35137618e-01 -8.42375755e-01 -6.69415474e-01
1.80809110e-01 -6.10788465e-01 -7.09770247e-02 -2.15088166e-02
-3.48058730e-01 -3.02411079e-01 7.47823775e-01 -6.03182793e-01
8.72807920e-01 -2.26441813e+00 -1.36491060e-01 1.72379598e-01
2.78700650e-01 2.57335424e-01 -3.59813660e-01 5.82113624e-01
4.37730029e-02 -4.27200943e-02 -2.11589664e-01 -5.45834661e-01
-1.34800985e-01 1.27784833e-01 -5.28255284e-01 3.46227318e-01
1.59795091e-01 8.94453645e-01 -7.70456970e-01 -6.03958964e-01
5.98182261e-01 3.94826978e-01 -5.05365849e-01 8.82033333e-02
-3.12046885e-01 2.75309314e-03 -2.81962126e-01 2.53228575e-01
5.03954709e-01 -3.14461321e-01 9.86553654e-02 -2.89681584e-01
-4.27568667e-02 4.59986180e-01 -8.31890821e-01 1.80418313e+00
-4.93939847e-01 8.52960169e-01 -5.37166834e-01 -1.26750791e+00
7.92811155e-01 3.29479139e-04 4.91100192e-01 -9.30349588e-01
3.05553406e-01 -8.16337764e-02 -1.55601606e-01 -6.61141813e-01
6.30829513e-01 -1.79086089e-01 1.44096520e-02 3.30625921e-01
1.05348252e-01 6.69843564e-03 3.68561953e-01 2.86579400e-01
1.02456117e+00 -1.30154993e-02 3.45644802e-01 -9.34336707e-02
5.88161111e-01 2.61515081e-01 -2.12485507e-01 1.03294683e+00
1.50077924e-01 8.10560107e-01 2.76780367e-01 -4.37380165e-01
-1.10562396e+00 -9.59963441e-01 5.32768145e-02 1.21403515e+00
1.30201131e-01 -6.99466646e-01 -6.22005165e-01 -6.97791219e-01
-1.24298833e-01 1.01125515e+00 -5.93902647e-01 -7.92535618e-02
-2.65645176e-01 -5.99715531e-01 4.17704791e-01 4.97228622e-01
5.51082551e-01 -9.95649934e-01 -6.59336627e-01 6.52526021e-02
-1.72771275e-01 -1.46008062e+00 -2.50540167e-01 1.03441052e-01
-5.91882288e-01 -7.59771228e-01 -6.30398810e-01 -8.22104573e-01
8.11808169e-01 5.63391507e-01 1.00548148e+00 1.11666963e-01
-5.44889808e-01 6.24058306e-01 -6.79798782e-01 -3.69606048e-01
-3.13048661e-01 -1.66182876e-01 -1.68623075e-01 2.94487119e-01
6.23663008e-01 -1.65106639e-01 -4.02902603e-01 -1.49003342e-01
-1.00575769e+00 4.75766629e-01 5.72689950e-01 7.86176026e-01
4.33378845e-01 1.76949427e-01 -3.08003067e-03 -6.33221745e-01
4.73855257e-01 -2.84603417e-01 -4.94608164e-01 3.68392229e-01
-2.88397461e-01 6.21829890e-02 7.40292132e-01 -4.37009841e-01
-9.24337745e-01 3.64515394e-01 -3.23262475e-02 -4.69703108e-01
-2.55943656e-01 8.00401390e-01 1.29436441e-02 1.98363543e-01
5.79159498e-01 7.13935018e-01 -3.78403276e-01 -3.95676047e-01
4.39513773e-01 7.76395738e-01 3.80499691e-01 -2.16648895e-02
5.11016667e-01 6.46798372e-01 -2.79994190e-01 -1.24819863e+00
-1.12257802e+00 -7.43572295e-01 -7.70217717e-01 -1.68341860e-01
1.00361443e+00 -8.71297359e-01 -4.25517350e-01 3.43327940e-01
-1.38841045e+00 7.21153198e-03 -2.22392663e-01 6.90015554e-01
-2.03319132e-01 6.67291403e-01 -2.01703280e-01 -7.05506265e-01
-1.33901179e-01 -7.38888264e-01 1.27122045e+00 -1.07531352e-02
-3.81400026e-02 -8.64912808e-01 6.77511038e-04 3.41191173e-01
2.65914530e-01 -3.72713298e-01 8.91499817e-01 -9.27834332e-01
-2.80932099e-01 -3.60682160e-01 -7.70731986e-01 2.89979249e-01
1.72854796e-01 -3.42221260e-01 -9.87273633e-01 -6.58955798e-02
6.88352436e-02 -1.94575533e-01 1.26611269e+00 2.67220616e-01
1.33240557e+00 -1.92107841e-01 -3.09174716e-01 2.94217497e-01
1.52545393e+00 -7.81638697e-02 5.75760305e-01 2.58313268e-01
1.02438545e+00 7.63042808e-01 2.66791373e-01 5.17721534e-01
5.31813562e-01 5.17918706e-01 2.16006413e-01 -1.45807698e-01
-2.51074851e-01 -3.98973703e-01 3.04509312e-01 7.28217125e-01
3.93035710e-01 -6.09928846e-01 -1.34424341e+00 5.60451508e-01
-1.77626657e+00 -8.11181128e-01 -1.44194186e-01 2.08059883e+00
5.85974693e-01 1.43515438e-01 -2.10437179e-01 -8.06330442e-02
6.72377825e-01 6.37556612e-02 -3.24204117e-01 -2.42464498e-01
-1.87915295e-01 2.17797998e-02 4.59029049e-01 3.75448674e-01
-1.18215513e+00 1.34262252e+00 6.22251081e+00 7.89155126e-01
-1.26232779e+00 -1.25215396e-01 4.43649650e-01 5.87729141e-02
-1.34509385e-01 -4.22279835e-02 -6.08649492e-01 2.68922806e-01
8.18503737e-01 -2.14072242e-01 3.93765002e-01 6.94802701e-01
1.33449659e-01 -3.69811594e-01 -1.16763842e+00 1.26108551e+00
7.77925909e-01 -1.23587465e+00 6.19695723e-01 -2.20973551e-01
3.12950104e-01 2.23947361e-01 -1.78440720e-01 1.92740455e-01
1.10032804e-01 -1.06817079e+00 6.68362677e-01 5.26359141e-01
4.17744279e-01 -4.49297130e-01 7.23551512e-01 3.78955960e-01
-1.04790497e+00 2.49613822e-02 -5.27732432e-01 -7.12697208e-02
-1.14191622e-01 5.71989775e-01 -1.00739431e+00 4.51224715e-01
6.33199930e-01 8.28352571e-01 -8.96397650e-01 9.13066745e-01
-6.16399534e-02 3.61170232e-01 -3.61233860e-01 -3.07899922e-01
2.95452863e-01 5.15388697e-02 3.28411788e-01 1.55488169e+00
2.03908890e-01 -6.26018345e-02 3.69327337e-01 7.57150948e-01
-1.43369168e-01 5.67133546e-01 -7.59898782e-01 -3.07376474e-01
1.59076806e-02 1.25253701e+00 -1.14066052e+00 -7.12148428e-01
-7.37737358e-01 1.35373223e+00 2.97523797e-01 2.29370520e-01
-5.92860639e-01 -2.51573384e-01 1.43874004e-01 -6.07416816e-02
4.82868791e-01 -2.07982019e-01 -3.84723336e-01 -1.41926014e+00
3.22552547e-02 -6.36457682e-01 2.58745164e-01 -1.10100865e+00
-1.30779016e+00 5.98014712e-01 -2.07738549e-01 -1.07605803e+00
4.07856144e-02 -8.68260860e-01 -2.36034438e-01 6.83957279e-01
-1.36565757e+00 -1.21344316e+00 -4.04212236e-01 6.87779367e-01
1.03494453e+00 -1.54479519e-01 6.74728274e-01 1.59968600e-01
-2.90179789e-01 2.96496779e-01 1.32201791e-01 3.11306804e-01
4.63434398e-01 -1.07544243e+00 4.02850181e-01 8.45635295e-01
6.79461718e-01 4.59518403e-01 6.86345935e-01 -7.15275168e-01
-1.41262496e+00 -1.16654134e+00 1.07865608e+00 -4.63485152e-01
8.41879487e-01 -6.55440211e-01 -8.64921093e-01 2.88282752e-01
2.89270550e-01 -1.60547823e-01 6.82985425e-01 3.08474153e-02
-5.65287828e-01 1.68943882e-01 -7.66244709e-01 9.36455905e-01
8.65463316e-01 -6.91073239e-01 -8.97214055e-01 4.79229569e-01
5.68023086e-01 -2.30178554e-02 -2.86693513e-01 2.36603245e-01
3.52553248e-01 -7.08800852e-01 8.63616705e-01 -4.61346954e-01
6.27894640e-01 -1.94513068e-01 -4.01932657e-01 -1.04066443e+00
-2.13172376e-01 2.25771099e-01 3.80102903e-01 1.04447234e+00
3.55722517e-01 -2.75697738e-01 5.70442498e-01 1.79125160e-01
1.92115437e-02 -2.27109939e-01 -7.97680438e-01 -5.08587062e-01
-1.24983445e-01 -6.84529841e-01 3.40512842e-02 9.62616026e-01
1.09573297e-01 8.86469185e-01 -3.00057530e-01 2.75839423e-03
3.63278598e-01 4.09645401e-02 5.52916169e-01 -1.17956030e+00
-1.07394703e-01 -5.20039320e-01 -5.80866396e-01 -1.02899098e+00
1.70666024e-01 -1.05475807e+00 3.64961416e-01 -1.76059544e+00
5.96149087e-01 2.05971345e-01 -2.78910190e-01 5.42411327e-01
-1.36067569e-01 4.27136421e-01 2.85641551e-01 3.26490641e-01
-8.95979345e-01 6.62812412e-01 7.13543475e-01 -2.17928439e-01
-9.63716209e-02 -5.50344288e-01 -5.23197949e-01 1.02435708e+00
7.73614526e-01 -3.89882058e-01 -3.87654752e-01 -7.56898224e-01
1.80696875e-01 -3.57558936e-01 5.12699127e-01 -1.09589624e+00
4.23517674e-01 1.36079639e-02 5.80962241e-01 -8.38865221e-01
3.85968447e-01 -9.55108285e-01 -4.04004902e-01 3.68328482e-01
-5.52335620e-01 -1.38959602e-01 2.65410811e-01 6.21942520e-01
-2.49550000e-01 -2.60627955e-01 7.11897671e-01 -9.78614092e-02
-8.04319620e-01 -1.39948666e-01 -7.27309763e-01 -2.69993663e-01
7.66459227e-01 -3.59096467e-01 -2.59641290e-01 -2.67337650e-01
-4.98733819e-01 5.34108579e-02 4.59469438e-01 6.39728844e-01
9.12547886e-01 -1.02008796e+00 -6.14506662e-01 1.44477218e-01
5.10492980e-01 -3.88393462e-01 1.20904915e-01 6.93083644e-01
-5.06707191e-01 6.30638421e-01 -2.02858541e-02 -7.28043139e-01
-1.39345157e+00 7.72243142e-01 2.38569543e-01 -1.33928239e-01
-7.44230449e-01 5.60799956e-01 4.89554018e-01 -2.81974822e-01
2.50912279e-01 -1.95484802e-01 -4.75619942e-01 1.84358746e-01
5.65748632e-01 -2.64061660e-01 7.48906508e-02 -8.59269321e-01
-4.94597316e-01 7.02323973e-01 -2.61366636e-01 -3.69024038e-01
1.16609776e+00 -2.68484861e-01 -2.06942201e-01 6.98830664e-01
1.20408714e+00 -1.58830300e-01 -7.35509694e-01 -4.99090195e-01
2.23524392e-01 -5.35837352e-01 3.69386345e-01 -7.43278086e-01
-7.59128869e-01 9.60746467e-01 6.73027754e-01 1.69360980e-01
1.10324609e+00 2.08097443e-01 3.84628564e-01 7.48750925e-01
-3.67345661e-02 -1.21991622e+00 4.41737831e-01 8.14556181e-01
7.29980052e-01 -1.32888746e+00 2.45735832e-02 -3.67305517e-01
-7.69902170e-01 1.16316152e+00 3.94206315e-01 -1.81971774e-01
4.68222290e-01 -5.96295670e-02 -7.31294155e-02 -2.24559635e-01
-6.91224039e-01 -7.57477045e-01 5.30836046e-01 3.46868664e-01
5.32237887e-01 -2.35354796e-01 -4.16992456e-01 3.51400286e-01
-1.00925021e-01 -5.62123246e-02 3.99614960e-01 9.16967034e-01
-7.78306901e-01 -8.90278161e-01 -2.58831412e-01 5.48049986e-01
-3.43955308e-01 -5.38446724e-01 -6.71947241e-01 3.62123579e-01
-1.95117965e-02 9.81638670e-01 3.66238445e-01 -2.58433789e-01
2.61479706e-01 1.15784720e-01 3.63719016e-01 -8.31150174e-01
-2.65626639e-01 1.79343641e-01 -1.42038916e-04 -2.83472121e-01
-3.46851379e-01 -5.44398546e-01 -1.22426569e+00 4.71154936e-02
-1.54167712e-01 -2.39835456e-01 9.52881396e-01 1.12119687e+00
2.83634692e-01 5.19793928e-01 5.29091656e-01 -6.32259786e-01
-1.77428976e-01 -9.23512757e-01 -3.37553203e-01 5.18679976e-01
7.90043175e-02 -3.36476833e-01 -9.72395986e-02 3.42452377e-01] | [11.764827728271484, 2.237490653991699] |
3d0f5fba-81de-48eb-bcb5-49b7e7847075 | a-tvsnet-aggregated-two-view-stereo-network | 2003.00711 | null | https://arxiv.org/abs/2003.00711v1 | https://arxiv.org/pdf/2003.00711v1.pdf | A-TVSNet: Aggregated Two-View Stereo Network for Multi-View Stereo Depth Estimation | We propose a learning-based network for depth map estimation from multi-view stereo (MVS) images. Our proposed network consists of three sub-networks: 1) a base network for initial depth map estimation from an unstructured stereo image pair, 2) a novel refinement network that leverages both photometric and geometric information, and 3) an attentional multi-view aggregation framework that enables efficient information exchange and integration among different stereo image pairs. The proposed network, called A-TVSNet, is evaluated on various MVS datasets and shows the ability to produce high quality depth map that outperforms competing approaches. Our code is available at https://github.com/daiszh/A-TVSNet. | ['Sizhang Dai', 'Weibing Huang'] | 2020-03-02 | null | null | null | null | ['stereo-depth-estimation'] | ['computer-vision'] | [ 2.49624655e-01 5.85356094e-02 1.43722609e-01 -4.48609859e-01
-8.91140401e-01 -4.03288275e-01 4.44867462e-01 -9.41468701e-02
-3.70741159e-01 6.56664193e-01 4.18750048e-01 4.52621467e-02
-6.39401318e-04 -1.00566065e+00 -6.95204318e-01 -6.22929037e-01
3.15587938e-01 4.44546759e-01 7.64051676e-01 -1.48482814e-01
4.01545256e-01 5.09324133e-01 -1.70541489e+00 3.13410699e-01
8.46942604e-01 1.25872493e+00 4.21809107e-01 6.24954939e-01
1.80664167e-01 8.75003338e-01 1.48984477e-01 -1.39097616e-01
5.75542271e-01 7.17188790e-02 -8.25784981e-01 -1.59139588e-01
1.03089416e+00 -8.21279287e-01 -4.32231307e-01 9.87349629e-01
7.49245048e-01 1.72842264e-01 3.29160690e-01 -1.00076902e+00
3.75031829e-02 -9.70937405e-03 -6.33425474e-01 3.06489170e-01
4.38219249e-01 1.30571440e-01 8.50431383e-01 -9.97875988e-01
8.73551965e-01 1.13959718e+00 5.59990525e-01 3.76372576e-01
-1.19904637e+00 -5.94726861e-01 5.02338782e-02 2.80385643e-01
-1.25682676e+00 -5.75806260e-01 8.72869372e-01 -3.41120303e-01
1.00047350e+00 -7.43286461e-02 1.00899744e+00 9.52694118e-01
3.51610452e-01 8.05991530e-01 1.20241833e+00 -5.29587418e-02
1.33730650e-01 -2.60137409e-01 -8.47120360e-02 7.48221278e-01
3.57708149e-02 3.66697699e-01 -9.69658554e-01 -2.83925124e-02
1.30165648e+00 2.49512345e-02 -5.11506081e-01 -8.23307037e-01
-9.44180965e-01 7.90636480e-01 5.21180570e-01 -2.74698108e-01
-4.79259461e-01 1.01158828e-01 2.44356409e-01 7.20594823e-02
8.29623878e-01 8.46417621e-02 -3.44704360e-01 -2.11139452e-02
-9.39630032e-01 1.45000413e-01 4.83779550e-01 8.97207141e-01
1.09999645e+00 -1.50400460e-01 3.34226310e-01 8.38142037e-01
2.31028363e-01 3.81203830e-01 1.14815176e-01 -1.63386631e+00
6.28498912e-01 5.78213513e-01 1.85909141e-02 -9.01343286e-01
-4.34553653e-01 -1.07630342e-01 -9.37884986e-01 5.78503847e-01
2.33128190e-01 -1.15934797e-01 -6.88357532e-01 1.59598172e+00
5.32875717e-01 1.47751942e-01 -3.80565971e-02 9.25749719e-01
1.24742532e+00 3.82920831e-01 -7.05600023e-01 1.19064905e-01
8.19515646e-01 -1.17529190e+00 -3.30300987e-01 -3.82017910e-01
1.72842056e-01 -6.31451964e-01 5.94580412e-01 5.53731263e-01
-1.61531305e+00 -4.39282268e-01 -1.08580303e+00 -5.65816283e-01
-1.69840157e-01 -2.65568495e-01 5.88746011e-01 3.66485156e-02
-1.76162219e+00 6.06420398e-01 -7.78235316e-01 -3.72467935e-01
4.64889437e-01 5.55355132e-01 -5.11308730e-01 -4.09881264e-01
-7.74663270e-01 4.04049128e-01 4.08639491e-01 -1.84650738e-02
-9.51841831e-01 -7.09007680e-01 -1.25589406e+00 -1.42261147e-01
3.92275959e-01 -1.38272071e+00 1.18609607e+00 -7.66835928e-01
-1.39863741e+00 9.07318652e-01 -5.84337413e-01 -1.00928284e-01
5.52985311e-01 -2.03493834e-01 1.18417494e-01 4.95352387e-01
2.76664436e-01 1.07428181e+00 6.79035008e-01 -1.47315240e+00
-9.78669465e-01 -6.42812133e-01 3.30845267e-01 6.74804747e-01
1.59928188e-01 -4.79361087e-01 -6.40153408e-01 -3.61235410e-01
7.23678768e-01 -6.79452896e-01 -2.65998423e-01 1.94308579e-01
-5.18683374e-01 1.18439525e-01 5.64501822e-01 -5.24319470e-01
7.28710294e-01 -1.76665139e+00 3.05757046e-01 1.74547240e-01
8.00821960e-01 -9.56892446e-02 -1.55286789e-02 5.46052694e-01
-2.94379350e-02 -3.64070982e-01 -3.81299760e-03 -6.65749192e-01
-4.61513281e-01 -4.69632709e-04 1.11230604e-01 4.72084969e-01
-4.41742718e-01 7.65870869e-01 -9.38333511e-01 -5.61414540e-01
7.83149123e-01 4.95391846e-01 -8.01707149e-01 3.40698451e-01
7.68574849e-02 7.91238666e-01 -1.57793626e-01 6.77446306e-01
9.70515013e-01 -5.67584813e-01 -4.09473851e-02 -5.20667732e-01
-4.49318200e-01 2.06418231e-01 -1.22681510e+00 2.13767338e+00
-3.60504359e-01 5.90901077e-01 2.93531299e-01 -4.21966881e-01
6.86072409e-01 8.67175311e-02 7.50551939e-01 -6.13995314e-01
1.38716534e-01 8.06055516e-02 -5.23052990e-01 -2.17635587e-01
7.53538966e-01 1.41936257e-01 3.83633167e-01 3.87785792e-01
2.82819629e-01 -3.45189035e-01 1.47887826e-01 4.09793884e-01
1.11909437e+00 2.08717942e-01 2.78560340e-01 -1.96108565e-01
7.38085032e-01 -3.53487015e-01 8.32607269e-01 6.04337990e-01
-1.09963037e-01 9.91940677e-01 1.78933054e-01 -7.50736058e-01
-1.17615831e+00 -1.43584847e+00 -4.21804190e-02 6.53216422e-01
6.74639523e-01 -4.74133044e-01 -4.32476759e-01 -3.78668547e-01
-2.69573182e-02 8.39706585e-02 -4.64145333e-01 3.85933131e-01
-3.14810604e-01 -1.63129583e-01 -1.14910021e-01 6.85486972e-01
8.15919340e-01 -8.17900479e-01 -4.58312571e-01 -2.02752456e-01
-5.67970037e-01 -1.25557673e+00 -6.99851930e-01 1.23266995e-01
-1.17794812e+00 -1.28876972e+00 -6.09913111e-01 -6.37650669e-01
7.01987445e-01 6.20221317e-01 1.26473081e+00 -1.20809227e-01
2.25852445e-01 6.63329542e-01 -9.22455266e-02 -1.68110341e-01
7.08212480e-02 8.30467939e-02 -1.30941972e-01 -2.25860346e-02
2.41795052e-02 -1.26382530e+00 -1.22189140e+00 3.11402380e-01
-7.39006698e-01 6.91761553e-01 4.29123014e-01 6.68885291e-01
9.43931758e-01 -3.24888527e-01 -1.72432750e-01 -8.40455234e-01
2.51344666e-02 -3.10247391e-01 -9.34780717e-01 -1.56331956e-01
-4.55215394e-01 -1.69364959e-01 2.75595695e-01 3.34916085e-01
-1.15358186e+00 2.49347866e-01 -4.10401165e-01 -5.21094084e-01
-2.25572154e-01 6.33364022e-02 -1.68367922e-02 -5.75083494e-01
3.70310336e-01 3.83746952e-01 6.81076199e-02 -3.15124720e-01
1.71206474e-01 3.48481506e-01 7.18779087e-01 -2.86760539e-01
7.17231393e-01 1.10570526e+00 7.79819265e-02 -7.56364465e-01
-9.50301945e-01 -8.30479026e-01 -1.00455952e+00 -4.36566472e-01
9.97298896e-01 -1.32413137e+00 -7.10990906e-01 7.62717187e-01
-1.31516802e+00 -4.33403432e-01 -4.41820249e-02 4.32823688e-01
-8.67573738e-01 5.28103948e-01 -8.76248121e-01 -4.52916414e-01
-5.50501287e-01 -1.08061659e+00 1.33550262e+00 3.09405863e-01
1.71150029e-01 -1.29598188e+00 2.07438976e-01 8.09240639e-01
1.69889361e-01 4.73349690e-01 4.19318676e-01 -1.33346871e-01
-1.24435723e+00 3.18391055e-01 -4.98984993e-01 3.69463086e-01
1.88529208e-01 -2.32731462e-01 -1.14910412e+00 -5.97303867e-01
2.21036170e-02 -5.11481225e-01 9.64751363e-01 8.83538902e-01
9.38298881e-01 7.15671899e-03 -2.16230646e-01 1.33627486e+00
1.79140031e+00 -1.13724209e-02 7.43862927e-01 5.23642063e-01
9.93954003e-01 5.28220534e-01 4.26832467e-01 4.84314382e-01
8.49218905e-01 6.75918698e-01 9.02499676e-01 -3.69523048e-01
-5.24778031e-02 -2.02028677e-01 -2.13660635e-02 1.08546102e+00
-3.45514387e-01 -1.57210037e-01 -9.35201406e-01 5.23712873e-01
-1.91375482e+00 -7.67029941e-01 -1.14297695e-01 2.11319685e+00
2.32996106e-01 1.05353400e-01 -8.31501000e-03 -2.64249414e-01
5.84908962e-01 4.85480905e-01 -8.46413016e-01 -8.91404152e-02
-2.03141704e-01 1.53187048e-02 6.60802007e-01 7.10287929e-01
-1.09933531e+00 8.43313694e-01 6.16556311e+00 4.95717317e-01
-7.92042792e-01 1.99169666e-01 5.75798929e-01 -3.04757237e-01
-3.84572864e-01 -1.76120684e-01 -7.21453071e-01 1.82622045e-01
2.55669504e-01 -1.05818555e-01 3.58342886e-01 6.20574832e-01
3.72819811e-01 -5.48638999e-01 -9.83122766e-01 1.41310441e+00
3.01321477e-01 -1.68819511e+00 1.98318318e-01 2.19669834e-01
1.04047024e+00 7.01584995e-01 -1.48442119e-01 -3.55661243e-01
6.73440456e-01 -5.72955310e-01 5.82371414e-01 4.39575344e-01
7.37803340e-01 -8.84061515e-01 8.36294353e-01 1.60380214e-01
-1.55297267e+00 -2.71164235e-02 -1.71934441e-01 3.14235361e-03
4.01302099e-01 5.95254719e-01 -3.00820738e-01 9.57725167e-01
9.70286071e-01 1.31689382e+00 -5.75771451e-01 1.15832269e+00
-1.79629400e-01 -2.00467810e-01 -1.18542857e-01 6.98500574e-01
2.64248669e-01 -4.06733602e-01 6.40344918e-01 6.54504061e-01
2.82179892e-01 4.83311713e-02 5.40434495e-02 7.12293863e-01
7.80673400e-02 -1.09062612e-01 -9.01561916e-01 7.43999541e-01
5.40891767e-01 1.33045495e+00 -7.06162632e-01 -3.18826705e-01
-5.48812091e-01 9.72720802e-01 6.23450994e-01 3.66321176e-01
-3.91807973e-01 -2.33510230e-02 8.40862870e-01 1.25365451e-01
3.70403588e-01 -8.08164701e-02 -2.24846408e-01 -1.38319850e+00
3.98435928e-02 -6.34680688e-01 5.13038039e-01 -1.32919729e+00
-1.12267959e+00 7.53963053e-01 -1.75973047e-02 -1.36713111e+00
-2.93630540e-01 -4.88806069e-01 -4.89488602e-01 9.03493881e-01
-1.75246882e+00 -1.04026639e+00 -9.92143929e-01 9.50198829e-01
6.03406429e-01 4.50194143e-02 3.01593274e-01 2.35210553e-01
-2.68896699e-01 1.80961952e-01 1.01005624e-03 -9.95404497e-02
6.25189960e-01 -1.34022856e+00 5.43039739e-01 7.12248623e-01
-2.28178009e-01 2.46824920e-01 1.51742607e-01 -5.55568278e-01
-1.02818370e+00 -1.08674073e+00 5.24633288e-01 -4.64531422e-01
2.53481656e-01 -2.68399715e-01 -6.97551250e-01 8.39187920e-01
3.68598580e-01 1.28406314e-02 3.51769686e-01 -1.44361347e-01
-5.82536720e-02 -3.63848686e-01 -1.03393829e+00 4.45591629e-01
1.28615248e+00 -6.10030115e-01 -2.08094299e-01 2.07084313e-01
5.99703014e-01 -9.87781048e-01 -9.70620513e-01 4.28211838e-01
6.52883589e-01 -1.93529320e+00 1.20838487e+00 3.24176401e-01
8.19415987e-01 -1.85014561e-01 -2.37518162e-01 -1.29279208e+00
-2.39475280e-01 -4.40881640e-01 7.10193738e-02 6.91639721e-01
7.09389374e-02 -8.79744947e-01 1.13949585e+00 3.91577363e-01
-3.72996569e-01 -9.76050735e-01 -1.03032649e+00 -5.83563924e-01
-2.82293677e-01 -4.09221828e-01 3.53698224e-01 6.63033903e-01
-4.96155083e-01 3.18289310e-01 -3.02129388e-01 3.01642805e-01
1.16344643e+00 8.21045190e-02 8.96700680e-01 -1.29512060e+00
-1.68437421e-01 -2.02038586e-01 -5.37617981e-01 -1.42405713e+00
-1.18163839e-01 -7.58811355e-01 -9.60458145e-02 -1.88202751e+00
4.22546685e-01 -4.03705016e-02 -9.84052569e-02 1.08867221e-01
2.17002884e-01 5.10387301e-01 5.19441403e-02 2.45871246e-01
-7.65183389e-01 6.05225801e-01 1.50494909e+00 3.45895112e-01
-1.85053930e-01 -6.18172847e-02 -3.75401378e-01 1.14522088e+00
6.96015000e-01 -1.76371321e-01 -5.61538100e-01 -5.56018889e-01
2.19723240e-01 5.14243901e-01 5.28575838e-01 -1.19153464e+00
5.83526790e-01 1.90215111e-02 3.80185604e-01 -1.12936854e+00
9.03208733e-01 -6.61294520e-01 1.20939851e-01 1.81139037e-01
1.25094503e-01 3.51459950e-01 8.09058100e-02 5.56574106e-01
-3.35803479e-01 1.36718467e-01 8.81216347e-01 -4.58228499e-01
-9.54104781e-01 7.68735647e-01 -1.04275011e-02 1.54698044e-01
6.34224415e-01 -5.80898523e-01 -4.84203011e-01 -7.05243886e-01
-5.78504026e-01 3.97137344e-01 9.61927116e-01 2.08504185e-01
1.02771020e+00 -1.47805560e+00 -4.21234548e-01 3.34588915e-01
9.83214080e-02 5.86626947e-01 4.97146189e-01 9.42992449e-01
-8.79425585e-01 3.92979175e-01 -3.97573322e-01 -9.95256305e-01
-1.47936904e+00 6.69703484e-02 4.43129778e-01 -2.97306031e-01
-9.19131637e-01 8.77300978e-01 8.28069866e-01 -6.29683018e-01
2.16149300e-01 -2.23112121e-01 -1.89087272e-01 -2.42778733e-01
3.80329013e-01 5.82371414e-01 -3.63323949e-02 -7.95976341e-01
-1.86898708e-01 9.25936878e-01 -6.04124367e-03 -2.98308223e-01
1.54839635e+00 -6.46481335e-01 -2.88305223e-01 4.68693286e-01
1.13769758e+00 -3.53854537e-01 -1.76773512e+00 -4.02684569e-01
-5.04490435e-01 -8.42159152e-01 4.29351687e-01 -3.64033282e-01
-1.38755643e+00 5.67755997e-01 5.35182357e-01 -2.24357814e-01
1.32721627e+00 -1.15128294e-01 1.00682831e+00 2.23376065e-01
7.36647844e-01 -9.77908611e-01 3.45967561e-01 6.76224351e-01
7.78065383e-01 -1.52073848e+00 1.32786557e-01 -5.58809638e-01
-4.06377614e-01 1.14155757e+00 8.84261966e-01 -1.03945710e-01
7.44522929e-01 2.62463361e-01 2.27004543e-01 -4.97813731e-01
-8.18451762e-01 -2.56629467e-01 2.50970900e-01 6.88868523e-01
7.43658096e-02 -6.04126930e-01 3.13006759e-01 -8.53902400e-02
-1.95641562e-01 1.17599711e-01 5.06494701e-01 8.56363773e-01
-3.72571349e-01 -8.53043556e-01 -2.13579416e-01 4.90503222e-01
-4.21517156e-02 -2.92361856e-01 -2.75825769e-01 7.03245640e-01
2.07537055e-01 8.15225959e-01 3.13366145e-01 -5.40284455e-01
2.78332025e-01 -4.06373739e-01 5.94119132e-01 -6.02097034e-01
-3.69869947e-01 4.34632897e-02 3.73241715e-02 -1.32236052e+00
-8.24326992e-01 -7.52175808e-01 -6.50455892e-01 -6.42176390e-01
6.23954982e-02 -2.64205217e-01 3.40337366e-01 7.25671887e-01
3.92670274e-01 3.16874027e-01 7.22191095e-01 -1.42271054e+00
3.42317820e-01 -6.90645635e-01 -6.96893573e-01 1.35442153e-01
5.67997932e-01 -6.30490541e-01 -4.63587970e-01 -2.09780797e-01] | [8.918947219848633, -2.612142324447632] |
b523757f-2cc2-494e-b20a-f4f876424ad1 | multi-granularity-contrastive-knowledge | null | null | https://openreview.net/forum?id=Xf7cE59PJuP | https://openreview.net/pdf?id=Xf7cE59PJuP | Multi-Granularity Contrastive Knowledge Distillation for Multimodal Named Entity Recognition | It is very valuable to recognize named entities from short and informal multimodal posts in this age of information explosion. Despite existing methods success in multi-modal named entity recognition (MNER), they rely on the well aligned text and image pairs, while a lot of noises exist in the datasets. And the representation of text and image with internal correlations is difficult to establish a deep connection, because of the mismatched semantic levels of the text encoder and image encoder. In this paper, we propose multi-granularity contrastive knowledge distillation (MGC) to build a unified joint representation space of two modalities. By leveraging multi-granularity contrastive loss, our approach pushes representations of matched image-text pairs or image-entity pairs together while pushing those unrelated image-text or image-entity pairs apart. By utilizing CLIP model for knowledge distillation, we can obtain a more fine-grained visual concept. Experimental results on two benchmark datasets prove the effectiveness of our method. | ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['multi-modal-named-entity-recognition'] | ['natural-language-processing'] | [ 1.80049688e-01 -8.51074047e-03 -1.13197885e-01 -3.72260928e-01
-1.03899896e+00 -4.86111164e-01 7.28159785e-01 1.06231887e-02
-5.36769509e-01 7.36872673e-01 5.51973641e-01 2.66659290e-01
-4.00712378e-02 -5.14136255e-01 -9.08181667e-01 -5.91088533e-01
5.84894359e-01 2.85148323e-01 -3.24708619e-03 3.67873646e-02
-8.20986852e-02 1.21841155e-01 -1.40101767e+00 6.60512805e-01
6.44585013e-01 1.09189057e+00 3.57877731e-01 3.50988567e-01
-5.02551198e-01 9.46757019e-01 -2.67739326e-01 -8.97589743e-01
1.47624433e-01 -3.39195669e-01 -9.15764272e-01 1.71635151e-01
7.54660666e-01 -2.78963774e-01 -4.59042609e-01 1.22727132e+00
6.00846827e-01 8.87764692e-02 6.09234869e-01 -1.35702610e+00
-1.18524492e+00 7.20980465e-01 -8.41150165e-01 -1.58883214e-01
3.35370630e-01 -1.62346065e-01 1.21995807e+00 -1.19837093e+00
7.73676574e-01 1.39684844e+00 6.22160077e-01 3.83339107e-01
-1.06207561e+00 -7.01218486e-01 2.53120035e-01 3.67095977e-01
-1.78679132e+00 -4.40815747e-01 6.84086680e-01 -3.68984938e-01
4.95139569e-01 6.65227100e-02 3.53603035e-01 1.37125814e+00
-2.69104689e-01 1.16437697e+00 1.11594796e+00 -2.75022238e-01
-3.27018470e-01 4.18962419e-01 -3.11569482e-01 7.43057787e-01
2.38083646e-01 -2.24338174e-01 -7.82276690e-01 2.64866829e-01
8.53040278e-01 4.67047989e-01 -4.53527093e-01 -5.21883845e-01
-1.43867373e+00 5.35761178e-01 6.46079838e-01 5.34015417e-01
-3.25884402e-01 1.95618033e-01 3.26081008e-01 9.22558755e-02
1.73662961e-01 2.25026459e-01 -3.63100827e-01 -8.84376094e-02
-1.06767607e+00 -2.70928204e-01 6.13306880e-01 1.30705130e+00
9.45438683e-01 -3.14573497e-01 -3.77477765e-01 8.27976227e-01
2.55508184e-01 6.68744981e-01 5.07485330e-01 -5.57100058e-01
7.89531529e-01 8.25017333e-01 -1.97480079e-02 -1.30902851e+00
-1.38039634e-01 -1.71863407e-01 -1.12019527e+00 -2.96121091e-01
2.00970575e-01 -6.61383718e-02 -8.47004712e-01 1.69953072e+00
1.13936067e-01 2.59268463e-01 2.29909316e-01 8.92264128e-01
8.63034189e-01 4.76059347e-01 3.43333751e-01 -1.42083168e-02
1.64863014e+00 -1.16558456e+00 -8.69775355e-01 -2.21536294e-01
3.78587246e-01 -7.25417733e-01 7.95536160e-01 -2.65020311e-01
-9.78827536e-01 -6.54480577e-01 -8.22486401e-01 -4.70364094e-01
-8.23483944e-01 4.09329295e-01 3.66294712e-01 4.01232034e-01
-7.37898946e-01 1.92650229e-01 -6.47863567e-01 -3.75076383e-01
5.74391663e-01 3.07344999e-02 -9.41783369e-01 -4.11071897e-01
-1.18002570e+00 8.77189577e-01 7.27943778e-01 2.56585747e-01
-7.01303661e-01 -6.29036784e-01 -1.05918634e+00 6.23293966e-02
4.62863624e-01 -9.18480515e-01 8.18442583e-01 -1.15618265e+00
-9.02487159e-01 1.08264792e+00 -7.35117272e-02 -1.85464680e-01
5.91263473e-01 -5.11729658e-01 -6.41471386e-01 3.48316282e-01
3.79268944e-01 8.65590632e-01 9.44593132e-01 -1.67442131e+00
-8.22829127e-01 -4.66679096e-01 3.15177850e-02 5.75888097e-01
-5.38414538e-01 -1.18043087e-01 -8.92869473e-01 -6.93180382e-01
8.75323713e-02 -6.34296119e-01 2.76957452e-02 5.38633801e-02
-5.73175550e-01 1.18986294e-01 6.91617429e-01 -8.33531082e-01
1.06956172e+00 -2.27224493e+00 2.51908869e-01 -1.39737793e-03
3.90114546e-01 9.57094729e-02 -3.24508041e-01 3.94740194e-01
-8.90279785e-02 1.75861493e-01 -8.56432095e-02 -5.61993957e-01
1.94898710e-01 2.27878168e-01 -4.89679217e-01 1.86284930e-01
4.22239989e-01 1.40319705e+00 -1.05637598e+00 -9.98420417e-01
9.73849818e-02 7.36192107e-01 -1.79819733e-01 2.36363709e-01
1.65472284e-01 3.84568483e-01 -4.91277635e-01 8.93995285e-01
8.20021808e-01 -5.51540136e-01 1.31099924e-01 -8.71134937e-01
2.14086831e-01 -4.23218459e-01 -1.10870218e+00 2.08757448e+00
-5.43168128e-01 6.07832313e-01 -5.03855646e-02 -1.03284883e+00
7.88834393e-01 4.20600921e-01 3.96341115e-01 -8.66621017e-01
6.05705492e-02 6.14722073e-02 -7.38171279e-01 -7.17179775e-01
7.09378779e-01 -4.30903107e-01 -1.71328112e-01 1.10678233e-01
4.52082366e-01 2.72318833e-02 -3.87319326e-02 3.55481476e-01
7.51742482e-01 2.16545999e-01 9.05799493e-02 3.83309066e-01
4.75157589e-01 -4.08982664e-01 4.38663185e-01 8.52033019e-01
-2.31337890e-01 8.55375826e-01 1.76072851e-01 -1.51359767e-01
-1.07193077e+00 -1.24976695e+00 -7.92176053e-02 1.09330034e+00
5.34695268e-01 -3.85006815e-01 -4.07913357e-01 -8.33100438e-01
-1.39190480e-01 3.45552266e-01 -7.22911656e-01 -5.40523641e-02
-3.44230145e-01 -4.44573790e-01 7.38886416e-01 6.21066570e-01
8.50793123e-01 -8.50948930e-01 -4.74266261e-02 6.07968345e-02
-5.59754670e-01 -1.69481075e+00 -5.25908828e-01 -1.04056306e-01
-5.23687899e-01 -9.81382370e-01 -1.14764845e+00 -1.02545416e+00
8.26275289e-01 3.39085639e-01 1.22620869e+00 -1.64587662e-01
-3.48681360e-01 8.11182857e-01 -4.80094999e-01 -1.17713086e-01
-1.39929904e-02 -1.55004740e-01 -1.67224154e-01 4.83270615e-01
2.98791438e-01 -4.35371280e-01 -6.89126253e-01 2.21208349e-01
-1.19793046e+00 3.11904669e-01 9.99533057e-01 9.41129506e-01
7.29647994e-01 -3.40305232e-02 3.58310133e-01 -8.08726728e-01
3.81617069e-01 -6.31781697e-01 -1.20004229e-01 9.15525258e-01
-3.28383982e-01 1.87763393e-01 3.04051131e-01 -4.30560321e-01
-1.39934218e+00 1.83387801e-01 2.95522958e-01 -8.40638757e-01
-1.46476671e-01 5.78260005e-01 -3.79190922e-01 1.52900741e-01
2.34917298e-01 4.29889977e-01 -2.90364146e-01 -4.93505895e-01
1.07347941e+00 6.42437577e-01 7.88821220e-01 -6.41228437e-01
8.49378049e-01 7.00190425e-01 -3.54928583e-01 -4.86738414e-01
-1.14635718e+00 -6.74886584e-01 -6.59880996e-01 -1.22101456e-01
1.27841854e+00 -1.34194458e+00 -4.21803534e-01 2.83600837e-01
-1.27246356e+00 5.26116312e-01 -3.90033811e-01 5.31152427e-01
-4.07980233e-01 5.54311156e-01 -4.05731827e-01 -6.15776777e-01
-1.91753805e-01 -9.53937173e-01 1.47790885e+00 5.11000454e-01
4.10096288e-01 -9.58435297e-01 -6.22527748e-02 7.56051123e-01
2.42458090e-01 1.27960309e-01 6.13366187e-01 -6.53542519e-01
-7.32760489e-01 -2.58104742e-01 -8.99363637e-01 4.30056006e-01
9.66077372e-02 -3.08555424e-01 -9.61100996e-01 -9.74087343e-02
-1.96599379e-01 -5.87075591e-01 1.06622577e+00 -1.42213136e-01
1.11972380e+00 -3.68190050e-01 -4.78534400e-01 5.56134760e-01
1.57310557e+00 -2.76769251e-01 7.93872952e-01 1.77644625e-01
1.23909152e+00 5.95069528e-01 5.56224823e-01 4.28107232e-01
8.07656586e-01 4.73746002e-01 3.45064491e-01 -3.27504784e-01
-3.07974547e-01 -5.79467475e-01 2.69344360e-01 9.99839127e-01
-1.36421889e-01 -1.71492681e-01 -8.55035365e-01 8.27390432e-01
-2.01975369e+00 -1.15138233e+00 2.96144366e-01 1.91239417e+00
1.14619946e+00 -3.52136344e-01 -4.35946137e-01 -4.98606145e-01
1.06851506e+00 6.93765730e-02 -4.09646451e-01 3.26505572e-01
-6.03949547e-01 -4.19031754e-02 5.17214775e-01 1.18044123e-01
-1.24986982e+00 9.27057445e-01 5.14098215e+00 1.13438439e+00
-8.15311015e-01 2.16046542e-01 4.76291686e-01 1.28343049e-02
-3.95810157e-01 6.54778928e-02 -8.80817950e-01 3.82206917e-01
4.69957978e-01 -1.01355547e-02 8.67027715e-02 5.88618279e-01
-3.74011755e-01 9.89182070e-02 -1.13998425e+00 1.58138597e+00
3.48701626e-01 -1.36027014e+00 3.92309308e-01 -2.59649664e-01
9.86965299e-01 -1.18744224e-01 4.68161777e-02 4.40788537e-01
1.94271743e-01 -1.08024979e+00 8.08710396e-01 1.02096319e+00
8.85228038e-01 -4.43105876e-01 7.67246962e-01 7.66533092e-02
-1.47805929e+00 -1.65608451e-02 -2.81002760e-01 5.21812081e-01
2.48009086e-01 4.85747665e-01 -6.28462315e-01 9.71638143e-01
8.07690501e-01 8.43159139e-01 -7.07981765e-01 7.93790936e-01
-1.11387901e-01 -2.28298977e-02 -4.75843847e-02 2.92565703e-01
8.41185227e-02 -1.52602032e-01 1.92711189e-01 1.40179193e+00
2.61290282e-01 1.32122254e-02 2.99620889e-02 1.02428305e+00
-5.09093404e-01 1.67696163e-01 -6.39195085e-01 -4.42820668e-01
5.66472054e-01 1.40302277e+00 -4.11119699e-01 -4.55028236e-01
-7.52210200e-01 1.45771706e+00 6.17125988e-01 5.60064495e-01
-9.23165619e-01 -4.21556562e-01 3.16772312e-01 -3.79180998e-01
5.56092024e-01 -1.71185359e-01 -6.63342997e-02 -1.61012304e+00
2.25496605e-01 -7.03795910e-01 4.46528107e-01 -1.06949997e+00
-1.77373731e+00 4.80147928e-01 -1.00113913e-01 -1.42339003e+00
2.40268394e-01 -4.45974916e-01 -1.32017449e-01 7.28946865e-01
-1.70269895e+00 -1.79869354e+00 -4.04300123e-01 1.04278350e+00
2.05250233e-01 -1.58520509e-02 7.76544452e-01 6.24012113e-01
-4.36736614e-01 7.84799039e-01 2.33817175e-01 4.93789554e-01
8.36272895e-01 -1.14782298e+00 -1.58866048e-01 6.82143927e-01
4.69730288e-01 5.61082482e-01 2.24355415e-01 -5.63254476e-01
-1.36112428e+00 -1.08969355e+00 7.67255485e-01 -6.77335501e-01
7.99675047e-01 -2.99414366e-01 -8.42899859e-01 6.06027961e-01
3.86288673e-01 1.71434253e-01 8.83511364e-01 1.71259269e-02
-8.85707915e-01 -2.72443853e-02 -8.59051704e-01 6.20977104e-01
9.39758539e-01 -1.16199279e+00 -7.79974937e-01 2.86748320e-01
8.61561239e-01 -1.86430231e-01 -1.18667483e+00 4.32500750e-01
5.34774721e-01 -7.60152698e-01 1.11597204e+00 -6.78853273e-01
6.36859715e-01 -3.64004016e-01 -5.22950828e-01 -8.82866323e-01
7.77883530e-02 -1.85197338e-01 -2.22298503e-02 1.72094023e+00
3.74385476e-01 -2.12129518e-01 4.63130921e-01 5.96307039e-01
1.89914718e-01 -4.62917835e-01 -1.04248774e+00 -5.77843130e-01
-1.43315569e-01 -2.00662464e-01 4.48425055e-01 1.39172435e+00
3.28465551e-02 6.53370678e-01 -7.46789515e-01 3.18738312e-01
5.45561731e-01 2.55279005e-01 5.41496515e-01 -8.04065645e-01
-2.67838985e-01 -3.45500857e-01 -6.11833453e-01 -1.04202950e+00
1.95350334e-01 -8.79980862e-01 4.64636832e-02 -1.64085913e+00
7.15594888e-01 -2.44390845e-01 -5.93121529e-01 6.90374315e-01
-3.65931988e-01 3.93466473e-01 3.66447270e-01 3.23288679e-01
-1.13705802e+00 8.53253603e-01 1.30156410e+00 -3.84281009e-01
3.04637551e-01 -6.93967581e-01 -7.46274412e-01 5.74066877e-01
3.03756982e-01 -3.65473002e-01 -2.90982693e-01 -6.57995701e-01
3.89631569e-01 1.01467781e-01 5.21799803e-01 -7.79399037e-01
4.33232307e-01 -2.38068476e-02 6.03024364e-01 -5.71147442e-01
4.60274905e-01 -1.19089234e+00 6.51292726e-02 -8.67798552e-02
-4.16648239e-01 -1.74021125e-01 5.32135740e-02 9.96164918e-01
-7.69178689e-01 -3.30460183e-02 5.59708774e-01 -1.75079703e-01
-1.08099961e+00 4.57798868e-01 4.37048078e-01 3.17505300e-01
9.06821549e-01 -2.05699131e-01 -4.48616683e-01 -3.01371962e-01
-8.22105527e-01 2.98787653e-01 3.11456472e-01 7.75532424e-01
7.82716453e-01 -1.73912179e+00 -5.55053592e-01 5.37556782e-02
5.57201266e-01 -4.73999605e-02 7.40870774e-01 9.97554898e-01
-2.31196687e-01 2.41037980e-01 -2.90306956e-01 -4.38527465e-01
-1.07996631e+00 5.86081624e-01 2.94075191e-01 -3.05949122e-01
-4.12165463e-01 9.88701046e-01 6.55471206e-01 -4.15229499e-01
2.85177618e-01 -1.30551902e-03 -1.46374136e-01 3.91627699e-01
5.33114672e-01 -2.43582558e-02 -2.27891892e-01 -9.78735924e-01
-3.18161786e-01 7.40506709e-01 -2.57227421e-01 -1.55245572e-01
1.09609890e+00 -6.02366328e-01 -9.21734720e-02 5.02692819e-01
1.52341783e+00 -1.38164595e-01 -1.17882621e+00 -7.05177486e-01
-8.18353370e-02 -5.23886085e-01 -1.09912910e-01 -6.35606766e-01
-1.11972404e+00 1.01307058e+00 7.70520210e-01 -7.86532611e-02
1.23354232e+00 2.78926969e-01 8.56877685e-01 4.90114540e-01
3.37953381e-02 -1.25412810e+00 4.04909402e-01 4.12627012e-01
8.12948108e-01 -1.63301706e+00 -8.64284486e-02 -1.57251969e-01
-1.00215471e+00 9.01313066e-01 5.39551675e-01 3.40663850e-01
5.21612585e-01 -2.21067578e-01 1.95032684e-03 -3.31303358e-01
-4.12060082e-01 -5.15140414e-01 5.80617368e-01 7.05730796e-01
2.77624667e-01 1.73321459e-02 5.79867177e-02 7.96715379e-01
1.95476845e-01 -2.21361443e-01 1.60906047e-01 1.03143346e+00
-2.64017642e-01 -9.58938003e-01 -2.02775925e-01 1.77184209e-01
-3.57394427e-01 -4.42179143e-01 -2.65612125e-01 7.62627125e-01
2.20502004e-01 7.16184556e-01 2.17181370e-02 -4.01655644e-01
2.81571954e-01 6.48509115e-02 4.95800525e-01 -3.76654536e-01
-1.39500827e-01 -1.07636563e-01 -1.53668150e-01 -4.61041540e-01
-7.67974257e-01 -4.14785653e-01 -1.21898222e+00 3.83523712e-03
-4.66619611e-01 -2.59084348e-02 7.54932404e-01 9.41033125e-01
5.23127735e-01 2.70354569e-01 4.69980717e-01 -5.77677846e-01
-3.19538683e-01 -6.97025478e-01 -8.59503448e-01 9.18714046e-01
2.88063735e-01 -6.12294614e-01 -1.38974801e-01 5.07931530e-01] | [10.786239624023438, 1.4174575805664062] |
056db6f4-a854-46be-b23d-38308a14ddb5 | building-robust-machine-learning-models-for | 2208.10784 | null | https://arxiv.org/abs/2208.10784v1 | https://arxiv.org/pdf/2208.10784v1.pdf | Building Robust Machine Learning Models for Small Chemical Science Data: The Case of Shear Viscosity | Shear viscosity, though being a fundamental property of all liquids, is computationally expensive to estimate from equilibrium molecular dynamics simulations. Recently, Machine Learning (ML) methods have been used to augment molecular simulations in many contexts, thus showing promise to estimate viscosity too in a relatively inexpensive manner. However, ML methods face significant challenges like overfitting when the size of the data set is small, as is the case with viscosity. In this work, we train several ML models to predict the shear viscosity of a Lennard-Jones (LJ) fluid, with particular emphasis on addressing issues arising from a small data set. Specifically, the issues related to model selection, performance estimation and uncertainty quantification were investigated. First, we show that the widely used performance estimation procedure of using a single unseen data set shows a wide variability on small data sets. In this context, the common practice of using Cross validation (CV) to select the hyperparameters (model selection) can be adapted to estimate the generalization error (performance estimation) as well. We compare two simple CV procedures for their ability to do both model selection and performance estimation, and find that k-fold CV based procedure shows a lower variance of error estimates. We discuss the role of performance metrics in training and evaluation. Finally, Gaussian Process Regression (GPR) and ensemble methods were used to estimate the uncertainty on individual predictions. The uncertainty estimates from GPR were also used to construct an applicability domain using which the ML models provided more reliable predictions on another small data set generated in this work. Overall, the procedures prescribed in this work, together, lead to robust ML models for small data sets. | ['Sundaram Balasubramanian', 'Sudarshan Behera', 'Shivanand K. Veesam', 'Nikhil V. S. Avula'] | 2022-08-23 | null | null | null | null | ['gpr', 'gpr'] | ['computer-vision', 'miscellaneous'] | [-6.21282123e-03 -2.17930868e-01 -7.16477111e-02 -3.92748207e-01
-8.73233140e-01 -5.32701910e-01 5.56243777e-01 4.69870806e-01
-5.91133356e-01 1.29296291e+00 -3.86964798e-01 -3.74984682e-01
-2.99844086e-01 -8.17984939e-01 -6.16593897e-01 -1.11334026e+00
-5.12501895e-02 6.24659419e-01 2.02737018e-01 2.41047591e-02
5.76579630e-01 7.48879969e-01 -1.58217764e+00 -3.72304201e-01
1.41824222e+00 8.13034356e-01 7.27472380e-02 8.45881939e-01
1.10215424e-02 2.85367161e-01 -5.38634717e-01 -3.23823810e-01
1.67869788e-03 -3.71667504e-01 -5.36011994e-01 -3.48968446e-01
3.63435954e-01 5.21494336e-02 3.07241201e-01 6.52421653e-01
6.64451897e-01 7.38869607e-01 1.12318289e+00 -9.49210346e-01
1.54732512e-02 2.63817251e-01 -9.63713452e-02 1.34797931e-01
3.89545590e-01 3.94452155e-01 5.57735682e-01 -6.01537108e-01
4.32560652e-01 1.13694048e+00 6.64857805e-01 4.55840468e-01
-1.41956341e+00 -6.67814076e-01 -2.04801053e-01 -2.44215414e-01
-1.27099717e+00 -3.17308545e-01 4.19460356e-01 -6.23414040e-01
1.01039481e+00 3.48742813e-01 6.00776732e-01 1.12473154e+00
4.95037705e-01 2.02498853e-01 1.55133998e+00 -4.44979519e-01
7.38103151e-01 4.68838274e-01 2.62103260e-01 2.74540544e-01
6.81428313e-01 4.70004499e-01 -1.07664354e-01 -5.48456132e-01
5.19193172e-01 -4.62691426e-01 -1.63551986e-01 -2.80018002e-01
-6.29930854e-01 1.18972027e+00 -6.79522455e-02 1.42093480e-01
-5.74016236e-02 1.97355449e-02 5.35857677e-01 1.27532825e-01
4.77396399e-01 9.32281375e-01 -7.32090473e-01 -3.74308944e-01
-9.49179232e-01 7.16353595e-01 1.33892095e+00 3.99668872e-01
3.96560431e-01 9.53971297e-02 1.08207189e-01 7.16348886e-01
4.37323034e-01 5.82456410e-01 5.27051687e-01 -7.48374522e-01
4.75839436e-01 2.81647533e-01 3.33663553e-01 -5.81926465e-01
-5.99952638e-01 -5.91513738e-02 -5.39472401e-01 3.46833825e-01
6.97256804e-01 -3.59432071e-01 -8.95884991e-01 1.46815085e+00
3.95283043e-01 1.98578555e-02 3.24559182e-01 5.45162857e-01
6.94546759e-01 6.98470056e-01 6.28848612e-01 -5.53369701e-01
8.42625558e-01 -4.97606903e-01 -4.23509806e-01 7.68296868e-02
8.18773508e-01 -7.73946583e-01 7.26465523e-01 5.62172711e-01
-9.74376559e-01 -6.20984316e-01 -1.11568666e+00 3.70010793e-01
-5.06003439e-01 -2.03458861e-01 7.78677225e-01 1.11544359e+00
-5.28941393e-01 1.35406816e+00 -9.69623685e-01 -3.97817791e-01
-6.17829673e-02 6.49178863e-01 -2.94796884e-01 3.48372340e-01
-1.17012918e+00 1.29134667e+00 5.87686181e-01 -1.96280982e-02
-4.58644748e-01 -7.35868752e-01 -9.87689078e-01 -3.06094646e-01
9.50027034e-02 -4.59105074e-01 9.74388838e-01 -4.55075592e-01
-1.80273461e+00 4.34986919e-01 -6.94148540e-02 -3.94802243e-01
6.65999711e-01 -1.58920825e-01 -4.43549961e-01 -2.15831593e-01
-3.84940386e-01 2.92735010e-01 5.35692155e-01 -1.34232259e+00
-1.39438868e-01 -1.87246397e-01 -3.78708750e-01 -5.63669130e-02
4.40982938e-01 -1.13975242e-01 1.50332823e-01 -3.06689709e-01
5.36603518e-02 -1.00230217e+00 -4.13103402e-01 -5.63749015e-01
-1.28978148e-01 -1.80152774e-01 4.43226814e-01 -5.18716514e-01
1.33852077e+00 -1.69562781e+00 1.52929947e-01 5.65496087e-01
-1.01954177e-01 4.07872856e-01 3.73811841e-01 6.58043325e-01
-9.90856513e-02 4.30279732e-01 -5.91321766e-01 -1.25321865e-01
-9.90938395e-02 1.60647973e-01 -1.66296199e-01 4.24056351e-01
2.75765836e-01 5.24888515e-01 -7.90493369e-01 -4.20177519e-01
5.10715425e-01 4.75447297e-01 -3.99751961e-01 3.01311553e-01
-1.44959241e-01 6.83901191e-01 -3.14180702e-01 5.83974063e-01
7.70775735e-01 5.64074256e-02 9.39237475e-02 -2.02324376e-01
-1.92374066e-01 -2.74017304e-02 -1.30196726e+00 9.60209012e-01
-3.73346657e-01 2.99239516e-01 -3.10958624e-01 -7.96928287e-01
1.24615240e+00 1.16587810e-01 3.12863469e-01 -3.04471076e-01
2.39046663e-01 6.75746918e-01 4.49128658e-01 -6.04034305e-01
3.29000592e-01 -6.13424838e-01 1.99818030e-01 6.35085329e-02
1.44168228e-01 -7.42473960e-01 2.72269219e-01 -3.12311053e-01
5.42640388e-01 4.19313878e-01 5.25865078e-01 -5.17406046e-01
7.62328625e-01 -1.64880410e-01 3.76506805e-01 6.97194278e-01
-2.91847974e-01 3.80296201e-01 4.61739302e-01 -3.96257460e-01
-1.27912104e+00 -8.45786095e-01 -7.57106543e-01 4.54993814e-01
-6.22517392e-02 -3.49170089e-01 -6.13505244e-01 -2.97912210e-01
3.83769304e-01 8.81377518e-01 -4.86821353e-01 -1.55854061e-01
-5.42325079e-01 -1.41729760e+00 2.76790738e-01 4.82920468e-01
4.58177254e-02 -9.56065714e-01 -4.23787117e-01 3.61813277e-01
5.62322557e-01 -8.65098894e-01 2.12410420e-01 5.40020883e-01
-1.10446894e+00 -1.17500031e+00 -3.56870919e-01 -9.53488350e-02
3.03722411e-01 -5.31689107e-01 9.44533646e-01 1.76628698e-02
-1.30856127e-01 2.22247124e-01 -2.71705151e-01 -4.31170285e-01
-7.56847143e-01 -1.27916513e-02 3.71046215e-01 -5.36836326e-01
3.64109397e-01 -4.10164803e-01 -4.90834802e-01 4.11748022e-01
-7.85129845e-01 -5.22170782e-01 3.15432817e-01 7.35897303e-01
5.10647297e-01 -1.06160827e-01 5.73092043e-01 -1.05246460e+00
9.11446214e-01 -4.62541163e-01 -7.29456663e-01 1.89212620e-01
-8.58267784e-01 2.62721717e-01 6.69919252e-01 -5.46425104e-01
-9.28573489e-01 -2.53304183e-01 -3.21376622e-01 -2.05925182e-01
-1.48747221e-01 5.02950490e-01 1.22492097e-01 -4.23223674e-01
6.91753447e-01 -1.32999152e-01 3.43400955e-01 -4.10875648e-01
-1.81775689e-01 4.68805313e-01 -2.24372093e-02 -1.12050891e+00
4.40324605e-01 -1.83729194e-02 6.04965270e-01 -8.06865454e-01
-3.49086136e-01 -2.00186163e-01 -7.91555047e-01 -1.28064424e-01
6.38621390e-01 -5.87004185e-01 -1.07018089e+00 3.79530817e-01
-5.96434414e-01 -4.93666679e-01 -1.90910310e-01 8.14552367e-01
-6.23014569e-01 5.40696800e-01 -3.98617715e-01 -1.31541121e+00
-3.27319711e-01 -1.47576594e+00 8.33955169e-01 5.67278504e-01
-5.03109872e-01 -1.52094650e+00 2.85518289e-01 1.88069642e-01
4.12616521e-01 7.00511456e-01 9.38649654e-01 -9.95124519e-01
-1.06904000e-01 -3.33876491e-01 2.73186952e-01 3.70975345e-01
-3.80726606e-02 7.18673468e-01 -1.12967074e+00 -5.86660206e-01
1.80951636e-02 -3.24140728e-01 8.06302786e-01 6.81704938e-01
1.05711138e+00 8.84889439e-02 -3.02110672e-01 4.44566727e-01
1.56368434e+00 5.70652008e-01 5.43558300e-01 3.03034216e-01
3.86263192e-01 6.51704848e-01 7.29985654e-01 3.32558930e-01
-2.21421793e-01 4.63729203e-01 -1.56378359e-01 2.03793600e-01
3.57899964e-01 -1.35855675e-01 3.20681572e-01 6.51978493e-01
-4.49817955e-01 -1.97667718e-01 -9.89938021e-01 -2.87714601e-01
-1.57349837e+00 -8.47525597e-01 -2.22223416e-01 2.80657125e+00
7.17920899e-01 2.65475601e-01 2.39069492e-01 8.02242234e-02
3.82023633e-01 -1.88012183e-01 -6.36665881e-01 -8.70469928e-01
2.51536835e-02 3.73614877e-01 7.64021218e-01 8.47703457e-01
-1.05009806e+00 6.35456085e-01 7.60781002e+00 9.66672719e-01
-1.12362182e+00 -3.12769741e-01 9.38376606e-01 1.24691315e-01
-1.16775885e-01 6.19667955e-02 -1.18359470e+00 7.36780524e-01
1.42922413e+00 -9.64680389e-02 1.34899348e-01 7.09125280e-01
4.51023787e-01 -7.21472085e-01 -1.05267799e+00 6.98868930e-01
-2.01100633e-01 -9.20095086e-01 -2.37872168e-01 7.97108784e-02
7.79766023e-01 -3.47714216e-01 -1.52059421e-01 5.87533891e-01
1.36837125e-01 -1.24276555e+00 2.04514205e-01 7.67008364e-01
6.09813929e-01 -8.66215348e-01 1.18675482e+00 2.98062384e-01
-9.33509707e-01 1.43575355e-01 -5.41862190e-01 -6.09671175e-02
2.17405278e-02 6.74817324e-01 -7.58019567e-01 5.40383339e-01
3.39011818e-01 1.92748368e-01 -4.67407793e-01 1.34889567e+00
1.87999859e-01 6.18759990e-01 -6.81864142e-01 -4.27355111e-01
1.14237033e-01 -8.26191843e-01 3.56493145e-01 1.19713342e+00
3.67577642e-01 -3.74943092e-02 -4.84097153e-02 9.05511141e-01
4.49075162e-01 2.66927928e-01 -4.73780870e-01 4.85271476e-02
4.67364788e-01 9.73799348e-01 -7.34961271e-01 -1.07608981e-01
-1.01872556e-01 3.17669034e-01 3.08855712e-01 4.42553669e-01
-7.43570685e-01 -2.12361574e-01 5.26762366e-01 -2.04491466e-02
-1.13415057e-02 -2.29348704e-01 -1.31389707e-01 -8.60553086e-01
-2.29655266e-01 -6.72306538e-01 2.57820368e-01 -4.27568823e-01
-1.52314711e+00 3.42917144e-01 5.83189309e-01 -1.06299853e+00
-4.18726861e-01 -1.09127140e+00 -7.73589611e-01 1.22904038e+00
-1.21716356e+00 -4.96033370e-01 -1.09795064e-01 -2.85343498e-01
2.76319772e-01 -6.12970255e-02 8.04354548e-01 -6.12286180e-02
-8.35824251e-01 3.69204879e-01 5.83148062e-01 -3.91110271e-01
6.81978047e-01 -1.37773275e+00 3.38362120e-02 3.93775046e-01
-6.23426795e-01 8.98813188e-01 1.17100704e+00 -9.03898716e-01
-1.17920804e+00 -6.25900865e-01 3.52750957e-01 -5.97090960e-01
6.43577695e-01 -1.19172789e-01 -1.15425146e+00 1.88791305e-01
-4.12309110e-01 -8.39708745e-03 9.79346037e-01 2.12267563e-01
9.93418917e-02 1.84716076e-01 -1.52212751e+00 3.05321157e-01
3.96442860e-01 -2.59463996e-01 -3.97919178e-01 1.25844017e-01
3.09908122e-01 -6.47700965e-01 -1.49144089e+00 6.37078762e-01
7.02773869e-01 -1.02317655e+00 9.16126728e-01 -7.61886299e-01
1.65469989e-01 -7.20439255e-02 -1.14893205e-01 -1.27862310e+00
1.15311228e-01 -4.92810756e-01 -3.49180877e-01 1.19313788e+00
6.20274723e-01 -8.84992659e-01 7.27140248e-01 1.43764567e+00
3.21038991e-01 -1.09115398e+00 -1.00027215e+00 -9.74780142e-01
6.02752268e-01 -4.75199133e-01 4.76458490e-01 4.77371782e-01
5.70350848e-02 5.91189861e-02 -1.40748441e-01 -8.10270682e-02
6.36160493e-01 -1.56701252e-01 6.55953765e-01 -1.41834867e+00
-2.81708211e-01 -3.38860363e-01 -4.41479325e-01 -4.27970171e-01
2.54215688e-01 -5.57752490e-01 -1.45060435e-01 -9.25187945e-01
1.74908504e-01 -7.62824714e-01 -9.08285603e-02 -2.97001421e-01
-4.34702963e-01 -3.96636903e-01 1.80433840e-01 1.61243439e-01
-4.20578569e-02 4.74479735e-01 1.32598388e+00 3.21623415e-01
-4.93328452e-01 3.52650225e-01 -2.08910495e-01 5.26221335e-01
9.90434825e-01 -2.88168818e-01 -2.51601934e-01 6.54841125e-01
2.38799468e-01 1.31652951e-01 9.48362648e-02 -1.14601791e+00
-1.25559643e-01 -3.25107634e-01 6.29953265e-01 -4.15516585e-01
3.34531337e-01 -7.75357783e-01 3.60468239e-01 4.93105918e-01
-1.80304736e-01 -1.19756028e-01 5.65591097e-01 5.82397103e-01
2.72145886e-02 -9.54232216e-01 9.21818733e-01 -1.37413934e-01
-2.87399024e-01 2.05510378e-01 -2.95066118e-01 -1.41140774e-01
1.11649323e+00 -4.58136916e-01 -1.66842386e-01 -1.02175660e-01
-8.86131227e-01 1.17785029e-01 5.61791897e-01 -1.41936630e-01
2.80971706e-01 -8.71317208e-01 -3.90100390e-01 1.31044850e-01
-1.16988689e-01 -6.90666884e-02 1.37765735e-01 8.48457456e-01
-8.42707634e-01 5.39964974e-01 -5.13268672e-02 -4.67544883e-01
-1.11980700e+00 5.27327120e-01 6.51869595e-01 -3.89719188e-01
-1.13608807e-01 6.91707015e-01 -3.62765461e-01 -5.19367695e-01
-1.92926005e-01 -5.58116853e-01 -3.38902533e-01 3.90679017e-02
2.35216975e-01 6.73584342e-01 1.14886843e-01 -5.35735130e-01
-1.38528407e-01 8.71709645e-01 3.20500284e-02 4.33336683e-02
1.21313918e+00 1.15679197e-01 -1.33041374e-03 7.65920877e-01
1.11916947e+00 -6.91244453e-02 -1.26046097e+00 3.94740015e-01
4.61097658e-02 -3.03655297e-01 1.02829952e-02 -6.29706800e-01
-5.20216286e-01 7.66439438e-01 7.20541716e-01 1.93426058e-01
5.19595027e-01 -4.19512331e-01 2.25853994e-01 3.87080729e-01
1.60335124e-01 -1.27109158e+00 -4.75971520e-01 3.57718080e-01
7.02145517e-01 -1.62661254e+00 4.81737763e-01 -6.06038213e-01
-7.67866194e-01 1.57609022e+00 7.51877308e-01 -3.99008542e-02
6.74675643e-01 2.13041812e-01 -3.03633511e-01 2.49986351e-02
-5.72913885e-01 5.74470945e-02 3.49655151e-01 6.12308145e-01
7.57779658e-01 2.04387531e-01 -5.90157330e-01 3.36038709e-01
-1.83751717e-01 -3.00979558e-02 4.46068048e-01 8.80351782e-01
-4.40975040e-01 -1.27548945e+00 -5.68453550e-01 7.88859904e-01
-2.86576360e-01 2.99261268e-02 -4.80098128e-02 1.15351391e+00
-9.13046449e-02 9.02233779e-01 3.89475562e-02 -2.35588491e-01
1.77699894e-01 3.42420667e-01 5.54787934e-01 -2.97669470e-01
-5.60409248e-01 -1.68455020e-01 2.88003892e-01 -2.04588950e-01
-4.96333718e-01 -7.95618713e-01 -1.11036277e+00 -5.81027091e-01
-6.65022314e-01 5.12436092e-01 7.01852798e-01 9.90478158e-01
-1.29652709e-01 1.58836231e-01 4.22320306e-01 -1.05687237e+00
-7.31213927e-01 -1.06555355e+00 -7.44852841e-01 4.77288663e-01
1.44472554e-01 -9.83822048e-01 -9.34153557e-01 -3.40232760e-01] | [6.192411422729492, 3.679269790649414] |
d21d87e3-de7d-486c-9adc-3e0d6a3b268d | extending-compositional-attention-networks | 2210.01191 | null | https://arxiv.org/abs/2210.01191v1 | https://arxiv.org/pdf/2210.01191v1.pdf | Extending Compositional Attention Networks for Social Reasoning in Videos | We propose a novel deep architecture for the task of reasoning about social interactions in videos. We leverage the multi-step reasoning capabilities of Compositional Attention Networks (MAC), and propose a multimodal extension (MAC-X). MAC-X is based on a recurrent cell that performs iterative mid-level fusion of input modalities (visual, auditory, text) over multiple reasoning steps, by use of a temporal attention mechanism. We then combine MAC-X with LSTMs for temporal input processing in an end-to-end architecture. Our ablation studies show that the proposed MAC-X architecture can effectively leverage multimodal input cues using mid-level fusion mechanisms. We apply MAC-X to the task of Social Video Question Answering in the Social IQ dataset and obtain a 2.5% absolute improvement in terms of binary accuracy over the current state-of-the-art. | ['Alexandros Potamianos', 'Georgios Paraskevopoulos', 'Christina Sartzetaki'] | 2022-10-03 | null | null | null | null | ['video-question-answering'] | ['computer-vision'] | [ 4.23405111e-01 1.84906557e-01 2.73627639e-01 -4.53117937e-01
-9.78162646e-01 -3.32028508e-01 7.51707673e-01 2.05263682e-02
-5.76231182e-01 3.11351091e-01 6.46933794e-01 -3.87558579e-01
-4.67046201e-02 -3.50724548e-01 -8.54477167e-01 -2.52169400e-01
1.31462306e-01 -7.52472505e-02 2.44097501e-01 -1.64439783e-01
6.71342388e-02 5.95688559e-02 -1.66470635e+00 1.22591984e+00
4.86308753e-01 1.31652308e+00 -1.09069869e-01 1.25714123e+00
3.78601290e-02 1.80424285e+00 -3.62282902e-01 -6.87200427e-01
-3.03337872e-01 -6.11112177e-01 -1.16877532e+00 -3.65154207e-01
6.69588685e-01 -7.31200933e-01 -7.58663833e-01 4.70846862e-01
5.12860000e-01 4.64403152e-01 5.82261682e-01 -1.33741677e+00
-6.37555420e-01 8.97296488e-01 -3.15769315e-01 2.68088371e-01
7.44631350e-01 2.38777593e-01 1.08900118e+00 -7.90107250e-01
4.73702163e-01 1.58634067e+00 6.15570784e-01 6.12543404e-01
-1.09541047e+00 -4.88671958e-01 2.42662266e-01 4.89695013e-01
-8.70201707e-01 -6.21394217e-01 4.75383222e-01 -2.93925643e-01
1.36041677e+00 2.05040678e-01 3.95814806e-01 1.39013517e+00
1.85624480e-01 1.38052404e+00 8.47799182e-01 -7.90779144e-02
8.31363872e-02 -5.90116858e-01 2.86053330e-01 9.99706089e-01
-6.86089277e-01 -3.71894717e-01 -1.08794236e+00 9.17179219e-04
5.34536898e-01 2.30968952e-01 -5.12483232e-02 2.10646451e-01
-1.41987729e+00 4.72439498e-01 9.21102405e-01 1.58528313e-01
-4.92666572e-01 7.86846459e-01 5.92172861e-01 4.36724931e-01
1.65473372e-01 4.38139200e-01 -2.93086231e-01 -2.96691298e-01
-9.28550601e-01 2.65702784e-01 4.97851312e-01 4.78661746e-01
2.02510133e-01 -2.28271574e-01 -7.69843578e-01 7.00709939e-01
4.74131376e-01 3.82675618e-01 1.87432274e-01 -1.77311730e+00
6.04331613e-01 5.84188163e-01 -1.65329263e-01 -6.54613793e-01
-6.63222551e-01 -1.31047040e-01 -6.03204012e-01 3.49725001e-02
3.35174799e-01 -3.10289860e-01 -9.31914568e-01 2.15067554e+00
-5.02275750e-02 2.05136538e-01 3.33204031e-01 8.53744507e-01
1.26264811e+00 6.45202875e-01 3.78186613e-01 1.22492939e-01
1.38871968e+00 -1.22318912e+00 -6.76153839e-01 -6.54623732e-02
3.71176571e-01 -2.55104184e-01 1.00056958e+00 2.68459380e-01
-1.66706717e+00 -5.29952645e-01 -8.60873699e-01 -5.29073358e-01
-2.60683030e-01 8.89611021e-02 5.95042050e-01 -8.19329694e-02
-1.40411723e+00 5.17897248e-01 -8.17453086e-01 -5.89720845e-01
5.59542775e-01 4.36489701e-01 -2.79337376e-01 9.19637308e-02
-1.18814373e+00 6.62734866e-01 -9.70422998e-02 2.00717032e-01
-9.02637303e-01 -5.41134834e-01 -9.90266263e-01 3.53558242e-01
5.40532947e-01 -1.29982698e+00 1.70454097e+00 -1.07040381e+00
-1.63298810e+00 7.50494063e-01 -3.75512332e-01 -5.48416376e-01
2.73879051e-01 -5.15471518e-01 -1.93972304e-01 8.84388089e-01
-1.55578420e-01 1.12548912e+00 9.97371852e-01 -9.07167554e-01
-4.74783987e-01 -9.23631415e-02 5.82300603e-01 2.41837725e-01
-8.19192678e-02 1.05041064e-01 -6.47457540e-01 -6.28702939e-01
-1.46104217e-01 -8.71739388e-01 7.80014396e-02 4.47311886e-02
-1.82809934e-01 -3.16843927e-01 6.44086480e-01 -8.61149549e-01
1.16737330e+00 -2.17362714e+00 6.77661657e-01 -2.43833840e-01
4.75885004e-01 1.37212142e-01 -5.53403914e-01 3.80391240e-01
-4.13949639e-02 -3.49910610e-04 -2.49919876e-01 -9.40061271e-01
1.65730283e-01 -7.92345777e-02 -5.57335973e-01 -8.89225826e-02
5.69649041e-01 1.53451538e+00 -8.85183990e-01 -4.42043275e-01
6.26617372e-02 5.86237490e-01 -7.43903995e-01 3.92402351e-01
-5.46055496e-01 9.27302092e-02 -1.40203640e-01 7.81817436e-01
1.29963592e-01 -4.78110224e-01 4.78254482e-02 -3.51169288e-01
3.15704048e-01 1.98711067e-01 -4.00728285e-01 2.14944863e+00
-3.54643613e-01 9.37176168e-01 1.99926153e-01 -6.17838979e-01
9.71246883e-02 5.56221902e-01 1.65908605e-01 -7.80713081e-01
4.60020065e-01 -2.42098808e-01 -1.10890903e-01 -7.06934392e-01
5.39198458e-01 1.99160367e-01 -2.23885894e-01 7.34086812e-01
5.45907617e-01 2.45323703e-01 8.46153051e-02 8.41622829e-01
1.48061466e+00 4.21778202e-01 -2.40109682e-01 2.77559131e-01
4.87222642e-01 -5.92791736e-01 -2.20944062e-02 9.25936282e-01
-2.38611236e-01 5.64416051e-01 9.22014654e-01 -1.66210998e-02
-6.77923679e-01 -1.10257888e+00 5.08813679e-01 1.77577746e+00
-1.04268111e-01 -4.19443935e-01 -8.61404359e-01 -8.10935557e-01
-4.36680429e-02 6.65741920e-01 -9.54324782e-01 -3.50037336e-01
-5.32124996e-01 -1.03302233e-01 9.32652473e-01 8.42258990e-01
6.88788772e-01 -1.41051912e+00 -9.53248739e-01 8.83454978e-02
-4.18199718e-01 -1.46566963e+00 -3.34172279e-01 -1.55674860e-01
-4.24211770e-01 -1.03539014e+00 -7.79706657e-01 -4.04124975e-01
3.04643631e-01 1.95170373e-01 1.18628657e+00 1.80383980e-01
9.61526036e-02 8.75223398e-01 -3.54691327e-01 9.71361399e-02
-1.18598849e-01 6.87947422e-02 -2.51625627e-01 2.96337664e-01
-9.65693668e-02 -4.22188103e-01 -6.04874730e-01 9.70911384e-02
-1.01263165e+00 3.43177766e-01 6.28493726e-01 8.60397160e-01
-1.66302081e-02 -7.57761061e-01 7.64072955e-01 -3.73807251e-01
7.03148961e-01 -5.20335019e-01 -4.43501547e-02 4.08934057e-01
3.00828904e-01 2.45238751e-01 2.72750884e-01 -4.63688433e-01
-1.18875539e+00 -2.69967109e-01 -2.12467998e-01 -7.81473637e-01
-7.79278530e-03 6.80424690e-01 1.90176979e-01 1.13850705e-01
2.58464187e-01 -1.48502588e-01 3.91357727e-02 -1.81912966e-02
6.07532680e-01 4.34965909e-01 9.70496655e-01 -5.32431304e-01
6.92107752e-02 3.77829403e-01 8.85703042e-02 -4.88006800e-01
-1.01664913e+00 -5.52498959e-02 -2.40974888e-01 -6.26015723e-01
1.41720450e+00 -9.43697155e-01 -1.74212050e+00 5.01965344e-01
-1.21830177e+00 -7.46374965e-01 5.03258407e-02 2.92932361e-01
-7.44218588e-01 1.86937436e-01 -1.23038161e+00 -9.31259394e-01
-4.90914732e-01 -1.12139595e+00 1.27723491e+00 2.70743072e-01
-3.57647896e-01 -6.58392429e-01 -3.67919743e-01 9.19465661e-01
4.01722014e-01 7.81076998e-02 9.18923616e-01 -3.76041710e-01
-6.79856420e-01 1.07724011e-01 -6.34320796e-01 -8.78518540e-03
-5.80078900e-01 9.65341479e-02 -1.23945892e+00 -8.88956338e-02
-3.82957280e-01 -9.25936460e-01 1.51909280e+00 2.62664616e-01
1.21906412e+00 4.68754135e-02 2.13630348e-02 3.67185026e-01
7.37870932e-01 -9.21650678e-02 6.04370236e-01 -2.05933884e-01
6.51573598e-01 4.75948304e-01 3.52299392e-01 1.78457037e-01
8.18953156e-01 4.46616828e-01 6.96410418e-01 -8.23794603e-02
-2.24655539e-01 -1.25038564e-01 7.47038066e-01 5.54060340e-01
7.39555061e-02 -3.41949165e-01 -9.54568386e-01 5.00706673e-01
-2.38965225e+00 -1.32548964e+00 3.00867677e-01 1.62262595e+00
4.73434716e-01 3.98457013e-02 2.51771182e-01 7.49783069e-02
3.73172820e-01 1.28778175e-01 -6.71550751e-01 -5.67211151e-01
1.94215868e-02 2.38602534e-02 -1.06775187e-01 5.08817255e-01
-1.19165993e+00 8.81964266e-01 6.94696093e+00 4.64398861e-01
-1.08338833e+00 1.61975846e-01 5.93094766e-01 -5.21951139e-01
-1.40959755e-01 -3.69171888e-01 -2.85056323e-01 1.43407896e-01
1.15834510e+00 4.48313117e-01 7.39796162e-01 -7.51130208e-02
-1.03279963e-01 -3.69649082e-01 -1.31435931e+00 1.13789248e+00
4.65967655e-01 -1.32327271e+00 6.69235662e-02 -4.29679751e-01
4.94311780e-01 1.99661955e-01 3.98074687e-01 7.21633673e-01
2.36308292e-01 -1.12796175e+00 9.28644121e-01 1.05316210e+00
7.04209507e-01 -6.05037689e-01 5.84747434e-01 1.30192712e-01
-1.05373931e+00 -5.59715211e-01 4.71455246e-01 -1.62372485e-01
4.40472364e-01 -1.17296331e-01 -3.92916441e-01 5.27230263e-01
8.16076398e-01 7.40400672e-01 -7.76158512e-01 4.87400174e-01
-1.69378906e-01 5.92056334e-01 -2.18272611e-01 8.43377113e-02
3.89231205e-01 4.20265228e-01 3.72258127e-01 1.18759072e+00
6.03733622e-02 3.31582099e-01 -3.04472744e-01 7.29113936e-01
-5.29577494e-01 -4.80195850e-01 -3.00550699e-01 -3.58382732e-01
6.72016516e-02 1.08575261e+00 -4.38120872e-01 -7.63179302e-01
-4.74486858e-01 1.14272439e+00 6.21503711e-01 7.30943501e-01
-1.10076833e+00 -3.39411139e-01 5.60381353e-01 -3.81202430e-01
4.37952876e-01 -1.52906120e-01 -6.66401982e-02 -1.20803380e+00
-7.23003671e-02 -9.10486877e-01 6.53461754e-01 -1.53113365e+00
-1.21019506e+00 6.68059409e-01 -3.10456883e-02 -7.40857661e-01
-4.80004698e-01 -5.02750635e-01 -4.68973875e-01 4.61885333e-01
-1.20149004e+00 -1.55205584e+00 -3.23848695e-01 7.68364608e-01
5.24022102e-01 -1.13515340e-01 6.92779958e-01 2.35547096e-01
-4.52586144e-01 5.83863199e-01 -5.38128138e-01 1.99855343e-01
6.93926156e-01 -1.10625362e+00 2.79062420e-01 6.74042404e-01
-1.40109584e-01 6.56946003e-01 2.48718485e-01 -3.40987116e-01
-1.54924929e+00 -9.74648237e-01 8.14982057e-01 -6.06874347e-01
6.37891531e-01 -3.64407301e-01 -5.99738538e-01 9.02824998e-01
8.17438245e-01 -1.66225061e-01 7.66896367e-01 1.37487546e-01
-7.70815969e-01 3.66308503e-02 -1.06794345e+00 9.73308802e-01
9.69071507e-01 -1.21946263e+00 -8.89130414e-01 -1.71757743e-01
1.07630980e+00 -2.19080836e-01 -8.00547719e-01 5.37573159e-01
8.18630040e-01 -1.04424345e+00 9.30572569e-01 -8.27806592e-01
8.61029744e-01 -1.66141450e-01 -3.17431718e-01 -7.17820704e-01
-1.41497090e-01 -8.05734754e-01 -4.94897455e-01 9.75912929e-01
2.93401152e-01 -2.36611828e-01 3.01084965e-01 6.96302772e-01
-6.96818382e-02 -8.18059444e-01 -8.74085188e-01 3.50746252e-02
-3.29644740e-01 -6.88698113e-01 2.60605603e-01 6.74082279e-01
4.13593888e-01 8.82833958e-01 -3.04437995e-01 -1.39753044e-01
3.30278546e-01 3.65644619e-02 4.96341646e-01 -7.22846150e-01
-3.94028574e-01 -7.55263150e-01 -1.51530057e-01 -1.15004051e+00
5.44669688e-01 -8.04871559e-01 -3.71702723e-02 -1.75872135e+00
3.56578708e-01 4.77451533e-01 -5.29754817e-01 7.56723344e-01
-2.80133575e-01 6.36799157e-01 7.49843299e-01 -2.05602020e-01
-1.52729094e+00 6.22394502e-01 9.40823019e-01 -2.72037715e-01
1.89377517e-02 -4.82893854e-01 -6.11193836e-01 7.12643266e-01
4.52438086e-01 9.35660005e-02 -2.84652501e-01 -7.79890656e-01
7.00139344e-01 6.34123564e-01 7.03388929e-01 -1.13427365e+00
4.75934625e-01 3.78827065e-01 3.75046819e-01 -7.86945701e-01
8.99310291e-01 -5.19720852e-01 -3.68197709e-01 2.72752434e-01
-1.01178169e+00 1.13501631e-01 3.23146850e-01 4.30119544e-01
-2.41604686e-01 3.29071999e-01 4.37761664e-01 1.78502992e-01
-4.84527230e-01 1.28651718e-02 -7.84442902e-01 -2.01624826e-01
7.44704127e-01 3.32367390e-01 -5.61691344e-01 -8.55829537e-01
-1.09410572e+00 5.75504303e-01 6.47825748e-02 4.93066311e-01
7.89860666e-01 -1.36479104e+00 -5.41256607e-01 -4.33732942e-02
-3.87703888e-02 -4.93999422e-01 4.96326566e-01 1.13224685e+00
-1.77562773e-01 4.53640401e-01 -9.16963965e-02 -5.70223093e-01
-1.38235915e+00 5.35397708e-01 3.49867284e-01 -1.58403903e-01
-2.06838250e-01 8.87664258e-01 1.08094350e-01 -2.04915747e-01
4.53221947e-01 -5.79174221e-01 -2.71176398e-01 1.49844795e-01
5.54409504e-01 2.55830765e-01 -2.53366917e-01 -5.11818051e-01
-3.80397558e-01 4.63254392e-01 1.97058797e-01 -6.03936851e-01
1.04685354e+00 -2.10465312e-01 -1.33205235e-01 7.22983718e-01
1.20928133e+00 -4.43609536e-01 -1.20356023e+00 -2.32837871e-01
-2.36535668e-01 8.70827362e-02 -6.56756246e-03 -1.17823005e+00
-9.12015736e-01 1.11242795e+00 2.62321234e-01 -3.38526331e-02
1.35576439e+00 1.65852219e-01 7.53128648e-01 7.66875863e-01
-2.38450661e-01 -7.49850571e-01 5.50428808e-01 9.13338900e-01
1.09149623e+00 -1.06773293e+00 -2.98547894e-01 1.45543784e-01
-8.16073358e-01 9.78813827e-01 5.35396278e-01 7.01481625e-02
4.02082264e-01 1.10933617e-01 -5.98206781e-02 -2.53403455e-01
-1.56478691e+00 -4.39798832e-01 3.66335183e-01 -9.17974412e-02
4.14247781e-01 -2.93876916e-01 1.39027521e-01 8.37797523e-01
3.19387704e-01 2.85543829e-01 5.20645902e-02 9.50383842e-01
-5.87600917e-02 -4.15975779e-01 -2.14873478e-01 3.75016928e-01
-5.11681855e-01 -1.79473072e-01 -5.13954580e-01 1.68147489e-01
-1.10918336e-01 1.20675671e+00 3.07215631e-01 -6.26980424e-01
1.27984554e-01 3.69173408e-01 6.89498961e-01 9.58975405e-04
-1.03941107e+00 -4.52489890e-02 2.92015672e-01 -1.05457366e+00
-6.93261981e-01 -5.62584639e-01 -1.44088304e+00 -2.59857088e-01
2.57916808e-01 -1.96693480e-01 4.09180433e-01 1.07803583e+00
6.87485933e-01 9.73771214e-01 1.03180856e-01 -1.13442922e+00
-7.89826140e-02 -9.94565606e-01 2.24796429e-01 3.91674906e-01
7.81367600e-01 -4.46997851e-01 -1.01459645e-01 1.12152383e-01] | [10.42888069152832, 1.0941272974014282] |
b6a68ea2-f559-45ae-8ee2-6d1624aa24bf | x2face-a-network-for-controlling-face-1 | null | null | http://openaccess.thecvf.com/content_ECCV_2018/html/Olivia_Wiles_X2Face_A_network_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Olivia_Wiles_X2Face_A_network_ECCV_2018_paper.pdf | X2Face: A network for controlling face generation using images, audio, and pose codes | The objective of this paper is a neural network model that controls the pose and expression of a given face, using another face or modality (e.g. audio). This model can then be used for lightweight, sophisticated video and image editing. We make the following three contributions. First, we introduce a network, X2Face, that can control a source face (specified by one or more frames) using another face in a driving frame to produce a generated frame with the identity of the source frame but the pose and expression of the face in the driving frame. Second, we propose a method for training the network fully self-supervised using a large collection of video data. Third, we show that the generation process can be driven by other modalities, such as audio or pose codes, without any further training of the network. The generation results for driving a face with another face are compared to state-of-the-art self-supervised/supervised methods. We show that our approach is more robust than other methods, as it makes fewer assumptions about the input data. We also show examples of using our framework for video face editing. | ['Andrew Zisserman', 'Olivia Wiles', 'A. Sophia Koepke'] | 2018-09-01 | null | null | null | eccv-2018-9 | ['talking-head-generation'] | ['computer-vision'] | [ 7.26649284e-01 6.10063851e-01 2.02662334e-01 -7.34156072e-01
-3.81190926e-01 -5.66524386e-01 7.90009856e-01 -6.62437499e-01
-1.99558496e-01 6.33537471e-01 5.29516041e-02 2.74341047e-01
4.99537557e-01 -6.24782205e-01 -1.26049495e+00 -7.25391865e-01
5.50032444e-02 2.96155840e-01 9.65434238e-02 -3.45154941e-01
-9.24326479e-02 8.02473068e-01 -2.13442063e+00 4.43007737e-01
-3.94519642e-02 1.16196132e+00 3.48817417e-03 9.67749536e-01
2.37611532e-01 8.89984548e-01 -6.92811310e-01 -2.18547970e-01
3.10443074e-01 -5.95683813e-01 -7.04591990e-01 4.84552443e-01
9.76445079e-01 -7.19330788e-01 -2.89535165e-01 7.87866652e-01
6.18968964e-01 2.34506968e-02 4.73419070e-01 -1.54536092e+00
-2.93378294e-01 4.14762944e-01 -2.39415929e-01 -4.47116822e-01
7.92987943e-01 -7.19394758e-02 2.28137985e-01 -8.96852076e-01
1.08292162e+00 1.45850611e+00 5.52358687e-01 1.15226090e+00
-1.18532515e+00 -8.33111286e-01 -1.12574017e-02 -1.20901279e-01
-1.41431069e+00 -1.19282448e+00 8.43299568e-01 -3.82348746e-01
5.13883710e-01 2.72831529e-01 7.59218395e-01 1.27288508e+00
-1.59859538e-01 4.61650699e-01 8.54277670e-01 -8.61010730e-01
1.38533324e-01 2.80033767e-01 -5.99014223e-01 9.31516171e-01
-5.72417140e-01 3.98768127e-01 -7.75056720e-01 -9.51894149e-02
1.12362945e+00 -4.50423658e-01 -3.49200606e-01 -3.95444304e-01
-1.11959302e+00 7.62229145e-01 -7.75955676e-04 1.84393242e-01
-7.07653835e-02 5.36514401e-01 2.90642858e-01 5.22725046e-01
4.23328608e-01 4.84532863e-02 -2.58341849e-01 -3.28129791e-02
-1.24835646e+00 2.36357823e-01 8.94845605e-01 1.11740530e+00
9.44087386e-01 2.86608785e-01 -1.21031046e-01 6.45119905e-01
3.27642977e-01 5.84066093e-01 2.48858958e-01 -1.54922116e+00
-7.94384349e-03 -1.27097458e-01 2.93787792e-02 -9.78246033e-01
-1.11916013e-01 4.13864017e-01 -5.25669336e-01 6.97809100e-01
1.64119691e-01 -5.91157734e-01 -8.49120557e-01 2.10971856e+00
4.74361241e-01 2.97187805e-01 3.42445411e-02 5.76573908e-01
9.00036037e-01 7.08866298e-01 -2.82481551e-01 -5.33225775e-01
9.89819765e-01 -7.89000273e-01 -9.33231473e-01 1.03090964e-01
1.67227089e-01 -9.09384310e-01 3.97822320e-01 4.12311614e-01
-1.44570291e+00 -9.10466194e-01 -8.76558304e-01 6.25177845e-02
-1.91532478e-01 4.73748118e-01 4.36724901e-01 6.03080988e-01
-1.70843661e+00 6.77957177e-01 -7.16279387e-01 -6.49390221e-01
1.36522800e-01 7.27073610e-01 -7.73997724e-01 2.79748350e-01
-1.02417195e+00 7.05736697e-01 2.15601310e-01 -9.50063318e-02
-1.15399659e+00 -4.99583006e-01 -1.04534113e+00 -2.42011979e-01
2.92241186e-01 -7.09014535e-01 1.54250014e+00 -1.95113254e+00
-2.00003695e+00 1.03516400e+00 -2.31693372e-01 -2.97021955e-01
7.14750588e-01 -1.31461427e-01 -4.36392903e-01 6.79143310e-01
-2.50266716e-02 1.60346889e+00 1.59149551e+00 -1.48046649e+00
-5.38121581e-01 -2.19495580e-01 2.88245052e-01 5.44435754e-02
-2.42041647e-01 4.38747555e-01 -6.97849751e-01 -7.12107420e-01
-4.67096955e-01 -1.20445979e+00 3.10716033e-01 6.81000412e-01
-4.68984514e-01 -8.39545056e-02 1.37062776e+00 -3.70749891e-01
7.86088169e-01 -2.33211803e+00 2.53982216e-01 2.86165655e-01
-1.83044031e-01 3.05525661e-01 -2.82814205e-01 4.54668730e-01
-5.83950698e-01 -8.58681500e-02 -4.20686826e-02 -7.04388499e-01
-3.27895641e-01 3.04521531e-01 -3.77486318e-01 4.98014271e-01
3.81865919e-01 4.53731954e-01 -6.02620006e-01 -5.74234664e-01
2.69548535e-01 1.04936922e+00 -5.76863289e-01 5.01401603e-01
-1.88915774e-01 7.11745203e-01 1.27555251e-01 3.67450327e-01
4.94293690e-01 2.93885648e-01 1.08832149e-02 -5.55392921e-01
-2.50761807e-01 -2.20650062e-01 -1.47777951e+00 1.69860625e+00
-4.90354925e-01 8.55787277e-01 5.04167676e-01 -6.56872928e-01
8.29998851e-01 8.45230758e-01 4.69013155e-01 6.68731034e-02
3.10688525e-01 4.60838750e-02 -3.88505995e-01 -6.28802955e-01
1.94223568e-01 -3.51314135e-02 3.04248005e-01 5.73484063e-01
3.79896402e-01 -4.98292476e-01 4.78885174e-01 1.89846918e-01
6.55912578e-01 5.85443914e-01 7.28801936e-02 1.96470562e-02
6.67116642e-01 -3.97969335e-01 2.30150074e-01 4.84115660e-01
1.13993607e-01 8.32856953e-01 3.83310914e-01 -3.92647564e-01
-1.08403552e+00 -6.18765473e-01 1.18781269e-01 1.26195335e+00
-2.20364988e-01 -4.31587040e-01 -1.25985146e+00 -3.87563050e-01
-1.52787924e-01 3.10873896e-01 -7.86452770e-01 -7.84590542e-02
-5.95798612e-01 1.27070934e-01 7.15352774e-01 5.07252395e-01
4.04421329e-01 -1.29789388e+00 -5.53980649e-01 -1.23903221e-02
-2.71442719e-02 -1.14957643e+00 -6.91792011e-01 -1.03848413e-01
-5.86266398e-01 -9.70013499e-01 -6.29190624e-01 -1.01021552e+00
1.07356703e+00 -8.68998691e-02 8.62774849e-01 2.73835391e-01
-8.22147876e-02 9.11929667e-01 -2.58512735e-01 -3.76097620e-01
-7.90823817e-01 -3.91824573e-01 3.68055433e-01 5.38726628e-01
-2.04096913e-01 -6.04527771e-01 -2.60831505e-01 3.04624945e-01
-1.14213741e+00 1.98432922e-01 1.56326238e-02 6.01087749e-01
3.93261433e-01 -1.15756877e-01 3.01638126e-01 -9.53777790e-01
2.32095405e-01 6.25775084e-02 -5.39270520e-01 -1.14219217e-02
1.18441403e-01 -1.10333122e-01 5.17995477e-01 -5.83482742e-01
-1.10173357e+00 9.56765413e-01 -1.16153382e-01 -5.83666027e-01
-4.40092057e-01 -1.43553793e-01 -2.16883808e-01 -5.00894547e-01
7.10348487e-01 -5.88546619e-02 5.52620888e-01 -1.03826620e-01
4.69466686e-01 6.22157693e-01 7.77023852e-01 -3.99437487e-01
9.55484509e-01 8.32432330e-01 2.09972635e-03 -9.33460295e-01
-5.22671282e-01 8.29224065e-02 -9.19694901e-01 -6.52506053e-01
7.09081531e-01 -1.00314128e+00 -8.34593654e-01 7.20958769e-01
-1.53781116e+00 -3.47502053e-01 -2.10699156e-01 2.75408059e-01
-9.84182954e-01 1.05773874e-01 -4.43794817e-01 -7.36443639e-01
-2.34127045e-02 -1.12255752e+00 1.47497475e+00 2.06848904e-01
-2.65906781e-01 -8.82890463e-01 -4.96844426e-02 1.36920527e-01
3.23997051e-01 3.96959573e-01 2.28417382e-01 -2.21887127e-01
-2.91892648e-01 -1.29745707e-01 1.49589956e-01 5.85307956e-01
2.59703547e-01 8.10189009e-01 -1.20606470e+00 -3.77660632e-01
-1.23480313e-01 -5.90626001e-01 4.08763081e-01 1.88670412e-01
1.24972713e+00 -4.80561972e-01 -3.19293261e-01 7.38316119e-01
1.05535650e+00 2.91586697e-01 7.36285388e-01 -7.61535019e-02
6.27832294e-01 7.34674037e-01 3.81488711e-01 3.31939608e-01
1.38205096e-01 7.93948531e-01 4.03400600e-01 -2.75653780e-01
-2.89801210e-01 -2.52356708e-01 7.03264236e-01 1.53475732e-01
-3.31039548e-01 -6.61091506e-03 -3.33764493e-01 2.93252796e-01
-1.57261133e+00 -1.18873024e+00 2.86847711e-01 2.05441618e+00
1.04381335e+00 -3.89100790e-01 2.75971800e-01 3.00237566e-01
1.14483833e+00 -1.23760141e-01 -3.62496793e-01 -4.18323457e-01
-2.52550859e-02 3.88272226e-01 1.52511492e-01 7.08627641e-01
-1.25700760e+00 8.86579931e-01 7.35430336e+00 4.70344037e-01
-1.49533534e+00 -8.76306742e-02 5.02188683e-01 -1.78774133e-01
-8.07757583e-03 -2.70449251e-01 -7.51561105e-01 1.96761563e-01
8.33987713e-01 1.00942604e-01 7.77725041e-01 7.72376537e-01
4.16855454e-01 -5.10110818e-02 -1.62660182e+00 9.93502140e-01
7.45946348e-01 -1.35304332e+00 6.73620552e-02 -1.78212315e-01
6.84285402e-01 -5.67793250e-01 -1.45943258e-02 -5.62702902e-02
9.72332656e-02 -9.27152038e-01 1.07754648e+00 4.89073962e-01
1.40678048e+00 -7.17577457e-01 2.33945161e-01 1.51293099e-01
-9.61748660e-01 6.99805245e-02 -3.76149416e-02 3.95082273e-02
1.65234357e-01 1.69764578e-01 -5.75301468e-01 5.99578544e-02
6.14060760e-01 5.45041680e-01 -3.76412600e-01 4.39207524e-01
-2.96370775e-01 3.51072431e-01 -4.07474697e-01 4.30435866e-01
-1.39530480e-01 2.17251047e-01 4.79927093e-01 1.10725451e+00
4.51766163e-01 8.57120007e-03 5.92758171e-02 6.04390085e-01
-2.00953409e-01 3.00227143e-02 -9.50871289e-01 1.41103655e-01
3.11950445e-01 1.36577868e+00 -3.39864373e-01 -4.78464276e-01
-2.23344430e-01 1.17423809e+00 4.29706313e-02 2.79101700e-01
-8.10894012e-01 -3.75505269e-01 3.30678940e-01 2.81736016e-01
4.84330148e-01 9.76781547e-02 5.74271262e-01 -8.91345024e-01
-8.64570662e-02 -9.23277617e-01 -1.78862829e-02 -1.36407328e+00
-8.38227212e-01 9.41234946e-01 1.73316285e-01 -1.23180687e+00
-9.13426816e-01 -6.39820993e-01 -6.42972767e-01 5.03766954e-01
-1.32387757e+00 -1.37901545e+00 -3.91039550e-01 9.87977386e-01
2.42719024e-01 -4.21442211e-01 9.91103232e-01 3.27567279e-01
-1.87608346e-01 5.48610091e-01 -4.64941829e-01 2.29209304e-01
1.05770409e+00 -8.84202600e-01 6.52193427e-02 4.42344546e-01
1.42686740e-01 5.10363221e-01 7.84865022e-01 -3.90179545e-01
-1.42659760e+00 -1.14654660e+00 8.41682494e-01 -2.76946932e-01
3.11788708e-01 -5.33210337e-01 -4.21253979e-01 9.10677314e-01
4.47749346e-01 3.16612870e-01 6.08785033e-01 -6.11239016e-01
-2.24985436e-01 -3.26794535e-01 -1.41705739e+00 3.81092578e-01
9.16024268e-01 -6.83934510e-01 -2.80819088e-01 4.14441645e-01
6.63257658e-01 -7.49505401e-01 -5.64223230e-01 2.25816831e-01
6.99029505e-01 -1.18459702e+00 7.53525913e-01 -3.85259956e-01
5.22318423e-01 -4.56261575e-01 -5.91656845e-03 -1.29443920e+00
6.29788339e-02 -1.02157116e+00 -3.83907147e-02 1.53087151e+00
1.73794702e-01 -3.04102868e-01 5.64110219e-01 5.11996329e-01
9.84972045e-02 -3.26779097e-01 -9.17147994e-01 -3.71794015e-01
-3.96571606e-01 -3.39544773e-01 4.97060388e-01 7.71010697e-01
-2.00009942e-01 2.11842880e-01 -7.17108428e-01 7.33597577e-02
3.42838258e-01 -2.92508245e-01 1.09041882e+00 -1.01362956e+00
-8.55895579e-02 2.26388842e-01 -6.19713664e-01 -7.44631112e-01
5.14716983e-01 -5.11226416e-01 1.55800104e-01 -1.01525784e+00
-2.41873935e-01 2.36578076e-03 5.12909055e-01 7.66020954e-01
3.99043977e-01 7.27182567e-01 4.29087132e-01 -4.30879034e-02
-1.41755432e-01 3.58997464e-01 1.11989450e+00 7.62937963e-02
-9.75441411e-02 -1.26283377e-01 -4.65324134e-01 9.00126100e-01
5.95985115e-01 -2.57689953e-01 -4.14881945e-01 -3.87712508e-01
1.15867779e-01 2.54449189e-01 5.50472200e-01 -1.13366616e+00
2.87772089e-01 1.39694378e-01 6.31431878e-01 -2.85925597e-01
6.57735109e-01 -1.10222292e+00 5.83361626e-01 1.53495282e-01
-4.47665125e-01 7.51647800e-02 1.78808317e-01 1.92214072e-01
-2.73560703e-01 -2.78529584e-01 8.61016810e-01 -9.31114331e-02
-3.73685747e-01 2.41230920e-01 -4.81937319e-01 -4.99055564e-01
1.20143998e+00 -2.99280286e-01 1.18783578e-01 -1.02448559e+00
-1.06210446e+00 -2.09882990e-01 5.31272054e-01 4.84059572e-01
5.04215181e-01 -1.64962018e+00 -5.99511266e-01 7.06658125e-01
-7.62098208e-02 -1.53639197e-01 -5.94667904e-02 3.82958770e-01
-5.43738425e-01 3.18866633e-02 -3.85368109e-01 -7.67174721e-01
-1.67436182e+00 3.60992312e-01 4.34920907e-01 5.66237032e-01
-1.69466063e-01 7.17482924e-01 1.17827289e-01 -3.31887662e-01
3.29548329e-01 1.40833616e-01 -1.52005821e-01 1.40021414e-01
8.69694710e-01 4.58025262e-02 -1.43563345e-01 -1.20746517e+00
-1.24233723e-01 7.48816431e-01 2.23246217e-01 -4.96340901e-01
1.33246756e+00 -1.54922204e-03 -3.54947776e-01 3.30323726e-01
1.25672150e+00 8.15368220e-02 -1.50146508e+00 6.04705848e-02
-9.18451428e-01 -4.71625328e-01 -1.71088651e-02 -5.19728363e-01
-1.48470616e+00 6.36110067e-01 6.10050440e-01 7.57019520e-02
1.28912377e+00 -3.54937240e-02 3.97188753e-01 3.35634530e-01
2.97802508e-01 -1.31743455e+00 1.47054583e-01 3.54735583e-01
1.25311279e+00 -9.50695336e-01 -2.28635594e-01 -6.06332660e-01
-5.98107636e-01 1.50475705e+00 6.29218161e-01 -9.01898369e-02
6.94679320e-01 5.51360071e-01 3.36629212e-01 -1.85892079e-02
-7.29777753e-01 9.75769758e-02 2.47469898e-02 9.69026268e-01
5.96009910e-01 -3.80140841e-01 2.29855761e-01 -1.46681100e-01
-3.70333999e-01 4.16530132e-01 6.77166224e-01 1.00713360e+00
-1.55866995e-01 -1.21627963e+00 -7.19457567e-01 4.48758639e-02
-4.66272593e-01 1.06566273e-01 -6.04313076e-01 7.17401505e-01
4.42203909e-01 9.12216544e-01 2.85478264e-01 -1.98279008e-01
2.22331375e-01 1.13563240e-01 8.59173000e-01 -8.05292249e-01
-4.94390100e-01 2.77342319e-01 9.72947180e-02 -6.90123856e-01
-1.13777018e+00 -7.61995494e-01 -1.14837182e+00 -2.61469513e-01
-2.72566319e-01 -1.69731021e-01 6.74830019e-01 7.40455985e-01
3.77731264e-01 1.37663230e-01 1.00997722e+00 -1.54986167e+00
1.64688118e-02 -7.92730629e-01 -4.23100889e-01 4.21463519e-01
6.28645360e-01 -6.18875682e-01 -4.53073084e-01 1.00956655e+00] | [13.08745002746582, -0.35870662331581116] |
9473cf4b-984b-4006-a99b-50c1def9829d | image-processing-methods-for-coronal-hole | 2201.01380 | null | https://arxiv.org/abs/2201.01380v1 | https://arxiv.org/pdf/2201.01380v1.pdf | Image Processing Methods for Coronal Hole Segmentation, Matching, and Map Classification | The paper presents the results from a multi-year effort to develop and validate image processing methods for selecting the best physical models based on solar image observations. The approach consists of selecting the physical models based on their agreement with coronal holes extracted from the images. Ultimately, the goal is to use physical models to predict geomagnetic storms. We decompose the problem into three subproblems: (i) coronal hole segmentation based on physical constraints, (ii) matching clusters of coronal holes between different maps, and (iii) physical map classification. For segmenting coronal holes, we develop a multi-modal method that uses segmentation maps from three different methods to initialize a level-set method that evolves the initial coronal hole segmentation to the magnetic boundary. Then, we introduce a new method based on Linear Programming for matching clusters of coronal holes. The final matching is then performed using Random Forests. The methods were carefully validated using consensus maps derived from multiple readers, manual clustering, manual map classification, and method validation for 50 maps. The proposed multi-modal segmentation method significantly outperformed SegNet, U-net, Henney-Harvey, and FCN by providing accurate boundary detection. Overall, the method gave a 95.5% map classification accuracy. | ['C. N. Arge', 'M. S. Pattichis', 'V. Jatla'] | 2022-01-04 | null | null | null | null | ['boundary-detection'] | ['computer-vision'] | [ 1.56747416e-01 -1.22305505e-01 2.94993937e-01 -3.48697126e-01
-7.76602149e-01 -5.23488581e-01 6.72947884e-01 4.29752588e-01
-1.36410728e-01 7.10947156e-01 -3.56189102e-01 -5.15736043e-01
-3.61738950e-01 -1.10090852e+00 -5.00798523e-01 -7.18789220e-01
2.07123011e-01 9.76371408e-01 6.96619391e-01 -2.00705498e-01
9.55099463e-01 7.13820696e-01 -1.59518933e+00 -2.78431699e-02
1.48699594e+00 1.15155983e+00 4.18067634e-01 6.62286401e-01
-1.39039502e-01 3.33187759e-01 -4.93898690e-01 3.50416005e-01
3.19110185e-01 -7.11101830e-01 -6.85826302e-01 -1.30503625e-02
3.63115191e-01 3.01995277e-01 4.93605852e-01 1.01517856e+00
4.03194010e-01 2.53832519e-01 1.02720785e+00 -1.04158068e+00
7.66580403e-02 1.15468271e-01 -8.32776070e-01 4.69497502e-01
-5.24996035e-02 -1.92024469e-01 5.88009775e-01 -4.87338930e-01
3.40014666e-01 7.05000639e-01 1.00111902e+00 -3.12521011e-01
-1.03927505e+00 -6.14878118e-01 -3.38376135e-01 4.43082154e-01
-1.39797449e+00 -2.20168516e-01 5.01937509e-01 -9.66323197e-01
8.83821309e-01 5.25952160e-01 7.46388435e-01 -1.73669249e-01
2.85290956e-01 -5.12453616e-02 1.79228783e+00 -9.20072854e-01
4.70825851e-01 1.45004839e-01 3.75837356e-01 5.82231045e-01
1.31472394e-01 5.27964532e-02 -3.91668260e-01 -3.80231082e-01
4.71303970e-01 -6.68700457e-01 -2.00737029e-01 -1.25437185e-01
-7.20598340e-01 1.05663240e+00 3.85605603e-01 3.99722666e-01
-6.93446636e-01 -3.57999444e-01 -4.23132032e-02 -2.99017042e-01
3.72225434e-01 5.30193925e-01 -2.38604590e-01 3.58035654e-01
-1.64518476e+00 3.73064071e-01 6.66140020e-01 2.03726187e-01
9.07760262e-01 -1.64120480e-01 1.33632362e-01 9.51679289e-01
5.99216819e-01 1.07709539e+00 3.04426283e-01 -8.84467006e-01
1.99195474e-01 5.56036830e-01 1.35479137e-01 -1.44154370e+00
-6.01230502e-01 -7.67037563e-04 -6.78516567e-01 4.87765282e-01
2.00531363e-01 -2.49162242e-01 -1.37114286e+00 1.06563509e+00
5.77187657e-01 -1.70003563e-01 -9.37113836e-02 8.09880853e-01
7.38620579e-01 9.01893914e-01 1.43645078e-01 -2.19444036e-01
1.27771533e+00 -8.21594954e-01 -5.14796972e-01 -6.94303364e-02
1.42116323e-01 -6.75928235e-01 2.42439702e-01 3.40368181e-01
-8.12156916e-01 -6.47492409e-01 -1.20609987e+00 6.23471141e-01
-7.96125412e-01 2.02504635e-01 3.63695204e-01 5.99246740e-01
-1.17254961e+00 3.97010505e-01 -8.23752403e-01 -6.94322467e-01
-7.88705721e-02 1.74481854e-01 1.73922241e-01 8.93839538e-01
-1.18548703e+00 1.17753708e+00 6.53600633e-01 1.14808127e-01
-5.67680001e-01 -3.83454770e-01 -7.19217420e-01 -2.90599167e-01
-2.32423976e-01 -5.93339384e-01 9.08060491e-01 -8.16165864e-01
-1.14077437e+00 1.07389152e+00 -3.87130678e-01 -7.33740628e-01
4.27316636e-01 4.52212036e-01 -5.30648649e-01 5.41798174e-01
5.59768975e-01 8.07804346e-01 6.32552564e-01 -1.75982106e+00
-1.23975885e+00 -3.39844525e-01 -3.97677064e-01 4.17303562e-01
3.39808524e-01 -3.35710645e-02 -5.82000732e-01 -2.31415287e-01
6.47569358e-01 -8.38463664e-01 -3.99242461e-01 -5.42062402e-01
-3.94675136e-01 -9.12093967e-02 7.95942664e-01 -1.54000378e+00
1.08936286e+00 -1.68668413e+00 -2.09841356e-01 1.16671622e+00
1.57875374e-01 -4.02807966e-02 7.45723128e-01 -2.25745011e-02
9.02791321e-02 4.26630396e-03 -7.95037627e-01 1.39226153e-01
-2.03905612e-01 -1.77626222e-01 -5.24038337e-02 3.88591230e-01
-3.93889785e-01 4.30247724e-01 -3.41981143e-01 -8.36761832e-01
6.70094073e-01 5.87735996e-02 3.06058917e-02 -5.02046198e-02
-1.41142085e-01 5.10893404e-01 -2.85813123e-01 6.75253034e-01
1.08277428e+00 1.52603807e-02 1.61013946e-01 -1.07409932e-01
-6.58077836e-01 -4.16106498e-03 -1.28717363e+00 8.78150821e-01
-1.16233461e-01 4.86643970e-01 2.82177269e-01 -1.08062935e+00
1.43166316e+00 -3.68214138e-02 5.28105140e-01 -1.26600862e+00
-2.03360356e-02 4.42900032e-01 -2.05375791e-01 -4.84753549e-01
6.24095976e-01 -3.23269397e-01 -1.66130979e-02 3.78061503e-01
-1.87524825e-01 -7.31218338e-01 4.02109206e-01 1.74509417e-02
5.50887167e-01 -4.49308157e-02 1.28710316e-03 -8.08692098e-01
4.88245338e-01 7.03665972e-01 5.49815416e-01 1.01633298e+00
-1.50987372e-01 7.92191386e-01 6.42981902e-02 -3.55940789e-01
-1.07173538e+00 -1.09457707e+00 -6.27964854e-01 4.83627945e-01
6.04809284e-01 8.47238153e-02 -1.10639501e+00 -2.03665763e-01
-6.31569475e-02 6.21395528e-01 -6.58020973e-01 4.79323119e-01
-2.25455776e-01 -1.62174261e+00 3.33061129e-01 1.69343695e-01
9.15137827e-01 -1.28384435e+00 -9.64759290e-01 2.19164997e-01
-5.43444037e-01 -6.33519292e-01 3.32586259e-01 4.74904627e-01
-7.45314956e-01 -1.26395833e+00 -4.35585976e-01 -7.53238499e-01
7.26509571e-01 -2.92954128e-02 9.88386631e-01 1.26057982e-01
-2.22451657e-01 1.36003584e-01 -3.16998869e-01 -4.55354065e-01
-3.78269136e-01 1.14594162e-01 -1.68368936e-01 -1.43639028e-01
2.33110383e-01 -3.57079744e-01 -4.48367029e-01 5.00696361e-01
-6.44907892e-01 3.35946411e-01 3.03238273e-01 3.55480343e-01
9.66291249e-01 6.51683569e-01 4.11107272e-01 -5.67459524e-01
3.66667539e-01 -5.19946575e-01 -9.10799325e-01 5.40136397e-01
-6.47534370e-01 -3.87639880e-01 1.78515628e-01 2.92314053e-01
-1.43404341e+00 1.46356493e-01 -1.55305639e-01 1.41651452e-01
-2.66387880e-01 5.17886460e-01 1.42486513e-01 -5.04909933e-01
8.73210847e-01 1.67953506e-01 -1.95980847e-01 -2.27913260e-01
1.02131285e-01 9.14382398e-01 7.68142045e-01 -4.69461739e-01
6.85835600e-01 6.07360661e-01 -2.13275895e-01 -1.09553158e+00
-6.47449017e-01 -6.70414567e-01 -8.05095613e-01 -6.73402309e-01
1.33079982e+00 -6.57521307e-01 -7.95680955e-02 8.72370601e-01
-8.47212613e-01 -4.25213933e-01 2.78638721e-01 3.39350164e-01
-4.79960233e-01 3.72395575e-01 -3.34328800e-01 -8.39181602e-01
-6.37228072e-01 -8.46258342e-01 7.95605659e-01 8.92394722e-01
-2.11451858e-01 -1.12029839e+00 5.13049185e-01 7.03979135e-01
3.27678800e-01 4.53898311e-01 7.90840507e-01 -2.50195771e-01
-3.55779201e-01 6.76182806e-02 -2.72644520e-01 2.03938540e-02
9.16040968e-03 2.37084553e-01 -8.70658278e-01 7.97082782e-02
5.85273206e-02 -5.22332005e-02 9.07704055e-01 9.10665810e-01
8.73669982e-01 3.25328588e-01 -7.43510306e-01 6.83045805e-01
1.70800018e+00 8.16141546e-01 7.29893327e-01 8.07354152e-01
3.85290265e-01 5.90720832e-01 7.12561429e-01 1.33001521e-01
3.42779696e-01 4.33523208e-01 2.16082484e-01 -5.32114148e-01
1.43036366e-01 -6.46415586e-03 -3.26555133e-01 6.76744461e-01
-9.45831686e-02 2.15219483e-01 -1.42752230e+00 8.45275164e-01
-1.98042166e+00 -8.52788389e-01 -5.94698250e-01 2.02414775e+00
4.38502043e-01 3.33288401e-01 6.14284500e-02 2.03019738e-01
1.19556296e+00 -9.55108255e-02 -4.66397524e-01 -5.51806331e-01
-3.92238885e-01 2.60995746e-01 8.32776904e-01 7.85501599e-01
-1.27067697e+00 7.33080268e-01 6.55561972e+00 7.48300552e-01
-1.16800606e+00 1.36194587e-01 8.84458244e-01 6.07220411e-01
-8.46977457e-02 1.50139213e-01 -8.91959786e-01 7.26500273e-01
7.20387638e-01 1.58296943e-01 2.90582657e-01 2.42608458e-01
4.08886999e-01 -1.05898905e+00 5.36022633e-02 7.87724376e-01
2.58232325e-01 -1.61631918e+00 -3.50722402e-01 -1.47717237e-01
1.24147022e+00 3.09246629e-01 -7.66709387e-01 -1.37452096e-01
2.81696618e-01 -8.87359977e-01 9.50270832e-01 1.01536822e+00
4.42330003e-01 -4.09522802e-01 6.07149780e-01 2.61930346e-01
-1.40486681e+00 1.63021967e-01 -2.08377734e-01 4.72987369e-02
4.12477672e-01 9.26344275e-01 -6.44215763e-01 9.11876023e-01
1.31433988e+00 1.25875220e-01 -7.27632344e-01 1.46864629e+00
-1.48580313e-01 6.56239688e-01 -9.53534365e-01 1.65407211e-01
2.39742503e-01 -7.00126886e-01 1.85614362e-01 1.22248316e+00
2.86885262e-01 1.17665215e-03 9.15156528e-02 8.76039207e-01
6.52041197e-01 2.27705866e-01 -3.40524346e-01 5.65064967e-01
5.99756241e-01 1.46034026e+00 -1.68040311e+00 -4.70122129e-01
1.35853216e-01 5.32638609e-01 -8.04035217e-02 9.56728235e-02
-9.65526879e-01 -5.61145365e-01 -1.76625267e-01 2.15752169e-01
2.71864563e-01 8.91542956e-02 -1.15588439e+00 -3.56367975e-01
-4.18245234e-02 -5.56450427e-01 5.72087765e-01 -1.07093167e+00
-8.01146865e-01 5.66364825e-01 2.88204640e-01 -6.67260587e-01
3.00891995e-02 -2.67180234e-01 -1.18068075e+00 1.17776382e+00
-1.15605223e+00 -8.58692467e-01 -5.17468870e-01 1.03568248e-01
1.01515926e-01 3.50681283e-02 5.03217161e-01 1.86930567e-01
-2.70390272e-01 -1.72397763e-01 3.45906585e-01 -2.36167476e-01
1.88660368e-01 -1.33820593e+00 2.89511008e-05 9.84466732e-01
-3.43174338e-01 -1.82776392e-01 6.94704711e-01 -1.03835380e+00
-2.36408338e-01 -9.62887466e-01 1.00544834e+00 -1.70336217e-01
4.52431440e-01 6.04673754e-03 -8.74969840e-01 2.25780502e-01
2.14569882e-01 -6.33988976e-01 2.60048628e-01 -3.49177003e-01
5.46814680e-01 5.75763509e-02 -1.40015757e+00 4.18996625e-02
3.65591615e-01 -2.46748865e-01 -8.35076511e-01 3.54206890e-01
-2.07909852e-01 -4.15513009e-01 -7.49976337e-01 8.31456602e-01
2.03408301e-01 -1.06446218e+00 6.65468276e-01 2.73803949e-01
2.21427307e-01 -8.94824147e-01 5.70979975e-02 -1.27091682e+00
-1.25240475e-01 -6.72580227e-02 6.11190021e-01 1.15400374e+00
7.15689659e-01 -4.82359022e-01 7.15204716e-01 1.71673015e-01
-1.48935005e-01 -4.58824247e-01 -8.12599778e-01 -2.54068762e-01
9.01939198e-02 -2.44173333e-01 3.75989437e-01 1.06106174e+00
-3.35172683e-01 7.53346980e-02 1.55591518e-01 6.79671347e-01
7.81687558e-01 6.40499890e-01 5.12916863e-01 -1.43501151e+00
8.24782997e-02 -3.90377849e-01 -1.19570181e-01 -5.41310728e-01
-8.15151706e-02 -7.10345626e-01 6.00539505e-01 -2.04390907e+00
2.44088486e-01 -1.08258760e+00 -1.08873174e-01 5.08469224e-01
-5.53932749e-02 6.35433197e-01 -6.57156035e-02 4.14642394e-01
-3.41373712e-01 2.85088360e-01 6.85272932e-01 -6.20929524e-02
-3.98001611e-01 -1.58913240e-01 -2.86770672e-01 7.83270895e-01
1.24005818e+00 -4.17072654e-01 -3.70760411e-02 -1.77578647e-02
1.42277926e-01 1.42008170e-01 2.89164484e-01 -1.51398396e+00
3.96030843e-01 -2.08969757e-01 5.87271810e-01 -1.10838079e+00
2.11614277e-02 -4.54944909e-01 5.26415169e-01 4.63801384e-01
3.91836613e-01 -9.50351655e-02 4.11101192e-01 -4.66299206e-02
-3.06862086e-01 -5.20813465e-01 1.19327164e+00 -8.43129605e-02
-9.92871165e-01 -2.86812901e-01 -5.42426169e-01 -2.52889127e-01
1.24537528e+00 -5.18829525e-01 -2.34071687e-01 -1.29585147e-01
-9.78148580e-01 6.64060533e-01 6.17773771e-01 1.03140175e-01
1.72633752e-01 -9.17837799e-01 -5.76488554e-01 1.69315040e-01
-1.44548804e-01 -2.71589328e-02 3.99777114e-01 9.96388495e-01
-1.15225017e+00 2.54209399e-01 -4.13819641e-01 -1.02102435e+00
-1.15407383e+00 -1.53192654e-01 1.07120061e+00 -2.13683352e-01
-4.48033154e-01 5.28843462e-01 -4.12484594e-02 -9.40285087e-01
-2.61570603e-01 -6.34808987e-02 -6.56214714e-01 4.53111410e-01
7.36585408e-02 4.78316933e-01 6.14013910e-01 -9.89366651e-01
-2.86008537e-01 1.03150296e+00 6.43561542e-01 -4.08391982e-01
1.04376268e+00 -2.50882089e-01 -3.91475439e-01 3.88700485e-01
4.44414675e-01 -2.40137219e-01 -9.85928714e-01 2.99115956e-01
-1.17984107e-02 -7.89099839e-03 2.98629135e-01 -1.23100829e+00
-1.11200571e+00 4.01512116e-01 9.47437942e-01 4.86003160e-01
1.15727293e+00 -7.83487707e-02 5.38131893e-01 -6.26282692e-02
3.57612878e-01 -1.68371320e+00 -7.85333455e-01 4.06434894e-01
6.63164973e-01 -1.12525773e+00 -9.47464432e-04 -3.51026207e-01
-3.97444874e-01 9.34234023e-01 6.22426987e-01 -4.83952463e-02
1.13339925e+00 1.38733581e-01 4.94938612e-01 -6.64240479e-01
-3.62123325e-02 -7.73623064e-02 1.91938072e-01 4.93906945e-01
-2.05321968e-01 3.94317299e-01 -7.53789842e-01 2.35067606e-01
-6.10096693e-01 -4.11635861e-02 2.06126779e-01 9.13154781e-01
-1.08516943e+00 -5.42345762e-01 -1.22851682e+00 9.40218270e-01
-1.59787431e-01 8.25588778e-02 -3.71006936e-01 5.93745112e-01
7.03660369e-01 1.13641095e+00 5.14376342e-01 -1.40084505e-01
1.01188436e-01 -6.19905582e-03 2.16868415e-01 -1.15349382e-01
-4.79070812e-01 -1.07976288e-01 7.49756843e-02 2.55769286e-02
-6.33553505e-01 -7.76589274e-01 -1.41943121e+00 -1.00301951e-01
-4.42696065e-01 7.82366753e-01 1.08077979e+00 1.16021407e+00
-3.07604317e-02 3.20876360e-01 6.57631457e-01 -1.11314368e+00
3.05794418e-01 -1.08729041e+00 -8.03029120e-01 1.21449053e-01
-2.41928056e-01 -9.17300940e-01 -6.17322206e-01 1.35914847e-01] | [9.303457260131836, -1.6414005756378174] |
dfcc3f0a-86d9-45d0-bbaf-f0c2490c210f | 25-years-of-criticality-in-neuroscience | 1903.05129 | null | http://arxiv.org/abs/1903.05129v1 | http://arxiv.org/pdf/1903.05129v1.pdf | 25 years of criticality in neuroscience -- established results, open controversies, novel concepts | Twenty-five years ago, Dunkelmann and Radons (1994) proposed that neural
networks should self-organize to a critical state. In models, criticality
offers a number of computational advantages. Thus this hypothesis, and in
particular the experimental work by Beggs and Plenz (2003), has triggered an
avalanche of research, with thousands of studies referring to it. Nonetheless,
experimental results are still contradictory. How is it possible, that a
hypothesis has attracted active research for decades, but nonetheless remains
controversial? We discuss the experimental and conceptual controversy, and then
present a parsimonious solution that (i) unifies the contradictory experimental
results, (ii) avoids disadvantages of a critical state, and (iii) enables
rapid, adaptive tuning of network properties to task requirements. | [] | 2019-03-12 | null | null | null | null | ['novel-concepts'] | ['reasoning'] | [ 2.35652253e-01 1.05147779e-01 -2.58206666e-01 -2.69584566e-01
6.18020184e-02 -6.56313241e-01 6.94136798e-01 1.65992469e-01
-6.48727536e-01 9.46914315e-01 -5.49451485e-02 -6.80906177e-01
-7.04230487e-01 -3.79645675e-01 -3.68960798e-01 -6.58600807e-01
-2.37758070e-01 2.47023389e-01 5.20908654e-01 -4.96209003e-02
5.66892982e-01 7.03317285e-01 -2.06156301e+00 -1.50666043e-01
7.64905930e-01 8.75832260e-01 6.04701817e-01 5.16220272e-01
2.38903522e-01 5.08277774e-01 -3.75854015e-01 -3.33466917e-01
5.83866835e-02 -6.45634234e-01 -8.53387237e-01 -8.07869993e-03
2.68042892e-01 1.09356001e-01 -9.54106003e-02 9.67292011e-01
3.40439022e-01 1.40014738e-01 4.48392868e-01 -1.18221414e+00
-7.87525177e-01 4.85371292e-01 -2.81348169e-01 5.82314134e-01
-3.32088709e-01 2.49066740e-01 1.14379883e+00 -4.98986334e-01
7.19392180e-01 9.66967881e-01 4.99419928e-01 7.82625020e-01
-1.43796897e+00 -4.95205641e-01 3.32909226e-01 2.60366499e-02
-1.27858043e+00 -5.43516994e-01 5.81949949e-01 -4.83973801e-01
9.25661564e-01 2.96550691e-01 1.20209026e+00 8.54130864e-01
5.03952384e-01 4.33454812e-01 1.37604260e+00 -5.16079664e-01
4.47824806e-01 6.38458729e-02 2.78638929e-01 4.04981852e-01
7.25707650e-01 3.48403752e-01 -3.59532714e-01 -4.85336557e-02
9.28343177e-01 -5.42000651e-01 -1.92582369e-01 -2.71790266e-01
-1.05255723e+00 8.28422725e-01 6.42858893e-02 8.04865479e-01
-2.30934978e-01 8.24603364e-02 3.97327840e-01 4.26979780e-01
1.16613947e-01 9.50213075e-01 -4.70122635e-01 -1.34677738e-01
-8.70330632e-01 4.14685309e-01 6.98129594e-01 4.88544464e-01
4.66921091e-01 1.81621760e-01 5.38236260e-01 6.71222210e-01
4.73844439e-01 -7.21809715e-02 7.51146674e-01 -1.11778545e+00
7.96487778e-02 2.04423830e-01 -5.78692071e-02 -1.18506312e+00
-6.68928206e-01 -4.92636353e-01 -8.42378616e-01 2.00786397e-01
5.37213326e-01 -2.99587429e-01 -4.01940048e-01 2.10155058e+00
-1.97435528e-01 -4.39679712e-01 -6.18240535e-02 8.22045326e-01
1.89697877e-01 2.74306715e-01 2.29478195e-01 -5.83701968e-01
9.43566322e-01 -4.69814837e-01 -5.55863798e-01 -5.50991714e-01
3.03939641e-01 -5.42685390e-01 9.12052929e-01 5.19087732e-01
-1.28059018e+00 -4.73899513e-01 -1.33041978e+00 3.82966816e-01
-2.19774216e-01 -3.82038832e-01 1.39237285e+00 9.78049636e-01
-1.47401202e+00 6.54951215e-01 -1.00798821e+00 -9.48096752e-01
-1.21385790e-02 6.29425764e-01 -3.35661501e-01 5.25310159e-01
-1.25523019e+00 1.19912267e+00 6.39441609e-01 1.49821207e-01
-3.19010675e-01 -1.26869753e-01 -3.81979376e-01 7.10429549e-02
3.30502808e-01 -9.18909073e-01 9.35903192e-01 -1.25727534e+00
-1.42483091e+00 8.25592399e-01 -4.78346497e-02 -5.38707614e-01
2.18497276e-01 3.69367212e-01 -2.98646361e-01 3.22365910e-01
-2.49289081e-01 8.05065751e-01 3.18624169e-01 -1.38121009e+00
-4.53731537e-01 -5.43860972e-01 2.54056752e-02 2.36786798e-01
-2.42129475e-01 2.27085631e-02 1.57683924e-01 -6.08856499e-01
7.59507596e-01 -9.48388100e-01 -2.90105045e-01 -1.35404974e-01
1.12468507e-02 -4.71241862e-01 1.23214759e-01 3.96199115e-02
1.13754332e+00 -1.96015429e+00 -1.42791972e-01 1.04674496e-01
5.51422238e-01 1.27459019e-01 5.68214841e-02 4.30793703e-01
-4.96202469e-01 6.07457340e-01 -2.69432843e-01 1.88239068e-01
-5.73675334e-02 8.99184495e-02 -1.96395114e-01 5.72739065e-01
1.10039473e-01 8.65840137e-01 -6.43598378e-01 -3.74642700e-01
1.21509828e-01 2.01245427e-01 -6.74043059e-01 -2.51853228e-01
7.93909505e-02 4.95814122e-02 -1.38440341e-01 4.30650473e-01
4.11979139e-01 -2.00734332e-01 6.00203335e-01 4.33636233e-02
-6.92021012e-01 3.49988699e-01 -1.08374143e+00 9.34410572e-01
1.80080742e-01 1.07305217e+00 2.48261929e-01 -1.38650537e+00
7.61376858e-01 4.39703375e-01 5.53869247e-01 -5.61863661e-01
5.34701526e-01 3.53740484e-01 5.41689754e-01 -4.57178742e-01
4.75521892e-01 -5.01339674e-01 2.35350937e-01 6.81269825e-01
2.49779373e-01 -8.28156620e-02 3.17905605e-01 6.07942939e-02
9.81848896e-01 -6.99182823e-02 1.84081957e-01 -7.65097618e-01
9.69604328e-02 -1.72710270e-01 5.86465716e-01 8.18664908e-01
-5.27593613e-01 3.28794003e-01 7.17610955e-01 -3.35968047e-01
-1.02539086e+00 -1.03411853e+00 -3.41050118e-01 8.63454044e-01
2.11450264e-01 -1.15610078e-01 -8.27617764e-01 8.26770663e-02
-1.05250597e-01 5.09857357e-01 -7.62196362e-01 -3.43987346e-01
-4.09363091e-01 -8.51300418e-01 4.99451637e-01 2.24494502e-01
4.60317522e-01 -1.10929549e+00 -9.65053082e-01 1.50399253e-01
6.33553490e-02 -7.81186998e-01 2.68320918e-01 7.96296597e-01
-1.47105157e+00 -7.82769978e-01 -4.43412632e-01 -9.07417476e-01
7.32770085e-01 5.43954313e-01 8.48126233e-01 5.07741868e-01
-1.21914364e-01 3.07695448e-01 -4.78614047e-02 -1.79848641e-01
-4.17085022e-01 2.63457477e-01 5.58589995e-01 -4.52386916e-01
2.73126125e-01 -1.10930002e+00 -4.04103816e-01 3.46121103e-01
-1.03298092e+00 -4.49070521e-02 6.71498597e-01 5.83591640e-01
2.92670250e-01 8.39793906e-02 1.20541871e+00 -7.24589348e-01
8.50667655e-01 -2.32453302e-01 -6.84922576e-01 7.86286816e-02
-1.19278264e+00 2.29936205e-02 4.44241196e-01 -2.97601640e-01
-8.96319389e-01 -1.16914988e-01 -1.70416117e-01 2.12970406e-01
-4.22093332e-01 6.11711025e-01 -2.38536410e-02 -2.39177734e-01
8.65288258e-01 2.21376941e-01 2.37608522e-01 -1.45233199e-01
2.17622295e-02 4.07207161e-01 2.54032761e-01 -6.63488746e-01
7.83819079e-01 2.31013700e-01 -2.79456340e-02 -1.15763760e+00
-4.59051281e-01 1.86703280e-02 -8.22009683e-01 -4.58796173e-01
6.24479771e-01 -4.40271705e-01 -8.59425902e-01 4.95503038e-01
-9.87236202e-01 -3.51168066e-01 -2.02507183e-01 6.64762139e-01
-9.96749580e-01 2.98734903e-01 -4.79147524e-01 -8.54647934e-01
2.64323145e-01 -9.55660462e-01 -1.38862103e-01 3.71347725e-01
-5.09418547e-01 -1.12990952e+00 1.02382824e-02 -7.26971254e-02
5.26050448e-01 -4.21976820e-02 1.13960576e+00 -5.15983760e-01
-5.15595317e-01 -1.88100398e-01 -1.19280681e-01 1.21184744e-01
-1.02997899e-01 3.39466780e-01 -9.69478786e-01 -1.70734942e-01
3.32094491e-01 -4.45615470e-01 8.62360120e-01 4.48913187e-01
7.61119962e-01 -1.19628847e-01 -2.84635454e-01 2.96178818e-01
1.43339360e+00 5.21102786e-01 7.52677441e-01 4.54362959e-01
-1.22236028e-01 9.03690338e-01 1.45823658e-01 1.33162295e-03
2.92694330e-01 2.76798844e-01 5.39735615e-01 1.69944271e-01
1.79709494e-01 5.61584421e-02 6.92786276e-03 9.54577446e-01
-3.29431236e-01 -2.45010778e-01 -8.00813556e-01 6.25684321e-01
-1.72885859e+00 -1.30972219e+00 1.12917423e-01 2.14834166e+00
8.19311440e-01 5.85363209e-01 2.81663418e-01 2.31040195e-01
8.33873868e-01 3.96295786e-01 -6.19991779e-01 -3.88467908e-01
-4.28919137e-01 -2.58111954e-01 2.98513114e-01 4.25041348e-01
-5.31052709e-01 9.93981540e-01 7.99899006e+00 3.52168560e-01
-1.12843502e+00 -1.73005894e-01 7.18436897e-01 1.94935545e-01
-2.46371701e-01 1.96420282e-01 -6.66960537e-01 1.57469586e-01
1.14212096e+00 -6.21295273e-01 3.48240793e-01 5.33637047e-01
2.12597270e-02 -6.67068601e-01 -9.51391876e-01 4.96723086e-01
1.36680603e-01 -1.12962496e+00 -2.68221319e-01 1.44489080e-01
5.10557771e-01 7.87015781e-02 1.65776700e-01 8.87261331e-03
1.92875549e-01 -8.04244816e-01 7.21182227e-01 2.49275833e-01
3.34077775e-01 -6.62541986e-01 2.84332842e-01 6.69830918e-01
-9.56862450e-01 -2.04982996e-01 -6.12636745e-01 -4.67816293e-01
-6.78928047e-02 5.02917111e-01 -2.74400055e-01 -4.61605601e-02
4.01393890e-01 2.39079386e-01 -7.61002243e-01 1.26877391e+00
-1.34084716e-01 7.36698747e-01 -3.06167126e-01 -4.56314951e-01
2.30460480e-01 -1.51040524e-01 4.20915037e-01 7.19334304e-01
-5.76251671e-02 1.51635617e-01 -5.38212538e-01 9.61064935e-01
3.58883858e-01 -9.27219838e-02 -6.86277986e-01 -4.02644396e-01
6.13919854e-01 1.24933887e+00 -1.59091580e+00 -2.75854915e-02
-1.73711032e-01 5.66744506e-01 3.54699045e-01 4.04168457e-01
-4.98923212e-01 -3.72642457e-01 5.65395832e-01 1.94269300e-01
-1.32224172e-01 -5.91407537e-01 -5.71811974e-01 -8.61788750e-01
-1.74958222e-02 -4.90239859e-01 -1.84249934e-02 -3.95510852e-01
-9.12437081e-01 7.32977748e-01 8.68515223e-02 -7.44508862e-01
-1.90410018e-01 -5.35211742e-01 -2.82333910e-01 5.42400897e-01
-9.78677154e-01 -4.76050556e-01 1.48562267e-01 9.45193246e-02
3.29949737e-01 -4.00777301e-03 8.15611064e-01 1.70617729e-01
-6.17582500e-01 4.46740717e-01 1.67201132e-01 -2.53173888e-01
5.29542685e-01 -1.11685634e+00 3.19722921e-01 4.48335171e-01
-1.10409722e-01 1.17959177e+00 9.51066017e-01 -5.39859653e-01
-1.08123660e+00 -4.52963054e-01 1.14098191e+00 -3.48085821e-01
7.64201522e-01 -3.07567745e-01 -8.81143451e-01 5.72478175e-01
3.55575830e-01 -4.98555094e-01 5.95657945e-01 1.71390921e-01
-1.24662600e-01 5.37360534e-02 -8.83120358e-01 8.88247013e-01
1.02289999e+00 -2.14617655e-01 -6.76874280e-01 1.96203351e-01
3.36322784e-01 2.90260226e-01 -5.83682060e-01 2.50965923e-01
9.64217544e-01 -1.48567510e+00 5.58843970e-01 -4.93008912e-01
5.34160845e-02 1.63034797e-01 5.23372963e-02 -1.20938885e+00
-5.28710842e-01 -6.69172943e-01 5.64886451e-01 1.03171945e+00
5.12708902e-01 -9.98702347e-01 7.04293549e-01 8.10815036e-01
-2.27323771e-01 -9.53136623e-01 -1.00454342e+00 -9.76556957e-01
4.14853632e-01 -2.39873245e-01 1.64466441e-01 9.69507992e-01
5.54270148e-01 4.03966188e-01 8.30188841e-02 -4.00204599e-01
2.93345511e-01 -1.75391108e-01 2.10989326e-01 -1.89832568e+00
1.08060166e-01 -9.97497201e-01 -3.05259228e-01 -9.22233701e-01
1.59837872e-01 -7.80866265e-01 6.12626262e-02 -1.55613065e+00
3.23159285e-02 -4.59961027e-01 -2.85831600e-01 1.83168381e-01
1.28847152e-01 1.91618279e-01 1.86082095e-01 3.35887343e-01
-4.83161747e-01 1.65855974e-01 9.73159671e-01 5.03820956e-01
-3.27026665e-01 -8.38590413e-02 -1.27055705e+00 1.04121983e+00
1.16623235e+00 -3.29634398e-01 -6.14674866e-01 -1.44905716e-01
8.07808697e-01 -1.41925335e-01 1.78143367e-01 -1.46445417e+00
4.24845874e-01 -3.33910465e-01 3.86844814e-01 -1.72333688e-01
3.47054929e-01 -7.61955678e-01 -4.87632975e-02 6.54677391e-01
-4.56680298e-01 3.17644775e-01 2.35744670e-01 3.90001267e-01
8.27001780e-02 -5.40649593e-01 1.14349771e+00 -3.68151665e-01
-3.75302255e-01 5.04761338e-02 -9.85738099e-01 1.10989911e-02
1.01010299e+00 -5.85985541e-01 -4.88435179e-01 -2.67211765e-01
-6.37990773e-01 6.94241151e-02 5.25550246e-01 2.22800732e-01
4.23427403e-01 -1.05422306e+00 -1.84680149e-01 2.02196568e-01
-2.83032596e-01 -4.85367358e-01 -1.93072520e-02 9.86081719e-01
-3.69087130e-01 7.10072935e-01 -5.05695164e-01 -3.47606003e-01
-8.55335414e-01 4.43533003e-01 4.32899714e-01 1.74567848e-01
-4.32855308e-01 9.92307663e-01 8.59241746e-03 -3.43286991e-03
1.35708287e-01 1.40673205e-01 -3.35061699e-01 5.44438362e-01
3.12112629e-01 1.95296183e-01 -8.85912329e-02 -5.59961677e-01
-3.19583327e-01 3.76523733e-01 -1.30422711e-01 -5.09155452e-01
1.42883325e+00 -1.86109111e-01 -1.50395423e-01 8.55813503e-01
7.59353518e-01 -3.79997671e-01 -1.04501271e+00 -2.36827731e-02
1.68742001e-01 -1.94849491e-01 -6.78375214e-02 -5.14464736e-01
-8.40957642e-01 1.01138854e+00 3.01856577e-01 9.44970906e-01
1.06558371e+00 -1.34905903e-02 2.05417916e-01 5.84130585e-01
3.10929567e-01 -1.51819885e+00 1.13181829e-01 5.97131550e-01
7.74388790e-01 -7.84788132e-01 2.13242471e-02 -1.25164181e-01
-5.71970940e-01 1.10882807e+00 7.14074016e-01 -2.58675963e-01
6.48260951e-01 3.05759698e-01 -1.08532287e-01 -2.51985371e-01
-1.20115495e+00 -1.81765944e-01 -3.63783181e-01 7.87665725e-01
7.41715670e-01 -4.16008651e-01 -9.07151401e-01 7.07900405e-01
-5.20339310e-01 -1.70695595e-02 5.93309283e-01 9.03331459e-01
-1.11694872e+00 -9.12071407e-01 -2.96083421e-01 5.03529787e-01
-3.42911154e-01 -1.71741620e-02 -6.52603626e-01 1.26815164e+00
5.61618879e-02 9.47581589e-01 7.51251429e-02 -3.81398499e-01
1.35755196e-01 2.92838901e-01 5.66621423e-01 -4.62453932e-01
-6.53832316e-01 2.24078417e-01 -1.47131115e-01 -1.94369286e-01
-4.89235938e-01 -7.80938745e-01 -1.03020263e+00 -3.71652126e-01
-5.88879585e-01 5.68223059e-01 6.05128169e-01 1.03521335e+00
6.10738695e-02 5.20678341e-01 4.15501267e-01 -7.43651330e-01
-4.42911386e-01 -1.00097680e+00 -8.87447059e-01 -3.49453986e-01
1.34420037e-01 -7.32819736e-01 -8.14474106e-01 -7.90529177e-02] | [8.08639907836914, 3.381420135498047] |
299eace6-d49e-4f5d-9455-1f06910a28cc | a-video-summarization-method-using-temporal | 2109.12581 | null | https://arxiv.org/abs/2109.12581v4 | https://arxiv.org/pdf/2109.12581v4.pdf | A Stacking Ensemble Approach for Supervised Video Summarization | Video summarization methods are usually classified into shot-level or frame-level methods, which are individually used in a general way. This paper investigates the underlying complementarity between the frame-level and shot-level methods, and a stacking ensemble approach is proposed for supervised video summarization. Firstly, we build up a stacking model to predict both the key frame probabilities and the temporal interest segments simultaneously. The two components are then combined via soft decision fusion to obtain the final scores of each frame in the video. A joint loss function is proposed for the model training. The ablation experimental results show that the proposed method outperforms both the two corresponding individual method. Furthermore, extensive experimental results on two benchmark datasets shows its superior performance in comparison with the state-of-the-art methods. | ['Guoqiang Zhang', 'Shenghui Zhao', 'Yubo An'] | 2021-09-26 | null | null | null | null | ['supervised-video-summarization'] | ['computer-vision'] | [ 4.83961046e-01 -1.21284217e-01 -3.30246031e-01 -2.61222929e-01
-9.18823242e-01 1.93254650e-02 5.89043915e-01 1.10492535e-01
-1.86539620e-01 8.30773592e-01 5.39457262e-01 2.78131187e-01
3.43883298e-02 -3.78835201e-01 -6.25246644e-01 -1.04051840e+00
5.81657290e-02 -2.20455125e-01 6.11292005e-01 2.38016367e-01
4.83370602e-01 -1.92551240e-01 -1.65221930e+00 4.38757092e-01
1.13785255e+00 1.13718390e+00 2.84124464e-01 6.16400778e-01
-1.17060542e-01 1.29131472e+00 -4.45958793e-01 -3.54969472e-01
1.17973305e-01 -8.05100441e-01 -4.45762426e-01 5.07719576e-01
3.19118381e-01 -4.68751460e-01 -4.66453612e-01 1.05076873e+00
6.05623603e-01 5.16476631e-01 5.75286210e-01 -1.01004064e+00
-1.91469401e-01 4.64902133e-01 -8.36758316e-01 3.90625298e-01
6.15895808e-01 -1.36911884e-01 9.31005716e-01 -9.87742305e-01
5.42719007e-01 1.10543239e+00 4.80633318e-01 1.69582278e-01
-7.16819108e-01 -3.67622852e-01 5.62406421e-01 6.39622629e-01
-1.01920021e+00 -5.50781131e-01 8.86677444e-01 -4.11958933e-01
7.87327707e-01 1.15552597e-01 6.03461146e-01 7.59775817e-01
2.98402160e-01 1.35988712e+00 7.96751976e-01 -3.18091899e-01
9.27629322e-02 -3.03447932e-01 1.79178119e-01 8.51276278e-01
1.50222018e-01 -3.08907241e-01 -7.41860330e-01 -2.95305159e-02
4.33742136e-01 8.90877098e-02 -4.92285997e-01 -1.95515513e-01
-1.31482792e+00 5.51667213e-01 5.46414182e-02 1.83515564e-01
-6.25697076e-01 -1.18841089e-01 7.36310542e-01 -2.57540606e-02
8.23334396e-01 -1.53086454e-01 8.96727890e-02 -1.42866895e-01
-1.47782135e+00 4.63758558e-01 6.32866502e-01 9.30006981e-01
4.86170352e-01 1.17661551e-01 -8.43205392e-01 9.00114477e-01
2.86755443e-01 -5.51155470e-02 5.18620074e-01 -1.06743991e+00
6.21835828e-01 4.66351539e-01 6.71137944e-02 -1.04203928e+00
1.27099067e-01 -2.77631223e-01 -8.74941528e-01 -1.91488117e-01
-1.16041929e-01 -3.11172426e-01 -8.13252270e-01 1.42853713e+00
3.13893020e-01 1.00468564e+00 5.09382747e-02 6.20839715e-01
1.15667331e+00 1.18960297e+00 2.68893200e-03 -9.34005082e-01
1.02926254e+00 -1.51125240e+00 -1.04571712e+00 1.96415499e-01
1.29350558e-01 -6.36641800e-01 3.23708177e-01 3.10035497e-01
-1.56653261e+00 -7.34518468e-01 -1.26676762e+00 1.44992903e-01
1.93898991e-01 2.04885691e-01 2.20121875e-01 2.50428230e-01
-9.75804150e-01 6.38035774e-01 -8.83577585e-01 -2.73108721e-01
5.22452056e-01 1.07297480e-01 9.25218035e-03 -2.60851383e-02
-1.04230440e+00 6.82582557e-01 6.63573146e-01 7.69193023e-02
-7.79076457e-01 -4.49500412e-01 -9.72850144e-01 1.60440475e-01
5.43860793e-01 -9.16318178e-01 1.40079141e+00 -8.08186114e-01
-1.69601250e+00 4.63843822e-01 -5.59958458e-01 -5.89016855e-01
6.66282535e-01 -3.47084731e-01 -1.06640704e-01 4.25809801e-01
7.48125091e-02 4.80325609e-01 1.02955878e+00 -1.29635322e+00
-1.29825258e+00 3.53825912e-02 -2.67732628e-02 7.00747788e-01
-2.25082025e-01 2.05437124e-01 -9.17878985e-01 -8.35010827e-01
1.28901862e-02 -4.07634974e-01 -1.43006206e-01 -6.38748109e-01
-4.42302167e-01 -4.54317003e-01 8.81451309e-01 -8.88986349e-01
1.91369355e+00 -1.94689786e+00 5.89935482e-01 -2.70810783e-01
3.37012298e-02 4.13387448e-01 1.71168566e-01 4.96835947e-01
-2.49846019e-02 -1.71025157e-01 -2.45431319e-01 -7.15503395e-01
-1.31275475e-01 -1.39177978e-01 -1.20966740e-01 3.43202353e-01
-3.40021104e-02 6.11771524e-01 -9.19211149e-01 -8.20913434e-01
4.59507883e-01 2.84932077e-01 -3.05856556e-01 4.81414050e-01
-2.74187140e-02 4.46227103e-01 -4.78168100e-01 5.68433702e-01
4.94329244e-01 3.56773660e-02 8.30434337e-02 -4.14854497e-01
-2.38881901e-01 1.86936166e-02 -1.14148593e+00 1.66726995e+00
1.47650823e-01 3.76901656e-01 -1.64350837e-01 -1.17847061e+00
5.70490301e-01 5.34902334e-01 6.87816083e-01 -1.70394436e-01
2.19368562e-01 -6.76669255e-02 -2.34019592e-01 -6.52584732e-01
8.43556821e-01 -1.22871622e-01 -3.66138294e-02 9.57461447e-02
3.20439756e-01 2.80723542e-01 6.18703187e-01 3.10468733e-01
8.10527682e-01 5.31117976e-01 7.43080378e-01 -1.99025706e-03
9.58011925e-01 -2.63760686e-01 9.20895875e-01 7.79975235e-01
-4.31177080e-01 8.53237808e-01 6.68724179e-01 -1.15130499e-01
-8.82924020e-01 -7.75723159e-01 1.78762749e-01 8.70706677e-01
6.14865303e-01 -6.71768963e-01 -1.07743347e+00 -6.97482824e-01
-4.89081860e-01 6.59907103e-01 -4.37018573e-01 -6.33845851e-02
-6.74214542e-01 -6.71206772e-01 1.18285008e-01 4.93664414e-01
8.78695190e-01 -9.47486281e-01 -5.28449118e-01 3.19680870e-01
-4.56330508e-01 -1.17591894e+00 -5.90253949e-01 -5.33349276e-01
-8.61929655e-01 -8.78588736e-01 -1.03854489e+00 -8.49170983e-01
4.08605546e-01 6.19516730e-01 7.49171615e-01 1.38806179e-02
2.54758060e-01 3.38815361e-01 -7.17842042e-01 -3.14149708e-01
-2.11020097e-01 -8.89622644e-02 9.56120377e-04 5.06368756e-01
1.68560781e-02 -5.12369335e-01 -7.45003700e-01 4.11315896e-02
-8.05825830e-01 3.50329876e-01 7.31912613e-01 7.10237920e-01
7.73200333e-01 1.81723729e-01 6.50538087e-01 -8.40600252e-01
5.54487109e-01 -5.70224643e-01 -1.88815564e-01 3.89657110e-01
-2.41558567e-01 -2.88118035e-01 3.84201497e-01 -1.47267908e-01
-1.47077155e+00 -4.30781245e-02 1.39197588e-01 -5.37745476e-01
-2.37219762e-02 7.67526269e-01 -2.73762256e-01 3.45103949e-01
-1.09555125e-01 5.77337384e-01 -1.94793344e-01 -3.64735633e-01
1.81778312e-01 4.65218663e-01 7.45805681e-01 -4.02821273e-01
5.35091341e-01 2.37691194e-01 -3.15974027e-01 -9.86173749e-01
-8.98112357e-01 -6.46999300e-01 -5.40463328e-01 -6.22734785e-01
1.12408197e+00 -1.06198943e+00 -2.64024109e-01 7.82436788e-01
-1.20871305e+00 1.11904316e-01 -2.01279297e-01 5.03073573e-01
-7.84410417e-01 9.09867883e-01 -6.19593740e-01 -8.95188153e-01
-3.80851924e-01 -1.30632412e+00 1.09652770e+00 6.73638821e-01
9.42664668e-02 -7.87562907e-01 1.40396699e-01 4.52906489e-01
-6.67915344e-02 3.91066879e-01 6.00928366e-01 -6.51750684e-01
-5.42700589e-01 -2.95422345e-01 -1.45479009e-01 4.71305639e-01
2.29905564e-02 2.68344522e-01 -5.99297106e-01 -1.89945295e-01
1.70291737e-01 3.33648138e-02 1.21288514e+00 8.07831824e-01
1.26555657e+00 -2.10843205e-01 -3.17000926e-01 3.15003127e-01
1.38555062e+00 3.66773665e-01 8.03081334e-01 2.21514404e-01
7.61644483e-01 3.38826597e-01 9.84977901e-01 6.44873023e-01
4.23994362e-01 6.90098584e-01 2.06685528e-01 2.17859611e-01
-4.43538204e-02 -2.10981101e-01 5.98500788e-01 1.16025710e+00
-5.40115297e-01 -5.96893430e-01 -2.83816397e-01 4.49972302e-01
-2.52925158e+00 -1.49121320e+00 -1.59711540e-01 2.19217443e+00
5.30503213e-01 1.46219626e-01 2.78724492e-01 1.11712441e-01
1.10772026e+00 8.08214784e-01 -4.55272079e-01 -1.31641895e-01
-1.82867423e-01 -1.81568772e-01 2.21230432e-01 3.31106931e-01
-1.55887961e+00 6.82798445e-01 6.40461063e+00 1.22967100e+00
-5.93999743e-01 1.18538462e-01 6.90843165e-01 -2.53052920e-01
1.05602801e-01 3.58095393e-02 -5.68657577e-01 7.31488168e-01
6.20873988e-01 -4.68594044e-01 -5.92397004e-02 3.44760954e-01
5.34548819e-01 -3.67811501e-01 -8.33462179e-01 1.09888923e+00
5.25705278e-01 -1.37058830e+00 2.41891906e-01 -2.45609626e-01
1.10139358e+00 -2.83285141e-01 -2.51764059e-01 3.39471370e-01
-3.64470147e-02 -4.30594504e-01 8.80494058e-01 9.53259587e-01
4.24862981e-01 -8.23926806e-01 9.44825470e-01 3.64612281e-01
-1.37022245e+00 -3.93807143e-02 -1.20562404e-01 -4.13200855e-02
7.49508858e-01 4.34703231e-01 -1.63883209e-01 1.32699132e+00
4.77502078e-01 1.21831751e+00 -3.85209948e-01 1.41839135e+00
-2.04840824e-01 7.06307173e-01 9.31105763e-02 2.25753516e-01
2.95491993e-01 -4.11450803e-01 9.39250350e-01 1.24734104e+00
4.77309287e-01 4.09568310e-01 4.34907287e-01 1.50271520e-01
-5.06771058e-02 1.74721390e-01 -3.44470650e-01 2.38491938e-01
2.53658980e-01 1.25285506e+00 -6.78469658e-01 -8.14573348e-01
-6.72866285e-01 1.07319355e+00 6.71788082e-02 3.42865616e-01
-1.09295106e+00 -3.14013749e-01 2.39209890e-01 -2.38741413e-01
6.05902553e-01 5.02593108e-02 -5.51870354e-02 -1.48589051e+00
1.16196610e-01 -6.47946894e-01 5.48221767e-01 -7.25491345e-01
-1.04803181e+00 3.38195413e-01 4.26119179e-01 -1.60243595e+00
-2.29512215e-01 -6.27642423e-02 -1.15167534e+00 4.63183373e-01
-1.49453795e+00 -9.56057429e-01 -3.68389428e-01 2.39972040e-01
1.32354152e+00 -3.20446610e-01 2.45545730e-01 1.12490572e-01
-1.13076067e+00 2.70495147e-01 2.69373953e-01 -1.93616748e-01
6.90444529e-01 -1.10286355e+00 -8.66912231e-02 1.23897028e+00
-2.88636416e-01 1.71259224e-01 7.79737294e-01 -7.55060434e-01
-1.05528259e+00 -1.14999986e+00 8.51372421e-01 3.70794870e-02
3.17574650e-01 3.01379412e-01 -9.21527863e-01 4.79205698e-01
7.05051661e-01 -4.00841892e-01 6.58653677e-01 -4.21941608e-01
2.02092111e-01 -7.48450831e-02 -8.68556857e-01 5.61579823e-01
8.87541354e-01 4.11099792e-02 -1.01940703e+00 1.90728560e-01
7.47069240e-01 -3.65653455e-01 -7.77206898e-01 6.16229057e-01
5.65107584e-01 -1.37103593e+00 7.81688809e-01 -3.89489532e-01
9.07808542e-01 -4.53155965e-01 -1.47819757e-01 -1.17525506e+00
-4.36359704e-01 -6.63042009e-01 -6.30329907e-01 1.61434126e+00
-5.06636240e-02 -2.36146703e-01 5.47886193e-01 1.17751636e-01
-3.85158479e-01 -9.94396865e-01 -7.30190694e-01 -5.07374942e-01
-4.63032365e-01 -7.53963217e-02 2.17066497e-01 5.94830513e-01
1.41442671e-01 6.64786816e-01 -6.70994222e-01 -8.42352808e-02
8.34191322e-01 1.02545090e-01 5.78060448e-01 -8.83869648e-01
-2.19527140e-01 -6.66971862e-01 -5.75508714e-01 -1.23852408e+00
2.82766312e-01 -5.31717181e-01 1.35354012e-01 -1.96705472e+00
7.43663251e-01 3.26778203e-01 -4.98359293e-01 -1.35987846e-03
-7.67647266e-01 -9.80043691e-03 3.78070295e-01 4.11621869e-01
-1.23316383e+00 8.30052912e-01 9.74470019e-01 -9.19927582e-02
-2.98032433e-01 1.25922501e-01 -6.73560679e-01 9.42338228e-01
5.44804633e-01 -1.03237234e-01 -5.31503260e-01 -2.60527551e-01
-4.93069470e-01 3.70613307e-01 6.01881333e-02 -1.29769659e+00
3.67774874e-01 -9.30900052e-02 1.53379038e-01 -1.16430259e+00
2.55346298e-01 -4.15175706e-01 3.99219319e-02 2.27124333e-01
-3.33597839e-01 -2.45703429e-01 -1.79246902e-01 1.04940343e+00
-6.54438436e-01 -2.41182774e-01 8.55189502e-01 -4.77507785e-02
-9.08306003e-01 4.78640378e-01 -2.70553052e-01 -2.57439435e-01
1.33812976e+00 -5.25767326e-01 -1.16811352e-04 -6.22943997e-01
-5.57724833e-01 4.32248622e-01 2.18270943e-01 2.95943886e-01
6.38423622e-01 -1.32755494e+00 -1.03320062e+00 -2.47170061e-01
-7.32530206e-02 -3.91884930e-02 6.50698066e-01 1.20568967e+00
-4.54554617e-01 2.74569303e-01 2.43314058e-02 -5.82254469e-01
-1.44002163e+00 4.71608371e-01 9.80968680e-03 -4.90210414e-01
-5.45604169e-01 6.76642120e-01 4.40455943e-01 4.17495131e-01
3.72628629e-01 -1.51162833e-01 -7.23305702e-01 3.22210938e-01
7.81066418e-01 7.19124794e-01 -2.60654986e-01 -1.03476703e+00
-1.10937543e-01 5.50350606e-01 -1.97451398e-01 4.53745835e-02
1.39249325e+00 -4.99774992e-01 -1.42908335e-01 5.92162430e-01
1.15036464e+00 -3.78772110e-01 -1.44964516e+00 -3.64811629e-01
-2.46962205e-01 -4.06067789e-01 -5.04281335e-02 -2.02843904e-01
-9.02866483e-01 5.99570572e-01 2.15053111e-01 2.72104740e-01
1.50941229e+00 -2.33188286e-01 1.04372001e+00 -1.15319043e-02
1.47247463e-01 -1.04982853e+00 -5.35089895e-02 3.37473810e-01
7.06632733e-01 -1.02117348e+00 5.08145809e-01 -5.87631464e-01
-8.50671053e-01 9.22059298e-01 4.15700972e-01 -3.65877301e-01
3.15738350e-01 -1.94240198e-01 -5.20108163e-01 1.50372893e-01
-9.96840835e-01 -1.78701982e-01 3.60911787e-01 2.58773476e-01
4.39224094e-01 -2.64330894e-01 -9.48328972e-01 8.37713957e-01
3.37309361e-01 2.61539817e-01 4.77109134e-01 9.92466152e-01
-7.86995709e-01 -9.72982943e-01 -3.46840620e-01 6.62170172e-01
-7.02858090e-01 1.54928729e-01 -5.24278805e-02 2.85013825e-01
-5.64691518e-03 1.09519589e+00 -4.48700115e-02 -5.31472147e-01
4.32857722e-01 1.40715569e-01 4.96161252e-01 -4.06672001e-01
-4.45548922e-01 4.30110544e-01 1.45927042e-01 -4.50735331e-01
-1.12235749e+00 -1.05698514e+00 -9.42447186e-01 -1.56313762e-01
-3.40269715e-01 1.40335321e-01 9.48494002e-02 9.67788875e-01
2.23022848e-01 7.89804041e-01 9.16621864e-01 -1.50751412e+00
-4.06053275e-01 -9.03752863e-01 -4.69169557e-01 2.79344261e-01
2.49329299e-01 -6.54256165e-01 -3.05845022e-01 5.29026806e-01] | [10.358960151672363, 0.41821005940437317] |
ec254f65-f0d7-4169-a70a-da9905f63b27 | open-set-semi-supervised-object-detection | 2208.13722 | null | https://arxiv.org/abs/2208.13722v1 | https://arxiv.org/pdf/2208.13722v1.pdf | Open-Set Semi-Supervised Object Detection | Recent developments for Semi-Supervised Object Detection (SSOD) have shown the promise of leveraging unlabeled data to improve an object detector. However, thus far these methods have assumed that the unlabeled data does not contain out-of-distribution (OOD) classes, which is unrealistic with larger-scale unlabeled datasets. In this paper, we consider a more practical yet challenging problem, Open-Set Semi-Supervised Object Detection (OSSOD). We first find the existing SSOD method obtains a lower performance gain in open-set conditions, and this is caused by the semantic expansion, where the distracting OOD objects are mispredicted as in-distribution pseudo-labels for the semi-supervised training. To address this problem, we consider online and offline OOD detection modules, which are integrated with SSOD methods. With the extensive studies, we found that leveraging an offline OOD detector based on a self-supervised vision transformer performs favorably against online OOD detectors due to its robustness to the interference of pseudo-labeling. In the experiment, our proposed framework effectively addresses the semantic expansion issue and shows consistent improvements on many OSSOD benchmarks, including large-scale COCO-OpenImages. We also verify the effectiveness of our framework under different OSSOD conditions, including varying numbers of in-distribution classes, different degrees of supervision, and different combinations of unlabeled sets. | ['Zsolt Kira', 'Zijian He', 'Peter Vajda', 'Junjiao Tian', 'Xiaoliang Dai', 'Chih-Yao Ma', 'Yen-Cheng Liu'] | 2022-08-29 | null | null | null | null | ['semi-supervised-object-detection'] | ['computer-vision'] | [-2.46536564e-02 2.49379665e-01 -2.82415003e-01 -2.76811957e-01
-5.12215137e-01 -5.04623830e-01 4.71188992e-01 -3.24013941e-02
-2.54613101e-01 3.59852344e-01 -8.09943527e-02 -7.39892945e-02
2.69782811e-01 -4.52683866e-01 -8.07312250e-01 -6.15253747e-01
1.37459591e-01 5.68982005e-01 8.48690033e-01 9.37819034e-02
-7.29403319e-03 2.90573180e-01 -1.90175891e+00 3.02735399e-02
8.66499305e-01 1.04347396e+00 2.89658129e-01 2.07371101e-01
-2.82163382e-01 5.95440865e-01 -6.31799042e-01 -2.00876713e-01
7.55355060e-01 -1.49860129e-01 -5.20012617e-01 5.21882474e-01
7.45978892e-01 -6.11717224e-01 -4.00132418e-01 1.38339305e+00
5.03449142e-01 -1.12879239e-01 7.44562566e-01 -1.55914390e+00
-7.81220496e-01 4.19085562e-01 -8.81721020e-01 3.07032645e-01
-1.52332289e-02 2.36205339e-01 8.03650916e-01 -1.19833708e+00
7.06261396e-01 1.39883626e+00 7.04562008e-01 4.60954964e-01
-9.97624695e-01 -6.51907384e-01 2.56433010e-01 3.80879082e-02
-1.39650035e+00 -2.50573009e-01 6.48695290e-01 -2.86856413e-01
7.58319557e-01 -6.93763494e-02 3.65439802e-01 8.16273510e-01
-8.87842476e-02 1.17190862e+00 1.17026663e+00 -6.35127664e-01
3.73003393e-01 6.23425961e-01 5.57070017e-01 5.76552808e-01
6.98016942e-01 3.25705647e-01 -2.73781359e-01 -6.47244900e-02
4.94362056e-01 2.05965415e-01 -9.35021117e-02 -6.43407583e-01
-8.92827272e-01 6.57311916e-01 4.14368361e-01 1.30224675e-02
7.14744329e-02 -3.81039947e-01 3.68850768e-01 2.04550520e-01
4.78031248e-01 6.01617768e-02 -5.17366171e-01 3.95145953e-01
-7.60699153e-01 -1.64021447e-01 7.94847310e-01 1.20350873e+00
8.37976158e-01 -1.52715996e-01 -1.73727959e-01 9.82646227e-01
4.95903999e-01 7.07055449e-01 7.26231515e-01 -6.97428882e-01
3.79735440e-01 7.91806996e-01 1.26091778e-01 -8.40124846e-01
-2.89294988e-01 -5.03087461e-01 -4.40029174e-01 2.37864375e-01
6.21645212e-01 8.28390270e-02 -1.33806252e+00 1.58952832e+00
8.25657070e-01 -2.40194164e-02 1.18419595e-01 1.15926099e+00
8.82286131e-01 4.48092133e-01 1.61939804e-02 -4.32037562e-01
1.35640156e+00 -1.40346074e+00 -8.48978519e-01 -4.27076280e-01
8.24429274e-01 -7.14463472e-01 1.17295325e+00 1.95207998e-01
-3.74527872e-01 -5.82180440e-01 -1.21168458e+00 1.51831880e-01
-4.92288440e-01 1.66878894e-01 4.71305907e-01 7.82087564e-01
-7.08330691e-01 2.21501872e-01 -4.83936727e-01 -6.72902524e-01
7.24388599e-01 9.74048078e-02 -2.58691519e-01 -3.48132193e-01
-9.56813872e-01 6.23373210e-01 7.36952841e-01 -3.02691627e-02
-1.16125059e+00 -3.43552679e-01 -6.44369245e-01 -1.30292714e-01
9.49088216e-01 -1.93181187e-01 1.12007678e+00 -1.21528399e+00
-9.43137646e-01 1.12231183e+00 1.49377674e-01 -4.00199711e-01
5.66270888e-01 -2.18370587e-01 -5.75113416e-01 1.68376222e-01
2.90225655e-01 7.29896605e-01 1.08306515e+00 -1.36110628e+00
-6.13862932e-01 -7.29975402e-01 -4.26103063e-02 3.63009214e-01
-5.43674827e-01 -8.40387493e-02 -6.25463426e-01 -5.90863883e-01
4.58138525e-01 -1.13291895e+00 7.42546096e-03 3.15630943e-01
-6.02609336e-01 -3.52259666e-01 1.34686983e+00 -7.48856068e-02
9.90271568e-01 -2.40234208e+00 -6.91500187e-01 -1.27651379e-01
2.83314139e-01 5.71943879e-01 -2.34291852e-02 3.36464867e-02
-8.29805899e-03 -5.50292917e-02 4.10210714e-03 -1.85020864e-01
-7.06836954e-02 3.81852239e-01 -2.25694448e-01 3.64547789e-01
1.38097346e-01 7.01811075e-01 -8.46470892e-01 -9.31493521e-01
6.99652284e-02 -2.12691829e-01 -2.69367665e-01 2.56865919e-01
-3.62268746e-01 -5.94182499e-03 -4.85972792e-01 1.12877405e+00
9.60084081e-01 -4.83453423e-01 -1.47344723e-01 -1.13318361e-01
1.47105023e-01 -3.38253200e-01 -1.58543098e+00 1.29984689e+00
1.78432450e-01 4.15446699e-01 -1.00276381e-01 -8.26405346e-01
7.65733778e-01 2.83187404e-02 3.39319021e-01 -5.32273710e-01
3.24034721e-01 4.81425554e-01 -1.48398399e-01 -5.65604031e-01
2.85837173e-01 -2.15548705e-02 4.04265970e-01 4.77600634e-01
3.52528811e-01 9.06085670e-02 2.37299085e-01 4.43804473e-01
1.05384207e+00 6.39625192e-02 3.55747044e-01 -2.82493323e-01
2.72757053e-01 3.30971003e-01 6.13178015e-01 9.98699605e-01
-8.98828030e-01 7.46319711e-01 2.79785216e-01 -2.40319967e-01
-8.48394215e-01 -1.08940327e+00 -6.21388435e-01 1.04428804e+00
6.92400813e-01 -3.42720300e-02 -8.26178133e-01 -1.16710901e+00
2.16905832e-01 3.67805362e-01 -4.24599797e-01 -1.21146210e-01
1.73756965e-02 -8.97239387e-01 4.22042459e-01 4.37028766e-01
5.86447716e-01 -8.41706872e-01 -2.90158063e-01 5.74999899e-02
1.03059318e-02 -1.40035510e+00 -6.67383552e-01 4.47416395e-01
-9.25335765e-01 -1.24465823e+00 -7.66498148e-01 -1.05864978e+00
9.48497772e-01 8.88778448e-01 8.16386163e-01 -4.46312204e-02
-3.85932416e-01 3.81092906e-01 -4.52785969e-01 -7.96143055e-01
-5.11260629e-01 -1.99582919e-01 2.88445234e-01 1.70103759e-01
7.17132688e-01 -4.31026965e-02 -4.64231580e-01 7.53760397e-01
-1.12019777e+00 -2.31836930e-01 6.22281671e-01 9.16373909e-01
7.40435243e-01 1.09459728e-01 6.71997309e-01 -9.93683279e-01
1.83449499e-02 -5.06252527e-01 -6.24004543e-01 2.86137253e-01
-9.52935994e-01 2.24607289e-02 4.43034828e-01 -7.64710665e-01
-1.08488345e+00 2.93491185e-01 2.40594164e-01 -8.58555377e-01
-3.40620965e-01 -2.30595749e-02 -3.79318029e-01 -5.31292111e-02
8.59984100e-01 1.32569879e-01 1.17788441e-01 -5.29247284e-01
1.80586338e-01 1.17761421e+00 3.13479334e-01 -3.10785204e-01
9.34194028e-01 7.83667207e-01 -3.33369017e-01 -7.70072699e-01
-1.37333560e+00 -8.74828100e-01 -5.42336524e-01 -1.67291194e-01
5.28029740e-01 -1.30715966e+00 -2.03655772e-02 7.96545446e-01
-7.31929064e-01 -8.93328264e-02 -4.77406830e-01 4.44604546e-01
-1.77696541e-01 7.00923800e-01 -4.63073522e-01 -7.94996083e-01
-2.53265619e-01 -1.12483180e+00 1.22890842e+00 4.16535795e-01
3.49904418e-01 -5.50155878e-01 -2.80199111e-01 5.35135984e-01
-2.74338517e-02 -1.02581054e-01 5.09214282e-01 -1.13483810e+00
-5.78614533e-01 -2.20105454e-01 -4.37764496e-01 5.94358623e-01
5.81760742e-02 -7.98074156e-02 -1.21594059e+00 -3.65093380e-01
1.30394101e-02 -6.54426932e-01 7.91564941e-01 1.18916698e-01
9.28037703e-01 5.82867973e-02 -4.04409319e-01 3.73275816e-01
1.44944286e+00 -6.76522627e-02 3.94284397e-01 2.63790548e-01
6.59707963e-01 5.38304687e-01 1.18260837e+00 3.90003979e-01
1.27053827e-01 5.00393748e-01 5.95726132e-01 -1.99757189e-01
-4.45083857e-01 -4.12661493e-01 4.03221637e-01 4.39095527e-01
4.64642197e-01 -2.58022487e-01 -8.43137145e-01 6.05190754e-01
-1.79863119e+00 -6.03954077e-01 -4.07631397e-01 2.13928699e+00
7.48818457e-01 4.77207601e-01 2.47019187e-01 1.08260296e-01
1.09609282e+00 -1.63615882e-01 -8.72569382e-01 2.16310412e-01
-2.86496341e-01 -3.65960330e-01 7.23110199e-01 -1.49648875e-01
-1.35089982e+00 9.04197037e-01 5.97135735e+00 9.74068522e-01
-8.35232437e-01 4.22442198e-01 5.48227966e-01 9.72738266e-02
2.11848527e-01 -7.83092603e-02 -1.28647923e+00 5.40166318e-01
3.11657965e-01 7.21043870e-02 -7.57573321e-02 1.34181011e+00
8.00144076e-02 -4.40975696e-01 -9.63810980e-01 1.01365042e+00
3.51666570e-01 -7.46493459e-01 -1.07043795e-02 -2.56660562e-02
1.01102865e+00 2.97376335e-01 -2.42490202e-01 5.28829753e-01
7.75599182e-02 -3.08077425e-01 8.05493474e-01 -2.01521784e-01
6.74273252e-01 -1.57850027e-01 9.48661745e-01 7.72480786e-01
-9.18578088e-01 -3.43982011e-01 -5.13469100e-01 1.60289153e-01
-1.11618273e-01 8.55346084e-01 -8.50025773e-01 1.85475662e-01
8.89207900e-01 5.65303206e-01 -8.32973599e-01 1.11012936e+00
-9.57591832e-02 6.63954020e-01 -7.50520647e-01 6.50652125e-02
1.01583123e-01 1.10148303e-02 5.38568854e-01 7.49590397e-01
5.30010927e-03 5.71045429e-02 4.65362102e-01 7.73471594e-01
-1.60458848e-01 1.70850635e-01 -5.85009992e-01 -3.07568163e-03
4.76616293e-01 1.08575463e+00 -1.04328907e+00 -5.03370523e-01
-5.97450316e-01 7.36093760e-01 2.23055393e-01 2.00802431e-01
-1.04711914e+00 -1.58546194e-01 1.61978602e-01 2.95409888e-01
4.61140484e-01 7.29817301e-02 -1.51522711e-01 -1.46565914e+00
1.70541093e-01 -7.88177073e-01 7.05674171e-01 -8.58681738e-01
-1.49729025e+00 2.73926705e-01 -3.71666588e-02 -1.43895292e+00
4.48879719e-01 -6.38545632e-01 -2.66392618e-01 1.69460908e-01
-1.64077032e+00 -9.77685511e-01 -4.93285269e-01 4.86199349e-01
7.75910437e-01 -7.22850189e-02 2.90551454e-01 4.70940799e-01
-7.71774888e-01 5.92590034e-01 2.80514121e-01 1.67815059e-01
9.93254185e-01 -1.17185926e+00 -1.62197798e-01 1.00774097e+00
2.99365819e-01 2.16760457e-01 5.76809943e-01 -6.62597537e-01
-1.30542064e+00 -1.28077805e+00 3.41757894e-01 -4.12418634e-01
6.00540519e-01 -3.49495620e-01 -1.00521123e+00 5.22161484e-01
-3.43540877e-01 7.42181242e-01 2.16461569e-01 -1.84906662e-01
-4.04422522e-01 -1.35895088e-01 -1.24987912e+00 2.39667818e-01
1.37346339e+00 -2.46093199e-01 -9.16298628e-01 6.72034502e-01
9.08916473e-01 -3.91567290e-01 -5.07869720e-01 5.42304516e-01
1.73696488e-01 -8.42679501e-01 8.06796610e-01 -3.11984658e-01
2.09223539e-01 -6.67499781e-01 -1.04479946e-01 -8.90874505e-01
9.11290199e-02 -1.76542535e-01 -4.97905105e-01 1.36871660e+00
1.37945846e-01 -6.12481773e-01 8.59887779e-01 3.29176813e-01
-1.41580954e-01 -5.71455061e-01 -7.35356629e-01 -1.28180981e+00
-3.83438051e-01 -1.95144281e-01 2.46666610e-01 9.10378516e-01
-2.44790226e-01 4.48359817e-01 -2.21086994e-01 4.60401893e-01
8.83185983e-01 1.80995122e-01 8.46105337e-01 -1.09788787e+00
-3.35034281e-01 6.76223710e-02 -5.78301609e-01 -1.25381470e+00
-1.66993812e-01 -7.37593174e-01 2.66935289e-01 -1.13333774e+00
5.23395360e-01 -6.14082396e-01 -3.12663406e-01 6.34439707e-01
-4.04229522e-01 4.80740726e-01 1.67392224e-01 6.34659171e-01
-1.05916572e+00 6.54459834e-01 1.30877721e+00 -1.56470150e-01
-1.37423843e-01 -8.48233327e-02 -4.66130853e-01 8.36940229e-01
3.61589402e-01 -8.21520686e-01 -4.83195901e-01 -3.19044709e-01
-4.41110730e-01 -3.86434853e-01 3.02228868e-01 -1.09636581e+00
7.58745596e-02 7.96517581e-02 1.15124300e-01 -8.31164598e-01
-1.39823273e-01 -1.04489601e+00 -2.79086471e-01 5.33474982e-01
-1.14350924e-02 -5.82361877e-01 4.39849421e-02 1.07892859e+00
-2.71311045e-01 -3.33520025e-01 1.01284671e+00 1.34481257e-02
-1.15512264e+00 3.10959071e-01 -1.77935977e-02 3.53141069e-01
1.51941466e+00 -5.78170955e-01 -3.90138537e-01 4.83536422e-02
-6.52827322e-01 4.55842257e-01 4.52252299e-01 6.00372672e-01
3.09104711e-01 -1.10948956e+00 -3.34934413e-01 3.30908269e-01
6.69923425e-01 2.17492938e-01 1.21516936e-01 8.50253582e-01
-4.59700406e-01 1.56117484e-01 1.23431183e-01 -1.01438034e+00
-1.18428183e+00 9.84655380e-01 3.76507103e-01 -8.21077228e-02
-5.32499552e-01 6.14697695e-01 3.57690871e-01 -5.69621682e-01
6.09082341e-01 -1.50793791e-01 2.11102217e-02 9.89665985e-02
3.89738530e-01 2.40054429e-01 1.66496366e-01 -4.24434870e-01
-2.85923630e-01 2.51265109e-01 -1.66475937e-01 4.97811943e-01
9.86902118e-01 -3.74028057e-01 3.52406859e-01 5.21580994e-01
1.06322062e+00 -2.98045129e-01 -1.37675071e+00 -6.37080848e-01
5.58435209e-02 -5.15177906e-01 -7.04603037e-03 -6.97426856e-01
-9.87247169e-01 6.28780365e-01 9.73826468e-01 2.05205008e-01
9.04787421e-01 2.99877524e-01 8.05644274e-01 4.25226331e-01
7.06946969e-01 -1.45870352e+00 3.73199880e-01 4.60623443e-01
2.65804052e-01 -1.96573532e+00 7.63720870e-02 -8.30186605e-01
-6.64535880e-01 8.40641022e-01 1.08168852e+00 1.70827881e-02
8.37321162e-01 1.87431887e-01 2.19463691e-01 -1.97572500e-01
-4.69910651e-01 -4.93306398e-01 -1.04613248e-02 5.29073656e-01
-1.86581403e-01 -9.35416520e-02 -1.70551702e-01 4.47529107e-01
4.58268762e-01 5.04582375e-02 5.33924699e-01 9.83629823e-01
-6.70378029e-01 -7.16633976e-01 -7.39843249e-01 6.39505684e-01
-1.69136271e-01 1.45839438e-01 -2.72904932e-01 9.50087070e-01
3.59673560e-01 8.84201586e-01 -6.94796890e-02 -7.04024658e-02
2.89946824e-01 1.13560492e-02 1.65034428e-01 -8.92307281e-01
-1.55635118e-01 1.86956689e-01 -1.98128074e-01 -3.68271083e-01
-5.33840537e-01 -5.83268166e-01 -1.39798498e+00 2.52786607e-01
-1.09397388e+00 -5.67408651e-02 3.64770472e-01 9.12461579e-01
4.36244160e-01 2.15298966e-01 7.05255747e-01 -4.46075261e-01
-1.25685322e+00 -9.13087130e-01 -9.96086478e-01 8.75140607e-01
1.17297232e-01 -1.01446116e+00 -7.18716860e-01 6.86143115e-02] | [9.272408485412598, 1.3872967958450317] |
e1049783-2e66-4ab6-b373-678d3a464c93 | robust-event-stream-pattern-tracking-based-on | 1803.06490 | null | http://arxiv.org/abs/1803.06490v1 | http://arxiv.org/pdf/1803.06490v1.pdf | Robust event-stream pattern tracking based on correlative filter | Object tracking based on retina-inspired and event-based dynamic vision
sensor (DVS) is challenging for the noise events, rapid change of event-stream
shape, chaos of complex background textures, and occlusion. To address these
challenges, this paper presents a robust event-stream pattern tracking method
based on correlative filter mechanism. In the proposed method, rate coding is
used to encode the event-stream object in each segment. Feature representations
from hierarchical convolutional layers of a deep convolutional neural network
(CNN) are used to represent the appearance of the rate encoded event-stream
object. The results prove that our method not only achieves good tracking
performance in many complicated scenes with noise events, complex background
textures, occlusion, and intersected trajectories, but also is robust to
variable scale, variable pose, and non-rigid deformations. In addition, this
correlative filter based event-stream tracking has the advantage of high speed.
The proposed approach will promote the potential applications of these
event-based vision sensors in self-driving, robots and many other high-speed
scenes. | ['Luping Shi', 'Hongmin Li'] | 2018-03-17 | null | null | null | null | ['event-based-vision'] | ['computer-vision'] | [-8.86213183e-02 -9.08774912e-01 3.05196673e-01 -2.85655651e-02
2.14619949e-01 -2.72105753e-01 6.75149202e-01 5.48516288e-02
-5.21540821e-01 4.17430490e-01 -9.38648358e-03 2.81092077e-01
-8.03865939e-02 -7.12685168e-01 -7.87447333e-01 -7.04214334e-01
-2.52350539e-01 -1.19288497e-01 1.14988816e+00 -7.29859844e-02
3.64577591e-01 1.00439441e+00 -2.07079720e+00 1.09997407e-01
2.70213395e-01 1.23390746e+00 4.15160447e-01 7.73938835e-01
-1.04090199e-01 9.01329994e-01 -3.97913694e-01 3.62717658e-02
2.49765813e-01 -2.05644399e-01 2.71315426e-01 4.97334190e-02
3.27567846e-01 -4.86190051e-01 -8.81205976e-01 1.07100284e+00
4.88476217e-01 1.64695546e-01 6.32895470e-01 -1.01974356e+00
-4.91097808e-01 1.82796866e-02 -8.83123219e-01 7.76426196e-01
2.55460024e-01 5.54217637e-01 1.95398200e-02 -6.51213467e-01
7.56528258e-01 1.50081873e+00 6.59590304e-01 6.04960680e-01
-6.17282510e-01 -8.24703574e-01 3.92705724e-02 4.85491484e-01
-9.51382041e-01 -1.65550232e-01 7.39965856e-01 -6.32807314e-01
6.70414746e-01 -1.95812941e-01 9.69254613e-01 9.00189996e-01
7.46917725e-01 4.86766696e-01 7.82033741e-01 2.46530958e-03
2.62829900e-01 -5.27385354e-01 1.11449175e-01 8.39597106e-01
5.21943212e-01 6.61172628e-01 -7.07275808e-01 1.82089448e-01
1.03640437e+00 4.29808676e-01 -1.45531594e-04 -1.30184606e-01
-1.50804543e+00 4.08295542e-01 3.66934896e-01 -2.33109500e-02
-4.27795768e-01 6.97427213e-01 4.58984524e-01 8.87771547e-02
-1.34225011e-01 -2.83602655e-01 -3.41200382e-01 -2.13947743e-01
-5.92361212e-01 4.32761252e-01 5.68099678e-01 7.74942756e-01
3.87137473e-01 4.05848533e-01 -4.08957005e-01 2.21980676e-01
6.95776880e-01 8.36647153e-01 6.31403744e-01 -8.00509453e-01
2.35804692e-02 4.96831745e-01 3.07728589e-01 -1.10049617e+00
-5.95955312e-01 -3.03921491e-01 -7.58688927e-01 6.30134463e-01
3.57990533e-01 -2.25710556e-01 -1.02686870e+00 1.38980746e+00
6.53756142e-01 5.42275369e-01 -1.28940074e-02 1.07679605e+00
9.46435034e-01 8.83864403e-01 8.15564618e-02 -5.22923112e-01
1.59475207e+00 -3.20138991e-01 -9.10195231e-01 2.20510602e-01
-3.09353739e-01 -8.35698724e-01 3.44307363e-01 7.05332682e-02
-9.01258111e-01 -8.93850207e-01 -1.11600220e+00 1.44487470e-01
-2.73158371e-01 8.91163275e-02 6.46660984e-01 2.76183903e-01
-5.76444089e-01 3.82015467e-01 -1.20091057e+00 -4.61167246e-01
4.34336811e-01 3.33281815e-01 -1.91483367e-03 8.15714151e-02
-8.65364134e-01 5.76828063e-01 3.67765844e-01 1.44304112e-01
-1.00922561e+00 -4.02342975e-01 -5.91585636e-01 -1.80468336e-01
-1.35275707e-01 -5.11663675e-01 1.02815044e+00 -7.25060463e-01
-1.43129790e+00 4.99930382e-01 -1.31510332e-01 -6.84553146e-01
4.90451217e-01 -1.62398852e-02 -4.82142210e-01 9.11596045e-02
-2.62785792e-01 5.01913071e-01 9.68694866e-01 -5.40829599e-01
-8.54989767e-01 -3.84137034e-01 -3.70074630e-01 2.07378000e-01
-2.91869938e-02 3.81358117e-01 -2.15659589e-01 -4.47554767e-01
1.69407815e-01 -5.69451034e-01 -1.10943966e-01 7.36689270e-01
3.31156284e-01 -5.05329788e-01 1.52532744e+00 1.65870658e-03
7.79531360e-01 -2.30457735e+00 -3.63394558e-01 -4.69273895e-01
-6.27881065e-02 4.02410269e-01 3.57266627e-02 2.08515465e-01
4.06642437e-01 -5.43457270e-01 2.53937423e-01 3.54615808e-01
-4.35825109e-01 9.74536166e-02 -3.48893404e-01 6.42601192e-01
5.17454147e-01 7.75071383e-01 -8.10442865e-01 -4.73188519e-01
6.52441919e-01 6.28993213e-01 -8.82533640e-02 -2.62544397e-02
-5.22310495e-01 7.17509866e-01 -5.36046565e-01 7.16816962e-01
7.07719564e-01 -3.48753221e-02 -6.28253937e-01 -4.66028988e-01
-7.40536392e-01 -3.72526795e-01 -1.40212452e+00 1.42526603e+00
2.56764531e-01 9.42642152e-01 -1.56162918e-01 -6.51832998e-01
1.29078901e+00 -1.54282656e-02 6.83953941e-01 -9.00842488e-01
4.93369818e-01 8.60036910e-02 1.06567815e-01 -8.03364158e-01
4.53793019e-01 1.41330436e-01 2.95429200e-01 3.18590328e-02
1.04059391e-01 4.06287909e-01 2.91678727e-01 -1.69133365e-01
1.32750976e+00 2.89646506e-01 -2.98412237e-02 6.33929148e-02
5.58026791e-01 -2.79283207e-02 8.51304531e-01 6.08887494e-01
-6.46643817e-01 3.73017013e-01 -1.03781044e-01 -8.41519773e-01
-1.00290728e+00 -1.03706884e+00 -3.25955719e-01 6.52184904e-01
8.81587684e-01 2.27899656e-01 -6.87469319e-02 9.69950929e-02
1.33351266e-01 3.47048603e-02 -3.74282330e-01 -1.29625127e-01
-7.18858540e-01 -8.10858071e-01 4.08568203e-01 5.84356487e-01
9.11679745e-01 -1.18883336e+00 -1.22960687e+00 7.81969965e-01
2.80714095e-01 -1.20372331e+00 -1.23171173e-01 6.26376569e-02
-9.34493959e-01 -1.18274844e+00 -5.47596753e-01 -1.03798187e+00
2.93473840e-01 3.71928006e-01 4.54146951e-01 -2.83106476e-01
-7.67177105e-01 5.64630628e-01 -1.92780182e-01 -8.44878316e-01
-3.59739065e-02 -7.56301045e-01 -4.29942980e-02 3.43081862e-01
5.17587423e-01 -3.64299446e-01 -9.76709545e-01 3.32834303e-01
-8.87705028e-01 -1.55014098e-01 1.72912717e-01 5.88870704e-01
7.99813271e-01 2.03958854e-01 3.67783070e-01 1.10142581e-01
1.17288508e-01 -4.76777703e-01 -1.21605289e+00 -3.11586726e-03
-7.75445327e-02 -1.71193734e-01 4.68907028e-01 -1.06332695e+00
-1.11163950e+00 4.62674141e-01 3.28676432e-01 -5.39854884e-01
-1.92286238e-01 6.78854957e-02 4.57446814e-01 -2.65173435e-01
7.99728811e-01 4.17520434e-01 8.43382850e-02 -8.99497718e-02
1.23871431e-01 3.60556781e-01 8.79382610e-01 -2.36646309e-01
7.56658852e-01 9.44394946e-01 3.40112925e-01 -6.89769089e-01
-1.61107510e-01 -4.57862258e-01 -2.45605081e-01 -8.24940562e-01
1.00248897e+00 -1.26218021e+00 -1.09347153e+00 1.00266767e+00
-1.54788876e+00 -6.51881993e-02 -1.53917417e-01 9.22383904e-01
-5.68245888e-01 8.70891735e-02 -7.82645881e-01 -1.13267863e+00
-2.67159432e-01 -8.11143398e-01 9.33170378e-01 1.07295215e+00
5.17055333e-01 -5.30701518e-01 -2.38817069e-03 -3.00933689e-01
5.85822940e-01 6.54132962e-01 2.23476738e-01 -1.33106843e-01
-1.18697977e+00 -3.83409232e-01 -3.97060066e-01 -2.07754299e-01
1.25687465e-01 4.79142815e-01 -6.37167513e-01 -1.11453578e-01
1.08344734e-01 -1.00170940e-01 9.40928578e-01 6.47182584e-01
8.93832326e-01 9.97778494e-03 -2.70120293e-01 7.28323340e-01
1.66399658e+00 6.93889439e-01 6.60699427e-01 2.44766131e-01
5.44203341e-01 3.67962271e-01 7.87092745e-01 7.11452603e-01
3.36302787e-01 3.40438545e-01 6.53521121e-01 1.93714231e-01
-4.70267147e-01 2.70676538e-02 5.51940501e-01 5.01282513e-01
-1.65760234e-01 -1.66118905e-01 -5.11173487e-01 5.15046775e-01
-1.95091021e+00 -1.37438917e+00 -5.83556712e-01 1.91566575e+00
4.20623004e-01 4.77338433e-02 6.33037984e-02 -1.50389388e-01
1.08241391e+00 -1.12965172e-02 -1.04415166e+00 -1.57836787e-02
-4.76978272e-01 -1.00824654e-01 6.44814849e-01 -2.94822127e-01
-1.13104272e+00 7.12853312e-01 5.52028704e+00 4.18709338e-01
-1.58917117e+00 -7.17363730e-02 8.95311832e-02 -1.04765147e-01
3.22736710e-01 -2.56490201e-01 -1.10970771e+00 8.59824002e-01
7.95430064e-01 -2.86522508e-01 2.37523898e-01 5.58597445e-01
4.59624678e-01 -3.10950458e-01 -5.78348875e-01 1.22334445e+00
-5.14997169e-02 -1.40525365e+00 -1.21334337e-01 -2.90735066e-01
5.67762733e-01 4.21408355e-01 7.10650086e-02 -7.66021535e-02
2.82072574e-01 -4.44503099e-01 9.14878428e-01 8.72828722e-01
6.09241903e-01 -4.07860339e-01 3.68218154e-01 2.31033504e-01
-1.58446825e+00 -2.84030080e-01 -6.64790750e-01 -1.24540843e-01
1.21312641e-01 4.57789779e-01 -2.55915821e-01 1.20290875e-01
8.77534747e-01 9.48974252e-01 -2.39901423e-01 1.59655344e+00
3.65211517e-01 4.05096292e-01 -6.32144749e-01 -4.40786958e-01
-4.59635258e-02 2.15183914e-01 8.02177012e-01 1.21413291e+00
3.65700811e-01 3.37676227e-01 7.47700036e-02 7.31278658e-01
3.10223579e-01 -3.37080270e-01 -7.51954615e-01 -2.99385358e-02
3.56701285e-01 1.06509542e+00 -9.94563997e-01 -2.96417624e-01
-5.52932680e-01 6.62208378e-01 -1.59253433e-01 2.29891956e-01
-8.67960095e-01 -4.18481857e-01 5.01294851e-01 1.10862121e-01
7.00428307e-01 -7.26665199e-01 2.75440756e-02 -1.07809889e+00
8.56377184e-03 -2.03029752e-01 2.18429565e-01 -8.63318920e-01
-1.19225287e+00 2.31746256e-01 -2.93593347e-01 -1.52582765e+00
1.47988155e-01 -6.49921834e-01 -7.31040299e-01 4.55775529e-01
-1.72103405e+00 -9.83513653e-01 -6.43804014e-01 8.93622160e-01
6.74097896e-01 -3.22410822e-01 1.82368457e-01 3.56757641e-01
-3.34859759e-01 9.15419012e-02 2.14281127e-01 1.48580804e-01
5.24922013e-01 -6.08558357e-01 1.30849063e-01 9.58296537e-01
-1.62025481e-01 2.56329477e-01 7.10009277e-01 -9.73949254e-01
-2.02421093e+00 -1.39299870e+00 1.88635573e-01 4.43801805e-02
8.04561496e-01 -1.17199989e-02 -8.15927982e-01 1.22193567e-01
-7.18570948e-02 6.48238778e-01 -7.86578283e-02 -7.85868168e-01
-2.02730387e-01 -5.45454025e-01 -1.06143141e+00 3.22183311e-01
1.02067459e+00 -1.74520925e-01 -5.75964987e-01 1.77493498e-01
6.21769369e-01 -6.94597960e-01 -5.07494807e-01 6.27585828e-01
8.42264891e-01 -8.50519657e-01 8.42354059e-01 -2.48510152e-01
1.28765360e-01 -9.63922203e-01 5.03662303e-02 -5.15653908e-01
-5.45425177e-01 -6.12550378e-01 -3.99695814e-01 9.51133132e-01
-3.71495217e-01 -5.05389631e-01 5.96433163e-01 -5.87559417e-02
-1.34203076e-01 -2.43233427e-01 -1.03417885e+00 -9.15387213e-01
-6.22380793e-01 -1.18834145e-01 3.23322535e-01 5.11017323e-01
-6.94415450e-01 -6.08580150e-02 -1.83066845e-01 4.27154124e-01
8.42373729e-01 -1.50757050e-02 5.48035085e-01 -1.49923921e+00
5.67484386e-02 -2.62529105e-01 -1.11364508e+00 -8.07425201e-01
-4.05356318e-01 -4.34796184e-01 2.88298965e-01 -1.34904456e+00
-6.85708448e-02 -2.67363489e-01 -3.59856248e-01 -9.10921581e-03
-1.53790286e-03 1.76569760e-01 1.49738774e-01 3.69327873e-01
-8.28506947e-01 6.84693217e-01 1.22361732e+00 -1.46823853e-01
-1.66126594e-01 -1.68192834e-01 2.28066444e-01 5.88244200e-01
6.05591416e-01 -3.96316826e-01 -1.65007323e-01 -4.54917043e-01
-3.58995982e-02 2.49683440e-01 8.16888094e-01 -1.62145281e+00
1.00893891e+00 -2.18278408e-01 8.08698118e-01 -9.80723441e-01
2.21667722e-01 -8.02714407e-01 5.00613511e-01 9.45934296e-01
2.95030028e-02 2.77338982e-01 4.38498676e-01 1.20588768e+00
-1.14994906e-01 1.66301936e-01 1.03846395e+00 -1.15057133e-01
-9.96765137e-01 6.21688366e-01 -5.69191217e-01 -3.36126089e-01
1.27764654e+00 -6.42756522e-01 -5.17007113e-01 1.52758107e-01
-3.27994466e-01 7.00457916e-02 2.39443570e-01 4.72990394e-01
6.56181931e-01 -1.41849887e+00 -6.01850510e-01 3.63786459e-01
9.89433080e-02 5.52969240e-02 2.36342236e-01 6.21956646e-01
-7.45957255e-01 1.29000962e-01 -6.21561110e-01 -9.68255997e-01
-1.11127996e+00 4.71382499e-01 2.44750157e-01 4.66825724e-01
-9.00044918e-01 6.80039823e-01 5.16036302e-02 4.27756935e-01
4.52488989e-01 -4.33751613e-01 -2.25825310e-01 -4.07508500e-02
7.69909918e-01 4.42191482e-01 -3.76319945e-01 -5.75541258e-01
-1.98702872e-01 9.91568863e-01 4.41158451e-02 6.41836822e-02
1.08093286e+00 -8.22370946e-02 7.60439113e-02 5.94246924e-01
7.23889291e-01 -4.05524671e-01 -1.79593611e+00 -3.00442994e-01
-2.04209790e-01 -2.10386142e-01 -6.02836022e-03 -5.28867185e-01
-1.14726818e+00 7.06107199e-01 1.16826797e+00 -6.64171809e-03
1.06376028e+00 -3.29320103e-01 8.00538540e-01 2.88179725e-01
5.12448609e-01 -8.48000824e-01 3.48923028e-01 7.35355377e-01
6.70436680e-01 -1.06953466e+00 -1.48565769e-01 -9.24165845e-02
-9.64020565e-02 1.59913659e+00 8.46254110e-01 -5.25248826e-01
7.67271638e-01 4.08160359e-01 -2.91000269e-02 -2.27932855e-01
-9.24366117e-01 -5.42669117e-01 -1.42489955e-01 8.49637866e-01
-7.04181641e-02 -2.56463736e-01 -4.96176541e-01 7.39732943e-03
2.65163988e-01 4.48801041e-01 4.47729349e-01 8.94470513e-01
-8.29360604e-01 -4.68044877e-01 -6.21906698e-01 3.36090475e-01
-3.90090346e-01 3.40963870e-01 1.77174956e-01 5.22751451e-01
4.58116412e-01 6.99598730e-01 5.30913293e-01 -3.91361266e-01
2.91956931e-01 -3.29359949e-01 6.10320449e-01 -8.70790556e-02
-3.67829144e-01 1.72741055e-01 -5.31641543e-01 -5.18101215e-01
-6.97421789e-01 -8.17849874e-01 -1.57399225e+00 -7.11721629e-02
-2.29757458e-01 -5.51731110e-01 1.00695312e+00 5.81728637e-01
4.83016551e-01 6.01721466e-01 5.43614686e-01 -6.68605328e-01
-1.20425105e-01 -6.77613020e-01 -2.69659609e-01 2.85518736e-01
5.56726456e-01 -1.05427146e+00 -3.47930759e-01 3.37942898e-01] | [8.596421241760254, -1.2518936395645142] |
f262baf4-b6f7-4b0b-b722-866ec98e2701 | a-cross-modal-distillation-network-for-person | 1810.11641 | null | https://arxiv.org/abs/1810.11641v3 | https://arxiv.org/pdf/1810.11641v3.pdf | Cross-Modal Distillation for RGB-Depth Person Re-Identification | Person re-identification is a key challenge for surveillance across multiple sensors. Prompted by the advent of powerful deep learning models for visual recognition, and inexpensive RGB-D cameras and sensor-rich mobile robotic platforms, e.g. self-driving vehicles, we investigate the relatively unexplored problem of cross-modal re-identification of persons between RGB (color) and depth images. The considerable divergence in data distributions across different sensor modalities introduces additional challenges to the typical difficulties like distinct viewpoints, occlusions, and pose and illumination variation. While some work has investigated re-identification across RGB and infrared, we take inspiration from successes in transfer learning from RGB to depth in object detection tasks. Our main contribution is a novel method for cross-modal distillation for robust person re-identification, which learns a shared feature representation space of person's appearance in both RGB and depth images. In addition, we propose a cross-modal attention mechanism where the gating signal from one modality can dynamically activate the most discriminant CNN filters of the other modality. The proposed distillation method is compared to conventional and deep learning approaches proposed for other cross-domain re-identification tasks. Results obtained on the public BIWI and RobotPKU datasets indicate that the proposed method can significantly outperform the state-of-the-art approaches by up to 16.1% in mean Average Precision (mAP), demonstrating the benefit of the distillation paradigm. The experimental results also indicate that using cross-modal attention allows to improve recognition accuracy considerably with respect to the proposed distillation method and relevant state-of-the-art approaches. | ['Frank Hafner', 'Amran Bhuiyan', 'Julian F. P. Kooij', 'Eric Granger'] | 2018-10-27 | null | null | null | null | ['cross-view-person-re-identification'] | ['computer-vision'] | [ 1.55468419e-01 -3.63280326e-01 2.19292432e-01 -3.80566806e-01
-6.90039635e-01 -5.05176067e-01 8.16573977e-01 -1.64144024e-01
-7.75581062e-01 4.65654492e-01 1.12820953e-01 3.02367121e-01
-4.18261588e-02 -4.43082660e-01 -6.65010035e-01 -8.54531825e-01
2.17930302e-01 3.64837557e-01 -1.14634678e-01 -1.72193646e-01
-9.86597613e-02 6.34268224e-01 -1.98286641e+00 1.65137097e-01
4.33573365e-01 9.74091709e-01 -1.97049707e-01 8.50634754e-01
3.46826106e-01 2.72583604e-01 -5.65807641e-01 -7.48345494e-01
6.81889236e-01 -1.28375784e-01 -5.87962806e-01 2.38004997e-01
1.06859863e+00 -5.33141315e-01 -6.78536713e-01 1.09824920e+00
8.48044515e-01 2.98547506e-01 5.58939755e-01 -1.45023751e+00
-9.43254590e-01 -1.94376573e-01 -5.96521676e-01 3.20547491e-01
6.99125588e-01 2.26981044e-01 3.19502980e-01 -6.24914646e-01
3.15965861e-01 1.30365884e+00 1.05095887e+00 1.11504447e+00
-1.12457502e+00 -7.48253047e-01 2.00979382e-01 4.11100745e-01
-1.53217411e+00 -4.13097203e-01 6.37952387e-01 -4.61523354e-01
9.79825318e-01 2.74557978e-01 6.49897933e-01 1.41566837e+00
-3.60042512e-01 7.64559150e-01 1.13171053e+00 -3.85169685e-01
-2.13680372e-01 5.06100118e-01 6.56848550e-02 5.63745379e-01
5.89830518e-01 2.84500837e-01 -5.01622558e-01 6.75947666e-02
4.85784680e-01 4.82016504e-01 -2.19445422e-01 -4.81437892e-01
-1.06477332e+00 4.11493868e-01 5.33017695e-01 3.06403935e-01
-2.88477629e-01 8.83882642e-02 3.55361521e-01 2.50751823e-01
2.20422342e-01 2.26816177e-01 -2.56309628e-01 -1.46833025e-02
-6.04044020e-01 2.19588697e-01 4.08726335e-01 7.86215484e-01
6.51445270e-01 -2.03020811e-01 -2.71677256e-01 7.27657616e-01
2.94921607e-01 9.57013249e-01 4.74345803e-01 -5.13911784e-01
5.30248761e-01 7.25802839e-01 4.22501087e-01 -1.04868472e+00
-5.39501607e-01 -1.78427741e-01 -9.77919281e-01 2.00125754e-01
6.10157371e-01 -1.89901248e-01 -8.54903102e-01 1.76099122e+00
4.10322160e-01 2.39421979e-01 2.88917363e-01 1.03939056e+00
9.70557213e-01 2.17625499e-01 2.29118049e-01 2.49602243e-01
1.58486402e+00 -7.97502816e-01 -2.39714801e-01 -2.93358773e-01
2.73871809e-01 -3.04602623e-01 4.34636980e-01 2.04597503e-01
-5.80434918e-01 -1.01104319e+00 -1.03067720e+00 1.26298055e-01
-7.25057840e-01 3.97810102e-01 4.49627966e-01 1.16688573e+00
-1.26122987e+00 2.48003110e-01 -5.94210386e-01 -1.13483000e+00
3.34200144e-01 5.13516963e-01 -9.59740639e-01 -3.55609655e-01
-1.04396141e+00 8.73618901e-01 2.83590853e-01 3.06276858e-01
-8.58238816e-01 -5.48776567e-01 -8.11875761e-01 -2.25099623e-01
5.14088608e-02 -8.79048765e-01 8.46158266e-01 -1.16819036e+00
-1.42677903e+00 1.32125568e+00 -1.32420495e-01 -3.42122763e-01
6.51066244e-01 -3.74291956e-01 -5.81912935e-01 9.63246599e-02
-1.27088800e-02 7.31552005e-01 9.02945280e-01 -1.30704796e+00
-9.19555664e-01 -9.01960850e-01 1.56680584e-01 3.16826224e-01
-6.07610524e-01 -2.94482186e-02 -5.47366083e-01 -3.68163317e-01
-2.72653520e-01 -1.07673657e+00 7.93742761e-02 7.13573843e-02
-2.84971446e-01 -3.41753244e-01 6.55699551e-01 -8.09869647e-01
4.71780717e-01 -2.03352690e+00 3.22159201e-01 8.33004434e-03
7.58458152e-02 5.03144264e-01 -1.54148936e-01 1.24976709e-01
-2.30815500e-01 -3.23501259e-01 4.02794331e-02 -8.03128779e-01
1.56001896e-02 -6.70419831e-04 1.24260768e-01 9.03876960e-01
-2.17405092e-02 8.95499527e-01 -9.23435271e-01 -4.35983054e-02
6.78066790e-01 8.67105365e-01 1.26606319e-03 4.59669858e-01
5.51922739e-01 6.61813736e-01 -2.72720214e-02 8.04951370e-01
8.45002055e-01 -1.66053546e-03 -3.95423695e-02 -6.10623002e-01
-6.30442426e-02 -2.90210515e-01 -1.19529986e+00 1.68530345e+00
-1.31781012e-01 6.66772008e-01 -8.24242309e-02 -1.09914100e+00
7.55492508e-01 3.26035887e-01 4.64229733e-01 -9.24700439e-01
3.21400702e-01 -6.67652711e-02 -3.77911597e-01 -6.00871503e-01
6.05412245e-01 2.72139981e-02 -1.58235475e-01 3.00244480e-01
1.54185489e-01 5.01655281e-01 -7.47433901e-02 -2.03360215e-01
6.98010325e-01 1.01017125e-01 1.12573475e-01 -5.81357107e-02
7.93602407e-01 -1.46668062e-01 2.86041409e-01 9.53369498e-01
-6.97593033e-01 5.61456084e-01 -2.08420157e-01 -7.39535451e-01
-9.30286586e-01 -9.35819924e-01 6.88918754e-02 1.19600761e+00
4.33452010e-01 1.93501234e-01 -7.78061152e-01 -6.38361931e-01
3.41514617e-01 1.74769595e-01 -1.03154564e+00 -2.44663537e-01
-4.95651245e-01 -1.07241774e+00 9.06979680e-01 7.40495443e-01
1.00515604e+00 -5.51124096e-01 -7.15682924e-01 -2.88525254e-01
-1.82659179e-01 -1.39077556e+00 -2.27072284e-01 -2.68282503e-01
-6.00923657e-01 -1.33931589e+00 -1.21688819e+00 -7.10637093e-01
6.79553509e-01 7.72584319e-01 7.66633809e-01 -2.26198301e-01
-5.04334569e-01 1.27716899e+00 -3.45817089e-01 -2.08745405e-01
-3.44806686e-02 -2.91662328e-02 5.74565053e-01 4.58219141e-01
9.56581354e-01 -1.20727066e-02 -8.62241745e-01 2.66805410e-01
-5.84563851e-01 -3.14331710e-01 3.49494785e-01 7.12295115e-01
1.63969815e-01 -1.54354453e-01 3.37919205e-01 -3.49776506e-01
3.24589610e-01 -2.43995368e-01 -5.18517673e-01 6.03177845e-01
-4.16787028e-01 -7.93960020e-02 1.70623407e-01 -4.75223750e-01
-1.10377347e+00 1.84307039e-01 8.07888955e-02 -3.63320827e-01
-6.45596921e-01 -2.74936706e-01 -2.27010563e-01 -4.01990414e-01
6.80814207e-01 3.18742126e-01 7.87695497e-02 -3.52849483e-01
3.14288825e-01 8.63952458e-01 9.41143334e-01 -2.56596506e-01
8.52428436e-01 8.91306460e-01 -1.99937671e-01 -8.46789658e-01
-4.28048849e-01 -7.14865804e-01 -8.88192773e-01 -3.49647790e-01
1.12604558e+00 -1.23461246e+00 -1.13828373e+00 1.08598948e+00
-1.08678460e+00 8.26854818e-03 -9.02233552e-03 3.19731295e-01
-1.26603276e-01 6.76982582e-01 -2.77803659e-01 -9.72619653e-01
-3.55039686e-01 -9.78834331e-01 1.21536624e+00 7.50948131e-01
1.03137627e-01 -9.62467134e-01 1.57549977e-01 6.64676726e-01
3.56745869e-01 3.32354546e-01 8.09333473e-02 -5.83452582e-01
-2.71601498e-01 -6.43782198e-01 -6.01519704e-01 1.96138278e-01
3.30290020e-01 -5.66324413e-01 -1.40394568e+00 -7.62488365e-01
-3.79848808e-01 -2.08323359e-01 8.59782696e-01 1.09789640e-01
7.01352596e-01 1.24002799e-01 -5.67456543e-01 7.92253435e-01
1.31420434e+00 4.46019731e-02 5.76547563e-01 7.88174927e-01
1.00846684e+00 7.55958915e-01 1.78211987e-01 4.81351882e-01
8.04912686e-01 9.13669348e-01 4.06217515e-01 -3.09884876e-01
-2.62708098e-01 -8.18699151e-02 3.23059171e-01 -3.87990288e-02
-4.38543409e-01 -1.10713169e-01 -7.70956635e-01 6.75574780e-01
-1.92601860e+00 -1.10514438e+00 2.72735476e-01 2.54546642e+00
2.67843425e-01 -5.11711776e-01 4.56154317e-01 4.05484773e-02
9.57263112e-01 -1.56822354e-01 -5.97805202e-01 8.71196911e-02
-3.10813755e-01 -2.43401811e-01 8.54635596e-01 2.82989025e-01
-1.57151508e+00 6.63502455e-01 5.71952057e+00 2.03033537e-01
-1.04500473e+00 1.95960104e-01 3.34870070e-01 5.53457113e-03
3.17325890e-01 -7.45937824e-01 -1.04766381e+00 3.58378530e-01
7.41897225e-01 3.79867673e-01 4.49858457e-01 6.36661768e-01
-3.11575741e-01 -9.00654793e-02 -1.34715426e+00 1.83516979e+00
8.23467374e-01 -7.32811272e-01 -1.60285458e-01 -1.94272861e-01
7.14854479e-01 3.59323196e-04 3.15196633e-01 2.78249592e-01
1.12365544e-01 -9.44481492e-01 5.41019022e-01 6.23832345e-01
8.81473124e-01 -6.78312659e-01 1.08189678e+00 -8.33760574e-02
-1.29914939e+00 -3.64453852e-01 -4.41758186e-01 9.18465033e-02
-4.45470549e-02 -8.42740759e-02 -7.06220269e-01 6.94420099e-01
1.29499829e+00 9.17093515e-01 -8.55189502e-01 9.78479207e-01
9.10278335e-02 -1.00749440e-01 -1.65551856e-01 2.07359269e-01
-1.23498626e-01 5.22631481e-02 5.04799187e-01 1.49292719e+00
2.35236704e-01 1.11411223e-02 -2.53757853e-02 4.91634279e-01
1.12875886e-01 -4.74343956e-01 -6.59006476e-01 3.15465510e-01
1.25436589e-01 1.05156946e+00 -2.92286903e-01 -3.58631551e-01
-5.54927349e-01 1.60411870e+00 2.30247259e-01 5.75376987e-01
-8.32507312e-01 -1.82653069e-01 1.06960559e+00 -1.27865016e-01
3.59931499e-01 -1.69971377e-01 1.13094620e-01 -1.27456450e+00
9.14906412e-02 -6.86083913e-01 6.74484789e-01 -6.29278898e-01
-1.67668796e+00 7.58804679e-01 2.07125902e-01 -1.21190667e+00
-3.11976820e-01 -1.01343369e+00 -6.64383685e-03 9.09771919e-01
-1.74836195e+00 -1.68081164e+00 -9.27066922e-01 1.03099144e+00
2.66870737e-01 -5.88935673e-01 8.77512038e-01 5.90187252e-01
-6.60076261e-01 1.11380100e+00 2.27027357e-01 4.19357240e-01
9.27403092e-01 -9.74821806e-01 3.02054197e-01 1.09920573e+00
-1.42669380e-01 5.13655663e-01 5.02551138e-01 -3.97585720e-01
-1.71690261e+00 -1.09319603e+00 5.85421324e-01 -8.52119684e-01
7.80828297e-02 -4.24226910e-01 -5.20082891e-01 5.95225930e-01
7.38176703e-02 1.66976854e-01 7.64289737e-01 9.85592697e-03
-7.72010088e-01 -4.79506224e-01 -1.45851851e+00 2.84544170e-01
1.06220567e+00 -7.06127167e-01 -5.66313505e-01 1.46121278e-01
2.09162086e-01 -2.77829885e-01 -4.98453707e-01 2.57157296e-01
9.13184881e-01 -1.10341895e+00 1.50484192e+00 -5.87157249e-01
-3.33379090e-01 -3.89139175e-01 -4.38610643e-01 -1.09089661e+00
-4.53849405e-01 -3.20757449e-01 -2.58200634e-02 1.29830611e+00
-2.33421460e-01 -8.22281599e-01 7.56563604e-01 1.01550400e+00
3.51623386e-01 7.22813681e-02 -1.21233046e+00 -6.79901004e-01
-2.58298188e-01 -3.22921127e-01 5.22681534e-01 7.03831196e-01
-3.56066853e-01 -1.28901601e-01 -7.32487619e-01 5.89765429e-01
9.71783459e-01 -9.86763388e-02 1.01183617e+00 -1.32740200e+00
-1.80004835e-01 -2.58018553e-01 -8.36698472e-01 -1.01065278e+00
1.49835378e-01 -7.01511443e-01 -4.56971563e-02 -1.37030005e+00
5.37257195e-01 -3.64901960e-01 -4.38064814e-01 3.70306343e-01
-2.84193456e-01 7.08796084e-01 3.78679156e-01 1.59352928e-01
-8.79366636e-01 5.44312298e-01 5.54150403e-01 -6.18431389e-01
-1.63647383e-01 -2.48104092e-02 -6.85296059e-01 3.65570575e-01
5.65681636e-01 -1.28901884e-01 -5.45155443e-02 -8.07612300e-01
-7.88339525e-02 -4.64424103e-01 1.06063831e+00 -1.35817385e+00
4.79892403e-01 3.07512164e-01 8.44651282e-01 -2.74468899e-01
6.43157482e-01 -1.07652760e+00 2.20686793e-01 5.44776678e-01
-2.45562382e-02 2.57339060e-01 4.12927896e-01 8.22973847e-01
2.57148147e-02 1.36304244e-01 7.23554671e-01 -2.06951156e-01
-1.09427094e+00 2.50671715e-01 -2.54768312e-01 -2.67550707e-01
1.11342037e+00 -7.58383512e-01 -3.35980296e-01 -2.37860605e-01
-5.26542902e-01 -1.38656106e-02 4.46682125e-01 8.57274175e-01
6.20610654e-01 -1.41028011e+00 -7.96234131e-01 4.34715003e-01
4.34164941e-01 -4.23135698e-01 5.69612026e-01 6.99686825e-01
-1.72176406e-01 7.05019891e-01 -5.70778608e-01 -8.34882438e-01
-1.47892988e+00 6.38141036e-01 7.10802019e-01 6.10099509e-02
-4.34587866e-01 9.69045341e-01 1.74880341e-01 -5.40378869e-01
3.80899251e-01 4.81996462e-02 -1.98642865e-01 -5.22447675e-02
9.19862926e-01 7.04205334e-01 -1.54340686e-02 -1.14716983e+00
-7.87646711e-01 9.58034277e-01 3.25601362e-02 5.92416450e-02
1.11526084e+00 -4.65720952e-01 7.04992861e-02 1.51999786e-01
1.10600770e+00 -5.57841301e-01 -1.37357783e+00 -3.25150937e-01
-3.45456183e-01 -6.36613309e-01 -3.02556664e-01 -7.39063919e-01
-9.98362303e-01 8.17083657e-01 1.58362329e+00 4.06248076e-03
1.13169408e+00 7.79305026e-02 5.98794997e-01 3.71804327e-01
5.63148499e-01 -9.42648709e-01 1.48306116e-01 2.67772108e-01
6.38036907e-01 -1.84370196e+00 -5.13378978e-02 -5.10540009e-02
-5.97802520e-01 8.74987721e-01 8.04052830e-01 6.60813600e-02
3.80862653e-01 -2.26005718e-01 1.43753111e-01 -5.19505776e-02
1.21536240e-01 -4.16772693e-01 3.94364119e-01 1.36747134e+00
1.42841756e-01 7.69788176e-02 5.58152020e-01 4.59033310e-01
2.66308218e-01 -1.82837948e-01 -2.21025031e-02 6.18919969e-01
-6.40868116e-03 -9.01078165e-01 -8.79693031e-01 -4.93325219e-02
-1.42165244e-01 3.20319176e-01 -4.00211453e-01 8.53646457e-01
5.07590830e-01 1.16413391e+00 2.32128054e-01 -6.81803465e-01
5.24706066e-01 -7.81002641e-02 7.63948619e-01 -8.46471861e-02
-7.59072065e-01 -6.14612639e-01 -2.53712803e-01 -5.10667503e-01
-9.77601230e-01 -8.46548676e-01 -5.17086625e-01 -3.04289758e-01
-3.75273712e-02 -2.90174276e-01 6.36181176e-01 9.52394843e-01
5.13489604e-01 2.74020731e-01 3.49866301e-01 -1.27655983e+00
-3.46999556e-01 -9.34681892e-01 -2.82758206e-01 8.28958631e-01
6.97918057e-01 -8.75653923e-01 -1.06636047e-01 8.36733282e-02] | [14.604158401489258, 0.9408527612686157] |
6ddb3392-6b1d-4714-8023-8f02bd17ba8c | exploring-conditional-text-generation-for | 2110.02334 | null | https://arxiv.org/abs/2110.02334v2 | https://arxiv.org/pdf/2110.02334v2.pdf | Exploring Conditional Text Generation for Aspect-Based Sentiment Analysis | Aspect-based sentiment analysis (ABSA) is an NLP task that entails processing user-generated reviews to determine (i) the target being evaluated, (ii) the aspect category to which it belongs, and (iii) the sentiment expressed towards the target and aspect pair. In this article, we propose transforming ABSA into an abstract summary-like conditional text generation task that uses targets, aspects, and polarities to generate auxiliary statements. To demonstrate the efficacy of our task formulation and a proposed system, we fine-tune a pre-trained model for conditional text generation tasks to get new state-of-the-art results on a few restaurant domains and urban neighborhoods domain benchmark datasets. | ['Thamar Solorio', 'Nedim Lipka', 'Franck Dernoncourt', 'Siva Uday Sampreeth Chebolu'] | 2021-10-05 | null | null | null | null | ['conditional-text-generation'] | ['natural-language-processing'] | [ 5.80972254e-01 3.53339821e-01 -1.22105077e-01 -7.18207002e-01
-1.22331297e+00 -6.34949923e-01 1.02266049e+00 3.89600128e-01
-2.92971972e-02 9.71369267e-01 7.11594701e-01 -5.94867051e-01
5.43495953e-01 -1.03596663e+00 -6.30029678e-01 -4.11739171e-01
5.02471626e-01 6.30905151e-01 -1.88756019e-01 -6.25458717e-01
2.97303468e-01 2.07767021e-02 -1.32240653e+00 8.26654017e-01
1.04739046e+00 9.09568012e-01 -2.40031064e-01 7.09747195e-01
-3.19484502e-01 8.59596968e-01 -1.01844656e+00 -8.69568825e-01
-1.24683365e-01 -7.85910487e-01 -6.49529874e-01 2.33544901e-01
1.97209075e-01 1.91396773e-01 6.24032080e-01 8.11675727e-01
3.37106347e-01 2.56617874e-01 1.13587487e+00 -1.09477603e+00
-8.29404950e-01 8.57522428e-01 -5.69327354e-01 -1.92929909e-01
5.23502469e-01 9.04188603e-02 1.56833994e+00 -1.01563513e+00
5.61460018e-01 1.19598019e+00 4.76814479e-01 6.06854320e-01
-1.07820189e+00 -4.69290197e-01 6.13908410e-01 -4.88479942e-01
-7.72082984e-01 -1.97590947e-01 8.75325084e-01 -3.15231532e-01
1.19506037e+00 2.95320898e-01 6.87350810e-01 1.22741222e+00
2.55473465e-01 9.17786419e-01 1.13330221e+00 -5.14168978e-01
4.80995715e-01 5.14849067e-01 3.04342389e-01 1.68715060e-01
2.97423631e-01 -4.01291817e-01 -5.46154857e-01 -3.04312974e-01
-1.24968514e-01 -6.39106393e-01 1.29687220e-01 1.62756935e-01
-1.24015594e+00 1.28945947e+00 6.41118810e-02 2.60646250e-02
-5.03596008e-01 -1.87077880e-01 3.24207872e-01 -1.01952218e-01
9.69417214e-01 8.54397774e-01 -6.71016216e-01 8.79653823e-03
-7.47159302e-01 6.28835738e-01 1.01095068e+00 1.05108154e+00
6.34123325e-01 1.63717449e-01 -7.43453145e-01 7.26305783e-01
4.62023944e-01 8.49382937e-01 3.57978076e-01 -3.06774020e-01
7.57669926e-01 8.43929529e-01 3.20153713e-01 -8.66765976e-01
-3.64789963e-02 -2.53959537e-01 -3.50808948e-01 -2.09568962e-02
5.86508028e-03 -5.96754611e-01 -1.06253982e+00 1.55624914e+00
4.34081793e-01 -4.35110688e-01 4.65156704e-01 4.66259956e-01
1.03575456e+00 1.11196661e+00 3.66742074e-01 -8.26199576e-02
1.65433788e+00 -1.32420838e+00 -6.64690316e-01 -7.31320798e-01
6.15890503e-01 -9.17115569e-01 1.30967093e+00 1.04322851e-01
-9.38357711e-01 -4.26110238e-01 -1.08813000e+00 -2.21900064e-02
-6.67164147e-01 4.77935135e-01 6.81334496e-01 7.35454619e-01
-8.94569695e-01 -6.62570447e-02 -4.51621413e-01 -8.99939463e-02
1.32271051e-01 -1.21239141e-01 -4.10398915e-02 2.92911142e-01
-1.32231259e+00 6.05072320e-01 1.93611577e-01 -1.48084402e-01
-7.01711118e-01 -7.82864749e-01 -1.26479936e+00 2.76837759e-02
-1.13396443e-01 -8.91783535e-01 1.40220189e+00 -1.08780181e+00
-1.47736394e+00 8.38840663e-01 -4.90744859e-01 -3.81483644e-01
5.02864458e-02 -1.97420597e-01 -5.13878107e-01 -1.06357001e-01
7.97634721e-01 7.32753038e-01 8.42354298e-01 -1.24203980e+00
-1.02169394e+00 -3.08703184e-01 2.92873681e-01 5.03227651e-01
-1.97251916e-01 2.29363993e-01 -1.42798364e-01 -8.84217143e-01
-5.48184574e-01 -8.43464077e-01 -4.15679693e-01 -6.41657591e-01
-7.77266085e-01 -4.22193378e-01 4.96429116e-01 -4.94612455e-01
1.20277131e+00 -2.01362896e+00 -3.40465069e-01 1.50509864e-01
-3.81780654e-01 1.07328817e-01 -3.51772130e-01 3.48949790e-01
-1.57187283e-01 1.60194188e-01 -3.52486134e-01 -4.26703244e-01
1.42747000e-01 -1.93111002e-01 -9.15025711e-01 -3.42709154e-01
6.48550570e-01 8.31712902e-01 -1.00009227e+00 -4.35598284e-01
-1.00369327e-01 4.10221577e-01 -6.30080700e-01 3.48398060e-01
-7.38459289e-01 2.55628735e-01 -6.99612916e-01 5.72538137e-01
3.84951502e-01 -1.91136420e-01 -4.70961928e-02 -3.05585772e-01
-5.38497558e-03 1.00406444e+00 -8.13301444e-01 1.13064766e+00
-7.59216845e-01 4.57594335e-01 -3.38360786e-01 -5.91595292e-01
1.19719613e+00 2.07783699e-01 1.61302388e-01 -5.18118978e-01
1.63302589e-02 1.01004625e-02 -4.14232939e-01 -9.00664926e-02
1.11857951e+00 -4.37752604e-01 -6.05234742e-01 9.11982775e-01
-6.53869063e-02 -8.12298357e-01 7.83192158e-01 4.56492424e-01
7.74139822e-01 2.64934689e-01 4.38752979e-01 -2.38272920e-01
6.57160640e-01 3.63641530e-01 2.36760452e-01 5.34859300e-01
7.79251158e-02 7.82704830e-01 7.92471290e-01 -2.84157127e-01
-8.08407903e-01 -1.04684949e+00 1.70467168e-01 1.15539551e+00
-2.70622373e-01 -6.18227541e-01 -6.13604903e-01 -1.07473195e+00
-3.85535777e-01 1.60864508e+00 -1.02698100e+00 4.17460203e-02
-2.37249717e-01 -1.09660399e+00 2.35846728e-01 4.93203014e-01
2.75642723e-01 -1.20905030e+00 -1.69329032e-01 1.85494795e-02
-5.86652577e-01 -1.14877069e+00 -7.18594074e-01 8.28055665e-02
-5.74322224e-01 -8.02124441e-01 -5.85453510e-01 -9.06858861e-01
9.15174007e-01 -6.32174015e-02 1.60724473e+00 -6.68582737e-01
3.74814093e-01 8.82073492e-02 -6.09111726e-01 -9.12138402e-01
-6.72246575e-01 2.29966089e-01 -4.65917826e-01 1.42804056e-01
4.96605515e-01 -1.86216712e-01 -4.36354727e-01 -6.87503070e-02
-8.24120998e-01 2.46638983e-01 4.88563657e-01 6.89297259e-01
7.38410890e-01 -1.38680115e-01 5.88217139e-01 -1.42056477e+00
1.25458670e+00 -2.65694946e-01 -3.40058833e-01 3.72720897e-01
-5.36862850e-01 -3.82940732e-02 6.35963976e-01 -1.83481783e-01
-1.49621415e+00 2.45650113e-02 -3.46649289e-01 6.92683458e-01
-1.62125811e-01 7.93241143e-01 -2.32957810e-01 9.13868129e-01
8.14460516e-01 2.69578069e-01 -3.99822474e-01 1.81729138e-01
6.84326470e-01 6.74661577e-01 1.81991681e-01 -4.82559085e-01
5.96997738e-01 4.53512043e-01 -2.99397737e-01 -5.20578384e-01
-1.68043816e+00 -3.71638387e-01 -1.71502471e-01 -8.38968158e-02
1.05298126e+00 -1.23862982e+00 -4.31764163e-02 2.87522703e-01
-1.36680496e+00 -2.63604075e-01 -5.24812996e-01 2.00158760e-01
-3.44225883e-01 -2.12125793e-01 -1.89013600e-01 -7.34685361e-01
-9.35937583e-01 -1.03924191e+00 1.43377841e+00 2.21674055e-01
-7.29849696e-01 -1.08172703e+00 4.21176314e-01 7.33127296e-01
3.63902718e-01 2.54746914e-01 1.06898928e+00 -9.25177872e-01
-9.04313102e-02 -3.07913631e-01 -6.81466013e-02 5.36974847e-01
4.56492782e-01 2.84237325e-01 -8.29982460e-01 1.36681765e-01
2.55722795e-02 -2.57455200e-01 6.59704089e-01 3.43413562e-01
4.59839404e-01 -5.22310615e-01 -1.13559693e-01 2.21214563e-01
9.92366433e-01 1.81532532e-01 5.37821352e-01 3.49772185e-01
6.01452470e-01 6.24949276e-01 1.07250786e+00 3.70257050e-01
6.90284252e-01 3.60215664e-01 2.18222998e-02 -1.53547093e-01
-9.01770517e-02 -4.52237844e-01 7.79063165e-01 8.07865500e-01
3.72482896e-01 -3.41193944e-01 -5.62508464e-01 9.64979708e-01
-1.52828896e+00 -8.55116606e-01 -2.78348416e-01 1.62261474e+00
1.32107711e+00 3.89331102e-01 -1.23809429e-03 -3.35545808e-01
4.97555584e-01 5.86902320e-01 -2.61747718e-01 -8.45092654e-01
-2.19586581e-01 3.59644890e-01 -1.90717369e-01 5.13801455e-01
-1.25235355e+00 1.14575064e+00 6.15252686e+00 7.68559992e-01
-9.60842073e-01 -8.15773383e-02 1.15755820e+00 1.45533949e-01
-9.46532547e-01 1.24116801e-01 -1.11767995e+00 2.26442143e-01
8.59970212e-01 -4.40306664e-01 -9.26545113e-02 1.19326890e+00
2.44273767e-01 -1.02499530e-01 -1.08659315e+00 2.70823032e-01
5.13419449e-01 -1.16281295e+00 8.10524821e-01 -4.88767475e-01
1.29487312e+00 -2.05791473e-01 1.75592899e-01 5.08739471e-01
7.02383637e-01 -8.96840096e-01 7.41063833e-01 2.12638322e-02
6.03597879e-01 -7.84042239e-01 8.46503258e-01 -1.52380586e-01
-1.06728292e+00 4.55055922e-01 -8.55087489e-02 -8.56920853e-02
4.15159941e-01 9.88338411e-01 -1.13854110e+00 3.85085136e-01
4.37764585e-01 6.50817692e-01 -5.40266275e-01 2.55100548e-01
-1.04889262e+00 8.23072672e-01 4.77415286e-02 -5.11374414e-01
5.01100838e-01 -3.07190210e-01 5.83591938e-01 1.31781352e+00
3.49598467e-01 -9.38453674e-02 -2.42523298e-01 1.01492798e+00
-1.94336548e-01 3.57605368e-01 -5.76056898e-01 -2.55939662e-01
1.93590559e-02 1.47019482e+00 -6.96986556e-01 -6.46839619e-01
-4.07384306e-01 7.12354600e-01 3.54440659e-02 6.00369692e-01
-7.60089397e-01 -7.73024738e-01 7.77528703e-01 -7.77000487e-02
5.70345581e-01 2.44157925e-01 -7.46602058e-01 -1.21804786e+00
6.94952533e-02 -1.08570158e+00 3.35105836e-01 -1.09312844e+00
-1.22225821e+00 9.33645785e-01 -1.90502122e-01 -1.09480274e+00
-5.50505817e-01 -5.30611873e-01 -1.14739537e+00 1.03032446e+00
-1.47917604e+00 -1.42730010e+00 6.93838522e-02 2.46075884e-01
6.81288898e-01 -1.86401099e-01 9.26177561e-01 -2.51009226e-01
-2.81519234e-01 4.61620122e-01 -3.65095913e-01 2.95669556e-01
7.30472624e-01 -1.30929697e+00 9.93314087e-01 9.96104300e-01
1.01582222e-01 6.05291903e-01 1.01486123e+00 -8.57014835e-01
-8.82578373e-01 -1.57333744e+00 1.53322673e+00 -7.41584778e-01
8.31935883e-01 -4.09506738e-01 -3.20746213e-01 7.35735953e-01
4.74234045e-01 -7.32214808e-01 1.02930617e+00 3.05930614e-01
-3.43883038e-01 -1.59625739e-01 -9.09726620e-01 9.41688240e-01
2.47901455e-01 -3.54665428e-01 -7.47141659e-01 3.84770185e-01
1.14901626e+00 -3.20466101e-01 -3.59082371e-01 3.77099246e-01
2.22815022e-01 -7.99364209e-01 6.10978007e-01 -7.94098198e-01
1.13189471e+00 -3.43878627e-01 -1.56740248e-01 -1.73844635e+00
-3.90614532e-02 -4.04936790e-01 3.97205204e-01 1.55885410e+00
1.45802629e+00 -4.95181084e-01 7.37798035e-01 7.75660574e-01
-2.16622218e-01 -9.71572876e-01 -4.41771001e-01 -5.53111658e-02
8.73420089e-02 -4.71502393e-01 7.54134178e-01 7.11212695e-01
1.54923126e-01 1.07995987e+00 -4.42587845e-02 -3.27690504e-02
1.17511213e-01 6.45122230e-01 8.11856508e-01 -5.96526265e-01
-2.05756858e-01 -3.58413279e-01 3.09636414e-01 -8.21681142e-01
2.20513046e-01 -6.56428576e-01 3.57790381e-01 -1.92868257e+00
9.11369473e-02 -1.26387328e-01 2.03168709e-02 4.24848497e-01
-8.42846572e-01 2.03942299e-01 -6.87435418e-02 -4.34306085e-01
-6.90765381e-01 8.01348984e-01 1.14493716e+00 -4.79814678e-01
-3.96925062e-01 5.15361726e-01 -1.57969677e+00 5.85201561e-01
8.20688605e-01 -3.64538640e-01 -5.85396349e-01 -3.77735794e-01
9.04378414e-01 -3.18991125e-01 -2.31605962e-01 -4.52641010e-01
-2.05925331e-01 -2.75044709e-01 2.31588066e-01 -7.89976597e-01
3.14481437e-01 -2.54164815e-01 -4.53063965e-01 1.92392826e-01
-7.76213229e-01 3.07651371e-01 2.31665269e-01 3.29230964e-01
-5.13842702e-01 -1.94321021e-01 3.77215326e-01 -6.26624823e-02
-1.16523743e-01 1.58943720e-02 -6.19786084e-01 4.46478903e-01
7.81304240e-01 3.37507039e-01 -4.47442859e-01 -6.38742447e-01
-2.88075954e-01 2.69100338e-01 2.15287760e-01 3.82916689e-01
6.58032477e-01 -1.22699523e+00 -1.07436967e+00 1.64206430e-01
4.36175019e-01 1.35266915e-01 -1.66991167e-02 3.43695641e-01
-3.74487162e-01 5.26910365e-01 2.38333553e-01 -8.25162679e-02
-1.00910461e+00 3.03513974e-01 -1.42723918e-01 -9.32941914e-01
-4.09537852e-02 9.00424659e-01 4.41877186e-01 -8.33909273e-01
-1.41238138e-01 -4.98535365e-01 -5.49917281e-01 2.19884664e-01
6.44751072e-01 -1.40912339e-01 8.69398788e-02 -7.68871188e-01
-1.92582816e-01 1.73707426e-01 -1.62279904e-01 -5.07150114e-01
1.16969383e+00 2.90035624e-02 -2.31419086e-01 4.11954373e-01
8.79543006e-01 3.95869195e-01 -7.73916423e-01 1.43242195e-01
-3.80989611e-01 8.07085820e-03 -2.41085157e-01 -1.17736816e+00
-8.36738467e-01 6.73747480e-01 -3.63845348e-01 1.73294708e-01
1.10738039e+00 7.25473017e-02 7.93559909e-01 2.15855911e-01
-1.40540645e-01 -1.22638714e+00 1.41156271e-01 8.37996781e-01
1.12028956e+00 -1.25105429e+00 7.33138323e-02 -3.79478663e-01
-1.35811543e+00 7.19638646e-01 7.06176877e-01 9.07952711e-03
6.57501876e-01 1.28185779e-01 4.69022602e-01 -2.22039118e-01
-1.13323867e+00 -1.32945821e-01 4.05357569e-01 7.24101722e-01
7.72168458e-01 2.68706799e-01 -4.53296542e-01 9.43553627e-01
-7.84514189e-01 -5.54755211e-01 5.32836139e-01 6.17619395e-01
-1.94836363e-01 -9.45621848e-01 -1.27006382e-01 5.34261167e-01
-5.88517725e-01 -6.31800950e-01 -7.54540265e-01 4.43454623e-01
-2.96343952e-01 1.35322380e+00 -9.48373079e-02 -1.07482739e-01
5.06462395e-01 2.65532639e-02 -9.43265706e-02 -1.05650389e+00
-8.76176000e-01 -1.35617048e-01 7.15676367e-01 -1.40555725e-01
-4.71928596e-01 -6.41444504e-01 -1.14041317e+00 1.76108003e-01
-2.00329468e-01 6.31991088e-01 8.47989678e-01 1.04758227e+00
4.05437320e-01 5.78100979e-01 7.72807300e-01 -4.58899647e-01
-3.84065539e-01 -1.33502102e+00 -3.26450765e-01 6.35751724e-01
9.93933901e-02 1.27674239e-02 -3.45544517e-01 2.49760404e-01] | [11.520869255065918, 6.823875904083252] |
de094260-23b2-4e48-b24c-2e2c5d340a59 | topic-guided-abstractive-multi-document | 2110.11207 | null | https://arxiv.org/abs/2110.11207v1 | https://arxiv.org/pdf/2110.11207v1.pdf | Topic-Guided Abstractive Multi-Document Summarization | A critical point of multi-document summarization (MDS) is to learn the relations among various documents. In this paper, we propose a novel abstractive MDS model, in which we represent multiple documents as a heterogeneous graph, taking semantic nodes of different granularities into account, and then apply a graph-to-sequence framework to generate summaries. Moreover, we employ a neural topic model to jointly discover latent topics that can act as cross-document semantic units to bridge different documents and provide global information to guide the summary generation. Since topic extraction can be viewed as a special type of summarization that "summarizes" texts into a more abstract format, i.e., a topic distribution, we adopt a multi-task learning strategy to jointly train the topic and summarization module, allowing the promotion of each other. Experimental results on the Multi-News dataset demonstrate that our model outperforms previous state-of-the-art MDS models on both Rouge metrics and human evaluation, meanwhile learns high-quality topics. | ['Le Hu', 'Peng Cui'] | 2021-10-21 | null | https://aclanthology.org/2021.findings-emnlp.126 | https://aclanthology.org/2021.findings-emnlp.126.pdf | findings-emnlp-2021-11 | ['graph-to-sequence'] | ['natural-language-processing'] | [ 1.26343086e-01 3.97791386e-01 -4.22766805e-01 -3.20686907e-01
-1.11274624e+00 -3.13937664e-01 7.79083550e-01 3.61262798e-01
1.01342224e-01 7.66832769e-01 1.14657712e+00 1.28324345e-01
-7.66268373e-02 -8.48741114e-01 -7.27665544e-01 -6.49742782e-01
2.44898811e-01 7.17181385e-01 3.02416205e-01 -2.07116142e-01
5.13151526e-01 -1.95068508e-01 -1.09863472e+00 5.02288520e-01
1.40698612e+00 3.90393525e-01 4.69512612e-01 4.05959785e-01
-5.64984381e-01 6.04833007e-01 -1.16631949e+00 -4.58544850e-01
-5.91107428e-01 -5.95837176e-01 -9.56075311e-01 4.09348220e-01
3.00518990e-01 -3.90317917e-01 -3.34837884e-01 9.95039999e-01
6.16891980e-01 3.13201129e-01 9.85435963e-01 -1.06248820e+00
-9.79201972e-01 1.13716936e+00 -7.81134069e-01 -6.25899658e-02
4.98476803e-01 -2.83926934e-01 1.38407958e+00 -5.68577826e-01
7.18723655e-01 1.58553350e+00 1.99594840e-01 4.45376068e-01
-1.12925720e+00 -3.89011681e-01 5.21049857e-01 2.61625070e-02
-9.58547652e-01 -3.37197483e-01 1.05023968e+00 -1.56383052e-01
7.91626573e-01 1.82200193e-01 5.63985288e-01 1.28105521e+00
3.73748541e-01 1.44593906e+00 2.72714078e-01 -1.18880188e-02
2.63298422e-01 -1.21541649e-01 3.17327380e-01 5.47453463e-01
5.47295809e-01 -9.52076018e-01 -7.19641149e-01 -2.31427848e-01
3.48615646e-01 2.64632851e-02 -2.66229808e-01 -2.91548193e-01
-1.46995234e+00 9.37864959e-01 3.97173762e-01 2.18369544e-01
-5.91768503e-01 5.52758388e-02 7.75221109e-01 -1.63351782e-02
9.30251837e-01 4.60337549e-01 4.65781912e-02 3.14003229e-01
-1.14061737e+00 4.84056652e-01 8.08457494e-01 1.05288219e+00
6.46432936e-01 -1.07076885e-02 -7.57391214e-01 9.36198413e-01
4.75908607e-01 3.42951417e-01 7.75816798e-01 -6.91569448e-01
9.30788696e-01 8.75046968e-01 -7.81944208e-03 -1.24931037e+00
-3.27131510e-01 -5.07091582e-01 -1.20233059e+00 -8.09678793e-01
-5.52059233e-01 -1.82584226e-01 -6.83498740e-01 1.51096332e+00
1.83828935e-01 2.84584612e-01 3.81486863e-01 6.32420838e-01
1.39629459e+00 1.31174815e+00 3.99945118e-02 -3.79153818e-01
1.29677105e+00 -1.24526107e+00 -8.00685287e-01 -2.53635824e-01
4.88163292e-01 -4.37631011e-01 9.96985137e-01 2.51183331e-01
-1.17059076e+00 -4.50899690e-01 -1.23279154e+00 -3.20002466e-01
-6.57295957e-02 2.91363180e-01 4.36925977e-01 -5.34872264e-02
-1.06023622e+00 5.32631159e-01 -8.32858086e-01 -4.13010031e-01
4.25587118e-01 -8.28794017e-02 1.41244223e-02 2.66575534e-02
-1.18100488e+00 5.53309500e-01 9.95220721e-01 -2.61194706e-01
-9.89156187e-01 -6.00032747e-01 -9.38761294e-01 4.11835492e-01
4.59454566e-01 -1.40087926e+00 1.23666823e+00 -2.26036385e-01
-1.35821140e+00 4.95811611e-01 -4.27564234e-01 -3.26539159e-01
2.07580864e-01 -3.28658134e-01 -3.31865340e-01 4.27474380e-01
6.51669204e-01 6.67007625e-01 7.33041048e-01 -1.54113173e+00
-7.59839535e-01 -2.39628211e-01 -3.81476171e-02 6.09109342e-01
-5.71130812e-01 -1.54498190e-01 -6.27077639e-01 -8.74195039e-01
1.42740756e-01 -4.89812970e-01 -2.29576454e-01 -8.80845189e-01
-1.19116724e+00 -7.50388145e-01 7.56165028e-01 -8.67129087e-01
1.65249932e+00 -1.68002617e+00 8.27431083e-01 -1.89870089e-01
3.71667594e-01 -1.20441057e-01 -1.35789812e-01 8.28927577e-01
2.83296019e-01 1.45277008e-01 -3.44099611e-01 -8.12004089e-01
2.03464583e-01 -1.11780688e-01 -7.34505296e-01 -5.13412170e-02
1.41377421e-03 9.68689561e-01 -1.10267162e+00 -7.81591654e-01
-2.35332564e-01 1.83284715e-01 -2.75245726e-01 2.73802012e-01
-6.82340860e-01 2.90227771e-01 -1.01439714e+00 2.00626716e-01
4.08639878e-01 -5.59091747e-01 2.74449848e-02 -2.12704077e-01
1.04309894e-01 5.61922193e-01 -8.29549313e-01 2.36531830e+00
-2.97719866e-01 2.57171243e-01 -4.42100137e-01 -1.10272264e+00
1.03220892e+00 3.23241502e-01 3.47756267e-01 -2.47088000e-01
-1.44191295e-01 -1.62137225e-02 -6.93213046e-01 -4.30173159e-01
1.18634200e+00 1.24954320e-01 -5.39277554e-01 9.01078045e-01
1.84598252e-01 -2.71942258e-01 5.35239220e-01 8.50441456e-01
7.30914116e-01 -5.65597042e-02 3.46225351e-01 -3.05551231e-01
4.41981673e-01 -2.72352118e-02 4.29444373e-01 6.78601742e-01
5.88062465e-01 6.28575027e-01 9.83101785e-01 7.03665391e-02
-9.73629057e-01 -1.09518778e+00 4.03266013e-01 1.00703335e+00
3.67465675e-01 -7.34728634e-01 -7.28442907e-01 -8.77876580e-01
-1.60377547e-01 1.20702589e+00 -3.21750045e-01 -4.98952895e-01
-4.91087288e-01 -8.82031739e-01 2.90817887e-01 4.84759569e-01
5.12868762e-01 -1.10015523e+00 -6.72437623e-02 4.20415848e-01
-6.65396810e-01 -7.39297628e-01 -7.69433200e-01 -3.90256584e-01
-1.00126064e+00 -6.61214411e-01 -9.81643498e-01 -8.91862392e-01
5.21292865e-01 6.10642970e-01 1.10708690e+00 -1.77827418e-01
3.18230510e-01 3.64108771e-01 -3.52488935e-01 -3.21823865e-01
-5.20011246e-01 7.67328799e-01 -2.45817795e-01 -2.91002560e-02
4.35870476e-02 -6.67619884e-01 -6.16972923e-01 -1.85873553e-01
-1.29386675e+00 4.02576625e-01 6.46862268e-01 7.56470025e-01
4.77766663e-01 1.47199824e-01 1.26508558e+00 -9.96142209e-01
1.39151585e+00 -7.56109059e-01 4.19309475e-02 5.79440057e-01
-4.29320246e-01 1.65835172e-01 6.50855720e-01 -8.71082395e-02
-1.48320222e+00 -5.35975873e-01 1.44048542e-01 -3.48784439e-02
1.64223284e-01 9.19511974e-01 -4.85326737e-01 9.01976645e-01
2.50846863e-01 8.02984476e-01 -3.13517421e-01 -4.51967508e-01
6.50306702e-01 7.36785293e-01 5.67994416e-01 -5.77498913e-01
4.62841868e-01 5.12666643e-01 -3.16110283e-01 -8.58426571e-01
-1.24570858e+00 -5.69688857e-01 -5.28404236e-01 2.68122330e-02
9.07104790e-01 -1.09912920e+00 -2.50769794e-01 3.83818090e-01
-1.67875099e+00 1.67970955e-01 -1.26024842e-01 2.82058477e-01
-6.22027278e-01 6.62632048e-01 -6.08102739e-01 -2.91335940e-01
-9.05223489e-01 -8.09769809e-01 1.63038397e+00 5.01456022e-01
-2.97710747e-01 -1.26912808e+00 1.56431168e-01 2.12691545e-01
4.08627139e-03 9.32677239e-02 1.11180234e+00 -1.00432575e+00
-5.30647218e-01 4.65852618e-02 -2.52785027e-01 1.32837981e-01
3.06164801e-01 -6.08435273e-02 -4.79209781e-01 -3.87041271e-01
1.21011585e-02 -2.90877521e-01 1.43184614e+00 6.04265630e-01
1.00700033e+00 -7.67882824e-01 -6.98316634e-01 2.90751666e-01
1.07295930e+00 -1.16789877e-01 5.23301125e-01 3.43341738e-01
9.38452601e-01 6.65300727e-01 5.01464069e-01 6.31499290e-01
1.07661271e+00 2.00069129e-01 4.74028103e-02 1.08555414e-01
-1.31379619e-01 -7.08620846e-01 3.40882480e-01 1.43109310e+00
3.83615792e-01 -7.92688012e-01 -5.86132824e-01 5.43396831e-01
-2.14885783e+00 -1.02287281e+00 9.22073722e-02 1.66612089e+00
9.15527940e-01 1.31175131e-01 1.38041213e-01 -4.04170454e-01
1.03630173e+00 8.28555107e-01 -5.52827418e-01 1.26725081e-02
-2.11162940e-01 -5.53966880e-01 -2.52661616e-01 2.05642253e-01
-1.07654762e+00 1.13812172e+00 5.44761276e+00 9.68884468e-01
-8.56965601e-01 -8.87671933e-02 5.21401227e-01 -8.02532770e-03
-8.56800675e-01 -6.17847629e-02 -8.94225717e-01 6.68361306e-01
4.93832648e-01 -8.99903178e-01 -1.42222971e-01 8.20624352e-01
2.12014511e-01 1.56885207e-01 -8.82335424e-01 5.72115898e-01
6.16077065e-01 -1.65000260e+00 9.01733398e-01 -2.38352060e-01
1.08188820e+00 -3.00767034e-01 -1.18177302e-01 3.55271131e-01
5.44514358e-01 -5.36137998e-01 5.13668835e-01 7.08787501e-01
3.22705954e-01 -8.04414511e-01 5.78322113e-01 4.63863075e-01
-1.18549252e+00 2.69395888e-01 -6.99546635e-01 5.65437913e-01
5.67785084e-01 5.98922491e-01 -7.80576587e-01 1.17924571e+00
2.64656097e-01 1.28006339e+00 -4.51513976e-01 9.94292915e-01
-4.73658085e-01 3.68301302e-01 2.10292175e-01 -3.13085556e-01
3.18241954e-01 -3.31021726e-01 8.40505540e-01 1.34355497e+00
5.99489272e-01 -5.40937111e-02 4.19469118e-01 8.14888060e-01
-3.52378547e-01 2.08162904e-01 -4.31733817e-01 -9.45028067e-02
5.98203063e-01 1.28291011e+00 -7.26358891e-01 -7.86515474e-01
-1.91315487e-01 1.11675012e+00 2.84609377e-01 5.55071235e-01
-6.27850533e-01 -6.16100132e-01 1.35003984e-01 -3.18257570e-01
3.21848333e-01 -5.82510829e-02 -2.17227846e-01 -1.57239771e+00
1.09490864e-01 -6.11158788e-01 6.40144825e-01 -9.37640190e-01
-1.21005177e+00 5.96125305e-01 1.95883170e-01 -1.21377861e+00
-3.45616221e-01 2.63650388e-01 -1.21411026e+00 6.39217973e-01
-1.40644360e+00 -1.48578942e+00 -7.89786577e-02 1.89254344e-01
9.42946494e-01 -3.06139261e-01 5.08107424e-01 -2.16841310e-01
-7.27537334e-01 1.29214928e-01 3.81690979e-01 -2.26856351e-01
7.04415560e-01 -1.36761594e+00 6.35100424e-01 9.17754829e-01
-1.81427542e-02 7.02419221e-01 7.02392876e-01 -1.12510157e+00
-1.07092953e+00 -1.37705421e+00 9.08780932e-01 -3.68357375e-02
6.77949190e-01 -1.74017027e-01 -1.09784126e+00 8.03918004e-01
5.91592312e-01 -1.12423754e+00 7.58709192e-01 2.03020662e-01
-1.10667326e-01 -4.78804335e-02 -4.79626834e-01 7.45600820e-01
9.50845063e-01 -3.03254783e-01 -1.21949708e+00 5.45931935e-01
1.44815826e+00 -2.63002038e-01 -6.78216398e-01 1.39643282e-01
2.86504291e-02 -7.13965893e-01 8.71498048e-01 -7.43845820e-01
9.45324183e-01 -1.60460934e-01 2.20545784e-01 -2.05786681e+00
-2.71308482e-01 -6.54779196e-01 -4.06572163e-01 1.83553863e+00
3.23476970e-01 -4.33419257e-01 6.12673640e-01 -6.45106286e-02
-6.99270368e-01 -5.85044384e-01 -5.48655033e-01 -4.80410039e-01
-3.97582240e-02 1.20622270e-01 9.17757273e-01 7.34064043e-01
2.90908277e-01 1.05226469e+00 -3.56232733e-01 8.11994225e-02
6.13231540e-01 5.71736872e-01 7.98846304e-01 -1.28234267e+00
4.64639775e-02 -6.88491046e-01 7.61552155e-02 -1.48584223e+00
6.29924774e-01 -1.13492858e+00 -2.67990120e-02 -2.43502879e+00
5.77261806e-01 9.17491689e-02 3.47275659e-02 4.94448356e-02
-6.03588164e-01 -6.62268996e-01 -9.82874855e-02 4.66885298e-01
-1.23393810e+00 1.26112854e+00 1.39806092e+00 -4.12962288e-01
-2.70851940e-01 1.04943067e-01 -1.36981869e+00 5.55068135e-01
5.99689245e-01 -5.08126795e-01 -6.88488066e-01 -7.20422268e-01
2.03777835e-01 3.46114665e-01 -4.73217620e-03 -6.56898856e-01
6.03072166e-01 -5.60285635e-02 2.03325033e-01 -1.17140698e+00
1.81116298e-01 -4.51167673e-02 -2.99354672e-01 2.86602467e-01
-7.39332974e-01 -9.13814157e-02 -1.69216245e-01 8.89685154e-01
-3.69944304e-01 -2.24907801e-01 1.62510961e-01 -1.40906855e-01
-6.07312024e-01 4.97746706e-01 -4.54361252e-02 2.10802510e-01
7.22201347e-01 2.05900922e-01 -7.70488977e-01 -4.97250795e-01
-3.00700098e-01 7.46461332e-01 3.25090885e-01 6.31454945e-01
7.26394594e-01 -1.49965620e+00 -1.16684425e+00 -2.34079137e-01
1.72644913e-01 5.61446369e-01 6.05935276e-01 3.39390963e-01
-2.40514681e-01 6.32850647e-01 1.08495377e-01 -6.23628497e-01
-8.21465373e-01 5.36526799e-01 -2.76254833e-01 -5.71787119e-01
-8.30057383e-01 6.11597776e-01 4.10416842e-01 -1.90747991e-01
1.88823253e-01 -1.57801032e-01 -6.26176894e-01 5.99470317e-01
6.99583769e-01 5.32622278e-01 -1.96600273e-01 -2.95685709e-01
3.80063467e-02 3.32152724e-01 -5.07634044e-01 -4.46872003e-02
1.42444539e+00 -5.19128501e-01 -5.33522487e-01 4.93195117e-01
1.24253750e+00 -1.56596065e-01 -1.00537395e+00 -4.64587003e-01
3.01578254e-01 6.02013357e-02 -1.74630105e-01 -3.60104024e-01
-6.83565557e-01 7.57839441e-01 -4.79650974e-01 5.60065567e-01
9.50698912e-01 4.97852743e-01 1.10248744e+00 5.36149323e-01
-1.98934257e-01 -1.10487449e+00 5.69636703e-01 3.72450948e-01
1.21921575e+00 -9.94767964e-01 2.21020356e-01 -2.85023272e-01
-1.07178271e+00 1.06922615e+00 6.94808483e-01 -1.35107949e-01
1.70103163e-01 -3.65003049e-01 -3.68226439e-01 -3.17634761e-01
-8.87294114e-01 1.09163605e-01 5.46909809e-01 4.43476796e-01
3.11291963e-01 -9.87581210e-04 -3.59434664e-01 9.87312973e-01
-3.29274267e-01 -2.32671216e-01 6.38591409e-01 4.99540657e-01
-8.94373715e-01 -7.12744772e-01 3.13857733e-03 6.58572316e-01
-2.40636259e-01 -1.08842269e-01 -4.06457782e-01 3.40666533e-01
-6.51248455e-01 9.39291120e-01 7.32073709e-02 -2.73623943e-01
2.73858786e-01 -4.80004475e-02 -6.50599077e-02 -1.10707843e+00
-2.79662579e-01 2.67644972e-01 1.24825485e-01 -1.34981871e-01
-3.72761488e-01 -7.08602428e-01 -1.28890944e+00 -2.61452079e-01
-1.16630033e-01 4.99098539e-01 5.02719045e-01 1.04437065e+00
5.65300941e-01 1.06999385e+00 7.82291353e-01 -7.89112031e-01
-6.86886787e-01 -1.19078255e+00 -6.44323289e-01 2.31597185e-01
2.36856595e-01 -2.57034868e-01 -1.58918872e-01 7.79638141e-02] | [12.585346221923828, 9.493476867675781] |
2e6d4eef-953f-443d-a8bb-a40581518f29 | decoupling-makes-weakly-supervised-local | 2201.02861 | null | https://arxiv.org/abs/2201.02861v2 | https://arxiv.org/pdf/2201.02861v2.pdf | Decoupling Makes Weakly Supervised Local Feature Better | Weakly supervised learning can help local feature methods to overcome the obstacle of acquiring a large-scale dataset with densely labeled correspondences. However, since weak supervision cannot distinguish the losses caused by the detection and description steps, directly conducting weakly supervised learning within a joint describe-then-detect pipeline suffers limited performance. In this paper, we propose a decoupled describe-then-detect pipeline tailored for weakly supervised local feature learning. Within our pipeline, the detection step is decoupled from the description step and postponed until discriminative and robust descriptors are learned. In addition, we introduce a line-to-window search strategy to explicitly use the camera pose information for better descriptor learning. Extensive experiments show that our method, namely PoSFeat (Camera Pose Supervised Feature), outperforms previous fully and weakly supervised methods and achieves state-of-the-art performance on a wide range of downstream tasks. | ['Longguang Wang', 'Yulan Guo', 'Kai Xu', 'Qing Ran', 'Li Liu', 'Kunhong Li'] | 2022-01-08 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Li_Decoupling_Makes_Weakly_Supervised_Local_Feature_Better_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Li_Decoupling_Makes_Weakly_Supervised_Local_Feature_Better_CVPR_2022_paper.pdf | cvpr-2022-1 | ['camera-localization', 'image-matching'] | ['computer-vision', 'computer-vision'] | [-1.25039602e-02 -1.41424417e-01 -5.26064634e-01 -5.90586603e-01
-1.55247319e+00 -8.81572723e-01 8.14584374e-01 1.76880166e-01
-5.62035978e-01 3.46114159e-01 1.82175010e-01 2.72741139e-01
-2.42281309e-03 -2.91356653e-01 -7.75205255e-01 -7.05315948e-01
2.24299893e-01 5.20239472e-01 4.26941216e-01 1.95911184e-01
2.95568258e-01 6.96123302e-01 -1.54589474e+00 2.69227028e-01
1.93033561e-01 7.99617529e-01 4.76731986e-01 3.95472139e-01
1.68713957e-01 6.28960848e-01 2.97756866e-02 -2.78787971e-01
4.75846171e-01 -3.37636590e-01 -7.74687409e-01 1.74550191e-01
9.82387006e-01 -6.09146535e-01 -2.45522648e-01 9.70541179e-01
5.51405966e-01 5.89733832e-02 4.04667675e-01 -1.04546189e+00
-1.44271255e-01 2.29264438e-01 -7.10335314e-01 4.68966402e-02
5.28240919e-01 1.86085314e-01 1.37359619e+00 -1.49367285e+00
7.82628655e-01 1.03522539e+00 9.50783372e-01 2.76685208e-01
-1.42333019e+00 -6.43087447e-01 3.35310549e-01 -1.19620152e-02
-1.70769048e+00 -6.01985514e-01 6.88057482e-01 -3.73958617e-01
1.13917720e+00 -1.19035527e-01 3.97250861e-01 9.86896694e-01
-3.82491499e-01 1.12705922e+00 7.71768570e-01 -2.00888857e-01
-7.94784799e-02 -1.32482164e-04 1.74990267e-01 1.16791642e+00
-2.11827811e-02 3.65025997e-01 -7.51852393e-01 -1.15574360e-01
6.95879221e-01 1.49262041e-01 9.73952264e-02 -8.79568696e-01
-1.39660275e+00 7.53875732e-01 5.11799037e-01 4.26451936e-02
-1.91937745e-01 3.42374146e-01 3.41947556e-01 2.77378470e-01
3.90176475e-01 4.74927902e-01 -6.89533532e-01 -2.91727275e-01
-1.23334801e+00 1.52673617e-01 5.11852860e-01 1.27203941e+00
1.24777651e+00 -5.26744068e-01 -1.59388244e-01 6.98698759e-01
1.62084281e-01 4.84918565e-01 -1.12145096e-01 -9.20498431e-01
4.31317776e-01 4.44155335e-01 -2.90729180e-02 -6.31351054e-01
-5.21840155e-01 -4.31239635e-01 -3.77597451e-01 -4.94675189e-02
6.14299536e-01 9.26958248e-02 -7.77016222e-01 1.28055537e+00
3.55392009e-01 2.20844045e-01 -3.36629510e-01 1.03212881e+00
7.40370154e-01 2.87972152e-01 -4.58956584e-02 4.77706343e-02
1.11263263e+00 -1.38262880e+00 -2.45631874e-01 -3.23764294e-01
8.83310556e-01 -1.07976556e+00 8.17422509e-01 2.41729930e-01
-9.73002017e-01 -7.07814276e-01 -8.10590982e-01 -5.29643536e-01
-1.42663896e-01 5.71229160e-01 7.05046237e-01 -7.07660466e-02
-9.01923120e-01 7.00205982e-01 -1.19853866e+00 -7.07997561e-01
5.73751390e-01 4.55648512e-01 -9.92007852e-01 -4.57525551e-01
-4.08939153e-01 8.16391349e-01 2.41726533e-01 -3.74955833e-02
-1.06983805e+00 -6.53917134e-01 -1.09854305e+00 -7.70259872e-02
4.62714881e-01 -6.38179421e-01 1.33892477e+00 -5.49771726e-01
-1.26616907e+00 1.15370178e+00 -3.69932085e-01 -1.47020251e-01
6.42048478e-01 -5.51589906e-01 3.62207383e-01 2.13048354e-01
2.84533471e-01 8.23619664e-01 8.68438482e-01 -1.16434169e+00
-8.87040734e-01 -2.91402608e-01 -2.69597222e-04 1.29155859e-01
-1.53516158e-01 1.90283626e-01 -1.01567233e+00 -4.41636056e-01
2.73728997e-01 -1.08457935e+00 -1.71890363e-01 5.28963029e-01
-2.58966178e-01 -3.20375144e-01 8.68410468e-01 -3.37763667e-01
8.14846992e-01 -2.34651399e+00 1.35281146e-01 1.90497309e-01
1.35311306e-01 1.28132731e-01 -4.58698183e-01 5.29498160e-01
-1.13229798e-02 -3.76774997e-01 3.29135880e-02 -9.21393394e-01
1.98724009e-02 2.53751397e-01 -6.12493642e-02 9.27271783e-01
5.85277021e-01 9.19046640e-01 -1.15386391e+00 -6.45217657e-01
3.82547587e-01 2.63846755e-01 -6.63335741e-01 3.53211224e-01
1.15871795e-01 3.38496536e-01 -4.83974963e-01 9.68154013e-01
7.27703810e-01 -2.41023049e-01 -1.38022423e-01 -4.22905833e-01
-2.00157866e-01 2.36939996e-01 -9.77920175e-01 2.24200130e+00
-5.85924387e-01 5.96750021e-01 1.28524706e-01 -1.07919300e+00
7.87220180e-01 -9.89281014e-02 4.69396591e-01 -4.99888420e-01
-1.15339398e-01 3.31482023e-01 -3.39133739e-01 -3.18870425e-01
2.76329011e-01 -1.16621301e-01 -1.97189659e-01 2.22609267e-01
5.53885341e-01 -1.89579293e-01 1.38751686e-01 2.33931601e-01
1.31874263e+00 7.11128950e-01 3.84558499e-01 -1.62056386e-01
4.37269717e-01 2.55274504e-01 5.77901244e-01 7.45113194e-01
-4.84862238e-01 1.04867828e+00 2.77548283e-01 -3.32354426e-01
-1.00982296e+00 -1.18509662e+00 -3.08041126e-01 1.39319146e+00
3.45924228e-01 -6.76788390e-01 -4.08769339e-01 -1.17412066e+00
3.48187715e-01 1.65914208e-01 -3.90623569e-01 -2.24055037e-01
-5.42391360e-01 -3.06113213e-01 4.32332546e-01 9.12241161e-01
3.60770583e-01 -6.58792377e-01 -1.19397275e-01 4.30711312e-03
1.29892528e-01 -1.34119785e+00 -6.62716448e-01 5.11448860e-01
-7.31458247e-01 -1.09350443e+00 -6.18183553e-01 -1.04713488e+00
9.15011942e-01 5.63714921e-01 1.00855553e+00 1.54104456e-01
-3.16503465e-01 4.34423178e-01 -3.82859826e-01 2.12053210e-01
9.38612148e-02 1.86447620e-01 -3.93726490e-03 -2.57992864e-01
5.48539460e-01 -1.89250767e-01 -6.21516466e-01 4.42411155e-01
-3.90764147e-01 -1.35622367e-01 1.07520151e+00 1.19634807e+00
8.60767365e-01 -5.15509069e-01 7.50687495e-02 -6.18514776e-01
-3.87289710e-02 -1.83155522e-01 -6.87286437e-01 1.67674869e-01
-5.84048629e-01 3.91960829e-01 4.66689229e-01 -2.27083936e-01
-1.00523055e+00 1.01756442e+00 -1.26140118e-01 -5.99502563e-01
-3.08445185e-01 3.70008945e-01 -1.89819902e-01 -3.24785203e-01
4.08031464e-01 2.74675816e-01 -1.51752174e-01 -7.56546021e-01
3.72850597e-01 3.99072915e-01 6.31941378e-01 -5.84531605e-01
1.30712748e+00 6.58992350e-01 1.50156477e-02 -5.65271199e-01
-1.24583399e+00 -1.21214843e+00 -1.26255751e+00 1.23731205e-02
6.64884627e-01 -1.38154590e+00 -3.91706377e-01 3.21583569e-01
-1.13024521e+00 -1.60559371e-01 -3.19296718e-01 6.41435385e-01
-6.94254458e-01 3.06308150e-01 -6.26525998e-01 -3.74555409e-01
3.78189795e-02 -1.18107581e+00 1.90808213e+00 -6.15642928e-02
-1.64085299e-01 -8.03532183e-01 1.50662839e-01 4.08118725e-01
-5.14882468e-02 -2.56957635e-02 7.86045343e-02 -7.53170550e-01
-8.52478206e-01 -6.90802634e-01 -5.69342434e-01 3.42858195e-01
-5.68096004e-02 -9.58888978e-02 -1.09974813e+00 -5.63728929e-01
-7.26099551e-01 -7.85577774e-01 1.16305435e+00 -5.50768971e-02
9.61535215e-01 1.69355854e-01 -5.12369573e-01 1.03488576e+00
1.33264172e+00 -6.75711691e-01 2.36375555e-01 3.93091887e-01
1.01188815e+00 5.17607510e-01 1.09620881e+00 3.89054030e-01
3.66463482e-01 8.24029744e-01 2.25793585e-01 -2.80805230e-01
-4.15382117e-01 -6.23735368e-01 6.56346381e-01 5.57900369e-01
3.86172086e-02 2.55839527e-01 -7.88067222e-01 6.07758403e-01
-1.98687184e+00 -8.54953945e-01 -2.19211895e-02 2.10319638e+00
7.31585562e-01 1.64163336e-01 1.20312937e-01 -2.64347762e-01
5.42677641e-01 2.28917167e-01 -4.30934459e-01 9.57968757e-02
-8.36527869e-02 4.50150296e-02 7.46282220e-01 6.02404535e-01
-1.40448236e+00 1.45082319e+00 5.91334391e+00 9.05551553e-01
-8.48108828e-01 3.49377804e-02 1.46964654e-01 -2.45197952e-01
2.31264308e-02 3.71101826e-01 -1.16299820e+00 2.50341129e-02
5.44396222e-01 2.78814912e-01 2.22036108e-01 1.07467222e+00
9.54031870e-02 -1.30962029e-01 -1.71667683e+00 1.16385508e+00
9.78616178e-02 -1.10639822e+00 -8.93848613e-02 1.90897398e-02
6.73495531e-01 4.73760694e-01 -5.30339256e-02 3.07389379e-01
3.14195603e-01 -6.46499038e-01 6.62020504e-01 4.02560741e-01
8.44901264e-01 -6.94268465e-01 7.60433018e-01 4.32085395e-01
-1.44461584e+00 6.85816929e-02 -3.95587921e-01 -2.35480592e-02
6.22340590e-02 6.63687527e-01 -8.24680924e-01 3.50797743e-01
6.91199064e-01 1.30349553e+00 -8.03837538e-01 1.02769613e+00
-4.01715547e-01 3.62193942e-01 -5.19968629e-01 2.04558983e-01
4.80093598e-01 1.79849356e-01 3.42206448e-01 1.51172626e+00
2.18596682e-01 -2.69135028e-01 6.99089289e-01 6.38774157e-01
-1.26911074e-01 -1.39870748e-01 -6.30070448e-01 1.78921714e-01
4.95418280e-01 1.46780360e+00 -7.32408345e-01 -1.27735481e-01
-7.18913853e-01 1.31192553e+00 6.65203810e-01 1.14685945e-01
-5.40548563e-01 -2.12576285e-01 6.29691660e-01 2.15876312e-03
6.65850580e-01 -3.34161758e-01 1.28986621e-02 -1.42397392e+00
1.74122706e-01 -5.15241683e-01 4.54802632e-01 -5.62506735e-01
-1.45064878e+00 1.14402384e-01 6.58239378e-03 -1.42864656e+00
-3.72812450e-01 -6.29094481e-01 -4.17324930e-01 6.63292110e-01
-1.90453184e+00 -1.74354696e+00 -4.18222040e-01 7.86659002e-01
6.93978488e-01 1.64212193e-02 6.94848001e-01 2.87495792e-01
-4.72305804e-01 8.62684309e-01 1.23364516e-01 3.94219965e-01
1.18550885e+00 -1.33673632e+00 2.95590013e-01 8.43117356e-01
4.01962459e-01 6.52335703e-01 4.18924809e-01 -4.09079522e-01
-1.52342713e+00 -1.01851344e+00 8.68937552e-01 -7.04219103e-01
8.06086957e-01 -6.59712434e-01 -6.03573084e-01 6.63848639e-01
-2.14678273e-01 8.98359835e-01 4.91423965e-01 3.07988048e-01
-7.22242713e-01 -1.94364324e-01 -9.47376788e-01 5.73551431e-02
1.33978283e+00 -9.89268541e-01 -5.00198662e-01 4.65231597e-01
4.39503580e-01 -3.84018391e-01 -7.53516734e-01 2.20409304e-01
5.90824544e-01 -7.71293402e-01 1.10804820e+00 -3.79848272e-01
2.92079568e-01 -4.32327777e-01 -1.27851352e-01 -1.05805540e+00
-4.76857275e-01 -6.20325506e-01 -9.93668139e-02 1.43273890e+00
3.29766929e-01 -1.43890843e-01 9.42266405e-01 1.93506584e-01
-1.31096095e-01 -5.26601076e-01 -8.07735801e-01 -1.03214288e+00
-1.00664183e-01 -2.60372996e-01 3.81810553e-02 8.30460608e-01
8.43089968e-02 4.22687531e-01 -2.93221503e-01 2.12044016e-01
8.61696839e-01 3.68467838e-01 1.01870584e+00 -1.09915793e+00
-3.35775405e-01 -2.13904768e-01 -6.42794549e-01 -1.49607706e+00
3.21477920e-01 -9.91100252e-01 4.53459233e-01 -1.30854809e+00
7.23516047e-01 -3.06374818e-01 -2.62541473e-01 7.85612643e-01
-2.02948526e-01 4.23761964e-01 1.82586625e-01 5.04204154e-01
-1.13497889e+00 4.59310442e-01 1.04231083e+00 -3.07956059e-02
1.09032154e-01 -1.40326530e-01 -4.32733148e-01 8.63017678e-01
3.92109394e-01 -5.90181172e-01 -9.20228958e-02 -4.10878450e-01
-7.95344412e-02 -2.29306757e-01 5.28762341e-01 -7.77567267e-01
4.72026557e-01 1.31974921e-01 4.40443486e-01 -7.93141782e-01
2.68180996e-01 -9.53372955e-01 -6.27755880e-01 3.08011949e-01
-3.98072630e-01 6.77083656e-02 -5.54866297e-03 7.21327066e-01
-4.47936565e-01 -7.16623738e-02 7.68020868e-01 -7.26208463e-02
-1.03189194e+00 4.56643373e-01 9.66160744e-02 -5.96161583e-04
1.07088697e+00 -7.63917491e-02 -1.18014425e-01 -2.48569265e-01
-7.26492226e-01 3.76421750e-01 6.46313608e-01 4.39999402e-01
5.64336956e-01 -1.29637635e+00 -6.57719374e-01 1.89547956e-01
7.71476567e-01 2.88313508e-01 -2.01354831e-01 1.00295341e+00
-4.27940875e-01 3.80833626e-01 1.58346280e-01 -8.77014339e-01
-1.39275920e+00 5.83468437e-01 1.18882395e-01 -3.33988458e-01
-6.65429056e-01 1.08912349e+00 2.19291925e-01 -6.80832148e-01
4.37188745e-01 -2.32424110e-01 2.24438712e-01 1.61542237e-01
3.97041142e-01 2.33792201e-01 1.30552500e-01 -8.60211134e-01
-6.50144815e-01 8.93719852e-01 -3.91826540e-01 1.71556845e-01
1.54227686e+00 -2.63388425e-01 -3.27590071e-02 1.54742554e-01
1.82753754e+00 -1.22626284e-02 -1.80488384e+00 -4.89268303e-01
3.38153213e-01 -5.61424136e-01 3.70840728e-01 -5.71673155e-01
-9.97316599e-01 8.01013589e-01 5.43275058e-01 -5.24419010e-01
8.46692264e-01 5.34829259e-01 8.68680418e-01 8.86532426e-01
4.75971669e-01 -1.20505178e+00 2.28009880e-01 5.44074476e-01
6.47913218e-01 -1.74012578e+00 2.60217816e-01 -4.49616879e-01
-4.65684980e-01 1.17131531e+00 6.14441216e-01 -3.35638314e-01
5.65079272e-01 5.02033234e-01 4.50228415e-02 -2.51251012e-01
-6.42696559e-01 -7.06500590e-01 4.75587755e-01 5.76964319e-01
4.17543620e-01 -4.12690371e-01 8.82457197e-02 2.61257946e-01
1.09816425e-01 -1.64177746e-01 1.51110617e-02 1.06800067e+00
-3.65876704e-01 -1.32037175e+00 -2.49687955e-01 4.93979231e-02
-2.27484956e-01 -1.73741877e-01 -6.11801863e-01 9.48174059e-01
1.66107133e-01 7.07071126e-01 8.81630331e-02 -4.25273240e-01
4.77649301e-01 -1.04773827e-01 6.50349021e-01 -8.31101477e-01
-4.54076022e-01 3.99899840e-01 1.35398924e-01 -1.18385875e+00
-5.67092896e-01 -1.14519382e+00 -1.28476679e+00 -3.42991427e-02
-5.86334825e-01 -9.20825303e-02 3.31017971e-01 1.05244172e+00
3.15517604e-01 1.26842394e-01 6.75125480e-01 -1.17579877e+00
-8.17995250e-01 -7.58120894e-01 -4.75897312e-01 5.97582102e-01
7.23996341e-01 -8.26800764e-01 -4.53884304e-01 1.55340776e-01] | [8.008402824401855, -2.1625423431396484] |
4c13294f-87c3-40fa-9db9-480bd61dce73 | commonsense-knowledge-augmented-pretrained | null | null | https://openreview.net/forum?id=51c-7iGxPri | https://openreview.net/pdf?id=51c-7iGxPri | Commonsense Knowledge-Augmented Pretrained Language Models for Causal Reasoning Classification | Commonsense knowledge can be leveraged for identifying causal relations in text. In this work, we convert triples in ATOMIC2020, a wide coverage commonsense reasoning knowledge graph, to natural language text and continually pretrain a BERT pretrained language model. We evaluate the resulting model on answering commonsense reasoning questions. Our results show that a continually pretrained language model augmented with commonsense reasoning knowledge outperforms our baseline on two commonsense causal reasoning benchmarks, COPA and BCOPA-CE, without additional improvement on the base model or using quality-enhanced data for fine-tuning. | ['Anonymous'] | 2021-09-17 | null | null | null | acl-arr-september-2021-9 | ['commonsense-causal-reasoning'] | ['natural-language-processing'] | [ 2.24131271e-01 5.69966376e-01 -6.02106988e-01 -3.21173817e-01
-4.52500433e-01 -5.16303420e-01 1.09839332e+00 4.61675793e-01
-3.03547710e-01 1.07284784e+00 8.88869047e-01 -4.64275718e-01
-2.82247633e-01 -1.16022122e+00 -8.83303285e-01 2.26839408e-01
1.38278887e-01 8.86894286e-01 4.93685216e-01 -7.98445344e-01
3.22570354e-01 3.99473533e-02 -1.01431775e+00 6.47816598e-01
1.24249399e+00 6.67067826e-01 -2.82668084e-01 4.79034722e-01
-2.57915437e-01 2.21204376e+00 -3.99718165e-01 -1.05389941e+00
-3.04814577e-01 -5.01421392e-01 -1.55067062e+00 -8.14549923e-01
5.90092063e-01 -1.51507497e-01 -7.31608927e-01 1.03843784e+00
-3.84947360e-02 2.42220849e-01 7.00166106e-01 -1.48299158e+00
-1.34778643e+00 1.65630698e+00 -1.51623681e-01 6.79447234e-01
7.58725822e-01 2.08318055e-01 1.61105597e+00 -2.73555338e-01
7.65114307e-01 1.75088799e+00 6.00338101e-01 5.59428990e-01
-1.43631506e+00 -6.68228030e-01 -7.63759464e-02 9.76999223e-01
-7.89645255e-01 -2.14207903e-01 1.02869713e+00 -2.56511360e-01
1.55362594e+00 1.97749570e-01 4.65041459e-01 1.66609013e+00
2.15099469e-01 7.47324228e-01 1.10527050e+00 -3.61581266e-01
1.48850113e-01 -4.31487948e-01 6.78365707e-01 9.72677231e-01
4.43049818e-01 7.92517364e-02 -6.23592436e-01 -3.88211757e-01
4.17747408e-01 -2.56180972e-01 -8.79596621e-02 4.74430621e-02
-1.39308023e+00 9.38463449e-01 9.73584712e-01 2.66856283e-01
-3.42964262e-01 9.36202109e-01 7.23079324e-01 4.33714360e-01
1.76834986e-01 1.01430738e+00 -4.51878607e-01 -2.92225838e-01
-5.65562546e-01 4.79699910e-01 1.10702968e+00 8.24755192e-01
1.62040815e-01 -1.69863760e-01 -4.89453286e-01 7.04901099e-01
1.73963830e-01 5.99558234e-01 3.75927448e-01 -1.03533459e+00
6.92431390e-01 9.46158588e-01 -1.93270817e-01 -1.10265255e+00
-4.67657626e-01 1.83743104e-01 -5.19476414e-01 -5.33226967e-01
2.42916062e-01 -8.65905285e-02 -7.81433523e-01 1.77907121e+00
-2.00228795e-01 4.45649147e-01 2.72458375e-01 7.28409767e-01
1.13229370e+00 3.06940198e-01 4.32566345e-01 2.32015867e-02
1.33134794e+00 -6.93326116e-01 -8.01241100e-01 -5.61372221e-01
3.80215317e-01 -1.06762521e-01 1.52989936e+00 2.58323193e-01
-7.43066370e-01 2.10386097e-01 -9.59340811e-01 -7.37043321e-01
-4.56428409e-01 -4.23655212e-01 1.16075408e+00 -6.14979081e-02
-5.83616316e-01 5.55262744e-01 -6.00578785e-01 -2.70914465e-01
7.87436664e-01 -3.05894822e-01 1.40291629e-02 -2.98552752e-01
-2.12321615e+00 1.31999850e+00 1.11436987e+00 -4.22209769e-01
-1.00057876e+00 -1.15280199e+00 -9.56468165e-01 6.61935583e-02
1.01662874e+00 -1.12046254e+00 1.00143659e+00 -1.57921359e-01
-1.21000159e+00 9.87304151e-01 1.94101691e-01 -9.40548122e-01
5.83202243e-01 -4.40248787e-01 -6.17226899e-01 7.37705380e-02
3.47019762e-01 2.56948709e-01 3.32435340e-01 -1.05057156e+00
-1.76255286e-01 -1.19180381e-01 7.74285197e-01 -2.88944989e-01
1.13398150e-01 8.29509348e-02 1.35452732e-01 -6.86270237e-01
-4.03565615e-01 -7.39887893e-01 1.40560672e-01 -4.11527008e-01
-9.64897633e-01 -7.31792986e-01 5.94468355e-01 -4.99568909e-01
1.17943585e+00 -1.50192201e+00 2.33985588e-01 1.19221032e-01
3.92428279e-01 -3.71601224e-01 -8.33760500e-02 2.45997623e-01
-2.93022037e-01 2.75259435e-01 -2.55501896e-01 2.99408168e-01
3.62526476e-01 6.52338386e-01 -9.90069687e-01 -1.12051308e-01
3.28671306e-01 1.50373328e+00 -1.74078465e+00 -7.89061844e-01
1.00352310e-01 2.89681531e-03 -7.74041176e-01 -6.61246702e-02
-1.00561082e+00 -3.24238658e-01 -2.62003362e-01 5.36683500e-01
-2.10076309e-04 -5.69129229e-01 1.63682103e-01 -3.62399369e-01
5.55778325e-01 1.04397655e+00 -4.66181964e-01 1.64820266e+00
-5.11521459e-01 7.04804361e-01 -8.66270423e-01 -6.43010914e-01
4.65300798e-01 1.42254353e-01 -5.97901940e-02 -6.84402227e-01
1.15616456e-01 -6.02510907e-02 1.47213802e-01 -6.18921220e-01
3.99126023e-01 -6.93243444e-01 -3.74950051e-01 5.32083988e-01
2.72118181e-01 -7.24386334e-01 5.82540989e-01 7.91511714e-01
1.43747485e+00 -1.44569293e-01 4.80419159e-01 -2.94297904e-01
3.00925225e-01 5.59944212e-01 1.29773110e-01 7.53563643e-01
-6.08919896e-02 -1.97976723e-01 1.10956109e+00 -2.50714540e-01
-6.55943394e-01 -1.28097165e+00 -4.23480049e-02 1.17138255e+00
7.87508339e-02 -6.11514568e-01 -9.65897515e-02 -9.79887247e-01
4.08534765e-01 1.89432931e+00 -9.73678708e-01 -5.21866143e-01
-4.65375006e-01 -5.92948973e-01 1.17521310e+00 6.40993059e-01
6.53637111e-01 -1.47770369e+00 -4.44798529e-01 7.30846673e-02
-6.37234926e-01 -1.35068977e+00 7.57368952e-02 1.89664304e-01
-6.11351609e-01 -1.63293791e+00 4.39733654e-01 -1.44855604e-01
-8.00991431e-02 -2.78049707e-01 1.81218612e+00 1.78548798e-01
-2.72739559e-01 2.20682278e-01 -2.46171027e-01 -4.92557943e-01
-6.82651460e-01 -1.61191106e-01 -9.93011668e-02 -9.30987537e-01
6.94280505e-01 -7.54794896e-01 1.18395081e-02 -4.22698021e-01
-7.30431020e-01 -1.61942467e-01 -7.25713372e-02 9.33325887e-01
7.04310983e-02 2.18494281e-01 5.09864748e-01 -1.04951775e+00
1.10760868e+00 -8.65787089e-01 -2.19859138e-01 3.82721931e-01
-6.31154180e-01 4.93383944e-01 7.47356117e-01 -1.05936535e-01
-1.14478910e+00 -8.03465128e-01 2.42541239e-01 -4.13226396e-01
1.74441546e-01 7.97683120e-01 1.01250067e-01 6.12965584e-01
1.22342503e+00 -1.63068295e-01 -5.70516884e-01 1.78438708e-01
1.10251403e+00 -7.52452686e-02 9.80562389e-01 -1.28902841e+00
8.21232319e-01 3.86341572e-01 2.38377694e-02 -1.32188067e-01
-1.77328467e+00 -1.81976616e-01 -3.23802263e-01 3.34445655e-01
8.87279034e-01 -7.51338482e-01 -8.48615706e-01 -2.11601734e-01
-1.42235959e+00 -6.71209157e-01 -4.68638927e-01 1.79141849e-01
-8.36400688e-01 -9.41779872e-04 -7.91340292e-01 -2.60733634e-01
-3.44008237e-01 -3.58469963e-01 6.62214637e-01 -1.19792446e-01
-6.19667470e-01 -1.44230878e+00 4.20628667e-01 6.30159676e-01
1.26293436e-01 5.83302677e-01 1.50255072e+00 -6.51052415e-01
-1.43336877e-01 1.31825566e-01 -6.23782992e-01 9.28896363e-04
2.20506981e-01 -3.54914144e-02 -7.87806332e-01 5.73191464e-01
-5.87383151e-01 -1.19371152e+00 1.32459331e+00 -2.30888017e-02
1.30562377e+00 -4.70731109e-01 -3.37953418e-01 2.38308504e-01
1.29575610e+00 -3.48374009e-01 7.30840623e-01 4.53463405e-01
8.88742626e-01 1.08431108e-01 2.20081225e-01 2.77011484e-01
9.29665983e-01 6.20790944e-02 5.69799304e-01 3.92109126e-01
-2.08787382e-01 -7.34099448e-01 1.49088979e-01 1.18950658e-01
-5.93314350e-01 -1.53524252e-02 -1.41198230e+00 8.14307511e-01
-1.85820639e+00 -1.88683295e+00 -1.86949179e-01 1.06691647e+00
1.72836137e+00 2.62622207e-01 -9.55895483e-02 -2.23809630e-02
2.53772855e-01 2.48374194e-01 -6.40898228e-01 -4.42232549e-01
-3.41188222e-01 4.41075563e-01 2.67449766e-01 6.80761337e-01
-9.86392438e-01 1.41601062e+00 6.49751759e+00 7.53215194e-01
-5.56358635e-01 1.33972734e-01 -9.52096134e-02 -2.87581921e-01
-8.48373950e-01 2.03194931e-01 -1.29319876e-01 2.30075970e-01
8.42619896e-01 -7.11185694e-01 8.12720418e-01 7.28878319e-01
-1.83342695e-01 -8.20030738e-03 -1.41915917e+00 7.35909462e-01
1.13260597e-01 -1.77648222e+00 3.56828272e-01 -3.58322978e-01
7.95693099e-01 2.53037274e-01 -4.11310971e-01 8.71427894e-01
1.51122165e+00 -1.19026554e+00 7.69311070e-01 6.35121226e-01
4.47015256e-01 -6.78199947e-01 6.48706317e-01 8.49368870e-02
-6.24700963e-01 -1.96841091e-01 -3.29773724e-01 -2.23565847e-01
1.82551458e-01 9.20125008e-01 -8.64603281e-01 5.86400986e-01
5.34256876e-01 8.87432277e-01 -7.90144563e-01 9.39557552e-02
-1.17599189e+00 9.21235085e-01 -1.29600912e-01 -2.69245207e-01
1.95285961e-01 4.63912964e-01 5.89768648e-01 1.38300586e+00
-5.11985123e-01 6.17466807e-01 6.98615517e-03 1.71861875e+00
-6.66286170e-01 -5.02436340e-01 -5.60930669e-01 -6.29418910e-01
4.91933525e-01 8.06578159e-01 -2.12545410e-01 -8.17106187e-01
-3.53779420e-02 9.49196696e-01 8.21832061e-01 9.39131528e-02
-1.33462250e+00 -2.34132260e-01 4.26415533e-01 -7.31836110e-02
-2.04510570e-01 1.54824555e-01 -4.17314380e-01 -1.41842079e+00
-4.69492793e-01 -7.14538753e-01 8.95877600e-01 -1.23487628e+00
-1.93121755e+00 1.42622039e-01 4.06660229e-01 -3.44474688e-02
-4.38292772e-01 -6.78955913e-01 -7.26479411e-01 4.58725035e-01
-1.47642398e+00 -1.23326290e+00 -3.32935482e-01 8.16652298e-01
1.83003068e-01 2.22427584e-02 6.84670806e-01 -3.47573578e-01
-3.44218165e-01 3.36401075e-01 -7.21628547e-01 4.00499195e-01
6.91080272e-01 -1.73816383e+00 4.02966142e-01 5.53715289e-01
1.08422093e-01 1.14498234e+00 1.05106235e+00 -1.03574979e+00
-1.20493579e+00 -1.08980513e+00 1.00182998e+00 -1.18899512e+00
1.78501153e+00 -6.12439355e-03 -8.01859915e-01 1.18185568e+00
4.45111126e-01 4.86519635e-02 6.67219877e-01 8.70173514e-01
-1.22067511e+00 4.73890364e-01 -1.10669839e+00 7.84197509e-01
1.58647954e+00 -9.89624560e-01 -1.93371189e+00 4.52725589e-01
1.20302808e+00 -3.36035728e-01 -9.50238049e-01 2.33204037e-01
2.04539627e-01 -4.16864455e-01 1.07749104e+00 -1.46572220e+00
1.52087820e+00 -1.45578384e-01 -3.08947623e-01 -1.66856003e+00
-3.57415259e-01 -2.69292235e-01 -5.89718938e-01 8.71259093e-01
6.24605536e-01 -3.98761511e-01 2.04635888e-01 8.18376184e-01
1.77846551e-01 -2.41993293e-01 -7.30876088e-01 -7.71911621e-01
3.46370399e-01 -7.00037360e-01 5.49392998e-01 1.72606599e+00
7.65084863e-01 1.09274161e+00 1.74866185e-01 -7.70195201e-02
8.14717233e-01 3.34054410e-01 3.63173544e-01 -1.25215065e+00
-2.89739221e-01 -6.04849100e-01 -1.68829143e-01 -2.00771108e-01
8.78359437e-01 -1.41368663e+00 -2.42201298e-01 -1.96703184e+00
7.67727554e-01 -1.46137699e-01 -3.38410050e-01 1.15313625e+00
-6.96825743e-01 -3.99826765e-02 1.68165520e-01 -2.43474126e-01
-7.26803899e-01 6.29620075e-01 1.26742077e+00 -5.13626933e-01
5.08550368e-02 -9.15581703e-01 -9.36559677e-01 8.21715534e-01
7.57370412e-01 -4.21572894e-01 -8.18606198e-01 -3.74098569e-01
9.80736613e-01 -1.39277205e-01 1.21485651e+00 -4.56469148e-01
2.73837477e-01 -8.60141695e-01 8.24561641e-02 -3.44452798e-01
-3.44004929e-02 -4.74075317e-01 -3.28750879e-01 3.99722695e-01
-8.75703752e-01 2.15453003e-02 5.36382854e-01 6.31751537e-01
5.12820028e-04 8.36459696e-02 4.28227723e-01 -1.14020929e-01
-9.13298965e-01 -2.81503409e-01 4.11581576e-01 9.82659280e-01
6.90882385e-01 3.76458555e-01 -1.15467274e+00 -1.92857087e-01
-5.69565773e-01 2.64626652e-01 2.29404926e-01 6.37494147e-01
5.33930957e-01 -1.29522967e+00 -8.86320889e-01 -8.43328774e-01
4.39797044e-01 -2.39609554e-01 -5.22655100e-02 6.76793277e-01
-4.90178764e-01 5.57860017e-01 -1.18024856e-01 6.32002950e-02
-5.48897862e-01 9.12492037e-01 4.38078195e-01 -5.64868033e-01
-6.01373672e-01 9.49615717e-01 -3.77684385e-01 -5.59349954e-01
-3.23414892e-01 -8.98855627e-01 -1.63530365e-01 -2.25953549e-01
4.53264356e-01 4.23913866e-01 -3.93489450e-01 2.93324776e-02
-7.28242636e-01 -1.85546190e-01 2.05481518e-03 -4.56876680e-02
1.40536022e+00 4.42632616e-01 -7.50982344e-01 6.70545161e-01
5.79802573e-01 2.61294246e-01 -4.93781984e-01 -2.75880933e-01
3.20201904e-01 -9.08244699e-02 1.56558201e-01 -1.56310749e+00
-6.22098029e-01 4.40169752e-01 -5.12603939e-01 8.65103379e-02
5.00815511e-01 5.09551704e-01 4.73263621e-01 8.85620773e-01
4.51791614e-01 -8.55987966e-01 3.16235512e-01 1.10059762e+00
1.55955446e+00 -1.25345647e+00 1.19197778e-01 -4.02166605e-01
-7.35697210e-01 9.41839814e-01 6.81355476e-01 -6.16696239e-01
4.44207579e-01 3.30412865e-01 -4.03176516e-01 -9.10679221e-01
-1.21158385e+00 -3.10554773e-01 3.85666549e-01 3.89490396e-01
6.48006201e-01 3.24807495e-01 -1.65888637e-01 6.84296966e-01
-9.01502192e-01 3.11401069e-01 5.22537768e-01 4.60091084e-01
-1.06154144e-01 -3.95893365e-01 -2.03638777e-01 5.88590443e-01
-2.44788099e-02 -8.62023532e-01 -9.97569203e-01 9.00628269e-01
4.05089185e-02 9.72180843e-01 -8.31075460e-02 -2.58369178e-01
4.27357644e-01 1.35636106e-01 1.04366672e+00 -5.68039238e-01
-4.36227024e-01 -1.09890485e+00 7.09250867e-01 -8.05942893e-01
-4.48867500e-01 -2.92462319e-01 -1.86346579e+00 -8.57659042e-01
-4.16216552e-02 -1.09647565e-01 -1.00187741e-01 1.23835146e+00
-9.78889596e-03 9.31343675e-01 -3.39704067e-01 2.15208784e-01
-5.41740656e-01 -9.97018933e-01 -6.23169281e-02 6.82132304e-01
2.68436372e-01 -8.88083935e-01 -3.51231813e-01 2.27998748e-01] | [10.029989242553711, 8.075284957885742] |
114f667f-5d60-4b6c-bc14-37441e18e315 | is-chatgpt-a-biomedical-expert-exploring-the | 2306.16108 | null | https://arxiv.org/abs/2306.16108v1 | https://arxiv.org/pdf/2306.16108v1.pdf | Is ChatGPT a Biomedical Expert? -- Exploring the Zero-Shot Performance of Current GPT Models in Biomedical Tasks | We assessed the performance of commercial Large Language Models (LLMs) GPT-3.5-Turbo and GPT-4 on tasks from the 2023 BioASQ challenge. In Task 11b Phase B, which is focused on answer generation, both models demonstrated competitive abilities with leading systems. Remarkably, they achieved this with simple zero-shot learning, grounded with relevant snippets. Even without relevant snippets, their performance was decent, though not on par with the best systems. Interestingly, the older and cheaper GPT-3.5-Turbo system was able to compete with GPT-4 in the grounded Q&A setting on factoid and list answers. In Task 11b Phase A, focusing on retrieval, query expansion through zero-shot learning improved performance, but the models fell short compared to other systems. The code needed to rerun these experiments is available through GitHub. | ['Udo Kruschwitz', 'Samy Ateia'] | 2023-06-28 | null | null | null | null | ['retrieval', 'answer-generation'] | ['methodology', 'natural-language-processing'] | [-1.69032782e-01 -1.06629334e-01 -2.31843516e-01 1.93703756e-01
-1.65079808e+00 -4.77425426e-01 8.88325930e-01 3.44042718e-01
-7.23004341e-01 8.06702793e-01 3.05741459e-01 -3.79736930e-01
-3.65918338e-01 -5.56807995e-01 -4.88385975e-01 -1.29366621e-01
-3.27646255e-01 1.14032769e+00 5.05950212e-01 -8.67106974e-01
4.53002036e-01 -3.20042133e-01 -1.27370274e+00 7.14464009e-01
8.11744750e-01 8.47569406e-01 1.56265214e-01 1.19399023e+00
-5.08699000e-01 1.02741718e+00 -6.81229889e-01 -6.05704725e-01
2.25796416e-01 -2.50760555e-01 -9.41938579e-01 -7.08727896e-01
7.82224119e-01 -1.17962688e-01 -6.21457338e-01 6.45369768e-01
1.09497952e+00 3.27022851e-01 4.32177633e-01 -1.18469524e+00
-7.79474199e-01 7.71651626e-01 -1.87351685e-02 5.58947384e-01
7.65852332e-01 3.09834003e-01 1.54757345e+00 -1.24640763e+00
8.89847934e-01 1.44655180e+00 7.35308051e-01 6.01276577e-01
-1.15812278e+00 -5.90052783e-01 -4.81958330e-01 4.37126607e-01
-1.38703108e+00 -7.68769443e-01 -1.98009804e-01 -2.34493852e-01
1.66538095e+00 4.01374191e-01 4.96921599e-01 1.13237870e+00
4.55537409e-01 1.09990823e+00 8.33234191e-01 -4.27868575e-01
4.44370359e-01 1.28527448e-01 4.45767850e-01 3.38910520e-01
-8.74192268e-02 2.89990515e-01 -9.75363016e-01 -4.71000314e-01
-7.76291713e-02 -3.32659334e-01 -2.02466160e-01 -2.17834398e-01
-1.18957198e+00 9.31574702e-01 4.23102647e-01 4.49147105e-01
-3.44961882e-01 2.10148647e-01 5.19714773e-01 7.03696012e-01
4.27540988e-01 1.02253282e+00 -6.17219865e-01 -4.22603250e-01
-1.25353754e+00 7.78909445e-01 1.13677990e+00 1.24779999e+00
3.37354809e-01 -9.67301428e-02 -1.18750310e+00 9.41355348e-01
-1.32212579e-01 5.50013244e-01 8.81599486e-01 -1.05828166e+00
6.63586080e-01 1.29181445e-01 2.72870243e-01 -5.07223547e-01
-5.44153571e-01 -5.72240591e-01 -4.33053166e-01 -6.14362121e-01
3.34249526e-01 -3.06756109e-01 -1.24215508e+00 1.32012904e+00
-3.07585448e-01 -6.12025149e-02 3.26516837e-01 7.72384882e-01
1.29908121e+00 9.81262863e-01 2.34103277e-01 -3.45694907e-02
1.43437648e+00 -1.08205700e+00 -7.62528539e-01 -3.93413842e-01
1.01150215e+00 -8.67468774e-01 1.24863470e+00 2.74667531e-01
-1.36737454e+00 -7.69745171e-01 -8.63615453e-01 -2.87340760e-01
-5.40470362e-01 -5.00197232e-01 6.98636293e-01 5.92576087e-01
-1.51954103e+00 5.23000002e-01 -1.66279763e-01 -6.99566662e-01
6.66570067e-02 1.31580934e-01 -6.77763000e-02 -3.99700969e-01
-2.08866549e+00 1.10899496e+00 4.17699218e-01 -4.60957378e-01
-9.68510926e-01 -9.24236417e-01 -7.89871216e-01 3.72030497e-01
5.01499593e-01 -8.06505799e-01 1.84073639e+00 9.03030112e-03
-1.23327827e+00 5.40075064e-01 -1.93189755e-02 -1.13543940e+00
2.12552890e-01 -3.39666426e-01 -4.44138885e-01 1.77853733e-01
2.64748871e-01 1.17859817e+00 4.24863189e-01 -5.11324883e-01
-5.94057381e-01 -3.42416465e-02 -1.00653246e-02 3.54361594e-01
-2.49524921e-01 1.95537768e-02 -5.72257698e-01 -1.12208933e-01
-3.06234688e-01 -8.46006274e-01 -1.84558660e-01 -7.16317236e-01
-9.05889422e-02 -6.09081089e-01 2.43726596e-01 -5.87348104e-01
1.46605325e+00 -1.66073906e+00 -2.05506179e-02 -2.29833871e-01
3.09132989e-02 6.06951118e-01 -5.68312228e-01 8.83034766e-01
1.90679595e-01 3.06619972e-01 1.09225720e-01 -1.72472015e-01
2.81084716e-01 -1.31731540e-01 -3.67957979e-01 -1.33342892e-01
2.51393542e-02 1.60323787e+00 -1.25323045e+00 -5.60511112e-01
-1.54510334e-01 -1.21015042e-01 -6.57297552e-01 -7.54273981e-02
-4.75923419e-01 -3.28801215e-01 -2.68333435e-01 7.64001548e-01
7.31411576e-02 -3.53003532e-01 -1.89616770e-01 3.29002827e-01
1.46554202e-01 4.39499229e-01 -6.05540335e-01 2.17922378e+00
-4.09557700e-01 5.16442418e-01 -3.09339345e-01 -4.65233386e-01
6.04540825e-01 5.85249543e-01 4.34851170e-01 -1.52904165e+00
-3.70616406e-01 5.39006054e-01 2.47540951e-01 -5.98221898e-01
1.03584135e+00 -2.94423550e-01 -4.05878007e-01 1.93278715e-01
6.33825541e-01 -6.82918787e-01 6.45811617e-01 9.50590491e-01
1.41169810e+00 -7.32758939e-02 8.93196911e-02 -3.30024868e-01
3.23051959e-01 3.19199145e-01 2.64548391e-01 1.40066481e+00
-1.37741983e-01 6.32104635e-01 2.78908789e-01 -1.38709590e-01
-1.01291490e+00 -9.46409345e-01 -3.96139473e-02 1.45969856e+00
-2.31786385e-01 -1.16001821e+00 -4.78434503e-01 -4.19953853e-01
3.20239305e-01 1.08073044e+00 -2.70521671e-01 -2.70029664e-01
-1.55104980e-01 -5.95631838e-01 6.44584656e-01 2.38623619e-01
2.37765104e-01 -1.18111610e+00 -4.71539617e-01 3.67095619e-01
-4.20732468e-01 -8.13673079e-01 -6.47494674e-01 2.95595229e-01
-7.46943116e-01 -6.81641459e-01 -1.39775848e+00 -3.13561767e-01
-2.24773422e-01 2.59913534e-01 1.67851710e+00 1.34415135e-01
-2.09069401e-01 6.12314701e-01 -5.52086592e-01 -6.09116256e-01
-1.58323362e-01 6.42804623e-01 -6.75497800e-02 -7.33410597e-01
4.84769940e-01 1.20070472e-01 -6.22024000e-01 -1.83897850e-03
-8.04362595e-01 -3.89048010e-01 6.29967332e-01 1.14189255e+00
1.58446670e-01 -4.35673684e-01 9.73221064e-01 -9.55615461e-01
1.24172807e+00 -6.57971025e-01 -3.13607484e-01 4.50902641e-01
-8.87412667e-01 1.38972297e-01 2.21361578e-01 2.17069201e-02
-8.91841292e-01 -5.91805339e-01 -3.51391315e-01 -2.82814413e-01
5.63023806e-01 6.88007534e-01 4.37134326e-01 2.00654387e-01
1.17827094e+00 1.84320703e-01 -2.68220276e-01 -4.95454997e-01
6.18889868e-01 7.55585790e-01 2.42563203e-01 -4.27235603e-01
4.08630848e-01 -1.96510911e-01 -5.32597601e-01 -7.59868443e-01
-8.09778214e-01 -1.11406636e+00 -1.37971953e-01 1.27423778e-01
5.61263561e-01 -9.93435323e-01 -2.36896113e-01 1.29953265e-01
-9.20855045e-01 -4.46706504e-01 -5.06868303e-01 2.38455057e-01
-7.19765961e-01 1.59556180e-01 -1.00307417e+00 -8.90851676e-01
-8.71222854e-01 -8.09615254e-01 1.28160560e+00 1.37378693e-01
-4.30501938e-01 -8.15554380e-01 4.00522500e-01 5.43326974e-01
7.32706428e-01 -6.69064343e-01 9.93206799e-01 -1.18270600e+00
-3.71008575e-01 -4.67075855e-01 -3.54528800e-02 -1.72302115e-03
-4.27677572e-01 -5.46133876e-01 -9.11384761e-01 -5.88184297e-01
-2.19240442e-01 -9.28134441e-01 1.31252885e+00 2.11407661e-01
6.66259766e-01 -2.29878560e-01 -2.20210686e-01 2.41700813e-01
1.04637396e+00 2.56338269e-01 7.37912774e-01 2.14765891e-01
3.22226994e-02 6.61682129e-01 6.45840526e-01 2.34116301e-01
2.16354743e-01 7.24797547e-01 -1.61628991e-01 4.13143396e-01
-4.63788331e-01 -3.80657703e-01 4.74560082e-01 9.96908605e-01
4.90175992e-01 -6.32614434e-01 -1.21027792e+00 6.45117283e-01
-1.79084229e+00 -9.54972148e-01 -1.39575884e-01 2.24139690e+00
8.29133213e-01 5.00697970e-01 1.35111675e-01 -2.19140887e-01
9.24290568e-02 1.89458773e-01 -4.28681195e-01 -4.20636415e-01
-3.44083160e-01 5.50323308e-01 3.53784561e-01 5.56649625e-01
-5.71474791e-01 1.24602914e+00 7.20603228e+00 1.33378053e+00
-4.67950344e-01 4.62606013e-01 4.93700683e-01 -5.27559638e-01
-5.14089346e-01 4.83113341e-02 -9.79733884e-01 4.31318104e-01
1.85508072e+00 -7.44729459e-01 1.04797274e-01 4.67829973e-01
-2.74119198e-01 -2.31469303e-01 -1.04718244e+00 7.98334956e-01
1.25126272e-01 -1.50586259e+00 1.25228148e-03 -1.15367204e-01
8.75599921e-01 6.28412902e-01 6.83513284e-02 1.53397274e+00
3.22392255e-01 -1.07121325e+00 4.93260592e-01 6.44550920e-01
7.02881932e-01 -5.44362903e-01 8.87563050e-01 6.05686605e-01
-7.02741325e-01 -2.99956620e-01 -6.81507289e-01 5.03225476e-02
1.39611304e-01 2.45464653e-01 -1.03812361e+00 7.09530294e-01
5.08220553e-01 3.83813471e-01 -8.20933938e-01 1.46279597e+00
2.73063226e-04 6.24865472e-01 -1.92637309e-01 -2.81915337e-01
4.13521260e-01 5.11206329e-01 6.37156844e-01 1.30479360e+00
4.04501766e-01 1.75637797e-01 3.65515873e-02 5.17003655e-01
-3.25422436e-01 2.73138076e-01 -5.62042475e-01 -3.80580544e-01
4.76257086e-01 1.19899821e+00 -1.31597772e-01 -8.67799461e-01
-2.58923709e-01 8.41185868e-01 1.50234461e-01 4.11798149e-01
-4.67219770e-01 -5.70234835e-01 2.67152563e-02 2.90363014e-01
-8.40769038e-02 1.68649986e-01 1.89590946e-01 -1.18138707e+00
-3.49227011e-01 -1.29172027e+00 7.04301894e-01 -1.03512847e+00
-1.37776113e+00 5.50901115e-01 1.61078289e-01 -9.51262951e-01
-7.67858505e-01 -3.17425847e-01 -3.43817472e-01 1.03804958e+00
-1.31260061e+00 -6.79303408e-01 2.31874168e-01 4.01815146e-01
8.69324803e-01 -4.10450399e-01 1.05933177e+00 4.44146454e-01
-1.50834739e-01 6.89689338e-01 3.25998425e-01 -3.07788968e-01
9.82929230e-01 -1.29314280e+00 8.06431890e-01 3.87914151e-01
6.03941679e-01 6.67505503e-01 6.19537592e-01 -8.43372464e-01
-1.53148234e+00 -5.47158957e-01 1.28650773e+00 -8.34532201e-01
8.94602001e-01 -4.80751157e-01 -9.48354304e-01 4.20960367e-01
5.23254216e-01 -2.33907074e-01 4.51956034e-01 5.27540982e-01
-1.78319141e-01 5.57186268e-03 -6.64307237e-01 4.37321573e-01
7.64685154e-01 -8.12024295e-01 -9.38517988e-01 8.73431861e-01
1.22346592e+00 -3.96930873e-01 -8.08154047e-01 4.65741068e-01
4.86848891e-01 -6.69663310e-01 1.13046122e+00 -9.35891688e-01
1.33086279e-01 2.68986523e-01 -2.66798347e-01 -1.32248271e+00
-4.58968669e-01 -1.02213502e+00 -4.59008008e-01 6.44257009e-01
7.19576776e-01 -3.55690867e-01 7.90099263e-01 4.70656455e-01
-1.86794907e-01 -6.85323060e-01 -1.01708210e+00 -1.03617787e+00
4.77711976e-01 -2.86770850e-01 2.45484337e-01 3.56757760e-01
2.63061970e-01 8.12153041e-01 -3.07181865e-01 -7.39358008e-01
3.53531331e-01 -2.83392757e-01 5.84507287e-01 -1.03641164e+00
-4.03851867e-01 -4.65999097e-01 8.25283825e-02 -9.60121512e-01
-1.12133943e-01 -1.08491170e+00 1.59992918e-01 -1.70941901e+00
2.12631509e-01 -2.72030562e-01 -6.94194615e-01 1.70418441e-01
-4.71050531e-01 7.71776289e-02 5.76252341e-01 3.58500600e-01
-1.18007886e+00 6.05371177e-01 1.09899747e+00 -4.16087508e-01
-1.37196571e-01 2.29155496e-02 -7.42193520e-01 9.58963484e-02
5.66189051e-01 -4.24844176e-01 -6.28811121e-01 -2.84964263e-01
4.15475845e-01 6.32445455e-01 1.40231594e-01 -1.20277596e+00
7.21367478e-01 3.84872586e-01 2.39783838e-01 -7.53656745e-01
6.32389486e-01 -4.64450866e-02 -1.28089130e-01 6.16556525e-01
-8.03487182e-01 1.12464517e-01 4.91255730e-01 5.03757417e-01
-2.12944463e-01 -5.34022510e-01 1.52443007e-01 -5.10351062e-01
-9.13760602e-01 2.92001784e-01 -3.35485518e-01 5.51515639e-01
4.84992296e-01 9.61389169e-02 -6.05313718e-01 -7.49129415e-01
-5.86359739e-01 8.31030369e-01 -8.66491944e-02 7.00436532e-01
7.90578365e-01 -1.30928636e+00 -1.05063546e+00 -1.55197293e-01
6.34115934e-01 -5.97550750e-01 5.38354874e-01 8.62607360e-01
-1.99855819e-01 1.51494086e+00 7.61635229e-02 -4.42695856e-01
-8.62179518e-01 5.77085137e-01 3.82517278e-01 -8.24383199e-01
-2.57171661e-01 1.19223487e+00 -8.20909515e-02 -5.35915732e-01
3.64173561e-01 -9.92483795e-02 -1.94617778e-01 3.97946864e-01
7.72033513e-01 3.61663342e-01 4.58342761e-01 6.36065975e-02
-2.33038768e-01 -1.91680491e-02 -2.84154117e-01 -5.52799821e-01
9.80963051e-01 8.47121514e-03 2.61706561e-01 5.75592279e-01
1.12142253e+00 -3.51516306e-01 -5.82361937e-01 -4.36549127e-01
6.05888367e-01 -1.11739703e-01 1.64403602e-01 -1.22819710e+00
-3.92441213e-01 9.92421031e-01 5.08411825e-01 1.68837339e-01
6.59914732e-01 -1.95102796e-01 1.20804298e+00 8.03876102e-01
7.57812738e-01 -1.25330961e+00 2.37403303e-01 1.01484334e+00
8.90991688e-01 -1.35697460e+00 -5.80400266e-02 4.96055186e-01
-5.72280109e-01 7.84329474e-01 6.17597699e-01 1.80473566e-01
2.98336059e-01 -2.27500200e-01 -1.96658745e-01 -5.09854436e-01
-1.51271784e+00 -3.58561218e-01 6.73784614e-01 2.71656185e-01
6.95915103e-01 -6.92960694e-02 -3.62552017e-01 4.45592344e-01
-1.97971940e-01 2.36448035e-01 1.97088107e-01 9.13495541e-01
-7.61719048e-01 -1.12618363e+00 -5.51134236e-02 6.77385271e-01
-4.20337409e-01 -8.47621024e-01 -3.43228757e-01 6.83803618e-01
-1.62508667e-01 1.14844394e+00 -1.07906483e-01 -5.00962138e-01
5.06586015e-01 5.54784298e-01 3.51015955e-01 -1.10864806e+00
-9.93833482e-01 1.16684064e-01 5.24629414e-01 -7.38514364e-01
1.96312681e-01 -4.79097486e-01 -9.67981100e-01 -3.13138306e-01
-2.81405121e-01 7.84548104e-01 3.51357132e-01 4.81240898e-01
5.12817681e-01 4.87531126e-01 -2.36629080e-02 -3.03520530e-01
-1.10435641e+00 -1.46693492e+00 -5.49478829e-01 1.37944326e-01
1.30934402e-01 -2.89348006e-01 -1.15904905e-01 -5.66987813e-01] | [11.454195976257324, 7.901678085327148] |
c4a4d557-56a6-4d37-af63-f10ad9fa36d5 | latent-correlation-based-multiview-learning | 2106.07115 | null | https://arxiv.org/abs/2106.07115v3 | https://arxiv.org/pdf/2106.07115v3.pdf | Understanding Latent Correlation-Based Multiview Learning and Self-Supervision: An Identifiability Perspective | Multiple views of data, both naturally acquired (e.g., image and audio) and artificially produced (e.g., via adding different noise to data samples), have proven useful in enhancing representation learning. Natural views are often handled by multiview analysis tools, e.g., (deep) canonical correlation analysis [(D)CCA], while the artificial ones are frequently used in self-supervised learning (SSL) paradigms, e.g., BYOL and Barlow Twins. Both types of approaches often involve learning neural feature extractors such that the embeddings of data exhibit high cross-view correlations. Although intuitive, the effectiveness of correlation-based neural embedding is mostly empirically validated. This work aims to understand latent correlation maximization-based deep multiview learning from a latent component identification viewpoint. An intuitive generative model of multiview data is adopted, where the views are different nonlinear mixtures of shared and private components. Since the shared components are view/distortion-invariant, representing the data using such components is believed to reveal the identity of the samples effectively and robustly. Under this model, latent correlation maximization is shown to guarantee the extraction of the shared components across views (up to certain ambiguities). In addition, it is further shown that the private information in each view can be provably disentangled from the shared using proper regularization design. A finite sample analysis, which has been rare in nonlinear mixture identifiability study, is also presented. The theoretical results and newly designed regularization are tested on a series of tasks. | ['Songtao Lu', 'Weiran Wang', 'Xiao Fu', 'Qi Lyu'] | 2021-06-14 | understanding-latent-correlation-based | https://openreview.net/forum?id=5FUq05QRc5b | https://openreview.net/pdf?id=5FUq05QRc5b | iclr-2022-4 | ['multiview-learning'] | ['computer-vision'] | [ 4.48979549e-02 1.92578688e-01 -3.42172235e-01 -2.34113887e-01
-6.75739408e-01 -6.14413619e-01 7.52717853e-01 -4.62028056e-01
2.17143625e-01 4.58730757e-01 4.20381486e-01 2.70322055e-01
-4.75728214e-01 -3.85906935e-01 -7.63301790e-01 -1.13333642e+00
1.71257313e-02 1.02071553e-01 -7.43587375e-01 4.49068025e-02
-9.45989937e-02 2.05850944e-01 -1.73048043e+00 1.21056505e-01
5.81087828e-01 8.59891653e-01 3.07130385e-02 4.94851291e-01
1.06868915e-01 5.06125093e-01 -3.03497702e-01 -4.70067054e-01
4.51858371e-01 -3.17821264e-01 -1.64895609e-01 6.12629235e-01
3.91345412e-01 -3.17580789e-01 -3.58729869e-01 1.19195545e+00
1.80972517e-01 -1.78209916e-01 7.58830130e-01 -1.56688106e+00
-9.55945432e-01 4.69009340e-01 -7.06656754e-01 -2.69272536e-01
1.19219929e-01 -1.47868738e-01 1.23229122e+00 -1.21933734e+00
5.96157253e-01 1.33506465e+00 9.81311277e-02 3.98153782e-01
-1.65449226e+00 -6.65168643e-01 2.09136158e-01 9.45173874e-02
-1.22103798e+00 -5.50441325e-01 1.31251657e+00 -6.79406762e-01
1.57280773e-01 3.27632844e-01 6.29634380e-01 1.58787990e+00
1.15065284e-01 9.37903404e-01 1.39285886e+00 -3.78063321e-01
1.79208651e-01 5.96524596e-01 5.67830577e-02 4.42822635e-01
3.35500628e-01 1.30766913e-01 -6.75006747e-01 -2.64364868e-01
5.48448145e-01 2.99222410e-01 -3.36191952e-01 -1.02441740e+00
-1.15601158e+00 9.07741427e-01 -1.76365804e-02 1.88356280e-01
-2.75452673e-01 -1.85275465e-01 3.42597753e-01 4.00927961e-01
4.48859960e-01 1.63055569e-01 -3.13334882e-01 3.74868870e-01
-7.11728036e-01 -5.19170202e-02 6.43471003e-01 1.12705469e+00
7.51375079e-01 4.07699257e-01 4.20310915e-01 8.20775449e-01
6.71753526e-01 6.91557944e-01 3.89151216e-01 -1.07027841e+00
4.64786440e-01 3.65503758e-01 -1.20094351e-01 -1.34344351e+00
-4.80776206e-02 -3.89867306e-01 -1.34314895e+00 3.51454854e-01
2.50443071e-01 1.75897405e-02 -4.25568998e-01 2.13993168e+00
9.49017331e-02 1.16435833e-01 3.35175723e-01 6.18081987e-01
5.69465935e-01 3.25871408e-01 -5.22611678e-01 -6.58414245e-01
1.31034243e+00 -5.08900166e-01 -1.11565912e+00 -2.77513973e-02
1.77095056e-01 -6.23463094e-01 7.69090176e-01 5.68766594e-01
-1.03147328e+00 -5.40125906e-01 -1.31601453e+00 1.22301869e-01
-1.14436381e-01 6.78272173e-02 4.20010895e-01 7.45147943e-01
-7.25911021e-01 2.52870083e-01 -8.84888470e-01 -2.60899197e-02
4.24242556e-01 1.41345963e-01 -8.79447639e-01 -2.62139708e-01
-8.68339956e-01 4.20203686e-01 -4.32339758e-02 1.47639588e-01
-9.95773733e-01 -5.65504491e-01 -8.78836215e-01 -6.47688061e-02
4.77194965e-01 -7.96692431e-01 4.28904504e-01 -1.21370268e+00
-1.25401807e+00 8.41184974e-01 -2.68793911e-01 -4.81998846e-02
3.58050436e-01 -3.36746752e-01 -4.59978431e-01 2.71242291e-01
9.37816948e-02 7.89561570e-02 1.48653591e+00 -1.86053741e+00
2.91553270e-02 -8.77733648e-01 2.58741416e-02 1.94251299e-01
-4.66238648e-01 -1.89347729e-01 -1.00738339e-01 -6.21374190e-01
5.72075546e-01 -1.01067197e+00 4.60882187e-02 1.41128069e-02
-5.94147742e-01 3.35842580e-01 9.31729257e-01 -7.24815071e-01
7.60321319e-01 -2.28675461e+00 6.83855057e-01 2.56294459e-01
6.51019394e-01 -1.79606020e-01 -1.15779815e-02 6.00909114e-01
-5.25115848e-01 1.44850954e-01 -1.61181949e-02 -6.47760272e-01
-4.48155813e-02 2.17527926e-01 -4.44014817e-01 8.24019194e-01
4.57339548e-02 7.11649954e-01 -5.44987500e-01 -1.44862980e-01
2.96403527e-01 6.43873334e-01 -4.72557366e-01 3.46260518e-01
1.46119833e-01 7.32391238e-01 -2.36565903e-01 4.78853136e-01
7.09498107e-01 -4.41489279e-01 3.54835719e-01 -6.45387471e-01
1.36063665e-01 -3.06799471e-01 -1.47030818e+00 1.69577539e+00
-4.81793612e-01 5.92025101e-01 2.30246380e-01 -1.18879294e+00
6.99564934e-01 6.57073498e-01 6.14240646e-01 -2.50671774e-01
-5.38541153e-02 -1.42151269e-03 1.25820130e-01 -6.04242861e-01
2.79305398e-01 -2.09592104e-01 9.61911231e-02 4.78195280e-01
4.78922337e-01 2.69566566e-01 -3.12274009e-01 2.68931240e-01
4.89348978e-01 1.60075608e-03 6.32908285e-01 -1.38679847e-01
3.59640211e-01 -5.89390218e-01 4.49520558e-01 4.47791606e-01
1.05344214e-01 6.05938137e-01 5.55000722e-01 1.18679469e-02
-1.33500969e+00 -1.20169425e+00 -8.64103734e-02 5.17239809e-01
5.63517585e-02 -4.78763819e-01 -4.28693533e-01 -4.03979450e-01
-2.44098634e-01 4.29556489e-01 -5.34578741e-01 -3.03059250e-01
-1.55089378e-01 -6.09931886e-01 2.88538426e-01 3.68346542e-01
2.17228577e-01 -3.21155339e-01 -2.00610057e-01 -1.74924746e-01
-2.06684798e-01 -1.20318174e+00 -1.41670674e-01 1.20847605e-01
-8.71480703e-01 -9.69548643e-01 -5.53716779e-01 -2.19985932e-01
7.09844172e-01 6.66756570e-01 7.18899310e-01 -4.88469243e-01
-2.29738224e-02 9.90696847e-01 -1.47978544e-01 -1.06109314e-01
-3.44697267e-01 -5.87817192e-01 5.70419788e-01 6.57749116e-01
1.71920508e-01 -1.15726399e+00 -2.71986365e-01 2.74940670e-01
-1.05616343e+00 1.98823437e-01 5.16834915e-01 1.17156851e+00
4.55129474e-01 -1.15678906e-01 4.95127738e-01 -7.04537094e-01
4.20757443e-01 -5.71973264e-01 -4.38282162e-01 2.34920457e-01
-5.80815554e-01 1.88047633e-01 6.94778442e-01 -6.50253356e-01
-9.67142224e-01 -2.39820972e-01 4.45192039e-01 -1.14017463e+00
-3.57131928e-01 6.09995544e-01 -7.50394881e-01 2.26711035e-01
4.14594173e-01 2.76957422e-01 4.36196119e-01 -3.63124579e-01
6.83791399e-01 3.60609114e-01 1.96668908e-01 -4.65083838e-01
1.04246736e+00 8.87337923e-01 1.72426805e-01 -1.08641732e+00
-6.77681386e-01 -3.53992909e-01 -8.72893631e-01 -3.84594709e-01
8.52636397e-01 -1.28562343e+00 -7.19582260e-01 3.36179048e-01
-9.28117871e-01 4.08343136e-01 -1.78020939e-01 8.09564412e-01
-7.13084400e-01 4.87537861e-01 -2.95639098e-01 -1.10481906e+00
2.39290208e-01 -1.20834160e+00 9.33998883e-01 -5.71367927e-02
-2.08807573e-01 -1.12277865e+00 1.42735345e-02 5.63194931e-01
-5.08518629e-02 2.60185182e-01 9.97738779e-01 -8.26788902e-01
-7.35697508e-01 -1.58821374e-01 8.53667408e-02 7.42133796e-01
3.36270571e-01 5.32296859e-02 -1.41335177e+00 -3.77738535e-01
6.17239952e-01 -2.83347011e-01 6.19827867e-01 5.66514134e-01
9.39033329e-01 -4.83473271e-01 -7.88066462e-02 7.29881406e-01
1.37234497e+00 6.54255301e-02 4.20439690e-01 -7.56095722e-02
1.03431070e+00 1.02550042e+00 1.64939910e-01 5.62108099e-01
2.26213083e-01 5.48466444e-01 6.50043190e-01 1.68877885e-01
4.02232379e-01 -1.60398126e-01 5.08157432e-01 1.35273528e+00
-1.96937725e-01 -2.35799342e-01 -3.17102194e-01 1.75553307e-01
-1.79525363e+00 -1.25956726e+00 -3.58708501e-02 2.49011993e+00
2.94990212e-01 -2.06271544e-01 -1.25709787e-01 1.87727079e-01
6.17090821e-01 4.64965045e-01 -5.71383357e-01 5.89324646e-02
-6.03600204e-01 -1.71218008e-01 1.27103657e-01 3.83338988e-01
-1.03535819e+00 3.17143530e-01 4.99040842e+00 5.73596835e-01
-8.42736363e-01 7.65755251e-02 4.68257010e-01 -1.49521515e-01
-5.44459045e-01 1.67190596e-01 -4.78298277e-01 1.89658970e-01
5.82416177e-01 -1.62998125e-01 4.89494383e-01 7.11066842e-01
1.94940031e-01 1.32317051e-01 -1.31811965e+00 1.30107689e+00
4.86454308e-01 -1.08793986e+00 7.52428770e-02 5.75571954e-01
6.80563986e-01 -5.08535504e-01 5.51541328e-01 3.69262397e-02
1.15221381e-01 -9.09719408e-01 4.76397008e-01 7.45394707e-01
5.69762290e-01 -6.32155359e-01 4.18378830e-01 5.28431654e-01
-8.76134753e-01 -1.09537676e-01 -2.19910786e-01 1.53242707e-01
2.49125779e-01 7.47050345e-01 -9.59456787e-02 1.02480710e+00
4.82524484e-01 1.10900414e+00 -3.92723322e-01 2.64461577e-01
6.73205033e-02 4.31583881e-01 -1.27863348e-01 6.58813596e-01
-2.33405679e-01 -9.43421960e-01 9.95597422e-01 4.55472112e-01
3.77390087e-01 -1.88957483e-01 -3.33805121e-02 1.19370830e+00
1.39142886e-01 -9.34063047e-02 -1.22533405e+00 -1.03235513e-01
1.70019791e-01 1.42753947e+00 -2.10775718e-01 -2.03026012e-01
-7.02561200e-01 9.14465070e-01 1.41858906e-01 6.58795297e-01
-5.70659459e-01 4.69715327e-01 9.08675551e-01 -8.34454894e-02
1.84725598e-01 -2.93819278e-01 -1.33985236e-01 -1.60162115e+00
2.46124998e-01 -1.08338344e+00 -1.81298628e-02 -6.39068723e-01
-1.67281675e+00 2.36423045e-01 3.58031064e-01 -1.76505792e+00
-4.04669493e-01 -5.33633709e-01 -3.61248046e-01 9.57009733e-01
-9.07358408e-01 -1.38318086e+00 -9.17426124e-02 6.37176931e-01
3.02598268e-01 -5.68327725e-01 8.22868168e-01 3.21539670e-01
-5.91610074e-01 5.15142381e-01 4.19910431e-01 -3.45294289e-02
6.68750882e-01 -1.13635767e+00 -2.36280203e-01 7.09881783e-01
4.06451523e-01 1.10462427e+00 6.79319561e-01 -4.22252953e-01
-1.85477626e+00 -7.54435539e-01 1.83903784e-01 -3.33845466e-01
7.74705648e-01 -6.75676942e-01 -1.00148201e+00 8.64316583e-01
3.52963239e-01 1.42801017e-01 1.19834948e+00 1.32066637e-01
-8.68580341e-01 -1.05183513e-03 -7.36654341e-01 6.73067451e-01
8.06474924e-01 -8.65344346e-01 -3.44388425e-01 6.21480457e-02
6.76189244e-01 6.90768361e-02 -9.66320693e-01 2.73919851e-01
9.21427131e-01 -1.22907293e+00 1.21643031e+00 -8.73273432e-01
6.76853538e-01 -1.27926379e-01 -8.53902161e-01 -1.49636173e+00
-8.20228904e-02 -3.68738174e-01 -5.11944771e-01 1.40315974e+00
1.61027595e-01 -7.19210923e-01 4.14189517e-01 5.14596105e-01
1.21178798e-01 -4.97161120e-01 -8.22105229e-01 -8.60611141e-01
-8.96179155e-02 -3.95956069e-01 9.40179452e-02 1.34670055e+00
-1.41331583e-01 5.28497517e-01 -9.85841632e-01 5.22859216e-01
1.05498397e+00 1.12915955e-01 1.00399685e+00 -1.22626638e+00
-6.18981183e-01 -2.50027239e-01 -5.75601459e-01 -8.23889315e-01
3.92539918e-01 -9.28695202e-01 -5.57698965e-01 -8.72439265e-01
3.46835613e-01 -2.28611603e-02 3.34277749e-03 -3.43871087e-01
-6.04969859e-02 -3.09924603e-01 4.00372654e-01 2.51728266e-01
-1.13075674e-01 8.77250195e-01 1.23210847e+00 -1.65803626e-01
-4.77473391e-03 1.61963552e-01 -7.32768714e-01 8.46260071e-01
5.37327945e-01 -3.02575856e-01 -7.07582057e-01 3.43726054e-02
4.54745919e-01 4.59132850e-01 5.44995248e-01 -5.79552174e-01
3.76024134e-02 -3.92295159e-02 2.34290779e-01 -4.33930337e-01
8.47268522e-01 -1.12286425e+00 5.19542813e-01 5.00030234e-04
-3.95861894e-01 1.04528233e-01 -2.72349060e-01 1.13432562e+00
-4.52886879e-01 -1.46581084e-01 4.94441926e-01 -1.52197331e-01
-9.03978795e-02 3.16626370e-01 -1.29888356e-01 4.44838172e-03
8.60135615e-01 -1.38675496e-01 -2.51968324e-01 -8.01994085e-01
-9.91884291e-01 -7.46647790e-02 2.46176749e-01 2.80220032e-01
7.02373564e-01 -1.63205719e+00 -4.86054629e-01 4.67785180e-01
4.41413343e-01 -2.03979641e-01 6.33819461e-01 1.15104890e+00
4.31750715e-01 4.61017638e-02 -7.45601058e-02 -8.60237598e-01
-1.41069031e+00 9.11208510e-01 6.23862445e-02 -1.05809227e-01
-5.52237630e-01 3.87894660e-01 7.95896888e-01 -4.22340721e-01
2.46619675e-02 -3.60089391e-02 -3.14022481e-01 5.32677829e-01
5.10164320e-01 4.17782217e-01 -3.96323740e-01 -9.19549227e-01
3.78299616e-02 5.43663919e-01 -8.95691849e-03 -3.69598299e-01
1.27131295e+00 -4.14416283e-01 -2.59832054e-01 1.13914728e+00
1.64430702e+00 2.75871158e-01 -1.11196327e+00 -4.51865673e-01
-2.08074465e-01 -5.02686799e-01 -2.72319555e-01 -6.36466406e-03
-1.09709764e+00 1.15042579e+00 5.63408375e-01 3.51969302e-01
8.80019844e-01 6.67724907e-02 5.50816804e-02 1.80126593e-01
2.59513468e-01 -7.07228005e-01 3.28885555e-01 -1.99150145e-02
1.04414701e+00 -1.48008811e+00 7.72295371e-02 -3.83869171e-01
-7.49331474e-01 1.03149903e+00 3.27407807e-01 -1.63250044e-01
9.23073471e-01 1.96060747e-01 -2.27885082e-01 -3.74610662e-01
-9.21650529e-01 1.06191687e-01 4.27811414e-01 6.64061546e-01
9.36432108e-02 2.72143781e-01 2.77894199e-01 6.82368636e-01
1.49262054e-02 -6.57964885e-01 6.87915683e-01 4.87428725e-01
2.24887818e-01 -8.51519465e-01 -5.42600095e-01 5.42844176e-01
-3.79793704e-01 5.29747494e-02 -2.83948064e-01 9.23661530e-01
-2.00503126e-01 8.65210712e-01 -1.58373117e-01 -4.45536584e-01
5.95779531e-02 2.59916931e-01 4.79371279e-01 -4.37838912e-01
1.96119323e-02 5.75130761e-01 -2.56311506e-01 -4.18308556e-01
-9.42155957e-01 -8.90415847e-01 -5.22070587e-01 5.24317920e-02
-4.16105360e-01 -2.72033155e-01 5.29676676e-01 1.00877917e+00
1.06791265e-01 4.16687489e-01 9.34581876e-01 -8.60579848e-01
-6.48045540e-01 -7.69856989e-01 -1.01737845e+00 5.63828826e-01
6.08347893e-01 -8.63600492e-01 -9.27124441e-01 2.17175022e-01] | [8.350378036499023, 4.555517196655273] |
2cb6c71a-6b3a-4c3b-b0a4-9941c96395d6 | continual-mixed-language-pre-training-for | 2105.03953 | null | https://arxiv.org/abs/2105.03953v1 | https://arxiv.org/pdf/2105.03953v1.pdf | Continual Mixed-Language Pre-Training for Extremely Low-Resource Neural Machine Translation | The data scarcity in low-resource languages has become a bottleneck to building robust neural machine translation systems. Fine-tuning a multilingual pre-trained model (e.g., mBART (Liu et al., 2020)) on the translation task is a good approach for low-resource languages; however, its performance will be greatly limited when there are unseen languages in the translation pairs. In this paper, we present a continual pre-training (CPT) framework on mBART to effectively adapt it to unseen languages. We first construct noisy mixed-language text from the monolingual corpus of the target language in the translation pair to cover both the source and target languages, and then, we continue pre-training mBART to reconstruct the original monolingual text. Results show that our method can consistently improve the fine-tuning performance upon the mBART baseline, as well as other strong baselines, across all tested low-resource translation pairs containing unseen languages. Furthermore, our approach also boosts the performance on translation pairs where both languages are seen in the original mBART's pre-training. The code is available at https://github.com/zliucr/cpt-nmt. | ['Pascale Fung', 'Genta Indra Winata', 'Zihan Liu'] | 2021-05-09 | null | https://aclanthology.org/2021.findings-acl.239 | https://aclanthology.org/2021.findings-acl.239.pdf | findings-acl-2021-8 | ['low-resource-neural-machine-translation'] | ['natural-language-processing'] | [-1.40796766e-01 -3.42133760e-01 -3.50950897e-01 -2.86442846e-01
-1.50775373e+00 -8.97395134e-01 7.07930267e-01 -6.00432515e-01
-5.27596295e-01 1.14954245e+00 2.23811269e-01 -7.82567561e-01
6.91658139e-01 -4.81788486e-01 -1.00544536e+00 -4.03615773e-01
6.50281072e-01 7.56205797e-01 -2.46514007e-01 -6.01268291e-01
-4.36614782e-01 6.06988259e-02 -7.28039145e-01 4.17208612e-01
1.17264903e+00 2.20241129e-01 5.35486877e-01 3.29556227e-01
-1.32281989e-01 5.34210727e-02 -4.20279443e-01 -8.19933355e-01
5.59088171e-01 -7.54167080e-01 -8.63916218e-01 -3.35335612e-01
4.80942458e-01 -2.09283799e-01 -3.95218045e-01 1.11311543e+00
7.02577293e-01 -1.50002167e-01 2.95079738e-01 -7.73582518e-01
-1.09300888e+00 9.77240920e-01 -5.22968173e-01 3.10030073e-01
8.30338746e-02 1.70737371e-01 8.82973015e-01 -1.54234946e+00
7.41127789e-01 1.31285751e+00 5.42986155e-01 8.28731060e-01
-1.13997257e+00 -8.49615693e-01 2.35210713e-02 7.08606793e-03
-1.43633676e+00 -9.06786799e-01 3.98947567e-01 -1.58779353e-01
1.09408474e+00 2.16201425e-01 2.25778118e-01 1.59023285e+00
9.11400095e-02 8.84570658e-01 1.18487549e+00 -8.29720736e-01
-2.72931069e-01 3.07502359e-01 -3.04966062e-01 3.17160219e-01
4.84967679e-02 2.23112926e-01 -5.61800957e-01 3.05027212e-03
6.28944337e-01 -3.54593277e-01 -2.93351978e-01 1.44456625e-01
-1.69857657e+00 5.93752325e-01 1.89794287e-01 6.37931526e-01
-1.56160459e-01 -2.10877463e-01 4.81687844e-01 7.02788234e-01
7.36286819e-01 5.10003507e-01 -8.19109321e-01 -9.91844106e-03
-7.97458649e-01 -7.16353059e-02 6.29356027e-01 1.41514862e+00
8.61256778e-01 7.44103193e-02 -2.86140800e-01 1.27757525e+00
5.76171204e-02 8.63551378e-01 8.22887182e-01 -3.66865903e-01
1.09099412e+00 2.18201086e-01 2.08293390e-03 -2.27341726e-01
1.01019643e-01 -6.36723876e-01 -9.14761007e-01 -5.07981181e-01
3.31617206e-01 -4.29084688e-01 -9.72020805e-01 1.96453166e+00
2.02836335e-01 -1.59176052e-01 4.55364883e-01 9.06987250e-01
7.28148460e-01 9.37041640e-01 -1.60902217e-01 -2.41474837e-01
1.13648212e+00 -1.55381584e+00 -7.47224808e-01 -6.04242980e-01
7.06233561e-01 -1.35598886e+00 1.57396948e+00 -2.34732088e-02
-9.50775325e-01 -6.56929910e-01 -8.74722362e-01 -3.14121217e-01
-3.45571935e-01 5.00680804e-01 3.47841620e-01 3.14185947e-01
-1.16263545e+00 3.18039358e-01 -8.68117154e-01 -6.59967661e-01
3.60745676e-02 1.65270358e-01 -4.73190367e-01 -6.75068796e-01
-1.64284611e+00 1.21531618e+00 4.73323375e-01 2.96499074e-01
-1.07494736e+00 -3.41212779e-01 -7.19698012e-01 -3.89294475e-01
1.86116904e-01 -7.21013308e-01 1.41597521e+00 -1.12527311e+00
-1.53894544e+00 9.08821166e-01 -3.15271139e-01 -6.09358288e-02
6.96608007e-01 -3.71791154e-01 -6.10510349e-01 -4.70889390e-01
4.17575300e-01 6.92540824e-01 5.63999236e-01 -1.04726410e+00
-5.11219323e-01 -1.15788087e-01 -2.02391565e-01 5.80794036e-01
-3.23466659e-01 3.89600784e-01 -8.18490803e-01 -7.98949778e-01
-7.49458969e-02 -1.01092947e+00 -2.36434892e-01 -6.30352616e-01
-4.16211635e-01 -9.58127528e-02 4.22024786e-01 -9.67106402e-01
1.06966305e+00 -2.03718686e+00 2.69760191e-01 -2.75981635e-01
-3.72565418e-01 4.77366626e-01 -7.49229133e-01 6.24988317e-01
1.36839628e-01 1.52811527e-01 -2.56267458e-01 -4.86777812e-01
-1.29388750e-01 3.39819282e-01 -4.83604729e-01 2.63046443e-01
3.26269865e-01 1.22518539e+00 -9.15157735e-01 -2.64304489e-01
-1.85217485e-01 3.95719916e-01 -1.62735656e-01 3.35668445e-01
-2.66875952e-01 8.33793879e-01 -2.70024896e-01 8.11805069e-01
5.67718267e-01 8.74701515e-02 2.89589584e-01 9.78683382e-02
-1.29076228e-01 7.05284238e-01 -5.16685605e-01 2.01017594e+00
-6.87498808e-01 6.10562563e-01 -1.25056311e-01 -5.59558451e-01
9.57742274e-01 5.49338043e-01 -1.91609468e-02 -8.22954297e-01
-5.46146370e-02 8.82819414e-01 1.09996155e-01 -3.31892103e-01
4.78129655e-01 -3.35395068e-01 -2.47971609e-01 5.41249335e-01
2.19817385e-01 -1.18756860e-01 2.41459206e-01 -1.90826043e-01
7.50035763e-01 5.14493823e-01 2.42135212e-01 -3.73434722e-01
5.21664560e-01 3.23532075e-01 8.67449760e-01 4.74178553e-01
-2.19985038e-01 6.75994694e-01 -1.31774426e-01 -3.75732929e-01
-1.34841394e+00 -1.14386868e+00 -4.69277613e-02 1.33930731e+00
-1.23444028e-01 -2.51192421e-01 -7.68606424e-01 -7.72995949e-01
-3.33708346e-01 6.87006056e-01 -1.85067818e-01 -7.34062195e-02
-1.00546408e+00 -9.21523154e-01 8.41433823e-01 1.89048544e-01
5.00833929e-01 -9.55431938e-01 4.83456194e-01 3.54373515e-01
-8.18142593e-01 -1.12362814e+00 -8.53450060e-01 1.51728079e-01
-7.50439703e-01 -5.01012743e-01 -9.09057856e-01 -1.24828303e+00
6.35618865e-01 2.89127380e-01 1.36426950e+00 -2.35292390e-01
3.50507528e-01 -3.63873839e-01 -3.21378231e-01 -2.30160534e-01
-8.12935293e-01 6.51549816e-01 3.62413079e-01 -2.93186933e-01
6.88629389e-01 -3.56607318e-01 -1.41180933e-01 5.24036765e-01
-6.48925185e-01 2.68988848e-01 7.97742069e-01 1.00654662e+00
8.25529754e-01 -5.00119686e-01 6.77141607e-01 -8.70489776e-01
5.82030296e-01 -5.27045846e-01 -5.45657575e-01 6.08867764e-01
-3.60104680e-01 -7.69359171e-02 8.53342414e-01 -7.13164270e-01
-9.37776029e-01 -1.03974387e-01 -1.30381852e-01 -3.79279435e-01
5.72422221e-02 6.12213016e-01 -4.75944668e-01 2.21931189e-01
8.08173895e-01 4.22823906e-01 -5.00793934e-01 -8.04626226e-01
4.28280771e-01 1.10445774e+00 5.83571494e-01 -8.93421054e-01
9.51857030e-01 -2.01944664e-01 -7.21797526e-01 -4.26819563e-01
-8.00779819e-01 -3.36653948e-01 -8.87740612e-01 2.19270527e-01
4.42403823e-01 -1.35378635e+00 4.04952675e-01 3.32903653e-01
-1.39628077e+00 -4.40212995e-01 1.17373601e-01 7.90522695e-01
-3.85782123e-01 -3.13684121e-02 -8.19732070e-01 -1.69967681e-01
-5.62076330e-01 -1.50240445e+00 9.83556747e-01 -1.47405341e-01
-9.32056084e-02 -9.14986670e-01 3.87940377e-01 5.22610962e-01
5.44384062e-01 -2.59605676e-01 9.40808535e-01 -7.39309430e-01
-4.84113336e-01 8.76461640e-02 -1.02838442e-01 5.61166048e-01
4.46035922e-01 -2.51810163e-01 -9.03482139e-01 -6.14118993e-01
-2.42169723e-02 -4.88159537e-01 5.73478401e-01 -1.22434139e-01
3.32903713e-01 -4.71054584e-01 -1.56518906e-01 6.54952824e-01
1.36537886e+00 -3.17346342e-02 3.53512824e-01 3.68903279e-01
7.75411785e-01 4.24314648e-01 6.78950608e-01 -3.57967228e-01
4.66566354e-01 7.11946249e-01 -2.35033929e-01 -4.03687239e-01
-3.68543088e-01 -4.83738095e-01 9.43143189e-01 1.75301015e+00
1.88358426e-02 -3.22055221e-01 -1.21324289e+00 7.67477155e-01
-1.83662188e+00 -4.37945187e-01 -5.66764809e-02 2.24655747e+00
1.46900034e+00 -2.65170693e-01 -2.08960682e-01 -6.52866900e-01
9.99912024e-01 -1.50555238e-01 -5.15786767e-01 -4.18113410e-01
-4.82528687e-01 1.66348100e-01 3.48211527e-01 7.98209965e-01
-9.12156045e-01 1.77488637e+00 5.99301720e+00 9.88257051e-01
-1.22356915e+00 6.64619327e-01 5.36614656e-01 -1.65409595e-01
-4.59139675e-01 4.63744067e-02 -1.08054936e+00 3.05713207e-01
1.11828351e+00 -3.63324523e-01 8.24883699e-01 5.35974145e-01
1.60353884e-01 4.76552218e-01 -1.24856532e+00 6.64760470e-01
2.05541953e-01 -8.37754071e-01 1.89819902e-01 -4.92876880e-02
1.13757789e+00 7.81916082e-01 -3.04010278e-03 7.61781454e-01
5.04946470e-01 -8.94443333e-01 5.78259170e-01 -1.23805009e-01
1.29697168e+00 -6.86792493e-01 7.96129107e-01 6.35691941e-01
-8.06044221e-01 5.13123512e-01 -6.94805264e-01 1.36771068e-01
8.47985893e-02 4.93545234e-01 -9.70954716e-01 7.34281838e-01
5.48434734e-01 6.42274797e-01 -4.68528181e-01 6.08867288e-01
-6.43328905e-01 8.05056810e-01 -2.70748258e-01 2.44290933e-01
3.86421651e-01 -3.68714094e-01 5.36665738e-01 1.30313671e+00
5.43251574e-01 -4.09605116e-01 2.08272770e-01 6.93058610e-01
-6.24070346e-01 5.06170332e-01 -7.91124761e-01 -2.04589054e-01
4.01359528e-01 1.12175405e+00 -8.63236934e-02 -3.69104713e-01
-6.60314441e-01 1.24906468e+00 7.98662186e-01 6.63031816e-01
-6.48034632e-01 -1.26347527e-01 6.67487144e-01 -1.97157353e-01
-8.57656077e-02 -1.91841707e-01 -9.61553454e-02 -1.67982340e+00
2.16155186e-01 -1.41135168e+00 1.90023482e-01 -6.30166471e-01
-1.48044837e+00 1.16572654e+00 -3.47530037e-01 -1.36869228e+00
-4.57248062e-01 -4.31977719e-01 -2.07999125e-01 1.50398016e+00
-1.54886389e+00 -1.70486462e+00 3.53825122e-01 5.72371066e-01
8.20776880e-01 -5.01631379e-01 9.83828783e-01 6.88606918e-01
-6.50021791e-01 8.82577598e-01 5.66703320e-01 3.34096968e-01
1.30891204e+00 -8.82005870e-01 8.90593648e-01 1.29850662e+00
3.59093904e-01 8.33947837e-01 3.50471944e-01 -8.43800962e-01
-1.35603762e+00 -1.49913168e+00 1.47423983e+00 -6.56267285e-01
7.18533933e-01 -8.46269369e-01 -9.94250715e-01 1.14619231e+00
6.14107609e-01 -1.47737414e-01 6.56154454e-01 2.81453222e-01
-4.00190026e-01 2.81724930e-02 -7.79063165e-01 8.89945388e-01
9.95096803e-01 -7.61667430e-01 -6.89950526e-01 6.34702682e-01
1.00020623e+00 -6.47051096e-01 -7.29844391e-01 4.16943163e-01
3.56705457e-01 -1.71518594e-01 6.72098458e-01 -8.54449928e-01
5.27468085e-01 -2.57528096e-01 -3.59352648e-01 -1.72797954e+00
-2.80480504e-01 -8.06098521e-01 3.22047561e-01 1.45730746e+00
9.78405595e-01 -6.82144642e-01 1.97325647e-01 1.04664847e-01
-2.14327559e-01 -6.41623139e-01 -9.99546349e-01 -1.05372500e+00
7.17591405e-01 -2.22649410e-01 7.12381899e-01 1.27282965e+00
-2.27253884e-01 7.72929132e-01 -5.42205155e-01 1.15026586e-01
2.51122057e-01 9.67736021e-02 8.09503376e-01 -4.16674763e-01
-4.03494090e-01 -2.18284428e-01 3.59320074e-01 -9.69963431e-01
3.29288393e-01 -1.43465149e+00 3.31377447e-01 -1.32356167e+00
2.71332860e-01 -5.58270931e-01 -3.99422020e-01 7.37116694e-01
-5.93600810e-01 4.21748668e-01 7.53512606e-02 7.08572328e-01
-1.77644461e-01 6.56678200e-01 1.49131262e+00 -9.78526399e-02
-2.33550355e-01 -1.44497501e-02 -6.42916799e-01 3.72065336e-01
1.02604663e+00 -7.20714211e-01 -1.28571689e-01 -1.21699190e+00
4.40256819e-02 -1.15704522e-01 -2.58613586e-01 -4.95672762e-01
-2.05811989e-02 -3.25624675e-01 2.38301858e-01 -4.09897387e-01
7.20011219e-02 -5.07978737e-01 1.34437323e-01 2.12880388e-01
-2.41611347e-01 5.23743629e-01 4.87749934e-01 -2.03550383e-02
-3.88702631e-01 -5.04146665e-02 6.99122906e-01 -3.16410869e-01
-3.22750002e-01 3.79931867e-01 -1.00208841e-01 2.81580418e-01
4.92866248e-01 3.23104680e-01 -4.96720344e-01 -1.01524398e-01
-3.18591744e-01 2.93925375e-01 6.02866113e-01 8.69518280e-01
1.60930321e-01 -1.67015898e+00 -1.31566656e+00 2.63371170e-01
3.51978511e-01 -1.07490839e-02 -2.04189941e-01 7.54431725e-01
-3.68319809e-01 5.26466012e-01 -1.73636839e-01 -4.33437616e-01
-9.12176728e-01 4.05856013e-01 3.57572734e-01 -4.23101068e-01
-2.56328940e-01 8.20794106e-01 2.81383902e-01 -1.26327896e+00
-1.44337595e-01 7.53078610e-03 2.91431755e-01 -2.83968657e-01
3.81188124e-01 -5.65566495e-03 2.40298823e-01 -8.87599111e-01
-3.73588264e-01 3.92895699e-01 -4.35955524e-01 -5.19104481e-01
1.18950009e+00 -4.28386182e-01 -3.74145567e-01 5.73220134e-01
1.18231189e+00 2.85704046e-01 -6.77667975e-01 -7.76200831e-01
-1.25065399e-02 -2.83346593e-01 -1.91648006e-01 -1.18211126e+00
-8.68141711e-01 1.01951480e+00 2.30870456e-01 -6.31947815e-01
9.40260112e-01 3.17635834e-02 1.05780721e+00 5.79277873e-01
5.27755082e-01 -1.04558480e+00 -4.15599018e-01 1.01508605e+00
1.08164918e+00 -1.51404965e+00 -3.42617214e-01 -2.08990484e-01
-6.09009624e-01 8.51771235e-01 7.49260068e-01 2.49976441e-01
8.28913450e-02 1.64808735e-01 7.19491422e-01 4.12725508e-01
-8.20642948e-01 -1.26333863e-01 4.22821790e-01 4.75459665e-01
7.32496142e-01 3.56145054e-01 -3.21219832e-01 5.54175913e-01
-6.12314463e-01 -1.37055635e-01 2.68942297e-01 6.96680129e-01
-5.13320751e-02 -1.55854905e+00 -3.63761008e-01 7.71728680e-02
-6.00898147e-01 -7.61235893e-01 -6.52308285e-01 7.01717913e-01
1.47734806e-01 8.21669638e-01 -2.82689482e-01 -3.91205519e-01
3.23275924e-01 2.52289116e-01 4.22988921e-01 -9.42782342e-01
-7.17872679e-01 3.17821980e-01 1.98640987e-01 -2.78935820e-01
-5.12169150e-04 -4.57053006e-01 -7.20632255e-01 -2.35431254e-01
-2.54071355e-01 1.99857548e-01 7.21606433e-01 9.01563466e-01
3.34411740e-01 6.95758462e-02 6.33979380e-01 -5.07018447e-01
-6.70977712e-01 -1.37260258e+00 1.10364780e-01 1.85703740e-01
3.72562259e-01 -1.17291301e-01 -1.83061376e-01 8.75900984e-02] | [11.637646675109863, 10.27037525177002] |
2969ea36-6c7e-46fe-b6c4-ed0fefd3892e | topics-as-entity-clusters-entity-based-topics | 2301.02458 | null | https://arxiv.org/abs/2301.02458v1 | https://arxiv.org/pdf/2301.02458v1.pdf | Topics as Entity Clusters: Entity-based Topics from Language Models and Graph Neural Networks | Topic models aim to reveal the latent structure behind a corpus, typically conducted over a bag-of-words representation of documents. In the context of topic modeling, most vocabulary is either irrelevant for uncovering underlying topics or contains strong relationships with relevant concepts, impacting the interpretability of these topics. Furthermore, their limited expressiveness and dependency on language demand considerable computation resources. Hence, we propose a novel approach for cluster-based topic modeling that employs conceptual entities. Entities are language-agnostic representations of real-world concepts rich in relational information. To this end, we extract vector representations of entities from (i) an encyclopedic corpus using a language model; and (ii) a knowledge base using a graph neural network. We demonstrate that our approach consistently outperforms other state-of-the-art topic models across coherency metrics and find that the explicit knowledge encoded in the graph-based embeddings provides more coherent topics than the implicit knowledge encoded with the contextualized embeddings of language models. | ['Tri Kurniawan Wijaya', 'Steven Derby', 'Manuel V. Loureiro'] | 2023-01-06 | null | null | null | null | ['topic-models'] | ['natural-language-processing'] | [-4.31928426e-01 4.99324292e-01 -5.90778530e-01 -1.19426996e-01
-3.69885117e-01 -5.37950695e-01 1.10468674e+00 8.74399841e-01
-2.30383158e-01 2.21169695e-01 8.28515291e-01 -2.57694095e-01
-3.55187446e-01 -1.22331607e+00 -5.04256427e-01 -4.11570787e-01
-3.97994995e-01 5.65037489e-01 1.98451594e-01 -2.19650224e-01
2.80811518e-01 -1.63316041e-01 -1.35109961e+00 1.70759439e-01
7.58223712e-01 7.71019042e-01 3.72007042e-01 3.39511633e-02
-7.88960159e-01 7.25338995e-01 -5.17884374e-01 -3.22379380e-01
-5.53574800e-01 9.02302489e-02 -6.32551372e-01 2.21173428e-02
1.13054626e-01 1.78792089e-01 -7.95093358e-01 9.28142786e-01
-2.42139563e-01 2.54547298e-01 1.02972448e+00 -1.29291105e+00
-1.07812011e+00 1.01360822e+00 -3.43430310e-01 2.06079468e-01
4.23462577e-02 -4.75268006e-01 1.80434811e+00 -1.23172426e+00
1.02965546e+00 1.25687408e+00 4.92230624e-01 8.72979499e-03
-1.38165605e+00 -3.03439707e-01 8.25242996e-01 1.87548429e-01
-1.59142137e+00 7.35855550e-02 1.19325793e+00 -7.72315025e-01
1.10644805e+00 -1.06256939e-01 8.64014566e-01 1.27311695e+00
1.49611861e-01 5.15465379e-01 5.45305252e-01 -3.46785575e-01
4.04201418e-01 4.91095811e-01 7.46001661e-01 4.70273912e-01
8.81655991e-01 -4.31976408e-01 -8.15808356e-01 -6.05472624e-01
4.58043724e-01 4.26042914e-01 -2.34521821e-01 -9.60961044e-01
-1.14303231e+00 1.43261671e+00 3.93976837e-01 5.49484789e-01
-5.81797004e-01 1.41904861e-01 4.07201290e-01 -9.33886468e-02
1.06192791e+00 8.67935181e-01 -3.74920428e-01 1.46143749e-01
-9.47617233e-01 2.85767466e-01 8.68538320e-01 1.03397143e+00
9.48887587e-01 -1.04273476e-01 5.69596849e-02 6.72690570e-01
6.88148379e-01 1.00579314e-01 7.08940208e-01 -2.02985659e-01
4.58639562e-01 1.05089235e+00 -1.96042642e-01 -1.59371173e+00
-3.87542814e-01 -5.31038523e-01 -3.90642732e-01 -5.96767664e-01
-2.51911789e-01 -1.51273385e-02 -7.61878490e-01 1.67346096e+00
2.21925884e-01 8.19704961e-03 5.57711124e-02 5.44725299e-01
9.82506990e-01 9.53338444e-01 4.06597167e-01 -2.00239681e-02
1.73070502e+00 -7.97325373e-01 -1.01059091e+00 -3.40632886e-01
6.18492723e-01 -3.45323324e-01 9.80925739e-01 -5.11948317e-02
-5.45748830e-01 -2.12301135e-01 -1.01612234e+00 -2.11565241e-01
-9.86754298e-01 -1.05502598e-01 9.37354624e-01 3.91796052e-01
-9.21280682e-01 1.92074567e-01 -8.55313599e-01 -6.93721950e-01
3.34809184e-01 -8.33758861e-02 -3.10451418e-01 -1.65918231e-01
-1.39829826e+00 7.28096485e-01 7.43080080e-01 -4.97820467e-01
-6.44715428e-01 -8.88404131e-01 -1.20377398e+00 3.95811856e-01
6.82471156e-01 -5.45165360e-01 5.74705422e-01 -3.63804922e-02
-8.21844995e-01 4.66168612e-01 -2.93114305e-01 -4.40388024e-01
-5.15816987e-01 -3.52072120e-01 -4.90108103e-01 1.88501135e-01
1.97297186e-01 3.89550626e-01 7.63755202e-01 -1.27418339e+00
-3.84709030e-01 -3.67404759e-01 1.59471959e-01 1.32610396e-01
-1.00912929e+00 -1.75858647e-01 -5.02144217e-01 -9.32370901e-01
2.98530400e-01 -7.33470380e-01 -7.32670203e-02 -2.30956540e-01
-6.81691706e-01 -7.55351782e-01 1.06371844e+00 -4.17909533e-01
1.64269352e+00 -1.99333119e+00 3.01759839e-01 2.23937839e-01
8.97018373e-01 -2.88149506e-01 8.78268927e-02 9.27424252e-01
-6.90969266e-03 3.84372681e-01 1.16565078e-01 -3.78908157e-01
1.91221058e-01 2.01216549e-01 -9.37587857e-01 2.84737349e-01
1.60395756e-01 9.76074159e-01 -9.16225135e-01 -5.35854280e-01
4.86576408e-02 6.09906077e-01 -7.20206559e-01 3.08119655e-02
-5.34395576e-01 -3.85251790e-01 -7.90070593e-01 4.83773381e-01
1.40320227e-01 -6.00508869e-01 5.32562077e-01 -3.32577050e-01
5.57834506e-02 8.55719268e-01 -6.16893411e-01 1.88141966e+00
-3.72171670e-01 1.04143691e+00 -4.97138709e-01 -1.09456539e+00
9.24403548e-01 4.47084576e-01 4.52356905e-01 -1.42838091e-01
-1.88580066e-01 -3.04753155e-01 -2.87544101e-01 -2.58068025e-01
1.09783697e+00 -1.37876496e-01 -2.97424108e-01 7.69328356e-01
4.45390910e-01 4.45683971e-02 9.41182449e-02 9.82640445e-01
9.02676702e-01 -3.01030487e-01 3.45552772e-01 -5.24924159e-01
-2.74840534e-01 2.40370601e-01 3.42359126e-01 5.20788550e-01
3.25522035e-01 2.74395317e-01 7.45712876e-01 -1.70035630e-01
-9.50686932e-01 -1.02990210e+00 -1.92754790e-01 1.23995423e+00
2.04419401e-02 -1.29088688e+00 -1.20817907e-01 -5.43585956e-01
3.14297378e-01 1.06416130e+00 -1.09236574e+00 -2.21002683e-01
-1.99530721e-01 -5.48685849e-01 2.28338480e-01 6.08397663e-01
-4.02607352e-01 -5.91060877e-01 -1.75763756e-01 2.07243815e-01
-2.73345202e-01 -1.07643592e+00 -3.04921687e-01 1.39700830e-01
-9.60013568e-01 -1.01061141e+00 -3.33218575e-01 -4.16328043e-01
6.79588974e-01 4.10343289e-01 1.40894890e+00 -2.15379670e-01
-1.73784226e-01 6.04923189e-01 -4.69348580e-01 -4.66524869e-01
4.04323637e-02 2.25488275e-01 1.99103639e-01 -3.07394773e-01
9.43296850e-01 -6.67840779e-01 -3.86020750e-01 -1.38932601e-01
-1.06368637e+00 -2.47784093e-01 3.00019830e-01 6.85707331e-01
2.20054284e-01 1.85553536e-01 3.90834719e-01 -9.66328025e-01
9.16514814e-01 -1.08313560e+00 -3.35710198e-01 1.60530806e-01
-8.71483028e-01 3.03887218e-01 1.14970200e-01 -5.85837662e-01
-8.41412902e-01 -7.62117147e-01 8.23634028e-01 -5.22666454e-01
1.73606858e-01 1.11978686e+00 4.68701608e-02 6.72125578e-01
6.19160652e-01 2.23341465e-01 -3.93681854e-01 -5.49774408e-01
1.02969241e+00 4.09979373e-01 5.26171811e-02 -7.60527372e-01
7.39179611e-01 6.17264390e-01 -3.11253548e-01 -1.13755560e+00
-8.84827435e-01 -8.79839361e-01 -5.56578279e-01 2.72616804e-01
7.89426267e-01 -1.29501784e+00 4.11077263e-03 -3.75226408e-01
-1.28019142e+00 3.97709370e-01 -4.09039706e-01 6.74725294e-01
-1.39324382e-01 1.52018830e-01 -5.18520892e-01 -6.12191081e-01
-1.47084773e-01 -7.45343745e-01 1.13269138e+00 -2.24075466e-01
-6.80330396e-01 -1.55710983e+00 4.52218741e-01 2.43826248e-02
3.25949073e-01 9.25801620e-02 1.55137920e+00 -1.14251816e+00
-5.65653920e-01 -2.49493048e-01 -2.98163503e-01 -2.74522990e-01
3.61006051e-01 -5.98032735e-02 -9.23726559e-01 -2.32274905e-01
-2.02108279e-01 -5.42572737e-02 1.11607063e+00 2.46886596e-01
8.37530971e-01 -6.03876472e-01 -8.99272680e-01 2.17369542e-01
1.41238737e+00 -1.33603856e-01 1.91882551e-01 4.06180948e-01
8.94136369e-01 9.51861203e-01 2.53060132e-01 5.06024539e-01
8.52757335e-01 5.63928246e-01 2.44761407e-01 2.20365644e-01
2.55598485e-01 -7.80704081e-01 1.32413223e-01 1.22624743e+00
1.00293271e-01 -2.89895892e-01 -1.09454501e+00 1.12625301e+00
-1.81317186e+00 -8.24670136e-01 1.55363724e-01 1.73831189e+00
7.65348315e-01 1.32398400e-02 -1.62608892e-01 -2.55654037e-01
5.60563266e-01 6.48290992e-01 -2.96145856e-01 -4.36880663e-02
-2.97990255e-02 -1.94930121e-01 4.94950041e-02 1.56000182e-01
-1.06269443e+00 1.12903488e+00 6.01883268e+00 6.18394375e-01
-8.01533282e-01 1.93601429e-01 1.22785829e-02 -1.06797442e-01
-9.69206870e-01 2.58502632e-01 -8.06407869e-01 2.13779286e-01
9.16106284e-01 -6.79864824e-01 -2.57552117e-01 1.19221449e+00
-1.55478209e-01 2.73372561e-01 -1.24720311e+00 5.23645937e-01
3.28448117e-01 -1.62484038e+00 5.41294634e-01 5.05561292e-01
6.96938574e-01 9.13784802e-02 2.05639631e-01 4.50903744e-01
5.60195923e-01 -1.01310241e+00 5.45732319e-01 5.12570739e-01
3.87132436e-01 -5.45302749e-01 5.09123206e-01 9.69995707e-02
-1.13548398e+00 1.96632087e-01 -8.24802995e-01 1.25731543e-01
6.83627948e-02 8.75953496e-01 -1.00766170e+00 5.55186510e-01
4.71832663e-01 8.82139921e-01 -4.62944299e-01 5.31685293e-01
-2.80934691e-01 8.37699771e-01 -1.75336465e-01 -2.05499813e-01
4.28713739e-01 -1.43590987e-01 6.25971377e-01 1.33413494e+00
2.92328328e-01 -3.53259631e-02 9.46476683e-02 1.27224159e+00
-2.74356216e-01 3.19543183e-01 -1.05571067e+00 -7.93667972e-01
6.55908465e-01 1.21159828e+00 -7.22640693e-01 -4.59140003e-01
-6.77196741e-01 2.86661685e-01 5.65557539e-01 6.15893841e-01
-2.68389583e-01 -2.84931660e-01 8.97456586e-01 1.55668035e-01
6.35162055e-01 -5.44979036e-01 -1.38542682e-01 -1.47997487e+00
-9.01857615e-02 -3.08591366e-01 3.80977511e-01 -6.01254880e-01
-1.48872757e+00 6.49336874e-01 3.56977940e-01 -8.87825906e-01
-4.34340358e-01 -4.10852909e-01 -6.58553600e-01 7.60359883e-01
-1.31993783e+00 -1.24116731e+00 2.63272524e-02 3.49792063e-01
5.81512630e-01 -3.21749717e-01 9.84693229e-01 -2.84451455e-01
-3.12603176e-01 8.41284543e-03 1.87259302e-01 8.56265351e-02
5.27695060e-01 -1.34530199e+00 5.94307065e-01 6.35637522e-01
8.13346922e-01 1.45186472e+00 8.05798411e-01 -8.24045002e-01
-1.41785407e+00 -1.12394273e+00 1.21630919e+00 -8.23388875e-01
1.23843396e+00 -7.65860200e-01 -1.26171029e+00 9.71330762e-01
3.05461556e-01 -5.85876070e-02 1.31097543e+00 1.13692844e+00
-8.62530172e-01 4.44959104e-01 -3.78498137e-01 5.61770260e-01
6.86865211e-01 -1.01469195e+00 -1.34729064e+00 3.33066225e-01
1.41250050e+00 3.62062782e-01 -9.49398220e-01 -1.71259910e-01
3.12461257e-01 -2.36904994e-01 1.01094306e+00 -1.04436052e+00
6.22062862e-01 -1.90532673e-02 -4.46498662e-01 -1.50661254e+00
-3.87847155e-01 -2.58306444e-01 -6.74709618e-01 1.18880570e+00
5.18098533e-01 -4.01784331e-01 7.40824401e-01 5.33083439e-01
1.66215956e-01 -6.49113536e-01 -6.27001107e-01 -6.44708991e-01
3.40383798e-01 -6.32827163e-01 5.20271540e-01 1.44779825e+00
7.00565159e-01 6.29414618e-01 -6.19274080e-02 2.07802773e-01
6.13941193e-01 3.21282685e-01 5.06559849e-01 -1.78439009e+00
7.50185549e-02 -3.30603302e-01 -6.40946448e-01 -8.18299830e-01
3.65143806e-01 -8.96270275e-01 -3.67108285e-01 -1.78070641e+00
4.00737375e-01 -4.76614773e-01 -3.79378766e-01 2.80918211e-01
-1.95942953e-01 -3.99307102e-01 5.46994843e-02 5.56184530e-01
-7.43041396e-01 9.89173472e-01 5.67749858e-01 -2.26603970e-01
-1.11653075e-01 -5.10608912e-01 -1.05902946e+00 8.57402146e-01
4.44635570e-01 -7.07056820e-01 -7.37659574e-01 -3.63850683e-01
6.40156388e-01 -3.77094865e-01 2.44791210e-01 -3.86743844e-01
5.62259972e-01 -1.08211100e-01 1.28987171e-02 -6.74017608e-01
6.07788920e-01 -7.13520110e-01 -2.83800602e-01 -7.31450319e-02
-6.00717187e-01 1.56849176e-02 8.96503031e-02 1.13560724e+00
-4.89895225e-01 -3.28114703e-02 -2.32872844e-01 -4.60207723e-02
-7.73119748e-01 3.87183875e-01 -4.79150563e-01 2.57793456e-01
5.71232557e-01 1.17711350e-01 -5.52611291e-01 -5.36738217e-01
-5.71230888e-01 3.80942151e-02 3.95493150e-01 7.96726942e-01
6.12054944e-01 -1.37916470e+00 -3.03286493e-01 -9.04126465e-02
8.26448977e-01 -3.54586914e-02 3.41174379e-02 3.43676656e-01
2.05920607e-01 1.16617441e+00 3.44867468e-01 -4.70188260e-01
-8.62947464e-01 9.03937042e-01 -3.93800944e-01 -4.24976707e-01
-9.59056199e-01 5.94616592e-01 8.16062331e-01 -3.23011905e-01
6.26686737e-02 -5.74027658e-01 -5.68643212e-01 5.20286739e-01
2.42684931e-01 7.66798854e-02 -2.07686588e-01 -6.41407311e-01
-3.02944452e-01 3.79949808e-01 -2.81326115e-01 -2.10497499e-01
1.53331459e+00 -2.51132190e-01 -3.02954882e-01 1.03344023e+00
1.31131220e+00 -6.99537322e-02 -7.93019176e-01 -8.22644413e-01
5.56144595e-01 -2.41559669e-01 4.88126427e-01 -2.23180324e-01
-5.62155187e-01 1.00646234e+00 -8.30387697e-02 6.09808922e-01
3.23613942e-01 5.62728643e-01 3.36948454e-01 6.73304141e-01
2.48369098e-01 -1.04124880e+00 1.59018546e-01 3.55294555e-01
8.34721684e-01 -1.12333202e+00 3.43442798e-01 -6.26122475e-01
-6.63514256e-01 1.06919956e+00 2.20175788e-01 7.58383274e-02
1.12753057e+00 -2.22015604e-01 -2.84555256e-01 -1.00846708e+00
-1.25946701e+00 -1.86886758e-01 7.89392889e-01 6.69212580e-01
5.37146628e-01 2.41279945e-01 -5.05502224e-02 8.76419723e-01
-3.83974344e-01 -6.92177594e-01 4.38379854e-01 7.48131752e-01
-5.67690670e-01 -7.13453889e-01 -5.07026091e-02 3.83768350e-01
-4.40068185e-01 -4.42611009e-01 -3.59990954e-01 1.02555645e+00
-3.00099283e-01 8.94742429e-01 4.84610528e-01 -1.62138015e-01
-1.78028315e-01 3.27030689e-01 -3.35213423e-01 -1.02521253e+00
1.08983539e-01 7.60802180e-02 7.59468153e-02 -4.06809986e-01
-3.04757446e-01 -4.55887437e-01 -9.08180356e-01 6.88351765e-02
-5.12209356e-01 6.43406808e-01 1.09723830e+00 1.10080385e+00
6.20419085e-01 6.80926263e-01 2.52311736e-01 -3.09572607e-01
-1.15455039e-01 -1.20981157e+00 -8.42139125e-01 1.69129103e-01
3.05177756e-02 -9.94828403e-01 -3.23912293e-01 -2.74638813e-02] | [10.419486045837402, 7.005243301391602] |
710435d3-d4c8-49fe-97a4-24dea6c1ddc1 | mapping-instructions-to-actions-in-3d | 1809.00786 | null | http://arxiv.org/abs/1809.00786v2 | http://arxiv.org/pdf/1809.00786v2.pdf | Mapping Instructions to Actions in 3D Environments with Visual Goal Prediction | We propose to decompose instruction execution to goal prediction and action
generation. We design a model that maps raw visual observations to goals using
LINGUNET, a language-conditioned image generation network, and then generates
the actions required to complete them. Our model is trained from demonstration
only without external resources. To evaluate our approach, we introduce two
benchmarks for instruction following: LANI, a navigation task; and CHAI, where
an agent executes household instructions. Our evaluation demonstrates the
advantages of our model decomposition, and illustrates the challenges posed by
our new benchmarks. | ['Andrew Bennett', 'Dipendra Misra', 'Valts Blukis', 'Yoav Artzi', 'Max Shatkhin', 'Eyvind Niklasson'] | 2018-09-04 | mapping-instructions-to-actions-in-3d-1 | https://aclanthology.org/D18-1287 | https://aclanthology.org/D18-1287.pdf | emnlp-2018-10 | ['action-generation'] | ['computer-vision'] | [ 3.17172974e-01 3.15341085e-01 -2.41262734e-01 -4.17190582e-01
-6.39042258e-01 -3.71895462e-01 1.05071819e+00 -2.85078049e-01
-5.20766795e-01 9.01372015e-01 5.44238985e-01 -7.83772767e-01
5.22653461e-01 -6.31020546e-01 -1.15215647e+00 -3.04119557e-01
-2.53814697e-01 4.42311168e-01 1.77170392e-02 -2.58871168e-01
5.13665974e-01 2.25916088e-01 -1.80340719e+00 5.06076396e-01
7.66263604e-01 5.28612018e-01 3.58860373e-01 1.13859725e+00
1.35123640e-01 1.70945036e+00 -7.17355490e-01 3.70715946e-01
1.79706402e-02 -6.42944872e-01 -9.01801527e-01 -8.54582991e-03
2.59580344e-01 -9.20900524e-01 -4.38002020e-01 8.70431960e-01
4.78259958e-02 4.00191784e-01 7.94719636e-01 -1.62942612e+00
-6.70309842e-01 9.32722330e-01 -1.91150978e-01 -4.68613096e-02
8.95493388e-01 8.88428986e-01 8.37027788e-01 -6.34669900e-01
9.08977032e-01 1.47490287e+00 -5.33584952e-02 9.74238157e-01
-1.34541202e+00 -5.80705345e-01 4.36174423e-01 3.31583053e-01
-9.18895662e-01 -6.11486554e-01 3.68422359e-01 -4.57332909e-01
1.66400659e+00 -1.83839947e-01 4.75688577e-01 1.59583092e+00
4.72087592e-01 1.20760942e+00 1.19381189e+00 -4.38723028e-01
4.23707455e-01 -4.77655917e-01 3.15379381e-01 1.02128196e+00
-8.10583755e-02 8.37950528e-01 -7.12969422e-01 3.60805362e-01
7.76969373e-01 -7.56005198e-02 -3.86720657e-01 -4.97657359e-01
-1.51496208e+00 7.15347290e-01 4.24086422e-01 -2.03211933e-01
-3.40569168e-01 8.12940538e-01 3.13890368e-01 4.89210933e-01
-2.73695469e-01 6.19713247e-01 -5.35057411e-02 -4.41254020e-01
-4.45566863e-01 3.29481781e-01 9.16850150e-01 1.31281400e+00
7.42532969e-01 4.84452069e-01 -1.76559910e-01 -3.32182795e-02
2.56212324e-01 6.39509022e-01 6.94114983e-01 -1.52360678e+00
3.80822331e-01 3.79148245e-01 2.61540741e-01 -3.07577342e-01
-5.26468754e-01 2.22220197e-01 -1.13511737e-02 1.12653375e+00
3.32692713e-01 -2.29873240e-01 -1.09127259e+00 1.94621611e+00
-2.14350253e-01 3.53605896e-01 5.31183660e-01 7.74987042e-01
8.74524176e-01 6.99830294e-01 4.13409889e-01 1.28986284e-01
8.56171608e-01 -1.45157766e+00 -4.69332755e-01 -5.15879035e-01
9.44935203e-01 -2.15104073e-01 1.43505216e+00 4.06617731e-01
-1.01078832e+00 -7.43926287e-01 -1.26670063e+00 -8.99554405e-04
-3.48607838e-01 3.53694893e-02 6.48354709e-01 8.35865960e-02
-1.55205262e+00 6.13814771e-01 -1.21745491e+00 -3.34278941e-01
1.46835625e-01 4.50927109e-01 -3.06472301e-01 -1.21725798e-02
-5.23262501e-01 9.83924687e-01 6.79663837e-01 -3.47546101e-01
-1.97326207e+00 -2.57809669e-01 -1.50247264e+00 1.64108396e-01
4.35169905e-01 -6.73132956e-01 1.61035478e+00 -7.62329519e-01
-1.83206904e+00 7.02544749e-01 -1.36507815e-02 -6.66649640e-01
1.47981048e-01 -3.42580020e-01 -4.94051240e-02 -3.06122061e-02
1.82391718e-01 1.29338408e+00 7.62463212e-01 -1.40340793e+00
-8.55905294e-01 -1.52969822e-01 5.96479118e-01 4.29831207e-01
3.48372579e-01 -1.99862733e-01 -3.09999168e-01 -1.28022909e-01
-5.75184166e-01 -1.07485282e+00 -3.47751975e-01 -5.96659601e-01
-5.65242767e-01 -2.25194693e-01 8.01614106e-01 -4.00135845e-01
7.41990924e-01 -2.17588782e+00 4.68434930e-01 -6.65884418e-03
2.84843743e-01 -3.88702244e-01 -6.25295758e-01 3.15064579e-01
-3.92572954e-02 -7.12140575e-02 1.47623166e-01 -3.84227514e-01
3.48771542e-01 4.11645025e-01 -4.44111705e-01 -5.22450060e-02
1.26913032e-02 1.19181824e+00 -1.24000275e+00 -4.91470814e-01
4.73724008e-01 1.42910525e-01 -7.28281677e-01 7.29073465e-01
-7.61001945e-01 6.42740726e-01 -3.68297219e-01 5.23898304e-01
-2.16954529e-01 -3.28979731e-01 4.31854695e-01 3.45839858e-02
-1.72700226e-01 5.92560768e-01 -5.28788149e-01 2.10599136e+00
-6.53898835e-01 6.34599328e-01 -6.11246340e-02 -7.38255322e-01
4.39549387e-01 -2.40162928e-02 -1.78236980e-02 -9.49182034e-01
8.54869857e-02 -7.82605447e-03 7.52281174e-02 -4.67940390e-01
3.92948776e-01 4.74187672e-01 -3.67883712e-01 7.28112519e-01
2.42022738e-01 -6.36672854e-01 5.16853154e-01 3.58569115e-01
1.38830745e+00 1.02847576e+00 5.42792797e-01 -1.78782344e-01
2.28266418e-01 3.49717170e-01 1.14080403e-02 1.03375816e+00
-2.51701385e-01 5.29734008e-02 4.48780298e-01 -4.95343506e-01
-7.50356376e-01 -1.18746150e+00 8.30566049e-01 1.45465398e+00
1.22159690e-01 -4.18510944e-01 -8.76218915e-01 -1.07088459e+00
-1.97104916e-01 1.50685847e+00 -7.64269471e-01 -2.62692690e-01
-9.93479311e-01 -6.73301443e-02 1.47861689e-01 8.69333208e-01
3.10138047e-01 -1.88128018e+00 -1.31768024e+00 -6.00490533e-02
-7.06883892e-02 -1.13147974e+00 -6.60156786e-01 5.45307875e-01
-6.53225482e-01 -1.29012537e+00 -5.77940606e-02 -1.05969441e+00
7.03184903e-01 5.26473485e-02 1.75104523e+00 3.38405609e-01
-1.91595286e-01 8.19747508e-01 -1.51257545e-01 -2.94982761e-01
-8.40564907e-01 1.65548362e-02 -1.27761379e-01 -8.21333945e-01
3.74467939e-01 -4.95592296e-01 -5.19980371e-01 5.07565886e-02
-5.10147274e-01 5.75495481e-01 6.07520282e-01 9.15113389e-01
4.41359192e-01 -5.74920416e-01 1.71994716e-01 -6.96440816e-01
7.22215176e-01 -1.62701279e-01 -8.60785186e-01 1.66247770e-01
-3.95378947e-01 4.60849583e-01 7.86670089e-01 -2.07710966e-01
-1.09420526e+00 4.47371751e-01 1.01329305e-03 -1.13791443e-01
-6.84265554e-01 3.27518970e-01 -9.43224281e-02 7.46242981e-03
7.55695224e-01 3.52878720e-01 -4.10515741e-02 6.84851557e-02
8.97803009e-01 6.05069585e-02 9.61707950e-01 -8.38494480e-01
3.99614245e-01 1.65574521e-01 -1.13003001e-01 -5.25169492e-01
-5.49999714e-01 2.35137478e-01 -4.94650155e-01 -1.38648838e-01
1.25266254e+00 -8.10069919e-01 -1.33591580e+00 1.48553431e-01
-1.08863413e+00 -1.49460828e+00 -4.87357259e-01 4.67688531e-01
-1.33689761e+00 1.04619130e-01 -8.48989427e-01 -2.77602673e-01
-1.84863299e-01 -1.92623925e+00 1.04176712e+00 2.40390360e-01
-4.82543290e-01 -8.06748152e-01 1.96520194e-01 -7.09123164e-02
2.60891140e-01 3.41305435e-01 1.09995043e+00 -4.85096335e-01
-9.62456584e-01 4.58594441e-01 -1.78806722e-01 1.97317656e-02
4.09436487e-02 -2.20261276e-01 -6.11011088e-01 -3.39197606e-01
-1.78966388e-01 -1.14603710e+00 6.81136549e-01 3.45437169e-01
1.20389211e+00 -1.25966102e-01 -4.48938340e-01 6.99882805e-01
1.22167742e+00 7.91002333e-01 6.28566027e-01 4.17613089e-01
6.90922141e-01 2.96325237e-01 7.04545736e-01 1.83402941e-01
5.32378852e-01 5.12489438e-01 9.29459751e-01 9.86053273e-02
-3.57291698e-01 -4.59730864e-01 8.69905770e-01 4.94467229e-01
-1.35866120e-01 -3.16125870e-01 -8.94360125e-01 3.81447107e-01
-1.81791949e+00 -7.80384123e-01 3.98725539e-01 1.78603268e+00
7.73140073e-01 2.89266586e-01 -9.17639285e-02 -4.79412347e-01
-1.09323226e-01 1.16224699e-01 -7.37599134e-01 -4.89252061e-01
3.95055979e-01 5.38275957e-01 3.12166303e-01 1.07095647e+00
-9.90524709e-01 1.42177248e+00 8.00102234e+00 3.95029843e-01
-8.59364569e-01 -2.12571159e-01 4.33460176e-01 -1.22555301e-01
-6.74841553e-02 -2.89931800e-02 -7.73788333e-01 9.87222865e-02
1.07199275e+00 -8.52506533e-02 8.18965018e-01 1.16225207e+00
8.40910822e-02 -2.13361606e-01 -1.95685661e+00 7.48793781e-01
2.96376318e-01 -1.27479458e+00 1.23808116e-01 -9.29332152e-02
6.38406873e-01 1.22635521e-01 1.44680068e-01 1.01089966e+00
1.44104421e+00 -1.45236361e+00 6.56355679e-01 1.88717976e-01
7.41288662e-01 -5.96163154e-01 9.79409739e-02 2.67797530e-01
-9.13583279e-01 4.41837637e-03 1.29398957e-01 -3.67762834e-01
9.76725593e-02 -6.44253254e-01 -1.11897457e+00 1.14345811e-02
3.38594347e-01 6.97097242e-01 -5.30063212e-01 3.11551541e-01
-9.86531496e-01 4.48223174e-01 2.93188337e-02 -1.08405583e-01
4.54464108e-01 -1.55064464e-01 1.28708735e-01 1.03549254e+00
3.16956013e-01 6.66920701e-03 7.32020438e-01 9.95949745e-01
2.67868564e-02 -2.88893551e-01 -1.09991908e+00 -3.43876542e-03
1.71022311e-01 8.87369215e-01 -5.63885570e-01 -7.42969096e-01
-6.24207497e-01 1.26893306e+00 7.07049191e-01 6.67898476e-01
-1.04236400e+00 -2.41598517e-01 8.08850765e-01 -2.91298777e-01
1.57236338e-01 -5.14365554e-01 -7.47612938e-02 -1.15369892e+00
-5.35500288e-01 -1.38617218e+00 6.85109869e-02 -1.14030921e+00
-3.69077027e-01 5.98783672e-01 2.28799507e-01 -8.25779915e-01
-1.10209846e+00 -1.02248037e+00 -8.23859215e-01 4.16609406e-01
-1.43323874e+00 -1.00353825e+00 -6.76517904e-01 6.02787197e-01
9.84707057e-01 -2.56790847e-01 1.06080437e+00 -2.54717916e-01
-5.34095943e-01 2.52535373e-01 -4.16646719e-01 -4.90867645e-02
5.06869316e-01 -1.63503087e+00 7.80990541e-01 7.19181955e-01
4.75419201e-02 5.85524619e-01 8.74043107e-01 -6.46787524e-01
-1.46893764e+00 -8.43234956e-01 3.43611032e-01 -6.87189579e-01
7.79812336e-01 -3.65983874e-01 -3.88129920e-01 1.48995209e+00
8.88305545e-01 -4.37518030e-01 4.12331611e-01 -1.85217008e-01
-2.48138875e-01 5.67507029e-01 -6.33592010e-01 1.20864809e+00
1.30132103e+00 -3.63432795e-01 -7.33728230e-01 3.61404568e-01
9.37252104e-01 -6.14006102e-01 -3.82545263e-01 1.48796923e-02
4.32759464e-01 -1.12383842e+00 1.03509367e+00 -9.55218434e-01
8.73502076e-01 -2.99744904e-01 -2.20887035e-01 -1.76712859e+00
-1.52055025e-01 -4.10645008e-01 -2.22554922e-01 4.88782644e-01
5.81033766e-01 -3.16744506e-01 8.96790683e-01 4.14943546e-01
-5.70556998e-01 -2.89645106e-01 -3.65153849e-01 -7.40120769e-01
3.12270857e-02 -4.00949091e-01 7.47712612e-01 3.26616675e-01
4.53907669e-01 5.25001585e-01 -3.42617482e-01 -1.95884615e-01
5.91388881e-01 2.69050241e-01 1.25931430e+00 -3.96900982e-01
-7.15388894e-01 -5.88451743e-01 2.08492018e-02 -1.51438344e+00
7.68382072e-01 -7.52277493e-01 5.89349389e-01 -1.37350357e+00
2.62353241e-01 2.02291131e-01 -1.26904339e-01 8.12539101e-01
-1.35238737e-01 -7.80747905e-02 3.95951360e-01 1.10926628e-02
-1.05200434e+00 5.37925541e-01 1.19840479e+00 -3.27440858e-01
-2.46529832e-01 -5.27262866e-01 -4.52452958e-01 9.07312036e-01
8.68417144e-01 -2.81900279e-02 -7.85634398e-01 -6.91663444e-01
1.79711040e-02 2.83939749e-01 2.99166918e-01 -1.23286879e+00
1.48219559e-02 -4.33676273e-01 2.02469155e-01 -5.97900331e-01
3.07763577e-01 -5.17347574e-01 -2.42626801e-01 7.80973256e-01
-6.09552026e-01 4.30766404e-01 1.93930298e-01 2.50399321e-01
-2.64965117e-01 -9.60848480e-02 4.64099467e-01 -3.43792170e-01
-1.42713487e+00 -3.35804150e-02 -7.89861381e-01 6.85153902e-02
1.19904768e+00 1.35206103e-01 -5.51218331e-01 -7.35528886e-01
-1.00481069e+00 6.61307037e-01 9.01102662e-01 4.79266971e-01
8.70941460e-01 -1.29010940e+00 -4.89321202e-01 3.03237289e-01
2.59999901e-01 -3.78431499e-01 -1.93585515e-01 3.62962425e-01
-7.45518506e-01 5.19847393e-01 -5.49077988e-01 -4.08612579e-01
-1.20907795e+00 1.03218329e+00 8.01385716e-02 -6.23635590e-01
-6.50499582e-01 5.49382031e-01 8.20394337e-01 -3.75591308e-01
3.79704356e-01 -8.39218974e-01 -3.81848007e-01 -8.64860475e-01
5.87981224e-01 2.45936308e-02 -5.27315319e-01 -2.88481623e-01
-1.95983782e-01 8.42924044e-02 -7.90729001e-02 -2.73333013e-01
9.24191296e-01 1.07736491e-01 1.85520828e-01 2.21746042e-01
8.74934971e-01 -4.52835739e-01 -1.69865811e+00 1.55099943e-01
-1.70285590e-02 1.63604375e-02 -3.35413247e-01 -7.59885192e-01
-6.78159833e-01 1.02042139e+00 3.40744168e-01 -2.53423780e-01
8.82856131e-01 -2.33556733e-01 4.74439323e-01 8.45342457e-01
6.63564503e-01 -9.10907626e-01 6.81336284e-01 8.76292825e-01
1.09961092e+00 -1.50351107e+00 -3.28815818e-01 -9.06514842e-03
-6.37312472e-01 8.72864246e-01 1.24204421e+00 -3.18583131e-01
1.83434442e-01 6.63231850e-01 3.48896757e-02 -3.04410428e-01
-1.03844261e+00 -2.66054511e-01 -1.50942996e-01 1.00028348e+00
5.13883948e-01 2.08414584e-01 1.24748006e-01 1.35447279e-01
-2.89786935e-01 2.44770512e-01 6.31251097e-01 1.35830939e+00
-3.76498938e-01 -1.02678502e+00 -3.19952250e-01 2.30909243e-01
5.98190061e-04 -2.76318461e-01 -3.75937909e-01 1.10272288e+00
-3.17772776e-01 8.05514514e-01 -4.93093170e-02 -5.40611923e-01
2.37562671e-01 2.31928289e-01 7.70535290e-01 -8.94628048e-01
-3.38279963e-01 -2.93066949e-01 1.98732093e-01 -1.42447424e+00
-1.14757754e-01 -3.56620193e-01 -1.66944861e+00 -1.05807157e-02
3.96318465e-01 -1.35792911e-01 4.59215224e-01 6.46834493e-01
2.50645757e-01 9.72442687e-01 1.52445197e-01 -1.33304524e+00
-6.54088497e-01 -7.61006415e-01 -2.33938880e-02 7.05198228e-01
2.57814884e-01 -5.95098436e-01 -2.40893677e-01 3.57634842e-01] | [4.381521701812744, 0.8763519525527954] |
c7ba0156-33ed-4585-9ca8-89e47513ba96 | atco2-corpus-a-large-scale-dataset-for | 2211.04054 | null | https://arxiv.org/abs/2211.04054v2 | https://arxiv.org/pdf/2211.04054v2.pdf | ATCO2 corpus: A Large-Scale Dataset for Research on Automatic Speech Recognition and Natural Language Understanding of Air Traffic Control Communications | Personal assistants, automatic speech recognizers and dialogue understanding systems are becoming more critical in our interconnected digital world. A clear example is air traffic control (ATC) communications. ATC aims at guiding aircraft and controlling the airspace in a safe and optimal manner. These voice-based dialogues are carried between an air traffic controller (ATCO) and pilots via very-high frequency radio channels. In order to incorporate these novel technologies into ATC (low-resource domain), large-scale annotated datasets are required to develop the data-driven AI systems. Two examples are automatic speech recognition (ASR) and natural language understanding (NLU). In this paper, we introduce the ATCO2 corpus, a dataset that aims at fostering research on the challenging ATC field, which has lagged behind due to lack of annotated data. The ATCO2 corpus covers 1) data collection and pre-processing, 2) pseudo-annotations of speech data, and 3) extraction of ATC-related named entities. The ATCO2 corpus is split into three subsets. 1) ATCO2-test-set corpus contains 4 hours of ATC speech with manual transcripts and a subset with gold annotations for named-entity recognition (callsign, command, value). 2) The ATCO2-PL-set corpus consists of 5281 hours of unlabeled ATC data enriched with automatic transcripts from an in-domain speech recognizer, contextual information, speaker turn information, signal-to-noise ratio estimate and English language detection score per sample. Both available for purchase through ELDA at http://catalog.elra.info/en-us/repository/browse/ELRA-S0484. 3) The ATCO2-test-set-1h corpus is a one-hour subset from the original test set corpus, that we are offering for free at https://www.atco2.org/data. We expect the ATCO2 corpus will foster research on robust ASR and NLU not only in the field of ATC communications but also in the general research community. | ['Dietrich Klakow', 'Khalid Choukri', 'Petr Motlicek', 'Alexander Blatt', 'Jan Černocký', 'Allan Tart', 'Pavel Kolčárek', 'Claudia Cevenini', 'Iuliia Nigmatulina', 'Seyyed Saeed Sarfjoo', 'Amrutha Prasad', 'Mickael Rigault', 'Martin Kocour', 'Igor Szöke', 'Karel Veselý', 'Juan Zuluaga-Gomez'] | 2022-11-08 | null | null | null | null | ['dialogue-understanding'] | ['natural-language-processing'] | [ 1.64280191e-01 3.57512623e-01 4.17972952e-02 -4.63412076e-01
-1.06811249e+00 -8.88979912e-01 5.47707140e-01 -7.95706138e-02
-3.36537480e-01 6.60937309e-01 6.28359079e-01 -7.62400508e-01
-1.71904340e-02 -3.37902635e-01 3.49805132e-02 -2.74723858e-01
1.23486910e-02 7.19481826e-01 3.21443751e-02 -5.81979275e-01
-1.35012612e-01 5.05292952e-01 -1.20240152e+00 3.98885101e-01
5.81722856e-01 1.11495864e+00 2.39019856e-01 9.97089744e-01
-2.19829932e-01 5.42521119e-01 -8.27850938e-01 -4.34115529e-02
4.06818748e-01 -4.90240723e-01 -1.13828683e+00 8.73302072e-02
-3.51575017e-01 4.23347987e-02 -3.66745293e-01 8.01433563e-01
7.59926856e-01 3.37238729e-01 1.82275519e-01 -1.17402625e+00
1.40689224e-01 5.76995432e-01 1.38972431e-01 3.52346003e-01
5.74783802e-01 3.88318628e-01 1.15708113e+00 -6.28031254e-01
5.23626089e-01 1.09340334e+00 2.54573107e-01 6.17128611e-01
-7.34349310e-01 -7.40588665e-01 -1.01235002e-01 -2.32847393e-01
-1.30829358e+00 -9.11267757e-01 3.63700062e-01 -5.40789902e-01
1.10921979e+00 5.87878764e-01 3.67118806e-01 1.30584693e+00
-3.12822521e-01 7.67733037e-01 8.84629905e-01 -5.04312396e-01
2.63757586e-01 4.41887289e-01 3.15467030e-01 4.49246138e-01
-4.12189066e-01 3.12249452e-01 -6.44396305e-01 -2.13740900e-01
4.09869134e-01 -5.04502118e-01 -6.18543088e-01 3.89725089e-01
-1.34326649e+00 6.32085264e-01 -2.32874855e-01 7.31091321e-01
-4.23607975e-01 -4.73004967e-01 5.29460609e-01 6.44302487e-01
3.42385501e-01 8.21299136e-01 -8.76011550e-01 -8.47533762e-01
-5.01908660e-01 3.46977562e-02 1.29019773e+00 1.53407490e+00
3.32302779e-01 2.86280304e-01 -1.19350418e-01 1.23699820e+00
1.97842538e-01 5.19194067e-01 6.99285269e-01 -7.13773906e-01
8.62563789e-01 4.19745654e-01 1.78763896e-01 -3.01444829e-01
-7.11734653e-01 -3.32856119e-01 -4.93449271e-01 -3.04442495e-01
2.95405835e-01 -7.94465780e-01 -6.52518213e-01 1.50763142e+00
1.89655319e-01 -4.30381387e-01 4.71424103e-01 8.03046882e-01
1.00554788e+00 9.66764927e-01 -2.97436923e-01 -2.75112629e-01
1.66421008e+00 -9.21858132e-01 -1.05287337e+00 -4.40351069e-01
8.82387578e-01 -9.35384810e-01 8.71977150e-01 3.71961832e-01
-8.57187688e-01 -5.91103256e-01 -8.24718475e-01 2.80012101e-01
-6.64600730e-01 3.20187271e-01 4.00914401e-01 8.40526819e-01
-8.41543376e-01 -1.52943403e-01 -3.89996946e-01 -6.77957356e-01
-1.04954205e-02 2.84011275e-01 -4.85912532e-01 3.09745461e-01
-1.69683838e+00 7.67965794e-01 3.80206436e-01 1.28746718e-01
-7.74692476e-01 -2.65659869e-01 -8.55816245e-01 -1.52505875e-01
8.17153096e-01 -1.69313103e-01 1.78687990e+00 -7.72363186e-01
-1.75364113e+00 6.90663397e-01 -7.81531632e-02 -5.63515544e-01
1.16178244e-01 -1.91626459e-01 -1.11792183e+00 -4.13837954e-02
8.86579379e-02 4.39608246e-01 6.85035408e-01 -7.73567915e-01
-1.02001429e+00 -2.34554917e-01 -1.40878439e-01 1.66659012e-01
-6.86267018e-02 4.24443960e-01 -2.39307255e-01 -7.10692585e-01
-2.84791261e-01 -1.24668419e+00 2.37342790e-02 -7.94617653e-01
-6.44994497e-01 -2.24065587e-01 7.47477114e-01 -8.13983679e-01
1.50859642e+00 -2.29206824e+00 -1.13104485e-01 1.28948703e-01
-4.90884623e-03 6.43710017e-01 -1.51317522e-01 6.91532791e-01
-3.61431926e-01 1.90941319e-01 -1.38285771e-01 1.59251578e-02
1.31342217e-01 -1.31451160e-01 -5.01914918e-01 1.59849804e-02
2.23937809e-01 5.63342988e-01 -7.52135217e-01 1.61973219e-02
1.72993496e-01 -1.48673549e-01 -2.92504370e-01 4.89821374e-01
-7.52845109e-02 7.34270275e-01 -5.68145335e-01 5.57025254e-01
1.81726858e-01 3.68576199e-01 6.88673481e-02 -4.09605503e-02
-5.41765153e-01 6.86093450e-01 -1.08054304e+00 1.40086687e+00
-6.38156116e-01 8.59927833e-01 5.02203763e-01 -6.74625635e-01
1.15278494e+00 8.73041689e-01 4.24988627e-01 -3.24949086e-01
4.08043087e-01 3.26075971e-01 2.44126946e-01 -5.54873765e-01
7.33631194e-01 -1.72090173e-01 -4.88058150e-01 2.39476949e-01
2.53723174e-01 -4.59618330e-01 2.34334543e-01 1.02682866e-01
1.37321961e+00 -5.06500185e-01 5.33367574e-01 -8.54946375e-02
7.71777809e-01 1.70400947e-01 5.46323955e-01 4.52490866e-01
-5.92086554e-01 3.85244340e-01 3.36922318e-01 -3.84640954e-02
-8.04934740e-01 -5.77262700e-01 -2.20641479e-01 8.69750619e-01
-5.67674100e-01 -5.98233938e-01 -7.82910407e-01 -5.77478468e-01
-3.09729457e-01 1.12266743e+00 3.87615785e-02 7.02009648e-02
-3.83243710e-01 -2.09963992e-01 9.69172001e-01 3.03042322e-01
6.29593670e-01 -1.08315897e+00 -1.41584918e-01 2.71895438e-01
-5.02539158e-01 -1.70482969e+00 -7.78011084e-01 5.07687211e-01
-2.44236395e-01 -8.82490337e-01 -4.39486891e-01 -5.89068055e-01
1.49921581e-01 1.35654017e-01 7.18836069e-01 -3.93627733e-01
-1.23395272e-01 6.59821808e-01 -6.79307163e-01 -6.59641385e-01
-9.21846449e-01 3.58967751e-01 3.81965697e-01 1.62884474e-01
5.90808094e-01 -6.54793531e-02 1.60291761e-01 7.78394580e-01
-5.59500456e-01 -1.09057263e-01 4.79786843e-01 8.02929699e-01
1.07534826e-01 1.05763867e-01 8.97934377e-01 -6.55359864e-01
7.85480559e-01 -3.70450944e-01 -6.35149121e-01 1.45334797e-02
-9.76437554e-02 -2.84722656e-01 5.84558785e-01 -8.02677423e-02
-1.15573215e+00 9.63361636e-02 -5.10216177e-01 -1.65774018e-01
-7.86583066e-01 4.77319241e-01 -5.45597672e-01 4.37932491e-01
6.63937569e-01 1.06970794e-01 -2.70822626e-02 -3.45003426e-01
1.54699475e-01 1.78275919e+00 4.86919433e-01 -2.68244356e-01
7.94727087e-01 -2.14383572e-01 -6.72556579e-01 -1.62702620e+00
-8.75358582e-01 -9.43528473e-01 -7.03977823e-01 -2.23084435e-01
1.08990669e+00 -9.14406955e-01 -5.68486214e-01 2.67037004e-01
-1.18250823e+00 -4.69083816e-01 -4.05855209e-01 8.57323289e-01
-5.09660959e-01 3.05179711e-02 -2.98513889e-01 -1.11412132e+00
-2.83227980e-01 -1.19629407e+00 1.03495324e+00 -4.09839563e-02
-7.18656838e-01 -6.63395166e-01 -1.32766113e-01 9.92306113e-01
2.93235093e-01 -2.61535376e-01 4.85214740e-01 -1.38816321e+00
-9.90650803e-02 -4.78677362e-01 2.34332323e-01 6.19749129e-01
2.90117025e-01 -3.57121080e-01 -1.24104798e+00 -2.35209227e-01
1.11927673e-01 -2.56255627e-01 1.49086699e-01 -3.75449331e-03
8.01676393e-01 -3.05160195e-01 -1.33052424e-01 -4.32989299e-02
3.83918196e-01 8.83418977e-01 3.67450893e-01 -8.20603445e-02
2.51971871e-01 1.02990294e+00 9.41787004e-01 5.55820525e-01
1.13748023e-02 8.87118340e-01 -1.42896250e-01 9.99963135e-02
1.29322046e-02 -9.07393321e-02 6.92203999e-01 1.42080820e+00
1.86785355e-01 -5.54308474e-01 -1.17527986e+00 5.22416651e-01
-1.54609287e+00 -8.94482970e-01 -2.16216207e-01 2.23104000e+00
8.23246479e-01 1.96436673e-01 2.94739574e-01 7.58876428e-02
8.77227426e-01 2.33595401e-01 -1.38640940e-01 -4.57466334e-01
1.50697291e-01 -7.40323290e-02 4.54767644e-01 6.83858335e-01
-1.18404472e+00 1.10207307e+00 5.24500322e+00 9.96798515e-01
-7.35554993e-01 1.83885634e-01 4.00942683e-01 -4.32974920e-02
8.74031782e-02 -1.39358148e-01 -1.13902724e+00 3.30283910e-01
1.67925191e+00 -2.75347829e-01 6.63534343e-01 6.84759796e-01
5.26109934e-01 2.57542487e-02 -9.73175466e-01 1.00413537e+00
-1.97429731e-01 -9.88605142e-01 -3.36219728e-01 2.13024855e-01
2.12540418e-01 1.69793457e-01 -2.33265623e-01 6.53543174e-01
1.87524617e-01 -7.75352776e-01 7.47081637e-01 1.82098106e-01
9.40518796e-01 -7.36043394e-01 1.03964972e+00 4.52622741e-01
-1.33694780e+00 -2.76985228e-01 -4.00925167e-02 2.27453202e-01
4.14592803e-01 3.68070424e-01 -1.30685794e+00 6.98351204e-01
5.36482930e-01 3.18125278e-01 -8.78889412e-02 5.58270395e-01
-1.50328383e-01 9.70425725e-01 -2.31471226e-01 -2.32983857e-01
4.03170168e-01 -2.77316183e-01 1.10330832e+00 1.31087303e+00
2.08646163e-01 5.60420632e-01 2.96184033e-01 5.67348957e-01
-1.59136057e-01 1.93777502e-01 -8.31432998e-01 -7.12434709e-01
7.68638372e-01 1.27487934e+00 -4.09532785e-01 -2.38781288e-01
-4.36737776e-01 6.24661267e-01 -2.88083225e-01 3.13010812e-01
-6.70153201e-01 -8.17683458e-01 7.07053304e-01 6.86769411e-02
-3.97696579e-03 -5.21273553e-01 9.92257968e-02 -7.62069106e-01
-3.03122103e-01 -1.42188239e+00 2.27194011e-01 -6.90926611e-01
-9.94987369e-01 1.12059271e+00 6.78456351e-02 -1.44109905e+00
-4.84169006e-01 -7.27638543e-01 -2.28809103e-01 9.59195673e-01
-9.39979970e-01 -7.73177981e-01 -2.23766387e-01 3.25563669e-01
1.15910852e+00 -9.18152988e-01 9.59049284e-01 4.57111657e-01
-7.34954119e-01 3.19190681e-01 -2.87975639e-01 3.79944056e-01
7.45850861e-01 -1.07404530e+00 5.60577333e-01 4.83832687e-01
6.10934310e-02 5.68742454e-01 5.13102055e-01 -5.78804016e-01
-1.38227928e+00 -1.16581976e+00 1.02181673e+00 -5.51696241e-01
9.59873497e-01 -5.86301684e-01 -5.47786057e-01 7.30035543e-01
1.98020786e-01 -4.40577179e-01 1.01317060e+00 -2.33117361e-02
2.58023888e-01 -2.50752755e-02 -9.65785742e-01 5.47862649e-01
9.16496873e-01 -8.36139739e-01 -7.74482012e-01 5.50788581e-01
9.60100293e-01 -5.22520959e-01 -9.68650401e-01 4.26608697e-02
3.58366743e-02 -4.22613651e-01 3.78607124e-01 -6.24048173e-01
-2.75253773e-01 -3.94919544e-01 -3.72180372e-01 -1.47035635e+00
-6.78017444e-04 -1.46135449e+00 5.16808331e-01 1.47742665e+00
9.00617361e-01 -8.86145413e-01 1.25228390e-01 5.94940960e-01
-6.70958877e-01 -2.82010645e-01 -1.03919232e+00 -8.84598076e-01
-4.01152551e-01 -1.08060205e+00 6.29070282e-01 8.60287488e-01
2.67498553e-01 7.80516148e-01 -1.93424016e-01 2.45746106e-01
-1.43218592e-01 -4.31845188e-01 9.44203734e-01 -1.07532692e+00
-1.15314901e-01 -2.50162154e-01 -2.19335228e-01 -1.14437521e+00
1.57687455e-01 -8.93039584e-01 3.65192264e-01 -1.09624505e+00
-6.69553995e-01 -3.25077236e-01 2.55565375e-01 6.42659307e-01
3.85718435e-01 -4.39194649e-01 1.24030851e-01 3.55286300e-02
-2.77337134e-01 6.44447446e-01 9.17448938e-01 -9.28496271e-02
-4.70118672e-01 4.55773413e-01 -4.61489916e-01 5.58902800e-01
1.01619422e+00 -3.29825193e-01 -3.19170237e-01 1.96672991e-01
-3.47408593e-01 4.27381366e-01 -8.82286206e-02 -1.00217354e+00
3.01892906e-01 5.56235947e-02 -2.43632197e-01 -5.71966946e-01
4.75196332e-01 -8.80526125e-01 1.38635840e-02 1.17694728e-01
-4.49815422e-01 4.07964587e-02 4.10058230e-01 3.53199840e-01
-5.15828311e-01 -4.84465897e-01 6.02423668e-01 -7.58597329e-02
-7.31541634e-01 3.46563309e-02 -1.16885424e+00 2.84673184e-01
8.71744335e-01 4.08761986e-02 -3.23715597e-01 -8.33580136e-01
-7.66963005e-01 3.20498854e-01 -1.86296612e-01 8.61409903e-01
3.90478522e-01 -1.05693817e+00 -5.47456145e-01 4.75524604e-01
2.81462938e-01 -2.28581890e-01 5.62657937e-02 8.62787902e-01
-1.89145297e-01 1.06604171e+00 2.12383077e-01 -2.88416624e-01
-1.34451139e+00 1.30644485e-01 3.53663743e-01 -2.97260135e-02
-3.00081253e-01 6.36261106e-01 2.45732069e-02 -9.08299744e-01
3.67951274e-01 -3.73990327e-01 -2.46449813e-01 1.78658321e-01
6.52785838e-01 2.84626633e-01 1.63070738e-01 -9.14851427e-01
-3.84779125e-01 -1.21035594e-02 5.02659678e-02 -4.93618816e-01
1.00516760e+00 -1.88962042e-01 1.44411743e-01 7.43552208e-01
1.13886678e+00 2.65331596e-01 -4.12456870e-01 3.21066380e-02
3.39876175e-01 -1.34862378e-01 1.39863476e-01 -8.82950187e-01
-7.11516857e-01 6.83448493e-01 5.11106610e-01 5.61232448e-01
8.39256048e-01 -4.75153401e-02 9.21406388e-01 7.05462039e-01
4.25566822e-01 -1.35121143e+00 -2.45651886e-01 1.09506714e+00
1.26104414e+00 -1.18832493e+00 -4.74648684e-01 -4.86622721e-01
-1.14353883e+00 1.03407288e+00 5.35920501e-01 8.30197692e-01
5.85357845e-01 3.29638213e-01 4.03259844e-01 -2.14108974e-01
-7.81208515e-01 -5.75397551e-01 2.46174321e-01 5.31608701e-01
6.44678593e-01 1.69515714e-01 8.37472156e-02 9.39168096e-01
-6.37386978e-01 -2.95692503e-01 4.46403027e-01 9.21043277e-01
-4.59121853e-01 -1.12063777e+00 -4.83257353e-01 3.40945989e-01
-2.43008122e-01 -1.30400151e-01 -8.19646776e-01 6.89887166e-01
-2.89855570e-01 1.61734474e+00 -1.34007096e-01 -6.49432898e-01
8.31949174e-01 4.19409037e-01 -3.49404395e-01 -9.30958569e-01
-8.56198728e-01 1.62191749e-01 1.05211496e+00 -3.99411410e-01
-2.14280300e-02 -4.96358484e-01 -1.35152686e+00 -2.14790013e-02
-5.34804523e-01 7.87397623e-01 7.39890397e-01 8.30058277e-01
4.95632529e-01 8.10902655e-01 8.28090966e-01 -5.08521438e-01
-5.00808954e-01 -1.30270505e+00 -7.02095687e-01 -2.13365644e-01
1.86479926e-01 -4.34060037e-01 -4.59098488e-01 -1.16540417e-01] | [14.234077453613281, 6.902078151702881] |
abd96b13-8a4a-4366-87a4-2834eb97bb8c | spi-gcn-a-simple-permutation-invariant-graph | null | null | https://hal.archives-ouvertes.fr/hal-02093451/ | https://hal.archives-ouvertes.fr/hal-02093451/document | SPI-GCN: A Simple Permutation-Invariant Graph Convolutional Network | A wide range of machine learning problems involve handling graph-structured data. Existing machine learning approaches for graphs, however, often imply computing expensive graph similarity measures, preprocessing input graphs, or explicitly ordering graph nodes. In this work, we present a novel and simple convolutional neural network architecture for supervised learning on graphs that is provably invariant to node permutation. The proposed architecture operates directly on arbitrary graphs and performs no node sorting. It also uses a simple multi-layer perceptron for prediction as opposed to conventional convolution layers commonly used in other deep learning approaches for graphs. Despite its simplicity, our architecture is competitive with state-of-the-art graph kernels and existing graph neural networks on benchmark graph classification data sets. Our approach clearly outperforms other deep learning algorithms for graphs on multiple multiclass classification tasks. We also evaluate our approach on a real-world original application in materials science, on which we achieve extremely reasonable results. | ['Jean-Claude Crivello', 'Nataliya Sokolovska', 'Asma Atamna'] | 2019-04-08 | null | null | null | hal-archives-ouvertes-2019-4 | ['graph-similarity'] | ['graphs'] | [ 2.72612125e-01 2.82836348e-01 -1.92620769e-01 -3.22079122e-01
-1.49640515e-01 -3.51836324e-01 3.48085910e-01 8.92809510e-01
-3.44153166e-01 4.89250749e-01 -3.60501736e-01 -8.84804666e-01
-1.27756909e-01 -1.52554035e+00 -8.63356769e-01 -5.56122959e-01
-4.77340788e-01 6.82261527e-01 3.17654699e-01 -1.28271088e-01
6.64323270e-02 7.04112411e-01 -9.79956567e-01 3.31893772e-01
5.79521120e-01 1.10230052e+00 -9.15010273e-02 1.00371850e+00
-2.22014800e-01 1.01040864e+00 -2.17774108e-01 -7.71344483e-01
1.75960332e-01 1.00813732e-01 -1.07313609e+00 -1.16160758e-01
1.05067182e+00 2.06222553e-02 -7.65577555e-01 1.19291949e+00
4.26647514e-01 1.07492842e-01 6.58556998e-01 -1.30584991e+00
-1.40586650e+00 7.82322764e-01 -3.12153399e-01 1.61912367e-01
3.20704818e-01 -4.56389226e-02 1.42164302e+00 -3.88117254e-01
4.24705893e-01 1.07733309e+00 1.13735294e+00 1.58732429e-01
-1.43163717e+00 -4.35629815e-01 2.16355413e-01 3.91461700e-01
-1.09445584e+00 1.41803682e-01 9.92837965e-01 -3.96710724e-01
1.33088124e+00 1.71920598e-01 8.25155854e-01 7.29400158e-01
5.86794853e-01 5.61253190e-01 8.52951407e-01 -2.40989313e-01
2.35280991e-01 -4.78921026e-01 5.00489771e-01 1.45028079e+00
5.97769856e-01 -1.10360853e-01 4.05989066e-02 -2.50591367e-01
5.84576666e-01 4.36345823e-02 -5.85983321e-02 -6.46036506e-01
-9.03073013e-01 8.69025528e-01 1.18040073e+00 1.80225015e-01
-2.38373220e-01 8.28559399e-01 5.69334686e-01 6.30268037e-01
5.23383141e-01 4.11003023e-01 -3.86149406e-01 6.22024179e-01
-3.77953857e-01 -5.02093248e-02 1.01116645e+00 1.01583648e+00
9.39068496e-01 1.41493812e-01 -1.72930896e-01 5.56632996e-01
1.50552318e-01 5.03093526e-02 -6.46483451e-02 -1.93636164e-01
2.79534638e-01 9.32928145e-01 -6.28445089e-01 -1.22698545e+00
-7.30541706e-01 -3.43728155e-01 -1.29603732e+00 1.80996880e-01
3.95566851e-01 1.80193678e-01 -1.28874242e+00 1.25325811e+00
-7.47265220e-02 1.57952532e-01 -7.64849931e-02 4.53162313e-01
1.48230708e+00 2.88170427e-01 1.15529105e-01 3.10986727e-01
9.72193003e-01 -1.12389290e+00 -3.47128630e-01 -2.23485842e-01
8.71633351e-01 -4.35048133e-01 1.12539268e+00 2.42497757e-01
-8.47120345e-01 -3.47445816e-01 -1.05863428e+00 -3.53030384e-01
-8.36118639e-01 2.05135234e-02 1.54782367e+00 7.40938365e-01
-1.61536872e+00 1.14429379e+00 -7.08140969e-01 -5.81996024e-01
7.32785702e-01 7.31319427e-01 -6.65771008e-01 -2.15363115e-01
-8.63651693e-01 5.22057354e-01 5.83796442e-01 8.45402554e-02
-6.04574680e-01 -5.73116124e-01 -1.21971619e+00 5.41828394e-01
3.80910963e-01 -7.70550251e-01 9.98826087e-01 -8.56936097e-01
-1.31943583e+00 1.05198550e+00 3.64954978e-01 -6.52505100e-01
6.87224269e-02 2.33407125e-01 -3.44702005e-01 2.11275816e-02
-2.42196232e-01 3.81520569e-01 6.79692209e-01 -8.68258655e-01
-1.21497526e-03 -1.84584990e-01 2.32722908e-01 -2.92244732e-01
-4.71394867e-01 -2.00426012e-01 -1.77815154e-01 -5.54869354e-01
-6.12932164e-03 -9.67109323e-01 -5.91837287e-01 1.01349913e-01
-7.18501210e-01 -4.35968906e-01 6.12872005e-01 -2.28247151e-01
9.10138667e-01 -1.70389628e+00 -3.49855199e-02 4.25354421e-01
8.78478646e-01 3.09118152e-01 -3.94540608e-01 5.75845242e-01
-4.13676977e-01 9.79059041e-02 -1.87934354e-01 7.54195973e-02
4.48129587e-02 1.80696428e-01 1.47996750e-02 6.34236455e-01
3.06694895e-01 1.50033605e+00 -9.26052272e-01 -3.10491115e-01
4.15824145e-01 2.18295798e-01 -6.49700344e-01 5.87209314e-02
-4.14379358e-01 -2.28162989e-01 -1.15750976e-01 7.98410833e-01
7.00495005e-01 -1.01330602e+00 5.15241206e-01 -3.32900524e-01
5.57404518e-01 8.15170929e-02 -8.18139195e-01 1.58038104e+00
-1.98964864e-01 7.34579504e-01 -1.97168067e-01 -1.63609540e+00
9.43091512e-01 -1.30351216e-01 3.75577092e-01 -5.28648138e-01
2.81849205e-01 4.67182249e-02 3.70837778e-01 -9.83669758e-02
2.04274431e-01 1.74095407e-01 -3.25703062e-02 3.25154692e-01
4.70196009e-01 -1.92796588e-02 3.67961884e-01 3.91732335e-01
1.70245790e+00 -4.83169854e-01 3.31934005e-01 -6.35435581e-01
4.88500834e-01 -2.62122989e-01 -1.11198850e-01 8.39427114e-01
-1.27985060e-01 2.51426905e-01 7.88100123e-01 -1.22460508e+00
-7.01189876e-01 -1.14813352e+00 3.37240249e-01 1.13201392e+00
9.90048274e-02 -7.78076768e-01 -5.06644130e-01 -8.85773063e-01
3.03454459e-01 2.09632486e-01 -7.42838323e-01 -3.03615808e-01
-4.85318214e-01 -7.16831982e-01 4.28564638e-01 7.74589717e-01
2.11259276e-01 -1.15856898e+00 3.15891325e-01 2.17773765e-01
6.66442871e-01 -1.27196229e+00 -2.86768615e-01 3.28480899e-01
-7.94797719e-01 -1.62915075e+00 -1.96280107e-01 -1.40409482e+00
8.51788044e-01 3.62840623e-01 1.58271992e+00 6.45713031e-01
-4.96040970e-01 2.40181714e-01 7.32970517e-03 -2.70646751e-01
-4.42024469e-01 4.52492684e-01 -3.77961725e-01 -2.06536859e-01
3.78466994e-01 -7.19025254e-01 -3.03264081e-01 -2.56681651e-01
-6.04494572e-01 -9.45798773e-03 6.56860292e-01 9.63776231e-01
5.13601363e-01 2.79556245e-01 4.08006340e-01 -1.35464573e+00
9.06636119e-01 -3.14136416e-01 -8.52435231e-01 4.90560949e-01
-7.46236563e-01 4.25432295e-01 1.11802375e+00 -2.65722156e-01
-2.23351121e-01 -2.71396525e-02 -1.79104343e-01 -3.80641907e-01
-1.63864419e-01 7.47204959e-01 7.19714910e-02 -8.50375295e-01
5.15802562e-01 -6.87970892e-02 -1.63864605e-02 -1.31567553e-01
4.98284638e-01 7.47454241e-02 2.82699347e-01 -4.34738725e-01
7.26029575e-01 3.89867216e-01 6.19401872e-01 -7.86302328e-01
-5.52892268e-01 -2.19352573e-01 -6.74034894e-01 -9.74086374e-02
7.10081518e-01 -5.79165816e-01 -1.09971333e+00 6.95888698e-01
-1.11341214e+00 -6.13448799e-01 8.26557800e-02 2.30501667e-01
-4.28849936e-01 5.38229048e-01 -1.09922779e+00 -2.67185152e-01
-6.63170218e-01 -9.76367593e-01 1.12571132e+00 -1.06018245e-01
1.66555420e-01 -1.65419674e+00 -9.24842060e-02 7.16611696e-03
5.02746403e-01 4.55440015e-01 1.42389584e+00 -9.38742936e-01
-8.43626499e-01 -3.66271734e-01 -6.66719317e-01 1.43874735e-01
2.56948352e-01 4.53915894e-02 -5.95094442e-01 -6.33002222e-01
-9.29443598e-01 -4.79339361e-01 1.28233111e+00 3.34381670e-01
1.73801386e+00 -2.59486079e-01 -4.67296809e-01 8.25842559e-01
1.58994508e+00 -1.89867109e-01 5.12467861e-01 7.07783401e-02
1.35845542e+00 1.88142404e-01 -1.40555769e-01 -2.07301363e-01
4.83557254e-01 2.98965760e-02 7.24789679e-01 -5.53711593e-01
-1.72684878e-01 -6.33769631e-02 -5.29925935e-02 9.35090661e-01
-4.41032238e-02 -6.29024565e-01 -1.06896985e+00 3.17242652e-01
-1.89175642e+00 -7.35732019e-01 -3.68461072e-01 1.77243936e+00
2.27785945e-01 3.16762805e-01 -9.52746347e-02 4.00631130e-02
8.26264739e-01 4.05733407e-01 -3.99117768e-01 -7.45486557e-01
3.43880616e-02 1.06648409e+00 8.96541774e-01 3.19488764e-01
-1.50729287e+00 1.15683103e+00 6.92747545e+00 6.17012680e-01
-1.15224636e+00 -2.15886593e-01 5.96907437e-01 2.46957868e-01
-3.89103681e-01 -1.68382898e-01 -4.93217967e-02 6.07472379e-03
7.99619913e-01 -2.01030478e-01 5.83379090e-01 9.91529047e-01
-6.61789417e-01 4.18098360e-01 -1.50551915e+00 1.01056087e+00
2.68791597e-02 -1.84224153e+00 2.28940442e-01 5.62920459e-02
6.30861342e-01 2.83481002e-01 -1.41104221e-01 4.38623667e-01
8.05724621e-01 -1.50297999e+00 9.26122889e-02 1.54938981e-01
7.14654684e-01 -6.73656106e-01 7.01261282e-01 -3.04203331e-01
-1.55228686e+00 4.02926207e-02 -5.52414119e-01 -3.02095890e-01
-2.37267330e-01 6.71167850e-01 -9.75311935e-01 7.15215504e-01
5.69725871e-01 9.17218268e-01 -8.34067643e-01 9.51891541e-01
-1.49094090e-01 4.92701709e-01 -1.01347424e-01 -3.50508600e-01
4.39564496e-01 -1.70143515e-01 1.01404198e-01 1.23234999e+00
9.01323408e-02 -2.86266625e-01 5.20287097e-01 6.98553920e-01
-6.90394521e-01 1.84976697e-01 -1.06186187e+00 -5.27091503e-01
2.64748335e-02 1.42108786e+00 -1.07178426e+00 -2.37275079e-01
-6.06096268e-01 9.31846797e-01 1.11018801e+00 3.27233434e-01
-6.98444009e-01 -7.44744778e-01 4.73101646e-01 4.40126657e-02
3.92384112e-01 -4.56229657e-01 -1.17715552e-01 -9.87948596e-01
-1.78226829e-03 -4.34905171e-01 7.47381151e-01 -4.96977925e-01
-1.73571646e+00 6.85493350e-01 -4.53539401e-01 -5.21688044e-01
1.79559186e-01 -1.45756364e+00 -8.80679488e-01 5.02511859e-01
-1.58249533e+00 -1.46394145e+00 -4.73587036e-01 6.99528575e-01
-2.28501782e-01 -2.70682454e-01 9.18586075e-01 3.35890234e-01
-3.89165312e-01 7.60752857e-01 -1.14128485e-01 5.85282028e-01
4.18560356e-01 -1.73144114e+00 1.25420642e+00 7.06835628e-01
4.54259336e-01 3.84143531e-01 1.13116868e-01 -6.58802450e-01
-1.97084522e+00 -1.39810181e+00 5.87123692e-01 -4.27852608e-02
1.13099778e+00 -6.34140670e-01 -9.84112680e-01 9.77269650e-01
2.98560858e-01 6.22313797e-01 8.40750813e-01 5.21074116e-01
-5.23841679e-01 -2.51947343e-01 -9.60184515e-01 6.01821303e-01
1.38074183e+00 -7.44923949e-01 -2.82241628e-02 8.73815060e-01
9.08305645e-01 -3.67043108e-01 -1.15231025e+00 4.64730799e-01
2.62479335e-01 -7.28410900e-01 9.86482441e-01 -1.32781601e+00
2.46590361e-01 -5.84864728e-02 2.46831663e-02 -1.46818876e+00
-9.23056066e-01 -5.90818584e-01 -3.19306344e-01 4.70626175e-01
3.46135378e-01 -7.92995930e-01 1.26900768e+00 3.07720900e-01
-2.51742661e-01 -9.96398866e-01 -6.41041875e-01 -7.55677402e-01
-1.01927845e-02 -2.07175463e-01 6.29141688e-01 1.15541232e+00
-2.58904934e-01 5.23064971e-01 3.54742520e-02 3.58366758e-01
8.60537350e-01 5.90334296e-01 7.85879493e-01 -1.48902035e+00
-2.99382806e-01 -7.70953476e-01 -1.14783978e+00 -5.24280369e-01
6.45001531e-01 -1.66094375e+00 -4.06302541e-01 -1.95542777e+00
1.05079509e-01 -4.23593611e-01 -5.65143645e-01 7.35681534e-01
-2.03964517e-01 2.33107895e-01 -1.30480364e-01 -5.22569239e-01
-6.80931568e-01 3.39634031e-01 1.24823523e+00 -7.47376561e-01
1.38279960e-01 -1.42301679e-01 -6.32717967e-01 4.99469846e-01
1.00357533e+00 -2.20885083e-01 -4.68933851e-01 -5.31169415e-01
3.96055967e-01 -4.11921948e-01 5.61578512e-01 -1.09014642e+00
3.04670513e-01 -3.41197178e-02 3.11473936e-01 -3.77930671e-01
-1.47248685e-01 -7.38250375e-01 5.55316843e-02 7.64510930e-01
-1.73927620e-01 2.78025538e-01 2.89064884e-01 6.91046834e-01
2.55839434e-02 1.36675745e-01 8.07049394e-01 -2.89449751e-01
-7.96361327e-01 1.01011956e+00 -1.38052493e-01 -2.25860924e-02
1.05040395e+00 -1.96874887e-01 -7.33137548e-01 -1.56516641e-01
-7.62649894e-01 1.63752735e-01 5.02909541e-01 3.51346016e-01
7.05414176e-01 -1.44072461e+00 -5.26443303e-01 2.09160641e-01
3.47470939e-01 -4.03869711e-02 -1.07554175e-01 4.93085623e-01
-1.07362366e+00 2.75208801e-01 -3.31115156e-01 -5.25520086e-01
-1.25675499e+00 1.12996519e+00 3.27783495e-01 -4.98429000e-01
-6.73094034e-01 8.87384832e-01 8.92734975e-02 -7.08602905e-01
1.42861307e-01 -7.88758457e-01 2.18281940e-01 -3.77236396e-01
-1.59209073e-02 1.36213601e-01 4.42580849e-01 -2.26598695e-01
-4.68510002e-01 2.55405188e-01 -2.84959286e-01 9.61399436e-01
1.38718069e+00 6.73044086e-01 -5.61855257e-01 1.36163995e-01
1.47473049e+00 -2.61790037e-01 -5.94280243e-01 -1.52347535e-01
1.54484704e-01 -1.02614567e-01 -1.53286094e-02 -3.40300620e-01
-1.37527680e+00 9.13825750e-01 3.10535401e-01 6.87156081e-01
1.01811218e+00 1.41262516e-01 8.93139660e-01 1.13134885e+00
3.33089113e-01 -7.90907085e-01 2.41159186e-01 5.69262028e-01
7.15865731e-01 -1.50162280e+00 2.67388344e-01 -8.65222454e-01
-4.47625816e-02 1.36991251e+00 7.93813050e-01 -6.25934780e-01
9.34111655e-01 2.98140019e-01 -3.82726640e-01 -7.21715689e-01
-7.78050303e-01 -3.72481436e-01 5.24237931e-01 7.70651758e-01
5.57481766e-01 4.05697703e-01 4.77202646e-02 1.43358007e-01
-1.44151583e-01 -2.47105137e-01 4.16998446e-01 7.54386723e-01
-2.61348248e-01 -1.15163791e+00 4.58992004e-01 1.00762773e+00
-2.93487132e-01 -5.09087205e-01 -7.84806311e-01 8.23338211e-01
-3.70845973e-01 4.96327639e-01 2.14087814e-01 -5.28149009e-01
7.16204122e-02 -4.04172391e-01 8.07424545e-01 -7.08853006e-01
-5.68837941e-01 -6.98129594e-01 3.12063247e-01 -7.16344357e-01
-2.50197440e-01 6.65242150e-02 -1.11683583e+00 -7.23250329e-01
-3.85182679e-01 -1.42144188e-01 2.20827803e-01 5.22098184e-01
4.17167753e-01 8.16633880e-01 2.49333262e-01 -8.08505118e-01
-1.17307536e-01 -6.72965407e-01 -6.15191996e-01 5.51546037e-01
2.11246073e-01 -5.20652950e-01 2.13362388e-02 -4.86179918e-01] | [6.905698776245117, 6.204734802246094] |
595d83e6-aa44-4bb3-b3b2-df7d32806e2e | novel-pipeline-for-diagnosing-acute | 2307.04014 | null | https://arxiv.org/abs/2307.04014v2 | https://arxiv.org/pdf/2307.04014v2.pdf | Novel Pipeline for Diagnosing Acute Lymphoblastic Leukemia Sensitive to Related Biomarkers | Acute Lymphoblastic Leukemia (ALL) is one of the most common types of childhood blood cancer. The quick start of the treatment process is critical to saving the patient's life, and for this reason, early diagnosis of this disease is essential. Examining the blood smear images of these patients is one of the methods used by expert doctors to diagnose this disease. Deep learning-based methods have numerous applications in medical fields, as they have significantly advanced in recent years. ALL diagnosis is not an exception in this field, and several machine learning-based methods for this problem have been proposed. In previous methods, high diagnostic accuracy was reported, but our work showed that this alone is not sufficient, as it can lead to models taking shortcuts and not making meaningful decisions. This issue arises due to the small size of medical training datasets. To address this, we constrained our model to follow a pipeline inspired by experts' work. We also demonstrated that, since a judgement based on only one image is insufficient, redefining the problem as a multiple-instance learning problem is necessary for achieving a practical result. Our model is the first to provide a solution to this problem in a multiple-instance learning setup. We introduced a novel pipeline for diagnosing ALL that approximates the process used by hematologists, is sensitive to disease biomarkers, and achieves an accuracy of 96.15%, an F1-score of 94.24%, a sensitivity of 97.56%, and a specificity of 90.91% on ALL IDB 1. Our method was further evaluated on an out-of-distribution dataset, which posed a challenging test and had acceptable performance. Notably, our model was trained on a relatively small dataset, highlighting the potential for our approach to be applied to other medical datasets with limited data availability. | ['Mohammad Hossein Rohban', 'Ali Sharifi-Zarchi', 'Amirhossein Askari-Farsangi'] | 2023-07-08 | null | null | null | null | ['multiple-instance-learning', 'specificity'] | ['methodology', 'natural-language-processing'] | [ 1.41547099e-01 2.36382633e-01 -2.00108394e-01 -3.31588507e-01
-8.14591527e-01 -2.72746295e-01 3.26970309e-01 7.75470853e-01
-5.76948404e-01 6.79115236e-01 -3.06448996e-01 -4.57033575e-01
-1.22926161e-01 -1.00521123e+00 -3.61845493e-01 -8.50761294e-01
1.66076854e-01 9.42623258e-01 3.46126735e-01 1.37822777e-01
4.84802499e-02 7.28107274e-01 -1.10554159e+00 4.36191618e-01
9.32368100e-01 7.68933535e-01 1.26982152e-01 6.47458553e-01
-3.75767231e-01 8.71948838e-01 -8.15958440e-01 -5.37783146e-01
-9.26016271e-02 -3.52238595e-01 -8.77850413e-01 3.89196910e-02
7.46062025e-02 -3.47873151e-01 -8.58342573e-02 5.64717650e-01
6.77867055e-01 -2.29518697e-01 6.63892210e-01 -1.01012743e+00
-1.69737309e-01 3.55080336e-01 -5.85026503e-01 1.16652876e-01
1.71292290e-01 2.39002317e-01 5.99513769e-01 -4.55432206e-01
4.61474687e-01 6.67744219e-01 6.33724272e-01 6.90154850e-01
-1.13958406e+00 -5.60834944e-01 -1.98097721e-01 6.39869496e-02
-1.33370864e+00 -1.30643964e-01 2.26507410e-01 -5.39075375e-01
8.88525188e-01 2.58983761e-01 8.78182411e-01 7.58418381e-01
1.20094612e-01 8.80167425e-01 1.01575887e+00 -5.39896369e-01
2.84328938e-01 1.63872570e-01 -1.39580816e-02 6.55379832e-01
4.05551851e-01 -2.08680585e-01 -1.27880558e-01 -8.19343552e-02
5.69062948e-01 3.87788802e-01 -2.58278161e-01 -7.08977431e-02
-8.33046913e-01 8.99975538e-01 3.75577182e-01 6.55311108e-01
-2.60598958e-01 -1.87131688e-01 2.63232291e-01 4.99712154e-02
3.93295467e-01 4.47421074e-01 -3.98500800e-01 -8.97917897e-02
-1.06791818e+00 -1.31953761e-01 8.47280741e-01 2.81141430e-01
3.44089061e-01 -4.96659309e-01 -2.65497029e-01 7.79373944e-01
-4.21461500e-02 2.32838094e-01 4.88494128e-01 -5.43553770e-01
-7.76170269e-02 9.87045050e-01 -9.62139964e-02 -6.79970860e-01
-8.39881122e-01 -6.46294177e-01 -1.03637826e+00 -2.16435335e-04
9.32751834e-01 1.58235490e-01 -1.05724382e+00 1.48224616e+00
2.95595884e-01 1.46320969e-01 -7.75609687e-02 8.45233738e-01
8.21919501e-01 3.13742846e-01 1.75675407e-01 -1.72436684e-01
1.49418259e+00 -6.43638849e-01 -5.99015892e-01 4.48097698e-02
8.67019892e-01 -6.64723873e-01 7.58902013e-01 7.63728380e-01
-8.93926322e-01 -6.83911070e-02 -9.73715365e-01 1.06259190e-01
-4.50544715e-01 6.01498187e-02 7.82886922e-01 7.90378869e-01
-9.62713242e-01 5.80635548e-01 -9.83006001e-01 -8.06598783e-01
7.29282081e-01 5.22884905e-01 -4.32348937e-01 -4.27292675e-01
-9.21488404e-01 1.03347921e+00 3.49121928e-01 -8.74345973e-02
-5.50226033e-01 -7.62301207e-01 -4.83710915e-01 2.89216060e-02
3.81871104e-01 -7.41319001e-01 1.03549170e+00 -5.89381039e-01
-1.26083207e+00 1.19342589e+00 -7.19438642e-02 -3.21704030e-01
7.02432454e-01 1.65281102e-01 -2.96012998e-01 4.45908397e-01
-9.55094323e-02 5.74459910e-01 3.04692239e-01 -7.53400743e-01
-8.57746124e-01 -3.48480791e-01 1.91666447e-02 -2.78828442e-01
-3.11037213e-01 -4.33989428e-02 -6.50164843e-01 -4.66079205e-01
4.48796228e-02 -7.87429392e-01 -4.26619470e-01 1.14603072e-01
-1.33829698e-01 -3.28310579e-01 3.35964620e-01 -6.19680524e-01
1.15743530e+00 -1.88652682e+00 -2.99568415e-01 1.96168125e-01
3.07525307e-01 5.79809308e-01 2.24543616e-01 3.17584515e-01
1.22708902e-01 3.47896248e-01 -2.58838415e-01 -2.69696176e-01
-3.05107147e-01 1.55955046e-01 1.27627835e-01 4.12224710e-01
4.36755151e-01 8.56752515e-01 -1.11439359e+00 -5.17266393e-01
2.11472824e-01 7.83194542e-01 -3.12694430e-01 4.22476172e-01
-2.19983876e-01 4.68292147e-01 -2.01271981e-01 7.03955531e-01
6.24499619e-01 -5.64010262e-01 2.70373464e-01 2.66839787e-02
1.81997284e-01 -8.39140862e-02 -9.19480145e-01 1.47525144e+00
-2.80848116e-01 3.46315026e-01 -1.46671891e-01 -1.07573783e+00
6.96935236e-01 4.22666341e-01 6.98956549e-01 -6.91049218e-01
1.88685045e-01 4.06403124e-01 2.55378485e-01 -7.73969591e-01
-1.53331324e-01 -4.02581245e-01 3.85607302e-01 4.65986907e-01
-1.09342903e-01 -1.99282631e-01 4.44299638e-01 2.22857222e-01
1.46811533e+00 -3.21167022e-01 4.43245649e-01 -3.81342806e-02
4.83649671e-01 5.57573438e-02 6.76931500e-01 7.97249615e-01
-1.79145664e-01 9.19595122e-01 5.74057400e-01 -6.34128451e-01
-5.47402382e-01 -7.90081859e-01 -5.24490178e-01 4.77728486e-01
-2.74414986e-01 -2.15609074e-01 -5.92251599e-01 -8.79298449e-01
2.40847226e-02 5.58157086e-01 -6.33472621e-01 1.23241670e-01
-3.86519343e-01 -1.09237170e+00 5.37239134e-01 3.98382962e-01
3.09778243e-01 -7.70940363e-01 -7.92512834e-01 4.77204621e-01
-9.20454487e-02 -1.06155872e+00 9.13898647e-02 2.24091426e-01
-7.13906586e-01 -1.64400196e+00 -1.01921594e+00 -6.31889999e-01
9.06226873e-01 -7.14044720e-02 1.11474121e+00 5.86684525e-01
-6.58612490e-01 1.03959635e-01 -2.95061409e-01 -5.58028936e-01
-4.31664556e-01 9.68530700e-02 -2.43009225e-01 3.56763490e-02
7.05576539e-01 -1.85907736e-01 -8.24385405e-01 1.51450247e-01
-1.15565431e+00 1.58451065e-01 8.54225278e-01 1.05174315e+00
5.44714510e-01 1.66964144e-01 6.20377600e-01 -1.20906126e+00
2.56866753e-01 -4.21764433e-01 -3.15816641e-01 3.84173483e-01
-6.17520213e-01 -2.82464206e-01 6.71689153e-01 -2.64982998e-01
-7.97529697e-01 4.72140722e-02 -4.17789012e-01 -4.08921298e-03
-5.78156412e-01 5.14644206e-01 6.38279393e-02 1.56354383e-01
4.49657202e-01 -8.46668035e-02 1.84016913e-01 -4.22815830e-01
-3.76396686e-01 7.08843768e-01 2.21393883e-01 -3.66537035e-01
2.27205053e-01 6.94053888e-01 4.31545407e-01 -7.81738639e-01
-8.76941144e-01 -6.22815847e-01 -5.22094905e-01 -4.21698242e-02
7.38450348e-01 -6.78148031e-01 -8.09281886e-01 6.31013930e-01
-8.77469897e-01 -3.77539188e-01 -8.17792267e-02 5.04060686e-01
1.47933699e-02 3.86637390e-01 -8.62895250e-01 -5.78398645e-01
-4.21658844e-01 -1.24846745e+00 7.79031992e-01 3.88219506e-01
-4.05475378e-01 -1.23936737e+00 6.13715127e-02 3.77349377e-01
6.22071505e-01 4.90521133e-01 1.01161349e+00 -1.10941660e+00
-2.83503801e-01 -5.02812088e-01 -2.72467166e-01 1.88871045e-02
3.84579748e-01 2.33897179e-01 -1.15613151e+00 -4.81736928e-01
-5.69997914e-02 -3.38269860e-01 9.55473661e-01 3.05130541e-01
1.13593006e+00 1.06210224e-01 -5.13392270e-01 4.91019368e-01
1.35353827e+00 3.52655709e-01 4.88876581e-01 4.39153522e-01
4.24111158e-01 7.66812980e-01 5.05906522e-01 3.62232178e-01
3.89784575e-01 5.43962061e-01 3.49807382e-01 -5.37574530e-01
-1.25973299e-01 -4.81066406e-02 -2.18904480e-01 3.87120694e-01
2.44196817e-01 -3.52335244e-01 -1.39590132e+00 6.98942065e-01
-1.69324470e+00 -6.89687252e-01 -2.49358878e-01 2.06855130e+00
7.62385666e-01 1.53972298e-01 9.43656936e-02 3.74577045e-01
5.39292753e-01 -5.41858852e-01 -4.74067181e-01 -3.78720105e-01
-7.08279982e-02 5.68944633e-01 5.17560840e-02 1.68028727e-01
-9.87307549e-01 5.07231176e-01 5.85266733e+00 8.42154801e-01
-1.49490547e+00 -7.70969838e-02 1.05257785e+00 -4.95921038e-02
-5.04666418e-02 -1.60775855e-01 -5.65613866e-01 6.32977247e-01
6.59106314e-01 1.34131879e-01 -1.84959471e-01 4.39118624e-01
1.58711299e-01 -4.69159633e-01 -1.35323405e+00 8.30784917e-01
5.95438220e-02 -1.25440025e+00 -2.30574131e-01 1.77071746e-02
4.71546412e-01 -2.74002463e-01 -4.65765670e-02 1.84168354e-01
1.58560425e-01 -1.40752268e+00 -5.92704266e-02 2.53770828e-01
8.68990660e-01 -6.98469818e-01 1.22294724e+00 5.92842519e-01
-6.68138921e-01 1.79122478e-01 -1.85308039e-01 -4.61286603e-04
4.03339826e-02 1.09224617e+00 -1.44739008e+00 4.24681485e-01
6.42589450e-01 4.69724119e-01 -6.14025831e-01 1.31482327e+00
-3.61650318e-01 6.39048398e-01 -5.54744482e-01 -1.99415833e-02
-5.30437231e-02 1.65522918e-01 -1.30928457e-01 1.33136261e+00
3.71879905e-01 2.98999310e-01 1.07049823e-01 3.60211819e-01
1.03425853e-01 2.54879564e-01 -2.41564676e-01 8.62999558e-02
2.15453744e-01 1.42691326e+00 -1.09142768e+00 -2.37396985e-01
-4.57836330e-01 5.58226466e-01 4.38090026e-01 1.26268312e-01
-6.26872540e-01 -3.54080409e-01 9.77429897e-02 2.15552568e-01
5.05861305e-02 5.66881657e-01 -3.23883742e-01 -1.01176643e+00
-2.52950311e-01 -7.72694767e-01 9.05446589e-01 -1.91150472e-01
-1.25226223e+00 3.31038803e-01 -3.12631547e-01 -1.06980503e+00
-1.44416913e-01 -6.19582713e-01 -6.61673009e-01 6.35869086e-01
-1.73909986e+00 -1.15006077e+00 -3.49092156e-01 1.63844079e-01
1.40198886e-01 2.30133142e-02 1.02467537e+00 3.62821996e-01
-8.72030199e-01 8.10337126e-01 -1.66742519e-01 2.33925924e-01
8.74428928e-01 -1.33948636e+00 -1.67050362e-01 7.01094091e-01
-2.41595700e-01 5.82488835e-01 5.24874091e-01 -3.02963018e-01
-1.15326118e+00 -8.41011167e-01 1.03658736e+00 -2.65957564e-01
3.57639700e-01 1.42959561e-02 -1.17484891e+00 3.36577386e-01
2.16164738e-02 2.76461523e-02 1.25444937e+00 -7.45461956e-02
1.56608094e-02 -1.05110042e-01 -1.68589151e+00 3.76690954e-01
5.68444610e-01 -1.22847207e-01 -2.12334394e-01 2.96359539e-01
1.46057621e-01 -5.77555954e-01 -1.14263964e+00 6.32974505e-01
5.25163114e-01 -1.17120469e+00 5.96813321e-01 -3.50606710e-01
4.42571908e-01 -2.09315389e-01 2.81629413e-01 -1.21806824e+00
-2.32195944e-01 -2.39103153e-01 -1.66998968e-01 1.37162602e+00
3.72073352e-01 -7.96278417e-01 1.07856345e+00 8.46807182e-01
1.49591520e-01 -1.27174485e+00 -7.01726377e-01 -4.02522057e-01
2.27666482e-01 -3.10040981e-01 7.46223152e-01 1.06843579e+00
7.46744052e-02 -3.50359939e-02 1.18987955e-01 5.94595820e-02
4.92383391e-01 1.14979655e-01 7.12360501e-01 -1.21093094e+00
-4.04456288e-01 -5.93203187e-01 -3.69698465e-01 -5.93119860e-01
-1.05977781e-01 -8.70248199e-01 -3.55884403e-01 -1.83006024e+00
6.30116165e-01 -7.81398356e-01 -5.09944916e-01 7.23260224e-01
-4.02143568e-01 3.15895408e-01 5.42720072e-02 8.60806406e-02
-2.19925731e-01 -1.74419150e-01 1.02671504e+00 -3.12553316e-01
9.01008546e-02 1.75688446e-01 -7.21497059e-01 6.68024063e-01
8.98313761e-01 -4.46818143e-01 -1.13390274e-01 -3.04062366e-01
8.95883441e-02 8.50196928e-02 1.09503642e-02 -7.86268532e-01
4.56119001e-01 -1.63715482e-01 6.11151278e-01 -5.64443529e-01
2.84025043e-01 -8.30742955e-01 3.14881623e-01 9.70514297e-01
-9.03530717e-02 -3.50557894e-01 1.24035574e-01 5.12432456e-02
-2.92895883e-01 -3.10412377e-01 8.59586954e-01 -1.68332994e-01
-5.51963627e-01 2.87071258e-01 -2.26270184e-01 1.81197505e-02
1.21161330e+00 -1.60846904e-01 -5.00730455e-01 -1.43219218e-01
-4.60244954e-01 4.70675319e-01 7.41088688e-01 -1.10464729e-02
5.85344195e-01 -7.51766145e-01 -8.59064937e-01 2.09388733e-01
2.08108425e-01 4.12863642e-01 3.04699719e-01 1.10735154e+00
-9.59431171e-01 5.54863632e-01 2.56528035e-02 -8.74952853e-01
-1.29022312e+00 4.90371794e-01 3.25288475e-01 -7.66007543e-01
-5.22932470e-01 8.57523739e-01 8.00817609e-02 -3.08176696e-01
3.62443447e-01 -2.24638328e-01 -3.78135353e-01 1.66232571e-01
7.04898119e-01 2.71474928e-01 4.62605923e-01 -3.04108143e-01
-4.86988544e-01 5.07438421e-01 -4.74574149e-01 2.58711785e-01
1.48689806e+00 4.39847827e-01 -1.35889724e-01 1.38121724e-01
9.17542696e-01 -6.17376296e-03 -8.47166300e-01 -3.78977954e-02
1.20261639e-01 -5.02191842e-01 -1.00872211e-01 -1.00698245e+00
-1.18979359e+00 1.14836907e+00 5.28914392e-01 2.09229290e-01
1.29422045e+00 -6.89255595e-02 7.60286212e-01 9.23189744e-02
3.64559710e-01 -8.52286339e-01 7.46263610e-03 2.66570866e-01
2.88885206e-01 -1.61828887e+00 2.40161389e-01 -3.57223898e-01
-4.13896143e-01 1.25889230e+00 6.12997234e-01 3.96895468e-01
4.66668427e-01 4.20897722e-01 3.58812481e-01 -1.08913340e-01
-8.96636009e-01 -2.28262007e-01 -6.81976005e-02 5.21772087e-01
7.10990965e-01 3.23736109e-02 -3.92211020e-01 6.40276372e-01
1.27459437e-01 3.33866656e-01 4.06233490e-01 1.08329499e+00
-2.61018097e-01 -1.31461680e+00 -3.04244727e-01 6.36893094e-01
-1.06186974e+00 2.54682183e-01 -2.85412282e-01 8.00150514e-01
2.42712706e-01 1.05565941e+00 4.18782188e-03 -2.83701625e-02
1.93771675e-01 -3.64953391e-02 5.17886102e-01 -6.48432612e-01
-7.15275705e-01 3.32032777e-02 -1.99730530e-01 -2.28596002e-01
-3.68953258e-01 -4.49821085e-01 -1.43292570e+00 -3.39144915e-01
-2.56933421e-01 2.18692467e-01 5.52621424e-01 1.00165164e+00
1.20141000e-01 5.92943192e-01 5.27988851e-01 -2.30739787e-01
-2.27831334e-01 -7.37503409e-01 -4.68173355e-01 6.99841499e-01
2.15251833e-01 -3.73916090e-01 -3.54130983e-01 -2.54994929e-01] | [15.0772066116333, -2.9741296768188477] |
04c2c551-138a-4a6b-b615-2e76eb0c8b83 | advance-prediction-of-ventricular | 1811.12938 | null | http://arxiv.org/abs/1811.12938v1 | http://arxiv.org/pdf/1811.12938v1.pdf | Advance Prediction of Ventricular Tachyarrhythmias using Patient Metadata and Multi-Task Networks | We describe a novel neural network architecture for the prediction of
ventricular tachyarrhythmias. The model receives input features that capture
the change in RR intervals and ectopic beats, along with features based on
heart rate variability and frequency analysis. Patient age is also included as
a trainable embedding, while the whole network is optimized with multi-task
objectives. Each of these modifications provides a consistent improvement to
the model performance, achieving 74.02% prediction accuracy and 77.22%
specificity 60 seconds in advance of the episode. | ['Marek Sirendi', 'Marek Rei', 'Joshua Oppenheimer'] | 2018-11-30 | null | null | null | null | ['heart-rate-variability'] | ['medical'] | [-4.55415025e-02 2.46959507e-01 -3.29941899e-01 -5.29723287e-01
-6.25521779e-01 -3.52386117e-01 -4.76616845e-02 2.27930501e-01
-5.04466832e-01 9.73461688e-01 1.62091419e-01 -5.66874504e-01
-2.76930243e-01 -4.88414645e-01 -1.42329544e-01 -6.19496703e-01
-6.79747045e-01 5.68019509e-01 -5.69240391e-01 1.47494629e-01
1.72988269e-02 4.31333005e-01 -5.54658294e-01 1.95916906e-01
6.93527102e-01 1.13657975e+00 -4.00827944e-01 1.08415723e+00
3.14262331e-01 7.44623840e-01 -7.80345440e-01 4.09038514e-02
8.71620253e-02 -4.23627675e-01 -5.05191982e-01 -7.56348550e-01
-1.41842943e-02 -3.06189299e-01 -3.20622385e-01 1.46945551e-01
1.33531904e+00 -3.97435904e-01 4.93495792e-01 -8.40690255e-01
-2.82524586e-01 7.97841728e-01 -8.01166669e-02 8.99728477e-01
-1.06183469e-01 9.10879001e-02 9.03214455e-01 -7.82605052e-01
1.74670875e-01 4.48407948e-01 1.14896679e+00 5.78800678e-01
-1.20390522e+00 -5.13115942e-01 -4.51036483e-01 -8.05580392e-02
-1.28417611e+00 -2.40037099e-01 8.99010241e-01 -4.91765797e-01
1.06789863e+00 3.31902146e-01 1.03717995e+00 8.84927094e-01
7.15877771e-01 3.06449383e-01 4.84066904e-01 -2.88362563e-01
-1.36470407e-01 1.37565181e-01 4.09865528e-01 4.94845599e-01
2.84365684e-01 4.16619420e-01 -3.12668502e-01 -6.10298991e-01
8.80613923e-01 1.33619592e-01 -3.34068030e-01 3.90915722e-02
-1.27796209e+00 7.03027606e-01 9.75926146e-02 3.96292478e-01
-5.91117859e-01 2.48929888e-01 7.53457725e-01 4.46676284e-01
4.20686513e-01 8.62393796e-01 -9.20602143e-01 -4.85710591e-01
-8.00148368e-01 2.08752528e-01 7.59014845e-01 1.53718725e-01
3.21573913e-01 7.15666831e-01 -4.32343453e-01 7.53422141e-01
1.77939400e-01 6.22619987e-01 5.21591127e-01 -8.71636748e-01
4.38757688e-01 3.78343374e-01 1.04441447e-02 -7.91502714e-01
-1.01382458e+00 -9.49602067e-01 -1.09626162e+00 -1.36565730e-01
1.32326573e-01 -5.78239977e-01 -9.97441113e-01 1.71400583e+00
-2.02453718e-01 4.41492170e-01 2.21838832e-01 6.17530704e-01
9.17153120e-01 4.12788391e-01 1.79868758e-01 -4.75840122e-01
1.04835212e+00 -6.25042498e-01 -8.22546661e-01 -1.27675146e-01
7.76160598e-01 -5.70329353e-02 3.68693590e-01 1.99385479e-01
-1.13526678e+00 -4.92191792e-01 -8.90798032e-01 3.02756608e-01
-5.42054921e-02 2.97782004e-01 5.69521785e-01 7.81808674e-01
-9.67967749e-01 9.03225064e-01 -8.04507971e-01 2.41232678e-01
2.99183905e-01 6.56277835e-01 1.01034306e-02 5.93905449e-01
-1.75565350e+00 1.11042643e+00 3.60046744e-01 2.78081417e-01
-4.64535028e-01 -1.30802751e+00 -8.75106871e-01 2.35562593e-01
-5.37223518e-01 -1.07621634e+00 8.12599063e-01 -4.63772386e-01
-1.42683840e+00 1.04304433e+00 1.63836516e-02 -8.06257844e-01
4.32927281e-01 -2.14986295e-01 -5.82594633e-01 4.74475212e-02
-1.58659667e-01 9.49539319e-02 7.32197165e-01 -3.82324636e-01
-2.80607730e-01 -3.65911305e-01 -4.14366663e-01 1.78632513e-01
-1.70596376e-01 -1.21646091e-01 -2.20324859e-01 -7.90739179e-01
-2.11765364e-01 -9.61706519e-01 -5.20374656e-01 -5.25455832e-01
-1.42802820e-01 8.05510655e-02 4.88447398e-01 -1.03432763e+00
1.83061707e+00 -2.19952655e+00 1.08718783e-01 3.41670096e-01
6.89477563e-01 5.62752187e-01 1.28016293e-01 1.70990616e-01
-4.82668340e-01 4.45965081e-01 -1.44027650e-01 -1.56200230e-02
-4.35191810e-01 3.01614497e-02 -2.72327065e-01 2.79652923e-01
1.53481841e-01 1.35649276e+00 -5.75967073e-01 -7.15563074e-02
4.41522926e-01 6.82450235e-01 -1.57666728e-01 1.47409067e-01
4.88373280e-01 4.54208225e-01 -4.25343037e-01 4.00167495e-01
2.88636267e-01 -4.97963965e-01 2.26063490e-01 -6.82991892e-02
2.28776902e-01 3.57323587e-01 -6.05872691e-01 1.43263137e+00
-3.67348820e-01 4.17459071e-01 -5.28475225e-01 -9.86215234e-01
1.06128883e+00 8.23695779e-01 9.74481106e-01 -5.43554902e-01
1.68577701e-01 2.05148056e-01 3.99830580e-01 -3.97272736e-01
-3.96286070e-01 -1.05977692e-01 -6.73702285e-02 5.17936289e-01
-1.69280842e-01 8.37029517e-02 -2.90578663e-01 -2.33431026e-01
1.32242596e+00 -3.20931822e-01 6.51866138e-01 -3.50154281e-01
3.25022787e-01 -4.82387871e-01 9.22042012e-01 1.00877202e+00
-2.73881733e-01 6.80655479e-01 4.84767705e-01 -1.49782598e+00
-9.77017760e-01 -1.20664620e+00 -4.73141730e-01 2.54188210e-01
-4.02433723e-01 -4.50192302e-01 -6.17975965e-02 -4.99796182e-01
2.31505424e-01 3.01122069e-01 -8.83518875e-01 -5.03140986e-01
-9.00768399e-01 -1.15077269e+00 1.00855505e+00 9.84166563e-01
-2.60004252e-02 -1.34516358e+00 -1.00663364e+00 6.80628479e-01
-1.16418384e-01 -7.58377612e-01 -1.06183194e-01 4.87611055e-01
-1.26915646e+00 -1.02323341e+00 -8.02585423e-01 -5.79747260e-01
9.01795700e-02 -8.97823155e-01 1.49813175e+00 2.88330376e-01
-8.01516354e-01 -2.30234466e-03 1.15116544e-01 -7.91169524e-01
-1.76374599e-01 2.09579691e-01 1.77481085e-01 -1.61948904e-01
1.48181513e-01 -7.92359591e-01 -7.48640239e-01 -2.85295695e-01
-2.92051196e-01 -1.72730938e-01 3.70097816e-01 1.05942512e+00
6.42978489e-01 -5.96293390e-01 1.05935502e+00 -9.12150562e-01
8.75262141e-01 -5.03958762e-01 -1.12446949e-01 6.27307445e-02
-1.28214240e+00 -2.39619777e-01 5.97594798e-01 -2.67961591e-01
-1.81505606e-01 -1.01183616e-02 -1.06596068e-01 -5.95350266e-01
7.26826265e-02 5.65247297e-01 4.21065778e-01 1.97777435e-01
5.69975555e-01 3.28604132e-01 6.01476692e-02 -1.45369798e-01
-2.72134095e-01 4.07344192e-01 4.81139183e-01 -1.11023635e-01
5.36874652e-01 2.39980295e-02 -1.26516093e-02 -5.70307970e-01
-5.88897049e-01 -3.06287020e-01 -4.96446639e-01 6.11693747e-02
9.10446763e-01 -8.41983020e-01 -6.16854548e-01 4.44482177e-01
-1.06062293e+00 -2.76152402e-01 -4.45235252e-01 7.26347268e-01
-5.81903279e-01 -2.43272949e-02 -7.49405742e-01 -6.57567739e-01
-1.22232294e+00 -6.36876106e-01 4.49818730e-01 -1.55192090e-03
-6.09791219e-01 -1.10714674e+00 5.64867437e-01 -1.61300078e-01
6.36116266e-01 7.60306180e-01 1.19307125e+00 -9.46268678e-01
-3.77716077e-03 -3.54756773e-01 9.38663781e-02 3.61592501e-01
5.05259991e-01 5.90504594e-02 -8.60555768e-01 -3.94216448e-01
-8.33878741e-02 8.36469010e-02 8.28796804e-01 1.09050095e+00
1.30670023e+00 7.66343263e-04 -3.32138360e-01 1.18792236e+00
1.23509777e+00 8.00254464e-01 6.95693552e-01 1.80639401e-01
6.26665890e-01 -3.00112367e-01 -5.14007290e-04 7.29071021e-01
3.57632041e-01 3.31199616e-01 9.62398648e-02 -5.28938234e-01
3.45291078e-01 3.65769535e-01 -2.18237430e-01 8.03516567e-01
-6.21155679e-01 1.44206837e-01 -1.35573769e+00 4.77842957e-01
-1.72903442e+00 -1.14454913e+00 -5.44482395e-02 2.12071276e+00
8.19286227e-01 6.63771108e-02 3.42214815e-02 3.11195940e-01
4.75336552e-01 4.02987659e-01 -7.02026129e-01 -8.10134172e-01
-8.77514631e-02 5.89556575e-01 2.18129054e-01 4.93301600e-01
-1.19483721e+00 4.45164949e-01 8.13799667e+00 -1.63958669e-01
-1.39773726e+00 -2.69812673e-01 9.89273787e-01 -1.42595142e-01
-9.14866179e-02 -3.01766515e-01 -4.86190856e-01 5.31016886e-01
1.40392685e+00 -4.83548760e-01 1.96153805e-01 6.17770195e-01
2.20232189e-01 6.74780905e-01 -1.11746931e+00 1.17253900e+00
5.16068526e-02 -1.60607898e+00 -3.35146278e-01 -1.26265094e-01
3.29526842e-01 3.13550055e-01 1.20055079e-01 4.27346051e-01
-5.58265924e-01 -1.22364271e+00 -1.95663143e-02 7.21893609e-01
1.23603642e+00 -8.56748462e-01 9.79486585e-01 8.30048099e-02
-9.80833709e-01 -2.09585726e-01 8.59114528e-02 -4.14564908e-02
2.67559230e-01 7.52328992e-01 -1.00705659e+00 1.66114941e-01
8.29674780e-01 9.15631831e-01 -3.38112652e-01 9.61772263e-01
4.58250642e-02 8.17954481e-01 -3.08537304e-01 3.83788466e-01
-1.80291370e-01 -1.20838480e-02 6.44758284e-01 1.17013323e+00
1.19647114e-02 2.54060507e-01 7.07445070e-02 6.85447276e-01
9.80959013e-02 4.16683257e-02 -8.54104519e-01 1.03515506e-01
4.73259926e-01 9.91653800e-01 -8.11552182e-02 -4.83647406e-01
-1.42056555e-01 5.61692297e-01 1.31922752e-01 3.01872075e-01
-1.03719044e+00 -9.91342247e-01 5.82238793e-01 -4.41609323e-02
-4.06528302e-02 3.63325179e-01 -8.62253606e-01 -9.73144650e-01
1.33882165e-01 -7.82820642e-01 5.36908329e-01 -2.96851963e-01
-7.55468428e-01 8.70487869e-01 -5.48827767e-01 -1.20818233e+00
-8.05305064e-01 -2.81624824e-01 -1.03844428e+00 1.33726132e+00
-1.39812195e+00 -5.72150707e-01 5.13346121e-02 2.36725852e-01
1.73361465e-01 -3.94209802e-01 1.61195946e+00 3.97135735e-01
-6.44398093e-01 8.24690223e-01 -2.93002993e-01 4.51136380e-01
5.94378769e-01 -1.56510353e+00 4.46926594e-01 4.61726218e-01
-1.26978621e-01 7.22196758e-01 2.72799104e-01 -4.19723511e-01
-7.56989419e-01 -1.17429197e+00 1.35848379e+00 -6.42313659e-01
2.73581684e-01 8.89326036e-02 -9.33509886e-01 5.98469794e-01
-1.56838968e-01 2.28382587e-01 9.58247960e-01 3.85918260e-01
1.28215468e-02 -3.25778604e-01 -8.59233201e-01 2.52132446e-01
3.85870218e-01 -5.99068940e-01 -6.85139656e-01 5.59528032e-03
4.10566688e-01 -6.54937625e-01 -1.57613456e+00 1.01835048e+00
8.77614319e-01 -6.41693413e-01 1.21467888e+00 -9.97890949e-01
2.37987146e-01 2.05602348e-01 5.83589673e-01 -1.25226808e+00
-5.98405778e-01 -9.88139391e-01 -6.24834538e-01 4.79248255e-01
8.84144902e-01 -7.61610448e-01 7.61168182e-01 7.82326698e-01
-2.89491415e-01 -1.29890907e+00 -8.07654202e-01 -3.56494486e-01
3.51522118e-02 -1.33001104e-01 2.97980398e-01 9.49902117e-01
1.09777942e-01 3.69541287e-01 -7.43131220e-01 9.83780250e-02
3.74404013e-01 1.44306496e-01 1.21142797e-01 -1.38376594e+00
-3.23607504e-01 -4.74038064e-01 -4.11755621e-01 -7.01546893e-02
5.71712106e-02 -8.73888254e-01 -4.69363093e-01 -1.38846064e+00
3.60609591e-02 -5.68240285e-01 -1.31266284e+00 6.48149490e-01
-5.85587800e-01 2.27321029e-01 -1.13642573e-01 -4.21191193e-02
4.54250202e-02 2.05144569e-01 7.01250255e-01 2.55167484e-02
-7.83667982e-01 6.71002388e-01 -5.65685749e-01 6.68727815e-01
1.34497142e+00 -6.71800256e-01 -4.43037987e-01 -3.20080727e-01
6.90896064e-02 7.66814828e-01 2.07057983e-01 -9.69623923e-01
-3.55799049e-01 2.53839552e-01 7.71875679e-01 -4.42447394e-01
3.60810280e-01 -4.04689163e-01 2.13545606e-01 9.53272283e-01
-4.24798161e-01 6.22480392e-01 3.21302325e-01 3.85374337e-01
-2.79166430e-01 3.24611217e-01 6.97392642e-01 6.70841560e-02
-6.46366999e-02 6.92979515e-01 -4.33129400e-01 3.89649242e-01
8.10061157e-01 -2.28397235e-01 3.51712741e-02 -1.16697907e-01
-1.34964609e+00 2.73180939e-03 -2.64000118e-01 6.01584017e-01
7.30478406e-01 -1.21448040e+00 -9.18218076e-01 3.09278786e-01
2.36044582e-02 -4.42151546e-01 3.30074787e-01 1.12042868e+00
-6.12279892e-01 6.40403092e-01 -3.63833010e-01 -4.99950826e-01
-1.28968406e+00 1.21469721e-01 8.16539824e-01 -5.41452706e-01
-1.00828779e+00 6.29929543e-01 -3.83968472e-01 -9.10925418e-02
3.80307496e-01 -4.52755988e-01 -4.31088686e-01 -2.11839318e-01
5.60755908e-01 3.42487603e-01 6.98985904e-02 2.98412982e-02
-5.91225624e-01 5.85108638e-01 4.63919081e-02 2.87129253e-01
1.33306968e+00 2.83344865e-01 -4.98830192e-02 8.05849850e-01
1.07312024e+00 -3.69793206e-01 -5.68879485e-01 -1.25684455e-01
-3.50994542e-02 1.68599397e-01 1.12455837e-01 -1.29028690e+00
-1.18620253e+00 9.14198279e-01 1.06152284e+00 4.69212458e-02
1.16304100e+00 -4.76762563e-01 6.66615784e-01 2.88033128e-01
-1.53647438e-02 -7.24918723e-01 -4.20932829e-01 4.80008811e-01
6.14659011e-01 -1.01903200e+00 -4.31124531e-02 2.73812562e-01
-7.08693564e-01 1.19749379e+00 3.37064862e-01 -2.89114386e-01
9.76009250e-01 3.37663233e-01 6.95109069e-01 -1.24823995e-01
-9.83905971e-01 4.77180332e-01 2.72386193e-01 4.05013829e-01
5.52477717e-01 2.02638939e-01 -5.80736220e-01 8.02987278e-01
5.39168194e-02 2.05402046e-01 2.45861903e-01 5.26571929e-01
-6.60140961e-02 -8.93901348e-01 2.91539729e-01 1.01783705e+00
-9.47416663e-01 -4.29587960e-01 7.72045255e-02 5.21169901e-01
-6.54739365e-02 3.75874460e-01 8.04365799e-02 -4.01811242e-01
2.28323951e-01 6.46521866e-01 7.47135803e-02 -6.84495211e-01
-1.01000977e+00 -2.52885491e-01 4.31320556e-02 -5.44509530e-01
1.76618025e-01 -4.07116205e-01 -1.41193283e+00 2.16904193e-01
-6.03534915e-02 2.51462966e-01 4.63763267e-01 5.30058861e-01
7.00925350e-01 8.78223836e-01 9.23216641e-01 -3.46005201e-01
-6.18021250e-01 -1.02471328e+00 -6.47752285e-01 1.63567707e-01
8.91675055e-01 -5.46811409e-02 -5.97223699e-01 -3.13322507e-02] | [14.374743461608887, 3.3461363315582275] |
5cccbd21-bbae-48ff-ad6a-79544782e4f9 | mkis-net-a-light-weight-multi-kernel-network | 2210.08168 | null | https://arxiv.org/abs/2210.08168v1 | https://arxiv.org/pdf/2210.08168v1.pdf | MKIS-Net: A Light-Weight Multi-Kernel Network for Medical Image Segmentation | Image segmentation is an important task in medical imaging. It constitutes the backbone of a wide variety of clinical diagnostic methods, treatments, and computer-aided surgeries. In this paper, we propose a multi-kernel image segmentation net (MKIS-Net), which uses multiple kernels to create an efficient receptive field and enhance segmentation performance. As a result of its multi-kernel design, MKIS-Net is a light-weight architecture with a small number of trainable parameters. Moreover, these multi-kernel receptive fields also contribute to better segmentation results. We demonstrate the efficacy of MKIS-Net on several tasks including segmentation of retinal vessels, skin lesion segmentation, and chest X-ray segmentation. The performance of the proposed network is quite competitive, and often superior, in comparison to state-of-the-art methods. Moreover, in some cases MKIS-Net has more than an order of magnitude fewer trainable parameters than existing medical image segmentation alternatives and is at least four times smaller than other light-weight architectures. | ['Erik Meijering', 'Antonio Robles-Kelly', 'Muhammad Arsalan', 'Tariq M. Khan'] | 2022-10-15 | null | null | null | null | ['skin-lesion-segmentation'] | ['medical'] | [ 2.80090332e-01 1.41725704e-01 -2.43972182e-01 -3.63323629e-01
-5.01800597e-01 -2.29479909e-01 1.08277850e-01 9.40894559e-02
-7.59856164e-01 3.94192308e-01 -8.81534144e-02 -3.99000406e-01
-1.81506440e-01 -4.70498711e-01 -3.78570706e-01 -8.21017146e-01
2.47798771e-01 1.94513679e-01 7.77576685e-01 9.38633233e-02
1.24777742e-02 7.81622648e-01 -9.73357737e-01 1.63411170e-01
1.10762727e+00 1.06214797e+00 2.15054020e-01 5.22811472e-01
-5.26869446e-02 6.75506353e-01 -1.74782813e-01 -3.40549737e-01
-6.29395340e-03 -1.66416958e-01 -9.51954424e-01 1.13011457e-01
4.11393195e-01 -3.12349230e-01 -1.90279305e-01 9.58985269e-01
5.14043093e-01 2.20768228e-01 7.32128918e-01 -5.06802678e-01
-6.63126647e-01 5.46194792e-01 -6.31892025e-01 3.68074894e-01
-4.38191324e-01 -8.16081613e-02 5.42938292e-01 -5.91133356e-01
2.45680973e-01 8.12717021e-01 7.05699086e-01 6.57779872e-01
-1.06474972e+00 -2.66639143e-01 -7.01936707e-02 -9.45667848e-02
-1.27386343e+00 -1.12050079e-01 3.09759825e-01 -3.62823308e-01
8.39001119e-01 2.62886643e-01 6.09954834e-01 5.46545208e-01
4.73339200e-01 9.25614536e-01 1.08389819e+00 -2.79062718e-01
1.68168068e-01 9.03361812e-02 2.82408327e-01 1.00610590e+00
1.34285614e-01 -2.29904145e-01 1.13724329e-01 -8.83095488e-02
1.33562911e+00 2.00815573e-01 -3.12687993e-01 -8.86976793e-02
-1.22657192e+00 7.12417245e-01 6.89943254e-01 5.55671871e-01
-2.73482352e-01 3.83068323e-01 1.93286717e-01 -3.17351967e-01
4.12951052e-01 4.35241848e-01 -1.70253336e-01 1.62537739e-01
-9.20066595e-01 -1.40032992e-01 3.76080304e-01 4.24586922e-01
4.90486771e-01 1.34678990e-01 -4.08821523e-01 1.07699919e+00
9.06838700e-02 2.70429522e-01 6.81270421e-01 -8.92268538e-01
1.52446870e-02 6.97193384e-01 -2.12358683e-01 -4.91397649e-01
-8.58705997e-01 -4.60905373e-01 -1.17545915e+00 2.98423022e-01
4.80850875e-01 -7.49097764e-02 -1.35718071e+00 1.23657179e+00
4.01748151e-01 4.40506577e-01 -1.01910263e-01 8.69838536e-01
1.16967869e+00 4.93437648e-01 2.51895964e-01 -2.16791942e-03
1.49931240e+00 -1.12709284e+00 -4.90039647e-01 -2.34076604e-01
5.56247115e-01 -8.71853173e-01 8.31167042e-01 3.12833101e-01
-1.11330700e+00 -5.75527668e-01 -7.23152637e-01 -4.99167927e-02
-1.45326301e-01 5.54667175e-01 1.11741149e+00 7.02178359e-01
-1.09891021e+00 6.42715991e-01 -9.64927554e-01 -2.07234994e-01
7.05089927e-01 6.48030102e-01 -3.36334586e-01 1.35906994e-01
-7.47659683e-01 7.59636641e-01 4.49723482e-01 3.13075073e-02
-2.17706397e-01 -9.33664560e-01 -8.40984404e-01 8.98738056e-02
3.26237053e-01 -5.65710783e-01 1.21524119e+00 -6.22980654e-01
-1.72569561e+00 9.84210372e-01 7.56219178e-02 -3.58653009e-01
2.27719769e-01 1.12096481e-01 -2.66415983e-01 5.94387949e-01
-1.97957277e-01 1.03516674e+00 9.45882738e-01 -7.95376003e-01
-6.09446824e-01 -1.81879491e-01 -7.95025751e-02 2.06180289e-02
-3.99613529e-01 7.08294148e-03 -8.16561520e-01 -8.48363876e-01
2.30809636e-02 -1.00156271e+00 -6.71071529e-01 1.25909001e-01
-3.42018962e-01 -4.00001734e-01 4.09253985e-01 -5.02463460e-01
1.23951030e+00 -2.24270940e+00 1.89631823e-02 1.58204451e-01
5.71865857e-01 7.86716163e-01 -1.62832290e-01 -1.76354721e-01
-7.16053396e-02 6.53933957e-02 -3.81165236e-01 7.02701956e-02
-3.57972413e-01 1.02657698e-01 1.38430372e-01 4.40328836e-01
2.40710095e-01 1.14785254e+00 -5.69998682e-01 -5.93206406e-01
5.05972326e-01 5.57828367e-01 -3.52373362e-01 -6.23064116e-02
1.51405916e-01 4.99978930e-01 -6.51297688e-01 6.46461368e-01
5.53676963e-01 -7.02714562e-01 -1.79208010e-01 -4.37672317e-01
-8.26219320e-02 -3.97352725e-01 -8.92918825e-01 1.63469958e+00
-2.97254324e-01 5.07336736e-01 -9.37229171e-02 -1.03138542e+00
5.27754843e-01 3.09693724e-01 8.07393312e-01 -6.51903331e-01
4.92358088e-01 2.43184701e-01 2.06368446e-01 -4.50345844e-01
1.56712651e-01 -8.69334862e-02 3.28071922e-01 2.78889358e-01
1.88266188e-01 7.94589892e-03 3.51296574e-01 -2.72253510e-02
9.64026988e-01 -2.63061911e-01 2.94045389e-01 -4.06060845e-01
6.27899766e-01 -3.47782522e-01 6.10982656e-01 6.70671225e-01
-3.39462399e-01 5.53931534e-01 3.03574562e-01 -6.70225501e-01
-7.43253529e-01 -1.00029457e+00 -5.67209721e-01 7.18880892e-01
3.07262868e-01 1.36788690e-03 -1.04987681e+00 -5.02102911e-01
-8.31388384e-02 1.98042050e-01 -5.96580863e-01 3.31653841e-02
-6.29201472e-01 -1.03218794e+00 7.56650329e-01 8.48459005e-01
7.31737792e-01 -1.12065661e+00 -7.45483279e-01 3.19462717e-01
-1.68812796e-02 -1.24416280e+00 -5.76097846e-01 1.86991654e-02
-1.12555730e+00 -1.13876951e+00 -1.25405574e+00 -1.04051912e+00
9.38465655e-01 1.66089758e-01 8.91527474e-01 -4.14452627e-02
-9.02200162e-01 1.27698645e-01 -1.88336834e-01 -4.47775513e-01
-2.44342864e-01 1.47343159e-01 -3.21551710e-01 1.24883272e-01
1.98538721e-01 -2.13608563e-01 -9.20519829e-01 1.90348178e-01
-1.18566442e+00 1.16166875e-01 9.94548261e-01 7.52670348e-01
7.99124241e-01 6.43646643e-02 4.40788537e-01 -1.18570209e+00
5.66831768e-01 2.30275490e-03 -6.69360757e-01 4.65531856e-01
-4.13794070e-01 -2.92456672e-02 5.17216861e-01 -6.04353011e-01
-1.11021781e+00 1.86738163e-01 -2.98389226e-01 -2.46399298e-01
-1.28006235e-01 3.24018240e-01 6.64143682e-01 -7.82608092e-01
7.24383831e-01 -9.31767523e-02 1.36499494e-01 -2.87463695e-01
3.05222541e-01 4.76503938e-01 4.17727768e-01 -2.74386317e-01
3.98676783e-01 5.78942418e-01 9.64590609e-02 -9.98927951e-01
-7.75863290e-01 -7.61523724e-01 -5.38067639e-01 -1.31312564e-01
1.23531651e+00 -5.56039751e-01 -9.02122021e-01 8.05407703e-01
-8.50840032e-01 -6.30270302e-01 -2.16590434e-01 7.49049008e-01
-3.03271621e-01 4.80475098e-01 -1.06273735e+00 -4.70018029e-01
-5.75067759e-01 -1.55380821e+00 8.38227034e-01 7.72975326e-01
1.42474353e-01 -1.18836832e+00 -3.31895590e-01 5.50449967e-01
5.01874030e-01 2.43690386e-01 1.14154780e+00 -3.45022351e-01
-4.73603040e-01 -1.88999012e-01 -6.43752098e-01 5.96785784e-01
2.50785381e-01 -1.71996083e-03 -7.63694048e-01 -3.07327330e-01
-2.76228100e-01 -3.42858762e-01 1.27923203e+00 1.10142255e+00
1.60244668e+00 2.29575232e-01 -4.83571380e-01 9.15955782e-01
1.38483214e+00 3.28059971e-01 6.55130863e-01 -6.07143342e-02
7.92214811e-01 3.89589012e-01 1.58813000e-01 1.02587014e-01
1.42389283e-01 2.85152704e-01 2.26085827e-01 -9.90798950e-01
-2.24877939e-01 5.69461167e-01 -2.96110660e-01 7.98350036e-01
-3.21134508e-01 3.31767835e-02 -8.02585602e-01 4.63288158e-01
-1.73374796e+00 -6.82764292e-01 -1.61989555e-01 2.01573110e+00
8.41314197e-01 -5.79048472e-04 9.44859385e-02 -3.16500962e-01
5.75351357e-01 -1.45568907e-01 -7.40909457e-01 -3.59273046e-01
1.64539665e-01 7.90285707e-01 8.75773132e-01 1.77482471e-01
-1.49885833e+00 9.88189995e-01 6.97790098e+00 1.16895878e+00
-1.24103200e+00 -3.46857309e-02 8.89632523e-01 2.99465418e-01
2.83811633e-02 -3.75246286e-01 -7.65296459e-01 2.33276665e-01
5.99416912e-01 4.04257804e-01 2.55967289e-01 7.51345277e-01
-1.59471527e-01 -3.99647146e-01 -6.35817647e-01 1.06114912e+00
1.10138327e-01 -1.68545985e+00 -8.10358971e-02 6.56793863e-02
6.82775676e-01 8.63245502e-02 1.76342010e-01 1.03325002e-01
-3.98851931e-02 -1.06992865e+00 7.35467374e-02 2.97641397e-01
1.03367782e+00 -7.44158447e-01 7.71210909e-01 4.96199541e-02
-1.05689728e+00 -1.21804783e-02 -5.20034373e-01 4.81310040e-01
1.68887690e-01 7.53409982e-01 -4.66858476e-01 2.72903174e-01
6.98370874e-01 4.19191182e-01 -5.21579087e-01 1.49849641e+00
-1.23595651e-02 7.23711014e-01 -2.29760855e-01 1.86988350e-03
6.16756558e-01 -4.45120633e-01 3.90674211e-02 1.36562121e+00
8.01882297e-02 2.30072826e-01 2.82429338e-01 7.18480825e-01
-2.96673715e-01 2.41967395e-01 -1.48471937e-01 -1.28133241e-02
-7.19941854e-02 1.64253092e+00 -1.32602262e+00 -4.46111143e-01
-5.51521778e-01 9.54682767e-01 2.09929079e-01 4.61333871e-01
-8.74698281e-01 -6.58090055e-01 4.42481071e-01 -1.99827105e-01
2.12086141e-01 9.64051560e-02 -1.35276183e-01 -9.24386263e-01
-1.27848178e-01 -5.77704728e-01 2.30115101e-01 -4.17880267e-01
-1.17648745e+00 7.77425110e-01 -1.05867974e-01 -9.94620085e-01
2.33801171e-01 -1.09745216e+00 -6.13956273e-01 6.89645588e-01
-1.93164384e+00 -1.32363939e+00 -3.37604314e-01 5.72708607e-01
3.87754053e-01 -1.65678188e-01 8.52996767e-01 3.08900952e-01
-6.05273068e-01 5.00840664e-01 1.76107809e-01 3.82399559e-01
6.81799769e-01 -1.25260341e+00 3.35978657e-01 6.72243595e-01
2.09828075e-02 5.21543086e-01 -1.57418415e-01 -3.82671416e-01
-6.98164046e-01 -1.15746331e+00 3.17415923e-01 1.26119778e-01
5.36835551e-01 2.00377956e-01 -9.51664329e-01 3.83959651e-01
2.28489846e-01 4.23360258e-01 1.10345960e+00 -1.57976523e-01
-1.11068733e-01 -3.43412459e-01 -1.10553563e+00 6.10150814e-01
3.99095774e-01 -7.83909932e-02 -2.51873970e-01 5.74928820e-01
4.65949237e-01 -7.14494348e-01 -1.03021002e+00 6.77148581e-01
5.26796758e-01 -9.37538683e-01 1.19234216e+00 -3.38785350e-01
2.65487760e-01 -1.61595374e-01 5.03287554e-01 -1.03923512e+00
-7.01542795e-01 -4.44979489e-01 1.95001125e-01 7.34298646e-01
4.15588319e-01 -7.82883227e-01 8.09848070e-01 4.66903836e-01
-3.55476201e-01 -1.22929537e+00 -8.44847739e-01 -4.84540939e-01
1.51377752e-01 -1.26514181e-01 1.17953211e-01 6.71902061e-01
-3.43748868e-01 9.09955576e-02 -4.08583730e-02 -1.66084804e-02
7.03350544e-01 -2.34040469e-02 1.08298741e-01 -1.19927585e+00
-3.92347097e-01 -7.18635321e-01 -4.21012998e-01 -9.42006230e-01
3.02203465e-02 -9.28144395e-01 -1.21595338e-01 -1.78393006e+00
1.34939879e-01 -5.55438876e-01 -5.83960593e-01 7.41672516e-01
-4.32991773e-01 7.18256176e-01 -8.20955411e-02 1.29626662e-01
-4.60877806e-01 3.23532410e-02 1.71441460e+00 -1.95609316e-01
-3.28993708e-01 2.84646690e-01 -5.82909942e-01 1.00337076e+00
9.37001109e-01 -1.20371409e-01 -3.85029823e-01 -3.30390453e-01
-1.60025179e-01 -1.66128948e-01 1.81869060e-01 -1.03967428e+00
2.44660273e-01 -1.79492399e-01 5.79160631e-01 -3.55347782e-01
2.39206031e-01 -5.32025456e-01 -3.49415362e-01 4.96178180e-01
-1.55538052e-01 -3.88113648e-01 3.72875035e-01 2.87311882e-01
-2.10248858e-01 -3.18435878e-01 1.22612512e+00 -2.99541146e-01
-8.43067408e-01 5.01631439e-01 -6.21933639e-01 -7.67895356e-02
1.25476468e+00 -4.35426265e-01 -1.29870832e-01 4.21475619e-01
-6.26215160e-01 1.02149040e-01 -9.68225077e-02 1.44583330e-01
7.40136921e-01 -1.03389359e+00 -4.36861962e-01 2.07444623e-01
4.68847044e-02 2.23999947e-01 6.02080941e-01 1.24351823e+00
-1.03894961e+00 3.95192087e-01 -1.77253723e-01 -5.86715102e-01
-1.22979939e+00 2.47786403e-01 3.56942356e-01 -3.81425172e-01
-8.53043318e-01 1.15418017e+00 4.44963455e-01 -3.08386207e-01
1.95272148e-01 -8.45097423e-01 -3.98001194e-01 -3.39867055e-01
6.09149516e-01 4.59254891e-01 1.23389162e-01 -3.60346258e-01
-2.41875693e-01 9.20546532e-01 -3.72231871e-01 4.55961376e-01
1.20394385e+00 3.65070701e-01 -3.70826304e-01 -1.86043158e-02
8.07036638e-01 -3.84155065e-01 -1.15154338e+00 -3.08160841e-01
-2.78816164e-01 -9.63679329e-02 4.27131325e-01 -8.80086124e-01
-1.35183644e+00 8.16819787e-01 6.92874491e-01 7.88780802e-04
1.30342197e+00 -3.83150205e-02 1.08964276e+00 3.10411751e-01
6.55612648e-02 -1.32293308e+00 1.28225401e-01 8.10077116e-02
4.10812348e-01 -1.41002476e+00 4.77626547e-02 -8.71008694e-01
-6.88432276e-01 1.31536794e+00 5.18561304e-01 -2.40625173e-01
7.26651132e-01 3.36393088e-01 3.45055610e-01 -1.06022507e-01
1.22623947e-02 -4.54385728e-01 7.43610919e-01 5.20558476e-01
5.44936299e-01 1.22059420e-01 -2.76020169e-01 2.21201748e-01
4.07280743e-01 1.29087985e-01 2.32973889e-01 7.00761080e-01
-5.29690146e-01 -1.10811591e+00 -9.55061540e-02 8.06262255e-01
-8.13717365e-01 -9.03865620e-02 3.04643493e-02 6.29434168e-01
2.89833471e-02 7.17886209e-01 -2.75641438e-02 9.59897339e-02
1.83035031e-01 -3.14632326e-01 5.91341078e-01 -5.78269422e-01
-7.60016382e-01 3.73197913e-01 -4.82678682e-01 -6.86438501e-01
-6.14197314e-01 7.56543949e-02 -1.62273443e+00 -5.68412282e-02
-4.88566220e-01 -1.58423916e-01 6.15561247e-01 1.00258827e+00
2.30180755e-01 9.84511375e-01 2.25751579e-01 -6.70199752e-01
-3.77137423e-01 -7.78327525e-01 -8.15433800e-01 2.68866390e-01
1.56292170e-01 -4.64612514e-01 1.59450978e-01 1.94951922e-01] | [14.685514450073242, -2.593090534210205] |
cadb981f-1541-45a1-833b-b5d14cef92b4 | temporal-knowledge-propagation-for-image-to | 1908.03885 | null | https://arxiv.org/abs/1908.03885v3 | https://arxiv.org/pdf/1908.03885v3.pdf | Temporal Knowledge Propagation for Image-to-Video Person Re-identification | In many scenarios of Person Re-identification (Re-ID), the gallery set consists of lots of surveillance videos and the query is just an image, thus Re-ID has to be conducted between image and videos. Compared with videos, still person images lack temporal information. Besides, the information asymmetry between image and video features increases the difficulty in matching images and videos. To solve this problem, we propose a novel Temporal Knowledge Propagation (TKP) method which propagates the temporal knowledge learned by the video representation network to the image representation network. Specifically, given the input videos, we enforce the image representation network to fit the outputs of video representation network in a shared feature space. With back propagation, temporal knowledge can be transferred to enhance the image features and the information asymmetry problem can be alleviated. With additional classification and integrated triplet losses, our model can learn expressive and discriminative image and video features for image-to-video re-identification. Extensive experiments demonstrate the effectiveness of our method and the overall results on two widely used datasets surpass the state-of-the-art methods by a large margin. Code is available at: https://github.com/guxinqian/TKP | ['Bingpeng Ma', 'Xilin Chen', 'Xinqian Gu', 'Shiguang Shan', 'Hong Chang'] | 2019-08-11 | temporal-knowledge-propagation-for-image-to-1 | http://openaccess.thecvf.com/content_ICCV_2019/html/Gu_Temporal_Knowledge_Propagation_for_Image-to-Video_Person_Re-Identification_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Gu_Temporal_Knowledge_Propagation_for_Image-to-Video_Person_Re-Identification_ICCV_2019_paper.pdf | iccv-2019-10 | ['image-to-video-person-re-identification'] | ['computer-vision'] | [-7.10272416e-02 -5.18138468e-01 -3.51355463e-01 -3.78172964e-01
-5.44529080e-01 -4.86359864e-01 4.60083187e-01 -4.46692109e-01
-3.99364144e-01 4.44582611e-01 1.50166333e-01 3.80863845e-01
-2.02606067e-01 -4.44813520e-01 -6.99578822e-01 -6.62955225e-01
-7.16381073e-02 4.45020087e-02 1.44935620e-03 2.22204506e-01
-6.60018548e-02 1.31823987e-01 -1.45250487e+00 2.67855406e-01
5.07740319e-01 1.17209077e+00 6.15009591e-02 3.56988102e-01
2.34776929e-01 9.83006537e-01 -3.21210921e-01 -6.32301033e-01
5.21156371e-01 -3.30089808e-01 -6.30549610e-01 3.61634910e-01
5.18720865e-01 -6.58233941e-01 -1.23855042e+00 1.29333675e+00
4.58765894e-01 4.95715946e-01 1.79896757e-01 -1.64448059e+00
-1.04084420e+00 2.59813070e-01 -7.18424618e-01 3.08212966e-01
5.57879150e-01 2.11917356e-01 5.95709085e-01 -8.64511609e-01
5.16749382e-01 1.32613289e+00 5.65652132e-01 6.74229383e-01
-8.65733683e-01 -1.03001261e+00 3.92397135e-01 7.82278895e-01
-1.72331333e+00 -5.09395123e-01 7.24134147e-01 -4.35443074e-01
4.74472821e-01 2.92035818e-01 8.10941398e-01 9.34273303e-01
-3.70101810e-01 9.26307440e-01 4.64523762e-01 6.72070682e-02
-5.93692899e-01 2.63741016e-01 5.10931276e-02 6.79272890e-01
1.24520166e-02 3.62016082e-01 -6.71935320e-01 3.82613763e-02
8.00934732e-01 7.46449351e-01 -5.51478565e-01 -1.47312105e-01
-1.17288637e+00 6.17427170e-01 5.45127988e-01 2.73684233e-01
-2.27105692e-01 2.26186395e-01 5.05862176e-01 5.04963875e-01
2.50175893e-01 -1.45381480e-01 -3.34805287e-02 8.31394196e-02
-8.53219748e-01 1.35371089e-01 2.65155584e-01 9.27629769e-01
8.83883834e-01 -2.60801077e-01 -4.66635793e-01 7.96390176e-01
2.37006098e-01 6.16938770e-01 6.36674225e-01 -9.45129037e-01
6.78675473e-01 5.37660599e-01 1.90142915e-01 -1.46764970e+00
1.18506653e-02 -1.07196160e-01 -9.88297880e-01 -4.27014410e-01
2.98443437e-01 -1.61104947e-02 -6.18640065e-01 1.73742807e+00
2.24417925e-01 6.71157539e-01 -9.84606668e-02 1.15622890e+00
7.67367959e-01 8.91189575e-01 -2.14025676e-02 -2.56346613e-01
1.26222837e+00 -1.09603977e+00 -6.69211090e-01 -1.34041741e-01
3.05548310e-01 -6.13135993e-01 4.57122207e-01 3.78906466e-02
-8.54121983e-01 -8.79511535e-01 -7.74618685e-01 4.41686325e-02
-8.52314457e-02 3.80466253e-01 2.88245410e-01 4.36321378e-01
-8.94063056e-01 5.23790419e-01 -5.13046026e-01 -3.86057466e-01
4.93482143e-01 5.22002459e-01 -7.39776850e-01 -6.13130748e-01
-1.29265928e+00 3.67237359e-01 5.16788304e-01 4.29762691e-01
-9.25808907e-01 -6.25689149e-01 -7.76786387e-01 -6.53807595e-02
4.45399255e-01 -6.31896496e-01 8.33830953e-01 -1.37001908e+00
-1.13692570e+00 8.40111136e-01 -1.94432363e-01 -3.13371181e-01
7.75551677e-01 -2.04189152e-01 -6.58687472e-01 3.18210810e-01
2.84284443e-01 6.47478580e-01 1.04471266e+00 -1.00141561e+00
-7.61301517e-01 -3.49065274e-01 1.70284480e-01 3.03874195e-01
-6.55763090e-01 1.44974098e-01 -1.14271510e+00 -7.53378510e-01
-2.98105747e-01 -1.06447494e+00 1.02206893e-01 2.42942780e-01
-2.98082650e-01 -1.68591008e-01 9.61365044e-01 -1.07412815e+00
1.15422535e+00 -2.49549842e+00 2.51360565e-01 3.20587277e-01
1.58167988e-01 2.73435831e-01 -2.81604916e-01 2.47276202e-01
-2.49652043e-01 -9.50358510e-02 8.14738795e-02 -4.28206891e-01
-1.44802600e-01 1.42828785e-02 -2.56285429e-01 7.57554531e-01
-2.07604855e-01 9.16886747e-01 -8.68311584e-01 -6.05961800e-01
3.04910243e-01 6.46835208e-01 -3.42090130e-01 4.27012503e-01
3.20398420e-01 5.98186851e-01 -5.21101356e-01 6.03917420e-01
8.09931457e-01 -4.16124225e-01 7.47177005e-02 -5.39417386e-01
7.46316686e-02 -3.78265262e-01 -1.03667903e+00 1.78099024e+00
-4.83762585e-02 5.59489310e-01 -2.02195734e-01 -1.23750734e+00
5.80801606e-01 3.45573425e-01 7.69791782e-01 -7.96471477e-01
5.84336668e-02 -6.38325885e-02 -4.19451892e-01 -7.13543892e-01
1.83202147e-01 2.53471941e-01 5.09494282e-02 3.56236219e-01
2.39219125e-02 7.50305414e-01 1.72870174e-01 1.20367885e-01
4.65656817e-01 1.09867081e-01 -1.39713287e-01 2.00063974e-01
9.03935909e-01 -3.09873164e-01 8.84567022e-01 5.54673254e-01
-3.95486265e-01 4.62351918e-01 -5.05156443e-03 -5.58753431e-01
-8.76110315e-01 -8.23540330e-01 1.74346030e-01 1.00288522e+00
6.36360526e-01 -3.67786556e-01 -5.86165667e-01 -6.29026234e-01
1.69855595e-01 -1.77721158e-02 -6.59396112e-01 -2.82185435e-01
-5.63877404e-01 -6.10564113e-01 4.70013618e-01 3.60266566e-01
9.25924420e-01 -7.07464576e-01 1.62180528e-01 -9.02593508e-02
-5.66419065e-01 -1.15362597e+00 -1.25653148e+00 -8.55654478e-01
-5.40687978e-01 -1.31106758e+00 -1.12891412e+00 -1.00624979e+00
8.97895992e-01 8.12919796e-01 5.67834020e-01 3.75915051e-01
-3.33470017e-01 7.55837917e-01 -3.98297399e-01 2.77281046e-01
1.44818753e-01 -3.75967741e-01 3.27974945e-01 7.25825727e-01
4.95186329e-01 -3.32092464e-01 -9.63517487e-01 5.04687309e-01
-8.49014521e-01 -1.02637373e-01 2.64477849e-01 1.07240820e+00
4.70848888e-01 3.05952489e-01 2.74405062e-01 -2.88787484e-01
1.75466895e-01 -4.98952121e-01 -5.17081201e-01 5.73090851e-01
-3.02877784e-01 -3.26977432e-01 5.03005385e-01 -6.74215198e-01
-9.81435180e-01 8.62751305e-02 2.92557359e-01 -1.03918779e+00
2.81613678e-01 2.33916581e-01 -2.76478469e-01 -3.64998460e-01
-7.21279234e-02 5.53665042e-01 1.05246037e-01 -3.88177961e-01
2.32334793e-01 5.22187948e-01 6.57571554e-01 -3.06949288e-01
9.71932650e-01 5.31960607e-01 -3.80603552e-01 -5.61539292e-01
-5.74011922e-01 -4.25766796e-01 -3.89108509e-01 -4.18906927e-01
7.91987956e-01 -1.29164410e+00 -9.43770885e-01 6.95029974e-01
-1.11284471e+00 1.49902672e-01 -8.87710452e-02 6.86741471e-01
-2.27925897e-01 8.47705960e-01 -6.64972186e-01 -6.69089973e-01
-2.08897412e-01 -1.08575463e+00 6.59320712e-01 5.37374437e-01
3.45883131e-01 -8.27945650e-01 -2.50583231e-01 5.08803606e-01
1.50291726e-01 -1.04189783e-01 3.15387815e-01 -3.35865945e-01
-8.51606905e-01 -4.33551967e-01 -5.11046767e-01 3.88215661e-01
3.32761377e-01 -2.05743060e-01 -8.13198924e-01 -7.13229299e-01
-9.13340673e-02 -2.22310781e-01 1.05205476e+00 3.47428411e-01
1.42355871e+00 -6.75379455e-01 -5.50561488e-01 8.92522693e-01
1.20701098e+00 2.58566439e-01 4.80141342e-01 3.45812857e-01
1.00189233e+00 7.53172338e-01 6.12963319e-01 5.18107474e-01
5.68888247e-01 9.12640512e-01 1.25690788e-01 7.60639459e-02
-9.14891064e-02 -5.84179580e-01 5.82164347e-01 7.03565776e-01
-1.60157293e-01 -2.02809572e-01 -4.01241750e-01 4.93735105e-01
-2.22085857e+00 -1.39708686e+00 4.31199729e-01 2.31922007e+00
4.18051124e-01 -4.99625117e-01 3.03055704e-01 -1.51420683e-01
1.25739014e+00 2.25650355e-01 -7.33710706e-01 5.07597208e-01
-1.93868101e-01 -7.58028567e-01 5.99789977e-01 3.49863231e-01
-1.27907157e+00 7.26161540e-01 5.46436262e+00 1.00862169e+00
-1.02220547e+00 2.43698969e-01 7.32672215e-01 -3.86729300e-01
-4.73845452e-02 -8.69062915e-02 -6.29447341e-01 9.58283424e-01
4.58822101e-01 -5.80665290e-01 7.69735217e-01 6.47956491e-01
6.95663989e-02 2.90453017e-01 -1.15821052e+00 1.78635192e+00
3.86499554e-01 -1.28019452e+00 2.22052321e-01 -9.34937969e-02
6.63586318e-01 -2.27666318e-01 2.92082459e-01 1.25074223e-01
-1.42039340e-02 -9.18921351e-01 5.28006554e-01 8.68384004e-01
9.84759867e-01 -7.12856412e-01 7.22585797e-01 -8.64574313e-02
-1.62445641e+00 -4.49761063e-01 -4.87943798e-01 2.19728261e-01
2.15968758e-01 2.63497382e-01 -1.56966701e-01 8.36414576e-01
1.11450016e+00 1.44720602e+00 -5.44528604e-01 1.08377147e+00
8.11684951e-02 1.26551121e-01 -1.24346338e-01 4.22007114e-01
5.99853396e-02 -3.66533875e-01 4.50347662e-01 1.03762782e+00
4.85869169e-01 8.21420625e-02 5.10575414e-01 5.97851753e-01
-2.60018796e-01 2.35257968e-02 -5.83191156e-01 -5.70004471e-02
5.15140116e-01 9.11609948e-01 -1.40248492e-01 -3.15180033e-01
-5.46994388e-01 1.45093119e+00 1.64579257e-01 6.22896791e-01
-9.05089617e-01 -1.90572634e-01 7.32598782e-01 -2.45565712e-01
2.84721255e-01 1.16934769e-01 6.21736705e-01 -1.52673721e+00
3.47856283e-01 -7.13551819e-01 7.37124801e-01 -5.82966208e-01
-1.61512518e+00 5.23353994e-01 -2.36240937e-03 -1.55096483e+00
-1.37143552e-01 -3.02379727e-01 -4.52357173e-01 6.69783354e-01
-1.52291203e+00 -1.34470689e+00 -5.75586975e-01 1.19131935e+00
2.84545153e-01 -4.92815226e-01 3.27937961e-01 8.85440230e-01
-9.36712503e-01 1.01354969e+00 2.92619288e-01 6.74258649e-01
8.43503833e-01 -5.35470963e-01 -2.27294862e-02 1.01257098e+00
-4.68491875e-02 4.66582745e-01 2.20451623e-01 -5.33391833e-01
-1.53740752e+00 -1.36065447e+00 5.20640075e-01 -1.22905396e-01
5.32389700e-01 -8.68604556e-02 -8.39185894e-01 8.76628816e-01
2.12828014e-02 2.69840807e-01 7.53199160e-01 -4.03048635e-01
-5.69028735e-01 -5.17161548e-01 -9.98736501e-01 3.06827396e-01
1.17340982e+00 -8.72803867e-01 -2.91843623e-01 4.86540496e-01
6.53901994e-01 -1.97218850e-01 -9.75313008e-01 1.33903906e-01
7.82989979e-01 -7.99596846e-01 1.16946661e+00 -5.39242744e-01
-1.34785231e-02 -4.77040201e-01 -1.61139503e-01 -9.40450847e-01
-5.14734030e-01 -6.78758144e-01 -6.82777390e-02 1.45124912e+00
-1.35511413e-01 -7.12039351e-01 5.94333172e-01 8.53169978e-01
4.86896008e-01 -2.61842251e-01 -1.11180067e+00 -8.88174534e-01
-3.96326542e-01 -7.96793550e-02 7.63207912e-01 9.43324506e-01
-1.28927752e-01 -1.65573329e-01 -1.16296101e+00 2.39302009e-01
8.47946465e-01 1.75007015e-01 6.29302144e-01 -7.75006413e-01
-2.87867278e-01 -3.31260324e-01 -6.62626028e-01 -1.29672611e+00
3.06528658e-01 -9.19076324e-01 -3.09413642e-01 -1.07139969e+00
6.64528847e-01 -3.92722189e-01 -5.52587748e-01 5.67590594e-01
-2.73553044e-01 4.00551409e-01 6.43725038e-01 7.73223877e-01
-8.55548680e-01 6.96357548e-01 1.11993229e+00 -5.62709987e-01
6.00788295e-02 -9.98030826e-02 -5.55184603e-01 3.29136342e-01
4.88374650e-01 -4.61516410e-01 -4.20855820e-01 -6.99902058e-01
-1.16683602e-01 1.86931819e-01 7.13726521e-01 -9.02185261e-01
6.58034682e-01 -1.62371621e-02 6.79490387e-01 -3.34972352e-01
4.91403818e-01 -8.76498044e-01 4.35467273e-01 5.54781020e-01
-2.87950218e-01 1.09790675e-01 3.11165079e-02 8.99142146e-01
-5.25108397e-01 5.15390523e-02 5.58470726e-01 -1.06055088e-01
-9.01486397e-01 1.06706429e+00 8.48158896e-02 -2.06471890e-01
1.17329407e+00 -3.39473933e-01 -3.93621713e-01 -5.13372123e-01
-5.29561400e-01 7.50806630e-01 4.79513019e-01 7.14150369e-01
8.82636905e-01 -1.72283566e+00 -7.94820249e-01 2.11944118e-01
1.94682986e-01 -4.36387479e-01 9.31921363e-01 8.81322861e-01
-2.09243193e-01 3.65844488e-01 -3.44142169e-01 -5.00991404e-01
-1.36734426e+00 8.78765702e-01 5.91492593e-01 8.55488516e-03
-7.13375926e-01 8.72157872e-01 5.53106666e-01 -1.43905640e-01
2.87870377e-01 4.65885192e-01 -2.73354650e-01 6.53273910e-02
9.31411564e-01 4.23838258e-01 -5.25193274e-01 -1.06292546e+00
-5.78144670e-01 8.30258310e-01 -3.73188615e-01 2.66866446e-01
1.07310545e+00 -5.47300994e-01 -1.55207619e-01 1.05457149e-01
1.59054005e+00 -3.87936980e-01 -1.46782398e+00 -4.74559873e-01
-5.15809238e-01 -9.87874448e-01 -1.55905783e-01 -3.09936404e-01
-1.65739405e+00 5.59841096e-01 9.79855418e-01 -1.01026364e-01
1.31040835e+00 -1.39124438e-01 9.86631572e-01 3.35053056e-01
1.99723721e-01 -1.09693503e+00 2.80038595e-01 -2.38792971e-02
7.12875009e-01 -1.44902515e+00 -3.74517366e-02 -2.22430050e-01
-7.44684994e-01 9.74245131e-01 5.47955513e-01 -2.13550538e-01
7.50533938e-01 -3.98077071e-01 -1.41227081e-01 6.41683340e-02
-3.88896793e-01 6.29602671e-02 3.37831259e-01 6.67364657e-01
-1.07366160e-01 -6.07275516e-02 6.28588721e-02 6.67189956e-01
3.15583825e-01 2.74280220e-01 1.32181138e-01 5.10374486e-01
4.77608144e-02 -1.05729115e+00 -3.43426585e-01 3.15135717e-01
-4.29461420e-01 4.48061265e-02 -3.43569890e-02 5.24541795e-01
2.69309908e-01 1.04994833e+00 7.68304244e-02 -7.02093422e-01
6.57028854e-02 -4.22237992e-01 3.93937707e-01 -6.44942597e-02
-2.93130249e-01 -1.19724728e-01 -3.07140768e-01 -7.79603183e-01
-7.50758588e-01 -6.29657328e-01 -7.54193962e-01 -5.74620366e-01
-2.29411826e-01 3.21918994e-01 1.08692557e-01 7.48860419e-01
5.27964413e-01 2.20820576e-01 1.04244685e+00 -8.62329364e-01
-3.39074939e-01 -5.74971616e-01 -6.30083144e-01 7.97487855e-01
5.03590941e-01 -7.22463012e-01 -2.63805449e-01 3.76853496e-01] | [14.643102645874023, 1.000072717666626] |
112f8e92-0174-4d5f-8da5-25c80cbbfc79 | zero-shot-cross-lingual-conversational-1 | null | null | https://openreview.net/forum?id=Qb6XuSBODsW | https://openreview.net/pdf?id=Qb6XuSBODsW | Zero-shot Cross-lingual Conversational Semantic Role Labeling | While conversational semantic role labeling (CSRL) has shown its usefulness on Chinese conversational tasks, it is still under-explored in non-Chinese languages due to the lack of multilingual CSRL annotations for the parser training. To avoid expensive data collection and error-propagation of translation-based methods, we present a simple but effective approach to perform zero-shot cross-lingual CSRL. Our model implicitly learns language-agnostic, conversational structure-aware and semantically rich representations with the hierarchical encoders and elaborately designed pre-training objectives. Experimental results show that our cross-lingual model not only outperforms baselines by large margins but it is also robust to low-resource scenarios. More importantly, we confirm the usefulness of CSRL to English conversational tasks such as question-in-context rewriting and multi-turn dialogue response generation by incorporating the CSRL information into the downstream conversation-based models. We believe this finding is significant and will facilitate the research of English dialogue tasks which suffer the problems of ellipsis and anaphora. | ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['semantic-role-labeling'] | ['natural-language-processing'] | [ 1.85955778e-01 4.38942075e-01 -2.79953778e-01 -7.34488308e-01
-1.39475250e+00 -7.99522281e-01 7.96336710e-01 -4.94245179e-02
-4.62658972e-01 1.03881276e+00 1.09404671e+00 -4.72660214e-01
3.02778155e-01 -4.55221474e-01 -3.40263337e-01 -2.96798348e-01
1.39929458e-01 7.02326119e-01 -5.66116124e-02 -9.09700572e-01
1.10193282e-01 5.93950860e-02 -1.04445744e+00 9.57496047e-01
8.50036383e-01 2.31708363e-01 3.47602248e-01 5.66380918e-01
-3.64056915e-01 1.39057028e+00 -6.79245055e-01 -6.56178534e-01
-1.01954393e-01 -8.05397689e-01 -1.43581390e+00 3.56973172e-03
2.05853079e-02 -2.43994832e-01 -4.89431471e-02 6.24719203e-01
6.06427431e-01 2.47368008e-01 1.26698658e-01 -8.49864542e-01
-7.49725223e-01 8.80479634e-01 -9.81307030e-02 2.12850377e-01
7.50185668e-01 1.28847152e-01 1.53965557e+00 -7.21334815e-01
1.02684569e+00 1.78470612e+00 4.88918424e-01 1.03890944e+00
-1.06836379e+00 -3.39998782e-01 2.12597370e-01 1.83663815e-01
-7.92579830e-01 -6.78997695e-01 7.35967875e-01 -7.98675790e-02
1.44445801e+00 1.93367839e-01 1.43965974e-01 1.32042742e+00
-1.10478796e-01 9.72717404e-01 1.20827234e+00 -6.99208021e-01
-2.36430302e-01 1.92436874e-01 -2.76970747e-03 6.13736272e-01
-3.41484487e-01 -2.74981707e-01 -7.02100456e-01 -9.78169888e-02
4.28794712e-01 -6.13554597e-01 -2.67251283e-01 2.46301323e-01
-1.11709571e+00 1.11429906e+00 9.95672643e-02 7.05048680e-01
-3.90246101e-02 -4.99481335e-02 8.43404830e-01 8.18363905e-01
6.21612847e-01 9.90227520e-01 -6.90712988e-01 -5.33713698e-01
-2.29247957e-01 2.85389423e-01 1.21005452e+00 1.08034611e+00
4.91949558e-01 -1.46204188e-01 -2.29572982e-01 1.37272942e+00
-2.62922905e-02 1.85793266e-01 5.66111445e-01 -1.35272014e+00
9.01447773e-01 7.23522365e-01 9.80899576e-03 -5.27521849e-01
-3.61568123e-01 2.88913418e-02 -1.65067792e-01 -6.46784008e-01
5.64732134e-01 -3.37825000e-01 3.89689580e-02 2.08455229e+00
2.83768386e-01 -3.57082784e-01 7.02804804e-01 9.80391383e-01
9.82365787e-01 6.78275585e-01 2.45807409e-01 -2.78004378e-01
1.62629080e+00 -1.38680577e+00 -9.32548642e-01 -3.48622054e-01
1.09968042e+00 -9.23212469e-01 1.58147550e+00 -2.93593794e-01
-1.22602236e+00 -1.65208668e-01 -7.03259110e-01 -5.86592972e-01
-1.01316705e-01 2.74762698e-02 9.05196488e-01 3.22784722e-01
-9.94326174e-01 1.80353150e-01 -4.53307450e-01 -6.85846150e-01
-1.36416212e-01 -1.15515739e-01 -3.63979101e-01 -1.33361831e-01
-1.77111042e+00 1.37691927e+00 -7.89575800e-02 6.20403327e-02
-5.65944195e-01 -4.51554567e-01 -1.16372740e+00 -1.26573086e-01
4.19689804e-01 -3.53656888e-01 1.82209158e+00 -9.27885532e-01
-2.00451875e+00 1.25888503e+00 -4.58485454e-01 -4.99716610e-01
1.63165241e-01 -2.26188511e-01 -1.92976192e-01 1.15547016e-01
5.06136060e-01 5.20981312e-01 1.53296158e-01 -8.98511648e-01
-6.42184556e-01 -1.41347244e-01 6.84988320e-01 9.26640332e-01
-3.81072573e-02 4.69803274e-01 -2.49325708e-01 -4.37465966e-01
-1.76011890e-01 -1.01411033e+00 -1.06716260e-01 -5.51458180e-01
-1.19966269e-01 -8.31118584e-01 4.57709730e-01 -7.35310733e-01
1.04806137e+00 -1.82144511e+00 1.60520211e-01 -6.41583323e-01
-3.61415207e-01 1.41445681e-01 -3.37279171e-01 1.00095248e+00
2.89172769e-01 1.15640864e-01 -3.33305687e-01 -3.54142278e-01
-1.16134018e-01 4.40652043e-01 -3.14432085e-01 1.45307332e-01
5.85135520e-01 1.02796853e+00 -1.09204328e+00 -4.70260590e-01
6.14803806e-02 1.24691084e-01 -6.24596834e-01 6.73876166e-01
-4.28830504e-01 7.95773208e-01 -4.43817288e-01 5.26651382e-01
2.77782995e-02 -1.09285392e-01 7.75544524e-01 2.56712645e-01
-2.99461037e-01 1.49132895e+00 -3.44342023e-01 2.04723644e+00
-1.00188577e+00 5.85737109e-01 3.19685459e-01 -9.03650701e-01
6.96065784e-01 6.80466354e-01 9.35511887e-02 -9.10682619e-01
-1.17796287e-01 2.57129252e-01 2.45847851e-01 -7.15349197e-01
4.83528793e-01 -4.35630500e-01 -6.57117844e-01 8.95375729e-01
1.38729215e-01 -2.27251470e-01 1.76217124e-01 3.30845296e-01
9.87102330e-01 2.55874038e-01 4.97645468e-01 -4.30794269e-01
9.60586905e-01 3.36378008e-01 5.52241862e-01 4.61312920e-01
-2.52161711e-01 2.44120732e-01 7.21777141e-01 -2.75733083e-01
-6.86990917e-01 -4.98630583e-01 -1.33605435e-01 1.76818562e+00
1.93689857e-02 -4.02541161e-01 -8.92446339e-01 -8.18246484e-01
-2.10818008e-01 1.06139398e+00 -2.86360115e-01 1.46805093e-01
-1.25867105e+00 -5.84424675e-01 8.95431280e-01 2.44337693e-01
5.18188596e-01 -1.51430941e+00 -9.99785066e-02 6.04864299e-01
-7.32729077e-01 -1.57017863e+00 -4.41566706e-01 -7.24999085e-02
-6.00745380e-01 -1.20280468e+00 -2.62164772e-01 -1.01229227e+00
2.43485540e-01 3.64372730e-01 1.30276191e+00 8.26575607e-02
9.83900428e-02 3.99006933e-01 -5.36321521e-01 -8.26317593e-02
-8.76575768e-01 4.05294716e-01 -2.90961057e-01 -1.43912852e-01
7.20285773e-01 -4.15678889e-01 -2.50452399e-01 1.66483909e-01
-2.39226729e-01 7.89829642e-02 2.52542526e-01 9.34322238e-01
1.63900312e-02 -7.59441257e-01 1.16864789e+00 -1.38019454e+00
1.24660778e+00 -4.46957082e-01 -2.05831200e-01 3.55013967e-01
-3.49149674e-01 1.61331475e-01 6.35981083e-01 1.34910485e-02
-1.84481812e+00 -3.38823229e-01 -3.00214916e-01 5.09294093e-01
-6.75982386e-02 4.68006402e-01 -2.69192964e-01 4.32802349e-01
6.90474451e-01 -4.12840471e-02 2.26873696e-01 -5.79737663e-01
6.93847537e-01 8.90052199e-01 4.02793109e-01 -9.67863798e-01
2.64632612e-01 1.54961959e-01 -6.57416701e-01 -8.65169644e-01
-1.25579906e+00 -6.19407117e-01 -4.31933790e-01 1.47331040e-02
1.12681413e+00 -1.25734615e+00 -7.11146772e-01 1.45136163e-01
-1.60346723e+00 -4.93798018e-01 -9.52515844e-03 1.92786545e-01
-5.54558218e-01 3.51292223e-01 -1.29628122e+00 -6.72839642e-01
-5.07354259e-01 -1.01742256e+00 1.03197682e+00 -8.99029970e-02
-6.75583959e-01 -1.23592567e+00 1.49812236e-01 9.75988388e-01
4.29051638e-01 -2.02347696e-01 1.16640079e+00 -9.18218613e-01
-4.82337475e-01 1.52886868e-01 -1.25955075e-01 2.46339858e-01
2.16347814e-01 -6.08687580e-01 -9.96578932e-01 -1.43964380e-01
2.60680765e-01 -9.07977879e-01 3.78638655e-01 -1.36696115e-01
4.62440550e-01 -6.73107028e-01 7.18348399e-02 8.53444859e-02
9.39377904e-01 1.39504552e-01 2.91400075e-01 3.63448530e-01
5.73433518e-01 1.30132711e+00 7.57983148e-01 -1.04870960e-01
1.06660092e+00 8.54861140e-01 -8.95883590e-02 5.28422324e-03
-3.20548326e-01 -2.87845582e-01 5.46283066e-01 1.33775985e+00
1.28052011e-01 -1.50031611e-01 -6.64774001e-01 6.76510096e-01
-1.98699617e+00 -9.12722588e-01 -2.29430035e-01 1.70714676e+00
1.53053558e+00 -2.30427623e-01 -1.28537044e-01 -5.57304084e-01
7.16697752e-01 3.80766809e-01 -2.00047374e-01 -9.32763040e-01
-2.28205860e-01 1.72575220e-01 -1.31478831e-01 1.07067394e+00
-8.75065327e-01 1.71189749e+00 5.82784128e+00 3.08986783e-01
-8.10119748e-01 6.44696176e-01 3.75516087e-01 2.34181091e-01
-3.68020892e-01 2.04130068e-01 -9.17069852e-01 1.52199715e-01
1.03068161e+00 -2.21060261e-01 5.84034503e-01 8.68900478e-01
2.83013552e-01 -9.33744851e-03 -1.26938510e+00 4.99775559e-01
1.83367163e-01 -1.30106235e+00 -2.52389342e-01 -3.19271922e-01
5.96747398e-01 1.06154449e-01 -6.62500322e-01 6.91675603e-01
6.24614179e-01 -8.73997629e-01 2.22380981e-01 -2.02098697e-01
7.21900463e-01 -5.15757978e-01 7.75655925e-01 4.02936250e-01
-9.17488217e-01 2.85125766e-02 -3.30498844e-01 -3.75364482e-01
4.08658534e-01 -7.20956177e-02 -1.29286599e+00 3.16849321e-01
2.23000750e-01 6.84419811e-01 -1.57569230e-01 2.33053695e-02
-7.71977305e-01 7.82319963e-01 1.28214017e-01 -2.53387034e-01
4.78208363e-01 -2.60156453e-01 5.58724463e-01 1.64144909e+00
-2.89681077e-01 2.64980108e-01 3.81529540e-01 6.52397096e-01
-3.13438296e-01 4.30412829e-01 -7.36864984e-01 -8.68598148e-02
7.71166027e-01 1.13739324e+00 -2.16094047e-01 -2.18158558e-01
-7.17658579e-01 1.02670097e+00 8.14679980e-01 2.05775648e-01
-3.85880440e-01 -3.07090711e-02 8.86614799e-01 1.76658668e-02
-2.88194805e-01 -2.54734963e-01 -1.74408928e-01 -1.26532781e+00
-6.19649291e-02 -1.35008192e+00 5.06034374e-01 -4.12330091e-01
-1.58047986e+00 8.24502707e-01 -1.02768198e-01 -8.20459485e-01
-9.42711353e-01 -5.13418794e-01 -6.41539216e-01 9.28094923e-01
-1.78547800e+00 -1.45067644e+00 2.47752160e-01 5.28748155e-01
1.34330499e+00 -3.07768404e-01 1.31936479e+00 2.31443897e-01
-4.76633638e-01 5.17551124e-01 -3.89342129e-01 3.17146689e-01
1.01140869e+00 -1.19765496e+00 4.69196260e-01 6.40143335e-01
-1.32519379e-01 8.32589149e-01 5.33629537e-01 -4.37122136e-01
-1.19720602e+00 -8.20116520e-01 1.80957365e+00 -6.68394268e-01
9.70100224e-01 -5.83090723e-01 -1.12055695e+00 7.97141790e-01
7.89597869e-01 -5.64470649e-01 8.82049143e-01 5.84318697e-01
-2.42027789e-01 1.47870585e-01 -8.78846645e-01 6.56953037e-01
1.08178449e+00 -1.01919603e+00 -1.18307185e+00 6.18651390e-01
1.10256624e+00 -4.88052338e-01 -7.85755217e-01 1.07818305e-01
1.87005952e-01 -6.51942372e-01 6.17637575e-01 -9.15384412e-01
6.52818859e-01 1.50414020e-01 -1.65629461e-01 -1.17630482e+00
-1.02217376e-01 -8.20568204e-01 3.17008317e-01 1.41473246e+00
7.50064135e-01 -5.21023452e-01 3.05638194e-01 7.29738474e-01
-4.51320380e-01 -5.24157763e-01 -9.75181341e-01 -4.07917678e-01
3.68460208e-01 -2.52814561e-01 3.93108577e-01 1.27734101e+00
6.66568637e-01 1.19780433e+00 -3.97685975e-01 -2.46237919e-01
1.34803236e-01 3.35041314e-01 7.61309147e-01 -9.60466623e-01
-3.07996213e-01 -2.11779401e-01 3.82189363e-01 -1.04783738e+00
9.55428302e-01 -1.12415564e+00 2.65834391e-01 -1.47239888e+00
-1.26829986e-02 -5.85583627e-01 2.58202374e-01 4.89150494e-01
-3.33349735e-01 -1.34683520e-01 9.91430357e-02 2.89084643e-01
-7.77964830e-01 6.79810941e-01 1.32545888e+00 2.73869544e-01
-2.91320920e-01 2.09700107e-03 -1.06386387e+00 6.33993506e-01
6.85143650e-01 -5.07664621e-01 -4.78669316e-01 -5.78107595e-01
-3.96506339e-02 4.46991265e-01 -1.90096140e-01 -4.70485501e-02
5.83413541e-02 -2.69601762e-01 -5.28815448e-01 2.99262106e-02
3.33426028e-01 -2.17574254e-01 -6.71355426e-01 6.86422810e-02
-9.56802785e-01 1.27486616e-01 6.22469857e-02 3.19004536e-01
-3.95225912e-01 -3.93705428e-01 5.47146738e-01 -6.34506941e-01
-6.28535390e-01 -3.43186557e-01 -8.43502402e-01 6.69008970e-01
5.82327425e-01 2.63088346e-01 -4.84426945e-01 -8.72405291e-01
-4.38943923e-01 4.93944108e-01 1.95899248e-01 8.57056797e-01
9.14745480e-02 -1.03027582e+00 -9.77206886e-01 -3.53395082e-02
2.89290816e-01 -4.51731589e-03 -3.43683288e-02 5.76127946e-01
-3.81143332e-01 8.00814629e-01 4.77802008e-02 -3.31043303e-01
-1.16927314e+00 8.53036344e-02 2.90020227e-01 -6.13131225e-01
-6.60020769e-01 8.44224691e-01 1.98914140e-01 -1.16404808e+00
1.92142919e-01 1.36465892e-01 -3.74193579e-01 -9.74663813e-03
4.45259333e-01 -2.65835896e-02 5.36610410e-02 -7.36599326e-01
-3.45734239e-01 -9.29722488e-02 -2.21946433e-01 -4.13788557e-01
1.21014321e+00 -4.84757692e-01 -4.86909777e-01 5.00878394e-01
1.04464936e+00 1.49055377e-01 -1.03203189e+00 -4.79979187e-01
6.15703166e-01 -5.10049611e-02 -4.97955263e-01 -8.65200639e-01
-3.58409017e-01 8.63451719e-01 -2.54695475e-01 7.45959282e-02
5.92047393e-01 2.92996734e-01 8.77320647e-01 7.43660152e-01
4.28952217e-01 -1.26825213e+00 2.13624135e-01 1.12224901e+00
1.27084410e+00 -1.29585147e+00 -3.26952875e-01 -6.30768955e-01
-1.22046101e+00 8.75372469e-01 1.05347013e+00 5.90430684e-02
1.29029214e-01 2.04615057e-01 5.72757423e-01 -1.85415030e-01
-1.34034526e+00 -3.25352401e-01 -1.57881126e-01 4.36159998e-01
1.32926691e+00 8.68351907e-02 -8.83529425e-01 5.87482691e-01
-4.06876683e-01 -4.89659399e-01 5.33309519e-01 8.17910790e-01
-1.82154179e-01 -1.64536250e+00 1.21384911e-01 -1.53349251e-01
-7.03264594e-01 -5.19465268e-01 -8.12621713e-01 9.78149116e-01
-4.35332179e-01 1.18820262e+00 4.20728959e-02 1.28048286e-01
3.81296694e-01 5.50588131e-01 2.43502885e-01 -1.20749974e+00
-1.04836690e+00 -6.25473866e-03 1.14890778e+00 -5.94484687e-01
-6.19375288e-01 -6.92117333e-01 -1.40001392e+00 -1.25416398e-01
-1.55402079e-01 4.16823596e-01 4.58484590e-01 1.20067585e+00
4.14135426e-01 2.31646612e-01 7.02675998e-01 -4.60418761e-01
-7.35540092e-01 -1.34279728e+00 7.20286146e-02 6.36319220e-01
-1.33057861e-02 -3.34066540e-01 -3.08737099e-01 -1.33259028e-01] | [12.400540351867676, 8.112032890319824] |
fc489ab6-8fe1-4b24-8e7f-f9dfa3015ee8 | cgof-controllable-3d-face-synthesis-with | 2211.13251 | null | https://arxiv.org/abs/2211.13251v1 | https://arxiv.org/pdf/2211.13251v1.pdf | CGOF++: Controllable 3D Face Synthesis with Conditional Generative Occupancy Fields | Capitalizing on the recent advances in image generation models, existing controllable face image synthesis methods are able to generate high-fidelity images with some levels of controllability, e.g., controlling the shapes, expressions, textures, and poses of the generated face images. However, previous methods focus on controllable 2D image generative models, which are prone to producing inconsistent face images under large expression and pose changes. In this paper, we propose a new NeRF-based conditional 3D face synthesis framework, which enables 3D controllability over the generated face images by imposing explicit 3D conditions from 3D face priors. At its core is a conditional Generative Occupancy Field (cGOF++) that effectively enforces the shape of the generated face to conform to a given 3D Morphable Model (3DMM) mesh, built on top of EG3D [1], a recent tri-plane-based generative model. To achieve accurate control over fine-grained 3D face shapes of the synthesized images, we additionally incorporate a 3D landmark loss as well as a volume warping loss into our synthesis framework. Experiments validate the effectiveness of the proposed method, which is able to generate high-fidelity face images and shows more precise 3D controllability than state-of-the-art 2D-based controllable face synthesis methods. | ['Hongsheng Li', 'Quan Wang', 'Zhaoyang Huang', 'Ning Zhang', 'Shangzhe Wu', 'Keqiang Sun'] | 2022-11-23 | null | null | null | null | ['face-generation'] | ['computer-vision'] | [ 2.70208687e-01 4.34267223e-01 1.97667077e-01 -2.79995322e-01
-2.87777424e-01 -3.80091757e-01 8.70931506e-01 -6.14605188e-01
2.92407334e-01 6.68818533e-01 1.04492173e-01 3.65604460e-01
-3.50107066e-02 -1.12619972e+00 -9.15133119e-01 -8.69928896e-01
3.71858895e-01 6.50551260e-01 -2.04125747e-01 -1.33584559e-01
6.75408728e-03 9.63734865e-01 -1.82010996e+00 -3.10148895e-02
8.24719727e-01 9.19990420e-01 -1.45844743e-01 3.18159759e-01
9.02828481e-03 2.49816447e-01 -2.89419085e-01 -2.91452885e-01
1.05757095e-01 -5.19473970e-01 -1.92108437e-01 6.40484691e-01
4.76141214e-01 -2.02830121e-01 -1.44054249e-01 8.69174421e-01
3.92034322e-01 -1.86299086e-02 9.71504450e-01 -1.23700309e+00
-9.01545882e-01 -7.31045678e-02 -4.94419336e-01 -6.76855922e-01
3.77204925e-01 3.08953404e-01 3.52264702e-01 -1.04093337e+00
9.28325832e-01 1.62510192e+00 3.13244641e-01 9.28328514e-01
-1.59372675e+00 -7.27043569e-01 1.12045191e-01 -5.10670364e-01
-1.54723942e+00 -6.19743764e-01 1.09650564e+00 -6.07769608e-01
5.26298344e-01 3.30522031e-01 7.01518297e-01 1.15989137e+00
3.97981554e-01 2.20904246e-01 1.24371898e+00 -4.25843328e-01
2.75731832e-01 -5.49081247e-03 -6.87496960e-01 8.84355843e-01
-1.20515805e-02 2.78503448e-01 -5.20980537e-01 -1.28861293e-01
1.44141996e+00 -1.15795754e-01 -4.21818197e-01 -6.30021632e-01
-9.99604762e-01 8.05640996e-01 2.75444448e-01 -2.23840661e-02
-3.38087738e-01 7.26724491e-02 -2.20292464e-01 -2.10569933e-01
8.03056121e-01 3.84614795e-01 2.30322685e-02 3.38890731e-01
-8.65060985e-01 6.35866404e-01 5.35484731e-01 1.09929478e+00
9.32437360e-01 4.00019169e-01 -7.37364233e-01 6.06199086e-01
5.73439360e-01 8.63415956e-01 -1.12208493e-01 -9.94538605e-01
-8.47748220e-02 7.05655456e-01 9.54068378e-02 -1.15698397e+00
8.73866454e-02 -2.66210020e-01 -8.85466933e-01 4.46942836e-01
-1.54723927e-01 9.49039832e-02 -1.39699566e+00 2.03935623e+00
6.65028095e-01 2.26096809e-01 -3.68364513e-01 8.28409433e-01
6.46137357e-01 5.47177494e-01 -1.98944267e-02 -5.83141029e-01
9.06891882e-01 -4.01406914e-01 -8.94299269e-01 1.46156251e-01
-2.79292613e-01 -6.72737122e-01 1.00384223e+00 3.84238991e-03
-1.37925780e+00 -5.14430761e-01 -8.62726748e-01 1.42208621e-01
4.36999053e-02 -1.22482747e-01 4.54308361e-01 5.46919942e-01
-1.16634142e+00 2.39494160e-01 -8.29601765e-01 7.96283409e-03
6.05332553e-01 3.56395125e-01 -5.70902944e-01 -2.86840368e-02
-9.70489144e-01 6.87438369e-01 -1.13340706e-01 2.21832126e-01
-1.34384847e+00 -9.65643048e-01 -1.07189345e+00 -6.54564947e-02
2.21568629e-01 -9.52623427e-01 6.77620947e-01 -6.69177711e-01
-1.97427452e+00 1.08181894e+00 -1.69685304e-01 3.10430288e-01
6.19774759e-01 1.56752408e-01 -6.89364672e-02 8.79910961e-03
-9.12000705e-03 9.87610519e-01 1.39997685e+00 -1.75066018e+00
2.81157672e-01 -4.23561126e-01 -1.76559255e-01 1.37985975e-01
-2.65473366e-01 -2.22135738e-01 -5.85420430e-01 -7.55703330e-01
9.86505649e-04 -1.01641011e+00 -2.10919797e-01 3.36960584e-01
-4.71101344e-01 1.02196649e-01 1.06582248e+00 -2.34286368e-01
8.60698760e-01 -2.00795984e+00 6.74741626e-01 3.42414439e-01
1.55753940e-01 1.78195074e-01 -1.52861223e-01 2.65339434e-01
3.53465974e-02 2.25536391e-01 -4.95041907e-01 -7.21451342e-01
1.19456358e-01 3.75895530e-01 -2.39839554e-01 4.75318372e-01
5.81959188e-01 8.30651522e-01 -5.28946221e-01 -4.52807873e-01
4.70980257e-01 1.20974088e+00 -1.03976250e+00 4.46270466e-01
-5.75642288e-01 1.04833257e+00 -6.36282682e-01 5.77118397e-01
9.28597510e-01 9.13416296e-02 1.76644355e-01 -1.14209324e-01
-2.02441067e-01 -3.57056618e-01 -9.41610217e-01 1.50643289e+00
-5.97600937e-01 -1.26447335e-01 5.04931152e-01 -5.12704968e-01
1.17531323e+00 4.46804881e-01 4.54883337e-01 -3.29069912e-01
3.20962608e-01 7.66296238e-02 -4.45363462e-01 -1.10203400e-02
4.31529246e-02 -3.85141879e-01 1.06096792e-03 7.73408934e-02
6.67546913e-02 -1.03272808e+00 -1.52089730e-01 -2.32421178e-02
4.63573754e-01 3.47675383e-01 -2.14868292e-01 -5.59566081e-01
6.77369654e-01 -5.98165691e-01 5.52093148e-01 1.63351357e-01
3.23824883e-01 1.01667488e+00 4.80285704e-01 -1.57900766e-01
-1.19021630e+00 -1.09313345e+00 -4.73187447e-01 1.23931840e-01
4.12876867e-02 -1.79333374e-01 -1.12966883e+00 -3.23558033e-01
5.55374958e-02 4.56707507e-01 -8.43177438e-01 -2.37064630e-01
-5.43030918e-01 -5.91532826e-01 9.03894603e-02 1.41218290e-01
6.05416775e-01 -1.08657539e+00 -3.38555843e-01 8.42645764e-02
7.91153088e-02 -1.05804753e+00 -6.49193406e-01 -4.44802493e-01
-5.21779597e-01 -8.26962471e-01 -5.81505895e-01 -6.66432321e-01
1.20903051e+00 -4.16714758e-01 9.66433167e-01 1.88499779e-01
-4.70614672e-01 4.69016820e-01 -2.13172324e-02 -1.35492980e-01
-5.80621958e-01 -2.47194856e-01 2.58569121e-01 5.20851851e-01
-4.28591520e-01 -8.27426910e-01 -5.59392035e-01 4.75250810e-01
-1.13739920e+00 2.92132974e-01 5.22913858e-02 8.91372144e-01
1.03364635e+00 1.72826082e-01 4.22194481e-01 -6.94548249e-01
3.58858794e-01 -7.12644607e-02 -7.43954122e-01 3.86143290e-02
-5.76888978e-01 4.09337021e-02 3.45857054e-01 -4.99602824e-01
-1.28733993e+00 2.66396552e-01 -1.89556777e-01 -9.09863532e-01
-8.59336928e-02 -6.43851608e-02 -7.80630529e-01 -3.02704275e-01
3.66127104e-01 1.93320587e-01 3.63314748e-01 -2.77984977e-01
4.97167259e-01 3.99732411e-01 3.31001580e-01 -9.45427835e-01
1.12372053e+00 6.02710366e-01 3.59202862e-01 -7.60284960e-01
-4.05013353e-01 4.63772178e-01 -6.68939531e-01 -5.15128374e-01
1.09265947e+00 -7.24188089e-01 -7.75024652e-01 7.61681139e-01
-1.01260436e+00 -3.57799470e-01 -1.87355220e-01 8.35070759e-02
-8.33818853e-01 2.94754561e-02 -3.67053449e-01 -8.15764725e-01
-2.51525104e-01 -1.46800745e+00 1.64557993e+00 1.30493477e-01
-1.51507914e-01 -6.87938511e-01 -1.31058216e-01 2.79585242e-01
6.35380805e-01 1.04674554e+00 1.04898882e+00 5.67335308e-01
-8.21241081e-01 1.18012272e-03 1.71155468e-01 2.49843642e-01
3.78896445e-01 3.61793220e-01 -8.50262105e-01 -3.25142562e-01
1.01138644e-01 -2.10292861e-01 5.26911497e-01 4.74033982e-01
1.04351151e+00 -3.83911550e-01 -4.37918693e-01 8.05089116e-01
1.29013026e+00 1.30019605e-01 8.09568167e-01 -3.56521457e-01
8.13865960e-01 6.26102924e-01 3.73477906e-01 6.20626569e-01
6.01360202e-02 1.03636456e+00 6.92947567e-01 -8.72037858e-02
-3.95605356e-01 -6.76794887e-01 1.69155359e-01 4.17461574e-01
-1.91926315e-01 -2.70794630e-01 -5.18873513e-01 2.04036862e-01
-1.62667549e+00 -7.09444821e-01 1.85148388e-01 2.19993043e+00
9.17698443e-01 -1.06256753e-01 -2.57386833e-01 -6.44030720e-02
8.03618610e-01 1.74020469e-01 -5.69829166e-01 -1.82953030e-01
-1.06367311e-02 5.24567127e-01 -3.48923095e-02 7.42579639e-01
-8.59473825e-01 1.00821936e+00 6.29157209e+00 8.25506270e-01
-1.23697913e+00 -9.79190990e-02 7.93242097e-01 -1.50988415e-01
-9.73215938e-01 -9.95736718e-02 -8.86538148e-01 4.77844924e-01
3.29649568e-01 -6.37182454e-03 3.69236082e-01 6.70856953e-01
4.00398612e-01 3.41193736e-01 -8.83425772e-01 7.83115149e-01
1.80375800e-01 -1.53483486e+00 5.50795078e-01 3.50718021e-01
1.09759104e+00 -8.33658278e-01 3.83006752e-01 -6.56486228e-02
5.24767116e-02 -1.32695675e+00 9.18648779e-01 8.00675929e-01
1.46523809e+00 -1.01738882e+00 1.01956390e-02 2.20352307e-01
-1.03316367e+00 3.41695011e-01 -5.34216724e-02 2.23549619e-01
3.86327535e-01 7.35744178e-01 -5.24541438e-01 2.89529622e-01
5.35004795e-01 4.68966722e-01 -1.13594078e-01 2.23341227e-01
-5.08861959e-01 1.70374617e-01 -2.58914113e-01 2.89839685e-01
-1.18342511e-01 -4.36808407e-01 7.85225570e-01 4.97158051e-01
4.24854785e-01 2.23494500e-01 1.85453489e-01 1.51481509e+00
-1.27728313e-01 -1.84265763e-01 -7.59561658e-01 7.62725854e-03
5.01762807e-01 1.22441256e+00 -5.38136721e-01 1.21454045e-01
1.16670333e-01 8.24675798e-01 2.01999366e-01 2.67371684e-01
-9.71818626e-01 1.93942502e-01 9.56594110e-01 6.97366536e-01
3.03390622e-01 -4.29523498e-01 -1.29772589e-01 -1.13725376e+00
-1.09376460e-02 -8.48771513e-01 -3.35982084e-01 -7.83841848e-01
-1.08436096e+00 7.31848598e-01 7.52718002e-02 -8.96260023e-01
-2.47080669e-01 -4.12925631e-01 -4.84849095e-01 8.88037384e-01
-1.38579834e+00 -1.62635195e+00 -2.24742219e-01 7.33408093e-01
1.87525302e-01 -3.52923907e-02 9.86085355e-01 1.43401787e-01
-4.99513507e-01 5.68510056e-01 -5.56832433e-01 -2.07274243e-01
5.36872506e-01 -6.87275112e-01 2.76643425e-01 6.87252283e-01
-2.79645443e-01 6.67687356e-01 5.93919814e-01 -7.71365821e-01
-1.79305434e+00 -1.29854643e+00 4.76048440e-01 -6.51523888e-01
-4.25536446e-02 -7.13209450e-01 -6.86145306e-01 5.45788229e-01
5.10869958e-02 2.79305935e-01 1.97810918e-01 -4.08621609e-01
-1.82408750e-01 -2.94314269e-02 -1.61085808e+00 6.35186434e-01
1.45773637e+00 -3.84628624e-01 7.84556847e-03 2.05955118e-01
6.91763341e-01 -6.06130540e-01 -9.55342531e-01 9.04857039e-01
4.54819471e-01 -8.66994739e-01 9.32260752e-01 -1.38903230e-01
4.13239479e-01 -5.27083516e-01 -4.97294441e-02 -1.35044336e+00
-3.71965885e-01 -9.82032359e-01 5.01847081e-02 1.46661997e+00
9.17513669e-02 -4.97532457e-01 6.18106306e-01 8.04735780e-01
-2.22156912e-01 -7.17391849e-01 -1.03515935e+00 -5.87272584e-01
3.96465920e-02 -4.66142073e-02 9.77857888e-01 7.96792209e-01
-3.93390864e-01 3.61192971e-02 -4.04663324e-01 2.25611985e-01
7.12455869e-01 1.77579254e-01 7.43401945e-01 -1.07795274e+00
-3.42584103e-02 -3.26830864e-01 -3.98292899e-01 -6.20523393e-01
5.13479948e-01 -7.76233494e-01 1.02269657e-01 -1.18210649e+00
-1.27249886e-03 -5.18237591e-01 4.07190233e-01 4.52899605e-01
5.60340583e-02 4.36401606e-01 3.13910353e-03 -1.10128291e-01
1.70206413e-01 1.17160714e+00 1.95999873e+00 4.29422371e-02
-8.88206288e-02 -3.25258106e-01 -5.38430572e-01 4.98955429e-01
3.34929347e-01 -1.23397604e-01 -6.65862799e-01 -3.06035966e-01
-1.31740347e-01 1.70289919e-01 3.79468679e-01 -7.46426046e-01
-2.54659265e-01 -3.61982465e-01 5.43976963e-01 -1.73066050e-01
6.42592430e-01 -7.31713295e-01 7.06143856e-01 1.52743414e-01
-6.80381358e-02 -2.87316591e-01 1.81017175e-01 4.05810982e-01
-2.52427220e-01 4.58088785e-01 1.29408443e+00 -1.54850140e-01
-9.14786756e-02 9.95602965e-01 -7.71997347e-02 -1.37837142e-01
1.11763811e+00 -5.92200905e-02 6.64862990e-02 -2.94207066e-01
-5.87757945e-01 -9.79130864e-02 8.71996641e-01 4.46526110e-01
7.62072980e-01 -1.72368085e+00 -7.45273471e-01 9.69463289e-01
-8.00635014e-03 4.74170417e-01 3.02942425e-01 3.54165137e-01
-4.10848022e-01 1.48860574e-01 -2.47897238e-01 -7.52171278e-01
-9.54819798e-01 4.81571883e-01 4.65764135e-01 6.31713793e-02
-5.21398664e-01 7.37007141e-01 5.71434855e-01 -6.43732190e-01
-1.52981952e-01 -4.28133905e-02 -7.37546897e-03 -2.57684469e-01
2.80078620e-01 -1.15679987e-01 -4.39680330e-02 -8.37055683e-01
-3.35599124e-01 1.12345338e+00 2.68900245e-01 -2.23531738e-01
1.32102549e+00 1.65022343e-01 -2.51779944e-01 -2.19764680e-01
9.21691000e-01 6.83583468e-02 -1.88821292e+00 3.75806570e-01
-8.56032491e-01 -7.98957527e-01 1.40935913e-01 -4.58506346e-01
-1.43083251e+00 6.17672622e-01 2.91807294e-01 -2.88933516e-01
1.13059056e+00 -1.15332872e-01 5.52728295e-01 -5.05186319e-01
6.40782535e-01 -7.27567434e-01 2.45448709e-01 2.57649481e-01
1.37966919e+00 -7.50693142e-01 -2.07270011e-01 -8.38819504e-01
-3.72454375e-01 7.96915174e-01 7.48544633e-01 -1.58603445e-01
8.34391892e-01 5.85683882e-01 -1.85344413e-01 -3.22172016e-01
-5.60883224e-01 2.55168796e-01 3.64224881e-01 7.31064558e-01
3.10323417e-01 5.02735302e-02 -1.33688703e-01 2.09526509e-01
-3.64967212e-02 8.31249505e-02 1.73138335e-01 6.23214781e-01
1.94173697e-02 -1.25721109e+00 -3.70291114e-01 -1.10858627e-01
1.89753007e-02 1.28712356e-01 -2.54693866e-01 7.33226895e-01
3.51883084e-01 7.35545158e-01 1.79798171e-01 -2.27651343e-01
4.04469937e-01 1.03148827e-02 8.79243553e-01 -7.96518624e-01
-1.58575997e-01 2.31125355e-01 -2.58299828e-01 -4.62006718e-01
-4.01129305e-01 -6.00195646e-01 -1.16959143e+00 -3.38673830e-01
-2.48790279e-01 -1.12307981e-01 4.39150900e-01 6.56697154e-01
7.74198055e-01 3.16832125e-01 9.03019428e-01 -1.31143868e+00
-1.42173901e-01 -7.31401026e-01 -6.11380875e-01 4.41432655e-01
1.17384061e-01 -1.16116345e+00 -3.61420453e-01 1.33290216e-01] | [12.630338668823242, -0.3666478991508484] |
fe696415-ab5e-4432-8755-30dbfdea7b90 | enhancing-feature-invariance-with-learned | 2002.01642 | null | https://arxiv.org/abs/2002.01642v4 | https://arxiv.org/pdf/2002.01642v4.pdf | Learning Test-time Augmentation for Content-based Image Retrieval | Off-the-shelf convolutional neural network features achieve outstanding results in many image retrieval tasks. However, their invariance to target data is pre-defined by the network architecture and training data. Existing image retrieval approaches require fine-tuning or modification of pre-trained networks to adapt to variations unique to the target data. In contrast, our method enhances the invariance of off-the-shelf features by aggregating features extracted from images augmented at test-time, with augmentations guided by a policy learned through reinforcement learning. The learned policy assigns different magnitudes and weights to the selected transformations, which are selected from a list of image transformations. Policies are evaluated using a metric learning protocol to learn the optimal policy. The model converges quickly and the cost of each policy iteration is minimal as we propose an off-line caching technique to greatly reduce the computational cost of extracting features from augmented images. Experimental results on large trademark retrieval (METU trademark dataset) and landmark retrieval (ROxford5k and RParis6k scene datasets) tasks show that the learned ensemble of transformations is highly effective for improving performance, and is practical, and transferable. | ['Clinton Fookes', 'Sridha Sridharan', 'Simon Denman', 'Osman Tursun'] | 2020-02-05 | null | null | null | null | ['trademark-retrieval', 'content-based-image-retrieval'] | ['computer-vision', 'computer-vision'] | [ 1.39618158e-01 -6.84556007e-01 -3.85111183e-01 -7.46015489e-01
-1.02996504e+00 -7.35930443e-01 7.19061017e-01 -1.23624668e-01
-9.05356407e-01 4.32300329e-01 6.55259192e-02 -1.44790777e-03
-4.69340771e-01 -6.54185295e-01 -8.32513213e-01 -6.87131703e-01
-2.06433728e-01 4.61598516e-01 7.39205554e-02 -2.02876925e-01
3.68237942e-01 9.24258053e-01 -1.71019256e+00 3.48782212e-01
5.51262915e-01 1.43590617e+00 2.42468864e-01 6.29027784e-01
2.74045728e-02 4.66018319e-01 -3.67920458e-01 -7.69286379e-02
8.28454614e-01 -6.55491054e-02 -6.90241516e-01 6.63307905e-02
7.91213036e-01 -5.00274420e-01 -5.99155962e-01 9.73346472e-01
5.45543313e-01 4.33151960e-01 6.96568191e-01 -1.32920027e+00
-1.00373220e+00 1.52496219e-01 -3.41075301e-01 2.49901250e-01
-1.74820468e-01 5.66428341e-02 1.07422376e+00 -1.03784525e+00
6.30650818e-01 9.95655000e-01 4.46476340e-01 3.42160612e-01
-1.01682198e+00 -8.01864624e-01 1.64053082e-01 4.22804713e-01
-1.43008029e+00 -5.37135422e-01 5.67549706e-01 -8.74075517e-02
1.01619053e+00 1.56460404e-01 4.27524209e-01 7.06034780e-01
3.06705236e-01 6.97067738e-01 9.21229899e-01 -3.82286668e-01
3.16482931e-01 2.94762533e-02 -1.59019053e-01 7.67267764e-01
1.00970030e-01 2.71502674e-01 -6.44152462e-01 -3.00303549e-01
7.50168681e-01 2.04871804e-01 -1.80741727e-01 -9.58281159e-01
-1.23820937e+00 7.87123024e-01 7.93958604e-01 1.48466066e-01
-5.19239366e-01 4.36241537e-01 4.19040501e-01 8.01258504e-01
3.23719442e-01 6.67451918e-01 -9.07292962e-01 1.05604984e-01
-7.05827773e-01 1.44180626e-01 4.61146802e-01 9.45258915e-01
1.28804743e+00 -2.20085736e-02 -2.24007532e-01 1.15181279e+00
1.56596407e-01 7.98869789e-01 8.28732550e-01 -1.03411484e+00
1.22877039e-01 5.26585996e-01 3.06919795e-02 -8.68335366e-01
-1.75657541e-01 -3.46566468e-01 -5.94395459e-01 1.24604486e-01
1.73275456e-01 1.00306690e-01 -1.37023830e+00 1.61047995e+00
3.40961039e-01 1.31446704e-01 -3.85224633e-02 7.55480230e-01
4.86714780e-01 5.68189740e-01 -8.99049491e-02 2.80654222e-01
9.34590220e-01 -1.05963922e+00 -2.61303037e-01 -2.65775949e-01
8.08228076e-01 -9.14894819e-01 1.22971332e+00 9.45208445e-02
-8.00097346e-01 -5.09967744e-01 -1.05744684e+00 -8.98168795e-03
-7.47928143e-01 1.98418453e-01 6.40406489e-01 2.94953763e-01
-1.25450099e+00 5.77513337e-01 -6.47739589e-01 -4.13898826e-01
4.76530939e-01 9.58409429e-01 -5.96760213e-01 -3.38933319e-01
-9.94908452e-01 9.12806749e-01 5.48826039e-01 1.30106375e-01
-9.58124459e-01 -7.36471951e-01 -8.48499000e-01 2.44180292e-01
3.42134237e-01 -3.76287431e-01 1.29567504e+00 -1.17126894e+00
-1.63194180e+00 7.00996757e-01 2.82048970e-01 -4.03705955e-01
2.28079364e-01 -3.55868191e-01 -2.34643191e-01 4.30725724e-01
1.66222304e-02 1.06187928e+00 1.20234311e+00 -1.01600718e+00
-8.76712859e-01 -1.98186845e-01 -1.17174285e-02 2.74980128e-01
-8.59845519e-01 1.05256423e-01 -6.92580163e-01 -5.48108459e-01
-8.35398361e-02 -1.20179844e+00 -1.84399500e-01 4.19320703e-01
1.48132324e-01 -8.62677544e-02 1.06794906e+00 -1.82497934e-01
8.15616667e-01 -2.16943336e+00 -1.67175129e-01 7.10445106e-01
-1.89064294e-01 3.98329049e-01 -9.34620321e-01 2.41273567e-01
-3.09551824e-02 -1.82253093e-01 3.03334673e-03 8.44675824e-02
-2.07269173e-02 3.71473283e-01 -4.50804085e-01 5.28323352e-01
2.35067651e-01 1.06314564e+00 -8.47660124e-01 -4.23200577e-01
2.96378374e-01 4.93375301e-01 -6.30223930e-01 2.60594934e-01
-1.71665087e-01 -1.83787383e-02 -6.09350562e-01 6.75747693e-01
4.33877259e-01 -4.86075550e-01 -1.26782447e-01 -2.83255547e-01
3.21083307e-01 -7.02258870e-02 -1.10826600e+00 1.81152320e+00
-5.97198248e-01 4.86962438e-01 -3.06397080e-01 -9.61109519e-01
1.03407323e+00 -1.39172003e-01 6.10046864e-01 -9.77593184e-01
8.36368427e-02 3.38705212e-01 4.50628102e-02 -2.64963478e-01
5.71292162e-01 6.33129478e-01 -3.63591760e-02 7.07745612e-01
3.65604937e-01 -7.27105215e-02 2.38799691e-01 1.26684040e-01
1.18730581e+00 1.69580549e-01 2.46029109e-01 -2.67150849e-01
3.96940202e-01 6.67955652e-02 3.62195283e-01 1.04131734e+00
-1.16176210e-01 5.96148968e-01 -8.61906484e-02 -8.56000841e-01
-1.16095793e+00 -8.17593634e-01 -1.17568359e-01 1.58147371e+00
9.57608074e-02 -2.71923274e-01 -4.41019714e-01 -8.72649729e-01
2.84541994e-01 1.62361339e-01 -9.14889932e-01 -5.44357836e-01
-7.37183869e-01 -5.79743505e-01 3.28064799e-01 6.23103917e-01
7.71245658e-01 -1.32067096e+00 -6.91647887e-01 1.51276052e-01
3.29700232e-01 -1.04606771e+00 -7.91360617e-01 2.74471462e-01
-6.40594482e-01 -1.13436043e+00 -6.40205324e-01 -9.02230561e-01
1.05314708e+00 6.28242731e-01 9.12952602e-01 3.26443791e-01
-3.79494876e-01 8.00441802e-01 -3.43701333e-01 -2.79096186e-01
-4.82935794e-02 2.15548411e-01 1.50023922e-01 1.69127628e-01
4.45329517e-01 -1.38194501e-01 -7.51849234e-01 4.84140277e-01
-1.11068130e+00 -4.29658562e-01 8.55296969e-01 1.25280201e+00
8.14588547e-01 -1.84986070e-01 2.68009275e-01 -5.67836285e-01
5.53918779e-01 -1.55369073e-01 -8.31042886e-01 6.50336385e-01
-8.01438451e-01 3.58257055e-01 5.59582293e-01 -8.10323060e-01
-9.28460002e-01 1.60806969e-01 5.90333641e-01 -8.15215647e-01
1.35584131e-01 4.65995133e-01 1.25778973e-01 -6.61211908e-01
6.95386112e-01 1.85400322e-01 1.53161110e-02 -2.23881409e-01
5.32068908e-01 5.29795647e-01 5.01228869e-01 -7.97823191e-01
9.94345784e-01 4.29143161e-01 -8.31327513e-02 -4.55886930e-01
-7.65582800e-01 -4.35451239e-01 -6.78204060e-01 -8.57751593e-02
2.67287523e-01 -7.74329960e-01 -5.00416875e-01 4.07129437e-01
-6.30438387e-01 -5.65649211e-01 -4.93007421e-01 5.27883887e-01
-5.50808430e-01 -6.01963289e-02 -4.28450316e-01 -1.99142307e-01
-5.57523847e-01 -1.14804029e+00 1.09589243e+00 2.45576546e-01
4.59687673e-02 -6.63085222e-01 -6.67837337e-02 -1.42458647e-01
7.93088615e-01 -2.17902243e-01 9.41591620e-01 -8.06975663e-01
-8.57403815e-01 -5.06142437e-01 -4.19544041e-01 4.23754156e-01
3.29513878e-01 1.02041841e-01 -7.43979454e-01 -6.95182979e-01
-4.45578337e-01 -6.25500798e-01 9.27971721e-01 3.21291715e-01
1.63294566e+00 -3.31826717e-01 -1.81494907e-01 8.17916214e-01
1.33813024e+00 1.65769041e-01 4.81875211e-01 8.31176341e-01
4.14639175e-01 2.63719022e-01 7.26993501e-01 3.04244727e-01
-5.08699454e-02 6.18652284e-01 3.18632096e-01 -9.26529318e-02
-1.25402719e-01 -1.41553387e-01 1.41204536e-01 4.72513646e-01
2.68883318e-01 -1.02028227e-03 -9.02915359e-01 5.96056581e-01
-1.91457224e+00 -6.57716691e-01 8.85299563e-01 2.28956628e+00
7.62769341e-01 -5.85970096e-02 -2.49997348e-01 -2.92356730e-01
4.49373126e-01 1.39097050e-01 -8.29374433e-01 -1.06450155e-01
-1.81758389e-01 3.57368588e-01 7.71354258e-01 2.61205196e-01
-1.12558544e+00 1.17101419e+00 6.71284723e+00 9.19264078e-01
-1.49429417e+00 -2.02206686e-01 5.06717741e-01 -4.29990679e-01
-1.01737365e-01 -1.46043167e-01 -6.14625633e-01 1.62622064e-01
7.50088990e-01 -2.36457035e-01 6.17191732e-01 9.04449224e-01
-2.67484516e-01 1.87642202e-01 -1.13104856e+00 8.78168702e-01
2.13022411e-01 -1.49923205e+00 3.70453209e-01 1.77413523e-02
8.64944041e-01 5.18088222e-01 4.81117010e-01 4.17154670e-01
6.41939938e-01 -7.22079098e-01 2.68802434e-01 3.39224041e-01
7.70113945e-01 -8.15104723e-01 5.70374668e-01 -2.40594223e-01
-9.59909439e-01 -3.16653043e-01 -6.71019316e-01 5.04146814e-01
-4.96745884e-01 5.17029501e-02 -8.46362591e-01 1.02264561e-01
9.94298995e-01 7.38751471e-01 -8.60262930e-01 1.05629468e+00
-2.78836042e-02 2.76756287e-01 -4.36693311e-01 -1.22262746e-01
5.72177112e-01 1.65343136e-02 1.73484474e-01 8.63399446e-01
4.16831702e-01 -1.15038574e-01 2.34744161e-01 2.23481685e-01
-4.79187548e-01 2.51066059e-01 -6.05451822e-01 1.09309042e-02
5.25135815e-01 1.43403244e+00 -5.79285324e-01 -3.54575157e-01
-3.58385295e-01 7.06484258e-01 4.45277601e-01 6.52759075e-01
-5.13007343e-01 -4.45678532e-01 6.98010504e-01 -2.28207946e-01
6.41816974e-01 -9.60790180e-03 2.22070575e-01 -1.14505732e+00
-7.54434466e-02 -1.03313780e+00 5.93977749e-01 -7.66449094e-01
-1.37332773e+00 6.43244386e-01 2.21580610e-01 -1.39890909e+00
-4.46823478e-01 -7.65512109e-01 -2.53064454e-01 5.24719954e-01
-1.79614210e+00 -1.17377102e+00 -3.48344564e-01 8.17514360e-01
4.40375298e-01 -6.44828796e-01 9.29705143e-01 2.65208155e-01
-3.03015143e-01 8.78793180e-01 5.76217651e-01 1.76848739e-01
1.12993515e+00 -9.24234927e-01 2.46817812e-01 5.21474898e-01
5.07310212e-01 5.78600705e-01 3.01442415e-01 -2.13774726e-01
-1.48586059e+00 -1.22379076e+00 5.05719364e-01 -2.34044865e-02
7.53126860e-01 -1.49035543e-01 -1.06293666e+00 6.88190758e-01
1.50912926e-01 6.76332772e-01 7.96472371e-01 7.36086816e-02
-7.87070572e-01 -5.66771805e-01 -1.16092479e+00 6.31757438e-01
8.62828135e-01 -5.58810532e-01 -2.87649989e-01 4.88799334e-01
4.25039947e-01 -4.61404622e-01 -8.51141572e-01 5.87476194e-01
8.58288586e-01 -4.36379671e-01 1.01249099e+00 -9.95384037e-01
2.05582753e-01 -2.62720227e-01 -3.75270486e-01 -1.33763266e+00
-4.88934547e-01 -4.35734421e-01 -1.01234227e-01 6.83630228e-01
3.52410704e-01 -7.12708950e-01 8.23616862e-01 8.46771955e-01
4.57801223e-02 -8.52462888e-01 -7.94224858e-01 -7.74294913e-01
-1.65669322e-01 -8.65563125e-05 8.65288436e-01 8.16885114e-01
-5.86353719e-01 -8.03899616e-02 -2.07012728e-01 1.48032814e-01
5.53937912e-01 2.98789591e-01 1.00384855e+00 -1.06716061e+00
-1.53044194e-01 -3.73083889e-01 -4.59124357e-01 -9.39222395e-01
3.79334778e-01 -7.88812637e-01 4.77447510e-02 -1.05678380e+00
1.00872964e-01 -6.86084449e-01 -1.12911642e+00 9.86116707e-01
-5.88959306e-02 4.97788072e-01 2.70600617e-01 4.83923405e-01
-8.45280826e-01 6.34423316e-01 1.08524644e+00 -3.36389035e-01
-2.01343730e-01 -2.87882507e-01 -5.26189566e-01 5.17402470e-01
8.32652211e-01 -5.10700464e-01 -5.33682942e-01 -6.82905257e-01
7.54088834e-02 -4.68882680e-01 2.24085540e-01 -8.88340354e-01
4.00068015e-01 -2.70553917e-01 7.49262393e-01 -5.94359279e-01
3.84258062e-01 -9.56810713e-01 -2.67541111e-01 3.47150773e-01
-6.62128806e-01 2.79820114e-01 3.22776586e-01 5.38851023e-01
-2.90534049e-01 -1.54442966e-01 7.50099361e-01 4.00260724e-02
-1.07076979e+00 7.54274964e-01 -5.42337000e-02 -1.23258911e-01
9.48177218e-01 -1.84899598e-01 -3.52342397e-01 -3.82136047e-01
-2.60303497e-01 3.17505419e-01 3.32482725e-01 7.81852245e-01
7.41254568e-01 -1.51633430e+00 -5.07234454e-01 4.25018817e-01
4.30873692e-01 -2.52031833e-01 -8.63956064e-02 1.06269352e-01
-4.01449919e-01 4.05847639e-01 -5.17880857e-01 -6.76591635e-01
-1.27799249e+00 5.46589315e-01 3.79599333e-01 -3.21607977e-01
-3.90458912e-01 8.03099215e-01 -2.34505422e-02 -5.26993334e-01
2.62037247e-01 -6.44678995e-02 1.18728958e-01 -1.10989548e-01
5.25193989e-01 7.61708319e-02 2.66051918e-01 -5.07358372e-01
-2.01076090e-01 6.38211310e-01 -7.79283404e-01 4.76622805e-02
1.56001425e+00 6.33164421e-02 -1.56213054e-02 -7.68087134e-02
1.60122955e+00 -4.77463245e-01 -1.39531744e+00 -8.42090249e-01
-3.51398066e-02 -6.77899361e-01 2.26942196e-01 -7.92346478e-01
-1.35012102e+00 4.34715390e-01 8.10931444e-01 -3.28730077e-01
1.28863716e+00 -3.07672441e-01 7.18178332e-01 1.21854174e+00
4.78908777e-01 -1.40132415e+00 4.98758852e-01 5.40197849e-01
9.89832222e-01 -1.25127029e+00 1.38429508e-01 7.68905953e-02
-2.96460986e-01 1.14066732e+00 8.60208213e-01 -5.15723407e-01
8.22078466e-01 2.89769899e-02 3.59968543e-01 -3.37628238e-02
-7.71637022e-01 -9.22347885e-03 5.44422567e-01 4.01673168e-01
-5.01768813e-02 -2.74976641e-01 1.21622421e-01 -4.12716538e-01
-4.02171351e-02 -1.13841981e-01 8.88716057e-02 1.23532128e+00
-3.93186986e-01 -1.28259039e+00 -2.79559851e-01 9.01861668e-01
-2.45491892e-01 -1.04964875e-01 -2.34492049e-01 9.25749540e-01
-2.56753683e-01 4.36763585e-01 3.00962240e-01 -1.76209509e-01
2.42015347e-01 -1.53000712e-01 5.06461740e-01 -4.07891154e-01
-5.92719793e-01 -1.02645205e-03 -4.25480694e-01 -7.78008759e-01
-4.83434051e-01 -4.24099624e-01 -8.93635094e-01 -2.06755064e-02
-4.45558429e-01 1.49826840e-01 6.89159691e-01 8.25733602e-01
7.08401084e-01 2.49820367e-01 9.22620773e-01 -8.65725458e-01
-8.54509175e-01 -6.88434660e-01 -3.32307488e-01 5.45365095e-01
3.41490716e-01 -7.20683217e-01 -1.46620944e-01 -6.27252609e-02] | [10.64837646484375, 0.728018045425415] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.