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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
f62e866b-1138-4d2c-8422-405c901a268a | is-syntax-structure-modeling-worth-leveraging | null | null | https://openreview.net/forum?id=L3TcOq4D2cD | https://openreview.net/pdf?id=L3TcOq4D2cD | Is syntax structure modeling worth? Leveraging pattern-driven modeling to enable affordable sentiment dependency learning | Is structure information modeling really worth in Aspect-based sentiment classification (ABSC)? Recent popular works tend to exploit syntactic information guiding sentiment dependency parsing, i.e., structure-based sentiment dependency learning. However, many works fall into the trap that confusing the concepts between syntax dependency and sentiment dependency. Besides, structure information (e.g., syntactic dependency tree) usually consumes expensive computational resources due to the extraction of the adjacent matrix. Instead, we believe the sentiment dependency mostly occurs between adjacent aspects. By proposing the sentiment patterns (SP) to boost the sentiment dependency learning, we introduce the Local dependency aggregating (Lena) to explore sentiment dependency in the text. Experiments show that Lena is more efficient than existing structure-based models without dependency matrix constructing and modeling expense. The performance on all five public ABSC datasets makes a big step compared to state-of-the-art models, and our work could inspire future research focusing on efficient local sentient dependency modeling. | ['Anonymous'] | 2021-12-17 | null | null | null | acl-arr-december-2022-12 | ['sentiment-dependency-learning'] | ['natural-language-processing'] | [-1.79063752e-01 -1.47253603e-01 -5.74404895e-01 -9.68392789e-01
-2.17854798e-01 -4.84910667e-01 4.21324164e-01 4.81847197e-01
-2.88230985e-01 4.38046336e-01 6.40019298e-01 -5.08244038e-01
2.24314943e-01 -8.41082752e-01 -4.12281275e-01 -6.20386004e-01
2.09580407e-01 3.80565941e-01 1.29749432e-01 -7.78102696e-01
4.42100018e-01 -1.29918575e-01 -1.21288753e+00 5.44932246e-01
8.00332546e-01 8.99207711e-01 8.38113949e-02 3.18931967e-01
-7.55302608e-01 1.02289665e+00 -7.27930784e-01 -1.06202853e+00
-2.28772074e-01 -3.91941994e-01 -9.15658593e-01 -5.26111312e-02
-3.09652001e-01 3.79484862e-01 6.81631923e-01 9.95821714e-01
4.81241718e-02 -1.88625351e-01 5.25436699e-01 -1.12679553e+00
-7.60959923e-01 1.15012705e+00 -8.36548924e-01 2.62490839e-01
2.83106625e-01 -5.71454525e-01 1.43309259e+00 -7.07302272e-01
3.70092422e-01 1.15236676e+00 6.38228118e-01 2.38053098e-01
-5.91697097e-01 -5.91769457e-01 8.39940965e-01 1.95577011e-01
-4.85159367e-01 9.65106264e-02 1.21617532e+00 -2.75940061e-01
1.23558760e+00 2.05475286e-01 7.86990941e-01 1.06473577e+00
6.47383749e-01 9.81191993e-01 1.28500307e+00 -4.71113026e-01
6.27653524e-02 3.65196198e-01 9.94174600e-01 5.06096065e-01
4.80658889e-01 -5.06607115e-01 -6.61493361e-01 5.38773760e-02
-3.68430018e-02 -4.10647579e-02 3.20660770e-01 1.67233348e-02
-8.58530760e-01 9.95822549e-01 2.74937719e-01 3.93603504e-01
-6.61865016e-03 -3.48465234e-01 6.97680295e-01 4.44796860e-01
8.89976859e-01 4.18246031e-01 -1.24118555e+00 -9.50891972e-02
-4.12470400e-01 -2.76658475e-01 9.64334846e-01 1.01654863e+00
7.93696642e-01 -2.73429692e-01 4.54158008e-01 8.18794787e-01
6.39989495e-01 3.09515595e-01 4.98998195e-01 -1.15914248e-01
6.29190505e-01 1.19439983e+00 -4.43268359e-01 -9.15273726e-01
-8.16715300e-01 -3.35899711e-01 -6.08366728e-01 -2.52917260e-01
7.55256116e-02 -5.84017098e-01 -7.53211498e-01 1.33662355e+00
4.89884824e-01 -3.01131368e-01 2.74990231e-01 5.78717709e-01
1.00640309e+00 5.02813160e-01 4.49218452e-01 -2.37128541e-01
1.79559815e+00 -1.31585598e+00 -7.16325045e-01 -5.93077123e-01
1.17546153e+00 -1.06955302e+00 1.15058398e+00 2.41625383e-01
-6.27918959e-01 -3.76432657e-01 -1.03223515e+00 -5.42150289e-02
-7.58121967e-01 -1.75484926e-01 1.61997318e+00 9.71049130e-01
-6.50344074e-01 1.43406168e-01 -8.32228005e-01 -3.34726781e-01
2.17841908e-01 5.38368762e-01 -3.96896333e-01 1.54583722e-01
-1.30678630e+00 9.23027098e-01 6.08148091e-02 1.43339662e-02
-3.40885669e-02 -6.03839517e-01 -1.15356815e+00 -3.90486941e-02
3.54716212e-01 -6.63924396e-01 9.31281745e-01 -1.16906965e+00
-1.26963377e+00 6.32441282e-01 -5.75392842e-01 5.08385897e-02
-3.46070319e-01 -3.57684702e-01 -4.50833738e-01 -3.07367146e-01
2.46270180e-01 1.04734242e-01 6.93747997e-01 -1.19086623e+00
-9.05716538e-01 -6.84943557e-01 5.40745974e-01 2.53705144e-01
-8.65490437e-01 5.72081149e-01 -6.32517159e-01 -6.42763317e-01
1.32415578e-01 -1.06446958e+00 -5.36154807e-01 -1.08883810e+00
-7.16738284e-01 -5.73836327e-01 6.45537257e-01 -1.95253178e-01
1.44881773e+00 -1.87582886e+00 -5.98484511e-03 6.06434159e-02
-2.46055111e-01 -2.04239637e-02 -4.44679409e-02 4.69349504e-01
-3.65769029e-01 4.92084056e-01 -1.80997774e-01 -6.71818793e-01
-1.92906395e-01 3.77244025e-01 -1.95535108e-01 -4.86256778e-02
1.51023448e-01 9.33732748e-01 -6.86937690e-01 -6.75449312e-01
2.72346754e-02 3.62802655e-01 -6.80511773e-01 -2.18829326e-02
-1.93017468e-01 4.29632574e-01 -9.80969608e-01 8.95853281e-01
6.69789553e-01 -2.77444065e-01 4.56958115e-01 -2.90724158e-01
-7.83212408e-02 6.78557456e-01 -7.34897256e-01 1.46885288e+00
-8.31217051e-01 1.03673980e-01 -2.05465585e-01 -9.60435927e-01
9.74362969e-01 2.04339102e-02 4.23082530e-01 -5.59695005e-01
4.71129507e-01 -7.62615725e-03 6.09445870e-02 -5.45937300e-01
5.47583699e-01 -4.47284997e-01 -6.10866189e-01 6.41329646e-01
9.47738625e-03 -1.13849811e-01 5.88891447e-01 2.22311854e-01
8.03628087e-01 9.92639139e-02 3.08582127e-01 -4.80801582e-01
9.08166051e-01 2.59137690e-01 6.81858242e-01 5.88833205e-02
-4.47287671e-02 2.38031507e-01 1.03025639e+00 -6.15717351e-01
-4.18036819e-01 -3.05587530e-01 -1.99114606e-01 1.58772922e+00
3.32823724e-01 -1.08766067e+00 -6.07770920e-01 -1.32355356e+00
-3.73051673e-01 6.49064243e-01 -8.41931760e-01 1.68403208e-01
-7.33094931e-01 -1.45123625e+00 -5.11623472e-02 8.63740206e-01
3.74657214e-01 -1.21490490e+00 -6.92218766e-02 1.44033283e-01
-1.50845900e-01 -1.14898407e+00 -2.87127048e-01 6.81110859e-01
-1.14605784e+00 -9.35953677e-01 1.34186838e-02 -9.99515891e-01
8.74740303e-01 2.94296563e-01 1.58720160e+00 4.15854037e-01
3.99805188e-01 -7.06521496e-02 -1.07653809e+00 -7.82476783e-01
-3.84494811e-02 5.36351979e-01 -1.80412412e-01 -2.98385978e-01
1.07121587e+00 -6.54126287e-01 -4.49355513e-01 3.38030636e-01
-7.67227769e-01 -6.70701042e-02 6.06489956e-01 6.63936853e-01
6.15302265e-01 1.81116804e-01 4.88560438e-01 -1.60276306e+00
6.89191520e-01 -6.24664664e-01 -1.47409022e-01 1.90609574e-01
-9.15323436e-01 1.12382047e-01 9.12992179e-01 -5.46031957e-03
-1.39792943e+00 -6.82028607e-02 -4.64637935e-01 4.93049741e-01
-4.86303344e-02 1.01096308e+00 -2.84652889e-01 3.63339692e-01
1.60889223e-01 -1.22548349e-01 -4.87122953e-01 -4.14327472e-01
1.97252274e-01 4.36505407e-01 -2.88680315e-01 -5.90788901e-01
4.33795094e-01 5.84024608e-01 4.50328477e-02 -5.38131475e-01
-1.63709795e+00 -6.13268197e-01 -7.83243537e-01 1.76126197e-01
9.18740630e-01 -7.76658654e-01 -4.81812328e-01 2.59272158e-01
-1.06942701e+00 2.05621645e-01 3.26838493e-02 3.66674662e-01
4.83101159e-02 4.25620079e-01 -7.92734742e-01 -7.75192738e-01
-5.38436294e-01 -1.05181050e+00 9.67698038e-01 3.82354707e-01
-4.68010783e-01 -1.51230574e+00 2.54105598e-01 8.25649261e-01
1.88742772e-01 -5.50785884e-02 1.17853272e+00 -7.60031104e-01
5.24811214e-03 -4.13529128e-01 1.34291410e-01 4.08421546e-01
2.99644977e-01 1.26518011e-02 -7.60915041e-01 2.24549428e-01
4.27798778e-01 -2.06918657e-01 8.23890746e-01 2.72125155e-01
7.20354915e-01 -8.31433758e-02 -2.90803909e-01 5.43045104e-01
1.18530798e+00 2.45808989e-01 3.71488214e-01 6.99683964e-01
1.00146246e+00 1.02794158e+00 9.45609450e-01 3.14442873e-01
1.06245410e+00 1.79843709e-01 2.33610287e-01 -1.31532282e-01
2.42752954e-01 -9.58236232e-02 5.48311949e-01 1.51630950e+00
-2.05835164e-01 -2.15185240e-01 -7.26343215e-01 4.48307335e-01
-1.74151504e+00 -4.37562913e-01 -6.04189515e-01 1.40741909e+00
7.88792193e-01 4.15181458e-01 -1.65448889e-01 1.78024203e-01
3.73090833e-01 4.65733945e-01 -2.24650919e-01 -9.03817892e-01
-2.44699255e-01 3.05154920e-01 1.20152131e-01 4.20999140e-01
-1.21598423e+00 1.36886168e+00 5.51122236e+00 4.60688084e-01
-9.70067203e-01 2.78321177e-01 8.43007803e-01 2.08775252e-01
-7.81608224e-01 3.51108104e-01 -1.25834751e+00 3.12635541e-01
8.47323179e-01 1.77321762e-01 -2.90368885e-01 1.19527316e+00
-1.54539302e-01 -2.75389075e-01 -7.26988077e-01 5.14513314e-01
3.55019033e-01 -8.74212980e-01 1.79362521e-01 -2.62085408e-01
7.97904670e-01 4.20443341e-02 -8.71400088e-02 5.29160261e-01
4.18353796e-01 -7.68302441e-01 2.94468433e-01 -8.71070176e-02
8.74458849e-02 -9.56848919e-01 1.36455381e+00 2.36509532e-01
-1.40596402e+00 9.02365744e-02 -5.21095753e-01 -3.17756295e-01
3.65819752e-01 9.65289533e-01 -2.42667809e-01 8.04661095e-01
1.00451589e+00 1.13992333e+00 -7.71049201e-01 1.53374061e-01
-9.14664626e-01 7.47475207e-01 -1.13485605e-01 -5.44703603e-01
3.12830955e-01 -5.56287706e-01 1.60883486e-01 1.37662435e+00
-1.08083181e-01 -6.61069602e-02 -6.24273866e-02 -4.14096110e-04
1.35420531e-01 5.97259283e-01 -2.87864506e-01 3.51152681e-02
-2.05760300e-01 1.65501308e+00 -1.13756680e+00 -2.71058559e-01
-1.03945923e+00 7.96292663e-01 2.90562987e-01 -6.95100799e-02
-6.56152368e-01 -1.21519394e-01 6.38205528e-01 -3.43740642e-01
5.16435266e-01 -1.21033087e-01 -1.03103030e+00 -1.51716828e+00
2.96956718e-01 -7.09414363e-01 7.04252481e-01 -3.74677747e-01
-1.54717195e+00 9.96321261e-01 -2.20425457e-01 -1.10998917e+00
3.95286381e-02 -8.27478945e-01 -8.77635956e-01 6.72408879e-01
-1.93034947e+00 -1.42467129e+00 7.48959780e-02 5.54917514e-01
7.82264352e-01 7.90418088e-02 9.73907650e-01 9.47646350e-02
-7.75687873e-01 5.60914397e-01 -5.73598683e-01 1.03965960e-01
6.61542177e-01 -1.23346102e+00 5.42946756e-01 7.28244603e-01
3.20406407e-01 1.06631565e+00 5.01293361e-01 -6.96515799e-01
-1.46737409e+00 -6.72710419e-01 1.54014635e+00 -1.08564675e+00
1.02764404e+00 -3.43266398e-01 -5.82250416e-01 6.96821570e-01
4.02682841e-01 -2.39706576e-01 1.38176024e+00 1.04358828e+00
-5.34202576e-01 -1.64007574e-01 -5.89804173e-01 5.99235296e-01
8.71944249e-01 -2.25761205e-01 -8.47258925e-01 2.31520310e-01
8.00041020e-01 5.16156852e-03 -8.80290985e-01 6.29557431e-01
5.74271798e-01 -1.11712897e+00 5.63877165e-01 -8.47044289e-01
9.10282314e-01 -1.90155298e-01 -1.35382488e-01 -1.27736890e+00
-1.44253448e-01 -1.94069684e-01 1.31086767e-01 1.50390482e+00
1.13504815e+00 -5.83385289e-01 9.17805791e-01 7.09973752e-01
-1.53358340e-01 -1.08919442e+00 -4.65737313e-01 -3.54258180e-01
2.61660367e-01 -8.50708187e-01 7.66956925e-01 1.06125152e+00
6.09710753e-01 1.06623590e+00 -1.51215270e-01 2.16368772e-02
9.04337019e-02 5.54301441e-01 6.31180763e-01 -1.18725049e+00
-2.17797205e-01 -2.11409718e-01 -8.82691219e-02 -9.15901184e-01
4.01000440e-01 -7.08375692e-01 -3.11449379e-01 -1.66921079e+00
3.70543480e-01 -6.79370701e-01 -5.85525870e-01 5.32413125e-01
-6.81102157e-01 1.30575404e-01 6.66499361e-02 1.63054585e-01
-7.84507155e-01 4.62056726e-01 1.55853653e+00 -1.31215990e-01
-1.43029973e-01 2.14771882e-01 -1.60001099e+00 1.28972471e+00
1.02520227e+00 -8.54497969e-01 -4.88756627e-01 -3.63271266e-01
1.01259983e+00 -2.40097806e-01 -6.58355117e-01 -3.94750059e-01
2.76616722e-01 -1.86491281e-01 2.29043812e-01 -7.32074261e-01
1.52896255e-01 -9.26053226e-01 -4.64721084e-01 1.30100057e-01
-8.16263109e-02 3.97978812e-01 5.31377420e-02 2.88111091e-01
-5.47445059e-01 -4.31871980e-01 3.05362582e-01 -2.11409837e-01
-6.51259303e-01 1.45050645e-01 -1.97696432e-01 -2.40013972e-02
7.31493175e-01 2.04345703e-01 -4.56736982e-01 -1.02025576e-01
-4.51083332e-01 2.19974905e-01 2.11038798e-01 6.29587293e-01
1.24960639e-01 -7.28166640e-01 -4.21651155e-01 1.23060577e-01
1.99908793e-01 -3.53664011e-02 4.33417708e-02 8.56769800e-01
-5.87151572e-02 4.37182307e-01 1.32798120e-01 -2.20642865e-01
-1.42406583e+00 5.89021504e-01 -3.10645401e-01 -8.61110091e-01
-2.24732608e-01 1.19169652e+00 2.87684709e-01 -6.21646702e-01
-2.90958256e-01 -2.67614007e-01 -9.92687166e-01 4.36858714e-01
4.38527226e-01 -1.93450540e-01 2.10295334e-01 -7.03831553e-01
-6.42990410e-01 8.56287718e-01 -4.95221019e-01 9.58205387e-02
1.39964724e+00 -2.19669193e-01 -7.13830233e-01 4.73768741e-01
1.09939134e+00 4.70808119e-01 -6.77033782e-01 2.33276978e-01
2.87402272e-01 -1.38072640e-01 -9.48508680e-02 -6.94443643e-01
-1.23266733e+00 7.02298820e-01 -2.61170387e-01 5.41103005e-01
1.20959747e+00 2.39170700e-01 8.27293098e-01 6.15082264e-01
3.97577435e-01 -1.12274492e+00 -1.08879805e-01 1.06446338e+00
5.12903154e-01 -1.47058868e+00 3.67349088e-01 -8.74774039e-01
-1.19603825e+00 1.03239369e+00 8.10489416e-01 -1.86316967e-01
1.28557158e+00 7.75056243e-01 6.27226114e-01 -4.43889081e-01
-8.16556871e-01 -3.52396011e-01 7.84377083e-02 4.22212213e-01
1.03079724e+00 1.42417237e-01 -7.54123986e-01 1.22117090e+00
-7.72598326e-01 -5.92062294e-01 2.61611581e-01 1.07429457e+00
-1.60987720e-01 -1.83927691e+00 -7.20204860e-02 4.98907238e-01
-8.70317817e-01 -6.63069069e-01 -6.20555937e-01 7.49244213e-01
1.68303072e-01 1.47927487e+00 -9.20112804e-02 -6.91115439e-01
4.73059356e-01 5.02618738e-02 -4.52214526e-03 -9.47451174e-01
-1.19815886e+00 9.18064490e-02 4.44292754e-01 -3.97335649e-01
-9.37653065e-01 -5.80077350e-01 -1.34864902e+00 -1.58833459e-01
-6.20534837e-01 4.42569852e-01 1.01881683e+00 1.26925886e+00
2.53544450e-01 4.69793558e-01 8.58092666e-01 -3.40520710e-01
1.92299113e-01 -1.08002436e+00 -4.31455642e-01 3.71679574e-01
-1.46852136e-01 -3.91903520e-01 -5.57095170e-01 5.45053147e-02] | [11.418614387512207, 6.714746952056885] |
30ec480b-967e-4dc2-bba3-15a476553d57 | from-knowledge-graph-embedding-to-ontology | 1805.10461 | null | http://arxiv.org/abs/1805.10461v3 | http://arxiv.org/pdf/1805.10461v3.pdf | From Knowledge Graph Embedding to Ontology Embedding? An Analysis of the Compatibility between Vector Space Representations and Rules | Recent years have witnessed the successful application of low-dimensional
vector space representations of knowledge graphs to predict missing facts or
find erroneous ones. However, it is not yet well-understood to what extent
ontological knowledge, e.g. given as a set of (existential) rules, can be
embedded in a principled way. To address this shortcoming, in this paper we
introduce a general framework based on a view of relations as regions, which
allows us to study the compatibility between ontological knowledge and
different types of vector space embeddings. Our technical contribution is
two-fold. First, we show that some of the most popular existing embedding
methods are not capable of modelling even very simple types of rules, which in
particular also means that they are not able to learn the type of dependencies
captured by such rules. Second, we study a model in which relations are
modelled as convex regions. We show particular that ontologies which are
expressed using so-called quasi-chained existential rules can be exactly
represented using convex regions, such that any set of facts which is induced
using that vector space embedding is logically consistent and deductively
closed with respect to the input ontology. | ['Víctor Gutiérrez-Basulto', 'Steven Schockaert'] | 2018-05-26 | null | null | null | null | ['ontology-embedding'] | ['knowledge-base'] | [-8.18497315e-02 8.56225312e-01 -1.02621123e-01 -3.45967382e-01
4.76674885e-01 -5.87288737e-01 8.46635461e-01 5.67525625e-01
-6.46995455e-02 5.22722721e-01 3.04717392e-01 -4.26184356e-01
-6.74196064e-01 -1.33890355e+00 -8.20179522e-01 -4.78944361e-01
-2.20334321e-01 5.70219100e-01 3.11169297e-01 -7.61163652e-01
-9.60282236e-02 5.93529999e-01 -1.78569627e+00 2.46107787e-01
7.09412277e-01 7.11071312e-01 -3.53140116e-01 4.12920833e-01
-3.01401675e-01 9.31195140e-01 -3.37545633e-01 -6.72222257e-01
-3.76109928e-02 -1.95280597e-01 -1.14240217e+00 -1.60199270e-01
2.05645144e-01 1.77838966e-01 -2.97588468e-01 1.20210278e+00
-2.00031415e-01 8.14099163e-02 8.20798397e-01 -1.38012505e+00
-1.07256734e+00 7.54956186e-01 3.81365031e-01 4.34491970e-02
5.89792848e-01 -6.07631624e-01 1.29379356e+00 -6.16698980e-01
7.65599370e-01 1.31534350e+00 5.37841320e-01 5.30879498e-01
-1.57105505e+00 7.85394683e-02 1.48587137e-01 5.60807884e-01
-1.31178904e+00 -6.86033443e-02 9.11620677e-01 -6.89583004e-01
9.78705108e-01 4.98440415e-01 7.24844337e-01 8.57398927e-01
3.40840578e-01 2.80206054e-01 1.06882501e+00 -8.14937294e-01
4.38240975e-01 7.15267360e-01 5.43496609e-01 7.17921793e-01
7.51266956e-01 6.44410104e-02 -1.80824146e-01 -4.04166356e-02
4.27195728e-01 -1.34375528e-01 -4.84103739e-01 -1.15837216e+00
-1.20210862e+00 1.26884663e+00 6.73141062e-01 1.02140629e+00
-1.53906271e-01 -1.13245822e-01 3.63370687e-01 5.46349227e-01
2.72506446e-01 6.45384669e-01 -5.79922378e-01 3.58036309e-01
-3.31467390e-01 1.62951991e-01 1.05033612e+00 8.26025367e-01
7.01833725e-01 -4.41090725e-02 6.05491161e-01 2.81921715e-01
3.95188361e-01 1.09292738e-01 4.63130325e-01 -5.29594183e-01
6.42394200e-02 1.03787279e+00 1.03189670e-01 -1.51089633e+00
-6.54285073e-01 -1.48404241e-01 -5.34266949e-01 2.88617551e-01
4.40498233e-01 3.00682276e-01 -2.28164345e-01 1.84879816e+00
2.80502707e-01 1.68316945e-01 5.81156135e-01 5.91434836e-01
5.68465233e-01 3.27003866e-01 -2.54619628e-01 -1.03884108e-01
1.49521148e+00 -1.69650599e-01 -7.03675270e-01 2.16269821e-01
9.11616206e-01 -1.48110281e-04 8.00271213e-01 1.63809791e-01
-8.75308156e-01 -2.91931540e-01 -1.24364626e+00 -1.90976664e-01
-1.16463041e+00 -3.32834601e-01 7.95548916e-01 7.31807947e-01
-1.03324187e+00 6.11661315e-01 -6.60860658e-01 -6.55057549e-01
1.95040792e-01 3.08336109e-01 -8.74354005e-01 9.77180600e-02
-1.54514933e+00 1.37589109e+00 8.01115215e-01 1.32780463e-01
-2.04442635e-01 -5.12858987e-01 -1.24860489e+00 2.31327191e-01
6.80673599e-01 -5.60770571e-01 6.70629859e-01 -8.19315612e-01
-1.16400135e+00 8.44364405e-01 1.40512496e-01 -5.69534421e-01
3.66786957e-01 1.72176942e-01 -8.01087499e-01 -1.85157266e-02
-2.52751112e-01 2.63441745e-02 4.95431930e-01 -1.45524478e+00
-3.42652708e-01 -7.38100827e-01 8.81323099e-01 -1.93378344e-01
-4.74430740e-01 -2.94678658e-01 3.40459317e-01 -2.63442814e-01
2.06414789e-01 -8.37641299e-01 3.04611325e-02 -4.15805094e-02
-4.54015642e-01 -4.47570801e-01 6.34132922e-01 -2.54499316e-01
1.07257688e+00 -1.94653165e+00 6.66095078e-01 3.93218696e-01
3.23877275e-01 2.85820484e-01 2.81967461e-01 5.64607680e-01
-4.29289639e-01 4.04450893e-01 -2.48395711e-01 2.66432554e-01
6.12244964e-01 7.67030537e-01 -4.84586954e-01 5.70150733e-01
2.57445604e-01 8.27832699e-01 -8.99450839e-01 -2.92222291e-01
5.08927941e-01 6.54372215e-01 -6.66654170e-01 -2.74881333e-01
-2.89135545e-01 1.10273272e-01 -4.89468485e-01 9.65431556e-02
4.10650045e-01 2.25954019e-02 6.23111844e-01 -3.00576419e-01
-8.72156247e-02 1.28319561e-01 -1.42860782e+00 1.43246007e+00
-4.83324885e-01 4.27169114e-01 -3.31677943e-01 -1.50802124e+00
8.42445016e-01 4.78530496e-01 1.63056701e-01 -3.54381472e-01
1.75594300e-01 1.20013468e-01 2.91178450e-02 -5.96960187e-01
2.76807159e-01 -6.05441689e-01 -9.09041092e-02 4.94367033e-02
5.88366948e-02 8.87014195e-02 1.64355822e-02 1.78930223e-01
8.36893320e-01 5.37913702e-02 6.86918974e-01 -4.86166477e-01
8.66337001e-01 -7.38284364e-02 5.01712739e-01 3.23201329e-01
6.80292994e-02 -5.47814146e-02 1.01128006e+00 -6.12197101e-01
-8.59061182e-01 -1.07425523e+00 -6.81266427e-01 6.21755183e-01
6.64143413e-02 -4.75474209e-01 -3.43146890e-01 -7.58532524e-01
3.16484660e-01 8.92108262e-01 -1.10209966e+00 -3.84256989e-01
-2.28342444e-01 -3.61346424e-01 4.30411696e-01 3.85466725e-01
-7.10236430e-02 -9.42629993e-01 -8.07829499e-01 1.39912471e-01
1.39484540e-01 -1.05608630e+00 4.29401159e-01 2.00801685e-01
-8.65016639e-01 -1.45182788e+00 2.16616970e-02 -7.05388725e-01
5.51421404e-01 -3.16069633e-01 1.14276755e+00 1.81677073e-01
3.63585749e-03 6.01819098e-01 -5.51698744e-01 -3.57588500e-01
-5.38696885e-01 -2.59192795e-01 4.00247514e-01 3.48720580e-01
6.08958483e-01 -8.02148640e-01 -6.11666553e-02 -4.39933725e-02
-1.45934498e+00 -3.46918613e-01 6.08193427e-02 6.91011131e-01
4.15864587e-01 3.64458233e-01 5.95685363e-01 -1.19725966e+00
4.52924877e-01 -5.79868734e-01 -5.36125362e-01 5.65149784e-01
-6.58265233e-01 6.20693028e-01 6.96300566e-01 -5.85569888e-02
-7.90355980e-01 -2.55439579e-01 7.13030696e-02 -1.12718798e-01
-2.10697964e-01 8.02026093e-01 -4.79370445e-01 -3.44810225e-02
8.29175472e-01 1.36700384e-02 -7.93927535e-02 -2.92325318e-01
7.67173588e-01 5.32687783e-01 2.75628090e-01 -5.87021112e-01
8.71975958e-01 6.45512879e-01 4.58411366e-01 -8.99899602e-01
-6.06630683e-01 -5.80374561e-02 -1.02975738e+00 2.19541878e-01
9.36198354e-01 -3.49922597e-01 -7.75873840e-01 -3.76943558e-01
-1.04806924e+00 7.77369142e-02 -6.26966238e-01 4.54026729e-01
-7.52967775e-01 4.21037018e-01 -2.99539506e-01 -7.56018519e-01
3.61479640e-01 -8.06223154e-01 5.11341870e-01 -2.50790268e-01
-2.37864137e-01 -1.48112381e+00 1.40305787e-01 2.53037866e-02
1.58495858e-01 4.26200211e-01 1.55795324e+00 -9.59445477e-01
-2.16710329e-01 -2.69686103e-01 7.30912760e-02 3.64040315e-01
1.46419987e-01 -2.22139675e-02 -8.16388309e-01 8.96780193e-03
1.80622101e-01 6.25320524e-02 5.28291881e-01 -6.53685480e-02
6.85475647e-01 -3.82085830e-01 -3.39523107e-01 2.65784979e-01
1.78870058e+00 -3.23905535e-02 7.62998581e-01 3.63741010e-01
5.97153187e-01 9.20366347e-01 2.42017061e-01 -4.80688848e-02
6.03084147e-01 1.01758707e+00 5.83324909e-01 5.75615346e-01
3.37061852e-01 -1.52696922e-01 1.04673140e-01 6.54596984e-01
-3.30633551e-01 1.04267545e-01 -8.60743940e-01 6.48408294e-01
-1.81429470e+00 -1.18643475e+00 -2.52834886e-01 2.23229718e+00
7.91416824e-01 2.58232448e-02 3.33920009e-02 5.71840525e-01
5.23859680e-01 7.59862810e-02 -3.73771004e-02 -7.35228539e-01
-2.60086000e-01 2.00290218e-01 1.14893392e-01 9.16332066e-01
-9.04665351e-01 6.39585793e-01 5.51890182e+00 -1.91321410e-02
-7.74184167e-01 1.86503768e-01 -3.86695325e-01 1.79061994e-01
-8.48917544e-01 3.52050513e-01 -2.74865061e-01 3.53317738e-01
9.62247968e-01 -2.08891094e-01 3.64610910e-01 6.24016821e-01
-2.89999783e-01 2.97626704e-01 -1.51492500e+00 6.57429814e-01
1.51926503e-01 -1.33712924e+00 2.04426900e-01 2.14669213e-01
4.56527680e-01 -5.77216327e-01 -2.11915582e-01 3.57734263e-01
2.50964314e-01 -1.12763667e+00 8.01882923e-01 7.91113257e-01
3.26266646e-01 -7.79147208e-01 9.25303876e-01 2.45481089e-01
-9.75928187e-01 -2.75047362e-01 -6.34328604e-01 -2.06583336e-01
-1.23367561e-02 6.01909220e-01 -3.39026511e-01 1.10110819e+00
3.05743873e-01 6.81804478e-01 -3.84127915e-01 5.55722535e-01
-2.79022604e-01 2.70620555e-01 -2.45532796e-01 1.37161970e-01
-2.04396620e-02 -2.07408711e-01 4.50248927e-01 1.01424897e+00
1.39568210e-01 3.16783078e-02 -2.13559031e-01 1.02334905e+00
6.56425506e-02 1.00940906e-01 -1.15587234e+00 1.54413536e-01
1.92512706e-01 8.00298393e-01 -5.00909269e-01 -2.00122893e-01
-7.27069199e-01 5.85491717e-01 4.67773974e-01 2.43278965e-01
-6.44835353e-01 -2.13838995e-01 8.33747208e-01 4.10719067e-02
4.42910880e-01 -2.85805147e-02 7.26225898e-02 -1.55739331e+00
2.49464184e-01 -4.12500322e-01 3.49287480e-01 -6.34794295e-01
-1.23795187e+00 5.52673519e-01 8.49260762e-02 -8.63396108e-01
-9.04131755e-02 -9.55217421e-01 -2.86073178e-01 5.24805903e-01
-1.64083958e+00 -9.84075904e-01 1.22746333e-01 8.74819279e-01
-2.44098261e-01 1.19863600e-01 1.30680728e+00 1.59252726e-03
-3.16283524e-01 3.49868566e-01 6.90652356e-02 -5.07637598e-02
7.10713044e-02 -1.67660391e+00 -3.35138857e-01 6.59110367e-01
6.85932875e-01 8.27292979e-01 1.02653611e+00 -2.92287886e-01
-1.49402869e+00 -9.02616441e-01 1.36175311e+00 -9.43773925e-01
9.50814843e-01 -3.54156405e-01 -1.18204868e+00 1.19595838e+00
-3.32745239e-02 5.25287211e-01 7.05619335e-01 3.71454775e-01
-7.65613317e-01 -2.07892179e-01 -1.17949760e+00 5.35703301e-01
9.99101460e-01 -7.30810106e-01 -1.39372265e+00 2.51831144e-01
7.71393418e-01 -4.82251216e-03 -1.29864419e+00 3.80615354e-01
3.31072837e-01 -9.99162495e-01 1.00605869e+00 -1.25717115e+00
3.16479117e-01 -4.06511098e-01 -4.69650000e-01 -1.32387269e+00
-1.67667851e-01 -7.56282508e-02 -3.47272843e-01 1.00591469e+00
3.13227952e-01 -1.05223453e+00 4.37567800e-01 4.30827707e-01
9.31430832e-02 -7.16204226e-01 -1.25637686e+00 -9.84846056e-01
4.36132073e-01 -4.61764425e-01 8.40959847e-01 1.34931636e+00
7.53305256e-01 2.84924716e-01 -1.58746734e-01 3.42251658e-01
2.59724379e-01 2.81481236e-01 2.91390985e-01 -1.91430748e+00
-8.74411464e-02 -3.67526650e-01 -1.47105849e+00 -2.31851131e-01
4.79673892e-01 -1.35485613e+00 -3.93503875e-01 -1.70945811e+00
-2.46451691e-01 -4.65100735e-01 -4.46370661e-01 4.26894665e-01
3.63067448e-01 -1.28698006e-01 9.66969803e-02 -1.04795277e-01
-2.72198707e-01 5.93353093e-01 7.67223060e-01 1.09771220e-02
1.37091860e-01 -3.16749573e-01 -9.29989398e-01 8.30218375e-01
6.50212407e-01 -4.17772919e-01 -6.64109468e-01 -2.45846435e-01
9.22233939e-01 6.63224491e-04 7.43609250e-01 -7.21153438e-01
1.45556569e-01 -1.70827255e-01 -8.44684765e-02 1.25651255e-01
2.21696466e-01 -1.32516730e+00 2.62724876e-01 2.15453774e-01
-4.01748002e-01 -8.52773264e-02 2.49004066e-02 8.11081648e-01
-3.09679121e-01 -4.41387743e-01 4.22017187e-01 3.30501050e-02
-7.98143983e-01 -2.61864364e-01 -1.62713602e-01 -1.15900859e-02
1.22399199e+00 -2.76716292e-01 -2.89858013e-01 -1.63671792e-01
-1.01961732e+00 -1.61878571e-01 5.98096490e-01 3.96448612e-01
5.38759112e-01 -1.42626286e+00 -4.31292921e-01 6.98489845e-02
4.85271156e-01 -1.80245340e-01 1.89366385e-01 7.92677879e-01
-3.66498142e-01 6.82115078e-01 -1.41352341e-01 -2.50457734e-01
-1.03460777e+00 1.20572615e+00 5.11143208e-01 1.80742592e-02
-9.28780258e-01 5.60134470e-01 -6.10122690e-03 -6.46096647e-01
-7.44931325e-02 -7.66158104e-01 -4.60667819e-01 5.88392988e-02
3.16086531e-01 5.25600836e-02 1.40254065e-01 -1.08594203e+00
-6.90192163e-01 7.06804037e-01 2.90599138e-01 1.59813792e-01
1.36763680e+00 1.10719167e-01 -3.54467511e-01 7.10140467e-01
1.16392016e+00 2.88801849e-01 -5.94092607e-01 -2.70373046e-01
1.61896467e-01 -2.89018124e-01 -1.10366389e-01 -5.58506846e-01
-8.35829914e-01 7.69794881e-01 3.08956474e-01 9.80513334e-01
8.23174179e-01 4.44466561e-01 9.39485803e-02 4.74539250e-01
5.26727676e-01 -6.83196247e-01 -4.83831733e-01 4.35392112e-01
9.71251369e-01 -8.90542209e-01 -1.88968942e-01 -6.53095782e-01
-3.56449008e-01 1.40561259e+00 -1.39180478e-03 -2.34460711e-01
8.37912261e-01 -1.76231619e-02 -2.24147499e-01 -5.41333973e-01
-5.96907437e-01 -4.27190751e-01 2.71378249e-01 9.37907994e-01
1.87386453e-01 3.71325076e-01 -3.76086831e-01 6.03772521e-01
-4.69015896e-01 6.77689537e-02 8.35981429e-01 9.54240143e-01
-3.63932669e-01 -1.12477183e+00 -2.63871014e-01 2.57693619e-01
-2.81559885e-01 1.67571679e-01 -3.94228220e-01 1.04304302e+00
3.75078112e-01 9.81063068e-01 -3.16083729e-02 -4.96334642e-01
6.30748570e-01 3.61835659e-01 6.80278659e-01 -7.76157022e-01
-3.35767083e-02 -9.81328964e-01 -1.33074848e-02 -5.01310170e-01
-6.50115430e-01 -5.26574612e-01 -1.30400825e+00 -3.29141766e-01
-4.79331970e-01 3.27374071e-01 5.32461882e-01 1.14138937e+00
1.11649372e-01 5.49426317e-01 9.93040353e-02 -2.55598843e-01
-5.36863089e-01 -5.42277694e-01 -9.51782763e-01 5.75058639e-01
4.33945775e-01 -1.15563917e+00 -5.29075205e-01 -1.80624038e-01] | [8.845259666442871, 7.567572593688965] |
db2ca23d-06ce-412f-bb3b-6bed6b51ef42 | unsupervised-domain-adaptation-for-clinician | 2108.11801 | null | https://arxiv.org/abs/2108.11801v4 | https://arxiv.org/pdf/2108.11801v4.pdf | Unsupervised domain adaptation for clinician pose estimation and instance segmentation in the operating room | The fine-grained localization of clinicians in the operating room (OR) is a key component to design the new generation of OR support systems. Computer vision models for person pixel-based segmentation and body-keypoints detection are needed to better understand the clinical activities and the spatial layout of the OR. This is challenging, not only because OR images are very different from traditional vision datasets, but also because data and annotations are hard to collect and generate in the OR due to privacy concerns. To address these concerns, we first study how joint person pose estimation and instance segmentation can be performed on low resolutions images with downsampling factors from 1x to 12x. Second, to address the domain shift and the lack of annotations, we propose a novel unsupervised domain adaptation method, called AdaptOR, to adapt a model from an in-the-wild labeled source domain to a statistically different unlabeled target domain. We propose to exploit explicit geometric constraints on the different augmentations of the unlabeled target domain image to generate accurate pseudo labels and use these pseudo labels to train the model on high- and low-resolution OR images in a self-training framework. Furthermore, we propose disentangled feature normalization to handle the statistically different source and target domain data. Extensive experimental results with detailed ablation studies on the two OR datasets MVOR+ and TUM-OR-test show the effectiveness of our approach against strongly constructed baselines, especially on the low-resolution privacy-preserving OR images. Finally, we show the generality of our method as a semi-supervised learning (SSL) method on the large-scale COCO dataset, where we achieve comparable results with as few as 1% of labeled supervision against a model trained with 100% labeled supervision. | ['Nicolas Padoy', 'Afshin Gangi', 'Vinkle Srivastav'] | 2021-08-26 | null | null | null | null | ['semi-supervised-human-pose-estimation', '2d-human-pose-estimation', 'semi-supervised-person-instance-segmentation', 'semi-supervised-person-bounding-box-detection'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision'] | [ 6.72356248e-01 5.44105113e-01 -3.45460773e-01 -5.76152146e-01
-1.33421564e+00 -7.89666653e-01 2.49940157e-01 -1.25179803e-02
-5.95875382e-01 6.19468987e-01 3.00223172e-01 -1.39637142e-01
1.09022325e-02 -9.48688537e-02 -7.96317399e-01 -8.19298983e-01
3.31500590e-01 6.53548360e-01 -1.99145377e-02 2.71465600e-01
-3.80671859e-01 4.54067439e-01 -1.01336372e+00 3.30912352e-01
7.10931659e-01 8.44092369e-01 1.07546374e-01 7.18250632e-01
4.91353661e-01 2.06834212e-01 -3.56151789e-01 -3.64511341e-01
8.09739470e-01 -5.06153643e-01 -6.93320215e-01 4.54643130e-01
9.27235186e-01 -4.73363698e-01 -3.62941861e-01 8.88403118e-01
7.27412581e-01 -1.56141892e-01 4.76964921e-01 -1.00780463e+00
-2.30275795e-01 1.64403282e-02 -8.75663102e-01 -5.56547120e-02
2.75470138e-01 2.37997711e-01 2.73045361e-01 -5.25946498e-01
8.24294508e-01 9.05034125e-01 8.46187949e-01 8.63539517e-01
-1.48938513e+00 -7.85314143e-01 8.46143290e-02 -2.87827253e-01
-1.37893057e+00 -5.35851657e-01 5.53914607e-01 -4.37762320e-01
1.58280849e-01 3.74396890e-01 5.26329696e-01 1.53212810e+00
-1.99127775e-02 4.93424863e-01 1.42612290e+00 -5.21246433e-01
8.93533826e-02 4.24523145e-01 1.04706436e-01 9.18043971e-01
6.18833721e-01 2.78521283e-03 -5.08945346e-01 -4.30100381e-01
1.01000404e+00 2.56976664e-01 -5.65969527e-01 -8.47484767e-01
-1.35693312e+00 5.31446517e-01 3.13066423e-01 -3.85633819e-02
-1.65037036e-01 -5.96330613e-02 2.32107237e-01 -2.66873926e-01
2.30195105e-01 5.64692676e-01 -3.53836685e-01 2.14642137e-01
-8.89654100e-01 1.10050201e-01 6.04911745e-01 1.12111497e+00
3.23448181e-01 -5.39030194e-01 -4.42293376e-01 5.43020725e-01
1.12820864e-01 5.63595355e-01 5.57302117e-01 -9.60969508e-01
4.53560978e-01 3.73371005e-01 3.48785669e-01 -5.62130690e-01
-5.45257807e-01 -4.71114635e-01 -7.07326591e-01 8.02819431e-02
9.48282599e-01 -3.33749205e-01 -1.34664929e+00 1.68797898e+00
6.27357423e-01 1.43740728e-01 -1.76145956e-01 1.08465326e+00
8.75554442e-01 -1.11041307e-01 3.94818280e-03 -1.46622986e-01
1.69293511e+00 -9.44691658e-01 -6.64366543e-01 -5.61835408e-01
6.04490340e-01 -5.13412893e-01 1.09262288e+00 2.07367733e-01
-8.94140959e-01 -3.45660746e-01 -9.86996710e-01 -5.49014807e-02
-6.85600638e-02 4.88574266e-01 4.66656119e-01 8.19301963e-01
-6.65326774e-01 3.33745033e-01 -1.31633496e+00 -3.99427652e-01
8.48326921e-01 4.77646500e-01 -8.43260467e-01 -4.21373665e-01
-7.11324275e-01 6.89479470e-01 4.94506247e-02 -2.00152956e-02
-7.38969207e-01 -9.64022040e-01 -1.08302128e+00 -3.75022769e-01
5.40438175e-01 -8.32725048e-01 1.15405846e+00 -8.15582991e-01
-1.04130030e+00 1.26266956e+00 -3.47519457e-01 -3.14434737e-01
9.14582074e-01 -1.41806439e-01 -6.87098876e-02 3.90293002e-01
4.05559957e-01 6.30526304e-01 8.15878630e-01 -1.39738631e+00
-4.67966259e-01 -8.74721885e-01 -3.09815675e-01 3.82375032e-01
-3.99022847e-02 -2.44429395e-01 -9.78361964e-01 -6.73715651e-01
1.93820044e-01 -1.39101744e+00 -5.64209402e-01 3.94098043e-01
-6.98431849e-01 5.41156054e-01 5.31399667e-01 -9.18712437e-01
6.74497545e-01 -2.22738457e+00 -2.57772624e-01 2.68214226e-01
5.41592896e-01 4.39671072e-04 3.46457437e-02 -4.21870708e-01
-3.17384869e-01 -2.20683590e-01 -2.58327276e-01 -5.94249725e-01
-4.40243483e-01 1.33190528e-01 5.56033365e-02 9.57778335e-01
-2.31938422e-01 8.47153127e-01 -8.40605021e-01 -7.04297364e-01
1.02973036e-01 3.93652678e-01 -5.18655062e-01 1.79166541e-01
3.10473770e-01 1.25899780e+00 -5.39018512e-01 5.73387623e-01
7.52777457e-01 -3.90786320e-01 3.48687351e-01 -4.14480925e-01
3.43957990e-01 -7.85094053e-02 -1.18579483e+00 2.27441573e+00
-2.23673299e-01 2.84797132e-01 4.06566501e-01 -7.00523674e-01
4.36986834e-01 3.49632651e-01 6.77415371e-01 -5.35536289e-01
1.80997089e-01 7.04322904e-02 -5.33795096e-02 -6.16730571e-01
1.92311391e-01 -3.29762280e-01 -2.17155278e-01 1.67480499e-01
-6.17624857e-02 1.08850310e-02 -2.13744923e-01 6.46197572e-02
1.02866876e+00 8.62197876e-02 2.04100013e-01 -3.86890620e-02
1.00177884e-01 1.16928078e-01 8.01118016e-01 8.80441666e-01
-5.41032195e-01 1.21911621e+00 3.72208208e-01 -2.55360782e-01
-8.81770134e-01 -1.12698829e+00 -5.17123461e-01 5.75697720e-01
2.80218273e-01 -1.34730205e-01 -9.12604213e-01 -1.23531973e+00
-1.13065422e-01 3.77492994e-01 -8.58483136e-01 -1.39464572e-01
-6.64007783e-01 -7.10328579e-01 5.50829291e-01 7.01725781e-01
3.35295528e-01 -4.81133223e-01 -6.47490144e-01 -1.28459617e-01
-4.31899041e-01 -1.55896246e+00 -8.15838218e-01 2.41440043e-01
-6.10140502e-01 -1.24042380e+00 -1.06584239e+00 -6.98301613e-01
1.20282042e+00 9.31596681e-02 7.44076431e-01 -6.29360855e-01
-6.51836574e-01 6.47333264e-01 -8.42339173e-02 -5.28801024e-01
-1.54595628e-01 -1.20903917e-01 1.79947764e-01 1.94330454e-01
2.32480884e-01 -3.01047295e-01 -1.03522384e+00 5.26200652e-01
-5.39472282e-01 1.00963920e-01 6.14993572e-01 9.04550433e-01
9.55427587e-01 -3.01226079e-01 1.89670891e-01 -1.32108319e+00
2.44600236e-01 -1.47606246e-03 -5.48288524e-01 3.30568641e-01
-5.54171085e-01 4.99691069e-02 2.84149647e-01 -5.08840382e-01
-1.08738875e+00 6.51280403e-01 1.96717948e-01 -7.03226805e-01
-3.37743193e-01 -1.45454049e-01 -1.96375459e-01 -2.52112269e-01
9.49059606e-01 -1.12814773e-02 2.58213580e-01 -3.73382568e-01
2.82080263e-01 6.38200462e-01 9.04501438e-01 -5.71943700e-01
7.97511637e-01 9.91298795e-01 1.26659483e-01 -4.39339310e-01
-9.57739532e-01 -6.77826285e-01 -7.70768642e-01 2.32064262e-01
1.18994892e+00 -1.20433450e+00 -3.53642195e-01 1.12812109e-01
-8.89818490e-01 -1.94924369e-01 -5.16232908e-01 7.46129215e-01
-6.12428188e-01 3.06827247e-01 -3.25848401e-01 -4.68777776e-01
-2.45157763e-01 -1.39974678e+00 1.48997736e+00 2.65572488e-01
-3.52672368e-01 -8.28049481e-01 -2.42008013e-03 8.28076482e-01
-1.96073614e-02 6.56097531e-01 4.03662890e-01 -8.70633185e-01
-4.86126095e-01 -5.16061187e-01 -2.05107361e-01 1.61714405e-01
2.66487569e-01 -7.49682903e-01 -1.02282488e+00 -6.25281692e-01
8.87862444e-02 -2.71831930e-01 5.47420800e-01 6.68774843e-01
1.43504691e+00 -1.64161935e-01 -7.73851335e-01 9.20932472e-01
1.02575600e+00 -1.35901406e-01 3.52019370e-01 2.05368355e-01
9.38945234e-01 7.20474243e-01 8.63851547e-01 3.16853553e-01
3.50412369e-01 7.85547137e-01 1.43559113e-01 -5.76463223e-01
-2.00314730e-01 -3.06788415e-01 -1.19595721e-01 -1.91911273e-02
-6.24748804e-02 3.99281859e-01 -9.76059318e-01 5.22848070e-01
-1.65468776e+00 -5.29635668e-01 1.34021202e-02 2.57891870e+00
8.52172792e-01 -2.45650798e-01 8.85981470e-02 -3.55346978e-01
5.68236709e-01 -1.06040567e-01 -5.90782642e-01 2.83080220e-01
2.92220503e-01 2.70770073e-01 1.15447342e+00 2.99164802e-01
-1.35869586e+00 5.46925366e-01 5.87006330e+00 5.41160643e-01
-1.17091179e+00 3.44761223e-01 8.56420338e-01 -4.95422989e-01
1.39978275e-01 -2.93676466e-01 -8.71384978e-01 2.55892485e-01
5.86564898e-01 2.14058831e-01 2.29041860e-01 9.71998990e-01
2.37305790e-01 -4.21847850e-02 -1.48980057e+00 1.36002052e+00
3.98261547e-01 -1.13099682e+00 -4.25881356e-01 3.05267453e-01
7.39972353e-01 -6.63304329e-02 2.75446087e-01 -4.46166322e-02
1.71724886e-01 -1.14679480e+00 4.50115979e-01 3.54894817e-01
1.37416232e+00 -4.43886280e-01 6.07860327e-01 3.23434561e-01
-7.48361409e-01 1.92550465e-01 -5.16532622e-02 7.10916698e-01
-5.67402318e-03 2.37092987e-01 -1.22786701e+00 4.36436474e-01
8.84077132e-01 4.51348007e-01 -6.94679558e-01 9.04432356e-01
-1.17879875e-01 2.66514301e-01 -3.39653730e-01 6.17183268e-01
-1.44170821e-01 4.17102389e-02 5.17479479e-01 1.05934298e+00
8.35815519e-02 1.79331034e-01 2.61987388e-01 6.70639455e-01
-1.44442320e-01 -1.68947622e-01 -5.33040047e-01 1.68370247e-01
2.04249084e-01 1.51443362e+00 -5.65671563e-01 -1.03544667e-01
-4.39979017e-01 1.19171369e+00 -8.55118211e-04 5.35668612e-01
-8.81743252e-01 -8.57194588e-02 5.10027647e-01 6.65170789e-01
1.23488165e-01 2.25200206e-02 -4.26810294e-01 -1.26806223e+00
2.66282111e-01 -1.00730741e+00 5.55536628e-01 -5.43013632e-01
-1.07614815e+00 4.51074272e-01 1.62666947e-01 -1.48670101e+00
-3.72044593e-01 -5.37468910e-01 -2.50322223e-01 1.06469810e+00
-1.41848350e+00 -1.44022202e+00 -6.55447662e-01 8.14491034e-01
2.83019483e-01 5.31655513e-02 9.78751302e-01 2.65930891e-01
-5.11930048e-01 1.10856819e+00 1.47566929e-01 3.78267050e-01
1.23208988e+00 -1.26786745e+00 -1.41934022e-01 8.00734162e-01
-4.99046706e-02 7.72054255e-01 5.23145437e-01 -4.19298768e-01
-1.23057461e+00 -1.25198615e+00 3.07665944e-01 -1.06599438e+00
-3.25489715e-02 -6.71343863e-01 -6.06463611e-01 1.09302557e+00
-2.72597611e-01 7.60478616e-01 9.62563634e-01 6.75382093e-02
-1.30834267e-01 -1.15302883e-01 -1.62070918e+00 6.20216429e-01
9.92215395e-01 -2.75574267e-01 -4.46243197e-01 4.90880579e-01
5.47781646e-01 -1.04775345e+00 -7.59021044e-01 4.75245386e-01
6.65557086e-01 -7.45033503e-01 9.99782741e-01 -5.53529978e-01
1.22558862e-01 -3.20936173e-01 -4.00763843e-03 -9.99855101e-01
-6.49213120e-02 -5.06531239e-01 2.03395858e-01 9.58588004e-01
3.15748066e-01 -6.64493024e-01 1.35767436e+00 1.07627809e+00
7.91789889e-02 -6.39084816e-01 -1.04508913e+00 -4.66008812e-01
-1.25616267e-01 -2.80998617e-01 2.17246324e-01 1.04046226e+00
-2.28320539e-01 7.92165026e-02 -2.38439396e-01 5.23545444e-01
9.22010899e-01 1.12438723e-01 9.55087662e-01 -9.28536117e-01
-6.01813018e-01 1.84569612e-01 -3.23845387e-01 -7.91455388e-01
-9.05479491e-02 -7.76113451e-01 6.95765903e-03 -1.34169710e+00
4.24918115e-01 -5.63581705e-01 -1.69256866e-01 5.66103339e-01
-2.73728818e-01 4.51835334e-01 -1.16567157e-01 2.16536283e-01
-5.07103622e-01 1.30096138e-01 1.54019380e+00 -6.18640035e-02
-1.90036833e-01 1.73929006e-01 -1.00328112e+00 8.88282359e-01
2.49999300e-01 -6.78746819e-01 -4.46388453e-01 -1.95538014e-01
-4.58096176e-01 1.10913202e-01 6.02545798e-01 -8.25916708e-01
1.20808527e-01 1.49978437e-02 8.61155927e-01 -4.22301620e-01
4.38576043e-01 -1.07974088e+00 1.05453823e-02 3.22543532e-01
-3.10555190e-01 -4.52804565e-01 2.99148858e-01 7.20802009e-01
7.12114498e-02 5.26806489e-02 9.86315608e-01 -2.59505123e-01
-3.04883242e-01 4.84966755e-01 2.43791983e-01 2.49714911e-01
1.22910714e+00 -3.96325737e-01 -1.01648025e-01 -2.46754810e-01
-1.06961071e+00 3.33729327e-01 6.79790378e-01 3.86095047e-01
3.98946077e-01 -1.04169810e+00 -5.82599282e-01 3.67990166e-01
5.25228918e-01 6.60797060e-01 3.38189065e-01 1.15346658e+00
-4.58021998e-01 4.13621664e-01 8.71515349e-02 -7.64170051e-01
-1.54243863e+00 5.04945457e-01 6.28073037e-01 -4.57161486e-01
-7.90903509e-01 7.37905562e-01 9.02590275e-01 -6.15079284e-01
2.58131862e-01 -2.09015384e-01 2.16461912e-01 -4.10433739e-01
5.54460764e-01 8.38694721e-02 1.37602553e-01 -6.46509707e-01
-5.09897470e-01 5.84794402e-01 -4.39985245e-01 -1.09995417e-01
1.04134738e+00 -8.75719413e-02 5.03193557e-01 -5.56173474e-02
1.10806000e+00 3.23358208e-01 -1.56753492e+00 -3.28340113e-01
-4.00757015e-01 -6.40962601e-01 -1.13591760e-01 -1.01014626e+00
-8.53371501e-01 6.26255155e-01 1.08392346e+00 -5.83870471e-01
9.70031977e-01 2.83963770e-01 5.29114664e-01 -1.15943536e-01
3.22018474e-01 -9.34171557e-01 -1.01539515e-01 -3.04015189e-01
7.31935561e-01 -1.57520080e+00 1.12748116e-01 -7.39035726e-01
-8.77502799e-01 6.75778568e-01 6.61554515e-01 1.90694869e-01
3.26573968e-01 2.85919160e-01 4.06517833e-01 -1.48398161e-01
1.19061470e-02 3.37468497e-02 6.10708892e-01 8.91124368e-01
2.29801685e-01 3.99544723e-02 -2.56754346e-02 9.47335660e-01
-1.77411333e-01 -4.20846464e-03 2.63938934e-01 9.81777132e-01
2.99088597e-01 -1.14956450e+00 -7.18635261e-01 4.49992806e-01
-6.08692825e-01 1.66441113e-01 -2.35799670e-01 8.71817708e-01
3.03345352e-01 6.18424535e-01 -1.04243129e-01 -6.23621687e-04
5.86306453e-01 -1.53108299e-01 5.94222963e-01 -8.96279633e-01
-2.99302548e-01 2.39560962e-01 -4.77014594e-02 -8.30599904e-01
-5.36141694e-02 -8.86833251e-01 -1.27652895e+00 3.22506368e-01
-4.51213792e-02 -1.34707510e-01 6.99685514e-01 7.94255257e-01
4.88377303e-01 6.27058506e-01 4.76809740e-01 -8.00892413e-01
-5.60406744e-01 -6.86555922e-01 -6.11083448e-01 6.87376320e-01
4.94446605e-01 -5.25220215e-01 -2.22404465e-01 1.39205322e-01] | [14.590875625610352, -1.995203971862793] |
a9504174-7ab1-4d8a-a434-d5195d2b5812 | aom-detecting-aspect-oriented-information-for | 2306.01004 | null | https://arxiv.org/abs/2306.01004v1 | https://arxiv.org/pdf/2306.01004v1.pdf | AoM: Detecting Aspect-oriented Information for Multimodal Aspect-Based Sentiment Analysis | Multimodal aspect-based sentiment analysis (MABSA) aims to extract aspects from text-image pairs and recognize their sentiments. Existing methods make great efforts to align the whole image to corresponding aspects. However, different regions of the image may relate to different aspects in the same sentence, and coarsely establishing image-aspect alignment will introduce noise to aspect-based sentiment analysis (i.e., visual noise). Besides, the sentiment of a specific aspect can also be interfered by descriptions of other aspects (i.e., textual noise). Considering the aforementioned noises, this paper proposes an Aspect-oriented Method (AoM) to detect aspect-relevant semantic and sentiment information. Specifically, an aspect-aware attention module is designed to simultaneously select textual tokens and image blocks that are semantically related to the aspects. To accurately aggregate sentiment information, we explicitly introduce sentiment embedding into AoM, and use a graph convolutional network to model the vision-text and text-text interaction. Extensive experiments demonstrate the superiority of AoM to existing methods. The source code is publicly released at https://github.com/SilyRab/AoM. | ['Xiaojie Yuan', 'Ying Zhang', 'Shenglong Yu', 'Xumeng Liu', 'Wenya Guo', 'Ru Zhou'] | 2023-05-31 | null | null | null | null | ['sentiment-analysis'] | ['natural-language-processing'] | [ 1.82989717e-01 -5.20903990e-02 -8.09174404e-03 -4.81317997e-01
-6.21205926e-01 -6.03731990e-01 5.60812533e-01 1.47337645e-01
-1.68481767e-01 1.53952213e-02 4.42828238e-01 -2.40168963e-02
3.12779516e-01 -6.89188004e-01 -5.74581504e-01 -6.07077956e-01
7.91281641e-01 1.67056412e-01 -5.16000111e-03 -3.83657873e-01
3.21161419e-01 6.65604845e-02 -1.06089687e+00 5.71273327e-01
4.46886778e-01 9.35942233e-01 3.01572055e-01 2.96471775e-01
-6.05429828e-01 7.59806454e-01 -7.27903664e-01 -5.91891170e-01
-1.71126544e-01 -6.05986416e-01 -4.90788758e-01 7.94762433e-01
3.56463373e-01 -5.62705062e-02 -3.48617226e-01 1.40359485e+00
2.91194588e-01 3.41642462e-03 6.30086362e-01 -1.43950796e+00
-9.91343677e-01 4.46309328e-01 -1.03144193e+00 3.47203389e-03
2.13826939e-01 3.68021429e-01 1.04817140e+00 -1.20262611e+00
3.85502845e-01 1.34124994e+00 1.85772538e-01 3.19909483e-01
-9.42144215e-01 -4.65914547e-01 6.41791761e-01 1.52282879e-01
-1.12468648e+00 -2.95188040e-01 1.22079372e+00 -2.23430514e-01
4.63851988e-01 3.64430785e-01 1.02392983e+00 9.74908292e-01
1.05057873e-01 1.25777948e+00 1.04089105e+00 1.72611047e-02
-3.02166231e-02 4.60824907e-01 1.54063284e-01 5.89026153e-01
4.39846992e-01 -4.66889799e-01 -6.09600067e-01 2.84315258e-01
3.23578805e-01 3.09719980e-01 -3.08997154e-01 -5.58575153e-01
-1.31578493e+00 5.79440415e-01 5.10480165e-01 3.00325871e-01
-3.92103493e-01 6.58616722e-02 3.56224567e-01 9.70181637e-03
3.84952366e-01 2.01845303e-01 -2.31114328e-01 7.35048503e-02
-6.19222164e-01 -1.25178158e-01 3.62365693e-01 1.08772123e+00
1.10669243e+00 1.59895912e-01 -3.33026677e-01 7.03613400e-01
5.54670513e-01 7.83300161e-01 3.46354783e-01 -5.60860336e-01
6.72933102e-01 1.19498754e+00 -1.35696337e-01 -1.36881626e+00
-3.61243606e-01 -3.20419341e-01 -8.18371952e-01 2.94600017e-02
8.97324383e-02 -8.68975092e-03 -9.72406864e-01 1.24649489e+00
4.25355822e-01 -3.61404866e-01 8.71122628e-02 1.26191461e+00
1.14464116e+00 7.04680741e-01 5.37039973e-02 1.29136100e-01
1.79462564e+00 -1.24262190e+00 -8.35811019e-01 -7.34068394e-01
2.15621233e-01 -1.13951135e+00 1.30197620e+00 -5.77962399e-02
-9.74998951e-01 -4.03637379e-01 -8.69169831e-01 -1.52360603e-01
-4.79831457e-01 2.67266989e-01 5.71131527e-01 2.91814327e-01
-8.84783626e-01 -3.50460023e-01 -6.93676412e-01 -2.38739312e-01
5.66771209e-01 -3.28650326e-02 -3.93516630e-01 -3.42981905e-01
-8.77252698e-01 4.26476866e-01 1.89742103e-01 3.11040699e-01
-7.17450202e-01 -3.03449929e-01 -1.20803678e+00 1.15019418e-01
5.79775691e-01 -7.94447362e-01 1.06238902e+00 -1.45160592e+00
-8.78869236e-01 7.46675909e-01 -5.25828004e-01 2.75602877e-01
5.19290790e-02 -1.02103084e-01 -4.60333228e-01 4.61682230e-01
4.50878590e-01 6.54168725e-01 9.84056056e-01 -1.68762112e+00
-5.75275958e-01 -6.39748216e-01 4.75647181e-01 5.60252190e-01
-5.71816504e-01 -1.37193520e-02 -1.22855496e+00 -7.90732265e-01
2.01361120e-01 -6.30281746e-01 -2.88606852e-01 -4.87038344e-02
-7.36409366e-01 2.68404365e-01 8.58854651e-01 -4.86858875e-01
9.29584861e-01 -2.16161036e+00 1.95609517e-02 6.25003800e-02
3.67587209e-01 -6.29897835e-03 -5.45203507e-01 3.20773154e-01
-6.96476270e-03 -1.85792185e-02 -4.24053878e-01 -3.23608339e-01
-1.23918518e-01 1.27059773e-01 -7.90960118e-02 3.22275072e-01
3.08548480e-01 1.36606908e+00 -8.34770441e-01 -6.63928926e-01
2.85924673e-01 6.25979006e-01 -3.77040118e-01 2.10526530e-02
-3.03271681e-01 4.78446871e-01 -7.22000182e-01 9.96965706e-01
7.43361056e-01 -3.45360100e-01 -1.57761648e-01 -6.83342338e-01
1.99655280e-01 3.81329632e-03 -8.60748887e-01 1.50743878e+00
-6.47448123e-01 9.14563835e-01 1.07686438e-01 -1.12659609e+00
7.81524658e-01 1.21721201e-01 5.16894341e-01 -7.54900277e-01
3.34734380e-01 -1.10312626e-01 -2.59482145e-01 -6.16074860e-01
7.50800848e-01 -1.00094914e-01 -1.30462065e-01 5.49962342e-01
-5.39520010e-02 -3.07834685e-01 8.44354704e-02 3.17368776e-01
3.70271176e-01 -1.54283732e-01 2.48985082e-01 5.23103364e-02
7.32973397e-01 2.59293973e-01 4.23804611e-01 4.15629148e-01
-1.60030931e-01 8.22804570e-01 7.80647039e-01 -3.72117937e-01
-8.36785614e-01 -8.34292412e-01 3.47288668e-01 9.00651157e-01
8.71558249e-01 -5.01047671e-01 -6.58314228e-01 -7.13901937e-01
-3.44712764e-01 6.08033538e-01 -6.53711498e-01 -1.84002951e-01
-2.22349688e-01 -6.81314528e-01 -4.27767076e-02 4.12623048e-01
5.55687726e-01 -1.14059746e+00 -2.61282951e-01 -6.80917427e-02
-5.12027323e-01 -1.26796818e+00 -1.03536928e+00 -1.66740119e-01
-4.78318751e-01 -1.21817076e+00 -6.76059127e-01 -9.15118456e-01
1.10426068e+00 8.61611962e-01 1.21424031e+00 1.88212484e-01
4.48685093e-03 7.43643105e-01 -4.11481440e-01 -4.62257802e-01
-9.17369276e-02 -1.22279875e-01 -4.74344075e-01 4.54558760e-01
8.50114644e-01 -2.82988757e-01 -9.69351590e-01 2.91093290e-01
-1.14473259e+00 2.92516768e-01 8.30567777e-01 7.46505022e-01
8.98177505e-01 1.90932706e-01 -9.37816650e-02 -9.43810165e-01
5.79563618e-01 -3.85068595e-01 -4.51249212e-01 1.77550837e-01
-4.47895378e-01 -3.83307278e-01 4.58836198e-01 -2.73609966e-01
-9.84013140e-01 2.07664490e-01 -2.81849299e-02 -4.63379234e-01
-2.55694836e-01 5.52377045e-01 -7.04292357e-01 2.43389785e-01
-5.02455533e-02 6.84108675e-01 -9.72989351e-02 3.96724157e-02
5.37265301e-01 6.38785899e-01 2.81313360e-01 -2.18993753e-01
7.99967647e-01 8.83807719e-01 -1.55889079e-01 -7.81245172e-01
-1.28428054e+00 -7.43791163e-01 -3.58889341e-01 -5.21106422e-01
8.51422131e-01 -1.08313751e+00 -4.05301303e-01 4.58043516e-01
-1.21166408e+00 1.19206078e-01 -1.81950450e-01 2.60995209e-01
-4.39938366e-01 5.29570997e-01 -2.94865161e-01 -5.86856008e-01
-3.46949130e-01 -1.53036928e+00 1.51880455e+00 4.11959231e-01
-7.79717118e-02 -9.19986427e-01 -4.03453231e-01 8.00708175e-01
2.65249819e-01 -7.66481459e-02 5.89800894e-01 -3.73787791e-01
-7.42351115e-01 -3.38977963e-01 -4.77019131e-01 2.82352328e-01
2.87748486e-01 -1.57139972e-01 -9.38257635e-01 -1.32824080e-02
2.01784253e-01 2.77891792e-02 9.20196712e-01 4.05731559e-01
8.97736311e-01 -3.84311229e-01 -1.86128870e-01 4.65137869e-01
1.43903542e+00 1.14119880e-01 6.04900599e-01 5.90732336e-01
1.28827310e+00 6.68436229e-01 8.65340471e-01 1.33157447e-01
6.59039199e-01 6.13412559e-01 5.81457794e-01 -4.78970468e-01
-1.91049591e-01 -2.42853507e-01 4.87382919e-01 8.08293164e-01
4.52046037e-01 -2.95854509e-01 -7.38842785e-01 7.24039912e-01
-1.80665267e+00 -7.60513008e-01 -3.63266230e-01 1.63634682e+00
4.38781947e-01 1.05550334e-01 -1.11833975e-01 -1.27000585e-01
7.66004562e-01 5.98440349e-01 -6.67310536e-01 -2.50650465e-01
-4.40067142e-01 -4.00812358e-01 1.61186516e-01 4.34290469e-01
-1.14230323e+00 9.76458192e-01 4.49749708e+00 7.97238231e-01
-1.05792463e+00 2.55733162e-01 9.20889497e-01 -1.37952745e-01
-9.47489023e-01 2.49516405e-02 -4.53091115e-01 3.71347576e-01
1.31843597e-01 -1.59928605e-01 1.81079924e-01 8.68871272e-01
3.18962157e-01 -5.20065017e-02 -5.67007899e-01 9.20645118e-01
5.23154378e-01 -9.72806633e-01 6.03081226e-01 -7.55350962e-02
9.07715857e-01 -1.75503746e-01 2.48081669e-01 9.87428874e-02
-1.84132382e-01 -7.53743470e-01 7.91894495e-01 5.09569705e-01
5.34016430e-01 -8.36053431e-01 8.53644133e-01 -1.13069408e-01
-1.51688802e+00 2.99247324e-01 -5.31542420e-01 2.54512519e-01
1.63245142e-01 8.39807153e-01 -1.53528735e-01 5.27972162e-01
7.54375935e-01 1.05623341e+00 -7.14589298e-01 5.98349512e-01
-5.43951035e-01 3.31863344e-01 1.61120042e-01 -9.61432755e-02
4.28345144e-01 -5.17286360e-01 7.59409904e-01 9.71729815e-01
1.74873561e-01 -2.11313784e-01 3.65986722e-03 1.02194440e+00
-3.06496657e-02 3.13155055e-01 -6.82963789e-01 -2.38205016e-01
2.58243214e-02 1.67523587e+00 -8.98901105e-01 -4.21450078e-01
-9.90777671e-01 1.29244292e+00 1.58085313e-03 7.38168299e-01
-5.94753921e-01 -3.51306140e-01 8.83560717e-01 -1.04607113e-01
4.18160737e-01 7.34705594e-04 -6.09094620e-01 -1.38193178e+00
4.16549265e-01 -8.41399491e-01 3.11621875e-01 -1.22289777e+00
-1.12768185e+00 7.15127766e-01 -5.31526804e-01 -1.64731812e+00
3.41841668e-01 -4.82879043e-01 -8.40830624e-01 7.92219877e-01
-1.68257535e+00 -1.34234643e+00 -5.66915929e-01 6.13458514e-01
7.94282556e-01 -1.39958352e-01 3.06392252e-01 1.52827069e-01
-4.41675514e-01 4.38973427e-01 -2.14951098e-01 2.20577285e-01
5.23186624e-01 -1.19616520e+00 3.48895252e-01 1.04358327e+00
3.12031060e-01 5.83667636e-01 4.76298481e-01 -5.03238261e-01
-1.63001847e+00 -1.46278656e+00 7.77052999e-01 -6.29713058e-01
8.07778895e-01 -2.96594173e-01 -8.14985573e-01 6.63142562e-01
4.67522889e-01 -3.40702683e-01 5.31378329e-01 -1.75838575e-01
-4.00276512e-01 -2.13289723e-01 -5.06111145e-01 1.02692139e+00
6.29383087e-01 -7.96090245e-01 -4.62630719e-01 1.35764614e-01
8.17330360e-01 -2.39741907e-01 -4.90836620e-01 2.85471231e-01
3.76273960e-01 -1.08865964e+00 9.62617636e-01 -2.27622285e-01
8.15336466e-01 -7.18636096e-01 -1.72831759e-01 -1.42082489e+00
4.37241755e-02 3.32476273e-02 1.29008725e-01 1.47495067e+00
4.52284902e-01 -4.61212665e-01 5.69791257e-01 3.50371957e-01
-3.45354690e-03 -7.74714351e-01 -6.14586055e-01 -2.47441307e-01
-2.70406604e-01 -6.28136575e-01 7.57930219e-01 8.95774961e-01
-1.22351842e-02 8.06330800e-01 -3.06855232e-01 4.45474178e-01
3.98604751e-01 7.43311167e-01 7.98698306e-01 -5.67682147e-01
-1.14377243e-02 -8.21449161e-01 -5.16552389e-01 -9.35391068e-01
4.73672785e-02 -7.33347654e-01 5.03036678e-02 -1.98763454e+00
6.93598568e-01 1.17950313e-01 -3.17370802e-01 4.61752713e-01
-7.65625536e-01 6.26523137e-01 2.38190413e-01 4.97575887e-02
-8.91183257e-01 8.59585404e-01 1.81334198e+00 -8.03673685e-01
1.60597235e-01 -1.02593377e-01 -1.25938940e+00 7.84350574e-01
1.03240466e+00 -3.26807737e-01 -4.45656300e-01 -4.67453867e-01
4.52198416e-01 -3.29450965e-02 3.55114818e-01 -5.96692979e-01
1.08662419e-01 -1.67423457e-01 5.00657856e-01 -8.24176192e-01
4.72010702e-01 -1.19714189e+00 -2.16686726e-01 3.35109681e-01
-2.44915769e-01 3.51008102e-02 1.66285500e-01 6.91782117e-01
-8.43226552e-01 -2.62587648e-02 4.14081693e-01 -1.96585447e-01
-8.95717978e-01 4.56997424e-01 -4.67970550e-01 2.37223711e-02
7.99578428e-01 -1.42715290e-01 -4.40991282e-01 -7.81362951e-01
-3.14969361e-01 4.51487213e-01 6.62119985e-01 7.03338683e-01
9.26396608e-01 -1.33876443e+00 -4.63180870e-01 1.75428137e-01
7.69183815e-01 6.78658783e-02 6.16335213e-01 9.92785096e-01
-1.37418777e-01 3.11577082e-01 1.55130267e-01 -5.62622488e-01
-1.34551072e+00 6.85777366e-01 2.77606398e-01 1.13857761e-01
-5.68835557e-01 5.98259151e-01 9.21324492e-01 -5.62118828e-01
-7.67198280e-02 2.04220340e-02 -5.04988253e-01 1.41923264e-01
4.95476842e-01 -2.71211088e-01 -3.12906623e-01 -1.03972054e+00
-4.21304047e-01 8.95318210e-01 -1.36709176e-02 -5.22257714e-03
1.02962732e+00 -5.76279521e-01 -3.09841454e-01 4.83800560e-01
1.23709857e+00 1.39329091e-01 -1.01044488e+00 -3.09507698e-01
-3.37225258e-01 -3.68476719e-01 8.11174735e-02 -6.68478549e-01
-1.70017636e+00 1.08874047e+00 2.04361364e-01 -4.45717499e-02
1.45586300e+00 1.98018417e-01 7.88924336e-01 1.54587865e-01
-1.78157076e-01 -9.10759866e-01 4.17517900e-01 3.28487903e-01
7.91985452e-01 -1.56668270e+00 -6.41803518e-02 -3.78376037e-01
-1.27238071e+00 1.00603402e+00 8.83424401e-01 2.73983628e-01
4.40595984e-01 1.34007856e-02 6.26303136e-01 -6.88989758e-01
-5.31182170e-01 -5.09209335e-01 5.54181039e-01 5.39458096e-01
3.00592899e-01 1.40912548e-01 -4.38479297e-02 7.34542787e-01
-1.46126589e-02 -5.58599293e-01 6.03493094e-01 5.65385938e-01
-3.03193569e-01 -7.23739803e-01 -4.62029278e-01 3.95519376e-01
-3.91616225e-01 -4.63447154e-01 -5.16110837e-01 5.66220164e-01
-2.16305360e-01 1.20055115e+00 2.52717528e-02 -3.30404341e-01
4.62676167e-01 -2.14874953e-01 -1.02026559e-01 -4.46990311e-01
-3.98381889e-01 5.02732098e-01 -1.06108226e-01 -5.92396080e-01
-6.68241858e-01 -4.42114323e-01 -1.20337152e+00 -4.78616431e-02
-5.83662093e-03 8.20135698e-02 7.41265714e-01 8.33296776e-01
2.82965273e-01 7.86791444e-01 7.01640666e-01 -4.46394086e-01
4.92469341e-01 -7.06459701e-01 -6.60618603e-01 5.12327492e-01
4.43735123e-01 -2.83361793e-01 -2.82433003e-01 1.01730928e-01] | [10.7708740234375, 1.633555293083191] |
ea54e8b0-a724-49c2-95ff-7511b78af1fc | efficientphys-enabling-simple-fast-and-1 | 2110.04447 | null | https://arxiv.org/abs/2110.04447v3 | https://arxiv.org/pdf/2110.04447v3.pdf | EfficientPhys: Enabling Simple, Fast and Accurate Camera-Based Vitals Measurement | Camera-based physiological measurement is a growing field with neural models providing state-the-art-performance. Prior research have explored various "end-to-end" models; however these methods still require several preprocessing steps. These additional operations are often non-trivial to implement making replication and deployment difficult and can even have a higher computational budget than the "core" network itself. In this paper, we propose two novel and efficient neural models for camera-based physiological measurement called EfficientPhys that remove the need for face detection, segmentation, normalization, color space transformation or any other preprocessing steps. Using an input of raw video frames, our models achieve strong performance on three public datasets. We show that this is the case whether using a transformer or convolutional backbone. We further evaluate the latency of the proposed networks and show that our most light weight network also achieves a 33% improvement in efficiency. | ['Daniel McDuff', 'Shwetak Patel', 'Ziheng Jiang', 'Brian L. Hill', 'Xin Liu'] | 2021-10-09 | null | null | null | null | ['photoplethysmography-ppg-heart-rate'] | ['medical'] | [ 1.65142000e-01 -3.43873829e-01 1.99242562e-01 -7.00965405e-01
-6.05128348e-01 -5.14532089e-01 2.30050176e-01 -7.39942193e-02
-1.00039899e+00 3.82074863e-01 -2.40834691e-02 -2.51735210e-01
2.73351312e-01 -3.33200812e-01 -9.92827833e-01 -5.07595778e-01
-3.24127935e-02 4.06574979e-02 1.56134039e-01 2.88356364e-01
1.67168051e-01 6.47349000e-01 -1.34547305e+00 1.83781698e-01
4.05981690e-02 1.35647905e+00 -1.17828608e-01 8.72485876e-01
1.65433273e-01 4.89226520e-01 -5.52063584e-01 -5.61549246e-01
5.19915521e-01 -5.16478941e-02 -4.61623967e-01 -2.00232536e-01
7.98994243e-01 -8.81114483e-01 -3.43784690e-01 1.00529742e+00
1.05613494e+00 -1.60382941e-01 -4.47720923e-02 -1.51715779e+00
-3.05101454e-01 4.85715121e-01 -6.31286860e-01 3.47562671e-01
7.80256018e-02 1.52691826e-01 4.32866007e-01 -8.42500091e-01
1.96941718e-01 1.19223559e+00 8.27718437e-01 6.82890534e-01
-1.30651093e+00 -1.09819210e+00 1.15063190e-02 3.20640653e-01
-1.47711539e+00 -1.09918165e+00 5.01002371e-01 3.07014119e-02
1.27308142e+00 -1.90853514e-02 7.08356678e-01 1.32880926e+00
1.03033558e-01 4.62440759e-01 1.22648728e+00 1.09495344e-02
3.42798948e-01 -7.68185854e-02 2.14482084e-01 6.16618752e-01
3.07026386e-01 -1.52043670e-01 -6.74493492e-01 -1.12059496e-01
8.60291362e-01 4.42326488e-03 -4.05088305e-01 -4.03519012e-02
-1.00387204e+00 4.67272550e-01 2.33895585e-01 -3.26239653e-02
-3.11502188e-01 8.03833604e-01 6.39266372e-01 2.62388259e-01
3.28140050e-01 1.58852249e-01 -6.07946575e-01 -5.02086520e-01
-1.39088488e+00 -1.70858130e-02 7.96702206e-01 1.05592918e+00
4.88862455e-01 -1.27366846e-02 9.55262110e-02 6.81605637e-01
9.62427109e-02 5.08019149e-01 4.61112678e-01 -1.32348394e+00
6.66114315e-02 1.84254378e-01 -3.94103192e-02 -8.60102952e-01
-8.24548602e-01 -2.35662311e-01 -6.84681833e-01 5.10284975e-02
4.59921747e-01 -5.63470244e-01 -8.90288591e-01 2.01711154e+00
1.10690296e-01 5.81663072e-01 -3.00486803e-01 7.99160302e-01
6.35390699e-01 3.02948266e-01 1.16754137e-02 -1.23706624e-01
1.39701486e+00 -9.13977444e-01 -3.98846477e-01 -3.19133729e-01
1.88715383e-01 -6.86691999e-01 7.67155826e-01 5.41966796e-01
-1.42997980e+00 -2.63269186e-01 -1.05249989e+00 -1.74489483e-01
-2.06884235e-01 1.48211822e-01 9.54873741e-01 1.04990065e+00
-1.69984424e+00 5.80370426e-01 -1.00043929e+00 -5.90835452e-01
5.96720040e-01 9.10877168e-01 -5.23305595e-01 7.99709037e-02
-5.88804185e-01 7.98859179e-01 3.69326444e-03 1.59140423e-01
-1.10101676e+00 -8.39252770e-01 -3.96718323e-01 4.08385128e-01
2.26898447e-01 -4.82504517e-01 1.29018772e+00 -1.08719862e+00
-1.71893787e+00 7.82831490e-01 -2.54644394e-01 -5.49365282e-01
2.65899330e-01 -8.23358595e-02 -1.28726050e-01 3.08912814e-01
-4.65805322e-01 1.12884593e+00 8.80094707e-01 -5.54169834e-01
-2.99362004e-01 -3.78241986e-01 2.14535877e-01 -9.69051942e-02
-8.45179021e-01 4.65253681e-01 -9.23187792e-01 -2.30977654e-01
-1.16081625e-01 -1.08776426e+00 -3.48627195e-02 4.93586719e-01
-1.21193081e-01 7.89021552e-02 6.03142440e-01 -5.84692001e-01
7.73651063e-01 -2.19275308e+00 -4.30885673e-01 2.12595508e-01
5.00032663e-01 1.36335209e-01 -3.78399193e-01 -2.31444761e-01
-1.85327426e-01 1.08551398e-01 1.72344387e-01 -4.79215741e-01
6.87960461e-02 -1.28382996e-01 1.43098190e-01 7.24313617e-01
-3.12137306e-02 7.77828336e-01 -3.37210923e-01 -3.84308040e-01
9.42945182e-02 9.28554893e-01 -8.17098796e-01 -1.66834239e-02
2.21421406e-01 -4.33720201e-02 2.69478410e-01 6.34621203e-01
7.43164122e-01 -2.34885871e-01 2.79764622e-01 -4.15044636e-01
5.11154123e-02 2.66949594e-01 -1.05182207e+00 1.81190026e+00
-2.73880333e-01 9.50049222e-01 4.30865437e-01 -8.71316016e-01
5.37222683e-01 4.56403375e-01 5.45723975e-01 -8.37829232e-01
6.71782434e-01 -1.78789590e-02 2.24761710e-01 -1.99242562e-01
2.53240675e-01 2.01783121e-01 2.75835067e-01 4.54971582e-01
1.95593461e-01 4.40001756e-01 4.92770150e-02 5.26233539e-02
1.45286953e+00 1.47258088e-01 -1.07851863e-01 -2.80908912e-01
1.33485839e-01 -4.14900482e-01 6.19244933e-01 5.70639849e-01
-6.90842986e-01 5.15364468e-01 5.94040334e-01 -3.84399921e-01
-1.03730917e+00 -8.32289279e-01 5.15868999e-02 1.20707536e+00
-1.96412891e-01 -4.86553907e-01 -1.37436700e+00 -1.58989847e-01
-3.04000288e-01 1.26555800e-01 -5.75044513e-01 -1.30803391e-01
-3.75424594e-01 -9.99641836e-01 1.08615887e+00 7.92476892e-01
7.33116210e-01 -6.74786508e-01 -1.12759352e+00 2.27108046e-01
-2.27931932e-01 -1.59573400e+00 -4.41660851e-01 2.19193056e-01
-9.93885934e-01 -7.80874729e-01 -6.57739818e-01 -4.20361608e-01
7.14295447e-01 4.50729877e-01 8.83636832e-01 3.91857475e-02
-3.50920737e-01 4.68869418e-01 1.42594740e-01 -5.13599038e-01
4.38157022e-01 -6.88057542e-02 1.55712590e-01 -2.48396620e-02
6.84653401e-01 -8.15780282e-01 -9.25781786e-01 1.26063555e-01
-5.45282841e-01 2.82926727e-02 5.36136866e-01 3.95901442e-01
4.30204213e-01 -3.15946996e-01 2.71771789e-01 -5.09405851e-01
6.10138237e-01 -1.95396021e-01 -8.21183205e-01 6.51312992e-02
-6.30267859e-01 -3.28085870e-02 4.46954757e-01 -5.97680569e-01
-6.83295846e-01 3.61251414e-01 -1.40516073e-01 -7.19107747e-01
-2.42345914e-01 2.81870514e-01 -1.41117778e-02 -5.67564905e-01
4.73669767e-01 1.16802566e-01 -1.37953162e-01 -3.77990127e-01
1.06013395e-01 5.05913317e-01 6.07302547e-01 -5.05788207e-01
4.94642615e-01 5.72967291e-01 -6.79788296e-04 -7.13601410e-01
-3.16921234e-01 -3.01607162e-01 -4.14776415e-01 -3.04358453e-01
8.04789364e-01 -1.17603397e+00 -1.27351689e+00 6.50485337e-01
-1.30361867e+00 -3.63861293e-01 2.29004204e-01 5.33922970e-01
-3.13240021e-01 1.92987770e-01 -8.19522500e-01 -5.85335493e-01
-8.09050322e-01 -1.21391523e+00 8.84672284e-01 3.27497482e-01
-4.34681624e-02 -5.61521292e-01 -1.68297008e-01 2.82438338e-01
8.55157673e-01 4.00909930e-02 4.95625526e-01 -3.54040653e-01
-5.35278082e-01 -2.75857121e-01 -6.20171010e-01 2.06468001e-01
-4.25563693e-01 3.75925116e-02 -1.55029058e+00 -4.31685030e-01
2.24856548e-02 -3.82792354e-01 5.75387776e-01 6.04123592e-01
1.55776620e+00 6.09810911e-02 -1.43746167e-01 1.03768170e+00
1.45565021e+00 6.13777414e-02 8.22423160e-01 5.74058034e-02
5.93698919e-01 2.40046337e-01 -2.77842611e-01 5.38050115e-01
3.42717230e-01 5.40448904e-01 4.44210172e-01 -1.73352987e-01
-5.17585054e-02 2.49162078e-01 7.88752437e-01 7.07547307e-01
-2.21243188e-01 -8.68649185e-02 -6.99244022e-01 5.87384582e-01
-1.57169640e+00 -9.89783406e-01 9.92934406e-02 2.26004052e+00
7.39696622e-01 5.76670729e-02 1.54768959e-01 4.17553335e-02
6.53313994e-01 -1.73807144e-01 -6.68111026e-01 -6.59635723e-01
1.32403582e-01 7.76841760e-01 1.01443911e+00 1.26889825e-01
-9.15174425e-01 9.26139235e-01 6.77657366e+00 3.69058937e-01
-1.52222741e+00 4.97867227e-01 5.72177529e-01 -9.04506207e-01
4.40774858e-01 -1.97551772e-01 -7.90235043e-01 3.81382495e-01
1.61952829e+00 6.47072941e-02 7.49245107e-01 8.50088060e-01
3.49231809e-01 -2.47597679e-01 -1.25415826e+00 1.57777178e+00
2.66487390e-01 -1.13908041e+00 -3.95345360e-01 1.32879317e-01
2.35268116e-01 5.63594460e-01 -1.08401522e-01 1.28563941e-01
7.33976439e-02 -1.03570402e+00 7.35553503e-01 2.16319010e-01
9.30532217e-01 -7.55856395e-01 4.66522217e-01 -8.64284411e-02
-1.02249408e+00 -6.95910752e-02 -7.42539167e-01 -7.29947761e-02
-1.28588602e-01 5.11770070e-01 -4.51586425e-01 -2.11995587e-01
1.07085943e+00 2.82734036e-01 -8.50127339e-01 1.27813160e+00
9.49253589e-02 7.53333509e-01 -7.04543233e-01 -5.18121831e-02
1.36876315e-01 3.82783830e-01 -8.20722282e-02 1.34004641e+00
3.46917331e-01 7.59772286e-02 -1.96392551e-01 5.55376530e-01
-6.90396845e-01 -2.28184531e-03 -2.68583685e-01 2.09873039e-02
4.32017118e-01 1.61856794e+00 -1.02322245e+00 -3.70628089e-01
-6.63797498e-01 1.17217636e+00 3.55158418e-01 4.08414938e-02
-1.37517965e+00 -4.58235651e-01 8.93625677e-01 -9.58343223e-02
1.87886402e-01 -2.50943303e-01 -1.01934806e-01 -1.16329205e+00
1.29964247e-01 -8.43416214e-01 -2.18953043e-01 -8.60152960e-01
-8.11740577e-01 4.31626052e-01 -1.39153510e-01 -5.91429472e-01
1.58589050e-01 -9.06263947e-01 -3.76350820e-01 5.79674184e-01
-1.52663291e+00 -8.05168569e-01 -6.26610219e-01 1.16975951e+00
2.21361056e-01 2.56623834e-01 9.50044632e-01 9.98749971e-01
-8.66439641e-01 1.10515380e+00 -1.42905012e-01 4.28378254e-01
9.74939227e-01 -8.45191121e-01 4.47894394e-01 1.06911302e+00
-8.20924640e-02 7.80247390e-01 3.88606757e-01 -5.90059794e-02
-1.86343765e+00 -8.83957148e-01 7.75867820e-01 -1.45838380e-01
4.10581261e-01 -8.61404777e-01 -3.89327914e-01 7.57000685e-01
4.67001498e-01 3.73112112e-01 7.14521289e-01 1.69020057e-01
-6.45357192e-01 -4.90244418e-01 -1.15956366e+00 7.25698948e-01
1.18838787e+00 -5.80875099e-01 5.69768921e-02 3.34724993e-01
3.07447255e-01 -2.74602294e-01 -7.16451406e-01 -6.01377152e-02
9.07624125e-01 -1.11054850e+00 7.43467867e-01 -3.25708628e-01
1.15980126e-01 -9.48930606e-02 -1.27888471e-01 -9.82010663e-01
-3.21321994e-01 -7.91059494e-01 3.87557857e-02 1.01419270e+00
3.97316366e-01 -5.98552585e-01 8.91282082e-01 1.09464335e+00
4.33852337e-02 -5.26873529e-01 -8.92321467e-01 -6.94775283e-01
-2.18539149e-01 -8.29791665e-01 6.30329072e-01 7.19705403e-01
4.77616750e-02 3.03053916e-01 -4.17113841e-01 6.71588704e-02
5.85455060e-01 -3.36442441e-01 7.28687406e-01 -1.07524633e+00
-1.72593430e-01 -4.78808790e-01 -6.94100261e-01 -8.45823824e-01
8.30339268e-03 -5.72465837e-01 1.16889462e-01 -1.11664236e+00
5.08028209e-01 7.68894842e-03 -3.76330644e-01 8.29941630e-01
2.63318539e-01 6.83963537e-01 1.78605705e-01 -2.59114116e-01
-6.61682189e-01 3.66269331e-03 5.61613560e-01 1.29975975e-01
1.68909088e-01 -6.50385141e-01 -7.65395522e-01 8.19867909e-01
9.41245675e-01 -5.09711862e-01 -5.27156472e-01 -9.60334361e-01
2.06189230e-01 -1.93760902e-01 5.44011772e-01 -1.47679842e+00
5.97158432e-01 2.65705466e-01 7.67893314e-01 -8.86027440e-02
6.70637488e-01 -1.01170492e+00 4.38144952e-02 3.44610751e-01
-2.47504294e-01 6.26419425e-01 4.57517475e-01 2.31247470e-01
3.13823611e-01 8.66240785e-02 1.02772248e+00 -6.88681304e-02
-4.55609560e-01 5.01660883e-01 -3.83088201e-01 -5.31571433e-02
9.27214026e-01 -2.43975520e-01 -5.51945090e-01 -3.44413161e-01
-4.94101375e-01 -2.58558303e-01 3.43306988e-01 9.12832916e-02
6.01909339e-01 -8.71734679e-01 -3.67242455e-01 1.26268730e-01
-2.29223654e-01 -6.65553153e-01 2.84707900e-02 1.26817524e+00
-6.14786327e-01 2.95223057e-01 -5.72041214e-01 -3.93929690e-01
-1.48838723e+00 4.59735483e-01 5.08157253e-01 1.73076332e-01
-3.97073984e-01 9.99727249e-01 -9.50317159e-02 -7.38831833e-02
4.94992286e-01 -4.78787959e-01 2.14091361e-01 -4.95834574e-02
6.81238353e-01 4.46445167e-01 2.51277208e-01 -4.13207978e-01
-5.72179317e-01 4.01158452e-01 -6.08047619e-02 -2.03484058e-01
1.43970072e+00 -5.34318313e-02 -1.95793688e-01 3.07780765e-02
1.27246571e+00 -3.32301676e-01 -1.26353014e+00 2.14221068e-02
-2.00527057e-01 -1.89383417e-01 5.22879124e-01 -7.37414598e-01
-1.78229272e+00 1.19030714e+00 1.08438182e+00 -3.24577153e-01
1.45893943e+00 -6.23237848e-01 9.81368005e-01 6.35604143e-01
4.74603713e-01 -1.30855632e+00 -3.75717849e-01 3.85676235e-01
3.29978377e-01 -9.48654532e-01 -1.43709341e-02 -2.09850132e-01
-1.19897053e-01 1.04908061e+00 7.40920007e-01 -1.59580529e-01
7.65772641e-01 8.22756886e-01 -3.42656821e-02 -1.94103718e-01
-9.64495182e-01 -1.99052449e-02 -2.95299679e-01 3.74557137e-01
6.09923303e-01 -1.05546199e-01 -1.23970397e-01 4.37693894e-01
-6.69551194e-02 4.00537640e-01 5.42858958e-01 8.93679857e-01
-2.68847764e-01 -8.57342005e-01 -1.70718342e-01 5.79292536e-01
-9.12032545e-01 -4.78866577e-01 -3.63354534e-01 6.13191426e-01
1.55156115e-02 1.08184838e+00 1.20555013e-01 -4.89468873e-01
2.24943399e-01 2.84723639e-01 7.43473172e-01 -2.55105913e-01
-9.24823046e-01 2.28408173e-01 1.25249650e-03 -1.18742442e+00
-5.28220177e-01 -4.47005630e-01 -1.12742329e+00 -1.05334818e+00
-2.76273161e-01 -4.17959154e-01 1.14019668e+00 8.16693246e-01
8.37574124e-01 3.50696862e-01 2.63938874e-01 -9.94164348e-01
-1.92679808e-01 -8.32532585e-01 -1.80955142e-01 8.24521706e-02
-8.28984077e-04 -4.64998007e-01 -1.35395482e-01 2.41654679e-01] | [8.636618614196777, 2.752849578857422] |
c87206aa-0ddf-426b-8d99-52f28d1f0a2e | deep-incomplete-multi-view-clustering-with | 2303.15689 | null | https://arxiv.org/abs/2303.15689v2 | https://arxiv.org/pdf/2303.15689v2.pdf | Deep Incomplete Multi-view Clustering with Cross-view Partial Sample and Prototype Alignment | The success of existing multi-view clustering relies on the assumption of sample integrity across multiple views. However, in real-world scenarios, samples of multi-view are partially available due to data corruption or sensor failure, which leads to incomplete multi-view clustering study (IMVC). Although several attempts have been proposed to address IMVC, they suffer from the following drawbacks: i) Existing methods mainly adopt cross-view contrastive learning forcing the representations of each sample across views to be exactly the same, which might ignore view discrepancy and flexibility in representations; ii) Due to the absence of non-observed samples across multiple views, the obtained prototypes of clusters might be unaligned and biased, leading to incorrect fusion. To address the above issues, we propose a Cross-view Partial Sample and Prototype Alignment Network (CPSPAN) for Deep Incomplete Multi-view Clustering. Firstly, unlike existing contrastive-based methods, we adopt pair-observed data alignment as 'proxy supervised signals' to guide instance-to-instance correspondence construction among views. Then, regarding of the shifted prototypes in IMVC, we further propose a prototype alignment module to achieve incomplete distribution calibration across views. Extensive experimental results showcase the effectiveness of our proposed modules, attaining noteworthy performance improvements when compared to existing IMVC competitors on benchmark datasets. | ['En Zhu', 'Xinwang Liu', 'Zhibin Dong', 'Siwei Wang', 'Jiaqi Jin'] | 2023-03-28 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Jin_Deep_Incomplete_Multi-View_Clustering_With_Cross-View_Partial_Sample_and_Prototype_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Jin_Deep_Incomplete_Multi-View_Clustering_With_Cross-View_Partial_Sample_and_Prototype_CVPR_2023_paper.pdf | cvpr-2023-1 | ['incomplete-multi-view-clustering'] | ['computer-vision'] | [ 3.38248909e-02 -1.67751729e-01 -2.26003736e-01 -5.43314338e-01
-8.04180682e-01 -6.64680243e-01 4.98843431e-01 -5.60552068e-02
1.44100636e-01 3.32757682e-01 1.94159687e-01 3.71897042e-01
-2.40018502e-01 -3.83859366e-01 -7.40363479e-01 -9.06905472e-01
3.54857355e-01 5.11430979e-01 3.29271075e-03 1.79350749e-01
-1.78289283e-02 1.52099520e-01 -1.66234541e+00 4.30043727e-01
7.62699783e-01 9.20555890e-01 1.11475922e-01 2.06168041e-01
7.38497300e-04 3.23609889e-01 -5.32234728e-01 -3.84363145e-01
3.86618823e-01 -3.21081936e-01 -2.86191076e-01 8.02856147e-01
4.04252231e-01 -1.21922478e-01 -8.96173529e-03 1.28492475e+00
5.02139866e-01 -2.38297716e-01 5.35621643e-01 -1.64933395e+00
-5.61198354e-01 5.20775378e-01 -1.03314757e+00 -2.06250772e-01
3.24182004e-01 -3.44562344e-02 8.86577427e-01 -1.00893164e+00
4.78295892e-01 1.22901571e+00 6.43226802e-01 3.77106190e-01
-1.33763576e+00 -6.56259418e-01 7.06855834e-01 1.73261419e-01
-1.53695571e+00 -6.08126342e-01 1.11323965e+00 -3.73046756e-01
1.68213248e-01 1.86713278e-01 6.67079270e-01 1.23531973e+00
-2.82404304e-01 8.07165444e-01 1.19655979e+00 1.84848160e-02
3.41706425e-01 4.57929134e-01 -3.79558690e-02 2.45824963e-01
4.00475472e-01 -7.28391111e-02 -3.90338600e-01 -1.83299050e-01
4.76628453e-01 3.81764978e-01 -5.45332789e-01 -1.11797309e+00
-1.35614014e+00 4.93373305e-01 4.10750329e-01 2.45654896e-01
-3.54866296e-01 -4.48871493e-01 4.60804135e-01 1.64333671e-01
2.56225020e-01 -1.06232159e-01 -3.52554768e-01 3.63684028e-01
-8.76003802e-01 -4.63647321e-02 5.23915529e-01 1.20792401e+00
7.84762383e-01 1.10072298e-02 1.42123327e-01 7.89150715e-01
3.85711908e-01 5.43103576e-01 4.83017564e-01 -7.53978670e-01
1.02204907e+00 8.38728368e-01 9.94963795e-02 -1.30994296e+00
-3.05517048e-01 -5.55312216e-01 -1.38634372e+00 -6.54478893e-02
2.39905268e-01 1.01349279e-01 -6.00318491e-01 1.60797882e+00
4.61252958e-01 3.08998019e-01 3.18054140e-01 9.29697096e-01
8.30345750e-01 2.67623276e-01 -3.65209848e-01 -5.86230576e-01
9.85434711e-01 -6.35853887e-01 -5.88179231e-01 -4.87131812e-02
2.36031577e-01 -5.89493155e-01 8.13216746e-01 6.48956001e-01
-8.55434299e-01 -8.41985285e-01 -1.16694617e+00 5.77880085e-01
-2.65281945e-01 1.64764717e-01 2.62311935e-01 6.33858979e-01
-6.83827400e-01 2.14109376e-01 -7.17107952e-01 -1.44826353e-01
4.93666053e-01 1.83965340e-01 -6.31921053e-01 -4.39610183e-01
-5.29298365e-01 2.08273873e-01 4.60950106e-01 4.36604828e-01
-8.51678431e-01 -7.59710133e-01 -6.54480696e-01 -9.46484506e-02
5.19220829e-01 -6.85565174e-01 6.19343221e-01 -9.72720504e-01
-8.98393691e-01 5.71982801e-01 -2.53667474e-01 -3.13208550e-02
6.45628929e-01 -8.06641206e-02 -7.79069543e-01 1.15551934e-01
1.87848851e-01 5.14193773e-01 7.90834188e-01 -2.18060112e+00
-6.32770240e-01 -6.43027246e-01 -8.94744396e-02 4.23187494e-01
-1.89383492e-01 -6.55654728e-01 -7.89776325e-01 -4.01310056e-01
6.77734494e-01 -7.89889395e-01 -9.89442840e-02 -1.06880888e-01
-5.27483463e-01 8.03175271e-02 1.00862765e+00 -4.00670648e-01
9.73171234e-01 -2.34000063e+00 2.88076669e-01 1.81060508e-01
2.01383919e-01 1.31718338e-01 2.62367334e-02 3.57016593e-01
-2.60318220e-01 -1.55707702e-01 -3.20039690e-01 -4.97687221e-01
-8.56118575e-02 3.72212648e-01 -2.49098465e-01 7.19484031e-01
-7.41103292e-02 5.33203006e-01 -8.63268554e-01 -5.23949504e-01
4.94512349e-01 3.40259522e-01 -3.54155451e-01 2.08289549e-01
3.70440371e-02 6.89271092e-01 -2.21001223e-01 7.34796882e-01
1.14750338e+00 -4.11161602e-01 4.11490977e-01 -7.86557615e-01
2.78487265e-01 -3.50233704e-01 -1.72118223e+00 1.78439820e+00
-2.54600912e-01 -9.32302400e-02 2.08031937e-01 -1.36187518e+00
6.90904260e-01 4.57460672e-01 7.87194729e-01 -4.71453190e-01
-1.22826025e-01 7.96885639e-02 -1.87238380e-01 -4.18620288e-01
1.82613909e-01 -2.02876199e-02 5.31561673e-02 2.23821297e-01
3.36116576e-03 1.25344411e-01 -6.61759898e-02 1.61096185e-01
3.76894325e-01 1.93691060e-01 1.54694661e-01 1.09799393e-01
7.17271507e-01 -2.99107939e-01 9.04580176e-01 5.47269523e-01
-2.32677892e-01 1.12563014e+00 2.99408674e-01 -3.32684100e-01
-9.28471088e-01 -1.27110374e+00 -2.05299780e-02 4.58463907e-01
4.54874754e-01 -3.11653495e-01 -5.11078894e-01 -8.04909587e-01
-1.49206951e-01 3.43398511e-01 -4.64846343e-01 4.67969067e-02
-3.28644305e-01 -9.76026773e-01 1.53388917e-01 5.89648128e-01
6.34552419e-01 -4.83193338e-01 -3.78309190e-01 2.18957841e-01
-5.39102495e-01 -1.31402457e+00 -3.90765905e-01 -1.55391404e-02
-8.51243019e-01 -1.51142788e+00 -5.93833506e-01 -4.45057988e-01
9.02227759e-01 8.88500869e-01 1.04976070e+00 -9.95723233e-02
1.08635023e-01 6.87400222e-01 -3.96803230e-01 -1.21682808e-01
-2.47540742e-01 -3.05065721e-01 2.63541192e-01 4.59825516e-01
3.11249644e-01 -8.21733117e-01 -7.25714862e-01 4.19651031e-01
-9.74637926e-01 -8.34873971e-03 4.76558834e-01 8.11586380e-01
8.73025417e-01 1.95401236e-01 6.78217828e-01 -9.47466075e-01
5.51380068e-02 -5.47317922e-01 -5.54938495e-01 5.75260162e-01
-7.07445145e-01 -3.98271561e-01 7.98982680e-01 -2.66517043e-01
-8.53480279e-01 1.46243259e-01 2.93969095e-01 -1.03773475e+00
-4.06378567e-01 2.57084519e-01 -8.00193250e-01 3.51073682e-01
1.25444353e-01 4.39658284e-01 1.21611208e-01 -3.82294327e-01
4.30581361e-01 5.30975223e-01 6.29637778e-01 -4.15965319e-01
8.90228510e-01 7.91679382e-01 -1.19133689e-01 -5.41786373e-01
-7.54372180e-01 -6.18735135e-01 -8.44473362e-01 -2.08731413e-01
8.07472408e-01 -1.54701197e+00 -5.83589017e-01 4.07678008e-01
-7.19295979e-01 4.89724338e-01 -2.12079689e-01 5.13624549e-01
-4.48564351e-01 8.41164947e-01 -1.15105927e-01 -6.73077643e-01
-1.47022992e-01 -1.19411182e+00 1.11023176e+00 1.36276960e-01
1.88932091e-01 -8.42198193e-01 -2.64720708e-01 5.11859000e-01
-1.06105790e-01 3.25313121e-01 6.67680681e-01 -6.63302839e-01
-5.31783104e-01 -1.46059379e-01 -1.18844241e-01 5.28405726e-01
4.98051941e-01 -3.05953119e-02 -1.11332798e+00 -7.04613984e-01
2.58795738e-01 -2.17372417e-01 4.85504597e-01 4.60634619e-01
1.14639175e+00 -2.24363968e-01 -5.60514987e-01 6.58203185e-01
1.56724226e+00 6.47797063e-02 2.81565726e-01 -1.40836695e-02
9.57869053e-01 7.55228758e-01 7.38165975e-01 6.95148945e-01
7.44011819e-01 5.89055479e-01 7.09105968e-01 -5.15262634e-02
2.02768236e-01 -2.17127413e-01 1.96903586e-01 1.25816214e+00
1.78560078e-01 -1.98171392e-01 -4.25564885e-01 7.71567881e-01
-1.90710342e+00 -9.93585110e-01 -2.60716259e-01 2.36450863e+00
6.78836554e-02 4.16524149e-02 2.87237823e-01 2.37233028e-01
9.91770029e-01 2.95409739e-01 -7.19010472e-01 3.91741067e-01
-3.47800016e-01 -5.25105000e-01 1.41844541e-01 -4.65576649e-02
-1.12151074e+00 2.68469155e-01 5.00394773e+00 5.06313324e-01
-8.97746861e-01 2.29333296e-01 6.02928102e-01 -1.98620379e-01
-4.44872826e-01 2.67749410e-02 -6.81628406e-01 6.28326893e-01
2.95825481e-01 1.34815410e-01 3.16029191e-01 7.07801342e-01
2.22711172e-02 5.15674464e-02 -1.21851790e+00 1.43999112e+00
3.41551512e-01 -1.00763917e+00 1.23384640e-01 1.38098300e-01
9.40649986e-01 -1.00946106e-01 9.12782028e-02 1.07965849e-01
1.85069010e-01 -4.46390092e-01 6.17227376e-01 4.91295636e-01
5.63925803e-01 -1.01245904e+00 6.55281842e-01 5.56206107e-01
-1.36877131e+00 -2.21002579e-01 -5.45009971e-01 4.03396755e-01
1.28186584e-01 8.51943314e-01 -5.32078326e-01 1.39585125e+00
8.05393636e-01 9.87465203e-01 -7.04382241e-01 7.95541883e-01
2.12042779e-01 2.57084757e-01 -2.42387503e-01 6.75045192e-01
8.23394302e-03 -4.26982403e-01 3.87919307e-01 5.62237024e-01
3.93131196e-01 -1.55157849e-01 3.90294403e-01 7.04614222e-01
1.34697154e-01 -1.96150929e-01 -7.51229405e-01 4.58353579e-01
7.74069369e-01 1.21227479e+00 -6.31491423e-01 -4.96468246e-01
-7.46282160e-01 8.98878872e-01 7.21761733e-02 4.42459941e-01
-8.50083470e-01 3.37920159e-01 4.92541671e-01 -9.77965221e-02
6.21968687e-01 7.70290643e-02 -2.01237455e-01 -1.57560933e+00
4.73336607e-01 -1.06225228e+00 5.13786316e-01 -6.07180834e-01
-1.73220074e+00 3.69769931e-01 1.34668946e-01 -1.93483126e+00
-1.69406295e-01 -3.24983075e-02 -4.56400156e-01 4.01831210e-01
-1.12320137e+00 -1.40078366e+00 -4.18946743e-01 7.72414386e-01
5.31778038e-01 -3.54801625e-01 5.84714949e-01 4.34281081e-01
-5.85588217e-01 6.61114991e-01 2.14882031e-01 -3.20646130e-02
7.92373717e-01 -1.09910619e+00 5.12641929e-02 8.86638165e-01
2.54794508e-01 4.78776217e-01 4.45008993e-01 -4.85782623e-01
-1.50311446e+00 -1.27373528e+00 9.37666446e-02 -4.74937588e-01
2.43354768e-01 -4.19889748e-01 -1.11845708e+00 7.23371565e-01
1.89753413e-01 3.08671683e-01 9.11662459e-01 -3.37983784e-03
-4.01496172e-01 -5.77880681e-01 -1.07040858e+00 3.22282493e-01
9.62236166e-01 -3.97050440e-01 -4.06450033e-01 2.01786518e-01
4.29112077e-01 -1.78687721e-01 -1.19223380e+00 7.44383633e-01
4.12510931e-01 -1.49855208e+00 1.10469162e+00 -2.08893776e-01
1.67472020e-01 -7.61465728e-01 -6.01913154e-01 -1.38154209e+00
-3.03070188e-01 8.26856345e-02 -1.11442424e-01 1.92362678e+00
-2.96620242e-02 -5.55853248e-01 7.05771685e-01 3.15124810e-01
-1.57733962e-01 -5.17149031e-01 -9.82240856e-01 -6.66751862e-01
-1.78923041e-01 -2.95555443e-01 9.80980873e-01 1.21913028e+00
-3.48402560e-01 3.16187620e-01 -5.45233965e-01 8.15563023e-01
8.55701864e-01 5.81808090e-01 9.94447410e-01 -1.26553857e+00
-5.01087844e-01 -2.10333228e-01 -2.99590856e-01 -8.67405713e-01
-9.74327177e-02 -7.03200340e-01 -8.89432803e-02 -1.41918397e+00
4.73724604e-01 -4.52415168e-01 -3.65556091e-01 -4.02775779e-02
-2.30467111e-01 1.51399240e-01 3.17273736e-01 4.51505005e-01
-8.30753744e-01 6.69976115e-01 1.04590058e+00 -2.23809674e-01
-2.01048721e-02 1.67982459e-01 -8.31687868e-01 6.67476118e-01
4.21213299e-01 -2.49725565e-01 -7.38056242e-01 -3.13489288e-01
-2.50927918e-03 3.52924049e-01 4.95854557e-01 -1.24844813e+00
2.26685807e-01 6.72274232e-02 7.03785002e-01 -1.22782850e+00
3.93666446e-01 -1.30148077e+00 5.60629427e-01 1.87754110e-01
1.87534124e-01 1.54439092e-01 -2.10119143e-01 1.03967535e+00
-5.08212388e-01 1.07523531e-01 6.59429789e-01 -3.11676621e-01
-5.42818606e-01 2.48485118e-01 1.43723041e-01 -1.40361357e-02
1.06349087e+00 -4.57717448e-01 -2.81402141e-01 -3.09352607e-01
-6.39771342e-01 4.15283382e-01 7.23698020e-01 5.38629711e-01
5.74076891e-01 -1.51229978e+00 -5.45864940e-01 3.82367760e-01
4.48405325e-01 5.72000384e-01 7.05585957e-01 1.02038002e+00
1.34442583e-01 1.73905164e-01 1.95423141e-02 -1.18022180e+00
-1.10059881e+00 1.11490810e+00 7.48916417e-02 -7.91530833e-02
-6.51853919e-01 3.32466125e-01 5.28482616e-01 -7.24549592e-01
2.24703372e-01 -7.43914247e-02 -1.92145795e-01 3.56632799e-01
2.99523622e-01 2.77348101e-01 7.21496046e-02 -7.55129933e-01
-3.51996988e-01 8.14302146e-01 -6.12068027e-02 2.29092702e-01
1.37839973e+00 -6.05997801e-01 5.09318151e-02 7.64647841e-01
1.19125772e+00 -1.65013596e-01 -1.43581939e+00 -3.68668914e-01
-3.51335764e-01 -4.15456235e-01 -5.40515125e-01 -3.78078997e-01
-1.36548543e+00 8.39383483e-01 7.53616989e-01 6.73357844e-02
1.29501617e+00 -1.20972775e-01 3.95774305e-01 1.30167767e-01
5.09590447e-01 -1.07768321e+00 1.21389292e-01 -1.62119538e-01
5.99726915e-01 -1.56419921e+00 2.60961562e-01 -4.40244406e-01
-8.57727528e-01 8.59912217e-01 7.85933375e-01 1.17711276e-01
6.45875394e-01 -1.12029597e-01 1.61609709e-01 -2.66949534e-01
-6.37129068e-01 1.41711816e-01 6.62783235e-02 7.36389220e-01
-2.60801781e-02 -2.90588848e-02 1.97327062e-01 4.38885033e-01
1.50841013e-01 -2.34410405e-01 5.10700822e-01 7.00734317e-01
8.49126931e-03 -8.52043152e-01 -6.12703979e-01 3.96390349e-01
-1.20233558e-01 4.35144067e-01 -1.55485928e-01 8.99245083e-01
2.56669790e-01 1.11479950e+00 3.88142988e-02 -4.93124217e-01
3.75741541e-01 -3.81529480e-02 3.91076565e-01 -4.06086534e-01
-2.80958444e-01 5.08432150e-01 -3.37745190e-01 -3.47323328e-01
-9.83974934e-01 -8.33612800e-01 -8.10174346e-01 2.83638900e-03
-4.06212807e-01 3.97933759e-02 3.05878371e-01 8.50779951e-01
4.80681807e-01 4.98954505e-01 1.08648872e+00 -8.60548019e-01
-5.83627820e-01 -7.50723124e-01 -7.30087221e-01 7.57494390e-01
2.88257331e-01 -7.28133321e-01 -3.67296576e-01 9.96314138e-02] | [8.413588523864746, 4.536715984344482] |
ae7495e8-0872-4716-b1b5-4ed474148c12 | one-shot-object-affordance-detection-in-the | 2108.03658 | null | https://arxiv.org/abs/2108.03658v1 | https://arxiv.org/pdf/2108.03658v1.pdf | One-Shot Object Affordance Detection in the Wild | Affordance detection refers to identifying the potential action possibilities of objects in an image, which is a crucial ability for robot perception and manipulation. To empower robots with this ability in unseen scenarios, we first study the challenging one-shot affordance detection problem in this paper, i.e., given a support image that depicts the action purpose, all objects in a scene with the common affordance should be detected. To this end, we devise a One-Shot Affordance Detection Network (OSAD-Net) that firstly estimates the human action purpose and then transfers it to help detect the common affordance from all candidate images. Through collaboration learning, OSAD-Net can capture the common characteristics between objects having the same underlying affordance and learn a good adaptation capability for perceiving unseen affordances. Besides, we build a large-scale Purpose-driven Affordance Dataset v2 (PADv2) by collecting and labeling 30k images from 39 affordance and 103 object categories. With complex scenes and rich annotations, our PADv2 dataset can be used as a test bed to benchmark affordance detection methods and may also facilitate downstream vision tasks, such as scene understanding, action recognition, and robot manipulation. Specifically, we conducted comprehensive experiments on PADv2 dataset by including 11 advanced models from several related research fields. Experimental results demonstrate the superiority of our model over previous representative ones in terms of both objective metrics and visual quality. The benchmark suite is available at https://github.com/lhc1224/OSAD Net. | ['DaCheng Tao', 'Yang Cao', 'Jing Zhang', 'Hongchen Luo', 'Wei Zhai'] | 2021-08-08 | null | null | null | null | ['affordance-detection'] | ['computer-vision'] | [ 9.55578387e-02 -1.55317858e-01 -1.79748893e-01 -3.07042271e-01
-7.93122202e-02 -2.81000793e-01 3.65072846e-01 -2.13118196e-01
-3.25916290e-01 2.30746865e-01 3.96059543e-01 6.87961206e-02
-2.85183191e-01 -3.56087863e-01 -7.51196504e-01 -4.43795770e-01
-5.78447580e-02 2.57043332e-01 3.77194107e-01 -4.25125897e-01
2.05413133e-01 4.47716773e-01 -1.65514207e+00 4.80771847e-02
9.38062072e-01 1.05683327e+00 8.32503796e-01 3.54050815e-01
4.39119279e-01 7.90483892e-01 -3.07963580e-01 2.18409449e-01
2.79006362e-01 -2.57403571e-02 -7.65393496e-01 3.78548741e-01
3.97975624e-01 -9.36618507e-01 -6.38972878e-01 1.08258390e+00
3.02469641e-01 5.31872571e-01 5.72060645e-01 -1.65330160e+00
-9.73966777e-01 4.60683703e-01 -2.39627451e-01 3.49621207e-01
6.22255206e-01 8.56419146e-01 1.18991327e+00 -1.16366863e+00
5.94569981e-01 1.69423759e+00 -1.86984062e-01 7.66563356e-01
-1.04082656e+00 -5.14485717e-01 4.78752047e-01 4.19150442e-01
-1.10924065e+00 -4.48999196e-01 6.99149787e-01 -3.50255191e-01
9.91103828e-01 2.90461391e-01 7.59271145e-01 1.20507562e+00
-1.77927330e-01 1.45722890e+00 7.67662346e-01 -7.43663684e-02
-5.31991236e-02 -4.49149758e-01 2.15161636e-01 8.95955682e-01
1.82034880e-01 1.05580129e-01 -4.02620852e-01 1.75932139e-01
1.13314891e+00 4.79294211e-01 -5.01441300e-01 -6.62714660e-01
-2.06187654e+00 3.24465275e-01 1.18604851e+00 1.59388661e-01
-2.94308841e-01 3.00265312e-01 2.40318194e-01 3.39313775e-01
-2.19506621e-01 7.27058470e-01 -3.89336616e-01 1.60089433e-01
2.76789546e-01 3.88184875e-01 4.59738612e-01 1.47014034e+00
8.01082969e-01 -3.10297787e-01 -5.30483782e-01 6.95464015e-01
4.71536756e-01 3.49169195e-01 3.91023725e-01 -1.10635006e+00
6.93489432e-01 1.01288390e+00 4.09007519e-01 -9.37639475e-01
-5.17649472e-01 2.36876711e-01 -6.34792566e-01 2.10553557e-02
4.32109386e-01 2.18563393e-01 -1.04650533e+00 1.57178545e+00
5.63933492e-01 3.60872597e-02 8.60995576e-02 1.64568806e+00
1.19588125e+00 4.97738481e-01 4.01260611e-03 2.77535617e-01
1.61380184e+00 -1.39856827e+00 -5.12212932e-01 -3.32260728e-01
7.30605781e-01 -3.94552499e-01 1.77038538e+00 1.67623580e-01
-3.67135882e-01 -7.37058818e-01 -1.10528445e+00 -5.69980264e-01
-2.03770250e-01 5.35429299e-01 1.03812850e+00 -1.07727259e-01
-4.59326565e-01 3.77400309e-01 -8.56511235e-01 -4.44396645e-01
7.71660209e-01 1.66536421e-01 -5.38000584e-01 -4.66004789e-01
-7.84962952e-01 6.16202116e-01 7.74619699e-01 2.60653406e-01
-1.60391641e+00 -2.49112174e-01 -1.20617759e+00 6.45709410e-03
1.08648860e+00 -7.06139505e-01 1.23178816e+00 -5.72000086e-01
-1.10366452e+00 5.68790913e-01 1.05999917e-01 -6.88215718e-02
4.66737092e-01 -5.90406656e-01 -2.03983158e-01 -6.47519082e-02
4.47633207e-01 9.72387433e-01 8.17173243e-01 -1.19294500e+00
-5.68372667e-01 -3.42757314e-01 6.01206779e-01 4.62709159e-01
-1.62906110e-01 -9.66783613e-02 -4.52144653e-01 -4.86821473e-01
5.92257917e-01 -7.85952389e-01 -1.77367121e-01 5.03233969e-01
-8.78885865e-01 -8.24345946e-01 7.60501862e-01 -2.66396731e-01
5.74886203e-01 -2.32073474e+00 4.27985370e-01 -3.11074436e-01
4.97201622e-01 -8.82645100e-02 -3.31367105e-01 2.54088759e-01
2.18767926e-01 -1.91050515e-01 -1.20277405e-01 -1.02025442e-01
2.20663369e-01 4.36546952e-01 -3.35106462e-01 4.57735538e-01
5.19804060e-01 1.13954890e+00 -1.28756106e+00 -4.52063143e-01
4.60239083e-01 -3.58275436e-02 -4.73477900e-01 5.70869982e-01
-5.07397711e-01 7.08528161e-01 -8.69093418e-01 1.14246881e+00
2.45569393e-01 -2.66525090e-01 -1.12785123e-01 -3.37725878e-01
1.97782695e-01 1.23333126e-01 -1.16935730e+00 2.13780880e+00
-1.43965513e-01 3.53752077e-01 -2.93692499e-01 -6.62722766e-01
8.71511459e-01 -1.15144379e-01 1.44788772e-01 -4.88717645e-01
2.24967659e-01 3.03754777e-01 4.04903084e-01 -1.02222550e+00
3.90797943e-01 4.51643139e-01 -3.11616838e-01 2.14613646e-01
4.18693244e-01 -1.75287902e-01 2.34797776e-01 7.63813257e-02
8.85787845e-01 4.48163569e-01 4.29090172e-01 -2.50809103e-01
3.64959985e-01 -1.69785377e-02 5.23519278e-01 6.77737951e-01
-6.71429574e-01 3.87719989e-01 4.07635510e-01 -6.88595593e-01
-9.01546896e-01 -9.55317616e-01 1.41598657e-01 1.28114021e+00
8.63641262e-01 -1.75352573e-01 -2.54361987e-01 -5.70459127e-01
5.09473160e-02 5.11902869e-01 -7.32156277e-01 -2.39000961e-01
-4.63305384e-01 -2.65622914e-01 1.28684267e-01 6.21219456e-01
6.44099534e-01 -1.59944546e+00 -9.57386553e-01 -1.45510897e-01
-2.93794960e-01 -1.20008826e+00 -5.63311458e-01 9.88189410e-03
-6.19654000e-01 -1.21682239e+00 -6.20257199e-01 -9.72318053e-01
7.02952087e-01 8.21129382e-01 8.23155701e-01 3.01267028e-01
-4.63112503e-01 2.48439133e-01 -5.42410314e-01 -4.54412282e-01
-6.32764120e-03 -6.67670965e-02 4.00978386e-01 8.31612721e-02
3.83986205e-01 -2.66065210e-01 -7.88596988e-01 5.23880005e-01
-7.32870579e-01 1.96160212e-01 7.46952891e-01 8.55473220e-01
7.11609542e-01 -3.43018413e-01 3.54477912e-01 -4.22465988e-02
5.13263166e-01 -4.06725556e-01 -4.62811470e-01 1.79976478e-01
-1.43498942e-01 -2.50435155e-02 2.20903918e-01 -4.91615623e-01
-7.88092911e-01 1.33115798e-01 2.79373169e-01 -6.69049561e-01
-4.89476353e-01 1.89669609e-01 -6.33115947e-01 1.80355608e-01
6.13154590e-01 1.63159683e-01 -1.76822990e-02 -5.85965991e-01
5.02031088e-01 5.41350543e-01 5.04444718e-01 -5.84195077e-01
5.83966911e-01 5.11825085e-01 -1.74663395e-01 -6.20152652e-01
-1.04575479e+00 -6.47312105e-01 -9.55927730e-01 -2.23893479e-01
1.04829764e+00 -1.04756439e+00 -8.52377236e-01 5.22138536e-01
-1.18614507e+00 -4.82685268e-01 -1.13823548e-01 5.83785892e-01
-7.67247617e-01 2.08543107e-01 -2.69805670e-01 -6.28841758e-01
-4.17595692e-02 -1.39402139e+00 1.28879690e+00 3.04392636e-01
-2.45039295e-02 -2.73058861e-01 -4.85058576e-01 2.48217463e-01
-3.40979844e-01 3.24299455e-01 6.85685217e-01 -5.43356359e-01
-1.02139723e+00 4.99792583e-02 -6.53057218e-01 1.33473709e-01
4.46842074e-01 -4.27702099e-01 -6.57302976e-01 -1.97490916e-01
-1.94660932e-01 -9.49492455e-01 1.09438348e+00 -1.73918940e-02
1.24532390e+00 -1.94122046e-01 -4.16802466e-01 5.87775528e-01
1.17064619e+00 -1.18634656e-01 3.87251347e-01 3.03632200e-01
1.30691743e+00 7.20364332e-01 1.28954232e+00 3.07971597e-01
4.90775466e-01 5.37303150e-01 1.06654167e+00 2.46427700e-01
4.52809269e-03 -4.45146501e-01 2.72371233e-01 3.25788558e-01
-2.94970274e-01 -1.57888025e-01 -8.54377031e-01 5.52204311e-01
-2.16027594e+00 -6.81208134e-01 -5.86662404e-02 1.67364585e+00
5.13767302e-01 7.20127523e-02 1.50149032e-01 -6.95221499e-02
4.70326036e-01 2.66035408e-01 -1.17259347e+00 2.69044608e-01
1.20538508e-03 -6.40051484e-01 9.03276429e-02 -9.78282541e-02
-1.36034393e+00 1.12057614e+00 4.76224613e+00 5.17784119e-01
-6.62996829e-01 4.74244244e-02 2.27905318e-01 3.37179787e-02
4.83965948e-02 -5.20161614e-02 -5.12155294e-01 1.52331069e-01
-1.35330394e-01 2.15781778e-01 5.84104717e-01 1.19640553e+00
1.21646702e-01 -1.09546743e-01 -1.43093538e+00 1.07218552e+00
1.56261116e-01 -7.75943458e-01 2.58884490e-01 -4.43396494e-02
4.86667901e-01 2.87628233e-01 -1.75889954e-01 3.90556455e-01
2.41475105e-01 -1.00483751e+00 8.65709662e-01 5.58239758e-01
6.90452933e-01 -2.51030833e-01 3.91784728e-01 6.78889275e-01
-1.26387787e+00 -5.81972480e-01 -7.92767107e-01 -3.41336310e-01
-4.18994389e-03 -3.08712963e-02 -6.68344796e-01 5.44830441e-01
8.57516706e-01 1.26835608e+00 -5.05582869e-01 9.76612628e-01
-5.57284951e-01 -1.32255390e-01 -2.74890393e-01 -3.41445893e-01
3.81909966e-01 -5.74852899e-02 7.70030499e-01 3.99014741e-01
2.64790028e-01 2.74073124e-01 5.53212821e-01 1.13267350e+00
-2.81971693e-02 3.56313661e-02 -5.93816876e-01 -3.90977859e-02
5.22218227e-01 1.19117010e+00 -7.59667337e-01 -3.58702779e-01
-3.59248996e-01 1.11307704e+00 6.07703507e-01 3.37246150e-01
-6.99590862e-01 -2.80156463e-01 8.99049878e-01 -3.47093135e-01
2.72325128e-01 -4.82985526e-01 1.40886739e-01 -1.46489406e+00
2.97590643e-01 -8.37063134e-01 2.92608470e-01 -1.16357613e+00
-1.14467025e+00 4.37517494e-01 8.33788738e-02 -1.67369246e+00
3.59243989e-01 -1.11394393e+00 -4.59392875e-01 5.70665061e-01
-1.48075891e+00 -1.35159588e+00 -8.82351518e-01 7.03523576e-01
1.03364360e+00 -1.04201302e-01 5.78680038e-01 -2.62585759e-01
-5.49624860e-01 1.02753974e-01 -5.04107356e-01 2.81188577e-01
4.48395014e-01 -1.14678741e+00 4.71168518e-01 7.02768981e-01
2.78772295e-01 6.89964175e-01 4.33641315e-01 -5.78776658e-01
-1.85884047e+00 -1.16900122e+00 2.48209789e-01 -8.79155457e-01
7.00465202e-01 -4.79327291e-01 -7.68811762e-01 7.06850767e-01
-2.21064866e-01 3.97238433e-01 1.34546503e-01 -8.93299803e-02
-3.47666323e-01 1.77292094e-01 -7.25249648e-01 8.26763153e-01
1.91842628e+00 -4.67069775e-01 -9.88753676e-01 5.71738720e-01
1.28194427e+00 -5.84892988e-01 -7.49949157e-01 4.51793075e-01
6.71839595e-01 -9.01850522e-01 9.46660042e-01 -8.42682064e-01
4.91560370e-01 -4.84579980e-01 -3.87730777e-01 -1.11885679e+00
-3.67262423e-01 -3.07040453e-01 -2.67666578e-01 8.36371362e-01
7.38314837e-02 -4.40192252e-01 1.81552351e-01 3.51797074e-01
-4.64006543e-01 -7.52050161e-01 -7.22255886e-01 -9.02024388e-01
-5.43221772e-01 -4.89785761e-01 1.00350392e+00 7.03023553e-01
-1.60236046e-01 4.46284503e-01 -3.27947199e-01 1.92478642e-01
4.47359800e-01 3.98165941e-01 1.18184304e+00 -1.21970761e+00
-7.03167990e-02 -3.56654406e-01 -5.80696523e-01 -1.67291880e+00
1.18239976e-01 -9.34666514e-01 5.12390018e-01 -1.63759243e+00
2.57097751e-01 -4.97390032e-01 -4.77762878e-01 7.81094134e-01
-4.34243262e-01 1.62036307e-02 3.24899465e-01 6.35179937e-01
-8.33311021e-01 9.85114515e-01 2.08013916e+00 -4.30941254e-01
-3.12297791e-01 -3.40713561e-01 -6.16564393e-01 7.05330133e-01
5.75583100e-01 -8.96489248e-02 -5.10890543e-01 -6.70789480e-01
-1.95999518e-02 -2.45616525e-01 9.76388216e-01 -9.71927106e-01
-2.07773373e-01 -4.79605615e-01 2.69647837e-01 -5.34649968e-01
3.96687001e-01 -8.84515464e-01 -5.53991258e-01 6.42852366e-01
-2.87227452e-01 -2.52105325e-01 -2.19484612e-01 9.90879059e-01
7.94817507e-02 3.49528641e-02 1.94609165e-01 -3.23815942e-01
-1.62003219e+00 7.51796007e-01 1.81109816e-01 -1.71842113e-01
1.15129435e+00 -1.33391052e-01 -3.50168020e-01 5.30678630e-02
-6.04944825e-01 7.04570591e-01 5.16286552e-01 9.90015090e-01
1.04943144e+00 -1.42661905e+00 -5.51985443e-01 3.45543891e-01
9.88825977e-01 5.06523252e-01 2.60157108e-01 8.50527644e-01
-3.46770376e-01 1.68488979e-01 -4.89204019e-01 -8.64500582e-01
-8.49630952e-01 9.62783337e-01 2.07528129e-01 4.85417366e-01
-9.23545480e-01 1.11985803e+00 5.55037439e-01 -7.06964493e-01
2.99166143e-01 -9.23427224e-01 -4.26253051e-01 -3.78337204e-01
4.69138622e-01 2.16658086e-01 -6.29255474e-01 -8.98703694e-01
-4.32024390e-01 3.11523139e-01 1.27562121e-01 4.13106829e-01
1.14579058e+00 -2.91990548e-01 -2.74433196e-01 6.53084934e-01
9.61848378e-01 -8.46983314e-01 -1.77541864e+00 -3.65442812e-01
3.16243209e-02 -7.06379712e-01 -2.86049485e-01 -4.00206923e-01
-7.13853419e-01 8.47832739e-01 3.87651831e-01 -3.14026773e-02
8.89412940e-01 5.35893679e-01 5.43519616e-01 1.09066248e+00
6.23238266e-01 -8.73704374e-01 6.92890465e-01 5.15473187e-01
1.42774725e+00 -1.80501819e+00 -1.64463848e-01 -5.45396566e-01
-7.35845387e-01 1.02105474e+00 1.17333055e+00 -4.61266398e-01
4.01845604e-01 -5.51790595e-01 -1.58238143e-01 -5.00545621e-01
-5.13482749e-01 -5.77532232e-01 4.97806221e-01 7.25201428e-01
-1.47254914e-01 3.07785362e-01 5.85740432e-02 6.26520991e-01
-9.49932411e-02 -1.20871283e-01 2.98459351e-01 8.43697309e-01
-5.95175564e-01 -3.50720495e-01 -3.85198034e-02 5.97303927e-01
3.10514510e-01 6.99562877e-02 -4.33335721e-01 5.77909231e-01
1.96475089e-01 9.15334344e-01 -1.19636588e-01 -4.95849162e-01
5.11146426e-01 -2.82687157e-01 5.48155844e-01 -8.95366132e-01
1.55715242e-01 -1.31476924e-01 2.57853307e-02 -9.48421896e-01
-6.25965595e-01 -6.95235133e-01 -1.36448121e+00 3.69298816e-01
-4.59758699e-01 -5.22982299e-01 2.28420720e-01 1.12700546e+00
1.81193933e-01 7.57341027e-01 4.33692992e-01 -1.22827911e+00
-6.93030298e-01 -1.27359784e+00 -5.57553530e-01 4.98847723e-01
5.52192867e-01 -1.26340699e+00 -7.29199052e-02 -1.06521167e-01] | [5.149901866912842, -0.0975063145160675] |
49645a75-75ea-403d-8a71-1e58a91fd052 | battle-royale-optimization-algorithm | null | null | https://link.springer.com/article/10.1007/s00521-020-05004-4 | https://link.springer.com/article/10.1007/s00521-020-05004-4 | Battle royale optimization algorithm | Recently, several metaheuristic optimization approaches have been developed for solving many complex problems in various areas. Most of these optimization algorithms are inspired by nature or the social behavior of some animals. However, there is no optimization algorithm which has been inspired by a game. In this paper, a novel metaheuristic optimization algorithm, named BRO (battle royale optimization), is proposed. The proposed method is inspired by a genre of digital games knowns as “battle royale.” BRO is a population-based algorithm in which each individual is represented by a soldier/player that would like to move toward the safest (best) place and ultimately survive. The proposed scheme has been compared with the well-known PSO algorithm and six recent proposed optimization algorithms on nineteen benchmark optimization functions. Moreover, to evaluate the performance of the proposed algorithm on real-world engineering problems, the inverse kinematics problem of the 6-DOF PUMA 560 robot arm is considered. The experimental results show that, according to both convergence and accuracy, the proposed algorithm is an efficient method and provides promising and competitive results. | ['Taymaz Rahkar-Farshi'] | 2020-06-02 | null | null | null | null | ['metaheuristic-optimization'] | ['methodology'] | [ 1.84976935e-01 -2.40393341e-01 1.23315662e-01 1.48195028e-01
2.32575104e-01 -2.22522959e-01 3.80262673e-01 3.02309424e-01
-7.15874910e-01 1.14197695e+00 -4.31588024e-01 8.15305263e-02
-8.82941604e-01 -8.74720156e-01 -2.89312243e-01 -1.01197410e+00
1.49310986e-02 5.94489753e-01 4.48691696e-02 -7.01317191e-01
8.07917774e-01 5.55509925e-01 -1.73577356e+00 -6.23802423e-01
1.17435813e+00 7.45297432e-01 5.47134101e-01 4.59899008e-01
2.00694606e-01 3.95111293e-01 -8.76004219e-01 -1.02340765e-01
2.21222058e-01 -6.56528413e-01 -5.17253697e-01 -1.19308837e-01
-8.17788482e-01 3.38565499e-01 1.52346924e-01 1.08224177e+00
5.73995173e-01 7.55158186e-01 7.05634177e-01 -1.29916763e+00
-2.63935268e-01 4.49968964e-01 -6.34799778e-01 3.03864121e-01
2.01732263e-01 -1.05047435e-01 6.59112215e-01 -3.05847436e-01
5.83427072e-01 1.03330731e+00 3.52348775e-01 3.63858193e-01
-6.14913583e-01 -4.95539486e-01 -3.47515821e-01 4.51645732e-01
-1.38233066e+00 2.65817016e-01 8.36296380e-01 -8.46528187e-02
7.63446152e-01 5.30246735e-01 9.31077003e-01 3.56721669e-01
8.85232985e-01 3.82414937e-01 1.05981016e+00 -4.71814275e-01
6.05033159e-01 -1.58908844e-01 6.41715303e-02 5.46797454e-01
6.85433567e-01 4.20991778e-01 -1.23532154e-01 -2.84191072e-01
1.68995723e-01 -2.39927128e-01 -2.04664394e-01 -4.86252695e-01
-9.66612458e-01 1.01955819e+00 4.53114271e-01 5.49199164e-01
-7.81383872e-01 -1.75406873e-01 2.40612954e-01 -8.00425187e-02
-1.41025454e-01 8.54221284e-01 -1.62155539e-01 -4.14901793e-01
-3.54974270e-01 5.12529254e-01 1.02594757e+00 3.65842491e-01
3.13977599e-01 2.52186090e-01 3.39036286e-01 7.01331019e-01
4.89359319e-01 1.63389474e-01 6.40288532e-01 -5.67270994e-01
1.28884166e-01 7.44001687e-01 2.84903586e-01 -1.58837819e+00
-7.31142044e-01 -6.32776141e-01 -6.50344491e-01 7.25448370e-01
6.18834831e-02 -4.31728691e-01 -4.01004434e-01 1.23331249e+00
5.86424828e-01 -5.18218838e-02 4.70224380e-01 1.05951774e+00
6.67463005e-01 9.86282706e-01 -1.86337963e-01 -4.31194663e-01
1.07474196e+00 -1.09934974e+00 -8.23810339e-01 -1.40189171e-01
-8.01877603e-02 -9.21855628e-01 4.06008989e-01 7.34222412e-01
-8.32011521e-01 -3.65650326e-01 -1.47926426e+00 8.16683531e-01
-4.45077688e-01 -1.74255207e-01 6.39272630e-01 1.02052557e+00
-5.68327487e-01 7.19951630e-01 -6.64910316e-01 -6.56733394e-01
-1.10873461e-01 5.48692286e-01 -6.21805042e-02 1.89703077e-01
-9.99478340e-01 1.12181580e+00 9.42871511e-01 2.79332966e-01
-4.32878107e-01 4.84236553e-02 -4.43421304e-01 -4.65032309e-02
6.21993661e-01 -6.02987707e-01 6.99604809e-01 -8.38374257e-01
-1.88524699e+00 3.76799524e-01 1.53332680e-01 -3.51940334e-01
3.03503335e-01 1.72730014e-01 -2.16785252e-01 -4.56533618e-02
-2.02796549e-01 8.10903385e-02 5.02770782e-01 -1.22223055e+00
-7.27571666e-01 -4.00920182e-01 1.20947286e-01 5.12694597e-01
-3.33510518e-01 1.21348046e-01 6.13774061e-02 -7.79228806e-01
1.40672885e-02 -9.75867271e-01 -5.69581687e-01 -6.33275270e-01
-1.32064760e-01 -1.03892565e-01 7.70736277e-01 -3.38820428e-01
1.27748728e+00 -1.70388842e+00 7.30583012e-01 6.12176776e-01
-3.67413312e-01 3.27784330e-01 1.71067238e-01 6.84427321e-01
2.37256631e-01 -8.55218694e-02 -2.86419511e-01 3.36191356e-01
-6.15242459e-02 4.83654112e-01 5.23457944e-01 5.02614737e-01
-3.49010527e-01 2.23650649e-01 -8.75895739e-01 -3.33347976e-01
3.04053247e-01 1.91821337e-01 -5.07292986e-01 1.62582919e-01
1.38048753e-01 4.80393976e-01 -1.04087067e+00 7.09802032e-01
7.66848624e-01 2.86425054e-01 2.36102208e-01 -1.18255310e-01
-4.74117041e-01 -6.63885891e-01 -1.32702446e+00 1.31544614e+00
-1.21595755e-01 2.16851845e-01 2.41753742e-01 -1.28818500e+00
1.30792141e+00 8.40029791e-02 6.88130558e-01 -4.84274119e-01
9.15685833e-01 4.07409042e-01 4.18040663e-01 -5.66930473e-01
8.45509589e-01 1.44280652e-02 9.78912264e-02 4.17098016e-01
-3.08551818e-01 -4.27792400e-01 7.00907528e-01 -5.51238537e-01
9.34214711e-01 1.75521016e-01 7.07451940e-01 -4.39789861e-01
9.61232364e-01 4.86793846e-01 7.11631000e-01 5.78739285e-01
-2.80199945e-01 7.52171874e-02 -3.30147110e-02 -2.73329079e-01
-7.85336137e-01 -6.67509019e-01 5.28723672e-02 6.15916431e-01
7.38133252e-01 -6.40663579e-02 -6.77385628e-01 -6.63537905e-02
-1.07005470e-01 7.39917934e-01 -3.87266606e-01 -2.74299502e-01
-7.02119589e-01 -1.28629386e+00 1.05010062e-01 -2.37653717e-01
7.35116243e-01 -1.25993896e+00 -1.26406288e+00 7.15223610e-01
1.08465344e-01 -6.48004413e-01 2.52941519e-01 -8.54958370e-02
-9.29108381e-01 -1.23758507e+00 -6.89546049e-01 -9.00862336e-01
5.28900445e-01 8.85460377e-02 6.86857641e-01 4.25883502e-01
-5.05815685e-01 3.31031866e-02 -9.83220756e-01 -6.45647705e-01
-5.90206623e-01 1.38411731e-01 1.29981758e-02 -8.51185098e-02
-8.23356584e-03 -5.43962061e-01 -2.90619463e-01 5.27900338e-01
-6.92342758e-01 -3.94161552e-01 8.68050694e-01 9.51324582e-01
5.41266918e-01 1.03383970e+00 4.59501147e-01 -3.46347958e-01
1.07905805e+00 -4.83804375e-01 -8.09158623e-01 1.54416040e-01
-5.17850757e-01 7.20510408e-02 8.60644102e-01 -2.65414298e-01
-8.83124769e-01 -9.22302902e-02 -1.46795437e-01 2.08373349e-02
1.00356124e-01 7.60648131e-01 2.84293313e-02 -6.19566619e-01
4.08987015e-01 3.65705103e-01 3.69053453e-01 -3.99380922e-01
-1.99183717e-01 6.53753579e-01 4.10321385e-01 -6.05571151e-01
7.15880811e-01 1.58227831e-01 3.21221352e-01 -8.08865786e-01
-3.97760868e-02 -2.34258011e-01 2.22715568e-02 -5.20170152e-01
9.70077753e-01 2.78491806e-02 -1.23840380e+00 6.92621231e-01
-8.75095606e-01 3.86420012e-01 1.16107002e-01 5.86032033e-01
-6.41554475e-01 3.36535871e-01 2.20847083e-03 -1.20901465e+00
-4.52395231e-01 -1.34888530e+00 2.32549056e-01 1.00190675e+00
-3.71202268e-02 -6.36274040e-01 4.08480950e-02 1.70701638e-01
5.86097836e-01 1.01454747e+00 7.37668991e-01 -4.18170333e-01
-1.44087061e-01 -2.90364593e-01 5.14481008e-01 -1.98164769e-02
2.39934042e-01 1.82572246e-01 1.05661869e-01 -3.46653014e-01
4.32118744e-01 1.68816656e-01 -4.77286382e-03 3.78394336e-01
6.67222381e-01 -4.98513915e-02 -5.29529095e-01 2.71942079e-01
1.60365963e+00 1.27670467e+00 6.62464321e-01 1.15036714e+00
3.43580395e-02 4.43971932e-01 1.29566765e+00 8.01618457e-01
-1.51822818e-02 6.01071298e-01 9.93856788e-01 1.35103285e-01
7.06164062e-01 4.35658693e-01 7.12180883e-02 7.22810388e-01
-7.32536435e-01 -6.73238814e-01 -7.89621592e-01 1.15907915e-01
-1.92401230e+00 -1.04920959e+00 -1.02440193e-01 1.85073256e+00
9.03704986e-02 2.38320932e-01 9.02187452e-02 5.36073148e-01
1.00344419e+00 -2.82208603e-02 -3.79166245e-01 -8.15275431e-01
-1.77812353e-01 5.14286935e-01 5.80314755e-01 1.40652925e-01
-9.79185820e-01 5.05116343e-01 5.42574215e+00 8.51962566e-01
-9.58744287e-01 -8.32202658e-02 1.75299793e-01 1.22740418e-01
2.02108994e-01 -1.21102087e-01 -3.07586044e-01 5.10674477e-01
4.53594893e-01 -7.51430213e-01 7.47322679e-01 7.68458247e-01
4.12612557e-01 -5.55779755e-01 -2.19177052e-01 8.46016526e-01
1.32532686e-01 -1.06284225e+00 -1.57988042e-01 -9.42628905e-02
7.70382941e-01 -5.15123248e-01 2.68302556e-05 -3.26056182e-02
4.04713675e-02 -9.25299287e-01 6.75378263e-01 5.64721704e-01
-4.40635622e-01 -1.44637966e+00 1.16619945e+00 5.53838789e-01
-1.02613366e+00 -5.05847335e-01 -4.40143496e-01 -1.04402065e-01
4.16586727e-01 5.18983230e-03 -4.52852070e-01 1.24408007e+00
6.96136355e-01 4.19525355e-01 -2.59778231e-01 1.67372274e+00
9.96943936e-02 8.27739313e-02 -3.36490571e-01 -9.83132601e-01
5.60871303e-01 -7.87042439e-01 1.18487573e+00 3.43613446e-01
7.04465389e-01 3.56590718e-01 -4.44515236e-02 6.66988790e-01
6.51904166e-01 6.15165830e-01 -3.35352391e-01 -2.89145321e-01
3.52413923e-01 1.15795422e+00 -1.01909184e+00 1.35276467e-01
1.62769482e-01 6.42766953e-01 -3.55246335e-01 -8.61245021e-02
-1.15841722e+00 -9.14793134e-01 3.07658076e-01 -3.34555626e-01
2.63202906e-01 -1.95853665e-01 -4.41132449e-02 -3.42186928e-01
-3.91416937e-01 -9.47947323e-01 4.55735862e-01 -7.10592628e-01
-9.01092470e-01 8.61768126e-01 4.87449765e-02 -1.28880048e+00
-2.42920130e-01 -8.51871431e-01 -6.77878439e-01 6.01020098e-01
-9.63169873e-01 -4.87350225e-01 -4.13325310e-01 2.79096395e-01
5.97377419e-01 -7.09533155e-01 6.61764920e-01 8.67791697e-02
-6.50756419e-01 1.75658762e-01 5.82890213e-01 -5.84072173e-01
2.30059382e-02 -7.39487410e-01 -4.03509885e-01 6.80660784e-01
-5.71178615e-01 5.27464926e-01 1.25695705e+00 -5.75092614e-01
-1.64007902e+00 -3.26153189e-01 3.56683940e-01 3.71561468e-01
5.24119616e-01 3.73264432e-01 -2.93833375e-01 -7.28276223e-02
3.99652570e-01 -5.57930112e-01 3.84338170e-01 -4.24494624e-01
1.04096556e+00 -6.82812184e-02 -1.57644272e+00 6.08287513e-01
6.29763961e-01 6.90451205e-01 -6.61167085e-01 4.99301143e-02
3.28908265e-02 -7.05735862e-01 -8.57637465e-01 5.85696936e-01
5.53672671e-01 -1.11281025e+00 9.09893692e-01 -3.11141759e-01
1.22340418e-01 -5.54982722e-01 8.04938897e-02 -1.41014361e+00
-1.88389361e-01 -6.88656509e-01 2.40356937e-01 9.82269704e-01
1.66734844e-01 -1.00746214e+00 8.73001456e-01 1.11155927e-01
-1.38486266e-01 -9.02838469e-01 -1.05081153e+00 -1.05697310e+00
-2.57984966e-01 2.41323978e-01 9.64386106e-01 6.75527155e-01
-2.16882050e-01 6.36952445e-02 -1.83629289e-01 8.55726302e-02
5.78979135e-01 1.56452611e-01 6.85320556e-01 -1.21675956e+00
-5.56579292e-01 -5.96021235e-01 -7.01626003e-01 -2.50893325e-01
-3.05486899e-02 -5.40010571e-01 6.68913275e-02 -1.59406066e+00
-1.59705594e-01 -5.84798455e-01 -3.37344170e-01 -1.66275725e-01
-1.19339362e-01 6.05859496e-02 1.93514958e-01 -1.19921621e-02
-1.22806512e-01 6.17758691e-01 1.32136834e+00 -2.01675743e-01
-4.62692857e-01 2.03782231e-01 -5.22600293e-01 6.96898103e-01
1.03409028e+00 -5.90206504e-01 -2.21405447e-01 1.66771747e-02
1.20182440e-01 3.02068084e-01 -1.15873046e-01 -1.40093768e+00
1.64081737e-01 -4.99420106e-01 -8.10497478e-02 -5.78368425e-01
3.32420319e-01 -1.11624885e+00 7.52988636e-01 1.14759457e+00
2.96945721e-01 4.82646704e-01 1.09914362e-01 5.23885071e-01
-3.22683394e-01 -8.86687696e-01 8.50399613e-01 -7.05436170e-02
-9.56844747e-01 -2.09919050e-01 -6.22984946e-01 -2.43386492e-01
1.72590446e+00 -5.93000412e-01 -1.91120997e-01 -1.22052722e-01
-6.26826227e-01 2.86069691e-01 4.03262705e-01 4.14063603e-01
4.14042920e-01 -1.15385628e+00 -5.19027054e-01 -3.75342339e-01
-2.22175673e-01 -3.76479030e-01 1.62360936e-01 6.98035181e-01
-1.30497122e+00 1.40647978e-01 -8.46708715e-01 -3.00835639e-01
-1.42466450e+00 4.31095779e-01 3.72427523e-01 -2.57207304e-01
-1.94991231e-01 6.60651386e-01 -6.53676331e-01 -3.02371383e-01
-2.17447564e-01 1.51765078e-01 -7.94996679e-01 -1.66200057e-01
9.87759829e-02 8.94498289e-01 -1.13087766e-01 -9.05953348e-01
-5.87105513e-01 7.89412081e-01 4.80365366e-01 -1.42920613e-01
1.56793451e+00 2.09554344e-01 -4.99974459e-01 -5.12454137e-02
6.92723095e-01 6.31263852e-02 -4.46839690e-01 6.18048966e-01
8.56897514e-03 -6.75682306e-01 -8.65827426e-02 -8.43600273e-01
-1.02072525e+00 7.17832819e-02 7.25089729e-01 2.70929962e-01
1.39650023e+00 -7.84692883e-01 7.45267272e-01 5.09148002e-01
7.25717187e-01 -1.36479211e+00 -9.42177847e-02 5.40300786e-01
9.53161716e-01 -8.97221267e-01 1.41002283e-01 -3.98262262e-01
-4.42990869e-01 1.35920477e+00 7.65149832e-01 -3.84251356e-01
5.47035813e-01 1.14157043e-01 -1.87452376e-01 -1.26499504e-01
-1.64952323e-01 -7.69461244e-02 -4.48837467e-02 5.34163237e-01
1.44677178e-03 -1.72336474e-01 -1.47046542e+00 6.34030163e-01
-5.01351535e-01 6.28660545e-02 5.71267426e-01 1.42105150e+00
-7.51280367e-01 -1.29213214e+00 -9.72232580e-01 2.17081115e-01
-4.31770682e-01 4.99583542e-01 -2.42871255e-01 1.15564072e+00
3.88859183e-01 1.33036268e+00 -1.88075259e-01 -4.44535613e-01
3.96752417e-01 -6.59239590e-01 5.08699477e-01 -1.05926670e-01
-8.73866320e-01 -3.45059127e-01 3.62645462e-02 -2.34476730e-01
-7.13034034e-01 -4.68805820e-01 -1.53391969e+00 -4.15803701e-01
-4.27394062e-01 9.05531943e-01 8.71557951e-01 8.18810582e-01
-6.44810405e-03 7.41889596e-01 7.39750922e-01 -9.67050970e-01
-4.41602349e-01 -5.31997561e-01 -6.58464134e-01 -1.49367660e-01
-4.39303339e-01 -1.27955103e+00 -1.97563246e-01 -6.36792004e-01] | [5.617990970611572, 3.480499505996704] |
f00aba91-e789-4a67-bec1-54cb670e7ef3 | automl-two-sample-test | 2206.08843 | null | https://arxiv.org/abs/2206.08843v3 | https://arxiv.org/pdf/2206.08843v3.pdf | AutoML Two-Sample Test | Two-sample tests are important in statistics and machine learning, both as tools for scientific discovery as well as to detect distribution shifts. This led to the development of many sophisticated test procedures going beyond the standard supervised learning frameworks, whose usage can require specialized knowledge about two-sample testing. We use a simple test that takes the mean discrepancy of a witness function as the test statistic and prove that minimizing a squared loss leads to a witness with optimal testing power. This allows us to leverage recent advancements in AutoML. Without any user input about the problems at hand, and using the same method for all our experiments, our AutoML two-sample test achieves competitive performance on a diverse distribution shift benchmark as well as on challenging two-sample testing problems. We provide an implementation of the AutoML two-sample test in the Python package autotst. | ['Bernhard Schölkopf', 'Krikamol Muandet', 'Simon Buchholz', 'Vincent Stimper', 'Jonas M. Kübler'] | 2022-06-17 | null | null | null | null | ['hypothesis-testing', 'hypothesis-testing'] | ['methodology', 'miscellaneous'] | [ 7.07052127e-02 -1.68369114e-01 -2.70715445e-01 -4.76269990e-01
-1.10289216e+00 -9.75104630e-01 3.67651641e-01 5.95675290e-01
-3.72096062e-01 1.18638384e+00 -5.87517023e-01 -7.77880907e-01
-2.41411328e-01 -8.00667584e-01 -8.47477257e-01 -7.93248653e-01
-2.77525753e-01 7.48634636e-01 4.91957188e-01 2.36988932e-01
5.98297060e-01 3.41563880e-01 -1.44461417e+00 -3.68121415e-02
1.04905319e+00 9.29626465e-01 -3.43817055e-01 8.48612547e-01
1.09122582e-01 -5.80596030e-02 -5.73849261e-01 -5.41239440e-01
2.87800044e-01 -7.18537331e-01 -7.65322506e-01 -3.25501084e-01
5.31607687e-01 1.17635049e-01 2.59531826e-01 1.23562300e+00
7.43511915e-01 -2.11221408e-02 6.53051615e-01 -1.34709656e+00
-2.72013545e-02 6.51029170e-01 -6.31161690e-01 2.61503279e-01
2.68600911e-01 1.39178813e-01 1.15975022e+00 -7.82907605e-01
2.34387219e-01 1.09005332e+00 7.51614034e-01 2.17042908e-01
-1.71921432e+00 -7.86147714e-01 -5.36904633e-01 2.32135326e-01
-1.19906509e+00 -1.89951107e-01 2.96027124e-01 -4.96208310e-01
5.06173730e-01 3.97037446e-01 5.47410905e-01 7.71089077e-01
1.92224294e-01 7.35739529e-01 1.42114019e+00 -5.32771587e-01
5.87984145e-01 2.56347477e-01 2.41628453e-01 7.44295716e-01
4.92335051e-01 -4.26942147e-02 -2.94749528e-01 -4.92537916e-01
4.01174754e-01 -2.68707305e-01 -2.03217849e-01 -4.51586574e-01
-9.37859952e-01 9.63823378e-01 -3.50679010e-01 1.91806599e-01
2.49113694e-01 -8.06486532e-02 4.51460719e-01 5.71595013e-01
5.03011405e-01 5.71530461e-01 -7.64630973e-01 -2.82303005e-01
-9.72417295e-01 4.29560691e-01 1.20490360e+00 5.19966900e-01
7.48237193e-01 -2.55728811e-01 -3.63475889e-01 5.49051762e-01
-1.34770855e-01 7.13131666e-01 2.38959506e-01 -8.20714951e-01
1.00905344e-01 3.77184063e-01 -4.05407175e-02 -5.06981671e-01
-1.41501561e-01 -5.56896210e-01 -5.87848485e-01 1.83086589e-01
7.89767623e-01 -2.70744655e-02 -5.52333295e-01 1.61603260e+00
4.26298976e-01 2.85829425e-01 -3.98346841e-01 3.85648370e-01
1.82395726e-01 2.82693505e-01 -4.78707284e-01 -2.89333045e-01
1.07098019e+00 -2.89336622e-01 -1.80852234e-01 1.69173092e-01
9.96425986e-01 -5.99411368e-01 1.24744403e+00 8.03066492e-01
-1.12536144e+00 -1.18362680e-01 -1.09045136e+00 1.73003212e-01
-5.29919088e-01 -3.88172895e-01 6.01127446e-01 8.47286463e-01
-7.84988821e-01 8.52079809e-01 -6.21892214e-01 -2.49903858e-01
7.63728619e-01 3.11031729e-01 -4.80353653e-01 -1.53886855e-01
-7.49381781e-01 5.50529361e-01 2.19242722e-01 -2.43394241e-01
-6.67399764e-01 -9.81996238e-01 -8.23967636e-01 7.42250979e-02
4.28363830e-01 -3.35514933e-01 1.17734909e+00 -4.62552339e-01
-1.20902562e+00 1.08700740e+00 -2.86805052e-02 -5.25053084e-01
7.14053512e-01 -3.93779427e-02 -3.26137692e-02 -2.82302290e-01
3.49138618e-01 3.57061476e-02 2.41749853e-01 -6.16325319e-01
-6.57223344e-01 -3.80355626e-01 -4.44743335e-01 -4.94999498e-01
-1.84890013e-02 -2.14833215e-01 -1.24894060e-01 -2.64606148e-01
-1.61416844e-01 -6.39970839e-01 -7.38975033e-03 -1.18894547e-01
-6.87767446e-01 -3.83255363e-01 7.26629376e-01 -2.70281911e-01
9.83410478e-01 -2.20093513e+00 -2.39503652e-01 7.60683179e-01
9.40854847e-02 2.47575998e-01 -1.65389463e-01 2.67377406e-01
-1.53070524e-01 1.46130428e-01 -5.84156811e-01 -4.31436533e-03
2.37479627e-01 -7.52692819e-02 -3.02489787e-01 7.80314147e-01
1.53904021e-01 8.68017554e-01 -8.26212227e-01 -3.32859665e-01
-1.16209306e-01 -7.90807158e-02 -6.92936599e-01 1.72197893e-01
-3.19751769e-01 5.77210784e-01 -2.09523693e-01 4.77846563e-01
7.85473228e-01 -3.71872753e-01 1.19675197e-01 3.94686967e-01
-2.44700145e-02 3.83965611e-01 -1.02276528e+00 1.22056174e+00
-6.10942245e-02 6.06311798e-01 -2.41903827e-01 -1.37690008e+00
8.53516340e-01 -1.77442074e-01 2.34981403e-01 -5.85902929e-01
6.89603314e-02 6.69205487e-01 1.73251778e-01 -4.57367390e-01
-4.27513033e-01 -4.09440815e-01 -1.68495968e-01 5.11169016e-01
-3.93407494e-02 -2.34538615e-01 4.88037229e-01 3.50873843e-02
1.45796883e+00 -1.18879832e-01 4.90500778e-01 -6.01475418e-01
5.27843833e-01 -2.48517618e-01 6.73312962e-01 9.82142925e-01
1.81162328e-01 2.90976197e-01 1.19354606e+00 -2.95633912e-01
-8.80445361e-01 -1.26252484e+00 -5.88064730e-01 7.35317945e-01
-2.90145814e-01 -2.45253623e-01 -4.01997805e-01 -9.20140505e-01
5.64431131e-01 7.45013237e-01 -6.89455509e-01 -3.49444943e-03
-2.80768126e-01 -6.31313324e-01 6.70091510e-01 3.16375017e-01
3.28197449e-01 -6.20534778e-01 -3.64091605e-01 -2.71475375e-01
3.91645372e-01 -7.11023390e-01 -4.16767418e-01 5.17723680e-01
-7.46230483e-01 -1.63128817e+00 -4.89733726e-01 -7.75532305e-01
5.27906954e-01 -2.64554828e-01 1.12856054e+00 6.24284074e-02
-6.36105120e-01 2.19197676e-01 -1.09448336e-01 -5.89169085e-01
-2.27042019e-01 9.54631642e-02 -9.99842510e-02 -2.92458475e-01
6.08434498e-01 -7.03252017e-01 -4.11550552e-01 3.47210616e-01
-9.27692413e-01 -5.40298760e-01 5.07494807e-01 1.00076032e+00
5.98390698e-01 5.68476655e-02 8.45849812e-01 -1.13385320e+00
5.10135472e-01 -6.62370980e-01 -1.25939655e+00 3.08235168e-01
-7.87004054e-01 3.70427907e-01 8.32648575e-01 -3.32111239e-01
-6.26198873e-02 -3.17949265e-01 -1.91126317e-02 -8.34982693e-02
5.65995164e-02 5.78465819e-01 -2.61280000e-01 -5.59496172e-02
5.41905880e-01 2.42401585e-01 1.40102670e-01 -3.46938014e-01
-2.58063257e-01 4.34363931e-01 3.80153447e-01 -7.64215112e-01
8.44734311e-01 8.27701688e-02 7.85036325e-01 -7.26082265e-01
-1.15941596e+00 -4.04849857e-01 -3.95300806e-01 2.08381698e-01
4.02180403e-01 -1.99248493e-01 -1.17013323e+00 2.17166945e-01
-7.24026144e-01 -6.82309330e-01 -4.34892118e-01 5.84481776e-01
-5.40975034e-01 3.22065026e-01 -1.22436188e-01 -8.57056201e-01
-1.19550250e-01 -9.53052878e-01 9.41064954e-01 1.38645843e-01
-5.69534525e-02 -1.25031137e+00 4.65791494e-01 -8.06504413e-02
3.43244463e-01 3.83794874e-01 1.34913373e+00 -1.34761941e+00
-2.85553813e-01 -5.82427621e-01 4.47606556e-02 4.73326564e-01
1.01132207e-02 -3.90833132e-02 -7.75763273e-01 -4.42783117e-01
-1.41586110e-01 -5.67251265e-01 8.70745897e-01 3.77268076e-01
1.74168062e+00 -2.02952936e-01 -3.16488862e-01 7.90927291e-01
1.20158935e+00 -2.29731038e-01 5.89963436e-01 1.95528179e-01
1.28708780e-01 4.99987960e-01 5.99881172e-01 4.50165361e-01
2.24307738e-02 4.17616338e-01 1.74522251e-02 1.04119562e-01
5.68082154e-01 -1.80016786e-01 1.75815076e-01 4.65581000e-01
3.99804711e-01 1.80018786e-02 -9.19308126e-01 2.98444033e-01
-1.67248321e+00 -1.14805067e+00 -2.97455370e-01 3.00306106e+00
1.31866860e+00 3.17182750e-01 3.48544419e-01 4.53167170e-01
6.51233912e-01 -3.78648996e-01 -8.18599939e-01 -3.55685055e-01
-2.40740374e-01 9.45264041e-01 4.32984501e-01 5.76772034e-01
-8.57010245e-01 4.98403013e-01 6.95708275e+00 9.77362394e-01
-1.01449132e+00 -1.46395773e-01 7.74246991e-01 -4.63238219e-03
-4.21071678e-01 2.31772527e-01 -7.56039381e-01 6.02265835e-01
8.88590693e-01 -4.30907667e-01 3.55426520e-01 7.35937536e-01
-6.13903580e-03 -4.21154946e-01 -1.58925354e+00 8.82998645e-01
-2.36317575e-01 -1.12812293e+00 -5.36604881e-01 2.23877758e-01
5.70029140e-01 -6.72538579e-02 1.73594683e-01 3.82549077e-01
2.07711309e-01 -1.23057854e+00 1.83241561e-01 3.15019429e-01
1.05324137e+00 -8.55598748e-01 9.34201241e-01 3.62506092e-01
-5.92286766e-01 3.26062292e-01 -4.67068940e-01 -5.02297357e-02
-4.29133594e-01 1.29691815e+00 -1.06970131e+00 1.36828393e-01
4.51758176e-01 4.25512761e-01 -5.67160428e-01 1.58813179e+00
-2.13277832e-01 1.13243473e+00 -5.14292777e-01 -1.62814319e-01
-2.96069503e-01 -2.78031141e-01 5.16981423e-01 1.21431804e+00
4.08071816e-01 -4.31589663e-01 1.38330951e-01 8.43283951e-01
-2.29649305e-01 1.48300171e-01 -6.29711509e-01 2.18923707e-02
4.69118446e-01 1.13897049e+00 -6.61154926e-01 -2.04637095e-01
-2.64802158e-01 6.64996743e-01 2.42377698e-01 1.91547588e-01
-7.18649268e-01 -8.33813071e-01 5.11646628e-01 4.19640601e-01
4.49681163e-01 -1.19875185e-01 -2.88915485e-01 -9.41179574e-01
1.57457903e-01 -1.09050179e+00 5.73816717e-01 -1.79429173e-01
-1.67727613e+00 -2.07056075e-01 1.12841979e-01 -8.38848174e-01
-3.28593165e-01 -9.01834071e-01 -9.46800172e-01 7.86212027e-01
-1.31407130e+00 -3.42888802e-01 -2.79972181e-02 5.02252698e-01
-6.57120049e-02 -1.47156805e-01 8.31288397e-01 3.07471484e-01
-4.05782163e-01 9.78792489e-01 2.66453356e-01 7.70907551e-02
1.03085148e+00 -1.55989099e+00 1.10790022e-01 6.85036898e-01
5.98897226e-02 4.53251362e-01 8.97594452e-01 -5.55740595e-01
-1.46119332e+00 -6.30128801e-01 6.44353807e-01 -3.64201039e-01
1.04784584e+00 -5.53401470e-01 -9.14853454e-01 4.61348951e-01
-1.56824306e-01 3.86707336e-01 1.06976008e+00 3.70541990e-01
-6.02464557e-01 -3.60162765e-01 -1.35650420e+00 2.67858952e-01
7.62348533e-01 -3.69175702e-01 -2.23915771e-01 2.90811449e-01
1.80615172e-01 -4.16440479e-02 -8.42056096e-01 4.12955493e-01
6.57520652e-01 -1.01282680e+00 4.36200500e-01 -8.83801103e-01
2.17856109e-01 -2.11593255e-01 -5.36000766e-02 -1.16734695e+00
4.55584787e-02 -8.35929215e-01 2.65331548e-02 1.04261792e+00
6.32169425e-01 -1.14133942e+00 6.91169918e-01 3.02534640e-01
1.31683290e-01 -9.17690814e-01 -9.91743922e-01 -9.75165486e-01
4.15294498e-01 -5.11574924e-01 3.43485832e-01 8.67782176e-01
9.71665606e-02 1.40331253e-01 1.37431681e-01 7.35357329e-02
8.79594207e-01 2.44474173e-01 1.09891140e+00 -1.59937060e+00
-6.67316496e-01 -6.81814075e-01 -8.21193218e-01 -8.50093126e-01
3.57474357e-01 -1.13163638e+00 1.01739794e-01 -9.46600556e-01
5.72364867e-01 -3.07719529e-01 -2.20960647e-01 5.72554886e-01
-1.76764399e-01 3.57141405e-01 -5.56382537e-01 -2.81226248e-01
-5.91607153e-01 3.83907109e-01 9.21021044e-01 5.63966185e-02
1.11890044e-02 3.22197616e-01 -7.67132342e-01 5.52103162e-01
8.22409809e-01 -7.59667456e-01 -2.67302752e-01 2.70683825e-01
5.84625125e-01 -2.29324713e-01 4.15838182e-01 -9.62411642e-01
9.97927487e-02 -3.07048231e-01 4.75612789e-01 -4.41584587e-01
-3.11477005e-01 -3.35943043e-01 -3.23300689e-01 5.33437610e-01
-3.75927657e-01 1.72766838e-02 3.78243327e-01 3.66266668e-01
-5.22366026e-04 -4.50897038e-01 9.40357387e-01 3.96456987e-01
1.34658620e-01 2.92698294e-01 1.25560999e-01 5.08672416e-01
1.13997483e+00 -3.93023109e-03 -5.44192493e-01 -2.74196476e-01
-3.79188657e-01 4.86133903e-01 5.10159969e-01 -2.78072238e-01
3.78469676e-01 -1.08865201e+00 -9.36972439e-01 4.89476264e-01
6.22836649e-02 1.45106176e-02 -2.98989236e-01 1.22947490e+00
-4.40009356e-01 3.39200884e-01 4.21377309e-02 -6.17165565e-01
-1.18461514e+00 2.31827870e-01 2.10881770e-01 -2.59095907e-01
-3.16421062e-01 8.49305451e-01 1.30011201e-01 -4.27654237e-01
1.52547404e-01 -1.28928989e-01 4.79177833e-01 -2.51578897e-01
7.32346952e-01 4.92570937e-01 -6.43816311e-03 3.50653052e-01
-4.50339526e-01 4.92270172e-01 3.25108059e-02 -2.42451519e-01
1.51529515e+00 4.04121101e-01 -3.79741192e-01 8.46871912e-01
1.27096057e+00 5.86702287e-01 -9.05659914e-01 7.30603784e-02
2.96649784e-01 -5.67090571e-01 -3.63646597e-01 -8.37183118e-01
-6.73374116e-01 8.12883675e-01 2.35388190e-01 6.28642380e-01
8.67325425e-01 1.72208279e-01 3.52218717e-01 5.92463255e-01
4.20656979e-01 -7.85800874e-01 1.35676712e-02 4.97206628e-01
4.41142112e-01 -1.30264795e+00 2.24538699e-01 -1.89920440e-01
-5.10385148e-02 1.06524956e+00 2.98148483e-01 -1.81561202e-01
7.08301127e-01 5.07555485e-01 -4.27348703e-01 -3.17213088e-02
-8.76419485e-01 -2.23456711e-01 5.25475323e-01 4.20359880e-01
5.62004685e-01 -2.38392875e-01 -5.34858823e-01 4.74742413e-01
-2.65197188e-01 3.31522636e-02 2.82464087e-01 7.49482810e-01
-6.24698341e-01 -1.38876152e+00 -2.06330776e-01 8.17347646e-01
-5.35307944e-01 -1.21052042e-02 -4.21920776e-01 7.63235629e-01
-7.42162913e-02 6.99447095e-01 1.37897387e-01 -2.18997717e-01
-4.40299511e-02 3.04726958e-01 7.73067176e-01 -6.68381512e-01
-1.25644729e-01 -1.85063303e-01 -2.21007988e-01 -5.07182181e-01
-4.10303101e-02 -6.85652912e-01 -1.18794882e+00 -6.60961032e-01
-2.97934175e-01 5.75172484e-01 6.30861938e-01 1.11149180e+00
1.93587899e-01 -7.63386190e-02 7.35973477e-01 -2.99626410e-01
-8.69940341e-01 -9.39070880e-01 -8.58274400e-01 6.78946748e-02
2.38508508e-01 -4.69652712e-01 -7.59853423e-01 -5.40336490e-01] | [7.58945894241333, 4.2512335777282715] |
0efea464-decb-47ca-8691-30a2ee9a2609 | semi-supervised-medical-image-segmentation-2 | 2112.04894 | null | https://arxiv.org/abs/2112.04894v2 | https://arxiv.org/pdf/2112.04894v2.pdf | Semi-Supervised Medical Image Segmentation via Cross Teaching between CNN and Transformer | Recently, deep learning with Convolutional Neural Networks (CNNs) and Transformers has shown encouraging results in fully supervised medical image segmentation. However, it is still challenging for them to achieve good performance with limited annotations for training. In this work, we present a very simple yet efficient framework for semi-supervised medical image segmentation by introducing the cross teaching between CNN and Transformer. Specifically, we simplify the classical deep co-training from consistency regularization to cross teaching, where the prediction of a network is used as the pseudo label to supervise the other network directly end-to-end. Considering the difference in learning paradigm between CNN and Transformer, we introduce the Cross Teaching between CNN and Transformer rather than just using CNNs. Experiments on a public benchmark show that our method outperforms eight existing semi-supervised learning methods just with a simpler framework. Notably, this work may be the first attempt to combine CNN and transformer for semi-supervised medical image segmentation and achieve promising results on a public benchmark. The code will be released at: https://github.com/HiLab-git/SSL4MIS. | ['Shaoting Zhang', 'Guotai Wang', 'Tao Song', 'Minhao Hu', 'Xiangde Luo'] | 2021-12-09 | null | null | null | null | ['semi-supervised-medical-image-segmentation'] | ['computer-vision'] | [ 1.24746390e-01 4.80986983e-01 -2.80125052e-01 -6.09905303e-01
-7.62835383e-01 -3.68913472e-01 1.69500619e-01 -7.63082877e-02
-5.04641891e-01 5.63958049e-01 -2.21656322e-01 -5.63492894e-01
2.31634691e-01 -6.46966338e-01 -8.06906044e-01 -7.06317246e-01
2.04503179e-01 6.99179113e-01 4.10057008e-01 6.86754808e-02
-4.32400554e-01 2.32425317e-01 -9.73468959e-01 3.76811713e-01
8.67244780e-01 1.06906629e+00 1.38550505e-01 4.35113043e-01
-1.97720677e-02 1.11727870e+00 -2.25679323e-01 -3.93092901e-01
1.50260106e-01 -3.81288677e-01 -1.35877609e+00 2.63960421e-01
3.01096112e-01 -2.54908174e-01 -7.35571384e-02 9.74470794e-01
5.63565195e-01 -2.69540757e-01 3.05297345e-01 -1.06352782e+00
-2.93451816e-01 8.42816055e-01 -4.28903580e-01 -4.83933911e-02
-1.88192606e-01 -9.25418513e-04 7.87108958e-01 -5.86437702e-01
4.86523658e-01 8.60961914e-01 9.51290011e-01 7.74086773e-01
-1.11778653e+00 -6.61243320e-01 1.16993953e-02 1.64183211e-02
-1.31250954e+00 -1.13211803e-01 7.51026690e-01 -3.82621378e-01
5.78876436e-01 1.68329850e-01 7.32773721e-01 7.93010652e-01
-5.31949624e-02 1.26208889e+00 1.18104482e+00 -2.13839367e-01
1.34595903e-02 2.03030661e-01 2.85166949e-01 1.12568104e+00
-1.17367819e-01 9.44749825e-03 -5.79388021e-03 5.34504615e-02
6.27171516e-01 1.65805370e-01 -2.41741091e-01 -5.62523186e-01
-1.23522735e+00 7.02824712e-01 8.10601711e-01 4.79208469e-01
-1.35722607e-01 1.87884539e-01 6.42917275e-01 1.29370928e-01
6.75306439e-01 1.43464819e-01 -7.56347001e-01 1.51090115e-01
-1.21853840e+00 -1.16186745e-01 7.31465697e-01 8.02500069e-01
7.00826049e-01 -2.01422334e-01 -2.33060285e-01 8.76486480e-01
1.74741134e-01 5.16731404e-02 5.72199941e-01 -8.81535649e-01
1.07809201e-01 6.57085955e-01 -3.83947313e-01 -4.39397842e-01
-5.60262263e-01 -7.45951355e-01 -1.09610724e+00 1.51731476e-01
5.52408993e-01 -2.62759686e-01 -1.21858215e+00 1.51921439e+00
4.31383044e-01 3.86501580e-01 -1.62539557e-01 7.82334149e-01
1.16426730e+00 2.64093846e-01 2.53573749e-02 6.17826693e-02
1.34923339e+00 -1.56091094e+00 -6.08752668e-01 -2.49625575e-02
1.04562497e+00 -6.33198738e-01 9.56442416e-01 4.63101953e-01
-1.18887687e+00 -5.70856094e-01 -9.60760951e-01 -1.77717179e-01
-3.32916528e-01 3.73647749e-01 7.44127750e-01 4.39934164e-01
-1.25095475e+00 7.82091260e-01 -1.29850829e+00 -2.66609877e-01
9.40763772e-01 5.66003680e-01 -2.64824808e-01 4.23374698e-02
-1.06207585e+00 6.45569980e-01 4.95283395e-01 2.59782016e-01
-8.93225253e-01 -8.27016830e-01 -8.25995862e-01 -2.49446794e-01
4.63704497e-01 -7.66312063e-01 1.39590180e+00 -1.24833810e+00
-1.51551306e+00 1.25661981e+00 1.69500366e-01 -6.91301942e-01
9.08634722e-01 -1.15270101e-01 1.10897370e-01 5.83400801e-02
1.94683924e-01 1.05445302e+00 4.47108120e-01 -1.17290628e+00
-6.19928181e-01 -1.66940451e-01 4.33928780e-02 -4.38896976e-02
-2.21161053e-01 -1.72265857e-01 -8.49142730e-01 -4.98495251e-01
4.54177000e-02 -1.11605978e+00 -4.44199175e-01 1.92698970e-01
-9.19043958e-01 -3.62930357e-01 8.07740867e-01 -4.05826479e-01
8.40880394e-01 -2.09519029e+00 4.77823522e-03 2.96544880e-02
4.58856165e-01 4.90342438e-01 1.31377444e-01 -1.10041924e-01
-2.83958435e-01 6.97304979e-02 -6.71371341e-01 -7.05640316e-01
-2.40480185e-01 3.74134481e-01 1.57308102e-01 5.70410967e-01
8.33863541e-02 1.15498567e+00 -9.43444908e-01 -7.71300614e-01
3.04212540e-01 6.57180727e-01 -4.98719066e-01 1.31087810e-01
-1.20356336e-01 7.48859286e-01 -3.44335288e-01 7.37189770e-01
6.81891084e-01 -7.54414678e-01 1.37294456e-01 -3.08252603e-01
1.64969325e-01 1.75110847e-01 -7.89543033e-01 1.86817467e+00
-3.66704196e-01 4.47383255e-01 1.61876336e-01 -1.43796003e+00
4.79137272e-01 4.72446024e-01 6.41663253e-01 -5.44874251e-01
3.49686295e-01 2.79612035e-01 -1.38500690e-01 -4.04454499e-01
-6.35368675e-02 -2.29106262e-01 1.60574511e-01 3.02277654e-01
2.43803784e-01 -1.33041859e-01 3.87648314e-01 1.58493742e-01
8.77724111e-01 3.16159427e-01 1.04695745e-01 -2.28587955e-01
5.80302119e-01 1.19074821e-01 6.16347849e-01 5.28335452e-01
-3.27219903e-01 7.99566269e-01 5.01963973e-01 -5.95090091e-01
-7.28840828e-01 -9.63928223e-01 -3.25766832e-01 9.32250381e-01
-1.31681293e-01 -3.00758421e-01 -1.03247738e+00 -1.27399743e+00
-2.58278430e-01 1.39783531e-01 -9.25288856e-01 8.14384669e-02
-5.09149730e-01 -8.06119621e-01 7.19023168e-01 6.95501626e-01
6.60759270e-01 -1.09608877e+00 -4.75502938e-01 1.85934171e-01
-2.46681556e-01 -1.33283293e+00 -4.36818779e-01 5.52716911e-01
-1.08427799e+00 -1.28234994e+00 -1.09444153e+00 -1.15753269e+00
9.96727407e-01 -5.82315028e-02 1.25120544e+00 4.55368459e-01
-3.38553160e-01 9.02507156e-02 -3.31809938e-01 -2.99135029e-01
-4.10131276e-01 4.04126644e-01 -4.84210670e-01 -7.71515518e-02
1.20932399e-03 -4.26254809e-01 -7.80642807e-01 3.78084987e-01
-8.51915061e-01 4.58042681e-01 6.22587323e-01 9.87131059e-01
8.80410850e-01 -1.83355972e-01 1.83152393e-01 -1.47911668e+00
9.25529376e-02 -3.02112281e-01 -3.57150644e-01 2.11581945e-01
-5.87116182e-01 -1.05528750e-01 5.65364480e-01 -1.29115447e-01
-8.32175314e-01 5.22828281e-01 -6.06002152e-01 -3.28991354e-01
-3.59669000e-01 5.07832885e-01 2.01829940e-01 4.79647592e-02
3.97086531e-01 2.11919755e-01 1.85219258e-01 -5.09953439e-01
1.68075174e-01 4.71686870e-01 4.78472710e-01 -3.06979209e-01
6.93267226e-01 8.70256424e-01 -1.73145145e-01 -3.33583623e-01
-1.28087234e+00 -4.90937531e-01 -8.96470070e-01 5.10520814e-03
1.15987051e+00 -9.30753767e-01 -4.16444361e-01 6.45571828e-01
-9.51562881e-01 -8.88838112e-01 -4.19364423e-01 3.16377223e-01
-4.87754077e-01 4.10850763e-01 -9.88168716e-01 -6.88195229e-02
-5.55038750e-01 -1.53999007e+00 1.09044576e+00 4.62726094e-02
-1.54373338e-02 -1.42127633e+00 -1.16748828e-02 7.66312122e-01
3.84719580e-01 4.01481122e-01 4.99296635e-01 -7.91216910e-01
-3.94852430e-01 -2.06897691e-01 -2.70896703e-01 7.34446645e-01
2.81444192e-01 -5.51968887e-02 -1.05157197e+00 -4.10893172e-01
-1.25681862e-01 -6.74586177e-01 1.09662414e+00 5.92015445e-01
1.56891572e+00 3.40727121e-02 -5.07125080e-01 1.05119205e+00
1.23786461e+00 -9.12654027e-02 6.41327977e-01 3.30877751e-01
9.80949759e-01 4.62697476e-01 4.57390994e-01 8.45577344e-02
4.43885058e-01 3.90531957e-01 4.42320019e-01 -9.48430240e-01
-2.55843133e-01 -1.57217251e-03 3.73517498e-02 9.23697889e-01
9.75057408e-02 9.73143354e-02 -1.00411725e+00 5.85714877e-01
-1.85282683e+00 -4.30841774e-01 -1.78541094e-01 1.78860486e+00
1.13464189e+00 1.63518667e-01 1.31104410e-01 1.52916774e-01
4.87764090e-01 -7.95794129e-02 -3.85866016e-01 -1.87851284e-02
1.59913436e-01 3.09158742e-01 4.24288929e-01 3.98462534e-01
-1.52892661e+00 9.80472982e-01 5.65788507e+00 1.00855005e+00
-1.40645123e+00 5.08065224e-01 1.12722981e+00 9.31172594e-02
8.57679769e-02 -1.39699429e-01 -5.84993839e-01 3.62545490e-01
7.84738481e-01 4.72509950e-01 -1.76947862e-01 8.52887511e-01
8.01959187e-02 8.03480390e-03 -1.05872214e+00 6.74440205e-01
-1.35056064e-01 -1.42708755e+00 -2.82032341e-02 -1.01117633e-01
7.52778709e-01 3.29939902e-01 4.31683473e-02 2.38719702e-01
3.56047183e-01 -1.01727498e+00 5.87253988e-01 1.30279258e-01
8.40132713e-01 -5.95457971e-01 1.02823913e+00 2.92571753e-01
-1.08906651e+00 3.69664401e-01 -1.91877633e-01 3.09568346e-01
3.67222615e-02 7.00651348e-01 -9.14077640e-01 5.55458665e-01
8.25809121e-01 1.01876104e+00 -6.02615058e-01 1.03862476e+00
-2.66196579e-01 8.08925688e-01 -6.08471110e-02 3.40351313e-01
4.99689519e-01 5.97041510e-02 4.68276739e-02 1.50210655e+00
-9.62408632e-02 -2.33761922e-01 4.57629919e-01 8.47321451e-01
-2.54974514e-01 1.03241988e-01 -2.81979799e-01 2.38629788e-01
-2.45661020e-01 1.53699780e+00 -1.17591882e+00 -5.73819160e-01
-3.71031314e-01 1.05988705e+00 3.43531936e-01 1.37778908e-01
-9.92739558e-01 -2.93254048e-01 8.52213427e-02 1.69601038e-01
2.67238051e-01 1.97042674e-01 -3.63762230e-01 -1.01290929e+00
-2.17595175e-01 -6.04833603e-01 5.75736701e-01 -6.00141406e-01
-1.13592911e+00 7.87532806e-01 -7.63883144e-02 -1.28364551e+00
1.24616697e-01 -6.58223271e-01 -5.68632722e-01 5.56488216e-01
-1.87065995e+00 -1.47701299e+00 -3.72982949e-01 8.23628724e-01
5.10495365e-01 9.41019729e-02 6.96726680e-01 5.40319264e-01
-6.60375297e-01 7.12110162e-01 1.96310785e-02 6.23385549e-01
6.40798390e-01 -1.57015252e+00 1.22532263e-01 5.58456361e-01
1.87440813e-02 3.60745758e-01 3.14398140e-01 -4.04798806e-01
-8.55806053e-01 -1.37245691e+00 6.59095466e-01 -2.65078455e-01
5.28695703e-01 -2.97169417e-01 -8.34540784e-01 8.77695501e-01
4.90099162e-01 4.95527446e-01 7.67829597e-01 -3.72146629e-02
-1.81711867e-01 -1.45437866e-01 -1.13050079e+00 2.58176863e-01
6.91185355e-01 -3.58678162e-01 -2.41023198e-01 7.65733123e-01
8.04655075e-01 -7.32806206e-01 -9.39859092e-01 6.53439343e-01
2.14381903e-01 -9.30058777e-01 8.36498141e-01 -1.90102711e-01
5.30925512e-01 -3.30716699e-01 3.68964404e-01 -1.12918365e+00
-8.80259275e-02 -3.99035782e-01 1.74700990e-01 1.01469541e+00
6.51402473e-01 -5.79502761e-01 1.06283832e+00 1.63867980e-01
-4.22479153e-01 -1.25881004e+00 -7.34732747e-01 -6.65610671e-01
4.19404805e-01 -4.45954293e-01 1.74515605e-01 1.05807769e+00
-1.72977895e-01 9.24187154e-02 -2.99134195e-01 -1.05615549e-01
7.41228461e-01 3.48017029e-02 5.13695002e-01 -1.01443326e+00
-4.28109616e-01 -3.99097711e-01 -1.44868091e-01 -1.03170609e+00
2.05651283e-01 -1.29129088e+00 1.88608408e-01 -1.65339267e+00
5.15644431e-01 -5.57912767e-01 -4.16219294e-01 9.05541301e-01
-5.89376092e-02 7.63832629e-01 -8.93075094e-02 2.05423489e-01
-8.91818523e-01 3.20321709e-01 1.56974435e+00 -3.89451772e-01
-4.57223915e-02 2.63369173e-01 -5.99254549e-01 8.40602756e-01
9.91760910e-01 -5.31752586e-01 -3.84207189e-01 -4.56949562e-01
-1.18803330e-01 -7.59633183e-02 4.72590387e-01 -9.58045065e-01
3.34364355e-01 3.57410640e-01 2.75930285e-01 -6.81746006e-01
-9.88576785e-02 -7.48925984e-01 -2.13614833e-02 8.41645062e-01
-4.31948721e-01 -2.53531128e-01 2.28204727e-01 1.77379146e-01
-4.55985248e-01 -8.11407045e-02 1.06576228e+00 -3.93375874e-01
-4.61177826e-01 6.63231015e-01 -1.77167654e-01 1.32515267e-01
9.56991851e-01 -1.21953607e-01 -3.18743405e-03 -9.94422808e-02
-1.22035682e+00 5.33374369e-01 2.43400022e-01 1.11645959e-01
5.98865628e-01 -1.00554466e+00 -5.67763507e-01 6.07231399e-03
-2.39016563e-02 5.36953032e-01 2.52190560e-01 1.36950326e+00
-7.66891062e-01 4.10525471e-01 -5.21323346e-02 -1.00138950e+00
-1.35015464e+00 4.01265800e-01 8.50428164e-01 -7.26681590e-01
-6.23973012e-01 1.05253112e+00 3.60107750e-01 -9.56275403e-01
4.74603176e-01 -5.76856136e-01 -7.16753528e-02 -1.87713936e-01
3.18854064e-01 -6.21372797e-02 2.05774754e-01 -4.66169864e-01
-4.53882843e-01 4.95605469e-01 -4.16116863e-01 2.44444638e-01
1.38788140e+00 1.00680150e-01 -2.97683805e-01 3.62179726e-01
1.51512134e+00 -5.07080913e-01 -1.30692422e+00 -3.87876511e-01
-3.58493812e-02 1.43291473e-01 1.65275171e-01 -9.21783328e-01
-1.73166621e+00 1.06588805e+00 8.20378721e-01 -3.61988172e-02
1.13590932e+00 2.66294301e-01 1.06277621e+00 1.86191395e-01
2.17982069e-01 -9.57495749e-01 1.03912830e-01 3.88920367e-01
5.39286494e-01 -1.50056052e+00 -7.36767054e-02 -5.34519613e-01
-6.43902123e-01 1.00128281e+00 6.15533769e-01 -1.73705325e-01
1.08613181e+00 3.88167799e-01 3.32209438e-01 -4.05016601e-01
-4.67883915e-01 -3.87874901e-01 3.68263185e-01 3.43996733e-01
8.35846126e-01 1.37295648e-01 -7.71737918e-02 3.24450284e-01
-1.89251341e-02 3.80138814e-01 2.29961127e-01 9.07895625e-01
-8.32240433e-02 -1.23884869e+00 9.98868421e-02 4.01377350e-01
-1.00293553e+00 -1.96965083e-01 -2.05558464e-01 9.13681388e-01
3.42832714e-01 6.24509513e-01 -4.31113429e-02 -2.55725682e-01
2.45772839e-01 -1.47182897e-01 3.00674438e-01 -7.97242999e-01
-9.28083420e-01 3.07842195e-01 -1.73386022e-01 -6.74298584e-01
-7.68362582e-01 -5.26964903e-01 -1.54681289e+00 -6.85840920e-02
-1.93553880e-01 2.73585737e-01 4.11400944e-01 1.01565444e+00
7.53833055e-02 8.30834508e-01 3.90753120e-01 -6.73483074e-01
-2.40385219e-01 -8.35990071e-01 -4.51806277e-01 4.55789834e-01
3.59449923e-01 -3.83215636e-01 -2.36521408e-01 1.61645576e-01] | [14.671899795532227, -2.409572124481201] |
eaa6c026-a74a-47ff-b0c7-01b53371b3f1 | viskop-visual-knowledge-oriented-programming | 2307.03130 | null | https://arxiv.org/abs/2307.03130v1 | https://arxiv.org/pdf/2307.03130v1.pdf | VisKoP: Visual Knowledge oriented Programming for Interactive Knowledge Base Question Answering | We present Visual Knowledge oriented Programming platform (VisKoP), a knowledge base question answering (KBQA) system that integrates human into the loop to edit and debug the knowledge base (KB) queries. VisKoP not only provides a neural program induction module, which converts natural language questions into knowledge oriented program language (KoPL), but also maps KoPL programs into graphical elements. KoPL programs can be edited with simple graphical operators, such as dragging to add knowledge operators and slot filling to designate operator arguments. Moreover, VisKoP provides auto-completion for its knowledge base schema and users can easily debug the KoPL program by checking its intermediate results. To facilitate the practical KBQA on a million-entity-level KB, we design a highly efficient KoPL execution engine for the back-end. Experiment results show that VisKoP is highly efficient and user interaction can fix a large portion of wrong KoPL programs to acquire the correct answer. The VisKoP online demo https://demoviskop.xlore.cn (Stable release of this paper) and https://viskop.xlore.cn (Beta release with new features), highly efficient KoPL engine https://pypi.org/project/kopl-engine, and screencast video https://youtu.be/zAbJtxFPTXo are now publicly available. | ['Juanzi Li', 'Lei Hou', 'Peng Zhang', 'Jianjun Xu', 'Hailong Jin', 'Jifan Yu', 'Amy Xin', 'Shulin Cao', 'Xin Lv', 'Yuanyong Chen', 'Zijun Yao'] | 2023-07-06 | null | null | null | null | ['program-induction', 'knowledge-base-question-answering', 'question-answering', 'slot-filling'] | ['computer-code', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [-9.91409123e-01 1.41355678e-01 -1.94604963e-01 -2.93121368e-01
-6.34786785e-01 -8.04618359e-01 -2.67650157e-01 4.90455367e-02
-2.01655328e-01 6.88155890e-01 -1.43522307e-01 -9.44512129e-01
-2.21028998e-01 -1.10371232e+00 -9.48485494e-01 -3.54341194e-02
7.12625310e-02 6.48019731e-01 4.72921431e-01 -4.25476283e-01
8.03623423e-02 5.88771552e-02 -1.53339636e+00 5.60909867e-01
1.29592311e+00 5.10023296e-01 5.79591990e-01 1.13301992e+00
-6.41162574e-01 1.32592082e+00 -7.44833410e-01 -6.42371118e-01
-9.60600525e-02 1.59804761e-01 -1.22948253e+00 -1.12645388e+00
1.79368794e-01 -3.09115469e-01 -1.77171037e-01 1.29431558e+00
5.14311910e-01 -1.00889150e-02 1.77836120e-01 -1.74203181e+00
-1.29951811e+00 8.18226397e-01 -6.60266504e-02 1.54909089e-01
8.99920344e-01 4.50745344e-01 1.05582213e+00 -1.06451547e+00
7.63408005e-01 1.21853840e+00 5.91852784e-01 3.19552988e-01
-7.20984578e-01 -7.43919194e-01 -3.38157922e-01 8.32230985e-01
-1.45381975e+00 -6.40831748e-03 4.16394442e-01 -6.73726082e-01
1.14785302e+00 5.93446672e-01 7.06807911e-01 3.53348762e-01
-1.56673744e-01 9.54199255e-01 4.62178677e-01 -5.79577148e-01
-9.20519009e-02 4.81107146e-01 7.54071891e-01 1.46605563e+00
-5.92047349e-02 -1.46837890e-01 -6.10693932e-01 -2.70258367e-01
8.01423550e-01 -4.60277855e-01 -4.51462448e-01 -1.14908092e-01
-1.00937402e+00 5.41367292e-01 6.01043940e-01 1.27755836e-01
5.28428853e-02 3.33659619e-01 4.68071491e-01 6.98370278e-01
-3.93365711e-01 6.20774269e-01 -6.70806944e-01 -4.17733192e-01
-2.59673804e-01 4.94503558e-01 1.06119025e+00 1.45933700e+00
1.24476731e+00 -1.00976586e-01 -3.92853558e-01 7.74848759e-01
2.60255724e-01 6.07818186e-01 3.59376013e-01 -1.01785111e+00
6.40038311e-01 1.21890295e+00 1.28224462e-01 -1.03147566e+00
-2.41673425e-01 2.09841594e-01 -3.66785496e-01 9.29183587e-02
1.99803531e-01 -1.91309661e-01 -6.77272916e-01 1.35213268e+00
4.95219409e-01 -2.59222567e-01 2.58755028e-01 7.86538839e-01
1.73886728e+00 1.10630119e+00 2.25545973e-01 3.15142632e-01
1.65515053e+00 -1.23243010e+00 -5.94393373e-01 -9.82155800e-02
1.12609923e+00 -3.59466493e-01 1.96783817e+00 2.84657836e-01
-1.15597641e+00 -6.89419866e-01 -8.91584516e-01 -5.59603751e-01
-9.03994441e-01 4.77779210e-01 8.15989316e-01 3.20195794e-01
-1.18172932e+00 8.34816694e-02 -5.35665572e-01 -1.39981464e-01
3.53920400e-01 2.03851372e-01 -2.94336915e-01 -2.38018646e-03
-1.66717720e+00 8.02351117e-01 1.17474651e+00 7.74679184e-02
-6.53196096e-01 -1.30613554e+00 -1.04006803e+00 2.34720588e-01
5.54491460e-01 -9.93750393e-01 1.76661444e+00 -7.66368091e-01
-1.33131909e+00 7.69258618e-01 -1.38568595e-01 -5.33684231e-02
2.20913514e-01 -3.08417588e-01 -5.53980947e-01 2.05773741e-01
7.47914240e-02 5.56432128e-01 3.34252745e-01 -7.92492509e-01
-6.35858059e-01 -1.24998786e-01 5.09614110e-01 2.60353535e-01
-2.53303707e-01 5.08180320e-01 -1.13297021e+00 -3.14432949e-01
-6.19565606e-01 -5.00093699e-01 1.53001323e-01 1.87528264e-02
-4.75326687e-01 -5.77084303e-01 6.83503687e-01 -1.16868329e+00
1.84842789e+00 -1.97951221e+00 -6.05291091e-02 2.31515512e-01
3.35802644e-01 5.91180265e-01 1.62465483e-01 6.32817686e-01
-1.92951977e-01 2.05721200e-01 -1.00779690e-01 4.63633835e-01
2.40996689e-01 1.59417018e-01 -2.50148654e-01 -3.31232578e-01
-3.36772017e-02 1.47466254e+00 -9.32223380e-01 -7.66568542e-01
2.43629813e-02 -7.29381815e-02 -8.14793050e-01 3.55109453e-01
-8.78347993e-01 -2.43609175e-02 -4.90877718e-01 9.58518088e-01
5.64712107e-01 -3.29345316e-01 3.45730856e-02 -9.22984853e-02
-1.39038220e-01 1.78142235e-01 -1.27838159e+00 1.71540260e+00
-4.73347366e-01 7.23633647e-01 -1.93455949e-01 -5.62216043e-01
9.09801245e-01 2.79620051e-01 -5.04298568e-01 -5.76917768e-01
-2.67670453e-01 6.35833740e-02 -3.86269212e-01 -1.18864286e+00
7.42392361e-01 5.14270604e-01 -3.56171340e-01 1.86249822e-01
2.34751806e-01 -6.76206276e-02 5.06937861e-01 6.27182186e-01
9.26560044e-01 2.25599602e-01 5.24522305e-01 -4.23214771e-02
6.43238187e-01 7.05255032e-01 4.11702573e-01 8.48576963e-01
3.32012802e-01 -1.59711972e-01 8.80757153e-01 -3.34789485e-01
-6.82384312e-01 -1.18765748e+00 1.32354692e-01 1.51392186e+00
1.23081384e-02 -9.14911628e-01 -7.74532259e-01 -4.63741332e-01
1.22820392e-01 9.53218699e-01 -1.71447888e-01 -6.83759898e-02
-5.72738767e-01 -1.89336032e-01 1.06098378e+00 7.06216037e-01
8.61337423e-01 -1.59583712e+00 -4.09428090e-01 -9.47098061e-02
-6.38457611e-02 -3.99179101e-01 -1.06236860e-01 -8.44860673e-02
-4.48483199e-01 -1.28432417e+00 -2.34775931e-01 -1.27707970e+00
7.61249781e-01 -2.30568007e-01 1.44645405e+00 4.80822116e-01
-3.72034401e-01 5.29226899e-01 -2.91933179e-01 -5.26931465e-01
-3.33278269e-01 6.15180321e-02 -4.68062699e-01 -8.03882480e-01
4.72210020e-01 -4.33153927e-01 -2.59734720e-01 1.74929932e-01
-7.43851840e-01 6.08080268e-01 6.69796765e-02 6.67790830e-01
3.69687915e-01 6.27697632e-02 3.42124254e-01 -1.13267946e+00
8.25280130e-01 -5.25249302e-01 -1.12993526e+00 7.58952439e-01
-4.09711748e-01 9.47472304e-02 8.54240239e-01 3.32375094e-02
-1.28282380e+00 -3.72490406e-01 -3.77105594e-01 -3.70346874e-01
-1.26988992e-01 1.21970570e+00 -3.61809254e-01 5.32351397e-02
9.98106837e-01 2.73170173e-01 -3.17172199e-01 -4.63278174e-01
6.18417859e-01 8.00340533e-01 1.01794517e+00 -9.95285392e-01
9.13436592e-01 -2.12668598e-01 -8.23155999e-01 -4.76996660e-01
-3.93965065e-01 -2.70294189e-01 -3.08284163e-01 -2.87503839e-01
5.74952781e-01 -8.93505454e-01 -1.33478582e+00 3.07091504e-01
-1.35462523e+00 -6.10082209e-01 -4.36636806e-02 -1.19958751e-01
-3.70434374e-01 2.04260647e-01 -6.97782159e-01 -3.55345070e-01
-6.94696784e-01 -9.81839538e-01 4.06112641e-01 7.41103947e-01
-2.28443921e-01 -9.75791812e-01 1.65075436e-01 4.19499576e-01
1.00967616e-01 -2.34253734e-01 1.49881089e+00 -4.11684304e-01
-8.31827581e-01 1.46480212e-02 -2.94668287e-01 4.22037214e-01
-5.87109149e-01 4.36113119e-01 -6.08127236e-01 1.70409814e-01
-6.87328458e-01 -3.40096653e-01 2.88377941e-01 -1.14614062e-01
1.57325733e+00 -3.02463949e-01 -3.88917208e-01 9.64668214e-01
1.45848215e+00 4.20116127e-01 9.05191720e-01 5.79496861e-01
7.57979751e-01 1.89450786e-01 6.49551213e-01 -2.41422690e-02
7.47667909e-01 4.63978916e-01 1.97312668e-01 1.82428047e-01
3.99518237e-02 -6.70131445e-01 3.95808488e-01 1.03209937e+00
1.68487001e-02 2.33219098e-02 -1.51737022e+00 5.91161549e-01
-2.11403990e+00 -8.28150630e-01 -3.65216702e-01 1.82104754e+00
1.07016432e+00 -2.44708732e-01 1.22606754e-02 -3.00402254e-01
5.99193454e-01 -2.24359512e-01 -5.14128268e-01 -6.35169148e-01
1.14553468e-03 3.42075408e-01 2.03071326e-01 7.52955377e-01
-6.35618687e-01 1.37371433e+00 5.13339472e+00 1.05403697e+00
-8.34509313e-01 8.48842561e-02 -1.42764553e-01 2.11870193e-01
-4.66151357e-01 1.99354798e-01 -8.86838257e-01 4.64740604e-01
7.53286242e-01 -8.61012697e-01 6.33147359e-01 1.29571867e+00
-1.90036654e-01 -3.43594849e-01 -7.87653387e-01 1.19550753e+00
-2.42683873e-01 -1.83127081e+00 1.03933528e-01 -7.58300543e-01
1.68711275e-01 -2.04737514e-01 -2.71120757e-01 1.00755799e+00
7.00610399e-01 -1.00520587e+00 8.14132512e-01 8.33886981e-01
8.49130630e-01 -8.11249375e-01 3.33670735e-01 4.52167004e-01
-1.22754300e+00 -1.64201766e-01 -4.94222254e-01 -1.11934237e-01
-6.48575202e-02 3.36445570e-01 -1.12717140e+00 8.39055061e-01
1.22693455e+00 4.53319311e-01 -9.34556961e-01 1.28531647e+00
-9.41694915e-01 6.10487044e-01 -1.67170569e-01 -3.54625553e-01
-2.70307988e-01 -9.93349552e-02 3.61740857e-01 1.28089309e+00
3.06678414e-01 1.96650535e-01 -1.78059623e-01 1.20694661e+00
-9.74949151e-02 3.34367901e-01 -2.75473446e-01 -1.77218705e-01
7.02546597e-01 1.13438463e+00 5.14665339e-03 -6.24299884e-01
-2.22611412e-01 8.52993309e-01 9.37638223e-01 6.21958613e-01
-1.11930108e+00 -1.03171241e+00 6.61925733e-01 -2.25492612e-01
1.44640207e-01 1.06448932e-02 -2.10989825e-03 -1.24329627e+00
3.03133905e-01 -1.03346419e+00 8.80513549e-01 -1.85013020e+00
-8.95494878e-01 4.42360342e-01 2.05575705e-01 -6.47916734e-01
-5.27451336e-02 -8.13469410e-01 -7.22785234e-01 1.12752211e+00
-1.11737025e+00 -8.87180090e-01 -6.53544903e-01 8.43804657e-01
1.45569012e-01 -1.86581090e-01 1.02271652e+00 5.05144835e-01
-7.63227403e-01 6.49045885e-01 -4.03662175e-02 5.07427335e-01
5.07702589e-01 -1.42534244e+00 3.46851140e-01 5.91268122e-01
-7.89583176e-02 1.02360082e+00 3.64879429e-01 -7.89349198e-01
-1.57490969e+00 -1.02528405e+00 8.33080411e-01 -6.68915093e-01
7.84719169e-01 -1.91174537e-01 -1.22586787e+00 1.17996395e+00
1.02029577e-01 -2.02859372e-01 6.85427368e-01 9.89207067e-03
-4.73629832e-01 -1.15515396e-01 -7.45261908e-01 8.43388319e-01
8.06544423e-01 -8.98893535e-01 -9.90116000e-01 4.89172131e-01
1.06332624e+00 -1.02697515e+00 -8.35858166e-01 1.32104635e-01
4.10663754e-01 -9.60103273e-01 9.09695804e-01 -7.72605479e-01
4.07358557e-01 -1.06207359e+00 2.43344963e-01 -1.16345680e+00
-1.66891843e-01 -4.92723614e-01 -3.09217036e-01 1.31766438e+00
6.52736187e-01 -7.42549539e-01 6.77710354e-01 5.83409786e-01
-3.84193301e-01 -6.97091460e-01 -4.62215364e-01 -5.66716909e-01
-2.75483966e-01 -1.11599398e+00 8.42117786e-01 9.03098285e-01
5.35161674e-01 2.58493453e-01 -1.15822069e-01 5.08150339e-01
-1.20980725e-01 7.24404529e-02 1.06495309e+00 -8.91118109e-01
-6.27115190e-01 -2.26150453e-01 -1.07860148e-01 -1.09334838e+00
-1.40521657e-02 -1.38345182e+00 -4.10281241e-01 -1.74777210e+00
-6.29715770e-02 -5.30667067e-01 1.11136951e-01 9.45365787e-01
-2.00897843e-01 -3.25316846e-01 1.63791209e-01 1.92205563e-01
-6.15931869e-01 1.43589288e-01 1.06206107e+00 -2.00805634e-01
-3.67841572e-01 -3.59043241e-01 -7.46673346e-01 7.67640054e-01
9.70464706e-01 -4.37085539e-01 -5.15683234e-01 -8.50959480e-01
1.09630716e+00 1.71852037e-01 7.49452829e-01 -9.85360622e-01
7.38591313e-01 -1.55481175e-01 1.41572312e-01 -7.26513743e-01
-2.14650296e-02 -4.13019121e-01 4.33915704e-01 3.44803780e-01
-8.61540362e-02 5.06155670e-01 6.13552570e-01 1.33217126e-01
-3.41895550e-01 -7.32118785e-01 1.77520677e-01 -4.04878765e-01
-1.46652424e+00 1.32786438e-01 -2.33862996e-01 2.40065426e-01
1.03798556e+00 -1.67644590e-01 -9.43262994e-01 -1.19629659e-01
-6.23343229e-01 9.49490726e-01 2.14381814e-01 2.30794653e-01
4.89352435e-01 -1.16765761e+00 -7.34357536e-02 6.71745688e-02
5.04368544e-01 3.17723453e-01 4.62300658e-01 6.04260981e-01
-1.67613947e+00 4.59420890e-01 -1.60299689e-01 -1.05304822e-01
-1.39806163e+00 5.05769849e-01 4.70054597e-01 -1.37319535e-01
-4.71505851e-01 1.21561325e+00 -9.74371284e-02 -1.07912445e+00
3.86115402e-01 -4.73544180e-01 -1.94066435e-01 -2.11905435e-01
8.97593915e-01 3.48208696e-01 1.81357972e-02 1.66375250e-01
-3.74640226e-01 2.28736743e-01 3.17352414e-02 2.62098014e-01
1.08384931e+00 3.88527006e-01 -6.04884386e-01 2.67890096e-01
9.56967175e-01 -1.49943167e-02 -3.54552537e-01 2.67611787e-04
-5.79131357e-02 -2.74047911e-01 -4.46222275e-01 -1.18669045e+00
-6.21397078e-01 6.60294294e-01 9.61424857e-02 3.00688148e-02
1.01012981e+00 3.09313834e-01 5.53631365e-01 1.06692708e+00
3.65392983e-01 -9.05932307e-01 -2.85290599e-01 8.48949015e-01
1.14733601e+00 -9.07420874e-01 -1.61463976e-01 -6.16152942e-01
-6.06998265e-01 1.16853237e+00 1.06363618e+00 1.73743531e-01
5.00009537e-01 1.93914950e-01 2.19936818e-01 -5.09384453e-01
-8.11836243e-01 -2.13880092e-03 -2.62646060e-02 8.42517674e-01
6.47311984e-03 1.00903369e-01 -9.13407952e-02 1.00360823e+00
-7.86845505e-01 4.72018063e-01 3.53502154e-01 1.01430368e+00
-3.04287076e-01 -9.39181805e-01 -5.07499576e-01 4.01199490e-01
1.56565718e-02 -5.29637575e-01 -6.52139336e-02 1.10209763e+00
2.28678480e-01 3.62140864e-01 1.07815713e-02 -3.64407986e-01
6.60568058e-01 3.97471726e-01 1.48817882e-01 -7.65503526e-01
-9.25410509e-01 -9.78843570e-01 4.14789498e-01 -4.46853995e-01
3.22607785e-01 7.53928795e-02 -1.64608312e+00 -4.79813874e-01
-2.24851575e-02 4.94827658e-01 2.64525145e-01 4.56480742e-01
5.51963806e-01 3.40311438e-01 -4.04964983e-01 -5.54895261e-03
2.76095625e-02 -5.67744970e-01 -2.86438495e-01 -9.87102762e-02
-8.04795548e-02 -3.37107837e-01 1.76603720e-03 2.84752041e-01] | [9.81902027130127, 7.63580322265625] |
20ea5327-53b0-4c87-923e-d20899fbb495 | ccmi-classifier-based-conditional-mutual | 1906.01824 | null | https://arxiv.org/abs/1906.01824v1 | https://arxiv.org/pdf/1906.01824v1.pdf | CCMI : Classifier based Conditional Mutual Information Estimation | Conditional Mutual Information (CMI) is a measure of conditional dependence between random variables X and Y, given another random variable Z. It can be used to quantify conditional dependence among variables in many data-driven inference problems such as graphical models, causal learning, feature selection and time-series analysis. While k-nearest neighbor (kNN) based estimators as well as kernel-based methods have been widely used for CMI estimation, they suffer severely from the curse of dimensionality. In this paper, we leverage advances in classifiers and generative models to design methods for CMI estimation. Specifically, we introduce an estimator for KL-Divergence based on the likelihood ratio by training a classifier to distinguish the observed joint distribution from the product distribution. We then show how to construct several CMI estimators using this basic divergence estimator by drawing ideas from conditional generative models. We demonstrate that the estimates from our proposed approaches do not degrade in performance with increasing dimension and obtain significant improvement over the widely used KSG estimator. Finally, as an application of accurate CMI estimation, we use our best estimator for conditional independence testing and achieve superior performance than the state-of-the-art tester on both simulated and real data-sets. | ['Sudipto Mukherjee', 'Himanshu Asnani', 'Sreeram Kannan'] | 2019-06-05 | null | null | null | null | ['mutual-information-estimation'] | ['methodology'] | [ 9.31375250e-02 -3.96780193e-01 -4.50037509e-01 -5.71467459e-01
-8.51869047e-01 -2.45750338e-01 6.29035473e-01 1.90182328e-01
-1.75694153e-01 9.86520886e-01 -1.27496436e-01 -4.01157886e-01
-5.92266560e-01 -8.95563722e-01 -5.12714446e-01 -8.20119977e-01
-4.30256993e-01 3.87203962e-01 3.36254053e-02 4.29837018e-01
3.57590258e-01 3.61728013e-01 -1.45156956e+00 -4.55317795e-01
1.09462738e+00 9.22153294e-01 -1.54320315e-01 7.09835231e-01
1.71847492e-01 6.42549038e-01 -5.49606144e-01 -3.21304739e-01
-3.04991424e-01 -7.99428582e-01 -5.47820151e-01 -3.42141122e-01
-5.31342402e-02 -7.34665617e-02 -6.58653453e-02 9.41165924e-01
1.26276881e-01 -4.95170876e-02 1.52254415e+00 -1.61579263e+00
-5.63715816e-01 6.34056211e-01 -7.68257678e-01 1.54760525e-01
2.76506484e-01 -4.70279276e-01 1.03020489e+00 -8.23375523e-01
1.92153275e-01 1.18600035e+00 6.71621323e-01 1.01098411e-01
-1.44528592e+00 -1.03408134e+00 -2.18675300e-01 2.66581655e-01
-1.63796604e+00 -5.45941247e-03 4.68398839e-01 -5.30241787e-01
5.73095739e-01 2.61487603e-01 3.89634877e-01 8.50142241e-01
5.64643383e-01 9.06210721e-01 1.09206533e+00 -5.52812159e-01
3.48249197e-01 2.04275146e-01 1.89695165e-01 5.63771605e-01
3.53555679e-01 2.45277911e-01 -6.57663882e-01 -5.89741349e-01
8.32482338e-01 2.28916138e-01 -2.27799758e-01 -4.83462930e-01
-9.39235330e-01 1.19837797e+00 1.00280844e-01 2.90766925e-01
-1.12894602e-01 4.19986546e-01 -7.30148107e-02 2.28589118e-01
4.38606679e-01 -7.50561655e-02 -4.14052457e-01 -2.46350303e-01
-9.16202724e-01 2.10471004e-01 1.08964980e+00 8.21456075e-01
6.76736057e-01 -2.50062317e-01 -1.23328947e-01 7.25895345e-01
6.53361619e-01 7.04519868e-01 2.64724195e-01 -5.35676420e-01
2.12613210e-01 2.70378530e-01 -7.36778304e-02 -1.03645754e+00
-1.78878918e-01 -4.67302173e-01 -1.12064290e+00 -1.10785015e-01
4.03530151e-01 -9.76673886e-02 -7.13555336e-01 1.93197978e+00
2.50939131e-01 3.33786577e-01 -1.88519418e-01 4.84832704e-01
1.53069198e-01 5.47817290e-01 2.09767148e-02 -4.04362440e-01
8.10441971e-01 -1.74361780e-01 -7.00560868e-01 1.55957237e-01
6.21345043e-01 -5.79188704e-01 4.62231934e-01 4.30033743e-01
-6.52983010e-01 -1.23465419e-01 -1.03731263e+00 6.43722475e-01
-1.56239361e-01 -9.75276828e-02 8.64862621e-01 7.65612543e-01
-6.82853580e-01 6.76760972e-01 -9.98846889e-01 -3.02795768e-01
2.91446149e-01 1.18216574e-01 -3.25705796e-01 -8.08738768e-02
-1.18682313e+00 6.76129222e-01 2.69845992e-01 -1.55804157e-01
-8.11740935e-01 -8.19391727e-01 -7.01246619e-01 1.27414390e-01
1.80747002e-01 -4.92482692e-01 9.74665821e-01 -3.60726953e-01
-1.33256447e+00 3.17003042e-01 -1.80384845e-01 -4.85385716e-01
3.32603127e-01 -2.15504646e-01 -4.46047544e-01 -2.02987537e-01
1.37557015e-01 3.95316333e-02 8.52349579e-01 -7.88595796e-01
-5.13598084e-01 -4.44117159e-01 -5.77766299e-01 -3.37812454e-01
-2.05082282e-01 -1.78107843e-01 -1.67213321e-01 -5.89278519e-01
4.47984695e-01 -8.87547672e-01 -8.69848803e-02 -1.01894960e-01
-6.89801037e-01 -2.86480486e-01 7.34917223e-01 -5.00716090e-01
1.38662183e+00 -2.13763285e+00 2.83980817e-02 5.63146532e-01
1.40281841e-01 -2.07436651e-01 3.73169005e-01 5.81088662e-01
-1.97794959e-01 3.19931507e-02 -6.03687286e-01 5.82271069e-02
-1.33281454e-01 5.12945168e-02 -7.47435242e-02 8.47348690e-01
3.19173306e-01 4.95602369e-01 -8.73182356e-01 -4.25539732e-01
1.86551541e-01 4.91977721e-01 -6.08115017e-01 2.75465161e-01
2.01160699e-01 3.71658772e-01 -2.81261176e-01 4.57110316e-01
6.36268497e-01 -4.89851117e-01 3.50959957e-01 2.81516998e-03
1.62517160e-01 4.35067564e-02 -1.30819988e+00 1.23006952e+00
-4.84966069e-01 7.35355973e-01 -5.05141020e-01 -1.43732893e+00
1.04249525e+00 2.31824085e-01 4.15460467e-01 -2.53359765e-01
1.36291031e-02 2.19516471e-01 1.16675813e-02 -1.15285203e-01
-1.18212275e-01 -1.92575082e-01 -3.62996489e-01 2.91428387e-01
2.98848510e-01 4.14250046e-02 1.87430102e-02 2.68904388e-01
1.13587332e+00 -1.06051743e-01 8.20434570e-01 -4.25497234e-01
3.62765282e-01 -6.32300556e-01 5.28349340e-01 9.97099757e-01
1.17464796e-01 3.97337258e-01 6.89420879e-01 1.70616090e-01
-6.18433177e-01 -1.60382366e+00 -6.86105132e-01 5.91207922e-01
-2.91930195e-02 -3.86983216e-01 -3.74643892e-01 -4.79448110e-01
3.70880872e-01 9.29944873e-01 -6.66363478e-01 -2.34078139e-01
8.23562518e-02 -1.10545671e+00 3.97141576e-01 6.36887252e-01
3.79595965e-01 -2.67956585e-01 -3.32525313e-01 1.44403145e-01
-3.71083952e-02 -5.86046576e-01 -1.79829940e-01 5.15224755e-01
-8.82870078e-01 -1.13046694e+00 -6.87632263e-01 -1.65802434e-01
4.24825400e-01 2.10106783e-02 1.03807521e+00 -3.77607733e-01
-4.85117406e-01 4.69887108e-01 -1.85314924e-01 -3.66065472e-01
-4.35326964e-01 -1.55291900e-01 1.25637621e-01 5.30538298e-02
4.80833650e-01 -8.41875732e-01 -5.30787647e-01 5.79913676e-01
-8.00162852e-01 -2.47601047e-01 7.25457370e-01 9.69077051e-01
5.37473083e-01 1.94066778e-01 6.42776430e-01 -8.70253563e-01
5.43603301e-01 -8.70105207e-01 -7.47900844e-01 2.79250383e-01
-9.45662737e-01 5.67018509e-01 2.05677852e-01 -3.97423267e-01
-1.00169599e+00 -2.54914910e-01 2.78937936e-01 -2.06713736e-01
3.23294811e-02 9.07130063e-01 -1.72883853e-01 1.96620107e-01
3.78713697e-01 2.05272257e-01 -7.86783472e-02 -5.02083302e-01
2.62314498e-01 8.12442362e-01 5.73826194e-01 -3.67230445e-01
5.98883569e-01 2.74237871e-01 4.24062461e-01 -7.36904919e-01
-8.38588715e-01 -5.80465078e-01 -6.40568256e-01 -4.00115661e-02
7.75440574e-01 -7.39695966e-01 -9.74348068e-01 4.09743071e-01
-8.35423768e-01 1.39432594e-01 2.43871674e-01 8.67693603e-01
-5.88509500e-01 1.31576464e-01 -3.86580467e-01 -1.35155499e+00
8.59591179e-04 -6.47596061e-01 8.67561519e-01 2.74463892e-01
-2.44980350e-01 -1.35221660e+00 4.12316352e-01 -1.14390031e-01
1.83478996e-01 3.31341565e-01 1.03032458e+00 -7.15613663e-01
-2.81330317e-01 -5.06661475e-01 -2.71960407e-01 2.65030473e-01
4.39980239e-01 2.56221384e-01 -5.71378708e-01 -2.00190604e-01
-1.50419906e-01 -1.16883079e-02 9.65139508e-01 6.14259183e-01
1.10667741e+00 -1.65021703e-01 -7.07284927e-01 4.09535706e-01
1.58301914e+00 2.14861363e-01 5.71211100e-01 -1.76897883e-01
4.76510704e-01 3.25606823e-01 7.34219551e-01 8.51054728e-01
2.49920696e-01 5.20301223e-01 1.33630380e-01 3.02199960e-01
3.07409972e-01 -5.07789135e-01 2.43211418e-01 7.61213660e-01
1.21853940e-01 -2.60439724e-01 -9.26700056e-01 4.49827015e-01
-2.00814128e+00 -1.04515815e+00 -4.22151744e-01 2.66867709e+00
1.00738621e+00 -1.29032796e-02 1.77036822e-01 2.56823093e-01
7.52822697e-01 -4.15160149e-01 -6.46663725e-01 6.98507801e-02
1.15603916e-01 2.52542615e-01 7.24667192e-01 3.36692154e-01
-1.06678843e+00 4.06254023e-01 7.07766867e+00 8.62428963e-01
-8.42106640e-01 4.21132371e-02 5.62039196e-01 1.58110887e-01
-1.78892404e-01 2.33316839e-01 -9.29369509e-01 6.45371497e-01
1.22231317e+00 -3.38438064e-01 6.55180134e-04 7.79209495e-01
-1.33602560e-01 -5.98058105e-01 -1.25514627e+00 1.12465227e+00
-4.78731245e-02 -9.64727998e-01 -3.57908607e-01 2.82813489e-01
8.11353683e-01 -2.43651465e-01 8.40004459e-02 7.23994970e-02
5.66046417e-01 -1.16132700e+00 3.91241491e-01 8.40280175e-01
6.06623650e-01 -9.90096092e-01 7.36754417e-01 3.16023201e-01
-9.96702969e-01 1.14397623e-01 -2.50753909e-01 -3.51272784e-02
-8.15288723e-02 1.37500978e+00 -8.80045414e-01 5.15643895e-01
5.96209347e-01 7.27880418e-01 -2.82988876e-01 1.27755749e+00
-2.50057548e-01 1.06010866e+00 -4.58040059e-01 -3.31632912e-01
-1.74041316e-01 -2.02321574e-01 4.26042020e-01 1.07265985e+00
5.14345169e-01 -2.09555760e-01 -1.92169756e-01 9.96835589e-01
2.91276544e-01 -1.69957832e-01 -6.11870170e-01 -1.66235548e-02
7.13772833e-01 8.59406173e-01 -6.64953768e-01 -1.95786223e-01
-4.31744605e-01 8.94713581e-01 2.59637177e-01 3.34220320e-01
-9.74041343e-01 -5.09332895e-01 7.15776384e-01 -4.60159630e-02
5.66601038e-01 -3.82376581e-01 -2.63900701e-02 -1.11217713e+00
-1.17663249e-01 -4.17693764e-01 5.14968872e-01 -3.15249711e-01
-1.54786289e+00 1.75515357e-02 4.76474792e-01 -1.25485396e+00
-9.01155531e-01 -5.01557410e-01 -5.17328858e-01 9.14838135e-01
-1.22302663e+00 -4.86320078e-01 -4.70221741e-03 5.44217944e-01
-9.06941593e-02 3.17157023e-02 8.53567481e-01 1.53321609e-01
-5.71905077e-01 6.49353683e-01 6.52242422e-01 1.44263819e-01
7.40833163e-01 -1.32269430e+00 8.57886299e-02 6.65609479e-01
2.57799387e-01 6.89041197e-01 7.86182880e-01 -7.34731317e-01
-1.31942844e+00 -7.98194766e-01 6.19940221e-01 -2.60318995e-01
8.44338477e-01 -3.27663988e-01 -7.04267859e-01 6.01187289e-01
-4.10596430e-01 2.49414872e-02 1.01534522e+00 4.84734744e-01
-3.70777458e-01 -4.97658737e-02 -1.02678835e+00 3.23836595e-01
7.79624581e-01 -3.30920160e-01 -2.09686771e-01 1.30556598e-01
2.58650184e-01 1.26681417e-01 -1.15605640e+00 3.54577392e-01
7.41817117e-01 -1.30781472e+00 7.88983583e-01 -5.01813054e-01
2.62140512e-01 -2.10781276e-01 -3.08018655e-01 -1.28870142e+00
-2.87373424e-01 -4.14121896e-01 -1.66684628e-01 1.33010387e+00
5.40742695e-01 -8.16618800e-01 5.06298661e-01 5.23901284e-01
6.19986176e-01 -5.64744651e-01 -1.04540992e+00 -1.13986266e+00
1.55045372e-02 -7.74721742e-01 2.96595037e-01 8.13005030e-01
2.55426943e-01 3.77405226e-01 -4.49654281e-01 1.68885022e-01
1.10304463e+00 1.99032485e-01 7.44457126e-01 -1.58877850e+00
-5.83653867e-01 -2.37067103e-01 -8.77410233e-01 -8.69005978e-01
9.26473737e-02 -7.43570268e-01 1.38430104e-01 -1.07962751e+00
6.84262931e-01 -4.54792976e-01 -3.79284084e-01 1.88036218e-01
-3.49908829e-01 -9.40539234e-04 -2.34577388e-01 2.40510851e-01
-2.69991547e-01 5.17180860e-01 6.42147243e-01 4.97808196e-02
-1.33457392e-01 2.46981308e-01 -4.55124080e-01 6.67243898e-01
6.03215456e-01 -7.53526807e-01 -5.25717199e-01 2.81225175e-01
2.31164023e-01 5.19303977e-01 5.19176424e-01 -9.95779097e-01
1.17567822e-01 -2.06404001e-01 5.78700542e-01 -7.02493787e-01
-4.18315083e-02 -5.12752175e-01 3.00778806e-01 4.42087591e-01
-2.70655453e-01 -2.76730239e-01 -6.35569617e-02 9.48396325e-01
-3.57348770e-01 -2.15596363e-01 7.39576817e-01 3.30123097e-01
-1.95536152e-01 1.70637563e-01 -4.59226370e-01 -3.71275656e-02
8.90852273e-01 1.33857056e-01 1.00301310e-01 -8.10177565e-01
-5.20566881e-01 9.21632443e-03 3.62612642e-02 7.38001689e-02
7.05979109e-01 -1.46734416e+00 -7.63673604e-01 1.74893335e-01
1.48278624e-01 -5.16227782e-01 -5.79365268e-02 1.13107443e+00
-2.42509916e-02 5.43792486e-01 2.09879160e-01 -8.30860019e-01
-1.02197349e+00 3.00858051e-01 -4.98794690e-02 -4.69135791e-01
-2.12616593e-01 7.77428925e-01 3.25165480e-01 -2.09879979e-01
7.68964179e-04 -2.24884227e-01 1.57525629e-01 -1.55726999e-01
5.44806302e-01 4.74055678e-01 -1.46448627e-01 5.09734033e-04
-6.36882246e-01 3.14770728e-01 -6.99547119e-03 -3.59468490e-01
1.14768052e+00 -8.57970119e-02 -6.36613443e-02 9.31757212e-01
1.58832729e+00 -2.21172199e-01 -8.98032248e-01 -1.73936933e-01
1.78260967e-01 -5.84327757e-01 1.49726301e-01 -6.93034768e-01
-7.61601448e-01 7.63292015e-01 6.25504375e-01 4.90390539e-01
1.01188135e+00 1.76910996e-01 2.55733639e-01 8.70249271e-02
4.60041136e-01 -7.71129429e-01 -3.91117707e-02 1.63915411e-01
4.50765401e-01 -1.26356411e+00 7.01650754e-02 -5.04538655e-01
-2.92931348e-01 1.03051722e+00 1.26960963e-01 -1.12158962e-01
1.23791349e+00 4.06067610e-01 -4.40935761e-01 8.22322965e-02
-7.82267511e-01 -8.51079971e-02 4.54709142e-01 7.43718982e-01
7.50192046e-01 3.29848677e-01 -3.72908562e-01 4.50838745e-01
-1.53855354e-01 -4.84817959e-02 2.91304618e-01 7.49858856e-01
-2.46313378e-01 -1.04286063e+00 -2.99048662e-01 7.55551219e-01
-3.85497481e-01 -1.77230805e-01 -5.37315197e-02 8.75933349e-01
-5.72516978e-01 9.24115598e-01 2.77911156e-01 -2.13693440e-01
-2.87519217e-01 2.84836460e-02 5.95987260e-01 -2.05080256e-01
3.37954372e-01 -5.41145615e-02 -1.55161455e-01 -4.49406475e-01
-4.92225409e-01 -1.02990377e+00 -9.37901080e-01 -3.52148771e-01
-7.54077554e-01 2.30111480e-01 8.15743268e-01 1.08325458e+00
1.41930103e-01 3.35745960e-01 8.43338430e-01 -3.90086412e-01
-5.98844290e-01 -9.45949376e-01 -9.82570767e-01 6.98644072e-02
9.75182950e-02 -9.98469949e-01 -6.72098100e-01 -1.78242505e-01] | [7.4122633934021, 4.268125534057617] |
191f8a40-23b4-4cb8-94e1-7eb8e31ebb18 | dimsum-distributed-and-multilingual | null | null | https://aclanthology.org/2022.fnp-1.9 | https://aclanthology.org/2022.fnp-1.9.pdf | DiMSum: Distributed and Multilingual Summarization of Financial Narratives | This paper was submitted for Financial Narrative Summarization (FNS) task in FNP-2022 workshop. The objective of the task was to generate not more than 1000 words summaries for the annual financial reports written in English, Spanish and Greek languages. The central idea of this paper is to demonstrate automatic ways of identifying key narrative sections and their contributions towards generating summaries of financial reports. We have observed a few limitations in the previous works: First, the complete report was being considered for summary generation instead of key narrative sections. Second, many of the works followed manual or heuristic-based techniques to identify narrative sections. Third, sentences from key narrative sections were abruptly dropped to limit the summary to the desired length. To overcome these shortcomings, we introduced a novel approach to automatically learn key narrative sections and their weighted contributions to the reports. Since the summaries may come from various parts of the reports, the summary generation process was distributed amongst the key narrative sections based on the weights identified, later combined to have an overall summary. We also showcased that our approach is adaptive to various report formats and languages. | ['Msp Raja', 'Sangeeth Keeriyadath', 'Raghu Katikeri', 'Amit Vaid', 'Neelesh Shukla'] | null | null | null | null | fnp-lrec-2022-6 | ['document-ai'] | ['natural-language-processing'] | [ 2.20727831e-01 5.02484083e-01 -9.99037474e-02 -7.63047263e-02
-1.18411517e+00 -9.67678010e-01 1.03107655e+00 5.94128370e-01
-1.80034742e-01 1.35005581e+00 1.11552906e+00 -9.99593511e-02
-1.68946996e-01 -6.32915318e-01 -1.99901104e-01 -1.27416745e-01
5.47237583e-02 1.53117478e-01 1.68264136e-01 -1.19497649e-01
1.00719666e+00 3.97526920e-01 -1.27609313e+00 7.38384604e-01
8.92848730e-01 4.90945011e-01 2.29236901e-01 1.08212256e+00
-7.04666555e-01 1.29417813e+00 -1.45772266e+00 -6.12406671e-01
-1.44483745e-01 -8.56096447e-01 -9.26856935e-01 1.61975935e-01
7.16752335e-02 -1.64659843e-01 -1.13626927e-01 6.89468861e-01
3.52231264e-01 1.62507072e-01 1.01805747e+00 -1.11965215e+00
-4.64497477e-01 1.28422725e+00 -3.63847911e-01 4.12996382e-01
8.77501190e-01 -3.83063585e-01 1.08545816e+00 -9.46110308e-01
8.89678538e-01 1.04672492e+00 5.31421721e-01 5.07989407e-01
-7.57741868e-01 -4.91877347e-01 3.24227124e-01 -7.19157606e-02
-9.58115339e-01 -5.20343184e-01 8.25192153e-01 -7.41495013e-01
1.21706259e+00 5.20101309e-01 7.41483688e-01 1.01396000e+00
2.91762382e-01 6.25333607e-01 7.59814024e-01 -5.97413599e-01
2.54618466e-01 4.21830744e-01 2.55364567e-01 2.22111851e-01
4.79673713e-01 -5.88277102e-01 -7.30941355e-01 -4.64884639e-01
1.75924182e-01 -6.57499373e-01 -8.00260305e-02 5.70220292e-01
-1.39026129e+00 9.36551809e-01 -2.85470843e-01 4.86972302e-01
-5.86542308e-01 -1.73036173e-01 8.90848935e-01 4.66938540e-02
8.05597603e-01 7.02100933e-01 -1.26827359e-01 -3.37869167e-01
-1.53211021e+00 8.80284965e-01 1.14755952e+00 9.97754455e-01
2.55158097e-01 6.91512004e-02 -6.62143528e-01 5.99999428e-01
-1.39100254e-01 6.72121868e-02 4.51868087e-01 -9.43268538e-01
1.26805437e+00 6.81627989e-01 5.80416143e-01 -1.21144497e+00
-3.08624566e-01 -1.92953706e-01 -6.86095893e-01 -9.30973589e-02
9.72197801e-02 -5.19486129e-01 -4.43234146e-01 1.07506680e+00
-2.30436549e-01 -5.08466482e-01 5.13960719e-01 4.67833817e-01
1.43269944e+00 1.10635841e+00 -6.52565584e-02 -7.55787849e-01
1.25494564e+00 -7.82237291e-01 -1.11577570e+00 -3.75820398e-02
4.80474293e-01 -1.13663113e+00 6.45074129e-01 3.30013871e-01
-1.63333678e+00 -2.84790665e-01 -1.34266901e+00 1.48927808e-01
-3.54891032e-01 3.34761471e-01 2.78949916e-01 3.23221236e-01
-9.29556012e-01 6.00250602e-01 -3.67617846e-01 -2.48185754e-01
2.54578441e-01 -1.77311704e-01 -8.83071646e-02 4.08787161e-01
-1.09411466e+00 9.18056309e-01 6.99260354e-01 -2.97752857e-01
-1.67589039e-01 -8.44288468e-01 -7.87028134e-01 1.36814406e-02
3.53680164e-01 -6.87813163e-01 1.47239006e+00 -4.54261303e-01
-1.25332856e+00 6.72484994e-01 -3.80840778e-01 -5.98156929e-01
7.07489312e-01 -2.33054444e-01 -3.39487433e-01 3.56447875e-01
2.81067699e-01 1.97391823e-01 2.49385014e-01 -1.11963677e+00
-8.48619640e-01 2.18130857e-01 2.57180184e-02 2.85610259e-01
-2.75180846e-01 6.41468644e-01 -8.94117821e-03 -1.14766586e+00
-1.71935588e-01 -3.11787039e-01 -1.55769899e-01 -1.00433910e+00
-6.49994314e-01 -4.51119095e-01 4.67923284e-01 -1.02714705e+00
1.84269547e+00 -1.66837907e+00 1.53673038e-01 -1.71086669e-01
1.40594482e-01 -4.14102338e-03 1.08734973e-01 1.13304937e+00
1.13940462e-01 5.58647275e-01 -3.44102472e-01 -2.36487627e-01
8.69675055e-02 -2.93805063e-01 -8.11018288e-01 -8.73454064e-02
4.46596533e-01 6.65399253e-01 -1.01291692e+00 -7.42506683e-01
-2.00644314e-01 6.62807077e-02 -1.15843713e-01 1.68286175e-01
-1.61887228e-01 4.08682190e-02 -4.50191081e-01 4.18615133e-01
4.30975556e-01 2.41450518e-01 -1.74201623e-01 -1.49605423e-01
-6.22200906e-01 9.01638448e-01 -1.17361450e+00 1.24653304e+00
-3.18628728e-01 7.82210350e-01 -3.85194510e-01 -9.46363151e-01
1.17493510e+00 6.98413730e-01 4.52659398e-01 -2.76646167e-01
-9.70711336e-02 3.84851813e-01 -3.73586923e-01 -4.77179766e-01
1.26873970e+00 -2.09538162e-01 -5.42133272e-01 8.55467677e-01
-1.88324433e-02 -7.68799126e-01 1.04602253e+00 5.55958986e-01
8.87431741e-01 7.68512627e-03 7.92078018e-01 9.29986592e-03
6.97309732e-01 5.01108587e-01 4.35803652e-01 8.11109126e-01
2.37187847e-01 1.17642784e+00 9.83904541e-01 -4.20307994e-01
-1.18256700e+00 -7.47584105e-01 2.43824899e-01 4.74759549e-01
-5.01916170e-01 -1.02634096e+00 -8.46618593e-01 -7.14497924e-01
-4.25103188e-01 1.15383828e+00 -5.45914948e-01 1.95632905e-01
-7.05781937e-01 -6.56096578e-01 5.68660915e-01 4.01382923e-01
1.21642552e-01 -1.42876732e+00 -8.73310804e-01 6.19491041e-01
-5.33702314e-01 -9.87522781e-01 -5.98915458e-01 -3.81321013e-02
-6.56603456e-01 -1.03492105e+00 -1.05144930e+00 -4.06458288e-01
5.31363130e-01 -1.03618570e-01 1.03526092e+00 -3.15531820e-01
-1.61722843e-02 2.48959780e-01 -5.59938908e-01 -7.86459923e-01
-9.23522234e-01 3.54198843e-01 -1.23809040e-01 -4.80075896e-01
-9.31841955e-02 -1.78876400e-01 -3.90155204e-02 -4.55143690e-01
-8.32641602e-01 1.94419339e-01 3.04091156e-01 6.05407715e-01
2.65834719e-01 1.92997411e-01 1.02788663e+00 -9.77500021e-01
1.43567991e+00 -4.21167254e-01 -1.50668100e-01 3.50874245e-01
-1.90920785e-01 -5.57303876e-02 8.04828584e-01 -3.42952281e-01
-1.19908655e+00 -4.34244901e-01 9.80028231e-03 4.57643241e-01
1.58768237e-01 9.55096006e-01 1.10226728e-01 8.10329676e-01
4.41461921e-01 2.50144333e-01 -3.47333342e-01 -1.98853314e-01
1.15622975e-01 6.99455142e-01 4.17845190e-01 -3.92252833e-01
6.53008103e-01 -1.39573604e-01 -3.34795356e-01 -7.19975650e-01
-9.13792908e-01 -2.64237195e-01 -6.66136742e-01 -4.25555646e-01
5.27835965e-01 -8.17225575e-01 -1.06592365e-01 1.09734766e-01
-1.75693524e+00 4.00113128e-02 -6.13774598e-01 4.51311469e-01
-4.89541203e-01 2.85947889e-01 -4.91719425e-01 -1.05171192e+00
-7.49161839e-01 -6.52874649e-01 5.54922342e-01 4.53261942e-01
-1.19158554e+00 -9.72738922e-01 1.78790972e-01 7.74215311e-02
9.52659696e-02 6.20306015e-01 8.33301306e-01 -7.84611583e-01
1.94598362e-01 -3.40648741e-01 -1.65158004e-01 1.55071348e-01
4.00864542e-01 5.23994625e-01 -5.90502083e-01 1.33839026e-01
1.91465952e-02 -1.59788892e-01 8.17575932e-01 3.55683655e-01
6.36406124e-01 -1.06937957e+00 -1.53118014e-01 -1.34414852e-01
1.02958226e+00 4.72082883e-01 4.92450267e-01 5.99223256e-01
2.73000628e-01 8.32089126e-01 7.67563403e-01 7.14576721e-01
4.98571575e-01 3.05462748e-01 -2.15692431e-01 3.83476615e-01
-1.18226200e-01 -1.68783993e-01 7.16618836e-01 1.20201230e+00
-5.58544286e-02 -4.66284961e-01 -7.46046901e-01 9.82978046e-01
-1.76802182e+00 -1.31207359e+00 -2.89408296e-01 1.82676792e+00
9.23446357e-01 4.92716879e-01 3.24392438e-01 3.05835277e-01
5.72297096e-01 4.37102109e-01 5.62209561e-02 -1.02659285e+00
-3.88749063e-01 1.09893773e-02 2.31437609e-01 3.92428398e-01
-8.26304793e-01 3.28026205e-01 6.85482740e+00 5.36853433e-01
-6.51488423e-01 -2.22577304e-01 4.95881498e-01 -2.90198803e-01
-6.25221252e-01 -3.44206169e-02 -8.89733255e-01 4.09928888e-01
1.14807487e+00 -1.08503771e+00 -2.86468208e-01 4.46702540e-01
5.59068084e-01 -3.34128588e-01 -8.00694644e-01 4.02780741e-01
2.26094738e-01 -1.86337578e+00 4.61738050e-01 -2.79518574e-01
1.02209139e+00 -5.75197697e-01 -3.79566163e-01 4.83605154e-02
3.92960638e-01 -8.10168803e-01 1.28771508e+00 6.76995099e-01
6.34662509e-01 -1.15364683e+00 1.04980910e+00 3.63772750e-01
-1.09659290e+00 5.72881140e-02 -3.86312068e-01 -3.08169514e-01
4.34681147e-01 5.71281970e-01 -9.89870667e-01 1.10896027e+00
3.52777630e-01 7.21793711e-01 -5.00594556e-01 8.78432155e-01
-1.74429849e-01 7.29914367e-01 8.41146559e-02 -4.62722987e-01
4.35111523e-01 -1.04116902e-01 8.57945502e-01 1.75871158e+00
7.45553792e-01 2.17066690e-01 5.41565008e-02 8.43649209e-01
-7.17864484e-02 2.74511904e-01 -5.67731500e-01 -4.38641906e-01
5.17755389e-01 1.11490333e+00 -1.01960337e+00 -5.80999792e-01
-2.01955467e-01 5.36810994e-01 -6.09332360e-02 1.74954101e-01
-5.76797366e-01 -8.29161584e-01 2.06109472e-02 3.08535129e-01
1.53469309e-01 -2.80754548e-02 -7.25955009e-01 -9.39633906e-01
3.19048285e-01 -6.00606322e-01 4.74859089e-01 -8.25343370e-01
-9.82218146e-01 8.65232468e-01 4.47752863e-01 -1.27935934e+00
-5.46694994e-01 1.14474133e-01 -1.13834333e+00 1.05753303e+00
-1.05535948e+00 -7.03378499e-01 6.14040121e-02 -3.81805688e-01
9.55022812e-01 -5.93705177e-01 8.40306878e-01 -1.98619202e-01
-3.23457122e-01 2.48991758e-01 -8.11802596e-02 8.51619169e-02
6.10609770e-01 -1.38696349e+00 5.83488166e-01 1.05232882e+00
-1.48294255e-01 4.83724385e-01 1.13923252e+00 -8.86847854e-01
-6.08026445e-01 -8.70936632e-01 1.81112313e+00 -2.34119490e-01
9.71258879e-01 -1.93753287e-01 -7.96122491e-01 2.63053924e-01
8.38065684e-01 -8.58270586e-01 5.41845500e-01 -3.72596383e-01
8.24386477e-02 1.41436100e-01 -8.40411365e-01 7.24185407e-01
4.64591026e-01 -8.71356428e-02 -1.21370506e+00 2.47002915e-01
8.96550477e-01 -4.39745158e-01 -6.50062919e-01 2.13900432e-02
2.94162989e-01 -5.84704936e-01 5.27876019e-01 -3.40298563e-01
1.22057772e+00 -2.92283207e-01 2.02959150e-01 -1.47186720e+00
-9.24411789e-02 -8.47398639e-01 -2.85777807e-01 2.04575253e+00
6.34496272e-01 -2.63542771e-01 4.63076264e-01 2.51470029e-01
-2.40240663e-01 -4.70417321e-01 -8.36668253e-01 -3.77630770e-01
1.37220159e-01 -2.45696992e-01 5.18003702e-01 6.24633908e-01
4.80582595e-01 4.12586480e-01 -3.06416929e-01 -3.66654545e-01
2.99521565e-01 -1.69569422e-02 5.48933208e-01 -1.07317710e+00
3.48590195e-01 -7.36715317e-01 8.57989565e-02 -2.53687531e-01
4.65421565e-02 -6.59030199e-01 -3.12149804e-02 -2.24629974e+00
3.47674429e-01 3.83810177e-02 7.86229372e-02 1.12601571e-01
-3.23110044e-01 -6.56462908e-02 3.90627742e-01 1.82191804e-01
-3.04190218e-01 1.37886807e-01 1.02865672e+00 -3.79563898e-01
-5.92010081e-01 2.98616678e-01 -1.15655529e+00 4.26268250e-01
8.58781517e-01 -6.64671600e-01 -3.26234639e-01 1.58351630e-01
4.89313066e-01 2.74473369e-01 8.22444633e-02 -7.74114788e-01
3.11293632e-01 -3.66524726e-01 2.54227340e-01 -1.45042157e+00
-1.69967115e-01 -7.01768324e-02 1.86718538e-01 3.27415824e-01
-7.18508482e-01 3.38801086e-01 4.37159359e-01 5.84304743e-02
-4.91610795e-01 -7.34061718e-01 2.22041532e-01 -4.10189033e-01
-6.56282902e-02 -4.65666085e-01 -9.65144813e-01 2.50200599e-01
9.58763897e-01 -5.29365003e-01 -2.38102853e-01 -4.02557909e-01
-3.34529221e-01 1.23540610e-01 5.95155656e-02 3.99156898e-01
7.43979871e-01 -1.17679799e+00 -1.43026543e+00 -5.47612548e-01
2.04669200e-02 9.26781669e-02 1.19871579e-01 5.80495477e-01
-7.00454354e-01 7.68717051e-01 -2.19205692e-02 1.76696301e-01
-1.32169092e+00 2.21531048e-01 -2.00404435e-01 -7.64615715e-01
-8.91983747e-01 3.96971077e-01 -2.30052829e-01 -2.13095471e-02
2.04472035e-01 -5.32883525e-01 -8.73700023e-01 9.82905328e-01
1.19012535e+00 5.56775391e-01 1.80958211e-01 -5.68586051e-01
-2.13947073e-01 3.34605873e-01 -3.06275725e-01 -6.09556139e-01
1.73979807e+00 -2.57457882e-01 -9.96549875e-02 8.45451176e-01
8.50974560e-01 4.47967768e-01 -8.96236718e-01 5.09675369e-02
5.54588139e-01 1.43914983e-01 -3.88349235e-01 -8.44899893e-01
-4.53614175e-01 3.27017188e-01 -6.35373354e-01 5.22438109e-01
9.23544526e-01 -6.77005649e-02 6.32551014e-01 1.69915661e-01
-1.52598873e-01 -1.37767375e+00 -7.81839937e-02 6.76695943e-01
1.43833911e+00 -8.05399656e-01 6.39710546e-01 -1.25397518e-01
-9.96887326e-01 1.68159473e+00 8.92515928e-02 -3.90860960e-02
2.02450552e-03 5.68205059e-01 -6.31147474e-02 -1.89965203e-01
-9.46329772e-01 4.15428221e-01 5.09303570e-01 4.95402694e-01
7.92872071e-01 -2.23415956e-01 -9.66024578e-01 1.16308844e+00
-7.40513921e-01 -1.46689355e-01 1.46493995e+00 9.58866417e-01
-4.35640961e-01 -8.39739799e-01 -7.06241906e-01 4.51528579e-01
-8.62099290e-01 -1.10309096e-02 -9.24072146e-01 8.21959376e-01
-2.95511514e-01 1.12129295e+00 3.79084684e-02 3.13123576e-02
6.13438606e-01 2.49520466e-01 1.68620825e-01 -8.54495406e-01
-1.07134461e+00 -6.70320764e-02 5.14100194e-01 1.97773382e-01
-5.60489893e-01 -1.09069049e+00 -1.27236676e+00 -2.86980718e-01
2.38342327e-04 6.66704774e-01 5.22540450e-01 9.67896998e-01
-8.99536237e-02 1.06190348e+00 5.06913602e-01 -8.00094247e-01
-2.26702780e-01 -1.28312087e+00 -3.47754121e-01 -4.14907895e-02
3.25366318e-01 -1.93836406e-01 -3.45074385e-01 4.90015358e-01] | [12.420710563659668, 9.523408889770508] |
905f62ad-414f-406f-ba62-9604d5ca266f | embracing-the-disharmony-in-heterogeneous | 2103.12857 | null | https://arxiv.org/abs/2103.12857v3 | https://arxiv.org/pdf/2103.12857v3.pdf | Embracing the Disharmony in Medical Imaging: A Simple and Effective Framework for Domain Adaptation | Domain shift, the mismatch between training and testing data characteristics, causes significant degradation in the predictive performance in multi-source imaging scenarios. In medical imaging, the heterogeneity of population, scanners and acquisition protocols at different sites presents a significant domain shift challenge and has limited the widespread clinical adoption of machine learning models. Harmonization methods which aim to learn a representation of data invariant to these differences are the prevalent tools to address domain shift, but they typically result in degradation of predictive accuracy. This paper takes a different perspective of the problem: we embrace this disharmony in data and design a simple but effective framework for tackling domain shift. The key idea, based on our theoretical arguments, is to build a pretrained classifier on the source data and adapt this model to new data. The classifier can be fine-tuned for intra-site domain adaptation. We can also tackle situations where we do not have access to ground-truth labels on target data; we show how one can use auxiliary tasks for adaptation; these tasks employ covariates such as age, gender and race which are easy to obtain but nevertheless correlated to the main task. We demonstrate substantial improvements in both intra-site domain adaptation and inter-site domain generalization on large-scale real-world 3D brain MRI datasets for classifying Alzheimer's disease and schizophrenia. | ['Christos Davatzikos', 'Pratik Chaudhari', 'Rongguang Wang'] | 2021-03-23 | null | null | null | null | ['auxiliary-learning'] | ['methodology'] | [ 4.54496205e-01 2.20073182e-02 -2.62759686e-01 -6.58800602e-01
-9.88490701e-01 -6.99448586e-01 3.37767810e-01 3.57638478e-01
-8.45204771e-01 8.15228164e-01 2.49828100e-02 -2.49575630e-01
-2.88717479e-01 -4.10591036e-01 -5.67174733e-01 -7.65506506e-01
-1.71605840e-01 8.18487763e-01 3.83580476e-01 -9.54342932e-02
1.99195929e-02 5.14496565e-01 -9.77252066e-01 3.11005682e-01
8.67315114e-01 7.65266418e-01 2.86690295e-01 3.84626210e-01
1.58310339e-01 1.02008604e-01 -5.54475665e-01 -1.16822928e-01
2.78972358e-01 -3.09711397e-01 -8.99914086e-01 5.43108806e-02
3.41187090e-01 -6.11701384e-02 -3.07518300e-02 9.77916300e-01
9.93629515e-01 -1.72914401e-01 1.09387922e+00 -1.11760437e+00
-3.12990189e-01 4.12670314e-01 -5.34147322e-01 6.57379508e-01
-1.63378105e-01 -1.53669894e-01 5.23676157e-01 -5.16711891e-01
8.48199308e-01 8.08378577e-01 8.73648703e-01 8.28009367e-01
-1.56457305e+00 -6.57901525e-01 9.53976735e-02 2.57147610e-01
-1.08166349e+00 -3.46463323e-01 6.42620027e-01 -8.56878042e-01
4.77306485e-01 5.64454645e-02 3.50126147e-01 1.53649807e+00
2.95875072e-01 3.60821992e-01 1.47533000e+00 -4.18459445e-01
3.79920036e-01 3.41393590e-01 1.24570034e-01 1.26302466e-01
2.02919051e-01 -4.56711045e-03 -2.60209918e-01 -2.97122389e-01
7.97207475e-01 -1.58986449e-01 -3.45996380e-01 -7.58681655e-01
-1.38753843e+00 8.21878314e-01 4.32440966e-01 5.90236247e-01
-4.84204680e-01 -5.11180043e-01 6.71940148e-01 5.72566211e-01
5.12383819e-01 5.46077490e-01 -6.63997114e-01 2.44153261e-01
-9.85915601e-01 6.97297752e-02 5.73155105e-01 5.82151830e-01
2.86560476e-01 -3.24990779e-01 -4.49891947e-02 1.09079146e+00
-1.91227362e-01 4.08787489e-01 9.61641252e-01 -6.01843774e-01
4.68517601e-01 2.10154876e-01 -1.98089540e-01 -6.78450346e-01
-9.29505050e-01 -7.10695922e-01 -1.02897775e+00 1.36789232e-01
6.44071579e-01 -2.61716217e-01 -1.06489396e+00 1.92748511e+00
3.26153964e-01 -1.18821241e-01 -8.83457512e-02 8.34565878e-01
3.69773865e-01 -1.83371216e-01 3.78521562e-01 -2.45653003e-01
1.31309211e+00 -5.08903682e-01 -2.64171124e-01 -4.90948498e-01
7.63587296e-01 -4.35052603e-01 8.82492602e-01 3.33589941e-01
-7.88476229e-01 -3.51702303e-01 -9.24036324e-01 3.78075272e-01
-3.74681234e-01 -9.89967510e-02 3.72294068e-01 7.47597635e-01
-9.74521875e-01 4.29545403e-01 -8.27838838e-01 -5.51384211e-01
5.70533693e-01 7.46282399e-01 -7.37076342e-01 -1.16551921e-01
-1.07537735e+00 1.29369056e+00 5.49146771e-01 -2.35392958e-01
-6.70438230e-01 -1.02926099e+00 -4.42605197e-01 -2.34862044e-01
4.40763265e-01 -8.84982884e-01 1.09518492e+00 -1.32605708e+00
-1.06271219e+00 1.34362757e+00 1.98285431e-01 -4.51045483e-01
7.21823037e-01 3.18153352e-01 -5.81029236e-01 7.19664097e-02
4.51205283e-01 4.02613193e-01 9.78636563e-01 -1.01454520e+00
-4.13494080e-01 -8.83266568e-01 -3.82103950e-01 1.49786055e-01
-2.61896819e-01 6.59748586e-03 6.32365346e-02 -7.50188112e-01
1.99513674e-01 -1.08660650e+00 -4.01380926e-01 -1.79809749e-01
-8.45034793e-02 2.24466771e-01 4.55959022e-01 -7.57427275e-01
8.96792352e-01 -2.05520344e+00 3.24890971e-01 3.72002900e-01
3.25268656e-01 1.60987645e-01 -3.32178026e-02 -2.74543129e-02
-5.94758332e-01 -1.31105497e-01 -4.30756718e-01 -1.47269582e-02
-2.85993993e-01 2.18015566e-01 1.86930820e-02 6.23884201e-01
1.02132745e-01 6.96898520e-01 -7.15410829e-01 -4.44801718e-01
-1.54349990e-02 2.20378444e-01 -6.15329266e-01 9.67895985e-02
2.42146105e-01 1.05190337e+00 -4.68307167e-01 3.16203624e-01
6.56436503e-01 -2.81719446e-01 3.31883579e-01 -6.92946389e-02
2.67056018e-01 3.53912413e-02 -1.03365815e+00 1.84031999e+00
-4.81815487e-01 2.06867784e-01 2.22114235e-01 -1.66345680e+00
8.40011001e-01 2.74274081e-01 7.08245933e-01 -7.29412019e-01
1.34563684e-01 6.41670763e-01 3.65218252e-01 -4.88955349e-01
-2.25844905e-01 -8.45153153e-01 -2.08908543e-01 2.31628865e-01
3.08797270e-01 -1.60123371e-02 -2.27467373e-01 -2.49822110e-01
1.15908587e+00 -3.09072286e-01 5.47318339e-01 -4.98689026e-01
4.69276965e-01 3.55305225e-02 5.52582204e-01 7.80469894e-01
-4.14603531e-01 7.04897344e-01 5.21223903e-01 -3.88286114e-01
-1.17312777e+00 -1.17429543e+00 -5.61872005e-01 1.08465064e+00
-4.06302452e-01 1.95176199e-01 -6.35600209e-01 -8.91368270e-01
-9.14960075e-03 3.40742975e-01 -9.16134357e-01 -4.33414429e-01
-6.32121444e-01 -1.23281860e+00 4.19835269e-01 6.17221296e-01
2.47728571e-01 -7.14603484e-01 -8.25348675e-01 3.27481717e-01
-6.57129437e-02 -1.17310297e+00 -2.29580596e-01 6.35992289e-01
-1.22607434e+00 -1.02330625e+00 -1.13685393e+00 -6.35035217e-01
6.70541227e-01 -1.46760255e-01 1.23143339e+00 -1.70677930e-01
-2.42715016e-01 4.23451096e-01 -3.10302764e-01 -4.88583833e-01
-6.26833916e-01 5.84939957e-01 1.84140161e-01 3.05610299e-02
2.42091253e-01 -7.59934902e-01 -6.89065397e-01 3.78559142e-01
-9.09051061e-01 -2.50022680e-01 7.70009875e-01 1.08089066e+00
4.63470757e-01 -2.92273939e-01 8.42694223e-01 -1.33791733e+00
5.59870720e-01 -6.95886970e-01 -2.91909784e-01 2.64856994e-01
-7.95096278e-01 1.55714467e-01 3.64149958e-01 -6.00590348e-01
-8.03868115e-01 7.43825212e-02 -1.47100881e-01 -1.79521292e-01
-3.39532524e-01 5.90984285e-01 -1.45217612e-01 -1.31769910e-01
1.01867056e+00 4.80430154e-03 2.38553464e-01 -6.31508231e-01
1.70539469e-02 6.83494747e-01 5.75918853e-01 -6.47743881e-01
5.09102464e-01 5.84225416e-01 1.29633114e-01 -5.73845863e-01
-6.35919809e-01 -3.90383512e-01 -1.25477111e+00 9.24316347e-02
7.48919070e-01 -7.90720344e-01 -3.40976976e-02 4.46125954e-01
-7.99276233e-01 -4.23703671e-01 -2.93327719e-01 4.99277771e-01
-6.95957601e-01 1.88704103e-01 -4.32840809e-02 -2.62569785e-01
-2.07244068e-01 -1.22171330e+00 9.10412788e-01 -1.29639402e-01
-2.83583790e-01 -1.32387781e+00 1.19616650e-01 1.17176704e-01
6.25583768e-01 3.85276973e-01 1.36791861e+00 -1.16743588e+00
1.36584565e-01 -3.28448787e-02 -1.54335365e-01 3.99720639e-01
2.54184723e-01 -7.42219269e-01 -8.51394236e-01 -4.90572840e-01
2.61370689e-01 -1.19968414e-01 7.26629555e-01 7.00999379e-01
1.20826781e+00 1.86174363e-01 -4.71943617e-01 5.33788502e-01
1.30792761e+00 2.06591263e-01 3.47012907e-01 6.57221973e-01
4.67314601e-01 7.18646109e-01 3.66073191e-01 2.83940196e-01
3.22344869e-01 1.12997913e+00 1.43311266e-02 -1.80845186e-01
-2.16155782e-01 1.51887536e-01 2.79273139e-03 4.62167829e-01
-1.44505545e-01 3.99645448e-01 -1.40845430e+00 6.51968896e-01
-1.64743686e+00 -6.36914432e-01 1.15805648e-01 2.48217058e+00
9.52722073e-01 5.08105420e-02 5.42739034e-01 1.14246450e-01
6.99028313e-01 -2.43843511e-01 -6.87095404e-01 -6.90065995e-02
-9.13575515e-02 3.03991467e-01 8.01382959e-01 2.40038663e-01
-1.05632997e+00 5.60167789e-01 6.89394617e+00 5.90779364e-01
-1.56597114e+00 7.36104548e-01 7.68248796e-01 -5.82915694e-02
1.32137090e-01 -3.93440068e-01 -3.50108594e-01 4.73146707e-01
1.17732811e+00 -1.45428434e-01 2.27906257e-01 7.88484037e-01
7.50985071e-02 -2.98509877e-02 -1.38855827e+00 1.00282311e+00
1.38710588e-01 -1.04690874e+00 -3.20235699e-01 2.68186070e-02
5.84298849e-01 2.47222677e-01 1.52573496e-01 2.29023978e-01
-9.54924524e-02 -1.10681045e+00 4.44146097e-01 2.26314202e-01
1.02041578e+00 -4.54665154e-01 7.19347477e-01 3.15568477e-01
-4.78145659e-01 -1.54756561e-01 -2.88474917e-01 2.87014306e-01
4.87927906e-02 4.85916704e-01 -9.44111347e-01 4.91429269e-01
7.87932336e-01 5.15185535e-01 -7.04377949e-01 1.06165266e+00
2.58011788e-01 3.66118670e-01 -2.10321426e-01 6.49103045e-01
-5.40399589e-02 8.17371011e-02 4.14131761e-01 1.07652283e+00
3.21029425e-01 -5.10830320e-02 1.41742602e-01 5.25747061e-01
1.80247068e-01 1.76971719e-01 -6.34042978e-01 3.97053123e-01
1.29247963e-01 9.72743094e-01 -8.03707063e-01 -2.16588840e-01
-4.35727745e-01 1.01261175e+00 3.61347020e-01 1.91860840e-01
-6.76542699e-01 7.83214942e-02 4.89391059e-01 4.67321754e-01
1.91496819e-01 -1.11796238e-01 -5.81625640e-01 -1.13868713e+00
7.23075643e-02 -1.11134470e+00 7.66154647e-01 -3.80787194e-01
-1.65077066e+00 6.47548139e-01 3.93062145e-01 -1.03922582e+00
-3.26044232e-01 -8.12541306e-01 -3.00598651e-01 9.84957695e-01
-1.54728746e+00 -1.13531077e+00 -1.60685793e-01 8.46245646e-01
2.54596710e-01 -4.73128825e-01 9.59189355e-01 6.70635879e-01
-2.52221346e-01 6.09781802e-01 3.38350505e-01 1.05894208e-01
1.07757664e+00 -1.13888359e+00 -7.98154995e-02 4.25612718e-01
-1.18533723e-01 3.80931854e-01 7.10701048e-01 -5.39923429e-01
-9.09156680e-01 -8.65213096e-01 6.85430586e-01 -6.27889633e-01
7.62693763e-01 -3.68482828e-01 -1.11715436e+00 7.15204835e-01
-3.26594651e-01 2.20346689e-01 8.16416025e-01 3.96179378e-01
-4.75917280e-01 -1.64625004e-01 -1.32139003e+00 2.78593987e-01
1.00003552e+00 -4.89883482e-01 -7.38878071e-01 3.84782314e-01
1.47484049e-01 -4.83890623e-01 -1.10747612e+00 3.81832421e-01
4.54779267e-01 -8.83377731e-01 1.02365625e+00 -9.77424800e-01
1.52969286e-01 1.96936026e-01 -4.91504818e-02 -1.67009640e+00
-3.17028254e-01 -1.04346268e-01 4.38306451e-01 9.22153890e-01
4.57818121e-01 -8.76487315e-01 7.70820558e-01 8.14284921e-01
-5.49759902e-02 -6.37879372e-01 -1.31775820e+00 -8.73716831e-01
8.42826188e-01 -2.50858754e-01 4.30571824e-01 1.05615079e+00
-4.66998704e-02 3.40887338e-01 -1.33846760e-01 1.20295458e-01
4.81779903e-01 -5.02161793e-02 3.92484754e-01 -1.58633876e+00
-3.54815215e-01 -4.43274796e-01 -6.09033763e-01 -3.48960966e-01
2.16641098e-01 -1.08238721e+00 -1.64983094e-01 -1.11258852e+00
4.53624964e-01 -8.64579678e-01 -5.82840145e-01 3.54998678e-01
4.50964365e-03 2.61942536e-01 6.70762286e-02 2.89793193e-01
-2.92566329e-01 1.25439927e-01 1.05989075e+00 -1.03583857e-01
-1.62635744e-01 1.99144572e-01 -8.06352079e-01 5.76610625e-01
7.42879629e-01 -8.93013060e-01 -4.17021066e-01 -4.47052598e-01
-1.47699371e-01 1.58511568e-02 5.76650620e-01 -1.00988543e+00
1.63389035e-02 -3.72640565e-02 5.02425551e-01 2.10850328e-01
8.15725252e-02 -1.03617609e+00 1.03527158e-01 5.41498363e-01
-3.20809305e-01 1.79481745e-01 2.19704390e-01 4.92400825e-01
-2.07744464e-02 -3.23256046e-01 1.11022091e+00 -2.00230807e-01
-5.56424141e-01 4.18567657e-01 -1.30846694e-01 4.59936351e-01
1.02524126e+00 -2.64631093e-01 -9.70765054e-02 -4.08869982e-02
-1.35374117e+00 -9.15707126e-02 4.39966530e-01 4.74122137e-01
1.41841203e-01 -1.21656275e+00 -8.09997857e-01 2.28797272e-01
1.90683007e-01 -2.48173058e-01 4.33611959e-01 1.36124539e+00
2.21561035e-03 3.92964214e-01 -7.19945490e-01 -6.87917113e-01
-1.21488488e+00 4.80577409e-01 6.52131617e-01 -4.43302900e-01
-5.55448771e-01 5.86081326e-01 3.42216700e-01 -4.79707360e-01
-8.39381814e-02 -1.85903147e-01 -8.48921910e-02 3.64726186e-01
2.86827892e-01 -2.80209370e-02 4.96752053e-01 -6.91390038e-01
-6.34302318e-01 5.15563130e-01 -3.13737124e-01 -1.70390263e-01
1.63681734e+00 -1.23255216e-01 1.78510815e-01 4.78207737e-01
1.17020822e+00 -4.00653481e-01 -1.09104443e+00 -4.19421494e-01
2.27799207e-01 -3.34804595e-01 3.48775275e-02 -1.06832755e+00
-1.09603989e+00 9.57543015e-01 1.05258095e+00 -6.96010664e-02
1.21787155e+00 2.21658811e-01 3.38489652e-01 2.83966325e-02
6.05673075e-01 -1.13707840e+00 -1.35172933e-01 2.16173753e-01
8.49820733e-01 -1.49581599e+00 -1.54573828e-01 -1.47860065e-01
-9.12113845e-01 9.01185930e-01 3.84700507e-01 -6.11351728e-02
7.33150065e-01 1.61226556e-01 2.64695197e-01 -1.80648148e-01
-3.55591863e-01 1.36872558e-02 3.36841226e-01 1.21435082e+00
3.49199861e-01 1.31883711e-01 -3.34545881e-01 9.05490279e-01
-7.58588836e-02 1.25528723e-01 3.43484968e-01 6.92574441e-01
-8.85142535e-02 -1.55158651e+00 -4.60877210e-01 7.13358045e-01
-6.03370309e-01 1.33347318e-01 -5.45233153e-02 9.03883457e-01
2.82136321e-01 2.90682197e-01 -4.20448333e-02 6.39449432e-03
5.35286665e-01 3.06658387e-01 6.63939893e-01 -8.29275131e-01
-5.68341553e-01 -9.94120762e-02 -1.86933890e-01 -2.93493658e-01
-3.84750932e-01 -9.15983498e-01 -1.02178776e+00 4.92747463e-02
1.11299753e-01 -5.32737114e-02 6.67069077e-01 9.84494984e-01
4.75574225e-01 5.48713923e-01 3.64625007e-01 -7.79499888e-01
-9.66877162e-01 -8.19784641e-01 -7.30739295e-01 6.56358838e-01
3.92965138e-01 -1.03906918e+00 -7.29905143e-02 -7.43209105e-03] | [14.526108741760254, -1.9041730165481567] |
5158fb26-7af0-48fc-a8f7-b64b92e23712 | pvgru-generating-diverse-and-relevant | 2212.09086 | null | https://arxiv.org/abs/2212.09086v4 | https://arxiv.org/pdf/2212.09086v4.pdf | PVGRU: Generating Diverse and Relevant Dialogue Responses via Pseudo-Variational Mechanism | We investigate response generation for multi-turn dialogue in generative-based chatbots. Existing generative models based on RNNs (Recurrent Neural Networks) usually employ the last hidden state to summarize the sequences, which makes models unable to capture the subtle variability observed in different dialogues and cannot distinguish the differences between dialogues that are similar in composition. In this paper, we propose a Pseudo-Variational Gated Recurrent Unit (PVGRU) component without posterior knowledge through introducing a recurrent summarizing variable into the GRU, which can aggregate the accumulated distribution variations of subsequences. PVGRU can perceive the subtle semantic variability through summarizing variables that are optimized by the devised distribution consistency and reconstruction objectives. In addition, we build a Pseudo-Variational Hierarchical Dialogue (PVHD) model based on PVGRU. Experimental results demonstrate that PVGRU can broadly improve the diversity and relevance of responses on two benchmark datasets. | ['Yifei Zhang', 'Hinrich Schütze', 'Daling Wang', 'Shi Feng', 'Yongkang Liu'] | 2022-12-18 | null | null | null | null | ['response-generation'] | ['natural-language-processing'] | [-2.80574355e-02 3.06567609e-01 7.09411874e-02 -5.83764851e-01
-8.99203956e-01 -4.35750306e-01 7.59640694e-01 -4.74875003e-01
1.90764830e-01 9.68220055e-01 8.02076638e-01 -1.04102507e-01
1.83344945e-01 -7.39111960e-01 -5.11219978e-01 -8.81624699e-01
6.65454447e-01 8.41677666e-01 -6.16746768e-03 -7.55602777e-01
2.22990766e-01 -3.64763975e-01 -1.04571784e+00 6.91511810e-01
1.05436885e+00 4.05585438e-01 6.10546231e-01 8.92974794e-01
-7.62715399e-01 1.10946190e+00 -1.05661738e+00 -4.42382008e-01
-1.99297398e-01 -1.45819569e+00 -9.53933656e-01 1.97917745e-01
-3.22481334e-01 -3.19915593e-01 -2.77876914e-01 8.79970253e-01
6.77709520e-01 5.54758847e-01 8.23622048e-01 -8.87873888e-01
-1.20295632e+00 1.30723035e+00 -2.73851901e-01 1.46305570e-02
4.33355689e-01 2.07921356e-01 1.44428372e+00 -5.87967217e-01
6.70494437e-01 1.94490552e+00 4.82818395e-01 1.06568575e+00
-1.27906775e+00 -2.20633969e-01 4.26956266e-01 3.26089352e-01
-9.06221926e-01 -2.12180316e-01 1.04201567e+00 -1.72392771e-01
9.10385609e-01 6.00097716e-01 3.62471312e-01 1.66180229e+00
5.62217422e-02 1.19077837e+00 9.33921516e-01 -8.19910914e-02
1.50736436e-01 9.08781961e-02 1.43265739e-01 3.58082712e-01
-5.83410263e-01 -4.32500541e-01 -5.92926800e-01 -1.07773013e-01
5.38607180e-01 8.88508337e-04 -2.66547769e-01 2.69061267e-01
-7.48125434e-01 1.25524712e+00 2.34193280e-01 1.50870621e-01
-5.05788863e-01 1.46224648e-02 3.00992966e-01 2.79573590e-01
5.48952997e-01 2.97083944e-01 -2.11948559e-01 -4.65410590e-01
-3.95768017e-01 3.20064992e-01 9.70064759e-01 1.14474642e+00
9.07385230e-01 2.48306885e-01 -9.68504488e-01 1.18486154e+00
4.68875438e-01 2.20843896e-01 7.70376980e-01 -9.18076813e-01
4.98887241e-01 6.27897680e-01 -1.18159026e-01 -6.99424982e-01
-1.48878187e-01 -2.46142685e-01 -9.31068182e-01 -5.41920125e-01
1.89851239e-01 -5.58651149e-01 -6.42291129e-01 1.68349254e+00
3.65066141e-01 -1.66213602e-01 2.39362434e-01 8.66534233e-01
1.45564079e+00 8.68464828e-01 -1.54594213e-01 -3.16666394e-01
1.06183434e+00 -1.14272141e+00 -1.10368240e+00 1.05379708e-02
4.67753470e-01 -5.45603275e-01 1.28577161e+00 -4.35686717e-03
-1.03685105e+00 -5.52112699e-01 -5.17961860e-01 -2.42471814e-01
3.55715491e-02 -1.42991990e-01 3.75456750e-01 4.72807646e-01
-9.94885266e-01 4.11207080e-01 -4.69528109e-01 -7.00078458e-02
-6.85003921e-02 -9.87899005e-02 3.95394683e-01 3.33379179e-01
-1.55418622e+00 6.79067254e-01 1.47041768e-01 4.50972706e-01
-9.40111041e-01 -4.64626759e-01 -7.37433493e-01 8.09646249e-02
4.46698189e-01 -8.81404042e-01 1.49376667e+00 -7.12780535e-01
-2.22155571e+00 2.52751023e-01 -4.52992231e-01 -3.78093541e-01
7.69535482e-01 -2.22131480e-02 1.97501361e-01 -2.38725826e-01
-1.82946220e-01 3.89021397e-01 6.91180110e-01 -1.41921949e+00
-3.48835409e-01 -1.41990647e-01 -8.33720788e-02 4.75524902e-01
2.02449918e-01 -2.11814031e-01 -2.93623179e-01 -4.59498525e-01
-5.83435595e-02 -8.98154438e-01 -3.14111054e-01 -1.20168424e+00
-7.54525959e-01 -7.51628757e-01 6.47376657e-01 -7.75876641e-01
1.33029914e+00 -1.60814130e+00 5.72552741e-01 -1.94393963e-01
2.57180274e-01 3.34764570e-02 -1.98642284e-01 6.96342826e-01
6.35027409e-01 1.13185272e-01 -7.29115978e-02 -6.31549358e-01
2.68311739e-01 6.57689571e-01 -4.83648598e-01 -1.86872318e-01
2.31160328e-01 1.32631314e+00 -9.74171698e-01 -3.52852851e-01
-2.32111886e-02 4.36439276e-01 -6.76468074e-01 7.18366563e-01
-7.54091859e-01 8.38359594e-01 -6.40057266e-01 2.57995307e-01
3.55945975e-01 -4.12877917e-01 6.95390880e-01 1.20484412e-01
1.66348815e-01 6.55603707e-01 -5.89176059e-01 1.44983923e+00
-5.57817578e-01 4.35959280e-01 -2.83829361e-01 -7.86251843e-01
1.27009726e+00 3.17033321e-01 1.23635709e-01 -6.67525649e-01
2.89369583e-01 -1.16910942e-01 8.01194087e-02 -5.62506616e-01
1.02834260e+00 -7.63304681e-02 -3.18243742e-01 7.06938267e-01
2.26677060e-01 -8.89920369e-02 1.39623180e-01 3.84016395e-01
7.92084396e-01 1.24676995e-01 -5.61173856e-02 1.97794050e-01
5.26024520e-01 -5.65558970e-01 7.02846348e-01 1.05424309e+00
1.40290046e-02 8.73725593e-01 1.01694691e+00 -1.54012740e-02
-9.18602765e-01 -9.55550492e-01 3.61969829e-01 1.50705147e+00
7.49218240e-02 -4.68783379e-01 -7.94192672e-01 -6.40955448e-01
-4.62950170e-01 8.92982125e-01 -8.83976042e-01 -1.19436316e-01
-5.36883473e-01 -9.73938823e-01 3.52071434e-01 5.07331789e-01
4.05192673e-01 -1.38003695e+00 -1.01804867e-01 4.91736144e-01
-9.04031515e-01 -8.33870292e-01 -4.95602012e-01 1.91773288e-02
-5.89078546e-01 -6.52226210e-01 -6.83820426e-01 -4.79053557e-01
2.60883451e-01 1.32372528e-01 1.25565398e+00 -2.04505175e-02
2.13504165e-01 2.72730410e-01 -7.64432192e-01 -2.64519542e-01
-9.13066745e-01 4.36701387e-01 -4.86323327e-01 6.01378381e-02
2.22198233e-01 -4.71310556e-01 -3.08724552e-01 3.20128500e-01
-5.65470815e-01 2.08438888e-01 3.80741566e-01 1.28916454e+00
5.37784874e-01 -5.73763371e-01 8.52048397e-01 -1.35674584e+00
1.26546228e+00 -7.10642338e-01 2.60460526e-02 5.40278792e-01
-3.99167985e-01 3.77416074e-01 5.58311105e-01 -6.13322735e-01
-1.57434118e+00 -5.41759491e-01 -3.35823029e-01 -2.16458604e-01
-5.82138076e-03 5.89016199e-01 -2.74461895e-01 8.08654785e-01
4.16058362e-01 6.74772084e-01 2.68160515e-02 -4.69786674e-01
6.06816351e-01 7.51936793e-01 2.45975837e-01 -6.52274847e-01
1.22487389e-01 4.42948639e-02 -4.74638879e-01 -9.15857613e-01
-8.44909012e-01 -2.45564029e-01 -3.16454649e-01 -3.05344164e-01
9.43714976e-01 -7.13925004e-01 -8.91275585e-01 4.79481727e-01
-1.33703101e+00 -5.93070328e-01 -1.56136826e-01 1.49665877e-01
-5.95290065e-01 4.55981553e-01 -9.26966369e-01 -1.31021309e+00
-4.36287731e-01 -1.09131312e+00 9.24368322e-01 5.94203115e-01
-4.38613176e-01 -1.25819731e+00 3.55373234e-01 5.36282122e-01
4.46364224e-01 4.88176197e-02 9.26613629e-01 -7.65402257e-01
-6.31699681e-01 3.62303257e-01 1.81166857e-01 2.96921313e-01
2.51102060e-01 1.33557439e-01 -9.60825443e-01 1.78294271e-01
1.34229779e-01 -3.16624641e-01 9.69084680e-01 5.43232083e-01
9.60011542e-01 -7.05005705e-01 1.04844049e-01 3.07092696e-01
7.53428996e-01 3.49321663e-01 8.17946970e-01 -1.11651845e-01
9.75277364e-01 8.36365759e-01 3.63169640e-01 6.48467958e-01
7.50473082e-01 5.56917965e-01 3.04287851e-01 2.58477658e-01
2.62893550e-02 -6.00767076e-01 5.44081390e-01 1.46373701e+00
-2.05456540e-01 -6.44260705e-01 -3.44302654e-01 4.31701630e-01
-2.22349787e+00 -1.17810249e+00 -4.21030045e-01 1.62856734e+00
9.92109120e-01 -1.79561049e-01 1.97113246e-01 -5.13038158e-01
9.06348884e-01 4.88283455e-01 -6.01029813e-01 -6.97604954e-01
-2.79098988e-01 -3.14886570e-02 -9.65347737e-02 7.44447768e-01
-3.44781071e-01 1.11114514e+00 6.04737520e+00 8.38861346e-01
-7.13078141e-01 2.03243613e-01 6.88104212e-01 -1.13013880e-02
-9.84641731e-01 -8.78712460e-02 -9.60167646e-01 6.77142739e-01
9.43705857e-01 -2.15110153e-01 4.05910939e-01 5.51937819e-01
3.52810144e-01 2.68336266e-01 -8.80667150e-01 6.42681420e-01
1.33212939e-01 -1.24372065e+00 3.63729537e-01 -6.16215169e-02
9.29442704e-01 -2.71333843e-01 3.44744712e-01 7.18977809e-01
1.05344367e+00 -1.11663043e+00 4.36890334e-01 7.22201169e-01
1.84486762e-01 -8.42433512e-01 8.93222034e-01 5.36619782e-01
-7.83607543e-01 1.83986649e-01 -5.74319601e-01 8.86019319e-02
3.35844398e-01 1.40828058e-01 -1.32645833e+00 7.80526876e-01
3.92501026e-01 6.15614474e-01 -3.11101675e-01 2.14026645e-01
-6.36328876e-01 8.76181602e-01 2.34865874e-01 -7.17145562e-01
4.09895867e-01 -6.11401618e-01 6.28635705e-01 1.15592718e+00
1.20823801e-01 9.64603201e-02 8.86518061e-02 1.34301257e+00
-6.33026510e-02 3.77053842e-02 -4.26665731e-02 -8.51061791e-02
5.82153499e-01 1.07157910e+00 -3.65255564e-01 -2.76981205e-01
-4.27830257e-02 1.10014009e+00 2.99266756e-01 5.58255374e-01
-9.37953353e-01 -2.30839923e-02 8.05931807e-01 -5.52769303e-01
4.36368704e-01 9.59559903e-02 -2.76614070e-01 -1.13568759e+00
-1.68302014e-01 -1.02149928e+00 3.04010361e-01 -6.56483591e-01
-1.57045984e+00 8.22285175e-01 -2.50488162e-01 -7.69071639e-01
-1.01456952e+00 3.12210582e-02 -7.56894767e-01 1.21829271e+00
-1.10996878e+00 -1.14140213e+00 -1.80405080e-01 5.76838851e-01
1.08811951e+00 -1.07731760e-01 7.54139423e-01 -2.54156470e-01
-7.21548259e-01 7.54977405e-01 2.01948151e-01 1.28463984e-01
5.28381407e-01 -1.50622654e+00 4.88922536e-01 4.50539023e-01
1.44144595e-01 7.75027931e-01 9.91906703e-01 -6.20370090e-01
-1.04580164e+00 -9.77580011e-01 8.13929200e-01 -7.04509437e-01
3.62452716e-01 -4.26588446e-01 -1.26957798e+00 6.73820019e-01
5.58885038e-01 -9.61426139e-01 8.90268326e-01 5.15650272e-01
-1.71434715e-01 4.08832937e-01 -7.43538022e-01 7.16713428e-01
8.65172327e-01 -6.64151609e-01 -7.77271092e-01 1.80851385e-01
9.56694126e-01 -5.62073231e-01 -5.32702744e-01 -5.80549911e-02
4.12154526e-01 -1.01130116e+00 6.57668293e-01 -9.38730896e-01
7.63675988e-01 2.57041529e-02 -7.23589510e-02 -1.66055012e+00
-2.77893454e-01 -1.02979076e+00 -2.55525529e-01 1.57975948e+00
4.41912711e-01 -5.57226062e-01 5.93043149e-01 4.36338574e-01
-2.54473150e-01 -6.94183171e-01 -5.54571092e-01 -4.69559550e-01
1.53660759e-01 -1.64405018e-01 7.97395468e-01 9.22784805e-01
-5.61755225e-02 9.18266833e-01 -9.36332166e-01 -2.43026569e-01
1.52370557e-01 2.05297351e-01 1.04272687e+00 -9.34811413e-01
-6.90989137e-01 -6.01101279e-01 9.81494114e-02 -1.55548179e+00
3.97724897e-01 -6.41051352e-01 5.18694699e-01 -1.67879391e+00
3.96867096e-01 7.06021562e-02 4.02815230e-02 4.29774635e-03
-8.28615487e-01 -2.42644057e-01 7.95918778e-02 9.35669765e-02
-7.24746048e-01 1.26218760e+00 1.42495573e+00 -1.19958065e-01
-5.05657613e-01 1.73038498e-01 -9.45017874e-01 3.83018464e-01
7.22925663e-01 -3.22360039e-01 -6.59114122e-01 -3.25570136e-01
1.84217647e-01 3.68660361e-01 1.04928657e-01 -9.65738446e-02
8.51593688e-02 -3.13951075e-01 1.53750358e-02 -7.82292068e-01
3.13352555e-01 7.51537532e-02 1.50331110e-01 4.04466456e-03
-8.36599946e-01 1.78915765e-02 -4.21563685e-01 7.43045032e-01
-2.41572559e-01 -1.37871861e-01 3.21618944e-01 -3.04682046e-01
-4.18129593e-01 -2.73692850e-02 -6.63553536e-01 3.08630556e-01
3.95534873e-01 -1.06701246e-02 -4.51083362e-01 -1.02052784e+00
-6.41034901e-01 4.59343553e-01 -8.78259987e-02 6.78428769e-01
6.46887839e-01 -1.29102468e+00 -1.01508391e+00 -1.80999041e-01
-1.41757652e-01 1.35925829e-01 1.03540564e+00 5.07927895e-01
-8.94363895e-02 2.11136118e-01 -1.40319078e-03 -7.21756995e-01
-1.40285361e+00 1.54087901e-01 3.42741519e-01 -6.81998968e-01
-4.54547703e-01 1.20594394e+00 4.00480419e-01 -8.03928077e-01
-7.62068406e-02 -1.39236093e-01 -4.41295236e-01 3.44094485e-01
3.88216972e-01 3.35360438e-01 -3.34165931e-01 -5.41043758e-01
1.44492641e-01 -5.42347804e-02 -2.95264959e-01 -2.75020301e-01
1.23679292e+00 -5.83911538e-01 -9.96433850e-03 9.33306694e-01
9.47591543e-01 -8.76639262e-02 -1.46482790e+00 -3.41531873e-01
-2.67857701e-01 -2.73294955e-01 -4.98727858e-01 -5.60100496e-01
-8.07802677e-01 9.56642270e-01 -1.52460784e-01 4.40124273e-01
5.90817094e-01 2.33062088e-01 9.20391083e-01 3.29891980e-01
1.22123815e-01 -1.22975957e+00 4.08951402e-01 1.09282219e+00
9.56231236e-01 -9.56495225e-01 -4.97768641e-01 -1.75509080e-01
-1.46612012e+00 8.10949087e-01 6.66129708e-01 1.21599570e-01
2.42448561e-02 -1.76531762e-01 3.99272591e-01 -7.04690218e-02
-1.29282784e+00 -2.90783465e-01 2.46147141e-01 5.92388868e-01
6.27169907e-01 2.62375206e-01 -5.41510642e-01 7.97086954e-01
-6.38933063e-01 -8.28418016e-01 7.50356019e-01 4.12655920e-01
-3.60922962e-01 -1.19647539e+00 3.24554220e-02 1.59246907e-01
-2.86030143e-01 -8.76512453e-02 -8.84462297e-01 3.52199018e-01
-3.43391120e-01 1.24279428e+00 1.32714078e-01 -7.23340929e-01
1.02301285e-01 3.61743599e-01 2.67213941e-01 -7.80933380e-01
-9.31169868e-01 2.83868194e-01 9.49330404e-02 -2.29048714e-01
-2.78541625e-01 -6.57009900e-01 -1.14757860e+00 -3.47420454e-01
-3.30951571e-01 4.86378491e-01 3.86430442e-01 1.00717175e+00
2.20885247e-01 1.16446495e+00 8.97261381e-01 -3.34752530e-01
-9.36224103e-01 -1.54424071e+00 -5.03004909e-01 4.21053797e-01
9.94072258e-02 -3.96958500e-01 -3.35239083e-01 -1.27590045e-01] | [12.562708854675293, 8.385823249816895] |
d7a2ebbf-4f57-403c-9790-53aafd96affe | deep-network-guided-proof-search | 1701.06972 | null | http://arxiv.org/abs/1701.06972v1 | http://arxiv.org/pdf/1701.06972v1.pdf | Deep Network Guided Proof Search | Deep learning techniques lie at the heart of several significant AI advances
in recent years including object recognition and detection, image captioning,
machine translation, speech recognition and synthesis, and playing the game of
Go. Automated first-order theorem provers can aid in the formalization and
verification of mathematical theorems and play a crucial role in program
analysis, theory reasoning, security, interpolation, and system verification.
Here we suggest deep learning based guidance in the proof search of the theorem
prover E. We train and compare several deep neural network models on the traces
of existing ATP proofs of Mizar statements and use them to select processed
clauses during proof search. We give experimental evidence that with a hybrid,
two-phase approach, deep learning based guidance can significantly reduce the
average number of proof search steps while increasing the number of theorems
proved. Using a few proof guidance strategies that leverage deep neural
networks, we have found first-order proofs of 7.36% of the first-order logic
translations of the Mizar Mathematical Library theorems that did not previously
have ATP generated proofs. This increases the ratio of statements in the corpus
with ATP generated proofs from 56% to 59%. | ['Geoffrey Irving', 'Christian Szegedy', 'Cezary Kaliszyk', 'Sarah Loos'] | 2017-01-24 | null | null | null | null | ['game-of-go'] | ['playing-games'] | [ 4.76342529e-01 3.71708304e-01 -2.90269911e-01 7.63854058e-03
-9.98184085e-01 -8.95912051e-01 6.27738297e-01 2.38991559e-01
-2.26431176e-01 5.76233327e-01 -2.07554460e-01 -1.48955619e+00
-8.26338306e-02 -1.02410579e+00 -1.35828388e+00 2.49348134e-02
-2.36553058e-01 4.97905731e-01 8.49161372e-02 -1.82663813e-01
3.91389579e-01 2.53232360e-01 -1.08182681e+00 8.03325117e-01
7.32301772e-01 7.86633134e-01 -2.46990956e-02 1.34135044e+00
-1.95747867e-01 1.53286111e+00 -5.21545887e-01 -4.14396524e-01
6.17882647e-02 -3.60530078e-01 -1.15128279e+00 -6.81504309e-01
7.81436086e-01 -8.15405607e-01 -6.05881751e-01 1.33491552e+00
-7.34399855e-02 -5.50971806e-01 2.27832273e-01 -1.68218386e+00
-4.48494256e-01 1.44942081e+00 -7.79779702e-02 2.82194257e-01
3.60367358e-01 3.02287459e-01 1.27845132e+00 -3.18547428e-01
6.72029018e-01 1.27397919e+00 5.88013828e-01 8.70343029e-01
-1.20433927e+00 -8.69816661e-01 -3.99744630e-01 7.13815868e-01
-9.97460306e-01 -5.62040508e-01 8.30322802e-01 -3.33594382e-01
1.41395330e+00 2.33477131e-01 5.49403608e-01 7.73118913e-01
3.87428492e-01 1.02974439e+00 6.07791126e-01 -7.05753386e-01
6.36478588e-02 -2.00354800e-01 7.10316449e-02 1.47269654e+00
1.37042001e-01 5.45037873e-02 -2.48433083e-01 -9.39391106e-02
4.94042307e-01 -4.96628642e-01 8.76417309e-02 -2.57852197e-01
-1.64103484e+00 1.03049123e+00 1.46630540e-01 1.80968657e-01
2.42868736e-01 8.49227369e-01 9.61022675e-01 5.19126475e-01
-2.60543436e-01 9.35314834e-01 -6.61558986e-01 -2.08680481e-01
-1.00791824e+00 6.98781908e-01 9.34200406e-01 7.63682902e-01
3.12470257e-01 1.23484597e-01 -5.88394701e-02 -1.59996226e-02
2.02162787e-02 8.63193512e-01 -9.75302607e-02 -1.46088493e+00
7.59186804e-01 6.78496957e-01 -1.58014074e-01 -8.64036560e-01
-1.96009606e-01 -1.57324538e-01 -5.04139006e-01 2.09080487e-01
6.19182646e-01 -1.42704204e-01 -5.89351118e-01 1.61414194e+00
-1.39400482e-01 -4.61272113e-02 2.86162943e-01 5.52269340e-01
6.23970449e-01 8.79166067e-01 -4.51323003e-01 1.62722811e-01
1.16594875e+00 -6.86602652e-01 -2.44213015e-01 -9.20527652e-02
1.03556359e+00 -4.37235147e-01 6.43892884e-01 7.64224350e-01
-1.22400534e+00 -3.03613413e-02 -1.42355239e+00 -1.49321511e-01
-2.33335152e-01 -6.18489943e-02 1.12101209e+00 8.16680267e-02
-1.03059793e+00 5.70721745e-01 -5.68290830e-01 3.89806181e-01
7.55327940e-01 5.78567743e-01 -1.10873885e-01 -2.27960557e-01
-1.18018854e+00 7.63776124e-01 5.44677317e-01 1.39401574e-02
-1.25871778e+00 -8.29543233e-01 -1.12326097e+00 4.16195840e-01
6.13459349e-01 -5.40098429e-01 1.78198516e+00 -6.88983798e-01
-1.04206324e+00 6.55296147e-01 -3.42457779e-02 -1.32248509e+00
5.20333648e-01 -1.06757715e-01 -7.73671567e-02 2.46168897e-01
1.16733290e-01 9.06780183e-01 5.86741924e-01 -4.91331041e-01
-7.80112684e-01 3.23380679e-02 5.96767306e-01 -4.99995738e-01
4.44574445e-01 1.54655904e-01 2.50903577e-01 2.19751060e-01
-8.49728882e-02 -7.08909810e-01 8.85995328e-02 7.15814531e-02
-3.55196059e-01 -5.38841009e-01 6.82741821e-01 -6.06923938e-01
5.87899089e-01 -1.94191456e+00 6.21939637e-02 1.85397625e-01
5.56593060e-01 2.46956661e-01 -7.90491849e-02 8.62152651e-02
-6.19535074e-02 2.73453891e-01 1.24507606e-01 4.46985155e-01
6.30656779e-01 4.31104489e-02 -6.76896811e-01 4.60242569e-01
5.46916544e-01 1.37335443e+00 -1.20349765e+00 -6.67247593e-01
2.22563967e-01 -6.38277456e-02 -8.62131298e-01 -7.47095644e-02
-1.11539912e+00 -2.07270473e-01 -1.41665295e-01 6.59542620e-01
4.36682671e-01 -3.67140770e-01 3.24898303e-01 2.03073189e-01
-2.99920198e-02 9.78655100e-01 -7.55446613e-01 1.70848870e+00
-3.33434463e-01 1.39462447e+00 1.60260260e-01 -1.12562108e+00
2.71515787e-01 2.47754559e-01 7.26885572e-02 -9.77950513e-01
2.77854055e-01 2.90381789e-01 6.27575994e-01 -1.94254875e-01
2.93624490e-01 1.36170998e-01 -4.36395109e-02 6.21996999e-01
-1.16683051e-01 -2.76589423e-01 5.17881572e-01 5.47564447e-01
1.45554304e+00 6.78866357e-02 -1.21755362e-01 5.60193621e-02
5.22432566e-01 3.88920188e-01 2.60985881e-01 1.00542176e+00
-2.01831274e-02 -2.69713737e-02 1.09756398e+00 -8.21564496e-01
-1.41649699e+00 -7.68628418e-01 3.37120593e-01 9.07851577e-01
-4.18737680e-01 -6.95246816e-01 -8.90139759e-01 -5.87727487e-01
-4.04393598e-02 8.34142208e-01 -1.53874591e-01 -4.25426543e-01
-7.84700751e-01 2.76911706e-01 1.17426777e+00 5.09629786e-01
5.35280406e-01 -1.30403852e+00 -9.26967859e-01 1.40179753e-01
-2.08489850e-01 -1.21710789e+00 -1.94687732e-02 4.02937531e-01
-5.21716893e-01 -1.50234294e+00 -8.10250640e-02 -8.19961309e-01
4.98635203e-01 -2.32958943e-01 1.11072266e+00 4.44676459e-01
-1.97521523e-01 -5.64326905e-02 5.50003611e-02 -5.13743520e-01
-1.26972532e+00 1.57801107e-01 2.16154419e-02 -9.72936690e-01
2.24911198e-01 -2.81145543e-01 1.22827880e-01 -4.03026313e-01
-6.04581773e-01 4.21477437e-01 6.30186856e-01 8.58008146e-01
3.20817716e-02 3.44673246e-01 -8.58044475e-02 -6.31588638e-01
5.84942698e-01 8.69266763e-02 -1.49613905e+00 3.37840855e-01
-6.89247251e-01 6.40997767e-01 8.57096314e-01 -4.46590096e-01
-7.06451982e-02 -1.81855679e-01 1.49607196e-01 -5.74886858e-01
1.13497436e-01 5.97525895e-01 1.76130205e-01 -1.82066374e-02
6.68944538e-01 3.75496030e-01 1.21830069e-02 4.10426080e-01
3.80580515e-01 1.47849098e-01 8.44988644e-01 -1.02303863e+00
9.55093622e-01 -1.45427644e-01 4.93028760e-01 -9.24655646e-02
-6.43545508e-01 2.36192241e-01 -1.54770792e-01 2.30862930e-01
3.94721657e-01 -6.50979996e-01 -1.59411693e+00 8.36335644e-02
-1.76733780e+00 -8.34231555e-01 4.70646955e-02 8.66098255e-02
-6.71840429e-01 2.89277643e-01 -7.39741087e-01 -9.12648380e-01
-8.10171485e-01 -1.30407345e+00 9.17199850e-01 -2.26065531e-01
-5.51391661e-01 -5.01697779e-01 -3.22114676e-01 2.93905526e-01
7.05731884e-02 6.01653382e-03 1.60995770e+00 -6.30990922e-01
-1.04429293e+00 -2.23098755e-01 -6.16574109e-01 3.37470740e-01
-3.34075987e-01 1.03725893e-02 -7.54074752e-01 7.49120861e-02
-5.01443267e-01 -4.06911641e-01 7.15016484e-01 3.18565786e-01
9.02323365e-01 -8.94135714e-01 -1.28720090e-01 4.81896639e-01
1.06294513e+00 1.11268736e-01 6.41957402e-01 4.64607507e-01
4.79727179e-01 -1.61584884e-01 2.17660442e-01 1.42757483e-02
1.42377660e-01 -9.25072294e-04 5.76696992e-01 3.22728872e-01
9.79935676e-02 -4.43051159e-01 5.87636530e-01 9.61145684e-02
3.68078232e-01 1.98253810e-01 -1.31969953e+00 6.33682013e-01
-1.61213708e+00 -1.38049436e+00 6.39030486e-02 1.67089891e+00
1.10850286e+00 7.57500887e-01 -4.74345945e-02 6.58641696e-01
3.20262522e-01 -3.51522774e-01 -2.44392350e-01 -9.76424336e-01
2.11931288e-01 5.39003432e-01 7.57141590e-01 8.14599752e-01
-1.18334460e+00 9.95162189e-01 6.38908100e+00 5.43987095e-01
-1.18722773e+00 -4.62123364e-01 2.54532546e-01 4.18636948e-02
-3.39440912e-01 1.72092207e-02 -3.96501154e-01 1.07400045e-02
1.21640897e+00 -1.82988256e-01 1.05466878e+00 1.06361723e+00
-2.30229571e-01 -1.54456198e-01 -1.85689938e+00 7.87223577e-01
-1.58894494e-01 -2.17035460e+00 -1.23010918e-01 4.40542847e-02
3.62239510e-01 7.85969421e-02 3.62567268e-02 5.23946702e-01
3.65279645e-01 -1.32028747e+00 1.08946621e+00 6.59685582e-02
9.73766983e-01 -9.67292666e-01 8.86243820e-01 2.99038470e-01
-6.05772734e-01 -3.68488908e-01 -1.00780139e-02 -6.23557389e-01
-5.70699632e-01 1.05171695e-01 -1.63557911e+00 1.36324570e-01
2.76520520e-01 3.61862510e-01 -2.90337801e-01 5.06547511e-01
-5.44387043e-01 6.54513776e-01 -4.26373035e-01 -5.08577585e-01
4.21659350e-01 2.45904967e-01 6.06830060e-01 1.10331750e+00
-2.32515335e-01 -1.31640032e-01 -1.11254595e-01 1.70321870e+00
-2.73446769e-01 -8.02845240e-01 -7.05963373e-01 -6.40346467e-01
1.53957531e-01 9.07263398e-01 -7.02292979e-01 -5.57654440e-01
-9.58943069e-02 6.62728190e-01 7.17703179e-02 3.02036181e-02
-1.11734378e+00 -5.69120467e-01 2.76027948e-01 2.02368442e-02
4.21639025e-01 -5.52320302e-01 -3.50906283e-01 -8.95355463e-01
2.43270591e-01 -1.46573532e+00 -8.40532258e-02 -9.09777164e-01
-4.43462133e-01 4.96064387e-02 7.48842210e-02 -4.13599461e-01
-7.02991903e-01 -1.01256406e+00 -6.48019552e-01 7.12893009e-01
-1.40571296e+00 -9.36105788e-01 4.06708151e-01 1.58534795e-01
5.08836031e-01 -3.35409492e-01 9.21477079e-01 1.85169712e-01
-1.96662143e-01 6.27324224e-01 -2.98171788e-01 8.55305254e-01
-4.77083139e-02 -1.34064841e+00 5.99274099e-01 1.12705052e+00
3.00355554e-01 7.67300487e-01 8.53026867e-01 -2.55632490e-01
-2.20426202e+00 -7.03129590e-01 1.03238440e+00 -3.13711822e-01
1.13782275e+00 -5.88674068e-01 -3.30501974e-01 6.74008846e-01
9.73464027e-02 -2.68410474e-01 -2.49212101e-01 1.94661483e-01
-7.62278497e-01 -9.62317735e-02 -8.16840947e-01 9.12965059e-01
5.89571595e-01 -1.03492713e+00 -6.76279724e-01 5.02200425e-01
8.78116071e-01 -7.54296839e-01 -2.72661656e-01 1.31368175e-01
8.56441677e-01 -5.15126526e-01 8.17703903e-01 -8.29265654e-01
1.12210536e+00 -2.26467907e-01 -2.58830100e-01 -5.28318048e-01
-3.75331603e-02 -9.09782290e-01 -7.27676272e-01 8.21680248e-01
6.57844782e-01 -1.95837051e-01 9.52011287e-01 4.97604549e-01
-1.18286774e-01 -7.13012695e-01 -9.80945230e-01 -5.75437725e-01
3.70924979e-01 -9.07102525e-01 4.91643846e-01 6.91376805e-01
5.50877392e-01 6.29740357e-01 3.61719400e-01 1.06748655e-01
3.81355286e-01 3.55000705e-01 8.41235757e-01 -9.56281304e-01
-2.99908131e-01 -9.28305805e-01 -3.44152510e-01 -7.38054097e-01
7.75707185e-01 -1.35346866e+00 2.32332125e-01 -1.58317077e+00
2.99511075e-01 -1.31169319e-01 -7.98757672e-02 9.27692831e-01
5.68857908e-01 -3.02938163e-01 9.22551975e-02 -4.57405269e-01
-6.83839202e-01 -1.31614223e-01 8.31992209e-01 -1.08298624e+00
3.25573355e-01 -2.56764859e-01 -5.43444037e-01 6.72494292e-01
6.22634828e-01 -3.16969842e-01 -1.13711685e-01 -5.39757133e-01
9.36228395e-01 1.76949054e-01 8.33699584e-01 -1.10065448e+00
2.96932727e-01 -2.86676407e-01 8.28855559e-02 -5.71353018e-01
-2.15696290e-01 -7.12317646e-01 -5.17418325e-01 8.63771617e-01
-7.10620761e-01 2.54060268e-01 6.34903848e-01 -5.60565554e-02
1.84914783e-01 -1.31623566e-01 3.89172226e-01 -2.62135923e-01
-7.10845828e-01 -1.40544221e-01 -5.66446543e-01 1.76687203e-02
3.93871903e-01 1.59058243e-01 -3.78298551e-01 -3.63128304e-01
-5.05646095e-02 2.95204937e-01 1.80796266e-01 1.18852057e-01
8.18644822e-01 -8.73707414e-01 -6.67486727e-01 1.21399082e-01
-1.98063254e-01 2.94270694e-01 -4.61163819e-01 9.08590496e-01
-9.48698103e-01 1.00042939e+00 -1.25378162e-01 -4.44002479e-01
-1.32135868e+00 5.58757305e-01 5.42328835e-01 -4.11514997e-01
-6.64453089e-01 7.39831448e-01 -2.22931370e-01 -3.68062466e-01
3.94893199e-01 -1.12123632e+00 5.80389798e-01 -7.40599096e-01
6.07941449e-01 1.03966154e-01 -5.37382253e-02 2.50062585e-01
-6.91806376e-01 -5.70244379e-02 -1.84692249e-01 -1.80543244e-01
1.18070436e+00 9.36042368e-01 -5.52392244e-01 1.00312695e-01
9.06151474e-01 -1.02227405e-01 -5.21464586e-01 8.65006596e-02
1.22935750e-01 3.38254482e-01 1.06073067e-01 -9.06776845e-01
-4.73984897e-01 1.02318227e+00 -2.20205598e-02 1.44023642e-01
7.40544379e-01 9.57828239e-02 9.29668784e-01 1.37726748e+00
2.33060047e-01 -8.47049475e-01 -1.68815777e-01 1.16659808e+00
3.77056837e-01 -1.13091505e+00 9.92191061e-02 7.32278675e-02
1.41204536e-01 1.39227343e+00 3.70414615e-01 -2.72890925e-01
-3.23892921e-01 7.60847151e-01 -4.38874066e-01 -1.54519051e-01
-9.71124232e-01 4.11130577e-01 1.54326171e-01 3.50407511e-01
3.50417078e-01 4.02688347e-02 3.46336335e-01 5.26610315e-01
-4.77587402e-01 2.67992020e-01 5.89297593e-01 8.64216506e-01
-4.36075985e-01 -7.91266859e-01 -4.25867856e-01 1.94029808e-01
-5.24483204e-01 -6.71638250e-01 -4.41729903e-01 9.04065371e-01
2.68598516e-02 8.01588714e-01 -8.10403526e-02 -3.07765901e-01
-2.62979388e-01 2.68540382e-01 1.25968015e+00 -4.23188835e-01
-1.68349624e-01 -4.80976313e-01 3.62038046e-01 -4.53540862e-01
8.06934983e-02 -4.75282162e-01 -1.68606341e+00 -8.52151275e-01
-3.62120718e-01 1.51345685e-01 6.14727795e-01 1.27711964e+00
4.10007276e-02 7.57644236e-01 1.55051365e-01 -3.65369916e-01
-6.86677516e-01 -5.76161563e-01 1.43934518e-01 -2.18215123e-01
8.34450305e-01 -1.28955498e-01 -8.04246441e-02 9.25424099e-02] | [8.920608520507812, 7.069777965545654] |
d79dde35-7b36-49b9-a3bc-6915d30d2f26 | unicon-unsupervised-intent-discovery-via | null | null | https://openreview.net/forum?id=-jZkAHbpHlk | https://openreview.net/pdf?id=-jZkAHbpHlk | UNICON: Unsupervised Intent Discovery via Semantic-level Contrastive Learning | Discovering new intents is crucial for expanding domains in dialogue systems or natural language understanding (NLU) systems. A typical approach is to leverage unsupervised and semi-supervised learning to train a neural encoder to produce representations of utterances that are adequate for clustering then perform clustering on the representations to detect unseen clusters of intents. Recently, instance-level contrastive learning has been proposed to improve representation quality for better clustering. However, the proposed method suffers from semantic distortion in text augmentation and even from representation inadequacy due to limitations of using representations of pre-trained language models, typically BERT. Neural encoders can be powerful representation learners, but the initial parameters of pre-trained language models do not reliably produce representations that are suitable for capturing semantic distances. To eliminate the necessity of data augmentation and reduce the negative impact of pre-trained language models as encoders, we propose UNICON, a novel contrastive learning method that utilizes auxiliary external representations to provide powerful guidance for the encoder. Neural encoders can be powerful representation learners, but the initial parameters of pre-trained language models do not reliably produce representations that are suitable for capturing semantic distances. To eliminate the necessity of data augmentation and reduce the negative impact of pre-trained language models as encoders, we propose UNICON, a novel contrastive learning method that utilizes auxiliary external representations to provide powerful guidance for the encoder. | ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['text-augmentation', 'intent-discovery'] | ['natural-language-processing', 'natural-language-processing'] | [ 2.66234398e-01 5.65382421e-01 1.15601113e-02 -5.65688610e-01
-5.33617735e-01 -4.31407481e-01 7.50496805e-01 2.46719301e-01
-4.93554741e-01 3.47507834e-01 4.44017440e-01 -3.29658717e-01
1.37680015e-02 -7.59198904e-01 -5.62489688e-01 -2.77045786e-01
-1.30928271e-02 7.70122230e-01 -3.47989500e-02 -4.85504359e-01
2.05368862e-01 1.66922629e-01 -1.75745678e+00 4.97021139e-01
1.24677718e+00 5.60185254e-01 5.73135257e-01 6.06594741e-01
-5.67084491e-01 8.15849960e-01 -9.33889151e-01 1.25568628e-01
-4.55714539e-02 -7.38581181e-01 -9.85466838e-01 2.87379384e-01
-1.31172523e-01 -2.00512365e-01 -1.77211314e-01 8.50167990e-01
2.64245987e-01 3.24378282e-01 1.01807690e+00 -1.10833180e+00
-9.84476566e-01 7.52989531e-01 -1.72091499e-01 -9.19446722e-02
3.39404732e-01 -5.70323318e-02 9.41440225e-01 -7.42394447e-01
3.93319041e-01 1.33619964e+00 5.14704823e-01 9.16845143e-01
-1.15027249e+00 -4.87295359e-01 2.25970432e-01 1.22998180e-02
-1.14019561e+00 -6.04008138e-01 8.93524230e-01 -2.48700872e-01
1.13094771e+00 1.38139322e-01 2.14711070e-01 9.65888679e-01
-5.10357738e-01 9.16086674e-01 5.13016164e-01 -6.35887265e-01
4.44283158e-01 5.08864582e-01 2.23018646e-01 6.75988972e-01
-1.55491024e-01 -1.16796225e-01 -2.03259021e-01 -2.89681405e-01
7.55491734e-01 8.78260359e-02 -1.37979880e-01 -2.79481709e-01
-7.65259981e-01 1.36653852e+00 6.03411198e-01 6.07232511e-01
-3.80668968e-01 -2.50423938e-01 5.11301160e-01 2.91135490e-01
4.30519402e-01 9.89921212e-01 -3.41759056e-01 -1.99116543e-01
-6.60773396e-01 -2.03757748e-01 5.37940323e-01 9.95979130e-01
9.12322223e-01 2.69724458e-01 -5.58898896e-02 1.08516014e+00
1.64075255e-01 1.00005986e-02 1.11886990e+00 -7.98285604e-01
4.51657504e-01 1.26292372e+00 -3.52275878e-01 -7.15512335e-01
-4.37209159e-01 -2.53721476e-01 -9.40196633e-01 -1.24555640e-01
-1.84965208e-02 -2.31573850e-01 -9.48978603e-01 1.66564858e+00
-9.39398929e-02 3.07033602e-02 6.64591610e-01 5.45361578e-01
8.55155528e-01 1.02968860e+00 -6.24986142e-02 -3.67964208e-01
1.02554679e+00 -8.07651341e-01 -6.06287599e-01 -3.86803091e-01
1.50969481e+00 -3.91813755e-01 1.38444054e+00 1.54548779e-01
-9.43664789e-01 -9.22506928e-01 -9.33007836e-01 9.18902904e-02
-6.11399472e-01 2.02647597e-01 4.76092964e-01 4.82206643e-01
-1.06648970e+00 2.74082094e-01 -4.89478976e-01 -3.34464222e-01
2.99196541e-01 5.33771634e-01 -2.29946718e-01 -2.07034618e-01
-1.28624606e+00 9.54101324e-01 7.67780900e-01 -3.14257801e-01
-7.80777395e-01 -4.17703927e-01 -1.22370005e+00 2.59304583e-01
2.47493565e-01 -2.72457093e-01 1.13183033e+00 -1.26401627e+00
-1.36344802e+00 3.11903566e-01 -5.32845780e-02 -6.10438168e-01
-1.14303201e-01 2.14435924e-02 -2.91397274e-01 1.45100132e-01
1.86209112e-01 1.07769799e+00 6.70113623e-01 -1.38852715e+00
-4.07651037e-01 -1.70903653e-01 6.34376332e-02 4.76266921e-01
-9.16199028e-01 -2.32681423e-01 -2.43297398e-01 -4.93956119e-01
1.24800183e-01 -7.29008257e-01 -4.46582586e-01 -6.46906853e-01
-1.66599616e-01 -6.16957664e-01 1.25009537e+00 -4.49030370e-01
1.18037927e+00 -2.08155942e+00 -7.46633112e-02 1.68864533e-01
1.25739509e-02 6.64255679e-01 -5.51013470e-01 2.48231560e-01
-1.57086074e-01 2.70661414e-01 -3.95995766e-01 -3.17726403e-01
-1.92138135e-01 5.97007751e-01 -1.34042904e-01 -8.84223208e-02
6.59718812e-01 5.42797506e-01 -1.05286515e+00 -1.70609415e-01
5.34628212e-01 4.68591094e-01 -7.75957465e-01 7.00529814e-01
-5.02715051e-01 3.55304241e-01 -2.00254515e-01 5.64493053e-03
3.07277888e-01 -1.60477236e-01 1.59629866e-01 -4.84353229e-02
1.98270842e-01 7.71048069e-01 -1.01462388e+00 1.54017997e+00
-8.39568675e-01 4.86386001e-01 -2.54064202e-01 -1.60024738e+00
1.44122148e+00 4.60121751e-01 3.62006903e-01 -6.29535258e-01
7.40038753e-02 3.95262428e-02 2.80669212e-01 -4.40859079e-01
5.67389429e-01 -3.45407337e-01 -2.22744584e-01 6.92861021e-01
3.92659515e-01 -1.25151142e-01 1.64661720e-01 4.05438930e-01
1.13207388e+00 -1.86535388e-01 1.33494794e-01 3.28156054e-02
6.28264904e-01 8.10769051e-02 2.91881293e-01 4.02559996e-01
2.65518785e-01 6.66675985e-01 2.25523978e-01 -3.54948342e-01
-9.62565780e-01 -5.85037827e-01 -3.45673226e-02 1.45958340e+00
-1.91996112e-01 -6.64437115e-01 -5.34719706e-01 -1.15976882e+00
-3.58299464e-01 1.08671117e+00 -4.42409396e-01 -6.95291817e-01
-2.68179208e-01 -5.87548256e-01 7.27784991e-01 7.55538881e-01
3.09233844e-01 -1.14255881e+00 -3.54125112e-01 3.63028586e-01
-3.15256268e-01 -1.01464772e+00 -1.28439829e-01 6.35838568e-01
-9.48116362e-01 -8.89567554e-01 -4.95536208e-01 -9.26717639e-01
1.18586409e+00 2.46321023e-01 9.08557177e-01 3.40443879e-01
-2.68331356e-02 1.69618309e-01 -7.06510663e-01 -2.77788430e-01
-1.04619813e+00 2.51133859e-01 2.57174194e-01 -1.54234126e-01
6.98912799e-01 -3.35203350e-01 5.48163094e-02 3.54382426e-01
-1.07375777e+00 1.03755251e-01 4.62796867e-01 1.00464511e+00
7.77391866e-02 2.81043917e-01 1.15480065e+00 -1.10471809e+00
8.99049282e-01 -6.21818841e-01 -2.06297234e-01 1.39447764e-01
-5.05708456e-01 5.60838640e-01 1.18076897e+00 -5.10806799e-01
-1.18310690e+00 1.79241821e-02 -3.62294436e-01 -4.55981106e-01
-4.64702368e-01 6.16888285e-01 -3.59509215e-02 5.45166552e-01
9.72063422e-01 2.41578177e-01 9.42186564e-02 -5.01823723e-01
4.92806286e-01 1.13168919e+00 4.04200315e-01 -4.63344604e-01
4.31335390e-01 -1.13751572e-02 -6.65671408e-01 -9.34715331e-01
-7.74027467e-01 -7.24647641e-01 -9.28390682e-01 2.00905517e-01
6.59185588e-01 -8.81437302e-01 -2.08802614e-02 -2.40726203e-01
-1.22204721e+00 -2.44312465e-01 -3.48946720e-01 5.34668148e-01
-5.08118093e-01 3.69664431e-01 -3.13102424e-01 -9.02684033e-01
-2.17480972e-01 -1.07451475e+00 6.37133121e-01 2.29604408e-01
-6.62107110e-01 -1.27737999e+00 -5.22205457e-02 4.02061194e-01
3.61458868e-01 -2.08456025e-01 1.05732727e+00 -1.37533891e+00
9.66040343e-02 -1.38790056e-01 -6.18167818e-02 6.53660655e-01
6.06119514e-01 -1.64593682e-01 -1.02416968e+00 -3.34720373e-01
3.04211527e-02 -6.86881423e-01 5.94275415e-01 3.34864445e-02
1.16502571e+00 -5.72299957e-01 -1.97007403e-01 2.55563229e-01
8.64734054e-01 5.40064156e-01 5.81135035e-01 1.81140050e-01
5.94663620e-01 8.54452491e-01 4.80652064e-01 3.68599176e-01
3.98302138e-01 4.32450265e-01 1.51845679e-01 -4.03417230e-01
-6.32380927e-03 -3.24008554e-01 5.56753635e-01 1.13468289e+00
2.13504806e-01 9.66667198e-03 -1.02928376e+00 7.05337644e-01
-1.71006262e+00 -6.41716957e-01 2.03750551e-01 2.00174594e+00
9.26236451e-01 1.14626125e-01 1.16243616e-01 2.00872406e-01
7.51641452e-01 -2.27558851e-01 -4.63831455e-01 -9.01981056e-01
2.06163913e-01 1.00990586e-01 -9.15542617e-03 3.08301240e-01
-8.63738120e-01 1.06292582e+00 5.35408449e+00 4.81406242e-01
-1.00278604e+00 9.28828865e-03 5.33742487e-01 3.02424997e-01
-4.48173553e-01 9.08049047e-02 -6.07036471e-01 2.72030920e-01
1.10214031e+00 8.05549175e-02 1.18818969e-01 1.02120745e+00
6.10160977e-02 1.21047199e-01 -1.32746875e+00 7.89897263e-01
4.06453907e-01 -1.22881842e+00 4.69157666e-01 -4.01848331e-02
8.31696093e-01 -1.61616042e-01 -1.72623873e-01 8.49908531e-01
5.07095337e-01 -1.13396168e+00 -1.60019159e-01 1.33084685e-01
6.25139654e-01 -9.99299765e-01 1.11097395e+00 6.88521922e-01
-7.62390733e-01 -1.86035097e-01 -7.57179558e-01 -2.87899166e-01
-1.34352922e-01 9.72922817e-02 -1.58221662e+00 3.40260953e-01
2.39894405e-01 5.86791456e-01 -6.44212246e-01 4.14996177e-01
-9.47185010e-02 6.04901731e-01 -1.49404541e-01 -2.66601499e-02
6.90439224e-01 -3.36771272e-02 1.06671736e-01 1.24444294e+00
1.98738307e-01 3.41705717e-02 4.15666074e-01 7.52905250e-01
-1.83347404e-01 1.65786251e-01 -1.15964580e+00 -3.88015330e-01
4.95758504e-01 8.65497649e-01 -4.24242377e-01 -6.66684449e-01
-4.82429534e-01 7.51096487e-01 5.32358587e-01 2.70214230e-01
-4.11956578e-01 -4.24444228e-01 3.11345279e-01 2.28119597e-01
8.27108771e-02 -5.01413085e-02 -1.80107892e-01 -1.00162280e+00
-4.44172144e-01 -1.05269837e+00 4.61683899e-01 -5.37059188e-01
-1.18303263e+00 8.39691043e-01 1.19742919e-02 -1.26435339e+00
-9.04829681e-01 -2.72191703e-01 -8.99136484e-01 6.98814690e-01
-1.24363863e+00 -9.47084069e-01 -7.48890103e-04 7.08889723e-01
1.01498067e+00 -8.01532567e-01 1.13218629e+00 -1.57045886e-01
-3.15312237e-01 5.91813624e-01 1.40105784e-01 4.13305998e-01
6.93974555e-01 -1.15654087e+00 1.16979912e-01 5.81986248e-01
2.80326307e-01 7.99199045e-01 3.21104676e-01 -4.44739878e-01
-1.01397860e+00 -1.23403859e+00 8.08100283e-01 -2.27210104e-01
1.70149475e-01 -3.58350754e-01 -1.42017913e+00 6.46501362e-01
1.52932301e-01 -4.22791064e-01 1.12806392e+00 4.82547373e-01
-2.33185336e-01 2.17905849e-01 -1.02919602e+00 4.15828466e-01
6.18018389e-01 -5.27569830e-01 -1.23681712e+00 2.83378065e-01
7.34224975e-01 1.91164538e-01 -6.45976007e-01 4.70417857e-01
-1.45867005e-01 -7.11285293e-01 7.87204385e-01 -6.04798436e-01
3.37574542e-01 -3.66970748e-02 4.93705161e-02 -1.52868426e+00
-2.89516032e-01 -3.17683339e-01 -6.30232096e-02 1.36047471e+00
5.35527349e-01 -4.03800040e-01 7.39754498e-01 6.14869177e-01
-4.21032429e-01 -5.00232577e-01 -4.88109350e-01 -7.47423708e-01
6.89670518e-02 -4.14785594e-01 4.77615744e-01 1.37340379e+00
5.10044277e-01 1.01235235e+00 -1.50911853e-01 -1.30982265e-01
1.37153774e-01 -1.86375514e-01 9.14903581e-01 -1.29847848e+00
-5.33841327e-02 -3.08373362e-01 -2.17947587e-01 -1.27926683e+00
4.12282228e-01 -1.02336490e+00 3.44746143e-01 -1.68815875e+00
-3.90508324e-02 -7.17213094e-01 -1.98352680e-01 6.69427454e-01
-8.94045234e-02 -1.11096568e-01 6.35887608e-02 2.06038430e-01
-5.75713456e-01 8.31749797e-01 9.04846907e-01 -2.67470717e-01
-6.68625176e-01 1.22895770e-01 -8.75160217e-01 9.38348174e-01
9.40478265e-01 -4.55327481e-01 -1.02821982e+00 -2.85124719e-01
-8.93016830e-02 -1.04848199e-01 -9.45845768e-02 -9.92630303e-01
2.99265474e-01 1.63583070e-01 4.90245312e-01 -5.78348935e-01
3.58024567e-01 -7.07124174e-01 -4.58982766e-01 4.17617083e-01
-7.57731915e-01 -1.25994785e-02 2.53544211e-01 2.55132705e-01
-5.26092887e-01 -7.46671557e-01 6.97954893e-01 -4.07043070e-01
-6.28248096e-01 -1.17466636e-01 -7.25190639e-01 1.22354358e-01
7.13958204e-01 -3.18371147e-01 -4.46665958e-02 -6.76526308e-01
-5.82835913e-01 4.30626899e-01 3.79734844e-01 6.43030882e-01
8.26512635e-01 -1.21437812e+00 -6.35637343e-01 5.12094498e-01
2.41115332e-01 3.76298338e-01 -4.92861532e-02 2.71070004e-01
-1.40446275e-01 5.19095361e-01 -1.47814691e-01 -5.87469041e-01
-1.19498074e+00 6.94965303e-01 -1.12468004e-02 -2.40703270e-01
-4.06128109e-01 7.66418695e-01 5.19149542e-01 -9.77881074e-01
4.76564378e-01 -3.97188276e-01 -5.27048945e-01 7.03843310e-02
5.31638741e-01 5.88944182e-03 1.24571837e-01 -6.14163220e-01
-1.61770105e-01 6.80347458e-02 -4.47949588e-01 2.82465462e-02
1.43861413e+00 -2.20642406e-02 2.53790885e-01 4.69400972e-01
1.39615059e+00 -5.05873144e-01 -9.86230671e-01 -3.24040860e-01
1.61565036e-01 -1.07366331e-01 3.36378329e-02 -5.47589898e-01
-6.68553531e-01 1.29454315e+00 3.67900848e-01 2.51416892e-01
1.04440737e+00 1.47325292e-01 7.10106850e-01 7.30312765e-01
4.55101114e-03 -1.41109836e+00 6.81682050e-01 8.39089632e-01
8.15912485e-01 -1.31420362e+00 -3.87157202e-01 -1.52296096e-01
-1.06344724e+00 1.15005457e+00 9.87709939e-01 -3.97515204e-03
3.15899283e-01 6.39443398e-02 1.24348007e-01 -1.61978751e-01
-7.90213764e-01 -3.76453608e-01 2.96866030e-01 9.55433488e-01
7.17818499e-01 -1.67386085e-01 1.12792112e-01 6.60315692e-01
-3.18540663e-01 -4.16230857e-01 4.97424752e-01 8.49370539e-01
-5.72439611e-01 -1.15336096e+00 -3.01356733e-01 6.23169720e-01
1.24457367e-01 -1.13856554e-01 -6.08669460e-01 4.40985650e-01
2.21989304e-01 1.12625420e+00 5.11693895e-01 -5.97878873e-01
1.11827910e-01 4.19767678e-01 -2.88742618e-03 -1.30455232e+00
-6.94367588e-01 -1.89797692e-02 2.56115466e-01 7.13604018e-02
-2.66076028e-01 -3.16870004e-01 -1.55825889e+00 2.00692579e-01
-7.25889564e-01 6.74890101e-01 4.91949588e-01 1.16244900e+00
3.88149232e-01 5.35245180e-01 9.60452795e-01 -4.90820199e-01
-5.47226667e-01 -1.37216604e+00 -1.74731940e-01 6.13254786e-01
8.55919570e-02 -6.39567316e-01 -3.05999190e-01 1.24777265e-01] | [12.492609977722168, 7.40785551071167] |
551c8064-f463-477b-85c6-f794286c0017 | pneumonia-detection-on-chest-x-ray-using | 2101.04269 | null | https://arxiv.org/abs/2101.04269v2 | https://arxiv.org/pdf/2101.04269v2.pdf | Pneumonia Detection on Chest X-ray using Radiomic Features and Contrastive Learning | Chest X-ray becomes one of the most common medical diagnoses due to its noninvasiveness. The number of chest X-ray images has skyrocketed, but reading chest X-rays still have been manually performed by radiologists, which creates huge burnouts and delays. Traditionally, radiomics, as a subfield of radiology that can extract a large number of quantitative features from medical images, demonstrates its potential to facilitate medical imaging diagnosis before the deep learning era. With the rise of deep learning, the explainability of deep neural networks on chest X-ray diagnosis remains opaque. In this study, we proposed a novel framework that leverages radiomics features and contrastive learning to detect pneumonia in chest X-ray. Experiments on the RSNA Pneumonia Detection Challenge dataset show that our model achieves superior results to several state-of-the-art models (> 10% in F1-score) and increases the model's interpretability. | ['Yifan Peng', 'Ying Ding', 'Ahmed H Tewfik', 'Chongyan Chen', 'Yan Han'] | 2021-01-12 | null | null | null | null | ['pneumonia-detection'] | ['medical'] | [ 2.21030042e-01 4.88861501e-02 -2.51689941e-01 -5.49276531e-01
-1.11407900e+00 -3.28262717e-01 6.74022362e-02 2.59398490e-01
-2.47485995e-01 3.38231087e-01 3.37736845e-01 -9.37192380e-01
-2.57650584e-01 -6.58301055e-01 -5.67758799e-01 -4.70453709e-01
1.28167793e-01 7.75131941e-01 4.57324572e-02 2.30223998e-01
-3.60689729e-01 5.35908520e-01 -8.63199234e-01 5.85264146e-01
4.25786108e-01 9.15967703e-01 2.88971275e-01 1.18233526e+00
-2.99484488e-02 1.36304188e+00 -4.22996998e-01 -1.15482651e-01
5.77143021e-02 -5.48874557e-01 -8.72742951e-01 -1.76069483e-01
2.37687230e-01 -1.09057093e+00 -8.48656833e-01 5.49650729e-01
6.61626518e-01 -3.56469780e-01 6.82962596e-01 -6.31431401e-01
-8.12004626e-01 5.37323058e-01 -6.37054741e-01 7.90355206e-01
2.89815590e-02 1.98126599e-01 8.98067176e-01 -6.33393526e-01
3.11242670e-01 6.81102097e-01 9.68898952e-01 6.86031044e-01
-6.01041317e-01 -4.94450808e-01 -6.05710089e-01 8.17530081e-02
-8.34998310e-01 2.16971338e-01 1.43765479e-01 -4.50291246e-01
8.65026236e-01 5.87578773e-01 6.49073184e-01 9.32817876e-01
4.10866350e-01 6.06788695e-01 9.42741156e-01 -3.44077021e-01
-7.57159889e-02 -2.32010961e-01 1.22345649e-01 9.80630398e-01
3.16966385e-01 1.93382967e-02 -1.14467822e-01 -3.09135377e-01
9.34453905e-01 8.51139963e-01 -3.57055992e-01 1.97411999e-02
-1.20568144e+00 7.82824218e-01 6.38554752e-01 3.38120550e-01
-4.03055757e-01 2.35257387e-01 3.28636020e-01 -2.25565866e-01
1.77487269e-01 5.66369236e-01 -3.80316615e-01 -8.79481062e-02
-9.11876738e-01 -6.97480142e-02 3.81444424e-01 3.24767739e-01
-1.58527419e-01 -4.45924282e-01 -2.31576130e-01 7.56864965e-01
6.36192486e-02 6.63635671e-01 6.07228100e-01 -6.95039570e-01
2.94451863e-01 7.04954445e-01 -3.44188929e-01 -4.97431755e-01
-9.02250588e-01 -6.41131580e-01 -1.22887492e+00 -1.33813456e-01
3.18075746e-01 -2.43693087e-02 -1.26760280e+00 1.25595105e+00
1.79079667e-01 1.04674697e-02 -3.38913053e-01 1.02863133e+00
9.98542011e-01 2.58734733e-01 1.31555974e-01 5.05068190e-02
1.57245505e+00 -8.26434314e-01 -4.49388593e-01 4.28569727e-02
7.33815849e-01 -6.40276492e-01 1.11096096e+00 2.97317952e-01
-1.06924975e+00 -3.14985007e-01 -9.38480556e-01 -2.19829127e-01
-7.04899291e-03 1.22944117e-01 6.93640947e-01 6.60677969e-01
-7.74745941e-01 3.55346352e-01 -1.19815719e+00 -3.17583203e-01
9.97049034e-01 4.30212706e-01 -1.68055698e-01 -3.76608610e-01
-1.01936710e+00 8.79910231e-01 3.18115503e-02 -2.69273013e-01
-6.62481725e-01 -1.17387712e+00 -2.47611240e-01 2.98370212e-01
4.10515696e-01 -1.18502557e+00 1.58709896e+00 -1.80363297e-01
-1.01964653e+00 9.19451535e-01 3.13716561e-01 -3.98356438e-01
5.68864107e-01 -2.95502543e-01 -2.44190916e-01 7.60825574e-01
-1.39259277e-02 6.58031046e-01 5.20335615e-01 -8.32300246e-01
-7.44916677e-01 -3.26783061e-01 -8.14746320e-03 -6.10447489e-02
-2.80772835e-01 1.52235152e-02 -2.63488144e-01 -5.98132312e-01
2.54974723e-01 -1.06909597e+00 -2.85087854e-01 1.49335548e-01
-2.64371425e-01 5.24671115e-02 6.82471097e-01 -7.03887761e-01
1.12634480e+00 -1.85328615e+00 -4.12516028e-01 1.43842399e-03
1.05406439e+00 3.25543284e-01 5.28813660e-01 -1.22960158e-01
-4.74088669e-01 5.88555992e-01 -2.64506996e-01 1.89952210e-01
-3.38867426e-01 8.35349187e-02 -2.83956826e-02 3.20667028e-01
2.12552428e-01 1.28261912e+00 -9.64823544e-01 -6.57078087e-01
3.41276526e-01 6.66802585e-01 -4.39570397e-01 3.65528733e-01
1.14162706e-01 5.40123165e-01 -6.34702802e-01 7.03500032e-01
3.03209692e-01 -1.48844242e+00 1.23170495e-01 2.19223052e-02
3.98167133e-01 3.41355234e-01 -2.46292844e-01 1.38400018e+00
-3.71232748e-01 4.02714103e-01 -3.26385766e-01 -5.95566332e-01
2.03464627e-01 6.00050390e-01 9.05482352e-01 -5.84882200e-01
2.59349078e-01 4.82196994e-02 1.89750671e-01 -9.72745240e-01
-2.11162090e-01 -3.77330393e-01 3.04003149e-01 8.28984797e-01
-5.18440902e-01 -2.40673468e-01 -3.51522774e-01 2.19340459e-01
1.78129959e+00 -3.43626142e-01 5.49422443e-01 2.03577891e-01
1.80336878e-01 9.39097404e-02 -1.73193235e-02 9.84709263e-01
-1.71686187e-01 1.15260327e+00 1.23090409e-01 -6.97015584e-01
-1.19729948e+00 -1.31309569e+00 -4.29567546e-01 7.97152281e-01
-3.40772629e-01 1.02511324e-01 -6.24388278e-01 -9.24614608e-01
-5.62634319e-02 2.29800463e-01 -7.54299402e-01 8.46240856e-03
-7.24263668e-01 -1.09522069e+00 7.87941635e-01 9.71217453e-01
3.60710144e-01 -7.29337513e-01 -1.08202136e+00 -1.31264152e-02
-2.37038627e-01 -9.66210842e-01 -2.45014086e-01 2.78363436e-01
-1.11453354e+00 -1.27185309e+00 -1.06410062e+00 -4.85729367e-01
5.93366623e-01 4.75701183e-01 1.37242508e+00 4.90101933e-01
-1.00242102e+00 4.28784877e-01 -3.29938591e-01 -6.47082746e-01
-6.44553244e-01 2.34563068e-01 -3.95814866e-01 -6.50070488e-01
4.87941384e-01 -2.85362482e-01 -1.02009535e+00 -1.69558719e-01
-1.10879612e+00 1.68671116e-01 1.12145662e+00 9.59105492e-01
7.43580997e-01 -5.86955696e-02 5.80557168e-01 -1.30344558e+00
3.11538130e-01 -7.25265920e-01 -1.67498991e-04 1.77084997e-01
-6.88100636e-01 -1.92258075e-01 5.77262640e-01 2.45728418e-02
-9.57397282e-01 -2.10457727e-01 -1.98420644e-01 -3.67046595e-01
-2.49566436e-01 4.63569015e-01 6.14179313e-01 2.35971138e-01
9.27219689e-01 -7.85521790e-02 -4.19916622e-02 -4.02350426e-01
4.15813141e-02 7.70943940e-01 6.75211370e-01 5.72248846e-02
5.87388575e-01 6.25487387e-01 3.34484726e-01 -6.62865579e-01
-1.28085315e+00 -6.64434016e-01 -4.19543415e-01 -1.09888233e-01
1.22930658e+00 -7.68485367e-01 -6.56631768e-01 5.77407330e-02
-8.85795534e-01 9.49289203e-02 -2.41802201e-01 7.93434322e-01
-5.63949756e-02 4.23917562e-01 -9.08187091e-01 -3.17732364e-01
-5.37564218e-01 -1.19348419e+00 1.02816904e+00 6.57859817e-02
-4.30709124e-01 -8.04143906e-01 4.06181030e-02 7.90657222e-01
3.85503381e-01 3.52026463e-01 1.51351357e+00 -6.85057163e-01
-8.01166356e-01 -3.71070713e-01 -6.62400603e-01 2.14805439e-01
3.62179130e-01 -3.08983773e-01 -1.17569971e+00 -5.63880391e-02
4.13373083e-01 -2.47667834e-01 9.22330558e-01 8.08378279e-01
1.94777107e+00 1.27055362e-01 -4.20703739e-01 6.95413411e-01
1.13507020e+00 2.29261681e-01 4.00344372e-01 1.32344425e-01
7.80531406e-01 6.45338818e-02 1.21132098e-01 3.83163065e-01
3.08454841e-01 -1.18904442e-01 6.37627542e-01 -5.74306786e-01
-2.94762820e-01 2.15877965e-02 -6.61270797e-01 8.97682130e-01
-3.44326831e-02 -3.63139898e-01 -1.41814888e+00 2.74061769e-01
-1.40953350e+00 -7.19083786e-01 -1.97537705e-01 1.63841808e+00
7.30874062e-01 1.55784205e-01 1.26042590e-02 1.96543097e-01
6.22978985e-01 -9.54239741e-02 -7.06294358e-01 6.70126602e-02
1.69690445e-01 5.91571510e-01 5.16289175e-01 -3.51477750e-02
-1.06385565e+00 1.79734658e-02 7.05541611e+00 1.27697781e-01
-1.41112566e+00 2.85839230e-01 1.18546665e+00 -2.50544578e-01
-1.34437412e-01 -6.92696869e-01 -2.54482955e-01 3.29610974e-01
7.94350445e-01 3.48953694e-01 6.74360804e-03 1.00609601e+00
3.97072881e-02 -5.37646823e-02 -1.30863762e+00 1.09712076e+00
1.43256821e-02 -1.71488476e+00 -2.61729360e-02 -8.80854670e-03
6.37264252e-01 4.17804122e-01 4.62303787e-01 -5.27026365e-03
6.76200911e-02 -1.42397547e+00 -3.79231125e-02 2.76606232e-01
1.14772654e+00 -3.88909072e-01 9.93637264e-01 7.63869360e-02
-5.65400541e-01 4.89675514e-02 -1.28007293e-01 2.71544129e-01
-1.88123271e-01 6.13156378e-01 -1.66562903e+00 2.66161352e-01
9.43349063e-01 3.17516446e-01 -5.27421176e-01 9.45383608e-01
1.05116302e-02 1.06529295e+00 -1.83927521e-01 -2.01368821e-03
2.05370858e-01 3.05569530e-01 1.29955307e-01 1.27162540e+00
2.37534031e-01 7.42758930e-01 4.50765975e-02 6.54565632e-01
-4.53666836e-01 6.97921887e-02 -4.52156574e-01 -6.75778612e-02
2.07590759e-01 1.18638408e+00 -8.84887218e-01 -5.87638617e-01
-6.00403488e-01 4.60541338e-01 -1.12918593e-01 -1.29948884e-01
-1.09260356e+00 1.43891707e-01 -1.01991594e-02 4.19378668e-01
4.90652323e-02 2.74274349e-01 -8.24917853e-01 -7.55908966e-01
-1.09760843e-01 -9.22197342e-01 4.68856096e-01 -9.60349381e-01
-1.21103036e+00 4.79093909e-01 -3.67887884e-01 -1.11008310e+00
-2.55876362e-01 -5.87887526e-01 -2.67806679e-01 6.08178914e-01
-1.62729204e+00 -9.74392116e-01 -6.09928071e-01 1.83177993e-01
4.40523237e-01 -1.09433651e-01 9.28408921e-01 4.29151207e-01
-2.16012359e-01 4.92532760e-01 2.41947159e-01 2.87627816e-01
6.13654912e-01 -1.40882528e+00 2.62848884e-01 4.45368439e-01
-1.86202645e-01 5.77916920e-01 2.89239109e-01 -3.87054533e-01
-1.28076899e+00 -1.22451687e+00 6.93676651e-01 -9.00106728e-01
2.88693577e-01 3.05338711e-01 -9.89585102e-01 6.24341428e-01
-2.29715228e-01 1.92056850e-01 1.13809812e+00 -2.36010134e-01
-4.10610467e-01 1.11218475e-01 -1.09743595e+00 3.20266813e-01
7.15506375e-01 -6.19090557e-01 -7.33726919e-01 6.95612907e-01
6.96111858e-01 -6.59288347e-01 -9.38720226e-01 5.96734107e-01
7.48834789e-01 -1.01684117e+00 1.13081551e+00 -8.18009198e-01
9.42692459e-01 1.36153951e-01 8.70747790e-02 -7.81537831e-01
-4.64349180e-01 -1.54223725e-01 2.09845472e-02 2.64859319e-01
4.41095591e-01 -3.42834264e-01 1.01331222e+00 4.90869522e-01
-1.99686810e-01 -1.08357036e+00 -5.23949564e-01 -1.70308605e-01
8.34650993e-02 -3.81003767e-01 5.42264700e-01 9.44453716e-01
-1.95028663e-01 3.48479003e-01 1.36087611e-01 1.98835775e-01
3.65329504e-01 -4.44422550e-02 3.88948053e-01 -1.19857919e+00
-8.15493405e-01 -4.80601043e-01 -8.62486288e-02 -7.31268585e-01
-3.61514360e-01 -1.11756229e+00 -2.04221681e-01 -1.84578979e+00
7.56141841e-01 -4.19199944e-01 -5.92494547e-01 4.94571120e-01
-6.21395230e-01 4.02365595e-01 1.76716840e-03 3.52446318e-01
-4.09616023e-01 -2.75585830e-01 1.38517094e+00 -2.00107858e-01
9.27923545e-02 2.75548231e-02 -6.28987849e-01 9.36454952e-01
7.53577650e-01 -7.06720054e-01 -3.59778434e-01 -6.18667364e-01
5.03018796e-01 5.64316988e-01 3.29445064e-01 -1.07848585e+00
-5.26657254e-02 2.55078748e-02 8.83708775e-01 -8.60551655e-01
1.71008289e-01 -7.76190460e-01 -5.90313673e-02 8.55417252e-01
-5.77484429e-01 1.46237433e-01 8.73970315e-02 5.66917717e-01
-4.30387370e-02 8.84202123e-02 1.01297605e+00 -4.15566057e-01
6.45816028e-02 3.72687340e-01 -4.47010279e-01 2.88395315e-01
7.10776091e-01 6.33884594e-02 -4.02588874e-01 -2.68204898e-01
-4.58613485e-01 -9.38466564e-03 4.76707518e-02 1.97469696e-01
6.99326336e-01 -7.82630682e-01 -7.98353136e-01 -6.55464604e-02
-3.79078053e-02 5.11352479e-01 4.97225881e-01 9.85580921e-01
-1.08155537e+00 7.27896869e-01 5.27486317e-02 -1.05207288e+00
-1.32870817e+00 5.59428751e-01 3.84119838e-01 -6.13516867e-01
-9.36522901e-01 9.41205263e-01 6.28438950e-01 -1.86821863e-01
3.23087536e-02 -8.17162812e-01 8.55754018e-02 -5.47744870e-01
6.16725147e-01 4.61107969e-01 3.32204312e-01 1.45585658e-02
-3.17206144e-01 3.03002119e-01 -5.09480774e-01 2.73301691e-01
1.56935346e+00 1.41104087e-01 1.07812388e-02 2.27570891e-01
1.18837249e+00 -3.23399514e-01 -6.91707611e-01 -5.94768813e-03
-2.23555729e-01 -2.18869969e-01 2.95331150e-01 -1.16459477e+00
-9.83825386e-01 1.30253482e+00 8.78896654e-01 1.95083380e-01
1.03417432e+00 2.73837119e-01 1.12294149e+00 6.12301052e-01
-2.55071163e-01 -6.75726354e-01 4.46267992e-01 1.32655904e-01
3.51634979e-01 -1.50927496e+00 2.67054349e-01 -3.04852635e-01
-4.02555466e-01 1.28024852e+00 3.60494405e-01 1.42346814e-01
6.53535187e-01 4.04920816e-01 3.11056137e-01 -4.85216290e-01
-5.43306768e-01 2.65045166e-01 5.99172711e-02 4.22247976e-01
8.04715812e-01 3.07015449e-01 1.55215338e-01 5.24107516e-01
-8.33794996e-02 2.71618396e-01 4.78616148e-01 9.50483024e-01
-5.78152359e-01 -6.95041239e-01 -4.58184689e-01 1.04810905e+00
-1.21483207e+00 -2.84159571e-01 -3.79703432e-01 9.10458088e-01
7.13169500e-02 6.80195808e-01 -1.84323356e-01 -1.48116559e-01
2.98787672e-02 7.40173087e-02 4.83011842e-01 -7.36821890e-01
-6.44940078e-01 5.13360389e-02 -2.99657702e-01 -2.61665314e-01
-2.87480086e-01 -2.82735735e-01 -1.79098618e+00 -2.45234966e-01
-2.97822148e-01 -2.51577824e-01 6.37224138e-01 9.66428578e-01
2.37942740e-01 1.21826565e+00 5.47368050e-01 4.04288387e-03
-7.83030033e-01 -7.82089472e-01 -2.68910319e-01 3.09900969e-01
5.49551904e-01 -1.94816604e-01 -1.24606274e-01 2.07998380e-01] | [15.204936027526855, -2.000410318374634] |
d0a2f71b-6af2-4991-bf5c-c2af75125faa | dialogue-act-sequence-labeling-using | 1709.04250 | null | http://arxiv.org/abs/1709.04250v2 | http://arxiv.org/pdf/1709.04250v2.pdf | Dialogue Act Sequence Labeling using Hierarchical encoder with CRF | Dialogue Act recognition associate dialogue acts (i.e., semantic labels) to
utterances in a conversation. The problem of associating semantic labels to
utterances can be treated as a sequence labeling problem. In this work, we
build a hierarchical recurrent neural network using bidirectional LSTM as a
base unit and the conditional random field (CRF) as the top layer to classify
each utterance into its corresponding dialogue act. The hierarchical network
learns representations at multiple levels, i.e., word level, utterance level,
and conversation level. The conversation level representations are input to the
CRF layer, which takes into account not only all previous utterances but also
their dialogue acts, thus modeling the dependency among both, labels and
utterances, an important consideration of natural dialogue. We validate our
approach on two different benchmark data sets, Switchboard and Meeting Recorder
Dialogue Act, and show performance improvement over the state-of-the-art
methods by $2.2\%$ and $4.1\%$ absolute points, respectively. It is worth
noting that the inter-annotator agreement on Switchboard data set is $84\%$,
and our method is able to achieve the accuracy of about $79\%$ despite being
trained on the noisy data. | ['Sachindra Joshi', 'Harshit Kumar', 'Riddhiman Dasgupta', 'Arvind Agarwal', 'Arun Kumar'] | 2017-09-13 | null | null | null | null | ['dialogue-act-classification'] | ['natural-language-processing'] | [ 2.93019474e-01 6.83093846e-01 -2.22156998e-02 -9.21680391e-01
-6.00842357e-01 -3.11323136e-01 6.96743309e-01 2.19892636e-01
-5.06163299e-01 9.19212580e-01 5.77200115e-01 -1.38947830e-01
4.94438529e-01 -7.73196161e-01 -1.81455195e-01 -4.48162079e-01
1.84446514e-01 7.71537364e-01 -4.03660424e-02 -4.69425023e-01
4.49462123e-02 -2.58797705e-02 -1.12136602e+00 6.63970470e-01
5.90907037e-01 9.70710397e-01 -1.51885077e-01 7.10233927e-01
-6.30257487e-01 1.44930816e+00 -9.46890116e-01 -4.07638043e-01
-2.67882049e-01 -6.30702317e-01 -1.49325883e+00 3.53774726e-01
-8.59743133e-02 -3.56735200e-01 -1.98999226e-01 8.44549000e-01
1.11572236e-01 6.01332068e-01 6.18373573e-01 -1.02175379e+00
-3.65770787e-01 8.66007388e-01 -1.25245256e-02 -2.00167730e-01
5.23663461e-01 -8.35411102e-02 1.17009008e+00 -5.84715664e-01
3.69549870e-01 1.48566258e+00 3.41678739e-01 8.61087382e-01
-1.13456285e+00 -6.60636365e-01 2.68635005e-01 6.52260110e-02
-1.07011807e+00 -6.49700463e-01 5.99146366e-01 -4.54049587e-01
1.41127229e+00 1.57626912e-01 1.29903525e-01 1.03542984e+00
-1.28552303e-01 7.60622382e-01 1.34326160e+00 -6.59257650e-01
1.72310755e-01 3.14665705e-01 8.40233207e-01 5.23706317e-01
-9.45893347e-01 -3.59675139e-01 -3.85449111e-01 -1.58148348e-01
5.46549141e-01 -3.49801295e-02 -5.65900207e-02 6.15913928e-01
-7.21853971e-01 1.01217628e+00 3.90963405e-01 4.79809791e-01
-3.36805522e-01 -9.22700018e-02 6.39857948e-01 4.07421947e-01
6.43516004e-01 2.71248072e-01 -6.28001809e-01 -3.85419637e-01
-3.08004349e-01 1.02336742e-02 1.24666822e+00 9.13279653e-01
7.97416091e-01 -5.62688299e-02 -3.54336113e-01 1.44372296e+00
3.97650510e-01 8.51145387e-02 4.97390062e-01 -8.93943727e-01
4.53076839e-01 8.02115798e-01 1.60683468e-01 -8.10966730e-01
-6.61004305e-01 1.30386442e-01 -1.00633132e+00 -2.48048052e-01
2.90158749e-01 -4.25077528e-01 -7.66261160e-01 1.89822316e+00
1.03315271e-01 8.89508650e-02 3.85927409e-01 6.20537639e-01
1.28595245e+00 9.18642700e-01 4.55616474e-01 -5.76800227e-01
1.53577161e+00 -1.12597096e+00 -1.18141663e+00 -2.60094047e-01
8.43970776e-01 -6.91115022e-01 7.33909667e-01 7.29513764e-02
-1.07953012e+00 -6.04595542e-01 -6.93145812e-01 -1.28871156e-02
-3.71746868e-01 2.49299586e-01 4.06432629e-01 3.05150777e-01
-9.81514215e-01 1.81470513e-01 -5.66877902e-01 -3.37750256e-01
-2.96841264e-01 3.97089690e-01 -2.54884183e-01 2.46522292e-01
-1.58538938e+00 1.17442513e+00 4.75946903e-01 1.73283949e-01
-7.29205787e-01 4.79962863e-02 -1.00963783e+00 2.09157392e-01
2.65237957e-01 -2.76743639e-02 1.64147532e+00 -6.76848233e-01
-2.21602416e+00 9.11217213e-01 -5.17331898e-01 -5.52402794e-01
4.87853698e-02 -8.03860649e-02 -4.12990987e-01 -2.80224502e-01
-1.66875169e-01 6.95827603e-01 3.57466042e-01 -1.13176835e+00
-8.06410193e-01 -1.76804572e-01 4.29448992e-01 4.59152102e-01
-1.19276680e-01 2.23008871e-01 -2.20656499e-01 -1.26095831e-01
1.67980641e-01 -1.09055185e+00 -1.07020922e-01 -9.45996940e-01
-5.60562849e-01 -1.02586615e+00 4.51388448e-01 -8.47365975e-01
1.19946086e+00 -2.00131226e+00 9.10340846e-02 -1.15115047e-01
1.01798333e-01 1.40461773e-01 2.04408705e-01 5.52791536e-01
5.12707569e-02 4.57456931e-02 -2.99376726e-01 -7.46911287e-01
1.43152118e-01 3.28453004e-01 -2.38332972e-01 2.18200386e-01
9.42233577e-02 5.76137602e-01 -7.45796323e-01 -3.36518109e-01
3.87632102e-01 2.20183238e-01 -1.55435562e-01 8.96356642e-01
-3.61736953e-01 6.28042698e-01 -2.87559628e-01 5.02199344e-02
3.60275358e-01 -2.60487378e-01 7.17188179e-01 2.03915429e-03
-4.58511710e-02 9.52985704e-01 -7.93122292e-01 1.63553059e+00
-8.53537440e-01 6.30435169e-01 1.53897390e-01 -1.10657632e+00
1.46431744e+00 9.19136822e-01 1.99807450e-01 -6.04552269e-01
2.31014833e-01 7.05004409e-02 6.06777193e-03 -3.58331382e-01
6.58518255e-01 -4.33734745e-01 -5.78533709e-01 6.85827971e-01
3.72804016e-01 -1.00485489e-01 4.31196541e-02 1.28066525e-01
1.01430619e+00 -1.96838588e-01 3.77364308e-01 -4.40403819e-02
9.00437236e-01 -2.30201408e-01 6.71660960e-01 4.58118618e-01
-1.36764854e-01 9.24887694e-03 5.62302291e-01 -4.93490040e-01
-6.68969214e-01 -5.93989432e-01 -2.85390727e-02 1.60863006e+00
-1.62221678e-02 -3.53898495e-01 -8.01289976e-01 -8.17337632e-01
-4.65428919e-01 8.94397557e-01 -4.57889289e-01 2.54059546e-02
-8.42830002e-01 -4.82293934e-01 5.96597791e-01 4.77843881e-01
8.66718650e-01 -1.55512249e+00 -2.62442052e-01 4.15382832e-01
-5.70838332e-01 -1.38554990e+00 -3.34968492e-02 3.41684014e-01
-3.86422575e-01 -7.68845081e-01 -1.93689883e-01 -9.60061789e-01
5.17904878e-01 -2.60343015e-01 1.17902994e+00 1.35494888e-01
3.80461365e-01 -5.72237819e-02 -6.45376146e-01 -7.79804289e-02
-9.37916815e-01 1.39133751e-01 -3.05305012e-02 1.24632351e-01
7.21104205e-01 -3.48373353e-01 -1.06927536e-01 5.48576534e-01
-4.89789218e-01 3.04068983e-01 9.39369202e-02 1.09998095e+00
-6.64733723e-02 -8.88329446e-02 8.15885007e-01 -1.15831411e+00
8.06777060e-01 -3.70549500e-01 -2.13967428e-01 2.15128452e-01
-1.19840227e-01 -1.64642856e-02 7.38674581e-01 1.41381785e-01
-1.51411343e+00 2.63238102e-02 -5.99842250e-01 2.12139487e-01
-7.57504523e-01 4.15920794e-01 -2.04284377e-02 6.45277381e-01
4.00237322e-01 1.78586319e-02 -1.69704482e-01 -5.75473845e-01
4.37888980e-01 1.29936326e+00 3.73399168e-01 -4.41167146e-01
7.38798752e-02 2.36023832e-02 -5.77425718e-01 -8.52754116e-01
-1.00776553e+00 -6.96456075e-01 -7.53221154e-01 -4.10641193e-01
1.18759513e+00 -7.52878964e-01 -1.37740409e+00 4.28847104e-01
-1.53071070e+00 -4.79757011e-01 3.17612216e-02 4.04032648e-01
-6.11040175e-01 8.05636197e-02 -1.25830233e+00 -1.21743870e+00
-3.33047062e-01 -1.20006716e+00 7.49844134e-01 3.10778797e-01
-6.91383660e-01 -1.04684401e+00 -1.50021121e-01 6.58125162e-01
3.43202800e-01 1.25208557e-01 1.01121175e+00 -1.31611562e+00
4.32203077e-02 -5.16276062e-02 -2.80482113e-01 6.17019176e-01
3.26835752e-01 -4.07129496e-01 -1.36004984e+00 -1.37355164e-01
3.73868942e-01 -9.31102514e-01 5.62388122e-01 5.23447767e-02
6.07084870e-01 -4.57061619e-01 -1.64913327e-01 -3.12251151e-01
7.77368069e-01 7.56337345e-01 6.28443539e-01 -1.25526428e-01
5.19989431e-01 1.15207815e+00 6.13547862e-01 4.48954880e-01
5.17994821e-01 9.12446022e-01 1.06953830e-01 -1.56658724e-01
1.20828293e-01 -8.75701681e-02 3.09049964e-01 1.41242218e+00
2.27133140e-01 -2.80346990e-01 -1.07949626e+00 2.31115744e-01
-2.05265927e+00 -6.27672255e-01 -7.21639916e-02 1.84344649e+00
1.24024189e+00 2.10351303e-01 4.40249741e-02 -9.69470218e-02
9.73175049e-01 2.96344936e-01 -7.57319704e-02 -8.72943819e-01
2.38120630e-01 1.50762171e-01 8.39824695e-03 9.35633898e-01
-1.18698657e+00 1.24721181e+00 5.62084484e+00 5.95814049e-01
-9.51672971e-01 2.09807172e-01 8.66203189e-01 4.16110694e-01
1.48144379e-01 -8.56621489e-02 -8.87545764e-01 4.44682777e-01
1.34184551e+00 1.11871086e-01 3.77942234e-01 6.28631532e-01
1.66120246e-01 -4.78265248e-02 -1.22329092e+00 7.18246937e-01
1.00882821e-01 -1.19218910e+00 -4.13868934e-01 -2.23714381e-01
4.65115815e-01 -1.35396898e-01 -4.78909820e-01 8.29890668e-01
9.91856396e-01 -1.31918526e+00 1.94377095e-01 3.88760239e-01
6.23023808e-01 -6.18111849e-01 1.20424616e+00 6.57415092e-01
-1.13589013e+00 1.48335382e-01 -1.22875467e-01 -4.06745672e-01
4.25084859e-01 1.67273715e-01 -1.45144629e+00 4.68718290e-01
5.33073545e-01 6.58065677e-01 1.54710039e-01 1.33994550e-01
-4.40106660e-01 7.84226775e-01 -2.09254339e-01 -2.09551677e-01
3.75441521e-01 -2.83402234e-01 7.03403121e-03 1.52615476e+00
-3.20688486e-01 5.81596076e-01 6.67385995e-01 4.92922962e-01
-3.98438632e-01 1.83359817e-01 -3.29438180e-01 7.06482977e-02
5.21412849e-01 1.07366168e+00 -4.97723520e-01 -6.37616754e-01
-3.44569266e-01 7.96315730e-01 6.01534843e-01 3.02253157e-01
-5.71443200e-01 -2.44521916e-01 4.60415274e-01 -6.24952257e-01
-2.37124905e-01 -1.21170029e-01 -9.37995613e-02 -8.55758011e-01
-1.53686911e-01 -8.20328832e-01 3.83009166e-01 -3.11099231e-01
-1.30125320e+00 9.99769568e-01 -1.94337890e-01 -9.04674590e-01
-7.70627618e-01 -4.20042098e-01 -5.51787257e-01 1.09809566e+00
-1.21993351e+00 -9.45467651e-01 -2.49542966e-01 4.19247746e-01
1.03109074e+00 -1.44659445e-01 1.52408493e+00 9.89491194e-02
-5.63752115e-01 3.98417145e-01 -9.63721797e-02 6.94551110e-01
6.36906445e-01 -1.12162757e+00 3.31203282e-01 1.40060544e-01
-5.36923856e-02 4.49165136e-01 5.27691066e-01 -4.14651990e-01
-6.60444677e-01 -7.70040631e-01 1.33360755e+00 -3.00172776e-01
4.64055508e-01 -5.88990688e-01 -1.10292351e+00 9.43263471e-01
6.00911379e-01 -5.92419922e-01 9.53749061e-01 6.11889899e-01
-1.44941390e-01 2.46044263e-01 -1.10227573e+00 3.43169600e-01
6.58940673e-01 -8.34294915e-01 -9.71700847e-01 4.61352229e-01
1.04878795e+00 -5.09083092e-01 -1.22352993e+00 3.51538897e-01
2.10778356e-01 -8.68060291e-01 4.16779548e-01 -6.01197600e-01
3.72704595e-01 1.12813599e-01 -1.78689510e-01 -1.29455757e+00
-1.19320266e-01 -5.07921457e-01 2.91818291e-01 1.61260355e+00
6.42024338e-01 -5.21589518e-01 6.46584630e-01 1.02372515e+00
-3.57705951e-01 -5.72501183e-01 -9.86678898e-01 -2.62320191e-01
1.13434186e-02 -4.11847353e-01 4.07789052e-01 1.21799874e+00
7.15122163e-01 1.07302415e+00 -6.44148469e-01 -3.15313667e-01
-8.60737711e-02 2.70339493e-02 6.43501639e-01 -1.11562657e+00
-2.10270658e-01 -1.95410624e-01 -1.13319397e-01 -1.52275765e+00
7.42820263e-01 -6.50738895e-01 6.23122454e-01 -1.55819654e+00
-1.77692354e-01 -6.29216135e-01 -1.39335826e-01 6.57609999e-01
-9.11877006e-02 -9.99869108e-02 1.85867012e-01 2.56658763e-01
-6.53503060e-01 7.54676342e-01 7.15735078e-01 -2.70643562e-01
-4.15598780e-01 1.75910234e-01 -1.29940137e-01 9.66059923e-01
9.93416727e-01 -3.65109414e-01 -9.36417878e-02 -2.83708006e-01
-3.38703811e-01 6.18840039e-01 -2.18600899e-01 -5.62703252e-01
3.23177755e-01 -1.75086394e-01 -1.94827184e-01 -3.59954089e-01
8.19297731e-01 -6.12033546e-01 -2.96123415e-01 2.98571348e-01
-1.06070602e+00 -2.45973498e-01 5.78348637e-02 3.25087458e-01
-5.81408799e-01 -3.52564305e-01 6.94292009e-01 -3.10446411e-01
-6.11331522e-01 -2.88067758e-01 -5.64129651e-01 -1.20006651e-01
6.40079737e-01 2.89787203e-01 -4.33552325e-01 -6.99503005e-01
-1.11226833e+00 3.99598509e-01 -9.59616825e-02 5.08054435e-01
3.71364474e-01 -1.07051599e+00 -7.85318911e-01 3.26288752e-02
-7.65178679e-03 2.58604195e-02 3.16416055e-01 4.50461000e-01
-1.67758867e-01 7.25728750e-01 -9.61932093e-02 -5.49253821e-01
-1.32647943e+00 5.37170619e-02 3.10894549e-01 -6.76942110e-01
-2.51300752e-01 9.75053072e-01 1.79302856e-01 -8.66470754e-01
5.83076477e-01 -1.82860538e-01 -8.20138216e-01 8.05361941e-02
3.31853092e-01 1.77181542e-01 -9.76577625e-02 -9.60831523e-01
-3.76002342e-01 1.59230500e-01 -3.11993301e-01 -2.44909614e-01
9.95609641e-01 -2.12157279e-01 -3.90084356e-01 8.68380070e-01
1.31835747e+00 -4.21319455e-01 -8.65310311e-01 -7.18707621e-01
2.04491213e-01 -1.23077957e-02 -3.56004387e-01 -8.11390638e-01
-6.12618566e-01 8.52196932e-01 2.85578221e-01 8.26065660e-01
8.01729739e-01 7.48881549e-02 8.31202030e-01 5.13248622e-01
4.05031830e-01 -1.25180471e+00 2.57633273e-02 1.14808607e+00
8.63974750e-01 -1.35196686e+00 -4.29258317e-01 -5.59627891e-01
-9.14645612e-01 9.59962964e-01 8.79379272e-01 -1.05582736e-01
5.46499074e-01 3.56789082e-02 4.07946169e-01 -1.73004881e-01
-9.70036566e-01 -6.43806607e-02 -1.02305777e-01 1.89617023e-01
9.90509987e-01 3.84286940e-01 -3.42783839e-01 5.08338332e-01
-2.70000398e-01 -2.12685928e-01 3.74665648e-01 7.61534095e-01
-5.21583974e-01 -1.28572345e+00 -1.20172173e-01 3.36866587e-01
-4.42719579e-01 -8.99245590e-02 -5.37898839e-01 3.94591451e-01
-1.23915754e-01 1.62141979e+00 3.24742556e-01 -5.67995548e-01
4.20809209e-01 6.16832376e-01 -2.98235506e-01 -1.04543674e+00
-1.12995279e+00 -1.16119131e-01 7.85339177e-01 -2.18305022e-01
-6.67903364e-01 -1.98935911e-01 -1.65820432e+00 -2.74482101e-01
-3.37495834e-01 7.20218480e-01 5.44245720e-01 1.48491228e+00
-1.52857704e-02 6.52478099e-01 8.79592121e-01 -5.86981177e-01
-4.29157317e-01 -1.61582923e+00 -3.79075587e-01 7.52432346e-01
1.80061787e-01 -4.48501348e-01 -1.92633182e-01 1.57729656e-01] | [12.745969772338867, 7.699461460113525] |
fd055e81-81ad-4f12-a4f8-91e65419330d | target-tracking-in-real-time-surveillance | 1506.06659 | null | http://arxiv.org/abs/1506.06659v1 | http://arxiv.org/pdf/1506.06659v1.pdf | Target Tracking In Real Time Surveillance Cameras and Videos | Security concerns has been kept on increasing, so it is important for
everyone to keep their property safe from thefts and destruction. So the need
for surveillance techniques are also increasing. The system has been developed
to detect the motion in a video. A system has been developed for real time
applications by using the techniques of background subtraction and frame
differencing. In this system, motion is detected from the webcam or from the
real time video. Background subtraction and frames differencing method has been
used to detect the moving target. In background subtraction method, current
frame is subtracted from the referenced frame and then the threshold is
applied. If the difference is greater than the threshold then it is considered
as the pixel from the moving object, otherwise it is considered as background
pixel. Similarly, two frames difference method takes difference between two
continuous frames. Then that resultant difference frame is thresholded and the
amount of difference pixels is calculated. | ['Nayyab Naseem', 'Mehreen Sirshar'] | 2015-06-22 | null | null | null | null | ['video-background-subtraction'] | ['computer-vision'] | [ 7.04306185e-01 -4.64586258e-01 1.60978973e-01 5.16680852e-02
7.69078061e-02 -4.33172286e-01 3.87757599e-01 9.99973044e-02
-5.69509804e-01 6.63501024e-01 -8.18283632e-02 -2.52676904e-01
5.08451641e-01 -9.48666453e-01 -3.33403237e-02 -9.63752866e-01
2.83756226e-01 -5.36876380e-01 1.06918001e+00 -4.17528898e-02
3.40794414e-01 5.26354849e-01 -1.34135759e+00 4.12317216e-02
2.46251628e-01 9.16755199e-01 2.53802478e-01 8.91298950e-01
-4.13130432e-01 9.27126527e-01 -8.50546479e-01 2.27879092e-01
6.47631347e-01 -9.37801659e-01 -4.62715447e-01 2.51987547e-01
1.49621934e-01 -7.17223704e-01 -1.94175676e-01 1.36085188e+00
2.85870761e-01 4.86214101e-01 2.34239668e-01 -1.10845852e+00
2.15009436e-01 -2.88162082e-01 -9.19866502e-01 8.64744365e-01
3.94666314e-01 -1.68581545e-01 -4.15099896e-02 -5.18642306e-01
3.61812443e-01 1.08648396e+00 1.29512668e-01 3.65821928e-01
-8.27485323e-01 -7.49402344e-01 -2.07684830e-01 2.27488726e-01
-1.20711100e+00 -7.21786991e-02 6.80992901e-01 -4.08572674e-01
6.99529827e-01 3.36153626e-01 7.75356531e-01 1.97172463e-01
5.80475330e-01 2.61495322e-01 9.59513843e-01 -8.03911030e-01
1.51630968e-01 -1.40704904e-02 3.25256228e-01 3.97977471e-01
5.04004478e-01 -2.89179124e-02 5.22455312e-02 -1.41196638e-01
7.18319774e-01 2.84299374e-01 -3.18584532e-01 1.97947770e-01
-9.23132241e-01 4.34435874e-01 -1.80128425e-01 6.93394005e-01
-5.52151382e-01 -6.52783364e-02 3.30573618e-01 8.93482119e-02
1.53767169e-01 -4.85938638e-01 -1.25987500e-01 -3.37610602e-01
-1.10392189e+00 1.30653501e-01 5.00910580e-01 2.28286743e-01
5.30568838e-01 7.97867402e-02 1.59269348e-01 1.76369503e-01
2.22682104e-01 5.05107582e-01 3.87900174e-01 -9.07262921e-01
1.05453707e-01 9.69195902e-01 3.40377182e-01 -1.40854478e+00
8.38369057e-02 3.20709348e-01 -4.96526003e-01 9.30971920e-01
3.20017159e-01 -2.85158426e-01 -9.49446380e-01 1.09677243e+00
6.10519528e-01 3.43301713e-01 1.32145777e-01 5.97249329e-01
6.63462639e-01 1.15910995e+00 1.30750582e-01 -7.81033695e-01
1.32116902e+00 -2.73438275e-01 -9.23449218e-01 -4.62193452e-02
7.32652247e-02 -1.08660710e+00 3.01265538e-01 5.30820489e-01
-8.45030248e-01 -7.55150616e-01 -1.19229531e+00 4.07103866e-01
-3.06170434e-01 -2.14789510e-01 -1.77861497e-01 6.08074665e-01
-4.80200976e-01 2.29334891e-01 -8.32451463e-01 -5.62594116e-01
-1.28849804e-01 3.26362133e-01 -3.38312477e-01 1.74917981e-01
-9.53091621e-01 9.69336867e-01 6.43682957e-01 1.95609033e-03
-3.87815803e-01 1.64567173e-01 -5.48794329e-01 -2.54442096e-01
1.71277270e-01 -1.41816944e-01 8.81831110e-01 -1.66015887e+00
-1.09876478e+00 7.19574153e-01 -3.56051713e-01 -3.88790101e-01
5.36505938e-01 9.12771095e-03 -8.27929139e-01 1.53831571e-01
-3.16630118e-02 -8.86030123e-03 8.56723070e-01 -8.78163517e-01
-1.15129375e+00 -3.40510488e-01 1.86464898e-02 3.15627009e-01
8.85508806e-02 6.58768296e-01 -3.80677134e-01 -2.39341825e-01
2.02413008e-01 -7.60538757e-01 2.31217239e-02 -1.64361775e-01
9.18380544e-02 1.49675876e-01 1.80740643e+00 -1.01879418e+00
1.45088887e+00 -2.43830991e+00 -5.99973023e-01 1.34186357e-01
-2.56318420e-01 7.27689981e-01 3.80325139e-01 7.64863417e-02
-1.96951166e-01 -1.31972998e-01 -1.79343462e-01 6.68877840e-01
-9.19064522e-01 6.56955689e-03 1.52988598e-01 3.36515576e-01
2.79567530e-03 -2.88519114e-01 -6.35648429e-01 -5.92128634e-01
4.09649462e-01 5.07956982e-01 1.81650668e-01 1.35809496e-01
1.57665923e-01 5.62611759e-01 -4.74519670e-01 4.07970041e-01
1.07473373e+00 4.86321747e-01 -9.50476453e-02 -1.84333473e-01
-6.52002871e-01 -1.39490321e-01 -1.61959624e+00 5.52624762e-01
1.05799951e-01 9.03457403e-01 3.47758532e-01 -8.98152530e-01
9.53041434e-01 6.40595257e-01 5.43863177e-01 -6.95620179e-01
5.96214235e-01 -5.65931275e-02 1.46025047e-01 -7.89936364e-01
4.30936307e-01 -1.17557764e-01 3.84464085e-01 8.87496173e-02
-7.91810095e-01 2.71182984e-01 4.75704342e-01 4.08719443e-02
1.00231600e+00 3.13788503e-01 6.89564943e-01 5.15857860e-02
9.79642928e-01 2.29332268e-01 7.99076259e-01 1.32462785e-01
-4.81730491e-01 1.37144849e-01 1.59715116e-01 -5.06719768e-01
-7.17668533e-01 -7.90466785e-01 1.81766719e-01 5.80492198e-01
5.52001595e-01 8.84357542e-02 -8.00735295e-01 -3.33909035e-01
-2.32181773e-01 3.47151875e-01 -2.29483813e-01 -2.11926606e-02
-8.00521374e-01 -4.26066130e-01 -1.03383243e-01 1.27142504e-01
1.07837510e+00 -1.01699317e+00 -1.32780564e+00 4.88166839e-01
-3.64467651e-02 -8.98946822e-01 -2.41096124e-01 -3.92086089e-01
-9.60103095e-01 -1.08644855e+00 -6.26747906e-01 -8.27418447e-01
6.52526200e-01 9.11408067e-01 2.53415227e-01 2.92333603e-01
-5.52110374e-01 5.28884120e-02 -3.22133392e-01 -7.27900207e-01
-5.94252765e-01 -8.10651779e-01 -2.89353907e-01 -2.48294212e-02
6.46432757e-01 -8.08450803e-02 -6.73138320e-01 1.39797062e-01
-9.47700202e-01 -4.03585695e-02 1.81627154e-01 2.44220980e-02
3.34851205e-01 7.79097319e-01 6.02075234e-02 -4.02964205e-01
3.32638234e-01 -2.60511518e-01 -9.64456260e-01 4.22425196e-02
8.34452435e-02 -4.32011336e-01 2.88198590e-01 -2.62695432e-01
-1.13424110e+00 3.29749361e-02 3.52276981e-01 2.69957893e-02
-3.42482418e-01 9.40311179e-02 -3.76735181e-01 5.20997494e-02
1.74046084e-01 2.25885779e-01 9.18002948e-02 -5.66052198e-02
-3.89734000e-01 8.64518464e-01 6.03834867e-01 3.97230625e-01
4.96392846e-01 6.48092210e-01 -5.32565871e-03 -1.17060840e+00
-3.81043777e-02 -6.93867087e-01 -3.97153884e-01 -7.03536749e-01
1.37424457e+00 -5.35621226e-01 -2.71797568e-01 6.74049318e-01
-1.34410858e+00 3.27287197e-01 5.39870203e-01 7.53258705e-01
1.96136475e-01 6.47620142e-01 -3.05096120e-01 -1.43148422e+00
-5.28620660e-01 -8.53449225e-01 1.72217786e-01 9.67103899e-01
-1.74708858e-01 -7.23932326e-01 9.28283408e-02 1.13515653e-01
2.52356201e-01 6.15959346e-01 5.87810636e-01 -1.84509367e-01
-5.91538906e-01 -6.56761646e-01 -1.08976915e-01 5.26722729e-01
7.91359007e-01 5.69445074e-01 -5.14305413e-01 -1.12555420e-03
5.00938535e-01 7.38518894e-01 3.48540932e-01 4.27258253e-01
4.94931042e-01 -1.04610607e-01 -5.69773495e-01 -2.91331820e-02
1.66389120e+00 1.40662050e+00 9.68537688e-01 6.56466544e-01
3.05516422e-01 3.52840245e-01 9.42125142e-01 1.25985861e-01
-2.06240207e-01 4.49251443e-01 1.14264153e-01 -3.96349370e-01
9.63755399e-02 3.53260756e-01 4.81405795e-01 3.00753534e-01
-3.12271923e-01 -3.04777801e-01 -7.50883520e-01 3.45441490e-01
-1.52396166e+00 -1.54234433e+00 -6.36821210e-01 2.63548160e+00
4.75282729e-01 2.15657100e-01 1.12072907e-01 5.28366864e-01
1.24137330e+00 -6.65757880e-02 -2.33754352e-01 -5.18121362e-01
1.08938850e-01 9.26275849e-02 5.29772818e-01 4.58699346e-01
-9.47094798e-01 5.23167253e-01 5.65187263e+00 2.35219404e-01
-1.39861238e+00 -8.69881245e-04 4.75856751e-01 7.26761818e-02
4.31670755e-01 2.84875870e-01 -6.80909932e-01 8.56973767e-01
5.46820581e-01 -2.91135699e-01 8.45911205e-02 4.75300759e-01
8.61501992e-01 -1.24039853e+00 -2.51917362e-01 9.49280977e-01
1.48324981e-01 -5.75250149e-01 -2.20586866e-01 1.55770406e-01
4.56995636e-01 -6.40231907e-01 -4.32061791e-01 -1.23997860e-01
-1.39038518e-01 -2.64132440e-01 4.30681407e-01 3.73720139e-01
7.99288526e-02 -7.36553013e-01 8.85032475e-01 4.30586755e-01
-1.25323784e+00 1.14157163e-01 -2.12722301e-01 -4.34879243e-01
5.01151562e-01 4.64038879e-01 -7.34209239e-01 1.96070686e-01
6.28390908e-01 1.25439856e-02 -2.25668028e-01 1.22326005e+00
3.79728049e-01 4.75199640e-01 -4.50428367e-01 1.32947966e-01
1.22919165e-01 -6.92285419e-01 6.03306413e-01 1.07844985e+00
5.53169847e-01 4.30855542e-01 6.18386380e-02 3.00221294e-01
5.13924778e-01 2.42257476e-01 -6.96071208e-01 1.01935789e-01
4.35248286e-01 1.10945237e+00 -1.00216341e+00 -7.25718319e-01
-7.42365777e-01 1.19728208e+00 -7.51581550e-01 6.46356493e-02
-9.68956947e-01 -6.05299056e-01 3.41438770e-01 4.73897785e-01
2.27575630e-01 -3.67939442e-01 8.92248079e-02 -5.00377953e-01
-6.99611008e-02 -6.63514674e-01 4.80506480e-01 -7.86252618e-01
-3.46110493e-01 3.49390715e-01 1.71537504e-01 -1.27623296e+00
-2.56307833e-02 -4.62357074e-01 -1.13112092e+00 9.27326977e-01
-8.61204565e-01 -3.68451983e-01 -5.91918528e-01 5.83767116e-01
5.34839094e-01 -6.06260858e-02 4.23391134e-01 2.87141979e-01
-5.59426963e-01 -2.39936993e-01 -2.98612062e-02 4.21071291e-01
4.28559840e-01 -5.25079191e-01 -8.76860246e-02 1.42952204e+00
-4.19042110e-01 4.83919501e-01 8.42561245e-01 -1.04663503e+00
-7.63523817e-01 -4.30701762e-01 7.02770650e-01 -1.43209677e-02
1.77673399e-01 1.75075233e-01 -8.95980060e-01 3.23259205e-01
4.36272055e-01 -2.54071206e-01 5.12642682e-01 -1.05340660e+00
3.74319941e-01 -1.79278269e-01 -1.51144946e+00 5.38724124e-01
2.10988700e-01 -6.05071560e-02 -5.99779487e-01 -2.28875890e-01
7.31701255e-02 -1.67936027e-01 -3.13440263e-01 1.79182768e-01
6.52150869e-01 -1.18611801e+00 6.05408728e-01 7.67056784e-03
-2.39787444e-01 -9.75105166e-01 6.06534146e-02 -6.25630498e-01
-2.42313847e-01 -3.94115210e-01 5.55749893e-01 1.39827752e+00
-1.19130118e-02 -4.58070278e-01 7.62001097e-01 8.68369520e-01
5.80920219e-01 1.36125103e-01 -7.52004087e-01 -5.33382773e-01
-5.79109132e-01 1.25214666e-01 1.47282436e-01 7.51346111e-01
-2.98987795e-02 1.92276120e-01 -1.64653137e-01 1.97823450e-01
5.40265918e-01 -3.24354291e-01 5.99834561e-01 -1.18473041e+00
1.33303821e-01 -7.87895098e-02 -7.21710861e-01 -4.09759551e-01
-5.80454767e-01 2.06572600e-02 -1.67915389e-01 -1.66291213e+00
1.54210851e-01 2.77276307e-01 -1.30292565e-01 7.60927377e-03
-3.91368389e-01 1.37629822e-01 2.94790387e-01 1.49135277e-01
3.99426408e-02 -3.96070600e-01 8.26603353e-01 2.32197866e-01
-5.04067242e-01 1.47842407e-01 -3.67543697e-02 8.40873659e-01
9.66888905e-01 -3.75300080e-01 -2.70058811e-01 6.41082674e-02
-3.37200373e-01 2.96367794e-01 2.12160513e-01 -1.35933864e+00
1.04707487e-01 -4.67564970e-01 6.85185611e-01 -8.59172046e-01
1.43192083e-01 -1.33689296e+00 5.62710881e-01 8.08067143e-01
3.87606919e-01 3.60917389e-01 3.61761212e-01 3.82183999e-01
-1.44097894e-01 -5.96145451e-01 9.30389047e-01 -4.48407769e-01
-1.12087500e+00 -1.11445419e-01 -1.02733374e+00 -5.23378611e-01
1.54183912e+00 -1.19994605e+00 3.50598828e-03 -3.24393988e-01
-3.73147398e-01 -2.36223459e-01 5.53470671e-01 1.38263091e-01
4.55229878e-01 -1.07514644e+00 -3.10481489e-01 2.25296333e-01
-3.03104460e-01 -3.26204658e-01 1.72589168e-01 7.59166777e-01
-1.02484763e+00 9.54573676e-02 -4.47188586e-01 -1.27895713e-01
-2.17052245e+00 6.31592095e-01 3.03338379e-01 2.33330935e-01
-4.90916699e-01 4.09345686e-01 -1.09380506e-01 1.00690961e+00
-3.28982137e-02 -4.05412227e-01 -5.15541375e-01 -1.55271396e-01
1.19089115e+00 8.47235024e-01 -1.72360882e-01 -9.13707972e-01
-4.47756827e-01 8.02910328e-01 7.52361566e-02 -5.35487592e-01
7.17177570e-01 -7.68006667e-02 -5.16302168e-01 3.42504770e-01
1.12514544e+00 4.77504224e-01 -1.17895639e+00 2.53112286e-01
4.95353341e-02 -7.57746100e-01 1.60382062e-01 -5.06885111e-01
-8.43837380e-01 4.13242936e-01 1.08960164e+00 5.90059042e-01
1.45525026e+00 -5.52663267e-01 8.13120961e-01 -1.31380960e-01
3.40303391e-01 -1.18017852e+00 -2.66266495e-01 2.31471866e-01
2.98199356e-01 -1.03838789e+00 1.33923352e-01 -5.22009611e-01
-6.65773094e-01 1.27595234e+00 5.36002517e-01 -2.10519388e-01
5.66409826e-01 6.43943071e-01 2.15636477e-01 1.09193288e-01
-2.45706573e-01 -1.79018900e-01 -2.66507775e-01 7.30054200e-01
6.85811460e-01 -3.53330344e-01 -8.15622985e-01 -8.14213604e-02
3.05793941e-01 4.27906483e-01 6.31884277e-01 1.36827433e+00
-1.06790066e+00 -9.35304403e-01 -1.09144688e+00 9.24699381e-02
-9.00105476e-01 3.12769294e-01 -5.49818814e-01 8.10682356e-01
5.21959901e-01 1.29733515e+00 4.43284154e-01 3.95928808e-02
2.63481468e-01 3.58082093e-02 4.72567648e-01 -2.18757749e-01
-4.56739843e-01 2.74771929e-01 -1.48250535e-01 -7.52186403e-02
-6.43401682e-01 -5.29857039e-01 -1.69645154e+00 -2.25688845e-01
-4.42666113e-01 1.03240035e-01 6.83738828e-01 6.09400928e-01
-8.89724344e-02 1.42884314e-01 4.09397393e-01 -3.48482609e-01
2.65140086e-01 -7.36221135e-01 -3.63642037e-01 4.78800714e-01
3.31276864e-01 -4.12275165e-01 -4.28061485e-01 3.33967000e-01] | [8.935999870300293, -0.9968603253364563] |
7379dacb-6cf9-4011-9986-f4b79cc151fd | scalable-multi-agent-model-based | 2205.15023 | null | https://arxiv.org/abs/2205.15023v1 | https://arxiv.org/pdf/2205.15023v1.pdf | Scalable Multi-Agent Model-Based Reinforcement Learning | Recent Multi-Agent Reinforcement Learning (MARL) literature has been largely focused on Centralized Training with Decentralized Execution (CTDE) paradigm. CTDE has been a dominant approach for both cooperative and mixed environments due to its capability to efficiently train decentralized policies. While in mixed environments full autonomy of the agents can be a desirable outcome, cooperative environments allow agents to share information to facilitate coordination. Approaches that leverage this technique are usually referred as communication methods, as full autonomy of agents is compromised for better performance. Although communication approaches have shown impressive results, they do not fully leverage this additional information during training phase. In this paper, we propose a new method called MAMBA which utilizes Model-Based Reinforcement Learning (MBRL) to further leverage centralized training in cooperative environments. We argue that communication between agents is enough to sustain a world model for each agent during execution phase while imaginary rollouts can be used for training, removing the necessity to interact with the environment. These properties yield sample efficient algorithm that can scale gracefully with the number of agents. We empirically confirm that MAMBA achieves good performance while reducing the number of interactions with the environment up to an orders of magnitude compared to Model-Free state-of-the-art approaches in challenging domains of SMAC and Flatland. | ['Aleksei Shpilman', 'Vladimir Egorov'] | 2022-05-25 | null | null | null | null | ['smac-1', 'smac'] | ['playing-games', 'playing-games'] | [-5.16523898e-01 1.46822199e-01 -1.60458907e-01 -8.33099335e-02
-5.72195947e-01 -6.56075776e-01 8.67933095e-01 3.34800392e-01
-7.21238315e-01 1.13473475e+00 -1.86421350e-01 -3.63810331e-01
-2.86208957e-01 -8.16558719e-01 -7.21066475e-01 -9.73900557e-01
-6.29189909e-01 8.02722216e-01 4.20022219e-01 -6.72137380e-01
-5.28836176e-02 4.01696295e-01 -1.47179925e+00 -2.54900336e-01
8.66424322e-01 3.42282474e-01 2.85538554e-01 8.53925705e-01
2.12236792e-01 1.27770782e+00 -7.64753997e-01 2.48552039e-01
4.98347342e-01 -2.71964878e-01 -6.36388958e-01 1.86432853e-01
-3.34289163e-01 -6.90342128e-01 -4.60363179e-02 5.19352257e-01
6.09389603e-01 1.46870494e-01 2.83873379e-01 -1.57046402e+00
3.36511403e-01 8.98918867e-01 -7.50237465e-01 -1.06558487e-01
1.64158225e-01 3.30082327e-01 7.34593213e-01 1.89719852e-02
5.76377094e-01 1.11160123e+00 2.73962140e-01 4.87751037e-01
-1.11901760e+00 -6.09536469e-01 3.98078769e-01 1.21914588e-01
-8.93819749e-01 -1.81184292e-01 5.97853959e-01 -1.32490605e-01
1.10885727e+00 -3.18127759e-02 8.49196613e-01 8.24752450e-01
2.27931634e-01 8.18321168e-01 1.40896523e+00 -5.82925141e-01
6.89548790e-01 3.14151682e-02 -3.64700496e-01 6.55185819e-01
2.96387076e-01 4.79120761e-01 -4.59544629e-01 -2.29017764e-01
6.93393767e-01 -2.44004756e-01 1.79080680e-01 -7.33820796e-01
-1.13969564e+00 9.70041633e-01 3.61068338e-01 1.75649360e-01
-8.29879820e-01 4.43956494e-01 5.66045225e-01 9.08087134e-01
3.91783305e-02 5.50756335e-01 -4.41675276e-01 -3.99310827e-01
-5.20819426e-01 5.72132051e-01 1.07833016e+00 6.30573094e-01
9.06890690e-01 4.36747223e-01 4.69314933e-01 4.60390806e-01
3.56361300e-01 3.97163957e-01 3.85510415e-01 -1.29765570e+00
3.29989940e-01 7.66980827e-01 3.37250352e-01 -6.68151319e-01
-5.71735919e-01 -3.66620481e-01 -5.81789970e-01 9.37267482e-01
2.32910961e-01 -6.96539521e-01 -3.02914977e-01 1.82724726e+00
8.64152730e-01 9.51691240e-04 7.12994397e-01 6.54942036e-01
-9.23365075e-03 4.05248433e-01 -2.79279170e-03 -1.79004937e-01
9.54342663e-01 -1.06454349e+00 -2.61748344e-01 -2.79611439e-01
9.07090843e-01 -3.44292909e-01 5.06570339e-01 3.71168226e-01
-9.68785465e-01 -7.05245361e-02 -1.20964384e+00 8.00070226e-01
-1.91339105e-01 -4.22013700e-01 7.90176570e-01 6.28396094e-01
-1.16455412e+00 5.48136652e-01 -1.34741712e+00 -3.20191890e-01
1.59113601e-01 7.67015636e-01 -4.31501150e-01 2.39045322e-01
-8.27166200e-01 1.17089856e+00 4.53402996e-01 -3.81635547e-01
-1.57565498e+00 -1.96978047e-01 -7.47674346e-01 -7.43145645e-02
7.43693054e-01 -5.40946305e-01 1.56345952e+00 -1.07919514e+00
-1.77416468e+00 2.01909617e-01 4.65821058e-01 -9.50148106e-01
7.60541618e-01 -5.72085418e-02 1.63334519e-01 2.32224852e-01
3.95558961e-02 7.04571128e-01 6.89356685e-01 -1.64111853e+00
-9.55682993e-01 -6.71348209e-03 7.90164232e-01 5.95312476e-01
-3.70324463e-01 -2.40214065e-01 2.50335872e-01 -1.22327141e-01
-3.60448867e-01 -1.16267622e+00 -5.55374444e-01 -5.37443519e-01
1.70917735e-01 -2.99173385e-01 1.03152835e+00 -4.50522788e-02
8.21896672e-01 -1.76681685e+00 1.68828130e-01 1.71238288e-01
1.97654769e-01 1.18244499e-01 -3.26573998e-01 1.12790620e+00
5.29650450e-01 -1.98653311e-01 1.31308213e-01 -3.95885944e-01
2.46209413e-01 7.37398922e-01 2.55708676e-02 5.44045269e-01
-1.13722689e-01 6.11343086e-01 -1.01022208e+00 -4.83376086e-01
2.50595361e-01 1.41760334e-01 -7.33704686e-01 3.71211648e-01
-4.26798373e-01 6.45073533e-01 -6.26077116e-01 3.66018355e-01
3.17060053e-01 -2.40630820e-01 8.33708048e-01 7.31872082e-01
-3.17080617e-01 7.08776526e-03 -1.36330223e+00 1.29993582e+00
-6.48737192e-01 3.10898244e-01 7.79375255e-01 -1.12482083e+00
6.88507497e-01 6.03103161e-01 8.66529286e-01 -6.72766387e-01
7.39916563e-02 2.72773981e-01 4.09215182e-01 -2.10672274e-01
3.44857961e-01 -9.11976304e-03 -8.87749493e-02 7.97477424e-01
-2.11620033e-01 -3.01413357e-01 3.99792373e-01 2.56711185e-01
1.39998174e+00 3.16789091e-01 4.75020707e-01 -2.11325824e-01
3.70283633e-01 3.65749896e-01 7.39841938e-01 1.10328424e+00
-3.30731869e-01 -5.10339022e-01 3.83199573e-01 -5.15183091e-01
-1.00089729e+00 -6.58331513e-01 5.79303920e-01 1.39265740e+00
2.31507003e-01 -4.21514362e-01 -6.31061733e-01 -6.36556089e-01
-1.67763196e-02 3.78063321e-01 -5.24308205e-01 2.53770322e-01
-7.69297242e-01 -4.70289648e-01 4.79591906e-01 3.11222315e-01
8.96056294e-01 -1.15132344e+00 -1.53312910e+00 6.39363468e-01
2.67730981e-01 -9.01440263e-01 3.79179493e-02 5.20238101e-01
-6.24283969e-01 -1.24223351e+00 -3.61086994e-01 -5.98767698e-01
5.10699332e-01 3.82206619e-01 8.32750142e-01 2.74403095e-01
1.08902752e-01 7.65122652e-01 -6.09224319e-01 -5.77448368e-01
-8.36951017e-01 3.25232953e-01 3.44258815e-01 -4.66137558e-01
-1.46588385e-01 -9.06132936e-01 -5.44803619e-01 3.31891358e-01
-9.60277975e-01 2.28243649e-01 7.46097565e-01 1.03287554e+00
-2.11672764e-02 1.94821477e-01 8.25541317e-01 -7.09152877e-01
7.14550138e-01 -4.33433861e-01 -9.05849457e-01 1.65098116e-01
-5.63346267e-01 1.42094806e-01 9.59146202e-01 -5.25404274e-01
-8.44845772e-01 2.68724747e-02 2.32461274e-01 -1.37524247e-01
-2.13594094e-01 4.81241763e-01 4.39692587e-01 -2.62533635e-01
5.71401894e-01 1.74545377e-01 4.15660471e-01 1.02842927e-01
8.55092183e-02 4.90150481e-01 -8.95246267e-02 -1.03249943e+00
9.04672742e-01 4.33721572e-01 1.33908227e-01 -6.47372246e-01
-3.25511068e-01 -3.56639236e-01 -2.99775779e-01 -2.82931030e-01
3.88054281e-01 -9.81776714e-01 -1.20720148e+00 4.27855134e-01
-6.46857083e-01 -9.78761911e-01 -1.33809939e-01 4.75423872e-01
-7.52705872e-01 1.51938334e-01 -4.86118197e-01 -1.00532913e+00
-2.43837848e-01 -1.24303186e+00 5.90164542e-01 4.22410011e-01
1.00405432e-01 -1.21697927e+00 4.86837357e-01 6.45169988e-02
8.28257263e-01 4.65432376e-01 6.00061953e-01 -8.02322268e-01
-5.71359158e-01 7.38089457e-02 4.65083688e-01 2.25100536e-02
1.70685157e-01 -1.99431330e-01 -5.49230099e-01 -8.35165262e-01
-3.10165107e-01 -8.58036399e-01 9.96014178e-02 5.12611680e-02
2.16567308e-01 -4.04359132e-01 -2.01640740e-01 -4.55274945e-03
1.51587355e+00 2.64474094e-01 1.24484971e-01 1.06156075e+00
1.69637546e-01 6.07777119e-01 7.23065197e-01 9.11286592e-01
7.59398103e-01 5.68347633e-01 8.81532848e-01 -6.75225556e-02
2.89007366e-01 1.56291723e-02 7.00205624e-01 7.45935678e-01
-1.50285766e-01 8.47510248e-02 -1.01751304e+00 4.83933747e-01
-2.26362205e+00 -1.00761402e+00 3.99737716e-01 1.96486843e+00
7.07870424e-01 1.31433755e-01 4.25539851e-01 -6.37699664e-02
3.05161536e-01 9.12581086e-02 -6.61199927e-01 -4.35760677e-01
1.63765922e-02 -1.23374730e-01 5.24119794e-01 5.25409639e-01
-8.88975739e-01 9.49382901e-01 5.82218695e+00 5.37665367e-01
-9.98225033e-01 2.09187761e-01 1.66168332e-01 -9.10761207e-03
1.21848218e-01 2.73963332e-01 -5.64053833e-01 1.89240351e-01
1.03277230e+00 -2.90748984e-01 7.55673051e-01 1.15941548e+00
5.48709929e-01 -6.07970655e-01 -9.40781713e-01 5.19723952e-01
-3.58486235e-01 -1.19851792e+00 -4.79711562e-01 3.56283367e-01
9.83983696e-01 3.72617692e-01 -3.09787899e-01 7.56287098e-01
1.20929790e+00 -7.53855467e-01 7.13700712e-01 -1.31858923e-02
3.63625772e-02 -9.50857699e-01 8.05293798e-01 9.49824631e-01
-1.18167138e+00 -4.24462765e-01 -8.91211256e-02 -4.93024170e-01
-1.26916081e-01 -1.73787460e-01 -1.09150279e+00 7.60426104e-01
6.21537983e-01 2.19304398e-01 -4.52474833e-01 8.04058850e-01
-1.27601802e-01 6.04162216e-01 -5.77293992e-01 -2.63520062e-01
8.39240432e-01 -2.04100311e-01 5.52149415e-01 6.00201070e-01
-9.12316442e-02 -1.22313574e-01 9.20087874e-01 2.14922935e-01
2.43778333e-01 -5.46131060e-02 -7.26228595e-01 1.21158846e-01
6.73333049e-01 1.22338724e+00 -6.19356632e-01 -2.98319489e-01
-3.72925937e-01 5.31879485e-01 7.37452388e-01 2.14351997e-01
-7.64940441e-01 3.46283265e-03 5.21364629e-01 -1.58637315e-01
2.82600313e-01 -6.61714315e-01 3.05725574e-01 -7.44036496e-01
-1.62818313e-01 -1.35599828e+00 2.81274766e-01 -1.68327406e-01
-9.16728318e-01 4.82666761e-01 4.62232567e-02 -1.28797364e+00
-7.90188372e-01 -2.34766901e-01 -6.50564611e-01 1.67210057e-01
-1.80474043e+00 -1.30896664e+00 -1.92505002e-01 6.05532348e-01
5.69366813e-01 -6.05734289e-01 9.89003778e-01 -9.25220847e-02
-5.35180032e-01 1.46752849e-01 3.80310446e-01 -9.21399668e-02
5.85031688e-01 -1.41160452e+00 -3.64555717e-01 5.88490903e-01
-1.77323177e-01 4.57999438e-01 8.11905921e-01 -4.62995350e-01
-1.73266697e+00 -7.29375184e-01 9.55382362e-03 1.51820052e-02
7.39024103e-01 -7.81590417e-02 -4.02660161e-01 6.82917953e-01
7.36752927e-01 -1.83714911e-01 5.32772779e-01 1.29034877e-01
-7.21029425e-03 -2.71847844e-01 -8.39446962e-01 7.73373842e-01
3.13003004e-01 6.73924983e-02 -2.32879490e-01 2.07894012e-01
4.28151697e-01 -3.83227050e-01 -8.28398883e-01 1.99779540e-01
3.13897282e-01 -1.15111601e+00 3.61351341e-01 -3.94916415e-01
6.20371141e-02 -4.05307591e-01 -1.34515092e-01 -1.63291037e+00
1.03480101e-01 -1.00532770e+00 -1.39917716e-01 8.69922698e-01
1.47923082e-01 -1.02206123e+00 9.41179335e-01 3.55025649e-01
1.86080369e-03 -6.83114350e-01 -1.06892300e+00 -1.04512942e+00
3.07073027e-01 2.91026058e-03 5.89381158e-01 8.60244334e-01
2.72807062e-01 1.99634030e-01 -4.78318781e-01 2.90428191e-01
6.20825112e-01 1.38861030e-01 1.33583486e+00 -9.00561631e-01
-7.77472794e-01 -2.14516580e-01 -1.12923935e-01 -7.52687693e-01
3.82246047e-01 -4.50543612e-01 9.49165300e-02 -1.51819825e+00
-2.12242343e-02 -9.51424301e-01 7.22874105e-02 7.95008361e-01
1.57652587e-01 -3.14408541e-01 4.30893481e-01 3.02103192e-01
-1.15597379e+00 6.38102114e-01 1.09567893e+00 -3.37545923e-03
-4.92274761e-01 -4.39498350e-02 -3.25982004e-01 6.39755905e-01
1.18485212e+00 -4.40029919e-01 -6.05047107e-01 -3.93878609e-01
2.03503042e-01 3.51421177e-01 1.77820161e-01 -1.22069955e+00
5.26748955e-01 -5.80557525e-01 -1.73990622e-01 -8.24335665e-02
2.71992892e-01 -1.09061158e+00 1.46522120e-01 9.67970312e-01
-2.22725764e-01 3.70182663e-01 1.82112694e-01 5.76129854e-01
-2.88264334e-01 -1.83542356e-01 6.84423208e-01 -3.54092687e-01
-6.62932515e-01 -1.93021670e-02 -7.69731343e-01 -1.85270339e-01
1.63490117e+00 -1.32066221e-03 -4.22958732e-01 -7.00843334e-01
-4.58444327e-01 1.00811481e+00 6.32761359e-01 4.18375432e-02
2.04833433e-01 -8.47822487e-01 -8.13260078e-01 -1.29649714e-01
-1.95692688e-01 -5.75347543e-02 9.29286852e-02 9.83847737e-01
-6.80672467e-01 1.76818103e-01 -4.46661860e-01 -5.09912550e-01
-1.36426139e+00 2.66398698e-01 5.72130084e-01 -8.02290857e-01
-6.44514501e-01 1.70307696e-01 -1.78236604e-01 -6.79049492e-01
2.88571894e-01 2.02806026e-01 -4.10353765e-02 -1.43784106e-01
3.07820886e-01 2.81834692e-01 -1.38877138e-01 -1.52156964e-01
-2.65141964e-01 6.97115213e-02 -3.37954283e-01 -4.80606258e-01
1.73460889e+00 -1.53739870e-01 4.46774736e-02 1.34504303e-01
4.07394886e-01 -2.50110552e-02 -1.75560784e+00 -1.08459517e-01
-2.33796295e-02 -1.32077858e-01 -4.87719737e-02 -7.38630593e-01
-7.29077041e-01 3.89588356e-01 3.33479226e-01 5.08838058e-01
9.47902858e-01 -2.58269072e-01 2.45117426e-01 7.99848676e-01
1.02666640e+00 -1.41239071e+00 4.19515938e-01 6.07209921e-01
5.96353948e-01 -1.43475974e+00 5.11193871e-02 7.20887259e-02
-9.89968598e-01 9.86689627e-01 8.97319138e-01 -3.84362489e-01
2.79673010e-01 5.65647244e-01 1.99973419e-01 -3.03720925e-02
-1.20101547e+00 -3.55308205e-01 -8.37939799e-01 7.97733963e-01
-1.38733625e-01 1.46109730e-01 -8.26124568e-03 3.63291651e-02
-5.23662567e-02 -3.67852926e-01 7.58557498e-01 1.63136327e+00
-7.27815628e-01 -1.48899138e+00 -4.55533952e-01 1.80930883e-01
-1.65229425e-01 4.45917815e-01 -1.46952286e-01 1.20791256e+00
-1.50236338e-01 1.19013667e+00 -2.07998633e-01 1.99850835e-02
6.58257753e-02 -2.45334953e-01 4.59746957e-01 -4.83723730e-01
-9.84258175e-01 1.95895121e-01 2.24828497e-01 -4.55454111e-01
-7.74716079e-01 -4.76067454e-01 -1.45964885e+00 -5.04106224e-01
-1.88988402e-01 5.04587531e-01 6.19569898e-01 9.45480824e-01
3.40524524e-01 5.19734085e-01 8.96961153e-01 -9.49374080e-01
-9.34583366e-01 -8.66498888e-01 -5.24324059e-01 -5.51085733e-03
4.23161596e-01 -7.94366598e-01 -1.90595016e-01 -3.18201840e-01] | [3.882767915725708, 1.9979344606399536] |
8dfb4e30-49d7-4488-a05d-dedaee49de01 | wavecrn-an-efficient-convolutional-recurrent | 2004.04098 | null | https://arxiv.org/abs/2004.04098v3 | https://arxiv.org/pdf/2004.04098v3.pdf | WaveCRN: An Efficient Convolutional Recurrent Neural Network for End-to-end Speech Enhancement | Due to the simple design pipeline, end-to-end (E2E) neural models for speech enhancement (SE) have attracted great interest. In order to improve the performance of the E2E model, the locality and temporal sequential properties of speech should be efficiently taken into account when modelling. However, in most current E2E models for SE, these properties are either not fully considered or are too complex to be realized. In this paper, we propose an efficient E2E SE model, termed WaveCRN. In WaveCRN, the speech locality feature is captured by a convolutional neural network (CNN), while the temporal sequential property of the locality feature is modeled by stacked simple recurrent units (SRU). Unlike a conventional temporal sequential model that uses a long short-term memory (LSTM) network, which is difficult to parallelize, SRU can be efficiently parallelized in calculation with even fewer model parameters. In addition, in order to more effectively suppress the noise components in the input noisy speech, we derive a novel restricted feature masking (RFM) approach that performs enhancement on the feature maps in the hidden layers; this is different from the approach that applies the estimated ratio mask on the noisy spectral features, which is commonly used in speech separation methods. Experimental results on speech denoising and compressed speech restoration tasks confirm that with the lightweight architecture of SRU and the feature-mapping-based RFM, WaveCRN performs comparably with other state-of-the-art approaches with notably reduced model complexity and inference time. | ['Yu Tsao', 'Hsin-Min Wang', 'Tsun-An Hsieh', 'Xugang Lu'] | 2020-04-06 | null | null | null | null | ['speech-denoising'] | ['speech'] | [ 4.73439693e-01 -1.55043706e-01 3.54200602e-01 -2.79384673e-01
-5.09341955e-01 1.77227352e-02 4.09047812e-01 -2.55405575e-01
-5.14187574e-01 3.32757831e-01 3.90270084e-01 -4.57998276e-01
-3.89620692e-01 -5.49652994e-01 -6.30725324e-01 -8.59837055e-01
9.08169001e-02 -4.98053372e-01 1.47014931e-01 -4.11296964e-01
-7.99079686e-02 2.67812282e-01 -1.78425300e+00 4.95864630e-01
8.37040722e-01 1.09529030e+00 8.76550972e-01 7.73783028e-01
-6.30538762e-02 8.24713469e-01 -5.57955146e-01 1.03933051e-01
-1.30053610e-03 -6.41091883e-01 -3.23459804e-01 1.34799378e-02
-1.43242292e-02 -2.65736222e-01 -4.72143292e-01 1.15903544e+00
9.22526538e-01 5.79568088e-01 1.19280703e-01 -6.02290273e-01
-4.16863114e-01 5.62074006e-01 -2.05193251e-01 3.32334965e-01
-2.91858166e-02 -2.04100221e-01 7.58855522e-01 -1.21608472e+00
2.88136125e-01 1.13100290e+00 8.32350731e-01 6.13427043e-01
-1.15764618e+00 -5.86459517e-01 2.44809970e-01 3.95116866e-01
-1.17453802e+00 -8.81798625e-01 7.96406627e-01 1.18736900e-01
1.31011343e+00 4.87420648e-01 5.61121285e-01 9.99842942e-01
1.59238651e-01 8.19961667e-01 1.10041630e+00 -5.29537857e-01
2.45184764e-01 -1.49419188e-01 9.09793898e-02 4.17365015e-01
-5.40764511e-01 4.70481634e-01 -8.70060503e-01 1.63852915e-01
6.51599348e-01 3.58215123e-02 -5.20101488e-01 2.20185608e-01
-1.00322354e+00 5.43393731e-01 4.48752224e-01 5.67089200e-01
-6.08085811e-01 7.82252848e-02 5.92889845e-01 4.80466962e-01
7.51182616e-01 -5.69015257e-02 -5.46492636e-01 -1.67022735e-01
-1.30639672e+00 -4.30065766e-02 4.72321004e-01 5.01319349e-01
5.01514733e-01 5.99429965e-01 -2.91833073e-01 1.20547855e+00
2.47243658e-01 2.93114245e-01 7.64735401e-01 -6.87147617e-01
3.49475414e-01 5.57383709e-02 -2.86881417e-01 -6.35316312e-01
-3.76785040e-01 -9.43255842e-01 -1.45537925e+00 2.52232164e-01
-9.19612199e-02 5.78131936e-02 -1.09197128e+00 1.67402208e+00
8.35236534e-03 6.54637039e-01 4.00641799e-01 9.12874281e-01
9.40832317e-01 8.64244521e-01 -3.08546443e-02 -4.64077502e-01
1.19161677e+00 -1.08021891e+00 -1.23064041e+00 -3.41720074e-01
4.01146859e-01 -9.72877860e-01 7.46124268e-01 3.25152367e-01
-1.14793491e+00 -8.60598207e-01 -1.08490145e+00 -1.64145932e-01
-3.48063231e-01 3.60633999e-01 3.36943805e-01 5.78530788e-01
-1.40287495e+00 1.00440276e+00 -9.44675624e-01 -4.54787388e-02
6.26633838e-02 4.92643237e-01 -2.66427338e-01 1.14011072e-01
-1.39073110e+00 8.50077927e-01 3.00184846e-01 9.12224472e-01
-6.42620802e-01 -6.55509174e-01 -9.31123078e-01 3.93655539e-01
2.39143997e-01 -5.41366339e-01 1.20966661e+00 -8.84805679e-01
-1.99765825e+00 7.60611370e-02 -6.59013093e-01 -6.67210042e-01
2.13693589e-01 -1.83337152e-01 -5.94681263e-01 2.34220158e-02
-4.79624957e-01 3.81896347e-01 1.23433220e+00 -1.00103307e+00
-4.13369536e-01 -6.04009926e-02 -3.67892146e-01 1.45525649e-01
-4.45845693e-01 3.40136886e-01 -2.34317362e-01 -1.03913522e+00
4.86350387e-01 -5.41878283e-01 -3.90873402e-01 -1.94247335e-01
-1.25100002e-01 1.15103640e-01 1.08497798e+00 -1.20234597e+00
1.54740679e+00 -2.59543633e+00 2.21609250e-01 7.29220361e-02
9.51507613e-02 9.30732191e-01 -2.44714290e-01 3.37024778e-01
-3.63274604e-01 -5.38403988e-02 -4.01566237e-01 -7.81129360e-01
-6.66230824e-03 2.34775186e-01 -4.29021508e-01 2.73194343e-01
1.96672469e-01 7.36531138e-01 -6.13332927e-01 -4.05694954e-02
5.07957816e-01 1.03062356e+00 -4.37746137e-01 1.73850715e-01
2.43033871e-01 3.67915958e-01 -3.40489782e-02 1.90240055e-01
8.52889717e-01 1.79331586e-01 4.68278080e-02 -2.35078901e-01
-3.93388778e-01 7.52078593e-01 -1.22214091e+00 1.40353215e+00
-9.03354704e-01 5.93325794e-01 5.04753768e-01 -9.89162564e-01
1.02648675e+00 8.79503369e-01 1.30847692e-01 -9.00722265e-01
-4.95457882e-03 4.36840326e-01 1.01989895e-01 -2.70556122e-01
4.53377903e-01 -2.60325968e-01 5.92163205e-01 1.70662284e-01
1.20378166e-01 1.03845805e-01 -1.22411035e-01 -1.25579700e-01
1.01789880e+00 1.01053782e-01 2.32683167e-01 -1.03774853e-01
6.97916746e-01 -6.55839443e-01 8.73843610e-01 6.45233274e-01
-1.65403932e-01 8.72734725e-01 1.63223185e-02 -1.17124505e-01
-8.85504365e-01 -7.99755633e-01 -1.23449504e-01 9.59290802e-01
-1.50184423e-01 -5.92380583e-01 -8.95876050e-01 -1.47426486e-01
-6.11559629e-01 5.20353794e-01 -1.74031541e-01 -3.16517204e-01
-8.64470482e-01 -7.36488223e-01 5.31761229e-01 5.65040827e-01
7.34754145e-01 -1.15695298e+00 -3.92800272e-01 7.50425875e-01
-3.24670881e-01 -1.13679683e+00 -5.27151823e-01 5.40414631e-01
-1.02977002e+00 -4.00715679e-01 -7.05540001e-01 -8.43715787e-01
3.03732008e-01 5.19908726e-01 6.37386918e-01 1.73711300e-01
1.29066363e-01 -1.21108234e-01 -5.95240235e-01 -1.98554769e-01
-3.86436015e-01 -1.10759988e-01 1.58621166e-02 2.61385977e-01
2.75728852e-01 -8.49501312e-01 -5.58308601e-01 1.46960005e-01
-1.08338583e+00 2.29962468e-01 6.90545857e-01 1.23902345e+00
5.76045930e-01 5.15897810e-01 7.51165867e-01 -4.87596065e-01
5.48102021e-01 -1.41928613e-01 -3.45526576e-01 3.94626185e-02
-3.85651320e-01 -1.18440248e-01 9.00924563e-01 -5.30619800e-01
-1.39318359e+00 -6.00162111e-02 -8.04650545e-01 -4.79317546e-01
-2.83339806e-02 6.68788552e-01 -1.86953336e-01 5.00823855e-02
2.14215547e-01 6.06256306e-01 1.76742524e-01 -9.30945814e-01
-1.18654966e-01 9.42355394e-01 3.94136667e-01 6.41570101e-03
6.09669209e-01 2.77277023e-01 -2.38699451e-01 -1.19933736e+00
-6.35688126e-01 -4.60056096e-01 -4.79298979e-01 -2.96213161e-02
7.72879422e-01 -9.90662396e-01 -4.82173324e-01 7.44782031e-01
-1.24300539e+00 -3.31033021e-01 -3.07359964e-01 8.23501170e-01
-3.53327870e-01 4.12490159e-01 -8.39345932e-01 -1.11091805e+00
-6.00914657e-01 -1.24070597e+00 9.86806870e-01 -2.27593984e-02
6.47451952e-02 -8.85882556e-01 -4.78841931e-01 2.09674537e-01
9.63160813e-01 -3.55760902e-01 8.44878972e-01 -3.73954564e-01
-3.57508212e-01 1.27317756e-01 -8.31641331e-02 8.91210258e-01
5.06379120e-02 -3.33766937e-01 -1.38744724e+00 -3.91806722e-01
7.13485122e-01 2.72375524e-01 1.27182293e+00 7.65867472e-01
1.11541271e+00 -1.29542023e-01 7.50152320e-02 7.73690403e-01
1.04438746e+00 2.14923590e-01 1.03784931e+00 3.45367342e-01
3.42628270e-01 5.87545097e-01 4.34106946e-01 3.21473956e-01
2.75838822e-02 6.33981884e-01 2.22795278e-01 -5.28860986e-01
-6.45462036e-01 -7.48557374e-02 7.83562422e-01 1.48863721e+00
-1.02955349e-01 -1.65840104e-01 -4.01184946e-01 4.38096493e-01
-1.83032835e+00 -1.04506278e+00 -6.85512051e-02 1.99574757e+00
6.77194178e-01 1.34189995e-02 -4.25588727e-01 7.02781200e-01
8.39264870e-01 5.12923717e-01 -1.73087046e-01 -5.42336166e-01
-4.06735063e-01 5.56725562e-01 1.72183231e-01 5.83604157e-01
-9.93464589e-01 7.61810124e-01 5.73084116e+00 1.18251586e+00
-1.25704432e+00 3.72824937e-01 4.73955244e-01 -8.25641453e-02
-1.36679709e-01 -6.09819330e-02 -7.90967345e-01 2.93192625e-01
1.27432418e+00 3.64355475e-01 7.18891680e-01 4.18831885e-01
7.72976041e-01 1.88432693e-01 -7.22189963e-01 9.07327712e-01
-1.79059029e-01 -1.05671251e+00 -3.02818745e-01 -1.00334696e-01
4.04209197e-01 8.20176303e-02 1.99674860e-01 3.20651412e-01
-2.64226764e-01 -9.65888917e-01 8.58068764e-01 4.39749151e-01
7.22784221e-01 -7.83413351e-01 8.20017040e-01 4.88680691e-01
-1.48940516e+00 -3.23681384e-01 -5.38862884e-01 -1.10395566e-01
4.14203048e-01 9.35087919e-01 -4.26606148e-01 7.30928779e-01
1.00914025e+00 6.11916482e-01 -1.34825438e-01 8.80708814e-01
-3.26307595e-01 8.12634349e-01 -2.71092236e-01 3.67035329e-01
3.08241546e-01 -2.61775404e-01 6.89195454e-01 1.37699711e+00
5.07726073e-01 -7.60498047e-02 -2.62116671e-01 6.65953100e-01
1.53759003e-01 -1.21081799e-01 -1.44059181e-01 9.53715593e-02
2.55246937e-01 1.04423749e+00 -3.53456587e-01 -1.86152264e-01
-5.21617413e-01 1.06608689e+00 8.73983428e-02 6.66366100e-01
-6.35980010e-01 -4.82957691e-01 6.93142831e-01 -2.03554090e-02
7.71876991e-01 -2.98350632e-01 -1.71238661e-01 -1.07799923e+00
2.95794845e-01 -9.77893054e-01 -1.50210619e-01 -6.29861116e-01
-1.03689003e+00 1.13657594e+00 -4.51550245e-01 -1.25495148e+00
-2.89698392e-01 -4.95931745e-01 -6.74508393e-01 1.18525589e+00
-2.09227610e+00 -1.24372256e+00 8.76411721e-02 6.46082222e-01
7.38132060e-01 -7.37698525e-02 8.40417862e-01 5.94498873e-01
-6.93648398e-01 4.22160745e-01 3.09796154e-01 -2.18105093e-01
4.95328486e-01 -9.52573359e-01 5.86183488e-01 1.25614429e+00
3.84675786e-02 8.20530415e-01 5.93533397e-01 -5.44338346e-01
-1.11186135e+00 -1.18188608e+00 1.25799370e+00 5.31772137e-01
4.83049393e-01 -5.01417696e-01 -1.19645452e+00 4.10652071e-01
2.33037919e-01 3.02945077e-01 5.38208246e-01 -3.99470329e-02
-2.26301938e-01 -2.42318705e-01 -7.78606713e-01 5.03266454e-01
1.08533943e+00 -9.09800529e-01 -4.96147841e-01 -5.49388081e-02
7.75189221e-01 -3.43902797e-01 -6.98911548e-01 5.52212298e-01
3.09776753e-01 -1.07944262e+00 1.08332646e+00 -1.19892120e-01
5.18141463e-02 -5.34585834e-01 -3.69583219e-01 -1.45124638e+00
-3.33560497e-01 -8.19113553e-01 -2.34233975e-01 1.38172472e+00
3.46485555e-01 -6.99187100e-01 3.32701683e-01 1.18433133e-01
-6.79547369e-01 -7.20988989e-01 -1.25221336e+00 -9.06935215e-01
-2.91949153e-01 -6.71599269e-01 6.19104207e-01 4.91668910e-01
-3.75534147e-01 9.13059786e-02 -6.59625411e-01 4.06363487e-01
2.56242812e-01 -1.51126429e-01 1.13606326e-01 -9.93455648e-01
-4.68818218e-01 -3.48370016e-01 -1.02977403e-01 -1.28792787e+00
2.31647030e-01 -7.42616475e-01 2.76346296e-01 -1.59005547e+00
-1.41453996e-01 -1.97127506e-01 -5.33352792e-01 3.41441959e-01
-2.23925009e-01 2.05331087e-01 1.55134141e-01 -4.94971545e-03
-1.33746937e-01 1.04412723e+00 1.11049497e+00 1.35666549e-01
-3.95319074e-01 1.79127812e-01 -3.16873103e-01 7.38512635e-01
6.99364722e-01 -3.40220332e-01 -2.65602082e-01 -5.69086671e-01
-1.45685256e-01 1.67975649e-01 5.43985844e-01 -9.91614163e-01
6.28881156e-01 3.67941499e-01 2.22312048e-01 -7.11839437e-01
6.70067132e-01 -8.17678571e-01 1.13081262e-01 3.98663521e-01
-9.58253890e-02 -2.16582090e-01 2.12557361e-01 3.84490520e-01
-8.66478980e-01 -2.08995312e-01 8.27863276e-01 6.71548173e-02
-6.87954843e-01 4.98929285e-02 -7.86447048e-01 -7.03251660e-01
3.42211634e-01 -2.70104975e-01 -7.90113509e-02 -4.04257506e-01
-8.48940909e-01 -2.92560786e-01 -2.58730024e-01 2.62080312e-01
9.99337137e-01 -1.10933149e+00 -7.47781634e-01 5.72370112e-01
-4.91878390e-01 -1.32577762e-01 8.35890472e-01 1.04997778e+00
-8.67098395e-04 3.43913823e-01 1.86782345e-01 -3.71337295e-01
-1.34048212e+00 3.80517960e-01 5.21253288e-01 -3.31523389e-01
-9.52713311e-01 7.26754665e-01 1.75457194e-01 -5.59271038e-01
2.81853467e-01 -1.79279283e-01 -3.69464427e-01 -1.40226498e-01
8.68523180e-01 4.62378740e-01 5.82094848e-01 -7.11423039e-01
-1.94871396e-01 3.57596308e-01 -4.83163558e-02 -1.37343094e-01
1.74246824e+00 -3.50378036e-01 -2.90604144e-01 1.19576901e-01
1.11193347e+00 4.87639196e-03 -1.30791807e+00 -4.25438464e-01
-4.11807522e-02 -2.04776883e-01 7.86261618e-01 -6.17094517e-01
-1.22853148e+00 1.07925665e+00 7.95965791e-01 5.55705279e-02
1.77781355e+00 -5.85315943e-01 1.10785556e+00 1.42361566e-01
-5.65123893e-02 -1.12063015e+00 -1.97660297e-01 8.78974497e-01
1.01050913e+00 -8.83622289e-01 -3.60078067e-01 -6.37571692e-01
-2.75871485e-01 1.20813382e+00 2.29810417e-01 1.37218967e-01
8.25805843e-01 6.57468557e-01 1.52287975e-01 1.13533728e-01
-6.79989994e-01 -4.14774060e-01 2.41733670e-01 4.22066897e-01
4.29736197e-01 -2.47500017e-01 -2.31882915e-01 8.03074241e-01
-1.39815211e-01 -1.18937738e-01 2.18059853e-01 8.37850511e-01
-4.08232272e-01 -1.36704552e+00 -4.61715609e-01 1.35804653e-01
-5.65167010e-01 -6.03600979e-01 3.06649476e-01 2.85572916e-01
7.03902617e-02 1.32268918e+00 -4.14706357e-02 -5.27355790e-01
3.51692528e-01 8.02535191e-02 -5.01641929e-02 -4.03685868e-01
-9.95324612e-01 6.33274138e-01 3.08290925e-02 -5.13949037e-01
-5.14231861e-01 -3.47447872e-01 -1.20200562e+00 -1.81305960e-01
-6.31741881e-01 -2.04596687e-02 9.80917752e-01 1.11512733e+00
4.69272047e-01 1.15596449e+00 6.55207157e-01 -1.01568758e+00
-5.36997974e-01 -1.20442116e+00 -7.46928036e-01 3.99824455e-02
7.41416514e-01 -4.40980136e-01 -5.42757571e-01 -1.10894039e-01] | [14.90800666809082, 5.952182292938232] |
9dd26821-6319-43fb-987c-8508ffd55732 | joint-learning-for-pulmonary-nodule | 1802.03584 | null | http://arxiv.org/abs/1802.03584v1 | http://arxiv.org/pdf/1802.03584v1.pdf | Joint Learning for Pulmonary Nodule Segmentation, Attributes and Malignancy Prediction | Refer to the literature of lung nodule classification, many studies adopt
Convolutional Neural Networks (CNN) to directly predict the malignancy of lung
nodules with original thoracic Computed Tomography (CT) and nodule location.
However, these studies cannot tell how the CNN works in terms of predicting the
malignancy of the given nodule, e.g., it's hard to conclude that whether the
region within the nodule or the contextual information matters according to the
output of the CNN. In this paper, we propose an interpretable and multi-task
learning CNN -- Joint learning for \textbf{P}ulmonary \textbf{N}odule
\textbf{S}egmentation \textbf{A}ttributes and \textbf{M}alignancy
\textbf{P}rediction (PN-SAMP). It is able to not only accurately predict the
malignancy of lung nodules, but also provide semantic high-level attributes as
well as the areas of detected nodules. Moreover, the combination of nodule
segmentation, attributes and malignancy prediction is helpful to improve the
performance of each single task. In addition, inspired by the fact that
radiologists often change window widths and window centers to help to make
decision on uncertain nodules, PN-SAMP mixes multiple WW/WC together to gain
information for the raw CT input images. To verify the effectiveness of the
proposed method, the evaluation is implemented on the public LIDC-IDRI dataset,
which is one of the largest dataset for lung nodule malignancy prediction.
Experiments indicate that the proposed PN-SAMP achieves significant improvement
with respect to lung nodule classification, and promising performance on lung
nodule segmentation and attribute learning, compared with the-state-of-the-art
methods. | ['Zhen Zhou', 'Yizhou Wang', 'Jianwei Wang', 'Botong Wu'] | 2018-02-10 | null | null | null | null | ['lung-nodule-segmentation', 'lung-nodule-classification'] | ['medical', 'medical'] | [ 2.04905067e-02 3.07174385e-01 -3.86187524e-01 -2.76899904e-01
-7.84821093e-01 -4.07235056e-01 3.30169648e-01 -1.20480597e-01
-1.89733952e-01 4.56364900e-01 1.12462148e-01 -6.19836569e-01
-2.88902372e-01 -8.32735538e-01 -6.12123489e-01 -1.02422643e+00
3.55816841e-01 8.93392205e-01 4.89788234e-01 1.49299160e-01
-3.72943252e-01 4.70759839e-01 -1.18513680e+00 6.74652338e-01
6.48195207e-01 1.42134976e+00 4.94035661e-01 4.60063100e-01
-2.22127751e-01 8.69440734e-01 -1.36468247e-01 -2.48974651e-01
1.86402574e-01 -3.64479274e-01 -9.69381750e-01 1.34271428e-01
2.56226391e-01 -4.40929830e-01 -3.76396120e-01 8.44388068e-01
1.04010887e-01 -5.07125378e-01 1.05993509e+00 -8.78373206e-01
-5.00594854e-01 7.67888010e-01 -2.90766448e-01 3.16552341e-01
-5.10365665e-01 2.24319220e-01 1.17235470e+00 -7.83234239e-01
2.98540175e-01 6.91513777e-01 7.27902770e-01 4.06837791e-01
-5.36424458e-01 -6.93421364e-01 -8.58062878e-02 8.78157169e-02
-1.25932205e+00 2.05618307e-01 1.66475460e-01 -4.40591007e-01
3.18532825e-01 6.78194463e-01 6.85613096e-01 1.12094247e+00
4.55362916e-01 7.45896459e-01 9.46891546e-01 6.92951381e-02
-3.76293510e-01 1.96024418e-01 -2.91799039e-01 1.12412107e+00
5.18246770e-01 4.28648181e-02 1.72620028e-01 -6.62070215e-02
9.95510519e-01 2.83204466e-01 -2.33857289e-01 -1.57091960e-01
-1.66545737e+00 5.83164930e-01 9.77459490e-01 5.65839946e-01
-3.06455493e-01 2.63919383e-01 2.90638566e-01 -2.69906372e-01
3.35204095e-01 3.09140742e-01 -2.20103741e-01 5.06953657e-01
-7.07972050e-01 -1.15355432e-01 7.93108821e-01 7.06253409e-01
3.81654054e-01 -1.29169092e-01 -7.51445055e-01 6.05767310e-01
2.98766702e-01 3.46818805e-01 7.30836153e-01 -4.86862868e-01
1.40070632e-01 5.72768927e-01 -2.01267362e-01 -3.21855396e-01
-5.96255302e-01 -6.60343826e-01 -1.28115463e+00 -1.90712344e-02
4.15010601e-01 6.19105436e-02 -1.27110052e+00 1.20670807e+00
-1.15601858e-03 3.13509881e-01 -1.33030847e-01 8.57980490e-01
1.47701180e+00 1.94794208e-01 2.96463460e-01 2.05601782e-01
1.81397271e+00 -1.13666224e+00 -5.17648876e-01 -1.42045379e-01
6.81080878e-01 -7.31401742e-01 7.31698632e-01 -2.02788919e-01
-7.38118410e-01 -5.84404349e-01 -7.61926949e-01 1.76568747e-01
-1.37516782e-01 7.91474879e-01 5.47667861e-01 4.47855085e-01
-9.82321322e-01 3.91750306e-01 -9.44526196e-01 -4.02909517e-01
6.75455153e-01 5.83970904e-01 -2.00440526e-01 1.48734720e-02
-1.05248809e+00 1.01951301e+00 7.02854037e-01 1.13259889e-01
-1.22138572e+00 -7.78131425e-01 -1.63363606e-01 2.19561979e-01
7.36597717e-01 -1.09513998e+00 1.38669252e+00 -8.29815805e-01
-9.48299050e-01 8.27108502e-01 1.36272296e-01 -4.90852714e-01
6.89395547e-01 3.81760120e-01 -2.51809359e-01 6.28933758e-02
1.55353755e-01 7.69594491e-01 8.97691607e-01 -9.61832762e-01
-9.87245500e-01 -2.46291801e-01 -2.55276382e-01 2.73909628e-01
-1.95667610e-01 -4.68025893e-01 -5.34339428e-01 -6.25040472e-01
2.07469672e-01 -1.07722402e+00 -3.97041619e-01 1.98153123e-01
-7.85613418e-01 -4.47302103e-01 1.12162483e+00 -5.11116505e-01
1.01615024e+00 -2.08150125e+00 -2.33737439e-01 2.70803332e-01
6.43281877e-01 1.67197585e-01 1.86816618e-01 -4.16389614e-01
-1.20044328e-01 5.72396517e-01 -9.43507403e-02 1.21402740e-01
-2.87369132e-01 5.16531885e-01 3.98559213e-01 2.78099507e-01
2.44255498e-01 1.39301014e+00 -4.64371026e-01 -9.30029750e-01
3.45881253e-01 1.31811574e-01 -5.59870116e-02 2.20940396e-01
-3.34359109e-01 6.37987077e-01 -9.02687907e-01 7.90101647e-01
2.73547232e-01 -8.39890122e-01 5.82082234e-02 -5.86488783e-01
9.77393612e-02 -1.90272897e-01 -6.42900765e-01 1.07307792e+00
-3.96128565e-01 3.02078217e-01 -8.05636048e-02 -7.19574928e-01
6.68879390e-01 8.42293203e-01 8.17691684e-01 -2.21090674e-01
3.20131809e-01 5.56068599e-01 5.13059974e-01 -7.43636250e-01
-6.55367672e-02 1.52769852e-02 2.52893627e-01 4.21058625e-01
-1.39856264e-01 -2.40213960e-01 2.13721618e-02 -2.49067336e-01
1.24209428e+00 -1.69445217e-01 6.68975055e-01 -2.32584387e-01
6.95896864e-01 1.35301843e-01 3.26705724e-01 7.60788977e-01
-3.12944502e-01 7.71382391e-01 5.17822802e-01 -4.65467781e-01
-1.19772983e+00 -1.08173466e+00 -4.29795593e-01 7.98855603e-01
-2.42391769e-02 3.74453694e-01 -4.46502417e-01 -1.15432787e+00
-4.08140086e-02 4.63012695e-01 -1.04049850e+00 2.44058408e-02
-6.13966346e-01 -1.02636600e+00 7.16579556e-01 8.04963291e-01
7.41660595e-01 -1.09022224e+00 -5.18026292e-01 6.35952801e-02
-5.73850632e-01 -1.19984138e+00 -3.58720154e-01 4.45439994e-01
-9.87946928e-01 -1.34362447e+00 -7.71686018e-01 -8.74248683e-01
7.61688292e-01 7.08683282e-02 1.16147101e+00 3.40575546e-01
-4.65816170e-01 4.88476530e-02 -1.56629503e-01 -5.36654890e-01
-6.62961423e-01 3.78445178e-01 -6.76218569e-01 -8.41811523e-02
2.47236326e-01 -7.22030252e-02 -7.02154338e-01 5.79065204e-01
-9.01533127e-01 1.57594621e-01 1.56774652e+00 7.97290802e-01
9.19041872e-01 8.22208226e-02 2.91406900e-01 -1.15636921e+00
1.93284333e-01 -6.67269528e-01 1.11827785e-02 5.33849001e-01
-5.13128698e-01 -7.92225152e-02 5.51629603e-01 -2.33211786e-01
-1.05194759e+00 1.67834654e-01 -6.48495182e-02 -5.15964985e-01
-5.25323927e-01 3.59002173e-01 3.11361521e-01 4.70644273e-02
7.37268984e-01 1.39829114e-01 7.81683847e-02 -8.57444573e-03
-1.87657643e-02 5.89086175e-01 5.85975409e-01 -6.05162531e-02
7.10471392e-01 6.03486776e-01 3.58869076e-01 -3.85581374e-01
-1.27027655e+00 -7.45429158e-01 -7.07893610e-01 -1.87047437e-01
1.33238125e+00 -8.32519531e-01 -5.26543319e-01 4.09601405e-02
-7.38701701e-01 -2.02741195e-02 -2.72239059e-01 6.94722712e-01
-4.97473687e-01 8.52288380e-02 -5.90883911e-01 -1.21848971e-01
-4.69424009e-01 -1.44313550e+00 9.93209422e-01 7.93432742e-02
-1.09229192e-01 -8.73315394e-01 -5.27390361e-01 6.15314841e-01
6.20441973e-01 2.30813101e-02 1.46380341e+00 -1.23078763e+00
-8.74330819e-01 -2.18970969e-01 -6.51208043e-01 2.21669480e-01
4.05144989e-01 1.32837340e-01 -8.79926026e-01 -2.04448123e-02
2.86310520e-02 -1.62911132e-01 9.28278506e-01 7.23801851e-01
1.72300386e+00 -1.78417146e-01 -7.16393828e-01 6.56935275e-01
1.27344453e+00 9.31297541e-02 4.10168052e-01 2.77450681e-01
8.41994524e-01 1.33484304e-01 4.69773859e-01 3.25130612e-01
-4.58623329e-03 2.85092950e-01 1.16942799e+00 -3.31736028e-01
-5.44489205e-01 2.29384020e-01 -3.03782672e-01 3.12813431e-01
-4.61293608e-01 -3.69821548e-01 -1.12336278e+00 4.01144087e-01
-1.62672925e+00 -6.95920825e-01 -3.09773237e-01 1.70080042e+00
3.81353378e-01 2.50964612e-02 -3.32809716e-01 -3.17062318e-01
8.05163264e-01 1.67736888e-01 -6.13621831e-01 2.60411322e-01
1.16539702e-01 1.69747397e-01 6.54443681e-01 -2.05433398e-01
-1.46023834e+00 3.68664622e-01 5.26367140e+00 1.02831292e+00
-1.19011152e+00 3.85960847e-01 1.05646873e+00 1.21955700e-01
-1.54085770e-01 -5.48661172e-01 -5.88810325e-01 1.13347903e-01
4.25353348e-01 6.90283105e-02 -6.70367256e-02 8.87558043e-01
-2.00654734e-02 5.54158613e-02 -1.26075709e+00 5.82410157e-01
-5.38515002e-02 -1.31769574e+00 2.12052420e-01 1.55456200e-01
6.42770648e-01 4.05259311e-01 4.02413934e-01 2.79098928e-01
4.62669656e-02 -1.33289504e+00 3.29821289e-01 5.35950184e-01
8.19849730e-01 -4.40033048e-01 1.52770221e+00 3.55642766e-01
-1.38940430e+00 -1.30915552e-01 -3.00675362e-01 9.63400245e-01
-4.27360117e-01 4.91295487e-01 -1.78323114e+00 7.89353669e-01
9.11149502e-01 4.69098777e-01 -9.75653112e-01 1.10967350e+00
-6.65049031e-02 7.92407870e-01 -2.45685980e-01 -2.14927733e-01
5.12398124e-01 -2.19233930e-02 3.45684677e-01 9.92652893e-01
8.82466376e-01 2.15535253e-01 3.03253561e-01 1.08602846e+00
-1.87792256e-01 2.98136503e-01 -4.08205092e-01 -2.27279723e-01
2.28647232e-01 1.64825261e+00 -1.02412832e+00 -3.21113557e-01
-5.05057335e-01 4.96744841e-01 -2.37003937e-02 1.26801699e-01
-1.13385570e+00 3.19041550e-01 6.96428679e-03 2.93952078e-01
4.57206517e-01 4.46884185e-01 -4.36482072e-01 -6.52560890e-01
-1.58651605e-01 -5.03661513e-01 5.28836727e-01 -6.69495046e-01
-1.30131567e+00 9.44019020e-01 -2.98279911e-01 -1.45940757e+00
-1.36591494e-01 -6.67471409e-01 -7.31165171e-01 6.47851229e-01
-1.36695945e+00 -1.54962277e+00 -7.87705898e-01 5.72768509e-01
5.34924984e-01 -2.98982531e-01 7.71825254e-01 -2.89386697e-02
-3.42866778e-01 3.08470577e-01 3.02755032e-02 3.83195281e-01
6.89220071e-01 -1.41509509e+00 -1.53977290e-01 2.01657683e-01
-1.13587804e-01 -1.60652623e-01 1.90489262e-01 -6.25545144e-01
-7.81385541e-01 -1.69681478e+00 3.58764023e-01 -4.83212173e-01
6.39476597e-01 3.17202717e-01 -8.76624584e-01 7.98425734e-01
-1.10359089e-02 5.08559048e-01 6.52392566e-01 -3.61319482e-01
6.23070262e-03 -2.75909398e-02 -9.93605494e-01 4.22032565e-01
5.93969107e-01 -1.97480038e-01 -2.77570367e-01 6.95775449e-01
5.49023807e-01 -5.04867077e-01 -1.07808089e+00 1.06820750e+00
4.04539704e-01 -1.04814148e+00 1.10189092e+00 -3.29052359e-01
5.42262316e-01 -1.29224122e-01 1.56036718e-02 -9.36421990e-01
-7.34276652e-01 5.98654628e-01 2.13779390e-01 5.14485419e-01
8.63640964e-01 -4.07148927e-01 1.06266332e+00 2.54472494e-01
-4.71092820e-01 -1.22432625e+00 -9.14443254e-01 -3.88539851e-01
2.97265444e-02 -1.81442186e-01 5.93211532e-01 6.77123487e-01
-9.00386274e-01 -2.22953126e-01 1.48984954e-01 2.64973670e-01
2.49326617e-01 2.42465243e-01 1.65979117e-01 -1.33631802e+00
-2.69199729e-01 -7.66049981e-01 -1.19274929e-01 -5.04648089e-01
-1.55690253e-01 -1.21585405e+00 9.54640359e-02 -1.87371039e+00
5.98874509e-01 -5.05333841e-01 -5.48533678e-01 6.79033577e-01
-2.06528142e-01 3.47741932e-01 -8.67829993e-02 5.93820333e-01
-4.73406017e-01 1.32082403e-01 1.92028630e+00 -3.52836162e-01
3.14046085e-01 8.10155571e-01 -4.00463998e-01 8.86288524e-01
6.66363239e-01 -5.11499465e-01 -2.70711690e-01 -1.44022256e-01
-2.22907260e-01 2.65266746e-01 6.84459686e-01 -1.09903538e+00
1.82975665e-01 -1.76486209e-01 7.11574018e-01 -9.99191999e-01
2.62050986e-01 -1.08132720e+00 2.33617082e-01 8.93824995e-01
-1.99787036e-01 -2.23662585e-01 -2.05756068e-01 6.71968877e-01
-3.30798537e-01 -4.21321839e-01 7.86204517e-01 -8.11197817e-01
-5.19255459e-01 9.84964550e-01 -5.16296387e-01 -2.66852319e-01
1.32660162e+00 -1.92417964e-01 -2.79767662e-01 -1.36720613e-01
-9.91755247e-01 1.66683733e-01 -3.50079313e-02 1.30667940e-01
4.90890205e-01 -1.28781247e+00 -9.16033089e-01 -1.09109227e-02
2.51878887e-01 4.07459825e-01 1.86967239e-01 1.32787561e+00
-6.42313182e-01 8.70122671e-01 -7.86762312e-02 -9.01454449e-01
-1.25787103e+00 3.52521300e-01 8.72409880e-01 -6.25823915e-01
-3.54851454e-01 8.94644737e-01 5.89334905e-01 -4.67790574e-01
1.84198245e-01 -7.07464159e-01 -2.99535513e-01 -1.09179355e-01
-2.28044018e-01 1.07594073e-01 2.68294901e-01 -2.89542139e-01
-4.62410711e-02 2.46222809e-01 -3.01136971e-01 4.23401326e-01
8.91507745e-01 1.74322098e-01 -3.85186136e-01 2.55281031e-01
9.11579847e-01 -5.35866559e-01 -7.08707452e-01 -3.15677434e-01
-6.30624890e-02 1.34277483e-03 -9.12853424e-03 -1.04415190e+00
-1.38404751e+00 6.89215302e-01 8.82942080e-01 6.27360195e-02
9.27372277e-01 4.11179334e-01 6.60902917e-01 6.90845549e-01
-1.22239649e-01 -6.08971477e-01 3.34028959e-01 2.10456431e-01
7.77244389e-01 -1.63048613e+00 6.05983175e-02 -7.21815884e-01
-7.71035731e-01 1.51398194e+00 1.09901917e+00 1.93412498e-01
8.65997851e-01 1.14754334e-01 1.28129631e-01 -3.46392781e-01
-7.02181160e-01 -3.85669947e-01 6.10951781e-01 1.44333214e-01
5.07241428e-01 5.14039397e-01 1.14333972e-01 6.44054770e-01
-4.01459672e-02 -5.44688813e-02 2.71197230e-01 4.58135754e-01
-7.46301889e-01 -8.31385374e-01 -3.83394688e-01 1.29164982e+00
-6.54532433e-01 -2.81600822e-02 -3.61039400e-01 1.27398181e+00
4.76744503e-01 3.90064746e-01 7.72230774e-02 -1.66337803e-01
-2.89730299e-02 3.05272024e-02 1.09161265e-01 -7.83262253e-01
-8.63399982e-01 3.51782143e-01 6.22200966e-03 -2.04068705e-01
-3.35037231e-01 -4.77356344e-01 -1.36301005e+00 1.87006101e-01
-4.90894735e-01 5.60535751e-02 3.71448934e-01 1.17343652e+00
-2.94484586e-01 1.05205691e+00 5.18007934e-01 -2.16981381e-01
-7.73985088e-01 -1.12867773e+00 -4.84984308e-01 2.66429156e-01
1.24233827e-01 -3.15020621e-01 -2.97754705e-01 -3.59086007e-01] | [15.343533515930176, -2.1540348529815674] |
8bd91d52-fe41-46a9-9929-7b0c2f9e88b8 | multilabel-automated-recognition-of-emotions | 1905.12629 | null | https://arxiv.org/abs/1905.12629v2 | https://arxiv.org/pdf/1905.12629v2.pdf | A New Multilabel System for Automatic Music Emotion Recognition | Achieving advancements in automatic recognition of emotions that music can induce require considering multiplicity and simultaneity of emotions. Comparison of different machine learning algorithms performing multilabel and multiclass classification is the core of our work. The study analyzes the implementation of the Geneva Emotional Music Scale 9 in the Emotify music dataset and investigates its adoption from a machine-learning perspective. We approach the scenario of emotions expression/induction through music as a multilabel and multiclass problem, where multiple emotion labels can be adopted for the same music track by each annotator (multilabel), and each emotion can be identified or not in the music (multiclass). The aim is the automatic recognition of induced emotions through music. | ['Natalia Pichierri', 'Daniele Casali', 'Marco Matta', 'Giovanni Costantini', 'Fabio Paolizzo', 'Daniele Giardino'] | 2019-05-29 | null | null | null | null | ['music-emotion-recognition'] | ['music'] | [ 1.01162821e-01 -3.78829017e-02 -4.10109907e-01 -3.62001002e-01
-7.95246065e-01 -1.08249521e+00 3.44325006e-01 1.08144358e-01
-3.97930026e-01 7.23489165e-01 3.09549779e-01 3.21320951e-01
-5.31277597e-01 -2.27719218e-01 -1.20155640e-01 -6.98085725e-01
7.83033073e-02 7.42749155e-01 -6.41638279e-01 -5.39974719e-02
-9.83964093e-03 5.18754780e-01 -1.89469397e+00 9.41466331e-01
3.08709830e-01 1.18630993e+00 -5.44130027e-01 7.88630605e-01
-1.74824640e-01 9.75404143e-01 -8.20977986e-01 -4.42644179e-01
8.81002396e-02 -5.16726017e-01 -9.62049484e-01 -8.59313384e-02
6.24996245e-01 6.76081359e-01 5.49286187e-01 1.06690407e+00
7.59415090e-01 3.56209959e-04 8.02672505e-01 -1.71502316e+00
-1.96758509e-01 9.50623631e-01 -1.91759512e-01 -5.21518648e-01
6.92662895e-01 -6.03053749e-01 1.35588741e+00 -5.65241694e-01
6.84180915e-01 1.15303338e+00 6.73380971e-01 3.70944798e-01
-1.22001517e+00 -1.05683577e+00 -3.28407139e-01 1.86643392e-01
-1.41852057e+00 -2.37362310e-01 8.03300321e-01 -7.78982580e-01
5.41108906e-01 6.62946224e-01 8.20961654e-01 8.31341386e-01
-2.42469192e-01 5.05079985e-01 1.74340439e+00 -7.47734427e-01
8.24696049e-02 5.44411302e-01 2.52883971e-01 2.74686366e-01
-2.54908115e-01 -6.11202009e-02 -8.16391826e-01 -3.48330528e-01
3.55562598e-01 -6.86657310e-01 1.33205637e-01 9.98096988e-02
-1.21237075e+00 5.06835699e-01 -2.92273574e-02 9.23778594e-01
-3.72513026e-01 1.82280898e-01 1.03233957e+00 7.46303618e-01
3.42863679e-01 6.32039070e-01 -6.68127537e-01 -3.75286520e-01
-1.04068983e+00 -2.51725197e-01 9.34011102e-01 4.57995564e-01
7.03499258e-01 -1.68020949e-02 -5.28929345e-02 1.11858547e+00
8.47784504e-02 2.32751146e-01 5.24346054e-01 -1.07205212e+00
-4.72283632e-01 7.51817822e-01 8.44566971e-02 -1.02923346e+00
-7.93271482e-01 -4.56281304e-01 -5.93076468e-01 5.65850139e-01
2.27867901e-01 -1.46281153e-01 -1.15409717e-01 1.75532663e+00
3.80890578e-01 2.65248269e-01 2.89688915e-01 8.06695104e-01
1.02357340e+00 3.20907444e-01 3.96607131e-01 -5.18426359e-01
1.50533748e+00 -9.40605998e-01 -9.85973060e-01 4.45288509e-01
9.70113516e-01 -1.18480325e+00 9.51636732e-01 1.19675887e+00
-6.04267299e-01 -6.77683532e-01 -8.07695389e-01 1.65044084e-01
-7.93407500e-01 7.89202869e-01 6.61278605e-01 7.53098369e-01
-8.99739146e-01 3.64206612e-01 2.67406791e-01 -2.21892282e-01
-2.22715855e-01 6.04917526e-01 -5.39708197e-01 7.57694960e-01
-1.28903282e+00 8.17640901e-01 5.77484369e-01 -2.09403276e-01
-2.53119677e-01 -3.41221899e-01 -3.59693587e-01 -3.53332877e-01
-1.70109391e-01 -1.53080985e-01 1.06231523e+00 -2.18223453e+00
-1.45391214e+00 1.69149673e+00 2.04189032e-01 1.60942674e-01
7.73487464e-02 9.92964581e-02 -1.13994527e+00 8.71480629e-02
9.39806700e-02 5.41638732e-01 6.85226262e-01 -1.49021637e+00
-8.16476762e-01 -1.73501864e-01 6.64217472e-02 1.75289258e-01
-3.29677433e-01 6.39613807e-01 -9.22209956e-03 -7.71547556e-01
-1.25380650e-01 -1.34525788e+00 2.86949366e-01 -5.71876645e-01
-3.22386771e-01 -6.48172379e-01 4.61711735e-01 -4.28079247e-01
1.35493755e+00 -2.42128062e+00 2.62371987e-01 3.56840253e-01
-3.03602010e-01 -1.33024931e-01 -2.52427369e-01 2.85763174e-01
-7.84926951e-01 1.03147067e-01 2.04127282e-01 -1.00284129e-01
5.82021952e-01 2.93560296e-01 -2.36018196e-01 2.88547248e-01
-4.99686271e-01 6.28175020e-01 -8.22517097e-01 -8.65282238e-01
1.50732011e-01 4.16888446e-01 -1.81062326e-01 -6.42588213e-02
3.29032317e-02 6.53093696e-01 -2.05505863e-01 8.29050183e-01
3.17293704e-01 1.67819619e-01 4.14522737e-01 -5.04586756e-01
-3.47145408e-01 -1.17062703e-01 -1.54100430e+00 1.74603784e+00
-6.29382491e-01 4.42765683e-01 1.69677332e-01 -6.92754745e-01
1.03243637e+00 1.11756110e+00 8.90374064e-01 -3.08566660e-01
2.43755937e-01 4.09791231e-01 -1.15388304e-01 -5.37474394e-01
4.42796469e-01 -8.04061770e-01 -6.69228196e-01 7.66694725e-01
3.82790953e-01 2.76828138e-03 2.60603309e-01 -2.34794915e-01
5.53003490e-01 2.39381149e-01 4.74154294e-01 -2.37718314e-01
6.51510954e-01 6.30985722e-02 5.18960357e-01 4.93410021e-01
-1.96395233e-01 -1.19743928e-01 6.08810961e-01 -4.15992618e-01
-6.40329301e-01 -5.66681087e-01 -4.78343278e-01 1.82109821e+00
-3.17368090e-01 -4.39534843e-01 -3.48828644e-01 -7.58352280e-01
-1.60708442e-01 4.18910414e-01 -7.68587530e-01 2.39075109e-01
9.64540988e-02 -4.46046680e-01 1.02701032e+00 -2.58097239e-02
-2.16550529e-01 -1.56857967e+00 -3.72109324e-01 2.21082613e-01
-3.70076030e-01 -8.87475073e-01 -3.37582678e-02 6.39735341e-01
-4.03135747e-01 -1.10107422e+00 -1.69451639e-01 -8.46720576e-01
1.54644540e-02 -6.14289522e-01 1.46619320e+00 -2.69952595e-01
-3.39211047e-01 7.67867923e-01 -7.60644913e-01 -6.37757540e-01
-4.48966295e-01 -1.71976462e-02 3.70413244e-01 4.61096436e-01
4.49148685e-01 -7.12781966e-01 -1.41779348e-01 4.21301126e-01
-1.08178127e+00 -3.63781542e-01 3.91581476e-01 3.05628538e-01
6.59632504e-01 1.04625784e-01 9.39229846e-01 -7.17282534e-01
5.25410950e-01 -3.22159767e-01 -6.87237829e-02 2.89183229e-01
-4.55880433e-01 -3.90231669e-01 2.37680405e-01 -8.48241389e-01
-6.83535278e-01 4.31480080e-01 -2.10328460e-01 -2.28084132e-01
-6.29915357e-01 5.82492948e-01 -7.10406899e-02 -4.50105608e-01
6.33673549e-01 -3.51927161e-01 -6.93149626e-01 -2.64280289e-01
6.56565309e-01 9.54878271e-01 4.37368423e-01 -9.02329803e-01
2.08146602e-01 2.65151232e-01 1.96210757e-01 -4.99230981e-01
-1.20396769e+00 -9.29256141e-01 -9.01952207e-01 -1.04056406e+00
1.04554856e+00 -1.04519439e+00 -1.01623774e+00 1.63232699e-01
-1.00261247e+00 -1.76817238e-01 -6.47496164e-01 6.10146642e-01
-7.49957979e-01 6.25100583e-02 -5.81713796e-01 -1.09765005e+00
-4.38742578e-01 -6.98685706e-01 1.15780699e+00 -5.24354205e-02
-1.03401446e+00 -9.63277459e-01 5.90595305e-01 3.88754964e-01
-2.36721203e-01 5.60963273e-01 1.06110549e+00 -8.18125308e-01
5.80754757e-01 -4.64463949e-01 1.37211755e-01 5.26874721e-01
-2.54366864e-02 -2.12487243e-02 -1.33315337e+00 1.87895298e-01
-2.60797858e-01 -9.94631708e-01 3.50970507e-01 -4.44844365e-02
6.71280563e-01 -1.05935140e-02 3.07677239e-01 1.81037828e-01
1.40596533e+00 2.66745478e-01 5.36635995e-01 5.53584754e-01
3.86158943e-01 8.63915324e-01 9.14761722e-01 4.15663600e-01
-1.81800872e-01 8.58677268e-01 2.67756283e-01 -1.96056828e-01
6.97852373e-02 2.32334673e-01 4.90704179e-01 1.13692999e+00
-3.03741723e-01 2.94040412e-01 -5.00264943e-01 2.79895306e-01
-1.77330136e+00 -1.00664413e+00 -4.86486614e-01 2.00532317e+00
9.04416025e-01 -5.03555000e-01 3.67984295e-01 5.87393701e-01
7.25990832e-01 -1.09856606e-01 4.45590541e-02 -9.57270145e-01
-6.55747116e-01 4.97113287e-01 -2.43195251e-01 4.37505454e-01
-1.25363910e+00 9.75325763e-01 6.80917120e+00 1.09101832e+00
-1.09760630e+00 2.66587198e-01 2.10110769e-01 -2.07917839e-01
1.89523831e-01 -1.59473419e-01 -3.27135324e-01 -6.19695894e-02
9.33711529e-01 3.06803435e-01 5.12994766e-01 7.26457059e-01
1.09319173e-01 8.53422731e-02 -8.76097977e-01 1.35463357e+00
3.77518803e-01 -6.00990057e-01 1.41227961e-01 -1.27302706e-01
7.84505308e-01 -2.17080265e-01 -5.61857969e-02 4.31426197e-01
-1.15605574e-02 -8.94007623e-01 9.78608370e-01 8.38340223e-01
1.02320492e+00 -9.27671611e-01 6.36368573e-01 -7.50972778e-02
-1.25711262e+00 -1.98936537e-02 1.30077705e-01 -1.57338992e-01
6.83621094e-02 4.98408049e-01 -1.27887353e-01 7.73365378e-01
6.19711339e-01 6.78434074e-01 -3.09869081e-01 7.88792789e-01
-3.69620144e-01 5.98549604e-01 -3.27703059e-02 1.62223205e-01
1.88905045e-01 -5.31003773e-01 6.45777464e-01 1.62948394e+00
4.66712296e-01 -1.07757442e-01 3.33878577e-01 3.12085867e-01
3.76833268e-02 1.05879474e+00 -3.22610438e-01 -2.74230033e-01
-1.24970652e-01 1.93976212e+00 -7.28936017e-01 -5.92595577e-01
-1.18984319e-01 1.02486360e+00 9.31397974e-02 1.11735411e-01
-7.26791263e-01 -8.88454616e-02 4.42596018e-01 -5.10778308e-01
-3.10640693e-01 4.66330409e-01 -2.39785776e-01 -8.63084793e-01
-4.88808155e-01 -1.26495016e+00 8.94570172e-01 -1.24692357e+00
-1.28137982e+00 7.50316024e-01 -5.81445873e-01 -1.40433109e+00
8.52072686e-02 -7.73291826e-01 -2.49682769e-01 4.72469658e-01
-9.51971889e-01 -1.43827069e+00 -7.29207695e-02 6.99088216e-01
-1.10411979e-02 -1.15206420e-01 1.64314055e+00 6.68750703e-01
-1.98649183e-01 3.24038714e-01 -3.01295012e-01 -1.97630376e-03
1.42521989e+00 -1.35114694e+00 -1.28276408e+00 -1.84796348e-01
7.33431339e-01 2.32119504e-02 7.24587321e-01 -3.39707285e-01
-7.81094730e-01 -8.95003200e-01 1.30288756e+00 -3.45032334e-01
9.63197112e-01 6.84354603e-02 -2.98610568e-01 6.05396569e-01
4.79706138e-01 -4.55211639e-01 1.65253592e+00 6.54494047e-01
-6.85906470e-01 1.13688752e-01 -8.98618400e-01 1.77620769e-01
4.72923934e-01 -9.48015213e-01 -4.24813539e-01 4.99299705e-01
1.63899392e-01 1.98154077e-01 -1.53790498e+00 5.16420722e-01
1.01710677e+00 -8.32528889e-01 6.83373868e-01 -7.36436248e-01
1.97716758e-01 -5.68764567e-01 -3.49651009e-01 -1.15454257e+00
-2.50964791e-01 -5.00158191e-01 3.97821218e-01 1.43184698e+00
5.23917913e-01 -2.75911778e-01 3.03207487e-01 4.42124084e-02
-5.09439372e-02 -1.75217703e-01 -1.12848997e+00 -6.59891963e-01
-6.39514625e-02 -8.90518486e-01 3.57870936e-01 1.83297038e+00
6.90106332e-01 5.70880830e-01 -7.02898502e-01 -1.62954494e-01
1.64089546e-01 6.12035930e-01 8.27607334e-01 -1.73319685e+00
-4.53047514e-01 -8.08043063e-01 -7.85011530e-01 1.07754253e-01
8.04624021e-01 -1.48910952e+00 -2.41638020e-01 -1.03715670e+00
2.15737939e-01 -3.84528250e-01 -9.30580437e-01 8.33747983e-01
5.19958735e-01 1.01277459e+00 1.51551411e-01 1.11740999e-01
-1.00368679e+00 -6.85471594e-02 8.21108401e-01 -1.80661753e-01
-1.24525324e-01 -2.03171611e-01 -5.38192511e-01 1.10435140e+00
7.39747465e-01 -7.61786580e-01 1.86427161e-01 3.43248159e-01
1.00249231e+00 -2.08052993e-02 1.93936259e-01 -9.58144724e-01
3.55730429e-02 1.08836509e-01 6.74128458e-02 -3.87890995e-01
4.00693476e-01 -1.04983330e+00 7.25316107e-01 1.56348616e-01
-5.22414625e-01 -1.49147421e-01 3.50628495e-01 7.43315667e-02
-5.59329629e-01 -5.78901887e-01 7.82017827e-01 4.72660363e-02
-6.39371097e-01 -2.65588343e-01 -5.52481294e-01 -2.24885538e-01
9.28779602e-01 5.39537631e-02 2.82481641e-01 -4.18113321e-01
-1.62209332e+00 -4.37773943e-01 2.93324023e-01 2.96167523e-01
-1.88841015e-01 -1.76953077e+00 -6.81045532e-01 -5.14382243e-01
5.75548828e-01 -1.21336210e+00 2.24191427e-01 1.09660220e+00
-1.72289517e-02 1.22618914e-01 -2.82270730e-01 -3.59596133e-01
-1.81866586e+00 6.28325343e-01 3.65194768e-01 -4.99707818e-01
1.40409991e-01 5.86844325e-01 -1.58483982e-01 -8.59385192e-01
3.09347242e-01 4.01178122e-01 -6.33600712e-01 8.81573319e-01
2.38547638e-01 3.07650834e-01 -1.15675725e-01 -1.40206790e+00
-1.03196926e-01 1.00973034e+00 7.97217309e-01 -3.46237510e-01
9.61997330e-01 9.82887670e-02 -1.06695497e+00 1.47701097e+00
8.76275957e-01 4.34224188e-01 -2.84177512e-02 7.97645152e-02
1.66356310e-01 1.94401622e-01 9.98495147e-03 -1.42829013e+00
-7.53309488e-01 3.86517435e-01 8.90992284e-01 4.67554331e-01
1.35431278e+00 -5.25426678e-02 1.34330824e-01 2.50725746e-01
2.81564683e-01 -1.52933776e+00 -1.78860873e-01 3.49873185e-01
9.69187975e-01 -8.12467575e-01 -2.19092935e-01 -3.51128042e-01
-8.44913423e-01 1.23290825e+00 7.65167177e-02 1.08846270e-01
6.03253126e-01 3.68271470e-01 6.92604899e-01 -4.79458600e-01
-4.91154402e-01 -4.74813879e-01 7.56885827e-01 1.62903503e-01
9.14839625e-01 6.30968988e-01 -8.52720916e-01 1.02005494e+00
-2.52536476e-01 -7.77043926e-04 5.14549166e-02 5.04930854e-01
-2.22952917e-01 -1.42328477e+00 -5.76471746e-01 1.36614919e-01
-8.14673424e-01 -3.62832798e-05 -1.04586279e+00 5.71146131e-01
9.88032043e-01 1.21121323e+00 -2.87977755e-01 -5.63311338e-01
3.88037413e-01 7.83822060e-01 4.94844735e-01 -3.95626426e-01
-1.33268464e+00 4.98077184e-01 5.57372749e-01 -3.29211801e-01
-1.34254587e+00 -6.36232197e-01 -1.17736948e+00 3.78895283e-01
-2.08067954e-01 6.79185569e-01 8.96106958e-01 8.70942354e-01
-5.42271174e-02 5.57102919e-01 6.14868581e-01 -8.25029194e-01
2.90028363e-01 -9.95831370e-01 -1.23372245e+00 7.39453793e-01
-2.22192302e-01 -5.93554854e-01 -7.19309926e-01 4.17626441e-01] | [15.880032539367676, 5.201945781707764] |
a2f5d081-dabe-43d9-967a-6e930a77f3e8 | receptive-field-regularization-techniques-for | 2105.12395 | null | https://arxiv.org/abs/2105.12395v1 | https://arxiv.org/pdf/2105.12395v1.pdf | Receptive Field Regularization Techniques for Audio Classification and Tagging with Deep Convolutional Neural Networks | In this paper, we study the performance of variants of well-known Convolutional Neural Network (CNN) architectures on different audio tasks. We show that tuning the Receptive Field (RF) of CNNs is crucial to their generalization. An insufficient RF limits the CNN's ability to fit the training data. In contrast, CNNs with an excessive RF tend to over-fit the training data and fail to generalize to unseen testing data. As state-of-the-art CNN architectures-in computer vision and other domains-tend to go deeper in terms of number of layers, their RF size increases and therefore they degrade in performance in several audio classification and tagging tasks. We study well-known CNN architectures and how their building blocks affect their receptive field. We propose several systematic approaches to control the RF of CNNs and systematically test the resulting architectures on different audio classification and tagging tasks and datasets. The experiments show that regularizing the RF of CNNs using our proposed approaches can drastically improve the generalization of models, out-performing complex architectures and pre-trained models on larger datasets. The proposed CNNs achieve state-of-the-art results in multiple tasks, from acoustic scene classification to emotion and theme detection in music to instrument recognition, as demonstrated by top ranks in several pertinent challenges (DCASE, MediaEval). | ['Gerhard Widmer', 'Hamid Eghbal-zadeh', 'Khaled Koutini'] | 2021-05-26 | null | null | null | null | ['instrument-recognition'] | ['audio'] | [ 2.10499868e-01 -2.69075871e-01 2.07910001e-01 -5.08910954e-01
-3.06389153e-01 -6.18724406e-01 1.52491838e-01 -2.70593971e-01
-6.47325456e-01 3.19717944e-01 -2.23416984e-01 -5.70252426e-02
-1.53672487e-01 -5.39925933e-01 -7.68411756e-01 -5.24738014e-01
-2.47649178e-01 5.53597994e-02 4.77665335e-01 -3.30746442e-01
8.52651671e-02 5.26414275e-01 -1.87653136e+00 6.11515999e-01
2.97700852e-01 1.56256878e+00 1.01993412e-01 7.16368616e-01
2.39232734e-01 6.27582610e-01 -7.68268883e-01 -1.31315678e-01
9.84364673e-02 -5.69596961e-02 -8.46467197e-01 -3.99715036e-01
6.96628153e-01 3.15312803e-01 -1.59754544e-01 9.57054019e-01
7.97379851e-01 1.76196784e-01 6.24629855e-01 -1.14167714e+00
-5.93765020e-01 7.36629963e-01 -1.08469963e-01 4.64862406e-01
-7.79777840e-02 -1.33137375e-01 9.38619494e-01 -7.94672370e-01
2.49936268e-01 1.06256270e+00 1.10844803e+00 8.60842407e-01
-1.16527641e+00 -8.97675335e-01 2.13460237e-01 -6.09512702e-02
-1.63352156e+00 -5.73558152e-01 7.44899154e-01 -4.67628002e-01
9.69681442e-01 2.31153235e-01 5.12879431e-01 1.11739111e+00
5.07587306e-02 3.68942767e-01 7.40336537e-01 -4.62373316e-01
2.55528271e-01 1.65474024e-02 -2.89725494e-02 2.62414753e-01
-1.09855095e-02 -4.72201370e-02 -6.63480759e-01 1.75864041e-01
7.07881868e-01 -2.26177156e-01 -2.67400652e-01 -6.27803802e-03
-1.01909256e+00 6.42560959e-01 6.01101220e-01 8.01442146e-01
-1.98373511e-01 4.23897028e-01 6.13482654e-01 4.20922458e-01
4.17161226e-01 9.18572605e-01 -8.32138360e-01 1.50241628e-01
-9.26305532e-01 2.58215159e-01 5.01797915e-01 6.54344559e-01
5.54432213e-01 3.78498405e-01 -2.01091930e-01 1.26738608e+00
-1.61180332e-01 -4.08188924e-02 8.16656649e-01 -6.62979066e-01
4.08645064e-01 4.73629117e-01 -3.48255992e-01 -8.27959418e-01
-6.99651062e-01 -1.10422289e+00 -1.09574759e+00 3.71378288e-02
3.75148058e-01 -2.18058914e-01 -9.68412161e-01 1.90447879e+00
-1.91097662e-01 1.66810647e-01 5.07334098e-02 8.67750108e-01
1.08172786e+00 5.26153386e-01 1.84132084e-01 1.06890202e-01
1.29839492e+00 -6.84894264e-01 -5.13156354e-01 -3.43281001e-01
4.81590599e-01 -8.37646484e-01 1.25039256e+00 7.14001238e-01
-1.00620854e+00 -1.28768051e+00 -1.13280571e+00 2.83723056e-01
-5.87325990e-01 5.40404141e-01 4.90605533e-01 6.58404469e-01
-1.15822041e+00 8.77136111e-01 -4.96068120e-01 -3.61482710e-01
6.22362494e-01 8.85716021e-01 -2.28556454e-01 3.60641152e-01
-1.12253165e+00 3.62151593e-01 7.07008719e-01 2.45649487e-01
-1.02987218e+00 -6.30495787e-01 -3.73861879e-01 3.52449387e-01
5.49899228e-02 -5.03146708e-01 1.30294669e+00 -1.07539856e+00
-1.52476931e+00 7.72294819e-01 5.23350298e-01 -6.38110697e-01
1.44599795e-01 -3.35136920e-01 -6.90874636e-01 -2.04916000e-01
-3.80031049e-01 1.21036446e+00 1.06447160e+00 -9.06562209e-01
-5.48884034e-01 -2.96302903e-02 1.40036941e-01 -2.02078402e-01
-8.51399481e-01 -2.65723723e-03 -1.14826664e-01 -8.48496914e-01
2.49262974e-01 -9.93880987e-01 -1.58777401e-01 -4.30886000e-01
-3.89289051e-01 -2.73321360e-01 7.67376959e-01 1.18387267e-01
1.29593682e+00 -2.48125410e+00 3.18800397e-02 4.66783978e-02
-3.60269882e-02 4.50541705e-01 -3.41291010e-01 7.28770867e-02
-3.30931544e-01 2.50671804e-01 1.23408534e-01 -7.74387941e-02
-3.50258239e-02 2.10185617e-01 -2.83662200e-01 2.35579982e-01
2.74145037e-01 6.32272720e-01 -4.01933134e-01 -1.99200079e-01
-5.90617247e-02 6.30020499e-01 -8.41377854e-01 2.93850929e-01
-1.02081597e-01 4.52189803e-01 -1.64622262e-01 4.25923437e-01
5.54754019e-01 -1.75564528e-01 -1.04601800e-01 -3.68437380e-01
-7.01223165e-02 4.41269606e-01 -1.14056885e+00 1.75873423e+00
-6.20299578e-01 7.91440904e-01 -4.37833220e-02 -1.30401850e+00
1.15268326e+00 5.31860948e-01 2.14089349e-01 -4.32833642e-01
2.29069039e-01 2.64973938e-01 4.14230645e-01 -3.36083591e-01
3.53217900e-01 -1.24671854e-01 -1.83335200e-01 7.56637380e-03
6.93991005e-01 3.25873457e-02 -6.63863914e-03 -3.95183176e-01
9.14376259e-01 -2.00534448e-01 -4.16942798e-02 -4.51136798e-01
6.73293948e-01 -4.96158898e-01 4.74280864e-01 1.00534308e+00
1.40060887e-01 6.98707163e-01 2.93136925e-01 -7.26064563e-01
-7.94147074e-01 -5.92540741e-01 -3.97354007e-01 1.70469737e+00
-5.60193658e-01 -3.58348548e-01 -9.55590308e-01 -5.79575896e-01
-2.84723610e-01 1.87056810e-01 -7.48805940e-01 -3.14569384e-01
-5.06670833e-01 -7.18893528e-01 1.18964171e+00 6.73264563e-01
4.70478415e-01 -1.43252420e+00 -7.20067918e-01 2.93473393e-01
9.74745974e-02 -1.21406627e+00 -5.95578179e-02 8.67629647e-01
-9.40926015e-01 -7.71742940e-01 -5.98406792e-01 -9.87173498e-01
3.11525255e-01 -1.78010449e-01 1.25547552e+00 4.74789254e-02
-2.28137910e-01 1.48847103e-01 -4.67771202e-01 -7.31375337e-01
-3.56438875e-01 7.13014841e-01 2.62049496e-01 1.07676379e-01
2.17748865e-01 -8.53736699e-01 -5.70436478e-01 4.19442087e-01
-1.14921439e+00 -5.43732584e-01 4.54286426e-01 8.64791095e-01
4.53845471e-01 1.99192435e-01 7.63544083e-01 -5.86856306e-01
6.49407744e-01 -5.04126120e-03 -3.96040231e-01 -1.73906479e-02
-2.97803253e-01 -1.28578283e-02 7.39913702e-01 -9.78840828e-01
-4.00688648e-01 1.55394316e-01 -3.40513319e-01 -6.39208853e-01
-5.32576144e-01 3.05873543e-01 1.72087550e-02 -2.69584894e-01
8.73770177e-01 -2.65052587e-01 -5.46607494e-01 -6.90165043e-01
-7.34608322e-02 5.95327437e-01 4.80282217e-01 -4.32513714e-01
5.13699174e-01 1.99514449e-01 6.58785999e-02 -8.14545453e-01
-1.03532231e+00 -3.86365682e-01 -7.32042670e-01 -2.61295825e-01
8.83485258e-01 -7.75778592e-01 -8.22599590e-01 3.99134576e-01
-1.11384857e+00 -1.82986781e-01 -2.92413175e-01 6.00023270e-01
-3.39079887e-01 -3.97002190e-01 -3.84583652e-01 -7.17749894e-01
-2.34250084e-01 -1.13479125e+00 9.76410806e-01 1.67165324e-01
-2.58726865e-01 -8.26044500e-01 2.03859225e-01 -1.69341132e-01
6.54925942e-01 -1.24843784e-01 9.44532394e-01 -9.37404811e-01
-3.88978533e-02 -1.07330509e-01 7.04066071e-04 8.14702690e-01
-2.50138164e-01 -2.02511817e-01 -1.88157785e+00 -2.11339846e-01
-1.79944366e-01 -4.86428320e-01 1.31321216e+00 5.84095836e-01
1.65797782e+00 -1.78512424e-01 -6.25824407e-02 7.51210928e-01
1.23996031e+00 1.79941580e-01 6.53437972e-01 3.45296890e-01
4.30656075e-01 5.51413059e-01 2.31654003e-01 2.74137139e-01
-4.35449064e-01 9.43189979e-01 6.47325814e-01 -2.19430551e-01
-1.56509995e-01 -9.74112004e-02 3.18472624e-01 8.87582719e-01
-3.61308068e-01 -2.31345162e-01 -8.23993504e-01 4.55014020e-01
-1.56005287e+00 -7.07910001e-01 1.15383305e-01 2.02089882e+00
6.82256222e-01 2.57841080e-01 2.05981940e-01 6.49382234e-01
5.51076829e-01 3.40431482e-02 -1.06663100e-01 -5.14854133e-01
-2.42286280e-01 7.26639152e-01 2.19293132e-01 3.86774465e-02
-1.39537454e+00 1.05170584e+00 6.74236965e+00 9.65061724e-01
-1.53985298e+00 2.63362259e-01 5.49970508e-01 -9.65219364e-02
3.24521959e-01 -3.86868536e-01 -8.40321422e-01 1.00773368e-02
1.08083248e+00 7.63317883e-01 3.05349648e-01 9.34267759e-01
-1.40972674e-01 4.06920075e-01 -1.23312783e+00 1.09002221e+00
-1.58697948e-01 -1.34782875e+00 2.19185855e-02 1.55361909e-02
5.84156811e-01 1.28445745e-01 4.11906928e-01 5.46905160e-01
-1.40003264e-01 -1.32983923e+00 8.54258835e-01 3.69541705e-01
7.92346597e-01 -8.41326058e-01 9.52219963e-01 -2.06328724e-02
-1.19374812e+00 -4.89885241e-01 -8.14017415e-01 -2.01088548e-01
-3.67564321e-01 4.35110360e-01 -9.08467293e-01 1.54932499e-01
1.25837517e+00 4.21637475e-01 -8.68775487e-01 1.06204212e+00
2.15538815e-01 8.55887711e-01 -1.80911928e-01 -9.77191105e-02
3.53311896e-01 3.57291043e-01 2.80129313e-01 1.52151394e+00
4.54856247e-01 -3.05106938e-01 1.89979021e-02 7.38846540e-01
-3.01555961e-01 2.76531339e-01 -4.42652076e-01 -3.08652706e-02
2.75631458e-01 1.21648240e+00 -7.26541400e-01 -8.73670354e-02
1.20843342e-03 4.65791166e-01 3.01383048e-01 4.02423918e-01
-7.22712159e-01 -4.06862676e-01 6.66953862e-01 2.15089396e-01
6.39264226e-01 1.56748369e-01 -1.81276157e-01 -7.77830124e-01
-6.03341907e-02 -7.36627877e-01 3.01959187e-01 -7.81029940e-01
-1.18700707e+00 1.25546420e+00 -2.27476403e-01 -1.41396248e+00
-1.28134906e-01 -9.75089014e-01 -5.53372920e-01 4.81251955e-01
-1.35057163e+00 -1.04070330e+00 -1.66289747e-01 7.19731271e-01
3.90971571e-01 -6.75664127e-01 1.12946451e+00 6.13435328e-01
-4.41939473e-01 7.52775133e-01 -2.39956155e-01 3.58544350e-01
7.62694836e-01 -8.92760456e-01 2.12314412e-01 3.36906254e-01
7.16585398e-01 5.30127704e-01 5.46022713e-01 3.81349511e-02
-7.84128666e-01 -1.14091802e+00 5.64289033e-01 2.53133569e-02
4.78538752e-01 -6.54261768e-01 -9.11324561e-01 3.39767128e-01
1.61448061e-01 2.65103430e-01 8.31187725e-01 6.63926363e-01
-7.21594036e-01 -5.62752128e-01 -6.64341986e-01 1.42533347e-01
9.92786944e-01 -6.82658195e-01 -3.11492354e-01 7.58265033e-02
5.71257472e-01 -1.11943468e-01 -7.54184544e-01 7.59328723e-01
7.81356454e-01 -1.17425406e+00 1.06690383e+00 -1.00574541e+00
2.63512194e-01 -2.08306938e-01 -3.13164085e-01 -1.25035632e+00
-4.67456669e-01 -5.05527496e-01 2.05365747e-01 1.17600965e+00
6.09736145e-01 -2.61348575e-01 6.00498140e-01 -3.46685141e-01
-4.08712894e-01 -7.05809236e-01 -1.10813463e+00 -8.05407226e-01
6.51941597e-02 -8.77242267e-01 5.55433452e-01 9.21459854e-01
-3.44049573e-01 4.99167472e-01 -1.77371666e-01 1.80454493e-01
-6.66524991e-02 -3.33812952e-01 6.10989809e-01 -1.72467506e+00
-3.38721126e-01 -5.19172549e-01 -8.61242056e-01 -8.04465473e-01
1.26085952e-01 -6.90200806e-01 -1.39904451e-02 -8.89774740e-01
-1.66718125e-01 -6.75096869e-01 -7.46202409e-01 6.50099933e-01
3.26683789e-01 8.41636658e-01 2.51621902e-01 9.39141139e-02
-7.62032211e-01 1.08993113e-01 8.97201598e-01 -1.66202828e-01
-3.44698936e-01 3.04988205e-01 -3.89940798e-01 8.52975786e-01
9.86286879e-01 -6.69258595e-01 -2.90871531e-01 -3.93165201e-01
4.93377358e-01 -3.92336220e-01 6.71086669e-01 -1.72100365e+00
6.23307861e-02 3.34819227e-01 6.30403697e-01 -4.00820762e-01
5.13571978e-01 -7.76664078e-01 -1.33982450e-02 1.75186440e-01
-7.12407351e-01 4.00406122e-02 5.96198678e-01 2.21565872e-01
-6.06786191e-01 -5.39557934e-01 8.69764686e-01 -6.39839470e-02
-3.73687088e-01 1.19859181e-01 -3.98733109e-01 4.20883037e-02
3.39967698e-01 -1.90004826e-01 2.26293672e-02 -3.21217030e-01
-1.08982098e+00 -5.45074522e-01 -1.21870607e-01 5.89412987e-01
4.03985858e-01 -1.35291052e+00 -5.59870481e-01 3.35353523e-01
1.14430532e-01 -8.60823691e-02 3.21028024e-01 7.36828089e-01
-1.72493860e-01 6.65713608e-01 -4.65408891e-01 -1.02998126e+00
-1.22492909e+00 3.73684555e-01 7.02052236e-01 -1.59914225e-01
-1.22746058e-01 1.19065821e+00 4.66081142e-01 -5.61988294e-01
8.20074499e-01 -7.70532608e-01 -5.35012603e-01 1.42322943e-01
3.96821231e-01 3.28980270e-04 4.85091716e-01 -4.08710361e-01
-2.11912125e-01 6.96770847e-01 -5.13893254e-02 7.70939365e-02
1.50637436e+00 3.61386687e-01 1.28760621e-01 5.67689836e-01
1.18157780e+00 -2.63739526e-01 -9.82921064e-01 9.14227217e-03
-3.50918993e-02 -2.81264167e-03 3.20903659e-01 -6.49820566e-01
-1.29372334e+00 1.39553869e+00 1.02907801e+00 6.06263816e-01
1.32559586e+00 4.41263393e-02 2.99462914e-01 5.45404911e-01
7.06214532e-02 -1.18793488e+00 3.19490105e-01 7.67039418e-01
8.48631322e-01 -8.33018124e-01 -3.73147398e-01 -1.42242551e-01
-5.94703496e-01 1.34824157e+00 5.76113641e-01 -2.43502259e-01
8.58014345e-01 3.96389306e-01 1.42677382e-01 -2.50438511e-01
-7.84212410e-01 -2.22031981e-01 7.23477840e-01 6.57546937e-01
7.65840769e-01 -1.26523882e-01 3.14890325e-01 9.22159016e-01
-5.01215100e-01 -6.22556172e-02 1.37568951e-01 5.08418739e-01
-3.86766106e-01 -1.00377142e+00 -3.88257653e-01 1.41425550e-01
-9.11485255e-01 -1.55852839e-01 -4.60636705e-01 7.71612942e-01
6.64976418e-01 8.30231428e-01 2.55466461e-01 -7.44376659e-01
4.42101449e-01 3.21777225e-01 4.32476342e-01 -6.16636693e-01
-1.15423989e+00 3.76145095e-01 -8.01148918e-03 -2.89150059e-01
-7.95435905e-01 -2.83312708e-01 -8.06844711e-01 2.99752116e-01
-6.95474207e-01 2.59949535e-01 6.55393064e-01 8.26251864e-01
2.51192898e-01 1.00198305e+00 3.61442864e-01 -9.20192301e-01
-5.05164385e-01 -1.39037418e+00 -5.27734995e-01 9.70009789e-02
3.19545984e-01 -7.38041401e-01 -2.04850271e-01 1.09220900e-01] | [15.384377479553223, 5.197680473327637] |
f68d1b3c-b7f6-45ba-a6bf-9b6efc80c5eb | learning-from-noisy-labels-with-noise | 2005.00596 | null | https://arxiv.org/abs/2005.00596v1 | https://arxiv.org/pdf/2005.00596v1.pdf | Learning from Noisy Labels with Noise Modeling Network | Multi-label image classification has generated significant interest in recent years and the performance of such systems often suffers from the not so infrequent occurrence of incorrect or missing labels in the training data. In this paper, we extend the state-of the-art of training classifiers to jointly deal with both forms of errorful data. We accomplish this by modeling noisy and missing labels in multi-label images with a new Noise Modeling Network (NMN) that follows our convolutional neural network (CNN), integrates with it, forming an end-to-end deep learning system, which can jointly learn the noise distribution and CNN parameters. The NMN learns the distribution of noise patterns directly from the noisy data without the need for any clean training data. The NMN can model label noise that depends only on the true label or is also dependent on the image features. We show that the integrated NMN/CNN learning system consistently improves the classification performance, for different levels of label noise, on the MSR-COCO dataset and MSR-VTT dataset. We also show that noise performance improvements are obtained when multiple instance learning methods are used. | ['Man-Hung Siu', 'Zhuolin Jiang', 'Herbert Gish', 'Jan Silovsky', 'William Hartmann', 'Sancar Adali'] | 2020-05-01 | null | null | null | null | ['multi-label-image-classification'] | ['computer-vision'] | [ 2.15957135e-01 -4.35980886e-01 1.12684362e-01 -6.29920006e-01
-1.26128852e+00 -5.37508905e-01 3.23072135e-01 2.35453218e-01
-6.81552827e-01 5.68972349e-01 -3.70934010e-01 6.17496297e-02
3.62836719e-02 -5.24510264e-01 -7.44724870e-01 -8.46610785e-01
6.15277648e-01 5.35273492e-01 -1.26608163e-02 1.13718964e-01
-1.58960879e-01 2.23026827e-01 -1.70983040e+00 5.96756399e-01
4.55711007e-01 1.47646296e+00 6.21248223e-02 8.27763855e-01
-1.83266699e-01 1.32829046e+00 -8.53440404e-01 -4.61727113e-01
2.90273577e-01 -2.54456669e-01 -5.88073492e-01 2.12691799e-01
7.75319815e-01 1.46012485e-01 -3.07213634e-01 1.31398976e+00
7.04592347e-01 6.76252227e-03 7.02410638e-01 -1.16726291e+00
-5.54687619e-01 4.31891590e-01 -5.95411718e-01 -1.04872555e-01
-3.37669820e-01 7.46550858e-02 8.53712380e-01 -8.10271800e-01
5.62612414e-01 1.24834216e+00 1.14273953e+00 7.50950873e-01
-1.43189740e+00 -8.19431663e-01 1.95197165e-01 1.68752253e-01
-1.37514353e+00 -8.93816128e-02 5.86052895e-01 -3.55092287e-01
7.21616805e-01 1.72791127e-02 6.89788014e-02 1.51609135e+00
-6.69991747e-02 8.55747044e-01 1.27561450e+00 -4.59577471e-01
1.76109761e-01 -6.37530442e-03 2.70435125e-01 5.22145748e-01
2.43237764e-02 1.07590951e-01 -2.66838700e-01 -1.47958949e-01
3.29288483e-01 1.36328593e-01 4.63217348e-02 -5.01830354e-02
-6.98605597e-01 7.69972444e-01 2.65898734e-01 4.09873456e-01
-1.62643597e-01 5.74080706e-01 6.58598900e-01 5.25296748e-01
8.29375982e-01 2.46286124e-01 -8.27955008e-01 -4.04762775e-02
-9.81115282e-01 1.01103060e-01 7.09285617e-01 8.03072870e-01
8.63060653e-01 -8.67926031e-02 -2.12206393e-01 1.40745735e+00
1.10356048e-01 5.06548882e-01 4.35697526e-01 -9.57886696e-01
3.63874972e-01 5.49561977e-01 -4.96808365e-02 -7.72234619e-01
-7.70216167e-01 -8.87323201e-01 -1.12143385e+00 2.40860894e-01
3.97520572e-01 -2.57261515e-01 -1.34959257e+00 1.82740438e+00
1.98987439e-01 4.92541879e-01 -6.65029660e-02 4.18253154e-01
8.93987894e-01 2.62475640e-01 3.25789154e-01 -8.91498700e-02
1.06333804e+00 -1.14101374e+00 -1.04425609e+00 -4.34608579e-01
9.16290760e-01 -7.94695556e-01 7.59676635e-01 6.44524574e-01
-6.74908102e-01 -6.83650732e-01 -8.62310708e-01 -9.20258537e-02
-5.49571514e-01 5.00858188e-01 1.04823247e-01 5.26139915e-01
-8.41214776e-01 7.18353927e-01 -6.39452457e-01 -2.64656749e-02
8.13645422e-01 3.51735353e-01 -5.04291236e-01 -6.03465140e-01
-9.25254285e-01 9.43084300e-01 2.80781865e-01 3.07245433e-01
-1.04580665e+00 -6.50417447e-01 -7.62169480e-01 -8.34606439e-02
6.13961041e-01 -4.32671040e-01 1.47021520e+00 -1.15879583e+00
-1.01985908e+00 9.01091039e-01 1.14860520e-01 -2.71178246e-01
7.62975335e-01 -6.44400343e-02 -4.07653362e-01 -1.64049163e-01
1.79367706e-01 5.66070437e-01 7.77274609e-01 -1.61090291e+00
-8.03104818e-01 -4.35755223e-01 -3.21498007e-01 -1.53873116e-01
5.84064201e-02 -6.93074390e-02 -2.63944775e-01 -8.76304507e-01
1.78239554e-01 -1.04794133e+00 -2.97618538e-01 -1.62056208e-01
-4.78504449e-01 -2.55878001e-01 8.07308912e-01 -2.97792554e-01
9.74114478e-01 -2.21702147e+00 -6.91291178e-03 2.07559556e-01
2.06845343e-01 3.26147705e-01 -5.19999921e-01 4.12538163e-02
-3.41741055e-01 3.04614633e-01 -1.56665623e-01 -9.58227873e-01
-1.41827866e-01 5.74984908e-01 2.50045985e-01 4.30528373e-01
2.90355116e-01 7.61442065e-01 -8.04403365e-01 -1.04570179e-03
7.42461085e-02 5.33473969e-01 -9.69669670e-02 1.90125704e-01
-4.94036287e-01 4.76136327e-01 5.77778742e-02 6.14011645e-01
8.61446559e-01 -5.08944750e-01 1.18619598e-01 -4.66768175e-01
3.91553909e-01 -1.50749654e-01 -1.55243540e+00 1.73683882e+00
-6.17573917e-01 3.17668587e-01 2.60607779e-01 -9.86079454e-01
7.72183716e-01 4.75332081e-01 3.79436553e-01 -6.94438219e-01
4.57824647e-01 5.08214056e-01 -2.58956790e-01 -6.83208644e-01
-1.02133751e-01 -4.06220853e-01 -2.99175028e-02 2.82186478e-01
5.36133230e-01 1.38753399e-01 3.16633612e-01 -1.40604585e-01
1.38429344e+00 -3.44901115e-01 8.02634060e-02 7.42128864e-02
2.90991575e-01 -2.41975829e-01 8.32065046e-01 1.17463422e+00
-2.19771564e-01 9.01781321e-01 4.19519484e-01 -5.53661168e-01
-1.00604618e+00 -5.70883095e-01 -2.84791380e-01 1.13799739e+00
-2.64885724e-01 -2.37946928e-01 -6.44603074e-01 -1.21471500e+00
8.46609995e-02 2.97573894e-01 -9.41616237e-01 1.67931174e-03
-4.09886032e-01 -1.17723989e+00 7.29589045e-01 5.94544888e-01
5.01717508e-01 -9.33344901e-01 1.63770139e-01 3.55433404e-01
-3.41075808e-01 -1.26325512e+00 -1.68780014e-01 1.03216982e+00
-4.13696647e-01 -1.34748757e+00 -2.80849963e-01 -7.63522506e-01
5.56247532e-01 1.08269706e-01 1.25402510e+00 4.16827500e-01
-4.93165970e-01 3.22973222e-01 -3.79377484e-01 -3.79836798e-01
-7.19467163e-01 7.17185065e-02 -2.31241792e-01 1.68182671e-01
4.65993315e-01 -4.28220570e-01 -1.99924529e-01 4.05982584e-01
-1.27268970e+00 -2.93486923e-01 4.27052170e-01 1.28399098e+00
9.02024806e-01 4.38256443e-01 7.37339914e-01 -1.45270073e+00
2.56045729e-01 -6.09517157e-01 -5.28958201e-01 2.98024237e-01
-4.92571592e-01 1.40724555e-01 5.84537864e-01 -6.52322590e-01
-7.87651122e-01 3.18056405e-01 -5.95512748e-01 -5.34945488e-01
-5.56102037e-01 2.72798359e-01 -2.98645914e-01 -1.21033981e-01
7.69750834e-01 -2.00646251e-01 -1.97713852e-01 -8.33075225e-01
3.68536115e-01 8.11800480e-01 4.04774904e-01 -4.35197234e-01
3.41920018e-01 4.25499886e-01 1.82564154e-01 -3.62748235e-01
-1.52028322e+00 -7.45114148e-01 -5.27259231e-01 -9.78683978e-02
7.53028333e-01 -1.17145336e+00 -5.75724363e-01 1.06948590e+00
-1.07278955e+00 -3.83906037e-01 -2.68084824e-01 1.32594645e-01
-4.82317984e-01 1.11866124e-01 -1.01706123e+00 -6.16542876e-01
-1.11025468e-01 -1.53907430e+00 1.10986876e+00 -3.18756998e-02
2.39903405e-01 -9.70659077e-01 -6.51418418e-02 4.84781623e-01
4.01837438e-01 2.73003459e-01 9.25848305e-01 -1.00549185e+00
-3.64821076e-01 -4.75336790e-01 -3.19842160e-01 9.98479486e-01
-9.91022773e-03 -3.36776823e-01 -1.39029264e+00 -2.96727836e-01
2.12251246e-01 -8.23201180e-01 1.30748546e+00 2.98466116e-01
1.35362053e+00 6.72886446e-02 -1.14262290e-01 4.92607325e-01
1.81744647e+00 -6.70722276e-02 3.65735292e-01 3.27279121e-01
8.92350316e-01 2.76858121e-01 2.48011068e-01 1.62307858e-01
1.84663743e-01 5.74166000e-01 6.87283993e-01 -2.12102070e-01
-4.71013248e-01 6.63838163e-02 -1.93217158e-01 6.43156648e-01
6.09281719e-01 -3.55468124e-01 -7.53040433e-01 3.73912454e-01
-2.06173301e+00 -5.54234803e-01 -2.92671084e-01 1.68074274e+00
8.69609773e-01 -1.15681015e-01 -2.43826181e-01 2.68025458e-01
7.52212942e-01 -1.08446680e-01 -6.85704410e-01 1.00662440e-01
-4.07454997e-01 2.55636513e-01 6.33167505e-01 4.05820072e-01
-1.57700789e+00 7.31253922e-01 6.17137480e+00 1.30911064e+00
-8.87256145e-01 5.51608622e-01 9.74282861e-01 -2.24561721e-01
2.03265220e-01 -5.22468507e-01 -9.89814758e-01 4.52135652e-01
9.28345501e-01 8.76363933e-01 3.47013861e-01 8.76682878e-01
-1.71316996e-01 -1.49972722e-01 -1.02023470e+00 1.15980446e+00
1.23676330e-01 -1.02233040e+00 -2.96963394e-01 -1.50203139e-01
1.05378544e+00 2.41305247e-01 1.75908998e-01 3.79935890e-01
6.84786081e-01 -1.08709908e+00 7.23819792e-01 4.38929230e-01
6.43739820e-01 -9.45251405e-01 1.30289602e+00 5.14921069e-01
-8.04777980e-01 -5.16834199e-01 -4.29891288e-01 2.14934573e-01
-2.52545238e-01 1.28729200e+00 -3.17826569e-01 3.81640881e-01
6.01093054e-01 6.43914461e-01 -8.65898788e-01 1.05157876e+00
-2.75286347e-01 6.50253952e-01 -3.85658562e-01 3.62519324e-01
1.69624105e-01 1.63185388e-01 -6.25809357e-02 1.27051389e+00
1.66569948e-01 -2.83244997e-01 4.63584691e-01 5.08333385e-01
-6.17352009e-01 6.38063475e-02 -3.17092836e-01 3.02807391e-01
3.29676867e-01 1.44292092e+00 -7.83159316e-01 -3.11518967e-01
-5.73615909e-01 8.85878742e-01 7.86147714e-01 2.82579929e-01
-6.14117563e-01 9.28520560e-02 7.19349384e-01 -2.50129938e-01
3.23276520e-01 2.41679192e-01 -4.34589356e-01 -1.13923955e+00
-3.23229581e-02 -1.23178411e+00 3.73934090e-01 -5.12057841e-01
-1.96993005e+00 6.54910564e-01 -6.48745477e-01 -8.46340358e-01
1.87800154e-01 -7.39976764e-01 3.27096693e-03 7.30132639e-01
-1.54515827e+00 -1.30184793e+00 -3.16312075e-01 2.99023956e-01
4.94764090e-01 -3.74259390e-02 9.41036582e-01 7.94444621e-01
-7.34338343e-01 7.81584144e-01 4.07292992e-01 3.53884578e-01
1.02470386e+00 -1.34910607e+00 1.16970360e-01 5.04767299e-01
2.74128497e-01 -3.11710555e-02 4.43103880e-01 -6.45108700e-01
-6.04903996e-01 -1.57786250e+00 7.75585771e-01 -4.86266553e-01
5.38076758e-01 -5.31429470e-01 -9.37325716e-01 5.55798173e-01
-6.65886179e-02 7.28329957e-01 8.34803343e-01 1.52405173e-01
-8.50120902e-01 -1.70339480e-01 -1.38880301e+00 1.13215536e-01
9.04616594e-01 -6.29615009e-01 1.10509172e-01 5.69325268e-01
6.26863360e-01 -4.69701529e-01 -8.05546165e-01 6.53221607e-01
8.50927979e-02 -7.93084800e-01 8.26728940e-01 -6.57216072e-01
2.22574428e-01 -2.83666968e-01 -4.76550877e-01 -1.56100464e+00
-3.83909523e-01 1.96249008e-01 1.65873140e-01 1.26963270e+00
7.01334655e-01 -1.32152185e-01 7.31913269e-01 4.16049570e-01
-3.11128218e-02 -6.96196198e-01 -1.16388071e+00 -5.07638276e-01
1.08367428e-01 -8.21857274e-01 3.15577269e-01 9.74468172e-01
-7.70361006e-01 3.30874324e-01 -6.68196738e-01 1.89642355e-01
6.28230035e-01 -4.63199437e-01 3.38471264e-01 -1.36314821e+00
-3.06473970e-01 -1.25697151e-01 -4.46829677e-01 -6.49915457e-01
4.42522168e-01 -9.96825695e-01 3.15620691e-01 -1.32331538e+00
2.71705687e-01 -5.70479929e-01 -5.64478219e-01 7.55192041e-01
-4.38818365e-01 6.58302724e-01 1.42940342e-01 -1.14231426e-02
-1.02891994e+00 3.72452646e-01 1.02077973e+00 -3.20289016e-01
3.76596540e-01 3.93449366e-02 -4.75364417e-01 7.53498137e-01
6.01400137e-01 -1.01876700e+00 -1.19011782e-01 -5.06800830e-01
5.09271562e-01 -2.93149978e-01 2.40403324e-01 -1.23548985e+00
2.04927742e-01 3.32886219e-01 5.28778434e-01 -4.56898034e-01
2.04345211e-01 -1.22284579e+00 1.91774890e-01 1.55810714e-01
-6.38320148e-01 -2.06667762e-02 1.62796736e-01 8.79188478e-01
-1.84428796e-01 -5.11545837e-01 1.22417080e+00 -4.13245827e-01
-2.51409173e-01 2.68475652e-01 -3.51976484e-01 2.22070277e-01
7.23412335e-01 4.52259123e-01 -3.98937762e-01 -9.48654041e-02
-1.27524006e+00 2.49311715e-01 2.38710418e-01 3.32034618e-01
2.06905216e-01 -1.46119022e+00 -5.75606287e-01 2.51393139e-01
2.32932031e-01 1.70058072e-01 4.76910233e-01 5.14382005e-01
-1.29546806e-01 -7.72513524e-02 3.20841908e-01 -6.64395511e-01
-1.25882852e+00 5.91633439e-01 7.42734134e-01 -7.11832821e-01
-2.82379538e-01 1.06519461e+00 -1.98643729e-01 -8.93485963e-01
6.73832834e-01 -2.51377672e-01 -5.64315431e-02 3.00182849e-01
6.53321862e-01 4.09056067e-01 6.06087387e-01 -5.17791390e-01
9.44636241e-02 6.08705342e-01 -2.74591625e-01 2.99562752e-01
1.35574102e+00 -1.32751241e-01 -1.32071674e-01 6.64322019e-01
1.50301719e+00 -5.91988564e-01 -1.19148207e+00 -5.44870734e-01
2.02944860e-01 -2.07798898e-01 2.94154197e-01 -1.30024612e+00
-1.43035054e+00 8.27719688e-01 1.08658350e+00 1.26446962e-01
1.06292427e+00 -1.54783428e-01 6.86409473e-01 3.24063271e-01
2.36640409e-01 -1.33698463e+00 3.50593448e-01 6.34508610e-01
3.19087505e-01 -1.77996242e+00 -5.52404046e-01 -3.98909539e-01
-3.80536377e-01 9.63677526e-01 5.64866960e-01 5.48496144e-03
1.04904568e+00 6.46894872e-01 6.65021956e-01 -1.32560894e-01
-8.86426747e-01 -3.83777916e-01 -4.11011279e-02 5.91868877e-01
7.95389339e-02 5.70055060e-02 9.22730044e-02 8.23396087e-01
4.01874632e-01 2.49615386e-02 3.91713321e-01 8.10001433e-01
-4.36845064e-01 -1.39466739e+00 -2.99202710e-01 6.26798928e-01
-8.55946124e-01 -5.88484444e-02 -8.04728791e-02 4.68552083e-01
9.25063133e-01 1.18308759e+00 -8.77803937e-02 -4.64276612e-01
5.65745950e-01 4.49662596e-01 3.41628999e-01 -7.48240411e-01
-9.83162701e-01 2.30688691e-01 1.41020038e-03 -5.11635840e-01
-4.10863847e-01 -3.71246159e-01 -9.04472589e-01 -8.69331881e-02
-6.75139248e-01 -9.48583707e-02 8.84068549e-01 1.01342750e+00
1.80242658e-01 8.01257610e-01 4.44428235e-01 -4.99466807e-01
-7.56035328e-01 -1.29680991e+00 -8.40770543e-01 8.86934757e-01
5.10300040e-01 -7.06738532e-01 -6.58694625e-01 -2.05092564e-01] | [9.406637191772461, 3.853140354156494] |
120855d9-ce91-44e6-ba21-ae38629b0b91 | generalized-multi-view-shared-subspace | 2005.06038 | null | https://arxiv.org/abs/2005.06038v1 | https://arxiv.org/pdf/2005.06038v1.pdf | Generalized Multi-view Shared Subspace Learning using View Bootstrapping | A key objective in multi-view learning is to model the information common to multiple parallel views of a class of objects/events to improve downstream learning tasks. In this context, two open research questions remain: How can we model hundreds of views per event? Can we learn robust multi-view embeddings without any knowledge of how these views are acquired? We present a neural method based on multi-view correlation to capture the information shared across a large number of views by subsampling them in a view-agnostic manner during training. To provide an upper bound on the number of views to subsample for a given embedding dimension, we analyze the error of the bootstrapped multi-view correlation objective using matrix concentration theory. Our experiments on spoken word recognition, 3D object classification and pose-invariant face recognition demonstrate the robustness of view bootstrapping to model a large number of views. Results underscore the applicability of our method for a view-agnostic learning setting. | ['Shrikanth Narayanan', 'Krishna Somandepalli'] | 2020-05-12 | null | null | null | null | ['robust-face-recognition', '3d-object-classification'] | ['computer-vision', 'computer-vision'] | [ 4.80468944e-02 -4.70985100e-02 -4.23222780e-02 -5.33976793e-01
-9.08305645e-01 -7.10428059e-01 8.03089201e-01 -2.79462576e-01
-2.31926084e-01 2.89437473e-01 4.64080334e-01 1.50039196e-01
-1.51262939e-01 -5.82602859e-01 -9.57513928e-01 -7.86513567e-01
-5.03167771e-02 4.87369120e-01 -9.88302454e-02 5.77655956e-02
6.58103079e-02 6.81015551e-01 -1.67985964e+00 6.28435552e-01
9.20777500e-04 6.47414267e-01 1.06963508e-01 7.85708904e-01
2.58906841e-01 4.55421954e-01 -4.07917887e-01 -2.07070351e-01
3.39865565e-01 -3.79484862e-01 -4.37181771e-01 5.74025571e-01
6.60730720e-01 -4.25024003e-01 -2.66675353e-01 7.86714017e-01
7.60467052e-01 1.80646449e-01 8.04750144e-01 -1.18004012e+00
-4.77916330e-01 3.35563689e-01 -6.09429181e-01 5.53237021e-01
-2.20757350e-02 -2.34728679e-01 1.00575221e+00 -1.30932140e+00
7.64270902e-01 1.19800925e+00 3.08946013e-01 7.02551425e-01
-1.41165936e+00 -7.49379158e-01 4.84452724e-01 4.94098179e-02
-1.16383266e+00 -6.91161454e-01 8.00180554e-01 -4.96316254e-01
7.55190372e-01 1.01642780e-01 6.92370296e-01 1.57483184e+00
2.20880732e-01 6.17118418e-01 1.06481612e+00 -1.76267758e-01
4.00564611e-01 2.98607945e-01 5.66504672e-02 6.64434195e-01
2.92387694e-01 1.26804084e-01 -1.21619189e+00 -3.66204172e-01
5.91131747e-01 2.73239106e-01 -9.31551605e-02 -8.24710429e-01
-9.98587668e-01 1.02605820e+00 -1.55992031e-01 -4.80984300e-02
-1.59909189e-01 -9.74608236e-04 4.51930135e-01 3.40073884e-01
6.98931456e-01 2.14529872e-01 -6.37288392e-01 5.74102066e-02
-6.67184949e-01 3.68752927e-02 8.24692190e-01 8.81677628e-01
5.28814375e-01 1.17240183e-01 2.42646173e-01 8.61083090e-01
2.86289275e-01 6.70705974e-01 3.24217170e-01 -9.73181844e-01
6.52283967e-01 4.44550812e-01 -2.63158917e-01 -5.08570492e-01
-2.99947590e-01 -3.37811083e-01 -7.30729043e-01 2.96041101e-01
4.40472782e-01 -1.73509344e-01 -7.12257802e-01 2.02827454e+00
5.01697540e-01 3.40364695e-01 8.75648037e-02 5.12087703e-01
4.55500543e-01 2.53129512e-01 -4.52257931e-01 -2.89326906e-01
1.28359711e+00 -6.16244674e-01 -1.49959430e-01 -3.92453045e-01
4.26829278e-01 -5.10222614e-01 7.18074977e-01 6.24901235e-01
-8.81953359e-01 -4.88040894e-01 -1.12879515e+00 1.77239850e-01
-2.81269997e-01 2.51136571e-02 5.43603003e-01 7.37311006e-01
-6.16900206e-01 4.63313282e-01 -1.01296914e+00 -2.65658468e-01
3.93382162e-01 3.26442271e-01 -7.74613202e-01 -3.83754939e-01
-6.51356637e-01 5.84784687e-01 -2.04030927e-02 -1.60779700e-01
-1.21526563e+00 -8.56042445e-01 -6.19646132e-01 5.02431318e-02
1.87680766e-01 -8.01917851e-01 6.76046729e-01 -6.80975020e-01
-1.12827384e+00 9.00733650e-01 -4.19773638e-01 -1.62394553e-01
1.78254202e-01 -2.50452518e-01 -2.84519047e-01 5.00183165e-01
-2.74521671e-02 2.75317103e-01 1.22564876e+00 -1.27776754e+00
-2.59098113e-01 -1.01308870e+00 1.70756299e-02 4.25473392e-01
-4.68581676e-01 -2.76741415e-01 -2.37625629e-01 -4.13228840e-01
4.10280526e-01 -1.15932751e+00 8.76730159e-02 4.02443558e-02
-2.29298193e-02 1.63435176e-01 6.83990777e-01 -2.57634461e-01
4.07070071e-01 -2.22076106e+00 5.43144047e-01 -4.68446016e-02
3.56707156e-01 -2.44803473e-01 -2.96269268e-01 7.55427063e-01
-4.57836092e-01 -7.57459551e-02 2.25708842e-01 -4.29749042e-01
-4.18625265e-01 1.98168278e-01 -6.40540421e-01 6.56677902e-01
1.45701468e-01 5.73545516e-01 -8.00842226e-01 -1.62332281e-02
-5.41388430e-03 4.60885942e-01 -7.58547664e-01 3.61746073e-01
7.84660429e-02 4.24821436e-01 -3.82631302e-01 3.99720967e-01
5.63055634e-01 -6.18562758e-01 2.80047059e-01 -3.92275214e-01
4.64861631e-01 -4.61691916e-02 -1.27424300e+00 1.71080673e+00
-6.78357542e-01 4.19988126e-01 4.80759479e-02 -1.10019624e+00
4.75613326e-01 3.85930479e-01 6.48177803e-01 -9.01540220e-02
-1.58947885e-01 -2.89107263e-01 -1.22033827e-01 -4.97432500e-01
-3.00300140e-02 -5.94336808e-01 3.38078961e-02 9.53308523e-01
5.49735487e-01 1.25659004e-01 -2.06668839e-01 4.08433795e-01
1.14878488e+00 -2.28244945e-01 3.66851211e-01 9.03136581e-02
5.79051562e-02 -6.00041986e-01 3.90612483e-01 6.87513411e-01
-3.09321121e-03 7.61238694e-01 4.51693296e-01 -2.86868900e-01
-1.12374341e+00 -1.36676180e+00 -1.60011828e-01 1.30271018e+00
-2.42574260e-01 -3.04460943e-01 -3.19973499e-01 -8.73554528e-01
4.58568037e-02 4.95665908e-01 -8.09160352e-01 -4.10054266e-01
-2.25498304e-01 -1.09555054e+00 3.79013926e-01 6.55084848e-01
-6.26782700e-02 -4.43370432e-01 -5.18887043e-01 -1.18095197e-01
2.52181906e-02 -1.31378031e+00 -5.00204742e-01 4.27753180e-01
-1.03581703e+00 -1.27381527e+00 -6.13299608e-01 -3.86924475e-01
8.20500731e-01 6.34672582e-01 1.01832330e+00 -4.12046403e-01
-2.97453076e-01 1.25225902e+00 -1.62365690e-01 -4.41473365e-01
-3.41753870e-01 -2.56240636e-01 5.25363922e-01 3.40695381e-01
4.09769654e-01 -9.80834663e-01 -4.75621581e-01 1.76066741e-01
-8.41382682e-01 -2.00042307e-01 5.53226769e-01 1.14635742e+00
3.69092047e-01 -3.95308882e-01 6.89591467e-01 -9.99821186e-01
2.72898585e-01 -4.82373714e-01 -5.12358785e-01 3.01252365e-01
-3.88566971e-01 2.91028291e-01 5.30579329e-01 -7.01393306e-01
-7.33571112e-01 1.50088310e-01 1.85152397e-01 -9.05210197e-01
2.41592359e-02 3.05135459e-01 -3.07891369e-01 7.76929930e-02
5.54583430e-01 4.84974921e-01 4.01759714e-01 -2.44227320e-01
7.09952533e-01 2.55086511e-01 -1.76029831e-01 -5.36211371e-01
7.75607407e-01 9.73202109e-01 2.59339243e-01 -1.10536468e+00
-1.10553575e+00 -5.68410039e-01 -8.82280409e-01 -2.78215706e-01
8.00902903e-01 -1.53261125e+00 -7.61649489e-01 2.26931214e-01
-9.26231444e-01 -1.17428079e-02 -2.91044354e-01 6.77338660e-01
-7.41690993e-01 3.88701141e-01 -3.12474340e-01 -7.98243165e-01
-8.35758075e-03 -9.38705683e-01 1.38955641e+00 -1.69177786e-01
1.38153005e-02 -9.88605261e-01 2.30693713e-01 4.66087461e-01
-1.75490052e-01 -1.59045741e-01 8.50357056e-01 -1.06569493e+00
-6.66233063e-01 -1.28848627e-01 1.86899021e-01 4.63483095e-01
1.70980230e-01 -3.05273175e-01 -1.38302207e+00 -6.67037725e-01
5.79802513e-01 -6.53551102e-01 1.01538384e+00 4.02609915e-01
9.94944811e-01 -1.19813152e-01 -3.42396528e-01 5.31206727e-01
1.19348300e+00 -8.35412592e-02 7.16643855e-02 -1.71998575e-01
5.48197627e-01 5.70480883e-01 2.02914327e-01 6.96063876e-01
7.58106634e-02 4.71893340e-01 3.59193265e-01 4.20709908e-01
9.74679440e-02 -2.74705082e-01 3.40668350e-01 1.00543785e+00
2.49623314e-01 -2.62696058e-01 -5.13386130e-01 5.67132950e-01
-1.51144266e+00 -1.33458889e+00 5.92242062e-01 2.36225939e+00
3.47647548e-01 1.00318655e-01 1.77657247e-01 -7.84072950e-02
4.39564675e-01 3.61178428e-01 -8.36008966e-01 5.94576746e-02
5.51288668e-03 4.81685437e-03 3.14442366e-01 2.64592588e-01
-8.71634424e-01 4.63211685e-01 6.68341303e+00 3.04376423e-01
-1.13967443e+00 2.49179587e-01 5.38702130e-01 -7.13106513e-01
-2.98366845e-01 -2.18340933e-01 -1.21843517e+00 8.65283236e-02
1.21150541e+00 -9.47217494e-02 5.34535766e-01 7.45423973e-01
-5.43014444e-02 1.01438291e-01 -1.57817292e+00 1.08927131e+00
6.42786384e-01 -1.47404325e+00 2.32440263e-01 4.35626477e-01
7.03535020e-01 3.31949890e-01 3.12408537e-01 1.76078245e-01
2.91905522e-01 -7.60114849e-01 3.15967947e-01 4.76957619e-01
5.12506425e-01 -6.29835010e-01 8.38024393e-02 6.47278666e-01
-9.65679109e-01 -3.52429867e-01 -4.02861536e-01 2.97790259e-01
4.00880463e-02 5.98627329e-01 -8.29230607e-01 5.11290967e-01
5.13419688e-01 7.01518536e-01 -5.33053517e-01 4.34507936e-01
2.88099349e-01 5.12827754e-01 -4.21476483e-01 3.37382197e-01
-1.74968973e-01 -2.09899902e-01 6.02285028e-01 5.90243757e-01
3.86685312e-01 1.00612387e-01 -1.57770187e-01 6.17549062e-01
-7.78181702e-02 -2.94091523e-01 -1.12645340e+00 -1.40940905e-01
5.24790704e-01 1.24528027e+00 -5.15775442e-01 -2.64331788e-01
-7.04733908e-01 1.01107657e+00 6.04370892e-01 3.84302557e-01
-5.04286885e-01 3.83344650e-01 6.17752612e-01 6.42895997e-02
7.31511712e-01 -3.48024428e-01 2.53459103e-02 -1.61101389e+00
6.58889711e-02 -8.65840852e-01 4.33610976e-01 -6.10731006e-01
-1.67445481e+00 2.35001832e-01 2.01661140e-01 -1.33676004e+00
-4.14907396e-01 -7.73319840e-01 -4.60936010e-01 6.45918846e-01
-9.91398633e-01 -1.20893359e+00 6.91214278e-02 4.36826319e-01
7.47618735e-01 -6.12719893e-01 9.78078127e-01 1.33887500e-01
-3.47145289e-01 4.75259006e-01 3.74468744e-01 4.89357486e-02
8.30636859e-01 -1.07104897e+00 2.70031333e-01 6.33772612e-01
7.79705465e-01 6.94413900e-01 5.36080182e-01 -3.95259947e-01
-1.77678299e+00 -1.05964494e+00 2.81613201e-01 -1.12408352e+00
5.14096797e-01 -8.00559580e-01 -7.47242451e-01 9.20765162e-01
-1.56657606e-01 5.36527336e-01 1.09758222e+00 5.81872642e-01
-8.95497322e-01 -2.72496164e-01 -8.62360179e-01 4.21157092e-01
1.13792169e+00 -1.02718592e+00 -4.88669604e-01 3.66954505e-01
4.20208871e-01 -1.17705986e-01 -9.63289201e-01 2.57323027e-01
7.91009903e-01 -1.11394882e+00 1.33820891e+00 -1.09341753e+00
4.26178068e-01 1.58624545e-01 -6.15135074e-01 -1.38113415e+00
-8.25278163e-02 -1.17755644e-01 -4.72165912e-01 1.09973776e+00
1.45231947e-01 -6.40821040e-01 8.73003125e-01 2.70888269e-01
2.76771396e-01 -8.68326306e-01 -1.23704100e+00 -8.96333694e-01
2.62030274e-01 -2.05606833e-01 1.68816075e-01 8.34243059e-01
-1.15705214e-01 8.43995273e-01 -4.69556093e-01 5.94550908e-01
8.89184773e-01 3.39273095e-01 7.22420335e-01 -1.22478509e+00
-7.42973030e-01 1.36346743e-01 -3.00654799e-01 -8.58277082e-01
4.34891552e-01 -9.83769298e-01 -5.01763880e-01 -8.43729436e-01
7.38623202e-01 1.10022508e-01 -1.64084226e-01 -8.91373958e-03
-7.26982346e-03 4.38296273e-02 1.89740255e-01 2.48033144e-02
-3.74404162e-01 7.16187000e-01 9.70052600e-01 -2.32933667e-02
1.01367377e-01 2.00984031e-01 -8.58398855e-01 7.19109833e-01
3.56529564e-01 -5.04645407e-01 -7.48978198e-01 -2.40413904e-01
3.48929226e-01 1.81620896e-01 4.17550057e-01 -7.41159081e-01
1.18070848e-01 -1.14067614e-01 7.39850760e-01 -6.57015681e-01
9.16895986e-01 -8.64204705e-01 1.18226390e-02 1.26476794e-01
-4.86867964e-01 1.32898733e-01 8.77043307e-02 1.33309197e+00
9.58542526e-02 -9.68463495e-02 7.74861693e-01 -3.95873219e-01
-2.05684841e-01 3.93375605e-01 -2.44964853e-01 3.68850797e-01
1.06481564e+00 4.59860265e-02 -1.22536503e-01 -4.70381200e-01
-1.06030023e+00 -1.29960835e-01 1.55767947e-01 5.97193778e-01
6.60296023e-01 -1.44696414e+00 -4.08226937e-01 5.83027542e-01
4.32202905e-01 -3.18352371e-01 2.66578585e-01 5.38091540e-01
3.53480130e-01 2.09724605e-01 -3.36715467e-02 -7.41047621e-01
-1.50946236e+00 6.14856124e-01 3.04169506e-01 -5.38045987e-02
-5.87687671e-01 7.79868782e-01 5.99903822e-01 -5.32300353e-01
2.01812685e-02 1.04392000e-01 1.72267556e-02 4.33485210e-01
5.94580948e-01 1.99937135e-01 5.19238785e-02 -4.06102359e-01
-1.94462359e-01 5.17972708e-01 -3.93773615e-01 -4.46536481e-01
1.57288599e+00 -2.53351510e-01 2.84584194e-01 1.13190770e+00
1.50546420e+00 5.40016405e-02 -1.55565000e+00 -2.38075942e-01
-3.82312626e-01 -3.70238245e-01 -1.73542827e-01 -3.42306346e-01
-9.49251175e-01 1.23249221e+00 7.61982501e-01 -6.56848401e-02
6.31387770e-01 2.67636657e-01 1.70716792e-01 6.32126153e-01
3.32071006e-01 -8.80449116e-01 6.25009596e-01 3.78848016e-01
9.15661573e-01 -1.25613880e+00 1.60352170e-01 -7.92014375e-02
-6.62180185e-01 1.16592956e+00 6.40269518e-01 -1.45882249e-01
9.73088562e-01 1.28886133e-01 -1.76569149e-01 -4.87942159e-01
-1.29224563e+00 1.58633620e-01 1.77287549e-01 4.99384433e-01
1.86731800e-01 -2.38812715e-01 5.79187214e-01 3.65023434e-01
2.14060396e-01 -3.57759535e-01 5.74051797e-01 7.85800219e-01
-2.85061777e-01 -9.56127107e-01 -1.30124182e-01 7.74650693e-01
-3.61025155e-01 1.76488177e-03 -3.43371421e-01 5.51107049e-01
-3.42067778e-01 6.31456196e-01 -1.22458059e-02 -3.35990548e-01
1.62932009e-01 5.94808936e-01 9.10360456e-01 -1.02744198e+00
-2.88144667e-02 1.04925558e-01 -1.47302210e-01 -4.03221518e-01
-4.87672240e-01 -9.91526484e-01 -6.27104878e-01 9.47784632e-02
-3.94560665e-01 -2.05260262e-01 4.29678559e-01 1.13696206e+00
5.48840940e-01 3.08398426e-01 9.81437802e-01 -9.06623006e-01
-1.21449971e+00 -7.29702592e-01 -8.54723096e-01 3.77933770e-01
4.92711753e-01 -6.78692698e-01 -7.73238122e-01 8.76004919e-02] | [8.43234920501709, 4.4713239669799805] |
f4fdfb90-2bfc-4405-946e-1678f98a054e | understanding-programs-by-exploiting-fuzzing | 2305.13592 | null | https://arxiv.org/abs/2305.13592v2 | https://arxiv.org/pdf/2305.13592v2.pdf | Understanding Programs by Exploiting (Fuzzing) Test Cases | Semantic understanding of programs has attracted great attention in the community. Inspired by recent successes of large language models (LLMs) in natural language understanding, tremendous progress has been made by treating programming language as another sort of natural language and training LLMs on corpora of program code. However, programs are essentially different from texts after all, in a sense that they are normally heavily structured and syntax-strict. In particular, programs and their basic units (i.e., functions and subroutines) are designed to demonstrate a variety of behaviors and/or provide possible outputs, given different inputs. The relationship between inputs and possible outputs/behaviors represents the functions/subroutines and profiles the program as a whole. Therefore, we propose to incorporate such a relationship into learning, for achieving a deeper semantic understanding of programs. To obtain inputs that are representative enough to trigger the execution of most part of the code, we resort to fuzz testing and propose fuzz tuning to boost the performance of program understanding and code representation learning, given a pre-trained LLM. The effectiveness of the proposed method is verified on two program understanding tasks including code clone detection and code classification, and it outperforms current state-of-the-arts by large margins. Code is available at https://github.com/rabbitjy/FuzzTuning. | ['Hao Chen', 'Yifeng He', 'Yiwen Guo', 'Yuyang Rong', 'Jianyu Zhao'] | 2023-05-23 | null | null | null | null | ['code-classification'] | ['computer-code'] | [ 1.59334600e-01 1.28902653e-02 -7.08963275e-01 -5.20246208e-01
-2.10064650e-01 -8.02473724e-01 5.22468865e-01 3.47205520e-01
8.87739733e-02 -1.90267675e-02 -1.00575767e-01 -7.51416445e-01
2.48498023e-01 -1.04890358e+00 -1.17377841e+00 -1.70664698e-01
5.48383892e-02 2.81131584e-02 2.93751091e-01 -2.95289487e-01
5.45903563e-01 2.64388639e-02 -1.71808338e+00 6.26679838e-01
1.24797165e+00 5.84410608e-01 5.41617453e-01 6.06440842e-01
-5.04431605e-01 9.82563257e-01 -6.81593299e-01 -4.79814976e-01
-3.65132064e-01 -4.82764900e-01 -9.21588242e-01 7.43626133e-02
1.64378002e-01 -3.00530624e-02 -3.10045809e-01 1.43111157e+00
-3.01914245e-01 -1.09894976e-01 4.28489447e-01 -1.25369751e+00
-8.19912910e-01 1.21213400e+00 -5.47815502e-01 2.85609607e-02
3.43000114e-01 3.12891841e-01 1.25528502e+00 -3.28958929e-01
4.07067835e-01 1.19038999e+00 2.24122435e-01 8.91453087e-01
-1.35767519e+00 -3.87605965e-01 1.43677756e-01 7.40958750e-03
-1.21356273e+00 -2.09948763e-01 7.09537208e-01 -6.68306649e-01
1.22998691e+00 9.65176895e-02 1.08932242e-01 1.01373339e+00
2.46014774e-01 9.19661283e-01 4.64288145e-01 -6.95516050e-01
2.45942980e-01 5.26506901e-01 5.41657805e-01 9.58659947e-01
2.10984200e-01 9.71560329e-02 -3.20241749e-02 -3.57596636e-01
3.63135308e-01 3.83943111e-01 -3.71614844e-01 -3.41157705e-01
-1.05770016e+00 9.12894130e-01 4.81067657e-01 6.84828222e-01
2.63836205e-01 4.83335733e-01 7.60260522e-01 5.27643442e-01
-2.23464876e-01 5.33634722e-01 -5.20829439e-01 -3.13640952e-01
-7.46645808e-01 1.43617854e-01 7.87546635e-01 1.03149140e+00
1.06493497e+00 -1.32313352e-02 1.98577851e-01 7.40134656e-01
2.41779223e-01 2.86168516e-01 8.52088034e-01 -5.92648208e-01
7.96241701e-01 1.34185684e+00 -3.34061116e-01 -8.68564129e-01
-8.49928111e-02 -3.52730274e-01 -5.78616440e-01 1.39826253e-01
1.90210372e-01 1.28477946e-01 -6.01425231e-01 1.90811241e+00
-7.00021237e-02 2.29166374e-01 9.04622227e-02 6.50974929e-01
5.67503512e-01 8.16814184e-01 -2.69295089e-02 4.18973356e-01
1.44593215e+00 -8.90529931e-01 -2.30033219e-01 -5.98707139e-01
1.26276422e+00 -2.60548413e-01 1.52380955e+00 2.85842597e-01
-5.52711248e-01 -8.36213768e-01 -1.13830233e+00 1.69383302e-01
-4.44406420e-01 3.04974318e-01 8.00903440e-01 8.35300207e-01
-7.91857362e-01 5.54257810e-01 -9.86369491e-01 -4.92474549e-02
3.94524306e-01 2.43977562e-01 -6.00002706e-02 -1.58048067e-02
-9.34958518e-01 4.75582361e-01 7.31824517e-01 -2.55213410e-01
-1.06427097e+00 -6.70359790e-01 -1.24948645e+00 4.72562134e-01
5.05176902e-01 -3.75015378e-01 1.39154875e+00 -1.31170309e+00
-1.06069350e+00 9.54022527e-01 -1.85467452e-01 -3.98040146e-01
7.87319988e-03 -2.82654576e-02 -3.90888959e-01 -1.85146704e-01
-1.60870388e-01 3.47243696e-01 7.75686204e-01 -1.34638500e+00
-5.10833502e-01 -2.31180012e-01 6.47489011e-01 -5.11607349e-01
-3.25124234e-01 1.27350137e-01 -3.40326875e-01 -4.66204345e-01
-1.80121064e-01 -8.32245708e-01 -3.09420153e-02 -3.80760849e-01
-4.59759504e-01 -2.59515613e-01 5.93855619e-01 -4.95721161e-01
1.50034261e+00 -2.48301315e+00 2.52761424e-01 1.67293921e-01
3.85696054e-01 2.86982387e-01 -1.98180765e-01 4.70021665e-01
-2.53782839e-01 4.68317956e-01 -3.81531715e-01 -6.76912665e-02
3.73669863e-01 2.57891953e-01 -5.82465291e-01 2.78080940e-01
3.20539504e-01 1.11694264e+00 -8.79491031e-01 -1.93986416e-01
2.19514310e-01 -7.02330172e-02 -8.63934755e-01 3.36262107e-01
-8.85417819e-01 1.48575678e-01 -6.84209704e-01 5.99205911e-01
3.80316049e-01 -4.46983308e-01 8.41167420e-02 3.22169930e-01
1.10055387e-01 4.70517904e-01 -7.50796616e-01 1.73113477e+00
-9.77409720e-01 6.16684139e-01 -2.80330241e-01 -1.33107519e+00
8.02185833e-01 3.08716059e-01 -2.32277557e-01 -3.50494504e-01
2.19151258e-01 2.85706401e-01 4.55172986e-01 -7.26735950e-01
2.33800396e-01 4.15124372e-02 -3.52559119e-01 6.64572179e-01
1.17847651e-01 -1.93764076e-01 2.58523613e-01 2.78653532e-01
1.22149920e+00 1.57666475e-01 4.17210788e-01 -2.14808434e-01
8.26684296e-01 1.84569117e-02 2.80074686e-01 7.97067940e-01
1.10796221e-01 2.60450423e-01 1.01998961e+00 -2.25299627e-01
-8.53458941e-01 -7.16579497e-01 2.17538625e-02 1.24228168e+00
1.28062695e-01 -6.69911325e-01 -9.53117490e-01 -7.95314550e-01
9.20016505e-03 1.06679285e+00 -6.68867409e-01 -6.26301229e-01
-6.76502824e-01 -3.54724377e-01 7.40350962e-01 6.06766760e-01
2.04973966e-01 -1.25564516e+00 -8.19666028e-01 4.31958921e-02
-1.19086467e-01 -8.45773876e-01 -4.83728647e-01 3.51034880e-01
-1.01900554e+00 -1.23275387e+00 -5.62960878e-02 -9.67094421e-01
9.10211921e-01 1.05291240e-01 1.42413843e+00 7.67938852e-01
-3.12656850e-01 1.68279171e-01 -5.09742618e-01 3.12282853e-02
-1.15449893e+00 1.00105606e-01 -3.45744550e-01 -1.67199269e-01
5.15073657e-01 -5.17047048e-01 -1.25843346e-01 2.47183800e-01
-1.36194181e+00 -2.28991248e-02 4.94245917e-01 8.74690711e-01
4.62822914e-02 3.60612571e-01 3.06070417e-01 -1.26709616e+00
4.23159987e-01 -6.77312374e-01 -8.36001217e-01 4.77940977e-01
-2.72486627e-01 4.42384630e-01 1.23659599e+00 -5.63914776e-01
-9.91406560e-01 -2.39423588e-01 -1.49027675e-01 -3.81018102e-01
-4.65524733e-01 6.50052190e-01 -4.46915805e-01 3.80566344e-02
9.85900998e-01 5.66312432e-01 -1.78193152e-01 -2.72325903e-01
4.32284862e-01 6.17756903e-01 3.73342752e-01 -1.19673443e+00
9.24464822e-01 -1.02351485e-02 -6.12412453e-01 -4.11323786e-01
-2.03995705e-01 -2.62695491e-01 -3.92916739e-01 1.61526293e-01
4.94316906e-01 -5.33239484e-01 -6.76951051e-01 3.85748714e-01
-1.17669582e+00 -4.67591792e-01 -3.23667713e-02 3.18244770e-02
-5.59743404e-01 3.25139284e-01 -5.40903389e-01 -5.51586449e-01
2.05503032e-02 -1.71185935e+00 9.60866630e-01 2.57344514e-01
-3.64351392e-01 -1.06258380e+00 -5.41811474e-02 6.20526038e-02
4.33347374e-01 -5.70114590e-02 1.66685069e+00 -6.90848410e-01
-7.68183649e-01 -1.31624013e-01 7.50798441e-04 4.76838112e-01
1.81229785e-01 2.15570226e-01 -8.33111286e-01 -1.47279650e-01
2.68935144e-01 -5.77901602e-01 6.54658854e-01 9.45178866e-02
1.47493982e+00 -2.90287465e-01 -5.57346880e-01 6.60440624e-01
1.65481198e+00 3.37043285e-01 5.12746096e-01 2.65870035e-01
4.62860703e-01 4.58004624e-01 2.06809655e-01 3.55547935e-01
2.00595453e-01 4.28022683e-01 6.45495713e-01 3.12192023e-01
2.45021939e-01 -3.95075381e-01 5.93587697e-01 5.07951319e-01
3.26050133e-01 -1.16977498e-01 -1.08162296e+00 5.20873070e-01
-1.68521333e+00 -7.06870496e-01 1.50687218e-01 2.16112304e+00
9.12068963e-01 2.40503103e-01 -1.89051643e-01 -2.90153828e-03
6.71334624e-01 1.84555888e-01 -6.71923101e-01 -5.75426161e-01
2.55799115e-01 3.43251109e-01 1.03291132e-01 3.36980015e-01
-8.50737274e-01 9.04768646e-01 4.62044382e+00 9.89129007e-01
-1.19950426e+00 -8.81735906e-02 5.32447219e-01 5.49986124e-01
-6.46208346e-01 2.53921747e-01 -6.27385259e-01 4.54638302e-01
1.03712106e+00 -3.40508193e-01 8.18429470e-01 1.15691411e+00
-1.27261013e-01 3.55362799e-03 -1.56393039e+00 6.19606435e-01
-8.42420533e-02 -1.20967305e+00 8.77889432e-03 -3.99592549e-01
5.95895886e-01 -4.82748970e-02 2.04445235e-03 7.75798500e-01
1.17852516e-01 -1.04933965e+00 7.51023531e-01 1.70271564e-02
5.12885213e-01 -5.55894136e-01 5.40015280e-01 7.71429241e-01
-1.23699427e+00 -3.21587414e-01 -4.28172976e-01 -5.91921806e-02
-5.46862185e-01 3.55691522e-01 -4.23090279e-01 4.01362091e-01
4.77495015e-01 6.48513675e-01 -8.29724073e-01 7.09323168e-01
-5.31352639e-01 7.32396126e-01 5.03183790e-02 -3.02906632e-01
2.13776335e-01 -1.03824213e-01 1.06736280e-01 1.22585559e+00
3.02748203e-01 -2.00509578e-01 3.73293795e-02 1.68108153e+00
-2.66722262e-01 -5.86793497e-02 -6.24290228e-01 -5.69461465e-01
4.09941733e-01 9.76741016e-01 -7.07845151e-01 -5.83965361e-01
-8.41036201e-01 6.49483979e-01 4.42343801e-01 3.01686674e-01
-1.12835550e+00 -6.00681603e-01 7.43755281e-01 -1.63676813e-01
2.77175277e-01 9.52693447e-03 -2.76871413e-01 -1.55115378e+00
2.69122839e-01 -1.36391068e+00 2.47835517e-01 -3.85873318e-01
-8.60345364e-01 5.72975159e-01 -9.28586498e-02 -1.02032363e+00
-1.92996740e-01 -6.86584353e-01 -9.86855745e-01 7.35730410e-01
-1.24459207e+00 -7.92446554e-01 -1.74411327e-01 3.24028820e-01
6.63435996e-01 -1.95995733e-01 8.06048155e-01 1.82273358e-01
-5.07414341e-01 7.88693964e-01 -2.31833328e-02 2.67781496e-01
1.13524728e-01 -1.28705227e+00 7.67619133e-01 9.93500233e-01
2.71528482e-01 1.31063390e+00 6.21188521e-01 -3.18478048e-01
-1.77772009e+00 -1.05560970e+00 6.68507755e-01 -4.43860739e-01
9.71475124e-01 -6.73478365e-01 -1.29399991e+00 9.40878034e-01
1.24124296e-01 -1.32719666e-01 4.37432349e-01 2.68116258e-02
-8.14692914e-01 5.42801544e-02 -8.89395356e-01 5.52165627e-01
7.16585994e-01 -8.54701757e-01 -8.41248751e-01 1.55282959e-01
8.93978596e-01 -3.90176445e-01 -5.59760571e-01 2.59400100e-01
2.05688670e-01 -1.07419813e+00 6.99257672e-01 -6.40823960e-01
8.41709137e-01 -3.85856509e-01 -1.79369703e-01 -1.19969380e+00
8.77942368e-02 -2.97552526e-01 -2.87001878e-01 1.17765117e+00
4.48607415e-01 -7.10875571e-01 7.24340200e-01 5.92961907e-01
-3.07112575e-01 -6.29036844e-01 -3.68664503e-01 -6.36785388e-01
2.31664121e-01 -6.83576882e-01 7.23579526e-01 8.64056289e-01
5.07552981e-01 1.85929686e-01 5.30974083e-02 2.33057395e-01
2.28827521e-01 6.66057050e-01 6.97441518e-01 -8.30193043e-01
-9.57006872e-01 -1.00228155e+00 -3.51904184e-01 -1.30760312e+00
6.75267875e-01 -1.17759216e+00 5.92069514e-02 -8.26120317e-01
3.78161132e-01 -3.69200438e-01 -1.32135868e-01 6.90786719e-01
-3.80476385e-01 -4.25655216e-01 -5.64985164e-02 1.35819092e-01
-2.49872625e-01 3.65181535e-01 7.50662267e-01 -5.86933255e-01
-1.21473573e-01 3.22787881e-01 -7.91962266e-01 6.15167499e-01
7.55967677e-01 -4.75166619e-01 -6.11054122e-01 -4.62236077e-01
3.48449707e-01 2.81327248e-01 2.66035527e-01 -8.94798100e-01
6.35183323e-03 -1.75572515e-01 -3.34776670e-01 -4.24762815e-02
-3.30642313e-01 -8.22585523e-01 -3.88848297e-02 8.24141741e-01
-6.55763090e-01 -9.13027003e-02 3.66176724e-01 3.46265197e-01
-5.36263406e-01 -9.45285678e-01 7.75555253e-01 -3.37421745e-01
-1.00279033e+00 9.10604671e-02 -3.01564932e-01 1.43350154e-01
8.83373618e-01 -7.38919422e-04 -5.77580392e-01 1.37729600e-01
-3.99072677e-01 1.00476801e-01 9.05680656e-01 7.78478980e-01
4.38847929e-01 -8.25155735e-01 -2.30971113e-01 4.48222131e-01
6.90956175e-01 -1.95551649e-01 1.71498045e-01 5.25684059e-01
-6.62918925e-01 3.86987805e-01 4.32810523e-02 -5.88831723e-01
-1.04908216e+00 8.51417542e-01 4.51758087e-01 -1.78352535e-01
-4.77776557e-01 7.35846221e-01 7.59703577e-01 -8.51237118e-01
2.62963206e-01 -9.73210037e-01 5.30712074e-03 -6.48300767e-01
6.66663766e-01 -2.59344429e-01 -5.53165600e-02 -1.14120543e-01
-3.76564026e-01 4.75386411e-01 -3.81519273e-02 6.03854477e-01
1.02811158e+00 3.31019819e-01 -4.87717569e-01 5.25178254e-01
1.25939798e+00 -1.42314406e-02 -8.70714366e-01 -3.48270416e-01
4.67043936e-01 -3.56331646e-01 -4.28306997e-01 -5.94844580e-01
-1.07736206e+00 1.27668357e+00 9.73430648e-02 3.49762529e-01
1.02782035e+00 2.98611045e-01 5.25887847e-01 6.08152092e-01
5.98325431e-01 -4.65734929e-01 1.53249055e-01 5.93958914e-01
4.01099801e-01 -1.22667813e+00 -5.68435669e-01 -2.74988383e-01
-3.01088184e-01 1.32189631e+00 9.50157285e-01 -1.61901936e-01
2.50464559e-01 4.75699395e-01 -1.27484545e-01 -1.59674585e-01
-8.32705975e-01 1.10824980e-01 1.65539578e-01 3.10758471e-01
7.23575711e-01 8.58304873e-02 2.11830456e-02 7.64951050e-01
3.71720307e-02 -7.63111338e-02 5.28501630e-01 1.04101288e+00
-7.02297449e-01 -1.23873198e+00 -2.44512811e-01 5.25352120e-01
-3.09684962e-01 -2.16838822e-01 -1.44671038e-01 7.74536014e-01
9.26715583e-02 7.84536421e-01 -3.24937969e-01 -4.87989604e-01
1.87444746e-01 2.59865344e-01 5.15640974e-01 -1.07069170e+00
-6.71199203e-01 -4.91964549e-01 -3.26894403e-01 -5.99194288e-01
7.11400583e-02 -2.45796457e-01 -1.66150904e+00 -2.90576190e-01
-4.30133462e-01 3.32376778e-01 4.08599943e-01 1.00317562e+00
2.03961179e-01 7.18540847e-01 6.45123720e-01 -4.99459743e-01
-7.86227584e-01 -7.96958029e-01 -4.88663554e-01 5.67986012e-01
3.87067795e-01 -4.23313677e-01 -5.16797721e-01 2.77497411e-01] | [7.599034786224365, 7.836158275604248] |
90df7b2d-1a08-46e6-ad59-2f3d6b9347c9 | gibbsddrm-a-partially-collapsed-gibbs-sampler | 2301.12686 | null | https://arxiv.org/abs/2301.12686v2 | https://arxiv.org/pdf/2301.12686v2.pdf | GibbsDDRM: A Partially Collapsed Gibbs Sampler for Solving Blind Inverse Problems with Denoising Diffusion Restoration | Pre-trained diffusion models have been successfully used as priors in a variety of linear inverse problems, where the goal is to reconstruct a signal from noisy linear measurements. However, existing approaches require knowledge of the linear operator. In this paper, we propose GibbsDDRM, an extension of Denoising Diffusion Restoration Models (DDRM) to a blind setting in which the linear measurement operator is unknown. GibbsDDRM constructs a joint distribution of the data, measurements, and linear operator by using a pre-trained diffusion model for the data prior, and it solves the problem by posterior sampling with an efficient variant of a Gibbs sampler. The proposed method is problem-agnostic, meaning that a pre-trained diffusion model can be applied to various inverse problems without fine-tuning. In experiments, it achieved high performance on both blind image deblurring and vocal dereverberation tasks, despite the use of simple generic priors for the underlying linear operators. | ['Stefano Ermon', 'Yuki Mitsufuji', 'Toshimitsu Uesaka', 'Yuhta Takida', 'Chieh-Hsin Lai', 'Koichi Saito', 'Naoki Murata'] | 2023-01-30 | null | null | null | null | ['deblurring', 'blind-image-deblurring'] | ['computer-vision', 'computer-vision'] | [ 4.11018133e-01 -8.59293267e-02 2.81406641e-01 -8.10845196e-02
-8.60686123e-01 -3.57049435e-01 7.77320802e-01 -6.42345250e-01
-2.44223565e-01 3.90237868e-01 5.47946274e-01 -1.03864729e-01
-2.38947704e-01 -3.07431370e-01 -5.54220200e-01 -1.11636806e+00
2.54537195e-01 5.87048769e-01 -9.86308325e-03 1.76577196e-02
2.95586824e-01 1.75447524e-01 -9.82317626e-01 -1.76013067e-01
8.79436791e-01 5.93878925e-01 5.72265863e-01 7.97320426e-01
2.61575937e-01 5.51483214e-01 -5.98959923e-01 -6.83605820e-02
2.47289717e-01 -8.39370251e-01 -6.06245816e-01 5.21193504e-01
1.21422544e-01 -3.45288903e-01 -5.30486822e-01 1.30761385e+00
7.29311645e-01 2.85168290e-01 8.36821139e-01 -6.15318537e-01
-9.66298938e-01 3.79391551e-01 -6.64819181e-01 2.85144597e-01
1.79317474e-01 -9.14619640e-02 7.97636807e-01 -9.14029837e-01
3.56179744e-01 1.31106710e+00 9.29839432e-01 5.63125312e-01
-1.60215676e+00 -3.09849381e-01 -1.71549320e-01 3.21548074e-01
-1.25688696e+00 -7.80003011e-01 7.86687911e-01 -5.40612102e-01
5.50343454e-01 7.91141242e-02 3.28903109e-01 1.11452246e+00
3.08147985e-02 6.73552036e-01 1.63691056e+00 -5.18874943e-01
5.26464045e-01 -1.93060890e-01 -2.61296518e-02 3.37562889e-01
5.08526452e-02 2.07925215e-01 -4.68061507e-01 -3.45820576e-01
8.89889121e-01 -1.61431044e-01 -7.57281959e-01 -3.94089669e-01
-1.02222669e+00 9.81688023e-01 2.39616513e-01 4.45962250e-01
-6.84417427e-01 1.50735050e-01 -1.05525389e-01 2.60917425e-01
8.43905807e-01 1.33827314e-01 -4.29539010e-02 1.18906163e-01
-1.29891765e+00 -1.21019833e-01 9.55228329e-01 6.20887637e-01
6.96273506e-01 1.15603045e-01 -2.22032383e-01 1.00293231e+00
6.88670278e-01 8.65620494e-01 7.12357461e-01 -1.21182442e+00
7.51785710e-02 -5.13668478e-01 2.65776008e-01 -7.74514616e-01
4.23672833e-02 -5.64949512e-01 -1.06192911e+00 2.75222547e-02
4.28054571e-01 -1.62047863e-01 -1.26903796e+00 1.67561257e+00
2.59308875e-01 5.74669123e-01 1.38766423e-01 1.14822638e+00
1.89521611e-01 8.02398980e-01 -4.44265038e-01 -4.42692488e-01
1.06142664e+00 -8.86979461e-01 -1.08112216e+00 -6.65670216e-01
-7.98545703e-02 -1.04519141e+00 8.38051736e-01 7.23152876e-01
-1.06303906e+00 -4.06812489e-01 -9.69413280e-01 -7.45799914e-02
3.49110901e-01 1.31088257e-01 8.49138796e-02 8.15308154e-01
-1.29502881e+00 6.72280252e-01 -9.61643815e-01 -3.03983480e-01
1.56114370e-01 1.39633879e-01 5.25868423e-02 -5.74688017e-01
-8.21390867e-01 9.78746891e-01 -2.03033715e-01 3.30891997e-01
-1.38997853e+00 -3.83190930e-01 -5.03663778e-01 -1.79396465e-01
2.32831359e-01 -7.42305696e-01 1.17791104e+00 -8.08881104e-01
-2.05153847e+00 4.76339430e-01 -3.38806391e-01 -3.36195439e-01
4.13891226e-01 -3.98831248e-01 -3.14275295e-01 1.63430214e-01
9.18902755e-02 -5.77442981e-02 1.91576147e+00 -1.40011191e+00
5.73370270e-02 -3.78391057e-01 -3.38585049e-01 8.97550285e-02
-5.85700497e-02 -3.56482506e-01 -4.84828055e-01 -9.54727173e-01
6.28202260e-01 -1.05608296e+00 -1.93918452e-01 -2.59138018e-01
-4.44744498e-01 2.07056731e-01 8.25792432e-01 -1.27413058e+00
8.70548368e-01 -2.42676592e+00 7.28397250e-01 3.01944762e-01
3.38588730e-02 9.17482302e-02 -3.31178248e-01 3.89772236e-01
-1.00332439e-01 -1.91852808e-01 -7.75464535e-01 -6.56888485e-01
5.26943058e-02 3.91677976e-01 -4.00098503e-01 1.05283010e+00
-3.28502208e-01 3.29344690e-01 -7.93791771e-01 -5.11686243e-02
1.50745317e-01 8.25670123e-01 -5.58328986e-01 4.23254550e-01
-4.37451825e-02 7.41451383e-01 -1.01248890e-01 4.50493246e-02
8.50870848e-01 -1.95847481e-01 2.61784881e-01 -2.84232557e-01
1.53161377e-01 9.80157778e-02 -1.34743810e+00 1.93448138e+00
-7.53054678e-01 7.32066751e-01 7.09130466e-01 -1.11270523e+00
6.02179885e-01 6.14247441e-01 3.40831965e-01 -2.59660929e-01
-6.48267791e-02 2.44447231e-01 -7.23737553e-02 -6.06226742e-01
7.93504044e-02 -7.41666138e-01 4.97044146e-01 7.51385987e-01
2.02340439e-01 -4.95633245e-01 -6.78297579e-02 1.87939793e-01
1.18844092e+00 6.60383552e-02 1.02347903e-01 -3.02815169e-01
4.48423475e-01 -4.15018260e-01 3.33624482e-01 8.38163972e-01
3.03498685e-01 9.16117787e-01 -8.78588855e-02 3.08846265e-01
-8.00621092e-01 -1.08423030e+00 -1.81712747e-01 7.57253885e-01
5.87387830e-02 -8.55151042e-02 -9.85183239e-01 -2.41518989e-01
-3.78751814e-01 8.25554967e-01 -2.86603272e-01 1.08555183e-01
-4.07787681e-01 -1.16313851e+00 2.16964677e-01 -7.74290189e-02
5.23480296e-01 -5.83551407e-01 2.08356505e-04 3.72703820e-01
-5.32026291e-01 -9.84933019e-01 -8.49681199e-01 -4.56509925e-02
-1.03273988e+00 -7.45963395e-01 -1.12028098e+00 -6.42976165e-01
8.79284024e-01 5.01641214e-01 7.34028161e-01 -2.44289681e-01
-1.02859154e-01 8.10814083e-01 -9.39349756e-02 6.87518865e-02
-5.77092946e-01 -3.73597443e-01 1.33385986e-01 4.74677324e-01
-1.41272515e-01 -8.51377606e-01 -6.03126585e-01 4.30482656e-01
-1.16146708e+00 -3.95345896e-01 7.04641700e-01 1.07383120e+00
4.67121661e-01 2.28199184e-01 4.03092235e-01 -6.67983830e-01
9.67872620e-01 -4.42891002e-01 -4.42692637e-01 -1.80010665e-02
-6.65992618e-01 3.38402033e-01 2.45882571e-01 -8.19566369e-01
-1.31518412e+00 -1.11934200e-01 -1.49925992e-01 -4.11471635e-01
8.31113234e-02 5.77230752e-01 -1.55708507e-01 -1.27917826e-01
7.38774180e-01 3.32337677e-01 1.49279296e-01 -1.03752398e+00
6.24774039e-01 7.14533746e-01 5.43820739e-01 -6.20813847e-01
8.50574911e-01 6.31900609e-01 -1.75699860e-01 -1.12403393e+00
-9.36253130e-01 -6.47733212e-01 -2.81814843e-01 1.43294841e-01
7.01807380e-01 -9.26207542e-01 -1.78059787e-01 6.90659940e-01
-1.15101659e+00 -4.04812932e-01 -4.78880018e-01 7.76619494e-01
-6.53356493e-01 9.00667191e-01 -6.96106136e-01 -7.92122602e-01
-1.44561365e-01 -1.13805676e+00 8.02789927e-01 -1.13799393e-01
-6.90415874e-02 -1.21101606e+00 3.16324532e-01 3.02365065e-01
6.63421094e-01 -4.78737444e-01 7.06972003e-01 -1.69565529e-01
-5.36940634e-01 -2.87552327e-01 1.10545848e-02 9.65794623e-01
3.17527592e-01 -7.05744505e-01 -1.06185508e+00 -6.61217928e-01
9.41342413e-01 6.16812296e-02 9.55437005e-01 7.75690675e-01
8.14192772e-01 -4.60525036e-01 -1.94763690e-01 5.28480113e-01
1.23307419e+00 -2.58296609e-01 6.72603309e-01 -9.79984105e-02
4.06824440e-01 3.41755748e-01 -1.75217669e-02 4.18077677e-01
2.53364816e-02 5.64808905e-01 2.88820297e-01 2.67388951e-03
-5.09707212e-01 -3.68142240e-02 5.44887125e-01 1.11713016e+00
-1.36512503e-01 -3.50423425e-01 -5.66389680e-01 7.82964051e-01
-1.55454290e+00 -8.71228516e-01 -8.37174281e-02 2.20273733e+00
9.64044511e-01 -3.14647496e-01 -2.40611702e-01 1.02509208e-01
7.91223526e-01 2.87567109e-01 -6.87198520e-01 1.79900199e-01
-2.69528888e-02 3.27079505e-01 6.27402782e-01 9.71020639e-01
-7.87563026e-01 6.02077305e-01 7.03813171e+00 8.42237353e-01
-8.36706638e-01 7.42558181e-01 -4.29229736e-02 1.80291384e-01
-2.27809057e-01 2.69611388e-01 -3.44695926e-01 3.11388165e-01
7.04388082e-01 2.54172869e-02 1.03715396e+00 2.25337654e-01
3.68994772e-01 -7.01977834e-02 -9.30543780e-01 9.44709837e-01
2.23651320e-01 -7.74709821e-01 -4.08825845e-01 1.76278546e-01
9.09793079e-01 1.95976898e-01 1.96151853e-01 -2.41037026e-01
6.12324119e-01 -7.64284790e-01 5.06961763e-01 8.12599897e-01
5.01603305e-01 -6.30061254e-02 5.41497469e-01 5.91747880e-01
-4.56708759e-01 -8.57905112e-03 -3.69101435e-01 1.19336598e-01
4.55156416e-01 1.20697260e+00 -6.71348214e-01 3.09453249e-01
6.06254041e-01 8.61340642e-01 1.25279829e-01 9.87981915e-01
-6.79724753e-01 1.10972404e+00 -4.33171153e-01 6.26056850e-01
-1.43133670e-01 -5.43228328e-01 9.75866616e-01 1.01734293e+00
6.69408023e-01 1.24956906e-01 -5.09209186e-02 8.66470873e-01
2.25464776e-02 -2.41100714e-01 -9.29662809e-02 8.46466608e-03
4.60924655e-02 1.06799269e+00 -3.89111280e-01 -1.92257732e-01
-2.05527872e-01 1.38321602e+00 -2.33433500e-01 9.36257482e-01
-5.74085414e-01 3.50203440e-02 4.30795610e-01 7.84006640e-02
6.58072531e-01 -4.87062007e-01 1.38318539e-03 -1.33640206e+00
-2.22921818e-01 -1.03990972e+00 -4.51804325e-02 -8.37901413e-01
-1.62725794e+00 3.52053076e-01 -1.08209267e-01 -9.73059714e-01
-2.86547691e-01 -3.14956009e-01 -4.07950878e-01 1.31472254e+00
-1.45765316e+00 -8.58958364e-01 -6.88311756e-02 7.02528000e-01
4.67837214e-01 5.43315224e-02 7.11376727e-01 1.75353244e-01
-3.64198416e-01 -1.63907483e-01 6.13984346e-01 -2.43813887e-01
8.32101047e-01 -1.27071536e+00 2.71776825e-01 1.06390429e+00
2.21822903e-01 6.94819510e-01 1.20244825e+00 -6.76809371e-01
-1.48938227e+00 -7.18725681e-01 4.50233579e-01 1.39742270e-02
7.21435964e-01 -7.11433366e-02 -9.47688639e-01 6.99939847e-01
4.80814964e-01 3.71855088e-02 5.28523088e-01 1.46921035e-02
-1.67475432e-01 2.73414105e-02 -1.12719905e+00 2.44042173e-01
8.75390053e-01 -6.93034351e-01 -6.47139549e-01 6.06427550e-01
1.94098830e-01 -4.33918208e-01 -7.27548838e-01 -5.57175167e-02
1.44896299e-01 -7.29689121e-01 1.22166407e+00 -1.96191873e-02
-1.81635562e-02 -4.15354431e-01 -2.39199474e-01 -1.79109991e+00
-4.21916366e-01 -9.59927022e-01 -3.63766640e-01 1.00329113e+00
1.20115794e-01 -6.83235884e-01 2.58053750e-01 2.34538645e-01
-1.14992805e-01 1.00098155e-01 -1.01880240e+00 -7.61530995e-01
-1.48747951e-01 -4.12700981e-01 1.10300377e-01 7.41669834e-01
-4.20856744e-01 4.83019888e-01 -8.94539952e-01 6.23711646e-01
1.11300766e+00 -3.62915024e-02 4.82532293e-01 -1.06187141e+00
-9.22146916e-01 -3.60684916e-02 9.25366655e-02 -1.92329168e+00
1.10209674e-01 -7.85168767e-01 6.12676382e-01 -1.74277413e+00
-1.22941345e-01 -2.42892578e-01 6.94865733e-02 1.12790763e-01
-1.11327901e-01 2.16486052e-01 -8.92892703e-02 5.89074075e-01
-4.14775051e-02 6.55060947e-01 1.37288368e+00 -3.18147838e-01
-1.55764893e-01 4.17192280e-01 -6.56414092e-01 8.06089222e-01
4.75717992e-01 -8.13713491e-01 -5.00651300e-01 -7.43977249e-01
-4.92583588e-02 1.61543623e-01 4.82694417e-01 -8.42485607e-01
4.84057546e-01 8.91684294e-02 5.44082448e-02 -1.78210691e-01
6.01943016e-01 -7.85312235e-01 3.81697327e-01 2.38504872e-01
-1.50134787e-01 -4.34388965e-01 -2.15758979e-02 1.06966579e+00
-1.63521484e-01 -6.19819224e-01 8.91916871e-01 1.30598843e-02
-3.93111259e-01 6.73526675e-02 -5.42944968e-01 3.77108790e-02
4.03261513e-01 1.19370550e-01 3.10865398e-02 -8.66979003e-01
-1.21586049e+00 -4.60073143e-01 3.01432520e-01 3.01579442e-02
6.34480417e-01 -9.54580843e-01 -8.25455844e-01 3.33077401e-01
-5.75134039e-01 -1.04213998e-01 3.02422762e-01 1.06732881e+00
-1.84049100e-01 4.58650663e-03 4.13859725e-01 -5.91403663e-01
-8.78449917e-01 4.85466331e-01 4.54373896e-01 -5.32534085e-02
-7.81083763e-01 8.57929409e-01 1.60121724e-01 -2.63568103e-01
9.86924991e-02 -1.02746539e-01 2.50529706e-01 -1.63395092e-01
5.24604857e-01 5.82195401e-01 -9.58548300e-03 -6.93440676e-01
3.15462202e-02 7.68231869e-01 1.45926565e-01 -6.70217156e-01
1.31269693e+00 -6.42759383e-01 -4.94558603e-01 2.51808733e-01
1.09693360e+00 3.15356374e-01 -1.45821118e+00 -6.29707754e-01
-1.33698568e-01 -4.94625539e-01 7.59027243e-01 -7.51433671e-01
-1.15121961e+00 9.67932343e-01 6.90111756e-01 2.50621527e-01
1.29442096e+00 -1.00379162e-01 6.32530451e-01 1.06995083e-01
9.76850912e-02 -8.02447975e-01 2.18141377e-01 3.26966435e-01
1.11140704e+00 -9.08802569e-01 2.48483658e-01 -3.04757237e-01
-2.33216211e-01 8.65627646e-01 -2.58907706e-01 -1.90276653e-01
1.01428533e+00 1.64718032e-01 8.76349807e-02 -8.04953801e-04
-7.00484440e-02 -1.73659056e-01 4.28868055e-01 7.56268620e-01
2.03804716e-01 -9.97324064e-02 -2.76751876e-01 1.92792550e-01
1.06934711e-01 8.72300789e-02 4.28518891e-01 7.34291375e-01
-4.17083591e-01 -1.19613123e+00 -8.34273398e-01 1.11882418e-01
-4.18419570e-01 -3.98690045e-01 -4.95372787e-02 1.80281594e-01
-1.12813808e-01 1.29531181e+00 -2.77419209e-01 4.17887233e-02
4.38708849e-02 -6.37905598e-02 5.17293453e-01 -5.52034140e-01
8.45389217e-02 6.96221769e-01 -1.58821732e-01 -2.25938633e-01
-7.66379833e-01 -8.61293614e-01 -7.74327099e-01 -2.34802783e-01
-5.70550799e-01 3.07432324e-01 8.59569490e-01 1.17824662e+00
3.32725376e-01 3.94770324e-01 4.86063838e-01 -1.02433181e+00
-1.03478682e+00 -1.37030113e+00 -9.04719532e-01 1.76047117e-01
7.16538787e-01 -4.68604475e-01 -7.76077092e-01 4.14564759e-01] | [11.740579605102539, -2.429654359817505] |
6181ff3c-f15e-49a6-b51a-222c86286f82 | end-to-end-multimodal-emotion-recognition | 1704.08619 | null | http://arxiv.org/abs/1704.08619v1 | http://arxiv.org/pdf/1704.08619v1.pdf | End-to-End Multimodal Emotion Recognition using Deep Neural Networks | Automatic affect recognition is a challenging task due to the various
modalities emotions can be expressed with. Applications can be found in many
domains including multimedia retrieval and human computer interaction. In
recent years, deep neural networks have been used with great success in
determining emotional states. Inspired by this success, we propose an emotion
recognition system using auditory and visual modalities. To capture the
emotional content for various styles of speaking, robust features need to be
extracted. To this purpose, we utilize a Convolutional Neural Network (CNN) to
extract features from the speech, while for the visual modality a deep residual
network (ResNet) of 50 layers. In addition to the importance of feature
extraction, a machine learning algorithm needs also to be insensitive to
outliers while being able to model the context. To tackle this problem, Long
Short-Term Memory (LSTM) networks are utilized. The system is then trained in
an end-to-end fashion where - by also taking advantage of the correlations of
the each of the streams - we manage to significantly outperform the traditional
approaches based on auditory and visual handcrafted features for the prediction
of spontaneous and natural emotions on the RECOLA database of the AVEC 2016
research challenge on emotion recognition. | ['Björn Schuller', 'Panagiotis Tzirakis', 'George Trigeorgis', 'Stefanos Zafeiriou', 'Mihalis A. Nicolaou'] | 2017-04-27 | null | null | null | null | ['multimodal-emotion-recognition', 'multimodal-emotion-recognition'] | ['computer-vision', 'speech'] | [ 1.17666729e-01 -1.72212541e-01 2.96388239e-01 -5.44989645e-01
-3.04047167e-01 -1.97872713e-01 5.89335740e-01 1.85793012e-01
-6.45455420e-01 5.31090915e-01 2.78059363e-01 3.33274215e-01
6.09739982e-02 -4.96647447e-01 -3.55315328e-01 -4.63513166e-01
-1.44649729e-01 1.24370512e-02 -2.24244773e-01 -3.19739848e-01
1.27425268e-01 6.48058534e-01 -1.91448462e+00 4.62073952e-01
3.31980646e-01 1.58623612e+00 -9.74007770e-02 8.52584183e-01
-2.79760361e-01 1.20417619e+00 -4.87997174e-01 -2.67434567e-01
-1.36322066e-01 -3.05041194e-01 -6.91061318e-01 -5.87050628e-04
-1.44686803e-01 -9.78109390e-02 -2.23769888e-01 6.95606470e-01
7.53366709e-01 3.81310016e-01 4.61923331e-01 -9.54147637e-01
5.38676642e-02 2.57380426e-01 -2.47084081e-01 2.00445861e-01
5.50462544e-01 -1.03731267e-01 9.03718054e-01 -1.13433802e+00
4.90674973e-01 9.32850122e-01 4.26263332e-01 4.48248982e-01
-9.34976816e-01 -4.18668717e-01 4.37884666e-02 6.33634448e-01
-1.06725538e+00 -6.88886583e-01 9.48910654e-01 -3.22534829e-01
1.19101453e+00 2.39044249e-01 7.08589733e-01 1.54418445e+00
-3.13460715e-02 8.14172983e-01 1.01278603e+00 -4.72069860e-01
3.25700492e-01 4.66685742e-01 -9.62502137e-02 3.81488234e-01
-7.14757085e-01 -6.73084846e-03 -7.82844782e-01 2.73907874e-02
1.41906992e-01 -1.56508572e-02 -2.47648001e-01 -1.70283183e-01
-8.92899513e-01 7.40841687e-01 5.30448139e-01 7.12723076e-01
-8.93254101e-01 5.23227863e-02 9.97928381e-01 4.94400084e-01
5.61792314e-01 3.84658009e-01 -5.45580924e-01 -5.67959964e-01
-9.30758059e-01 -1.90856650e-01 9.48500335e-01 1.24392115e-01
5.63569009e-01 2.77737558e-01 2.64678318e-02 1.15928650e+00
1.51309177e-01 7.15881735e-02 8.38522673e-01 -3.70777100e-01
-2.81439722e-02 6.05044425e-01 -1.78889289e-01 -1.11342978e+00
-5.88768542e-01 -3.37962538e-01 -1.12079573e+00 2.16950715e-01
1.57601181e-02 -1.53316721e-01 -8.05695355e-01 1.61701357e+00
2.74500072e-01 2.23974898e-01 2.89943099e-01 8.73874962e-01
7.80599058e-01 7.96358645e-01 2.22533837e-01 -3.02267671e-01
1.23843324e+00 -4.49907929e-01 -7.48646677e-01 -3.31503958e-01
2.70169765e-01 -8.07453275e-01 7.79775739e-01 7.64814496e-01
-8.78320456e-01 -5.25709867e-01 -1.05852020e+00 1.11293487e-01
-8.28788161e-01 6.62145913e-02 3.34463656e-01 5.27328908e-01
-9.40896451e-01 4.72520053e-01 -6.28854692e-01 -4.50048715e-01
2.45850340e-01 3.65969121e-01 -5.43376029e-01 2.10293666e-01
-1.38475192e+00 1.22596586e+00 4.53943640e-01 4.92698491e-01
-5.83768249e-01 -1.99872345e-01 -8.18185508e-01 2.67758697e-01
1.85296595e-01 -4.07786578e-01 8.88781011e-01 -1.87775230e+00
-1.85995150e+00 6.18016541e-01 -1.47467420e-01 -6.09747767e-01
2.59679168e-01 -1.42725646e-01 -5.50042331e-01 5.12147605e-01
-5.15431225e-01 3.08819890e-01 1.13773620e+00 -9.78069663e-01
-3.72840047e-01 -3.55363548e-01 -3.59493256e-01 1.15132838e-01
-7.72116959e-01 4.42573160e-01 -2.25421235e-01 -5.06652236e-01
-2.07373455e-01 -8.91345263e-01 -1.40415028e-01 -4.59585011e-01
-2.74153262e-01 -9.97295678e-02 8.86704504e-01 -6.84415698e-01
9.33333576e-01 -2.39326978e+00 3.40170860e-01 5.25180519e-01
2.66975034e-02 4.27285701e-01 -1.59107402e-01 3.67884010e-01
-3.74910742e-01 -2.81339288e-01 8.45973287e-03 -5.65734088e-01
8.54055583e-02 1.14225902e-01 -2.62051702e-01 3.20757449e-01
3.79739761e-01 6.46144331e-01 -4.69205737e-01 -2.82192260e-01
5.39503336e-01 9.10890579e-01 -2.20210940e-01 6.04149997e-01
-1.47313058e-01 3.72209370e-01 -2.83375353e-01 1.58134043e-01
3.22412103e-01 1.76623821e-01 -1.20851643e-01 -1.32857993e-01
-7.09237605e-02 5.83330318e-02 -1.09954214e+00 1.53393352e+00
-8.87106299e-01 7.33246744e-01 2.33487889e-01 -1.24532056e+00
1.11320496e+00 8.44031036e-01 5.78017652e-01 -8.66344273e-01
5.23471475e-01 1.65306285e-01 -1.84792325e-01 -6.33520126e-01
4.02430415e-01 -4.40244734e-01 -4.95704971e-02 2.36346051e-01
3.57377440e-01 -7.77945966e-02 -4.56292406e-02 2.19162889e-02
9.14286017e-01 -2.79126197e-01 3.41882825e-01 3.92561853e-01
7.25780189e-01 -5.25495350e-01 3.07712406e-01 2.01216489e-01
-2.01341465e-01 5.28745294e-01 3.75116855e-01 -6.33225739e-01
-6.48445547e-01 -4.99246418e-01 1.12201005e-01 1.25356030e+00
-4.45291251e-01 -2.96756357e-01 -4.81905937e-01 -3.89900595e-01
-4.15798247e-01 6.67013943e-01 -8.39645982e-01 -4.52216566e-01
-2.56261021e-01 -3.67324889e-01 4.33942527e-01 4.48240101e-01
2.25525767e-01 -1.67842126e+00 -9.73839819e-01 5.62482953e-01
-1.21635556e-01 -1.31392717e+00 2.56245613e-01 6.25689209e-01
-4.30394024e-01 -6.95676446e-01 -4.69516456e-01 -3.80145699e-01
-7.79990479e-02 -3.41139466e-01 9.50399697e-01 -5.36544807e-02
-4.66058940e-01 7.56538153e-01 -6.01341248e-01 -6.01543665e-01
-3.45549524e-01 1.83124952e-02 -2.46108547e-02 7.36163020e-01
4.01199162e-01 -7.32941031e-01 -3.49645287e-01 -1.05826646e-01
-1.16015446e+00 -2.74351954e-01 6.01546824e-01 8.15366745e-01
3.05831224e-01 -1.44434005e-01 8.02305043e-01 -3.64183515e-01
6.20482147e-01 -4.87257957e-01 -1.05805077e-01 -1.26635181e-02
-1.33500621e-01 -8.63260180e-02 9.21506524e-01 -4.60802853e-01
-1.11759102e+00 3.27275217e-01 -6.33807600e-01 -5.24520695e-01
-6.23136282e-01 8.35940063e-01 -9.33511779e-02 8.08195919e-02
4.17594224e-01 1.65879652e-01 -7.90468678e-02 -1.80582941e-01
3.15654308e-01 8.30039263e-01 3.61096382e-01 -1.14686623e-01
1.65618733e-01 2.07427159e-01 -9.36044082e-02 -1.26890051e+00
-6.87493920e-01 -5.48515916e-01 -5.24696589e-01 -6.43661916e-01
9.90054309e-01 -7.87443161e-01 -6.69377625e-01 4.63207930e-01
-1.06401944e+00 1.50529779e-02 -2.35369757e-01 5.88019490e-01
-5.83362937e-01 9.30697545e-02 -7.00005472e-01 -1.04112363e+00
-5.77502489e-01 -8.77947211e-01 6.86916530e-01 3.15463722e-01
-4.26908344e-01 -8.64137471e-01 1.38497770e-01 5.33229932e-02
7.43098855e-01 4.15911496e-01 6.67186081e-01 -8.07209134e-01
1.90694645e-01 -5.27251840e-01 -7.12230057e-03 6.63254559e-01
-7.88513273e-02 1.01688176e-01 -1.34047472e+00 -9.83880460e-03
2.19755262e-01 -7.65738487e-01 1.06553292e+00 2.18833506e-01
1.08319199e+00 -1.04471162e-01 2.52630174e-01 2.80643344e-01
1.18731749e+00 9.69628170e-02 7.16860831e-01 3.39111239e-01
4.71065164e-01 7.94869661e-01 4.45471168e-01 9.01943982e-01
2.62781233e-01 5.67032576e-01 7.56305933e-01 -2.65178949e-01
3.29183638e-01 2.04529792e-01 4.58541989e-01 7.00413108e-01
-1.40601555e-02 -2.89752316e-02 -7.44342029e-01 4.60630357e-01
-1.69211030e+00 -1.08510280e+00 9.86050740e-02 2.08229542e+00
6.43415570e-01 7.45671764e-02 3.21858935e-02 4.77341771e-01
2.89267659e-01 2.96430290e-01 -3.58596921e-01 -9.95338321e-01
-1.28572673e-01 4.36761171e-01 -1.93320900e-01 1.50170535e-01
-1.10912192e+00 8.39665592e-01 4.64624119e+00 5.83225012e-01
-1.71353304e+00 -9.05590728e-02 6.19778633e-01 -2.75880069e-01
1.70941770e-01 -4.97849911e-01 -9.26183388e-02 2.51992822e-01
1.41311216e+00 2.46074855e-01 3.85652423e-01 7.09743381e-01
2.84255922e-01 -1.65469453e-01 -8.97543311e-01 1.21518302e+00
2.95128912e-01 -6.33129656e-01 -3.28784168e-01 -2.63968915e-01
2.07152307e-01 9.51204672e-02 1.88183799e-01 4.20337856e-01
-2.87736446e-01 -1.24925888e+00 7.02865958e-01 8.97879660e-01
4.81431782e-01 -1.17620265e+00 8.57655942e-01 3.28563988e-01
-1.02533829e+00 -2.49746338e-01 -2.43815899e-01 -1.46422356e-01
9.33023617e-02 7.55884290e-01 -8.72751772e-01 4.11275923e-01
8.07684600e-01 5.72787583e-01 -3.52113038e-01 7.94527292e-01
-9.18111503e-02 5.98804533e-01 -3.56718600e-01 -1.93000987e-01
2.39758208e-01 8.89995769e-02 4.20577019e-01 1.48974538e+00
3.61941695e-01 -2.39696503e-02 -1.27107784e-01 3.94084364e-01
-6.91267326e-02 4.63217944e-01 -6.07817113e-01 -1.99656829e-01
-7.88499042e-02 1.59397459e+00 -4.72150177e-01 -2.89394200e-01
-3.09457481e-01 1.33951402e+00 4.38363701e-01 2.49899164e-01
-6.58805728e-01 -3.53785574e-01 6.37345314e-01 -4.37366098e-01
4.83792305e-01 -1.82551607e-01 4.59254533e-02 -1.05970383e+00
-1.39785483e-01 -9.68180954e-01 1.73878223e-01 -7.68070340e-01
-1.14326799e+00 1.09751964e+00 -5.77889383e-01 -1.01328397e+00
-6.34128869e-01 -5.45909524e-01 -6.00121915e-01 7.96274841e-01
-1.57867157e+00 -1.03284264e+00 -1.93691701e-01 9.38266218e-01
4.92999256e-01 -3.22460085e-02 1.19255841e+00 2.38088682e-01
-6.19791687e-01 1.88440353e-01 -1.81817219e-01 4.65954654e-02
9.34357345e-01 -1.21472335e+00 -2.34333575e-01 6.07078552e-01
3.45665395e-01 1.60891816e-01 7.29655087e-01 -5.28780222e-02
-1.30758071e+00 -8.97897780e-01 1.04914081e+00 7.87165612e-02
6.37120545e-01 -3.33968341e-01 -9.36434925e-01 3.61826122e-01
5.73632956e-01 2.25864395e-01 9.01086271e-01 1.57368600e-01
-3.92112911e-01 -2.31144264e-01 -9.41884875e-01 2.88473248e-01
1.49120271e-01 -7.79097378e-01 -4.82651711e-01 -2.12914750e-01
1.98115736e-01 -1.10889331e-01 -7.30951846e-01 3.85239482e-01
6.63720310e-01 -1.24663866e+00 8.70132983e-01 -5.37220657e-01
5.97461045e-01 9.54941064e-02 -8.80189389e-02 -1.64892781e+00
3.95198986e-02 -1.97332919e-01 -3.47555289e-03 1.30677724e+00
2.55873829e-01 -4.06836689e-01 3.82222950e-01 6.88883543e-01
8.71572942e-02 -6.49979532e-01 -1.05993736e+00 -9.78905037e-02
-3.87928963e-01 -8.29874814e-01 1.39262393e-01 8.37557495e-01
2.45556891e-01 6.77493691e-01 -8.39236557e-01 -8.21296424e-02
-2.95701660e-02 1.04030140e-01 6.32697046e-01 -1.27625275e+00
-1.86073944e-01 -5.09623528e-01 -5.84407568e-01 -4.14053559e-01
4.01536435e-01 -5.89798331e-01 1.68890029e-01 -1.26665962e+00
-6.48195520e-02 1.29810438e-01 -7.10619390e-01 4.27387863e-01
-5.69873080e-02 3.07747185e-01 3.57881874e-01 -4.11700785e-01
-7.34111726e-01 9.26985085e-01 6.08798623e-01 -3.29523422e-02
-2.33028889e-01 2.03702040e-02 -2.82517046e-01 8.33459616e-01
8.78032327e-01 -2.75472313e-01 6.43827096e-02 3.63430753e-02
4.55476135e-01 1.86661437e-01 3.32814664e-01 -1.10923970e+00
2.64258146e-01 3.18990022e-01 7.24334538e-01 -4.43772435e-01
8.35021257e-01 -1.21766710e+00 -1.53667420e-01 1.30829617e-01
-4.14073616e-01 -8.26258212e-02 3.51037413e-01 3.77899349e-01
-7.25718200e-01 -1.84554845e-01 7.99342871e-01 -2.95355842e-02
-7.87387192e-01 1.73080429e-01 -8.49398255e-01 -3.86178076e-01
8.15929055e-01 1.02857381e-01 2.53992110e-01 -7.43180215e-01
-9.92832959e-01 -2.51568742e-02 -4.20853458e-02 5.28091073e-01
8.83679152e-01 -1.14357448e+00 -4.96529549e-01 2.73700029e-01
1.65581614e-01 -5.99686563e-01 5.07802784e-01 9.91183937e-01
-6.96587656e-03 8.23909231e-03 -3.22168022e-01 -3.26674283e-01
-1.45567393e+00 5.32478273e-01 3.58627379e-01 -4.08220232e-01
-1.50883675e-01 7.82649159e-01 -2.98536867e-01 -2.49555677e-01
4.28453058e-01 -4.74090539e-02 -7.46659815e-01 6.68566525e-01
6.94145083e-01 6.49302602e-02 3.60156834e-01 -8.60627174e-01
-3.70503098e-01 2.62595028e-01 1.18590929e-01 -2.57766336e-01
1.72106934e+00 -1.58560440e-01 -1.84681937e-01 7.34603882e-01
1.48120773e+00 1.60998777e-02 -8.62424076e-01 -2.03985050e-01
1.38938457e-01 -3.74768191e-04 3.28845799e-01 -8.70834112e-01
-1.10661161e+00 1.23924637e+00 7.94227242e-01 5.98852396e-01
1.54544878e+00 -3.93809736e-01 5.97033083e-01 2.95347899e-01
-3.59941088e-02 -1.32892048e+00 2.46057987e-01 8.36832047e-01
1.01817989e+00 -1.30668259e+00 -4.81994927e-01 2.37350434e-01
-9.11892474e-01 1.58782148e+00 1.10034756e-01 -3.63525711e-02
6.58438444e-01 2.83127904e-01 3.18095416e-01 -1.64239600e-01
-8.92591119e-01 -5.30543685e-01 3.37346882e-01 1.99801639e-01
5.18942118e-01 -7.81697482e-02 7.73260836e-03 8.09712231e-01
-1.25361040e-01 4.79646400e-02 2.65344679e-01 6.30986392e-01
-3.11895907e-01 -1.01103604e+00 -3.19571406e-01 3.38751167e-01
-7.18891799e-01 -6.05029538e-02 -6.45717084e-01 3.77507031e-01
-3.23894955e-02 1.21797669e+00 -1.07802898e-01 -5.04628003e-01
5.29420972e-01 4.52685058e-01 9.98396873e-02 -2.71814823e-01
-1.03621674e+00 2.10441813e-01 1.71691850e-01 -7.98685193e-01
-6.24148667e-01 -4.79395837e-01 -1.09274113e+00 1.65591508e-01
-1.45029739e-01 5.34442887e-02 9.96672928e-01 1.09357703e+00
2.29471192e-01 7.12615192e-01 9.57062125e-01 -1.13089812e+00
-3.19231242e-01 -1.13123846e+00 -6.12880170e-01 5.39397657e-01
4.34470564e-01 -2.87652284e-01 -4.22310501e-01 -6.43936545e-02] | [13.395405769348145, 5.2911601066589355] |
475ec35f-3a39-483b-92a7-6e26aeb29951 | collaborative-representation-for | 1403.1353 | null | http://arxiv.org/abs/1403.1353v1 | http://arxiv.org/pdf/1403.1353v1.pdf | Collaborative Representation for Classification, Sparse or Non-sparse? | Sparse representation based classification (SRC) has been proved to be a
simple, effective and robust solution to face recognition. As it gets popular,
doubts on the necessity of enforcing sparsity starts coming up, and primary
experimental results showed that simply changing the $l_1$-norm based
regularization to the computationally much more efficient $l_2$-norm based
non-sparse version would lead to a similar or even better performance. However,
that's not always the case. Given a new classification task, it's still unclear
which regularization strategy (i.e., making the coefficients sparse or
non-sparse) is a better choice without trying both for comparison. In this
paper, we present as far as we know the first study on solving this issue,
based on plenty of diverse classification experiments. We propose a scoring
function for pre-selecting the regularization strategy using only the dataset
size, the feature dimensionality and a discrimination score derived from a
given feature representation. Moreover, we show that when dictionary learning
is taking into account, non-sparse representation has a more significant
superiority to sparse representation. This work is expected to enrich our
understanding of sparse/non-sparse collaborative representation for
classification and motivate further research activities. | ['Michihiko Minoh', 'Yang Wu', 'Vansteenberge Jarich', 'Masayuki Mukunoki'] | 2014-03-06 | null | null | null | null | ['sparse-representation-based-classification'] | ['computer-vision'] | [ 2.72032231e-01 -2.35289142e-01 -2.51983464e-01 -5.47785282e-01
-6.01725817e-01 -1.09635159e-01 4.37605768e-01 4.68589220e-04
-1.82607621e-01 7.55058348e-01 2.34655797e-01 -1.94638029e-01
-4.32711065e-01 -6.35519385e-01 -3.84046942e-01 -8.85964513e-01
6.22963086e-02 2.48780221e-01 -1.03022501e-01 -2.74542809e-01
5.14686763e-01 6.45635068e-01 -1.96561003e+00 1.39067709e-01
7.58121610e-01 1.01228821e+00 2.28665844e-01 -4.21068259e-02
-1.99636221e-01 7.78770208e-01 -5.22074103e-01 -3.41543168e-01
3.69120330e-01 -6.53963268e-01 -6.02289796e-01 1.34986401e-01
5.41300654e-01 2.15063348e-01 5.14376983e-02 1.10047328e+00
4.96158183e-01 1.33900777e-01 5.27943313e-01 -9.88603711e-01
-6.19993627e-01 4.24097151e-01 -5.40995061e-01 4.17409152e-01
4.89774644e-01 -1.45844817e-01 7.73249865e-01 -9.89818633e-01
6.37465656e-01 1.02563047e+00 6.81033850e-01 2.29453251e-01
-1.06629491e+00 -7.44242430e-01 2.65526861e-01 1.74789369e-01
-1.66202927e+00 -6.30075097e-01 1.01350582e+00 -3.83009136e-01
6.92654550e-01 4.09550935e-01 4.83193994e-01 9.57297683e-01
-4.33781818e-02 3.30772817e-01 1.45053363e+00 -6.62163556e-01
2.74479270e-01 4.36059535e-01 1.96216762e-01 6.51894033e-01
5.87405026e-01 8.84850416e-03 -5.81547081e-01 -2.19990835e-01
5.17510235e-01 -7.99693912e-03 -4.36836749e-01 -4.73514646e-01
-7.11299062e-01 1.08299148e+00 1.28199041e-01 8.97969544e-01
-4.29474980e-01 -2.08617792e-01 3.06056738e-01 5.13997495e-01
3.07678640e-01 3.50654453e-01 -1.87888786e-01 -1.08027279e-01
-1.20150030e+00 -1.07582239e-02 6.95734739e-01 5.01808286e-01
6.44423723e-01 4.89707917e-01 7.42598549e-02 9.75945175e-01
2.48978332e-01 2.60142058e-01 6.49698436e-01 -7.62046218e-01
1.45142943e-01 5.57143033e-01 2.02967897e-02 -1.45859337e+00
-2.82471180e-01 -5.32037318e-01 -8.89820337e-01 1.58071473e-01
4.72988009e-01 3.78342308e-02 -4.72402066e-01 1.67922342e+00
1.16439208e-01 3.82132590e-01 -1.75804973e-01 1.28452575e+00
7.78747916e-01 3.94732863e-01 -1.24533717e-02 -6.43722355e-01
1.33189142e+00 -3.63359958e-01 -7.40834653e-01 -1.78913064e-02
5.46689510e-01 -1.04484272e+00 7.65415192e-01 5.50161541e-01
-7.85971880e-01 -5.55246413e-01 -9.79524493e-01 3.25061053e-01
-3.36535573e-01 2.14373767e-01 1.12366116e+00 1.00320935e+00
-8.48396242e-01 3.68139416e-01 -6.78731024e-01 -6.55269980e-01
1.26578569e-01 4.13242877e-01 -7.06567228e-01 -3.32813382e-01
-9.92538393e-01 8.99570286e-01 3.94974612e-02 1.22011386e-01
-4.30708319e-01 -4.90022779e-01 -7.76053548e-01 -6.82177395e-02
3.34787279e-01 -3.63207728e-01 4.90362763e-01 -1.18675840e+00
-1.32006371e+00 8.67509305e-01 -3.12269449e-01 -3.73348534e-01
-1.48061765e-02 1.41276598e-01 -5.18708646e-01 -7.27137998e-02
-3.75946984e-02 1.18833080e-01 1.09351695e+00 -1.25536573e+00
-2.53166169e-01 -3.63803416e-01 3.35786417e-02 6.62150420e-03
-4.30471808e-01 2.60419637e-01 -1.38356432e-01 -9.46013987e-01
4.50856090e-01 -7.97186792e-01 -2.13602513e-01 -1.17739066e-01
-3.32153663e-02 -1.94597140e-01 7.56171405e-01 -4.29426521e-01
1.32468224e+00 -2.32978439e+00 1.37863860e-01 5.20942926e-01
-1.17380761e-01 4.11976367e-01 -2.07998097e-01 4.82719511e-01
-4.80158180e-01 -3.64547186e-02 -3.19262832e-01 -1.24471962e-01
-3.23779523e-01 1.61836028e-01 -1.98469326e-01 7.98426986e-01
1.36605591e-01 2.24149197e-01 -6.81294143e-01 -3.96144569e-01
2.84764290e-01 7.96847701e-01 -6.44974649e-01 -7.92844370e-02
3.76239151e-01 4.94749933e-01 -5.29508531e-01 7.76890755e-01
7.43084550e-01 -1.92105874e-01 3.01851988e-01 -2.82414138e-01
-1.54083058e-01 -1.27711045e-02 -1.62328660e+00 1.47688472e+00
-4.44900274e-01 4.76556242e-01 2.45193824e-01 -1.70023656e+00
1.30072773e+00 4.00256097e-01 6.19919538e-01 -7.71074355e-01
2.68969744e-01 3.32589597e-01 1.37388811e-01 -4.12582606e-01
2.62828052e-01 -4.68890756e-01 1.58457771e-01 2.68516362e-01
9.81156677e-02 1.48142919e-01 3.40551049e-01 -1.28816530e-01
9.74117041e-01 -1.69511393e-01 5.17466605e-01 -5.00321567e-01
8.82153153e-01 -1.70007363e-01 6.39139712e-01 5.48227072e-01
-4.54675294e-02 6.53615415e-01 3.14613521e-01 -4.16360259e-01
-4.76862222e-01 -5.66285789e-01 -4.04394299e-01 8.83214414e-01
-1.31536514e-01 -4.93831843e-01 -4.28445786e-01 -5.00541747e-01
1.28952563e-01 4.75191444e-01 -5.22817552e-01 -7.82025978e-02
-5.41160524e-01 -8.17332327e-01 1.33639514e-01 1.35831863e-01
3.28515172e-01 -8.39227676e-01 -2.68058538e-01 2.74237171e-02
9.90931243e-02 -8.08739781e-01 -2.01277733e-01 2.02041224e-01
-1.09994388e+00 -1.08880436e+00 -8.01606059e-01 -7.93361962e-01
9.35137928e-01 5.68105698e-01 8.70613813e-01 3.06928277e-01
-2.02349186e-01 3.17673624e-01 -6.95702732e-01 -1.29402801e-01
2.61888723e-03 -1.63600400e-01 1.67838797e-01 3.24711978e-01
4.78678972e-01 -7.02627182e-01 -5.67904651e-01 2.16962367e-01
-8.39506567e-01 -5.24458706e-01 5.69803178e-01 7.42356539e-01
4.82988894e-01 1.98477820e-01 6.00566864e-01 -1.01718831e+00
6.84367836e-01 -5.35336792e-01 -3.11294824e-01 7.40261823e-02
-6.13101900e-01 -9.92856771e-02 7.69376040e-01 -2.98616290e-01
-8.19246113e-01 2.07038730e-01 -3.09137285e-01 -5.74644089e-01
-7.33274892e-02 7.07281113e-01 4.13774997e-02 -4.89325643e-01
6.77446842e-01 3.67117703e-01 9.61327031e-02 -5.33946514e-01
1.43453613e-01 6.35369718e-01 -3.35174464e-02 -5.16837001e-01
8.24709594e-01 6.04263127e-01 -2.40195747e-02 -1.13791740e+00
-7.77717531e-01 -6.90274119e-01 -3.73340666e-01 3.96262519e-02
4.71180260e-01 -8.49500597e-01 -4.50340748e-01 -1.71054341e-02
-7.23488927e-01 3.76742750e-01 -2.89116502e-01 6.66341126e-01
-3.19439888e-01 7.04931319e-01 -6.58474192e-02 -9.24266160e-01
-6.99913800e-02 -1.09846354e+00 6.42097056e-01 1.02611268e-02
-3.06783140e-01 -6.94758713e-01 2.54631788e-02 5.16297936e-01
6.99377060e-01 1.11381106e-01 7.61704087e-01 -5.80362320e-01
-2.22481251e-01 -3.22888017e-01 -1.64820194e-01 4.86902386e-01
3.27480793e-01 -1.98173448e-01 -7.89648890e-01 -5.47444522e-01
4.62037504e-01 -1.95611790e-01 7.47704208e-01 3.36653203e-01
1.14591193e+00 -3.80919009e-01 -1.18382163e-02 5.57809234e-01
1.68680406e+00 2.66267538e-01 9.31064725e-01 1.76596150e-01
3.84932667e-01 5.67289412e-01 5.96514761e-01 6.57732010e-01
7.61987409e-03 8.58720064e-01 1.74739853e-01 -5.56245781e-02
-2.99934894e-01 1.19464859e-01 3.01580161e-01 8.82546544e-01
-3.17053944e-01 6.56994060e-02 -5.48584700e-01 3.15559208e-01
-1.48011374e+00 -1.20943201e+00 -1.34147003e-01 2.47546434e+00
6.87687814e-01 -9.19064581e-02 1.73182741e-01 6.02905810e-01
6.12835646e-01 1.29813820e-01 6.11859141e-03 -4.50918674e-01
-2.14224428e-01 6.39984488e-01 1.25224978e-01 3.91484171e-01
-9.23131168e-01 6.68710589e-01 5.92633820e+00 9.95696187e-01
-1.56286919e+00 1.47059381e-01 4.81486142e-01 9.51140001e-02
-2.00079799e-01 1.25215143e-01 -7.63204932e-01 6.70502305e-01
7.09445775e-01 -8.62686429e-03 3.56929481e-01 8.47617507e-01
2.23706961e-01 -1.96827903e-01 -8.17221284e-01 1.31566083e+00
4.69342023e-01 -1.19194925e+00 1.99685261e-01 -5.06946407e-02
5.55001736e-01 -2.96051830e-01 6.39151335e-02 2.36660972e-01
-2.86845803e-01 -1.16674554e+00 4.90597188e-01 5.68792284e-01
4.63284463e-01 -5.98219573e-01 8.55996132e-01 2.54739523e-01
-1.14971256e+00 -9.78084207e-02 -4.05365735e-01 -1.35305226e-01
-1.54151157e-01 8.27338398e-01 -3.15605193e-01 5.76281369e-01
4.65444922e-01 6.93870068e-01 -5.09342968e-01 1.06970799e+00
1.10438593e-01 7.42319226e-01 -2.20635623e-01 1.61917329e-01
2.01947447e-02 -3.60915452e-01 5.27233422e-01 1.24971986e+00
6.62523031e-01 3.05960178e-01 1.06660806e-01 4.26576912e-01
3.12056541e-01 5.91211498e-01 -9.28626537e-01 -7.77995959e-02
3.73078883e-01 1.19870818e+00 -9.03527439e-01 -3.16305645e-02
-7.25553393e-01 7.82948911e-01 1.92336500e-01 1.96955115e-01
-5.32471418e-01 -2.30944380e-01 5.69193006e-01 3.45896155e-01
5.14994562e-01 -2.88777411e-01 -3.73593926e-01 -1.18907809e+00
-4.52929549e-02 -1.06289589e+00 4.86572057e-01 -4.17298049e-01
-1.42684996e+00 6.33226037e-01 -1.04036950e-01 -1.43507981e+00
7.12231034e-03 -4.71901417e-01 -4.12017077e-01 8.61453056e-01
-1.70953560e+00 -7.23280966e-01 -1.49752632e-01 6.55368686e-01
4.49546218e-01 -4.03901488e-01 9.41875279e-01 6.01023376e-01
-5.92398047e-01 5.84407449e-01 -1.35039672e-01 -1.34743050e-01
7.05391169e-01 -6.47128880e-01 -7.52972722e-01 7.10338712e-01
3.36721778e-01 1.02292287e+00 8.64931285e-01 -3.14134091e-01
-1.72234559e+00 -6.30947411e-01 1.01235807e+00 -2.35196277e-01
4.13214862e-01 -1.88288428e-02 -1.06588197e+00 1.98013708e-01
-4.70307693e-02 9.06433612e-02 9.82662678e-01 2.85608560e-01
-3.63660455e-01 -4.06444579e-01 -1.28316534e+00 2.19704792e-01
9.25832093e-01 -4.70282197e-01 -5.94757378e-01 3.20268005e-01
-3.00135766e-03 -8.30633491e-02 -9.51931238e-01 5.26081324e-01
3.30223203e-01 -1.19154179e+00 1.09947395e+00 -2.87342042e-01
2.13540524e-01 -4.17721659e-01 -5.44886947e-01 -8.12386870e-01
-4.35661942e-01 -2.68010557e-01 1.92157343e-01 1.16112745e+00
2.01986998e-01 -6.67188585e-01 1.00938499e+00 2.52163619e-01
3.28859240e-02 -8.95922720e-01 -1.29269695e+00 -9.22763467e-01
-1.49459347e-01 -3.43519598e-01 2.63475895e-01 1.32992291e+00
-4.90914136e-02 1.70669422e-01 -4.80531901e-01 9.84464288e-02
4.59354073e-01 4.60578710e-01 5.89680433e-01 -1.35355651e+00
-1.38876051e-01 -4.37554061e-01 -6.89563513e-01 -6.05445385e-01
1.77795291e-01 -1.00618696e+00 -3.76092762e-01 -1.20845056e+00
1.49128675e-01 -7.27149487e-01 -4.86874491e-01 4.80810881e-01
7.00730532e-02 5.20146608e-01 1.49954960e-01 3.32252324e-01
-3.51484984e-01 4.10479873e-01 1.18614352e+00 -6.98459819e-02
-1.63962021e-01 -5.54002114e-02 -1.07857001e+00 5.80361009e-01
7.69221961e-01 -4.50687796e-01 -4.87771451e-01 -7.96004757e-03
8.54148436e-03 7.59504214e-02 3.68818119e-02 -1.02054799e+00
1.67894542e-01 -2.95734465e-01 2.25037783e-01 -2.52189934e-02
5.82800686e-01 -9.82297719e-01 3.42955440e-01 4.91906822e-01
-6.56621754e-02 -1.61870971e-01 1.17825277e-01 4.60179776e-01
-4.36824173e-01 -5.76136470e-01 8.31598938e-01 -2.33219326e-01
-7.82435298e-01 -8.96607619e-03 -2.47390598e-01 -2.18314022e-01
9.22517121e-01 -6.03162229e-01 2.29478002e-01 -2.92312175e-01
-7.03791678e-01 -4.33820844e-01 3.22798342e-01 2.88102180e-01
6.60738170e-01 -1.19309771e+00 -6.80939555e-01 4.02169496e-01
2.96407361e-02 -7.38800764e-01 2.65334606e-01 9.05022085e-01
-2.39012122e-01 5.98595798e-01 -2.03837112e-01 -4.10896391e-01
-1.40108216e+00 5.20727098e-01 7.33585795e-03 -4.14136350e-02
-5.26136041e-01 8.35229754e-01 -2.06247047e-01 -1.43517613e-01
2.66881824e-01 -8.72460306e-02 -5.21906078e-01 2.31382266e-01
3.72654140e-01 2.91397512e-01 3.73150736e-01 -9.26477134e-01
-5.25227666e-01 8.02605987e-01 -1.16947470e-02 3.13875377e-01
1.37217367e+00 1.25525594e-01 -3.33487332e-01 2.65307397e-01
1.05305207e+00 2.73759782e-01 -7.23822236e-01 -2.19074432e-02
-1.04806498e-01 -9.04324532e-01 1.50767103e-01 -4.45678413e-01
-1.42059886e+00 4.48944539e-01 6.79760218e-01 3.66758823e-01
1.26401258e+00 -1.18031494e-01 4.43353295e-01 3.37863475e-01
8.03620994e-01 -9.11956549e-01 -9.19608921e-02 3.06525946e-01
1.23273993e+00 -1.35813832e+00 4.11527038e-01 -7.13488877e-01
-4.81758207e-01 9.91327643e-01 4.32968557e-01 -3.70088190e-01
1.01868320e+00 1.10781088e-01 -8.70935619e-02 -2.52780795e-01
-4.22001302e-01 -2.42131695e-01 1.61661953e-01 6.09977782e-01
7.77435243e-01 -1.39980823e-01 -8.61981809e-01 4.96873498e-01
-2.45659024e-01 1.79755494e-01 1.42207712e-01 9.07841980e-01
-4.75647599e-01 -1.49559450e+00 -6.36804521e-01 7.41853952e-01
-4.09968108e-01 4.68052626e-02 -1.20035835e-01 6.79095984e-01
3.47814977e-01 9.91987169e-01 -2.65131205e-01 -2.50672787e-01
3.63944143e-01 -1.65387411e-02 5.99865615e-01 -8.26660752e-01
-5.43052018e-01 -2.29955420e-01 6.20116107e-03 -5.70139885e-01
-7.48003602e-01 -6.80138409e-01 -8.86936665e-01 -3.78140628e-01
-3.73264611e-01 4.34086710e-01 7.39917099e-01 7.74773538e-01
3.13005984e-01 7.45154768e-02 7.49666750e-01 -6.35368645e-01
-5.94030201e-01 -7.01510787e-01 -7.78201580e-01 4.42588955e-01
-7.70776533e-03 -8.94407511e-01 -6.18175685e-01 -5.39759807e-02] | [12.425637245178223, 0.4367225468158722] |
6249b5b9-2ad9-4d7b-8698-dff6da419df1 | second-language-acquisition-of-neural | 2306.02920 | null | https://arxiv.org/abs/2306.02920v1 | https://arxiv.org/pdf/2306.02920v1.pdf | Second Language Acquisition of Neural Language Models | With the success of neural language models (LMs), their language acquisition has gained much attention. This work sheds light on the second language (L2) acquisition of LMs, while previous work has typically explored their first language (L1) acquisition. Specifically, we trained bilingual LMs with a scenario similar to human L2 acquisition and analyzed their cross-lingual transfer from linguistic perspectives. Our exploratory experiments demonstrated that the L1 pretraining accelerated their linguistic generalization in L2, and language transfer configurations (e.g., the L1 choice, and presence of parallel texts) substantially affected their generalizations. These clarify their (non-)human-like L2 acquisition in particular aspects. | ['Taro Watanabe', 'Hiroki Ouchi', 'Tatsuki Kuribayashi', 'Miyu Oba'] | 2023-06-05 | null | null | null | null | ['language-acquisition', 'cross-lingual-transfer'] | ['natural-language-processing', 'natural-language-processing'] | [-2.41238073e-01 1.95726901e-01 -5.95090330e-01 -3.20572883e-01
-4.80272055e-01 -8.41151655e-01 8.63495708e-01 3.24611604e-01
-6.28485143e-01 3.82598728e-01 3.55593860e-01 -1.05209064e+00
2.62566824e-02 -7.28479564e-01 -8.98501337e-01 -2.90969580e-01
1.71970725e-02 3.81603539e-01 -3.41875665e-02 -5.63494563e-01
1.53299510e-01 4.69858378e-01 -1.19385135e+00 3.48260760e-01
1.33937013e+00 6.67360350e-02 5.31647623e-01 1.11213967e-01
-4.88921463e-01 5.69256902e-01 -2.71852970e-01 -5.51460683e-01
2.62729734e-01 -7.14596152e-01 -8.83242965e-01 -2.56824046e-01
6.41205251e-01 -2.98477888e-01 -1.46259546e-01 1.02673924e+00
1.99769527e-01 1.05977664e-02 7.58676469e-01 -7.89577603e-01
-1.26872647e+00 1.39670193e+00 -4.70562428e-01 5.18880665e-01
3.01219583e-01 3.06956410e-01 9.88135040e-01 -1.10984313e+00
7.19188988e-01 1.36231768e+00 6.15882277e-01 5.97501159e-01
-1.39949465e+00 -6.17460310e-01 3.77490163e-01 3.79825830e-02
-1.38409138e+00 -5.13976812e-01 4.91111368e-01 -6.88931406e-01
9.12019014e-01 -6.04855180e-01 9.67557967e-01 1.10792530e+00
1.88891422e-02 1.03248286e+00 1.83952987e+00 -9.77448404e-01
-4.26392943e-01 7.02942669e-01 2.27155864e-01 3.86218488e-01
4.00371850e-01 4.90586072e-01 -8.01979661e-01 7.58815050e-01
9.47127521e-01 -6.64354622e-01 -1.78461686e-01 1.03478342e-01
-1.34220791e+00 6.96432531e-01 3.32342148e-01 1.16337526e+00
-1.08040236e-01 -3.56375337e-01 4.18594748e-01 8.63431990e-01
4.78948019e-02 5.65173447e-01 -5.80348909e-01 -4.64571007e-02
-7.90270746e-01 -3.16389054e-01 4.24914420e-01 1.04702783e+00
7.78609574e-01 2.25503251e-01 -5.59107140e-02 1.01256442e+00
1.24424629e-01 6.37676001e-01 8.44898701e-01 -1.18458793e-01
6.74684882e-01 4.70073342e-01 -4.60848480e-01 -6.03043437e-01
-3.64596993e-01 -5.71138620e-01 -4.21978086e-01 -2.17592850e-01
8.36378276e-01 -1.59930989e-01 -4.38788831e-01 2.34714961e+00
-1.30334973e-01 -5.08144379e-01 2.28140160e-01 5.99823534e-01
4.53420341e-01 5.64124525e-01 5.95506549e-01 -3.56399268e-01
9.29014802e-01 -8.37226450e-01 -4.67623770e-01 -2.84511745e-01
1.14908564e+00 -7.58760571e-01 1.71110260e+00 2.56647587e-01
-1.48351383e+00 -9.47374046e-01 -8.62604856e-01 -1.87343568e-01
-5.32767951e-01 7.33240247e-02 7.88305640e-01 9.78139877e-01
-1.47742927e+00 4.21741575e-01 -7.08256066e-01 -7.63169825e-01
1.13458239e-01 1.75942391e-01 -4.10508901e-01 -6.76759034e-02
-1.58381104e+00 1.35468614e+00 6.28702044e-01 -9.83221680e-02
-9.37051535e-01 -6.88291192e-01 -9.08388615e-01 -1.86677128e-01
9.74349454e-02 -4.09630924e-01 1.22957039e+00 -1.37262380e+00
-1.54273450e+00 1.54551232e+00 2.26122700e-03 -1.00280710e-01
1.44671008e-01 -2.36939177e-01 -6.33110523e-01 -2.20854841e-02
1.53290346e-01 7.33961701e-01 3.64082545e-01 -1.17900479e+00
-6.87909722e-01 -4.26507950e-01 1.33030683e-01 3.70623291e-01
-4.89375889e-01 2.44393796e-01 1.87853798e-01 -8.82363260e-01
-5.77331297e-02 -8.19932461e-01 2.08397925e-01 -4.73489761e-01
2.47202232e-01 -4.69225198e-01 -2.05574527e-01 -7.73269892e-01
1.43020260e+00 -2.12887692e+00 1.28234565e-01 2.87926376e-01
-3.13641131e-02 2.96970934e-01 -3.36907268e-01 9.29174244e-01
-8.33259746e-02 5.97602129e-01 1.33059487e-01 -4.52852175e-02
1.26945466e-01 -7.04974756e-02 -5.13473228e-02 3.78592372e-01
4.17666323e-02 1.27744246e+00 -8.81098151e-01 -5.20925522e-01
-8.65719467e-02 2.83781171e-01 -7.06550121e-01 1.53909564e-01
-7.50765800e-02 8.93025696e-01 -2.37315223e-02 4.38199371e-01
2.69298136e-01 2.26856824e-02 4.50429469e-01 1.71327576e-01
-7.44808793e-01 5.96439958e-01 -4.49898243e-01 1.86934042e+00
-9.26535070e-01 7.42634654e-01 6.66039959e-02 -8.76130998e-01
6.30203187e-01 3.18509966e-01 1.07101891e-02 -9.25069392e-01
1.92349777e-01 5.55096805e-01 9.97836292e-01 -4.27153230e-01
1.80372790e-01 -5.30324876e-01 1.39470592e-01 6.40692770e-01
1.66002020e-01 -2.50105739e-01 2.34840766e-01 1.11438423e-01
2.10155487e-01 2.55767196e-01 6.76991880e-01 -1.05256784e+00
6.94234133e-01 -8.05813372e-02 2.39444762e-01 7.64525294e-01
-3.40516046e-02 -1.36192396e-01 9.39544812e-02 9.26434845e-02
-8.40873301e-01 -1.23246944e+00 -4.26476598e-01 1.53657246e+00
-1.26182623e-02 -2.26862758e-01 -9.19508457e-01 -4.31616306e-01
-2.55628228e-01 1.55915892e+00 -4.16407228e-01 -2.69138485e-01
-1.15747499e+00 -2.71288723e-01 6.20450914e-01 4.84380186e-01
3.23220998e-01 -1.34653747e+00 -2.97473133e-01 2.37750202e-01
3.15786116e-02 -9.10963535e-01 -5.02802491e-01 7.05788983e-03
-9.02371407e-01 -7.33721316e-01 -5.35734653e-01 -1.18748510e+00
6.75429761e-01 6.47330061e-02 1.27006841e+00 3.49588692e-01
3.52511019e-01 6.12723827e-01 -3.17178249e-01 -4.54551548e-01
-8.96712542e-01 5.00740409e-01 6.35141194e-01 -4.60053563e-01
6.99369013e-01 -6.15744531e-01 -9.81201231e-02 1.11475535e-01
-9.06584203e-01 2.91668903e-02 8.28942835e-01 6.55772328e-01
9.42643061e-02 -2.75613844e-01 7.69415319e-01 -8.60319614e-01
8.14948738e-01 -5.58695078e-01 -5.84069610e-01 5.42109072e-01
-6.20060503e-01 1.09320488e-02 7.15832353e-01 -6.97753906e-01
-1.33894861e+00 -6.34075582e-01 -2.27078483e-01 2.23920748e-01
-3.27353716e-01 8.94601643e-01 -1.32306769e-01 2.98359990e-02
6.38238013e-01 4.76496220e-01 -1.93943396e-01 -5.71766853e-01
3.95263731e-01 4.57192332e-01 2.79354513e-01 -1.28603005e+00
8.56294990e-01 -3.55587482e-01 -5.20234048e-01 -1.24550879e+00
-8.19542468e-01 1.78313583e-01 -1.31073105e+00 -7.06385225e-02
8.75743032e-01 -1.13413584e+00 -2.15293407e-01 5.57221293e-01
-1.04757977e+00 -8.87587845e-01 -1.88779309e-01 9.82366681e-01
-3.34639788e-01 -3.77263688e-02 -6.65583432e-01 -5.48074722e-01
1.86093107e-01 -1.24800432e+00 3.70947331e-01 3.79440226e-02
-5.26113749e-01 -1.43408191e+00 1.88632250e-01 -1.15640247e-02
6.15607798e-01 -4.98760909e-01 1.57864809e+00 -8.79931211e-01
-4.87497479e-01 5.78193963e-01 8.25802982e-02 2.32282817e-01
2.33184308e-01 -1.54899776e-01 -6.16137862e-01 -4.29967672e-01
9.76268128e-02 -5.36868155e-01 3.63384157e-01 2.26733327e-01
3.97521436e-01 -2.51117021e-01 -9.06318650e-02 6.28106654e-01
1.26055610e+00 2.58449405e-01 1.36377618e-01 3.36754441e-01
4.42946643e-01 8.55301023e-01 3.27840477e-01 -3.64761293e-01
4.10735428e-01 5.71433544e-01 -4.41400707e-01 6.72670954e-04
-4.15859580e-01 -6.07234955e-01 9.85586524e-01 1.68922734e+00
6.12961547e-03 -1.59815736e-02 -1.25110626e+00 9.22000289e-01
-1.05031347e+00 -5.00820696e-01 1.08437590e-01 2.25686526e+00
1.37246406e+00 9.78127569e-02 9.54341516e-02 -2.89584905e-01
6.83637500e-01 -6.35138154e-02 -3.17768753e-01 -4.75780457e-01
-4.90893036e-01 2.27312729e-01 9.14710239e-02 7.82058120e-01
-3.17324579e-01 1.66284549e+00 7.06855059e+00 6.64606750e-01
-1.36483741e+00 1.85347468e-01 4.59651172e-01 1.69913381e-01
-7.40586519e-01 -1.90432936e-01 -1.15843117e+00 2.43921623e-01
9.33181465e-01 -4.58232194e-01 5.85540771e-01 3.18975955e-01
9.48740914e-03 -1.99669483e-03 -1.51759803e+00 5.83918393e-01
-9.40728039e-02 -5.76710045e-01 3.53208542e-01 -6.18850477e-02
7.81378210e-01 1.61142424e-01 2.58617252e-01 8.42604041e-01
2.11746946e-01 -1.17885876e+00 9.98173356e-01 4.89431359e-02
1.16405594e+00 -3.93262923e-01 2.40401417e-01 6.66523039e-01
-8.74638379e-01 1.57984272e-02 -7.22093806e-02 -3.49526584e-01
4.09172148e-01 6.24379441e-02 -8.68373632e-01 4.34754677e-02
3.72303009e-01 4.23559785e-01 -9.84543622e-01 2.00358972e-01
-7.90819407e-01 1.13970661e+00 -1.96884498e-01 -1.33901253e-01
3.51284057e-01 -3.07574987e-01 4.83531862e-01 1.20453382e+00
3.19928527e-01 3.04335803e-01 1.45629242e-01 8.01002443e-01
1.41754806e-01 8.88028979e-01 -8.57385278e-01 -4.56182480e-01
5.52248061e-01 5.70349932e-01 -7.14593291e-01 -1.98389277e-01
-9.25927997e-01 5.46705306e-01 8.02475274e-01 3.11275184e-01
-3.40250701e-01 2.71715987e-02 3.84231716e-01 4.18821216e-01
-1.41566128e-01 -6.87635839e-01 -3.18178535e-01 -1.17626286e+00
-2.50728697e-01 -1.11946249e+00 8.44529644e-02 -1.83662578e-01
-1.35803378e+00 6.44921601e-01 2.21744433e-01 -6.17786407e-01
-2.13886321e-01 -8.32813621e-01 -3.58318746e-01 9.32171464e-01
-1.20791554e+00 -1.32593966e+00 3.04749876e-01 6.15346551e-01
5.17785907e-01 -5.45705438e-01 5.59819400e-01 4.05334473e-01
-4.88523334e-01 1.09154427e+00 -1.33253843e-01 1.98440954e-01
7.75878549e-01 -1.01484597e+00 4.08526629e-01 8.87264609e-01
4.40543860e-01 1.45532835e+00 3.01920146e-01 -6.95809305e-01
-1.11906159e+00 -5.75706542e-01 1.26653016e+00 -3.77714276e-01
9.81343925e-01 -4.06579047e-01 -8.31489205e-01 1.19516706e+00
7.49833107e-01 -9.57354069e-01 8.66816878e-01 3.29277247e-01
-8.17839950e-02 1.50939837e-01 -7.04833329e-01 1.14471614e+00
1.61220706e+00 -7.14612484e-01 -8.76833081e-01 1.36830732e-01
9.08654153e-01 9.05470103e-02 -1.02371621e+00 3.64301920e-01
3.40253413e-01 -1.06030881e+00 7.54410088e-01 -6.85505331e-01
4.54608887e-01 5.17392382e-02 -7.19082579e-02 -1.57713246e+00
-5.11432648e-01 -3.99748296e-01 3.74594659e-01 1.21294880e+00
7.26788163e-01 -7.35774517e-01 -1.50675112e-02 3.41702431e-01
-2.85570353e-01 -5.75241566e-01 -5.25885284e-01 -8.35770488e-01
9.69816267e-01 -1.95112184e-01 4.78173375e-01 1.33662724e+00
2.31322318e-01 5.43163121e-01 1.21386118e-01 9.09361616e-02
8.84685665e-02 1.20573854e-02 6.65562212e-01 -1.01425922e+00
-2.48324245e-01 -1.01991916e+00 1.22140005e-01 -1.21504831e+00
5.98177671e-01 -1.63813972e+00 -1.43930525e-01 -1.04720926e+00
-2.76186973e-01 -6.27075374e-01 -2.82788008e-01 3.39906991e-01
-4.13979948e-01 -3.06922525e-01 4.40118283e-01 5.66083044e-02
-1.21127054e-01 4.88267779e-01 1.23617375e+00 2.94710040e-01
-5.25213718e-01 -1.42840087e-01 -9.53002512e-01 6.98392630e-01
9.33329701e-01 -2.98491597e-01 -6.01259947e-01 -1.13197899e+00
3.54330122e-01 1.08283851e-02 -3.75816435e-01 -6.49906695e-01
1.04260184e-01 -2.59406060e-01 8.29701796e-02 -1.48218423e-01
-3.39843869e-01 -6.94612145e-01 -2.72001237e-01 4.80840981e-01
-4.77328807e-01 4.59331274e-01 5.73984861e-01 -4.19009984e-01
-2.37085268e-01 -5.18761933e-01 8.34497333e-01 -3.50602597e-01
-6.59761012e-01 1.93889886e-01 -6.21455193e-01 5.17657816e-01
8.66452515e-01 -1.23434365e-01 2.34996766e-01 -1.46933109e-01
-6.83741808e-01 1.67605385e-01 8.03360522e-01 5.76724768e-01
6.24400824e-02 -1.28206182e+00 -9.27872896e-01 3.72205138e-01
1.71692017e-02 -2.80060977e-01 -1.73958614e-01 1.03109682e+00
-3.26447845e-01 6.54093444e-01 -4.85367268e-01 -4.74182665e-01
-5.78712523e-01 7.59071350e-01 1.06029622e-01 -2.07494944e-01
-2.70791024e-01 1.00514555e+00 6.43925905e-01 -7.73316562e-01
-3.86239253e-02 -1.28138244e-01 -3.54890645e-01 6.63809031e-02
3.36892635e-01 3.49091172e-01 -4.12829936e-01 -8.79896700e-01
-1.26543686e-01 8.03279281e-01 -4.22305316e-01 -4.40356463e-01
1.03007197e+00 -3.29869419e-01 -3.97801220e-01 1.10578430e+00
9.61255968e-01 9.08089280e-01 -5.47905505e-01 -4.76579279e-01
3.14820081e-01 -4.00229394e-02 -3.86522710e-01 -8.04152250e-01
-8.04876447e-01 1.08116555e+00 -3.28081660e-02 -1.95283368e-01
8.13638687e-01 1.54114574e-01 5.05574167e-01 3.26293409e-01
5.63535035e-01 -1.12626779e+00 -3.11641425e-01 9.69760120e-01
8.48950982e-01 -1.31421256e+00 -2.55104542e-01 -2.76990563e-01
-6.05299413e-01 8.89245868e-01 1.06403172e+00 1.45956293e-01
7.35076725e-01 6.02082610e-02 3.87834042e-01 -9.70749035e-02
-3.98343235e-01 -8.37478116e-02 2.61715114e-01 4.66641575e-01
1.11658680e+00 7.71240368e-02 -8.65724266e-01 8.80625844e-01
-9.67917383e-01 -2.93558151e-01 1.46211579e-01 6.47075653e-01
-1.87538937e-01 -1.26226354e+00 -1.54074028e-01 4.92616221e-02
-2.83157676e-01 -7.02299714e-01 -3.57773840e-01 1.13976514e+00
3.92812520e-01 7.64208376e-01 -2.29982864e-02 -2.25748662e-02
3.02907646e-01 4.43702370e-01 9.92778361e-01 -1.01983988e+00
-8.78118873e-01 -2.16072857e-01 -2.27675691e-01 -1.59214288e-01
-4.27512020e-01 -9.02619481e-01 -1.11590922e+00 -1.77769631e-01
-2.21602604e-01 1.55149937e-01 4.13189292e-01 1.12401378e+00
-2.79625744e-01 1.68309718e-01 2.07991555e-01 -4.87822473e-01
-4.90164787e-01 -1.25840938e+00 -6.55823231e-01 2.16487482e-01
9.22563970e-02 -4.74349916e-01 -3.58497232e-01 4.48292214e-03] | [10.822381973266602, 9.930713653564453] |
c7bdbca5-2b3f-4a0b-a4f0-3d6f84cade0b | subgroup-fairness-in-graph-based-spam | 2204.11164 | null | https://arxiv.org/abs/2204.11164v2 | https://arxiv.org/pdf/2204.11164v2.pdf | Are Your Reviewers Being Treated Equally? Discovering Subgroup Structures to Improve Fairness in Spam Detection | User-generated reviews of products are vital assets of online commerce, such as Amazon and Yelp, while fake reviews are prevalent to mislead customers. GNN is the state-of-the-art method that detects suspicious reviewers by exploiting the topologies of the graph connecting reviewers, reviews, and target products. However, the discrepancy in the detection accuracy over different groups of reviewers can degrade reviewer engagement and customer trust in the review websites. Unlike the previous belief that the difference between the groups causes unfairness, we study the subgroup structures within the groups that can also cause discrepancies in treating different groups. This paper addresses the challenges of defining, approximating, and utilizing a new subgroup structure for fair spam detection. We first identify subgroup structures in the review graph that lead to discrepant accuracy in the groups. The complex dependencies over the review graph create difficulties in teasing out subgroups hidden within larger groups. We design a model that can be trained to jointly infer the hidden subgroup memberships and exploits the membership for calibrating the detection accuracy across groups. Comprehensive comparisons against baselines on three large Yelp review datasets demonstrate that the subgroup membership can be identified and exploited for group fairness. | ['Sihong Xie', 'Xi Zhang', 'Yuefei Lyu', 'Jiaxin Liu'] | 2022-04-24 | null | null | null | null | ['spam-detection'] | ['natural-language-processing'] | [-9.40315127e-02 3.68198842e-01 -5.65431058e-01 -8.50241363e-01
-3.23684335e-01 -8.86917233e-01 6.83299482e-01 1.96212515e-01
1.09094836e-01 2.59213001e-01 -1.72240455e-02 -5.43461204e-01
-5.21179587e-02 -7.55962491e-01 -3.49034458e-01 -1.79784387e-01
1.27217725e-01 4.02429581e-01 3.55734080e-01 -2.14504287e-01
6.16751730e-01 1.25929609e-01 -1.14660060e+00 5.05448520e-01
1.15251660e+00 8.51449907e-01 -7.06281364e-01 3.88558239e-01
-1.19281910e-01 6.23150408e-01 -8.13348174e-01 -1.18930030e+00
7.14248180e-01 -6.27593696e-01 -2.30135977e-01 3.00419390e-01
8.25329483e-01 -1.94345355e-01 8.27689320e-02 1.26266587e+00
5.63594885e-02 -3.18181545e-01 7.65310824e-01 -1.67949355e+00
-1.07156527e+00 9.54352319e-01 -1.08498299e+00 2.42068559e-01
-1.50150452e-02 7.10397810e-02 1.74207139e+00 -6.34664118e-01
5.90720415e-01 1.50638342e+00 7.41127431e-01 3.29284757e-01
-1.17888045e+00 -7.72553742e-01 5.31023383e-01 -1.58416927e-02
-8.92429650e-01 -3.17187339e-01 5.29517055e-01 -4.68249142e-01
2.70058721e-01 3.08037907e-01 3.14923048e-01 8.38375926e-01
2.49511719e-01 5.42428136e-01 1.15561163e+00 -2.80386093e-03
1.93281248e-01 6.53557897e-01 8.85322690e-01 5.31205595e-01
1.03065550e+00 -5.61441332e-02 -6.24569237e-01 -7.89077997e-01
2.79913813e-01 -1.12586953e-01 -7.40878433e-02 -5.68321884e-01
-5.88243186e-01 1.46703172e+00 4.36174452e-01 -1.83759823e-01
-2.49222323e-01 7.20181763e-02 5.75960696e-01 5.13762295e-01
8.18578124e-01 7.04359114e-01 -4.27060246e-01 3.94858897e-01
-8.95115733e-01 9.57393721e-02 1.30708623e+00 7.26071119e-01
7.42787182e-01 -2.14370355e-01 -1.10703290e-01 5.70049405e-01
5.00759780e-01 3.42883199e-01 5.01995534e-02 -8.86818051e-01
4.29556936e-01 9.34194326e-01 1.55920953e-01 -1.56362379e+00
-5.13670221e-02 -5.81532419e-01 -3.70886922e-01 3.00836600e-02
7.65661061e-01 -5.61884716e-02 -5.98225594e-01 1.43237650e+00
3.16907614e-01 -1.99291661e-01 -5.24722993e-01 1.04862702e+00
4.93340850e-01 -1.50026130e-02 4.88230698e-02 -6.31283096e-04
1.31349814e+00 -1.12806010e+00 -6.71178997e-01 -4.67778087e-01
6.47496581e-01 -8.53752315e-01 8.18533182e-01 3.56604159e-01
-8.21191967e-01 -1.66862637e-01 -1.03887653e+00 7.91920051e-02
-4.08918798e-01 -8.49865451e-02 1.03758788e+00 1.20638978e+00
-9.09094274e-01 7.65526652e-01 -6.67488649e-02 -3.05752128e-01
5.68677545e-01 4.19557273e-01 -8.04858133e-02 5.55808805e-02
-1.26358414e+00 1.06034243e+00 -6.45008862e-01 1.79582775e-01
-4.17316079e-01 -7.80018628e-01 -5.07132113e-01 1.14531070e-01
4.93995756e-01 -3.41521770e-01 1.12497652e+00 -1.39305449e+00
-8.22095275e-01 8.62344503e-01 -1.79000199e-02 -4.64426845e-01
6.02294803e-01 1.89135134e-01 -3.97074521e-01 -1.00146439e-02
3.27271283e-01 2.33161345e-01 1.05889869e+00 -1.39202666e+00
-5.92759132e-01 -6.36080563e-01 -1.39028370e-01 -7.57239163e-02
5.07574296e-03 2.32407108e-01 1.23579420e-01 -1.26137361e-01
1.21198282e-01 -8.78091156e-01 -2.60443181e-01 -3.56625691e-02
-5.09038985e-01 -2.37528831e-01 7.80339062e-01 -5.81848145e-01
1.27633214e+00 -1.86550200e+00 -5.82184494e-01 8.88615489e-01
8.43598962e-01 2.94178635e-01 -2.27921113e-01 8.29409957e-02
8.74006525e-02 7.02900171e-01 2.39624411e-01 -2.30441049e-01
2.18000650e-01 1.62189126e-01 -3.83221507e-01 8.25232089e-01
5.47544025e-02 1.04334331e+00 -9.09867287e-01 9.37409047e-03
-3.01764131e-01 8.46003592e-02 -4.13314700e-01 -8.55050534e-02
1.34076029e-01 -1.57625571e-01 -3.16184074e-01 6.82842255e-01
1.06755078e+00 -4.21638578e-01 4.19918180e-01 -1.62555292e-01
2.69313365e-01 7.94456542e-01 -8.56146812e-01 2.90509194e-01
2.08527595e-01 4.53856796e-01 6.28620088e-01 -6.49277568e-01
1.31160843e+00 -3.09070915e-01 1.25295654e-01 -2.76559830e-01
2.88261950e-01 3.50848615e-01 3.85836184e-01 5.32220909e-03
8.10579956e-01 -3.08318045e-02 -9.61843878e-02 1.16119540e+00
-3.56642514e-01 -1.40517429e-02 2.58033276e-01 6.92968190e-01
1.08557439e+00 -8.10872674e-01 2.11792439e-01 -2.48156786e-01
2.96506137e-01 -7.07187429e-02 5.31529963e-01 1.14978242e+00
-8.23824584e-01 4.62090224e-01 1.29979515e+00 -4.29783344e-01
-9.64803159e-01 -1.00606990e+00 1.69180900e-01 1.13895726e+00
2.58446783e-01 -3.41100186e-01 -5.24361849e-01 -1.38622189e+00
6.83083177e-01 7.12789655e-01 -7.86494493e-01 -4.83878106e-01
-2.33863667e-01 -9.71128702e-01 4.51237589e-01 8.37370306e-02
1.48731828e-01 -5.43118596e-01 -1.48002210e-03 -9.26698223e-02
4.69050836e-03 -1.01710403e+00 -9.30695117e-01 -7.51070604e-02
-5.82453847e-01 -1.62665057e+00 1.05768181e-01 -3.57009232e-01
9.83865976e-01 7.49220312e-01 1.58805633e+00 3.91561091e-01
-6.81878254e-02 3.70270818e-01 -2.93414444e-01 -4.79878604e-01
-6.51511908e-01 1.10095479e-01 -3.47046182e-02 1.89258203e-01
1.04646111e+00 -1.81301385e-01 -5.71388483e-01 9.33651745e-01
-5.15173078e-01 -5.98608255e-01 3.45030069e-01 5.76132655e-01
-1.72761694e-01 -1.61638960e-01 9.37548161e-01 -1.68937016e+00
1.14876127e+00 -7.21220613e-01 -4.75897372e-01 3.26976418e-01
-1.18768334e+00 -2.06944779e-01 2.40551874e-01 -4.61034268e-01
-8.30417395e-01 -3.67572755e-01 7.02431142e-01 7.03954995e-02
4.75164205e-01 1.38172999e-01 5.06015494e-02 -2.51552045e-01
7.89952219e-01 -7.03016222e-01 5.31197846e-01 -1.91939980e-01
5.74880421e-01 6.95921838e-01 -5.54600060e-02 -1.33921534e-01
7.30098784e-01 4.23454523e-01 -3.46224993e-01 -3.91614959e-02
-1.10191607e+00 -6.77496433e-01 -4.35940474e-01 -9.82652307e-02
3.44643056e-01 -6.59799159e-01 -7.61278331e-01 1.20984554e-01
-1.17877769e+00 2.45789990e-01 -1.65827215e-01 1.78779047e-02
1.82654724e-01 5.88334382e-01 -9.69437361e-01 -8.61370683e-01
-5.01176655e-01 -9.38519716e-01 6.40614450e-01 1.15615658e-01
-5.98790109e-01 -1.00945020e+00 -1.00043319e-01 5.97311378e-01
5.39906085e-01 -2.11371660e-01 7.09694147e-01 -1.33905709e+00
-6.68710768e-01 -5.48259735e-01 -6.61265612e-01 5.65469325e-01
1.37060151e-01 4.22556728e-01 -7.23886371e-01 -9.11344141e-02
5.62191568e-02 3.66714299e-02 9.11431313e-01 3.15035999e-01
5.12687981e-01 -4.83352780e-01 -2.24229351e-01 -5.51376492e-02
7.84634233e-01 -2.85419285e-01 4.82165635e-01 3.73505019e-02
6.12022936e-01 1.08835042e+00 5.88072360e-01 3.15616578e-01
4.91166830e-01 2.76903957e-01 4.95353222e-01 -2.05295626e-02
3.06336939e-01 -2.92813659e-01 5.56551695e-01 6.79674685e-01
5.30813456e-01 -6.07932433e-02 -4.67027396e-01 3.50501835e-01
-1.92625046e+00 -8.26239109e-01 -6.97317779e-01 2.24945045e+00
5.50196648e-01 2.41771564e-01 2.24948287e-01 -2.33128548e-01
1.38453877e+00 2.78577566e-01 -4.44950223e-01 -8.74737203e-01
-5.01947850e-02 -2.57719219e-01 9.99027431e-01 7.48119473e-01
-7.30667293e-01 8.31598818e-01 7.05795956e+00 4.50337142e-01
-6.21556938e-01 6.78373054e-02 1.01175344e+00 -9.50935334e-02
-6.58047438e-01 2.95543194e-01 -1.08717442e+00 4.25673723e-01
9.33346391e-01 -3.81375849e-01 3.23877215e-01 8.57997835e-01
4.41619754e-01 -1.30000398e-01 -1.04433250e+00 3.78973544e-01
3.67094219e-01 -9.69004810e-01 5.06086349e-02 3.22072417e-01
9.60245073e-01 -1.67233571e-01 2.00938329e-01 3.02011341e-01
8.75860035e-01 -7.34425128e-01 3.40082020e-01 1.11573458e-01
2.24685103e-01 -5.71329057e-01 8.51309180e-01 -1.17438577e-01
-6.32008135e-01 -1.69892490e-01 -6.98211789e-01 -5.96543588e-02
1.96346551e-01 1.34518003e+00 -1.01172018e+00 5.26744165e-02
3.15795600e-01 8.50273311e-01 -1.02640235e+00 7.80737638e-01
-4.89519387e-01 6.37476265e-01 -7.29191862e-03 -1.12798326e-01
-4.11920026e-02 -6.32744670e-01 4.21814084e-01 8.02067637e-01
-7.31024370e-02 -3.68896604e-01 -1.03646833e-02 1.30084515e+00
-4.30953979e-01 -4.73999865e-02 -5.91162980e-01 -3.41752529e-01
4.09627020e-01 1.69058895e+00 -7.67514646e-01 -2.53397226e-01
-4.32170302e-01 5.97316206e-01 3.19135904e-01 3.37467641e-01
-8.42422247e-01 -8.52812976e-02 8.12572777e-01 5.77713013e-01
2.22670481e-01 7.87365511e-02 -7.65638471e-01 -1.15959823e+00
7.44765333e-04 -1.15329313e+00 4.15883780e-01 -4.28989261e-01
-2.10368466e+00 1.79037705e-01 -5.40253341e-01 -8.29267561e-01
6.65506348e-02 -5.66179574e-01 -8.33815694e-01 8.79579782e-01
-1.59274995e+00 -9.12659645e-01 -1.23864837e-01 3.25371400e-02
7.41818771e-02 -6.66710585e-02 2.93459952e-01 1.88176408e-01
-4.53902811e-01 8.23506355e-01 -3.36435378e-01 2.74392925e-02
1.21107912e+00 -1.29255891e+00 7.42537618e-01 6.97834551e-01
-3.96559052e-02 9.56413448e-01 6.77172363e-01 -1.02932584e+00
-8.06090593e-01 -8.82596552e-01 1.21889400e+00 -9.22666669e-01
1.20095611e+00 -6.67502105e-01 -9.92452204e-01 6.09759629e-01
4.08703880e-03 7.65974373e-02 8.50971699e-01 5.00161350e-01
-1.13752449e+00 -1.15852728e-01 -1.32540154e+00 4.24985915e-01
8.82880092e-01 -6.71082675e-01 -2.55428463e-01 2.12701634e-01
5.79607606e-01 1.70068219e-01 -4.43964839e-01 1.38871431e-01
5.40278494e-01 -1.32690811e+00 6.00234270e-01 -7.39343584e-01
5.06754994e-01 -1.41730875e-01 3.82866800e-01 -1.39875293e+00
-4.30809289e-01 -6.42119408e-01 4.73260991e-02 1.42793930e+00
7.47285187e-01 -9.18227792e-01 1.05019403e+00 1.32717514e+00
4.20042694e-01 -5.74164152e-01 -4.77854788e-01 -4.68990773e-01
-1.44601427e-02 1.30021766e-01 4.55255508e-01 9.58588660e-01
7.00806528e-02 5.42871356e-01 -3.69624376e-01 1.07006021e-02
7.60182559e-01 5.06563000e-02 8.05376291e-01 -1.57212842e+00
-1.22936651e-01 -4.85915810e-01 -2.57654935e-01 -6.77029610e-01
2.31943160e-01 -8.29702854e-01 -1.66442588e-01 -1.22711718e+00
2.74385482e-01 -7.17318892e-01 -1.88379791e-02 2.28696272e-01
-3.91553283e-01 1.00998461e-01 2.85150766e-01 3.46671820e-01
-9.46467042e-01 1.25867233e-01 1.22022331e+00 -7.53224045e-02
-8.73259529e-02 2.87563950e-01 -1.63739836e+00 5.38616419e-01
8.65521669e-01 -5.93792379e-01 -3.00740361e-01 4.58975770e-02
6.90364301e-01 -5.98403335e-01 2.42496952e-01 4.58169496e-03
2.37863764e-01 1.18343219e-01 1.45613819e-01 -4.20170516e-01
-2.67031610e-01 -6.08264208e-01 -2.44825050e-01 2.38311440e-01
-6.84447885e-01 2.40470797e-01 -4.71362978e-01 8.92704308e-01
-1.88150927e-02 -2.72462666e-01 7.48197556e-01 -3.81133765e-01
-1.17622444e-03 8.46587270e-02 -3.13596845e-01 2.48223767e-01
8.02672446e-01 -1.06481649e-02 -1.17690706e+00 -7.55117178e-01
-3.45438242e-01 4.42529798e-01 5.68666637e-01 5.88289917e-01
2.15321377e-01 -9.85137939e-01 -6.49292469e-01 2.10589603e-01
3.13290730e-02 -7.62931347e-01 -1.82893813e-01 9.27186430e-01
-4.32724692e-02 7.28668645e-03 -2.08842810e-02 -2.98319340e-01
-1.36592352e+00 5.26682734e-01 3.64323616e-01 -5.82464218e-01
2.75104225e-01 7.06677973e-01 1.16312556e-01 -6.95844948e-01
-6.35723770e-02 2.77304947e-02 -1.32703841e-01 4.49571788e-01
5.39397061e-01 6.41508102e-01 -4.62881066e-02 -5.67358971e-01
-2.16051206e-01 9.22828093e-02 -6.53888047e-01 2.56589323e-01
9.48876500e-01 -4.65373695e-01 -6.32972300e-01 1.32126197e-01
1.07111692e+00 3.50867808e-01 -8.61569047e-01 -9.38787535e-02
2.42865950e-01 -9.34874356e-01 -2.59197503e-01 -7.57583439e-01
-1.21121156e+00 5.92894971e-01 -1.01783335e-01 7.76114941e-01
2.52814889e-01 -1.50642172e-01 5.65961897e-01 -6.98111877e-02
3.35224241e-01 -1.07489634e+00 -1.58673134e-02 2.79447168e-01
4.32031482e-01 -1.49134648e+00 4.44412291e-01 -1.12217772e+00
-1.03275025e+00 7.96646953e-01 8.02360415e-01 -3.37945729e-01
7.09704816e-01 -8.99273604e-02 3.71145964e-01 -6.29415751e-01
-8.43854487e-01 2.57323325e-01 3.01116526e-01 5.21569133e-01
3.90045732e-01 2.12682739e-01 -8.44464123e-01 8.51096213e-01
-7.83643350e-02 -5.63076854e-01 1.08710361e+00 3.94793868e-01
-4.76516336e-01 -1.35366428e+00 -2.67164618e-01 1.13920009e+00
-5.98662734e-01 -1.94660008e-01 -1.14329827e+00 5.80064297e-01
-3.64030674e-02 1.31977272e+00 2.37601716e-02 -4.39592332e-01
3.01761806e-01 -2.40934953e-01 -1.41697273e-01 -9.15060163e-01
-1.31527460e+00 -2.28022546e-01 5.18453658e-01 -5.68511009e-01
-2.83097655e-01 -6.23566985e-01 -7.18009114e-01 -7.80522466e-01
-8.59590530e-01 5.76576710e-01 4.91439998e-01 5.38632035e-01
5.91811061e-01 -7.62129053e-02 8.91285658e-01 -3.68381947e-01
-1.15047050e+00 -5.64074397e-01 -1.00643647e+00 7.91412413e-01
9.48702171e-02 -6.38579726e-01 -1.32842159e+00 -5.08722365e-01] | [7.905447006225586, 10.048912048339844] |
76991b19-f2ee-4e09-bd1e-7f9de154264e | variable-viewpoint-representations-for-3d | 2002.03131 | null | https://arxiv.org/abs/2002.03131v1 | https://arxiv.org/pdf/2002.03131v1.pdf | Variable-Viewpoint Representations for 3D Object Recognition | For the problem of 3D object recognition, researchers using deep learning methods have developed several very different input representations, including "multi-view" snapshots taken from discrete viewpoints around an object, as well as "spherical" representations consisting of a dense map of essentially ray-traced samples of the object from all directions. These representations offer trade-offs in terms of what object information is captured and to what degree of detail it is captured, but it is not clear how to measure these information trade-offs since the two types of representations are so different. We demonstrate that both types of representations in fact exist at two extremes of a common representational continuum, essentially choosing to prioritize either the number of views of an object or the pixels (i.e., field of view) allotted per view. We identify interesting intermediate representations that lie at points in between these two extremes, and we show, through systematic empirical experiments, how accuracy varies along this continuum as a function of input information as well as the particular deep learning architecture that is used. | ['Tengyu Ma', 'Maithilee Kunda', 'Joel Michelson', 'Deepayan Sanyal', 'Xiaohan Wang', 'James Ainooson'] | 2020-02-08 | null | null | null | null | ['3d-object-recognition'] | ['computer-vision'] | [ 1.28310889e-01 -4.13643979e-02 1.46602169e-01 -5.08567393e-01
-5.26492178e-01 -9.02112961e-01 1.03444695e+00 5.65692745e-02
-1.09963395e-01 2.69133002e-01 4.51596320e-01 -1.51830211e-01
-3.01591039e-01 -9.80368316e-01 -6.38574898e-01 -7.07796454e-01
9.44717303e-02 6.90557420e-01 1.43928200e-01 -7.62943625e-02
4.48737264e-01 9.74214315e-01 -1.59336948e+00 5.13258874e-01
1.68850124e-02 9.17452097e-01 4.11259383e-02 6.96708918e-01
-1.69387206e-01 5.77614307e-01 -4.78224516e-01 -2.97217488e-01
4.49396223e-01 -2.05749974e-01 -8.58317196e-01 5.59938908e-01
5.94886065e-01 -2.68559158e-01 -5.40987194e-01 9.91411150e-01
1.79936707e-01 2.51358328e-03 9.31534946e-01 -7.50579119e-01
-7.32485354e-01 9.68960300e-02 -7.76162803e-01 4.38326955e-01
4.73492891e-01 2.26785943e-01 1.16360152e+00 -8.58157873e-01
8.55219722e-01 1.26644230e+00 4.84608471e-01 2.22201422e-01
-1.47013950e+00 -2.98323878e-03 1.82416171e-01 -2.28367507e-01
-9.59895194e-01 -4.25289750e-01 9.75738406e-01 -8.80650997e-01
7.80703545e-01 1.66217014e-01 7.18431473e-01 8.42894375e-01
3.66440654e-01 6.20548904e-01 1.40031493e+00 -3.91232282e-01
3.16601753e-01 1.74868226e-01 2.60633856e-01 4.92387861e-01
4.34718102e-01 1.10620886e-01 -3.94679099e-01 -1.71323612e-01
1.23366416e+00 3.32737774e-01 -3.99662584e-01 -1.05533767e+00
-1.05916905e+00 7.64989853e-01 4.31108475e-01 3.11251253e-01
-2.96110094e-01 1.08936451e-01 2.67987609e-01 4.02674079e-01
4.48042065e-01 5.23965776e-01 -2.15956911e-01 1.18852876e-01
-6.39107406e-01 3.40561628e-01 6.92984045e-01 7.45480120e-01
8.56764317e-01 -1.09532261e-02 3.63591433e-01 8.06318700e-01
1.55331016e-01 9.61490273e-02 3.06002915e-01 -9.84189808e-01
3.51050258e-01 8.07396889e-01 3.25536877e-01 -9.32311416e-01
-4.03440893e-01 -3.36052120e-01 -5.91738224e-01 8.98019969e-01
7.77448475e-01 -1.20902779e-02 -9.97498572e-01 1.76708722e+00
1.20503172e-01 -5.24093390e-01 -1.76487878e-01 1.07018697e+00
5.58511972e-01 4.75224048e-01 -4.54749942e-01 1.23820700e-01
1.38952732e+00 -5.37940681e-01 -5.66189624e-02 -3.27705979e-01
1.91452459e-01 -5.38209081e-01 9.94077027e-01 3.91369998e-01
-1.49819028e+00 -4.64032203e-01 -1.18895066e+00 -1.15519404e-01
-4.08854932e-01 -3.63884896e-01 6.15547299e-01 5.27210951e-01
-1.21489084e+00 7.89149582e-01 -7.90309250e-01 -4.34480667e-01
2.47466743e-01 2.34786093e-01 -5.95870674e-01 -9.29877982e-02
-4.91288543e-01 9.12105501e-01 -1.37435287e-01 -9.96837914e-02
-1.01904035e+00 -5.37408233e-01 -6.34261966e-01 2.42487714e-01
-4.76904660e-02 -8.02766085e-01 1.19675767e+00 -1.21080434e+00
-9.92185950e-01 1.48209095e+00 -7.40627348e-02 -3.67152244e-02
5.62780619e-01 -6.73424304e-02 -8.62060189e-02 1.82868674e-01
-1.22279406e-01 4.23120052e-01 6.95491731e-01 -1.58182847e+00
-3.87529552e-01 -1.06048024e+00 6.83500350e-01 4.83361751e-01
2.05245823e-01 4.83333878e-02 -1.08195312e-01 -5.00946820e-01
5.59971035e-01 -9.05353665e-01 -2.57701427e-01 4.73526418e-01
-2.30910704e-01 -1.19785103e-03 6.95154905e-01 -1.24981500e-01
4.20897007e-01 -2.00389528e+00 2.65273094e-01 -1.79799087e-02
5.70655882e-01 -1.44403875e-01 4.15424816e-02 5.78666568e-01
-3.17287385e-01 1.07578516e-01 -5.90075031e-02 -8.27136785e-02
-1.61905676e-01 7.26521239e-02 -2.90880591e-01 8.21936369e-01
8.95344615e-02 6.34446979e-01 -8.65942955e-01 -9.26415902e-03
1.70997992e-01 4.67895478e-01 -2.43637428e-01 1.58504680e-01
-1.64260894e-01 3.62882167e-01 -6.02885425e-01 2.95342594e-01
7.45698094e-01 -4.71924841e-01 1.07825615e-01 -2.63607018e-02
-1.68172434e-01 6.35499120e-01 -1.23659849e+00 1.58039486e+00
-2.84166485e-01 7.86520958e-01 1.84376314e-01 -1.04936659e+00
9.58957732e-01 3.15465927e-01 5.62728047e-01 -4.00308311e-01
7.26247905e-03 1.78231392e-02 -2.78537441e-02 -3.76269490e-01
5.34972370e-01 -5.18518507e-01 1.67593509e-01 8.87716234e-01
-5.73910810e-02 -1.07815012e-01 -1.99610218e-01 -8.19826350e-02
1.17455924e+00 9.90251154e-02 5.04993558e-01 -3.15145433e-01
3.10256649e-02 5.41936681e-02 1.24916434e-01 6.85502648e-01
-2.31026679e-01 1.07835436e+00 7.03938961e-01 -9.69452262e-01
-1.23106623e+00 -1.26329303e+00 -3.74411404e-01 7.47846365e-01
3.24453771e-01 6.22345544e-02 -4.32130486e-01 -5.01838088e-01
2.37818584e-01 3.23902607e-01 -8.38752329e-01 3.10074538e-02
-4.53299791e-01 -4.33698118e-01 1.19076006e-01 7.03182936e-01
-6.94139227e-02 -9.84345317e-01 -1.18657386e+00 -1.28233209e-01
3.89628798e-01 -7.20814884e-01 4.71531739e-03 5.98120153e-01
-1.25347531e+00 -1.07931876e+00 -8.18980634e-01 -5.10001838e-01
8.35912704e-01 9.10843551e-01 1.58590996e+00 -1.53230831e-01
-1.05443105e-01 7.49084055e-01 -1.77912861e-01 -3.04862201e-01
-3.13339531e-01 -2.72831202e-01 -1.44106328e-01 1.41635155e-02
3.87315333e-01 -9.36238587e-01 -7.42557168e-01 7.14414120e-02
-8.30312371e-01 -5.29418588e-02 5.62662840e-01 5.56753755e-01
4.81774300e-01 -2.88945913e-01 1.14894703e-01 -9.90197957e-01
3.39725763e-01 -6.05149984e-01 -5.53109884e-01 1.76256835e-01
-1.46780193e-01 2.23996252e-01 5.48977911e-01 -1.18660256e-01
-8.21042478e-01 -8.54086056e-02 -5.21044657e-02 -6.32456601e-01
-4.88131076e-01 3.01986545e-01 -4.53731269e-02 -1.23252431e-02
6.48520052e-01 2.45765418e-01 7.74770230e-02 -6.88856125e-01
3.87704492e-01 3.04890305e-01 2.28913054e-01 -5.69295883e-01
5.25482535e-01 7.55713522e-01 -1.31137013e-01 -7.72582710e-01
-7.21441865e-01 -2.42088750e-01 -8.77187252e-01 -7.04795569e-02
7.75094688e-01 -6.06074095e-01 -4.81452405e-01 2.25288510e-01
-1.07514858e+00 2.85274219e-02 -6.86066806e-01 3.39858592e-01
-9.24428940e-01 6.65735155e-02 -4.44878221e-01 -8.17597628e-01
2.28597432e-01 -1.27636552e+00 1.17898822e+00 8.06767400e-03
-1.31764621e-01 -9.53793228e-01 2.91515946e-01 5.04655875e-02
1.92430213e-01 3.11638296e-01 1.28106380e+00 -4.30162460e-01
-7.92858481e-01 -3.53415608e-01 -3.03131640e-01 1.92244843e-01
1.82771042e-01 -1.68928221e-01 -1.15928125e+00 -3.73500913e-01
3.72974694e-01 -3.73661965e-01 7.88606286e-01 7.79207945e-01
9.92792130e-01 -1.21460177e-01 -2.83613980e-01 6.21414244e-01
1.75261366e+00 2.97027081e-01 6.49629712e-01 2.38448665e-01
5.35113752e-01 6.58385098e-01 -8.39912891e-02 1.69952288e-01
2.06319243e-01 7.85901129e-01 7.25344181e-01 -5.46719544e-02
-8.87263007e-03 -8.78747329e-02 5.97885735e-02 4.45152074e-01
-2.57792503e-01 -1.30555674e-01 -8.80744159e-01 5.00272870e-01
-1.35928440e+00 -1.19087839e+00 3.54032278e-01 2.51030540e+00
3.71742994e-01 3.66462916e-01 3.91447186e-01 9.85133424e-02
5.62627673e-01 7.37110257e-01 -8.56791615e-01 -5.53762734e-01
1.60172105e-01 -1.63407937e-01 2.47716591e-01 5.01484036e-01
-8.85749221e-01 3.12841713e-01 7.45548964e+00 1.73265383e-01
-1.32564008e+00 -1.77272007e-01 7.04055309e-01 -1.23879105e-01
-5.93980908e-01 -3.70511413e-02 -5.00262856e-01 2.99254686e-01
6.32983208e-01 1.30592445e-02 3.18925947e-01 8.00195158e-01
-2.16708705e-01 -4.53393236e-02 -1.53813589e+00 8.04567218e-01
-4.85746861e-02 -1.44080710e+00 1.26461044e-01 3.38427395e-01
5.28555632e-01 3.19461733e-01 1.07121840e-01 -2.92976320e-01
5.28484523e-01 -1.09332824e+00 8.54297340e-01 5.52056193e-01
6.36817753e-01 -3.26550633e-01 1.73881412e-01 4.67920005e-01
-1.03990114e+00 -9.84652340e-02 -3.84475410e-01 -1.36132970e-01
9.53628402e-03 4.55130398e-01 -4.50979233e-01 2.87674576e-01
6.53162837e-01 5.19708037e-01 -4.46266949e-01 7.93808460e-01
2.59693801e-01 4.65016216e-02 -1.41841605e-01 1.60882398e-01
2.16452226e-01 -3.22895259e-01 7.70129681e-01 9.72290516e-01
2.38750264e-01 2.28798032e-01 -4.66170497e-02 1.06164563e+00
1.21646933e-02 -5.23315489e-01 -1.18010151e+00 4.34113853e-02
4.64406043e-01 1.07491219e+00 -9.18590248e-01 -1.74654678e-01
-6.61345840e-01 4.18924540e-01 4.48302567e-01 4.04050082e-01
-3.14287722e-01 -1.06368214e-01 8.83169591e-01 5.46146512e-01
3.93646181e-01 -4.86172110e-01 -4.28707272e-01 -1.18384409e+00
2.17477247e-01 -6.24872923e-01 1.20267309e-01 -1.03374159e+00
-1.26918721e+00 6.42886162e-01 1.62490439e-02 -1.52383482e+00
-2.91204184e-01 -8.30967605e-01 -5.35142779e-01 9.51249599e-01
-1.05068564e+00 -8.12882781e-01 -2.21675813e-01 2.39277050e-01
5.22001386e-01 7.77017996e-02 7.09931850e-01 -1.27073765e-01
3.71049084e-02 1.85538381e-01 2.91866541e-01 2.56137084e-02
8.25396739e-03 -1.44683957e+00 6.50658429e-01 3.13477159e-01
5.18356204e-01 6.48538947e-01 6.92833543e-01 -2.09458917e-02
-1.58983374e+00 -5.47980309e-01 4.14532363e-01 -9.71530080e-01
2.99059451e-01 -2.08716094e-01 -9.08982396e-01 1.04761338e+00
2.25640401e-01 1.89167649e-01 4.55607057e-01 3.05361152e-01
-7.14926004e-01 2.95373406e-02 -1.06457567e+00 4.43855882e-01
1.07975209e+00 -8.07485878e-01 -6.93080127e-01 1.42694354e-01
4.85195190e-01 -4.70257998e-01 -7.77086437e-01 2.29169503e-01
7.64421225e-01 -1.83735108e+00 1.20374835e+00 -7.71646082e-01
7.05682576e-01 -7.27190524e-02 -5.44846416e-01 -1.38888621e+00
-5.47162414e-01 9.44932271e-03 1.92413863e-05 7.60185122e-01
2.24261105e-01 -6.59017563e-01 1.00170445e+00 3.97166997e-01
-1.76434696e-01 -1.42005229e+00 -7.25580513e-01 -5.25189877e-01
2.62632221e-01 -1.87127322e-01 5.28563142e-01 7.86981702e-01
-2.06040084e-01 4.32245016e-01 2.54179031e-01 5.18577993e-02
4.33139145e-01 6.45187080e-01 7.04703808e-01 -1.34857976e+00
-2.88045734e-01 -7.26656437e-01 -8.07363570e-01 -1.31872964e+00
-4.34951782e-01 -8.16934228e-01 -3.08389157e-01 -1.54009354e+00
3.87780011e-01 -4.82565463e-01 -2.00752705e-01 2.23125084e-04
2.49027222e-01 -6.52487054e-02 1.40209496e-01 6.31718755e-01
-1.45192102e-01 1.79690942e-01 1.24219024e+00 1.68177765e-02
-7.38219218e-03 5.52971438e-02 -1.01270294e+00 1.11025524e+00
3.63471806e-01 -1.42468184e-01 -5.48212171e-01 -1.02797937e+00
1.61025435e-01 3.80140841e-01 5.98826647e-01 -9.54156458e-01
-1.01263940e-01 -3.49534392e-01 8.42492640e-01 -6.65028870e-01
7.64440835e-01 -7.06080317e-01 1.87857836e-01 9.66982692e-02
-6.64844930e-01 3.78533721e-01 -2.02798724e-01 7.10747063e-01
-9.18763205e-02 -2.15168476e-01 1.06290686e+00 -7.19422936e-01
-4.76526529e-01 2.39966840e-01 -1.15203910e-01 3.81806232e-02
9.57867324e-01 -8.00919712e-01 -3.54643732e-01 -5.30558288e-01
-7.03730762e-01 -3.31779599e-01 9.50697184e-01 2.29858935e-01
5.14772594e-01 -1.28770149e+00 -4.77110982e-01 4.28432912e-01
1.19001970e-01 1.11469679e-01 6.95320666e-02 5.00123858e-01
-4.79964435e-01 2.83055335e-01 -5.59688091e-01 -8.09593678e-01
-8.18839431e-01 5.23779869e-01 5.57874799e-01 -2.00819001e-01
-1.02590525e+00 8.67402673e-01 7.51518726e-01 -2.99681127e-01
1.75089225e-01 -3.03558022e-01 -1.53842017e-01 5.46269305e-02
3.49804878e-01 1.18102975e-01 1.23382643e-01 -6.63031161e-01
-2.33294964e-01 8.53861332e-01 -2.05540493e-01 -1.45328343e-01
1.32922280e+00 -8.81991833e-02 4.21585850e-02 8.43216538e-01
1.22369826e+00 -1.09447718e-01 -1.72992373e+00 -2.36525446e-01
-2.42331341e-01 -8.44897628e-01 -2.76526004e-01 -4.95040834e-01
-1.13236809e+00 1.28419626e+00 4.67590034e-01 9.08909202e-01
8.97300303e-01 2.43275195e-01 4.59403604e-01 1.76013336e-01
5.84572494e-01 -5.28464973e-01 2.66664743e-01 3.62499028e-01
1.08482206e+00 -1.03989625e+00 3.41607332e-01 -6.34613447e-04
-4.57945883e-01 1.14414883e+00 4.18200791e-01 -7.89727747e-01
6.13610625e-01 1.86348140e-01 3.40757892e-02 -7.11659610e-01
-7.94939220e-01 6.35197237e-02 1.86634541e-01 6.07528627e-01
6.02397084e-01 9.34679881e-02 3.51540864e-01 1.34458348e-01
-2.99663842e-01 -2.82281637e-01 6.14531159e-01 1.00022972e+00
-7.03826904e-01 -5.39257109e-01 -4.03634518e-01 5.53749084e-01
-2.27029994e-01 4.23858315e-01 -6.36112988e-01 8.96486104e-01
-9.60400626e-02 3.41464251e-01 3.67082477e-01 -2.59213239e-01
3.64751667e-01 4.78551388e-02 8.22534919e-01 -8.30312371e-01
-1.48196533e-01 -1.29886433e-01 -1.23966038e-01 -4.85964268e-01
-3.55201364e-01 -6.18983328e-01 -9.05035555e-01 -3.84019196e-01
-1.43081516e-01 -1.18033163e-01 4.63407338e-01 9.41072941e-01
1.36777937e-01 2.23263696e-01 8.30569386e-01 -1.58434975e+00
-8.34631383e-01 -6.33844197e-01 -8.07826638e-01 4.94401157e-01
7.98716843e-01 -7.67193556e-01 -5.10418475e-01 -1.60165712e-01] | [8.472466468811035, -3.0203824043273926] |
4d35aea4-98d8-4fb9-8730-fbdcb8ace2cb | finding-mirror-symmetry-via-registration | 1611.05971 | null | http://arxiv.org/abs/1611.05971v2 | http://arxiv.org/pdf/1611.05971v2.pdf | Finding Mirror Symmetry via Registration | Symmetry is prevalent in nature and a common theme in man-made designs. Both
the human visual system and computer vision algorithms can use symmetry to
facilitate object recognition and other tasks. Detecting mirror symmetry in
images and data is, therefore, useful for a number of applications. Here, we
demonstrate that the problem of fitting a plane of mirror symmetry to data in
any Euclidian space can be reduced to the problem of registering two datasets.
The exactness of the resulting solution depends entirely on the registration
accuracy. This new Mirror Symmetry via Registration (MSR) framework involves
(1) data reflection with respect to an arbitrary plane, (2) registration of
original and reflected datasets, and (3) calculation of the eigenvector of
eigenvalue -1 for the transformation matrix representing the reflection and
registration mappings. To support MSR, we also introduce a novel 2D
registration method based on random sample consensus of an ensemble of
normalized cross-correlation matches. With this as its registration back-end,
MSR achieves state-of-the-art performance for symmetry line detection in two
independent 2D testing databases. We further demonstrate the generality of MSR
by testing it on a database of 3D shapes with an iterative closest point
registration back-end. Finally, we explore its applicability to examining
symmetry in natural systems by assessing the degree of symmetry present in
myelinated axon reconstructions from a larval zebrafish. | ['Marcelo Cicconet', 'David G. C. Hildebrand', 'Hunter Elliott'] | 2016-11-18 | null | null | null | null | ['line-detection'] | ['computer-vision'] | [ 4.39269245e-01 -2.14951321e-01 4.44200873e-01 -2.34942883e-01
-4.70318854e-01 -6.28881872e-01 7.43237793e-01 -1.74537674e-01
-4.85754371e-01 1.38628095e-01 8.34957510e-02 -1.68266818e-01
-5.33611238e-01 -4.78753924e-01 -6.19850218e-01 -5.69338620e-01
-2.09790424e-01 5.17029047e-01 3.11216861e-01 -2.66310394e-01
7.34384537e-01 8.91976297e-01 -1.47371352e+00 1.09183773e-01
4.31571633e-01 7.92806149e-01 3.24841738e-02 3.58506173e-01
2.26116315e-01 -2.09264725e-01 -2.54812658e-01 -2.75357693e-01
6.64002597e-01 -3.74161929e-01 -7.94090986e-01 1.14978790e-01
4.79299456e-01 -5.25725149e-02 1.14448898e-01 9.67885911e-01
5.61394870e-01 5.87270129e-03 9.04416203e-01 -1.09432006e+00
-2.46526107e-01 -1.94082037e-02 -7.55540669e-01 -6.56349286e-02
4.82732892e-01 -1.57855198e-01 9.09572303e-01 -1.01225543e+00
9.73411620e-01 8.09502065e-01 8.14198434e-01 3.40413839e-01
-1.50666046e+00 -4.57948625e-01 -5.71088731e-01 -1.20718487e-01
-1.18037117e+00 -6.09333873e-01 5.79282165e-01 -8.46208870e-01
7.56387234e-01 1.67294174e-01 6.27478659e-01 4.10069585e-01
2.11525112e-01 -1.39859602e-01 1.16754484e+00 -4.95023727e-01
2.21833080e-01 -5.43819070e-01 -2.48974949e-01 5.05432367e-01
8.28385055e-02 1.40997395e-01 -5.59970498e-01 -2.09736273e-01
1.07204640e+00 -1.85666889e-01 -3.72256160e-01 -7.65508473e-01
-1.21983075e+00 5.63065469e-01 2.09261835e-01 8.23005736e-02
-1.84425250e-01 -3.52972090e-01 5.04523180e-02 4.33123291e-01
1.62698463e-01 7.37593114e-01 -4.47236933e-02 3.66965123e-02
-7.98843563e-01 1.81577787e-01 5.83357275e-01 6.86919630e-01
5.44046402e-01 -2.63029516e-01 4.13001448e-01 8.63737464e-01
3.57052863e-01 3.75899673e-01 6.47583783e-01 -8.64321589e-01
1.50527820e-01 6.32025182e-01 -3.24352324e-01 -8.32230270e-01
-8.88622284e-01 -3.45374197e-01 -7.36442029e-01 7.05129623e-01
7.20326424e-01 3.15635324e-01 -5.01098275e-01 1.61484718e+00
5.07297993e-01 -6.30029216e-02 4.86331666e-03 9.69450831e-01
4.55162942e-01 -2.49633510e-02 -7.17002451e-01 5.45424670e-02
1.31137514e+00 -8.72893110e-02 1.71307251e-02 1.01015218e-01
3.94066364e-01 -1.17262828e+00 4.48901027e-01 2.02724367e-01
-9.84928370e-01 -8.18576887e-02 -1.04112589e+00 -1.11488573e-01
2.55173002e-03 1.39344022e-01 1.26741109e-02 2.82581508e-01
-8.80044818e-01 7.91547418e-01 -6.48805380e-01 -5.98744392e-01
1.91805109e-01 4.68233019e-01 -1.04060280e+00 3.64487529e-01
-3.56737375e-01 8.63123953e-01 -1.66134462e-01 2.60028113e-02
-6.42143637e-02 -8.00524294e-01 -6.45615876e-01 -3.01357388e-01
-1.21952452e-01 -5.79650760e-01 9.68758941e-01 -4.52924490e-01
-1.47452772e+00 1.43791759e+00 -1.08920746e-01 -1.95399165e-01
6.92079008e-01 1.82915643e-01 -1.73189163e-01 1.15842775e-01
2.40248546e-01 3.28915179e-01 7.00580060e-01 -1.10578048e+00
-3.58154088e-01 -9.27893937e-01 -4.26562577e-01 1.37317523e-01
1.46903962e-01 9.24399421e-02 -1.09747693e-01 -4.78258371e-01
1.16275752e+00 -1.04749227e+00 5.69660626e-02 4.65821475e-01
-2.86696792e-01 2.08986223e-01 7.50285506e-01 -6.70795262e-01
5.49236357e-01 -2.22769475e+00 2.84919441e-01 8.00290227e-01
2.70268708e-01 -4.12807055e-02 -3.55905563e-01 5.62755764e-01
-5.11405885e-01 -2.53693938e-01 -2.51637429e-01 -8.81747250e-03
-9.70152989e-02 -2.21966490e-01 -2.70148337e-01 1.14561522e+00
1.23984747e-01 5.63596308e-01 -4.90448713e-01 6.08393885e-02
1.62129197e-02 4.85172391e-01 -5.66509545e-01 -1.31667815e-02
5.66979110e-01 5.27437806e-01 6.06095232e-02 2.44815186e-01
6.44847274e-01 -7.55887702e-02 -3.12560275e-02 -5.11346757e-01
-6.38255000e-01 1.43706366e-01 -1.41095281e+00 1.51021338e+00
-3.34181301e-02 7.39740491e-01 -6.97567090e-02 -9.12331522e-01
1.19797087e+00 1.97029799e-01 8.18352103e-01 -8.30402970e-01
7.73977861e-02 4.33632970e-01 3.86316925e-01 -2.55813330e-01
7.38264620e-02 -1.37463674e-01 3.66807133e-01 7.57115841e-01
-9.81839187e-03 -4.12100405e-01 1.73292235e-02 -2.82756299e-01
9.74398494e-01 1.85863271e-01 4.44723636e-01 -5.39722264e-01
4.31003541e-01 -3.01494122e-01 3.85394961e-01 1.88512057e-01
4.88408618e-02 1.14414775e+00 4.07241523e-01 -5.11182010e-01
-1.44613838e+00 -1.30989170e+00 -5.95457256e-01 2.73985654e-01
2.48793978e-02 -2.04775184e-01 -8.03370535e-01 6.70187399e-02
4.37205322e-02 1.01734981e-01 -4.87094015e-01 -1.28847539e-01
-5.55605054e-01 -6.79378510e-01 5.34003198e-01 1.11802027e-01
5.09678960e-01 -6.39533818e-01 -1.08267844e+00 -1.34188697e-01
-4.76666242e-02 -1.22954559e+00 -3.97885233e-01 -8.17093551e-02
-8.78100634e-01 -1.44956195e+00 -8.28211486e-01 -6.96211100e-01
8.61013532e-01 4.41146523e-01 5.67826450e-01 -4.95657921e-02
-4.74914014e-01 4.28514004e-01 -1.03547603e-01 -1.51430190e-01
-2.92434722e-01 -4.55633730e-01 4.38466668e-01 1.35723297e-02
1.48209184e-01 -9.31013763e-01 -4.18604523e-01 8.81106138e-01
-8.38185668e-01 3.43770459e-02 3.75647068e-01 6.78987145e-01
6.15765750e-01 -1.85598776e-01 2.22682208e-01 -1.17865041e-01
6.58278465e-01 5.23217805e-02 -9.48801816e-01 7.43423402e-02
-2.52285630e-01 3.00450176e-01 2.04551399e-01 -2.58974254e-01
-3.98112476e-01 2.41156384e-01 6.13734708e-04 -1.29132733e-01
-1.86124034e-02 3.57136309e-01 1.34595335e-01 -6.11926556e-01
7.79430091e-01 2.98688233e-01 7.14957416e-01 -4.26342875e-01
1.59205738e-02 5.94829857e-01 7.86145091e-01 -4.46492672e-01
9.12838280e-01 9.38288212e-01 7.42866874e-01 -1.26200449e+00
-1.37796104e-01 -6.44813180e-01 -9.69828963e-01 -2.56113648e-01
6.28043950e-01 -4.71605211e-01 -9.54650462e-01 7.19681084e-01
-1.15153873e+00 -1.32234395e-01 -2.86268324e-01 6.50419652e-01
-1.00446689e+00 6.57752097e-01 -5.20503744e-02 -4.57837611e-01
-2.28486910e-01 -1.18848836e+00 1.10870576e+00 -1.16331011e-01
-3.97159249e-01 -5.94326377e-01 4.32957858e-01 2.05010504e-01
1.97926953e-01 1.62797228e-01 9.48900104e-01 -6.80396378e-01
-2.66431749e-01 -2.30324313e-01 -1.99137449e-01 -1.02919050e-01
2.19521448e-01 2.01459169e-01 -8.39781046e-01 -2.13715777e-01
7.91034624e-02 8.87674168e-02 4.51626301e-01 3.01231056e-01
4.18559402e-01 1.15427136e-01 8.82943049e-02 7.36739635e-01
1.16348362e+00 -5.07197678e-02 7.12187946e-01 6.46182835e-01
3.01772088e-01 1.19100463e+00 1.24570578e-01 4.93232965e-01
2.68861908e-03 1.15159583e+00 3.23617190e-01 -4.80693914e-02
-1.88945755e-01 3.51060368e-02 9.76732299e-02 5.29569447e-01
-5.56617320e-01 5.64835250e-01 -9.62559402e-01 1.24170199e-01
-1.67513251e+00 -1.15061069e+00 -3.24144840e-01 2.81741667e+00
4.12694931e-01 -1.36293456e-01 3.10783148e-01 3.78168732e-01
7.46694982e-01 -5.71553528e-01 -4.54404444e-01 -8.93675759e-02
-2.46686712e-01 3.35577205e-02 4.08857077e-01 3.21513563e-01
-8.59355450e-01 3.87435406e-01 6.47266912e+00 2.42016584e-01
-1.32500243e+00 -4.62635756e-01 -1.46480247e-01 4.03382838e-01
-9.56063643e-02 9.28572267e-02 -4.43550766e-01 1.20525949e-01
1.43589526e-01 -1.33964270e-01 4.84464705e-01 3.76176238e-01
2.83968240e-01 -3.57939303e-02 -1.16677797e+00 1.13356876e+00
8.86078998e-02 -1.25526822e+00 1.14847701e-02 4.10832614e-01
5.13938308e-01 2.14863271e-01 3.46546397e-02 -7.18270659e-01
-2.47986600e-01 -9.00676489e-01 6.98049486e-01 4.87918198e-01
8.69017899e-01 -2.77081579e-01 4.12206292e-01 2.31257722e-01
-1.12366390e+00 7.19117597e-02 -2.38858461e-01 -1.51324496e-02
1.52130097e-01 4.72773015e-01 -8.85448813e-01 5.70131600e-01
6.80238843e-01 4.71676618e-01 -2.92800725e-01 1.54546988e+00
4.01066579e-02 -1.77560493e-01 -6.15192831e-01 3.67366493e-01
-2.06040144e-01 -7.71598756e-01 7.93889642e-01 7.84348071e-01
5.75463653e-01 1.37729853e-01 -4.69749421e-01 8.63303721e-01
3.90347272e-01 1.69448346e-01 -5.85092425e-01 3.03855091e-01
2.71706790e-01 1.13065410e+00 -1.02228773e+00 4.50513721e-01
-3.79884273e-01 7.46919036e-01 2.71704584e-01 2.22350031e-01
-1.02973007e-01 -3.57040912e-01 9.31861341e-01 3.64416569e-01
2.03834072e-01 -5.02536356e-01 -2.26910219e-01 -1.01183581e+00
3.49877119e-01 -7.30306685e-01 7.89784268e-02 -8.44148338e-01
-1.26273274e+00 3.67173463e-01 -5.75947911e-02 -1.58579373e+00
-2.68123835e-01 -9.12333488e-01 -6.93242371e-01 8.45348418e-01
-1.04200351e+00 -8.21253955e-01 -2.14597061e-01 5.36452949e-01
-5.72267473e-02 -2.09897131e-01 8.13681901e-01 1.40336975e-01
-4.29490469e-02 2.41777509e-01 1.71755210e-01 2.55338579e-01
7.16462970e-01 -1.02133691e+00 5.50337613e-01 7.90137112e-01
3.26455176e-01 7.32099712e-01 8.54507148e-01 -5.06276190e-01
-1.27508748e+00 -5.10524988e-01 8.33759367e-01 -2.19360977e-01
6.58180118e-01 -1.21470161e-01 -7.02656448e-01 6.25407696e-01
-4.47752714e-01 -1.14601627e-01 6.21497810e-01 -2.74365991e-01
-5.43583930e-01 7.07323253e-02 -1.22677231e+00 6.64433897e-01
1.21425748e+00 -5.33025801e-01 -5.79343379e-01 2.02239826e-01
-1.51932061e-01 -3.88129622e-01 -9.14364994e-01 4.97138619e-01
1.09546244e+00 -1.15083575e+00 9.15860713e-01 -5.11198580e-01
2.80150294e-01 -5.59756279e-01 -1.69972837e-01 -1.21957529e+00
-1.89452976e-01 -7.07282245e-01 6.92899406e-01 7.98597276e-01
1.37795195e-01 -9.08141792e-01 7.79551804e-01 5.14563501e-01
-1.29044995e-01 -5.19190311e-01 -1.36643958e+00 -9.94847000e-01
-1.17542464e-02 -1.29092440e-01 4.14915174e-01 6.46095932e-01
2.28879452e-01 3.44661511e-02 1.89584196e-01 3.78812462e-01
8.05840135e-01 3.65115225e-01 1.04191852e+00 -1.59039676e+00
-4.22060154e-02 -7.02946723e-01 -1.22784209e+00 -9.80557799e-01
-2.89574824e-03 -9.89089131e-01 -1.65765658e-02 -1.18712997e+00
6.77942717e-03 -3.55912417e-01 4.04533029e-01 1.05393231e-01
6.74883544e-01 3.75835985e-01 8.96176025e-02 4.87516075e-01
6.73842654e-02 3.98260355e-01 1.25638628e+00 2.74010777e-01
-9.00906771e-02 2.15428755e-01 -1.93509772e-01 8.29333067e-01
6.54801667e-01 -2.53003985e-01 -3.88352983e-02 -2.37302110e-01
3.15489024e-01 -2.47969165e-01 6.09107614e-01 -1.08073950e+00
5.25969386e-01 1.71184838e-01 2.76514471e-01 -2.40620196e-01
2.75506020e-01 -8.37678611e-01 4.60663885e-01 4.71856236e-01
-1.09016038e-01 4.33208466e-01 -6.88128620e-02 1.38086915e-01
2.52069277e-03 -4.19694871e-01 8.55061054e-01 1.79981247e-01
-4.03491944e-01 1.42739192e-01 -2.47511178e-01 1.41113937e-01
8.86013448e-01 -5.99514544e-01 -3.18534553e-01 -2.35136077e-01
-5.49929738e-01 -1.64517060e-01 1.00147486e+00 7.22974241e-02
9.67803717e-01 -1.20615089e+00 -7.40415514e-01 8.13738286e-01
2.46299833e-01 -1.29324883e-01 -5.69511615e-02 1.02910542e+00
-6.99758112e-01 1.94028661e-01 -6.98988199e-01 -9.37626362e-01
-1.59433711e+00 3.43200713e-02 6.28988862e-01 3.29917163e-01
-5.42381048e-01 4.53933060e-01 9.03282464e-02 -7.80343652e-01
-2.62573332e-01 -2.31547698e-01 -2.92691827e-01 -1.74892440e-01
4.12590116e-01 5.81371546e-01 3.87390196e-01 -1.09806561e+00
-3.78216207e-01 1.47627389e+00 2.92775542e-01 -5.69915771e-01
1.40903437e+00 -3.76505703e-02 -4.31858599e-01 2.09343866e-01
1.29661691e+00 1.98258027e-01 -1.02852464e+00 -9.12521482e-02
-1.01207290e-02 -3.09647977e-01 -3.73386949e-01 -2.08295673e-01
-8.20429862e-01 6.62638903e-01 5.16405463e-01 1.33400783e-01
9.43668246e-01 -1.14686303e-02 1.41369268e-01 4.02493656e-01
3.69195819e-01 -8.49622905e-01 -1.50193244e-01 4.08725083e-01
1.20306683e+00 -1.02429914e+00 2.94423968e-01 -5.62577486e-01
-1.33399799e-01 1.61240411e+00 1.12768643e-01 -4.03686404e-01
8.56355965e-01 2.17566088e-01 3.16071302e-01 -3.60015988e-01
-1.82970688e-01 -9.57860947e-02 6.59456372e-01 7.63852775e-01
3.72944742e-01 -3.05502445e-01 -3.72346640e-01 1.03867479e-01
-5.86262524e-01 -1.76090330e-01 5.22261739e-01 7.21807361e-01
-3.47093046e-01 -1.06855774e+00 -4.29122329e-01 1.83109701e-01
-1.42365873e-01 1.74586773e-01 -5.75939000e-01 6.32767975e-01
-2.72235870e-01 6.41284049e-01 3.76823246e-01 -3.56230527e-01
5.33445537e-01 -2.52876073e-01 7.94356406e-01 -3.13532799e-01
-3.23644578e-01 -6.77201599e-02 -1.81087464e-01 -7.12368309e-01
-5.69086373e-01 -9.32650864e-01 -1.31046748e+00 -9.32974219e-02
-1.74623385e-01 -2.22440064e-02 1.13077307e+00 9.40204799e-01
3.57474864e-01 -1.33076936e-01 5.29708743e-01 -9.73476589e-01
-8.34391594e-01 -5.92592537e-01 -7.41722941e-01 7.30172932e-01
2.54094601e-01 -7.00599790e-01 -5.45106828e-01 -6.02113046e-02] | [8.383743286132812, -2.347079038619995] |
d4ff37b8-acd5-4d03-a36a-b4156b6a8dca | learning-with-privileged-information-for | 1703.09911 | null | http://arxiv.org/abs/1703.09911v1 | http://arxiv.org/pdf/1703.09911v1.pdf | Learning with Privileged Information for Multi-Label Classification | In this paper, we propose a novel approach for learning multi-label
classifiers with the help of privileged information. Specifically, we use
similarity constraints to capture the relationship between available
information and privileged information, and use ranking constraints to capture
the dependencies among multiple labels. By integrating similarity constraints
and ranking constraints into the learning process of classifiers, the
privileged information and the dependencies among multiple labels are exploited
to construct better classifiers during training. A maximum margin classifier is
adopted, and an efficient learning algorithm of the proposed method is also
developed. We evaluate the proposed method on two applications: multiple object
recognition from images with the help of implicit information about object
importance conveyed by the list of manually annotated image tags; and multiple
facial action unit detection from low-resolution images augmented by
high-resolution images. Experimental results demonstrate that the proposed
method can effectively take full advantage of privileged information and
dependencies among multiple labels for better object recognition and better
facial action unit detection. | ['Xiaoxiao Shi', 'Tanfang Chen', 'Shiyu Chen', 'Shangfei Wang'] | 2017-03-29 | null | null | null | null | ['action-unit-detection', 'facial-action-unit-detection'] | ['computer-vision', 'computer-vision'] | [ 6.53996706e-01 2.95514357e-03 -7.67403126e-01 -8.32134426e-01
-6.09401882e-01 -3.30399662e-01 3.55499148e-01 1.17131531e-01
-5.69654405e-01 6.38899744e-01 4.29848284e-02 2.36319005e-01
-4.55501646e-01 -4.38348919e-01 -2.15858176e-01 -8.77865374e-01
1.50609806e-01 2.16362774e-02 2.45755777e-01 1.25290811e-01
4.30237532e-01 7.97215939e-01 -2.07535577e+00 4.50168610e-01
6.43359721e-01 1.22646654e+00 1.93258807e-01 1.42186955e-01
-5.66344447e-02 8.77763093e-01 -4.79845494e-01 -2.18655113e-02
1.21285081e-01 -9.72541943e-02 -5.96589267e-01 3.72263342e-01
5.72918475e-01 -3.91162783e-01 5.04411280e-01 9.61837471e-01
3.73511642e-01 2.18803525e-01 7.29368508e-01 -1.24633360e+00
-2.28716850e-01 1.15727901e-01 -6.49860799e-01 7.54873678e-02
3.70573282e-01 -4.09119517e-01 1.08573306e+00 -9.05532062e-01
4.01871115e-01 1.50082839e+00 4.82163914e-02 6.57299399e-01
-9.98591661e-01 -9.55062807e-01 5.01276135e-01 3.71665329e-01
-1.50229037e+00 -5.39582729e-01 8.78107846e-01 -4.93331105e-01
3.29547524e-01 3.94853503e-01 2.79735357e-01 6.92471802e-01
-1.78712327e-02 7.75708199e-01 1.36895275e+00 -6.06166959e-01
-6.20648153e-02 5.62507868e-01 4.61198330e-01 1.05973685e+00
1.14227295e-01 2.13116661e-01 -6.83128834e-01 -5.40471375e-01
4.60980535e-01 2.02932999e-01 -5.86223491e-02 -3.27526420e-01
-5.30134261e-01 8.33767891e-01 1.65779933e-01 3.36791039e-01
-7.08319321e-02 -4.37257767e-01 6.73447475e-02 -9.52439681e-02
4.02752072e-01 9.72668156e-02 -3.20724636e-01 4.72143769e-01
-6.82091236e-01 -1.67171836e-01 3.75319749e-01 7.92928398e-01
1.16485870e+00 -3.74102592e-01 -9.36974287e-02 1.12635744e+00
6.82617188e-01 5.34536183e-01 4.17042136e-01 -8.34534645e-01
-1.66598856e-02 7.35810459e-01 -4.87093143e-02 -9.32750881e-01
-4.27986234e-01 3.31475884e-02 -4.93801683e-01 4.11557913e-01
1.60053089e-01 1.38611868e-01 -6.92905843e-01 1.60304415e+00
4.38738495e-01 3.33586782e-01 -7.33876005e-02 7.83501863e-01
6.71295822e-01 2.94649363e-01 3.15651506e-01 -7.24021912e-01
1.36927891e+00 -5.19419372e-01 -8.06630433e-01 -1.97179377e-01
5.96460819e-01 -6.30259693e-01 7.66808212e-01 3.32981437e-01
-5.85710347e-01 -6.62560582e-01 -1.12340796e+00 1.75371543e-01
-3.68271202e-01 5.22628129e-01 6.37840688e-01 5.81789434e-01
-5.33891857e-01 4.34364170e-01 -4.59620923e-01 -5.49965613e-02
5.62086284e-01 7.02316463e-01 -4.33249474e-01 -2.09869742e-01
-1.05589867e+00 8.26677144e-01 5.49007118e-01 1.47991225e-01
-5.54980695e-01 -2.86119550e-01 -1.03225756e+00 -9.99628603e-02
5.52861989e-01 3.04525252e-02 8.14966679e-01 -1.20682395e+00
-1.29447353e+00 9.80934620e-01 -2.03761205e-01 3.15163761e-01
-5.37477946e-03 -2.76451349e-01 -5.50110400e-01 2.36298367e-01
9.24347937e-02 5.87387860e-01 9.32909787e-01 -1.11307609e+00
-1.06782162e+00 -6.27134144e-01 5.68796694e-02 3.92217994e-01
-6.81264222e-01 4.07571882e-01 -2.67737418e-01 -2.63962179e-01
1.55111507e-01 -8.67036045e-01 1.80135131e-01 1.56203583e-01
-3.07166874e-02 -4.15473819e-01 1.00373328e+00 -3.14708441e-01
1.19963074e+00 -2.07841754e+00 -6.05559200e-02 5.54757118e-01
-1.13734035e-02 3.70539904e-01 -2.27077797e-01 -2.56057352e-01
4.85506244e-02 -1.33285046e-01 7.68282935e-02 3.58213075e-02
-2.82403231e-01 4.27545577e-01 2.69875620e-02 4.20941204e-01
2.03824565e-01 2.41563603e-01 -7.04286575e-01 -1.00538099e+00
3.04819524e-01 4.68863040e-01 -1.88134924e-01 2.52986044e-01
-1.81950461e-02 3.70526463e-01 -7.75383651e-01 9.32945311e-01
7.33032763e-01 -1.25119656e-01 4.39627022e-01 -3.33770186e-01
8.45813826e-02 -2.60215819e-01 -1.36697733e+00 1.31208622e+00
-3.25251967e-01 1.14228986e-01 1.83688477e-01 -7.06310928e-01
1.00440824e+00 2.11181477e-01 6.34739280e-01 -5.38522720e-01
2.60277629e-01 9.87279192e-02 -3.09067518e-01 -5.43394625e-01
1.51583806e-01 -1.79450870e-01 9.46667790e-02 4.37240928e-01
1.08956806e-02 4.23386723e-01 -7.40941539e-02 -1.72946706e-01
4.72403258e-01 1.29052207e-01 7.69022167e-01 -1.49406135e-01
1.10804832e+00 -5.45249701e-01 8.55967820e-01 4.57093239e-01
-2.24632069e-01 -1.34575590e-01 1.25577405e-01 -3.70954365e-01
-3.25285435e-01 -7.09678650e-01 -4.02965218e-01 1.71954632e+00
2.94373602e-01 -3.22651953e-01 -3.37726355e-01 -1.06949794e+00
1.42658174e-01 2.51535624e-01 -5.73068440e-01 -1.26522720e-01
-5.15779972e-01 -7.07218349e-01 1.62054479e-01 5.17450750e-01
3.36198002e-01 -9.11674440e-01 -3.39513659e-01 -9.31756124e-02
6.66590780e-02 -9.78439808e-01 -3.47375959e-01 -9.90232155e-02
-6.96918011e-01 -1.29375041e+00 -3.02770287e-01 -8.02022934e-01
9.60745752e-01 3.27870965e-01 4.18832302e-01 3.12287837e-01
-3.96662086e-01 2.48895183e-01 -1.56822011e-01 -3.18448216e-01
-2.48588845e-01 -2.01054260e-01 2.00174034e-01 6.03829801e-01
3.76756191e-01 -3.22467595e-01 -1.09310441e-01 7.18627036e-01
-8.59461069e-01 -3.11446134e-02 9.36730325e-01 9.27671015e-01
6.01000845e-01 -4.14451696e-02 4.71837282e-01 -1.07289708e+00
3.33104759e-01 -1.25988007e-01 -7.57097542e-01 6.34478807e-01
-7.71880805e-01 2.38035187e-01 1.84439048e-01 -5.76648057e-01
-1.37353325e+00 4.24597055e-01 1.21793561e-01 -2.86276639e-01
-2.30995581e-01 1.94429144e-01 -7.46345520e-01 -5.03928304e-01
4.07409906e-01 -4.29651029e-02 2.06776764e-02 -4.95309442e-01
3.10797185e-01 9.27171052e-01 2.31610775e-01 -6.34295404e-01
2.51333326e-01 4.53203112e-01 3.15710813e-01 -6.55367911e-01
-1.25241554e+00 -7.53018200e-01 -9.40339029e-01 -4.84485030e-01
7.25488484e-01 -7.60864675e-01 -9.31311369e-01 4.69450206e-01
-8.28378856e-01 4.65085596e-01 3.11596304e-01 6.98630035e-01
-3.26319277e-01 4.98521298e-01 -2.88339943e-01 -8.94204080e-01
-1.53375581e-01 -1.18603933e+00 1.07746637e+00 3.92994195e-01
1.70795962e-01 -9.32748139e-01 -2.81586975e-01 6.59616351e-01
-8.37518498e-02 1.51983082e-01 1.00340760e+00 -8.29383194e-01
-6.46046877e-01 -3.02516520e-01 -3.40404749e-01 4.09888595e-01
5.19677937e-01 -1.48744643e-01 -1.15444112e+00 -3.31298083e-01
-6.93254545e-02 -5.27549803e-01 6.93620503e-01 1.19329326e-01
1.10847473e+00 -4.28669930e-01 -5.77437460e-01 2.84018070e-01
1.32740200e+00 2.56692499e-01 2.46564344e-01 2.95076277e-02
7.41728127e-01 9.48390961e-01 1.29048908e+00 6.28059804e-01
4.82855029e-02 8.16556156e-01 4.14537579e-01 8.46354738e-02
2.45960414e-01 9.94097367e-02 1.59322187e-01 4.19319510e-01
-1.40072152e-01 1.30391404e-01 -5.24200618e-01 -5.19338548e-02
-1.87524831e+00 -8.62365007e-01 3.36875767e-01 2.21426964e+00
7.78067172e-01 5.27830869e-02 -4.43269946e-02 -2.30999887e-02
1.02630317e+00 1.29804358e-01 -6.25642598e-01 -6.34668469e-02
7.01321289e-02 -5.22271954e-02 2.57956952e-01 6.42637312e-01
-1.28786814e+00 8.32571507e-01 6.42815781e+00 1.13233626e+00
-1.01637125e+00 -2.99090538e-02 3.99820536e-01 -2.23706365e-01
-9.09340940e-03 -1.54664695e-01 -1.38269186e+00 3.28319371e-01
4.71021295e-01 1.36234999e-01 6.95015714e-02 9.75728035e-01
-1.41450837e-01 -3.35733175e-01 -1.08169496e+00 1.05307949e+00
5.18079519e-01 -8.87947261e-01 1.45040631e-01 1.64239898e-01
2.94582307e-01 -7.10361660e-01 7.25836083e-02 8.81201494e-03
8.28445256e-02 -6.68423593e-01 3.24439257e-01 7.59414494e-01
7.46345222e-01 -6.98239625e-01 5.72176695e-01 3.83362651e-01
-1.27899325e+00 -3.43155622e-01 -2.70738453e-01 6.40057847e-02
-3.15485120e-01 3.18392247e-01 -9.19568717e-01 3.71928364e-01
1.79114416e-01 8.03238213e-01 -7.52663195e-01 6.00065529e-01
-3.71208757e-01 1.08520113e-01 -7.36579970e-02 -1.11339033e-01
-1.85195267e-01 -6.10429905e-02 2.60037512e-01 9.10814464e-01
-6.17561005e-02 3.02105576e-01 4.66671824e-01 2.39404723e-01
2.30470940e-01 4.57540900e-01 -4.34333324e-01 3.78900290e-01
4.20875311e-01 1.57811260e+00 -4.28678274e-01 -2.89973110e-01
-5.99115431e-01 5.85861623e-01 4.56424206e-01 4.09378223e-02
-6.83373630e-01 -1.07873745e-01 6.05346203e-01 -1.14024512e-01
5.93862496e-02 1.56598568e-01 1.60001293e-01 -9.32034492e-01
-1.39916793e-01 -6.37989879e-01 8.55837643e-01 -7.02032566e-01
-1.06637478e+00 3.87865901e-01 2.97833294e-01 -1.17128837e+00
-2.34449968e-01 -8.53867352e-01 -1.42535672e-01 7.88386285e-01
-1.69380820e+00 -1.45137823e+00 -2.23764345e-01 7.26050377e-01
1.78695723e-01 -4.15936619e-01 9.53368008e-01 2.96129495e-01
-6.24301732e-01 6.81243658e-01 -1.03667729e-01 -8.41300189e-03
1.05514562e+00 -6.93751037e-01 -9.11305964e-01 4.23921496e-01
1.49263844e-01 5.09914100e-01 1.28293261e-01 -6.27720475e-01
-7.81458676e-01 -1.00264049e+00 7.87632883e-01 -1.69740841e-01
2.74144828e-01 -2.21520975e-01 -7.93636143e-01 5.59571624e-01
-2.41760358e-01 1.01695359e-01 1.37966788e+00 2.82729834e-01
-6.34416163e-01 -4.40319300e-01 -1.25355852e+00 7.21867755e-02
7.34580815e-01 -4.40614223e-01 -4.69563514e-01 3.00370365e-01
2.53781676e-01 -1.92691490e-01 -8.84975493e-01 6.07075274e-01
9.19854343e-01 -6.10477328e-01 8.47799718e-01 -4.58401561e-01
-1.14386618e-01 -2.40019843e-01 -1.95022181e-01 -8.40350807e-01
-4.35731798e-01 -8.10678080e-02 9.51147676e-02 1.29621398e+00
2.60948062e-01 -6.29085124e-01 7.15263128e-01 7.54779339e-01
1.14543699e-01 -6.89210594e-01 -8.54482353e-01 -5.29275179e-01
-5.63582599e-01 1.59435198e-01 5.12334406e-01 9.54916656e-01
-2.23184600e-02 2.35656142e-01 -6.04083955e-01 4.02136534e-01
7.75450766e-01 3.16009670e-01 2.22214505e-01 -1.61562395e+00
-2.05206677e-01 -6.03838153e-02 -6.01020992e-01 -6.14874601e-01
6.98413074e-01 -7.75899231e-01 -1.50848821e-01 -8.87238979e-01
5.26287854e-01 -5.64384997e-01 -7.88754761e-01 9.31969404e-01
-2.53509641e-01 2.36236215e-01 6.60177693e-02 4.02354985e-01
-8.14293087e-01 1.91077471e-01 1.07540476e+00 -1.05099671e-01
-5.97678833e-02 2.61331320e-01 -7.51368046e-01 9.87339318e-01
6.02159441e-01 -4.20976251e-01 -4.07288909e-01 -5.93046732e-02
-2.46522829e-01 1.14146508e-01 2.64746308e-01 -8.58534515e-01
2.01901555e-01 -5.87332904e-01 5.33549249e-01 -4.14381415e-01
7.47347534e-01 -9.84510899e-01 -2.01405481e-01 2.34817907e-01
-6.41441166e-01 -5.72988987e-01 1.48352697e-01 5.41336715e-01
-1.13324858e-01 -3.44953150e-01 1.06253278e+00 3.90243833e-03
-1.07419384e+00 3.28867167e-01 -2.08102807e-01 -4.61057425e-01
1.24617016e+00 -1.08842008e-01 -3.01665753e-01 1.12663232e-01
-1.05204225e+00 1.52264774e-01 1.73516214e-01 5.57286322e-01
8.57994556e-01 -1.52369297e+00 -3.33803654e-01 5.59696972e-01
5.35805881e-01 -6.80361509e-01 1.07137524e-01 4.27809149e-01
2.02601790e-01 4.21623945e-01 -6.17680788e-01 -4.86729771e-01
-2.21646428e+00 4.77264017e-01 1.76755741e-01 -1.11966014e-01
1.07657097e-01 7.13597775e-01 2.77321339e-01 -2.14045927e-01
2.98853487e-01 1.54648736e-01 -8.28348339e-01 3.70948702e-01
6.82251453e-01 3.01180005e-01 -8.24436471e-02 -1.02453375e+00
-4.42214489e-01 9.43992913e-01 -5.03314316e-01 -1.66639499e-02
1.00799751e+00 -1.28253222e-01 -1.94726214e-01 2.84285039e-01
1.19534409e+00 4.44229832e-03 -9.98933136e-01 -5.98123312e-01
4.74899597e-02 -8.72483313e-01 4.99965064e-02 -8.26078892e-01
-1.00581753e+00 7.62016952e-01 8.09374809e-01 -3.47841501e-01
1.41882670e+00 3.84871848e-02 1.00933246e-01 3.65850925e-01
5.14891803e-01 -1.17407310e+00 2.51603991e-01 8.24687853e-02
5.41235626e-01 -1.44142377e+00 3.33359480e-01 -8.80994499e-01
-5.75806856e-01 1.03707922e+00 1.02271831e+00 4.10257071e-01
7.38231182e-01 8.28919839e-03 1.03208460e-01 -4.00693305e-02
-5.34738719e-01 -3.30762446e-01 5.87721765e-01 4.66339588e-01
2.38639474e-01 -6.66670501e-02 -4.44714576e-01 4.53683704e-01
5.10134161e-01 -1.52692080e-01 1.03224255e-01 1.03749156e+00
-9.06277835e-01 -1.40177631e+00 -4.97184038e-01 5.32992125e-01
-3.88306528e-01 1.51029497e-01 -3.77097368e-01 4.59853381e-01
6.23634756e-01 1.10769463e+00 -1.62242696e-01 -4.21936631e-01
1.22455485e-01 1.55961633e-01 5.61749697e-01 -8.02530825e-01
-2.35516116e-01 2.43938074e-01 4.50575538e-02 -4.95228082e-01
-8.07125628e-01 -5.98523974e-01 -1.13353884e+00 2.11979464e-01
-8.90510738e-01 2.01087594e-01 6.42502248e-01 1.00403273e+00
2.91412979e-01 1.84296176e-01 8.84406745e-01 -5.16125023e-01
-4.35927868e-01 -7.65873194e-01 -8.96336734e-01 3.76667589e-01
1.60565972e-01 -1.05365050e+00 -2.79526860e-01 3.10996477e-03] | [9.368966102600098, 4.009395122528076] |
55f0492e-97bb-4fdc-90f4-8502b30778cf | look-before-you-leap-bridging-model-free-and | 1803.07729 | null | http://arxiv.org/abs/1803.07729v2 | http://arxiv.org/pdf/1803.07729v2.pdf | Look Before You Leap: Bridging Model-Free and Model-Based Reinforcement Learning for Planned-Ahead Vision-and-Language Navigation | Existing research studies on vision and language grounding for robot
navigation focus on improving model-free deep reinforcement learning (DRL)
models in synthetic environments. However, model-free DRL models do not
consider the dynamics in the real-world environments, and they often fail to
generalize to new scenes. In this paper, we take a radical approach to bridge
the gap between synthetic studies and real-world practices---We propose a
novel, planned-ahead hybrid reinforcement learning model that combines
model-free and model-based reinforcement learning to solve a real-world
vision-language navigation task. Our look-ahead module tightly integrates a
look-ahead policy model with an environment model that predicts the next state
and the reward. Experimental results suggest that our proposed method
significantly outperforms the baselines and achieves the best on the real-world
Room-to-Room dataset. Moreover, our scalable method is more generalizable when
transferring to unseen environments. | ['William Yang Wang', 'Wenhan Xiong', 'Hongmin Wang', 'Xin Wang'] | 2018-03-21 | look-before-you-leap-bridging-model-free-and-1 | http://openaccess.thecvf.com/content_ECCV_2018/html/Xin_Wang_Look_Before_You_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Xin_Wang_Look_Before_You_ECCV_2018_paper.pdf | eccv-2018-9 | ['vision-language-navigation'] | ['computer-vision'] | [-1.64846033e-01 -3.32873518e-04 -1.83497250e-01 -4.65670675e-01
-5.31616926e-01 -3.84253442e-01 6.95621729e-01 -4.06096816e-01
-7.24494100e-01 7.85387933e-01 6.06525242e-02 -6.65607095e-01
2.82232165e-01 -7.61969507e-01 -1.04922485e+00 -4.08639252e-01
-5.92528097e-02 4.29590821e-01 3.74448299e-01 -8.09698641e-01
1.44262120e-01 1.31962717e-01 -1.42111707e+00 -7.94638246e-02
8.09376657e-01 4.27109569e-01 8.49759459e-01 8.20928931e-01
1.24061622e-01 1.24779379e+00 2.95149665e-02 2.45525181e-01
4.21400309e-01 -4.72758353e-01 -7.37533867e-01 -1.06321685e-01
2.14963540e-01 -7.26013184e-01 -7.47516394e-01 8.95675421e-01
4.72168595e-01 4.23443466e-01 3.14251602e-01 -1.42084825e+00
-8.39885592e-01 3.45440596e-01 -3.34461063e-01 -2.68301945e-02
3.97971153e-01 7.64975846e-01 6.23367012e-01 -7.42471576e-01
7.15410948e-01 1.64313996e+00 5.70523858e-01 1.06401205e+00
-1.00657725e+00 -4.45585757e-01 7.37100184e-01 4.07562345e-01
-7.21879900e-01 -3.02064955e-01 5.93750775e-01 -3.04992527e-01
1.29325199e+00 -4.92031276e-01 7.60609508e-01 1.67634368e+00
4.63375062e-01 1.01070487e+00 1.26219535e+00 -3.73997539e-01
5.22381008e-01 -3.26224267e-01 -2.22779438e-01 1.06532419e+00
5.68503290e-02 7.71580338e-01 -4.04677063e-01 2.99288809e-01
8.55944097e-01 -6.19661510e-02 -8.05876553e-02 -1.12675071e+00
-1.52869833e+00 8.51041317e-01 7.54026890e-01 -1.19893178e-01
-4.36881661e-01 6.83342338e-01 3.17869812e-01 3.79402190e-01
-2.30929151e-01 5.18620610e-01 -5.65798461e-01 -2.66754359e-01
-5.14713705e-01 7.41342247e-01 6.79775357e-01 1.19483817e+00
8.96150827e-01 3.07306409e-01 -2.06296965e-01 5.59490681e-01
6.00778401e-01 6.90356612e-01 7.37886906e-01 -1.52247918e+00
4.50146168e-01 3.61709557e-02 6.26656950e-01 -5.63065410e-01
-6.71830595e-01 -1.77993476e-01 -3.08785349e-01 5.47106504e-01
3.06571305e-01 -3.56580645e-01 -1.12817192e+00 1.96821523e+00
3.57757419e-01 3.44307303e-01 5.15842080e-01 8.68457437e-01
6.60294533e-01 6.17139578e-01 1.21802352e-01 9.62817594e-02
7.84829438e-01 -1.87165189e+00 -6.03434920e-01 -9.31888938e-01
7.63048589e-01 -1.70357466e-01 1.40010786e+00 3.21513921e-01
-8.48830462e-01 -8.45830977e-01 -1.13526690e+00 1.40539363e-01
-2.82288283e-01 -2.23793626e-01 6.37182891e-01 1.78548947e-01
-1.42805231e+00 3.25898260e-01 -1.14499640e+00 -6.01016521e-01
-1.31070197e-01 6.34622648e-02 -3.59138042e-01 -2.85120815e-01
-1.17524779e+00 1.13569963e+00 5.14805198e-01 2.32878506e-01
-1.42172289e+00 2.39035077e-02 -1.25413215e+00 -3.59908342e-01
5.89040041e-01 -9.85773623e-01 2.06289053e+00 -5.92051506e-01
-1.90252399e+00 4.73027349e-01 -2.16030866e-01 -7.20488250e-01
4.69831616e-01 -3.31298590e-01 -8.13816395e-03 -3.33987802e-01
3.86078119e-01 9.47595239e-01 5.81923008e-01 -1.68969929e+00
-8.22840512e-01 -1.50735483e-01 3.77520621e-01 5.56464434e-01
3.96714628e-01 -8.25384736e-01 -4.53365058e-01 -1.57161206e-01
7.05822259e-02 -1.23924887e+00 -8.21007371e-01 -5.79298660e-02
3.28045748e-02 5.11254109e-02 5.54397047e-01 -1.44643992e-01
7.30836391e-01 -1.85435736e+00 1.45008452e-02 -4.35059309e-01
-2.05756977e-01 1.25869647e-01 -5.67619503e-01 7.41070032e-01
3.68664414e-01 -4.62981045e-01 4.76552099e-02 -3.62367332e-01
1.74906060e-01 7.73639619e-01 -5.75904071e-01 1.25408545e-03
-1.34887815e-01 1.19183123e+00 -1.60322130e+00 -2.80340254e-01
3.94805789e-01 -5.27385175e-02 -7.82871246e-01 2.20485747e-01
-6.04046345e-01 7.78245032e-01 -5.18440306e-01 3.57296914e-01
4.35221642e-01 -4.07037735e-02 2.22745359e-01 5.46500862e-01
-1.34557813e-01 1.34015247e-01 -9.46266055e-01 2.26062036e+00
-7.57447422e-01 4.34262514e-01 -1.49540817e-02 -8.24419916e-01
8.77113461e-01 -1.52898729e-01 9.83632077e-03 -1.30015147e+00
-3.37415934e-01 1.42975211e-01 5.68080433e-02 -7.02210724e-01
8.09022427e-01 -2.06567701e-02 -1.75639093e-01 -2.02137195e-02
1.84221454e-02 -4.89523351e-01 2.20646132e-02 -4.91374684e-03
1.19171441e+00 1.00524044e+00 2.14795291e-01 1.39468059e-01
4.06190574e-01 3.16181988e-01 5.99835634e-01 1.37248552e+00
-7.98582971e-01 2.49937043e-01 8.78406018e-02 -4.54856902e-01
-8.38846028e-01 -1.21214402e+00 5.70940018e-01 1.23329163e+00
6.07242525e-01 -5.83681092e-02 -5.56423724e-01 -7.81218708e-01
-6.06672280e-02 1.04656363e+00 -6.27555311e-01 -3.27614993e-01
-7.33183444e-01 -2.57744402e-01 2.56263703e-01 5.80819786e-01
7.36461759e-01 -1.50119019e+00 -1.03605258e+00 5.67620397e-01
-3.17337960e-01 -1.26612854e+00 -1.27523080e-01 2.45668203e-01
-7.73294687e-01 -9.06356037e-01 -4.84807163e-01 -1.01528525e+00
2.43775293e-01 5.70532382e-01 1.03528094e+00 -2.86101699e-01
9.67577770e-02 7.73405373e-01 -4.66273844e-01 -3.78921181e-01
-6.25363469e-01 -1.83058918e-01 1.57000393e-01 -4.26050276e-01
1.16330199e-01 -1.22808196e-01 -7.67321706e-01 1.19504280e-01
-3.57368380e-01 2.24168763e-01 5.63517690e-01 1.26628304e+00
5.29299855e-01 -4.48341727e-01 7.27592349e-01 -3.41701299e-01
8.07368100e-01 -3.16661328e-01 -6.98525727e-01 1.88293695e-01
-9.42643344e-01 4.29351479e-01 5.86102545e-01 -3.92251432e-01
-1.17451048e+00 1.28821164e-01 -1.47923127e-01 -2.04195470e-01
-2.24341109e-01 3.23239058e-01 1.40363619e-01 -7.73358764e-03
6.63168371e-01 5.71474552e-01 -2.51259524e-02 -1.77298789e-03
7.81903267e-01 1.60629228e-01 7.60740280e-01 -6.33526266e-01
6.70616925e-01 5.42446375e-01 -2.91548055e-02 -4.15252805e-01
-6.36723757e-01 -3.67282659e-01 -4.08538431e-01 -1.81202874e-01
8.37678730e-01 -1.25815177e+00 -7.48836517e-01 6.49533391e-01
-9.46426153e-01 -1.35147107e+00 -4.07371759e-01 6.48851514e-01
-1.36302245e+00 4.78213727e-01 -5.44988990e-01 -8.98077190e-01
7.31284767e-02 -1.39525151e+00 1.04256833e+00 4.45134699e-01
8.33458155e-02 -9.63205457e-01 5.46144962e-01 1.46494200e-02
4.39378500e-01 3.30582373e-02 5.67437768e-01 -1.27024010e-01
-6.82797074e-01 2.34910890e-01 -2.49614809e-02 -5.10237738e-03
-7.17912242e-02 -4.25186485e-01 -7.44868755e-01 -5.07683456e-01
-2.05658212e-01 -9.27661657e-01 1.08079803e+00 4.32678849e-01
8.58165979e-01 -1.19861946e-01 -3.49939823e-01 6.09464705e-01
1.36074185e+00 5.69668174e-01 5.52186906e-01 1.04033351e+00
4.91143197e-01 3.83714765e-01 1.22889018e+00 3.86990488e-01
1.14475405e+00 5.78058839e-01 1.07165265e+00 -3.56669389e-02
9.70038176e-02 -8.83199215e-01 6.66867137e-01 5.11768341e-01
3.15570563e-01 -2.56290793e-01 -8.89544904e-01 6.88902855e-01
-2.45423722e+00 -9.72561359e-01 1.89924449e-01 1.89694858e+00
4.71426427e-01 1.66355416e-01 -6.75359368e-02 -5.99172056e-01
9.91197973e-02 2.66687632e-01 -9.47702706e-01 -5.02464175e-01
1.28689468e-01 -1.25049978e-01 4.64183897e-01 7.60436714e-01
-9.96643960e-01 1.62087715e+00 6.59536266e+00 2.88215309e-01
-1.10199738e+00 -8.83395523e-02 -2.55931099e-03 1.68485805e-01
-8.01631361e-02 1.67538702e-01 -7.36623883e-01 8.53856280e-02
8.84419560e-01 1.23911850e-01 6.59082115e-01 1.20151627e+00
5.15729964e-01 -3.75508964e-01 -1.21650064e+00 9.24380898e-01
-2.26340204e-01 -1.04010093e+00 -1.82257190e-01 3.43539263e-03
6.96226895e-01 6.24953210e-01 2.67916083e-01 1.18376184e+00
1.09742141e+00 -9.04877186e-01 9.12884355e-01 5.63800037e-01
2.94262707e-01 -5.03864527e-01 4.01885033e-01 8.97636831e-01
-8.58366072e-01 -5.05936563e-01 -3.66597295e-01 -4.37420994e-01
2.29392648e-01 -5.20592332e-01 -1.16253626e+00 4.05877560e-01
7.04154134e-01 7.57746994e-01 -4.59871978e-01 8.57574463e-01
-3.72880310e-01 2.87758380e-01 1.78945646e-01 -3.57434005e-01
9.64477599e-01 -1.63038269e-01 2.33977318e-01 1.02430606e+00
3.12078327e-01 -4.04418081e-01 5.44704795e-01 7.33280361e-01
4.04110461e-01 -2.84515411e-01 -1.15472460e+00 3.65465462e-01
3.38128150e-01 9.27162051e-01 -4.66673881e-01 -1.57316864e-01
-3.86916399e-01 1.21889651e+00 6.99105382e-01 7.93463826e-01
-8.66900086e-01 -1.11486636e-01 6.01651251e-01 -3.95220935e-01
5.88762403e-01 -7.08152652e-01 1.24047667e-01 -1.07799947e+00
-2.12398469e-01 -9.54962313e-01 -1.31084010e-01 -1.19125795e+00
-1.03755438e+00 5.44143736e-01 -7.33742863e-02 -1.17398953e+00
-8.79058778e-01 -8.31039846e-01 -4.28108811e-01 6.53477073e-01
-2.09514093e+00 -1.22027576e+00 -4.09628898e-01 6.15051270e-01
1.06556106e+00 -2.02592254e-01 9.50264752e-01 -4.18744147e-01
-2.28051096e-01 3.05269599e-01 3.00861090e-01 3.63520766e-03
7.03006148e-01 -1.33991039e+00 9.93767619e-01 8.26571226e-01
1.16892438e-02 5.06175876e-01 8.56035650e-01 -4.34196919e-01
-1.48014247e+00 -1.10642052e+00 3.80945683e-01 -5.83455205e-01
6.38787031e-01 -3.24577302e-01 -7.14874923e-01 9.31439459e-01
3.23115230e-01 8.44909921e-02 4.47220681e-03 -3.91146131e-02
-3.37769747e-01 1.16986141e-01 -9.03486907e-01 1.08760047e+00
1.26571286e+00 -3.57266337e-01 -6.40504897e-01 9.98106152e-02
9.40207422e-01 -7.16797352e-01 -1.39502779e-01 4.23064947e-01
6.76928937e-01 -1.11365390e+00 9.71090257e-01 -6.60363197e-01
3.14373255e-01 -4.16081905e-01 -3.37992311e-01 -1.61790729e+00
-5.40943742e-01 -4.93811548e-01 -2.33517215e-01 4.61496711e-01
3.55427712e-01 -6.28543973e-01 6.99296415e-01 2.19689190e-01
-3.44230294e-01 -7.17538476e-01 -9.25030529e-01 -1.02022612e+00
1.68293014e-01 -5.06038725e-01 4.57683057e-01 5.35914063e-01
-9.37950760e-02 3.40737313e-01 -6.09102488e-01 1.95164502e-01
4.82464343e-01 -3.81467007e-02 1.28151822e+00 -6.69619381e-01
-5.55251539e-01 -2.03499362e-01 -1.29319623e-01 -1.68262339e+00
5.19382060e-01 -5.06967247e-01 7.62651205e-01 -2.20758820e+00
-4.58790511e-02 -3.50844949e-01 -2.76929259e-01 4.08111244e-01
-1.56436488e-01 -4.51427668e-01 3.30525994e-01 -1.40428366e-02
-9.67602968e-01 1.05206656e+00 1.66119194e+00 -1.96140990e-01
-4.63647723e-01 -1.83391348e-02 -5.45181453e-01 7.85740376e-01
8.32550228e-01 -1.51665807e-01 -9.39334393e-01 -6.50926948e-01
1.45701289e-01 2.31471360e-01 4.98637289e-01 -1.12618136e+00
3.16297889e-01 -6.27167344e-01 1.01696312e-01 -6.08065963e-01
4.23475742e-01 -6.77896321e-01 -5.98324656e-01 7.89363205e-01
-3.36169392e-01 3.54062140e-01 3.24289262e-01 1.17859411e+00
1.20795764e-01 -1.01888157e-01 6.08508229e-01 -5.22405684e-01
-1.58839452e+00 1.74288899e-01 -6.20343804e-01 1.36913925e-01
1.03119171e+00 -2.52954930e-01 -4.26070482e-01 -7.57992864e-01
-7.59443164e-01 8.23478580e-01 5.29631317e-01 8.92504930e-01
8.62647116e-01 -1.18724394e+00 -4.11431104e-01 2.31189132e-01
4.27929550e-01 1.37721971e-01 2.92854398e-01 3.12718123e-01
-5.93257666e-01 5.05213499e-01 -3.39510947e-01 -6.80783987e-01
-6.50776982e-01 9.39615488e-01 6.80980265e-01 -5.57282627e-01
-5.74593365e-01 7.51141608e-01 4.74985003e-01 -1.14393735e+00
3.07956994e-01 -6.05661392e-01 -7.28923753e-02 -6.55081153e-01
2.03053579e-01 7.31191561e-02 -4.77006644e-01 -4.51220214e-01
-1.43859908e-01 3.73141170e-01 2.37298682e-02 -5.88496983e-01
1.00000930e+00 -5.46709180e-01 5.55554509e-01 9.14411962e-01
5.15891731e-01 -5.28983295e-01 -1.92342234e+00 -2.79555291e-01
1.01347417e-01 -9.86990184e-02 -2.16277361e-01 -9.53902006e-01
-2.42022052e-01 9.27405417e-01 7.62855589e-01 -3.41983289e-01
6.58393443e-01 -3.28018218e-01 7.17477143e-01 1.18862069e+00
9.53448713e-01 -1.36707723e+00 5.08636534e-01 1.21556544e+00
9.30135250e-01 -1.72808588e+00 -5.15144110e-01 3.79387408e-01
-9.35542703e-01 8.50183308e-01 1.39309311e+00 -1.93813339e-01
3.66111755e-01 -2.98602097e-02 3.95876616e-01 1.11934990e-01
-1.09466982e+00 -4.88246620e-01 -3.27699602e-01 1.20180821e+00
-6.05970360e-02 -5.04428297e-02 2.29604870e-01 1.65145740e-01
-2.35575289e-01 2.67049909e-01 6.73119843e-01 1.18153024e+00
-7.03150332e-01 -1.04556429e+00 -1.21678658e-01 -2.55827993e-01
3.10592085e-01 9.93557870e-02 -2.07661483e-02 1.14311719e+00
-4.01306115e-02 1.10967743e+00 -1.29145071e-01 -4.19608712e-01
5.48965573e-01 -5.96884973e-02 7.07951486e-01 -6.66189492e-01
-7.61989579e-02 -7.30563775e-02 -1.12822585e-01 -9.06085432e-01
-3.23973417e-01 -4.30164307e-01 -1.67081726e+00 8.98170248e-02
9.55861062e-02 -2.19392627e-01 6.97259843e-01 8.74675989e-01
4.12762940e-01 7.52908409e-01 5.90524614e-01 -1.07474327e+00
-9.41968203e-01 -6.77410007e-01 -3.31495404e-01 1.11355141e-01
8.90127778e-01 -7.35744953e-01 1.66362301e-01 -2.50408947e-01] | [4.498002529144287, 0.6220483183860779] |
c1d52fe7-1a22-49a2-8a91-ec8d92a72dd9 | shift-memory-network-for-temporal-scene | 2202.08399 | null | https://arxiv.org/abs/2202.08399v1 | https://arxiv.org/pdf/2202.08399v1.pdf | Shift-Memory Network for Temporal Scene Segmentation | Semantic segmentation has achieved great accuracy in understanding spatial layout. For real-time tasks based on dynamic scenes, we extend semantic segmentation in temporal domain to enhance the spatial accuracy with motion. We utilize a shift-mode network over streaming input to ensure zero-latency output. For the data overlap under shifting network, this paper identifies repeated computation in fixed periods across network layers. To avoid this redundancy, we derive a Shift-Memory Network (SMN) from encoding-decoding baseline to reuse the network values without accuracy loss. Trained in patch-mode, the SMN extracts the network parameters for SMN to perform inference promptly in compact memory. We segment dynamic scenes from 1D scanning input and 2D video. The experiments of SMN achieve equivalent accuracy as shift-mode but in faster inference speeds and much smaller memory. This will facilitate semantic segmentation in real-time application on edge devices. | ['Jiang Yu Zheng', 'Guo Cheng'] | 2022-02-17 | null | null | null | null | ['scene-segmentation'] | ['computer-vision'] | [ 8.37717950e-01 -6.15686141e-02 -5.01396596e-01 -5.56045890e-01
-2.61927575e-01 -6.31277323e-01 -7.00319558e-02 -3.64777029e-01
-5.09860933e-01 3.74323010e-01 -5.83714060e-02 -8.00592899e-01
-2.97570438e-03 -9.55794632e-01 -8.22993219e-01 -3.69859278e-01
-1.08951643e-01 3.01205926e-02 9.71475899e-01 3.82341266e-01
5.26162028e-01 5.40571272e-01 -1.43343425e+00 9.06223238e-01
5.44789672e-01 1.12875998e+00 7.74498403e-01 1.03475583e+00
-4.00439948e-01 8.58909070e-01 -6.20841086e-01 3.13226998e-01
4.03008014e-01 -2.14381278e-01 -1.13115180e+00 -2.19188645e-01
5.49221277e-01 -6.23531044e-01 -6.27436280e-01 8.43676925e-01
4.68797147e-01 3.13607633e-01 1.15196332e-01 -1.12697911e+00
-3.16209525e-01 7.34166622e-01 -6.24831617e-01 4.25209552e-01
5.56813665e-02 2.26476073e-01 4.38304394e-01 -5.86490393e-01
7.21170962e-01 1.19112301e+00 7.49930322e-01 4.79777217e-01
-9.31506693e-01 -7.27592111e-01 3.71151686e-01 2.89813727e-01
-1.38308835e+00 -5.97416401e-01 5.57445586e-01 1.25104487e-01
1.58345103e+00 2.68330842e-01 7.13470221e-01 7.47170150e-01
-4.91294637e-02 1.21570706e+00 4.48444754e-01 -6.98029548e-02
3.45488787e-01 -6.70725822e-01 3.04959059e-01 5.30855536e-01
-1.29711300e-01 -2.66610861e-01 -8.37541342e-01 4.46839958e-01
1.31456459e+00 9.23603997e-02 -2.86815733e-01 2.89191622e-02
-1.12143970e+00 1.51169240e-01 5.61276436e-01 1.45711988e-01
-2.97371410e-02 8.17054331e-01 6.44923151e-01 1.54856026e-01
3.47929806e-01 9.47899148e-02 -6.97724044e-01 -5.06571770e-01
-1.58607817e+00 -5.71485795e-02 5.06480575e-01 1.52207577e+00
4.89631653e-01 1.44069582e-01 -2.01417923e-01 7.85165071e-01
-9.93177518e-02 5.70978045e-01 4.05716836e-01 -1.57385385e+00
6.55588269e-01 4.16180789e-01 -2.83068687e-01 -9.57828343e-01
-7.34609902e-01 -2.84938127e-01 -8.39809835e-01 -4.77491282e-02
3.10350686e-01 2.26306021e-01 -1.26924169e+00 1.49186635e+00
1.45780206e-01 7.04813361e-01 -1.30386814e-01 1.03298807e+00
7.05493867e-01 8.89307916e-01 4.28422317e-02 2.16432318e-01
1.22664070e+00 -1.42281139e+00 -6.16032362e-01 -4.76470441e-01
1.02459025e+00 -4.87007409e-01 1.12998855e+00 5.38684249e-01
-1.29311609e+00 -8.07142317e-01 -1.33644271e+00 -6.59850895e-01
-3.68948966e-01 8.88421535e-02 8.11812580e-01 5.52455783e-01
-1.45027971e+00 9.38008428e-01 -1.36244237e+00 -3.07774663e-01
7.41970301e-01 6.21735156e-01 1.81278318e-01 -8.02171454e-02
-8.91651750e-01 3.73438336e-02 6.41889751e-01 2.52036333e-01
-5.58864772e-01 -8.55215430e-01 -8.69436324e-01 1.56958774e-02
4.72328216e-01 -5.34407198e-01 1.44609845e+00 -1.21643519e+00
-1.42206061e+00 6.69258356e-01 -4.95101631e-01 -6.58207119e-01
9.15873870e-02 -2.52873957e-01 -3.77771407e-01 4.56062734e-01
9.65054408e-02 1.43682384e+00 6.23791993e-01 -8.24104130e-01
-6.98667288e-01 -3.72431785e-01 -7.45047862e-03 2.20689476e-01
-2.33856037e-01 -3.46986473e-01 -1.23518407e+00 -7.38004804e-01
5.91129363e-01 -9.23176885e-01 -2.76152771e-02 4.65140462e-01
-4.26586121e-01 9.69200581e-02 1.44184661e+00 -5.90226293e-01
1.67174041e+00 -2.39313388e+00 -4.08356905e-01 1.66781127e-01
8.65488723e-02 2.94024408e-01 -1.36251047e-01 -3.03856879e-01
7.48886392e-02 -4.91873501e-03 -3.51310700e-01 -3.43975455e-01
-2.84488827e-01 5.18235087e-01 -2.92024225e-01 3.62590738e-02
5.64678684e-02 1.13112891e+00 -7.94254482e-01 -6.31207466e-01
6.13207407e-02 1.19087346e-01 -8.16035032e-01 -2.05404535e-01
-3.23191106e-01 -5.67172095e-03 -3.66463244e-01 8.20426404e-01
9.11676884e-01 -5.35160601e-01 3.18098366e-01 -5.31242311e-01
-8.96018669e-02 4.23735797e-01 -1.10508859e+00 2.52559566e+00
-3.98021668e-01 1.06562304e+00 9.63836685e-02 -8.47164929e-01
4.91061330e-01 -1.84176385e-01 2.66772032e-01 -1.46105921e+00
-1.47572666e-01 -9.82017908e-03 -2.15058342e-01 -3.92795771e-01
9.93773580e-01 9.28373873e-01 -2.26668678e-02 6.10199749e-01
-4.07233268e-01 1.60698205e-01 2.36472823e-02 2.85914660e-01
1.23256731e+00 5.65817237e-01 -3.54072034e-01 -3.60710621e-01
-3.30775119e-02 3.23679507e-01 6.22155845e-01 9.11150932e-01
-2.87258178e-01 6.92381203e-01 3.47128153e-01 -4.83373672e-01
-1.11448157e+00 -1.14346182e+00 -1.54306054e-01 1.19600189e+00
6.59627080e-01 -4.62845773e-01 -9.69252527e-01 -3.65795672e-01
-2.48397380e-01 6.55256510e-01 -1.70491621e-01 -2.95855990e-03
-1.12153292e+00 -4.97085571e-01 8.82968128e-01 1.33966815e+00
1.16843462e+00 -8.44339132e-01 -1.22435808e+00 2.98152536e-01
-9.86872688e-02 -1.27525544e+00 -8.24371994e-01 3.69286209e-01
-1.25537002e+00 -8.25347722e-01 -4.84338462e-01 -9.93234277e-01
5.73778510e-01 6.61634564e-01 1.13552094e+00 1.56743065e-01
-3.29261243e-01 -1.10690035e-01 -6.66194037e-02 2.00662509e-01
1.39318183e-01 3.75203192e-01 -5.59333384e-01 -7.05781937e-01
4.38340634e-01 -6.15494430e-01 -1.07543385e+00 4.87619996e-01
-1.04991698e+00 7.98405468e-01 1.91659167e-01 4.76026952e-01
6.79886699e-01 1.73933342e-01 1.36261404e-01 -9.00222421e-01
1.20606191e-01 -1.89470947e-01 -5.92183709e-01 7.73672946e-03
-4.94870365e-01 1.30684406e-01 5.02444625e-01 -4.46003795e-01
-9.91394699e-01 4.73263636e-02 -2.29303408e-02 -5.12560427e-01
-1.08290292e-01 9.45115834e-02 -6.27427653e-04 2.20317245e-01
4.31767344e-01 1.58759356e-01 -2.36611217e-01 -2.76695848e-01
3.58591557e-01 6.10323846e-01 8.61936808e-01 -6.90374494e-01
-6.38941228e-02 6.44585192e-01 -8.76170918e-02 -6.29614234e-01
-6.05596781e-01 -3.22090775e-01 -6.51072204e-01 -1.22773126e-01
1.07093263e+00 -9.53328967e-01 -8.19331050e-01 6.02061272e-01
-1.37896633e+00 -1.10181546e+00 -6.54046610e-02 1.14967134e-02
-4.80972409e-01 1.18905194e-01 -9.83679950e-01 -5.02520740e-01
-2.41575301e-01 -1.26715791e+00 1.35594535e+00 2.92870373e-01
-3.29174966e-01 -7.99362421e-01 -5.79354525e-01 4.33851257e-02
4.46970284e-01 -2.62319535e-01 9.06869948e-01 -1.49352670e-01
-1.29159880e+00 5.67012906e-01 -9.05483305e-01 -6.94354028e-02
-2.65224069e-01 -1.08566135e-01 -1.01521790e+00 8.05279538e-02
-3.23158473e-01 1.54602244e-01 9.44841504e-01 8.42423856e-01
2.07180810e+00 -7.89906271e-03 -7.14037657e-01 1.18561661e+00
1.48675025e+00 7.97561884e-01 9.26300406e-01 2.84749150e-01
8.89975905e-01 2.40844473e-01 4.58953947e-01 1.67961910e-01
2.79914647e-01 6.09295607e-01 1.35064512e-01 -1.02069788e-01
-4.63071436e-01 -2.13298753e-01 1.16464838e-01 6.78841293e-01
3.69543403e-01 -5.90074360e-01 -1.10392201e+00 6.38093233e-01
-2.04663587e+00 -8.79048586e-01 -1.74439296e-01 2.00278473e+00
6.14097655e-01 5.69595397e-01 -9.37275961e-02 2.41942145e-02
6.59041703e-01 3.98625493e-01 -8.27326477e-01 -6.52267098e-01
-2.68776268e-01 3.23058009e-01 1.20457876e+00 4.70385313e-01
-9.30313468e-01 1.31162310e+00 6.51587725e+00 1.21929169e+00
-1.20446992e+00 1.91160306e-01 1.06154692e+00 -4.93456513e-01
-3.18477452e-01 -3.54111046e-02 -9.03254390e-01 5.90464532e-01
1.07438838e+00 5.89283228e-01 4.41271037e-01 7.61769593e-01
4.57000136e-01 -6.21266186e-01 -1.07379782e+00 1.26552856e+00
-1.21887989e-01 -1.67806363e+00 -6.99471608e-02 -2.05221519e-01
6.88351214e-01 1.19786430e-02 8.81943181e-02 -3.62984203e-02
-9.95413437e-02 -1.07357717e+00 7.86446273e-01 3.48720044e-01
1.12674665e+00 -7.70293653e-01 2.47686088e-01 1.69858217e-01
-1.51761198e+00 1.93304587e-02 -2.12463245e-01 -1.80349603e-01
4.86006409e-01 7.15585649e-01 -5.76532304e-01 2.93241609e-02
8.27319264e-01 8.71831596e-01 -4.48312044e-01 6.33272707e-01
2.81177700e-01 5.72773755e-01 -5.80485940e-01 5.73748862e-03
4.73397374e-01 -4.44636978e-02 1.76570732e-02 1.39158380e+00
3.35503727e-01 1.52819186e-01 -2.91308928e-02 8.36502135e-01
-1.51485472e-03 -4.57989365e-01 -5.21227539e-01 -6.25288859e-02
7.83465087e-01 7.10479677e-01 -1.48047793e+00 -6.13237917e-01
-2.70276517e-01 1.69866061e+00 2.49202196e-02 5.26783288e-01
-1.17801559e+00 -4.69994187e-01 6.35401845e-01 2.14662895e-01
3.13270539e-01 -6.36582613e-01 -1.03867567e+00 -6.31686807e-01
-3.65252271e-02 -2.94655144e-01 1.56346112e-01 -1.16389132e+00
-4.68402088e-01 3.15421343e-01 6.12733737e-02 -9.20523942e-01
6.12383783e-02 -8.71173620e-01 -2.58585453e-01 4.43400145e-01
-1.27885163e+00 -8.14105213e-01 -3.83694291e-01 5.30772209e-01
1.13539708e+00 3.93137962e-01 3.36942077e-01 7.13635921e-01
-7.54194021e-01 6.95246398e-01 -2.20536098e-01 2.33937025e-01
4.88073885e-01 -9.36262488e-01 1.01277971e+00 9.00229216e-01
-9.70169082e-02 5.05035758e-01 3.84120978e-02 -8.99274111e-01
-1.35200882e+00 -1.22752428e+00 3.57301831e-01 -9.61963311e-02
1.16516918e-01 -3.70979726e-01 -8.71336520e-01 6.56616032e-01
1.05632573e-01 4.67376178e-03 3.73773396e-01 -2.22752631e-01
-2.59312928e-01 2.75598635e-04 -8.64322722e-01 9.34828341e-01
1.96424866e+00 -7.01206386e-01 -1.31637389e-02 1.32270396e-01
1.25092185e+00 -9.28007782e-01 -4.53957379e-01 2.95720309e-01
5.66886008e-01 -9.16478455e-01 1.12955761e+00 -2.26544514e-01
5.37223756e-01 -4.69070852e-01 -1.35611475e-01 -4.71996218e-01
-1.49093062e-01 -4.85445827e-01 -2.10617483e-01 6.81461155e-01
4.21401352e-01 -2.81160086e-01 1.42721283e+00 6.44180298e-01
-5.14907360e-01 -8.11414599e-01 -8.11355412e-01 -6.73421204e-01
-4.72944081e-01 -1.10605741e+00 5.58254421e-01 6.35404706e-01
-2.56471336e-01 4.55868132e-02 -2.45071109e-03 2.73641586e-01
3.73597622e-01 2.85524070e-01 4.22459841e-01 -4.99298573e-01
-2.36675471e-01 -5.54009140e-01 -3.15770328e-01 -2.20633578e+00
4.45090607e-02 -7.81726480e-01 1.29904538e-01 -1.34033918e+00
-2.64538657e-02 -5.38786173e-01 -2.88193338e-02 4.80360836e-01
1.57423630e-01 5.36479890e-01 -2.90104132e-02 7.37452209e-02
-9.34932649e-01 -6.68142512e-02 1.38624167e+00 -6.87846690e-02
-4.87612277e-01 -4.44528699e-01 -2.73730367e-01 7.52585709e-01
7.24088311e-01 -3.34141463e-01 -9.74033177e-01 -1.15218461e+00
2.45641664e-01 7.34486431e-02 4.71092135e-01 -1.37366724e+00
7.45336771e-01 -6.48744553e-02 6.61338747e-01 -1.27938259e+00
3.12301546e-01 -6.99060082e-01 2.29762182e-01 3.76071393e-01
-2.89573848e-01 4.14936483e-01 5.63083172e-01 3.28539073e-01
2.85254233e-02 -1.98152199e-01 5.15478194e-01 -1.12475678e-01
-1.60658634e+00 2.96709239e-01 -2.91218281e-01 9.16150287e-02
7.57575631e-01 -1.06831074e+00 -4.90448385e-01 -1.38738841e-01
-8.05732310e-01 2.36898884e-01 4.70335782e-01 2.44214699e-01
8.01539719e-01 -1.17017484e+00 2.14298904e-01 3.65383178e-01
-3.66712034e-01 7.06014156e-01 7.10364223e-01 7.52029955e-01
-1.26497364e+00 6.14848733e-01 -7.94520136e-03 -1.01034391e+00
-9.58348095e-01 4.24617976e-01 3.29059571e-01 1.54992998e-01
-7.39390552e-01 9.82546985e-01 2.78842032e-01 1.22775055e-01
3.29036236e-01 -7.14410067e-01 5.66878319e-01 -2.73136824e-01
5.71834385e-01 4.65909839e-01 2.11612359e-02 4.39985394e-02
-2.61001796e-01 6.15165055e-01 1.29163280e-01 -2.91980952e-01
8.54369283e-01 -3.17084163e-01 2.33696308e-02 3.34755421e-01
1.46242678e+00 -4.56336051e-01 -1.74834526e+00 -7.85459951e-03
-4.49116379e-02 -3.91577482e-01 3.15080434e-01 -5.39281785e-01
-1.36286581e+00 8.81817937e-01 6.99747026e-01 -1.13928050e-01
1.51057231e+00 -3.64833742e-01 1.37419975e+00 3.08900833e-01
5.14502525e-01 -1.67909050e+00 -3.13433707e-02 6.39464080e-01
1.72180459e-01 -7.73006260e-01 -1.60236806e-01 -7.82230556e-01
-1.49589017e-01 1.25478089e+00 9.74279106e-01 3.80330719e-02
6.94968402e-01 1.03735065e+00 -1.23182677e-01 -1.50471598e-01
-5.36423802e-01 3.43152657e-02 -3.43700312e-02 6.84007108e-01
1.38609752e-01 -1.86114416e-01 6.34826124e-02 3.11899632e-01
-2.54784692e-02 1.74015105e-01 1.45643592e-01 1.14260113e+00
-4.07956392e-01 -7.88840890e-01 -9.19313505e-02 3.79034758e-01
-4.70615104e-02 -3.56428772e-01 1.37620211e-01 5.72508931e-01
2.68399026e-02 6.37891829e-01 9.79108453e-01 -3.31092447e-01
9.44870412e-02 6.89271092e-02 4.19006169e-01 -2.61200666e-01
-3.82612258e-01 1.93352908e-01 8.29566121e-02 -1.31243789e+00
-4.57061350e-01 -1.73721611e-01 -1.89663041e+00 -4.61336255e-01
4.78840731e-02 -5.05421877e-01 7.76846766e-01 7.10420370e-01
8.74273360e-01 1.06854975e+00 3.30230519e-02 -8.06025922e-01
1.81505755e-01 -3.84070933e-01 -3.82107586e-01 7.39901736e-02
8.69874656e-02 -1.24329142e-01 1.44382015e-01 2.36234128e-01] | [9.328585624694824, -0.03744969144463539] |
792d61df-7211-451f-88fa-3671109912bb | unsupervised-hyper-alignment-for-multilingual | null | null | https://openreview.net/forum?id=HJe62s09tX | https://openreview.net/pdf?id=HJe62s09tX | Unsupervised Hyper-alignment for Multilingual Word Embeddings | We consider the problem of aligning continuous word representations, learned in multiple languages, to a common space. It was recently shown that, in the case of two languages, it is possible to learn such a mapping without supervision. This paper extends this line of work to the problem of aligning multiple languages to a common space. A solution is to independently map all languages to a pivot language. Unfortunately, this degrades the quality of indirect word translation. We thus propose a novel formulation that ensures composable mappings, leading to better alignments. We evaluate our method by jointly aligning word vectors in eleven languages, showing consistent improvement with indirect mappings while maintaining competitive performance on direct word translation. | ['Armand Joulin', 'Marco Cuturi', 'Edouard Grave', 'Jean Alaux'] | null | null | null | null | iclr-2019-5 | ['multilingual-word-embeddings'] | ['methodology'] | [ 3.72669280e-01 6.79876655e-02 -5.22075593e-01 -3.59894127e-01
-1.22044885e+00 -9.59940493e-01 7.43924320e-01 -3.83546464e-02
-6.74564838e-01 9.36650455e-01 4.55128610e-01 -5.96925855e-01
2.63386369e-01 -7.24975824e-01 -7.21221149e-01 -3.69678319e-01
4.16496933e-01 7.40708292e-01 -1.49376601e-01 -5.38291574e-01
4.40912470e-02 4.11100209e-01 -8.25409889e-01 2.79805452e-01
7.48937786e-01 -6.95796087e-02 3.19273680e-01 5.55368662e-01
-5.83094597e-01 -5.76808341e-02 -4.49649692e-01 -5.62533975e-01
5.08494496e-01 -6.03589118e-01 -1.04582679e+00 5.67940958e-02
8.19431901e-01 2.33850613e-01 6.27868921e-02 1.03215098e+00
1.08397044e-01 -1.10460617e-01 5.00492692e-01 -1.15203881e+00
-8.99637043e-01 6.75576031e-01 -3.83452088e-01 1.05460502e-01
2.40818858e-01 -3.57924700e-01 1.50726092e+00 -8.41502845e-01
7.03969121e-01 1.08112121e+00 5.03742397e-01 5.80127537e-01
-1.66646731e+00 -6.42808437e-01 2.69007683e-01 -2.95959860e-01
-1.44973791e+00 -5.24740219e-01 3.97132099e-01 -4.27361459e-01
1.36833549e+00 1.04601957e-01 2.48628587e-01 1.07577753e+00
3.99516732e-01 3.08255851e-01 1.01052153e+00 -7.29985058e-01
-1.94480032e-01 2.68034548e-01 9.22923535e-02 5.60037136e-01
5.78005612e-01 -6.31110668e-02 -5.09700596e-01 -2.56760687e-01
7.92165518e-01 -2.69595891e-01 -1.48231030e-01 -4.82396126e-01
-1.48880601e+00 9.38489854e-01 4.08839971e-01 6.00288749e-01
-6.21028990e-02 1.55521229e-01 1.45053834e-01 6.64765000e-01
5.23989499e-01 8.73649776e-01 -4.17414337e-01 1.88563630e-01
-6.81452215e-01 2.35098302e-01 8.59796822e-01 1.05011463e+00
8.85142922e-01 7.85296485e-02 1.25780970e-01 6.30974054e-01
2.10130975e-01 7.51553118e-01 5.15920699e-01 -5.27118564e-01
9.52335060e-01 2.65502900e-01 2.38184169e-01 -7.15489984e-01
-2.62840003e-01 -5.51828086e-01 -5.00373065e-01 -4.02882621e-02
2.82745361e-01 -1.57646403e-01 -6.37558341e-01 2.24380350e+00
-4.11180891e-02 8.12249109e-02 3.38645428e-01 7.82060623e-01
1.95614576e-01 8.18377912e-01 -8.50600377e-02 -6.61642104e-02
1.10777020e+00 -1.05078661e+00 -6.44427180e-01 -5.85825741e-01
7.98558533e-01 -1.20839167e+00 1.03736794e+00 7.82877430e-02
-1.03975606e+00 -5.45907259e-01 -1.14375794e+00 -1.58396050e-01
-1.29709214e-01 -9.02846605e-02 6.23611212e-01 5.42758584e-01
-1.25512719e+00 5.18801868e-01 -8.63545835e-01 -7.00125515e-01
-2.12879285e-01 5.61066985e-01 -7.43052661e-01 9.60560814e-02
-1.24666607e+00 1.44207990e+00 2.22101837e-01 -3.36547673e-01
-3.44081044e-01 -3.69216621e-01 -9.22837436e-01 -1.43904150e-01
7.18439929e-03 -8.06390703e-01 1.27894568e+00 -1.14452899e+00
-1.43283761e+00 9.10391271e-01 -4.88041639e-01 -3.69761646e-01
1.52397215e-01 -3.03087801e-01 -3.94241929e-01 -4.79978532e-01
3.90268177e-01 6.44495964e-01 4.67805386e-01 -1.05751705e+00
-4.86410320e-01 -1.86421126e-01 1.08155496e-01 4.47383761e-01
-6.00123525e-01 1.47131354e-01 -5.04011989e-01 -6.90855861e-01
6.99678436e-02 -1.22396243e+00 -4.32424188e-01 -2.30833322e-01
-2.57933021e-01 -8.89245942e-02 -2.87140720e-03 -6.96786940e-01
1.01342285e+00 -1.97213590e+00 9.78337824e-01 2.72686601e-01
-6.91025108e-02 6.23063818e-02 -7.67898917e-01 6.43730581e-01
-2.63834357e-01 1.64062724e-01 -4.24598843e-01 -5.76434731e-01
-6.60180077e-02 6.84038997e-01 -6.96228623e-01 6.21746778e-01
3.74449730e-01 1.10629654e+00 -8.08920681e-01 -2.70483911e-01
-2.53725588e-01 2.91105628e-01 -6.43843889e-01 3.38104069e-01
9.46313664e-02 2.14560330e-01 -1.55541018e-01 1.55051902e-01
4.07613009e-01 -9.14568156e-02 6.79393053e-01 4.57642078e-02
-4.22587730e-02 6.72560096e-01 -1.11620367e+00 1.96842885e+00
-8.95329475e-01 5.86694121e-01 -1.32322013e-01 -1.00022912e+00
1.14044702e+00 4.94447947e-01 3.46250117e-01 -4.66506243e-01
-1.80984065e-01 4.68265593e-01 1.65656861e-02 1.29666612e-01
6.35292113e-01 -5.57278752e-01 -3.29334497e-01 7.09021628e-01
3.30495030e-01 -4.31402296e-01 -1.10890809e-02 -8.66877586e-02
6.52671695e-01 2.38576978e-01 5.14588356e-01 -4.73768383e-01
2.86094636e-01 8.22030231e-02 3.42379600e-01 4.98009384e-01
3.29874843e-01 6.08662724e-01 1.87068447e-01 -3.77823055e-01
-1.33186233e+00 -1.43292987e+00 6.22871779e-02 1.03307307e+00
3.55979204e-02 -5.98723471e-01 -6.90549552e-01 -5.06983161e-01
-6.46880344e-02 6.96044147e-01 -4.46943104e-01 -3.62987556e-02
-1.12101638e+00 -5.44988990e-01 6.92721665e-01 4.92134213e-01
-2.01565415e-01 -7.33490109e-01 -8.09198320e-02 4.34506446e-01
-2.30161190e-01 -1.08509421e+00 -9.04151678e-01 3.88219923e-01
-1.06054044e+00 -6.71573639e-01 -6.95037127e-01 -1.08761275e+00
7.93076813e-01 3.86453718e-01 1.22568917e+00 2.14584526e-02
2.44580120e-01 7.14377090e-02 7.12802336e-02 3.43141705e-03
-7.08370268e-01 7.31132627e-01 4.39699948e-01 -1.32716805e-01
3.01664710e-01 -4.31727797e-01 2.13047311e-01 1.64550170e-01
-1.00359750e+00 2.80275524e-01 5.12489438e-01 9.16698098e-01
5.95858693e-01 -8.12035620e-01 2.86315620e-01 -8.26978624e-01
7.56692350e-01 -5.32119572e-01 -7.42048383e-01 4.49084848e-01
-7.01737583e-01 5.41976929e-01 8.13118160e-01 -4.38146442e-01
-5.75185776e-01 1.10810280e-01 -9.91608799e-02 -6.11675307e-02
1.39654294e-01 4.44642872e-01 -8.24289694e-02 -7.63365701e-02
7.58794010e-01 1.25848159e-01 -2.16452545e-03 -5.68911791e-01
7.88131952e-01 6.06389105e-01 5.04648209e-01 -7.87336051e-01
1.18981493e+00 1.26268864e-01 -8.71759281e-02 -5.60076296e-01
-7.14516342e-01 -4.12076741e-01 -8.35337996e-01 5.67510247e-01
7.24370539e-01 -1.01053786e+00 2.79740155e-01 -1.89037681e-01
-1.59367573e+00 -2.33318418e-01 -2.53418088e-01 5.14657021e-01
-6.46742761e-01 2.09304914e-01 -3.02906245e-01 -7.23493770e-02
-2.21161410e-01 -1.12370384e+00 1.05325162e+00 -2.19389185e-01
-8.50783885e-01 -1.29390419e+00 7.20872104e-01 2.71689203e-02
3.65037680e-01 -3.27592283e-01 8.81823242e-01 -8.32087517e-01
-3.54875892e-01 6.54779896e-02 7.09719062e-02 2.75510937e-01
6.68438673e-01 -3.71174484e-01 -4.84704494e-01 -5.71957767e-01
-2.63360679e-01 -4.06476595e-02 6.09801590e-01 -1.13268055e-01
2.98432142e-01 -3.49158406e-01 -2.80776322e-01 7.74845719e-01
1.55833876e+00 -1.40115306e-01 2.72253603e-01 4.97732967e-01
7.72326589e-01 4.74679798e-01 2.53564656e-01 -2.82331616e-01
2.12714374e-01 1.26266456e+00 -9.48287323e-02 -3.40250254e-01
-2.09737808e-01 -2.88565397e-01 5.78021646e-01 1.45670378e+00
4.93417978e-02 -2.69556493e-01 -1.04130626e+00 8.77343059e-01
-1.76812077e+00 -5.45735538e-01 9.19345617e-02 2.15121818e+00
1.10118401e+00 -9.37873200e-02 -1.37617603e-01 -4.39584553e-01
5.58828354e-01 8.04742053e-02 -2.09675692e-02 -8.89246762e-01
-2.71938473e-01 7.24127531e-01 7.67510891e-01 1.27937675e+00
-8.33913326e-01 1.28359675e+00 7.36823750e+00 2.69628704e-01
-1.11123991e+00 4.48701411e-01 1.24834836e-01 -9.47138965e-02
-9.75202382e-01 2.19244555e-01 -1.00744808e+00 -4.24328074e-02
1.01798809e+00 -5.31203747e-01 6.62428200e-01 5.54438949e-01
-1.77797586e-01 5.97304165e-01 -1.32777011e+00 7.39595175e-01
2.64856160e-01 -1.30754781e+00 4.88705784e-01 -1.11924978e-02
1.09712732e+00 2.48635337e-01 -6.32035825e-03 1.08007111e-01
5.94892442e-01 -1.20755994e+00 4.57999974e-01 1.81891650e-01
9.02132034e-01 -7.41255164e-01 6.02497339e-01 1.55018657e-01
-1.12244391e+00 4.65697914e-01 -4.25103188e-01 -2.12737128e-01
3.63449752e-01 8.74786377e-02 -9.04078066e-01 7.92640686e-01
9.04511884e-02 6.48330808e-01 -4.60695833e-01 6.34612024e-01
-3.52154672e-01 4.17596936e-01 -2.08061114e-01 1.00962877e-01
3.67314190e-01 -4.81367260e-01 4.70183432e-01 1.33633912e+00
6.83578134e-01 -2.57401675e-01 4.48624492e-01 6.89631879e-01
-1.28469989e-01 4.31339085e-01 -1.03898001e+00 -1.42031103e-01
2.84273565e-01 9.44241703e-01 -4.12737966e-01 -3.14133078e-01
-7.02706397e-01 1.19785535e+00 7.13967800e-01 3.89175683e-01
-5.97307086e-01 -2.76461214e-01 1.21303248e+00 -1.89382099e-02
1.02793597e-01 -7.47698843e-01 -4.09799546e-01 -1.35436475e+00
6.20867731e-03 -9.86633539e-01 1.97383463e-01 -4.10122484e-01
-1.23749697e+00 9.14556563e-01 3.98955084e-02 -1.01880395e+00
-4.38204825e-01 -4.71859694e-01 -3.42326164e-01 1.48469853e+00
-1.30185556e+00 -1.06963181e+00 3.39569777e-01 3.73295605e-01
6.20288491e-01 -5.34368098e-01 1.20366514e+00 1.98019058e-01
-1.23723321e-01 7.79789507e-01 7.95027241e-02 -8.14430267e-02
9.86304224e-01 -1.27326739e+00 1.19384754e+00 9.60403025e-01
1.01422131e+00 9.77469087e-01 6.71282589e-01 -6.46142900e-01
-1.43249333e+00 -1.19101393e+00 1.59943318e+00 -6.19791687e-01
9.18930471e-01 -4.72082376e-01 -1.03291643e+00 1.07321239e+00
6.46786213e-01 -2.82542616e-01 8.31520319e-01 2.88679689e-01
-5.77290952e-01 4.75122184e-02 -5.53777814e-01 8.36253941e-01
1.05693007e+00 -8.66924882e-01 -9.90326345e-01 4.04211491e-01
9.97002423e-01 -3.33396405e-01 -6.64267361e-01 -4.76544052e-02
2.70714492e-01 -2.86178496e-02 9.30403352e-01 -1.22285306e+00
3.15957159e-01 -5.06063581e-01 -5.52156031e-01 -1.66487205e+00
-4.03258532e-01 -6.55316532e-01 2.62201399e-01 1.03151512e+00
7.69137263e-01 -8.20202112e-01 4.04063195e-01 1.54595524e-01
1.61447562e-02 -5.15446126e-01 -1.19323266e+00 -1.17753470e+00
6.81734562e-01 -2.54341632e-01 6.61515415e-01 1.07606220e+00
-2.90311757e-04 6.98042810e-01 -5.12060583e-01 2.78866082e-01
2.57618010e-01 2.91375250e-01 8.75113249e-01 -8.91422272e-01
-5.37403584e-01 -4.17092472e-01 -3.62024546e-01 -1.16752815e+00
5.58209896e-01 -1.50436270e+00 1.06183216e-01 -1.50339103e+00
1.36554062e-01 -4.43675339e-01 -4.75741625e-01 6.08664572e-01
-2.15819523e-01 4.83504981e-01 3.00977468e-01 1.88193351e-01
-5.20645306e-02 2.12490022e-01 9.29774940e-01 -2.11729288e-01
-7.71711394e-02 -3.55893552e-01 -9.52173889e-01 2.25894257e-01
9.72491860e-01 -7.29682565e-01 -3.74022007e-01 -9.85579550e-01
3.86699378e-01 -2.34080136e-01 -1.07880466e-01 -6.61391854e-01
8.05121288e-02 -4.84701782e-01 -8.29285532e-02 5.44907562e-02
2.47032553e-01 -5.86159527e-01 3.09115320e-01 3.98085117e-01
-5.48184991e-01 6.41224444e-01 3.45210552e-01 2.24392712e-01
-2.80322820e-01 -3.43146831e-01 6.53567433e-01 3.18792015e-02
-5.02688706e-01 9.55918133e-02 -1.31168723e-01 1.39644608e-01
8.11276615e-01 1.03503592e-01 4.45532501e-02 -1.56050220e-01
-5.61748981e-01 -1.19903728e-01 7.62428105e-01 7.89557219e-01
2.63506711e-01 -1.68605042e+00 -1.03638387e+00 3.21880519e-01
1.59542590e-01 -5.61930478e-01 -6.75558925e-01 4.61314231e-01
-4.20213938e-01 6.28664434e-01 -3.22362334e-01 -4.63667035e-01
-1.33060586e+00 3.59521896e-01 3.12854648e-01 -2.63890535e-01
-4.39526111e-01 6.66667998e-01 2.11066529e-01 -8.36840868e-01
-2.36351296e-01 -1.64601415e-01 4.18484583e-02 -9.53753367e-02
3.15713704e-01 -1.07173629e-01 1.38786986e-01 -9.61405754e-01
-5.10949314e-01 1.04615366e+00 -1.44591093e-01 -6.50809705e-01
1.31231010e+00 -6.95427507e-03 -3.12855810e-01 5.56183636e-01
1.33130586e+00 4.16209549e-01 -6.92593873e-01 -5.53598404e-01
2.64722735e-01 -4.14831579e-01 -4.25993115e-01 -1.82610169e-01
-8.36666286e-01 7.64239311e-01 4.06638116e-01 -1.83773950e-01
7.27837801e-01 6.44244850e-02 6.52653456e-01 7.38554955e-01
4.43381488e-01 -7.41058290e-01 -1.80512950e-01 7.59433866e-01
8.40027213e-01 -1.03255641e+00 5.48678823e-02 -1.44000575e-01
-5.49859524e-01 1.14518666e+00 4.69893903e-01 -3.47682804e-01
1.49518788e-01 3.54289502e-01 3.82932186e-01 2.63489097e-01
-7.73407340e-01 -2.37618610e-02 5.22715390e-01 5.32410681e-01
7.31016099e-01 3.24720204e-01 -3.40569258e-01 2.32553616e-01
-4.03127849e-01 -3.06788623e-01 5.25446653e-01 7.22073853e-01
-2.29791075e-01 -2.04724884e+00 -3.80692035e-01 -7.90414773e-03
-2.91476309e-01 -3.79986048e-01 -5.23568511e-01 7.56738305e-01
-2.92519834e-02 7.04853833e-01 2.58977532e-01 -1.00401573e-01
3.04112583e-01 2.98207402e-01 6.02264643e-01 -1.16638720e+00
-5.90185285e-01 -8.07572976e-02 2.24008888e-01 -4.57670510e-01
-2.35493422e-01 -5.55094779e-01 -1.03474927e+00 -1.44931689e-01
-2.23304078e-01 4.00949508e-01 6.87134802e-01 9.93889630e-01
1.61217704e-01 2.94558525e-01 4.71893817e-01 -5.65223515e-01
-6.93093598e-01 -5.44907391e-01 -2.15032101e-01 5.21577775e-01
5.11610150e-01 -3.37393492e-01 1.76682249e-02 8.05337876e-02] | [11.21617317199707, 10.194050788879395] |
441d7bbe-136e-4897-9ae2-4d51ae07f9df | pose-oriented-transformer-with-uncertainty | 2302.07408 | null | https://arxiv.org/abs/2302.07408v1 | https://arxiv.org/pdf/2302.07408v1.pdf | Pose-Oriented Transformer with Uncertainty-Guided Refinement for 2D-to-3D Human Pose Estimation | There has been a recent surge of interest in introducing transformers to 3D human pose estimation (HPE) due to their powerful capabilities in modeling long-term dependencies. However, existing transformer-based methods treat body joints as equally important inputs and ignore the prior knowledge of human skeleton topology in the self-attention mechanism. To tackle this issue, in this paper, we propose a Pose-Oriented Transformer (POT) with uncertainty guided refinement for 3D HPE. Specifically, we first develop novel pose-oriented self-attention mechanism and distance-related position embedding for POT to explicitly exploit the human skeleton topology. The pose-oriented self-attention mechanism explicitly models the topological interactions between body joints, whereas the distance-related position embedding encodes the distance of joints to the root joint to distinguish groups of joints with different difficulties in regression. Furthermore, we present an Uncertainty-Guided Refinement Network (UGRN) to refine pose predictions from POT, especially for the difficult joints, by considering the estimated uncertainty of each joint with uncertainty-guided sampling strategy and self-attention mechanism. Extensive experiments demonstrate that our method significantly outperforms the state-of-the-art methods with reduced model parameters on 3D HPE benchmarks such as Human3.6M and MPI-INF-3DHP | ['Hongkai Xiong', 'Junni Zou', 'Chenlin Li', 'Min Guo', 'Yu Sun', 'Botao Wang', 'Hongwei Zheng', 'Wenrui Dai', 'Bowen Shi', 'Han Li'] | 2023-02-15 | null | null | null | null | ['3d-human-pose-estimation'] | ['computer-vision'] | [-5.25410175e-01 4.72864211e-01 7.39881629e-03 -2.40162820e-01
-6.41003847e-01 1.96393073e-01 3.13098073e-01 -1.73862875e-01
-3.31844956e-01 5.79104185e-01 5.96044779e-01 2.99160838e-01
-2.53485858e-01 -6.93701386e-01 -1.02037370e+00 -4.91069913e-01
-2.46587664e-01 1.07713187e+00 4.87917185e-01 -5.03726184e-01
3.76789942e-02 3.23811710e-01 -1.16314256e+00 -3.66752326e-01
8.69397402e-01 9.03762579e-01 -8.36940855e-02 4.17960584e-01
3.40964556e-01 3.73327374e-01 -4.34503764e-01 -3.15797627e-01
1.90394431e-01 -1.27498940e-01 -9.53300357e-01 -3.47269103e-02
2.71135867e-01 -2.96585619e-01 -4.69437182e-01 6.92220449e-01
8.58818889e-01 2.74764329e-01 6.77726328e-01 -1.29790092e+00
-4.32531893e-01 3.74123514e-01 -8.04500580e-01 9.30476040e-02
3.41824710e-01 2.58026779e-01 1.02058935e+00 -9.30126607e-01
4.78617668e-01 1.61190593e+00 8.30717802e-01 5.47661901e-01
-1.15887761e+00 -4.60857868e-01 4.33403969e-01 3.21528196e-01
-1.47716689e+00 7.36040771e-02 1.07580805e+00 -4.49572861e-01
9.82586980e-01 -5.56458682e-02 9.19845343e-01 1.16667747e+00
4.95083749e-01 9.14381981e-01 6.44912779e-01 -9.42615196e-02
-8.82333145e-02 -5.91817558e-01 1.31415486e-01 9.50323403e-01
-6.65056109e-02 1.80458933e-01 -6.79446280e-01 -5.40696159e-02
1.18034053e+00 -2.16153413e-01 -1.01527609e-01 -6.66039169e-01
-1.36377370e+00 6.33941293e-01 8.44682515e-01 -2.78728753e-01
-4.22414452e-01 5.37523627e-01 3.90003085e-01 -2.40277231e-01
6.98690414e-01 3.35972399e-01 -8.04280519e-01 -3.29710022e-02
-5.87784171e-01 5.76543212e-01 5.17822981e-01 1.04794216e+00
4.61524814e-01 -2.00065449e-01 -3.73986840e-01 7.55941510e-01
7.14004159e-01 2.51558185e-01 2.84990460e-01 -8.98879647e-01
5.61305881e-01 6.57194138e-01 -1.14252038e-01 -1.04295003e+00
-8.06550920e-01 -6.17651105e-01 -1.02312756e+00 3.28586400e-02
2.19686776e-01 5.73470928e-02 -1.20892155e+00 1.86214542e+00
8.32330465e-01 1.44935727e-01 -5.05397975e-01 1.16646469e+00
6.39265537e-01 3.89080882e-01 3.16681713e-02 3.47405791e-01
1.49280262e+00 -1.32370865e+00 -3.67274493e-01 -2.04472244e-01
2.95229018e-01 -3.25073540e-01 9.46708918e-01 1.41980976e-01
-1.12361562e+00 -8.48873615e-01 -9.73728061e-01 -3.73732358e-01
5.12142554e-02 -8.02805275e-02 4.22212362e-01 1.14494234e-01
-6.08893156e-01 8.97278190e-01 -1.26861930e+00 -2.41759360e-01
1.97212294e-01 5.30003309e-01 -3.68908942e-01 1.99780256e-01
-1.48920691e+00 9.51177776e-01 1.16580233e-01 5.83342552e-01
-7.80373096e-01 -8.35829735e-01 -1.06634736e+00 -2.36091286e-01
5.41855931e-01 -1.23652029e+00 1.05644667e+00 -1.38315931e-01
-1.73801470e+00 5.25208652e-01 1.66227203e-02 -3.19839269e-01
8.45741510e-01 -8.69976401e-01 1.52711779e-01 -3.33604813e-02
2.55454153e-01 9.22670126e-01 7.45650768e-01 -9.62213814e-01
-2.36790389e-01 -7.63511002e-01 1.55255459e-02 5.01096129e-01
2.35326603e-01 -3.43216687e-01 -9.96053100e-01 -8.71056616e-01
2.98739254e-01 -1.13459444e+00 -5.67299128e-01 2.49140054e-01
-7.35237479e-01 -4.66966122e-01 4.32888508e-01 -7.88290799e-01
1.25689328e+00 -1.78113019e+00 9.86873031e-01 2.78484374e-01
3.74593377e-01 -1.14310287e-01 7.35997595e-03 9.68773514e-02
1.85975268e-01 7.25286007e-02 -2.92180002e-01 -6.50705040e-01
1.63091213e-01 4.87458885e-01 2.55970448e-01 4.61266160e-01
5.04943192e-01 1.03755176e+00 -8.37422371e-01 -8.27325583e-01
3.05451035e-01 8.96103859e-01 -9.30121064e-01 3.72753084e-01
-3.40891480e-01 7.92793036e-01 -8.72918487e-01 7.18360722e-01
3.99093151e-01 -3.01595688e-01 -1.70632631e-01 -5.67668617e-01
1.38907462e-01 3.12470376e-01 -1.15945911e+00 1.89322352e+00
-2.27179080e-01 -1.38284683e-01 -8.08226764e-02 -5.24426341e-01
7.46728063e-01 2.24747315e-01 5.79331279e-01 -4.26959276e-01
9.91391838e-02 -1.15948588e-01 -6.96399063e-02 -2.60819912e-01
3.52010131e-01 7.85374194e-02 -2.89902508e-01 1.04765579e-01
1.06718436e-01 -2.80555785e-02 -1.12977266e-01 4.66952994e-02
8.71356785e-01 7.47408211e-01 1.78540587e-01 -2.80132115e-01
5.14619768e-01 -3.49232674e-01 9.59998608e-01 1.96572140e-01
-2.98813730e-01 1.00202394e+00 4.67535973e-01 -5.32015324e-01
-8.52941155e-01 -1.19096327e+00 -2.21783966e-02 1.05121970e+00
2.53647536e-01 -5.48129618e-01 -6.49217010e-01 -7.83968627e-01
3.44678223e-01 1.49897680e-01 -1.00250053e+00 -1.83009893e-01
-9.71058369e-01 -6.67990446e-01 4.80055690e-01 1.05265737e+00
4.77834165e-01 -8.85428905e-01 -6.27352595e-01 3.41661155e-01
-3.66865188e-01 -7.17817068e-01 -7.97201991e-01 1.30867839e-01
-9.82361257e-01 -1.18802726e+00 -1.01984417e+00 -3.77027065e-01
6.13726616e-01 -4.50920194e-01 1.16557455e+00 9.27957222e-02
-3.54616165e-01 3.31647038e-01 -1.80203348e-01 -7.78022259e-02
2.40755245e-01 3.41807395e-01 1.51878878e-01 -3.22515488e-01
-7.85829052e-02 -5.54545403e-01 -7.72403121e-01 7.66895592e-01
-3.76473367e-01 1.91648543e-01 5.30300260e-01 9.31588769e-01
9.29798484e-01 -2.16565967e-01 2.69666910e-01 -5.28251290e-01
3.20875168e-01 -4.02842522e-01 -1.01740249e-01 1.48443490e-01
-3.49408954e-01 5.82973123e-01 1.78193569e-01 -3.71927142e-01
-8.93901527e-01 1.56154662e-01 -5.22575438e-01 -5.38848281e-01
4.54913527e-02 5.08096337e-01 -5.10215223e-01 1.23103067e-01
2.80848831e-01 -1.11556821e-01 9.83943492e-02 -8.36785138e-01
2.48970509e-01 -2.72138976e-04 6.32096708e-01 -1.14919746e+00
7.43884087e-01 2.44390190e-01 2.58828849e-01 -6.26279294e-01
-8.72051656e-01 -2.21617520e-01 -9.97940063e-01 -1.13164291e-01
9.67518866e-01 -8.99199426e-01 -7.02079058e-01 6.47991598e-01
-1.20229578e+00 -2.96321511e-01 -2.15090692e-01 4.05295134e-01
-7.82314718e-01 5.85881233e-01 -8.14329028e-01 -5.81938624e-01
-5.43493152e-01 -1.43150151e+00 1.62253284e+00 -1.92036964e-02
-4.41443801e-01 -7.13744283e-01 2.64116704e-01 4.49840695e-01
7.75274448e-03 5.26381552e-01 9.26701546e-01 -3.15900862e-01
-4.82694626e-01 -2.16848627e-02 -4.67439555e-02 7.44296163e-02
-1.27307177e-01 -1.74565405e-01 -4.47658062e-01 -7.12146237e-02
-3.73692334e-01 -2.22993597e-01 7.60375559e-01 5.83471656e-01
1.08702397e+00 3.01586203e-02 -4.24860328e-01 7.07876265e-01
8.70564699e-01 -5.00856102e-01 6.86207473e-01 2.78517067e-01
1.11891222e+00 7.03154087e-01 7.46720314e-01 5.02242744e-01
8.46673250e-01 1.03997743e+00 5.46747804e-01 5.79253910e-03
-1.29665628e-01 -6.07173443e-01 1.31606609e-01 8.83542895e-01
-6.18457198e-01 -5.20128980e-02 -7.61129141e-01 3.60163867e-01
-2.03204966e+00 -2.92258054e-01 -3.61233205e-02 2.10605383e+00
8.48429799e-01 2.79462278e-01 4.32267308e-01 5.09319119e-02
5.91759384e-01 1.89114422e-01 -7.06357241e-01 1.78188339e-01
3.93295646e-01 1.69259250e-01 2.85983503e-01 4.85843092e-01
-1.07601130e+00 9.61768627e-01 5.40505600e+00 6.74142659e-01
-5.19694924e-01 -3.75141129e-02 3.14864010e-01 -4.71296161e-03
-5.61199971e-02 -2.75368661e-01 -1.07449377e+00 4.66266513e-01
5.30051649e-01 1.88208222e-01 -7.74472803e-02 7.75181234e-01
1.45910736e-02 1.06506363e-01 -1.23284853e+00 5.78335464e-01
-1.67322591e-01 -7.04952598e-01 -8.40129424e-03 2.12591559e-01
2.62334645e-01 -9.19665620e-02 -8.80716294e-02 3.46759379e-01
2.07008898e-01 -8.77084851e-01 8.73916388e-01 7.33054578e-01
4.43287492e-01 -9.76925790e-01 6.76280975e-01 2.88405031e-01
-1.60274792e+00 1.67882115e-01 -2.44582772e-01 6.02226555e-02
4.88881946e-01 5.27780533e-01 -4.48499113e-01 7.76007533e-01
9.56528604e-01 7.59225190e-01 -5.76231241e-01 9.94705677e-01
-4.83486295e-01 2.33215302e-01 -6.17153406e-01 2.65678227e-01
-1.63158365e-02 1.44152895e-01 7.41140604e-01 7.16453671e-01
1.61528051e-01 1.15190789e-01 2.91020930e-01 8.34290266e-01
1.90534949e-01 -1.81599885e-01 2.70626806e-02 3.18618923e-01
1.97729096e-01 7.69978166e-01 -5.06449699e-01 8.93473104e-02
-3.56543884e-02 1.10965967e+00 4.56633151e-01 8.37868601e-02
-1.08214569e+00 -1.23196535e-01 8.61840785e-01 5.08389473e-01
1.83282286e-01 -5.00667691e-01 9.58516076e-03 -9.74453330e-01
8.17279220e-02 -7.20635772e-01 4.43889439e-01 -7.02437520e-01
-1.36435974e+00 5.17814875e-01 2.96672493e-01 -8.22736621e-01
-2.02574581e-01 -4.19890732e-01 -4.10386026e-01 7.31681824e-01
-1.10649133e+00 -1.45601618e+00 -1.56608954e-01 5.30466080e-01
6.72690928e-01 4.63871628e-01 5.62128782e-01 1.88929617e-01
-6.70367956e-01 6.11316264e-01 -6.13763571e-01 2.11009026e-01
7.77776957e-01 -1.32353449e+00 7.95719385e-01 3.33935738e-01
-2.12482616e-01 7.58166552e-01 6.64224684e-01 -1.09265614e+00
-1.36598265e+00 -1.00460076e+00 6.96260095e-01 -6.23625696e-01
3.65643531e-01 -3.22751522e-01 -8.80002439e-01 7.88896024e-01
-2.62880296e-01 2.74325699e-01 2.86346734e-01 4.81495678e-01
-2.62096167e-01 1.61516875e-01 -9.40427125e-01 5.74695826e-01
1.44359112e+00 -1.82939887e-01 -9.74470615e-01 6.81679249e-02
1.05878794e+00 -8.28252196e-01 -1.19739437e+00 9.19298232e-01
8.39807987e-01 -5.82004428e-01 1.49739909e+00 -5.09344220e-01
4.12201017e-01 -4.05038238e-01 4.24893275e-02 -1.27409577e+00
-5.74086428e-01 -4.46344256e-01 -4.85484093e-01 9.75887895e-01
1.66129231e-01 -2.77632654e-01 1.04504132e+00 5.99522591e-01
-2.71274954e-01 -1.16300654e+00 -1.11353207e+00 -5.89141309e-01
1.37772948e-01 -2.91844070e-01 5.62672973e-01 3.54334593e-01
-2.66657472e-01 2.81758845e-01 -8.33688259e-01 3.19149047e-01
8.02345634e-01 -2.83724189e-01 1.01434970e+00 -1.45361304e+00
-5.96884429e-01 -1.47479624e-01 -6.31509602e-01 -1.43837714e+00
9.76802260e-02 -3.84658396e-01 2.82529533e-01 -1.46874595e+00
8.92563257e-03 -2.50610590e-01 -2.59167314e-01 3.91208708e-01
-3.70254099e-01 7.76202977e-02 1.21391490e-01 -3.70832048e-02
-6.94891334e-01 1.05067623e+00 1.51763225e+00 -8.57515186e-02
-1.88324839e-01 1.78460732e-01 -3.48680347e-01 9.24288929e-01
3.14919591e-01 -4.24539328e-01 -2.34288156e-01 -4.08956379e-01
1.84988409e-01 1.26055747e-01 6.80071175e-01 -1.18499112e+00
2.68399060e-01 8.44255984e-02 5.90408802e-01 -1.11279273e+00
4.96347338e-01 -6.68115258e-01 1.94064409e-01 5.51339865e-01
-1.98090494e-01 6.22669905e-02 8.74569267e-03 9.23075438e-01
2.30204314e-02 2.29161561e-01 5.99719107e-01 -1.90506905e-01
-5.76241910e-01 6.72335267e-01 -3.84088941e-02 2.53618300e-01
6.62802339e-01 -1.39321581e-01 2.78128982e-01 -2.27932297e-02
-1.03925228e+00 6.97591901e-01 2.95553029e-01 5.53828120e-01
6.71923101e-01 -1.42533612e+00 -6.09598160e-01 2.72729278e-01
1.02524668e-01 6.46443963e-01 4.74936426e-01 8.37151647e-01
-3.63962889e-01 2.12772295e-01 -2.45173335e-01 -7.86591887e-01
-1.16260052e+00 3.38399142e-01 3.53149652e-01 -6.87379718e-01
-8.26631665e-01 1.16503692e+00 2.75873333e-01 -7.27882504e-01
4.00099576e-01 -5.54055989e-01 1.47748766e-02 -2.38840282e-01
1.16380692e-01 7.02305436e-01 -1.59948558e-01 -8.33001196e-01
-6.28956199e-01 1.00750422e+00 6.17302880e-02 6.75939098e-02
1.30748403e+00 -1.39151216e-01 2.99593620e-02 2.98178107e-01
1.03144813e+00 -3.09469581e-01 -1.63428569e+00 -2.32664302e-01
-1.17449202e-01 -3.70253623e-01 -1.25350386e-01 -6.57845974e-01
-1.19516611e+00 9.54529345e-01 3.94642949e-01 -5.92259645e-01
7.29150653e-01 1.28067672e-01 9.98462439e-01 1.15047015e-01
7.66387939e-01 -1.16910839e+00 2.99192548e-01 5.91386318e-01
1.24232531e+00 -1.16933239e+00 2.70814180e-01 -5.41047454e-01
-5.81764340e-01 7.58321047e-01 1.08973277e+00 -2.83532977e-01
8.46851945e-01 4.01226059e-02 -4.02600348e-01 -3.23500365e-01
-5.82017422e-01 -1.75441176e-01 9.23080564e-01 6.30582690e-01
5.78508496e-01 -1.49665341e-01 -2.95420855e-01 8.98436785e-01
-7.80772865e-02 -6.46079630e-02 -1.92290932e-01 9.13610280e-01
-3.68556261e-01 -1.20536315e+00 -4.71172214e-01 1.19246900e-01
-3.68187904e-01 1.23906314e-01 -2.51660883e-01 1.12592936e+00
9.10337716e-02 3.45304370e-01 -2.65244767e-02 -5.95706940e-01
5.53705692e-01 2.99866498e-02 8.00097704e-01 -6.34612679e-01
-6.09509230e-01 3.88772666e-01 1.05228033e-02 -8.97126615e-01
-2.13922471e-01 -4.81426716e-01 -1.47430897e+00 -3.45537439e-02
-3.27822387e-01 8.50705057e-02 3.09533596e-01 9.65513587e-01
5.26965678e-01 7.22799778e-01 1.55147567e-01 -1.45671427e+00
-6.08699083e-01 -9.92469192e-01 -4.86666292e-01 2.16582671e-01
1.68966189e-01 -1.28344774e+00 -9.95611995e-02 -4.28720385e-01] | [7.0936479568481445, -0.668847918510437] |
4ad47bcf-d1e6-4834-9f70-c73479c8cc37 | realy-rethinking-the-evaluation-of-3d-face | 2203.09729 | null | https://arxiv.org/abs/2203.09729v2 | https://arxiv.org/pdf/2203.09729v2.pdf | REALY: Rethinking the Evaluation of 3D Face Reconstruction | The evaluation of 3D face reconstruction results typically relies on a rigid shape alignment between the estimated 3D model and the ground-truth scan. We observe that aligning two shapes with different reference points can largely affect the evaluation results. This poses difficulties for precisely diagnosing and improving a 3D face reconstruction method. In this paper, we propose a novel evaluation approach with a new benchmark REALY, consists of 100 globally aligned face scans with accurate facial keypoints, high-quality region masks, and topology-consistent meshes. Our approach performs region-wise shape alignment and leads to more accurate, bidirectional correspondences during computing the shape errors. The fine-grained, region-wise evaluation results provide us detailed understandings about the performance of state-of-the-art 3D face reconstruction methods. For example, our experiments on single-image based reconstruction methods reveal that DECA performs the best on nose regions, while GANFit performs better on cheek regions. Besides, a new and high-quality 3DMM basis, HIFI3D++, is further derived using the same procedure as we construct REALY to align and retopologize several 3D face datasets. We will release REALY, HIFI3D++, and our new evaluation pipeline at https://realy3dface.com. | ['Linchao Bao', 'Chun Yuan', 'Xuefei Zhe', 'Zhengzhuo Xu', 'Di Kang', 'Jing Ren', 'Haoxian Zhang', 'Zenghao Chai'] | 2022-03-18 | null | null | null | null | ['3d-face-reconstruction', 'face-reconstruction'] | ['computer-vision', 'computer-vision'] | [-1.75401315e-01 -1.52641714e-01 1.35989159e-01 -5.77170670e-01
-7.72543430e-01 -5.61377287e-01 5.65685511e-01 -4.30767119e-01
2.78581113e-01 2.30848372e-01 1.10159710e-01 2.74898827e-01
7.44059086e-02 -7.09496737e-01 -6.99170768e-01 -5.03968477e-01
2.62360722e-01 1.09770930e+00 -8.09034929e-02 -1.19844206e-01
2.53744144e-02 1.14180720e+00 -1.62550187e+00 8.90204906e-02
4.76638466e-01 1.03831971e+00 -3.91451508e-01 2.40691796e-01
-1.58996105e-01 -2.79699117e-01 -1.69542342e-01 -7.89953351e-01
4.92175430e-01 -3.15012813e-01 -6.60358608e-01 2.16627598e-01
1.12143064e+00 -3.96619827e-01 1.52934819e-01 9.11447167e-01
6.83158457e-01 -2.77442396e-01 7.53059149e-01 -1.14394748e+00
-1.65643603e-01 -7.19276667e-02 -9.28444684e-01 -4.00905848e-01
6.22268021e-01 2.83649951e-01 4.74450946e-01 -1.47960126e+00
1.09959149e+00 1.62108016e+00 1.06898880e+00 6.07130110e-01
-1.44300997e+00 -9.56043661e-01 -8.99033174e-02 -2.41297409e-01
-1.80072629e+00 -9.43458498e-01 9.87764597e-01 -4.62434679e-01
6.99187458e-01 2.71915108e-01 7.82162249e-01 9.39562500e-01
4.37383689e-02 1.35633469e-01 1.17439950e+00 -1.21226385e-01
-1.15701698e-01 -2.91709751e-01 -4.62589294e-01 1.01993406e+00
2.81923339e-02 3.37184399e-01 -5.99350929e-01 -3.96979928e-01
1.16025591e+00 -1.15485698e-01 -2.86741853e-01 -6.52169049e-01
-9.68947947e-01 3.56496304e-01 2.80581892e-01 2.12019458e-01
-4.18735027e-01 -3.23217958e-02 4.96489033e-02 1.35955811e-01
8.61961305e-01 4.79792198e-03 -3.36565495e-01 1.28658824e-02
-1.15269566e+00 3.79503965e-01 4.98427600e-01 8.79209280e-01
9.16475236e-01 -3.84462141e-02 4.08547558e-02 9.06783402e-01
5.45281529e-01 8.02854300e-01 -1.59158275e-01 -1.26871431e+00
1.21879064e-01 5.28992712e-01 -1.68447182e-01 -1.21624231e+00
-3.78564686e-01 -3.33379537e-01 -7.81147957e-01 3.66227448e-01
3.89548659e-01 3.86278868e-01 -9.06916261e-01 1.46109629e+00
9.28008258e-01 5.12293279e-01 -4.24036801e-01 1.02940714e+00
1.06127322e+00 1.11071773e-01 -4.13792282e-01 -5.03644794e-02
1.18303287e+00 -5.67555487e-01 -5.28528631e-01 2.05981225e-01
1.65826842e-01 -1.18744254e+00 8.35218251e-01 2.27848127e-01
-1.43298173e+00 -5.62104642e-01 -5.78650832e-01 -3.16416062e-02
1.95418999e-01 1.62337832e-02 1.13936566e-01 4.13941443e-01
-1.22313452e+00 6.87937140e-01 -7.79103994e-01 -2.72188932e-01
7.57148743e-01 3.23567599e-01 -9.25357580e-01 -2.58600622e-01
-4.09679770e-01 8.76158655e-01 -2.78130740e-01 5.70689179e-02
-9.45287526e-01 -1.00210690e+00 -7.95239389e-01 -4.37587142e-01
1.19946688e-01 -7.65181482e-01 1.12192285e+00 -6.79398060e-01
-1.56975770e+00 1.67521358e+00 -5.03945291e-01 2.46005028e-01
7.16164052e-01 2.51835138e-01 -3.02059978e-01 9.09570456e-02
3.18325404e-03 6.67717099e-01 1.08897877e+00 -1.61361504e+00
7.95316994e-02 -8.92170727e-01 -5.56634724e-01 -1.07408769e-01
2.08473817e-01 1.29192844e-01 -7.88963556e-01 -6.08612657e-01
5.29247999e-01 -7.92351544e-01 1.70711219e-01 5.70761204e-01
-2.51439780e-01 -7.82080367e-02 7.52402067e-01 -8.29568386e-01
7.03528106e-01 -2.11877298e+00 3.19610804e-01 6.97150707e-01
3.47179592e-01 2.50051413e-02 -3.73484731e-01 -1.62436925e-02
-2.33009338e-01 4.35613208e-02 -3.81893784e-01 -8.60396028e-01
-1.16290303e-03 1.75980464e-01 2.78550331e-02 7.92026877e-01
1.57860726e-01 8.85258436e-01 -5.59957087e-01 -5.91245174e-01
1.85816333e-01 9.55969810e-01 -7.36060917e-01 2.00054660e-01
-1.03800081e-01 9.20062304e-01 -1.50316805e-01 1.21349323e+00
1.27432334e+00 -1.35280594e-01 9.82696190e-02 -5.55608332e-01
-3.99030112e-02 1.00275546e-01 -1.21439028e+00 2.06492162e+00
-4.11198616e-01 1.97368324e-01 6.34377301e-01 -3.90149087e-01
1.10349667e+00 2.74019390e-01 7.69912720e-01 -7.37421453e-01
2.52108783e-01 3.48480105e-01 -3.47515792e-01 -4.73991930e-02
1.22026309e-01 -1.72478855e-01 5.00548124e-01 5.24695992e-01
1.20156609e-01 -6.50827527e-01 -4.14283216e-01 -2.01967612e-01
5.14336348e-01 5.65602720e-01 8.61109123e-02 -3.62944722e-01
4.18444872e-01 -1.59101129e-01 5.85269392e-01 -8.43752101e-02
4.09004465e-02 1.12762654e+00 2.18198478e-01 -4.94909465e-01
-1.08366561e+00 -1.10628414e+00 -7.13782072e-01 3.78450811e-01
5.05492240e-02 -6.37901604e-01 -1.02420580e+00 -5.64830244e-01
3.58285338e-01 2.59472817e-01 -6.72944725e-01 2.20228598e-01
-6.39316380e-01 -4.33284938e-01 5.43640792e-01 2.14848116e-01
4.18656975e-01 -6.47360981e-01 -8.76994208e-02 -1.29313618e-01
-1.05558755e-02 -1.07964063e+00 -6.57796860e-01 -7.00873375e-01
-9.65861380e-01 -1.25777781e+00 -5.98543763e-01 -5.07219136e-01
1.03044009e+00 4.63956222e-02 1.59215176e+00 6.49405181e-01
-1.87604487e-01 5.11627078e-01 2.36831680e-02 -1.30213469e-01
-5.18304646e-01 -3.54779929e-01 1.96449280e-01 8.96125212e-02
-4.17022966e-02 -7.96695530e-01 -6.69008553e-01 8.46375883e-01
-5.60808241e-01 6.23080954e-02 1.36498185e-02 5.78551888e-01
1.19330466e+00 -2.13876918e-01 3.87492925e-02 -6.67835474e-01
1.32497057e-01 -2.35578030e-01 -7.43543744e-01 2.01580301e-01
-5.18522859e-01 -1.06277898e-01 1.27148807e-01 -8.95442441e-02
-8.27284694e-01 1.18025653e-01 -6.72479272e-01 -9.92186368e-01
-1.28291547e-01 -1.04388677e-01 -4.60157335e-01 -4.72892702e-01
5.99169016e-01 -5.30936895e-03 4.95740145e-01 -8.68369043e-01
2.10856944e-01 3.09605449e-01 6.43796265e-01 -7.92025506e-01
1.00156403e+00 7.99537659e-01 2.47486159e-01 -6.18872106e-01
-3.10770988e-01 -1.54446229e-01 -1.00705326e+00 -4.11541045e-01
4.68098283e-01 -8.22133780e-01 -7.27403343e-01 6.68623209e-01
-1.30704021e+00 -4.15018141e-01 -1.97606429e-01 1.06741697e-01
-5.89979768e-01 2.38097519e-01 -2.34217852e-01 -4.54277962e-01
-4.42369133e-01 -1.45760190e+00 1.80384421e+00 -1.60179645e-01
-2.90073991e-01 -8.00712883e-01 1.91566154e-01 4.04926628e-01
2.40392268e-01 5.45619011e-01 6.14849985e-01 1.03440203e-01
-4.49582309e-01 9.26749632e-02 -1.17478848e-01 1.38079926e-01
2.47108266e-01 4.55011487e-01 -1.15669632e+00 -3.02007794e-01
-1.50397584e-01 -3.39916605e-03 4.23163205e-01 1.67606562e-01
1.23868191e+00 -1.62178546e-01 -3.17492843e-01 1.07410872e+00
1.16010273e+00 -3.95223856e-01 6.75912559e-01 -2.41699666e-01
8.68799925e-01 6.81087852e-01 6.34020388e-01 4.22932625e-01
4.23797429e-01 1.17914927e+00 6.91844881e-01 -8.44562650e-02
-6.18613422e-01 -4.13519889e-01 1.37590334e-01 8.34214330e-01
-4.03781712e-01 4.15200382e-01 -9.69137549e-01 1.25461638e-01
-1.35468006e+00 -7.76949048e-01 -1.50385380e-01 2.35803533e+00
9.84522820e-01 -4.29386586e-01 1.67277351e-01 -4.87247333e-02
5.44428349e-01 -1.08974785e-01 -4.90469545e-01 -5.23457527e-02
-1.10550717e-01 6.14446461e-01 7.79473186e-02 7.27813303e-01
-6.56372190e-01 1.00961125e+00 6.31742716e+00 8.48551571e-01
-1.28445065e+00 2.82786459e-01 6.01968408e-01 -1.66319609e-01
-6.75455630e-01 -1.46256581e-01 -8.97654474e-01 3.20561737e-01
5.40162683e-01 2.40068823e-01 4.98594046e-01 6.77175879e-01
2.00108051e-01 1.03423826e-01 -1.10183632e+00 1.43224299e+00
2.37529546e-01 -1.53121972e+00 1.05989069e-01 2.92538494e-01
8.45678151e-01 9.65642091e-03 -7.09928051e-02 -3.78089368e-01
1.97707880e-02 -1.37221193e+00 8.69917989e-01 8.21012139e-01
1.41012001e+00 -6.96522236e-01 3.15322489e-01 1.23826504e-01
-1.27310205e+00 7.38193452e-01 -2.82967091e-01 4.98569757e-01
1.08587243e-01 7.85269499e-01 -8.05695534e-01 6.70459986e-01
8.73607099e-01 9.13314342e-01 -5.13789296e-01 8.27305555e-01
-9.58312526e-02 2.31015861e-01 -5.13044298e-01 7.90640831e-01
-5.10026157e-01 -4.72372293e-01 6.78153932e-01 7.22453833e-01
5.12996078e-01 1.31811783e-01 6.55020401e-03 1.10177088e+00
-3.72380316e-01 5.09715863e-02 -6.30882025e-01 3.54957849e-01
4.63158011e-01 1.37334311e+00 -5.51166236e-01 -4.09667306e-02
-2.54103303e-01 9.03198063e-01 2.50849336e-01 2.80406065e-02
-8.05608928e-01 6.16178393e-01 1.07684910e+00 7.15527296e-01
-3.76404747e-02 -3.52236241e-01 -2.78722614e-01 -1.09359384e+00
3.48816179e-02 -9.50960398e-01 7.11940899e-02 -8.38519573e-01
-1.26701927e+00 7.49082625e-01 6.57064021e-02 -1.13255858e+00
-2.07125291e-01 -4.59322304e-01 -4.53309715e-01 9.95606363e-01
-1.61962545e+00 -1.25443089e+00 -6.07667685e-01 8.15138400e-01
2.68010616e-01 -1.93910412e-02 9.25751090e-01 4.46989596e-01
-4.49725360e-01 8.16915393e-01 -3.13059479e-01 -3.66622955e-02
7.77818561e-01 -6.31936550e-01 6.56297624e-01 4.33178812e-01
1.55111715e-01 5.32250404e-01 4.20480132e-01 -7.17849731e-01
-1.71649396e+00 -1.14244258e+00 4.76327091e-01 -7.48948634e-01
2.87710335e-02 -2.19267890e-01 -1.01301944e+00 5.96908629e-01
-2.24407598e-01 4.35050726e-01 3.56206357e-01 -8.03641379e-02
-5.65933168e-01 -2.22263038e-01 -1.71090066e+00 3.98933738e-01
1.70519876e+00 -5.08340240e-01 -2.00557753e-01 2.47199297e-01
2.06039667e-01 -8.72754157e-01 -1.27955830e+00 7.06782699e-01
8.75340343e-01 -1.15837264e+00 1.19932902e+00 -2.25126907e-01
9.78120640e-02 -3.90396208e-01 -1.69314504e-01 -1.30837679e+00
-9.63026136e-02 -4.96981263e-01 -9.42986757e-02 1.22884238e+00
-2.37404630e-02 -6.71603203e-01 7.65118122e-01 3.88928562e-01
-8.55969116e-02 -7.25968778e-01 -1.16408587e+00 -6.94594502e-01
2.47443318e-02 -3.73259872e-01 1.25973153e+00 1.02189863e+00
-7.60161102e-01 -2.81396896e-01 5.81613928e-02 1.43861309e-01
8.51221144e-01 4.16800350e-01 1.18489814e+00 -1.49772882e+00
2.19826192e-01 -5.01336277e-01 -4.27315116e-01 -7.86504149e-01
5.08022964e-01 -1.11053276e+00 -2.35925958e-01 -1.09017491e+00
4.34467895e-03 -7.76192427e-01 4.19041723e-01 6.06240630e-01
3.44660193e-01 8.61746430e-01 -1.22316949e-01 2.56197274e-01
-3.56656574e-02 6.22907758e-01 1.66043544e+00 7.38408118e-02
7.66477138e-02 -2.36659050e-01 -2.92141944e-01 9.00747478e-01
5.06393731e-01 -3.40177506e-01 -5.35049150e-03 -6.15192175e-01
6.43529510e-03 -3.46098356e-02 6.54654503e-01 -7.21639216e-01
-1.38144165e-01 -1.17984533e-01 6.94370508e-01 -6.83329880e-01
7.08284914e-01 -1.00096941e+00 8.44443738e-01 1.19848371e-01
3.35977256e-01 2.48148128e-01 2.73341656e-01 -6.96759671e-02
-5.58827706e-02 1.69455722e-01 1.14685965e+00 -1.00190431e-01
-3.65296543e-01 8.61746550e-01 6.21141851e-01 2.49274690e-02
7.17417419e-01 -4.29480851e-01 1.07168362e-01 -1.93400219e-01
-6.49661064e-01 -6.92655593e-02 1.27220845e+00 3.44008148e-01
9.68254089e-01 -1.71631241e+00 -1.02538204e+00 8.45029354e-01
9.25691277e-02 3.19115609e-01 1.62300333e-01 9.41725671e-01
-6.40401423e-01 -2.98512895e-02 -2.60470510e-01 -1.06303620e+00
-1.63186455e+00 -5.02469726e-02 7.37657845e-01 2.39956930e-01
-6.37433350e-01 7.33062387e-01 6.65877983e-02 -9.48802769e-01
-4.53684069e-02 -2.29065016e-01 1.41784891e-01 -4.64958809e-02
4.21664327e-01 3.65146041e-01 5.62699258e-01 -1.33846045e+00
-5.77079177e-01 1.47409749e+00 4.69768822e-01 -2.30338722e-02
1.33359730e+00 4.00799885e-02 -5.13932347e-01 -1.46295220e-01
1.27306592e+00 3.47252250e-01 -1.29491973e+00 -1.56541541e-01
-4.92638797e-01 -9.52028155e-01 2.76267994e-02 -5.84684253e-01
-1.60491693e+00 7.29979813e-01 5.89712143e-01 -4.73725080e-01
1.10490763e+00 1.79099470e-01 6.59637094e-01 -4.03471589e-01
7.73957312e-01 -5.25390148e-01 -1.65961713e-01 3.12206864e-01
1.41216290e+00 -1.10176671e+00 1.82392478e-01 -7.82330811e-01
-1.47339523e-01 9.87396598e-01 5.81442773e-01 3.30399238e-02
8.34062994e-01 3.91349882e-01 -4.44358326e-02 -5.68629086e-01
-3.12119037e-01 6.91857189e-02 5.80963910e-01 5.85619748e-01
3.33601773e-01 -6.66717952e-03 1.77572712e-01 2.08657682e-01
-5.05779028e-01 -1.05269335e-01 -1.80927753e-01 4.38789874e-01
1.50354177e-01 -1.34134078e+00 -8.91796231e-01 1.47054449e-01
-1.28001943e-01 1.82531893e-01 -4.16065663e-01 7.49508679e-01
1.44181058e-01 4.64224726e-01 3.15627754e-01 -3.09037328e-01
5.43205261e-01 7.96077307e-03 1.04883897e+00 -4.86354172e-01
-4.33003962e-01 1.43793926e-01 -9.08237696e-02 -1.06718957e+00
-4.36664850e-01 -8.47782195e-01 -1.32384849e+00 -8.05382848e-01
-5.36525697e-02 -1.05216905e-01 9.02937412e-01 5.56122780e-01
8.05949509e-01 -3.48576754e-02 8.07611048e-01 -1.27542269e+00
-1.83910310e-01 -6.79496348e-01 -3.98096085e-01 5.63609600e-01
1.46502629e-01 -1.05857217e+00 -2.34614938e-01 -1.59629509e-01] | [13.196359634399414, 0.031733930110931396] |
c75ab7e2-dd7f-4cd9-a742-71283d358134 | properties-of-winning-tickets-on-skin-lesion | 2008.12141 | null | https://arxiv.org/abs/2008.12141v1 | https://arxiv.org/pdf/2008.12141v1.pdf | Properties Of Winning Tickets On Skin Lesion Classification | Skin cancer affects a large population every year -- automated skin cancer detection algorithms can thus greatly help clinicians. Prior efforts involving deep learning models have high detection accuracy. However, most of the models have a large number of parameters, with some works even using an ensemble of models to achieve good accuracy. In this paper, we investigate a recently proposed pruning technique called Lottery Ticket Hypothesis. We find that iterative pruning of the network resulted in improved accuracy, compared to that of the unpruned network, implying that -- the lottery ticket hypothesis can be applied to the problem of skin cancer detection and this hypothesis can result in a smaller network for inference. We also examine the accuracy across sub-groups -- created by gender and age -- and it was found that some sub-groups show a larger increase in accuracy than others. | ['Sherin Muckatira'] | 2020-08-25 | null | null | null | null | ['skin-lesion-classification'] | ['medical'] | [ 4.42668140e-01 5.75447381e-01 -5.88301063e-01 -1.93739280e-01
-3.43968034e-01 9.77029130e-02 3.08325738e-01 2.44964287e-01
-5.35384476e-01 1.02362382e+00 -3.32683437e-02 -4.29760039e-01
-4.62507546e-01 -1.14675403e+00 -1.84899271e-01 -7.62453496e-01
-1.01951703e-01 2.66276956e-01 3.23449820e-01 -1.08215578e-01
-7.35191777e-02 2.99305111e-01 -1.23699355e+00 3.16682756e-01
8.43604028e-01 5.46861529e-01 -4.44856882e-01 6.26996517e-01
9.13143381e-02 5.84980905e-01 -6.17189229e-01 -6.32302642e-01
2.08336905e-01 -4.10428524e-01 -7.39496946e-01 -3.60447258e-01
7.71391928e-01 -2.32559204e-01 -2.20771119e-01 7.65309930e-01
5.04896522e-01 -2.92421728e-01 5.67987680e-01 -9.13834274e-01
1.15083322e-01 8.07037830e-01 -6.70273840e-01 2.62214206e-02
1.07823305e-01 -7.06649125e-02 7.19674885e-01 -2.76306318e-03
7.98714936e-01 1.09326625e+00 1.50409937e+00 1.00150681e+00
-1.43553603e+00 -8.52021158e-01 -1.28697470e-01 1.66038312e-02
-1.46434081e+00 -2.37854362e-01 3.90144259e-01 -3.43285464e-02
1.14965343e+00 5.95879793e-01 1.27254319e+00 1.21749246e+00
1.80754974e-01 3.63312125e-01 1.30083990e+00 -6.74984157e-01
8.02899450e-02 3.07219833e-01 3.18023056e-01 1.01776874e+00
8.01779866e-01 1.90436423e-01 -2.44938612e-01 -3.40694845e-01
8.66793811e-01 -4.41403314e-02 7.41050020e-02 2.06259489e-01
-3.75322580e-01 9.48203564e-01 3.28130156e-01 3.33669603e-01
-1.60361126e-01 1.82161719e-01 3.96158159e-01 2.10252807e-01
6.52914286e-01 6.10143125e-01 -5.08070350e-01 2.07934409e-01
-1.22383392e+00 1.23759158e-01 1.15959179e+00 7.05043301e-02
3.00258309e-01 -3.36520731e-01 -9.42020565e-02 1.08841908e+00
-1.22105181e-01 -2.74022192e-01 2.46592894e-01 -6.59484863e-01
4.99727465e-02 6.99216664e-01 -3.30264091e-01 -8.77535582e-01
-9.10144389e-01 -4.74734753e-01 -1.17938948e+00 2.40178525e-01
7.45248795e-01 -5.01545429e-01 -1.24690425e+00 1.57724428e+00
2.40745068e-01 6.78805411e-02 -3.68044406e-01 3.07728708e-01
6.06236279e-01 -1.57996967e-01 4.55809116e-01 -1.28984526e-01
1.23508728e+00 -6.04933143e-01 -4.78829592e-01 -1.74790457e-01
6.43228948e-01 -2.30086803e-01 3.84507477e-01 7.36392736e-01
-9.75266933e-01 -1.25332087e-01 -9.37097251e-01 2.65307844e-01
-4.29003686e-01 -4.64640893e-02 1.01545048e+00 1.22534359e+00
-1.21128953e+00 9.52432275e-01 -7.85490692e-01 -1.03898227e+00
7.26048112e-01 7.00002015e-01 -3.75016659e-01 -2.65813589e-01
-9.16432083e-01 1.29530466e+00 4.18258399e-01 -2.16286957e-01
-3.61576140e-01 -6.03413820e-01 -4.84618902e-01 -1.01948366e-01
3.69565904e-01 -1.10571086e+00 8.22418392e-01 -9.63416576e-01
-1.00151730e+00 8.80576015e-01 -3.22415620e-01 -8.11310530e-01
5.74107885e-01 8.74614567e-02 -3.04535806e-01 2.89681792e-01
-2.04840869e-01 7.43500888e-01 6.41773760e-01 -9.55782890e-01
-6.51936889e-01 -5.52892804e-01 -1.43860430e-01 -8.85247514e-02
-6.53955936e-01 -1.41744405e-01 -2.47457206e-01 -4.59996492e-01
4.16094437e-03 -8.46020579e-01 -6.94360197e-01 1.17591396e-01
-7.82775164e-01 -3.38355601e-01 3.45064193e-01 -8.26988757e-01
1.56045222e+00 -1.58437514e+00 -4.25178081e-01 6.31377697e-01
6.71950281e-01 2.24134520e-01 -1.16980802e-02 3.65817517e-01
-3.59553337e-01 6.79603755e-01 -8.04408938e-02 -1.00309789e-01
-6.48847580e-01 4.00056511e-01 5.08177519e-01 2.79881477e-01
3.00859451e-01 5.61716080e-01 -5.39568722e-01 -8.70967746e-01
5.11428006e-02 5.77693105e-01 -4.09243852e-01 -6.13967955e-01
5.88441901e-02 -3.71296972e-01 -2.53444165e-01 8.41552079e-01
5.57734430e-01 -2.56798416e-01 4.91531909e-01 9.15536210e-02
1.93230748e-01 -1.15094466e-04 -5.13955176e-01 9.16757047e-01
-1.12108454e-01 7.02425003e-01 2.27521151e-01 -8.15572739e-01
8.52491200e-01 1.42206311e-01 5.89468718e-01 -3.11406821e-01
1.80828765e-01 -6.30307794e-02 5.75474858e-01 -4.39138174e-01
1.81584552e-01 -3.66299659e-01 4.37919617e-01 1.14113770e-01
-1.94393158e-01 8.54541883e-02 2.36782834e-01 8.11356381e-02
1.52116573e+00 -4.26051766e-01 4.37668204e-01 -1.20715804e-01
2.06860509e-02 3.27890545e-01 6.45987809e-01 1.05936301e+00
-2.07687110e-01 4.82999563e-01 8.67084444e-01 -5.55863857e-01
-9.78982389e-01 -7.15742946e-01 -5.29138207e-01 6.53834701e-01
-3.82209510e-01 -3.33534181e-01 -1.00473607e+00 -8.84441495e-01
3.64254147e-01 6.58527493e-01 -1.08415365e+00 -1.65049896e-01
-2.88628340e-01 -1.56361187e+00 1.07629812e+00 5.25751829e-01
4.85677302e-01 -6.71767235e-01 -4.00599658e-01 8.75410531e-03
1.69560224e-01 -7.29138017e-01 4.81112748e-01 3.38625848e-01
-1.12192118e+00 -1.41280162e+00 -6.11236751e-01 -6.57923996e-01
8.73425841e-01 -2.40530908e-01 1.03265691e+00 5.40083647e-01
-1.00743318e+00 7.12518468e-02 -1.89342976e-01 -7.76405036e-01
-5.85571826e-01 3.42751205e-01 -2.06616595e-01 -4.72143024e-01
8.58249962e-01 -3.63560349e-01 -6.03538215e-01 -9.60307196e-02
-5.89914680e-01 -3.64583507e-02 9.32149529e-01 9.78026628e-01
2.37474054e-01 4.96821851e-01 7.73688734e-01 -1.31340063e+00
7.86870360e-01 -3.87152553e-01 4.67359237e-02 2.02494368e-01
-8.73263180e-01 -2.27510884e-01 2.70748407e-01 -5.39633095e-01
-7.61671901e-01 1.18056059e-01 -1.36956841e-01 2.69663334e-02
-2.35215709e-01 4.19016808e-01 3.63250226e-01 -3.49425137e-01
7.90938973e-01 -2.72411376e-01 3.93270165e-01 -4.05438691e-01
-2.71609902e-01 4.02723253e-01 6.50176629e-02 -1.23241074e-01
4.29007679e-01 5.22976160e-01 5.52088857e-01 -1.03373098e+00
-7.26872146e-01 -2.90968090e-01 -5.59525967e-01 -2.14613631e-01
6.31094873e-01 -5.35479486e-01 -5.58350623e-01 5.11581302e-01
-7.02829242e-01 -3.42690468e-01 -1.61138386e-01 2.69842893e-01
1.23199128e-01 3.91691804e-01 -6.81950510e-01 -1.05348313e+00
-4.20013845e-01 -5.28228760e-01 6.10164404e-01 4.81674552e-01
-6.93648279e-01 -1.36312509e+00 3.72194052e-02 3.17696661e-01
2.14006811e-01 6.68400824e-01 1.07313919e+00 -8.80128384e-01
1.64135963e-01 -6.11117601e-01 -3.25748891e-01 2.58085400e-01
8.54306519e-02 4.44818825e-01 -1.01425254e+00 -2.54711241e-01
-4.81610298e-01 -1.78719148e-01 1.47079480e+00 7.47931063e-01
1.44971991e+00 -2.71615714e-01 -9.87212062e-01 3.41009527e-01
1.62785828e+00 -1.57853350e-01 8.38999152e-01 1.75315827e-01
3.34710211e-01 6.94641232e-01 2.09747151e-01 1.13951020e-01
6.36535063e-02 6.45261854e-02 4.18094188e-01 -4.98658836e-01
-2.59144247e-01 -2.37028152e-01 -1.91646039e-01 1.53191730e-01
-5.52153707e-01 -6.43788651e-02 -9.77454662e-01 7.07531333e-01
-1.51038778e+00 -1.10111105e+00 -2.83070743e-01 2.18753362e+00
9.96710896e-01 5.24825275e-01 5.29179633e-01 3.18114966e-01
7.01624334e-01 -1.69270232e-01 -5.05458891e-01 -5.47480166e-01
-5.95904477e-02 7.29758918e-01 7.58499146e-01 3.16985577e-01
-9.28606272e-01 6.43299878e-01 8.08049679e+00 9.00325656e-01
-9.97793913e-01 -1.58921614e-01 1.02441132e+00 -3.06677699e-01
1.43943995e-01 -1.73192769e-01 -9.48854506e-01 2.11173341e-01
8.54300737e-01 3.45215499e-02 -7.78592303e-02 6.61318958e-01
1.38638794e-01 -6.42141163e-01 -9.55348372e-01 4.25316662e-01
8.65475312e-02 -1.19705403e+00 -1.14773266e-01 3.67640674e-01
5.74376523e-01 -1.66296512e-01 -2.50905782e-01 9.91186053e-02
5.06912172e-01 -1.35876441e+00 -2.46338427e-01 7.22943306e-01
1.00777674e+00 -6.79381847e-01 8.33947778e-01 2.96906233e-01
-5.73866725e-01 -3.16312313e-01 -4.54808623e-01 -1.80890784e-01
-3.62200737e-01 8.85541320e-01 -1.41818202e+00 3.07183295e-01
7.74624705e-01 3.57599199e-01 -8.93708169e-01 1.47130382e+00
9.70285088e-02 8.87431443e-01 -6.82227910e-01 -5.84301293e-01
-1.16103441e-01 9.96469706e-02 2.30887607e-01 1.24198818e+00
5.18477336e-02 -8.68846253e-02 -2.24221885e-01 4.39016104e-01
3.05387020e-01 -1.38441861e-01 -5.27599514e-01 7.55824000e-02
4.28917378e-01 1.31781745e+00 -8.83400083e-01 -1.72424912e-01
-2.79526949e-01 4.78156626e-01 2.82480985e-01 2.80953068e-02
-4.65435535e-01 -5.12917280e-01 4.51249778e-01 5.26496589e-01
-1.84316300e-02 3.00566584e-01 -7.11337864e-01 -5.58626115e-01
-4.03659135e-01 -6.93296134e-01 6.26119435e-01 -3.40982974e-01
-1.49314380e+00 8.66206810e-02 -7.65130222e-02 -4.48076040e-01
-4.86053899e-02 -7.61719406e-01 -9.05160725e-01 7.05446422e-01
-1.08876395e+00 -1.20277619e+00 -4.55186814e-01 2.75519341e-01
1.01611421e-01 1.10283300e-01 1.04670751e+00 7.62000633e-03
-9.72458601e-01 1.03540850e+00 -9.56033692e-02 2.93734819e-01
6.02351546e-01 -1.19841099e+00 -1.41814174e-02 4.26483601e-01
-3.14317524e-01 6.67346001e-01 4.81954336e-01 -1.00651801e+00
-6.44369960e-01 -1.00577521e+00 8.53740454e-01 -3.95672530e-01
4.17188764e-01 -8.02877024e-02 -6.51773632e-01 3.88753742e-01
2.13819191e-01 -4.13400888e-01 1.10096383e+00 7.67664135e-01
-2.42811546e-01 -1.26570627e-01 -1.65027809e+00 6.92676425e-01
9.73794401e-01 -1.85441777e-01 -3.40961814e-01 3.08582723e-01
2.34981805e-01 7.58759454e-02 -9.88965392e-01 5.33976316e-01
1.02521122e+00 -1.19182014e+00 9.29120958e-01 -7.71713376e-01
5.46198606e-01 5.13866723e-01 4.62467968e-01 -1.03907800e+00
-3.58335823e-01 -1.75270155e-01 -1.58594139e-02 9.00911868e-01
6.15385771e-01 -7.48602450e-01 1.58160436e+00 9.08076048e-01
3.02925527e-01 -1.43432426e+00 -1.01330006e+00 -5.82236886e-01
2.67387331e-01 -4.59995680e-02 2.40328684e-01 9.95107472e-01
2.06537724e-01 2.22348005e-01 -8.46021101e-02 -1.68610692e-01
8.04293454e-01 -5.76909840e-01 4.65462357e-01 -1.60522926e+00
-1.70444995e-01 -5.96128821e-01 -3.68664563e-01 -4.96306196e-02
-2.23092020e-01 -6.48009419e-01 -5.18565953e-01 -1.74293804e+00
5.84413052e-01 -5.79841733e-01 -1.94348082e-01 1.11126506e+00
-2.49099731e-01 3.86173397e-01 -1.94438592e-01 -3.10683399e-01
-1.37766272e-01 -4.12735254e-01 1.07902372e+00 -8.07929505e-03
-2.07643807e-01 1.65406033e-01 -8.99402738e-01 8.58571649e-01
1.11733866e+00 -4.79085773e-01 -6.45117685e-02 6.69156164e-02
2.38860652e-01 -9.09965858e-02 5.72738647e-01 -1.06538498e+00
2.16337457e-01 -6.66774958e-02 9.82981086e-01 -5.27493358e-01
5.59548080e-01 -5.02468407e-01 -3.58386338e-02 1.03337371e+00
-2.20822275e-01 -5.24932325e-01 2.46573642e-01 4.76555854e-01
2.24273995e-01 -5.03460169e-01 8.46506834e-01 -3.91908169e-01
-2.75635451e-01 -2.30227001e-02 -4.75873888e-01 -3.06867212e-01
1.05782664e+00 -6.92027867e-01 -5.03841758e-01 -1.89101458e-01
-9.72533286e-01 1.14867516e-01 4.28346336e-01 -1.10725582e-01
3.09084326e-01 -7.19874978e-01 -7.26512790e-01 1.05013631e-01
-3.08843583e-01 -1.19455338e-01 1.37165383e-01 9.57304478e-01
-6.52743518e-01 1.75738990e-01 -2.30954051e-01 -2.58146524e-01
-1.80760586e+00 6.66930377e-02 6.22617126e-01 -5.60873449e-01
-3.74700844e-01 1.14593339e+00 -4.67450052e-01 -2.37420201e-01
2.62263209e-01 -1.57229394e-01 -2.39658356e-01 1.90216273e-01
3.15798283e-01 8.27047527e-01 -1.73240259e-01 5.43758832e-02
-1.57199606e-01 4.72998232e-01 -3.96654814e-01 4.24009651e-01
1.41478002e+00 4.01509464e-01 -3.99828523e-01 -2.32103821e-02
8.77795935e-01 -1.12803675e-01 -6.28148735e-01 2.98794746e-01
1.11999139e-02 -2.62519509e-01 7.87063912e-02 -1.34349561e+00
-1.02009022e+00 5.60090661e-01 6.16104782e-01 4.85005558e-01
1.16623974e+00 -1.37641206e-01 4.35503840e-01 3.85099113e-01
3.15936983e-01 -1.37093341e+00 -4.20058250e-01 9.37811844e-03
2.69089669e-01 -1.12035501e+00 6.87113881e-01 -7.26986587e-01
-2.05737501e-01 1.12283027e+00 8.89697433e-01 -2.13476926e-01
6.80098474e-01 3.14398527e-01 -2.24627063e-01 -1.89702019e-01
-8.80568981e-01 -1.46515697e-01 -4.73215096e-02 9.32904482e-01
3.87456775e-01 2.02496067e-01 -6.57952249e-01 6.98863506e-01
-2.82002002e-01 3.29767734e-01 4.24019635e-01 5.61879039e-01
-3.04284275e-01 -1.22796905e+00 -3.93627703e-01 1.57264149e+00
-6.09887183e-01 -2.09560804e-02 -1.16798234e+00 1.25785923e+00
5.43841183e-01 8.89775157e-01 2.30106801e-01 -6.31827056e-01
-9.85547667e-04 1.30756915e-01 7.88224995e-01 -4.64713812e-01
-9.63829160e-01 -1.24613032e-01 8.00857902e-01 -4.29651678e-01
-3.23556751e-01 -5.76374352e-01 -8.44233990e-01 -4.21601236e-01
-6.96766198e-01 -1.31394058e-01 5.09799302e-01 5.74133456e-01
3.49161588e-02 4.11468148e-01 7.98463598e-02 -2.25623414e-01
-5.35666406e-01 -1.12258410e+00 -8.06949914e-01 1.05465561e-01
1.31510839e-01 -3.07862908e-01 -4.08073872e-01 -5.02832949e-01] | [15.621345520019531, -2.937554359436035] |
57b3e88f-748f-4f02-a6ab-decb2fffa784 | ost-efficient-one-stream-network-for-3d | 2210.08518 | null | https://arxiv.org/abs/2210.08518v1 | https://arxiv.org/pdf/2210.08518v1.pdf | OST: Efficient One-stream Network for 3D Single Object Tracking in Point Clouds | Although recent Siamese network-based trackers have achieved impressive perceptual accuracy for single object tracking in LiDAR point clouds, they advance with some heavy correlation operations on relation modeling and overlook the inherent merit of arbitrariness compared to multiple object tracking. In this work, we propose a radically novel one-stream network with the strength of the Transformer encoding, which avoids the correlation operations occurring in previous Siamese network, thus considerably reducing the computational effort. In particular, the proposed method mainly consists of a Template-aware Transformer Module (TTM) and a Multi-scale Feature Aggregation (MFA) module capable of fusing spatial and semantic information. The TTM stitches the specified template and the search region together and leverages an attention mechanism to establish the information flow, breaking the previous pattern of independent \textit{extraction-and-correlation}. As a result, this module makes it possible to directly generate template-aware features that are suitable for the arbitrary and continuously changing nature of the target, enabling the model to deal with unseen categories. In addition, the MFA is proposed to make spatial and semantic information complementary to each other, which is characterized by reverse directional feature propagation that aggregates information from shallow to deep layers. Extensive experiments on KITTI and nuScenes demonstrate that our method has achieved considerable performance not only for class-specific tracking but also for class-agnostic tracking with less computation and higher efficiency. | ['Xiuping Liu', 'Jian Liu', 'Shengjing Tian', 'Yinan Han', 'Xiantong Zhao'] | 2022-10-16 | null | null | null | null | ['3d-single-object-tracking'] | ['computer-vision'] | [ 7.82812908e-02 -3.56862903e-01 -1.76444381e-01 -3.40027273e-01
-2.92198956e-01 -6.05136514e-01 7.29200125e-01 -1.86343268e-02
-4.74006295e-01 3.84573877e-01 -4.43384737e-01 -8.65551308e-02
-4.26318854e-01 -8.10856283e-01 -6.80568635e-01 -7.47354150e-01
-8.21904317e-02 5.54456115e-01 8.47940743e-01 -9.42050517e-02
5.26416115e-02 9.58586693e-01 -1.84542072e+00 -1.18473187e-01
9.20100033e-01 1.53429973e+00 2.75552392e-01 1.78814322e-01
-2.37043291e-01 3.83190393e-01 -5.31919479e-01 -4.27953482e-01
4.85440582e-01 1.04147896e-01 -3.42811309e-02 2.05617938e-02
9.05671000e-01 -1.71074316e-01 -3.52007061e-01 1.01218534e+00
4.71945256e-01 -2.07190868e-02 3.61374885e-01 -1.49257922e+00
-2.65658885e-01 3.32015544e-01 -6.90664649e-01 2.13989288e-01
-1.81665510e-01 5.07619739e-01 8.62929642e-01 -9.12555158e-01
3.24947685e-01 1.16731596e+00 8.36911321e-01 4.21278119e-01
-1.08642876e+00 -1.05143654e+00 2.95837969e-01 1.00138985e-01
-1.41313291e+00 -2.72474527e-01 8.01229060e-01 -3.35082650e-01
5.39040506e-01 1.32024646e-01 9.91822124e-01 9.79322791e-01
9.29481909e-02 7.04803228e-01 7.82647610e-01 2.47757167e-01
1.91039205e-01 2.72132107e-03 -3.94608863e-02 6.75288558e-01
7.24658489e-01 4.76269633e-01 -7.80363619e-01 1.56557396e-01
7.87589669e-01 1.19143516e-01 -1.32634595e-01 -9.00409997e-01
-1.32493031e+00 5.82064509e-01 8.92808259e-01 1.52272478e-01
-3.68304342e-01 3.07968199e-01 2.68659294e-01 -7.49283358e-02
1.27888411e-01 2.12921217e-01 -3.71954888e-01 1.77909106e-01
-1.22374797e+00 1.73127875e-01 4.97006238e-01 1.12345433e+00
9.67919290e-01 2.84379542e-01 -2.76409656e-01 3.82049739e-01
4.55595255e-01 8.50583136e-01 9.64181870e-02 -7.42539227e-01
5.05046248e-01 8.07842255e-01 3.90012525e-02 -1.09328759e+00
-4.68296021e-01 -1.19760704e+00 -7.62146533e-01 4.19829398e-01
3.92575175e-01 9.50404927e-02 -9.46126163e-01 1.79830706e+00
6.58775926e-01 3.46661568e-01 -1.98294535e-01 9.90553856e-01
5.78444302e-01 1.43725216e-01 8.88534114e-02 1.50774807e-01
1.45472181e+00 -7.86751151e-01 -4.01814073e-01 -2.62887508e-01
2.14166641e-01 -5.10531247e-01 7.49884903e-01 1.24343164e-01
-9.26948905e-01 -7.25756466e-01 -1.24923420e+00 8.06361884e-02
-4.71653491e-01 9.52879116e-02 8.07126641e-01 4.89748627e-01
-9.13417101e-01 5.48783004e-01 -8.44524860e-01 -3.02604288e-01
6.95968568e-01 6.56524658e-01 -2.24992454e-01 1.24318734e-01
-9.00908589e-01 6.70961440e-01 4.97979730e-01 4.57492948e-01
-7.41639912e-01 -9.95926559e-01 -8.27013552e-01 2.27255955e-01
6.57997489e-01 -8.12837660e-01 9.39681351e-01 -6.57725692e-01
-1.34687996e+00 2.41206467e-01 -1.42212927e-01 -5.98786592e-01
6.27173483e-01 -8.73722434e-02 -4.46130008e-01 1.61447927e-01
1.66612133e-01 9.33141470e-01 1.13377786e+00 -1.25081921e+00
-1.07563651e+00 -5.29322505e-01 -5.22255450e-02 1.61902323e-01
-4.08560812e-01 -2.48382732e-01 -5.27002692e-01 -6.67135298e-01
3.04057509e-01 -8.22816789e-01 -1.12062626e-01 6.66740656e-01
-2.25931361e-01 -2.61511803e-01 1.31483769e+00 1.09529443e-01
9.43825960e-01 -2.26710582e+00 -8.31705332e-02 3.66887420e-01
5.00516355e-01 4.98780191e-01 -1.79241970e-01 1.65156066e-01
2.23974630e-01 -2.28062540e-01 -1.08595006e-01 -5.12224197e-01
3.12791020e-02 3.11168790e-01 -2.91117162e-01 5.14770329e-01
5.85586786e-01 1.17162251e+00 -1.01892364e+00 -7.09259272e-01
3.43350202e-01 6.05068445e-01 -4.37595874e-01 -1.83741465e-01
-3.23249340e-01 3.84633571e-01 -6.96770072e-01 9.29067135e-01
9.78798807e-01 -3.35529596e-01 -8.41309354e-02 -4.65354979e-01
-4.88348961e-01 1.09635353e-01 -1.35731959e+00 1.63485277e+00
-2.60680169e-01 2.88366765e-01 2.40075946e-01 -5.51791251e-01
9.87781167e-01 -6.03653863e-02 5.74266076e-01 -6.20530248e-01
2.29016855e-01 3.01201552e-01 1.60521850e-01 5.61918989e-02
5.41861832e-01 5.50934710e-02 -5.72341541e-03 -3.57951783e-02
8.73977765e-02 3.78522873e-02 2.74991900e-01 1.88949808e-01
9.12720442e-01 3.12514454e-01 -1.54222205e-01 -2.65334755e-01
7.08720267e-01 3.65391783e-02 8.53869915e-01 6.86264157e-01
-3.47408175e-01 3.44480217e-01 -3.73472832e-02 -3.28518480e-01
-6.24435723e-01 -1.34636843e+00 -2.45200038e-01 6.63105488e-01
5.26471078e-01 -3.51082027e-01 -1.67221621e-01 -7.72136748e-01
4.55806136e-01 2.94776469e-01 -4.86093760e-01 -1.61731556e-01
-6.01830423e-01 -3.39238018e-01 5.80988765e-01 7.90476620e-01
7.57707179e-01 -7.14091182e-01 -8.14088404e-01 1.89375594e-01
1.39430165e-01 -1.32349658e+00 -4.62354928e-01 3.29894394e-01
-8.81467402e-01 -9.90153134e-01 -2.44020551e-01 -3.98295373e-01
5.10907471e-01 5.70340574e-01 6.77795649e-01 6.82455674e-02
-1.59561902e-01 1.54325515e-01 -1.05902158e-01 -4.27455366e-01
2.29524732e-01 7.59635866e-02 1.68404207e-01 3.40188265e-01
3.85645002e-01 -7.72936225e-01 -6.52920365e-01 3.80289286e-01
-8.00342560e-01 -2.00533912e-01 8.51649165e-01 6.91498995e-01
5.95992386e-01 4.35855538e-02 3.30221772e-01 -2.53376514e-01
9.65429470e-02 -3.33637357e-01 -1.07617080e+00 2.99982335e-02
-6.51231587e-01 -3.83172706e-02 7.42295861e-01 -4.88905549e-01
-8.73900294e-01 3.56714785e-01 2.70881385e-01 -8.39414179e-01
7.09556863e-02 4.90877926e-02 -2.82684773e-01 -5.53378105e-01
3.26907665e-01 2.72346884e-01 1.20480768e-01 -4.18649942e-01
3.47796828e-01 1.89929798e-01 7.87979066e-01 -4.38506663e-01
1.66100812e+00 8.43086123e-01 3.91873747e-01 -4.85787541e-01
-8.12568486e-01 -4.66935039e-01 -8.23228002e-01 -4.58939224e-01
7.99275815e-01 -1.03652155e+00 -1.07579851e+00 4.58743811e-01
-1.03300798e+00 2.20572248e-01 -6.97515726e-01 5.65954983e-01
-2.08286911e-01 2.43067399e-01 -2.06845522e-01 -6.46401227e-01
-1.26689941e-01 -1.05607474e+00 1.28556478e+00 5.38837492e-01
2.28778943e-01 -7.00274229e-01 -3.60459954e-01 1.84309766e-01
6.46370053e-01 9.92077142e-02 4.39660370e-01 -6.33056998e-01
-1.35171056e+00 -1.74787432e-01 -6.17547452e-01 1.23750150e-01
4.35365103e-02 2.47010365e-02 -1.02822244e+00 -4.97190058e-01
-1.23503007e-01 -4.32115868e-02 8.47619534e-01 8.57767239e-02
1.02478421e+00 7.98019841e-02 -4.77189571e-01 8.16774666e-01
1.43459857e+00 4.40156944e-02 3.19148034e-01 2.27147385e-01
8.39800239e-01 3.61502826e-01 7.11197555e-01 3.23355526e-01
4.72712606e-01 8.40820193e-01 9.07288730e-01 -1.13470614e-01
-4.33341950e-01 -3.68575037e-01 2.53461957e-01 6.02602541e-01
9.62814465e-02 -6.92411214e-02 -6.01909518e-01 5.37374377e-01
-1.73425627e+00 -9.59173143e-01 -1.16153151e-01 2.26550889e+00
4.17865425e-01 3.16948354e-01 -1.32616028e-01 -2.33951807e-02
5.92536867e-01 3.65284473e-01 -7.76110291e-01 2.28702277e-01
-2.00711176e-01 2.70891935e-01 7.11722195e-01 1.55509397e-01
-1.08753657e+00 1.00897622e+00 5.06893206e+00 1.02637327e+00
-1.24564195e+00 3.74981239e-02 -2.97262043e-01 -1.98507607e-01
-1.13170490e-01 1.02380171e-01 -1.19507110e+00 4.57659870e-01
5.81092179e-01 6.14880323e-02 9.23237354e-02 8.06572199e-01
7.61271119e-02 -4.75959517e-02 -9.27274823e-01 6.60010993e-01
-2.64211625e-01 -1.33697498e+00 1.45509675e-01 2.10534140e-01
2.77770519e-01 2.77429372e-01 1.34588495e-01 1.24802560e-01
1.33920059e-01 -4.63195682e-01 1.01990128e+00 3.89098018e-01
6.08243108e-01 -5.73996067e-01 5.59376359e-01 3.08679938e-01
-1.89564478e+00 -3.29572856e-01 -1.58307582e-01 2.20356494e-01
1.92151085e-01 5.50140142e-01 -6.52765512e-01 9.08144176e-01
8.03371966e-01 8.23484838e-01 -7.84016788e-01 1.34230793e+00
-2.08660841e-01 1.50707349e-01 -7.00488865e-01 1.02447078e-01
4.12680566e-01 -4.96923663e-02 8.58701706e-01 8.96397531e-01
2.86041290e-01 -3.62644434e-01 3.15707177e-01 9.36214387e-01
-1.18505973e-02 -3.01893681e-01 -4.33518499e-01 2.20791385e-01
7.80852020e-01 1.48654389e+00 -8.56098413e-01 -3.26257199e-01
-3.09660822e-01 5.39457023e-01 1.68668017e-01 1.63156226e-01
-1.10017359e+00 -3.39449972e-01 7.93523550e-01 1.75796345e-01
9.11345780e-01 -3.93784970e-01 -3.53959024e-01 -1.05149436e+00
2.89141357e-01 -4.54553366e-01 1.30136028e-01 -4.04095918e-01
-1.24070907e+00 5.53399682e-01 -9.74179804e-02 -1.62881470e+00
1.98783875e-01 -5.34327745e-01 -5.16567290e-01 7.60262430e-01
-1.97735262e+00 -1.48572648e+00 -5.68432212e-01 7.28453875e-01
2.96229661e-01 1.75105575e-02 1.50680378e-01 6.67455018e-01
-6.71366096e-01 6.41573131e-01 -2.72605807e-01 -3.24968994e-02
5.32646477e-01 -1.02643430e+00 2.98779488e-01 1.09603572e+00
6.42005876e-02 8.14772844e-01 3.74176115e-01 -8.26731265e-01
-1.44292819e+00 -1.39215398e+00 6.67559206e-01 -3.25345993e-01
8.62998486e-01 -4.98786181e-01 -8.57344210e-01 4.01630104e-01
-2.45171428e-01 3.18059236e-01 1.95382163e-01 -1.76459596e-01
-6.05229735e-01 -6.33018315e-01 -1.04775381e+00 4.45122719e-01
1.38740969e+00 -3.89472067e-01 -3.32485944e-01 1.11858644e-01
7.70994484e-01 -3.24719548e-01 -8.15419495e-01 8.13766062e-01
5.63848555e-01 -9.80820835e-01 1.15509462e+00 -1.92522421e-01
-1.99691996e-01 -1.11720145e+00 -6.74131438e-02 -7.49318719e-01
-4.27870482e-01 -5.37223577e-01 -3.22008759e-01 1.26178026e+00
2.32335001e-01 -7.99671888e-01 8.13967049e-01 2.53335774e-01
-3.74345601e-01 -6.55111551e-01 -1.31796730e+00 -1.22718060e+00
-4.44752812e-01 -3.96679699e-01 8.86986434e-01 6.95385396e-01
-7.05313742e-01 2.19456241e-01 -1.44379139e-01 5.30692756e-01
1.07066929e+00 2.79349953e-01 8.22169244e-01 -1.72386324e+00
-3.98998596e-02 -5.16676009e-01 -5.43637693e-01 -1.28865409e+00
-4.35488969e-02 -9.08031404e-01 7.26780146e-02 -1.20674646e+00
-1.24745995e-01 -8.28480721e-01 -2.81721234e-01 5.47006428e-01
1.08399183e-01 3.96729082e-01 5.38835526e-01 3.74311417e-01
-6.70175612e-01 8.43585074e-01 1.29670954e+00 -8.90091881e-02
-3.88240032e-02 1.66074738e-01 -5.24170637e-01 5.98585010e-01
4.80796546e-01 -5.78165889e-01 -3.73334199e-01 -5.15879750e-01
-6.22674227e-02 -4.51756448e-01 9.15440023e-01 -1.53971088e+00
7.06124485e-01 5.99837452e-02 4.68084335e-01 -1.08053756e+00
6.49860859e-01 -1.27885270e+00 1.48864016e-01 5.27904153e-01
1.76028207e-01 4.62777615e-02 2.37455353e-01 8.03789973e-01
-2.54390061e-01 1.79478005e-01 7.98067391e-01 1.99064821e-01
-8.35806787e-01 7.34061599e-01 3.20393927e-02 -2.50970125e-01
1.04312432e+00 -6.10264122e-01 -4.47043836e-01 1.20353557e-01
-3.11172664e-01 6.18060708e-01 5.46905637e-01 6.00807250e-01
4.98731405e-01 -1.34597218e+00 -2.35420957e-01 3.98299217e-01
1.54709116e-01 2.35330254e-01 1.83481500e-01 1.11981785e+00
-1.10694757e-02 3.95337194e-01 -2.47347608e-01 -1.08506846e+00
-8.82935941e-01 5.25868893e-01 3.21652800e-01 -1.68534830e-01
-7.08034098e-01 6.93040073e-01 1.91132486e-01 -2.28814438e-01
3.53098363e-01 -4.20038342e-01 -3.86297295e-04 3.21529299e-01
2.77269721e-01 1.87936395e-01 1.17355417e-02 -7.52168000e-01
-5.42405546e-01 8.70383620e-01 4.39580344e-02 2.14415103e-01
1.10797620e+00 -7.71977007e-02 2.26269197e-02 1.24862261e-01
8.00895214e-01 9.69673619e-02 -1.58433211e+00 -4.07182455e-01
1.45652875e-01 -5.52004278e-01 3.34154889e-02 -6.37902439e-01
-1.45136356e+00 8.82461786e-01 5.79924881e-01 5.65744005e-02
1.09549391e+00 -1.26696855e-01 9.72923100e-01 2.29519397e-01
5.85407257e-01 -8.47902536e-01 3.57467122e-02 5.52671313e-01
4.53187406e-01 -1.02072883e+00 -2.55682245e-02 -5.18369138e-01
-2.48537809e-01 9.80787516e-01 9.61482584e-01 2.37375926e-02
5.47208607e-01 4.76091594e-01 -7.64242336e-02 -4.24305886e-01
-5.59250355e-01 -5.76760471e-01 3.25953633e-01 7.22307384e-01
-3.55214775e-01 -2.38214791e-01 -8.69911350e-03 3.06874722e-01
-8.80398229e-02 5.27917705e-02 -6.54148906e-02 9.16842341e-01
-5.12477875e-01 -1.05710149e+00 -2.81944424e-01 3.06203991e-01
4.43876982e-02 3.05163278e-03 -1.86985016e-01 1.12003529e+00
5.16097069e-01 6.93752825e-01 1.97947979e-01 -5.47263026e-01
4.84393626e-01 -2.51179755e-01 3.00783187e-01 -3.68725151e-01
-8.78378391e-01 6.71378002e-02 -1.50661081e-01 -9.33106720e-01
-4.67484057e-01 -6.39562905e-01 -1.24730325e+00 -6.98875338e-02
-6.09755397e-01 1.02428459e-01 7.52090037e-01 8.61211121e-01
6.37504399e-01 6.54485583e-01 4.42204505e-01 -1.08016503e+00
-6.16351902e-01 -5.39043248e-01 -2.83040851e-01 1.50128826e-02
4.89203095e-01 -1.16316926e+00 -3.17584604e-01 -4.39630955e-01] | [6.467754364013672, -2.2567920684814453] |
a40e1a84-d7fa-41bd-ae3f-c8efbd7f64ec | scene-consistency-representation-learning-for | 2205.05487 | null | https://arxiv.org/abs/2205.05487v1 | https://arxiv.org/pdf/2205.05487v1.pdf | Scene Consistency Representation Learning for Video Scene Segmentation | A long-term video, such as a movie or TV show, is composed of various scenes, each of which represents a series of shots sharing the same semantic story. Spotting the correct scene boundary from the long-term video is a challenging task, since a model must understand the storyline of the video to figure out where a scene starts and ends. To this end, we propose an effective Self-Supervised Learning (SSL) framework to learn better shot representations from unlabeled long-term videos. More specifically, we present an SSL scheme to achieve scene consistency, while exploring considerable data augmentation and shuffling methods to boost the model generalizability. Instead of explicitly learning the scene boundary features as in the previous methods, we introduce a vanilla temporal model with less inductive bias to verify the quality of the shot features. Our method achieves the state-of-the-art performance on the task of Video Scene Segmentation. Additionally, we suggest a more fair and reasonable benchmark to evaluate the performance of Video Scene Segmentation methods. The code is made available. | ['Linlin Shen', 'Weicheng Xie', 'Haozhe Liu', 'Bo Ren', 'Ruizhi Qiao', 'Yanan Luo', 'Keyu Chen', 'Haoqian Wu'] | 2022-05-11 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Wu_Scene_Consistency_Representation_Learning_for_Video_Scene_Segmentation_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Wu_Scene_Consistency_Representation_Learning_for_Video_Scene_Segmentation_CVPR_2022_paper.pdf | cvpr-2022-1 | ['scene-segmentation'] | ['computer-vision'] | [ 2.08377764e-01 -2.82445520e-01 -5.71475863e-01 -6.57695472e-01
-6.48044884e-01 -6.13871276e-01 5.00810921e-01 -1.15585119e-01
-1.88530818e-01 3.67279232e-01 3.42330635e-01 -1.73752457e-02
7.02798069e-02 -6.13618851e-01 -1.06165922e+00 -5.65270305e-01
-1.28049999e-01 1.07494041e-01 6.22008264e-01 1.26305833e-01
2.88137019e-01 -8.62569883e-02 -1.38463223e+00 6.18557096e-01
5.59692860e-01 1.13905752e+00 5.79403400e-01 6.80313051e-01
-3.68080795e-01 1.15832651e+00 -3.06322932e-01 -3.96640301e-01
1.57061949e-01 -7.20069647e-01 -1.19076800e+00 6.65682793e-01
6.06067717e-01 -5.77046096e-01 -5.26959240e-01 1.17435324e+00
-7.98998848e-02 2.34127894e-01 5.02822578e-01 -1.38075757e+00
-3.62679869e-01 5.97042143e-01 -5.98267198e-01 3.26800615e-01
6.50573611e-01 2.08899558e-01 1.04366648e+00 -6.78400457e-01
8.70384693e-01 9.55785573e-01 4.76059526e-01 3.41993243e-01
-8.50025833e-01 -5.12896836e-01 6.13768935e-01 6.06795967e-01
-1.36795783e+00 -5.24042726e-01 8.67174506e-01 -6.51890099e-01
3.71224344e-01 2.26304159e-02 7.53122747e-01 1.15620422e+00
-1.21542312e-01 1.12014818e+00 6.72291815e-01 -2.39361152e-01
2.33381286e-01 -1.99377507e-01 4.08186466e-01 7.77731776e-01
-1.74549595e-01 -4.24454659e-01 -4.48740810e-01 1.50581628e-01
8.20035219e-01 4.33870196e-01 -3.26089025e-01 -6.89712524e-01
-1.17646694e+00 6.00037694e-01 2.94474512e-01 4.23443228e-01
-2.58179367e-01 1.07906416e-01 5.29331625e-01 7.18355030e-02
3.54094207e-01 3.12593251e-01 -4.38889503e-01 -3.34351569e-01
-1.26570201e+00 2.44537070e-01 5.72775006e-01 1.08924878e+00
9.02939200e-01 -2.51354992e-01 -3.90858084e-01 7.44999528e-01
1.98003799e-01 4.07732986e-02 4.23130035e-01 -1.27655530e+00
4.91499245e-01 5.13384998e-01 1.43067449e-01 -9.05994177e-01
4.05716747e-02 -1.03451833e-01 -6.25861466e-01 -3.97836268e-01
4.68118340e-01 -8.99233595e-02 -1.13950622e+00 1.67732620e+00
3.09720039e-01 9.32993472e-01 -1.44359812e-01 9.17942107e-01
8.17366064e-01 1.12394285e+00 5.50732985e-02 -4.05528307e-01
1.26457131e+00 -1.31623900e+00 -8.98974121e-01 -3.74162197e-01
5.76840818e-01 -4.91604686e-01 1.04560316e+00 1.06368445e-01
-8.88221085e-01 -7.01369107e-01 -1.00665522e+00 -1.30368128e-01
-2.15594918e-01 -1.46726072e-01 5.43495059e-01 1.41749755e-01
-5.97993016e-01 5.49548090e-01 -8.80141616e-01 -5.97818971e-01
4.28620696e-01 -2.36126617e-01 -2.07408518e-01 -4.52297330e-01
-1.07086778e+00 3.53673786e-01 4.77465898e-01 8.85537416e-02
-1.27925837e+00 -7.46583760e-01 -1.21279979e+00 -6.57788245e-03
8.78435671e-01 -3.78826827e-01 1.40361464e+00 -1.31112671e+00
-1.20717609e+00 9.57953751e-01 -4.31396246e-01 -3.65034580e-01
4.02169973e-01 -3.81787390e-01 -2.36441895e-01 3.08235914e-01
3.43762904e-01 6.67351425e-01 8.78413737e-01 -1.22224236e+00
-6.81534290e-01 -1.58474773e-01 5.42130113e-01 5.86272143e-02
-1.67159960e-01 2.19707221e-01 -1.24543619e+00 -6.20177269e-01
1.78887784e-01 -8.61292124e-01 -2.85916716e-01 -1.89076234e-02
-4.29959327e-01 -4.76164334e-02 8.45512688e-01 -6.21406317e-01
1.42179358e+00 -2.39903641e+00 3.05385739e-02 -2.13671654e-01
-6.23498708e-02 1.66511182e-02 -5.49796857e-02 3.17636341e-01
-2.91723758e-02 1.11852720e-01 -4.02903944e-01 -2.46407926e-01
-2.99105316e-01 2.05905572e-01 -5.44895291e-01 3.31869423e-01
-7.46223703e-02 9.10048723e-01 -9.58130896e-01 -6.28001690e-01
2.10031256e-01 -2.88553238e-02 -4.27660495e-01 5.25901020e-01
-5.42110145e-01 4.57110107e-01 -4.93444711e-01 3.93022627e-01
5.06668270e-01 -6.46079183e-01 6.73978999e-02 -2.64569074e-01
-8.56553987e-02 -1.62884817e-02 -1.16966689e+00 2.36161113e+00
-8.13156217e-02 7.04507291e-01 -1.78935900e-01 -1.02130437e+00
4.53954041e-01 2.18537256e-01 6.89721107e-01 -6.51141703e-01
1.96358040e-02 -2.61430323e-01 -5.34276545e-01 -9.66824949e-01
5.93899310e-01 2.21952591e-02 -3.42362314e-01 3.27934831e-01
1.65097892e-01 -5.39192418e-03 4.36342359e-01 3.56861830e-01
7.94756472e-01 3.90501052e-01 2.12585837e-01 -1.00562900e-01
3.27102631e-01 2.12452859e-01 9.28287983e-01 7.17794240e-01
-3.46553952e-01 9.78313684e-01 5.52904665e-01 -3.32989126e-01
-1.02088869e+00 -8.26191545e-01 5.39055020e-02 1.14990258e+00
7.81223953e-01 -6.55753434e-01 -8.81629288e-01 -6.51420891e-01
-3.88402879e-01 6.57231688e-01 -5.60346484e-01 -9.90417539e-05
-5.56935370e-01 -4.17980522e-01 1.94128141e-01 5.13871074e-01
7.10780978e-01 -8.51862371e-01 -4.15659904e-01 2.53111143e-02
-6.10468805e-01 -1.44085240e+00 -8.35275233e-01 -2.29603618e-01
-6.50369883e-01 -1.30141413e+00 -7.89509237e-01 -9.49009478e-01
5.04775286e-01 6.61948562e-01 8.82541955e-01 -1.33531481e-01
-8.07805825e-03 3.87039602e-01 -6.14286482e-01 1.61089778e-01
-2.23645326e-02 -1.33111924e-01 -2.28084847e-01 2.72970796e-01
2.73216307e-01 -2.61463314e-01 -6.44179285e-01 4.97215062e-01
-1.03141344e+00 4.49051321e-01 2.35886090e-02 4.38638359e-01
6.77500904e-01 5.65209463e-02 1.57989964e-01 -1.00017679e+00
5.53098172e-02 -7.59305775e-01 -3.13068211e-01 4.63605076e-01
-1.61499634e-01 -2.02693477e-01 6.13919437e-01 -3.52020621e-01
-1.16857743e+00 8.49752128e-02 -3.45361754e-02 -7.61077762e-01
-3.02184403e-01 4.94474322e-01 -1.79138646e-01 4.12635803e-01
3.43397766e-01 3.01149338e-01 -3.23076487e-01 -5.18211722e-01
4.94291455e-01 4.75556523e-01 6.60089910e-01 -4.12377059e-01
5.50563574e-01 5.49430728e-01 -6.02784336e-01 -7.41838396e-01
-1.35772359e+00 -8.96955729e-01 -8.77107024e-01 -4.97997582e-01
1.18297148e+00 -1.11267662e+00 1.47611403e-03 5.84507048e-01
-1.14295220e+00 -6.42876327e-01 -2.21291125e-01 4.56648141e-01
-7.07542896e-01 4.39042240e-01 -7.41034269e-01 -2.82223672e-01
1.92940906e-01 -1.12616682e+00 1.00459850e+00 2.74427623e-01
-1.89322889e-01 -8.18620920e-01 1.92818627e-01 5.25230467e-01
-1.37198612e-01 1.03763595e-01 6.34359181e-01 -4.91244882e-01
-7.38805771e-01 6.59911409e-02 -2.90793628e-01 1.71393439e-01
2.24508747e-01 2.38941163e-01 -9.33070362e-01 -2.68199056e-01
5.09309061e-02 -3.65330487e-01 9.51258123e-01 5.89261949e-01
1.52297378e+00 -2.92544246e-01 -3.07766944e-01 8.47631514e-01
1.39049697e+00 3.12214434e-01 6.89005911e-01 3.45478415e-01
9.89399731e-01 6.50543690e-01 9.61893082e-01 4.79683280e-01
4.37236130e-01 6.05895400e-01 1.52545780e-01 -8.47440511e-02
-9.90146324e-02 -6.06243253e-01 4.06893373e-01 7.12880909e-01
3.29546958e-01 -2.55755514e-01 -8.10320497e-01 6.84133530e-01
-2.12624526e+00 -1.24827421e+00 2.50741746e-02 1.99236143e+00
6.05977595e-01 2.14640543e-01 1.42293379e-01 -1.42977014e-01
9.39654410e-01 8.97424996e-01 -6.97406709e-01 2.21396983e-02
1.19974345e-01 -5.59679985e-01 1.98076874e-01 2.58325249e-01
-1.26954556e+00 1.24253464e+00 6.63304996e+00 1.02483404e+00
-1.10021245e+00 6.89299032e-02 9.29007411e-01 -2.81415172e-02
-1.36432320e-01 9.82747152e-02 -6.19233489e-01 5.80565870e-01
7.63329327e-01 -1.76175922e-01 2.84030199e-01 1.01768708e+00
5.03861308e-01 -2.74991304e-01 -1.19242489e+00 1.05898201e+00
4.57277596e-01 -1.60130298e+00 6.32118583e-02 -3.75542849e-01
9.81233239e-01 -5.50741963e-02 -3.10246110e-01 4.61543977e-01
-6.37042290e-03 -7.10658848e-01 9.45499361e-01 5.20937383e-01
7.89921403e-01 -4.60029930e-01 3.73669326e-01 3.79984051e-01
-1.52497303e+00 -6.79466799e-02 -1.84567690e-01 -1.20098256e-02
4.89162326e-01 3.98153901e-01 -3.49109352e-01 4.78457093e-01
7.65640855e-01 1.19394350e+00 -7.27522492e-01 1.28092730e+00
-1.82270199e-01 6.76665843e-01 9.37780216e-02 3.87264937e-01
4.06913251e-01 -2.79566586e-01 3.13778639e-01 1.24083066e+00
1.22917853e-01 2.66583890e-01 5.80660462e-01 7.61797488e-01
-2.39204749e-01 8.73506814e-02 -6.44788504e-01 -9.82976034e-02
1.71134159e-01 1.03699303e+00 -9.11863565e-01 -6.09719694e-01
-6.15478933e-01 1.18789053e+00 8.89170840e-02 6.00517869e-01
-1.05131328e+00 -1.31893650e-01 5.57812870e-01 1.93899482e-01
4.52231437e-01 -2.54055768e-01 3.11310496e-02 -1.53172696e+00
1.14239134e-01 -6.53421998e-01 5.36676347e-01 -1.12039781e+00
-1.20869052e+00 3.96034241e-01 2.14148149e-01 -1.39931214e+00
-2.64811903e-01 -1.09084979e-01 -8.38648021e-01 1.70556948e-01
-1.47804320e+00 -9.17516172e-01 -5.02300680e-01 6.98892951e-01
1.27619183e+00 2.38333806e-01 2.69886047e-01 2.67770827e-01
-7.75548935e-01 2.33006522e-01 1.10704795e-01 3.73879910e-01
6.34851336e-01 -8.58750939e-01 2.99630076e-01 1.16815102e+00
3.05166632e-01 3.35923463e-01 7.39043713e-01 -7.63161480e-01
-1.25659835e+00 -1.35847986e+00 4.54111338e-01 -1.23872399e-01
7.80623794e-01 -3.49368304e-01 -1.16016686e+00 9.11123335e-01
2.10018441e-01 1.18553102e-01 6.06377423e-01 1.28828302e-01
-2.97745615e-01 -2.04652056e-01 -7.32933402e-01 6.35078907e-01
1.24714351e+00 -6.53634906e-01 -7.74703860e-01 4.24054444e-01
1.03240633e+00 -2.42042005e-01 -5.81510127e-01 3.87371451e-01
4.85012025e-01 -9.14451659e-01 7.13940501e-01 -5.71716249e-01
7.44612575e-01 -3.39728206e-01 -1.25214159e-01 -8.72148871e-01
-2.91418135e-01 -6.34584367e-01 -1.39200315e-01 1.34050059e+00
1.52347118e-01 9.17003453e-02 9.74822998e-01 6.89382255e-01
-1.09786995e-01 -6.48570836e-01 -5.37929475e-01 -8.25019598e-01
-3.75926912e-01 -7.02905118e-01 5.24060547e-01 9.65191722e-01
5.38234338e-02 2.63831586e-01 -6.82312131e-01 1.31827012e-01
5.03830075e-01 6.30287230e-01 8.28772545e-01 -7.95085371e-01
-7.44730011e-02 -2.15433404e-01 -3.07309359e-01 -1.64833879e+00
3.39103132e-01 -6.56098902e-01 3.87472212e-01 -1.56907070e+00
7.26465940e-01 -8.74286890e-02 -3.64275455e-01 2.21707717e-01
-2.76247352e-01 9.88252908e-02 2.87720472e-01 3.48154306e-01
-1.48736739e+00 5.46856225e-01 1.31829381e+00 -1.79366723e-01
-2.30714709e-01 -1.00131504e-01 -4.58113521e-01 8.61296594e-01
5.54852009e-01 -4.39255178e-01 -6.68335319e-01 -4.56735849e-01
-9.31784287e-02 3.04524928e-01 2.49696180e-01 -9.42983866e-01
5.28021097e-01 -5.50817609e-01 6.77682236e-02 -5.87080002e-01
2.77538776e-01 -6.18797183e-01 1.48337379e-01 9.02535841e-02
-5.15252352e-01 -2.24836603e-01 -1.22769438e-01 7.94477224e-01
-6.03317976e-01 -3.29258710e-01 7.07053185e-01 -3.11241299e-01
-1.42950237e+00 8.40721250e-01 -2.02354640e-01 2.11827606e-01
1.34672379e+00 -3.91911238e-01 -1.47687182e-01 -4.12685633e-01
-5.92685044e-01 5.47968566e-01 7.08373547e-01 5.80002725e-01
6.52386129e-01 -1.15979505e+00 -4.19000417e-01 8.20477456e-02
3.77198488e-01 5.42289168e-02 6.25180602e-01 5.33669889e-01
-4.59298819e-01 3.23579580e-01 -1.05112202e-01 -7.75713980e-01
-1.28815567e+00 7.98576295e-01 1.25753284e-01 -4.86284029e-03
-7.12961137e-01 7.73292482e-01 5.76266646e-01 9.18760488e-04
3.93184125e-01 -3.30976754e-01 -4.40857172e-01 2.12133527e-01
6.91891909e-01 2.32194275e-01 -4.27836895e-01 -7.24922121e-01
-1.35119751e-01 6.91435754e-01 -1.05478272e-01 7.92515278e-02
1.19463837e+00 -4.72955227e-01 3.76346037e-02 9.40562725e-01
1.47158074e+00 -3.91333878e-01 -1.75616145e+00 -2.60418653e-01
4.10617478e-02 -7.38062441e-01 -9.76468772e-02 -3.07373762e-01
-1.04760373e+00 6.75028920e-01 2.80686378e-01 6.71858639e-02
1.01062906e+00 2.10781112e-01 1.06845582e+00 2.55614847e-01
4.79267329e-01 -1.00755739e+00 2.35450178e-01 5.92154801e-01
5.67356408e-01 -1.45921373e+00 4.83345473e-03 -4.22113091e-01
-9.40580130e-01 9.92405593e-01 6.22485101e-01 -6.02327064e-02
5.83387733e-01 -1.24227040e-01 -1.95255846e-01 -3.26038033e-01
-6.45744085e-01 -3.42975497e-01 2.79069036e-01 9.28583667e-02
1.31853744e-01 -2.12621599e-01 1.47851883e-02 5.30567288e-01
2.58657396e-01 9.94710997e-02 5.04340708e-01 8.99014652e-01
-5.22730827e-01 -6.47841215e-01 -6.35626689e-02 4.00790751e-01
-4.14802790e-01 1.24126308e-01 -1.53246477e-01 4.44799900e-01
7.00927526e-02 9.28854525e-01 2.29449511e-01 -3.80252242e-01
1.61529571e-01 9.29862112e-02 2.49138340e-01 -8.80643189e-01
-8.10378045e-02 1.56315833e-01 -8.23579878e-02 -8.74428749e-01
-7.40034580e-01 -7.40466118e-01 -1.19735408e+00 -3.14209349e-02
-2.97714099e-02 1.94245562e-01 3.52483660e-01 1.17699671e+00
2.21440718e-01 4.14397269e-01 8.12973857e-01 -8.16616356e-01
1.48962826e-01 -6.42897367e-01 -5.91957748e-01 8.27816188e-01
3.73575091e-01 -5.58739185e-01 -2.54737020e-01 6.09632850e-01] | [9.384827613830566, 0.5332509875297546] |
f09551c0-543b-481f-8e51-f851f0a72161 | kodf-a-large-scale-korean-deepfake-detection | 2103.10094 | null | https://arxiv.org/abs/2103.10094v2 | https://arxiv.org/pdf/2103.10094v2.pdf | KoDF: A Large-scale Korean DeepFake Detection Dataset | A variety of effective face-swap and face-reenactment methods have been publicized in recent years, democratizing the face synthesis technology to a great extent. Videos generated as such have come to be called deepfakes with a negative connotation, for various social problems they have caused. Facing the emerging threat of deepfakes, we have built the Korean DeepFake Detection Dataset (KoDF), a large-scale collection of synthesized and real videos focused on Korean subjects. In this paper, we provide a detailed description of methods used to construct the dataset, experimentally show the discrepancy between the distributions of KoDF and existing deepfake detection datasets, and underline the importance of using multiple datasets for real-world generalization. KoDF is publicly available at https://moneybrain-research.github.io/kodf in its entirety (i.e. real clips, synthesized clips, clips with adversarial attack, and metadata). | ['Gyeongsu Chae', 'Sungwoo Park', 'Gyuhyeon Nam', 'Jaeseong You', 'Patrick Kwon'] | 2021-03-18 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Kwon_KoDF_A_Large-Scale_Korean_DeepFake_Detection_Dataset_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Kwon_KoDF_A_Large-Scale_Korean_DeepFake_Detection_Dataset_ICCV_2021_paper.pdf | iccv-2021-1 | ['face-reenactment'] | ['computer-vision'] | [-2.10434079e-01 8.97127092e-02 -8.57637674e-02 -3.37569475e-01
-6.17496789e-01 -7.51849949e-01 7.25552976e-01 -8.08676660e-01
-9.70248207e-02 6.85271502e-01 7.64199376e-01 3.62622857e-01
1.60707861e-01 -5.10437727e-01 -6.05346203e-01 -4.75263059e-01
-9.03014019e-02 1.94506180e-02 -3.83862853e-01 -2.68815488e-01
9.28134322e-02 3.63312155e-01 -1.56426466e+00 4.80828106e-01
4.78184551e-01 1.01752043e+00 -3.69890451e-01 4.62061733e-01
3.15774739e-01 7.77647078e-01 -8.40460122e-01 -1.33993077e+00
5.26294172e-01 -4.68276739e-01 -5.56625962e-01 -4.69755977e-02
1.05150640e+00 -9.38809156e-01 -8.08208168e-01 1.17995644e+00
9.74917829e-01 -1.83750674e-01 1.70845836e-01 -1.80763578e+00
-9.08127129e-01 6.09382629e-01 -6.17367268e-01 3.24056715e-01
7.20979333e-01 4.44472343e-01 6.68946981e-01 -1.23373270e+00
1.08822525e+00 1.54215610e+00 7.70364821e-01 9.35068071e-01
-8.78395617e-01 -1.39168274e+00 -3.06271881e-01 3.54880363e-01
-1.47007918e+00 -8.99468184e-01 8.66304696e-01 -5.37240744e-01
6.86213255e-01 1.94985956e-01 8.06998968e-01 2.15226722e+00
-1.34048909e-01 6.62407696e-01 8.88340831e-01 -1.86206967e-01
-1.39062867e-01 2.11489603e-01 -4.22195673e-01 5.62547743e-01
8.11183229e-02 1.34420037e-01 -7.40087688e-01 -3.25421304e-01
5.57333291e-01 -2.96380490e-01 -5.04335046e-01 7.82867223e-02
-7.70968318e-01 1.00142288e+00 1.95279241e-01 2.09068447e-01
-3.13740969e-02 6.18899390e-02 7.54462898e-01 4.08100814e-01
4.92263228e-01 1.97961062e-01 -1.33577824e-01 -2.62458891e-01
-8.31079483e-01 5.50579250e-01 6.55325413e-01 7.43989110e-01
3.52078408e-01 1.20770894e-01 -2.18040403e-02 1.02135468e+00
1.78770348e-02 4.32279021e-01 6.53636098e-01 -1.21940374e+00
3.58391792e-01 1.60463750e-01 -1.39674291e-01 -1.54949391e+00
-7.53164068e-02 -4.20605391e-02 -5.92632651e-01 4.91477288e-02
4.28564996e-01 -3.04483980e-01 -6.35025561e-01 1.96259487e+00
2.58132219e-01 4.31093037e-01 -1.36160418e-01 1.02785540e+00
1.01243627e+00 5.35405278e-01 -2.64927477e-01 1.42067196e-02
1.27318048e+00 -7.94053674e-01 -8.97597969e-01 1.13926739e-01
4.01676476e-01 -8.72050941e-01 1.26363552e+00 6.93873763e-01
-9.68170464e-01 -4.04386222e-01 -8.54855657e-01 -3.60731520e-02
-3.41885686e-01 -3.57305035e-02 4.34266955e-01 1.06995547e+00
-1.20755339e+00 4.30142581e-01 -2.22998187e-01 -4.91892844e-01
8.71017277e-01 9.41931903e-02 -9.41871047e-01 -1.79703027e-01
-1.37352848e+00 7.17515886e-01 -1.89936571e-02 1.96865559e-01
-1.24048758e+00 -7.87504137e-01 -4.89783466e-01 -1.79456517e-01
4.72654253e-01 -2.80033290e-01 1.32605600e+00 -1.52728033e+00
-1.27643764e+00 1.37651730e+00 2.67529786e-01 -2.36672610e-01
7.38118112e-01 -5.64978123e-01 -9.43396330e-01 3.78057331e-01
-5.87814860e-02 7.34756172e-01 1.31569529e+00 -8.99883866e-01
3.78471264e-03 -2.73373991e-01 1.62244141e-01 -2.03878731e-01
-7.31269658e-01 5.55940390e-01 -2.10800663e-01 -1.09612644e+00
-7.15155780e-01 -8.81491065e-01 4.36483622e-01 2.35354900e-01
-4.88351941e-01 -9.81933475e-02 1.02497280e+00 -9.10872519e-01
1.18579769e+00 -2.47473598e+00 -1.95756331e-02 -1.35663033e-01
5.50520062e-01 4.92830127e-01 -4.44347084e-01 5.86400807e-01
-4.94168371e-01 4.51648712e-01 7.21481740e-02 -3.03023517e-01
-7.07674995e-02 -3.93451229e-02 -4.58314121e-01 6.58566654e-01
2.26132244e-01 8.13956916e-01 -7.20927715e-01 -2.17920735e-01
-1.92268550e-01 5.77451229e-01 -9.01087165e-01 1.54631361e-01
8.53036940e-02 1.87199548e-01 -1.16373993e-01 9.84572470e-01
1.07822382e+00 1.73226774e-01 1.20487832e-01 -2.32018337e-01
9.60446522e-02 -2.66884983e-01 -7.61081874e-01 1.63862956e+00
8.56032968e-02 1.01383054e+00 4.02403861e-01 -5.89001179e-01
6.31519139e-01 4.36679393e-01 4.00012553e-01 -7.05226302e-01
3.88168097e-01 1.07664853e-01 -8.19007903e-02 -6.51692510e-01
4.93915170e-01 -1.85396120e-01 2.90547982e-02 1.88846543e-01
4.71669912e-01 2.73813605e-01 4.14798632e-02 5.65303981e-01
1.03508723e+00 -1.40002385e-01 -1.91972237e-02 -1.83417603e-01
3.40645671e-01 -4.06706095e-01 6.69029713e-01 3.05278242e-01
-6.87938929e-01 7.47179806e-01 8.37515712e-01 -4.34497535e-01
-8.53787899e-01 -1.00535917e+00 -8.44315812e-02 8.88826847e-01
-2.29290545e-01 -8.35255146e-01 -1.18999302e+00 -7.32585907e-01
2.00796172e-01 1.87558293e-01 -9.06239748e-01 -4.25703466e-01
-4.44684386e-01 -4.40509260e-01 1.07146621e+00 1.44557133e-01
7.14458406e-01 -1.26644075e+00 -1.89430535e-01 -1.57524198e-01
-4.41189051e-01 -1.17636967e+00 -7.04126418e-01 -7.82310665e-01
-9.19254422e-02 -1.27928495e+00 -8.56174946e-01 -4.53742594e-01
3.17651570e-01 2.10185468e-01 1.20739496e+00 5.30406758e-02
-6.25939667e-01 5.37374735e-01 -4.94081587e-01 -1.14767134e-01
-3.57721835e-01 -4.56121176e-01 4.75766331e-01 2.91009665e-01
3.40930164e-01 -5.87616980e-01 -5.80722272e-01 3.66298586e-01
-9.02513862e-01 -2.06282884e-01 1.87747136e-01 6.68888569e-01
-1.70206070e-01 -1.35810316e-01 9.38774943e-01 -8.69910300e-01
5.92179596e-01 -7.00619638e-01 -3.37656379e-01 4.77696322e-02
-1.48303911e-01 -6.41505420e-01 2.44251311e-01 -6.66634738e-01
-1.13009357e+00 -4.82325047e-01 -2.63980567e-01 -9.78942692e-01
-9.46773887e-02 2.14205086e-01 -4.60083842e-01 -1.68704331e-01
6.88800216e-01 -1.76435247e-01 1.23832695e-01 -4.57286060e-01
4.65294719e-01 7.01052308e-01 6.99001908e-01 -4.93094802e-01
7.66200662e-01 4.25933331e-01 -5.48804641e-01 -7.95703888e-01
-4.68589544e-01 1.93145260e-01 -2.75526702e-01 -8.57689202e-01
6.76208973e-01 -1.18241298e+00 -7.44808674e-01 1.14018679e+00
-1.03018188e+00 -2.11334005e-01 -1.19791642e-01 2.43892998e-01
-2.82165617e-01 3.43769461e-01 -7.60369241e-01 -4.37825471e-01
-1.51840135e-01 -9.62014139e-01 7.07482159e-01 2.57859617e-01
-3.03668618e-01 -5.62101483e-01 1.58783942e-01 6.39775038e-01
3.56899112e-01 6.87535942e-01 3.11598808e-01 -3.54852915e-01
-2.98033535e-01 -1.65993840e-01 -2.70232171e-01 6.25025272e-01
8.97671133e-02 4.50526118e-01 -1.34960759e+00 -4.21550363e-01
-2.89262477e-02 -8.41132343e-01 6.81055844e-01 1.46920249e-01
1.17658174e+00 -6.98121011e-01 -1.03849337e-01 6.34737074e-01
1.06176341e+00 8.08581617e-03 8.54177117e-01 7.95905571e-03
4.63991463e-01 6.84598565e-01 3.91612023e-01 7.70967543e-01
-8.03380609e-02 6.27280235e-01 4.95556802e-01 8.39507207e-02
-3.08955312e-01 -5.15794694e-01 7.17703700e-01 5.80943823e-01
6.38705269e-02 -4.61421877e-01 -7.55907774e-01 3.31712544e-01
-1.41420352e+00 -1.34375215e+00 2.23411635e-01 1.80114853e+00
7.51947045e-01 -1.79216817e-01 4.38597113e-01 -1.51694149e-01
9.74291384e-01 3.32765490e-01 -4.89637822e-01 -3.93263668e-01
-3.54666859e-01 1.56641066e-01 2.87232921e-02 1.61641464e-01
-8.78406048e-01 1.06220996e+00 6.24904060e+00 9.14234519e-01
-1.33457017e+00 3.26295495e-01 7.52148807e-01 -4.14034426e-01
-6.97125718e-02 -4.24231946e-01 -5.16421139e-01 7.99530864e-01
9.85529721e-01 -3.72935712e-01 5.57149172e-01 9.76964474e-01
6.75303191e-02 1.38023093e-01 -8.82710278e-01 1.31608415e+00
4.94804204e-01 -1.32821822e+00 2.24413630e-03 1.54019445e-01
5.89641511e-01 -1.50329235e-03 4.11367655e-01 3.11054200e-01
-1.98759567e-02 -1.12158680e+00 8.91498089e-01 2.67891854e-01
1.19593441e+00 -7.76762247e-01 4.08550918e-01 -2.45827258e-01
-7.20296562e-01 -2.21014068e-01 -2.19994158e-01 -9.14537534e-02
1.36176776e-02 5.84420145e-01 -5.26044667e-01 2.53815413e-01
1.05799448e+00 9.10881877e-01 -6.72338188e-01 6.90416098e-01
1.06357113e-02 5.91536164e-01 -2.01434404e-01 4.64335322e-01
-2.14193121e-01 1.12996269e-02 6.18674874e-01 1.09223115e+00
4.99706417e-01 1.16834370e-02 -2.57253289e-01 7.30639398e-01
-6.76139593e-01 -1.00087106e-01 -7.62154758e-01 -4.78506535e-01
5.07480443e-01 1.50167954e+00 -4.16218728e-01 -1.81858659e-01
-5.87866902e-01 9.91079926e-01 1.77314192e-01 2.07231537e-01
-1.29573786e+00 -9.96225923e-02 1.31139433e+00 3.21273625e-01
-1.38305768e-01 1.70529515e-01 4.09303129e-01 -1.50563598e+00
1.05638383e-02 -1.25618410e+00 3.80134732e-01 -1.00004721e+00
-1.51193964e+00 7.81562448e-01 5.98619832e-03 -1.10586882e+00
-6.13419339e-02 -5.41877687e-01 -4.60135490e-01 3.80153894e-01
-1.01187253e+00 -8.43178928e-01 -4.55766112e-01 1.06632078e+00
4.09909666e-01 -3.04871798e-01 5.79698801e-01 7.93847978e-01
-8.82377923e-01 1.00774777e+00 -2.58577704e-01 3.26925188e-01
1.26370442e+00 -5.28911352e-01 3.36819112e-01 7.23114192e-01
2.38563009e-02 4.83826548e-01 5.90622187e-01 -4.67663795e-01
-1.22669101e+00 -8.69950116e-01 4.10793960e-01 -4.85929221e-01
7.99208820e-01 -7.99641013e-01 -6.44598246e-01 7.52243519e-01
4.74289179e-01 5.08320145e-02 7.88217783e-01 -2.18833253e-01
-7.97876000e-01 -2.36251369e-01 -1.56990814e+00 4.95401025e-01
1.34938502e+00 -7.13967800e-01 -2.84775972e-01 4.38985437e-01
5.22105455e-01 -1.36512473e-01 -6.02783680e-01 3.60360444e-01
7.79529870e-01 -1.41241252e+00 9.28831637e-01 -5.63768744e-01
6.93109691e-01 2.10096642e-01 -1.46466404e-01 -1.22405636e+00
-2.19927981e-01 -9.18913603e-01 -1.91050604e-01 1.67710781e+00
-5.67487329e-02 -7.31178045e-01 8.04020822e-01 6.15329921e-01
1.97476968e-02 -3.91761005e-01 -1.05164719e+00 -7.20467806e-01
-2.66030449e-02 -2.28878155e-01 4.88555998e-01 1.45866513e+00
-6.52519390e-02 -1.08835131e-01 -9.68560636e-01 -1.84514776e-01
4.39293295e-01 -3.22104186e-01 9.20439661e-01 -1.00881708e+00
-1.21928424e-01 -4.53398734e-01 -5.98562598e-01 -2.86873460e-01
3.73027503e-01 -7.30736315e-01 -5.92470765e-01 -4.98812497e-01
4.17594820e-01 2.11477771e-01 -2.63835210e-02 6.61507428e-01
1.68485999e-01 8.35496664e-01 3.83564264e-01 2.55041793e-02
-4.34231311e-02 4.86560285e-01 1.22962916e+00 1.01264484e-01
3.04799348e-01 -4.84670460e-01 -9.05976951e-01 7.04656839e-01
1.11985612e+00 -3.64240557e-01 -3.55036855e-01 -2.24053845e-01
3.31358671e-01 -1.50627896e-01 6.51918471e-01 -9.98166800e-01
-4.26418900e-01 8.04997887e-03 5.08252859e-01 -3.22420560e-02
6.11647785e-01 -6.08798265e-01 3.85396242e-01 3.55124742e-01
-2.43646160e-01 1.43382505e-01 2.74546236e-01 3.11418891e-01
-2.75153667e-01 1.04796313e-01 9.57527518e-01 4.35410216e-02
-7.57041633e-01 4.29280311e-01 -1.94518492e-01 3.05362284e-01
1.11016035e+00 -2.92303115e-01 -7.39346325e-01 -7.23996520e-01
-7.64258742e-01 -3.75669524e-02 3.96049082e-01 9.67899919e-01
7.40132391e-01 -1.51579809e+00 -8.92768383e-01 4.32274014e-01
8.72824639e-02 -8.26760709e-01 7.77561426e-01 6.09973431e-01
-5.26693344e-01 -1.21264413e-01 -6.78852856e-01 -7.08584338e-02
-1.53084397e+00 6.86679900e-01 1.56948566e-01 3.09503824e-01
-6.57538176e-01 1.17019415e+00 4.25809830e-01 -1.90888554e-01
6.68786690e-02 8.44801366e-02 1.10095426e-01 4.75038439e-01
8.73701632e-01 4.46350276e-01 -2.12058216e-01 -8.84905100e-01
-4.50329959e-01 8.35002512e-02 -1.47961140e-01 2.92871278e-02
1.20721364e+00 1.55688478e-02 -1.24616005e-01 -1.56447083e-01
1.45058167e+00 3.94834667e-01 -1.15481675e+00 1.37889922e-01
-4.64523613e-01 -9.99937773e-01 -2.27305427e-01 -1.01178265e+00
-1.65432727e+00 7.34441817e-01 6.25448763e-01 8.99863020e-02
1.28796482e+00 6.88944608e-02 9.88727748e-01 -1.55638292e-01
3.02720487e-01 -8.80976439e-01 5.15660644e-01 1.55091137e-01
1.49829316e+00 -8.84295225e-01 -2.43700981e-01 -4.11850214e-01
-6.70970321e-01 8.91545773e-01 8.41431975e-01 -1.37643591e-01
6.81749701e-01 4.20350939e-01 8.30655470e-02 3.47957388e-02
-7.27941275e-01 4.13877249e-01 -2.56434053e-01 7.49300539e-01
1.60614163e-01 -1.55374790e-02 -8.69223773e-02 8.40928733e-01
-5.39603829e-01 1.80420533e-01 7.96325982e-01 4.89641488e-01
4.91362773e-02 -9.66152012e-01 -3.76337498e-01 2.95533657e-01
-7.40047157e-01 6.53195307e-02 -8.50460231e-01 8.19626212e-01
2.98281819e-01 7.74032891e-01 -4.77030575e-02 -7.24176049e-01
1.22452825e-01 1.78226270e-02 5.47441959e-01 -1.73226923e-01
-6.10635996e-01 -2.10396022e-01 2.81465054e-01 -9.50325191e-01
-2.43979812e-01 -5.35984099e-01 -6.12131655e-01 -1.03423512e+00
-8.29633623e-02 -4.43262942e-02 3.01348448e-01 3.15316498e-01
5.52820146e-01 -1.11656338e-01 6.98580146e-01 -1.09616601e+00
-3.19060028e-01 -9.20766592e-01 -7.76018679e-01 5.51167130e-01
3.02771479e-01 -7.23243117e-01 -5.75527012e-01 -2.19885949e-02] | [12.6832914352417, 1.0215203762054443] |
fd611bae-8b71-4413-bd76-0f72b7f6a294 | split-embed-and-merge-an-accurate-table | 2107.05214 | null | https://arxiv.org/abs/2107.05214v3 | https://arxiv.org/pdf/2107.05214v3.pdf | Split, embed and merge: An accurate table structure recognizer | Table structure recognition is an essential part for making machines understand tables. Its main task is to recognize the internal structure of a table. However, due to the complexity and diversity in their structure and style, it is very difficult to parse the tabular data into the structured format which machines can understand easily, especially for complex tables. In this paper, we introduce Split, Embed and Merge (SEM), an accurate table structure recognizer. Our model takes table images as input and can correctly recognize the structure of tables, whether they are simple or a complex tables. SEM is mainly composed of three parts, splitter, embedder and merger. In the first stage, we apply the splitter to predict the potential regions of the table row (column) separators, and obtain the fine grid structure of the table. In the second stage, by taking a full consideration of the textual information in the table, we fuse the output features for each table grid from both vision and language modalities. Moreover, we achieve a higher precision in our experiments through adding additional semantic features. Finally, we process the merging of these basic table grids in a self-regression manner. The correspondent merging results is learned through the attention mechanism. In our experiments, SEM achieves an average F1-Measure of 97.11% on the SciTSR dataset which outperforms other methods by a large margin. We also won the first place in the complex table and third place in all tables in ICDAR 2021 Competition on Scientific Literature Parsing, Task-B. Extensive experiments on other publicly available datasets demonstrate that our model achieves state-of-the-art. | ['Jun Du', 'Jianshu Zhang', 'Zhenrong Zhang'] | 2021-07-12 | null | null | null | null | ['table-recognition'] | ['computer-vision'] | [ 1.40902445e-01 9.75209773e-02 -3.22395653e-01 -3.71608287e-01
-9.25351143e-01 -8.61035764e-01 2.34761953e-01 7.08272457e-01
-2.36680433e-01 5.84815025e-01 2.13523984e-01 -4.38507348e-01
1.62473798e-01 -9.99456286e-01 -1.15924072e+00 -5.01810014e-01
2.79682964e-01 7.69480944e-01 1.34293750e-01 1.00968853e-01
1.72186181e-01 3.72388631e-01 -1.33049667e+00 9.18721855e-01
1.10574579e+00 1.32952392e+00 4.19744134e-01 4.48679388e-01
-7.92396367e-01 8.32741916e-01 -5.18954039e-01 -8.05414319e-01
-1.60508007e-02 -2.55842805e-01 -9.77942765e-01 1.87124953e-01
3.84722292e-01 -1.64785787e-01 -7.14070573e-02 1.23146629e+00
9.91289914e-02 -3.85593086e-01 5.87599814e-01 -9.39741433e-01
-5.99474669e-01 1.17773080e+00 -7.40038872e-01 8.19045305e-02
3.99625450e-01 -1.24606550e-01 1.28301287e+00 -1.01641190e+00
8.20694268e-01 1.36178613e+00 2.85362303e-01 2.90842235e-01
-8.84823799e-01 -6.83087468e-01 3.37059587e-01 3.20791185e-01
-1.12881875e+00 -3.32425267e-01 4.85175937e-01 -3.99616808e-01
7.84101784e-01 3.71500075e-01 3.32215905e-01 6.75539553e-01
3.20779383e-01 9.89348412e-01 9.26891863e-01 -2.45172501e-01
1.52132034e-01 3.14911783e-01 2.77813852e-01 9.18976367e-01
5.92306256e-01 -8.31261396e-01 -4.52249110e-01 2.88696468e-01
5.34869015e-01 -9.81940553e-02 -1.14466809e-01 -3.33487898e-01
-1.40750945e+00 5.26424587e-01 7.07492292e-01 2.12204531e-01
-1.46093294e-01 -4.39638734e-01 3.67808104e-01 -1.54118076e-01
-6.66702464e-02 3.50967228e-01 -5.20690560e-01 3.35059166e-02
-5.06434560e-01 1.29944170e-02 8.75327051e-01 1.14321709e+00
5.44670224e-01 -7.20827103e-01 -2.36656472e-01 6.93786919e-01
1.12690054e-01 2.64035434e-01 3.23007435e-01 -6.20561719e-01
1.25029469e+00 1.16766846e+00 -1.80323184e-01 -9.90584970e-01
-4.96178567e-01 -4.08488661e-01 -1.29867196e+00 -2.84818530e-01
6.30440474e-01 1.78124726e-01 -9.64428484e-01 1.26315510e+00
3.93838614e-01 -4.15273428e-01 8.92674029e-02 5.75275242e-01
1.37563050e+00 8.17167163e-01 2.70245168e-02 -1.20272942e-01
2.11506128e+00 -1.05266607e+00 -9.50671315e-01 -4.11535025e-01
5.45608938e-01 -7.10288644e-01 1.02081215e+00 5.43977618e-01
-1.07662213e+00 -5.54767430e-01 -1.10658419e+00 -7.36889243e-01
-6.54589772e-01 4.61624146e-01 6.52858436e-01 3.15677106e-01
-5.10073245e-01 4.33347970e-01 -7.40422606e-01 -1.72512725e-01
6.79774582e-01 2.76169747e-01 -6.74065590e-01 -7.32637718e-02
-9.46789980e-01 5.60271680e-01 5.63419461e-01 2.68377900e-01
-1.84330642e-01 -6.98720872e-01 -1.09457970e+00 3.89606744e-01
6.66597307e-01 -6.09096467e-01 8.71818066e-01 -4.33642030e-01
-8.93914938e-01 1.19397509e+00 -5.56595683e-01 -3.97006512e-01
4.04937387e-01 -4.11427319e-02 -1.07254334e-01 -7.23806815e-03
3.24072063e-01 7.16397941e-01 2.35528484e-01 -1.21056473e+00
-8.02204669e-01 -7.06979215e-01 7.45887682e-02 2.63730824e-01
2.66525289e-03 -1.97741419e-01 -1.06418109e+00 -4.92096215e-01
4.41664666e-01 -4.97851729e-01 -7.84248114e-03 4.31787968e-02
-9.69628692e-01 -2.90873230e-01 1.65464938e-01 -8.55988741e-01
1.46841359e+00 -2.13028884e+00 1.62424341e-01 5.41636422e-02
4.65703547e-01 -7.24610910e-02 2.91960865e-01 8.95871297e-02
-6.52952790e-02 4.83176440e-01 -4.15682256e-01 -3.40565950e-01
-1.09462433e-01 5.20720296e-02 -4.29285198e-01 -7.11219832e-02
3.90488565e-01 9.50592518e-01 -4.99854386e-01 -7.67055988e-01
-1.23961754e-01 9.61947292e-02 -5.68542898e-01 6.63971677e-02
-2.87983596e-01 1.83298767e-01 -4.98319447e-01 8.75366867e-01
8.50928605e-01 -7.01680720e-01 4.57220256e-01 -4.68025774e-01
2.68134773e-02 5.70399523e-01 -1.46057618e+00 1.52704978e+00
-2.35018283e-01 1.90663725e-01 -4.77565918e-03 -9.14151430e-01
8.50942671e-01 -1.77071109e-01 1.50305286e-01 -7.08136320e-01
1.89600796e-01 2.97935724e-01 1.53575182e-01 -2.31933683e-01
3.67609382e-01 3.08036745e-01 -3.08957577e-01 -3.07789445e-03
2.78670676e-02 1.33077547e-01 6.58138394e-01 3.01745683e-01
8.30296218e-01 -8.09998512e-02 5.93669534e-01 -1.28492475e-01
9.62122262e-01 2.97969137e-03 6.61072493e-01 6.39692605e-01
2.22395407e-03 7.76781917e-01 1.14881063e+00 -5.16397357e-01
-7.58502007e-01 -9.41506445e-01 -2.04700708e-01 5.46718001e-01
3.57141137e-01 -7.07263231e-01 -9.04968083e-01 -8.94704103e-01
9.39287543e-02 4.43737715e-01 -8.52704287e-01 8.48845094e-02
-6.99614108e-01 -8.11428547e-01 8.04463699e-02 7.08439529e-01
7.43732333e-01 -1.25509679e+00 -2.33462617e-01 1.57133386e-01
-5.59962869e-01 -1.59647310e+00 -6.04107082e-01 5.24519444e-01
-7.62562096e-01 -1.24286425e+00 -1.60613254e-01 -1.01972723e+00
7.89878309e-01 -6.72089309e-02 1.12527108e+00 1.24614969e-01
-1.52539358e-01 -4.74384785e-01 8.05209577e-03 -5.08620560e-01
-1.58214316e-01 3.08884561e-01 -5.26320755e-01 1.31681472e-01
2.43450835e-01 -4.23855558e-02 -3.39672297e-01 1.89021766e-01
-7.30231881e-01 4.60928470e-01 7.35677481e-01 6.97066188e-01
1.13044989e+00 2.19082862e-01 5.68172857e-02 -1.26012051e+00
3.82447839e-01 -3.37931812e-01 -7.04198420e-01 5.91055036e-01
-4.36869472e-01 4.24821377e-01 1.01330376e+00 -1.09902613e-01
-8.60571802e-01 2.64750361e-01 -3.19275856e-01 -1.08796759e-02
-1.35982007e-01 5.49320817e-01 -9.04030561e-01 5.77124059e-01
1.42438591e-01 3.97264302e-01 -2.15290114e-01 -6.04962230e-01
1.57199621e-01 6.28318071e-01 6.96412742e-01 -5.44703484e-01
7.17309833e-01 4.42982644e-01 2.12663077e-02 -2.52351761e-01
-1.12668204e+00 -2.22826377e-01 -7.28578329e-01 3.38251799e-01
1.15854156e+00 -9.00960684e-01 -1.05695450e+00 3.22052926e-01
-1.10293424e+00 3.58353890e-02 5.51620387e-02 5.81369847e-02
-1.83206186e-01 1.91935673e-01 -6.81420624e-01 -4.47792143e-01
-4.84089553e-01 -1.41288209e+00 1.19532752e+00 4.70233440e-01
3.48085421e-03 -6.37394428e-01 -4.73433793e-01 7.36717939e-01
-1.85450241e-01 7.19269887e-02 1.48712111e+00 -9.21378136e-01
-9.12220299e-01 6.54130280e-02 -4.89125222e-01 -1.70978770e-01
2.29296982e-01 -1.67463324e-03 -7.12042153e-01 2.79595733e-01
-2.27869153e-01 -2.37542823e-01 1.17328644e+00 2.00665519e-01
1.75058949e+00 -3.05382133e-01 -4.74082679e-01 7.78350353e-01
1.36982620e+00 5.12055159e-01 7.60952532e-01 5.90029299e-01
8.13722551e-01 5.81260800e-01 6.36678219e-01 1.05032340e-01
7.74280369e-01 3.97862583e-01 3.58171105e-01 -1.18818298e-01
2.09318430e-04 -4.96697396e-01 5.45880087e-02 6.54932916e-01
4.15461332e-01 -4.40900862e-01 -9.33905244e-01 2.62403220e-01
-1.63722301e+00 -6.95971489e-01 -1.71912804e-01 2.08181381e+00
1.07885146e+00 6.71959758e-01 -1.06215388e-01 2.94809490e-01
7.28164434e-01 1.08302692e-02 -5.24819076e-01 -4.91625160e-01
-2.40293220e-01 1.30990312e-01 3.93109947e-01 3.52680504e-01
-1.41131234e+00 9.17648315e-01 5.12065601e+00 9.35946286e-01
-7.99003303e-01 -6.01004481e-01 1.32031000e+00 3.34031135e-01
-2.95640677e-01 -2.27168530e-01 -1.33672822e+00 5.29054105e-01
5.90422213e-01 4.09526229e-02 5.19020677e-01 6.61406398e-01
-2.84275115e-01 -1.67622849e-01 -1.27365935e+00 1.14368761e+00
2.30883788e-02 -1.60779369e+00 4.74987954e-01 -7.59340003e-02
2.37126812e-01 -6.16464734e-01 -9.85182524e-02 3.62067491e-01
-7.26219639e-02 -1.25730455e+00 8.49550784e-01 2.39343926e-01
7.80703306e-01 -6.93182528e-01 8.14048469e-01 3.83608133e-01
-1.47785759e+00 -1.47780590e-02 -3.20888996e-01 3.29427153e-01
-2.63185889e-01 6.17235005e-01 -6.61016345e-01 7.05206513e-01
8.68463397e-01 7.11857855e-01 -8.34939659e-01 8.85747790e-01
-2.36320317e-01 2.91887671e-01 -2.03597933e-01 -5.00450134e-02
-5.18843494e-02 -3.06914717e-01 1.07757688e-01 1.01906717e+00
7.17111900e-02 1.70510009e-01 4.27650288e-02 9.55641747e-01
-7.16990709e-01 2.63377249e-01 -2.06636876e-01 -2.07398906e-01
4.75437731e-01 1.38313651e+00 -1.15566218e+00 -5.02312660e-01
-3.95576566e-01 7.74579287e-01 5.80824137e-01 -1.39467329e-01
-7.79222786e-01 -5.78878522e-01 3.93768728e-01 -1.66760817e-01
5.99536419e-01 2.48019338e-01 -1.09028637e+00 -1.32519507e+00
4.99303639e-01 -1.14635706e+00 7.69299090e-01 -7.31799066e-01
-1.05755413e+00 7.71530986e-01 -3.47136348e-01 -9.60193455e-01
1.19101569e-01 -8.46658170e-01 -3.39889795e-01 7.71179378e-01
-1.28920841e+00 -9.08164442e-01 -3.63834113e-01 3.17038357e-01
7.06776619e-01 1.10642828e-01 7.51819074e-01 2.77925313e-01
-9.22467291e-01 7.53049731e-01 -5.27628325e-03 6.62443697e-01
6.42212033e-01 -1.65847540e+00 6.44538343e-01 7.40819216e-01
2.21918523e-01 7.41430223e-01 3.61243099e-01 -7.00808942e-01
-1.56784761e+00 -9.44662392e-01 1.31826687e+00 -4.22112674e-01
5.36116183e-01 -9.14988995e-01 -1.10288858e+00 6.46345377e-01
-2.95607839e-02 -7.02594742e-02 4.22719598e-01 1.68692708e-01
-5.44051111e-01 -4.40009892e-01 -1.05765021e+00 5.75121045e-01
9.30470347e-01 -2.55846292e-01 -5.61964452e-01 2.65005291e-01
7.14667261e-01 -8.44139218e-01 -8.46102953e-01 4.70441490e-01
5.13547361e-01 -7.48156011e-01 9.88952100e-01 -4.58043367e-01
1.10846841e+00 -3.98404837e-01 6.28984859e-03 -8.06589663e-01
-2.83267766e-01 -2.49335825e-01 -1.19954199e-01 1.41573393e+00
7.86108255e-01 -3.03645164e-01 8.91270220e-01 4.32092816e-01
6.80676252e-02 -1.10167062e+00 -7.31462777e-01 -3.67814302e-01
1.71780601e-01 -6.96968809e-02 8.43802869e-01 7.05682695e-01
-3.82140726e-02 5.17257273e-01 7.60746822e-02 1.69672608e-01
4.70211178e-01 6.35528028e-01 6.39335215e-01 -1.12483597e+00
-1.08001724e-01 -5.93319476e-01 -1.05678335e-01 -1.13240445e+00
-1.47898182e-01 -1.06860292e+00 -1.13912269e-01 -1.92777979e+00
5.94389915e-01 -3.31310660e-01 -3.01859289e-01 5.88238537e-01
-6.12106800e-01 -6.03488758e-02 3.42733055e-01 1.35070354e-01
-7.17668355e-01 1.68744400e-01 1.35590708e+00 -4.40832943e-01
-5.50546311e-02 -1.83912203e-01 -1.29676580e+00 6.98307216e-01
6.46890938e-01 -3.92545521e-01 -5.16214296e-02 -3.95165056e-01
1.76142141e-01 2.37844095e-01 -1.52135938e-01 -7.99955964e-01
2.43394718e-01 -1.31876335e-01 8.43680382e-01 -1.02433622e+00
-2.61486266e-02 -7.18679070e-01 -1.05903231e-01 4.31446880e-01
-4.27566797e-01 1.38393059e-01 3.74176681e-01 3.36873353e-01
-2.47332454e-01 -4.17521186e-02 5.72532356e-01 -1.86445609e-01
-4.29554462e-01 3.70085359e-01 -1.90477539e-02 2.39589974e-01
7.12464094e-01 -7.98873901e-02 -5.56249738e-01 1.24061771e-01
-6.11989260e-01 5.30509293e-01 2.56093323e-01 4.69310015e-01
5.01734912e-01 -1.08837354e+00 -4.98166293e-01 3.63619894e-01
2.68495083e-01 5.38701415e-01 1.15276173e-01 6.30891502e-01
-7.30897963e-01 5.32979250e-01 -1.55146658e-01 -5.50423682e-01
-1.32103205e+00 7.32234299e-01 1.82672128e-01 -8.62569690e-01
-5.33152342e-01 8.04140866e-01 6.50953591e-01 -4.14085150e-01
4.51618850e-01 -8.14977884e-01 -6.08943760e-01 1.65810660e-01
7.56207228e-01 -2.51378447e-01 3.29393119e-01 -3.72866511e-01
-6.24706924e-01 5.81231654e-01 -4.40949470e-01 3.65318984e-01
1.16451073e+00 3.51633392e-02 -3.56179386e-01 4.36478496e-01
1.03235662e+00 3.01281035e-01 -8.13701332e-01 -1.06148347e-01
1.29012749e-01 -2.31624782e-01 -3.63925040e-01 -1.12464488e+00
-1.05279529e+00 9.33100522e-01 2.04069316e-01 5.10641821e-02
1.12851393e+00 1.10385917e-01 7.63573527e-01 3.78064752e-01
1.09951109e-01 -5.91606498e-01 -3.55964392e-01 5.12958705e-01
8.81593108e-01 -1.41312099e+00 1.01259649e-01 -9.80308533e-01
-6.50991678e-01 1.25924814e+00 8.00287545e-01 1.95480421e-01
4.47493672e-01 5.16612709e-01 -1.38273537e-01 4.47728997e-03
-8.63635004e-01 -1.25548705e-01 6.62584066e-01 1.74839765e-01
4.60049272e-01 4.02725860e-02 -3.40998620e-01 1.15872896e+00
-4.22321230e-01 -2.83105701e-01 4.49639469e-01 7.20248759e-01
-5.18367112e-01 -1.08863962e+00 -4.62439001e-01 5.89626372e-01
-7.81294823e-01 -2.74912894e-01 -5.35046160e-01 7.76476741e-01
2.39909947e-01 8.65506887e-01 2.76108623e-01 -2.78962731e-01
3.69051516e-01 -2.97360122e-02 3.69806021e-01 -4.49387163e-01
-7.03086734e-01 9.27739590e-02 2.64192317e-02 -5.04692256e-01
2.57740259e-01 -4.93084967e-01 -1.70761895e+00 -2.22395480e-01
4.29867059e-02 1.51998803e-01 4.53707159e-01 9.30861712e-01
3.71850222e-01 7.93857455e-01 3.61116558e-01 3.15834396e-03
-2.83027917e-01 -6.58371329e-01 -3.71663243e-01 4.67058092e-01
2.59737074e-01 -4.52714592e-01 -2.08564010e-02 1.98203713e-01] | [11.700858116149902, 3.0511741638183594] |
e978db9f-3e3b-4924-bfd1-3b2d0665b0ae | text-classification-for-azerbaijani-language | 1912.13362 | null | https://arxiv.org/abs/1912.13362v1 | https://arxiv.org/pdf/1912.13362v1.pdf | Text Classification for Azerbaijani Language Using Machine Learning and Embedding | Text classification systems will help to solve the text clustering problem in the Azerbaijani language. There are some text-classification applications for foreign languages, but we tried to build a newly developed system to solve this problem for the Azerbaijani language. Firstly, we tried to find out potential practice areas. The system will be useful in a lot of areas. It will be mostly used in news feed categorization. News websites can automatically categorize news into classes such as sports, business, education, science, etc. The system is also used in sentiment analysis for product reviews. For example, the company shares a photo of a new product on Facebook and the company receives a thousand comments for new products. The systems classify the comments into categories like positive or negative. The system can also be applied in recommended systems, spam filtering, etc. Various machine learning techniques such as Naive Bayes, SVM, Decision Trees have been devised to solve the text classification problem in Azerbaijani language. | ['Umid Suleymanov', 'Behnam Kiani Kalejahi', 'Rashid Badirkhanli', 'Elkhan Amrahov'] | 2019-12-26 | null | null | null | null | ['text-clustering'] | ['natural-language-processing'] | [-4.99274969e-01 -2.52472788e-01 -2.65302807e-01 -5.22191703e-01
-1.06729768e-01 -3.88622850e-01 7.75559127e-01 7.48823285e-01
-3.20693821e-01 4.68438655e-01 1.98417470e-01 -6.13281846e-01
-3.39278579e-02 -1.04374027e+00 7.30154589e-02 -5.92293382e-01
3.96063775e-01 6.33802354e-01 2.41208509e-01 -7.29120851e-01
9.46668446e-01 2.78370440e-01 -1.80104840e+00 6.54902041e-01
8.04013073e-01 7.18247890e-01 3.00319940e-01 2.63701826e-01
-8.32129478e-01 9.47991192e-01 -7.84590542e-01 -3.53078336e-01
2.61434391e-02 -4.52093154e-01 -1.02266240e+00 1.54431298e-01
-1.61468834e-02 1.94598511e-01 3.41256171e-01 1.26821625e+00
1.09760642e-01 1.19216770e-01 8.98619413e-01 -1.06499553e+00
-2.97803789e-01 7.00661182e-01 -6.25406504e-01 1.41604230e-01
3.56901675e-01 -7.85060644e-01 8.05270970e-01 -6.79334104e-01
5.43953896e-01 1.35924911e+00 3.34868997e-01 1.57866806e-01
-3.27138156e-01 -7.58723140e-01 1.57414317e-01 4.18760002e-01
-9.09786344e-01 1.90513343e-01 5.56513846e-01 -6.68197095e-01
6.48403645e-01 4.22017306e-01 6.76296949e-01 3.46716404e-01
4.38967049e-01 6.58654869e-01 1.30870044e+00 -6.37022555e-01
3.37310493e-01 1.05073738e+00 8.93090427e-01 3.54777008e-01
3.61057550e-01 -8.19473326e-01 -2.64641911e-01 -1.53917044e-01
-1.49811000e-01 3.76203567e-01 1.30253613e-01 2.00960100e-01
-5.49050987e-01 1.42250705e+00 2.49206379e-01 8.84270668e-01
-2.34330911e-02 -7.09359527e-01 7.63933539e-01 6.50680244e-01
8.48982036e-01 5.25502384e-01 -7.57183135e-01 -2.37864200e-02
-6.25981927e-01 1.96343511e-01 1.05328643e+00 5.85316956e-01
6.06323540e-01 -3.78501594e-01 5.76867282e-01 1.15727425e+00
6.84614062e-01 1.12284876e-01 1.12731707e+00 -2.76982367e-01
2.55262703e-01 1.00593078e+00 -8.12109783e-02 -1.29277265e+00
-5.32401979e-01 -1.54040039e-01 -6.31568968e-01 8.57565850e-02
1.14177607e-01 -2.53629714e-01 -7.58039355e-01 5.30068576e-01
3.68786991e-01 -6.43106341e-01 5.04759513e-02 8.65339100e-01
1.24146521e+00 1.06208575e+00 1.14047393e-01 -4.16814834e-01
1.72187626e+00 -9.31573391e-01 -7.88566232e-01 -4.00083363e-02
8.37842286e-01 -1.40750873e+00 9.07950163e-01 1.01193750e+00
-4.79860097e-01 -4.43398654e-01 -6.28802836e-01 3.30522209e-01
-1.09662437e+00 2.47763604e-01 8.01759124e-01 8.74375522e-01
-5.22615016e-01 3.68370056e-01 -3.36799055e-01 -1.10801816e+00
-4.93847542e-02 3.50068599e-01 -9.12259370e-02 7.38358796e-02
-1.23708308e+00 9.68955696e-01 3.81786346e-01 -3.98056030e-01
4.23160009e-02 4.95950803e-02 -6.85256779e-01 -1.00957379e-01
5.75036816e-02 -6.06575832e-02 9.32046413e-01 -1.42349577e+00
-1.13408566e+00 1.06310201e+00 -4.71975505e-02 -2.42060810e-01
-2.34562308e-02 1.71035618e-01 -9.08045769e-01 -5.02255782e-02
4.08600837e-01 -4.35492881e-02 4.35090423e-01 -7.71471858e-01
-1.20717096e+00 -6.51899159e-01 -1.90138847e-01 1.82198465e-01
-6.56524658e-01 5.51831901e-01 -6.77421838e-02 -7.69618273e-01
3.34854722e-01 -6.99588716e-01 -2.40682021e-01 -8.55434895e-01
-1.09229967e-01 -5.61772048e-01 1.01842082e+00 -6.65090501e-01
1.19943309e+00 -1.87379611e+00 -5.59234440e-01 6.02159560e-01
-2.17532918e-01 2.82206267e-01 5.50028384e-01 5.97466111e-01
-2.89821289e-02 2.60507375e-01 4.09364074e-01 4.09169853e-01
-4.24120843e-01 2.41628122e-02 -1.21499278e-01 2.13432118e-01
-3.93330455e-01 1.18117826e-02 -6.26868606e-01 -6.06573164e-01
1.06995746e-01 9.39569548e-02 -1.15407050e-01 -3.81228089e-01
3.69230695e-02 -4.11352292e-02 -6.76271021e-01 5.87915957e-01
5.38517177e-01 -1.39262319e-01 2.25979224e-01 2.19382811e-02
-2.70008564e-01 5.73959708e-01 -1.05122650e+00 8.18012476e-01
-4.78322089e-01 8.39738727e-01 -7.03999447e-03 -1.49876249e+00
1.33358705e+00 1.24909125e-01 5.81350207e-01 -3.29124987e-01
7.60762572e-01 1.77322790e-01 2.23479182e-01 -7.92284369e-01
6.54594719e-01 -2.43282735e-01 3.52351293e-02 4.36791778e-01
-2.63215959e-01 -9.17038098e-02 5.77629983e-01 2.98488200e-01
5.24083316e-01 -2.51806170e-01 5.54391742e-01 -4.64739442e-01
8.84558976e-01 7.09485590e-01 1.20045654e-01 3.92523319e-01
-9.43591120e-04 1.25185177e-01 3.45341891e-01 -5.05972862e-01
-6.86020970e-01 -5.62390052e-02 -3.11538458e-01 1.25204885e+00
2.20029324e-01 -6.72090590e-01 -5.46766102e-01 -7.71378934e-01
-1.23901568e-01 5.64083338e-01 -1.53876930e-01 9.71280262e-02
4.18650620e-02 -7.14105427e-01 -4.00359213e-01 -2.27994427e-01
6.04100227e-01 -8.83919060e-01 1.76017750e-02 2.82889664e-01
-1.22411415e-01 -5.07238328e-01 5.23937605e-02 1.93298116e-01
-9.08691168e-01 -9.19830978e-01 -5.81753552e-01 -1.29460919e+00
6.05541289e-01 5.44081271e-01 8.99452984e-01 3.12873721e-01
-1.06310025e-01 6.37653545e-02 -1.08766496e+00 -1.00257838e+00
-6.42558336e-01 4.94879067e-01 1.73416838e-01 -1.13097802e-01
1.03255224e+00 1.54374734e-01 -1.36581764e-01 5.41094840e-01
-6.60150230e-01 -3.07836264e-01 2.36173272e-01 6.85801506e-01
-1.07527703e-01 8.77837241e-01 8.68305326e-01 -1.63232970e+00
9.21647251e-01 -6.85134411e-01 -3.35876346e-01 7.88622871e-02
-8.42279971e-01 -3.98929477e-01 7.18673885e-01 -3.39620620e-01
-1.18077052e+00 -5.16392365e-02 -5.18661559e-01 7.61968791e-01
-3.63823891e-01 8.24529350e-01 -6.13097735e-02 -4.46133763e-02
9.07707751e-01 -9.80397239e-02 1.44990487e-03 -3.69034410e-01
-1.61584586e-01 1.71629608e+00 -3.74950230e-01 -5.90883307e-02
3.59275132e-01 2.77084440e-01 -4.10209209e-01 -1.02137458e+00
-7.95038640e-01 -1.35041082e+00 -2.06559017e-01 -4.47463989e-01
6.11398995e-01 -7.73490131e-01 -7.03352094e-01 2.34842673e-01
-8.07919264e-01 5.34447372e-01 2.13773459e-01 8.23727608e-01
4.94488366e-02 3.61986250e-01 -7.33354449e-01 -8.90058637e-01
-4.78041351e-01 -8.01045120e-01 4.45216268e-01 5.83863854e-01
-4.70074654e-01 -1.04095411e+00 -2.21310571e-01 6.54909968e-01
1.97433025e-01 -5.59419751e-01 9.74720776e-01 -1.23427069e+00
1.59602061e-01 -4.92878616e-01 1.01447858e-01 5.26442647e-01
1.85242563e-01 1.34063020e-01 -4.72387016e-01 -4.95416336e-02
3.24322462e-01 -3.24193686e-02 7.21498668e-01 3.53330702e-01
9.35087979e-01 -3.46069247e-01 -5.84622025e-01 -2.14746714e-01
1.16477621e+00 8.69388819e-01 5.95754147e-01 8.47928047e-01
2.43387103e-01 1.02855313e+00 1.31676126e+00 4.05961335e-01
9.53537002e-02 2.57697850e-01 -1.54609559e-02 -7.75640607e-02
5.12097895e-01 1.36949629e-01 3.53433818e-01 1.10214245e+00
3.08564365e-01 -7.42864907e-02 -8.54254723e-01 3.74902964e-01
-1.91697848e+00 -8.54625344e-01 -9.02597427e-01 1.86847496e+00
6.15768194e-01 2.71537185e-01 2.71606207e-01 5.48856676e-01
7.48780191e-01 -3.10363203e-01 9.91207659e-02 -8.22399855e-01
2.18971893e-01 2.11891621e-01 5.44798076e-01 2.69031465e-01
-1.32985783e+00 1.02275419e+00 5.35941744e+00 1.09004033e+00
-1.29664874e+00 5.82634471e-02 8.29200506e-01 4.53521699e-01
3.12912725e-02 1.43514782e-01 -1.06374669e+00 6.25956357e-01
6.65153742e-01 -3.42947304e-01 -2.85097182e-01 1.30350280e+00
3.14817548e-01 -6.74353123e-01 -2.24086881e-01 9.56690133e-01
3.01776797e-01 -1.36688137e+00 1.93594396e-02 -2.65079513e-02
7.65319169e-01 -1.90812454e-01 -2.29270294e-01 4.61487859e-01
2.36252606e-01 -5.48950791e-01 7.30150342e-02 -2.82355328e-03
-5.71704917e-02 -1.02165389e+00 1.25599492e+00 6.31315351e-01
-7.18618393e-01 -1.69940338e-01 -6.93388760e-01 -6.56439126e-01
-1.52372360e-01 7.86118031e-01 -1.00224805e+00 4.54687417e-01
1.11276424e+00 8.40813994e-01 -5.24811685e-01 1.16245818e+00
1.05630040e-01 7.21592426e-01 -2.16877207e-01 -9.66109097e-01
2.89424926e-01 -7.45129347e-01 3.71013060e-02 1.00238776e+00
3.17425191e-01 7.17648119e-02 1.91140920e-01 -5.94613291e-02
4.74848539e-01 1.25112522e+00 -5.55436671e-01 6.51205629e-02
-7.44720101e-02 1.41296053e+00 -1.46640754e+00 -6.71228528e-01
-3.35355073e-01 5.82857013e-01 -4.68182176e-01 -9.73765403e-02
-3.37656498e-01 -1.05607224e+00 3.03034633e-01 5.96058071e-01
-1.34107113e-01 1.39317557e-01 -4.66835350e-01 -9.98003066e-01
-4.11057949e-01 -1.08489537e+00 6.34124756e-01 -6.33330703e-01
-1.29650390e+00 5.83731115e-01 -1.74182385e-01 -1.30013096e+00
-1.52810723e-01 -9.00153160e-01 -5.28758228e-01 4.75889891e-01
-9.27236974e-01 -6.83183849e-01 -4.57908865e-03 4.00445729e-01
8.45327914e-01 -8.55643153e-01 6.44724905e-01 6.18499517e-01
-2.90124118e-01 -2.13511288e-01 4.76935357e-01 1.74961671e-01
9.17214870e-01 -1.16753054e+00 -8.46113563e-02 4.88683224e-01
1.34633988e-01 6.49809599e-01 7.84226954e-01 -6.58332586e-01
-4.69742507e-01 -6.71040535e-01 1.41282785e+00 -2.01100037e-01
6.66290045e-01 -1.65697783e-01 -6.80167496e-01 2.29085371e-01
4.21328396e-01 -9.03763413e-01 9.73540962e-01 3.48129451e-01
2.28375047e-01 -2.96285272e-01 -1.16914487e+00 4.36078936e-01
-2.36508306e-02 3.11886314e-02 -6.61668658e-01 9.99898374e-01
1.91696569e-01 -6.28139302e-02 -5.11296034e-01 -3.87188822e-01
2.49460652e-01 -8.80232930e-01 4.58990574e-01 -8.06453288e-01
4.01583582e-01 -4.04782653e-01 3.50799114e-01 -1.36260629e+00
-1.14495419e-01 -1.54311374e-01 8.65130782e-01 1.48829150e+00
6.38075352e-01 -8.49197567e-01 9.42906082e-01 3.18668932e-01
2.10139662e-01 -3.98718119e-01 -3.16019744e-01 -3.43705386e-01
5.85736893e-02 -4.11692351e-01 1.97824642e-01 1.35701334e+00
5.66046894e-01 8.17735374e-01 -2.03893363e-01 -4.31990296e-01
1.23230055e-01 1.35889769e-01 7.71636963e-01 -1.63590360e+00
2.98686177e-02 -4.47249681e-01 -4.59084392e-01 -8.48886907e-01
1.21512651e-01 -8.10932100e-01 -3.18881750e-01 -1.86757183e+00
-6.09444194e-02 -6.02553129e-01 1.94868669e-01 2.98979998e-01
1.44629300e-01 1.22225560e-01 5.17560951e-02 4.15302098e-01
-2.40409434e-01 -1.99696630e-01 9.46868598e-01 -1.65462211e-01
-2.56311715e-01 7.45184243e-01 -1.08408880e+00 8.83936942e-01
1.00255692e+00 -6.74410999e-01 -2.66328186e-01 2.81882346e-01
7.99194634e-01 -2.91620463e-01 -6.23395324e-01 -6.38534486e-01
3.00213963e-01 -4.35959727e-01 2.24025205e-01 -7.58509099e-01
-1.58473447e-01 -1.12567282e+00 6.71970844e-02 5.74952841e-01
-2.48211622e-01 -1.27186313e-01 -2.65936494e-01 1.61091954e-01
-6.32359445e-01 -9.43881869e-01 7.75565326e-01 -4.57731724e-01
-6.98771656e-01 -2.41651252e-01 -1.19291234e+00 -4.68726426e-01
1.29548371e+00 -2.46452212e-01 -4.01897013e-01 -6.92721546e-01
-7.21960843e-01 4.15538222e-01 2.48476759e-01 5.23024619e-01
1.58858761e-01 -1.03682578e+00 -4.64157790e-01 8.49774703e-02
2.55254805e-01 -5.29287100e-01 -3.13347876e-02 7.07176983e-01
-1.14067018e+00 4.14051503e-01 -3.39379936e-01 -2.68813968e-01
-1.82849801e+00 4.69138145e-01 -8.69546458e-02 -2.47833773e-01
-1.62125394e-01 6.16897047e-01 -2.26847813e-01 -4.81525987e-01
2.10831463e-01 -1.30597204e-01 -1.12833261e+00 5.95570743e-01
5.37632346e-01 3.14431340e-01 3.57967257e-01 -6.83932543e-01
-3.47972065e-01 4.97787356e-01 -4.18741524e-01 6.38058707e-02
1.08490980e+00 -1.61911353e-01 -5.83296239e-01 5.62202930e-01
1.18531084e+00 1.38497472e-01 2.93742448e-01 1.24112681e-01
3.44781250e-01 -3.96381706e-01 1.40287116e-01 -7.99144626e-01
-8.69982481e-01 7.04518855e-01 5.85876524e-01 1.06853914e+00
9.62870359e-01 2.09028926e-02 5.35932660e-01 5.88326812e-01
3.33652258e-01 -1.73346269e+00 -3.83635044e-01 7.33590364e-01
3.49394709e-01 -1.50236952e+00 3.55628252e-01 -5.56753933e-01
-1.08810711e+00 1.40745318e+00 3.36888820e-01 -2.60102637e-02
1.28084874e+00 7.28151277e-02 4.69793379e-01 -2.15093032e-01
-4.89165395e-01 -3.75283867e-01 1.71640322e-01 5.00257611e-01
1.09682298e+00 3.83968912e-02 -1.38922417e+00 4.92583096e-01
-4.61647868e-01 -1.62624568e-01 8.45033407e-01 8.11325550e-01
-1.00643909e+00 -1.38337767e+00 -7.52367973e-01 1.17872739e+00
-9.60578918e-01 8.41769949e-02 -5.82090080e-01 7.94108748e-01
1.48143902e-01 1.43923318e+00 1.29358575e-01 -4.12437052e-01
3.65155600e-02 4.64954264e-02 -2.28879988e-01 -7.02600598e-01
-1.05297518e+00 3.78756464e-01 3.25122356e-01 2.84432858e-01
-7.43111193e-01 -6.10450089e-01 -1.25197947e+00 -3.92765850e-01
-7.22369790e-01 1.02102542e+00 1.26051390e+00 8.74646842e-01
-1.70418397e-01 1.06351031e-03 7.84917116e-01 -2.20916823e-01
3.02774981e-02 -1.14101827e+00 -9.62899864e-01 5.27844012e-01
-3.05447310e-01 -4.84394789e-01 -3.54425639e-01 9.18597504e-02] | [10.910591125488281, 7.036771774291992] |
25dc2574-6d31-43a8-b2d3-039c195fa484 | deep-reinforcement-learning-for-asset-1 | 2301.05300 | null | https://arxiv.org/abs/2301.05300v1 | https://arxiv.org/pdf/2301.05300v1.pdf | Deep Reinforcement Learning for Asset Allocation: Reward Clipping | Recently, there are many trials to apply reinforcement learning in asset allocation for earning more stable profits. In this paper, we compare performance between several reinforcement learning algorithms - actor-only, actor-critic and PPO models. Furthermore, we analyze each models' character and then introduce the advanced algorithm, so called Reward clipping model. It seems that the Reward Clipping model is better than other existing models in finance domain, especially portfolio optimization - it has strength both in bull and bear markets. Finally, we compare the performance for these models with traditional investment strategies during decreasing and increasing markets. | ['Bo-Kwan Jeon', 'HyungJun Moon', 'KangHun Lee', 'Moon-Ju Kang', 'Jiwon Kim'] | 2023-01-02 | null | null | null | null | ['portfolio-optimization'] | ['time-series'] | [-7.84853756e-01 1.79751337e-01 -5.57514191e-01 6.09798096e-02
-1.54680446e-01 -4.00065154e-01 4.73991781e-01 -8.42569321e-02
-4.41063464e-01 1.50865245e+00 1.06540598e-01 -4.60920334e-01
-5.28814316e-01 -1.08570421e+00 -1.76171958e-01 -6.30511940e-01
-4.80024785e-01 7.72599161e-01 9.33080167e-02 -6.75009310e-01
8.80322576e-01 4.83034104e-01 -9.67834771e-01 -1.81868106e-01
6.28971219e-01 9.99587357e-01 -8.78062993e-02 5.33318758e-01
-2.91209370e-01 1.32997847e+00 -8.92315686e-01 -6.08500540e-01
7.43950546e-01 -5.34290195e-01 -4.50798631e-01 -3.70876819e-01
-5.98116279e-01 -6.43209219e-01 -9.90686938e-02 7.16931939e-01
7.46252000e-01 5.29962108e-02 6.58829629e-01 -1.36900127e+00
-7.45351911e-01 1.04619813e+00 -9.13926303e-01 6.71004593e-01
2.26033777e-01 1.54592499e-01 1.03582871e+00 -3.69294763e-01
3.92467886e-01 1.19367802e+00 3.95140290e-01 5.57249665e-01
-5.73132873e-01 -6.25082552e-01 1.98014453e-01 2.19173968e-01
-4.54950273e-01 1.46029189e-01 7.40704417e-01 -2.63471156e-01
1.41691196e+00 9.28398669e-02 1.19096494e+00 4.78799373e-01
4.91938114e-01 9.64838088e-01 1.32796073e+00 -6.05358660e-01
4.28946376e-01 3.01664025e-01 -2.10979879e-01 -2.18017437e-02
5.20432532e-01 9.33806777e-01 -4.16820586e-01 -2.58406043e-01
1.02336514e+00 -1.23792149e-01 2.81395525e-01 -2.96886444e-01
-8.16962838e-01 1.19308269e+00 5.79036493e-03 2.05896556e-01
-9.96642649e-01 4.40168977e-01 2.36861333e-01 8.55300188e-01
3.11652452e-01 6.48932159e-01 -5.62558830e-01 -4.62504566e-01
-8.76085818e-01 8.56348634e-01 1.06229281e+00 7.11590409e-01
8.67261961e-02 1.04530764e+00 -1.85862467e-01 4.27540958e-01
5.75754285e-01 4.46911752e-01 7.09310949e-01 -1.12903643e+00
6.54598475e-01 4.20584530e-02 7.04033852e-01 -5.48537910e-01
-4.22673076e-01 -5.02396762e-01 -3.51372540e-01 1.02469301e+00
4.60850835e-01 -7.56579697e-01 -1.13571130e-01 1.22024655e+00
1.34066299e-01 -6.75750747e-02 5.28408408e-01 4.20471817e-01
2.91590691e-01 7.06093609e-01 3.20263468e-02 -7.47647583e-01
8.26505899e-01 -1.27191949e+00 -1.02620947e+00 2.81377673e-01
2.44825687e-02 -6.45419955e-01 2.74899542e-01 7.41277337e-01
-1.60568273e+00 -3.70599687e-01 -9.72368658e-01 1.06376588e+00
-3.01095963e-01 -3.97512555e-01 7.75542140e-01 9.35633957e-01
-9.26887512e-01 1.30389798e+00 -3.53959292e-01 2.35912383e-01
3.33939046e-01 4.68269408e-01 4.87726986e-01 8.81774962e-01
-1.40500367e+00 1.24442935e+00 4.31244373e-01 -2.26538137e-01
-1.00301385e+00 -4.79714662e-01 -1.35700315e-01 2.42719114e-01
3.42307031e-01 -4.73024637e-01 1.53937900e+00 -1.05948448e+00
-2.17280531e+00 1.87686294e-01 6.95833564e-01 -1.12351990e+00
8.79949391e-01 -3.46866608e-01 -4.23439533e-01 5.90700172e-02
-4.54824328e-01 4.00413245e-01 8.17837000e-01 -9.19109285e-01
-9.72857893e-01 6.28691688e-02 2.85579562e-01 4.33405310e-01
1.36514250e-02 1.98757336e-01 7.05346942e-01 -1.01672375e+00
-7.03082919e-01 -3.85001063e-01 -4.82520491e-01 -4.31471139e-01
3.07685971e-01 -5.45825839e-01 3.20236355e-01 -6.03654921e-01
1.41544831e+00 -1.80235529e+00 -1.74596712e-01 3.22759449e-01
-2.71649718e-01 3.38400006e-01 1.28578126e-01 7.27865934e-01
-4.00007546e-01 9.52871144e-02 3.83936971e-01 4.56385165e-01
2.78048486e-01 2.70843178e-01 -3.73916566e-01 8.53252113e-02
7.39817768e-02 1.07008469e+00 -8.58569801e-01 -3.55258822e-01
6.28048033e-02 -4.02235657e-01 -4.36913997e-01 3.32780451e-01
-3.02070498e-01 2.98258420e-02 -8.37469995e-01 9.68479276e-01
6.67989135e-01 3.08433026e-01 9.07391869e-03 6.50298297e-01
-2.87416279e-01 -1.68149948e-01 -1.48119533e+00 7.29738891e-01
-1.68691203e-02 2.06342310e-01 -3.53171676e-01 -1.39543033e+00
1.26755154e+00 6.51837766e-01 7.87053943e-01 -8.05210173e-01
9.68130752e-02 5.93035877e-01 2.04259261e-01 -4.56784785e-01
3.43314946e-01 -7.06953704e-01 4.14144307e-01 8.24271500e-01
-7.33395964e-02 -5.32004237e-03 4.19755340e-01 -2.05051392e-01
8.43035281e-01 6.00493073e-01 5.35829782e-01 -4.45295841e-01
5.67494333e-01 -1.34863153e-01 6.52548730e-01 8.52023542e-01
-5.95798492e-01 -1.88485458e-01 1.02289748e+00 -6.73845589e-01
-1.12696075e+00 -1.00080180e+00 9.32309404e-02 9.00922120e-01
-3.60571146e-02 4.45995927e-01 -4.73837972e-01 -7.54535854e-01
6.78383827e-01 8.67586374e-01 -5.34067512e-01 8.40232521e-02
-5.98746598e-01 -1.03266668e+00 1.80734441e-01 6.15738094e-01
6.48013651e-01 -1.91587818e+00 -9.74822760e-01 7.16212213e-01
6.58451498e-01 -1.98242530e-01 -4.56916867e-03 2.20575899e-01
-1.11807740e+00 -1.01205814e+00 -1.31330538e+00 -2.73928255e-01
5.67994900e-02 -2.14585051e-01 1.36925352e+00 -1.61126312e-02
5.38256615e-02 4.46837485e-01 -7.98622489e-01 -1.22463822e+00
-1.95898056e-01 -2.88097203e-01 -8.15960690e-02 -3.15051049e-01
3.28565389e-01 -4.96433437e-01 -7.38451540e-01 1.48405492e-01
-4.25752014e-01 -7.85138726e-01 6.05687857e-01 9.43333745e-01
1.51762024e-01 1.76909789e-01 1.44396770e+00 -8.54323208e-01
1.16959822e+00 -6.89248621e-01 -9.16464448e-01 3.94447446e-01
-1.17429340e+00 1.40437648e-01 5.09809554e-01 -3.36835206e-01
-1.33835471e+00 -5.74827313e-01 1.94795094e-02 -8.93971771e-02
5.39069355e-01 3.20917636e-01 3.72871459e-01 -1.70538619e-01
3.17796767e-01 -3.71730141e-02 2.45700166e-01 -4.09886986e-01
9.94133949e-03 4.62391287e-01 -1.02030672e-01 -6.07611716e-01
6.66892946e-01 -6.31079972e-02 1.13778144e-01 -1.95716575e-01
-4.87493634e-01 1.09345771e-01 -2.03092992e-01 -5.02239406e-01
5.09250104e-01 -4.40835506e-01 -8.23661685e-01 5.67107320e-01
-6.91873610e-01 -4.65313017e-01 -9.70525980e-01 7.88519919e-01
-1.00574720e+00 5.33936955e-02 -6.01230741e-01 -1.50494587e+00
-5.40249765e-01 -6.49113417e-01 3.61146964e-02 8.57980788e-01
1.80583254e-01 -1.26892555e+00 6.56600893e-01 -2.06549224e-02
7.67195106e-01 5.59120059e-01 5.50089955e-01 -8.60209048e-01
-2.88576066e-01 -6.08444251e-02 2.01976448e-01 3.88169289e-01
-2.99602836e-01 -2.46283785e-02 -3.93810928e-01 -3.23367566e-01
1.80149689e-01 -4.89247650e-01 5.67613423e-01 5.61011910e-01
5.74826300e-01 -2.47208595e-01 1.18977576e-01 1.82766035e-01
1.56270850e+00 1.13449657e+00 7.55092561e-01 1.21330810e+00
-4.04489934e-01 6.48082018e-01 1.27125907e+00 1.01038957e+00
-1.17700368e-01 2.89122552e-01 6.51637256e-01 1.96980253e-01
4.05315220e-01 -2.42289435e-02 4.59878534e-01 4.12645191e-01
-6.70775175e-01 -1.94510147e-01 -5.02155423e-01 1.53858975e-01
-2.18915796e+00 -1.47917116e+00 3.72756809e-01 1.99017608e+00
6.32800162e-01 6.53887987e-01 9.02342141e-01 -3.56635563e-02
5.31846344e-01 5.30198291e-02 -7.36213684e-01 -1.03923094e+00
-2.78545111e-01 6.47520721e-01 7.41106153e-01 3.86912018e-01
-7.55786896e-01 7.96420217e-01 8.09624004e+00 8.26618195e-01
-5.87305486e-01 -8.12014639e-02 7.25376368e-01 -1.57330096e-01
-3.14797610e-01 2.13304862e-01 -8.66095364e-01 5.20614624e-01
1.02322793e+00 -7.97919869e-01 3.54245186e-01 1.19492292e+00
2.70484596e-01 -1.89861491e-01 -5.88799298e-01 4.91020888e-01
-2.90682584e-01 -1.19879317e+00 1.74040988e-03 1.17022976e-01
6.74655914e-01 -5.72461903e-01 -2.66921148e-02 7.49720454e-01
7.07544565e-01 -9.10772979e-01 8.87467206e-01 8.80563498e-01
-8.33165869e-02 -1.29806781e+00 1.17787421e+00 1.57467380e-01
-9.27192628e-01 -5.11581838e-01 -8.12317073e-01 -4.63036150e-01
3.37089211e-01 3.89999151e-02 -2.02485174e-01 8.49503934e-01
4.97737110e-01 6.12616956e-01 -2.03308240e-01 1.32204127e+00
-2.84700215e-01 4.69938546e-01 -1.73695721e-02 -6.91325247e-01
6.30284727e-01 -5.75648248e-01 3.38205278e-01 6.82459772e-01
5.21422505e-01 2.49576941e-01 -7.35618696e-02 5.70224762e-01
4.59028244e-01 2.27376804e-01 -4.78712916e-01 5.08660637e-02
3.75696748e-01 9.31648374e-01 -5.03647506e-01 -3.39581549e-01
-4.66221631e-01 2.39016727e-01 2.96640582e-02 3.35187137e-01
-9.82449830e-01 -6.03989542e-01 1.72322914e-01 -4.44935597e-02
4.96587843e-01 1.15992479e-01 -1.22631378e-01 -8.35827410e-01
-3.38776529e-01 -9.20024514e-01 6.47374511e-01 -4.81919438e-01
-1.22265196e+00 2.04863891e-01 2.75081515e-01 -1.29038346e+00
-5.89594424e-01 -8.95327389e-01 -1.04819536e+00 6.95445240e-01
-1.82225001e+00 -4.31961447e-01 5.00590384e-01 4.62528050e-01
5.54277182e-01 -1.31096351e+00 3.65200430e-01 6.09748438e-02
-4.70369995e-01 4.12976086e-01 7.90053308e-01 -1.83407620e-01
2.87718445e-01 -1.69306552e+00 1.38162181e-03 2.49506891e-01
-1.12429336e-01 4.55792174e-02 6.63104773e-01 -6.39901698e-01
-8.42267871e-01 -3.43101680e-01 2.97312349e-01 -7.80672729e-02
8.86831522e-01 5.94534039e-01 -4.40965474e-01 5.42817116e-01
1.05492198e+00 -5.84166944e-01 5.13640225e-01 -2.83538312e-01
4.54664290e-01 -4.18515146e-01 -1.26568818e+00 3.35031480e-01
3.39970022e-01 5.60658634e-01 -9.88341153e-01 1.80321246e-01
4.64858502e-01 -5.76791503e-02 -1.05684996e+00 1.59600064e-01
6.11933529e-01 -1.20972848e+00 1.14901733e+00 -9.66421902e-01
2.90932387e-01 2.92865068e-01 3.13222378e-01 -1.45758319e+00
-5.32716095e-01 -1.16111326e+00 -6.01859391e-01 1.25807106e+00
5.66204786e-01 -9.95646119e-01 1.11747956e+00 3.50802362e-01
3.51326317e-01 -1.11843085e+00 -1.05367541e+00 -1.24669039e+00
7.22363174e-01 3.92414659e-01 9.78843451e-01 8.17736089e-01
3.27106148e-01 -4.55445684e-02 -7.72716820e-01 -7.17981100e-01
8.90901327e-01 2.20608175e-01 4.19367522e-01 -8.15774620e-01
-8.13750088e-01 -8.46086860e-01 2.59261094e-02 -4.38542306e-01
-1.06183618e-01 -3.56912881e-01 -7.03628778e-01 -1.12646592e+00
-2.89704166e-02 -3.82860184e-01 -8.84121478e-01 1.85220852e-01
-9.62116348e-04 -4.49936718e-01 5.43281138e-01 2.26400211e-01
-3.34888905e-01 6.19745076e-01 1.38796926e+00 3.68799344e-02
-3.00362855e-01 5.27652144e-01 -6.04364634e-01 5.71035624e-01
1.41471493e+00 -5.15240788e-01 -2.41656482e-01 1.24779344e-01
4.60214138e-01 5.99907935e-01 -2.67345548e-01 -7.02760339e-01
-1.66124180e-01 -7.21037030e-01 5.02832115e-01 -6.21966481e-01
-2.62749314e-01 -6.45625830e-01 8.23451858e-03 1.10559130e+00
-2.42594212e-01 5.76708019e-01 3.89851406e-02 3.92546862e-01
-4.29340482e-01 -1.19211805e+00 8.57323766e-01 -6.34842932e-01
-4.09923702e-01 7.31813908e-02 -4.44359601e-01 2.31560618e-01
1.65048194e+00 -2.36477032e-01 -1.98515102e-01 -5.87395608e-01
-8.17697406e-01 7.14500725e-01 -2.94597596e-01 1.10924929e-01
5.80930054e-01 -1.48417485e+00 -8.76208246e-01 -3.56698036e-01
-7.40081012e-01 -7.52326906e-01 -1.04577571e-01 5.53014755e-01
-1.14303184e+00 5.00845373e-01 -9.48081315e-01 3.93648207e-01
-8.00596893e-01 7.65672684e-01 7.43863702e-01 -1.11481869e+00
-2.37713814e-01 6.19708657e-01 -3.02805126e-01 4.42680269e-02
3.55060458e-01 2.59772778e-01 -9.44299698e-01 2.87007838e-01
4.81934756e-01 9.09591734e-01 -6.69509768e-01 1.97705403e-01
4.87132780e-02 5.75851977e-01 -1.16516843e-01 -4.43616301e-01
1.73617041e+00 2.39041522e-01 1.61465585e-01 2.44297355e-01
1.55139834e-01 -4.65061873e-01 -1.34799778e+00 1.80613428e-01
6.03863001e-01 -7.01680005e-01 -3.05566221e-01 -6.98302031e-01
-1.18196332e+00 7.69223213e-01 6.25243008e-01 8.04493248e-01
9.58278060e-01 -7.27612555e-01 5.68586230e-01 3.86704266e-01
4.94665921e-01 -1.90653181e+00 4.97073382e-01 2.27453098e-01
1.03840768e+00 -9.34725702e-01 4.60221499e-01 4.63359892e-01
-1.09569681e+00 1.44689560e+00 6.30235374e-01 -9.68094647e-01
1.00574565e+00 3.26720744e-01 -1.15215510e-01 3.49422581e-02
-1.07101202e+00 -2.09990572e-02 -4.36012447e-01 7.21399963e-01
3.21747571e-01 -6.23850413e-02 -8.55989695e-01 6.01365149e-01
-2.38352105e-01 1.42344236e-01 7.37283945e-01 1.18673027e+00
-9.07341421e-01 -1.69214678e+00 -5.13557673e-01 4.31418926e-01
-7.82908499e-01 1.36599794e-01 -3.31324846e-01 1.27398431e+00
-1.96737900e-01 6.17527783e-01 -6.48271153e-03 2.52008319e-01
5.75976670e-01 -6.82477430e-02 5.73693216e-01 -9.93582904e-02
-1.10156572e+00 4.66902286e-01 1.11657858e-01 -3.50843161e-01
-7.09343076e-01 -1.03702700e+00 -9.28526103e-01 -4.31066483e-01
-3.63227814e-01 6.43163681e-01 6.75124079e-02 6.76060498e-01
-2.70910472e-01 6.47363305e-01 1.26449716e+00 -4.64840978e-01
-1.54206049e+00 -9.38355744e-01 -1.50480843e+00 -1.67422831e-01
-8.54357406e-02 -8.45551372e-01 -3.14913750e-01 -4.65134561e-01] | [4.4214396476745605, 3.840658664703369] |
9e279e84-ecd1-4b44-8f26-9633fc38a277 | bridging-the-gap-between-indexing-and | 2206.10128 | null | https://arxiv.org/abs/2206.10128v3 | https://arxiv.org/pdf/2206.10128v3.pdf | Bridging the Gap Between Indexing and Retrieval for Differentiable Search Index with Query Generation | The Differentiable Search Index (DSI) is an emerging paradigm for information retrieval. Unlike traditional retrieval architectures where index and retrieval are two different and separate components, DSI uses a single transformer model to perform both indexing and retrieval. In this paper, we identify and tackle an important issue of current DSI models: the data distribution mismatch that occurs between the DSI indexing and retrieval processes. Specifically, we argue that, at indexing, current DSI methods learn to build connections between the text of long documents and the identifier of the documents, but then retrieval of document identifiers is based on queries that are commonly much shorter than the indexed documents. This problem is further exacerbated when using DSI for cross-lingual retrieval, where document text and query text are in different languages. To address this fundamental problem of current DSI models, we propose a simple yet effective indexing framework for DSI, called DSI-QG. When indexing, DSI-QG represents documents with a number of potentially relevant queries generated by a query generation model and re-ranked and filtered by a cross-encoder ranker. The presence of these queries at indexing allows the DSI models to connect a document identifier to a set of queries, hence mitigating data distribution mismatches present between the indexing and the retrieval phases. Empirical results on popular mono-lingual and cross-lingual passage retrieval datasets show that DSI-QG significantly outperforms the original DSI model. | ['Daxin Jiang', 'Guido Zuccon', 'Ming Gong', 'Jian Pei', 'Linjun Shou', 'Houxing Ren', 'Shengyao Zhuang'] | 2022-06-21 | null | null | null | null | ['passage-retrieval'] | ['natural-language-processing'] | [ 2.48892739e-01 -2.07591116e-01 -6.02896988e-01 1.91684105e-02
-1.38661659e+00 -9.94599044e-01 1.13270354e+00 3.15658063e-01
-4.54285830e-01 3.85050207e-01 4.35049385e-01 -1.03344150e-01
-7.99242556e-01 -6.20341361e-01 -4.84949708e-01 -1.19729526e-01
4.72475290e-02 1.10779691e+00 5.40371716e-01 -4.38760579e-01
6.05299115e-01 2.66352862e-01 -1.73447335e+00 3.90433311e-01
6.95712686e-01 9.56057191e-01 3.69468987e-01 4.82543528e-01
-5.39781094e-01 3.73038143e-01 -8.67238879e-01 -1.80900604e-01
2.32440323e-01 -3.09495926e-02 -1.12442231e+00 -6.85747206e-01
6.34406209e-01 -2.97659308e-01 -5.62866628e-01 9.05612171e-01
6.15924239e-01 7.92170502e-03 6.95701838e-01 -1.24169946e+00
-8.19166601e-01 8.71336401e-01 -2.24194512e-01 3.15834165e-01
7.28776157e-01 -7.24946558e-01 1.50040233e+00 -1.07812834e+00
8.17669809e-01 1.44684625e+00 3.64980549e-01 3.35193992e-01
-1.05096304e+00 -5.67504525e-01 7.22315907e-02 1.51578739e-01
-1.72190309e+00 -5.37820876e-01 4.19953972e-01 -2.14090720e-01
1.17490554e+00 5.00463545e-01 1.12025537e-01 8.40490043e-01
4.99334037e-02 1.01199257e+00 4.01633471e-01 -6.62612259e-01
-8.59338716e-02 1.15561433e-01 4.83959883e-01 1.11985192e-01
5.54599911e-02 2.69700646e-01 -8.55350375e-01 -5.72320580e-01
2.54571259e-01 -9.66218486e-02 -2.71961272e-01 2.17588991e-02
-1.13077259e+00 8.57684493e-01 2.30359435e-02 5.33792496e-01
-2.66658485e-01 -6.69221058e-02 6.52005434e-01 6.05041087e-01
1.95651010e-01 5.40317714e-01 -4.69604284e-01 1.42740959e-03
-1.27426839e+00 6.02312505e-01 9.65294421e-01 1.31727457e+00
6.18305445e-01 -7.12193072e-01 -6.92243934e-01 1.11665773e+00
6.02960587e-01 7.32496202e-01 8.73705149e-01 -6.32565439e-01
5.51362991e-01 5.24761319e-01 9.20498371e-02 -8.80319655e-01
-4.89696674e-02 -3.86413693e-01 -4.06805187e-01 -4.49216336e-01
-4.16143052e-02 6.07669175e-01 -6.97175860e-01 1.69089985e+00
-5.85074015e-02 -3.92717808e-01 2.42469415e-01 5.56359887e-01
8.37531686e-01 8.09499979e-01 -1.11346312e-01 -2.32360110e-01
1.37622011e+00 -8.31690133e-01 -7.24396050e-01 -3.02298754e-01
7.70234823e-01 -1.21433711e+00 1.05523825e+00 -2.89129727e-02
-1.32631278e+00 -5.74021459e-01 -7.97196507e-01 -4.66832399e-01
-5.72359562e-01 4.41751368e-02 1.04773618e-01 1.47256568e-01
-1.43897450e+00 2.07352862e-01 -3.03669721e-01 -4.45896775e-01
-2.41678074e-01 2.85417885e-01 -5.78222275e-02 -1.19099356e-01
-1.65851605e+00 7.24150956e-01 5.25084913e-01 -2.03937709e-01
-6.74872279e-01 -8.66390526e-01 -5.74069083e-01 2.44345829e-01
3.53827208e-01 -4.31690633e-01 1.44078994e+00 -2.83491999e-01
-9.14138913e-01 8.93634617e-01 -4.97944862e-01 -2.51987457e-01
1.33689865e-01 -3.25264931e-01 -6.00493729e-01 1.75794706e-01
5.39719701e-01 4.65968311e-01 8.01925421e-01 -1.18670475e+00
-8.28664899e-01 -4.00941581e-01 3.50696370e-02 4.27386045e-01
-1.69194832e-01 4.00750130e-01 -1.13288629e+00 -7.04392552e-01
1.91030428e-01 -8.32644403e-01 6.31207585e-01 -2.12201312e-01
-4.61221159e-01 -8.33513677e-01 8.97777498e-01 -4.89843607e-01
1.91804481e+00 -2.27108908e+00 1.36713147e-01 4.52713072e-01
6.64372987e-04 4.61563736e-01 -4.38076437e-01 1.08584547e+00
1.79063275e-01 3.27326119e-01 2.27726460e-01 -3.40805441e-01
3.62132818e-01 3.86637926e-01 -8.77396286e-01 -5.93803301e-02
-3.00793648e-01 9.65918005e-01 -1.04893136e+00 -6.93775654e-01
-4.12218273e-01 4.91428226e-01 -5.09279013e-01 3.71414810e-01
-1.77280352e-01 -8.91656503e-02 -5.41621983e-01 5.50604820e-01
3.89379472e-01 -2.51621366e-01 -9.19636339e-02 -3.70087624e-01
-4.04910594e-02 9.48036909e-01 -9.86958444e-01 1.81783259e+00
-4.51941460e-01 3.87956411e-01 -3.88003699e-03 -6.62038803e-01
7.35824645e-01 5.75072825e-01 5.60124636e-01 -1.29369020e+00
-4.45349038e-01 7.33418167e-01 -4.05935049e-01 -2.06247225e-01
1.01059425e+00 2.64790624e-01 -2.38258526e-01 7.55400836e-01
6.34169281e-02 -1.37200907e-01 8.33662450e-01 5.75787663e-01
9.42970276e-01 -4.94452938e-02 -2.61890292e-01 -2.40886658e-01
7.73028493e-01 -8.37666243e-02 2.60530502e-01 1.45391273e+00
2.39412725e-01 5.64379573e-01 -6.29074723e-02 8.95304903e-02
-9.14979100e-01 -9.52919126e-01 -3.88078362e-01 1.46362007e+00
2.46515974e-01 -8.05649757e-01 -4.27479237e-01 -5.71552753e-01
2.84629256e-01 5.94228029e-01 -1.99412823e-01 -4.75752890e-01
-7.87545741e-01 -2.41337761e-01 6.29800737e-01 2.10692346e-01
2.60360427e-02 -1.06760001e+00 -3.35475236e-01 4.52167183e-01
-5.36465764e-01 -7.22359776e-01 -1.03437495e+00 2.32288077e-01
-4.88328695e-01 -8.08414459e-01 -8.54698420e-01 -8.91072214e-01
2.55533904e-01 5.16093552e-01 1.58756602e+00 4.18534577e-01
-2.00020626e-01 7.69069493e-01 -6.43300235e-01 -2.27305323e-01
-5.33084810e-01 7.60207415e-01 -3.77313718e-02 -4.11939979e-01
6.03537023e-01 -1.60012871e-01 -6.84918761e-01 4.57003146e-01
-1.53266799e+00 -3.99077624e-01 3.20974469e-01 8.36217821e-01
5.67213595e-01 -3.32268625e-02 6.75419331e-01 -6.83143497e-01
8.92123103e-01 -6.05036378e-01 -6.80988967e-01 8.71173382e-01
-1.06892776e+00 4.47355986e-01 1.30170926e-01 -3.79011780e-01
-7.58693457e-01 -8.29688549e-01 -1.08104154e-01 -2.12964982e-01
4.26400185e-01 1.01711917e+00 1.59016386e-01 2.14058086e-01
4.01204646e-01 4.64554250e-01 -7.47477785e-02 -9.05407131e-01
3.42366934e-01 8.04225445e-01 2.87341952e-01 -7.64762700e-01
7.51718521e-01 1.29937813e-01 -4.63616610e-01 -5.54255247e-01
-8.97882760e-01 -9.82595384e-01 -4.44975853e-01 9.42127500e-03
3.78287792e-01 -7.37149656e-01 -4.15204406e-01 3.53835583e-01
-1.21471441e+00 2.76430398e-02 -5.31210005e-01 2.60703772e-01
-3.78859460e-01 2.63509482e-01 -6.08696699e-01 -4.93084699e-01
-6.84458494e-01 -1.23736930e+00 1.57890272e+00 -4.12374400e-02
-4.33529854e-01 -1.05620372e+00 3.26688856e-01 7.25149512e-02
8.01158726e-01 -8.07756722e-01 1.31181407e+00 -1.15262735e+00
-5.36813855e-01 -4.61935580e-01 -2.95772552e-01 -3.62760969e-03
2.48557732e-01 -1.89777672e-01 -7.99292445e-01 -7.36632228e-01
-1.14592068e-01 -5.25702119e-01 5.32297432e-01 4.68168147e-02
7.85641909e-01 -4.41466391e-01 -5.75305164e-01 3.99877161e-01
1.48284364e+00 4.73726034e-01 4.14603293e-01 4.50564086e-01
2.95263588e-01 6.54602587e-01 5.95130622e-01 2.74785250e-01
4.79007035e-01 1.07752299e+00 1.52682345e-02 1.58640012e-01
-3.98149043e-01 -5.52256644e-01 1.77860171e-01 1.27855301e+00
6.63033843e-01 -5.98435879e-01 -8.74304354e-01 6.57717526e-01
-1.57983100e+00 -9.22980845e-01 4.02146548e-01 2.58636856e+00
1.21400118e+00 2.13985965e-02 -2.26247370e-01 1.17774777e-01
3.18198413e-01 1.02891274e-01 -4.60551322e-01 -6.43646717e-02
-2.75810868e-01 3.86920631e-01 4.64975655e-01 8.27970505e-01
-7.00939834e-01 7.81357527e-01 7.09358215e+00 9.86333013e-01
-9.66842771e-01 -1.32336766e-01 -7.54953697e-02 -1.92028448e-01
-6.80306017e-01 1.28686056e-01 -1.49469995e+00 5.47501028e-01
1.10828197e+00 -7.64096737e-01 5.33213675e-01 4.21390951e-01
-2.79333711e-01 1.31910086e-01 -1.45806074e+00 9.35642004e-01
2.17435762e-01 -1.24565995e+00 6.15763903e-01 1.20297603e-01
2.89639056e-01 2.15021119e-01 2.54400253e-01 6.37433290e-01
2.45089173e-01 -7.88560748e-01 9.89004493e-01 6.97364509e-01
8.13189566e-01 -2.66448706e-01 5.02139270e-01 2.11496249e-01
-1.31633902e+00 -9.22982320e-02 -2.51680762e-01 6.63539648e-01
1.35870844e-01 1.81996524e-01 -4.52132434e-01 7.04943538e-01
6.24626219e-01 5.87776601e-01 -6.69865906e-01 1.04339826e+00
3.16814333e-01 2.78180242e-01 -5.16034186e-01 1.88160047e-01
3.35334927e-01 -1.80727765e-01 6.22291505e-01 1.20616627e+00
5.00024676e-01 -4.44673657e-01 2.99617440e-01 5.59879184e-01
-2.63885129e-02 1.13698840e-01 -7.19516337e-01 -1.42115191e-01
1.10464251e+00 5.87733269e-01 -1.88328698e-01 -5.90277135e-01
-5.07472634e-01 9.96395469e-01 1.83117270e-01 6.90895617e-01
-2.92968184e-01 -5.19409955e-01 5.31407535e-01 1.08227015e-01
2.32550919e-01 -1.70581825e-02 4.90186602e-01 -1.04846191e+00
3.30887645e-01 -1.17189217e+00 8.24711978e-01 -6.45934403e-01
-1.30957425e+00 5.91518164e-01 4.83813077e-01 -1.19997191e+00
-9.03437376e-01 -1.82751387e-01 7.80999660e-02 1.30512369e+00
-2.06489468e+00 -9.66142237e-01 2.67462939e-01 6.80554926e-01
5.96786201e-01 -1.55125782e-01 1.06382620e+00 8.56628060e-01
-3.30050997e-02 8.23739052e-01 6.58142090e-01 6.56811967e-02
8.75639319e-01 -1.03893495e+00 3.50606382e-01 3.70194137e-01
3.83823872e-01 1.12340677e+00 2.91363090e-01 -6.61342025e-01
-1.59267795e+00 -6.85765803e-01 1.65020084e+00 -4.61733371e-01
6.31315351e-01 -9.64726582e-02 -1.24833012e+00 5.44007897e-01
2.57849127e-01 -3.50617349e-01 6.87619746e-01 2.07046699e-03
-5.42443931e-01 -2.08910793e-01 -7.25390434e-01 3.09770823e-01
8.80348921e-01 -1.24129415e+00 -8.67865145e-01 6.16894245e-01
7.78544307e-01 -2.89166361e-01 -9.43767726e-01 3.00674826e-01
5.98376274e-01 -5.38053989e-01 1.39665091e+00 -4.09650087e-01
-1.11180224e-01 -2.02762797e-01 -4.14085627e-01 -9.42227840e-01
-2.51842499e-01 -5.93628466e-01 -2.78837442e-01 1.32338393e+00
4.51436907e-01 -5.82108617e-01 1.99079007e-01 4.97530282e-01
1.28891366e-02 -3.27863812e-01 -1.03468108e+00 -1.03975713e+00
3.66795152e-01 -2.06203595e-01 7.78721333e-01 4.27966863e-01
-3.95862386e-02 3.81015629e-01 5.57898544e-02 -9.39738750e-02
5.72011888e-01 1.20730437e-01 3.51186305e-01 -1.49605715e+00
-7.23131672e-02 -6.55868828e-01 1.06818393e-01 -1.51812136e+00
1.45797282e-01 -1.33316565e+00 2.05031806e-03 -1.43432474e+00
1.20786063e-01 -8.83435190e-01 -4.08285111e-01 2.85507530e-01
-8.85802805e-02 -8.08384120e-02 5.88316210e-02 9.96213913e-01
-5.17622530e-01 4.56084043e-01 8.49627852e-01 -3.90722364e-01
-1.06294081e-01 -5.92392311e-02 -5.19695759e-01 8.91517550e-02
2.03552827e-01 -6.56897128e-01 -8.58533025e-01 -8.87362957e-01
5.54090977e-01 1.55652836e-01 -3.87299471e-02 -6.60241246e-01
7.58940995e-01 2.03722313e-01 -2.47435719e-01 -1.05756998e+00
1.98758811e-01 -7.73828387e-01 4.87661026e-02 1.86219275e-01
-9.26259518e-01 4.68441993e-01 1.63980071e-02 3.31010789e-01
-7.20659971e-01 -6.96530581e-01 2.54341215e-01 -1.10544637e-01
-3.90604764e-01 4.25737113e-01 -3.92543256e-01 5.87675989e-01
2.13599458e-01 1.90874543e-02 -2.83419013e-01 -2.08121359e-01
-3.76091570e-01 3.66727710e-01 2.81907707e-01 8.14048052e-01
5.61159313e-01 -1.43050766e+00 -7.16395736e-01 4.26131636e-01
5.80226243e-01 -7.37980083e-02 -2.84328312e-01 3.65143448e-01
-1.88956838e-02 1.16766655e+00 4.45503026e-01 -4.44460928e-01
-1.02709174e+00 5.48907042e-01 5.20772785e-02 -7.04981089e-01
-5.43439031e-01 7.55228102e-01 8.77763256e-02 -4.63090718e-01
6.98982775e-01 -1.40218243e-01 -2.82156467e-01 3.98588151e-01
7.93990493e-01 1.08899251e-01 4.27815408e-01 -8.09095800e-01
-2.58385330e-01 6.35388732e-01 -7.38151908e-01 -5.68198204e-01
8.92943382e-01 -5.10377288e-01 -4.22911495e-01 5.33728063e-01
1.64673233e+00 -7.31702447e-02 -3.19427222e-01 -1.11718357e+00
6.35220766e-01 -4.07947689e-01 1.32824838e-01 -8.43642056e-01
-7.42154658e-01 5.84118366e-01 5.87230980e-01 4.91227031e-01
1.02555895e+00 2.26716816e-01 9.91901577e-01 4.16137010e-01
4.64336962e-01 -1.25858736e+00 3.95904519e-02 9.78764892e-01
1.06987143e+00 -9.48512852e-01 -2.34265998e-01 1.35581553e-01
-9.93645787e-02 9.33656693e-01 7.17446133e-02 4.09210414e-01
8.10415268e-01 6.74807355e-02 2.34145314e-01 -5.97134292e-01
-7.03318536e-01 -1.05593860e-01 7.79317975e-01 3.00536990e-01
5.39477825e-01 -4.67498511e-01 -4.59684819e-01 1.65103823e-02
-2.79334068e-01 -4.31149378e-02 -1.87901765e-01 1.36772823e+00
-1.52565166e-01 -1.54665077e+00 -2.76550591e-01 4.69684869e-01
-5.54439127e-01 -5.78243017e-01 -1.79611072e-01 5.15434682e-01
-4.81239319e-01 9.22204018e-01 2.26128101e-01 -9.85478535e-02
2.85747826e-01 4.30415452e-01 3.67201194e-02 -6.80069506e-01
-6.33038461e-01 7.71741346e-02 -1.59848183e-01 -6.37405396e-01
-4.16327894e-01 -5.05015433e-01 -8.78669441e-01 1.47037134e-01
-5.43842196e-01 7.71106482e-01 8.05713713e-01 7.70176411e-01
6.02255940e-01 1.70275658e-01 7.27836370e-01 -3.45571637e-01
-1.05961061e+00 -9.46452081e-01 -5.19229770e-01 5.83046556e-01
4.93321985e-01 -5.89992762e-01 -4.68986392e-01 -1.14652842e-01] | [11.540877342224121, 7.64747953414917] |
d358f74d-22e9-4f35-a633-d1a288ed31fa | assessing-rate-limits-using-behavioral-and | 2305.03297 | null | https://arxiv.org/abs/2305.03297v1 | https://arxiv.org/pdf/2305.03297v1.pdf | Assessing Rate limits Using Behavioral and Neural Responses of Interaural-Time-Difference Cues in Fine-Structure and Envelope | The objective was to determine the effect of pulse rate on the sensitivity to use interaural-time-difference (ITD) cues and to explore the mechanisms behind rate-dependent degradation in ITD perception in bilateral cochlear implant (CI) listeners using CI simulations and electroencephalogram (EEG) measures. To eliminate the impact of CI stimulation artifacts and to develop protocols for the ongoing bilateral CI studies, upper-frequency limits for both behavior and EEG responses were obtained from normal hearing (NH) listeners using sinusoidal-amplitude-modulated (SAM) tones and filtered clicks with changes in either fine structure ITD or envelope ITD. Multiple EEG responses were recorded, including the subcortical auditory steady-state responses (ASSRs) and cortical auditory evoked potentials (CAEPs) elicited by stimuli onset, offset, and changes. Results indicated that acoustic change complex (ACC) responses elicited by envelope ITD changes were significantly smaller or absent compared to those elicited by fine structure ITD changes. The ACC morphologies evoked by fine structure ITD changes were similar to onset and offset CAEPs, although smaller than onset CAEPs, with the longest peak latencies for ACC responses and shortest for offset CAEPs. The study found that high-frequency stimuli clearly elicited subcortical ASSRs, but smaller than those evoked by lower carrier frequency SAM tones. The 40-Hz ASSRs decreased with increasing carrier frequencies. Filtered clicks elicited larger ASSRs compared to high-frequency SAM tones, with the order being 40-Hz-ASSR>160-Hz-ASSR>80-Hz-ASSR>320-Hz-ASSR for both stimulus types. Wavelet analysis revealed a clear interaction between detectable transient CAEPs and 40-Hz-ASSRs in the time-frequency domain for SAM tones with a low carrier frequency. | ['Deborah Vickers', 'Birger Kollmeier', 'Stephan Ewert', 'Hongmei Hu'] | 2023-05-05 | null | null | null | null | ['eeg', 'eeg'] | ['methodology', 'time-series'] | [ 4.33870964e-02 -5.52940607e-01 5.20806015e-01 1.52540803e-01
-8.89964044e-01 -6.89295590e-01 2.09230796e-01 4.53535408e-01
-7.73702025e-01 5.95276535e-01 5.15091896e-01 -2.27623180e-01
-2.36384585e-01 -1.59795508e-01 -5.37373722e-01 -6.14168644e-01
-5.32497764e-01 -3.54684770e-01 5.58778226e-01 -4.85283509e-02
2.50931293e-01 3.35770071e-01 -1.82599306e+00 1.69762343e-01
6.01920009e-01 1.30467522e+00 5.10378003e-01 7.49955475e-01
3.53537321e-01 3.98612283e-02 -1.05817139e+00 1.05571873e-01
2.16720715e-01 -5.83259463e-01 -2.88513571e-01 -6.15967393e-01
3.37152854e-02 -4.92078513e-02 -1.55017689e-01 8.30738306e-01
1.13129783e+00 -1.80775329e-01 7.44192421e-01 -1.11343491e+00
-4.28115636e-01 6.41984582e-01 -4.18360204e-01 8.63314927e-01
6.57292366e-01 1.28337638e-02 2.30186626e-01 -9.63036895e-01
2.60600597e-01 8.59953701e-01 5.66029429e-01 3.17747742e-01
-1.50075877e+00 -1.07354248e+00 -1.43637806e-01 2.86488622e-01
-1.56537223e+00 -4.39319253e-01 6.60149813e-01 -6.22484386e-01
1.25135827e+00 2.77650774e-01 9.48828399e-01 8.94309103e-01
5.83855569e-01 -1.72379434e-01 1.22669983e+00 -5.29795289e-01
3.42687488e-01 5.83870858e-02 8.36923998e-03 -4.95820820e-01
1.41624361e-01 4.38784301e-01 -4.89808142e-01 -3.72260123e-01
8.03660095e-01 -8.44491780e-01 -9.50463235e-01 8.19634914e-01
-8.54701817e-01 3.65790576e-01 2.47584749e-02 9.42457259e-01
-5.82197666e-01 1.47897601e-01 -3.49027477e-02 4.24274385e-01
1.77954122e-01 7.02717841e-01 -3.24254960e-01 -7.26948798e-01
-6.02431297e-01 5.05718142e-02 3.95109266e-01 4.11333591e-01
1.49305865e-01 3.86376649e-01 -3.09520036e-01 1.19917488e+00
-1.12131536e-01 6.29527271e-01 3.96180063e-01 -6.23113751e-01
-1.13004602e-01 -6.32432997e-01 3.81795876e-02 -7.25540459e-01
-5.93719840e-01 -3.59014899e-01 -1.94641918e-01 2.51153260e-01
3.67774129e-01 -1.22320205e-01 -9.62348521e-01 2.01178312e+00
-2.59481698e-01 1.60749823e-01 -3.50889593e-01 8.60348165e-01
3.64298373e-01 4.69364375e-01 3.15020442e-01 -9.14474428e-01
1.80741298e+00 2.47363836e-01 -1.11772454e+00 -3.71133327e-01
2.08176263e-02 -1.01048577e+00 1.17505097e+00 1.72810495e-01
-1.20072079e+00 -6.60669088e-01 -9.08322990e-01 5.70391297e-01
1.80622190e-01 -6.00243092e-01 -3.26823860e-01 1.03311539e+00
-1.07660842e+00 3.21052819e-01 -6.73591793e-01 -7.94789121e-02
-1.94161117e-01 7.47086257e-02 -2.19384238e-01 4.67708647e-01
-1.32819498e+00 1.15724659e+00 -2.25168422e-01 -7.12417662e-02
-7.52781630e-01 -1.26090086e+00 -4.33195055e-01 1.50317088e-01
-3.18017781e-01 -2.20492229e-01 1.43629968e+00 -3.96687537e-01
-1.19876444e+00 5.68075478e-01 -2.56562442e-01 -2.58335322e-01
-1.84108853e-01 1.50867105e-01 -9.25274968e-01 3.73195380e-01
-9.29493085e-02 3.44291806e-01 7.27044761e-01 -1.01121616e+00
-3.32161561e-02 -2.46875897e-01 -8.17657053e-01 -1.02246784e-01
-3.29328440e-02 6.27623677e-01 4.02770102e-01 -7.07662284e-01
2.83452332e-01 -6.71236217e-01 1.19353257e-01 -5.42575717e-01
-2.83174068e-02 -1.26391798e-01 3.37353855e-01 -6.86688364e-01
1.33070815e+00 -2.62010384e+00 -5.95372558e-01 8.70859995e-02
-4.10419814e-02 -8.47649500e-02 -4.18522835e-01 5.34613073e-01
-5.30506074e-01 3.61029059e-02 -3.33326638e-01 4.21112329e-01
1.12335436e-01 -6.05865896e-01 -2.60613501e-01 3.76980275e-01
-1.80660430e-02 3.32693934e-01 -5.86648881e-01 1.20290823e-01
-2.44454533e-01 7.24714458e-01 -3.42227936e-01 6.45536333e-02
4.76703346e-01 3.27942878e-01 3.75270247e-01 5.89736640e-01
9.25917745e-01 7.23093510e-01 -4.54648316e-01 -6.05610788e-01
-6.35260284e-01 5.12335658e-01 -8.14801574e-01 1.01936889e+00
-2.36355737e-01 7.91650295e-01 2.68573850e-01 -4.03996706e-01
1.04021716e+00 7.16877639e-01 4.44820076e-02 -1.27637327e+00
2.60865122e-01 1.03624308e+00 9.72417891e-01 -4.43426698e-01
-2.06447124e-01 -9.62380588e-01 9.83460806e-03 2.88509607e-01
-7.50125572e-02 -8.32779586e-01 -9.56508368e-02 1.72366705e-02
8.23508620e-01 -6.14717484e-01 8.98445919e-02 -3.40313613e-01
8.65871161e-02 -8.66526127e-01 2.11676970e-01 5.67789912e-01
-1.81373790e-01 6.72288299e-01 4.01180387e-01 3.33218962e-01
-4.99466717e-01 -1.53158522e+00 -9.30156767e-01 7.88913369e-01
1.45143896e-01 -1.54839873e-01 -6.13348544e-01 4.06665266e-01
1.32882848e-01 1.12520719e+00 -1.18189536e-01 -5.13586402e-01
-2.37578794e-01 -4.62544739e-01 6.24131799e-01 5.25755167e-01
-7.93015286e-02 -1.35810161e+00 -6.10065758e-01 5.06726384e-01
-5.10521531e-01 -1.26802862e+00 -7.31665134e-01 5.77170312e-01
-5.07636666e-01 -2.94514507e-01 -8.18839610e-01 -7.80211747e-01
1.92708492e-01 -5.52515924e-01 6.60485804e-01 -7.04603851e-01
-7.32332110e-01 8.53962183e-01 -1.76284388e-01 -1.08479548e+00
-2.02420503e-02 -1.01519501e+00 6.23965740e-01 -4.57528293e-01
2.46709928e-01 -1.23302758e+00 -6.11348152e-01 3.78829390e-01
-7.48109639e-01 -7.11202025e-01 6.19718492e-01 3.04703236e-01
5.11979163e-01 7.74526224e-02 1.36117089e+00 5.10407746e-01
1.21920109e+00 -1.61270604e-01 -3.88603099e-02 -4.10031915e-01
-2.93683916e-01 -4.35573190e-01 3.75432760e-01 -1.07865274e+00
-1.04730737e+00 -8.36993217e-01 -2.85146207e-01 -1.59520283e-01
-5.40678561e-01 5.44519603e-01 2.29984567e-01 -7.67114088e-02
1.14336157e+00 4.57276613e-01 -2.39223257e-01 -3.87037516e-01
-5.51675975e-01 9.72067475e-01 1.01865590e+00 -5.47456682e-01
4.34772134e-01 -4.16545346e-02 -5.40315270e-01 -9.53069925e-01
-1.38979435e-01 5.61764017e-02 1.41893432e-01 -2.57119238e-01
8.85856271e-01 -6.98908031e-01 -5.32834172e-01 9.13103461e-01
-8.57160389e-01 -2.94597954e-01 1.20630533e-01 1.48830640e+00
-5.90185285e-01 6.36731461e-02 -1.05024183e+00 -1.13528109e+00
-4.66653436e-01 -9.82892752e-01 3.77404511e-01 3.75043690e-01
-6.16612554e-01 -9.60611776e-02 2.68561721e-01 -1.02839194e-01
8.99445713e-01 -7.82458391e-03 1.32737553e+00 -6.88418865e-01
7.91158527e-02 -2.98201442e-01 3.55890781e-01 3.77203554e-01
4.34140593e-01 -1.89080343e-01 -1.16426396e+00 -3.44595909e-02
6.18235826e-01 -1.81764901e-01 1.16885100e-02 1.22867692e+00
8.53465617e-01 3.18168476e-02 -1.44651666e-01 2.79532999e-01
1.31248975e+00 1.21443844e+00 8.97266269e-01 -1.11782916e-01
-5.19635499e-01 4.26463723e-01 2.41384163e-01 1.64337158e-01
-4.91939247e-01 3.06768298e-01 3.98556739e-02 -1.10061489e-01
-2.97663063e-01 -9.08687115e-02 2.13867351e-01 9.24337506e-01
-8.60029086e-02 -5.62149286e-02 -8.05240214e-01 8.44099700e-01
-4.04336065e-01 -8.11410904e-01 -4.68769550e-01 2.53414083e+00
1.08783627e+00 2.64533877e-01 -6.17679656e-02 4.81677562e-01
7.64332533e-01 -5.18469930e-01 -3.22595060e-01 -7.86351204e-01
-6.89360127e-02 1.21760392e+00 -1.72211565e-02 3.23654294e-01
-5.27680442e-02 3.28214467e-02 6.47275066e+00 5.68489373e-01
-1.52656198e+00 5.60887866e-02 3.76164205e-02 -2.48800546e-01
-6.06794417e-01 -3.33641112e-01 -3.45031649e-01 5.06919444e-01
1.37977493e+00 -6.83727145e-01 3.23111743e-01 1.39465496e-01
5.93866110e-01 -7.33825505e-01 -1.04253793e+00 1.02457893e+00
-4.08225983e-01 -7.71102548e-01 -4.23489422e-01 -4.12439629e-02
7.40313083e-02 1.24176256e-01 1.67316109e-01 1.91507917e-02
-6.27546310e-01 -1.16381419e+00 1.06451738e+00 4.66159582e-01
1.56699073e+00 -7.33182788e-01 5.98371625e-01 1.05306566e-01
-1.39498794e+00 -1.36769325e-01 -2.98561990e-01 -3.58644873e-01
2.63211399e-01 5.50020278e-01 -9.84233975e-01 -4.32997793e-02
9.03890610e-01 -2.89928734e-01 1.19397558e-01 1.55014348e+00
4.67978083e-02 1.08607829e+00 -7.02067852e-01 9.83144790e-02
-3.41837347e-01 2.92642236e-01 6.48759782e-01 1.18277752e+00
8.07907045e-01 4.81905639e-01 -7.42682338e-01 9.85414982e-01
4.86248136e-01 -2.28122085e-01 -2.55967468e-01 -2.11592019e-01
1.24139977e+00 9.10860121e-01 -6.76395357e-01 3.50100279e-01
-3.40554863e-01 3.80970299e-01 -8.31982434e-01 8.55718255e-01
-7.57586181e-01 -1.24323547e+00 4.13678646e-01 3.92468631e-01
3.05224536e-03 6.87813014e-02 -1.37539327e-01 -1.57894641e-01
1.01913728e-01 -7.35248208e-01 -1.41633116e-03 -1.07135785e+00
-1.17898214e+00 7.63618052e-01 2.39243403e-01 -1.14601314e+00
-1.57372668e-01 -1.30599231e-01 -9.25491452e-01 1.45112491e+00
-1.03116262e+00 -1.15344696e-01 9.11228210e-02 9.45281744e-01
4.50990237e-02 6.87331855e-01 8.52576315e-01 3.63851190e-01
2.30096489e-01 5.11392474e-01 -3.85984361e-01 -2.63916790e-01
7.46776521e-01 -8.72939587e-01 -1.88383058e-01 4.65493828e-01
-4.65784520e-01 6.48522675e-01 9.19450700e-01 -4.74916458e-01
-7.19550192e-01 -4.26463008e-01 9.83532667e-01 2.94168144e-01
5.31595945e-01 -6.09245360e-01 -8.43394995e-01 2.24290743e-01
5.27350344e-02 -2.59966314e-01 7.52079189e-01 -4.97474611e-01
-3.32789153e-01 -2.54145384e-01 -1.14586592e+00 5.39678991e-01
8.35529208e-01 -9.12896097e-01 -8.42267334e-01 -2.82156825e-01
8.74545813e-01 9.19792801e-02 -7.94103622e-01 6.94149077e-01
8.72498751e-01 -8.58885944e-01 6.88052773e-01 1.81870744e-01
-9.98634100e-02 -3.97120237e-01 -2.99870279e-02 -1.57241488e+00
-4.77471739e-01 -6.89936280e-01 7.93672800e-01 1.04867136e+00
3.95265788e-01 -1.07867062e+00 -4.85415161e-01 2.12658092e-01
-6.50369525e-01 -1.14279650e-01 -1.07165527e+00 -1.07289910e+00
7.89622888e-02 -7.30572999e-01 -1.19378835e-01 1.80125773e-01
5.81506133e-01 8.91909078e-02 4.51132596e-01 3.37342203e-01
3.72370303e-01 -4.41194832e-01 -4.66893107e-01 -8.84563923e-01
-2.34402537e-01 -4.16787654e-01 -5.13617754e-01 -3.07907134e-01
-5.59933901e-01 -5.66739917e-01 4.30624783e-01 -1.48006892e+00
-2.21375450e-01 5.89574799e-02 -5.75467765e-01 1.04629621e-01
-1.51957065e-01 1.26825303e-01 4.00476865e-02 -2.69422472e-01
1.02229130e+00 3.73515278e-01 8.54897141e-01 1.07051851e-02
-6.06748283e-01 2.17062354e-01 -4.17986393e-01 6.51665926e-01
2.87135780e-01 -7.81369269e-01 -6.27580643e-01 1.44933119e-01
-2.98271608e-02 6.99381649e-01 2.31907830e-01 -1.28081322e+00
1.81061938e-01 1.85697064e-01 4.60176975e-01 -7.16328979e-01
4.03090954e-01 -3.73781115e-01 7.73586750e-01 7.11813390e-01
-1.05825789e-01 -1.98875442e-01 8.23868513e-01 2.22730383e-01
-1.98034987e-01 -1.60189018e-01 9.24727559e-01 1.81421012e-01
1.03104368e-01 -2.55385339e-01 -1.25727248e+00 1.08624727e-01
1.02014363e+00 -4.55329031e-01 -3.84302795e-01 -7.31816292e-01
-9.74792540e-01 -6.46291316e-01 -3.79182473e-02 2.13385925e-01
8.55321407e-01 -1.12818909e+00 -7.43740320e-01 4.20470595e-01
-2.22980697e-02 -5.85143507e-01 7.26481259e-01 1.54812682e+00
5.41500412e-02 3.63853782e-01 -4.68716204e-01 -6.35925531e-01
-1.13792741e+00 1.73858836e-01 6.27539873e-01 6.35949433e-01
-3.17194790e-01 1.55820322e+00 6.30592823e-01 7.02369094e-01
4.60420400e-01 -3.87357920e-01 -1.16295181e-01 2.94072360e-01
7.72565067e-01 2.60037780e-01 5.26539862e-01 1.32844985e-01
-7.24423885e-01 7.94074357e-01 4.31903720e-01 -7.50244200e-01
1.02043581e+00 -7.52258897e-02 4.50767726e-02 9.31658983e-01
1.24135840e+00 4.95323807e-01 -5.58983088e-01 3.40789914e-01
-2.20344558e-01 -4.17801887e-01 -1.19705029e-01 -1.34980714e+00
-6.59712613e-01 9.48259711e-01 9.92584825e-01 3.20100307e-01
1.66563272e+00 4.97487605e-01 7.42310762e-01 -5.57518125e-01
7.65214145e-01 -8.79720211e-01 2.99471080e-01 -1.97888732e-01
1.22062242e+00 3.52154762e-01 -5.74558794e-01 -1.80050537e-01
-3.74925882e-01 1.04729378e+00 4.45340306e-01 1.94283247e-01
9.56664205e-01 7.80293226e-01 5.75458333e-02 7.11595342e-02
-7.42716253e-01 -2.04217359e-01 1.88227624e-01 9.14349914e-01
6.31107330e-01 6.58892691e-02 -9.29434836e-01 9.33977723e-01
-7.87698030e-01 1.74017549e-01 7.31648624e-01 6.34025574e-01
-4.69696730e-01 -4.33647215e-01 -7.66197085e-01 7.25250661e-01
-4.84705597e-01 -3.48705024e-01 -1.45995602e-01 5.69256902e-01
9.70387235e-02 1.22709739e+00 6.43811226e-01 -5.72231114e-01
8.42745900e-01 4.90013957e-01 8.67457628e-01 -2.79557288e-01
-4.78426874e-01 1.03494990e+00 1.06705464e-01 -3.02594304e-01
5.27703539e-02 -6.05190039e-01 -1.57930827e+00 3.23883832e-01
-5.94680667e-01 6.91995621e-02 9.11742210e-01 4.75544363e-01
6.47295043e-02 7.12216318e-01 4.72497612e-01 -5.45772552e-01
-3.39445055e-01 -1.44258583e+00 -1.28455877e+00 -4.55028154e-02
5.60891509e-01 -4.89688814e-01 -1.16660213e+00 -1.22664884e-01] | [13.35891056060791, 3.57037353515625] |
93cc23d7-4023-4306-a7e0-aff0dcaff1ef | human-in-the-loop-how-to-effectively-create | 2212.09422 | null | https://arxiv.org/abs/2212.09422v1 | https://arxiv.org/pdf/2212.09422v1.pdf | Human in the loop: How to effectively create coherent topics by manually labeling only a few documents per class | Few-shot methods for accurate modeling under sparse label-settings have improved significantly. However, the applications of few-shot modeling in natural language processing remain solely in the field of document classification. With recent performance improvements, supervised few-shot methods, combined with a simple topic extraction method pose a significant challenge to unsupervised topic modeling methods. Our research shows that supervised few-shot learning, combined with a simple topic extraction method, can outperform unsupervised topic modeling techniques in terms of generating coherent topics, even when only a few labeled documents per class are used. | ['Benjamin Säfken', 'Christoph Weisser', 'Anton Thielmann'] | 2022-12-19 | null | null | null | null | ['document-classification'] | ['natural-language-processing'] | [ 1.52473629e-01 1.19616084e-01 -7.44490743e-01 -4.96406138e-01
-1.10325289e+00 -6.33872002e-02 9.71677482e-01 6.12473607e-01
-2.26820290e-01 7.19033480e-01 3.53542417e-01 1.65648147e-01
8.95389691e-02 -7.98613787e-01 -1.29540533e-01 -5.15735865e-01
-1.75590850e-02 7.28738487e-01 3.03627759e-01 -1.44565515e-02
4.21428382e-01 -3.77875656e-01 -1.67104399e+00 1.44358382e-01
6.79490328e-01 4.40948188e-01 2.27278978e-01 5.92024803e-01
-1.00844252e+00 8.00884783e-01 -6.76957846e-01 -1.51359960e-01
-3.96428049e-01 -3.03033352e-01 -7.63320386e-01 4.17511672e-01
3.08185726e-01 -3.31379443e-01 -2.59982288e-01 7.94235051e-01
2.47239038e-01 8.40976655e-01 9.33114827e-01 -1.12035155e+00
-7.67934322e-02 8.41726720e-01 -5.12398064e-01 2.41548851e-01
2.32243970e-01 -3.83147508e-01 1.27357638e+00 -1.23937702e+00
9.51192498e-01 1.16812742e+00 4.88619894e-01 3.83152872e-01
-1.17642772e+00 -5.23767889e-01 3.42747658e-01 6.43305928e-02
-1.30782533e+00 -5.23877680e-01 7.41115689e-01 -5.94657421e-01
1.24315894e+00 -2.19081640e-01 4.07387137e-01 9.57880020e-01
2.58670211e-01 8.28628898e-01 6.36920512e-01 -7.07054317e-01
7.32318759e-01 4.75676805e-01 1.06226218e+00 2.71953374e-01
3.87198210e-01 -5.41802764e-01 -7.57772028e-01 -5.52479386e-01
5.08473441e-02 4.74600166e-01 2.54604727e-01 -2.90209234e-01
-6.87756896e-01 1.52592456e+00 -2.30739936e-01 6.48512423e-01
-4.07097608e-01 1.13330901e-01 6.47687137e-01 1.52714392e-02
1.27123916e+00 6.28312469e-01 -2.23458201e-01 -4.25627708e-01
-1.58159614e+00 4.54043835e-01 1.11399770e+00 1.34418166e+00
9.73499477e-01 1.75700143e-01 -5.01481116e-01 1.10862911e+00
1.18802063e-01 2.35265642e-01 8.78883302e-01 -6.71905518e-01
1.21666335e-01 4.04704332e-01 1.50464863e-01 -8.55838060e-01
-1.57649815e-01 -1.41414002e-01 -2.93204665e-01 -4.45509225e-01
-2.69799203e-01 -2.13392884e-01 -1.26209283e+00 1.21186924e+00
2.37757415e-01 2.15103790e-01 -8.54244232e-02 1.30838096e-01
7.29145467e-01 9.38744605e-01 5.14648914e-01 -8.04024041e-01
1.46732640e+00 -9.22180057e-01 -1.21338105e+00 -4.66405332e-01
9.27461088e-01 -7.09004939e-01 9.06882763e-01 2.99165666e-01
-7.75777876e-01 -1.51584327e-01 -8.82272601e-01 3.39594707e-02
-6.79286361e-01 -4.12205160e-01 1.06054795e+00 7.98075974e-01
-5.36001325e-01 3.64031553e-01 -9.19608057e-01 -8.67090166e-01
6.17604554e-01 -8.14678445e-02 1.26191461e-02 -5.07800281e-01
-1.09028983e+00 6.99606836e-01 5.60543299e-01 -9.76377904e-01
-1.00644314e+00 -7.49691784e-01 -1.07408357e+00 2.67206401e-01
7.36566901e-01 -3.77411783e-01 1.61268985e+00 -2.27217376e-01
-1.15306532e+00 4.32820499e-01 -7.47258723e-01 -6.64685607e-01
-1.99496508e-01 -4.50870812e-01 -2.88004190e-01 3.30794811e-01
3.52202058e-01 5.84815979e-01 7.99598873e-01 -1.06230617e+00
-6.61339819e-01 -2.02753738e-01 -3.43256027e-01 2.60669142e-01
-7.95830607e-01 3.88483971e-01 -2.89860964e-01 -7.31378555e-01
4.96210940e-02 -5.31178355e-01 -6.06202662e-01 -3.87110084e-01
-2.74802089e-01 -6.44935727e-01 1.21164548e+00 -1.59720004e-01
1.23213673e+00 -1.92331624e+00 -3.32714736e-01 -1.52642712e-01
2.53627062e-01 3.22023802e-03 3.18024784e-01 6.01128459e-01
1.63510531e-01 2.13904545e-01 -4.56408336e-04 -7.66847491e-01
-8.23153853e-02 9.03577581e-02 -7.97854304e-01 2.53690094e-01
-1.16553806e-01 5.48744261e-01 -1.25455987e+00 -9.45002794e-01
3.55876416e-01 2.13424057e-01 -2.33956620e-01 9.65933874e-02
-5.51544368e-01 -3.59254003e-01 -3.68566781e-01 5.97647607e-01
2.18370661e-01 -4.14529681e-01 2.07635611e-02 4.52252507e-01
-1.39158936e-02 4.22958821e-01 -1.04920566e+00 1.89472616e+00
-3.53337616e-01 8.62234056e-01 -3.99191171e-01 -9.42115009e-01
7.26179719e-01 7.76843846e-01 6.71255171e-01 8.65186080e-02
2.16243893e-01 -1.51803315e-01 -3.96174103e-01 -2.54019707e-01
1.01128304e+00 -7.38969624e-01 -1.72709838e-01 9.65896070e-01
7.12529302e-01 -4.09137696e-01 6.98143184e-01 8.33496213e-01
9.85416055e-01 -4.17521715e-01 5.67134321e-01 -9.44771096e-02
-5.12820125e-01 5.38972795e-01 3.60010356e-01 1.07815266e+00
-1.30624816e-01 7.05047786e-01 2.23047704e-01 -1.77834347e-01
-1.04174721e+00 -6.73731923e-01 -2.23109558e-01 1.67112589e+00
-5.86185232e-02 -1.04042172e+00 -5.40880859e-01 -4.59910214e-01
-1.95923820e-01 1.19732559e+00 -5.68949580e-01 -1.58092484e-01
4.40582670e-02 -8.50019038e-01 2.60462642e-01 4.54214007e-01
-9.37288478e-02 -7.20985115e-01 -6.11917734e-01 6.07072830e-01
-6.19814657e-02 -1.02288032e+00 -2.51931578e-01 5.17225444e-01
-1.09142196e+00 -6.07609808e-01 -9.21466589e-01 -6.07925951e-01
3.22263300e-01 8.67348313e-01 9.80978429e-01 -3.14727038e-01
-4.79154646e-01 4.25397336e-01 -6.80718839e-01 -7.72847354e-01
-3.43771987e-02 1.55570507e-01 1.31859899e-01 -4.03342307e-01
9.65942025e-01 -4.25477445e-01 -3.80396582e-02 1.90908683e-03
-7.22030103e-01 -1.24860354e-01 1.25509113e-01 8.62712204e-01
2.81733572e-01 3.38667959e-01 5.99567056e-01 -1.52781308e+00
7.79569209e-01 -7.90159941e-01 1.60445407e-01 2.52883643e-01
-9.41413105e-01 -2.12293088e-01 6.69377223e-02 -7.03515589e-01
-1.37671983e+00 -3.78037989e-01 4.57811773e-01 -5.33757567e-01
-3.73687476e-01 6.75801456e-01 3.61475289e-01 4.71038580e-01
9.03566241e-01 2.50921965e-01 -2.74733007e-01 -4.64333117e-01
6.20645583e-01 7.51364231e-01 -1.00010835e-01 -2.10070118e-01
3.47208440e-01 6.27121687e-01 -5.23793578e-01 -1.33795106e+00
-1.30284083e+00 -1.36953795e+00 -5.57372034e-01 -1.14877060e-01
6.07823908e-01 -1.15714216e+00 4.94863689e-01 2.29113027e-01
-9.82542336e-01 6.22516870e-02 -7.42043734e-01 6.20870531e-01
-5.24262667e-01 1.76054806e-01 -5.25490105e-01 -1.08111179e+00
-3.86026055e-01 -4.85403389e-01 1.18278050e+00 2.46433124e-01
-7.32460797e-01 -1.18898880e+00 4.16466802e-01 5.41299284e-02
3.68612826e-01 -3.38095099e-01 8.21759522e-01 -1.23486495e+00
-6.35088161e-02 -7.88199782e-01 1.25889078e-01 -1.01446971e-01
4.21617299e-01 -1.05012543e-01 -1.06201589e+00 -2.63700753e-01
2.77680367e-01 -5.16381741e-01 1.10625720e+00 6.81246161e-01
6.87386513e-01 -1.31262422e-01 -7.83594072e-01 -6.60947710e-02
1.26863492e+00 2.09735215e-01 1.29577175e-01 -1.57682985e-01
5.29358685e-01 5.05835474e-01 8.72099757e-01 8.33959222e-01
9.64233801e-02 4.33372706e-01 -4.78597939e-01 3.64409804e-01
-2.97164591e-03 -1.42029792e-01 4.13856804e-02 8.48087609e-01
3.65734041e-01 -2.62789696e-01 -1.19250786e+00 8.20202112e-01
-1.83140278e+00 -1.30518866e+00 2.19702765e-01 1.82688391e+00
8.92581224e-01 3.92354369e-01 -1.92534868e-02 1.44748449e-01
7.84774780e-01 5.21330893e-01 -3.00702631e-01 -1.09231479e-01
1.75344139e-01 4.94932868e-02 1.79100439e-01 2.60396570e-01
-1.26794159e+00 1.18279815e+00 7.62651300e+00 1.03205276e+00
-7.65540957e-01 4.58994478e-01 3.35048318e-01 -2.54214555e-01
-2.49473840e-01 3.23622406e-01 -1.27201223e+00 4.12931293e-01
1.24031067e+00 -7.97638237e-01 -4.24275368e-01 1.40433836e+00
1.40399739e-01 -5.24711490e-01 -9.29365635e-01 7.63456106e-01
6.29667938e-01 -1.38515854e+00 3.58131886e-01 3.78974387e-03
1.11165774e+00 -1.19149357e-01 -2.21252084e-01 7.40572929e-01
5.42495549e-01 -8.29451680e-01 6.73719198e-02 3.44327211e-01
4.37532723e-01 -7.52435207e-01 7.28265703e-01 6.94283128e-01
-9.97740328e-01 7.88479298e-03 -7.61930704e-01 -1.75623342e-01
5.10292470e-01 9.68009770e-01 -1.19340456e+00 1.31997600e-01
4.05517787e-01 8.64779830e-01 -6.52688742e-02 1.13026762e+00
7.39287212e-02 1.01694393e+00 -2.22721726e-01 -9.23355222e-02
3.59863430e-01 3.14155072e-01 5.75870752e-01 1.51203620e+00
1.32491201e-01 2.72567689e-01 6.86239421e-01 5.30627072e-01
-1.79704893e-02 5.31476915e-01 -8.26776683e-01 -5.22784591e-01
6.14188433e-01 1.03845417e+00 -1.26626372e+00 -1.07543683e+00
-3.42211485e-01 4.58238721e-01 1.29763976e-01 2.23587394e-01
-1.69218734e-01 -6.95360839e-01 3.12561452e-01 7.69755095e-02
3.11768204e-01 -4.04753208e-01 -4.93689746e-01 -1.35913777e+00
-7.39140749e-01 -3.26831073e-01 4.17183667e-01 -6.27763927e-01
-1.17511690e+00 1.80921242e-01 4.91790920e-01 -1.05253136e+00
-6.98681474e-01 3.11397389e-02 -1.03968418e+00 4.34869438e-01
-9.59909379e-01 -9.62623298e-01 -7.67862201e-02 2.20122054e-01
1.59481823e+00 -3.89577448e-01 1.12065399e+00 -5.33354878e-02
-3.37411910e-01 7.63777718e-02 3.61805648e-01 -2.28678152e-01
7.80532241e-01 -1.17947304e+00 2.98708528e-01 7.14473844e-01
4.81805772e-01 8.71591926e-01 1.11737967e+00 -9.02657628e-01
-8.58480096e-01 -1.07025969e+00 1.19998431e+00 -3.75888735e-01
6.85455322e-01 -5.24493456e-01 -1.07372105e+00 6.78107619e-01
2.68532634e-01 -2.84191102e-01 1.27611458e+00 8.41923773e-01
-3.98951828e-01 4.06529665e-01 -8.02356362e-01 2.42495149e-01
4.16818738e-01 -6.55707598e-01 -1.09684455e+00 7.80793011e-01
1.04947340e+00 8.45123008e-02 -5.93387783e-01 -1.69223413e-01
2.43714049e-01 -3.58375348e-03 7.34961152e-01 -9.43722725e-01
4.01702732e-01 1.93512842e-01 -2.41276801e-01 -1.30951703e+00
-1.37919441e-01 -3.53628725e-01 -5.85639834e-01 1.26740217e+00
1.61021844e-01 2.36868132e-02 9.64083135e-01 7.64268041e-01
1.62961096e-01 -3.44995201e-01 -5.78198075e-01 -9.04175520e-01
-7.66493082e-02 -6.68832779e-01 -6.73608929e-02 1.05425835e+00
6.91282511e-01 6.90135658e-01 -6.54865623e-01 -5.25637746e-01
8.61155331e-01 1.84532642e-01 6.28191233e-01 -1.55235469e+00
7.63533488e-02 -1.67427391e-01 -3.66421580e-01 -6.35892153e-01
2.66269088e-01 -5.42128623e-01 6.39729619e-01 -1.70302165e+00
7.74376810e-01 -1.01168707e-01 -1.75811693e-01 4.06344950e-01
-2.78944075e-01 -3.54855880e-03 2.63054315e-02 4.42875326e-01
-9.90387559e-01 6.87673211e-01 4.54506189e-01 -3.41202945e-01
-3.48552674e-01 3.84258851e-02 -6.18950784e-01 8.22406650e-01
5.37377000e-01 -9.56138492e-01 -6.89749777e-01 8.75865147e-02
-1.32386982e-01 -1.61205769e-01 -4.24421281e-01 -9.48802769e-01
7.61610210e-01 -2.60801196e-01 -3.69867198e-02 -7.02272594e-01
5.44769228e-01 -3.75965208e-01 -3.23178738e-01 1.56915128e-01
-7.28503466e-01 -7.37881780e-01 -9.15028527e-02 9.36507285e-01
-3.97029012e-01 -6.24390423e-01 6.90532446e-01 -4.30014938e-01
-9.12821412e-01 2.89876163e-01 -9.67272937e-01 2.77960211e-01
1.13964939e+00 -2.22321749e-01 -1.16891861e-01 -6.03698015e-01
-9.31868792e-01 1.33783162e-01 2.67106473e-01 3.26728284e-01
5.19573867e-01 -1.03716671e+00 -5.01824737e-01 -2.62887597e-01
5.77237487e-01 -5.78354411e-02 2.78039843e-01 3.50354344e-01
2.92514622e-01 7.93434203e-01 1.70180812e-01 -4.90196079e-01
-1.12802458e+00 5.47283947e-01 -4.62902039e-01 -2.89778024e-01
-8.58218193e-01 8.32963824e-01 4.87953350e-02 1.09858915e-01
4.72276360e-01 1.50944099e-01 -5.06168246e-01 7.98936605e-01
9.42687631e-01 3.38921636e-01 -8.32299814e-02 -3.41718793e-01
-4.03979979e-02 1.77432057e-02 -7.88019538e-01 -3.97789359e-01
1.41774774e+00 -1.91842258e-01 2.92022496e-01 1.43571174e+00
9.30760443e-01 -5.71590960e-01 -6.83245301e-01 -7.26604700e-01
3.52627248e-01 -5.30296862e-01 3.81511241e-01 -2.67386287e-01
-2.86563814e-01 1.05987334e+00 1.81518763e-01 3.88315380e-01
3.29386115e-01 2.98620611e-01 6.19917572e-01 7.52118289e-01
5.52810550e-01 -1.41312420e+00 2.97278017e-01 7.49037206e-01
1.43160656e-01 -1.55747640e+00 4.02494758e-01 -3.74824733e-01
-6.49048686e-01 8.84922206e-01 4.67214644e-01 1.33441361e-02
1.22828674e+00 2.05880314e-01 -2.62205929e-01 -4.03576612e-01
-1.31438053e+00 -2.44069174e-01 2.12735444e-01 4.31043595e-01
7.01533973e-01 6.58974871e-02 -2.86170483e-01 6.26618266e-01
1.65270753e-02 -2.23746244e-02 7.59874761e-01 1.43782246e+00
-1.27105737e+00 -8.11885238e-01 -8.05800185e-02 1.02368021e+00
-4.95915353e-01 -2.83566236e-01 -1.40010923e-01 4.83437628e-01
-3.73126894e-01 1.13043094e+00 3.79226267e-01 7.66775385e-02
-3.50722522e-01 7.18761981e-01 -1.45828664e-01 -1.81234622e+00
-8.66944417e-02 4.07454282e-01 5.30040860e-02 -2.27258220e-01
-4.05462116e-01 -8.72796416e-01 -1.00597811e+00 -5.06063811e-02
-8.69578540e-01 6.14467859e-01 8.53607059e-01 1.38788092e+00
1.29068971e-01 3.62843871e-01 4.27848667e-01 -7.12053895e-01
-4.82865334e-01 -1.51893139e+00 -1.01693130e+00 9.70659032e-02
-1.97505299e-02 -8.33881617e-01 -4.37125206e-01 4.12046283e-01] | [10.418275833129883, 7.014618396759033] |
c32880fb-f6c9-41d1-bc2c-4f2bb8a7591c | bedlam-a-synthetic-dataset-of-bodies-1 | 2306.16940 | null | https://arxiv.org/abs/2306.16940v1 | https://arxiv.org/pdf/2306.16940v1.pdf | BEDLAM: A Synthetic Dataset of Bodies Exhibiting Detailed Lifelike Animated Motion | We show, for the first time, that neural networks trained only on synthetic data achieve state-of-the-art accuracy on the problem of 3D human pose and shape (HPS) estimation from real images. Previous synthetic datasets have been small, unrealistic, or lacked realistic clothing. Achieving sufficient realism is non-trivial and we show how to do this for full bodies in motion. Specifically, our BEDLAM dataset contains monocular RGB videos with ground-truth 3D bodies in SMPL-X format. It includes a diversity of body shapes, motions, skin tones, hair, and clothing. The clothing is realistically simulated on the moving bodies using commercial clothing physics simulation. We render varying numbers of people in realistic scenes with varied lighting and camera motions. We then train various HPS regressors using BEDLAM and achieve state-of-the-art accuracy on real-image benchmarks despite training with synthetic data. We use BEDLAM to gain insights into what model design choices are important for accuracy. With good synthetic training data, we find that a basic method like HMR approaches the accuracy of the current SOTA method (CLIFF). BEDLAM is useful for a variety of tasks and all images, ground truth bodies, 3D clothing, support code, and more are available for research purposes. Additionally, we provide detailed information about our synthetic data generation pipeline, enabling others to generate their own datasets. See the project page: https://bedlam.is.tue.mpg.de/. | ['Jinlong Yang', 'Joachim Tesch', 'Priyanka Patel', 'Michael J. Black'] | 2023-06-29 | bedlam-a-synthetic-dataset-of-bodies | http://openaccess.thecvf.com//content/CVPR2023/html/Black_BEDLAM_A_Synthetic_Dataset_of_Bodies_Exhibiting_Detailed_Lifelike_Animated_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Black_BEDLAM_A_Synthetic_Dataset_of_Bodies_Exhibiting_Detailed_Lifelike_Animated_CVPR_2023_paper.pdf | cvpr-2023-1 | ['synthetic-data-generation', 'synthetic-data-generation'] | ['medical', 'miscellaneous'] | [ 1.02116078e-01 -1.43182173e-01 2.84057826e-01 -3.09659839e-01
-6.19377911e-01 -6.13389373e-01 4.65544015e-01 -7.07507670e-01
-1.34717867e-01 6.89527810e-01 3.02974492e-01 9.26035792e-02
4.38806713e-01 -5.68753064e-01 -1.12783742e+00 -4.32172805e-01
1.80355348e-02 5.59824705e-01 -4.66119200e-02 -3.93826246e-01
-5.05120814e-01 3.39746326e-01 -1.44874644e+00 3.63743246e-01
2.70503223e-01 9.08366680e-01 -1.86310500e-01 1.02320194e+00
6.81383014e-01 5.15946209e-01 -4.88604486e-01 -7.20182538e-01
8.05738509e-01 -4.00691807e-01 -5.03670573e-01 4.85395610e-01
9.81349409e-01 -7.34370828e-01 -4.35044676e-01 6.71272933e-01
8.07343721e-01 7.07985610e-02 5.20057201e-01 -1.17451704e+00
-6.99776947e-01 2.62883812e-01 -4.85373080e-01 -3.22490513e-01
6.06693327e-01 7.25147903e-01 4.93306279e-01 -8.34961832e-01
8.22296858e-01 1.38871229e+00 1.11179233e+00 8.97462726e-01
-1.53658295e+00 -5.42181253e-01 -2.39520818e-02 -4.57810193e-01
-1.26347923e+00 -6.43775105e-01 6.34466290e-01 -5.58942258e-01
6.91669822e-01 4.34461772e-01 1.22324622e+00 1.73368084e+00
-3.45441513e-02 7.67888427e-01 1.33848453e+00 -3.53975475e-01
-4.85776812e-02 -1.45169884e-01 -2.59949267e-01 7.74150550e-01
2.67030746e-01 2.25366980e-01 -2.84449071e-01 1.16835386e-02
1.19236231e+00 -3.49022776e-01 -3.21912646e-01 -6.40088737e-01
-1.47358453e+00 4.68284488e-01 3.60822618e-01 -2.23962530e-01
-8.74325261e-02 5.25932312e-01 2.89712697e-01 1.51926115e-01
4.43220019e-01 4.17657852e-01 -5.29296100e-01 4.41922024e-02
-8.95822048e-01 9.56843495e-01 8.33285332e-01 1.14765728e+00
2.05253333e-01 2.97554135e-01 2.02961359e-02 8.12817931e-01
1.13859318e-01 8.48889589e-01 3.18425186e-02 -1.53704154e+00
1.89943641e-01 1.72589183e-01 5.08234441e-01 -1.02408445e+00
-5.08030176e-01 -3.48425210e-01 -9.42232788e-01 4.89143729e-01
8.13506961e-01 -3.23148161e-01 -1.15140224e+00 1.71671033e+00
4.82515395e-01 -1.79349780e-01 -2.23229811e-01 1.53485203e+00
1.08981252e+00 5.34682453e-01 -3.47245783e-02 2.83137202e-01
1.17229831e+00 -9.74631310e-01 -3.48532677e-01 -3.59601170e-01
2.16809347e-01 -9.11406577e-01 1.26430118e+00 4.57237750e-01
-1.46058226e+00 -7.39816487e-01 -8.97192717e-01 -2.19808832e-01
-1.18175618e-01 2.97312021e-01 6.96068943e-01 5.31032622e-01
-1.07882690e+00 7.81720400e-01 -1.00106514e+00 -7.40477443e-01
1.99723795e-01 2.69872189e-01 -3.84188741e-01 -4.63883914e-02
-1.13833416e+00 7.62606800e-01 -3.31194066e-02 2.98444569e-01
-1.05059052e+00 -5.87748706e-01 -1.18998075e+00 -6.51878953e-01
2.25503922e-01 -1.15964472e+00 1.45460618e+00 -1.17934978e+00
-1.28723919e+00 1.25572383e+00 3.67294252e-01 -2.80458212e-01
1.00056934e+00 -3.94012719e-01 -3.39594960e-01 -1.44647494e-01
-9.88279358e-02 1.05319905e+00 7.86342978e-01 -1.63948572e+00
9.88459289e-02 -1.01938933e-01 3.00582033e-02 2.23829642e-01
3.54040265e-01 3.61137465e-02 -6.65293872e-01 -9.01244998e-01
-2.37510264e-01 -1.33262777e+00 -1.07951477e-01 4.38185990e-01
-6.43275738e-01 5.72521985e-01 3.26735914e-01 -1.08855474e+00
6.05816364e-01 -1.90110350e+00 3.19877118e-01 -6.84688613e-02
1.04352213e-01 -1.72375180e-02 -1.35440335e-01 2.24240750e-01
8.48770235e-03 -4.88213412e-02 -2.93430328e-01 -6.41093254e-01
3.35219949e-01 1.29603401e-01 -1.45192854e-02 5.79785228e-01
1.55341581e-01 1.00602734e+00 -6.06564701e-01 -5.30952036e-01
4.53228295e-01 7.02838182e-01 -5.19160807e-01 2.18546703e-01
-3.39171588e-01 7.20096886e-01 2.09422514e-01 9.42951858e-01
5.95576942e-01 -3.46744925e-01 1.47493631e-01 -6.07363820e-01
1.14934482e-01 -1.56146824e-01 -1.22833622e+00 1.78170884e+00
-2.08217070e-01 5.76068759e-01 3.39197695e-01 -9.32601541e-02
6.31082356e-01 4.51348945e-02 4.55862939e-01 -4.48393941e-01
1.62669510e-01 -2.16198601e-02 -9.66775119e-02 -2.96880186e-01
4.93005067e-01 -1.93612412e-01 -1.03626929e-01 1.20815299e-01
-3.76366042e-02 -4.51307386e-01 2.00800017e-01 -4.70743403e-02
6.97552025e-01 8.70048463e-01 -4.24087867e-02 -1.53535560e-01
-3.12492639e-01 3.88158798e-01 3.17077428e-01 4.74334657e-01
-7.43277557e-03 1.10592532e+00 9.62093398e-02 -7.94050992e-01
-1.74784935e+00 -1.30623555e+00 -3.27057466e-02 9.42594588e-01
1.40025124e-01 -1.92795679e-01 -9.25081134e-01 -7.22397566e-02
2.14496747e-01 5.07219851e-01 -8.92207682e-01 2.61597961e-01
-7.32757449e-01 -7.60853469e-01 5.52292705e-01 6.42954946e-01
5.17886639e-01 -1.03293979e+00 -7.09494114e-01 3.16679180e-02
-1.40800089e-01 -1.15982759e+00 -5.67943096e-01 -2.36029759e-01
-5.22412002e-01 -9.72641945e-01 -1.05935144e+00 -3.38189572e-01
6.93125308e-01 -1.30852032e-02 1.53808272e+00 7.25995451e-02
-5.83494782e-01 4.26632822e-01 -1.39753565e-01 -2.06941456e-01
-6.38792634e-01 -2.75839597e-01 2.55294621e-01 -3.69904041e-01
-1.69822082e-01 -3.20562661e-01 -9.25749421e-01 5.44768274e-01
-6.18935764e-01 7.16580391e-01 1.70989290e-01 7.30735183e-01
5.44474185e-01 -3.62945765e-01 -1.97999924e-01 -5.45971632e-01
2.89940536e-01 -7.62079209e-02 -5.06160438e-01 -7.26323277e-02
8.18924606e-02 -2.87992001e-01 4.71074969e-01 -5.21738231e-01
-8.99006963e-01 1.47049248e-01 -3.00745755e-01 -6.67540550e-01
-3.47913861e-01 -2.51948923e-01 -1.06125854e-01 -3.63533013e-02
8.75472784e-01 -1.80664584e-02 -1.90306418e-02 -5.27897656e-01
4.35649693e-01 1.65794343e-01 7.02524602e-01 -8.89956057e-01
7.13107228e-01 5.11195421e-01 -1.05723210e-01 -4.95502174e-01
-8.41419816e-01 1.88784391e-01 -7.27831304e-01 -4.84873444e-01
8.81621599e-01 -1.21051550e+00 -7.67446697e-01 7.56660163e-01
-8.13335836e-01 -1.16477966e+00 -2.53379256e-01 2.98161149e-01
-7.75481582e-01 5.56930676e-02 -1.05733061e+00 -6.46112621e-01
-2.71216631e-01 -1.06259322e+00 1.56207132e+00 -7.72368908e-02
-5.86373031e-01 -7.61542380e-01 -5.40656857e-02 4.91618782e-01
1.93377078e-01 9.82367337e-01 4.33170855e-01 2.18813360e-01
-5.01987278e-01 7.94417560e-02 -1.45610839e-01 2.71437019e-01
2.54988503e-02 3.20405930e-01 -9.22096729e-01 -5.56249797e-01
-4.62103784e-01 -6.52881801e-01 6.25752091e-01 5.84165812e-01
9.69046891e-01 -4.56022084e-01 -7.45897964e-02 9.01831985e-01
1.31517196e+00 -3.14235538e-01 4.35479969e-01 2.66234607e-01
1.05722678e+00 5.90733230e-01 6.31780744e-01 4.51426536e-01
4.27573383e-01 8.67183864e-01 3.55337977e-01 -6.04527652e-01
-4.69771653e-01 -3.50780755e-01 4.19057012e-01 5.44888914e-01
-4.15830195e-01 -1.04082771e-01 -8.71116757e-01 4.39551890e-01
-1.55477619e+00 -8.97582412e-01 -1.03794001e-01 2.17341328e+00
9.44883287e-01 1.48053408e-01 7.53008306e-01 -1.11830495e-01
4.99405205e-01 7.27632195e-02 -6.90934062e-01 -2.81007975e-01
-2.32439503e-01 -1.18339047e-01 6.94182634e-01 4.45430040e-01
-1.22224712e+00 8.97937775e-01 6.50694275e+00 2.81447619e-01
-1.06978416e+00 5.38571253e-02 8.31301391e-01 -6.11553907e-01
-6.38157278e-02 -4.90936488e-01 -6.74898446e-01 4.90117103e-01
5.56136310e-01 4.04598385e-01 7.46537447e-01 8.26320291e-01
3.50626379e-01 9.81177092e-02 -1.05149412e+00 1.10522306e+00
8.69393945e-02 -1.09427547e+00 -1.50646806e-01 2.11499240e-02
9.26816761e-01 2.80019641e-02 5.20044155e-02 8.89020339e-02
7.62854695e-01 -1.15235066e+00 1.19784224e+00 7.04871833e-01
1.21457648e+00 -3.47499430e-01 4.12789702e-01 5.96889742e-02
-9.26347375e-01 3.40675950e-01 -3.04592401e-01 -7.75850937e-02
4.36583668e-01 3.22149456e-01 -5.57644725e-01 2.45731741e-01
8.69483113e-01 5.47928691e-01 -6.69299424e-01 6.85205340e-01
-1.13056405e-02 4.32487309e-01 -4.94197667e-01 1.40386462e-01
-7.10964873e-02 -1.42524308e-02 2.87314981e-01 1.31889117e+00
2.54428387e-01 4.14914675e-02 1.90740153e-01 9.23005462e-01
3.23546082e-02 -2.44386375e-01 -6.65737629e-01 1.93103999e-01
-2.03970261e-03 1.12183130e+00 -5.80811024e-01 -2.47287765e-01
-1.77446142e-01 1.11126435e+00 4.39895354e-02 4.13734615e-01
-1.07449782e+00 2.12882921e-01 7.93153048e-01 5.90370774e-01
2.27992088e-02 -5.39189041e-01 -1.54432669e-01 -1.24780393e+00
-4.43572439e-02 -1.27151418e+00 -9.66278184e-03 -1.42147565e+00
-1.37369716e+00 5.41555166e-01 2.62574941e-01 -1.14980018e+00
-3.49696040e-01 -8.84649098e-01 -4.16206159e-02 5.76864302e-01
-6.66605413e-01 -1.69724572e+00 -7.61656046e-01 5.22424161e-01
5.69743574e-01 3.34042609e-01 7.05937922e-01 3.08007509e-01
-4.24190283e-01 4.18360740e-01 -1.83678076e-01 4.25967813e-01
8.84058118e-01 -1.20821834e+00 1.12732899e+00 5.83826900e-01
-9.54964682e-02 4.67706770e-01 1.11234534e+00 -8.03147376e-01
-1.65233111e+00 -1.19779503e+00 6.91591799e-02 -8.58302832e-01
4.05277103e-01 -7.34557092e-01 -4.42205638e-01 9.94334579e-01
1.78487062e-01 3.69680077e-02 4.72577184e-01 -1.69146538e-01
-8.87437314e-02 5.13285771e-03 -1.14165890e+00 9.73176360e-01
1.51380301e+00 -5.85900526e-03 -2.35162407e-01 4.37616438e-01
7.25785553e-01 -1.00194883e+00 -1.10195100e+00 5.46560764e-01
1.18233025e+00 -1.08039236e+00 1.16416395e+00 -5.05319953e-01
7.09596038e-01 -4.16965783e-01 -4.82928038e-01 -1.39607334e+00
-3.00530344e-01 -4.95975882e-01 -2.24286050e-01 8.41439545e-01
2.09539697e-01 -1.29847363e-01 8.07175040e-01 9.19021547e-01
1.06921196e-01 -7.82787979e-01 -2.77515799e-01 -6.22642338e-01
-2.77907494e-02 -3.80755961e-01 5.36264658e-01 8.78261745e-01
-7.31327295e-01 8.43326673e-02 -8.96682084e-01 -4.32652272e-02
9.74154770e-01 2.25822777e-01 1.40519249e+00 -7.76378572e-01
-6.35990381e-01 -4.66428697e-02 -1.73608467e-01 -1.01009572e+00
-1.11661464e-01 -3.76778811e-01 -3.47396098e-02 -1.64011598e+00
1.33845285e-01 -4.16500956e-01 3.52926403e-01 5.48685014e-01
6.43073395e-02 8.80955696e-01 4.71364647e-01 3.88699360e-02
-4.33330297e-01 2.72646099e-01 1.67445707e+00 -9.55970287e-02
-1.59866251e-02 -1.55982345e-01 -4.45838064e-01 8.98716211e-01
7.98493743e-01 -3.84722352e-02 -1.17448062e-01 -7.23034918e-01
1.21303000e-01 3.91950868e-02 1.06562376e+00 -1.16856670e+00
-3.09514940e-01 -1.03080302e-01 1.21055067e+00 -4.25816298e-01
9.63354588e-01 -6.59617126e-01 9.51131165e-01 4.06690538e-01
-2.06861824e-01 1.93434283e-01 3.83995593e-01 3.33147012e-02
4.50413585e-01 3.54281068e-01 8.82312536e-01 -6.22914553e-01
-4.92988855e-01 2.08653450e-01 5.03428020e-02 3.23155038e-02
6.20097518e-01 -4.07735795e-01 -2.60059416e-01 -6.39947951e-01
-1.07285213e+00 -8.12135339e-02 1.33598077e+00 3.89927804e-01
4.08495635e-01 -1.50026512e+00 -8.13093662e-01 2.17375353e-01
-1.08743265e-01 7.68183321e-02 1.91994146e-01 5.14550567e-01
-9.25080419e-01 8.27043690e-03 -4.91946220e-01 -6.32731140e-01
-1.31685686e+00 5.55277586e-01 5.87683141e-01 2.68139057e-02
-5.60860753e-01 6.88652098e-01 2.82838553e-01 -7.22753048e-01
-6.22071810e-02 -4.23931092e-01 6.46923780e-01 -5.91851354e-01
1.55415833e-01 3.20927083e-01 -2.26654544e-01 -9.06852424e-01
-2.13759392e-01 8.17161202e-01 4.76826400e-01 -2.80726373e-01
1.22125137e+00 -5.25068156e-02 2.63322979e-01 4.70553190e-01
9.04047966e-01 4.35187891e-02 -1.86175048e+00 3.13598961e-01
-8.35326970e-01 -6.50237322e-01 -4.24645573e-01 -1.10160708e+00
-1.04861867e+00 6.88659251e-01 6.91246450e-01 -7.38819689e-02
1.06481159e+00 1.51206665e-02 8.02216887e-01 4.32892106e-02
6.73909903e-01 -8.05673599e-01 -1.04358710e-01 1.13650151e-01
1.23500657e+00 -1.23127031e+00 3.57068956e-01 -4.10944968e-01
-7.43062854e-01 7.21136451e-01 9.36223865e-01 -2.94195622e-01
1.83213383e-01 5.64166427e-01 2.18364507e-01 -1.08870879e-01
-5.49749494e-01 -7.91950449e-02 2.91041017e-01 6.73548162e-01
4.97602731e-01 2.90999293e-01 1.73531294e-01 3.44500005e-01
-6.90211117e-01 6.10821228e-03 3.17474812e-01 6.77754700e-01
7.90961310e-02 -8.55119944e-01 -7.09841132e-01 3.11391801e-01
-3.72229725e-01 5.09351902e-02 -5.42849123e-01 1.11830771e+00
2.66129941e-01 5.52849829e-01 1.16901565e-03 -4.05162990e-01
4.98169631e-01 -3.20039630e-01 1.10376298e+00 -3.45150083e-01
-4.58532929e-01 2.57847637e-01 5.49638867e-01 -6.61273360e-01
-4.69769329e-01 -5.70513010e-01 -8.24074388e-01 -6.61328495e-01
3.50349583e-02 -4.75454986e-01 6.67578757e-01 4.66673136e-01
2.28780091e-01 3.63163561e-01 6.20797053e-02 -1.55533934e+00
-2.22573653e-01 -9.31743264e-01 -4.48430032e-01 7.73696840e-01
2.91311502e-01 -7.31964648e-01 -1.09433539e-01 5.08592129e-01] | [7.206432342529297, -1.2211016416549683] |
2ac9fcf3-fa9c-46b8-8ed1-0807a44a8536 | deep-learning-for-video-game-genre | 2011.12143 | null | https://arxiv.org/abs/2011.12143v1 | https://arxiv.org/pdf/2011.12143v1.pdf | Deep learning for video game genre classification | Video game genre classification based on its cover and textual description would be utterly beneficial to many modern identification, collocation, and retrieval systems. At the same time, it is also an extremely challenging task due to the following reasons: First, there exists a wide variety of video game genres, many of which are not concretely defined. Second, video game covers vary in many different ways such as colors, styles, textual information, etc, even for games of the same genre. Third, cover designs and textual descriptions may vary due to many external factors such as country, culture, target reader populations, etc. With the growing competitiveness in the video game industry, the cover designers and typographers push the cover designs to its limit in the hope of attracting sales. The computer-based automatic video game genre classification systems become a particularly exciting research topic in recent years. In this paper, we propose a multi-modal deep learning framework to solve this problem. The contribution of this paper is four-fold. First, we compiles a large dataset consisting of 50,000 video games from 21 genres made of cover images, description text, and title text and the genre information. Second, image-based and text-based, state-of-the-art models are evaluated thoroughly for the task of genre classification for video games. Third, we developed an efficient and salable multi-modal framework based on both images and texts. Fourth, a thorough analysis of the experimental results is given and future works to improve the performance is suggested. The results show that the multi-modal framework outperforms the current state-of-the-art image-based or text-based models. Several challenges are outlined for this task. More efforts and resources are needed for this classification task in order to reach a satisfactory level. | ['Lukun Zheng', 'Yuhang Jiang'] | 2020-11-21 | null | null | null | null | ['genre-classification'] | ['computer-vision'] | [-5.50029762e-02 -8.89244795e-01 -2.64946043e-01 7.19670132e-02
-7.60823965e-01 -5.49681127e-01 5.49709141e-01 -2.42718175e-01
-2.78267086e-01 4.84018832e-01 1.27905771e-01 3.35772224e-02
-1.08386710e-01 -8.23823929e-01 -3.68467808e-01 -6.53295577e-01
1.53683394e-01 3.94062787e-01 3.26898128e-01 -6.52552724e-01
4.58670408e-01 1.48049500e-02 -2.02734017e+00 5.59385121e-01
5.27777612e-01 1.37690699e+00 3.89056146e-01 3.54793191e-01
-4.93429691e-01 6.61607027e-01 -6.99311614e-01 -7.63901711e-01
5.19877933e-02 -4.73113179e-01 -6.49228096e-01 3.91532332e-01
2.65936434e-01 -2.21581444e-01 -5.71293235e-01 1.26240134e+00
4.28803444e-01 8.96796212e-02 9.65494096e-01 -1.12377918e+00
-6.28023744e-01 4.91985738e-01 -9.16798472e-01 6.05586432e-02
4.47733492e-01 -5.69446236e-02 1.08581924e+00 -5.41575253e-01
4.22168046e-01 1.04991961e+00 5.38358986e-01 5.06831408e-01
-4.83501107e-01 -9.31990802e-01 9.77010876e-02 6.24376535e-01
-1.68014395e+00 -1.31403152e-02 9.28819656e-01 -5.85985720e-01
5.37194371e-01 3.79965305e-01 9.22523916e-01 1.27649915e+00
1.43411875e-01 7.15547264e-01 1.05020618e+00 -4.04801548e-01
-1.58309504e-01 2.85068214e-01 -2.63251394e-01 6.37800038e-01
-6.50553107e-02 -2.98616558e-01 -5.53486168e-01 1.58251286e-01
6.17212236e-01 9.90257971e-03 -3.09883714e-01 -1.39050573e-01
-8.11675787e-01 9.22158360e-01 -4.85431515e-02 5.47237933e-01
4.97275144e-02 -1.45908147e-01 8.96200776e-01 1.66792095e-01
6.41228080e-01 1.61013186e-01 -1.48619533e-01 -7.64956117e-01
-9.22254503e-01 3.28391910e-01 6.54469132e-01 9.71533716e-01
3.57437998e-01 7.28103817e-02 1.91440985e-01 1.42274559e+00
2.46659502e-01 3.86692733e-01 7.85895705e-01 -3.95869166e-01
6.42675519e-01 5.26054084e-01 -3.54202151e-01 -1.28892004e+00
-2.65808344e-01 -2.98774779e-01 -8.93024266e-01 -2.45265830e-02
3.49499047e-01 1.75959945e-01 -6.01216435e-01 1.32094026e+00
-2.65598714e-01 -1.25008414e-03 -2.53986150e-01 9.56989110e-01
1.39903414e+00 7.61234462e-01 1.14063658e-02 -1.33233085e-01
1.70302773e+00 -7.56810009e-01 -7.13910222e-01 -1.41182141e-02
3.41120541e-01 -9.20658648e-01 1.11463189e+00 6.06906533e-01
-8.79416466e-01 -5.30094981e-01 -9.68613863e-01 3.31259161e-01
-5.64598024e-01 3.02641481e-01 6.03898704e-01 8.89059424e-01
-5.98703444e-01 1.90884694e-01 -2.72906661e-01 -4.21365470e-01
3.58982950e-01 1.74259767e-01 -3.12227756e-01 -2.71149352e-02
-1.32270086e+00 5.10632753e-01 5.46131968e-01 -1.04222074e-01
-5.49578547e-01 -3.12205642e-01 -9.07700479e-01 -1.26686707e-01
5.57963669e-01 -2.65252531e-01 9.95773554e-01 -1.18197811e+00
-1.36377132e+00 1.15661335e+00 2.52050966e-01 -3.23177762e-02
5.12956083e-01 2.39693627e-01 -8.07372689e-01 8.40786025e-02
1.83443919e-01 2.05015630e-01 8.81951213e-01 -9.60233152e-01
-1.08998215e+00 -4.15571898e-01 4.96992290e-01 4.88581330e-01
-8.35766256e-01 1.39913321e-01 -1.07813704e+00 -8.59673679e-01
-1.76038206e-01 -9.31360960e-01 2.32422754e-01 -4.62165177e-01
-2.69934118e-01 -2.43266582e-01 7.97144234e-01 -5.11618733e-01
1.58377719e+00 -2.37985706e+00 2.69731462e-01 -1.57941580e-01
2.99785674e-01 1.21937186e-01 1.21357642e-01 3.93759280e-01
1.51197031e-01 3.28010112e-01 1.38896927e-01 -1.77383378e-01
4.89849923e-03 -1.97537601e-01 -1.52183637e-01 4.26852077e-01
-3.15643072e-01 5.92016518e-01 -4.97666538e-01 -6.09566569e-01
3.47764760e-01 1.81290388e-01 -1.96114779e-01 -1.87708065e-01
-9.61361006e-02 2.27377966e-01 -5.19395828e-01 9.49097872e-01
5.42408168e-01 -2.47455407e-02 -1.07946947e-01 -2.54114717e-01
-7.08853602e-02 -1.49452448e-01 -1.28409052e+00 1.78532648e+00
-5.17276824e-01 8.53411436e-01 -9.20266360e-02 -1.14474094e+00
6.66295826e-01 3.38640600e-01 6.94310963e-01 -6.82501733e-01
4.79773045e-01 3.60060990e-01 -5.56853823e-02 -6.23846591e-01
8.81201148e-01 -2.98337251e-01 -5.63447535e-01 1.74811438e-01
7.64410794e-02 -1.15202956e-01 5.37182391e-01 -5.18735796e-02
6.62164927e-01 2.73043122e-02 1.69670597e-01 6.75659031e-02
8.23478043e-01 1.03274025e-01 4.17545676e-01 5.26904464e-01
-2.26594225e-01 7.77031243e-01 4.79769439e-01 -3.76794010e-01
-9.56333220e-01 -6.91521049e-01 -3.22580814e-01 1.08235502e+00
5.77123761e-01 -5.73887289e-01 -7.20367491e-01 -3.93582731e-01
-3.66425425e-01 -1.83032393e-01 -3.70405883e-01 -3.70246246e-02
-1.88455895e-01 -7.81533003e-01 7.23382473e-01 3.08640033e-01
9.49044824e-01 -1.05402303e+00 1.30240358e-02 1.03468888e-01
-4.58579212e-01 -1.26477301e+00 -5.26724219e-01 -4.68712389e-01
-4.11970258e-01 -1.18462646e+00 -9.81728256e-01 -1.06286538e+00
1.08772635e-01 3.84417564e-01 1.00906491e+00 8.63757953e-02
-1.75359100e-01 3.49115342e-01 -8.03582191e-01 -4.30915803e-01
-2.13340536e-01 3.68940085e-01 7.09066391e-02 1.11230761e-01
6.50280654e-01 -3.59315902e-01 -2.74705142e-01 3.78062636e-01
-9.47973549e-01 1.62971169e-01 3.04553330e-01 7.99092352e-01
4.99314576e-01 6.15033507e-01 3.82712156e-01 -7.27366328e-01
7.71570802e-01 -4.36161816e-01 -3.83316278e-01 4.61307988e-02
-4.23353054e-02 -6.51611686e-01 6.80822432e-01 -6.93314731e-01
-7.10582316e-01 -3.20440054e-01 -3.65640223e-01 -5.14503896e-01
-2.28657350e-01 6.67713106e-01 -4.62821007e-01 -1.61837608e-01
4.94329363e-01 5.27496576e-01 -7.89125413e-02 -4.09267485e-01
-7.04656467e-02 1.27004290e+00 2.86163896e-01 -4.89074469e-01
6.58025682e-01 3.12170565e-01 -2.16553107e-01 -1.11570621e+00
-6.10854983e-01 -6.40401304e-01 -2.68858790e-01 -7.22284198e-01
8.02503705e-01 -1.08074737e+00 -8.06499183e-01 1.08507299e+00
-9.03084040e-01 1.83345556e-01 2.05933124e-01 5.28837800e-01
-5.26666582e-01 4.87756848e-01 -6.54242396e-01 -6.48074210e-01
-1.13227718e-01 -1.51889670e+00 9.88279521e-01 3.19852114e-01
7.54931942e-02 -9.42887127e-01 -2.81074494e-01 7.56445706e-01
1.18869230e-01 1.56939641e-01 6.75305724e-01 -4.17376965e-01
-3.67518693e-01 -2.79169649e-01 -2.73598880e-01 1.29954726e-01
2.49112815e-01 -1.91358235e-02 -8.39702725e-01 -1.27969012e-01
-1.03612699e-01 -4.19169843e-01 7.65685856e-01 4.81102258e-01
1.42254424e+00 -1.52787361e-02 -1.65302277e-01 6.74204290e-01
1.49464428e+00 4.90851879e-01 6.72451258e-01 7.30611205e-01
9.82161164e-01 4.32606518e-01 7.76860476e-01 5.95430970e-01
3.91542703e-01 9.63900447e-01 4.04025435e-01 1.58191949e-01
9.17431340e-02 -3.08962986e-02 2.33376741e-01 7.92902946e-01
-4.06175196e-01 -3.53184789e-01 -7.44240642e-01 5.03907859e-01
-1.69469345e+00 -1.35634387e+00 -9.18891281e-02 1.97231543e+00
6.16910100e-01 2.41845638e-01 5.69495082e-01 4.40086931e-01
9.07936513e-01 2.90962547e-01 -3.69940192e-01 -3.94573808e-01
-3.00855964e-01 -1.25967890e-01 3.97333682e-01 -1.52792633e-01
-1.50081766e+00 9.55774844e-01 5.30777979e+00 1.88516128e+00
-1.24383199e+00 7.60268718e-02 6.79441273e-01 -9.50431079e-02
1.09948758e-02 -6.15342498e-01 -8.13280523e-01 7.70731747e-01
4.92709905e-01 -1.89215004e-01 6.38711870e-01 8.80365431e-01
-5.37416749e-02 2.22310182e-02 -3.96440983e-01 1.73913276e+00
5.24495542e-01 -1.47927547e+00 9.73010361e-02 3.26978862e-01
5.44945300e-01 -2.71526754e-01 3.20756853e-01 5.12739182e-01
-3.33136559e-01 -8.87376070e-01 9.90251660e-01 3.43150012e-02
1.27939427e+00 -1.06188500e+00 8.06666553e-01 2.45697111e-01
-1.64842057e+00 -1.45602107e-01 -3.34601969e-01 -2.31063962e-01
5.99535787e-03 7.18569681e-02 8.76236409e-02 6.27400100e-01
8.99582982e-01 1.04507840e+00 -4.77675110e-01 1.03911567e+00
1.80255607e-01 4.34551835e-01 2.73245089e-02 -3.78068238e-01
3.52862626e-01 -3.90676588e-01 5.75041652e-01 1.04839170e+00
3.61717463e-01 4.30226363e-02 2.90282607e-01 4.90095168e-01
-6.92804307e-02 5.09087980e-01 -6.77377343e-01 -2.78397292e-01
1.21805251e-01 1.34717309e+00 -8.54158580e-01 -5.82777672e-02
-7.29164243e-01 7.86287963e-01 -1.45808667e-01 8.98701698e-02
-1.05798924e+00 -4.25522774e-01 7.70138741e-01 2.98644960e-01
1.85405046e-01 2.69698631e-02 -1.43562779e-01 -1.23478031e+00
8.13495368e-02 -1.08357775e+00 3.47043902e-01 -4.08923715e-01
-1.42103851e+00 8.31033766e-01 -2.12010238e-02 -1.88629282e+00
-1.67709365e-01 -8.19035947e-01 -2.11573273e-01 6.56691194e-01
-1.10752165e+00 -1.33288705e+00 -2.72470504e-01 6.40721262e-01
1.03097188e+00 -8.39693606e-01 6.72784865e-01 7.45293438e-01
-5.23676872e-01 7.66391397e-01 3.71544331e-01 5.68829298e-01
5.12338400e-01 -9.42727625e-01 -6.52744919e-02 4.50988024e-01
3.69215965e-01 9.61439982e-02 5.10762393e-01 -3.06511879e-01
-1.27024138e+00 -7.86600113e-01 2.20489085e-01 -2.50858843e-01
9.43199039e-01 -3.83759379e-01 -4.73696202e-01 1.22497953e-01
1.02361257e-03 -5.00669539e-01 8.40106130e-01 2.45815247e-01
-1.32586792e-01 -1.81600913e-01 -7.52570152e-01 6.07361853e-01
9.03039694e-01 -5.56744277e-01 -2.02880800e-01 2.30729118e-01
1.84056401e-01 -6.31218135e-01 -7.52680361e-01 1.19780585e-01
7.32412696e-01 -9.50496495e-01 6.92305446e-01 -1.60423011e-01
7.21628249e-01 -1.91904917e-01 -2.80067474e-01 -1.11457360e+00
-1.75895005e-01 -2.47024760e-01 2.36722887e-01 1.43344855e+00
2.24206194e-01 -2.14114010e-01 9.23965752e-01 1.51887089e-01
-6.46637604e-02 -7.82383919e-01 -9.76656556e-01 -6.87089741e-01
8.33216310e-02 -8.67685974e-01 4.81344849e-01 6.79330170e-01
1.56215757e-01 4.41139996e-01 -6.37575805e-01 -3.67505759e-01
3.11136901e-01 2.15217024e-01 7.08182633e-01 -1.21867573e+00
-2.09302798e-01 -8.67604733e-01 -1.02412784e+00 -9.06183958e-01
2.21246228e-01 -5.06400585e-01 -2.40919948e-01 -1.34765971e+00
5.10654569e-01 -3.99479151e-01 -3.13288793e-02 4.14931923e-02
9.36184376e-02 8.56876731e-01 3.51006120e-01 3.30210268e-01
-7.91827798e-01 4.61032480e-01 1.42691839e+00 -6.53043628e-01
-2.48220544e-02 2.38979101e-01 -8.31260622e-01 7.99625099e-01
7.20872343e-01 -1.30229682e-01 -4.28080589e-01 -1.68698981e-01
2.47627795e-01 1.17713675e-01 -5.64014278e-02 -1.01490510e+00
-2.22644862e-02 -1.83356255e-01 1.66639328e-01 -5.78883886e-01
6.55131400e-01 -7.13837326e-01 6.53653294e-02 8.72316882e-02
1.01973601e-01 -1.48703560e-01 3.26213539e-01 5.75790644e-01
-5.76313198e-01 -3.14656526e-01 6.76881373e-01 -3.37798208e-01
-1.08718789e+00 5.72340250e-01 -6.73454583e-01 2.21378773e-01
1.05771184e+00 -6.88745141e-01 -4.12035473e-02 -6.06292248e-01
-4.74924386e-01 -8.44261646e-02 5.54785311e-01 8.93439710e-01
5.23720503e-01 -1.42574382e+00 -7.42642999e-01 -9.68438461e-02
4.68479693e-01 -2.98669428e-01 7.05654383e-01 5.32742381e-01
-6.54717922e-01 3.40356797e-01 -4.63223636e-01 -5.93737185e-01
-1.57853675e+00 4.46615398e-01 2.58525193e-01 -2.64581949e-01
-3.91951591e-01 6.11303687e-01 4.34780508e-01 -7.30394647e-02
2.77695715e-01 1.92845762e-01 -9.16524410e-01 3.89992446e-01
8.13330472e-01 1.38744414e-01 4.10877578e-02 -1.27979016e+00
-1.82948068e-01 9.37530041e-01 -2.13962123e-01 -5.55103198e-02
1.03589094e+00 -1.60392985e-01 4.17343155e-02 5.61678290e-01
1.03939450e+00 -5.19504808e-02 -7.11499512e-01 -7.25679919e-02
-5.19136250e-01 -6.60382569e-01 2.92272288e-02 -5.55271327e-01
-1.38322103e+00 7.66939282e-01 5.34968972e-01 4.66330528e-01
1.35550952e+00 -7.04154521e-02 7.82829881e-01 -1.89492375e-01
6.77646399e-01 -1.33112109e+00 3.40738483e-02 5.86836159e-01
6.31896734e-01 -1.30673730e+00 -8.76368582e-02 -4.01324540e-01
-7.62930572e-01 1.04014874e+00 6.66049421e-01 1.45264775e-01
5.85682392e-01 6.27980605e-02 1.08220413e-01 -1.03845842e-01
-2.78155208e-01 -3.78769219e-01 4.71355200e-01 6.82535231e-01
4.78532672e-01 -3.11612356e-02 -6.28608942e-01 1.00933659e+00
-3.19089323e-01 -2.66573608e-01 5.71947157e-01 5.39780319e-01
-2.17279658e-01 -1.19727445e+00 -6.09656453e-01 6.48741901e-01
-8.85719776e-01 -3.25246304e-02 -3.12507391e-01 9.06896949e-01
3.81042302e-01 1.14651668e+00 4.91818674e-02 -8.87402117e-01
2.48105720e-01 -4.39726919e-01 5.31610012e-01 -6.09624445e-01
-5.28010607e-01 3.13276023e-01 -2.68034693e-02 -3.04981545e-02
-4.47122246e-01 -6.54656649e-01 -8.35874975e-01 -7.29731977e-01
-5.37099302e-01 1.43560082e-01 7.11075604e-01 1.11364365e+00
-1.34358197e-01 5.90958536e-01 5.85811734e-01 -8.61491561e-01
-2.61173248e-02 -1.03930902e+00 -1.18049800e+00 4.79862660e-01
-2.15894774e-01 -1.01041901e+00 -2.36313701e-01 5.47090918e-02] | [11.77140998840332, 2.1921300888061523] |
6f35d7b0-fbbd-43a5-b5ca-7fb8aeb902b5 | a-report-on-the-vardial-evaluation-campaign | null | null | https://aclanthology.org/2020.vardial-1.1 | https://aclanthology.org/2020.vardial-1.1.pdf | A Report on the VarDial Evaluation Campaign 2020 | This paper presents the results of the VarDial Evaluation Campaign 2020 organized as part of the seventh workshop on Natural Language Processing (NLP) for Similar Languages, Varieties and Dialects (VarDial), co-located with COLING 2020. The campaign included three shared tasks each focusing on a different challenge of language and dialect identification: Romanian Dialect Identification (RDI), Social Media Variety Geolocation (SMG), and Uralic Language Identification (ULI). The campaign attracted 30 teams who enrolled to participate in one or multiple shared tasks and 14 of them submitted runs across the three shared tasks. Finally, 11 papers describing participating systems are published in the VarDial proceedings and referred to in this report. | ['Marcos Zampieri', 'Yves Scherrer', 'Christoph Purschke', 'Niko Partanen', 'Nikola Ljubešić', 'Krister Lindén', 'Tommi Jauhiainen', 'Heidi Jauhiainen', 'Radu Tudor Ionescu', 'Dirk Hovy', 'Mihaela Gaman'] | null | null | null | null | vardial-coling-2020-12 | ['dialect-identification'] | ['natural-language-processing'] | [-3.70766848e-01 -2.48679817e-01 -8.98671970e-02 -5.16142607e-01
-1.31558323e+00 -1.15833795e+00 1.08172417e+00 4.19320285e-01
-5.90477645e-01 5.18654406e-01 6.18330538e-01 -2.35639498e-01
2.42339447e-01 -5.04807770e-01 -2.46927693e-01 -1.69282705e-01
-1.83278188e-01 1.03751266e+00 -1.41185150e-01 -3.33179206e-01
3.90415400e-01 5.76641262e-01 -1.14096785e+00 5.82662940e-01
7.93816149e-01 2.84356534e-01 -3.63085605e-02 7.57441938e-01
-4.22063738e-01 5.40311754e-01 -4.70970869e-01 -4.48545367e-01
1.66988239e-01 -5.81117682e-02 -1.25195193e+00 -5.97896874e-01
7.81774282e-01 1.78611532e-01 7.80622512e-02 9.86526370e-01
7.73219287e-01 2.88230125e-02 6.61560774e-01 -1.19339228e+00
-3.98750722e-01 1.31764388e+00 -5.34648001e-01 2.36044079e-01
8.24294627e-01 -3.12877536e-01 1.19754040e+00 -1.27701473e+00
1.02705038e+00 1.60381818e+00 1.03807044e+00 1.61713898e-01
-1.33240008e+00 -5.49781799e-01 -6.48932233e-02 -6.35095239e-02
-1.99391806e+00 -9.89642322e-01 4.40889180e-01 -6.01475120e-01
1.18389916e+00 4.02984560e-01 8.84148926e-02 7.92850912e-01
-5.97074866e-01 1.10850894e+00 1.18087327e+00 -6.61222458e-01
-2.52609234e-02 6.32095158e-01 4.61796075e-01 2.20235288e-01
-9.23635662e-02 -3.72109324e-01 -3.63105118e-01 -3.84742618e-01
3.14553022e-01 -1.08612430e+00 -5.95502146e-02 1.75940454e-01
-1.17225969e+00 9.67616856e-01 -1.71352968e-01 7.47581482e-01
-2.18094021e-01 -5.70477605e-01 8.93259585e-01 5.11424124e-01
9.01436687e-01 6.49346292e-01 -4.42138255e-01 -1.95660025e-01
-7.29214013e-01 4.78839010e-01 9.71883953e-01 9.41454947e-01
6.03411973e-01 -7.82597214e-02 -1.89411640e-01 1.63449681e+00
3.66258651e-01 5.48016965e-01 5.58926642e-01 -5.92775881e-01
7.32278168e-01 3.17237765e-01 4.03762572e-02 -7.49428332e-01
-6.34561539e-01 2.28227541e-01 -4.12267148e-01 -1.76204011e-01
7.47686565e-01 -6.36223853e-01 -4.60362256e-01 1.60186589e+00
3.05523217e-01 -3.11160594e-01 -1.82434283e-02 5.27778804e-01
1.13068640e+00 8.61363113e-01 6.13237582e-02 1.04177676e-01
1.49774992e+00 -4.82032478e-01 -4.54293609e-01 -4.74516638e-02
9.34368789e-01 -1.47186124e+00 9.22968328e-01 2.64060259e-01
-1.08605504e+00 -4.04190809e-01 -6.60551429e-01 -1.70951799e-01
-7.89256811e-01 -7.91931376e-02 5.26396513e-01 1.01130188e+00
-1.46345520e+00 -4.51636091e-02 -4.96525243e-02 -8.92094731e-01
-9.29045901e-02 5.68294525e-02 -3.85675669e-01 -4.64002602e-02
-1.47483480e+00 7.26752698e-01 -2.05811355e-02 -1.65915683e-01
-3.33986700e-01 -8.26360762e-01 -9.10903513e-01 -4.77849633e-01
-3.36427212e-01 1.10194944e-01 1.10764766e+00 -8.72841775e-01
-1.39511418e+00 1.96281052e+00 -2.08683580e-01 -3.07561338e-01
6.27523065e-01 -2.93741077e-01 -8.85624588e-01 -4.10572499e-01
5.13501585e-01 5.02982080e-01 6.40910268e-02 -1.03973341e+00
-8.52985620e-01 -4.84817356e-01 -3.38965565e-01 3.36037487e-01
4.22230735e-02 1.05685270e+00 -4.51828003e-01 -7.55230486e-01
-1.62417054e-01 -8.85870337e-01 1.21196151e-01 -1.05100870e+00
-5.64382553e-01 -7.69950032e-01 4.72479463e-01 -1.21573603e+00
9.92496550e-01 -2.23940873e+00 -1.15741551e-01 5.02577901e-01
-9.06049460e-02 3.55456650e-01 -3.01160425e-01 9.10552442e-01
-1.23928972e-01 2.72185117e-01 1.66701749e-01 -6.61641479e-01
2.58084983e-01 -2.66816050e-01 -3.70845586e-01 6.68773055e-01
-3.82842302e-01 7.13181913e-01 -6.90577865e-01 -3.19124103e-01
-2.82634161e-02 2.91738033e-01 2.15246845e-02 -1.60516024e-01
-1.91699211e-02 1.76604792e-01 7.01805651e-02 6.13199770e-01
1.13833463e+00 5.97598910e-01 4.65763956e-01 1.96718648e-01
-9.13855791e-01 6.89166546e-01 -1.32323444e+00 1.32671940e+00
-5.42642117e-01 9.03218865e-01 5.73777676e-01 -4.90947008e-01
9.72803771e-01 2.73572326e-01 3.08650732e-01 -5.29925644e-01
-3.12324557e-02 4.61468607e-01 -4.17983025e-01 -2.37798616e-01
9.80286479e-01 3.31067085e-01 -5.85370481e-01 7.06194580e-01
2.34361999e-02 6.25917166e-02 2.83460766e-01 2.80219823e-01
5.02000988e-01 -1.29999921e-01 4.41925138e-01 -9.42605257e-01
8.62120867e-01 1.32886618e-01 5.12126744e-01 8.85992646e-01
-4.15133357e-01 5.44647276e-01 3.33275616e-01 -3.23163956e-01
-1.05578327e+00 -1.28372061e+00 -3.23494166e-01 1.49482882e+00
-3.53428543e-01 -4.85210687e-01 -7.28316545e-01 -4.83721972e-01
-1.10170342e-01 7.10527062e-01 -3.20118040e-01 6.99704766e-01
-9.51137900e-01 -6.88702703e-01 1.44342303e+00 5.60653321e-02
4.36425060e-01 -1.20608461e+00 5.57688355e-01 -7.62108937e-02
-6.46352947e-01 -9.97603536e-01 -8.10794771e-01 -2.64142245e-01
6.23449171e-03 -8.46889734e-01 -9.70293164e-01 -1.61808288e+00
1.86715312e-02 3.38606350e-02 1.28985977e+00 -4.46371615e-01
-3.92833620e-01 3.75407815e-01 -3.32994133e-01 -4.05198365e-01
-6.47511065e-01 6.44128203e-01 6.38576001e-02 1.11702040e-01
8.58097315e-01 -1.08468540e-01 1.03314161e-01 1.35837242e-01
-2.60190666e-01 -4.35883999e-01 -1.81029364e-01 3.59650344e-01
2.01904401e-01 -3.95317465e-01 6.67592883e-01 -1.38326705e+00
8.34682643e-01 -6.23664618e-01 -7.03725994e-01 4.00814325e-01
-2.86892086e-01 -4.72837895e-01 4.52455640e-01 1.20362956e-02
-1.36540735e+00 1.12481624e-01 -6.72477603e-01 5.66031218e-01
-5.90575337e-01 4.47838843e-01 -5.55478811e-01 6.89591244e-02
7.08363295e-01 2.13577628e-01 -1.91831633e-01 -7.43347585e-01
5.83605826e-01 1.16470206e+00 5.43620110e-01 -5.49537480e-01
4.88151193e-01 -2.53705098e-03 -8.99762988e-01 -1.42656076e+00
-2.10956320e-01 -1.00390816e+00 -7.36047387e-01 -1.96323007e-01
8.15888703e-01 -1.28896093e+00 -4.31086332e-01 1.23697162e+00
-1.14630878e+00 -4.53700244e-01 -1.75199267e-02 3.41088980e-01
-1.31714329e-01 3.35840493e-01 -9.20290709e-01 -8.09826910e-01
-3.45365673e-01 -6.42777085e-01 8.49239707e-01 -1.40689155e-02
-7.34438300e-01 -1.33607531e+00 7.29340851e-01 3.01665068e-01
3.67324471e-01 1.14993530e-03 8.45382333e-01 -9.78661478e-01
1.46895602e-01 -1.76883996e-01 -2.80716598e-01 -9.96553246e-03
-1.23233796e-04 8.29717219e-02 -1.18341136e+00 -1.59602433e-01
-5.29080868e-01 -5.68002343e-01 6.75966203e-01 3.58238578e-01
2.33374432e-01 -1.24253094e-01 -2.55198538e-01 4.61495727e-01
1.31940079e+00 3.91936637e-02 5.57404578e-01 3.65240872e-01
6.97142184e-01 1.07161367e+00 2.90410370e-01 5.59973478e-01
9.44648027e-01 7.29411602e-01 -4.54157978e-01 -2.17517689e-02
-1.51432812e-01 -1.49570599e-01 6.10106528e-01 9.71235871e-01
1.74276903e-01 -1.63333416e-01 -1.47274721e+00 9.99931455e-01
-1.50211775e+00 -9.66015339e-01 -6.88156664e-01 2.27006984e+00
9.53002632e-01 -4.43970650e-01 8.56895268e-01 -2.02685773e-01
1.17086279e+00 2.70182282e-01 4.38028798e-02 -8.99070740e-01
-7.87055075e-01 1.74372748e-01 5.77816188e-01 1.15910470e+00
-1.08607340e+00 1.48055649e+00 7.03878164e+00 8.88071060e-01
-9.58748579e-01 1.31323934e-01 6.82836533e-01 4.41898137e-01
-4.72342938e-01 -9.01030451e-02 -1.29190433e+00 1.50897071e-01
9.87008333e-01 -5.20492136e-01 7.41400421e-01 5.93693674e-01
-1.19980790e-01 -5.70885725e-02 -6.50399327e-01 8.35717559e-01
2.77983159e-01 -1.03063595e+00 -2.87158698e-01 -4.44094390e-02
5.16817033e-01 9.21245933e-01 -4.62129973e-02 4.38678473e-01
7.23163903e-01 -9.22658086e-01 7.21151292e-01 1.51386052e-01
9.02232051e-01 -9.50325847e-01 5.14154434e-01 9.09253582e-02
-1.20737147e+00 3.54409188e-01 -1.64524809e-01 1.58277869e-01
3.00126791e-01 3.55412900e-01 -9.99391675e-01 3.68126512e-01
5.92305899e-01 4.60063368e-01 -6.33051813e-01 8.81012797e-01
5.88458888e-02 8.26080203e-01 -5.14581800e-01 -1.67444527e-01
1.37890935e-01 -1.74776375e-01 8.88137519e-01 1.84806454e+00
-3.91458310e-02 -4.06231999e-01 3.61146390e-01 3.07137638e-01
-8.42657462e-02 7.43171871e-01 -6.19988799e-01 -6.21150769e-02
6.95295751e-01 1.31085861e+00 -6.91469729e-01 -3.08735967e-01
-2.91508734e-01 9.47507620e-01 3.63940597e-01 2.71006405e-01
-3.07188869e-01 -6.95423603e-01 8.61189902e-01 -3.08194794e-02
-7.78259560e-02 -1.79405689e-01 -5.76834619e-01 -1.05544007e+00
-2.46655524e-01 -9.58714604e-01 9.18487966e-01 -1.03936166e-01
-1.75897253e+00 9.26458538e-01 -1.54400870e-01 -5.91780543e-01
-5.39279222e-01 -4.66577291e-01 -5.74179053e-01 1.35337925e+00
-1.08549678e+00 -1.20269191e+00 1.29388064e-01 7.86138713e-01
4.41682637e-01 -6.83754146e-01 1.21019220e+00 6.83743358e-01
-5.11759639e-01 1.05718076e+00 3.54254007e-01 4.92654115e-01
1.32372510e+00 -1.47287810e+00 9.39433157e-01 5.24685085e-01
1.95477888e-01 5.31831443e-01 4.30018246e-01 -6.11272573e-01
-9.44377661e-01 -1.02605641e+00 2.14444709e+00 -3.88138831e-01
9.09691036e-01 -9.55703914e-01 -5.02546608e-01 8.57198000e-01
6.61571741e-01 -7.87304282e-01 7.52543032e-01 6.96293235e-01
-5.22154689e-01 1.08657457e-01 -1.31107259e+00 5.03696978e-01
7.85716593e-01 -8.15096557e-01 -2.33591795e-01 6.39390051e-01
4.74738061e-01 -2.42202714e-01 -8.23860884e-01 -1.68110371e-01
4.81666356e-01 -6.00092590e-01 8.26978862e-01 -3.31052065e-01
-5.00313453e-02 5.08018807e-02 -4.85550582e-01 -1.46786678e+00
-3.59722733e-01 -7.09596276e-01 1.16199648e+00 2.17451453e+00
6.62207007e-01 -8.80479932e-01 4.53767896e-01 3.34106714e-01
-3.69634554e-02 4.32044983e-01 -8.30686212e-01 -6.39019132e-01
5.92260122e-01 -4.81794685e-01 5.31659961e-01 1.51787317e+00
2.61140287e-01 7.16809988e-01 -2.73942918e-01 -6.32536933e-02
5.52036345e-01 -1.22364707e-01 7.82316267e-01 -1.08750117e+00
-2.38269851e-01 -7.46676147e-01 -2.36144051e-01 -5.91817975e-01
2.81886280e-01 -1.36803854e+00 8.43864456e-02 -1.23712718e+00
-3.47618461e-01 -8.26877356e-01 1.41186431e-01 4.38880116e-01
6.54958189e-02 3.95346373e-01 2.74144500e-01 2.13002503e-01
-5.14572263e-01 -1.09962106e-01 1.11611731e-01 -1.33997172e-01
-3.66242170e-01 1.68768600e-01 -6.61139429e-01 4.89360303e-01
9.39175010e-01 -1.27729595e-01 1.02884769e-01 -4.97139841e-01
2.64735013e-01 -1.83004200e-01 -2.09973186e-01 -6.43689454e-01
1.00666158e-01 1.92903787e-01 2.23205596e-01 -7.17479765e-01
1.66399390e-01 1.89306773e-02 -1.89805269e-01 6.19369522e-02
-3.41920316e-01 4.55304474e-01 4.13462937e-01 -3.92077506e-01
-3.63167673e-01 -1.62727490e-01 7.41358995e-01 -1.17640302e-01
-9.43154514e-01 1.16377078e-01 -1.06460571e+00 5.65666497e-01
8.26375723e-01 2.53644168e-01 -2.12897986e-01 -2.63539493e-01
-5.81290066e-01 4.56537277e-01 3.24631184e-01 5.22367835e-01
8.28298181e-02 -1.28955162e+00 -1.57204258e+00 2.35775724e-01
3.18951577e-01 -7.56117880e-01 3.15432370e-01 4.63075042e-01
-7.38076806e-01 5.52103281e-01 -1.37309805e-01 -2.74879724e-01
-1.54474819e+00 1.64460018e-01 1.72903344e-01 -4.20830131e-01
-1.97838262e-01 1.13812888e+00 -6.15903661e-02 -1.49375463e+00
1.94009781e-01 7.12213099e-01 -5.99612117e-01 5.52749395e-01
7.21731424e-01 4.90307420e-01 2.27490552e-02 -1.14688909e+00
-6.67311311e-01 3.16037625e-01 -3.00636113e-01 -6.47583425e-01
1.12279129e+00 -3.95152956e-01 -5.93688428e-01 7.74052560e-01
1.34859776e+00 5.91607749e-01 -2.02351362e-01 -5.92260718e-01
6.14789426e-01 -6.45917207e-02 -2.22952023e-01 -1.03499949e+00
-4.49676424e-01 4.12794322e-01 3.97244811e-01 3.36602032e-01
5.84714174e-01 -7.41431937e-02 7.91891277e-01 -7.22777992e-02
2.74780691e-01 -1.56642497e+00 -8.76011133e-01 1.14991939e+00
8.30341160e-01 -1.09222734e+00 -5.28170347e-01 -6.00000918e-01
-7.62290776e-01 8.35370600e-01 4.06383872e-01 3.36235352e-02
9.66469109e-01 4.52022403e-01 5.44940054e-01 -3.67144030e-03
-3.06205392e-01 -1.89732686e-01 1.52319774e-01 9.60014343e-01
9.64828908e-01 4.07181233e-01 -5.85989714e-01 3.99279803e-01
-4.72358912e-01 -5.12835503e-01 2.52404898e-01 4.55763251e-01
-1.66015670e-01 -1.38021183e+00 -5.31225324e-01 2.43240222e-02
-3.80444020e-01 -6.26204848e-01 -1.03260195e+00 8.05177689e-01
-5.53731527e-03 1.06573570e+00 2.99794793e-01 -2.92863756e-01
2.48670623e-01 2.61486471e-01 3.56602669e-01 -5.34683406e-01
-1.18261373e+00 -3.35047091e-03 9.94808555e-01 -2.04735503e-01
9.26763227e-04 -1.28105164e+00 -9.93388474e-01 -8.63798201e-01
3.92486125e-01 2.92418510e-01 7.26188540e-01 6.19420350e-01
1.16964124e-01 -2.72437811e-01 7.64363885e-01 -5.26466787e-01
-8.65563899e-02 -1.12755191e+00 -9.51336443e-01 2.80637801e-01
-1.18890226e-01 7.96777606e-02 -1.58426419e-01 -2.30401665e-01] | [10.19018840789795, 10.733905792236328] |
47126b6b-538e-4548-8ca3-675d6cce9f12 | end-to-end-learning-of-keypoint-detection-and | 2104.01085 | null | https://arxiv.org/abs/2104.01085v1 | https://arxiv.org/pdf/2104.01085v1.pdf | End-to-end learning of keypoint detection and matching for relative pose estimation | We propose a new method for estimating the relative pose between two images, where we jointly learn keypoint detection, description extraction, matching and robust pose estimation. While our architecture follows the traditional pipeline for pose estimation from geometric computer vision, all steps are learnt in an end-to-end fashion, including feature matching. We demonstrate our method for the task of visual localization of a query image within a database of images with known pose. Pairwise pose estimation has many practical applications for robotic mapping, navigation, and AR. For example, the display of persistent AR objects in the scene relies on a precise camera localization to make the digital models appear anchored to the physical environment. We train our pipeline end-to-end specifically for the problem of visual localization. We evaluate our proposed approach on localization accuracy, robustness and runtime speed. Our method achieves state of the art localization accuracy on the 7 Scenes dataset. | ['Marco Paladini', 'Nikola Sivacki', 'Luca Del Pero', 'Antoine Fond'] | 2021-04-02 | null | null | null | null | ['camera-localization'] | ['computer-vision'] | [-1.84467621e-03 -2.36359611e-01 4.30733487e-02 -4.13125277e-01
-1.25916576e+00 -1.05539095e+00 5.77045441e-01 2.96307147e-01
-6.14600420e-01 1.15973633e-02 -2.38373861e-01 -1.04454964e-01
-2.00702213e-02 -1.45252243e-01 -1.16624308e+00 -1.26094759e-01
-1.58639923e-02 7.27723002e-01 5.02243876e-01 1.83840185e-01
6.29958689e-01 1.00409925e+00 -1.22744477e+00 -4.17674899e-01
1.83915421e-01 1.31522775e+00 4.89192307e-01 9.99360621e-01
3.51657391e-01 5.63287795e-01 -2.47176260e-01 8.38121120e-03
4.97294337e-01 3.28720599e-01 -5.25423229e-01 9.07949582e-02
1.17930448e+00 -5.49671590e-01 -4.72338378e-01 9.04635429e-01
4.87483710e-01 1.16638735e-01 5.13191283e-01 -1.25569820e+00
-2.49949813e-01 1.49682388e-02 -8.58884871e-01 -2.36310974e-01
7.74593055e-01 -2.76205316e-02 7.46057808e-01 -1.36792660e+00
6.86345458e-01 1.22671092e+00 8.21883202e-01 7.22175539e-02
-1.05268431e+00 -4.79024738e-01 6.18440844e-02 -4.31940705e-02
-1.72496212e+00 -7.26098835e-01 6.23013854e-01 -5.73562443e-01
6.50421798e-01 9.01683513e-03 2.49819487e-01 5.79540431e-01
4.75603603e-02 4.36320782e-01 5.99509120e-01 -3.77259493e-01
2.21011922e-01 1.89295709e-02 -1.30239502e-01 9.94744122e-01
6.49476051e-02 -1.26265110e-02 -6.37681305e-01 -1.59885198e-01
9.02714610e-01 2.08986476e-01 -6.96185827e-02 -1.26917756e+00
-1.58229828e+00 3.79462957e-01 6.94854856e-01 -4.17424142e-01
-2.65431076e-01 7.45238304e-01 -9.02567282e-02 5.75909168e-02
-1.11262836e-02 3.46975088e-01 -4.67063934e-01 -5.40313916e-03
-5.67852139e-01 2.38869816e-01 5.72319090e-01 1.45179033e+00
9.04120684e-01 -5.99657536e-01 1.23538658e-01 3.12758505e-01
6.38500750e-01 9.18305516e-01 -3.97704728e-02 -1.23945177e+00
3.17764550e-01 3.39334637e-01 6.48221374e-01 -1.15218413e+00
-2.61672765e-01 -2.15384990e-01 -2.23185107e-01 2.39836290e-01
2.01701596e-01 2.26753145e-01 -8.25965106e-01 1.35810554e+00
6.11221850e-01 7.00537562e-02 8.54562297e-02 9.89547253e-01
5.37080348e-01 3.27996641e-01 -3.20586830e-01 3.23244840e-01
1.09734952e+00 -1.00882351e+00 -2.63088197e-01 -7.24917829e-01
2.82498777e-01 -1.13347578e+00 5.96208751e-01 1.44160077e-01
-7.85580397e-01 -5.04198253e-01 -9.52391922e-01 -5.16298115e-01
-1.45058960e-01 6.17693067e-01 4.03779179e-01 1.49597991e-02
-1.05406189e+00 1.31051436e-01 -9.34370935e-01 -6.36264384e-01
6.38127476e-02 5.65920353e-01 -8.39361191e-01 -1.69322044e-01
-2.20032871e-01 8.43506515e-01 1.83828741e-01 1.88190013e-01
-1.03524375e+00 -4.80917245e-01 -1.16446459e+00 -2.54461974e-01
2.88360059e-01 -8.80964577e-01 1.33959174e+00 -2.67389238e-01
-1.26258266e+00 9.68086720e-01 -3.72515440e-01 -3.29518050e-01
7.04522252e-01 -6.61753595e-01 3.32688361e-01 9.77025107e-02
4.12818283e-01 8.28855634e-01 8.20615649e-01 -1.42982137e+00
-7.91484833e-01 -6.14949524e-01 -1.10405251e-01 3.88881803e-01
4.57930326e-01 -8.64597112e-02 -1.04532683e+00 -3.45867239e-02
8.23564053e-01 -1.12640619e+00 -3.35966110e-01 7.20394433e-01
-3.83000255e-01 4.73877192e-02 1.05118215e+00 -3.95594269e-01
2.50824392e-01 -2.24983144e+00 8.35187957e-02 3.29831570e-01
5.35259619e-02 -4.91761565e-01 1.57003608e-02 4.73913699e-01
3.43682379e-01 -4.63088304e-01 2.16523543e-01 -7.48045385e-01
1.21961899e-01 -2.72880584e-01 -5.14798641e-01 1.02864611e+00
-1.61920816e-01 8.34376991e-01 -8.13173056e-01 -3.35635513e-01
4.86805230e-01 5.07778227e-01 -3.76232415e-01 5.73394775e-01
-1.26118109e-01 5.24148285e-01 -3.41235369e-01 8.91000152e-01
7.29452074e-01 -9.96316075e-02 -2.71267831e-01 -5.09675801e-01
-2.00705051e-01 -6.21763207e-02 -1.35064793e+00 2.35635448e+00
-4.47875291e-01 5.87879837e-01 1.92209825e-01 -2.40887001e-01
8.95645559e-01 -2.28297487e-01 2.60481298e-01 -2.89383143e-01
-1.90775166e-03 2.41327018e-01 -4.83578920e-01 -1.80424284e-02
7.48588026e-01 6.55886352e-01 -3.37494761e-01 1.45487025e-01
-1.03765331e-01 -4.11046565e-01 -2.28294134e-01 1.70870066e-01
1.13739753e+00 5.29039979e-01 4.26283032e-01 1.25500068e-01
2.84303874e-01 3.12094033e-01 3.50584686e-02 7.48344123e-01
3.02831698e-02 6.95585370e-01 -1.34768814e-01 -3.24160427e-01
-1.05886078e+00 -1.28571618e+00 7.14874640e-02 9.48305309e-01
7.60487735e-01 -4.15616125e-01 -2.28811070e-01 -5.06162226e-01
2.75010765e-01 3.09580658e-02 -2.78986096e-01 1.30743578e-01
-3.97746652e-01 3.90625894e-01 1.30480051e-01 5.89772284e-01
4.06615198e-01 -5.31179845e-01 -1.13281453e+00 -1.85906410e-01
7.84018636e-02 -1.42336380e+00 -6.86876655e-01 1.68463737e-01
-5.74922442e-01 -1.19703853e+00 -4.04018879e-01 -9.78642106e-01
9.25140560e-01 6.80305243e-01 8.16291094e-01 -2.22105190e-01
-4.29312259e-01 9.68422234e-01 7.81243071e-02 -3.03624749e-01
6.87868297e-02 1.22042582e-03 2.40245417e-01 -2.18605638e-01
1.12468265e-02 -3.13143253e-01 -9.54629719e-01 5.31646907e-01
-2.61528403e-01 -8.76051113e-02 8.17070067e-01 2.92825550e-01
1.12544215e+00 -4.48876351e-01 -3.01554829e-01 -2.91963577e-01
2.29642005e-03 4.34270091e-02 -1.11807203e+00 2.21364528e-01
-1.67389005e-01 4.84988764e-02 4.99218479e-02 -3.06475967e-01
-4.67715889e-01 1.15147007e+00 2.35382840e-01 -8.33744884e-01
-2.05090389e-01 2.39556208e-01 -1.15176819e-01 -7.18119860e-01
5.27004600e-01 9.32806879e-02 -4.02204804e-02 -3.97785842e-01
6.88127995e-01 5.39024413e-01 1.07485688e+00 -4.58816350e-01
1.08714104e+00 6.30569816e-01 2.45626628e-01 -6.01506174e-01
-5.39975286e-01 -9.02218461e-01 -1.06219411e+00 4.95692864e-02
6.72111988e-01 -1.38026559e+00 -1.04035306e+00 1.37015745e-01
-1.40053308e+00 -1.08355030e-01 7.21140578e-02 5.57128191e-01
-8.01316082e-01 1.46508068e-01 1.09730766e-03 -6.45916998e-01
-3.61149579e-01 -1.20988178e+00 1.90228033e+00 1.35995988e-02
-1.06163183e-03 -5.69808125e-01 2.06849560e-01 -1.88970454e-02
6.03583679e-02 3.12174588e-01 2.88824260e-01 -2.06681654e-01
-1.20376348e+00 -5.91462970e-01 -3.61180395e-01 -3.26820970e-01
-5.17008342e-02 -2.23974347e-01 -9.24429059e-01 -5.43409288e-01
-4.22053456e-01 -3.56820494e-01 6.29836857e-01 1.06921002e-01
7.78225958e-01 -3.98461074e-02 -7.57614553e-01 8.17303896e-01
1.63593674e+00 8.89895633e-02 1.88979089e-01 4.61446851e-01
9.12384391e-01 4.01755810e-01 1.05525589e+00 2.24469900e-01
5.57564259e-01 9.68371630e-01 7.60768294e-01 5.47225699e-02
7.13193342e-02 -5.24985790e-01 3.23098242e-01 1.02434471e-01
5.46374977e-01 9.70666036e-02 -1.08027804e+00 4.39052522e-01
-2.04282904e+00 -3.05515110e-01 1.79831356e-01 2.43411922e+00
2.70320505e-01 -1.14609554e-01 -1.54783994e-01 -4.62698489e-01
5.18997252e-01 -5.26062399e-02 -7.43074954e-01 2.41461441e-01
4.03766513e-01 -3.56292397e-01 1.06512308e+00 7.13070035e-01
-1.26829398e+00 1.16553652e+00 6.13348198e+00 1.73786804e-01
-1.12216926e+00 -1.47389844e-01 1.32903531e-01 1.05109096e-01
2.21111909e-01 2.81112224e-01 -1.00051153e+00 -1.30933091e-01
3.52282286e-01 1.90495670e-01 1.64109066e-01 1.26287806e+00
-1.49195150e-01 -3.53911757e-01 -1.65268064e+00 1.47429478e+00
2.17394263e-01 -1.17256320e+00 -2.82650739e-01 -8.74292031e-02
3.81711513e-01 3.17251772e-01 1.42397165e-01 -9.59632918e-02
1.90109611e-01 -8.69161367e-01 1.14611745e+00 5.12506783e-01
9.17037904e-01 -7.70420671e-01 4.58379507e-01 4.42424446e-01
-1.30598342e+00 8.28127787e-02 -5.58441520e-01 1.28859848e-01
1.26480550e-01 -5.47950948e-03 -1.18320704e+00 2.12232605e-01
5.63805938e-01 6.03936017e-01 -9.37872112e-01 1.21015024e+00
-2.12151945e-01 -3.44670713e-01 -4.75918919e-01 2.65754282e-01
3.64430808e-02 1.55157521e-01 4.73623574e-01 8.03151011e-01
4.08747584e-01 -1.65290713e-01 6.66397572e-01 5.58346152e-01
-2.70088911e-02 -1.98374584e-01 -9.54595268e-01 3.14142704e-01
7.78083324e-01 1.32540739e+00 -8.16133022e-01 1.91184491e-01
-7.24542663e-02 1.33520114e+00 5.12803197e-01 6.12649806e-02
-6.83076322e-01 -4.76476848e-01 5.96397996e-01 1.11940257e-01
4.37348008e-01 -8.33936155e-01 8.76663774e-02 -8.39084864e-01
2.99730808e-01 -3.86011153e-01 -8.86960607e-03 -1.08837759e+00
-8.00178468e-01 4.88204718e-01 -4.96902689e-02 -1.24923611e+00
-3.98530394e-01 -6.06013536e-01 -1.88443333e-01 6.53619587e-01
-1.26914036e+00 -1.54666698e+00 -7.32860327e-01 6.53912008e-01
4.56867784e-01 2.71785744e-02 7.49601603e-01 -6.06633537e-03
7.39069358e-02 3.90694886e-01 1.01593556e-02 2.11737201e-01
1.03078830e+00 -1.23225605e+00 6.68440402e-01 9.17139709e-01
4.42229897e-01 7.85827994e-01 6.63350463e-01 -5.72688162e-01
-2.21743965e+00 -1.08736205e+00 5.03078043e-01 -1.04407203e+00
4.92377698e-01 -8.03363919e-01 -1.70635477e-01 1.01569390e+00
-1.27721235e-01 4.06429052e-01 -4.13074605e-02 -1.96896926e-01
-5.46328187e-01 -3.04477125e-01 -1.03112519e+00 4.69396472e-01
1.06728220e+00 -6.93852067e-01 -3.08631390e-01 4.74571049e-01
7.76187479e-01 -1.14962924e+00 -5.39336085e-01 1.98105171e-01
7.89004922e-01 -5.53658009e-01 1.30608034e+00 6.43268675e-02
-1.94197491e-01 -9.43851709e-01 -5.76982379e-01 -8.41094732e-01
-1.33428380e-01 -6.79325640e-01 -1.83857977e-02 9.11182523e-01
8.59058872e-02 -1.68189943e-01 8.29313815e-01 5.73317289e-01
7.49794394e-02 -2.23532915e-01 -8.57710600e-01 -5.01274288e-01
-8.15404415e-01 -4.43504959e-01 2.27829725e-01 3.12764347e-01
-5.09633899e-01 5.95142603e-01 -3.60479712e-01 8.68704200e-01
8.59169781e-01 4.65446442e-01 1.50402665e+00 -1.12697995e+00
-1.77412927e-01 -4.01290599e-03 -9.26547170e-01 -1.51142383e+00
1.83949977e-01 -5.19883454e-01 5.56494117e-01 -1.53536737e+00
1.63487062e-01 -4.30325419e-01 2.38306373e-01 3.09754372e-01
2.81040609e-01 4.34663475e-01 1.71769291e-01 4.05136108e-01
-1.07241356e+00 1.59097582e-01 5.63538551e-01 -4.23105843e-02
-1.20141551e-01 -8.96264315e-02 -2.90984392e-01 6.79396510e-01
3.48434836e-01 -3.98899198e-01 -3.54437470e-01 -6.72726035e-01
1.69789031e-01 1.78318039e-01 7.88916588e-01 -1.16950083e+00
8.26651573e-01 1.65689159e-02 7.36891627e-01 -1.12569022e+00
6.75704241e-01 -1.28427410e+00 1.96720526e-01 3.52153718e-01
-2.17254624e-01 5.75798512e-01 8.30326676e-02 8.09648395e-01
3.20990793e-02 1.13784492e-01 5.10229886e-01 -6.02851585e-02
-1.16384912e+00 5.40737629e-01 2.67804176e-01 -3.45787168e-01
1.24396086e+00 -1.46533519e-01 -2.42686868e-01 -4.65836674e-01
-5.04696727e-01 3.11792105e-01 1.00479662e+00 4.50435102e-01
9.68731105e-01 -1.17572987e+00 -3.49750400e-01 3.05658758e-01
6.60302281e-01 5.29712915e-01 -1.44561917e-01 8.25047016e-01
-9.84237492e-01 3.23294759e-01 1.83076724e-01 -1.22502697e+00
-1.46608937e+00 6.09828532e-01 2.75587440e-01 3.98700684e-01
-4.79680449e-01 8.27572227e-01 2.89577723e-01 -5.59548795e-01
7.32608497e-01 -2.87196249e-01 3.34460944e-01 -4.12402242e-01
4.49334681e-01 1.82452023e-01 6.51640724e-03 -8.14937711e-01
-7.89655507e-01 1.14264596e+00 -1.23661719e-01 -3.17234337e-01
1.04321194e+00 -4.37298924e-01 -1.85773656e-01 3.73204529e-01
1.32526255e+00 3.40844393e-01 -1.44437909e+00 -3.38498145e-01
4.97947559e-02 -7.61193037e-01 -2.00393815e-02 -6.62282169e-01
-4.79401052e-01 6.12270713e-01 9.74840522e-01 -6.83662772e-01
5.92214584e-01 4.33314204e-01 2.82035470e-01 9.95799720e-01
8.49671781e-01 -5.71397543e-01 2.44421378e-01 6.63469911e-01
9.35349107e-01 -1.48426223e+00 1.92727402e-01 -2.70850152e-01
-2.90064573e-01 1.09859633e+00 5.64602613e-01 -3.80800247e-01
4.28736985e-01 3.89535904e-01 1.74662948e-01 -1.73906863e-01
-3.69551986e-01 -1.10075407e-01 4.51083362e-01 7.14973748e-01
6.05875859e-03 -2.29322746e-01 5.77463925e-01 -3.27336155e-02
-1.41446173e-01 -4.16345239e-01 7.02929348e-02 1.11326873e+00
-5.32603383e-01 -6.72381699e-01 -4.69482243e-01 -3.09668422e-01
-3.54226790e-02 2.22970378e-02 -7.11155534e-01 8.23798060e-01
-2.24546149e-01 6.60338998e-01 1.80700436e-01 -2.89373755e-01
4.51330125e-01 -2.58720398e-01 7.36894131e-01 -5.72128952e-01
-3.25062573e-02 1.63396150e-01 -2.84077018e-01 -1.11188447e+00
-2.77719051e-01 -6.11505270e-01 -1.15144682e+00 6.40340969e-02
-1.61389187e-01 -1.05917454e-01 1.19996536e+00 6.39564157e-01
5.52641869e-01 -4.03458364e-02 5.86293280e-01 -1.26797926e+00
-5.64038098e-01 -6.30096793e-01 -2.01025859e-01 7.78409392e-02
8.31533611e-01 -5.31319439e-01 3.96988876e-02 -6.02275245e-02] | [7.615121841430664, -2.301884412765503] |
23262fa0-9e2b-4163-880d-87e81f3e38b0 | dense-captioning-events-in-videos-sysu | 2006.11693 | null | https://arxiv.org/abs/2006.11693v2 | https://arxiv.org/pdf/2006.11693v2.pdf | Dense-Captioning Events in Videos: SYSU Submission to ActivityNet Challenge 2020 | This technical report presents a brief description of our submission to the dense video captioning task of ActivityNet Challenge 2020. Our approach follows a two-stage pipeline: first, we extract a set of temporal event proposals; then we propose a multi-event captioning model to capture the event-level temporal relationships and effectively fuse the multi-modal information. Our approach achieves a 9.28 METEOR score on the test set. | ['Teng Wang', 'Mingjing Yu', 'Huicheng Zheng'] | 2020-06-21 | null | null | null | null | ['dense-captioning', 'dense-video-captioning'] | ['computer-vision', 'computer-vision'] | [ 1.07761994e-01 -1.23847112e-01 -3.56514424e-01 -4.86153036e-01
-1.16451085e+00 -4.19802725e-01 8.66625667e-01 3.63293886e-02
-5.07565856e-01 7.04060793e-01 9.42573607e-01 3.40266943e-01
2.29082808e-01 -2.93941081e-01 -8.46661627e-01 -1.94904074e-01
-4.22596663e-01 4.18029308e-01 6.92701399e-01 5.52851856e-02
-2.08557427e-01 1.77870855e-01 -1.31206036e+00 9.29652333e-01
1.32880524e-01 1.24243641e+00 1.99200062e-04 9.56832349e-01
2.40934432e-01 1.26944232e+00 -5.86521327e-01 -2.27208108e-01
-7.69772828e-02 -4.82816368e-01 -9.20630097e-01 -4.09968011e-02
3.96070540e-01 -6.59986436e-01 -1.04953778e+00 6.50010586e-01
4.48193550e-01 3.75308782e-01 3.19705784e-01 -1.47999406e+00
-1.46613985e-01 6.49523854e-01 -5.86603820e-01 1.11620593e+00
7.58957922e-01 1.89738110e-01 1.12799263e+00 -9.90928113e-01
6.75484836e-01 8.74420881e-01 4.40832168e-01 4.72044528e-01
-6.36028230e-01 -4.90107715e-01 3.12167466e-01 6.99500203e-01
-1.38768351e+00 -4.63168114e-01 5.55298090e-01 -3.91572237e-01
1.24625099e+00 3.33247185e-02 8.31472874e-01 1.64888310e+00
-4.62741144e-02 1.25166214e+00 4.68282551e-01 2.30574667e-01
-3.54553945e-02 -5.36828220e-01 1.64784297e-01 4.68090624e-01
-2.43057340e-01 3.77994552e-02 -1.06583226e+00 -4.30882573e-02
5.97103834e-01 -2.10970372e-01 -2.86967039e-01 2.72404671e-01
-1.53152001e+00 3.72922063e-01 9.64485556e-02 2.19835296e-01
-9.23282743e-01 5.65194488e-01 6.70825124e-01 -2.62938946e-01
4.43828017e-01 -6.43013269e-02 -2.75409967e-01 -5.42717993e-01
-1.37008309e+00 4.24982220e-01 5.20797312e-01 8.90530288e-01
2.62856513e-01 -1.86248690e-01 -9.33746755e-01 6.15839303e-01
4.48674500e-01 2.16652751e-01 3.12168390e-01 -9.72163498e-01
8.18751633e-01 -1.24971688e-01 1.95058540e-01 -8.84846032e-01
-2.17630491e-01 2.73840241e-02 -4.27992821e-01 -6.90029979e-01
6.71693608e-02 -1.44455984e-01 -9.05537367e-01 1.72491407e+00
5.76117001e-02 1.09618437e+00 -1.11698046e-01 9.89177346e-01
1.15055406e+00 1.12529635e+00 6.48371518e-01 -3.80640686e-01
1.53816998e+00 -1.46096110e+00 -1.07825685e+00 -4.77698922e-01
-1.58701111e-02 -4.49797630e-01 3.55091929e-01 7.56930262e-02
-1.45303464e+00 -4.94626164e-01 -9.95209575e-01 1.89490199e-01
-4.99378704e-03 1.61883831e-02 5.34179628e-01 -1.06462426e-01
-1.02073777e+00 2.58506984e-01 -1.00546265e+00 -3.91987532e-01
4.73274618e-01 7.67190903e-02 -3.01593870e-01 -9.82921347e-02
-1.44937181e+00 8.55069935e-01 9.54182506e-01 -2.72750854e-02
-1.48877847e+00 -7.67129958e-01 -9.93138671e-01 1.10921346e-01
2.71315634e-01 -6.05354905e-01 1.74794662e+00 -7.62878597e-01
-1.07427168e+00 8.26757550e-01 -5.83704829e-01 -7.65154541e-01
4.73295778e-01 -4.64068085e-01 -8.31443191e-01 7.35085964e-01
1.96686268e-01 9.14342463e-01 3.25247169e-01 -8.25141966e-01
-9.35745478e-01 3.43584925e-01 4.87649255e-02 4.21242952e-01
-1.61904112e-01 8.61800730e-01 -1.16857004e+00 -7.95061409e-01
-4.91503596e-01 -6.83608651e-01 -1.81323722e-01 -2.07380980e-01
-3.04819504e-03 -3.57212305e-01 7.96369314e-01 -8.46678495e-01
1.35410249e+00 -2.03691244e+00 -3.34173888e-02 -2.76995212e-01
2.48629928e-01 8.46779421e-02 -1.60374612e-01 3.01117599e-01
-1.90183103e-01 -1.44803107e-01 -7.58903772e-02 -6.80986643e-01
-1.29057899e-01 5.48397787e-02 -4.40757632e-01 2.24783346e-01
4.28197354e-01 1.19676733e+00 -1.37665331e+00 -9.08518255e-01
3.90332758e-01 5.37298262e-01 -3.26358587e-01 4.42330509e-01
-4.79390353e-01 3.39230746e-01 -4.46691275e-01 6.53540850e-01
2.23910928e-01 -5.73464811e-01 3.02962717e-02 -4.66433913e-01
2.66889222e-02 4.48058039e-01 -8.11485112e-01 2.04674721e+00
5.43605201e-02 9.74011719e-01 -3.78577471e-01 -6.05428100e-01
2.58700490e-01 9.89694118e-01 1.23521388e+00 -4.96369809e-01
2.35784380e-03 -2.16527402e-01 -5.32993913e-01 -6.30034029e-01
5.47749639e-01 6.78893402e-02 -3.96729767e-01 3.23070318e-01
4.48502094e-01 4.55301821e-01 4.64472353e-01 4.37672228e-01
1.47337961e+00 5.46305776e-01 3.25472742e-01 2.08674699e-01
4.64087069e-01 -5.05453385e-02 7.59776533e-01 4.95269805e-01
-7.57713735e-01 1.04592633e+00 3.73705536e-01 -6.01886094e-01
-9.02490616e-01 -1.22854543e+00 4.47314829e-01 9.38528001e-01
1.58973709e-01 -7.76656389e-01 -4.32560384e-01 -8.67069185e-01
-6.14650667e-01 6.82085693e-01 -6.09453917e-01 -8.28035101e-02
-8.94454181e-01 -8.86340380e-01 8.99235070e-01 8.38286340e-01
7.63111651e-01 -1.21398103e+00 -4.61025894e-01 4.17694986e-01
-1.08252931e+00 -1.84113801e+00 -7.87824452e-01 -3.40864062e-01
-3.79865766e-01 -1.03734362e+00 -7.50343740e-01 -7.04678118e-01
-7.47378767e-02 1.73565432e-01 1.50672078e+00 -3.66342574e-01
-1.74245669e-03 5.54169118e-01 -3.15353274e-01 -2.23715410e-01
-4.00891192e-02 -2.33169705e-01 -1.73074216e-01 2.41472259e-01
5.87243259e-01 -5.57344258e-01 -7.30565012e-01 2.64369786e-01
-6.68735504e-01 2.71917373e-01 3.72486681e-01 1.58807307e-01
7.32657194e-01 -3.71634275e-01 5.61603487e-01 -2.21350621e-02
2.81380385e-01 -1.07011414e+00 -2.82372117e-01 3.45556736e-01
1.23924457e-01 -3.42314422e-01 1.58649996e-01 -4.65899706e-01
-1.28959179e+00 4.02053654e-01 -1.89748436e-01 -7.10089386e-01
-2.97181189e-01 5.18286645e-01 -1.19044401e-01 3.85438353e-01
2.47423187e-01 4.05327439e-01 -6.41249776e-01 -1.86143667e-01
3.74486476e-01 3.20298046e-01 1.16376936e+00 -6.37873650e-01
6.00151479e-01 6.43903792e-01 -3.03647429e-01 -5.57373762e-01
-1.31331694e+00 -8.18860352e-01 -4.48800594e-01 -8.43090773e-01
1.38074994e+00 -1.43876839e+00 -5.03768981e-01 4.41131324e-01
-1.61197042e+00 -3.33654165e-01 -2.10780308e-01 7.26888597e-01
-7.59541035e-01 4.01709110e-01 -8.18072140e-01 -3.34758818e-01
-3.53604406e-01 -8.27082753e-01 1.22934747e+00 3.32219243e-01
-2.99680322e-01 -7.94142902e-01 5.13813138e-01 5.25530040e-01
9.81426910e-02 5.46316803e-01 -1.14130653e-01 -7.91176975e-01
-7.52494752e-01 -1.90560564e-01 -3.70840788e-01 -8.11842978e-02
-2.42616653e-01 -2.90831421e-02 -9.01250601e-01 2.28613298e-02
-4.01047468e-01 -3.37823182e-01 9.47035015e-01 5.27525127e-01
1.30567944e+00 -1.88786071e-02 -4.45463806e-01 5.56939602e-01
1.13487005e+00 1.08940348e-01 8.77295494e-01 1.73841119e-01
6.20809674e-01 1.36045560e-01 7.82600999e-01 6.45179749e-01
8.84163678e-01 8.46844792e-01 4.35211867e-01 7.43078515e-02
-2.95960426e-01 -3.96831095e-01 5.04801989e-01 4.85776067e-01
-2.49083534e-01 -7.43250072e-01 -9.63194072e-01 1.01527154e+00
-2.28752804e+00 -1.54851854e+00 -1.74188569e-01 1.60459065e+00
6.16354704e-01 1.68023720e-01 3.51237148e-01 -3.10821563e-01
8.96155179e-01 6.91754401e-01 -1.99485824e-01 1.48104399e-01
-2.33690500e-01 -3.79524827e-02 4.99257892e-01 1.16495430e-01
-1.58289468e+00 8.69088829e-01 7.86164761e+00 5.04955649e-01
-5.70269227e-01 4.21209782e-01 3.09618026e-01 -5.44086576e-01
1.18332490e-01 -2.21423864e-01 -7.65453696e-01 5.90689182e-01
1.69952428e+00 -5.39270282e-01 1.71235763e-02 3.86669546e-01
5.53223491e-01 -9.09894630e-02 -1.16698492e+00 1.09182417e+00
4.20611054e-01 -1.65376759e+00 4.04344685e-02 -1.47205010e-01
5.74828088e-01 5.84629714e-01 -4.34961647e-01 3.94764304e-01
2.97472656e-01 -7.42735624e-01 1.00196767e+00 7.23532498e-01
4.62100267e-01 -4.96800572e-01 6.05492175e-01 6.30482286e-02
-1.75438011e+00 2.22430706e-01 4.64468122e-01 4.83661927e-02
1.23149157e+00 4.56601351e-01 -6.53140366e-01 6.90643251e-01
8.96457613e-01 1.19860637e+00 -4.00694847e-01 1.54132903e+00
-3.41888756e-01 8.56422961e-01 -4.01159406e-01 3.48710984e-01
3.67334515e-01 3.67695123e-01 8.09817731e-01 1.61346781e+00
2.49502465e-01 4.34488028e-01 2.90859640e-01 5.11898637e-01
-3.17997456e-01 -3.19460481e-01 -2.92168319e-01 -1.21224865e-01
3.30991685e-01 1.20671356e+00 -5.90084434e-01 -7.67568111e-01
-7.81469107e-01 1.19989598e+00 1.24792531e-01 2.78956383e-01
-1.76337171e+00 9.82086062e-02 6.51536167e-01 -1.71679661e-01
3.52220684e-01 -2.89363742e-01 2.91747004e-01 -1.40848708e+00
1.44061252e-01 -5.25578380e-01 1.01448560e+00 -1.16248250e+00
-8.50935459e-01 6.23938382e-01 5.25237322e-01 -1.13160634e+00
-2.11559817e-01 -3.55245620e-02 -8.15965056e-01 5.71848333e-01
-1.50510383e+00 -1.32531369e+00 -6.13732696e-01 6.33695364e-01
8.80551815e-01 1.05301246e-01 4.20708805e-01 9.63088155e-01
-7.11363375e-01 2.17452079e-01 -6.08337343e-01 3.59548151e-01
7.11722851e-01 -1.02176917e+00 7.05390215e-01 1.24943066e+00
3.41133982e-01 -4.58848067e-02 7.68583417e-01 -6.63066328e-01
-8.64893854e-01 -1.43736911e+00 1.19018829e+00 -6.34947300e-01
8.26882839e-01 -5.92427254e-02 -7.60039747e-01 1.05192459e+00
5.17485917e-01 2.60794938e-01 6.27182007e-01 -2.80096740e-01
-3.44429702e-01 1.48940161e-01 -7.99423754e-01 3.14185590e-01
1.05931294e+00 -6.14223719e-01 -9.99489009e-01 5.97041965e-01
8.47418964e-01 -4.74994779e-01 -9.88146007e-01 3.95589769e-01
4.23202097e-01 -4.73122448e-01 1.21436429e+00 -7.01278567e-01
6.25911415e-01 -4.92098093e-01 -8.63400400e-02 -7.89223850e-01
-2.75226533e-01 -7.67018497e-01 -8.25732470e-01 1.12974381e+00
3.11325103e-01 2.52416402e-01 6.80390835e-01 2.89922744e-01
-3.06632012e-01 -2.13328585e-01 -9.36487675e-01 -6.34112477e-01
-6.80528939e-01 -8.03336740e-01 1.56267375e-01 6.77702963e-01
-1.27087757e-01 4.31450069e-01 -8.08271945e-01 4.18884248e-01
6.71047628e-01 -2.90314496e-01 2.12593600e-01 -8.54603887e-01
-3.34989905e-01 -2.42250964e-01 -3.54462028e-01 -1.20110536e+00
2.79286325e-01 -5.88551223e-01 3.20374161e-01 -1.79806530e+00
5.19093871e-01 4.66982812e-01 -7.48474956e-01 5.67681670e-01
-2.02445596e-01 5.17875612e-01 8.79818425e-02 2.61426568e-01
-1.62527537e+00 5.54738402e-01 6.42909765e-01 -1.37015611e-01
5.51863993e-03 -3.86082679e-02 -2.28433371e-01 6.16400838e-01
4.98388141e-01 -5.22539139e-01 -2.57329166e-01 -4.63623524e-01
2.46942975e-03 4.22410429e-01 6.38936102e-01 -1.28019786e+00
2.85907626e-01 -4.07051176e-01 1.70243189e-01 -1.17463672e+00
6.97338343e-01 -5.40104866e-01 2.66895622e-01 1.56994641e-01
-2.78805345e-01 1.77334607e-01 5.20900905e-01 8.98252606e-01
-5.21918178e-01 2.56342620e-01 5.22142708e-01 -3.03511024e-02
-1.14260447e+00 7.81532884e-01 -6.57665133e-01 1.54150277e-01
1.37366104e+00 1.43406853e-01 -2.61502892e-01 -5.64196825e-01
-8.86545181e-01 7.42651105e-01 -1.07849486e-01 9.77790773e-01
4.90548670e-01 -1.68609846e+00 -9.36367393e-01 -5.65350890e-01
2.47135252e-01 -2.93398768e-01 4.18519527e-01 7.81984925e-01
-3.34191084e-01 4.78661060e-01 -2.09530309e-01 -6.12567782e-01
-1.30753648e+00 4.38576162e-01 4.59134489e-01 -4.22579646e-01
-6.64934814e-01 8.26178491e-01 6.47147000e-03 4.92584825e-01
2.51578510e-01 2.15138849e-02 -3.95252645e-01 1.36176586e-01
1.05578661e+00 2.18701825e-01 -1.83059171e-01 -9.15631592e-01
-7.04788208e-01 1.41662732e-01 5.66178076e-02 -5.08984625e-01
1.35220444e+00 -1.07648313e-01 1.73182592e-01 2.44598597e-01
1.16402364e+00 -5.13114989e-01 -1.40807128e+00 -2.37828851e-01
2.83478379e-01 -2.17778474e-01 1.43864617e-01 -8.05393875e-01
-8.72608125e-01 5.19736767e-01 2.32331693e-01 -1.21720701e-01
1.34111774e+00 3.52666229e-01 1.07225430e+00 5.10638319e-02
3.76839675e-02 -1.06975353e+00 2.57486731e-01 5.92215419e-01
8.70171726e-01 -1.19652033e+00 -1.51326939e-01 -3.46400321e-01
-7.33084083e-01 7.91470885e-01 7.02909470e-01 9.25662667e-02
4.92346525e-01 2.58330971e-01 -2.95749903e-01 -3.74797165e-01
-1.25768518e+00 -3.46880406e-01 5.85090518e-01 3.79940987e-01
2.45505586e-01 -2.83376426e-01 -2.67972231e-01 8.07622135e-01
4.64125305e-01 4.17699367e-01 4.82920915e-01 7.99055994e-01
-3.62514794e-01 -6.80628896e-01 -1.03785522e-01 5.40355034e-02
-6.59167588e-01 -8.42144340e-02 -1.19541079e-01 3.52063596e-01
-8.33398197e-04 9.77854609e-01 2.72802263e-01 -4.00667757e-01
4.99621361e-01 2.19752282e-01 3.83990854e-01 -5.19405007e-01
-5.37886977e-01 6.73702136e-02 4.86608148e-01 -1.09221625e+00
-8.70229125e-01 -9.99364555e-01 -1.24518585e+00 3.05931512e-02
3.70872915e-01 2.39343628e-01 4.64515954e-01 1.01132131e+00
3.46171618e-01 8.26379299e-01 3.10331285e-01 -1.05201638e+00
2.26911560e-01 -8.94386888e-01 1.57455385e-01 4.27242190e-01
3.14576477e-01 -5.22755086e-01 -2.22880036e-01 5.21564722e-01] | [10.449316024780273, 0.6853233575820923] |
438add8f-b9a5-4b37-a9bb-ef3269c41fed | softgym-benchmarking-deep-reinforcement | 2011.07215 | null | https://arxiv.org/abs/2011.07215v2 | https://arxiv.org/pdf/2011.07215v2.pdf | SoftGym: Benchmarking Deep Reinforcement Learning for Deformable Object Manipulation | Manipulating deformable objects has long been a challenge in robotics due to its high dimensional state representation and complex dynamics. Recent success in deep reinforcement learning provides a promising direction for learning to manipulate deformable objects with data driven methods. However, existing reinforcement learning benchmarks only cover tasks with direct state observability and simple low-dimensional dynamics or with relatively simple image-based environments, such as those with rigid objects. In this paper, we present SoftGym, a set of open-source simulated benchmarks for manipulating deformable objects, with a standard OpenAI Gym API and a Python interface for creating new environments. Our benchmark will enable reproducible research in this important area. Further, we evaluate a variety of algorithms on these tasks and highlight challenges for reinforcement learning algorithms, including dealing with a state representation that has a high intrinsic dimensionality and is partially observable. The experiments and analysis indicate the strengths and limitations of existing methods in the context of deformable object manipulation that can help point the way forward for future methods development. Code and videos of the learned policies can be found on our project website. | ['David Held', 'Jake Olkin', 'YuFei Wang', 'Xingyu Lin'] | 2020-11-14 | null | null | null | null | ['deformable-object-manipulation'] | ['robots'] | [-1.19939387e-01 -9.62196440e-02 -2.93544918e-01 -1.13002785e-01
-2.25786984e-01 -8.10050130e-01 5.14648914e-01 -4.52172965e-01
-3.74518782e-01 9.32822764e-01 3.35874036e-02 8.50533098e-02
-4.42997009e-01 -5.81938922e-01 -9.36888039e-01 -9.88021433e-01
-6.62305593e-01 6.81090415e-01 4.77671742e-01 -7.32492030e-01
1.06596760e-01 7.12970853e-01 -1.67480576e+00 -8.70636851e-02
4.18984473e-01 5.32862723e-01 4.36629981e-01 9.18154001e-01
4.80336875e-01 6.22040212e-01 -3.44320029e-01 2.01217964e-01
5.35072446e-01 -1.71351600e-02 -9.25951421e-01 -1.58515617e-01
5.65186322e-01 -7.44240105e-01 -7.35901177e-01 6.80484354e-01
6.72133744e-01 6.30126238e-01 3.95947218e-01 -1.42754424e+00
-7.55765676e-01 5.36787927e-01 2.03064576e-01 2.29717121e-01
4.18569118e-01 8.58772814e-01 6.32885396e-01 -3.10907871e-01
9.53463614e-01 1.54670513e+00 2.52172291e-01 1.14634871e+00
-1.08763528e+00 -4.18258131e-01 3.87876004e-01 1.08243212e-01
-6.10344291e-01 -1.59254774e-01 4.94529158e-01 -5.18775523e-01
1.20143747e+00 2.29374886e-01 1.02567685e+00 1.55153453e+00
4.26056296e-01 8.43817413e-01 1.20318770e+00 9.65067372e-02
4.14415956e-01 -3.94583672e-01 -3.15074205e-01 8.20741773e-01
3.32016677e-01 6.75083578e-01 -3.28669429e-01 -1.90640159e-03
1.17104137e+00 -1.03644498e-01 -8.96755829e-02 -1.09477055e+00
-1.53018188e+00 5.15673757e-01 5.74897408e-01 4.75611165e-03
-1.05007380e-01 8.48529458e-01 1.32488534e-01 3.22454214e-01
-1.61999002e-01 8.21241081e-01 -6.90845966e-01 -5.38823426e-01
-1.14983544e-01 1.08689106e+00 1.23294687e+00 1.12654269e+00
3.21276248e-01 3.94814879e-01 -2.94029921e-01 3.27761322e-01
3.32076013e-01 6.40675008e-01 1.54600307e-01 -1.46512854e+00
3.48505735e-01 2.61830449e-01 4.46500033e-01 -5.86323321e-01
-5.85056961e-01 1.20960481e-01 -2.36613542e-01 9.33524072e-01
4.85864490e-01 -4.23015743e-01 -9.88712311e-01 1.70896912e+00
7.13623106e-01 7.87456408e-02 1.20575823e-01 1.12608731e+00
6.78172588e-01 4.19157952e-01 4.10550274e-02 3.26705098e-01
8.62777591e-01 -9.30346727e-01 -6.54911160e-01 -4.65374738e-02
4.84621674e-01 -3.81376714e-01 1.15582824e+00 1.12694897e-01
-1.21629775e+00 -2.85567075e-01 -8.70407879e-01 -3.52513865e-02
-4.26623851e-01 -2.70418197e-01 9.02707934e-01 1.14881709e-01
-9.68748569e-01 1.15740192e+00 -1.62839830e+00 -4.75794107e-01
4.18231726e-01 6.80209041e-01 -2.62175322e-01 2.83793241e-01
-9.86447930e-01 1.41476750e+00 5.06167352e-01 8.49672034e-02
-1.47267807e+00 -6.68423474e-01 -7.80321479e-01 -2.82298476e-01
6.38780832e-01 -5.94710290e-01 1.54828811e+00 -2.56351173e-01
-2.05428982e+00 4.36733097e-01 6.85500324e-01 -2.20484674e-01
8.26160848e-01 -5.22077441e-01 1.08950987e-01 -4.88658659e-02
-2.46852815e-01 8.60173106e-01 7.54173577e-01 -1.20612669e+00
-3.41313630e-01 -1.11718424e-01 5.99197090e-01 4.45765942e-01
-1.49254903e-01 -2.58424938e-01 -1.91152804e-02 -7.54678369e-01
-3.40402335e-01 -1.61425042e+00 -4.63542521e-01 4.44946021e-01
-8.60446841e-02 -2.30390415e-01 1.21476638e+00 -1.90197662e-01
5.49993753e-01 -1.87318313e+00 9.51043725e-01 -2.66715735e-01
-1.54242562e-02 4.37771559e-01 -1.92850113e-01 6.98154390e-01
1.67982206e-01 -8.28831494e-02 -8.00688863e-02 2.51716256e-01
3.29882830e-01 6.70298398e-01 -2.45413572e-01 3.56588006e-01
3.96045506e-01 1.11003125e+00 -1.27812064e+00 -2.33816653e-01
4.17313039e-01 2.44631633e-01 -8.00431550e-01 3.31421882e-01
-7.29860604e-01 1.05448842e+00 -8.35122585e-01 6.63975775e-01
3.13613564e-01 1.18453585e-01 -8.30742903e-03 -8.56852233e-02
-2.02050522e-01 9.28142965e-02 -1.38630855e+00 1.90397060e+00
-4.65062112e-02 3.36808085e-01 2.09480643e-01 -9.18558240e-01
6.77096307e-01 1.03735574e-01 7.64647901e-01 -2.48738974e-01
1.93867594e-01 2.28522569e-02 3.92708659e-01 -1.22478938e+00
4.16177928e-01 1.01206809e-01 5.84046505e-02 2.62705535e-01
2.14477167e-01 -1.04732156e+00 3.83688688e-01 -1.35224864e-01
1.37973702e+00 1.00095046e+00 -1.77533790e-01 -4.99818057e-01
1.30456155e-02 1.78826302e-01 2.99380869e-01 8.44668150e-01
-4.46710765e-01 1.63465217e-01 2.47887582e-01 -6.68752730e-01
-1.04604805e+00 -1.21823740e+00 -1.59642085e-01 1.09331810e+00
3.92141402e-01 -6.18903115e-02 -4.93721455e-01 -2.79000998e-01
5.34977317e-01 5.05720675e-01 -6.51395440e-01 -3.98272336e-01
-9.68824327e-01 -4.82918888e-01 3.58712316e-01 8.07818472e-01
2.42002025e-01 -1.68182957e+00 -9.34966683e-01 3.07722628e-01
2.41452962e-01 -1.08637536e+00 1.05919272e-01 2.44347766e-01
-8.89263868e-01 -1.14065528e+00 -5.88167667e-01 -9.59407568e-01
4.24002081e-01 -8.09255056e-03 8.67684126e-01 6.02833647e-03
-6.13488793e-01 9.07425761e-01 -3.58937085e-01 -4.57736552e-01
-4.52502489e-01 8.30122828e-02 3.78045321e-01 -7.54735470e-01
-2.76902944e-01 -4.67025131e-01 -5.25544882e-01 4.21977639e-01
-1.06708109e+00 -1.13803208e-01 2.80560821e-01 6.32588863e-01
4.28614229e-01 -3.98698181e-01 2.72443026e-01 -3.29018146e-01
4.89356935e-01 -3.65907937e-01 -9.01188254e-01 2.85375724e-03
1.25069723e-01 3.35119933e-01 2.37786606e-01 -9.68472362e-01
-1.01587403e+00 2.99494803e-01 -1.17891066e-01 -2.88316250e-01
-1.31389052e-01 1.70353383e-01 2.38051727e-01 -3.91180426e-01
8.01806986e-01 -1.27446204e-01 2.22137272e-01 -2.39802271e-01
4.17024732e-01 1.38360322e-01 1.93012804e-01 -1.25944519e+00
9.64585245e-01 5.48851788e-01 3.21908593e-01 -5.20806432e-01
-6.19797051e-01 1.36031330e-01 -7.02818692e-01 -3.10085297e-01
8.58975172e-01 -6.49852037e-01 -1.24120104e+00 6.74849212e-01
-5.74926734e-01 -1.32423532e+00 -6.47551656e-01 6.62931561e-01
-1.28442514e+00 9.62617621e-03 -7.61847913e-01 -5.08878291e-01
5.54386675e-02 -1.41946352e+00 1.20580971e+00 3.50224316e-01
-1.45560920e-01 -8.81208897e-01 3.64448518e-01 4.49787714e-02
7.56152809e-01 6.38044119e-01 5.90382040e-01 -2.39298284e-01
-8.83219659e-01 7.61085823e-02 4.67239738e-01 1.96626455e-01
2.31563359e-01 2.83381522e-01 -4.99887049e-01 -8.28194499e-01
-2.89909631e-01 -9.50362742e-01 5.02287388e-01 3.20065916e-01
1.38173723e+00 -1.30401388e-01 -3.84710371e-01 4.20391649e-01
1.17537689e+00 1.50149405e-01 3.46216589e-01 3.94749045e-01
7.96021283e-01 4.30025160e-01 8.74725044e-01 5.13256073e-01
4.05951858e-01 7.55482733e-01 9.85263407e-01 1.43764779e-01
-5.19843176e-02 1.75752774e-01 4.56045926e-01 5.88965297e-01
-4.88922894e-01 -1.10994421e-01 -8.69201601e-01 3.31915647e-01
-1.97802627e+00 -9.87779379e-01 3.87173146e-02 1.94229269e+00
9.44987893e-01 -6.11446425e-02 1.72428638e-01 -2.45969594e-01
3.08504134e-01 3.81850312e-03 -1.03185344e+00 -3.50244075e-01
2.42707431e-01 3.27869177e-01 2.88067371e-01 4.11203235e-01
-1.27628100e+00 1.22957242e+00 6.63285732e+00 1.84333324e-01
-1.17231679e+00 -4.45546120e-01 -2.11732298e-01 -4.42709535e-01
1.97895125e-01 -1.13739997e-01 -9.11799133e-01 3.93731505e-01
5.37419260e-01 -1.32790371e-03 7.87527025e-01 9.91882741e-01
2.21555069e-01 -5.74937388e-02 -1.44400263e+00 4.26891744e-01
-3.04939270e-01 -1.18156195e+00 -1.31308526e-01 9.53189880e-02
7.96603441e-01 3.57527852e-01 1.07702643e-01 6.92039013e-01
9.31388557e-01 -8.25590014e-01 7.02807367e-01 6.50890708e-01
3.85930628e-01 -2.39150077e-01 7.54717663e-02 3.30692351e-01
-9.53681529e-01 -3.50662917e-01 -4.54457730e-01 -3.79458964e-01
-2.13253722e-02 -2.18861982e-01 -3.67118686e-01 1.37400836e-01
1.11085725e+00 1.00096977e+00 -1.56201422e-01 1.00559306e+00
-8.71878266e-02 3.31829906e-01 -5.41247547e-01 -4.94009584e-01
1.52194560e-01 -1.16436161e-01 9.09741342e-01 8.52141917e-01
1.34202227e-01 2.14499220e-01 5.71013570e-01 7.30018854e-01
1.89797312e-01 -4.96590793e-01 -9.26877499e-01 -3.82625386e-02
7.47908652e-02 1.25887835e+00 -6.32041276e-01 -2.95414090e-01
-4.72559854e-02 5.19521952e-01 5.22384644e-01 2.37917587e-01
-1.13552868e+00 -1.20672219e-01 1.13340771e+00 -8.89003575e-02
5.04096210e-01 -9.98234630e-01 4.11533356e-01 -1.46226275e+00
-2.61774622e-02 -9.55438972e-01 -2.10127085e-02 -6.29124224e-01
-1.03825891e+00 4.47304770e-02 4.43801343e-01 -1.39400423e+00
-2.31374145e-01 -8.37659001e-01 -1.56345084e-01 2.43876949e-01
-1.34440827e+00 -1.02028465e+00 -5.59783638e-01 7.01489627e-01
5.69096744e-01 2.44983584e-02 9.84121323e-01 -7.81884603e-03
-4.38159108e-01 7.09521621e-02 2.80924201e-01 9.08825696e-02
6.42772019e-01 -1.43339634e+00 3.91731143e-01 3.05618554e-01
-2.49176905e-01 3.88657182e-01 7.60001004e-01 -7.20287085e-01
-2.16458988e+00 -1.01883841e+00 -3.15556794e-01 -8.58491123e-01
9.46402431e-01 -4.44860458e-01 -9.30151880e-01 8.96866083e-01
7.42127225e-02 3.80927712e-01 7.13380426e-02 -2.92388350e-01
2.64593154e-01 1.19559437e-01 -1.08328533e+00 8.96783054e-01
1.51922429e+00 4.21430767e-02 -5.31735718e-01 6.25535727e-01
6.41203165e-01 -1.26608670e+00 -1.20708978e+00 7.45757341e-01
7.54861593e-01 -3.38637859e-01 9.90600169e-01 -1.22829521e+00
5.60530424e-01 -4.32677388e-01 -1.55632477e-02 -1.58386219e+00
-4.69795287e-01 -7.61813104e-01 -6.07160926e-01 7.94439375e-01
-1.12727180e-01 -3.85835975e-01 7.06391931e-01 6.34601831e-01
-2.04801142e-01 -8.02685976e-01 -7.69924700e-01 -1.06138933e+00
3.72771472e-01 -1.13678528e-02 3.01904887e-01 8.05759132e-01
-1.08935639e-01 -2.91618615e-01 -2.73684621e-01 1.82283714e-01
3.78736436e-01 1.68136030e-01 1.04121578e+00 -9.88364041e-01
-4.35228944e-01 -3.04698437e-01 -6.04417920e-01 -8.56073320e-01
3.40014219e-01 -8.57673883e-01 3.70313495e-01 -1.59783316e+00
-1.11919880e-01 -7.20807791e-01 3.24155390e-03 7.79124558e-01
-5.88154569e-02 -8.09828267e-02 4.02486533e-01 1.24939680e-01
-5.50574362e-01 8.19240868e-01 2.06929898e+00 -2.00637981e-01
-2.55816102e-01 1.13119073e-02 -1.35560468e-01 6.85287058e-01
9.62027669e-01 -4.53134477e-01 -3.13774467e-01 -7.59106815e-01
8.26867446e-02 -4.41205427e-02 5.68409801e-01 -1.27256489e+00
-8.58458430e-02 -8.30716014e-01 2.92060435e-01 -3.62037383e-02
6.36905134e-01 -8.85422766e-01 -4.01998349e-02 1.03462958e+00
-5.68316281e-01 8.16215426e-02 4.08603668e-01 6.34017527e-01
3.42486352e-01 6.37432486e-02 8.00655723e-01 -2.19527826e-01
-7.85904467e-01 5.00233352e-01 -5.90564370e-01 2.58580834e-01
1.52199709e+00 1.37635782e-01 -3.63083094e-01 -6.96896613e-02
-1.15888798e+00 5.04394233e-01 5.81766963e-01 9.18078125e-01
3.29424113e-01 -1.35293329e+00 -5.77510238e-01 -4.33625653e-02
-9.22704041e-02 2.33064398e-01 -1.45087525e-01 4.03766274e-01
-7.38072455e-01 -1.39042273e-01 -9.30207133e-01 -6.42403185e-01
-1.00758147e+00 4.84744400e-01 4.46520597e-01 1.44284414e-02
-7.78540730e-01 5.60563028e-01 -1.66518807e-01 -7.76430190e-01
3.46836776e-01 -9.21386302e-01 -1.98782329e-02 -4.94385123e-01
4.05543298e-02 4.69758809e-01 -1.15628786e-01 -2.90737689e-01
-3.02435756e-01 4.98255998e-01 9.13712680e-02 1.64308831e-01
1.72748542e+00 4.63645011e-01 4.56096865e-02 4.54929799e-01
5.82047522e-01 -5.86519003e-01 -1.90037966e+00 3.02985776e-02
-2.45056897e-01 -3.09115142e-01 -3.45332503e-01 -6.49976671e-01
-9.55563426e-01 6.40458226e-01 6.71516418e-01 1.18315458e-01
4.13114130e-01 2.75566746e-02 4.90459025e-01 1.18710351e+00
6.38466358e-01 -1.31473851e+00 6.00617707e-01 9.22420382e-01
1.41530466e+00 -1.34166956e+00 1.10974886e-01 -1.44578442e-01
-5.27866364e-01 1.20333922e+00 1.21892107e+00 -7.70918012e-01
8.46985400e-01 7.32472181e-01 -1.58576652e-01 -1.27773941e-01
-7.32960820e-01 -2.72026062e-01 -1.52361058e-02 8.00706387e-01
1.64173335e-01 1.06324799e-01 -5.91435246e-02 -2.78609693e-01
-2.75663704e-01 7.86481500e-02 6.37507856e-01 1.65307283e+00
-3.92532349e-01 -1.24021542e+00 -1.56435564e-01 3.55622798e-01
-1.60903797e-01 5.69286108e-01 -2.33868301e-01 1.02896738e+00
-2.11040780e-01 5.50449848e-01 -1.09648526e-01 2.45927926e-02
5.19276857e-01 -3.53155285e-01 1.14714599e+00 -7.66150534e-01
-5.37318468e-01 -3.89913738e-01 -1.65285498e-01 -1.03847873e+00
-5.58377862e-01 -7.64947414e-01 -1.44026160e+00 -1.56674206e-01
-1.65943921e-01 -4.38493192e-01 7.74299860e-01 8.84482622e-01
3.63100886e-01 6.92806602e-01 1.74773395e-01 -1.73100376e+00
-1.09989405e+00 -7.39595711e-01 -2.33786076e-01 5.45304060e-01
5.62376440e-01 -1.24666905e+00 2.87745316e-02 8.60346183e-02] | [4.739785671234131, 0.6384881138801575] |
f53c3588-e39a-4a72-ab07-2dda73b0d7f6 | doric-domain-robust-fine-tuning-for-open | 2303.09827 | null | https://arxiv.org/abs/2303.09827v1 | https://arxiv.org/pdf/2303.09827v1.pdf | DORIC : Domain Robust Fine-Tuning for Open Intent Clustering through Dependency Parsing | We present our work on Track 2 in the Dialog System Technology Challenges 11 (DSTC11). DSTC11-Track2 aims to provide a benchmark for zero-shot, cross-domain, intent-set induction. In the absence of in-domain training dataset, robust utterance representation that can be used across domains is necessary to induce users' intentions. To achieve this, we leveraged a multi-domain dialogue dataset to fine-tune the language model and proposed extracting Verb-Object pairs to remove the artifacts of unnecessary information. Furthermore, we devised the method that generates each cluster's name for the explainability of clustered results. Our approach achieved 3rd place in the precision score and showed superior accuracy and normalized mutual information (NMI) score than the baseline model on various domain datasets. | ['Gary Geunbae Lee', 'Yunsu Kim', 'Seungyeon Seo', 'Jihyun Lee'] | 2023-03-17 | null | null | null | null | ['dependency-parsing'] | ['natural-language-processing'] | [-1.50614213e-02 4.29202408e-01 -1.70412734e-01 -7.53015995e-01
-9.15978014e-01 -7.60026872e-01 1.04260409e+00 -1.45884648e-01
-1.55777052e-01 7.76088357e-01 8.13740671e-01 -1.38086393e-01
-8.04480538e-02 -1.57325149e-01 -8.88848454e-02 -1.63180344e-02
2.43997604e-01 1.04980159e+00 2.77701795e-01 -7.03593671e-01
2.03286633e-01 -1.33254603e-01 -1.18184221e+00 7.69014060e-01
9.25802469e-01 6.67463303e-01 1.85157746e-01 5.15628576e-01
-3.51109028e-01 9.34121788e-01 -7.90815592e-01 -5.50103247e-01
5.94532266e-02 -6.69903040e-01 -1.23171437e+00 2.26526424e-01
-7.31560886e-02 -3.45421523e-01 -6.19643889e-02 6.03862703e-01
4.90486950e-01 4.45160329e-01 1.00248611e+00 -1.39767516e+00
-5.38046896e-01 8.33935380e-01 -5.95588684e-02 -4.34132129e-01
7.99856961e-01 8.18448365e-02 1.08132911e+00 -1.04734302e+00
1.02949941e+00 1.49858093e+00 4.09119159e-01 9.44306672e-01
-1.33586013e+00 -4.02557939e-01 -1.25130668e-01 1.04265483e-02
-1.05633152e+00 -7.39987195e-01 6.18147194e-01 -5.67018509e-01
1.11704755e+00 1.33290797e-01 -1.08398028e-01 1.62229598e+00
-4.25698280e-01 9.08983290e-01 9.41099823e-01 -7.61380315e-01
4.20524180e-01 7.18497574e-01 6.05799377e-01 2.47742072e-01
-2.54822046e-01 -2.44830564e-01 -4.04806912e-01 -3.93117547e-01
3.76432925e-01 -4.26128149e-01 -7.54461884e-02 -3.05562079e-01
-9.49677348e-01 1.19277346e+00 -1.27681255e-01 5.65989196e-01
-1.81038916e-01 -6.46822333e-01 5.58099091e-01 3.89964342e-01
5.29323339e-01 7.76499987e-01 -8.31944585e-01 -4.31315094e-01
-4.17766064e-01 2.67182231e-01 1.42070973e+00 1.21129930e+00
4.16655511e-01 -5.08559167e-01 -6.26061618e-01 1.30388999e+00
2.24097967e-01 5.79079241e-03 6.20658696e-01 -1.19734406e+00
5.96793354e-01 8.20257187e-01 4.51581776e-01 -4.43945885e-01
-4.57430094e-01 3.64091039e-01 -5.24280131e-01 -2.26948828e-01
4.48510259e-01 -5.90195835e-01 -5.82448959e-01 1.79082048e+00
1.94315925e-01 -3.32570434e-01 6.33745372e-01 7.55898476e-01
1.08410084e+00 3.62214625e-01 1.65378779e-01 -2.89653897e-01
1.43938863e+00 -8.48753750e-01 -9.61964250e-01 -2.55107135e-01
9.26360905e-01 -7.26315916e-01 1.36249936e+00 7.53140077e-02
-7.20156848e-01 -5.75762808e-01 -8.17660213e-01 -3.92976217e-02
-4.10642445e-01 3.71231139e-02 3.51032883e-01 6.90408468e-01
-8.12892973e-01 4.40280735e-02 -2.45975688e-01 -7.07064569e-01
-2.54015058e-01 3.15221995e-01 -2.16710478e-01 2.01787546e-01
-1.59648621e+00 1.12915778e+00 5.75308502e-01 -8.60673904e-01
-6.38860703e-01 -6.89011872e-01 -8.32220018e-01 -1.18442208e-01
2.77707547e-01 -3.74991417e-01 1.66742277e+00 -7.18398929e-01
-1.90400302e+00 8.84074450e-01 -2.29072452e-01 -5.77171028e-01
2.90115327e-01 -2.30844408e-01 -4.35989201e-01 -1.54847726e-01
2.64285713e-01 8.15782309e-01 2.99598873e-01 -1.39061630e+00
-5.62594771e-01 -2.14160115e-01 5.26576936e-02 4.49959368e-01
-3.57212782e-01 1.41963512e-01 -6.72739267e-01 -1.36787385e-01
-2.12325633e-01 -1.05179214e+00 -2.16878697e-01 -7.18475521e-01
-4.40317452e-01 -7.27376699e-01 9.29996908e-01 -6.42354727e-01
1.24555135e+00 -2.14835572e+00 -3.78879206e-03 -2.03602150e-01
6.81725740e-02 2.64173716e-01 -6.52503106e-04 7.34734595e-01
2.31760710e-01 6.74928352e-02 -6.13605715e-02 -6.12921119e-01
3.80004615e-01 1.76039100e-01 -2.96991855e-01 -1.65398031e-01
9.75485817e-02 5.85770071e-01 -7.74285793e-01 -5.98130822e-01
4.60677385e-01 3.91005129e-02 -6.62683964e-01 5.98233521e-01
-6.49259329e-01 5.10513425e-01 -3.71122867e-01 1.43917769e-01
4.72718894e-01 -2.83185780e-01 5.71119964e-01 -3.67796756e-02
1.40714170e-02 5.99687815e-01 -8.04958463e-01 2.00434113e+00
-7.21846223e-01 3.91406298e-01 5.09387478e-02 -6.43553972e-01
1.10696340e+00 7.05391228e-01 6.04806125e-01 -6.26642585e-01
1.13009781e-01 -9.22552869e-02 -6.03960939e-02 -6.93706870e-01
6.30509734e-01 -2.45760027e-02 -8.34108889e-01 5.37856936e-01
4.70253378e-01 -1.64700553e-01 7.69077276e-04 5.03706932e-01
9.91243541e-01 -2.11930260e-01 4.40256238e-01 -2.51388103e-01
4.91446406e-01 3.38036150e-01 2.44350910e-01 7.35879123e-01
-2.67830700e-01 6.34306490e-01 3.36464435e-01 -1.28655583e-02
-8.25249493e-01 -8.89731824e-01 -7.96475336e-02 1.52128756e+00
8.25272426e-02 -3.22963387e-01 -7.19965339e-01 -1.14154744e+00
-1.74785376e-01 1.28864837e+00 -3.75299513e-01 -1.04372188e-01
-1.57371238e-01 -3.75533789e-01 8.21839511e-01 1.85981572e-01
6.91956162e-01 -8.73144627e-01 -6.16387501e-02 2.95436233e-01
-7.77009904e-01 -1.49223387e+00 -5.80813289e-01 2.59944439e-01
-2.63017833e-01 -9.31186795e-01 -2.67956287e-01 -9.03881788e-01
8.65862072e-02 6.71865093e-03 1.35995650e+00 -4.35042262e-01
1.57336101e-01 4.31872040e-01 -6.81326151e-01 -3.30913037e-01
-7.40655005e-01 1.30166844e-01 2.01824918e-01 -2.00856298e-01
1.01531303e+00 -1.33607343e-01 -1.75622866e-01 5.09909809e-01
-4.66877967e-01 -4.36484590e-02 1.42853573e-01 9.43673193e-01
-3.09173644e-01 -2.16433391e-01 1.08034182e+00 -9.43417788e-01
1.30234504e+00 -5.62048554e-01 -1.96505338e-01 4.63875234e-01
-3.61903280e-01 1.13950990e-01 4.73205954e-01 -3.46938074e-01
-1.62821305e+00 2.15687275e-01 -1.58831298e-01 8.43833759e-02
-4.68810201e-01 2.34986812e-01 -3.80533487e-01 4.79648650e-01
8.41341913e-01 -7.19782189e-02 -9.62773710e-03 -5.44527948e-01
7.03097284e-01 1.42791390e+00 3.94846857e-01 -6.15284979e-01
1.81337908e-01 -1.25119435e-02 -9.27128851e-01 -8.80067706e-01
-8.80052269e-01 -9.40677881e-01 -9.73083973e-01 -2.23313123e-02
1.14890945e+00 -9.44417238e-01 -5.91368556e-01 -7.94701427e-02
-1.49699366e+00 -3.93068045e-01 -2.84759589e-02 2.86373168e-01
-4.25270706e-01 2.63373137e-01 -4.63603348e-01 -1.24196851e+00
-4.38641995e-01 -9.26956952e-01 9.50098872e-01 1.04291677e-01
-9.25469875e-01 -1.29183805e+00 2.10159674e-01 6.05332673e-01
2.51009434e-01 -1.36780828e-01 8.28483880e-01 -1.68092370e+00
2.27834284e-01 1.01326838e-01 -1.06747389e-01 3.06792319e-01
3.37323517e-01 -4.79300082e-01 -1.16204751e+00 6.35968819e-02
2.95518607e-01 -7.51204431e-01 2.54117012e-01 1.36220455e-01
5.19500196e-01 -4.49101895e-01 -3.48366022e-01 -3.09689194e-01
8.81982505e-01 5.54360092e-01 3.70229214e-01 1.28084302e-01
2.95798182e-01 9.23617542e-01 8.21411312e-01 7.08343863e-01
6.87954962e-01 1.15068054e+00 -3.12551290e-01 6.19407482e-02
-1.20894514e-01 -2.34163001e-01 3.43501657e-01 6.35654449e-01
5.06531775e-01 -3.64132762e-01 -9.11373556e-01 7.39041269e-01
-1.97133780e+00 -8.78547192e-01 -3.05610523e-02 2.03606606e+00
1.12333775e+00 1.61518008e-01 4.50484335e-01 -3.41194689e-01
7.30009496e-01 -2.45108515e-01 -1.82205677e-01 -5.40758073e-01
4.68522310e-02 3.73473503e-02 -3.20403662e-04 6.53412700e-01
-1.04944527e+00 1.31403399e+00 6.41485071e+00 5.94270885e-01
-4.73745167e-01 3.50226492e-01 5.28332174e-01 3.53238672e-01
-1.76149741e-01 -4.00077105e-02 -9.86850500e-01 4.69906509e-01
1.26645482e+00 -3.39226037e-01 1.21326752e-01 1.01122355e+00
2.40203381e-01 -6.30778000e-02 -1.26118720e+00 4.32614595e-01
-1.42584145e-02 -9.81997848e-01 -1.64695561e-01 -5.69524011e-03
7.08680511e-01 -9.09632593e-02 -2.49468341e-01 9.40257668e-01
1.04822779e+00 -8.35268199e-01 -3.07577476e-02 7.42281079e-02
6.38663232e-01 -5.25136650e-01 7.21133113e-01 6.23127103e-01
-7.36181021e-01 9.80880558e-02 -2.08718657e-01 1.09086717e-02
2.31651068e-01 1.59753814e-01 -1.86129761e+00 5.05274057e-01
2.81580359e-01 3.68655354e-01 -1.53254360e-01 3.86886537e-01
6.29309565e-02 4.66883719e-01 -9.20580849e-02 -2.50436813e-01
2.01481387e-01 -1.62035748e-01 3.92235667e-01 1.52279341e+00
-7.37894699e-02 2.69044042e-01 5.38768172e-01 6.47564709e-01
-2.07664907e-01 4.07723188e-02 -8.18246722e-01 1.04687624e-01
9.49401200e-01 1.02796292e+00 -2.62271941e-01 -5.66607535e-01
-4.61037815e-01 1.17506778e+00 1.05359867e-01 1.30245522e-01
-7.46369243e-01 -3.00860286e-01 8.41731906e-01 -3.28376710e-01
6.74296990e-02 -1.19009629e-01 -5.92941105e-01 -1.08289635e+00
-3.59995395e-01 -1.10772979e+00 5.49798071e-01 -4.21510488e-01
-1.60324740e+00 7.10904360e-01 2.21992359e-01 -1.10288906e+00
-9.03421283e-01 -4.28098500e-01 -4.75986063e-01 7.21230209e-01
-8.34368765e-01 -1.13931751e+00 -1.62762806e-01 7.42516339e-01
1.21744108e+00 -3.84578943e-01 1.30359197e+00 1.04872219e-01
-4.04945940e-01 6.78150773e-01 1.71927884e-02 3.22762161e-01
1.24014509e+00 -1.32648170e+00 3.99594545e-01 4.02940065e-01
-1.11820623e-01 8.35497439e-01 9.85426366e-01 -8.28901291e-01
-1.05797970e+00 -9.57060754e-01 1.04446089e+00 -9.61447537e-01
7.30136156e-01 -5.80574036e-01 -9.13393736e-01 6.89664781e-01
6.73563957e-01 -7.21391261e-01 9.47131217e-01 4.86615777e-01
-2.42757156e-01 3.90652090e-01 -1.34383142e+00 4.68176097e-01
7.45455205e-01 -6.26050234e-01 -1.15944302e+00 5.45435607e-01
1.12759113e+00 -3.14931542e-01 -1.11509180e+00 3.04405481e-01
3.19075465e-01 -7.59410918e-01 9.24577475e-01 -6.05337918e-01
3.04593563e-01 1.14423349e-01 -4.02443975e-01 -1.23561025e+00
-9.46485996e-02 -6.24046028e-01 8.10028166e-02 1.68273139e+00
8.28836620e-01 -7.78410062e-02 5.50963104e-01 1.24701536e+00
-2.67208159e-01 -2.35032011e-02 -7.47407854e-01 -6.99377716e-01
6.41143844e-02 -3.87107342e-01 4.01420534e-01 1.04389369e+00
7.82746375e-01 1.07900214e+00 -6.06939137e-01 -5.63181080e-02
4.26508486e-01 -4.81753163e-02 9.61207747e-01 -1.33025503e+00
-1.71472743e-01 -2.09283382e-01 2.09571809e-01 -1.20275092e+00
4.06765789e-01 -5.77773571e-01 3.28236699e-01 -1.51385641e+00
9.17191654e-02 -4.01988655e-01 1.97538629e-01 3.94793034e-01
-2.09311292e-01 -2.18778968e-01 1.15794227e-01 2.20575094e-01
-9.33372200e-01 6.77024066e-01 7.45291531e-01 4.55968045e-02
-7.59985626e-01 1.51753247e-01 -9.35907423e-01 4.47584659e-01
8.36683691e-01 -2.70475209e-01 -6.17166877e-01 -9.60031003e-02
-7.27941275e-01 2.65198082e-01 -1.09312743e-01 -8.12479854e-01
3.51418912e-01 -1.68920860e-01 -4.60889973e-02 -5.70055187e-01
7.58911669e-01 -7.91931689e-01 -1.63327515e-01 1.93317726e-01
-8.15833271e-01 -4.20830518e-01 1.40046790e-01 3.43116820e-01
-1.37941673e-01 -3.45121980e-01 6.41162455e-01 -1.39361814e-01
-1.00533950e+00 -4.02321905e-01 -5.06788909e-01 3.04670125e-01
9.81374741e-01 1.64686680e-01 -3.84045631e-01 -6.07962549e-01
-6.19040430e-01 4.30389553e-01 3.89001638e-01 8.52415681e-01
3.53799313e-01 -1.18211842e+00 -7.17723548e-01 -1.01172425e-01
4.97127682e-01 -4.59623277e-01 3.90361138e-02 4.45128173e-01
2.58308589e-01 7.53110707e-01 -3.85571383e-02 -5.49526811e-01
-1.30516160e+00 2.94337004e-01 8.15649256e-02 -5.81254482e-01
-2.75867403e-01 6.78179920e-01 8.64963904e-02 -1.00005233e+00
4.25380945e-01 1.62943587e-01 -4.99574155e-01 1.34224758e-01
5.31505048e-01 4.74438258e-02 1.42830284e-02 -4.90018964e-01
-4.82154936e-01 -1.48001313e-01 -2.76161909e-01 -7.07205892e-01
1.10771215e+00 -5.37612796e-01 2.86843061e-01 6.68895960e-01
1.17789638e+00 -3.16632152e-01 -1.00745499e+00 -4.20552582e-01
7.39672959e-01 -7.67770633e-02 -2.78053552e-01 -1.16320050e+00
-9.44534838e-02 4.88815308e-01 4.69760388e-01 5.49482107e-01
7.18084812e-01 2.37600416e-01 6.90835416e-01 5.11029661e-01
4.05309051e-01 -1.36789262e+00 3.50244582e-01 1.12981403e+00
8.17975640e-01 -1.80984950e+00 -4.33554471e-01 -3.70528281e-01
-1.61110175e+00 8.57563555e-01 9.72684145e-01 4.11909223e-01
4.21848148e-01 1.46187693e-01 2.65050113e-01 -2.35311478e-01
-9.41492558e-01 -3.79549414e-01 3.56950045e-01 8.07117879e-01
9.20556724e-01 6.33769780e-02 -3.50714922e-01 1.12714684e+00
-2.31094077e-01 -1.51305258e-01 2.73162514e-01 7.22654819e-01
-5.48624218e-01 -1.24045253e+00 9.94570740e-03 8.63819122e-02
-1.69402733e-01 -1.10158712e-01 -1.06633019e+00 8.36272657e-01
-2.49036089e-01 1.61999965e+00 -1.94002196e-01 -8.95024240e-01
4.09343868e-01 5.23726046e-01 -1.09808356e-01 -8.83814514e-01
-7.48285770e-01 -5.25688492e-02 9.22930777e-01 -2.33547211e-01
-4.17624474e-01 -4.94155079e-01 -1.17046881e+00 2.14235783e-02
-3.26241165e-01 5.48582375e-01 6.56935692e-01 1.22128069e+00
6.36162341e-01 1.54985756e-01 8.13620567e-01 -4.75586921e-01
-5.28867185e-01 -1.49887347e+00 -1.04780264e-01 7.63120651e-01
-6.01348765e-02 -6.03300989e-01 -2.60563284e-01 1.97918639e-01] | [12.657414436340332, 7.71690034866333] |
0019d61c-4a5c-498c-9d9e-751b05492998 | from-learning-to-relearning-a-framework-for | 2101.02647 | null | https://arxiv.org/abs/2101.02647v2 | https://arxiv.org/pdf/2101.02647v2.pdf | From Learning to Relearning: A Framework for Diminishing Bias in Social Robot Navigation | The exponentially increasing advances in robotics and machine learning are facilitating the transition of robots from being confined to controlled industrial spaces to performing novel everyday tasks in domestic and urban environments. In order to make the presence of robots safe as well as comfortable for humans, and to facilitate their acceptance in public environments, they are often equipped with social abilities for navigation and interaction. Socially compliant robot navigation is increasingly being learned from human observations or demonstrations. We argue that these techniques that typically aim to mimic human behavior do not guarantee fair behavior. As a consequence, social navigation models can replicate, promote, and amplify societal unfairness such as discrimination and segregation. In this work, we investigate a framework for diminishing bias in social robot navigation models so that robots are equipped with the capability to plan as well as adapt their paths based on both physical and social demands. Our proposed framework consists of two components: \textit{learning} which incorporates social context into the learning process to account for safety and comfort, and \textit{relearning} to detect and correct potentially harmful outcomes before the onset. We provide both technological and societal analysis using three diverse case studies in different social scenarios of interaction. Moreover, we present ethical implications of deploying robots in social environments and propose potential solutions. Through this study, we highlight the importance and advocate for fairness in human-robot interactions in order to promote more equitable social relationships, roles, and dynamics and consequently positively influence our society. | ['Abhinav Valada', 'Laura Londoño', 'Juana Valeria Hurtado'] | 2021-01-07 | null | null | null | null | ['social-navigation'] | ['robots'] | [ 2.31938124e-01 7.87870467e-01 -2.64194980e-02 -3.82761508e-01
2.34766960e-01 -3.12261015e-01 4.76028085e-01 1.36812210e-01
-6.66836977e-01 9.77938056e-01 2.96864323e-02 -2.48976544e-01
-5.11619091e-01 -6.59675479e-01 -5.94694436e-01 -5.62747240e-01
2.79567838e-02 2.06042528e-01 -1.82228923e-01 -4.53628898e-01
1.35474578e-01 5.70939124e-01 -1.80231190e+00 -4.54796284e-01
1.23012602e+00 6.32472456e-01 4.73499924e-01 2.56269693e-01
5.08126616e-01 6.70542419e-01 -2.54723549e-01 -4.03194904e-01
2.27532297e-01 -2.47436419e-01 -6.06562138e-01 -1.19958036e-01
-4.79280084e-01 -4.95758981e-01 7.54898414e-02 1.05072939e+00
5.37290871e-01 2.26916060e-01 9.35617268e-01 -1.73618650e+00
-6.05652034e-01 7.60623634e-01 -2.09940270e-01 -6.31606281e-01
6.52563095e-01 3.59238416e-01 7.94527352e-01 -1.88205630e-01
5.75750351e-01 1.36355889e+00 4.28904861e-01 6.69985950e-01
-1.06499958e+00 -8.97354186e-01 1.34292066e-01 1.43174797e-01
-9.76710498e-01 -4.08589780e-01 6.81821227e-01 -4.75900024e-01
4.76916224e-01 1.62521720e-01 9.54048455e-01 1.34385991e+00
5.02760410e-01 3.53188634e-01 8.13266933e-01 -3.00591230e-01
4.49841589e-01 5.21160245e-01 -2.59927064e-01 2.38118917e-01
8.22830975e-01 2.11276069e-01 -9.91955101e-02 -1.12164117e-01
5.04394531e-01 -4.83388267e-02 9.83401164e-02 -8.53601515e-01
-1.06918013e+00 5.92837811e-01 5.57581544e-01 1.78594694e-01
-9.53078866e-01 1.64925635e-01 3.59968126e-01 1.80990010e-01
-6.70772120e-02 6.96627915e-01 -2.21334562e-01 -2.79116541e-01
2.81463325e-01 3.83543819e-01 6.63799286e-01 8.30467463e-01
5.88577986e-01 -3.80416423e-01 1.50580630e-01 6.53187931e-01
5.24435401e-01 4.77421790e-01 1.79882482e-01 -1.75146818e+00
9.87279341e-02 6.60619497e-01 5.75638831e-01 -1.42507482e+00
-7.09091663e-01 -1.99307770e-01 -5.43616056e-01 5.32365084e-01
2.84278959e-01 -2.04964921e-01 1.55251130e-01 1.93354416e+00
6.05255187e-01 -3.46864283e-01 2.89745301e-01 7.68616140e-01
4.63368408e-02 3.02392006e-01 4.87718076e-01 -1.38085410e-01
1.07685649e+00 -4.45726871e-01 -7.01991916e-01 -3.30834031e-01
9.68242645e-01 -3.46480846e-01 1.31833124e+00 3.79489124e-01
-5.92015386e-01 -1.31525680e-01 -9.23853874e-01 2.74263650e-01
-1.46298975e-01 -6.13979697e-02 3.61980528e-01 8.37186337e-01
-7.59003103e-01 7.92480886e-01 -3.92685354e-01 -8.66735160e-01
3.24257553e-01 3.57134312e-01 -3.34193140e-01 9.20759216e-02
-1.08646870e+00 1.13904595e+00 4.66540568e-02 2.69833028e-01
-4.93008822e-01 -3.58070545e-02 -8.11956644e-01 -2.26548433e-01
2.63729870e-01 -5.93105018e-01 1.08963716e+00 -9.61151659e-01
-1.66885006e+00 6.25399828e-01 5.92562079e-01 -5.19970834e-01
7.48456299e-01 -3.19704354e-01 -4.22399566e-02 -1.55107439e-01
4.29224342e-01 8.44506681e-01 3.66076022e-01 -1.61929142e+00
-5.93304694e-01 -2.69423634e-01 3.10391247e-01 5.65415800e-01
-4.79072481e-01 -3.21107954e-02 7.81267703e-01 -1.30682737e-01
-2.27171242e-01 -1.42168546e+00 -6.18505597e-01 9.54559669e-02
-2.85053343e-01 -9.87313837e-02 6.61097109e-01 -3.92041683e-01
5.63882351e-01 -2.20389462e+00 7.06958994e-02 1.85126856e-01
6.24322854e-02 -7.17224702e-02 -2.45382842e-02 2.23988160e-01
3.19091022e-01 -2.65631899e-02 4.70986366e-02 -3.02657515e-01
4.51170415e-01 2.84236014e-01 -4.02129954e-03 6.18796766e-01
-5.74773587e-02 4.55592245e-01 -1.10574818e+00 -3.69665891e-01
4.14991528e-01 2.90886015e-01 -9.31642473e-01 2.14633197e-01
2.34402761e-01 9.39749837e-01 -4.84556228e-01 2.74803668e-01
3.76244158e-01 3.38806063e-01 4.29513603e-01 4.20979410e-01
-1.57066345e-01 1.71588153e-01 -7.87023246e-01 1.03170896e+00
-6.24164283e-01 2.99098730e-01 4.74451005e-01 -7.49319732e-01
8.97481740e-01 7.94651881e-02 4.81723398e-01 -7.77004004e-01
6.84024930e-01 3.76712114e-01 1.66772738e-01 -6.66180789e-01
7.60632813e-01 -8.70135985e-03 -5.32438338e-01 5.40460587e-01
-5.11169016e-01 -5.08645773e-01 -3.58152598e-01 1.26040364e-02
7.74956942e-01 3.54338884e-01 3.88483793e-01 -2.72826225e-01
4.33284730e-01 -2.26664051e-01 6.12309277e-01 5.43205261e-01
-8.62576485e-01 -2.76452124e-01 5.08842707e-01 3.10077220e-02
-8.75723958e-01 -9.06699300e-01 2.98906982e-01 1.19084418e+00
6.58240318e-01 3.60641211e-01 -7.64715731e-01 -5.07118285e-01
-4.19536792e-02 1.33841491e+00 -3.85163695e-01 -6.84747696e-01
-2.91865677e-01 -2.12356523e-01 2.77548820e-01 1.33073023e-02
1.94089681e-01 -1.29750884e+00 -1.23093605e+00 4.09188168e-03
-7.19877779e-02 -9.56588984e-01 -7.22615933e-03 8.57846513e-02
-3.75510097e-01 -9.83382583e-01 -3.82085264e-01 -6.88925445e-01
7.54501283e-01 3.41187179e-01 2.08241701e-01 -9.71636549e-02
3.46990898e-02 6.52329922e-01 -5.71680486e-01 -6.04366064e-01
-8.31128716e-01 -4.23563980e-02 6.27948046e-01 -1.99935138e-01
-1.85644627e-02 -9.09559429e-01 -6.26241207e-01 6.58886909e-01
-4.62661803e-01 1.59511995e-02 5.44820189e-01 4.79839444e-01
-2.51625150e-01 -1.37062860e-03 9.01290536e-01 -2.76170969e-01
8.52439821e-01 -8.30060720e-01 -2.13993579e-01 4.31600213e-02
-6.70313478e-01 -3.72997850e-01 4.39641178e-01 -5.30570030e-01
-1.22845709e+00 3.77663993e-04 2.82339245e-01 2.67969370e-01
-4.64921355e-01 -2.68766023e-02 -2.74043620e-01 2.08337940e-02
7.56645977e-01 -3.60052854e-01 3.74619871e-01 1.61623314e-01
6.56822860e-01 1.05990124e+00 8.71300250e-02 -7.84551501e-01
5.91605127e-01 3.02889526e-01 -1.02941364e-01 -7.95863152e-01
3.61475572e-02 1.09444104e-01 -5.19665003e-01 -8.76801252e-01
7.39230394e-01 -6.47081554e-01 -1.21595657e+00 3.47841620e-01
-1.04284811e+00 -5.78754544e-01 -4.03387845e-01 6.02097273e-01
-8.25502157e-01 3.24804336e-01 -2.96800137e-01 -1.45134056e+00
1.49124730e-02 -1.18662071e+00 7.31090426e-01 3.47541779e-01
-6.76395237e-01 -4.68910307e-01 -2.66432762e-01 5.24639726e-01
4.64147151e-01 2.67583579e-01 7.73568451e-01 -4.84463960e-01
-3.68637800e-01 -2.36484837e-02 -8.32545459e-02 2.56540865e-01
2.97822565e-01 -2.87820667e-01 -6.79983675e-01 -3.78641076e-02
-1.02534905e-01 -5.02756059e-01 -2.59806454e-01 1.53130025e-01
3.35996538e-01 -4.29539531e-01 -3.42716336e-01 -2.80243486e-01
5.20822823e-01 7.00199664e-01 6.40701473e-01 5.70465386e-01
3.33453596e-01 1.65279496e+00 1.19039679e+00 4.91415501e-01
8.99534047e-01 4.49228168e-01 7.64162540e-01 2.95419991e-01
5.15939415e-01 -2.23916277e-01 5.85441530e-01 3.41473371e-01
-5.53647503e-02 5.17196283e-02 -7.69121170e-01 4.34096307e-01
-2.10706162e+00 -7.37453103e-01 -1.83331314e-02 2.18120575e+00
3.53355795e-01 7.48499334e-02 1.53712973e-01 1.33034006e-01
9.37772036e-01 -3.78082842e-01 -4.40041512e-01 -7.01866150e-01
3.20135623e-01 -7.59908915e-01 5.21806836e-01 2.85445988e-01
-7.39684880e-01 7.47645438e-01 5.55580902e+00 2.65480161e-01
-6.49558842e-01 1.66323945e-01 8.16319883e-01 6.18796423e-02
-3.00882399e-01 -8.94085243e-02 -1.23251788e-01 1.65282860e-01
6.43947363e-01 -2.00961530e-01 5.64386487e-01 1.03863716e+00
6.64739430e-01 -5.21539986e-01 -1.00331819e+00 5.63103974e-01
-3.58392984e-01 -3.26830149e-01 -4.87469256e-01 1.51691690e-01
3.93156856e-01 -4.13835794e-01 9.46532339e-02 2.89619416e-01
5.09634674e-01 -8.59356523e-01 1.27152705e+00 3.68414402e-01
2.98772991e-01 -7.61107922e-01 7.81456113e-01 5.60827255e-01
-8.05282414e-01 -6.20239973e-01 -4.79926430e-02 -6.84046924e-01
3.57881606e-01 2.03159064e-01 -9.09270644e-01 3.00675392e-01
7.22729266e-01 4.47106272e-01 -9.77930427e-02 6.97089791e-01
-2.85401016e-01 -2.02834100e-01 -5.78177907e-02 -7.07137048e-01
-8.97653624e-02 -4.69949156e-01 7.09255397e-01 2.98399538e-01
5.01953840e-01 -8.23723990e-03 -6.57742396e-02 7.95852721e-01
4.93122250e-01 2.29526252e-01 -1.06344700e+00 7.34927729e-02
7.28201985e-01 9.07606959e-01 -7.78508723e-01 3.63979459e-01
1.06507547e-01 9.11339879e-01 1.72786668e-01 1.82040334e-01
-1.00641239e+00 -4.47208554e-01 9.51452136e-01 2.74662133e-02
-4.12585467e-01 -2.25928009e-01 -4.78393704e-01 -4.68021542e-01
-6.20912798e-02 -5.57711542e-01 -2.01052174e-01 -7.10055113e-01
-1.02658701e+00 2.70757407e-01 1.24512814e-01 -1.24039483e+00
-1.15329534e-01 -4.33597356e-01 -3.59781563e-01 1.44839913e-01
-1.22005045e+00 -1.03077114e+00 -1.09631926e-01 2.21607238e-02
-3.54331918e-02 4.26055342e-02 5.44451118e-01 2.97790736e-01
-4.63101953e-01 2.47951105e-01 -7.50942454e-02 -6.55292869e-01
7.62081683e-01 -5.63377082e-01 -3.48036699e-02 3.36982489e-01
-9.60381210e-01 4.91554648e-01 9.87831116e-01 -7.45287895e-01
-1.10921729e+00 -8.37291181e-01 7.19949722e-01 -2.38186985e-01
5.82557559e-01 -4.20258135e-01 -2.95541584e-01 5.46106815e-01
-1.19796269e-01 -6.55832529e-01 5.44008672e-01 -2.51301634e-03
2.10196570e-01 -1.66242972e-01 -1.76903319e+00 1.29681373e+00
1.45121288e+00 -3.04960012e-01 -3.93344045e-01 -4.57087196e-02
9.85931695e-01 1.97097257e-01 -4.20284063e-01 2.53597528e-01
7.25554824e-01 -1.09072411e+00 7.11070716e-01 -2.24840213e-02
1.87262267e-01 -5.48473150e-02 -2.43341461e-01 -1.12430406e+00
-2.46070892e-01 -7.60899007e-01 6.78247154e-01 1.28040993e+00
3.77376825e-01 -1.13731039e+00 5.82360506e-01 1.15920281e+00
-6.71771094e-02 -4.15200353e-01 -1.09766567e+00 -7.52302408e-01
6.51610047e-02 -5.67106247e-01 6.15808070e-01 8.36342752e-01
6.06678724e-01 1.28984340e-02 -5.07477403e-01 2.59556264e-01
6.00590706e-01 -5.96859455e-01 1.00711083e+00 -1.37303901e+00
1.92795768e-01 -5.13799906e-01 -3.91821682e-01 -4.61034924e-01
4.89085853e-01 -5.27379036e-01 4.88516390e-01 -1.42581308e+00
4.05809507e-02 -9.67249334e-01 5.84257469e-02 1.07157655e-01
2.93848962e-01 -1.26336142e-01 2.48763263e-01 5.49879372e-02
-5.85390806e-01 8.63139570e-01 1.05005801e+00 1.78502336e-01
-5.12804806e-01 -3.57951075e-02 -9.25943136e-01 8.66028130e-01
1.01679063e+00 -4.03637201e-01 -4.43431854e-01 1.62920117e-01
3.99045050e-01 3.25469859e-02 4.40998912e-01 -9.49432254e-01
1.90767497e-01 -6.00025296e-01 -3.02090108e-01 2.21423253e-01
2.98151076e-01 -1.36316359e+00 3.35235745e-01 1.01468492e+00
-3.89807463e-01 -3.41945171e-01 -2.19920605e-01 6.47986114e-01
5.75746298e-01 -1.46346807e-01 6.98680818e-01 2.55171776e-01
-1.95467263e-01 -4.48840469e-01 -1.03541934e+00 -6.82664812e-01
1.82756865e+00 -2.62027383e-01 -2.98053473e-01 -7.73477018e-01
-4.24045026e-01 7.47335672e-01 8.56444180e-01 6.86332285e-01
3.32074583e-01 -1.13283873e+00 -1.04948655e-01 -1.73563138e-01
5.88376559e-02 -1.77866518e-01 4.36916530e-01 7.44570196e-01
-5.23901522e-01 1.20812789e-01 -5.18454671e-01 -1.28303394e-01
-1.01997197e+00 4.21097159e-01 2.78595895e-01 2.47251198e-01
-2.83786237e-01 5.14702916e-01 3.22637618e-01 -1.05992734e+00
4.98314351e-01 -3.16185325e-01 -4.27725226e-01 -7.28978291e-02
-1.68637279e-02 7.83026397e-01 -7.39685893e-01 -8.57536077e-01
-3.63749832e-01 3.35830808e-01 5.44797599e-01 -8.21285471e-02
1.03655195e+00 -8.09334517e-01 -2.08567962e-01 3.45950902e-01
4.21371996e-01 -6.21377975e-02 -1.25243258e+00 4.43312973e-01
1.30126983e-01 -3.81935567e-01 -5.33298433e-01 -6.75208747e-01
-3.73041809e-01 3.76284599e-01 6.44866765e-01 3.39565814e-01
8.09301615e-01 -5.53349555e-02 6.65390909e-01 6.57799006e-01
1.13845956e+00 -1.41957104e+00 2.91296929e-01 2.17154354e-01
8.33371103e-01 -1.08709157e+00 -1.99593037e-01 -6.88151300e-01
-9.58236754e-01 5.81202626e-01 7.69549727e-01 3.52954082e-02
4.85718817e-01 -1.60517871e-01 -5.76864108e-02 1.59005553e-01
-2.17540696e-01 3.41176093e-02 -5.14600694e-01 1.00125754e+00
2.21135877e-02 4.68654066e-01 -5.19662857e-01 9.25495684e-01
-3.65857542e-01 -1.18373595e-01 8.02187324e-01 1.01136613e+00
-7.51232803e-01 -1.03882337e+00 -5.92840433e-01 1.81683406e-01
1.79725960e-02 7.33270824e-01 -4.14614707e-01 5.41503310e-01
4.40747559e-01 1.42826986e+00 -1.80122226e-01 -5.70680797e-01
5.62408090e-01 -1.92770794e-01 5.85561246e-02 -3.74543786e-01
-3.59412670e-01 -5.25218248e-01 5.08121610e-01 -6.70895398e-01
-3.97028178e-01 -9.09759223e-01 -1.39907050e+00 -2.36703813e-01
-2.07148194e-01 9.28381085e-02 7.78447568e-01 8.17195415e-01
4.61588681e-01 3.90174150e-01 8.06972325e-01 -1.41652215e+00
-4.91525501e-01 -6.29875839e-01 -5.77581942e-01 4.70158011e-02
4.60957475e-02 -9.53892410e-01 -4.58571702e-01 -5.33711493e-01] | [4.895188808441162, 1.0220483541488647] |
905e3e69-f145-4c36-ae93-edaae7359a71 | integrative-feature-and-cost-aggregation-with | 2209.08742 | null | https://arxiv.org/abs/2209.08742v2 | https://arxiv.org/pdf/2209.08742v2.pdf | Integrative Feature and Cost Aggregation with Transformers for Dense Correspondence | We present a novel architecture for dense correspondence. The current state-of-the-art are Transformer-based approaches that focus on either feature descriptors or cost volume aggregation. However, they generally aggregate one or the other but not both, though joint aggregation would boost each other by providing information that one has but other lacks, i.e., structural or semantic information of an image, or pixel-wise matching similarity. In this work, we propose a novel Transformer-based network that interleaves both forms of aggregations in a way that exploits their complementary information. Specifically, we design a self-attention layer that leverages the descriptor to disambiguate the noisy cost volume and that also utilizes the cost volume to aggregate features in a manner that promotes accurate matching. A subsequent cross-attention layer performs further aggregation conditioned on the descriptors of both images and aided by the aggregated outputs of earlier layers. We further boost the performance with hierarchical processing, in which coarser level aggregations guide those at finer levels. We evaluate the effectiveness of the proposed method on dense matching tasks and achieve state-of-the-art performance on all the major benchmarks. Extensive ablation studies are also provided to validate our design choices. | ['Stephen Lin', 'Seungryong Kim', 'Seokju Cho', 'Sunghwan Hong'] | 2022-09-19 | null | null | null | null | ['geometric-matching'] | ['computer-vision'] | [ 2.23992437e-01 -1.18314378e-01 -8.77719074e-02 -4.86822993e-01
-9.77600098e-01 -1.98902726e-01 6.65509880e-01 3.96643102e-01
-3.34708840e-01 3.36285293e-01 4.83083099e-01 5.91900684e-02
-9.58133712e-02 -1.07524586e+00 -7.39404142e-01 -7.31856287e-01
2.01914102e-01 3.20695132e-01 3.79000753e-01 -3.05919945e-01
2.16934934e-01 5.49407184e-01 -1.77701747e+00 6.29107237e-01
9.82340753e-01 1.58462048e+00 1.48525864e-01 2.25350067e-01
-1.52071074e-01 7.77161777e-01 -2.64211535e-01 -4.35130477e-01
6.41799688e-01 -2.25370467e-01 -6.22189403e-01 -5.39501533e-02
8.26659083e-01 -4.75699186e-01 -4.46212023e-01 9.94736969e-01
4.33605671e-01 6.56118914e-02 4.36420441e-01 -8.87472868e-01
-9.24294770e-01 4.25417036e-01 -6.90620065e-01 2.44893417e-01
2.83031285e-01 3.69899273e-01 1.42551267e+00 -1.04619396e+00
3.56113791e-01 1.33802938e+00 6.05971813e-01 4.97162854e-03
-1.19182706e+00 -7.57766962e-01 3.48485619e-01 1.04148440e-01
-1.36413634e+00 -5.13194621e-01 9.05545056e-01 -3.96803021e-01
1.04118311e+00 1.86706688e-02 6.24960899e-01 6.65335298e-01
1.68160155e-01 7.99055815e-01 1.12979770e+00 -2.86566228e-01
-5.50318174e-02 -9.14849788e-02 2.23383978e-01 7.37437963e-01
1.15381710e-01 3.70330870e-01 -5.77256203e-01 4.68922369e-02
6.25197351e-01 3.58100116e-01 -7.68045560e-02 -4.54744995e-01
-1.26506662e+00 8.08515370e-01 1.08881748e+00 3.27553362e-01
-7.18035281e-01 1.48372844e-01 2.18449622e-01 3.61492991e-01
4.60020095e-01 4.53361481e-01 -1.34910280e-02 2.46368229e-01
-1.17778468e+00 1.49593458e-01 4.14108127e-01 8.31420302e-01
1.17957127e+00 -2.85521537e-01 -6.50217235e-01 7.32742131e-01
2.04653755e-01 1.52810156e-01 3.37080270e-01 -6.25827193e-01
5.21398723e-01 1.02742493e+00 -6.02088086e-02 -1.16825557e+00
-2.82180667e-01 -7.50791788e-01 -8.25658739e-01 2.42225587e-01
2.46435687e-01 3.99637789e-01 -1.30011773e+00 1.82441378e+00
1.09790556e-01 1.96441174e-01 -2.42075533e-01 1.07702255e+00
9.68269527e-01 4.03991699e-01 5.62109277e-02 3.49599570e-01
1.26138508e+00 -1.18632269e+00 -4.34913814e-01 -1.59050897e-01
3.20951730e-01 -7.86851168e-01 1.00800681e+00 -1.09966859e-01
-1.28019917e+00 -8.96834493e-01 -1.27478898e+00 -4.75027829e-01
-4.77319092e-01 1.11633867e-01 4.39236134e-01 2.50399739e-01
-1.21674836e+00 7.57763684e-01 -8.48227382e-01 -2.44376540e-01
5.47787905e-01 5.21234572e-01 -2.11571157e-01 -9.45667550e-02
-1.24150658e+00 8.65973949e-01 -6.88865408e-02 1.29931578e-02
-5.03268600e-01 -8.62443030e-01 -9.93851960e-01 3.72694671e-01
6.94124699e-02 -8.91742885e-01 9.21404898e-01 -6.60402477e-01
-1.01950133e+00 9.12522078e-01 -2.01156080e-01 -4.99791712e-01
4.75637168e-01 -1.57540426e-01 -1.74808304e-03 2.27734551e-01
3.06415468e-01 1.12795091e+00 7.83465564e-01 -1.12698448e+00
-8.58689249e-01 -5.21814704e-01 2.95596421e-01 2.43806764e-01
-4.70800996e-01 -2.30581105e-01 -5.95427096e-01 -6.87912881e-01
2.30548695e-01 -6.61928177e-01 -2.18232736e-01 3.46612036e-01
-2.19711065e-01 -1.84459656e-01 6.33723557e-01 -4.64064032e-01
1.29721141e+00 -2.21953392e+00 1.97061881e-01 4.16079730e-01
4.94799584e-01 9.68806371e-02 -3.51126701e-01 4.19810027e-01
1.76247701e-01 -7.41872862e-02 -2.27525264e-01 -6.62829936e-01
1.38053164e-01 5.22764074e-03 -2.17323735e-01 3.98760617e-01
6.54549956e-01 1.13778651e+00 -8.29505920e-01 -4.89349425e-01
3.44123632e-01 6.58766150e-01 -7.10441172e-01 2.40794748e-01
8.91967639e-02 2.54936725e-01 -4.62722331e-01 7.13832021e-01
7.18307972e-01 -3.66398066e-01 -1.92883670e-01 -8.64663184e-01
-1.92594767e-01 4.63838518e-01 -1.01604617e+00 1.89697564e+00
-7.02513874e-01 3.75961065e-01 2.88233310e-02 -9.97143686e-01
1.07268500e+00 -1.32909730e-01 6.63136482e-01 -1.00978982e+00
4.60653193e-02 3.54218662e-01 1.45531490e-01 1.69254802e-02
6.22155964e-01 1.68849438e-01 -6.80252835e-02 4.80231404e-01
7.09131062e-02 2.22944412e-02 1.75701514e-01 2.60710239e-01
1.14741433e+00 6.49050176e-02 3.62538546e-01 -2.52386600e-01
5.70453703e-01 -1.96140930e-01 3.92266631e-01 9.00637865e-01
-1.08501449e-01 7.55906641e-01 2.48439729e-01 -4.59938914e-01
-1.02443409e+00 -1.03405869e+00 -1.68927163e-01 1.08462715e+00
5.07548451e-01 -3.61818045e-01 -4.12424475e-01 -6.07745588e-01
4.60936427e-01 1.11292645e-01 -9.22534108e-01 -3.02818358e-01
-5.92607379e-01 -4.60312068e-01 2.44498834e-01 8.80842865e-01
8.34438920e-01 -9.10964608e-01 -6.24092698e-01 2.61245430e-01
4.97014076e-02 -1.14478898e+00 -7.00888574e-01 2.36587837e-01
-6.64097369e-01 -8.52743328e-01 -5.35836458e-01 -6.69925630e-01
6.88235402e-01 4.36034948e-01 1.32368708e+00 3.67456853e-01
-1.73834458e-01 1.00814596e-01 -4.03907031e-01 2.25662831e-02
4.06805165e-02 2.44243324e-01 -2.94375509e-01 2.96811819e-01
3.77462775e-01 -7.41713762e-01 -8.67362142e-01 3.76934141e-01
-8.79212677e-01 2.65815686e-02 9.98711884e-01 1.15539753e+00
6.54584110e-01 -3.00240159e-01 3.73428792e-01 -6.94692194e-01
5.10848463e-01 -2.94546962e-01 -4.78897184e-01 2.53905922e-01
-4.83999968e-01 3.71194422e-01 5.03106713e-01 -1.32786855e-01
-7.86298871e-01 2.16177985e-01 -3.27974647e-01 -6.17138803e-01
1.41700149e-01 3.95987004e-01 -2.10012197e-01 -3.56442273e-01
3.28521907e-01 2.00843737e-01 -7.41612166e-02 -5.42282820e-01
4.19322312e-01 4.93359149e-01 7.24551857e-01 -7.25569785e-01
8.20518970e-01 6.37572229e-01 -1.08312875e-01 -2.47840002e-01
-9.51676548e-01 -6.04290724e-01 -7.22462296e-01 4.92531285e-02
7.38934755e-01 -9.89632249e-01 -5.58027089e-01 4.49679613e-01
-1.09629679e+00 6.82883933e-02 -3.73688668e-01 2.72849500e-01
-3.38298142e-01 2.45868415e-02 -6.32813811e-01 -2.92331845e-01
-4.96986389e-01 -1.51846766e+00 1.57902491e+00 1.84177384e-01
-1.02354728e-01 -7.12088346e-01 -1.00703225e-01 2.24080622e-01
7.00486362e-01 1.61221072e-01 7.43237078e-01 -7.09604204e-01
-9.35640693e-01 -2.53783733e-01 -9.03155088e-01 1.43500209e-01
1.94043875e-01 -2.41592824e-01 -1.04095852e+00 -2.50796437e-01
-2.72542775e-01 -2.52719164e-01 1.41447461e+00 2.51747668e-01
1.07749712e+00 -1.62283570e-01 -3.35175395e-01 7.89479792e-01
1.48687065e+00 -8.66894796e-02 6.74770772e-01 4.28675175e-01
8.40375185e-01 5.09533405e-01 5.05467951e-01 4.08888251e-01
5.89074373e-01 8.84635210e-01 4.87753361e-01 -6.06028259e-01
-4.45808768e-01 -2.19276890e-01 4.21627462e-02 5.63089252e-01
9.48446095e-02 1.38648242e-01 -6.20245695e-01 6.44568324e-01
-1.87504756e+00 -9.03730154e-01 2.72187531e-01 2.04732060e+00
8.41012895e-01 2.00288042e-01 7.43602738e-02 2.96353977e-02
6.47706091e-01 4.14837271e-01 -5.00930011e-01 -2.74633437e-01
-1.78123519e-01 5.64422667e-01 3.42363030e-01 5.74222803e-01
-1.23291647e+00 8.11487377e-01 5.36033058e+00 7.12816954e-01
-1.03651226e+00 -2.37670131e-02 6.06456518e-01 -2.90603638e-02
-5.51417470e-01 -5.31463362e-02 -7.51182854e-01 4.29932833e-01
2.48262122e-01 1.45507539e-02 1.44854993e-01 5.84222734e-01
-3.54417413e-02 -1.14108376e-01 -1.43931293e+00 7.92471111e-01
-8.89895111e-02 -1.45323563e+00 2.47083902e-01 1.95479244e-01
7.79385269e-01 7.15433657e-02 1.88049376e-01 2.29495689e-01
2.87803024e-01 -8.30153525e-01 7.38125861e-01 5.51120102e-01
6.14278257e-01 -5.67129731e-01 9.80739653e-01 -7.14992508e-02
-1.66810167e+00 -9.77154672e-02 -1.53365940e-01 2.37603616e-02
8.22839141e-02 8.24902356e-01 -1.83694452e-01 6.85024679e-01
5.95204890e-01 9.32210386e-01 -6.88749313e-01 1.04939473e+00
-2.51244120e-02 4.72993813e-02 -4.64320600e-01 3.26918691e-01
5.60481787e-01 -5.07718101e-02 3.44415337e-01 1.34636819e+00
2.70855844e-01 -1.03846498e-01 5.04768014e-01 1.10938346e+00
-1.97417855e-01 -7.61602521e-02 -3.91972899e-01 3.60767365e-01
6.50077522e-01 1.31225860e+00 -4.17891830e-01 -5.22963166e-01
-7.82746255e-01 8.41235936e-01 6.46861792e-01 1.41451597e-01
-6.96402788e-01 -4.77269500e-01 9.49934244e-01 2.62445211e-01
5.23935080e-01 3.91254127e-02 -4.93673503e-01 -1.02780950e+00
2.45406687e-01 -6.32606447e-01 4.93814588e-01 -5.47206223e-01
-1.45260286e+00 8.64882290e-01 -2.60289192e-01 -1.25717843e+00
-1.31859660e-01 -3.90091002e-01 -6.24173760e-01 1.15879202e+00
-1.90390217e+00 -1.43563890e+00 -6.55908883e-01 5.57272434e-01
3.60936314e-01 2.55951118e-02 5.30818224e-01 5.94373465e-01
-4.73479867e-01 9.12939787e-01 -2.74036348e-01 1.85804069e-01
7.08200634e-01 -1.07970989e+00 6.35080159e-01 8.58111143e-01
7.16085956e-02 6.09997988e-01 2.43292987e-01 -5.10270178e-01
-1.18653905e+00 -1.06272423e+00 9.35162127e-01 -2.40061935e-02
7.02353716e-01 -3.25071305e-01 -9.20048356e-01 2.92127699e-01
1.07524849e-01 3.78206670e-01 1.92084938e-01 -1.01869069e-02
-6.88066483e-01 -5.13065219e-01 -1.03097200e+00 3.22422385e-01
1.16386378e+00 -7.41666794e-01 -5.06766498e-01 -1.98937505e-01
8.10468674e-01 -4.16598618e-01 -9.56602156e-01 7.02973545e-01
6.36595786e-01 -1.40056264e+00 1.04516637e+00 -4.32998478e-01
6.16742074e-01 -3.86825085e-01 -4.17328179e-01 -1.20429730e+00
-5.83774686e-01 -3.10185343e-01 1.85268763e-02 1.24165094e+00
2.70634681e-01 -5.70966721e-01 7.48462915e-01 5.00086784e-01
-3.88299853e-01 -1.06273985e+00 -7.53844857e-01 -7.52374113e-01
-1.39788836e-01 -1.56849250e-01 8.83745492e-01 7.49391496e-01
-1.74885690e-01 2.94993222e-01 -1.46098986e-01 -2.07908843e-02
6.59294009e-01 5.16720951e-01 5.58491647e-01 -1.00117707e+00
-8.79558623e-02 -6.77401364e-01 -7.12672591e-01 -1.35087454e+00
8.50476697e-02 -9.59555030e-01 -1.74840819e-03 -1.54122996e+00
3.57360393e-01 -6.49133086e-01 -5.41475296e-01 5.53003252e-01
-4.20461237e-01 5.34021974e-01 3.77324134e-01 2.68647909e-01
-5.73964477e-01 6.19000971e-01 1.23143649e+00 -3.60347539e-01
-2.06999391e-01 -3.67435813e-01 -8.59941959e-01 2.88787872e-01
5.17595470e-01 -1.98252082e-01 -2.29617760e-01 -6.04523599e-01
-2.72777349e-01 -2.93834090e-01 5.75518072e-01 -1.11867857e+00
4.34815973e-01 2.84583092e-01 5.22600710e-01 -6.14848316e-01
3.74886751e-01 -8.37787986e-01 -1.77989379e-01 3.03159416e-01
-5.43993711e-01 2.23343268e-01 5.78389689e-02 3.06346923e-01
-6.02456868e-01 3.05107206e-01 9.16288018e-01 -7.67604113e-02
-6.34238780e-01 6.85949922e-01 2.39315063e-01 -3.72009911e-02
7.33532369e-01 -3.52965027e-01 -3.30334872e-01 -3.12944919e-01
-5.25707364e-01 2.94703573e-01 5.33215344e-01 4.43907946e-01
6.61353052e-01 -1.68455195e+00 -8.53050292e-01 4.54718292e-01
3.28008622e-01 -1.17451802e-01 9.42094103e-02 1.09030652e+00
-1.50868624e-01 4.17914331e-01 -4.66122657e-01 -8.02865505e-01
-9.12313640e-01 6.33012891e-01 3.38483691e-01 -7.24839032e-01
-6.22086525e-01 7.53206491e-01 6.57520533e-01 -5.73308393e-02
2.94348508e-01 -4.30120111e-01 -9.34335515e-02 2.71103352e-01
4.76155818e-01 4.15435918e-02 3.57216388e-01 -7.94814289e-01
-5.09744227e-01 8.27726305e-01 -3.75317693e-01 8.61244723e-02
1.29599679e+00 -1.41211614e-01 -1.34360239e-01 -7.29534552e-02
1.54987884e+00 -2.41855741e-01 -1.26350212e+00 -7.17583001e-01
-1.78180218e-01 -8.35069239e-01 2.23099008e-01 -6.48532391e-01
-1.46651411e+00 8.92130971e-01 4.87232208e-01 -1.15816789e-02
1.52749908e+00 1.01177782e-01 9.86167848e-01 9.22452100e-03
2.09381178e-01 -7.92853713e-01 8.26346651e-02 2.65189052e-01
9.38778102e-01 -1.37545276e+00 3.77828330e-02 -3.83492112e-01
-3.63492608e-01 9.11212564e-01 8.50628316e-01 -3.07478219e-01
6.56608522e-01 4.78922248e-01 -1.30906060e-01 -2.96138644e-01
-8.53997469e-01 -7.22764969e-01 6.33825839e-01 3.81631047e-01
5.01325846e-01 -1.30066618e-01 -2.88946182e-01 4.91566062e-01
7.80437738e-02 -1.88131914e-01 -1.15270458e-01 8.16428065e-01
-3.77028733e-01 -1.10329950e+00 -3.29429656e-01 7.68469214e-01
-2.42852062e-01 -4.10955459e-01 -5.40327251e-01 6.18056536e-01
1.48226216e-01 7.16434658e-01 5.07008731e-01 -6.86710298e-01
5.85034966e-01 -3.85759026e-01 4.71224546e-01 -4.10592884e-01
-9.46113646e-01 5.90156615e-02 2.36948766e-03 -8.77405107e-01
-5.11117041e-01 -5.39076626e-01 -1.02577746e+00 -4.41246241e-01
-2.58553535e-01 -7.04323798e-02 2.62728184e-01 8.15352499e-01
5.02268493e-01 5.34268975e-01 8.01857710e-01 -1.10048580e+00
-5.17679751e-01 -8.55169058e-01 -2.36945242e-01 5.54380417e-01
6.25716507e-01 -9.27632987e-01 -2.05786586e-01 -2.30444163e-01] | [10.1117525100708, 0.2077493667602539] |
85fa53af-07c1-4fd7-881c-1b801a112add | cloning-outfits-from-real-world-images-to-3d | 2204.02611 | null | https://arxiv.org/abs/2204.02611v2 | https://arxiv.org/pdf/2204.02611v2.pdf | Cloning Outfits from Real-World Images to 3D Characters for Generalizable Person Re-Identification | Recently, large-scale synthetic datasets are shown to be very useful for generalizable person re-identification. However, synthesized persons in existing datasets are mostly cartoon-like and in random dress collocation, which limits their performance. To address this, in this work, an automatic approach is proposed to directly clone the whole outfits from real-world person images to virtual 3D characters, such that any virtual person thus created will appear very similar to its real-world counterpart. Specifically, based on UV texture mapping, two cloning methods are designed, namely registered clothes mapping and homogeneous cloth expansion. Given clothes keypoints detected on person images and labeled on regular UV maps with clear clothes structures, registered mapping applies perspective homography to warp real-world clothes to the counterparts on the UV map. As for invisible clothes parts and irregular UV maps, homogeneous expansion segments a homogeneous area on clothes as a realistic cloth pattern or cell, and expand the cell to fill the UV map. Furthermore, a similarity-diversity expansion strategy is proposed, by clustering person images, sampling images per cluster, and cloning outfits for 3D character generation. This way, virtual persons can be scaled up densely in visual similarity to challenge model learning, and diversely in population to enrich sample distribution. Finally, by rendering the cloned characters in Unity3D scenes, a more realistic virtual dataset called ClonedPerson is created, with 5,621 identities and 887,766 images. Experimental results show that the model trained on ClonedPerson has a better generalization performance, superior to that trained on other popular real-world and synthetic person re-identification datasets. The ClonedPerson project is available at https://github.com/Yanan-Wang-cs/ClonedPerson. | ['Shengcai Liao', 'Xuezhi Liang', 'Yanan Wang'] | 2022-04-06 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Wang_Cloning_Outfits_From_Real-World_Images_to_3D_Characters_for_Generalizable_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Wang_Cloning_Outfits_From_Real-World_Images_to_3D_Characters_for_Generalizable_CVPR_2022_paper.pdf | cvpr-2022-1 | ['unsupervised-person-re-identification', 'generalizable-person-re-identification'] | ['computer-vision', 'computer-vision'] | [ 1.26695752e-01 -1.90196812e-01 2.10749283e-01 -3.15310061e-01
-1.04382738e-01 -7.01911926e-01 4.74752665e-01 -4.07657772e-01
-9.24487188e-02 6.49445415e-01 -3.47424634e-02 4.74073142e-01
4.66830462e-01 -1.04980099e+00 -8.07284534e-01 -6.05243504e-01
3.38465512e-01 7.58406043e-01 -6.06813729e-02 -2.79421002e-01
-1.43512651e-01 4.70660955e-01 -1.50807953e+00 9.51179862e-02
9.28314030e-01 4.28264111e-01 1.22828081e-01 5.45421422e-01
-5.29022627e-02 -7.88266491e-03 -8.70702147e-01 -1.09610033e+00
5.50620914e-01 -5.46909451e-01 -2.74907231e-01 5.07424772e-01
8.95698488e-01 -3.80099386e-01 -3.85146469e-01 1.22732735e+00
6.21112108e-01 1.57894462e-01 6.80865645e-01 -1.20792925e+00
-1.45508480e+00 4.06004548e-01 -8.57960403e-01 -5.55895448e-01
8.62453580e-01 1.33241266e-01 3.53832483e-01 -1.05393124e+00
8.64423692e-01 1.53154659e+00 9.79156852e-01 1.05796885e+00
-1.54353976e+00 -8.61996174e-01 1.13230601e-01 -1.60285249e-01
-1.59983873e+00 -1.61684111e-01 9.89260674e-01 -2.86953270e-01
1.30380705e-01 6.33888543e-01 1.25573409e+00 1.80394197e+00
-2.34447926e-01 8.40512931e-01 1.38667893e+00 -3.20268065e-01
-1.20276302e-01 8.20859611e-01 3.12312581e-02 7.40853548e-01
4.46016431e-01 7.80409276e-02 -1.71331316e-01 -3.19819063e-01
1.07837451e+00 2.43200049e-01 -5.53028047e-01 -5.30484259e-01
-1.24791551e+00 4.74221230e-01 2.92031616e-01 -1.06512062e-01
-1.60469130e-01 -2.75037706e-01 2.00112462e-01 5.15908450e-02
3.62324178e-01 3.81948084e-01 1.16892688e-01 3.64317834e-01
-5.64384758e-01 6.09124959e-01 4.82277095e-01 1.38716698e+00
6.89072907e-01 4.68864925e-02 -2.79519171e-01 1.36414528e+00
-5.49322702e-02 1.09955645e+00 4.13478047e-01 -7.73743510e-01
1.91287711e-01 7.61114717e-01 1.32216692e-01 -1.13167548e+00
-1.14153698e-02 -2.87279099e-01 -1.23126626e+00 -3.25693376e-02
4.77365375e-01 -1.61289815e-02 -9.65311766e-01 1.69324720e+00
4.31264281e-01 2.27790266e-01 1.05039719e-02 1.14787424e+00
9.11893070e-01 5.29810488e-01 -2.28473842e-01 1.96513787e-01
1.63554776e+00 -8.54732454e-01 -5.36090672e-01 -2.10594423e-02
-1.79506898e-01 -7.97670662e-01 1.41952288e+00 3.20251793e-01
-1.09501529e+00 -1.08866739e+00 -8.55009377e-01 2.52777398e-01
-3.86006206e-01 4.17743951e-01 5.08527935e-01 1.08359575e+00
-8.20687115e-01 4.35902953e-01 -7.09456354e-02 -7.17937052e-01
3.29768807e-01 -9.98398736e-02 -3.89341593e-01 -2.62914330e-01
-1.16080177e+00 3.76997381e-01 1.69421881e-01 2.22406201e-02
-7.92253911e-01 -6.31113589e-01 -8.60439181e-01 -4.59503353e-01
1.80722959e-02 -9.21776950e-01 5.90344310e-01 -9.53614652e-01
-1.43093109e+00 1.22542560e+00 7.52174929e-02 5.84835634e-02
1.02078986e+00 1.94586068e-01 -6.83277965e-01 -1.08472124e-01
2.04751223e-01 6.02724373e-01 1.05656314e+00 -2.01685357e+00
-2.86784112e-01 -5.93636990e-01 -1.87183663e-01 2.09154159e-01
-6.64011776e-01 -9.97513086e-02 -9.72551525e-01 -1.15363765e+00
8.36747661e-02 -1.15494478e+00 2.62506194e-02 -2.52751708e-02
-5.78207076e-01 1.29523918e-01 4.29143578e-01 -9.21338618e-01
6.84347034e-01 -1.96998000e+00 2.74764597e-01 5.27921259e-01
1.39405593e-01 2.90582534e-02 -1.86893731e-01 2.15539560e-01
3.98945510e-02 1.77459344e-02 -1.70395166e-01 -6.28320038e-01
9.45658758e-02 6.12308830e-02 -1.95505351e-01 3.57845753e-01
-2.24606410e-01 8.43335390e-01 -8.80914271e-01 -3.97874266e-01
2.20617294e-01 5.38344681e-01 -1.34719685e-01 1.58063278e-01
2.28533402e-01 6.12490952e-01 -1.76027849e-01 9.86345291e-01
1.13519990e+00 1.61120027e-01 2.10338965e-01 -2.95140833e-01
3.84227246e-01 -8.64514172e-01 -1.28702962e+00 1.58779275e+00
-7.05586448e-02 3.26409101e-01 -7.80053958e-02 -3.25755209e-01
1.36842453e+00 3.87100168e-02 3.37507993e-01 -5.14730513e-01
1.79422498e-02 1.96323693e-02 -7.40288496e-01 -3.13135564e-01
6.97661757e-01 1.42787382e-01 -1.14511073e-01 2.45892331e-01
-2.45207205e-01 -1.27553299e-01 1.57840297e-01 6.71356022e-02
3.19820821e-01 2.95944959e-01 -3.50769520e-01 -6.96950480e-02
4.84556079e-01 -8.31212327e-02 4.64380890e-01 7.44318247e-01
-1.13634430e-01 1.01611161e+00 -6.85546994e-02 -6.39884412e-01
-1.53408015e+00 -1.46393263e+00 -2.76529014e-01 5.61422288e-01
6.74089193e-01 -1.23444244e-01 -1.12427974e+00 -5.14081299e-01
2.14369103e-01 4.38242286e-01 -8.49970102e-01 -9.67325494e-02
-5.97219944e-01 -8.78301620e-01 8.32058489e-01 3.11852604e-01
9.27484751e-01 -9.61547852e-01 2.21593574e-01 -2.36085132e-02
-3.87975007e-01 -1.09073520e+00 -9.03202593e-01 -8.89882684e-01
-3.06354523e-01 -9.86537397e-01 -1.65524960e+00 -1.07382476e+00
1.02532005e+00 5.11893928e-01 1.04036629e+00 -4.09948789e-02
-4.49821591e-01 4.95447934e-01 -3.65111202e-01 -1.73753813e-01
-5.42287111e-01 -4.12943363e-01 5.48006833e-01 5.45218766e-01
2.77248174e-01 -3.00945491e-01 -6.34201944e-01 7.01047838e-01
-4.73455936e-01 3.76838297e-01 2.76876569e-01 8.63082826e-01
5.70079207e-01 3.22622135e-02 2.51658171e-01 -7.27206290e-01
8.61876845e-01 -5.26643805e-02 -3.13114226e-01 5.46322763e-01
-3.06289494e-01 -7.28602827e-01 6.69880271e-01 -1.08880091e+00
-1.09326041e+00 -9.46477056e-02 1.53130740e-01 -8.28465462e-01
-3.24852318e-01 -4.17617083e-01 -3.16618711e-01 -1.75637975e-01
8.19456995e-01 6.14961684e-01 4.05041464e-02 -4.20293212e-01
4.25278425e-01 6.35972679e-01 7.48473108e-01 -8.75942588e-01
1.12467444e+00 6.08167529e-01 -5.92870653e-01 -8.83808911e-01
-1.05375074e-01 -1.16582133e-01 -5.22680938e-01 -4.91045147e-01
9.85591590e-01 -1.00721717e+00 -6.86850369e-01 9.85447049e-01
-1.00009930e+00 -2.91200370e-01 -4.57744837e-01 3.72566909e-01
-2.29264081e-01 6.39862180e-01 -7.65831292e-01 -7.66800880e-01
-2.84527153e-01 -8.40768337e-01 1.05832362e+00 4.77201253e-01
-2.11539552e-01 -6.21346712e-01 3.40924524e-02 6.09782994e-01
-2.21473314e-02 4.42912072e-01 6.08507872e-01 -6.96514323e-02
-3.42422456e-01 -4.95919734e-01 -3.50884974e-01 3.57591897e-01
2.42636561e-01 -7.46401846e-02 -9.70658302e-01 -5.07478893e-01
-3.44900846e-01 -1.06363803e-01 5.52515864e-01 1.37258545e-01
1.10518861e+00 -8.97497684e-02 -3.93784791e-01 8.36748183e-01
1.14478433e+00 1.00295387e-01 4.79409158e-01 9.86288413e-02
1.05600965e+00 8.07932377e-01 3.26941639e-01 5.13678968e-01
5.58090866e-01 7.44575918e-01 -1.32216230e-01 -4.52943414e-01
-4.35234904e-01 -6.33921444e-01 3.13411504e-01 5.04004121e-01
-6.85950935e-01 -1.85980126e-01 -4.99479353e-01 2.42388695e-01
-1.51922905e+00 -1.13757122e+00 -1.30710930e-01 2.51199269e+00
6.86914265e-01 -8.88813958e-02 3.91662776e-01 -2.74446249e-01
1.31773114e+00 -7.91501477e-02 -4.88442570e-01 4.86062951e-02
-6.97024822e-01 -2.10356236e-01 4.73716766e-01 2.72750258e-01
-7.89190233e-01 9.73098814e-01 5.17820883e+00 8.85291338e-01
-6.57862604e-01 1.80940032e-02 6.00325823e-01 2.24098358e-02
-4.45494056e-01 -4.27004814e-01 -8.95403087e-01 7.79100239e-01
-3.22809555e-02 1.26735240e-01 8.55856419e-01 6.86961651e-01
5.81936119e-03 4.10658479e-01 -8.25096250e-01 1.58163607e+00
4.64075089e-01 -1.04553056e+00 3.77078891e-01 1.68576539e-01
1.06761801e+00 -7.57940948e-01 2.35350862e-01 3.24612439e-01
3.32683772e-01 -7.74054229e-01 8.59638453e-01 6.58173561e-01
1.18044031e+00 -6.50091767e-01 5.64982355e-01 1.16754673e-01
-1.40227389e+00 1.65121347e-01 -7.73505330e-01 4.54736501e-01
9.65208188e-02 1.56887442e-01 -6.03494287e-01 6.40976429e-01
9.14230108e-01 4.65629637e-01 -7.75274098e-01 7.10704565e-01
7.76067823e-02 2.16467828e-02 -9.73109603e-02 -6.18515760e-02
-3.68351549e-01 -7.79596388e-01 5.74140370e-01 1.11185408e+00
4.49639112e-01 -3.42513658e-02 3.60855907e-01 1.21660709e+00
-1.53270647e-01 1.44208968e-01 -7.12580562e-01 2.87237167e-01
4.04315948e-01 1.21810579e+00 -5.52149177e-01 -5.80182314e-01
-2.08675876e-01 1.73441064e+00 2.50814278e-02 7.76456118e-01
-1.20448279e+00 -1.62527263e-01 6.00670457e-01 1.82354629e-01
-2.34639913e-01 -1.93990786e-02 -3.23353484e-02 -1.46672392e+00
2.51422167e-01 -1.13017702e+00 1.15310028e-01 -9.53263819e-01
-1.80559814e+00 8.20863843e-01 -5.66414446e-02 -1.24858272e+00
4.06376660e-01 -4.56632674e-01 -4.37954813e-01 1.09952414e+00
-7.84321606e-01 -1.75779808e+00 -9.53593552e-01 9.08280611e-01
5.41523397e-01 -5.28173983e-01 8.73586774e-01 3.96397352e-01
-6.92144573e-01 9.50936675e-01 1.56128839e-01 4.92029279e-01
9.18457985e-01 -1.17603719e+00 7.83770442e-01 6.69756651e-01
1.31545871e-01 8.02284956e-01 4.09527034e-01 -1.10305572e+00
-1.42799032e+00 -1.14706767e+00 3.47199798e-01 -6.17470026e-01
1.22864440e-01 -8.94919991e-01 -8.47181261e-01 5.11676192e-01
6.77424017e-04 -2.00879619e-01 4.16846693e-01 -1.51876643e-01
-4.55677599e-01 -1.17708139e-01 -1.27282929e+00 1.03642297e+00
1.43590295e+00 -4.21846837e-01 -2.47754648e-01 4.18396086e-01
3.83894145e-01 -5.13839960e-01 -9.30639863e-01 3.07972610e-01
9.13386583e-01 -9.11229134e-01 1.48542190e+00 -3.16564292e-01
6.11243919e-02 -3.48775536e-01 -1.25825807e-01 -1.27541220e+00
-4.82506186e-01 -5.93632936e-01 8.11868161e-02 1.50366282e+00
7.08129629e-02 -7.31188893e-01 1.12504494e+00 4.93490309e-01
1.70001999e-01 -5.05476117e-01 -2.76028305e-01 -8.78657877e-01
-1.77344516e-01 -5.75736212e-03 1.10194612e+00 1.13008928e+00
-6.12911820e-01 -5.01252487e-02 -7.77486861e-01 2.31842458e-01
1.13954639e+00 2.47750267e-01 1.38027358e+00 -1.17415774e+00
-3.40047389e-01 -3.62474710e-01 -2.48148680e-01 -1.06930459e+00
6.96111470e-02 -8.60396266e-01 -4.03592616e-01 -1.28675556e+00
5.33769369e-01 -8.32495153e-01 1.75580621e-01 4.49682102e-02
-3.57733995e-01 7.14318991e-01 3.48406672e-01 3.42108637e-01
3.91402617e-02 6.57638192e-01 1.83456695e+00 -5.38436353e-01
-2.27771312e-01 2.21056178e-01 -4.93704438e-01 6.35773122e-01
6.76593840e-01 1.14903592e-01 -4.27321792e-01 -2.66166836e-01
-3.26620340e-01 -1.59622490e-01 8.47371757e-01 -1.02134562e+00
-1.04729697e-01 -2.15303507e-02 1.09939694e+00 -4.61748362e-01
7.77811706e-01 -7.27383256e-01 6.23648882e-01 3.55897665e-01
-4.04010043e-02 1.11491904e-01 -7.11282864e-02 5.92859566e-01
2.83161163e-01 -1.01712190e-01 7.52490819e-01 -4.76300031e-01
-6.21056080e-01 6.59553409e-01 8.90391245e-02 -5.98125495e-02
1.22634470e+00 -8.40384305e-01 -2.38298818e-01 -2.25944757e-01
-8.22274029e-01 -6.55242950e-02 1.13673544e+00 6.16700232e-01
9.39790428e-01 -1.63448584e+00 -1.03530681e+00 6.21981323e-01
2.58425117e-01 -2.74664015e-01 5.74723184e-01 1.98357448e-01
-5.42969167e-01 -2.89631307e-01 -4.17714953e-01 -6.41592622e-01
-1.43563330e+00 7.87541509e-01 4.38595861e-01 2.86796391e-01
-9.07512248e-01 9.17165101e-01 4.10450846e-01 -9.44180727e-01
1.67857632e-01 1.47968218e-01 -7.16645494e-02 -2.08762549e-02
5.10242999e-01 3.76457363e-01 -4.58173603e-01 -8.10933173e-01
-2.71875784e-02 1.05803514e+00 9.42282155e-02 -4.86679375e-02
7.47567534e-01 -2.32814357e-01 -7.71683380e-02 1.15620784e-01
9.34251368e-01 3.86589020e-01 -1.25343800e+00 -2.83551842e-01
-6.82154179e-01 -9.14957941e-01 -8.61819625e-01 -5.16020119e-01
-1.03806877e+00 4.09921765e-01 9.27069366e-01 2.33529359e-01
9.07253742e-01 -1.66097492e-01 9.88089919e-01 -5.14860563e-02
7.16833711e-01 -9.53705192e-01 2.46561021e-01 -7.62936398e-02
9.51692283e-01 -1.16102505e+00 -1.32061884e-01 -6.78890288e-01
-9.58136797e-01 8.35323215e-01 8.54369283e-01 -1.36783943e-01
6.24914765e-02 -1.35527298e-01 8.29945058e-02 4.94156629e-02
2.34195650e-01 8.51744711e-02 2.46503636e-01 1.04207921e+00
-2.08259776e-01 3.88604164e-01 2.73358196e-01 6.37884200e-01
-4.65405166e-01 -4.90740359e-01 3.12231362e-01 3.80741268e-01
9.87092555e-02 -1.12378168e+00 -9.43869174e-01 1.30803794e-01
1.43562943e-01 9.39204916e-02 -5.64607978e-01 8.29527318e-01
4.86147702e-01 6.78997815e-01 5.77410460e-02 -6.57389522e-01
6.45539999e-01 -2.37412632e-01 8.17963898e-01 -2.84760654e-01
-5.15841424e-01 -3.73908132e-02 -6.86355382e-02 -7.08020553e-02
-1.13152035e-01 -5.89274704e-01 -7.61731505e-01 -8.54320109e-01
-1.09356649e-01 5.16381813e-04 4.48075652e-01 2.17782587e-01
-3.68460491e-02 2.97475010e-01 7.63686359e-01 -1.05394948e+00
-3.30292642e-01 -6.92392409e-01 -8.08759868e-01 1.02634418e+00
-1.95858285e-01 -5.20486176e-01 5.62460441e-03 2.32163906e-01] | [12.120467185974121, -0.7940322160720825] |
558ca1a0-51f5-401b-9fca-0bb0456bdd4f | user-simulation-for-evaluating-information | 2306.08550 | null | https://arxiv.org/abs/2306.08550v1 | https://arxiv.org/pdf/2306.08550v1.pdf | User Simulation for Evaluating Information Access Systems | Information access systems, such as search engines, recommender systems, and conversational assistants, have become integral to our daily lives as they help us satisfy our information needs. However, evaluating the effectiveness of these systems presents a long-standing and complex scientific challenge. This challenge is rooted in the difficulty of assessing a system's overall effectiveness in assisting users to complete tasks through interactive support, and further exacerbated by the substantial variation in user behaviour and preferences. To address this challenge, user simulation emerges as a promising solution. This book focuses on providing a thorough understanding of user simulation techniques designed specifically for evaluation purposes. We begin with a background of information access system evaluation and explore the diverse applications of user simulation. Subsequently, we systematically review the major research progress in user simulation, covering both general frameworks for designing user simulators, utilizing user simulation for evaluation, and specific models and algorithms for simulating user interactions with search engines, recommender systems, and conversational assistants. Realizing that user simulation is an interdisciplinary research topic, whenever possible, we attempt to establish connections with related fields, including machine learning, dialogue systems, user modeling, and economics. We end the book with a detailed discussion of important future research directions, many of which extend beyond the evaluation of information access systems and are expected to have broader impact on how to evaluate interactive intelligent systems in general. | ['ChengXiang Zhai', 'Krisztian Balog'] | 2023-06-14 | null | null | null | null | ['user-simulation'] | ['natural-language-processing'] | [-7.11452365e-02 5.07913120e-02 -4.88435656e-01 -2.63740569e-01
-2.58797407e-01 -7.92037249e-01 6.10069871e-01 -5.41099422e-02
-6.19596004e-01 5.94862103e-01 2.41052993e-02 -8.78993273e-01
-5.45661688e-01 -4.66556996e-01 1.26012787e-01 -9.60430652e-02
-2.13928744e-02 6.26810849e-01 -7.86182284e-02 -8.09966326e-01
3.62143397e-01 4.32682931e-01 -2.04965925e+00 4.28220332e-02
1.21102214e+00 5.57482660e-01 5.06185949e-01 9.62641239e-01
-2.77372211e-01 5.95335424e-01 -6.88134253e-01 -3.58909428e-01
-1.86302111e-01 -1.91770509e-01 -9.34998214e-01 -3.54371846e-01
-8.88721794e-02 -6.83450341e-01 -4.79183376e-01 4.93900001e-01
7.29644418e-01 7.43620038e-01 7.59319425e-01 -1.31675017e+00
-3.84717345e-01 3.59127522e-01 2.73873270e-01 2.07691342e-01
1.01168787e+00 -4.23984714e-02 9.54313576e-01 -5.11835635e-01
2.42085084e-01 1.10744202e+00 4.45339620e-01 6.47884250e-01
-1.02615595e+00 -7.16697276e-02 3.08607578e-01 1.73680440e-01
-1.11302650e+00 -6.41850710e-01 9.95341167e-02 -3.59809428e-01
1.11923480e+00 9.84895706e-01 6.55293405e-01 8.57412219e-01
-3.33807506e-02 1.12533760e+00 6.11626148e-01 -6.86541021e-01
1.02615044e-01 1.08567405e+00 7.78852940e-01 5.41940272e-01
-8.91636088e-02 5.55721819e-02 -6.53541610e-02 -5.75468183e-01
5.43078125e-01 1.55728981e-01 -3.69179040e-01 -2.41612494e-01
-7.21530676e-01 8.05219114e-01 5.52568175e-02 3.56861651e-01
-4.96291727e-01 -4.92937177e-01 -4.73703071e-03 5.57559192e-01
9.48142484e-02 8.78651321e-01 -5.17257631e-01 -5.93806565e-01
-4.37577158e-01 6.49403572e-01 1.75389302e+00 1.04888487e+00
9.41703916e-02 -3.75126034e-01 -5.38108826e-01 1.28780687e+00
6.26809120e-01 5.49584627e-01 5.14632344e-01 -9.85148728e-01
-2.65130013e-01 2.62762994e-01 6.91080570e-01 -7.61218429e-01
-4.55008954e-01 -2.24544629e-01 -2.97888488e-01 -1.61786526e-01
4.27070647e-01 -4.74250585e-01 -2.29210377e-01 1.43724835e+00
1.13536313e-01 -7.65832424e-01 -4.25118245e-02 7.38733768e-01
1.09148681e+00 6.31430924e-01 2.97607988e-01 -4.59016144e-01
1.49767315e+00 -9.05100107e-01 -1.04870450e+00 1.83800623e-01
5.88653028e-01 -7.86846519e-01 1.33962095e+00 1.28890842e-01
-1.51480269e+00 -3.01378101e-01 -5.92729270e-01 3.59268263e-02
-5.82470298e-01 -1.96557164e-01 8.38617980e-01 1.17063665e+00
-1.08263481e+00 3.17840368e-01 -6.75236881e-01 -8.93209457e-01
-3.79799545e-01 4.68457669e-01 4.40149158e-01 1.49412826e-01
-1.54023933e+00 1.23680210e+00 -3.17058325e-01 -4.47257489e-01
-2.41257504e-01 -5.87833464e-01 -4.61971819e-01 3.74362797e-01
3.50844145e-01 -9.27830517e-01 2.05515218e+00 -3.53132099e-01
-1.68701720e+00 2.47723266e-01 -1.94179013e-01 9.66182724e-02
4.03569013e-01 -1.50456637e-01 -4.39897716e-01 -4.10861343e-01
-3.66591722e-01 1.67621121e-01 3.60051207e-02 -1.13533425e+00
-1.01488435e+00 -3.05067897e-01 4.69536811e-01 1.09577167e+00
-6.09901965e-01 4.64359701e-01 -7.23956347e-01 -1.22674242e-01
-4.86300319e-01 -7.64301598e-01 -5.91911912e-01 -2.85407811e-01
-3.05492375e-02 -4.62118000e-01 5.33919692e-01 -4.09792095e-01
1.94878566e+00 -1.73489201e+00 -4.39581610e-02 3.37511778e-01
1.79327160e-01 4.01717514e-01 2.52768129e-01 8.66888583e-01
6.49686933e-01 1.86904177e-01 3.95081997e-01 -1.60486281e-01
3.57293993e-01 -4.03304994e-02 8.78332648e-03 -2.40639955e-01
-8.38227272e-01 6.23090506e-01 -9.11962450e-01 -1.17665656e-01
4.18061405e-01 5.47768116e-01 -5.84156454e-01 5.78764617e-01
-1.71993285e-01 7.66332522e-02 -7.09568322e-01 3.86234224e-01
3.57299373e-02 -4.59865600e-01 2.31373370e-01 2.54193813e-01
-3.95086080e-01 4.24284637e-01 -1.06757009e+00 1.00555074e+00
-8.21699977e-01 3.47465366e-01 3.29746157e-01 -4.57823962e-01
3.14498574e-01 4.19272602e-01 5.26423633e-01 -8.84280324e-01
1.34709924e-01 -1.55351600e-02 1.76052079e-01 -7.40223646e-01
9.46063101e-01 5.11441231e-01 2.22892582e-01 1.02367735e+00
-3.15066189e-01 1.13593519e-01 1.51386634e-01 4.39690232e-01
9.38516855e-01 -3.67597044e-01 4.28708106e-01 -3.92899632e-01
4.43184376e-01 -2.04890385e-01 -3.11609358e-01 1.43017328e+00
-2.95509100e-01 -1.45649180e-01 -5.93342781e-02 -1.70641258e-01
-5.13764560e-01 -7.51255691e-01 -3.23448479e-01 1.81652677e+00
3.89372438e-01 -6.89775586e-01 -1.09192646e+00 -4.78119135e-01
2.09721029e-01 1.10346365e+00 -4.72469963e-02 4.27033752e-02
1.10589392e-01 -6.19393110e-01 2.35983610e-01 2.08027318e-01
4.93707597e-01 -1.12738776e+00 -5.71766436e-01 7.22710937e-02
-5.11442304e-01 -6.37414455e-01 -5.45430243e-01 -2.59411663e-01
-7.17896879e-01 -8.94838214e-01 -7.85372794e-01 -4.96203393e-01
4.02423739e-01 8.83947730e-01 1.18708944e+00 5.40994167e-01
-1.48876488e-01 1.42995870e+00 -2.73235440e-01 -3.61538619e-01
-3.22894484e-01 3.73295605e-01 3.53478789e-01 -5.95362127e-01
5.52284896e-01 -1.39344305e-01 -8.78793836e-01 9.41577196e-01
-6.05753481e-01 -1.60318166e-01 3.06071907e-01 6.75188541e-01
-1.57690316e-01 -2.49926280e-03 4.91674542e-01 -1.31197047e+00
1.66735256e+00 -8.31995070e-01 -1.45130351e-01 5.68977118e-01
-1.31297803e+00 -3.13604236e-01 3.01191300e-01 -2.22353920e-01
-1.49843800e+00 -4.36077088e-01 -3.37745994e-01 4.81647253e-01
-2.54055530e-01 6.79785430e-01 -1.23903766e-01 -1.13579750e-01
1.08912265e+00 9.65123624e-02 3.03226888e-01 -5.51846981e-01
5.23911119e-01 1.42395854e+00 -7.34395608e-02 -5.73492229e-01
2.22695097e-01 -2.67922133e-01 -8.01559865e-01 -1.24546659e+00
-1.74936503e-01 -6.46834612e-01 -9.85723455e-03 -4.14397806e-01
4.03597116e-01 -3.43016922e-01 -1.20565999e+00 2.03410715e-01
-7.68557549e-01 -5.30898511e-01 6.85467646e-02 3.96746963e-01
-5.64618826e-01 2.32729599e-01 -7.37797976e-01 -1.47941661e+00
-4.64351237e-01 -1.23775351e+00 3.80639255e-01 6.85415268e-01
-7.47070193e-01 -1.32101631e+00 -2.02420875e-02 5.90365469e-01
9.43391502e-01 -8.89403880e-01 1.04447174e+00 -8.41517329e-01
-2.93200076e-01 -4.23643768e-01 -1.68223586e-02 -2.19593465e-01
1.87936217e-01 -4.98889498e-02 -8.54848206e-01 -3.87116849e-01
-5.12001812e-02 1.29168143e-03 -8.25600252e-02 4.11428481e-01
1.06084955e+00 -3.93041164e-01 -7.94821560e-01 -6.42655939e-02
8.21308374e-01 7.83410311e-01 3.84582400e-01 1.67379543e-01
7.67356455e-02 8.31391752e-01 6.78871155e-01 7.86793172e-01
6.17262602e-01 8.83540034e-01 -2.00298160e-01 -4.93599623e-02
3.70177239e-01 -2.43136019e-01 1.15186214e-01 4.05604869e-01
-4.42246497e-01 -6.85505807e-01 -5.97592950e-01 -7.18339831e-02
-1.94046915e+00 -9.86966372e-01 1.72110170e-01 2.75868344e+00
6.54858947e-01 -2.84337759e-01 5.92409015e-01 -4.02921438e-02
5.61215937e-01 -3.74533147e-01 -7.05815434e-01 -3.40054810e-01
5.00584722e-01 -7.27396756e-02 2.66177297e-01 9.35343921e-01
-6.24746680e-01 7.65123248e-01 8.00780487e+00 4.40613687e-01
-4.89910990e-01 -1.34561330e-01 6.20272934e-01 -5.55875674e-02
-3.54735732e-01 -3.23616415e-01 -8.30194652e-01 5.65974593e-01
1.20567787e+00 -6.98963940e-01 1.03480625e+00 9.04947996e-01
4.43995893e-01 -3.95669818e-01 -1.14855504e+00 9.06794369e-01
-1.72626808e-01 -8.29847336e-01 -2.22506136e-01 2.09485933e-01
3.56834322e-01 -2.75443345e-01 2.54386425e-01 7.75478840e-01
5.75158119e-01 -7.70596087e-01 -2.84542479e-02 6.68923557e-01
5.58746219e-01 -4.37075406e-01 4.35943276e-01 7.75621831e-01
-7.85670578e-01 -1.84273496e-01 -5.29485941e-02 -3.90104741e-01
2.27104098e-01 -4.06349264e-02 -4.82059002e-01 3.27226855e-02
6.99612439e-01 1.27169386e-01 -1.55206099e-01 1.24834383e+00
3.94650906e-01 3.69091094e-01 -3.83711159e-01 -7.02162862e-01
-3.23947579e-01 -2.74322748e-01 4.58669662e-01 1.10409057e+00
7.54036009e-02 7.34046400e-01 1.88661814e-01 5.55801630e-01
2.42370829e-01 4.40630227e-01 -7.79519916e-01 -1.39082789e-01
8.55764031e-01 1.36481011e+00 -3.06847721e-01 -3.40354145e-01
-6.35448158e-01 8.53556037e-01 1.37350364e-02 6.04692042e-01
-5.21450877e-01 -5.83095968e-01 9.64963794e-01 3.91262054e-01
-6.80892825e-01 -8.98222625e-02 -2.85747766e-01 -9.42987323e-01
-3.71933043e-01 -1.16084528e+00 4.23531085e-01 -5.05819023e-01
-1.00720167e+00 5.04121006e-01 2.58777201e-01 -6.75331891e-01
-5.87842584e-01 -3.66112113e-01 -4.37751681e-01 1.09910250e+00
-9.04435992e-01 -3.07395488e-01 -4.57516044e-01 4.21275437e-01
6.06051087e-01 -2.59403229e-01 1.19815874e+00 6.17714524e-01
-6.66695178e-01 9.93481934e-01 5.70649385e-01 -7.09526896e-01
8.03950310e-01 -1.21869981e+00 2.61048168e-01 -9.37703922e-02
-3.07084084e-01 1.38210821e+00 7.41927028e-01 -3.45648140e-01
-1.66199636e+00 -2.83481807e-01 7.50487685e-01 -8.60488951e-01
9.05898027e-03 -1.35713264e-01 -6.93349063e-01 4.38132405e-01
3.39763045e-01 -1.10440159e+00 1.10803890e+00 7.95571804e-01
3.97267282e-01 2.80903876e-01 -1.37773824e+00 1.04452598e+00
1.15492845e+00 -4.27827597e-01 4.10582572e-02 5.05204201e-01
4.95304137e-01 -4.35201645e-01 -9.63802636e-01 -7.54092447e-03
1.07434762e+00 -9.07806277e-01 9.08020020e-01 -7.63724089e-01
-1.44898698e-01 2.47753009e-01 -4.55358177e-02 -1.32796252e+00
-5.56834161e-01 -1.03066897e+00 -3.29114616e-01 9.08999085e-01
6.92333817e-01 -8.15511763e-01 7.15067029e-01 2.04599524e+00
1.17283918e-01 -8.02006066e-01 1.57393843e-01 -3.38320196e-01
-1.35609642e-01 -1.17358141e-01 4.45653558e-01 7.24861145e-01
9.13699269e-01 7.18391299e-01 -3.23194742e-01 -3.17412764e-01
2.75774866e-01 -3.96966130e-01 8.24375749e-01 -1.42613649e+00
-3.73823285e-01 -7.81301379e-01 3.38478565e-01 -1.78352845e+00
-2.94555098e-01 -4.78820682e-01 -3.50160233e-04 -1.61340690e+00
2.14037731e-01 -5.39987326e-01 -2.06839457e-01 -1.20862508e-02
-3.13750893e-01 -3.20320070e-01 4.09105327e-03 3.85262698e-01
-8.23815882e-01 3.96682262e-01 1.10981894e+00 2.19456747e-01
-7.31411397e-01 1.00067842e+00 -1.13921916e+00 4.97150242e-01
5.87531328e-01 3.17595363e-01 -8.67261231e-01 -8.96923989e-02
1.64831474e-01 5.30896962e-01 -4.70454931e-01 -5.58570325e-01
5.51112652e-01 -3.09853226e-01 1.29557371e-01 -1.82428852e-01
2.38153815e-01 -9.75128770e-01 1.37188407e-02 2.10302532e-01
-7.39397109e-01 1.81870475e-01 1.71037644e-01 3.11594754e-01
3.91599029e-01 -3.99343729e-01 3.22221905e-01 -3.72478692e-03
-4.65636581e-01 2.20299199e-01 -1.22205126e+00 -2.48009235e-01
7.76032150e-01 -1.28924355e-01 -1.80031702e-01 -1.18794060e+00
-8.69443893e-01 7.54839003e-01 2.85218924e-01 5.20803869e-01
2.94801742e-01 -9.07742441e-01 -2.45195553e-02 2.06942514e-01
5.70683740e-02 -7.75837600e-01 3.60506803e-01 3.99181336e-01
-2.04661012e-01 8.29506576e-01 -2.79189739e-02 -4.97884750e-02
-1.59800553e+00 2.58556962e-01 3.58785987e-01 -2.34886482e-02
2.18808800e-02 4.59860742e-01 -2.52666268e-02 -6.15122557e-01
9.25264597e-01 2.52088249e-01 -7.07331121e-01 -2.75633633e-01
7.66160011e-01 7.82500863e-01 -1.53857380e-01 -1.04634680e-01
7.00931177e-02 -1.81495219e-01 -5.86935639e-01 -3.22353005e-01
6.00896358e-01 -5.66345811e-01 1.38080046e-01 2.99477160e-01
5.30580878e-01 -2.33962253e-01 -4.78963733e-01 -3.79977137e-01
-1.91588074e-01 -5.17011344e-01 5.28679267e-02 -1.28198206e+00
-5.30227363e-01 3.90221685e-01 7.30702221e-01 8.00516248e-01
8.48482490e-01 -2.35299870e-01 5.44395149e-01 1.04281747e+00
4.44692999e-01 -1.20828664e+00 -1.95307627e-01 4.82882470e-01
6.67288363e-01 -1.38541198e+00 -1.60023883e-01 -2.95633525e-01
-7.18003213e-01 6.64194286e-01 8.49598050e-01 6.48875773e-01
1.25152516e+00 -6.13420568e-02 9.86415669e-02 -5.88773713e-02
-8.43151450e-01 -2.61581302e-01 2.68238097e-01 5.86711943e-01
1.28716183e+00 8.41345787e-02 -7.19172001e-01 8.71650696e-01
-1.91857770e-01 -1.73765235e-02 2.37478256e-01 9.56422687e-01
-6.20373487e-01 -1.49682474e+00 -9.30447280e-02 9.64422405e-01
-2.92524546e-01 -4.29366007e-02 -5.48493445e-01 5.02287865e-01
-8.54827523e-01 1.53492248e+00 -2.40264699e-01 -5.68349838e-01
7.25650489e-01 1.68391734e-01 2.48919830e-01 -5.27690828e-01
-8.69105697e-01 -2.46018961e-01 3.84309769e-01 -4.66244310e-01
1.45199582e-01 -5.57023048e-01 -8.27480614e-01 -8.79232705e-01
-5.89165509e-01 7.44226873e-01 8.55056942e-01 6.83236837e-01
5.99291623e-01 3.35650802e-01 7.10181475e-01 -6.83284998e-01
-8.72583807e-01 -1.08996773e+00 -5.69233716e-01 1.88305214e-01
-1.13901414e-01 -7.31524289e-01 -1.55020505e-01 -4.73484546e-01] | [12.271880149841309, 7.6825432777404785] |
6f38bd60-ae7c-4f54-ade0-21847451c5e9 | stable-and-compact-face-recognition-via | 2111.02847 | null | https://arxiv.org/abs/2111.02847v1 | https://arxiv.org/pdf/2111.02847v1.pdf | Stable and Compact Face Recognition via Unlabeled Data Driven Sparse Representation-Based Classification | Sparse representation-based classification (SRC) has attracted much attention by casting the recognition problem as simple linear regression problem. SRC methods, however, still is limited to enough labeled samples per category, insufficient use of unlabeled samples, and instability of representation. For tackling these problems, an unlabeled data driven inverse projection pseudo-full-space representation-based classification model is proposed with low-rank sparse constraints. The proposed model aims to mine the hidden semantic information and intrinsic structure information of all available data, which is suitable for few labeled samples and proportion imbalance between labeled samples and unlabeled samples problems in frontal face recognition. The mixed Gauss-Seidel and Jacobian ADMM algorithm is introduced to solve the model. The convergence, representation capability and stability of the model are analyzed. Experiments on three public datasets show that the proposed LR-S-PFSRC model achieves stable results, especially for proportion imbalance of samples. | ['Haolin Chen', 'Yiming Xu', 'Licheng Jiao', 'Huan Wu', 'Zheng Wang', 'XiaoHui Yang'] | 2021-11-04 | null | null | null | null | ['sparse-representation-based-classification'] | ['computer-vision'] | [ 1.88426912e-01 -1.55573264e-01 -4.94098634e-01 -6.21935368e-01
-8.70955944e-01 7.31580928e-02 2.91703820e-01 -8.01464915e-01
2.40607131e-02 8.59785557e-01 3.92003566e-01 2.51892328e-01
-3.16077858e-01 -3.53374749e-01 -2.96161026e-01 -9.14008319e-01
3.76928598e-01 3.86973083e-01 -6.13273919e-01 -1.85706411e-02
1.45250380e-01 4.10068870e-01 -1.55951571e+00 3.47307026e-01
9.10970211e-01 1.02508950e+00 2.86668777e-01 -1.25801280e-01
-2.65998155e-01 1.13881385e+00 -1.32007450e-01 1.19014382e-01
3.60669792e-01 -5.34785509e-01 -3.58049333e-01 5.47486722e-01
4.69891489e-01 2.68373992e-02 -6.78065717e-01 1.16505015e+00
3.44243795e-01 2.80532569e-01 8.73503268e-01 -1.24005914e+00
-9.79746759e-01 2.76593298e-01 -1.02598739e+00 2.57877469e-01
1.58742502e-01 -3.05628151e-01 5.64952731e-01 -1.57370412e+00
5.47543824e-01 1.43698692e+00 6.40243173e-01 6.02209032e-01
-1.18794382e+00 -9.40470815e-01 3.00339162e-01 4.11929339e-01
-1.71866572e+00 -4.88186061e-01 1.05465734e+00 -5.88254333e-01
3.47962648e-01 3.87164295e-01 4.18573707e-01 1.10171211e+00
-7.95577392e-02 8.51025164e-01 1.35107839e+00 -1.66168466e-01
3.25608313e-01 3.34812880e-01 5.50887406e-01 5.03702641e-01
4.98028368e-01 3.92999351e-02 -6.94642901e-01 -1.69649512e-01
6.10202849e-01 5.23379147e-01 -2.88670272e-01 -4.26515073e-01
-8.29237521e-01 1.07588124e+00 2.28606567e-01 2.25316569e-01
-5.15666068e-01 -4.91117656e-01 -1.21692140e-02 1.23813547e-01
5.61404586e-01 -1.21830948e-01 -2.49249324e-01 3.08828264e-01
-1.25298131e+00 -7.27961734e-02 5.24936974e-01 9.35134113e-01
8.18441629e-01 7.21036732e-01 -2.81797163e-02 1.24743593e+00
6.37321234e-01 9.06852424e-01 9.13282037e-01 -9.29442763e-01
5.47389507e-01 7.43820250e-01 -2.62109280e-01 -1.53483653e+00
-1.66291729e-01 -8.08009982e-01 -1.36592162e+00 -1.15543440e-01
1.03361286e-01 -1.09713815e-01 -8.14633012e-01 1.55743909e+00
4.08797294e-01 5.45418859e-01 1.79285824e-01 1.31712651e+00
8.80272985e-01 8.97068739e-01 -2.56438166e-01 -7.48794913e-01
1.11973131e+00 -8.08515251e-01 -9.67753887e-01 -3.12546492e-01
2.75739759e-01 -6.56753898e-01 6.92765713e-01 5.00275731e-01
-7.35113323e-01 -5.96821368e-01 -8.64113986e-01 9.17961150e-02
6.75987825e-02 6.96737945e-01 6.68604732e-01 5.56928456e-01
-4.84510839e-01 1.44585714e-01 -4.84190762e-01 4.94972132e-02
6.42080843e-01 4.89587426e-01 -7.48488903e-01 -7.98247218e-01
-8.15527618e-01 5.17328143e-01 3.42216552e-03 7.28663385e-01
-7.85208702e-01 -6.60867989e-01 -8.86081398e-01 -1.46283165e-01
2.63765275e-01 -2.74173856e-01 5.00226080e-01 -1.08348036e+00
-1.33857191e+00 5.67901194e-01 -5.23515642e-01 3.39494571e-02
1.39647156e-01 -1.47669554e-01 -4.87933248e-01 1.43961550e-03
1.03039749e-01 2.00364754e-01 1.30162001e+00 -1.23606479e+00
-1.78952366e-01 -1.02153695e+00 -7.59797573e-01 2.50726610e-01
-4.38299775e-01 -1.16553798e-01 -1.24573641e-01 -7.57253170e-01
8.13688815e-01 -8.81039679e-01 -3.90502721e-01 -1.73616797e-01
-1.66678324e-01 8.24626461e-02 1.18563521e+00 -8.74066114e-01
8.57560456e-01 -2.21665263e+00 3.43477458e-01 2.98963696e-01
4.14090268e-02 1.60856843e-01 -3.24559391e-01 -8.53913277e-02
-5.27096808e-01 -4.33047891e-01 -4.03666079e-01 -2.31294289e-01
-4.36524063e-01 2.14360654e-01 -4.36490476e-01 9.18683827e-01
1.35588884e-01 5.58482409e-01 -6.38300359e-01 -4.59728181e-01
1.36044234e-01 7.96941936e-01 -4.79929388e-01 2.13725343e-01
3.44249874e-01 6.95386291e-01 -6.67718828e-01 1.03077900e+00
1.16882145e+00 -2.75467724e-01 1.11219332e-01 -5.10303736e-01
2.02543348e-01 -3.33565891e-01 -1.73776817e+00 1.73673022e+00
-3.82914633e-01 2.73431689e-01 3.22864056e-01 -1.32294595e+00
1.14512777e+00 1.46125913e-01 6.59126639e-01 -6.47713363e-01
9.24565420e-02 2.58764029e-01 -2.56260902e-01 -3.99986148e-01
9.85583290e-02 -3.24109465e-01 3.70605975e-01 2.92597979e-01
8.40962380e-02 2.13051915e-01 -1.10467665e-01 7.73690939e-02
6.50856972e-01 4.00766544e-02 8.41080025e-02 -2.35244453e-01
9.99919474e-01 -6.73185140e-02 1.13892651e+00 1.47435203e-01
8.09950531e-02 8.12162697e-01 5.60159683e-02 -3.86946589e-01
-6.93648517e-01 -6.24462247e-01 -3.40729535e-01 6.06692612e-01
6.78060576e-02 -4.97337244e-02 -3.91274631e-01 -6.43020391e-01
3.38002928e-02 4.30912495e-01 -5.96679151e-01 -3.20408016e-01
-5.37330270e-01 -1.01037061e+00 -1.02464885e-01 1.93223611e-01
4.23499852e-01 -8.58217001e-01 2.23010153e-01 -1.33954706e-02
-1.67473242e-01 -9.08687770e-01 -3.54348242e-01 -2.55578756e-01
-1.11169529e+00 -1.19058895e+00 -9.61858094e-01 -8.79006207e-01
1.28240240e+00 7.23876774e-01 4.55935746e-01 -3.37881148e-01
-4.85458255e-01 1.54490203e-01 -2.03050286e-01 1.96993768e-01
1.66166082e-01 -3.81847739e-01 3.43971044e-01 7.61870265e-01
3.99097413e-01 -4.77236927e-01 -3.93675238e-01 6.07090533e-01
-6.24131620e-01 -9.55343619e-02 6.60602450e-01 1.23680580e+00
8.32298815e-01 -1.17079029e-02 7.79317021e-01 -1.02095997e+00
3.44910413e-01 -8.51045072e-01 -5.24033070e-01 8.23337957e-02
-7.19150662e-01 -1.43886194e-01 6.67698026e-01 -6.36032939e-01
-1.23578465e+00 3.51228327e-01 2.00843096e-01 -1.09415007e+00
2.38594770e-01 7.14908838e-01 -4.69854116e-01 -2.06555873e-01
4.97166783e-01 7.36816466e-01 4.68100369e-01 -8.19203079e-01
3.01782012e-01 8.35857272e-01 3.34130019e-01 -4.96562570e-01
8.54925811e-01 5.24186790e-01 4.69667576e-02 -1.01222789e+00
-1.06433451e+00 -6.16228640e-01 -4.32043970e-01 8.68330821e-02
2.70496249e-01 -1.40207040e+00 -3.54844838e-01 2.46166393e-01
-6.47407174e-01 4.18601692e-01 -4.70827222e-01 8.40836823e-01
-3.67437541e-01 5.12871921e-01 -5.14170945e-01 -1.02503562e+00
-2.86402166e-01 -1.17979336e+00 8.48446071e-01 1.37717262e-01
2.76213825e-01 -4.87382233e-01 -2.12848373e-03 8.93987656e-01
2.70265549e-01 5.70227653e-02 4.73516017e-01 -5.24391711e-01
-3.88513625e-01 -4.17574763e-01 -2.87672013e-01 7.13076234e-01
2.01033622e-01 -5.07870853e-01 -8.03672075e-01 -5.38303912e-01
7.30834484e-01 -4.53297645e-01 7.84938514e-01 2.83849508e-01
1.34419799e+00 -7.20349312e-01 -1.82171017e-01 7.58415937e-01
1.42764890e+00 5.99898770e-02 5.39066970e-01 -2.62026101e-01
9.76862729e-01 7.17716455e-01 8.61960590e-01 5.50579846e-01
7.09552169e-02 4.40604031e-01 7.97573179e-02 8.76301080e-02
-1.28397360e-01 -1.91930085e-01 4.02793080e-01 1.27052736e+00
1.51748940e-01 3.46018702e-01 -5.06268084e-01 2.47171089e-01
-1.85573399e+00 -1.16616917e+00 -2.49808893e-01 2.00895786e+00
3.82842481e-01 -5.96040905e-01 -1.78613827e-01 3.86070251e-01
9.25051093e-01 1.24785237e-01 -8.34453583e-01 6.40759319e-02
-3.48791510e-01 -1.30945612e-02 2.12314203e-01 2.27816001e-01
-7.06666291e-01 6.93173647e-01 6.27858686e+00 1.05912447e+00
-1.22977185e+00 3.77060533e-01 8.10411930e-01 -5.99489808e-02
-1.73421696e-01 -2.20243931e-02 -9.82937157e-01 6.65447354e-01
7.00666189e-01 3.55717316e-02 6.10891581e-01 1.09076655e+00
4.16311055e-01 1.92257836e-01 -7.86573529e-01 1.70351720e+00
6.31788135e-01 -1.17227709e+00 3.04783046e-01 1.62119284e-01
9.68012094e-01 -1.77968785e-01 2.52282351e-01 4.00244623e-01
-1.04732089e-01 -1.21267653e+00 3.80366027e-01 6.53697133e-01
9.39399660e-01 -5.56588948e-01 5.87222338e-01 4.67326641e-01
-9.12342668e-01 -4.57264692e-01 -7.65071869e-01 2.48205271e-02
-1.43518314e-01 8.84070992e-01 -1.98930725e-01 5.10962605e-01
1.73238114e-01 1.30486095e+00 -3.37685764e-01 6.98949754e-01
9.96596292e-02 6.98252022e-01 -2.44126037e-01 5.15793145e-01
-9.66035202e-02 -9.62007105e-01 4.35938448e-01 7.13462353e-01
4.08896714e-01 4.33289409e-01 2.99746335e-01 8.60495746e-01
5.94512559e-02 3.85476977e-01 -6.70372069e-01 -5.83455488e-02
3.03418100e-01 1.38668501e+00 -2.50143349e-01 -7.36159086e-02
-4.62641716e-01 6.99722230e-01 2.72504240e-01 5.19141316e-01
-5.34049749e-01 2.06927940e-01 4.83053267e-01 7.38165900e-02
2.71185726e-01 -2.25075692e-01 -4.16491926e-01 -1.60479164e+00
2.98672188e-02 -1.25120676e+00 5.04335225e-01 -6.07213378e-01
-1.34430289e+00 3.96564633e-01 -2.91722834e-01 -1.43752146e+00
2.52293982e-02 -4.63568628e-01 -3.58452439e-01 9.33347225e-01
-1.58853841e+00 -1.18915820e+00 -3.06541204e-01 9.16205645e-01
7.44412899e-01 -9.16627765e-01 5.55191636e-01 6.19951487e-01
-1.07534719e+00 5.11023521e-01 4.03009653e-01 -3.82039174e-02
3.21919799e-01 -4.83411580e-01 -7.50848770e-01 8.34261417e-01
7.47256428e-02 8.47006977e-01 4.35318887e-01 -5.90215683e-01
-2.14751983e+00 -1.15376091e+00 5.78057468e-01 4.47117910e-02
3.39284927e-01 -4.63259518e-02 -9.25387383e-01 5.64395070e-01
-4.75180566e-01 6.93273544e-01 9.54481065e-01 1.02309927e-01
-5.02040684e-01 -4.66037393e-01 -1.28509533e+00 1.13412656e-01
8.55103195e-01 -4.67473328e-01 -5.86367428e-01 6.01793170e-01
2.08016962e-01 -1.20648466e-01 -7.69230783e-01 3.63005996e-01
4.43719804e-01 -5.44083297e-01 1.19725609e+00 -5.79248726e-01
3.01952899e-01 -4.68662649e-01 -6.04625583e-01 -1.04009914e+00
-5.86402416e-01 -1.44194394e-01 -3.56062531e-01 1.11155224e+00
1.49441779e-01 -6.15981579e-01 1.15033948e+00 5.68240583e-01
8.63284096e-02 -7.03822494e-01 -1.07643759e+00 -7.43973255e-01
-2.74470091e-01 -2.05461264e-01 2.63805330e-01 1.25869119e+00
-1.79963470e-01 5.37518322e-01 -7.88981497e-01 1.02557987e-01
1.00997651e+00 4.52716887e-01 5.27873397e-01 -1.26905608e+00
-1.00372948e-01 1.12641752e-01 -4.65507805e-01 -7.75364757e-01
6.28130019e-01 -1.18501210e+00 -3.14429551e-01 -1.32529211e+00
4.89991248e-01 -5.95512450e-01 -3.86289567e-01 4.80018944e-01
-6.95642922e-03 4.13189173e-01 1.25206009e-01 6.81364179e-01
-3.37283611e-01 1.04787910e+00 1.21598375e+00 -4.94424403e-01
-4.90118237e-03 -1.14713386e-01 -8.12497556e-01 5.39610326e-01
3.55565012e-01 -4.65564877e-01 -6.87398493e-01 -2.60247886e-01
-1.77022994e-01 2.63694584e-01 9.05692950e-03 -9.94634449e-01
2.24939629e-01 -4.61445689e-01 7.04953551e-01 -5.68036079e-01
7.01300323e-01 -1.01142907e+00 5.21077931e-01 4.33606684e-01
-2.83351302e-01 -3.86994332e-01 -3.15490693e-01 7.94503927e-01
-5.02886355e-01 -1.55159965e-01 8.84103179e-01 -3.07234935e-02
-4.68981862e-01 6.46389902e-01 -6.55848607e-02 7.30774477e-02
9.83434081e-01 -4.94683415e-01 8.02862868e-02 -1.42220601e-01
-8.79177511e-01 9.09559950e-02 2.45200172e-02 3.74362916e-01
1.06498146e+00 -1.68398023e+00 -9.71235573e-01 6.38232887e-01
-2.49561686e-02 -4.86094169e-02 6.23588085e-01 8.52848947e-01
-9.24431160e-02 4.91808057e-01 -2.20664456e-01 -4.87951964e-01
-1.14971089e+00 6.93053186e-01 6.86237141e-02 8.47097710e-02
-6.16341650e-01 6.39107108e-01 1.83379203e-01 -6.28977478e-01
-9.21611022e-03 2.99089879e-01 -5.76438546e-01 3.60255718e-01
7.35681951e-01 5.91145694e-01 -4.05328274e-02 -1.26197028e+00
-3.98504853e-01 7.82976210e-01 -1.23502396e-01 2.74316847e-01
1.58896530e+00 -1.50070518e-01 -2.81040519e-01 3.90713185e-01
1.55376148e+00 -1.78116858e-01 -1.00829375e+00 -3.75152588e-01
-3.16966325e-01 -9.20901895e-01 4.49355662e-01 -3.96952093e-01
-1.58167386e+00 7.84637451e-01 8.18136990e-01 -6.14413977e-01
8.42487156e-01 -6.19406879e-01 6.08294904e-01 3.50508481e-01
5.40320277e-01 -9.56287205e-01 3.19856517e-02 3.75983894e-01
1.10207033e+00 -1.42501128e+00 3.67008835e-01 -6.82430327e-01
-5.86595893e-01 9.00040686e-01 7.48690009e-01 -3.34257096e-01
9.05623436e-01 -1.65516749e-01 -2.40678087e-01 -1.00426845e-01
-5.67858219e-01 2.96459943e-01 3.85992616e-01 5.54357350e-01
1.87935129e-01 -9.89051387e-02 -4.95770842e-01 7.21852005e-01
7.47113898e-02 2.25424513e-01 2.68894345e-01 8.54059398e-01
-2.28161767e-01 -9.16227221e-01 -8.13271880e-01 7.35675514e-01
-2.96610087e-01 1.53804734e-01 1.76298025e-04 1.69499129e-01
-9.70826745e-02 9.59136605e-01 -2.76496321e-01 -2.23855183e-01
4.00362425e-02 -8.72587413e-02 2.82075465e-01 -6.92768455e-01
3.88048813e-02 9.43189561e-02 -4.17903274e-01 -5.56420267e-01
-3.01751226e-01 -7.91010141e-01 -9.95590448e-01 8.68205503e-02
-4.96058524e-01 5.23910463e-01 7.97134459e-01 7.49255836e-01
5.46906888e-01 4.24207076e-02 1.17378390e+00 -7.67817616e-01
-9.36258137e-01 -9.03310537e-01 -1.09198427e+00 3.58751148e-01
1.20955117e-01 -8.89842153e-01 -6.50407970e-01 -6.37530684e-02] | [12.473637580871582, 0.387382835149765] |
0dcf471a-f6f5-42b5-a4d9-b457de8290dc | low-resource-speech-to-text-translation | 1803.09164 | null | http://arxiv.org/abs/1803.09164v2 | http://arxiv.org/pdf/1803.09164v2.pdf | Low-Resource Speech-to-Text Translation | Speech-to-text translation has many potential applications for low-resource
languages, but the typical approach of cascading speech recognition with
machine translation is often impossible, since the transcripts needed to train
a speech recognizer are usually not available for low-resource languages.
Recent work has found that neural encoder-decoder models can learn to directly
translate foreign speech in high-resource scenarios, without the need for
intermediate transcription. We investigate whether this approach also works in
settings where both data and computation are limited. To make the approach
efficient, we make several architectural changes, including a change from
character-level to word-level decoding. We find that this choice yields crucial
speed improvements that allow us to train with fewer computational resources,
yet still performs well on frequent words. We explore models trained on between
20 and 160 hours of data, and find that although models trained on less data
have considerably lower BLEU scores, they can still predict words with
relatively high precision and recall---around 50% for a model trained on 50
hours of data, versus around 60% for the full 160 hour model. Thus, they may
still be useful for some low-resource scenarios. | ['Adam Lopez', 'Karen Livescu', 'Sameer Bansal', 'Sharon Goldwater', 'Herman Kamper'] | 2018-03-24 | null | null | null | null | ['speech-to-text-translation'] | ['natural-language-processing'] | [ 1.86595857e-01 8.91130865e-02 -3.08901846e-01 -3.02334189e-01
-1.48011363e+00 -5.68738759e-01 4.47916925e-01 -5.60656637e-02
-5.66074073e-01 8.70600462e-01 2.35740736e-01 -1.02266896e+00
5.83812296e-01 -4.72451389e-01 -5.68907917e-01 -4.55076724e-01
2.75567144e-01 5.26955009e-01 8.26421902e-02 -3.20402682e-01
-2.98830301e-01 3.08702707e-01 -1.19334781e+00 3.63316983e-01
8.45360637e-01 4.64501858e-01 4.08432692e-01 8.42642784e-01
-1.60013720e-01 4.71332848e-01 -7.44215965e-01 -5.29459417e-01
2.79227436e-01 -6.06574714e-01 -7.44907081e-01 8.31277817e-02
1.27150849e-01 -3.39470863e-01 -4.30863559e-01 6.19745255e-01
6.53137326e-01 2.60480470e-03 3.90308052e-01 -5.92250347e-01
-5.71692050e-01 5.17514288e-01 -2.07083791e-01 3.20460320e-01
4.28246081e-01 2.71789968e-01 8.65216374e-01 -1.04309261e+00
3.40011865e-01 1.10731184e+00 4.79130238e-01 5.99114954e-01
-1.05572283e+00 -4.53862995e-01 5.55715188e-02 -1.27067432e-01
-1.41908431e+00 -1.14481819e+00 1.02275930e-01 2.39549316e-02
1.67024112e+00 2.88609654e-01 3.97240281e-01 1.21200192e+00
2.72252947e-01 6.05892479e-01 8.42908740e-01 -5.67257345e-01
8.50539356e-02 2.38081679e-01 -3.35490704e-01 4.78493720e-01
-4.66897115e-02 -1.28920078e-01 -7.09721208e-01 -2.62742005e-02
4.17704552e-01 -1.16691425e-01 -2.25555032e-01 3.44513208e-01
-1.38837445e+00 6.90728307e-01 -7.56495893e-02 5.49839139e-01
-2.85557032e-01 -1.44076645e-01 3.94465983e-01 8.43454599e-01
9.03260887e-01 3.54672968e-01 -5.48069775e-01 -6.75939381e-01
-1.22222984e+00 -8.92696306e-02 1.13236856e+00 1.15967083e+00
6.94807768e-01 3.06860447e-01 -1.88305527e-02 1.01958203e+00
-5.64056411e-02 7.58304834e-01 6.68680072e-01 -4.15328443e-01
1.01501977e+00 -6.43390939e-02 1.59167107e-02 -4.70099241e-01
-1.68825805e-01 -4.45382148e-01 -7.24922776e-01 -4.83315766e-01
3.69383067e-01 -6.18180215e-01 -1.02519226e+00 1.48081517e+00
-1.46992663e-02 -8.51139948e-02 2.62640148e-01 8.10469329e-01
2.35829279e-01 1.18898487e+00 -1.87650517e-01 -6.11014545e-01
1.20742142e+00 -1.25630844e+00 -6.44265592e-01 -8.53575706e-01
9.35226858e-01 -1.02134466e+00 1.33943653e+00 1.38810575e-01
-1.31281197e+00 -3.43453884e-01 -9.40826952e-01 1.44110797e-02
-1.61377251e-01 1.65026218e-01 4.38323557e-01 7.38415182e-01
-1.36274540e+00 3.07897896e-01 -8.96600008e-01 -6.03906035e-01
-7.97861516e-02 3.08929712e-01 -2.77997851e-01 -2.24478409e-01
-1.11834073e+00 1.22618139e+00 3.77305672e-02 7.93909058e-02
-8.47259939e-01 -2.56752551e-01 -6.37373209e-01 1.64918229e-01
2.71958739e-01 -5.08293331e-01 1.63944185e+00 -1.08025730e+00
-1.86325574e+00 4.77329701e-01 -6.97949290e-01 -4.37254757e-01
3.72530609e-01 -9.27945152e-02 -5.12040675e-01 -6.09978959e-02
-1.39938846e-01 4.40372109e-01 6.86824560e-01 -6.03658020e-01
-6.03706360e-01 -3.86980455e-03 -1.30462497e-01 4.20911998e-01
-6.57668829e-01 4.31317419e-01 -6.39150143e-01 -5.28522730e-01
-1.31894592e-02 -1.03927839e+00 -1.61169961e-01 -4.42992061e-01
-9.74709447e-03 -1.12672567e-01 4.46991384e-01 -9.70605552e-01
1.33098316e+00 -1.91034710e+00 -4.36014570e-02 -6.52552322e-02
-3.38396579e-01 5.03618479e-01 -2.90099025e-01 7.26483405e-01
2.18842044e-01 2.12628394e-01 -2.33763218e-01 -4.84872729e-01
-3.90521109e-01 2.89371341e-01 -3.82223189e-01 1.69470206e-01
3.30022752e-01 9.26970184e-01 -6.21707261e-01 -1.69198394e-01
-1.30282834e-01 3.97441238e-01 -4.24710274e-01 2.26728484e-01
-1.46734193e-01 3.33047420e-01 -1.20027430e-01 6.03725374e-01
2.47327566e-01 -2.11525440e-01 3.71796697e-01 5.77161610e-01
-2.91656274e-02 1.04850650e+00 -6.37154996e-01 1.59173441e+00
-1.10221016e+00 9.08804595e-01 1.05926223e-01 -8.52730632e-01
9.08381402e-01 6.95369303e-01 2.24506017e-02 -9.66812015e-01
-1.93858728e-01 5.70826471e-01 2.21962526e-01 -3.77931327e-01
4.01173830e-01 -4.30353701e-01 -7.83447083e-03 6.01765573e-01
-6.29753321e-02 -1.94090202e-01 1.15612611e-01 -7.88470730e-02
1.31786275e+00 -2.65702248e-01 7.03709498e-02 -3.44136432e-02
2.48580143e-01 -2.05565151e-03 4.01951671e-01 5.94979465e-01
5.04567884e-02 5.09556293e-01 1.66839957e-01 -3.18231612e-01
-1.38973713e+00 -6.96139634e-01 1.73827127e-01 1.27304649e+00
-2.88854420e-01 -4.01244462e-01 -9.31467414e-01 -5.13185501e-01
-4.18269038e-01 7.70990610e-01 2.88552701e-01 -1.46172673e-01
-7.68549144e-01 -8.26717019e-01 7.50670075e-01 3.65005612e-01
1.06740415e-01 -9.03717220e-01 -1.42931700e-01 4.67007101e-01
-3.50354940e-01 -1.34313369e+00 -5.85829139e-01 3.12481135e-01
-8.57919335e-01 -1.09883398e-01 -9.73044097e-01 -9.02853429e-01
5.82342684e-01 4.49198544e-01 1.17036319e+00 1.93449780e-01
2.71157205e-01 -1.17512241e-01 -4.38215107e-01 -2.04250202e-01
-8.28239739e-01 6.78194165e-01 2.95579821e-01 -1.51545763e-01
4.73314673e-01 -4.79085535e-01 -2.16565534e-01 3.80920976e-01
-6.65766776e-01 1.27727732e-01 1.03941929e+00 8.35118651e-01
1.40072852e-01 -4.09876794e-01 8.74148130e-01 -5.24206758e-01
6.63380146e-01 -4.35433239e-01 -3.38042289e-01 3.27602297e-01
-7.86883652e-01 1.34392187e-01 9.67925489e-01 -5.04428625e-01
-8.79938006e-01 -1.26600727e-01 -5.71863532e-01 -1.97080821e-01
8.17946997e-03 5.77102602e-01 -1.32472545e-01 1.81253508e-01
6.71360552e-01 3.96620989e-01 -2.72092242e-02 -6.12549126e-01
2.21414715e-01 1.34950733e+00 3.39845419e-02 -3.33741188e-01
6.73977435e-01 -2.81936210e-02 -5.14434278e-01 -8.97928238e-01
-6.05690718e-01 -4.27129537e-01 -4.82745230e-01 2.25785911e-01
5.50793886e-01 -1.26408577e+00 -1.69648588e-01 9.04293656e-02
-1.31116450e+00 -6.11432850e-01 6.85932264e-02 7.17502177e-01
-5.07906258e-01 1.83907390e-01 -8.22113156e-01 -7.40093648e-01
-3.91462862e-01 -1.27853441e+00 1.08036578e+00 -1.55218467e-01
-2.42679611e-01 -8.71773720e-01 -6.58522621e-02 4.09526169e-01
6.16060495e-01 -6.81241572e-01 9.44738448e-01 -7.95737386e-01
-4.41098779e-01 -2.01971948e-01 -4.03139042e-03 4.68594968e-01
2.73751676e-01 -2.83345670e-01 -9.27237332e-01 -6.46761537e-01
-1.23426378e-01 -3.51498455e-01 6.55657232e-01 -9.01956633e-02
8.19443822e-01 -5.57946861e-01 -1.46710262e-01 2.83916622e-01
1.00589263e+00 3.26724052e-01 4.70640063e-01 -1.15855569e-02
5.13245881e-01 4.89290863e-01 4.10941690e-01 6.87173828e-02
3.49729866e-01 9.38488543e-01 -2.61382610e-01 -1.20239466e-01
-2.47395754e-01 -3.52702856e-01 1.00899208e+00 1.75590730e+00
1.52856991e-01 -6.88549876e-01 -1.30261636e+00 7.85897613e-01
-1.67186403e+00 -6.75427794e-01 2.69823283e-01 2.40043521e+00
1.24694800e+00 2.01135352e-01 1.34716049e-01 -1.59395009e-01
5.91577113e-01 1.40577018e-01 -3.28922480e-01 -8.03481519e-01
-4.99670336e-04 3.70464563e-01 6.56019568e-01 7.52685666e-01
-6.01555586e-01 1.12740302e+00 7.46105623e+00 9.59591806e-01
-1.40162194e+00 4.42207217e-01 7.81859040e-01 -4.33149755e-01
-4.45866346e-01 2.05622584e-01 -8.51479292e-01 5.16174555e-01
2.10225201e+00 -2.26856858e-01 7.40884781e-01 7.19413459e-01
4.54299301e-01 1.48761973e-01 -1.21582842e+00 9.23859239e-01
1.42377570e-01 -1.18187201e+00 6.42836839e-02 1.28608361e-01
7.44591534e-01 3.84540141e-01 -1.02003612e-01 4.68341887e-01
1.39782026e-01 -1.19137681e+00 5.77492237e-01 1.23828024e-01
1.13112319e+00 -8.94694507e-01 6.89352393e-01 8.75020325e-01
-9.21141505e-01 1.04257435e-01 -6.33911669e-01 -2.92899013e-01
3.18749517e-01 5.62970757e-01 -1.24350023e+00 3.56996328e-01
2.07551718e-01 2.69796491e-01 -3.79194736e-01 5.57683945e-01
-1.37296468e-01 9.60017800e-01 -3.58729511e-01 -3.82274657e-01
3.44366729e-01 2.84831524e-02 2.36373380e-01 1.51353514e+00
8.03300321e-01 -1.24041833e-01 1.67453542e-01 1.33927181e-01
-1.92728609e-01 3.77530992e-01 -7.32017934e-01 -2.27219865e-01
5.86711943e-01 8.32801342e-01 -5.04596770e-01 -4.27468121e-01
-7.92092860e-01 1.22332060e+00 4.22323495e-01 4.22399670e-01
-6.69856846e-01 -3.34803015e-01 6.68799520e-01 2.41170958e-01
1.58112556e-01 -7.21767068e-01 -2.39006326e-01 -1.34609878e+00
2.87905127e-01 -1.39327300e+00 -1.73401013e-01 -6.62218392e-01
-8.61493230e-01 8.76271248e-01 -4.14449304e-01 -1.08114982e+00
-9.13325071e-01 -5.09049058e-01 -3.45167339e-01 1.20506120e+00
-1.37582397e+00 -9.33888495e-01 3.06179941e-01 2.37473369e-01
1.15057433e+00 -3.17645431e-01 9.56365705e-01 5.31639755e-01
-5.65371573e-01 8.76898468e-01 4.09770131e-01 4.87351827e-02
7.76128769e-01 -8.03113341e-01 9.96867537e-01 1.13618767e+00
6.82073355e-01 6.29718244e-01 5.68627477e-01 -4.78973120e-01
-1.65687704e+00 -9.28660810e-01 1.64834774e+00 -4.95612442e-01
5.64917266e-01 -6.99172974e-01 -8.39485168e-01 6.05907738e-01
3.99156690e-01 -1.98818177e-01 6.35606706e-01 2.70379394e-01
-1.54850200e-01 -6.35539442e-02 -6.38707638e-01 8.99539888e-01
1.11887980e+00 -1.01524198e+00 -3.35904747e-01 4.79482144e-01
9.76924539e-01 -3.66819531e-01 -7.35255182e-01 4.64521758e-02
3.68312180e-01 -4.12257671e-01 4.93940771e-01 -5.91094315e-01
2.46102616e-01 -9.98405069e-02 -1.98596612e-01 -1.59228098e+00
-6.80920929e-02 -1.11207068e+00 7.99840037e-03 1.04907453e+00
1.15914834e+00 -7.08203614e-01 4.97145027e-01 5.61605990e-01
-2.03254014e-01 -8.36973429e-01 -1.13870239e+00 -1.06912732e+00
3.00178081e-01 -5.14365852e-01 6.41578197e-01 5.85072994e-01
1.91911355e-01 6.95375443e-01 -4.92272466e-01 5.89132458e-02
-2.27896035e-01 -1.74069554e-01 6.53548956e-01 -6.23735070e-01
-4.56239283e-01 -1.61231428e-01 5.13300970e-02 -1.57991111e+00
3.42425346e-01 -9.19478297e-01 3.61804008e-01 -1.42423570e+00
3.99582349e-02 -4.10524547e-01 1.89722294e-03 6.85198247e-01
-2.03498483e-01 2.21331447e-01 1.28502801e-01 3.84084523e-01
-2.40380123e-01 4.63222414e-01 9.56294537e-01 -4.45548370e-02
-1.47983819e-01 7.02341199e-02 -6.71020091e-01 3.33877891e-01
7.12398887e-01 -5.51960111e-01 -4.72390383e-01 -9.73622382e-01
1.64024711e-01 2.15677500e-01 -1.62072793e-01 -8.59665334e-01
2.24896267e-01 -1.84674621e-01 2.82230787e-02 -1.79041147e-01
3.08949202e-01 -5.54730356e-01 -2.27436353e-03 2.65162498e-01
-3.08790296e-01 4.78960425e-01 1.27347320e-01 2.87629426e-01
-3.58815193e-01 -2.85645843e-01 6.02386177e-01 -7.02713951e-02
-2.71292865e-01 1.74057826e-01 -7.90619969e-01 -7.59282708e-02
7.13040352e-01 -3.53286304e-02 -1.66665494e-01 -7.25970626e-01
-4.36373532e-01 -1.97948724e-01 5.37310123e-01 6.14118934e-01
3.95846844e-01 -1.15543854e+00 -7.92570591e-01 3.41208875e-01
7.97302723e-02 -2.64872938e-01 -2.47466460e-01 7.82354116e-01
-4.74124283e-01 7.56565869e-01 2.66326040e-01 -5.30544937e-01
-1.01333451e+00 3.96703541e-01 1.86324850e-01 -1.82080358e-01
-5.60842216e-01 7.14769244e-01 -4.72300127e-02 -2.91720748e-01
9.85280871e-02 -3.96161944e-01 3.83475274e-01 1.61830951e-02
5.96992791e-01 -3.39026153e-02 6.54599011e-01 -6.88045859e-01
-3.57818455e-01 2.20842630e-01 -2.34372303e-01 -6.48008227e-01
1.09604347e+00 -2.94459820e-01 2.60568559e-01 5.16550362e-01
1.20980704e+00 2.07972452e-01 -1.19061363e+00 -3.36999953e-01
4.52055149e-02 -2.28559211e-01 1.30490651e-02 -6.58588231e-01
-6.79140151e-01 1.16113150e+00 2.59426951e-01 1.08487330e-01
1.03007805e+00 -1.73760012e-01 1.14380276e+00 7.17783332e-01
6.83368504e-01 -1.11094511e+00 -1.60642356e-01 8.46199989e-01
6.03406310e-01 -1.07153559e+00 -3.66177082e-01 -1.25508547e-01
-6.69898033e-01 9.65114355e-01 2.17469797e-01 3.55829626e-01
2.36517236e-01 4.03638929e-01 3.46463978e-01 4.14592624e-01
-1.23190773e+00 -2.81740010e-01 1.08343981e-01 4.46415633e-01
6.95390344e-01 1.45262852e-01 -1.95760384e-01 2.89249033e-01
-4.25036579e-01 -2.70399213e-01 5.24514079e-01 7.01333344e-01
-4.83678281e-01 -1.41753983e+00 -1.22889899e-01 3.85638833e-01
-6.10219896e-01 -5.64314246e-01 -3.22832257e-01 2.84999311e-01
-2.86531895e-01 1.36143887e+00 1.62708968e-01 -4.24680740e-01
2.46142864e-01 5.35246372e-01 1.49379790e-01 -1.00362754e+00
-5.47708750e-01 2.03664824e-01 4.90947932e-01 -3.98586571e-01
-7.23638609e-02 -6.40672445e-01 -9.30829942e-01 -5.67055643e-01
-3.37993860e-01 2.95530140e-01 8.70317101e-01 1.08876920e+00
5.93939543e-01 2.13528737e-01 8.82954895e-01 -6.95126474e-01
-7.37278998e-01 -1.19389105e+00 -1.69894606e-01 -2.28896916e-01
4.72756147e-01 2.24630572e-02 -3.07205737e-01 8.49842355e-02] | [14.449347496032715, 7.139906406402588] |
6e6201bc-e80c-4a85-b9c3-19c4bbef7aa9 | airex-neural-network-based-approach-for-air | 2108.07120 | null | https://arxiv.org/abs/2108.07120v1 | https://arxiv.org/pdf/2108.07120v1.pdf | AIREX: Neural Network-based Approach for Air Quality Inference in Unmonitored Cities | Urban air pollution is a major environmental problem affecting human health and quality of life. Monitoring stations have been established to continuously obtain air quality information, but they do not cover all areas. Thus, there are numerous methods for spatially fine-grained air quality inference. Since existing methods aim to infer air quality of locations only in monitored cities, they do not assume inferring air quality in unmonitored cities. In this paper, we first study the air quality inference in unmonitored cities. To accurately infer air quality in unmonitored cities, we propose a neural network-based approach AIREX. The novelty of AIREX is employing a mixture-of-experts approach, which is a machine learning technique based on the divide-and-conquer principle, to learn correlations of air quality between multiple cities. To further boost the performance, it employs attention mechanisms to compute impacts of air quality inference from the monitored cities to the locations in the unmonitored city. We show, through experiments on a real-world air quality dataset, that AIREX achieves higher accuracy than state-of-the-art methods. | ['Makoto Onizuka', 'Shohei Yamasaki', 'Kei Harada', 'Yuya Sasaki'] | 2021-08-16 | null | null | null | null | ['air-quality-inference'] | ['miscellaneous'] | [ 9.29788314e-03 -4.22639847e-01 -1.92826644e-01 -9.36284885e-02
-9.56461370e-01 -4.16593701e-01 4.77785736e-01 1.51812449e-01
-2.47331351e-01 9.17969048e-01 2.28135273e-01 -7.35449493e-01
-5.15283763e-01 -1.57226014e+00 -7.33909309e-01 -7.12677181e-01
5.00406623e-01 3.90429586e-01 -8.81162286e-03 2.42973015e-01
-3.80119532e-01 5.20829022e-01 -1.42359865e+00 1.07722335e-01
1.26961601e+00 7.52324700e-01 -1.52494460e-01 7.81486213e-01
-1.64953358e-02 3.70998442e-01 -8.50689888e-01 -6.48884848e-02
2.46651217e-01 -2.81353444e-01 -6.44612491e-01 -1.09615512e-01
4.79654849e-01 -1.83969766e-01 -6.53252378e-02 1.26020074e+00
2.71866143e-01 2.96795398e-01 7.38592565e-01 -1.20112288e+00
-5.95516145e-01 2.36601159e-01 -1.58432871e-01 5.55895984e-01
-1.15270853e-01 4.55808848e-01 1.06391549e+00 -2.54273564e-01
-4.51124042e-01 1.19111657e+00 5.06916881e-01 -2.31014774e-03
-8.80990088e-01 -8.68761122e-01 7.14487076e-01 -3.30513506e-03
-1.49539447e+00 -5.55169843e-02 2.90242821e-01 -3.89123350e-01
1.02019024e+00 8.05492520e-01 8.84678662e-01 5.38287938e-01
3.31047863e-01 3.90063196e-01 1.22392166e+00 -1.91090144e-02
3.97606701e-01 -1.39967397e-01 -4.52034175e-02 4.80214328e-01
6.00854874e-01 6.05303288e-01 2.34153286e-01 -2.69584090e-01
4.75124657e-01 6.63850844e-01 -8.52722079e-02 7.70422995e-01
-9.30796504e-01 7.40058124e-01 1.08479035e+00 2.03882217e-01
-5.84979177e-01 4.67223972e-01 -2.20629930e-01 2.60592364e-02
1.10539687e+00 2.25695640e-01 -5.43350101e-01 3.32012117e-01
-9.07101095e-01 3.48176211e-01 3.00788343e-01 6.65633500e-01
8.92310798e-01 -2.34403282e-01 -3.34631979e-01 3.30974400e-01
8.42987359e-01 1.39767730e+00 -1.74730256e-01 -7.31161237e-01
6.94689751e-01 6.42622411e-01 4.99547333e-01 -1.10233617e+00
-3.12675625e-01 -6.14322186e-01 -9.13203657e-01 2.58584946e-01
3.24623317e-01 -5.59197962e-01 -7.28872776e-01 1.29917932e+00
6.62049830e-01 7.74078190e-01 -3.96052867e-01 7.10998118e-01
4.10233229e-01 1.02226925e+00 4.21548367e-01 1.50245294e-01
1.27580130e+00 -9.49375570e-01 -6.91123009e-01 -6.37059286e-02
1.14913121e-01 -7.78722242e-02 1.02423263e+00 2.89530069e-01
-7.34071910e-01 -7.01377034e-01 -6.13798022e-01 6.02731824e-01
-1.31538641e+00 -1.32405356e-01 3.14619750e-01 9.40095782e-01
-6.53619885e-01 5.48272014e-01 -8.43676031e-01 2.28957571e-02
5.12674749e-01 -1.47876581e-02 2.03218535e-01 5.05425222e-02
-1.56816983e+00 8.66727650e-01 2.24488497e-01 2.65519768e-01
-1.04934990e+00 -8.75747979e-01 -7.00084329e-01 2.53616124e-01
1.65946141e-01 -9.86964345e-01 1.14456689e+00 -5.97448468e-01
-1.16977525e+00 1.81532457e-01 -2.17924342e-01 -1.83562636e-01
-6.90499693e-03 -4.89962578e-01 -1.13155293e+00 -6.94992393e-02
3.44459862e-01 2.45186329e-01 6.18830204e-01 -1.11675298e+00
-1.14297783e+00 -3.61314237e-01 2.09437162e-01 -1.87836573e-01
-2.35005066e-01 7.92660043e-02 -2.00014748e-02 -2.93213636e-01
-4.65678036e-01 -7.94657290e-01 -6.46873474e-01 -7.94652775e-02
-3.70878875e-01 -4.89383847e-01 1.82884723e-01 -5.88570714e-01
1.67842972e+00 -1.65577114e+00 -5.48760235e-01 3.66257876e-01
2.29198620e-01 5.84604323e-01 -4.01360951e-02 1.81920137e-02
2.29916535e-02 3.03177565e-01 -5.21959007e-01 -3.28569591e-01
3.77661377e-01 4.36053187e-01 -2.70421922e-01 6.76963508e-01
4.35270131e-01 1.00407147e+00 -1.35525894e+00 -3.10731620e-01
4.27134335e-01 4.31619018e-01 -3.08118880e-01 2.21996740e-01
-6.03317142e-01 6.94816709e-01 -7.38433242e-01 3.71621460e-01
8.29115212e-01 -1.43590108e-01 -3.82242203e-01 2.19534755e-01
-4.15998876e-01 4.58382457e-01 -1.11914599e+00 1.27661431e+00
-1.00719249e+00 3.66419643e-01 -3.86918969e-02 -4.62209851e-01
4.05050129e-01 7.51219630e-01 1.46726161e-01 -7.95419574e-01
2.75625493e-02 -4.87021171e-02 -4.08777624e-01 -6.97939813e-01
1.29937693e-01 -4.13563669e-01 -4.33685556e-02 6.19116127e-01
-7.81322002e-01 -1.17919907e-01 -8.31523016e-02 -4.82000351e-01
7.96947956e-01 -8.20470005e-02 1.79594249e-01 -9.24767479e-02
5.93665421e-01 -9.18160826e-02 4.95157868e-01 6.25348747e-01
-2.83065528e-01 3.22715998e-01 -1.98436737e-01 -4.95988756e-01
-5.50241888e-01 -1.35514474e+00 -3.01612139e-01 1.11166728e+00
-1.55292436e-01 -1.15097277e-01 -8.24016035e-01 -8.28830063e-01
3.14111374e-02 1.14459360e+00 -7.50839651e-01 -5.66618107e-02
-2.88579702e-01 -1.19807541e+00 6.21839881e-01 5.88660538e-01
7.92175114e-01 -7.19743729e-01 -5.92577279e-01 1.57320663e-01
-2.76745975e-01 -7.24233508e-01 -2.73099571e-01 -2.69015044e-01
-7.57278562e-01 -1.23181236e+00 -3.97832185e-01 1.85515016e-01
7.08146870e-01 2.44178534e-01 1.52412093e+00 -3.14652063e-02
3.57890725e-02 2.77494431e-01 8.88438821e-02 -1.06865704e+00
-2.30902940e-01 8.79111066e-02 -1.13544419e-01 2.53417462e-01
6.86082840e-01 -6.41755044e-01 -8.08710217e-01 3.74206096e-01
-1.15198994e+00 -4.95609999e-01 5.25090158e-01 2.92470660e-02
8.47590446e-01 1.07196856e+00 5.84053338e-01 -7.12235153e-01
5.23936272e-01 -1.03204048e+00 -9.53096926e-01 1.67646319e-01
-7.08950162e-01 -1.29531905e-01 7.29991496e-01 -1.33191571e-01
-1.14611685e+00 -1.10743895e-01 -4.21674252e-01 -3.00719198e-02
-9.39199030e-01 2.56694824e-01 -5.30441463e-01 6.58485532e-01
3.99810046e-01 3.77512351e-02 -9.39559996e-01 -7.36622274e-01
5.54780185e-01 9.64803159e-01 2.06667185e-01 -6.25419199e-01
1.01424479e+00 6.71377063e-01 -6.88885227e-02 -6.30055308e-01
-1.54187894e+00 -7.61850834e-01 -7.02098370e-01 -2.78282434e-01
1.34341240e+00 -1.17001951e+00 -5.69384754e-01 2.04694029e-02
-1.11798406e+00 -2.95453936e-01 -1.79202482e-01 4.66819823e-01
9.42914635e-02 -1.77689627e-01 -4.74160612e-02 -1.11538792e+00
-2.51210153e-01 -8.71790230e-01 1.39539027e+00 2.19423726e-01
-4.57231142e-02 -1.62004089e+00 5.64739823e-01 4.40566748e-01
6.42538190e-01 5.47022521e-01 7.78190196e-01 -6.16099238e-02
-7.25602984e-01 -1.99579135e-01 -4.35541183e-01 2.96625286e-01
6.26624882e-01 -2.68794775e-01 -1.26789367e+00 -4.33204696e-03
-1.52094126e-01 3.34625900e-01 8.54367137e-01 2.80580819e-01
1.23460734e+00 -6.38025939e-01 -3.94234449e-01 3.49478006e-01
1.41866767e+00 1.17070667e-01 5.44426143e-01 1.91328242e-01
5.71639597e-01 4.05325502e-01 3.11745852e-01 1.75781220e-01
5.09092450e-01 2.22412303e-01 1.00109315e+00 -4.27772999e-01
3.56162526e-02 -2.58042634e-01 2.14510515e-01 4.62643951e-01
-1.09454408e-01 -4.66879785e-01 -7.28418469e-01 8.76399100e-01
-1.65448165e+00 -1.01945090e+00 -6.52744830e-01 2.14908051e+00
7.19130516e-01 -8.71170238e-02 2.67884880e-01 9.41034704e-02
4.59136069e-01 3.35309893e-01 -5.05953372e-01 -6.89732805e-02
4.49042320e-01 4.08589929e-01 6.40761673e-01 7.37215519e-01
-1.36377716e+00 4.59919423e-01 6.51788330e+00 5.28192937e-01
-7.32592940e-01 5.57731748e-01 5.04914463e-01 -2.23815948e-01
-4.61380899e-01 -2.89750785e-01 -9.12072957e-01 6.37704372e-01
1.38741171e+00 4.79248255e-01 3.99312109e-01 4.77634013e-01
7.88010716e-01 2.14794874e-02 -8.47621918e-01 4.94834691e-01
-5.78818977e-01 -8.47088277e-01 -1.04123335e-02 1.96399853e-01
1.11095262e+00 2.82173812e-01 1.70410976e-01 1.48862228e-01
7.74914205e-01 -1.35037887e+00 3.59485179e-01 1.06438613e+00
4.93813366e-01 -9.08385873e-01 6.56851888e-01 6.15085065e-01
-1.49111664e+00 -5.53424880e-02 -5.31156361e-01 -3.09107453e-01
9.59857181e-02 1.22625732e+00 -2.17958823e-01 7.59676158e-01
8.31616282e-01 5.46283543e-01 -5.17410696e-01 1.02370262e+00
-8.42878878e-01 1.08484793e+00 -4.41210061e-01 -7.37404302e-02
4.92899626e-01 -2.92452365e-01 2.97834069e-01 1.24715149e+00
3.59438211e-01 1.19574666e-01 5.29758260e-02 1.48696935e+00
1.91131249e-01 -2.89905131e-01 -7.63917446e-01 2.85675228e-01
2.09269494e-01 8.78388047e-01 -2.32601658e-01 -5.50120592e-01
-4.78551149e-01 6.39773250e-01 2.03476250e-02 5.73452473e-01
-1.29039276e+00 -4.74941224e-01 1.17326272e+00 1.16291091e-01
2.13249877e-01 -1.05475374e-01 -2.31008634e-01 -7.75802076e-01
-2.92601109e-01 -5.76076150e-01 2.91474193e-01 -6.53381467e-01
-1.39374638e+00 2.64177173e-01 2.84921676e-01 -1.04381311e+00
2.71847218e-01 -6.05688453e-01 -1.12945676e+00 1.12441993e+00
-2.10349846e+00 -8.35195243e-01 -3.06056172e-01 5.03008485e-01
2.78175920e-01 4.63868856e-01 8.30317497e-01 5.77117085e-01
-8.82468522e-01 5.73962890e-02 2.66842037e-01 2.07651302e-01
1.04876533e-01 -1.35556304e+00 6.71322286e-01 9.28697586e-01
1.17963575e-01 4.80116040e-01 3.22944105e-01 -6.54660285e-01
-7.75258660e-01 -2.25143695e+00 1.08233786e+00 -8.93195987e-01
6.04742050e-01 3.25398929e-02 -6.28393054e-01 6.48200572e-01
3.11820477e-01 1.06702298e-02 9.71319318e-01 4.52477932e-02
-2.35486463e-01 -4.37754422e-01 -1.18706334e+00 5.28468251e-01
7.72042811e-01 -6.53093338e-01 -6.58072054e-01 6.18523955e-01
9.20265496e-01 3.19247305e-01 -1.04947460e+00 1.09398561e-02
1.73820168e-01 -8.56954515e-01 1.11433470e+00 -8.10138047e-01
3.15952152e-01 -8.43853593e-01 -1.84649929e-01 -1.72371292e+00
-6.69919014e-01 1.15711860e-01 -9.79368314e-02 1.11039615e+00
8.01686943e-01 -9.58779335e-01 4.16830480e-01 7.20906615e-01
-3.06646843e-02 -3.73968273e-01 -7.39320695e-01 -8.05269122e-01
5.51786363e-01 -8.44935954e-01 1.74048066e+00 5.57769179e-01
-6.96814656e-01 1.76129103e-01 -1.11867495e-01 1.15235686e+00
7.53391623e-01 3.20874721e-01 3.57915729e-01 -1.38276696e+00
-3.15660745e-01 -5.81825912e-01 4.11020756e-01 -9.49246585e-01
3.34187090e-01 -6.76526606e-01 2.01501340e-01 -2.02658534e+00
-7.33041912e-02 -4.56820428e-01 -7.33588219e-01 3.72526258e-01
-7.29106307e-01 1.99823499e-01 -2.82317519e-01 -3.55941147e-01
-6.69892967e-01 6.34337425e-01 1.32757485e+00 -7.15645313e-01
-1.60612717e-01 6.10889852e-01 -6.98949277e-01 7.86054194e-01
1.09908187e+00 -8.90181363e-01 -4.55051839e-01 -8.59330058e-01
6.46739960e-01 -3.90343398e-01 5.14200151e-01 -1.24537575e+00
9.93788391e-02 -6.90577507e-01 4.90156233e-01 -8.07922482e-01
1.77168101e-01 -1.38871753e+00 4.81327593e-01 6.21116638e-01
-1.28429785e-01 1.26508132e-01 1.91145763e-01 8.79375458e-01
-4.32531312e-02 -4.22845110e-02 5.63705742e-01 -3.89774770e-01
-1.04953252e-01 6.98495567e-01 -6.90087199e-01 -7.52292648e-02
7.80898213e-01 3.82188827e-01 -2.81169981e-01 -8.90692100e-02
-4.96484011e-01 3.22018206e-01 -3.28186825e-02 6.68555573e-02
3.92273366e-01 -1.50625777e+00 -7.71354377e-01 1.38748080e-01
3.73147465e-02 3.89570564e-01 3.93178463e-02 6.36765957e-01
-2.57136613e-01 8.11685979e-01 5.79049230e-01 -3.63123894e-01
-8.05349827e-01 8.45700681e-01 8.30347836e-01 -6.64501548e-01
-2.91332930e-01 4.75086957e-01 2.38079026e-01 -9.28358078e-01
-5.58048412e-02 -1.03484547e+00 -1.36130065e-01 -1.21920571e-01
8.72870624e-01 8.62191200e-01 1.52590647e-01 -4.57016319e-01
-2.41981164e-01 6.80024803e-01 9.61054921e-01 9.50667188e-02
1.14914131e+00 -2.42734671e-01 -1.53319120e-01 5.59170306e-01
1.07390869e+00 1.41568422e-01 -1.15446532e+00 -1.91138387e-01
-1.89481303e-01 -5.64645588e-01 4.16926891e-01 -1.04409015e+00
-1.22631443e+00 1.15486753e+00 6.35288775e-01 4.78317916e-01
1.21976531e+00 -2.00684950e-01 1.15570617e+00 4.17850226e-01
5.54544851e-02 -9.11642134e-01 -3.91881019e-01 4.15960133e-01
4.76737022e-01 -1.30435884e+00 -8.40055943e-02 -6.94780350e-02
1.36346534e-01 4.87893432e-01 4.21485275e-01 3.37924063e-02
1.17441094e+00 -2.95094699e-02 9.60113853e-02 -3.71149689e-01
-5.70463538e-01 -6.50330842e-01 3.98942858e-01 7.05187917e-01
1.79044008e-01 6.62088454e-01 5.84887922e-01 4.02057797e-01
-1.66509394e-02 -1.59887761e-01 -3.65426630e-01 4.48784024e-01
-7.31935799e-01 -7.45893776e-01 -1.12292802e+00 4.32289809e-01
-4.33504641e-01 -3.27314347e-01 -2.06567213e-01 4.10229445e-01
8.41163456e-01 1.69332910e+00 2.08642989e-01 -1.38553428e-02
6.86855376e-01 2.47837175e-02 6.36487082e-03 -5.62078595e-01
-8.16786110e-01 -4.12585348e-01 -8.97515472e-03 -5.04070461e-01
-6.55991137e-01 -3.20002139e-01 -9.69158173e-01 -4.02544528e-01
-2.61460096e-01 4.29940730e-01 6.35536611e-01 1.16370773e+00
2.22925380e-01 1.13391399e+00 8.00657332e-01 -5.85619032e-01
-2.03888178e-01 -9.10175443e-01 -6.93909824e-01 1.82543069e-01
7.52003849e-01 -4.47951734e-01 -4.41285282e-01 -2.67805457e-01] | [6.254217624664307, 2.50225567817688] |
267043ed-0d6f-4483-b7dc-5d6cdde2175a | the-open-corpus-of-the-veps-and-karelian | 2206.03870 | null | https://arxiv.org/abs/2206.03870v1 | https://arxiv.org/pdf/2206.03870v1.pdf | The Open corpus of the Veps and Karelian languages: overview and applications | A growing priority in the study of Baltic-Finnic languages of the Republic of Karelia has been the methods and tools of corpus linguistics. Since 2016, linguists, mathematicians, and programmers at the Karelian Research Centre have been working with the Open Corpus of the Veps and Karelian Languages (VepKar), which is an extension of the Veps Corpus created in 2009. The VepKar corpus comprises texts in Karelian and Veps, multifunctional dictionaries linked to them, and software with an advanced system of search using various criteria of the texts (language, genre, etc.) and numerous linguistic categories (lexical and grammatical search in texts was implemented thanks to the generator of word forms that we created earlier). A corpus of 3000 texts was compiled, texts were uploaded and marked up, the system for classifying texts into languages, dialects, types and genres was introduced, and the word-form generator was created. Future plans include developing a speech module for working with audio recordings and a syntactic tagging module using morphological analysis outputs. Owing to continuous functional advancements in the corpus manager and ongoing VepKar enrichment with new material and text markup, users can handle a wide range of scientific and applied tasks. In creating the universal national VepKar corpus, its developers and managers strive to preserve and exhibit as fully as possible the state of the Veps and Karelian languages in the 19th-21st centuries. | ['Aleksandra Rodionova', 'Nataliya Pellinen', 'Irina Novak', 'Andrew Krizhanovsky', 'Natalia Krizhanovskaya', 'Nina Zaitseva', 'Tatyana Boyko'] | 2022-06-08 | null | null | null | null | ['morphological-analysis'] | ['natural-language-processing'] | [-4.84660059e-01 1.76768750e-02 -4.28775474e-02 -1.18513443e-01
-4.91247743e-01 -8.42245996e-01 6.28530741e-01 1.65764004e-01
-6.98508620e-01 5.58151424e-01 5.24457455e-01 -7.55964339e-01
-2.40836173e-01 -6.21629655e-01 1.63277656e-01 -2.95243144e-01
1.06117278e-02 6.32308662e-01 9.73474458e-02 -5.82781017e-01
2.51513362e-01 2.75271893e-01 -1.55768752e+00 -1.12689836e-02
7.14981675e-01 3.96778166e-01 8.14802349e-01 6.70823872e-01
-4.45338368e-01 2.02711344e-01 -7.36579180e-01 -4.49085176e-01
1.03816679e-02 -5.31738400e-01 -1.01272357e+00 -2.59198666e-01
1.12238213e-01 7.44878128e-02 -2.33554214e-01 8.13957512e-01
4.66285408e-01 -5.78302005e-03 3.66104603e-01 -4.62863773e-01
-5.64062357e-01 1.31987906e+00 -6.27501234e-02 3.12750787e-01
4.67863649e-01 -1.77817449e-01 1.04165947e+00 -8.79923046e-01
9.50599968e-01 1.21298909e+00 5.71491838e-01 4.32537168e-01
-9.53920722e-01 -6.00109398e-01 -4.73597854e-01 2.18251273e-01
-1.57813823e+00 -4.30935144e-01 3.81699324e-01 -6.39425337e-01
1.01402307e+00 3.39322895e-01 8.65124762e-01 7.75136769e-01
-2.22554982e-01 2.14863628e-01 8.66873562e-01 -1.09724486e+00
-1.06482003e-02 4.57541168e-01 1.16707996e-01 4.11596596e-01
1.94056228e-01 -1.93251997e-01 -3.64190429e-01 -8.58407989e-02
6.33019328e-01 -6.91463590e-01 -3.51431556e-02 4.24580961e-01
-1.21652889e+00 8.73309374e-01 -6.07021034e-01 1.03952062e+00
-1.63878754e-01 -4.63958442e-01 9.76442099e-01 4.18890566e-01
2.94975758e-01 3.39293718e-01 -6.79853618e-01 -4.70393836e-01
-1.13201857e+00 2.77880996e-01 1.02108502e+00 7.92204261e-01
2.97693491e-01 2.72942841e-01 2.91845918e-01 1.42046440e+00
5.02257705e-01 6.78634048e-01 9.37939942e-01 -7.83961713e-01
3.45445067e-01 4.82867271e-01 -4.02803928e-01 -7.81365752e-01
-4.74858254e-01 1.18198849e-01 -4.51050103e-01 -2.56659538e-01
4.92085427e-01 -2.56160796e-01 -6.18122756e-01 1.48450267e+00
4.00976270e-01 -7.40143597e-01 1.64959341e-01 3.97741735e-01
1.06830359e+00 9.27837074e-01 1.45472437e-01 -4.88412917e-01
1.66101646e+00 -2.58230567e-01 -9.21000123e-01 2.46164992e-01
7.64630795e-01 -1.23160577e+00 1.10503280e+00 5.03570735e-01
-1.29615390e+00 -4.59992498e-01 -9.09990072e-01 -2.95700505e-02
-5.32368898e-01 1.81120634e-01 6.26965165e-01 9.04823482e-01
-1.06744206e+00 4.01266962e-01 -5.81468403e-01 -6.42673075e-01
-4.72540148e-02 1.45834923e-01 -4.48116809e-01 5.01819372e-01
-1.44062889e+00 1.02130663e+00 7.31258273e-01 -1.91097289e-01
-1.53854057e-01 -5.41341126e-01 -1.13455415e+00 -2.87422687e-01
-5.77110983e-02 5.58716916e-02 1.00279129e+00 -8.03880930e-01
-1.38084698e+00 1.41789687e+00 3.25275660e-01 -3.21640015e-01
2.24663585e-01 3.18849236e-01 -9.57979262e-01 2.10267201e-01
4.44732666e-01 1.18334614e-01 3.08496691e-02 -4.64798272e-01
-9.83916521e-01 -2.73089379e-01 -2.56312668e-01 5.56541432e-04
-3.18414748e-01 8.22447777e-01 -3.61067861e-01 -9.07942295e-01
-9.07870103e-03 -5.87808549e-01 1.32950440e-01 -7.96056151e-01
1.22169472e-01 -4.62828934e-01 6.17022932e-01 -1.41189301e+00
1.78232658e+00 -2.23343587e+00 -1.09413713e-01 3.95602465e-01
-2.80073017e-01 3.77926677e-01 2.49791995e-01 8.35180223e-01
-1.87635720e-02 1.98589638e-01 -4.00691479e-03 8.13213736e-02
1.80451006e-01 5.47622263e-01 -1.58206925e-01 4.10683930e-01
-2.40236953e-01 3.83490533e-01 -8.44326735e-01 -6.58128560e-01
4.32193100e-01 3.82072777e-01 -2.96005875e-01 -4.52866644e-01
5.46463318e-02 2.30168104e-01 -9.98341218e-02 5.20794332e-01
3.52547735e-01 5.83518445e-01 3.83976907e-01 5.16420066e-01
-8.21110845e-01 8.60019743e-01 -1.20325053e+00 1.43111050e+00
-7.46262908e-01 7.79329240e-01 3.78180802e-01 -8.29560757e-01
1.08606207e+00 7.10005045e-01 3.64957154e-01 -4.81797189e-01
1.78316444e-01 6.73781097e-01 2.93045193e-01 -6.99822724e-01
9.59571838e-01 -5.79589121e-02 -4.34083879e-01 3.25060785e-01
5.80897570e-01 -3.13316494e-01 8.19429457e-01 1.21940963e-01
6.87919796e-01 2.08183289e-01 6.35099530e-01 -7.96148241e-01
7.86151469e-01 2.54686058e-01 4.53263760e-01 3.84825617e-02
8.28451365e-02 2.28861377e-01 3.97481620e-01 -2.45558873e-01
-1.43543351e+00 -6.65439069e-01 -9.50502753e-01 1.30849540e+00
-6.42943621e-01 -7.49986231e-01 -8.35256457e-01 1.53708845e-01
-3.61261845e-01 6.25209212e-01 -2.08293676e-01 4.46404546e-01
-9.18133318e-01 -3.23048353e-01 1.03756177e+00 -1.19829871e-01
2.41558999e-01 -1.63528788e+00 -4.89858925e-01 3.66219103e-01
-2.33124971e-01 -1.01997149e+00 -2.22113088e-01 4.87077199e-02
-3.19271713e-01 -8.46274793e-01 -5.39251685e-01 -1.44138026e+00
-4.27540131e-02 -4.80611324e-01 8.41214895e-01 -1.34900436e-01
-3.66295695e-01 1.84892148e-01 -5.90443492e-01 -5.30694366e-01
-1.08840072e+00 1.73796624e-01 -1.99172311e-02 -4.91965204e-01
1.95035905e-01 -4.82728690e-01 1.82157725e-01 -1.33232996e-01
-7.86387742e-01 -2.20963553e-01 -3.19667906e-02 5.38099647e-01
3.72855514e-01 -9.47098508e-02 8.19219112e-01 -9.78697062e-01
3.71597826e-01 -3.90197664e-01 -6.76746964e-01 -1.42116874e-01
-2.66356766e-01 -3.66318762e-01 6.91892505e-01 -2.08318293e-01
-1.07755637e+00 -3.15054357e-01 -8.27919602e-01 4.16349143e-01
-2.31985971e-01 8.00553203e-01 -3.00987333e-01 2.31482074e-01
6.16766214e-01 1.58938736e-01 -2.24334970e-01 -6.53254449e-01
4.48605210e-01 1.36961257e+00 8.90462935e-01 -5.39656281e-01
4.41038787e-01 7.54352845e-03 -3.62567127e-01 -1.40096676e+00
-1.50919765e-01 -6.75755739e-01 -6.57150984e-01 -3.65674466e-01
7.84412980e-01 -8.92224669e-01 -4.11165237e-01 3.23817044e-01
-1.07013309e+00 -2.32956216e-01 -5.13623953e-01 6.78453922e-01
-3.95934850e-01 5.05650103e-01 -6.79150701e-01 -8.75460029e-01
-4.67117816e-01 -4.78877366e-01 6.09171271e-01 1.56325907e-01
-7.05758154e-01 -1.34610176e+00 3.90036047e-01 -4.27002423e-02
1.26866121e-02 2.22386748e-01 1.16015744e+00 -6.76449835e-01
3.82392496e-01 -1.85894072e-02 3.23541433e-01 4.90565628e-01
-1.63860973e-02 4.20105129e-01 -6.04983509e-01 5.46523817e-02
-2.12912280e-02 -2.88840893e-05 3.46527040e-01 2.55570143e-01
4.07028258e-01 -4.35938030e-01 1.37889653e-01 3.07390273e-01
1.37052870e+00 5.05739450e-01 6.17026031e-01 5.32908261e-01
2.24113494e-01 7.69271016e-01 3.52605760e-01 3.96924257e-01
5.90854406e-01 6.06915474e-01 -3.98242593e-01 2.71120995e-01
-3.20516288e-01 -4.81494665e-02 9.05878007e-01 1.61794329e+00
-1.45241782e-01 2.02278778e-01 -1.10761154e+00 1.22130930e+00
-1.41124463e+00 -1.00608826e+00 -6.31036401e-01 2.34173417e+00
1.15894651e+00 1.49211204e-02 3.58965904e-01 4.87599850e-01
8.38131368e-01 -1.05882242e-01 5.86924434e-01 -1.05983031e+00
-4.04175282e-01 9.70735610e-01 3.52211356e-01 7.05301762e-01
-8.94775808e-01 1.24291766e+00 6.60116673e+00 1.16727459e+00
-1.03043056e+00 6.25060499e-02 -2.12051988e-01 1.86791956e-01
-2.77711958e-01 1.15037762e-01 -1.00658596e+00 4.91982937e-01
1.34439015e+00 -5.19640088e-01 4.12529618e-01 5.71664751e-01
5.21780074e-01 -1.88794345e-01 -5.89985013e-01 7.14822650e-01
9.34552103e-02 -1.45330608e+00 -2.94240445e-01 2.03736931e-01
5.12098730e-01 3.81800681e-01 -4.93205607e-01 4.37225044e-01
2.79957026e-01 -7.10874200e-01 1.23500252e+00 1.96805999e-01
1.16682637e+00 -6.83577478e-01 7.16624200e-01 2.23275319e-01
-1.19978285e+00 1.23115733e-01 -4.89181429e-01 -2.66829461e-01
2.76925892e-01 3.41211200e-01 -5.56241810e-01 7.07064033e-01
5.71857035e-01 3.47999305e-01 -4.65244502e-01 8.33917081e-01
-2.91928411e-01 1.18361807e+00 -4.81427222e-01 -1.78858057e-01
2.32625753e-01 -4.92739439e-01 8.64811122e-01 1.64975393e+00
3.14837307e-01 -1.64395511e-01 5.26136756e-02 4.80972707e-01
3.75677645e-01 9.11723137e-01 -3.21120620e-01 -2.37798795e-01
6.67675912e-01 1.36532331e+00 -9.16985154e-01 -3.03929657e-01
-3.52460772e-01 2.04366967e-01 -1.52744606e-01 -8.13520104e-02
-3.49439144e-01 -9.10551727e-01 3.54080498e-01 3.98276597e-01
2.44784534e-01 -3.81759584e-01 -4.19282854e-01 -7.20461071e-01
-4.11407538e-02 -9.05083716e-01 6.45741224e-01 -3.16174656e-01
-1.00023115e+00 8.73910666e-01 1.70730352e-01 -5.45611322e-01
-5.68490088e-01 -6.31978810e-01 -3.63962650e-01 1.06522107e+00
-6.82207227e-01 -1.17623353e+00 4.60512578e-01 4.47095305e-01
4.99455124e-01 -4.94581372e-01 1.27563190e+00 5.61609507e-01
-2.31972978e-01 2.93972671e-01 2.67768979e-01 5.13030052e-01
5.21314085e-01 -1.11761498e+00 2.25209698e-01 4.61559594e-01
2.75294900e-01 3.54595900e-01 7.17845738e-01 -5.78488648e-01
-6.81217372e-01 -7.10008085e-01 1.75875854e+00 -1.59512788e-01
1.33486426e+00 -6.01855814e-01 -6.40690565e-01 5.24945855e-01
4.22750831e-01 -7.89917290e-01 8.21280599e-01 1.06696285e-01
-4.24201377e-02 1.63144261e-01 -9.22223687e-01 5.72773576e-01
7.31399953e-01 -4.91278619e-01 -9.42368567e-01 3.55677128e-01
5.29031396e-01 -2.81320274e-01 -1.03945589e+00 -2.16524586e-01
6.28878236e-01 -5.82276642e-01 3.78925830e-01 -6.58681318e-02
9.62705091e-02 -2.24130958e-01 -5.20854704e-02 -1.11119449e+00
-9.28520113e-02 -1.02960861e+00 8.37849975e-01 1.84233284e+00
5.28078854e-01 -7.22342491e-01 2.62128979e-01 -2.23069973e-02
-6.05637908e-01 -2.10333951e-02 -1.19555473e+00 -6.24726057e-01
3.47051650e-01 -8.11185360e-01 3.04544389e-01 1.03600836e+00
6.79093540e-01 4.83293623e-01 -4.80062328e-03 -2.87971765e-01
6.52875453e-02 -3.53289276e-01 3.63337547e-01 -1.34537840e+00
-2.94318318e-01 -8.31716895e-01 -6.20142043e-01 -1.51273444e-01
1.73093587e-01 -1.30082941e+00 -3.01803708e-01 -1.44888377e+00
-4.13181096e-01 -5.90421200e-01 4.30359840e-01 5.36453009e-01
4.17216599e-01 1.72803923e-01 2.97718257e-01 1.87995270e-01
2.06950575e-01 -6.43889010e-02 8.24203968e-01 3.62172872e-01
-5.67695320e-01 -2.16371536e-01 -5.12246311e-01 8.05123925e-01
8.22392404e-01 -4.42973673e-01 1.87406823e-01 -1.12597838e-01
5.71859598e-01 -2.16345757e-01 6.46763295e-03 -8.82275760e-01
-1.26537353e-01 2.62595583e-02 -9.96116772e-02 -7.24529743e-01
-1.87066928e-01 -4.49225664e-01 4.68385607e-01 2.53616244e-01
9.98574216e-03 6.35247901e-02 5.12935221e-01 -5.12129009e-01
-3.09800625e-01 -6.00958526e-01 8.77756715e-01 -1.47529706e-01
-7.24674225e-01 -1.30403310e-01 -9.58051026e-01 2.61272788e-01
1.00595295e+00 -4.20317590e-01 -4.22137156e-02 -1.09185837e-01
-7.48449922e-01 1.67639688e-01 3.49794924e-01 2.16507480e-01
2.57981811e-02 -1.14483678e+00 -1.03639376e+00 4.72241431e-01
-4.36906293e-02 -3.14220220e-01 1.49880558e-01 6.38850570e-01
-1.16422069e+00 4.86368537e-01 -4.21924680e-01 6.79300260e-03
-1.12057388e+00 1.80897728e-01 -1.21114329e-02 -3.39696676e-01
-7.33898818e-01 4.32579458e-01 -1.94472075e-01 -6.82668090e-01
1.02346316e-01 -5.92374466e-02 -7.52501011e-01 4.32653904e-01
4.93696988e-01 3.75496000e-01 1.75886050e-01 -1.32482314e+00
-3.08686733e-01 3.20330709e-01 3.06425184e-01 -6.54359996e-01
1.35514295e+00 -3.01019609e-01 -4.68949974e-01 8.40226948e-01
8.18185568e-01 7.52614975e-01 -1.84159741e-01 1.68870479e-01
1.63180143e-01 1.53670266e-01 7.16222217e-03 -7.66238213e-01
-4.98006552e-01 5.33408523e-01 5.97223174e-03 4.95713204e-01
8.58954728e-01 2.01281548e-01 6.29110873e-01 -4.32545170e-02
2.52918541e-01 -1.56847632e+00 -9.29311991e-01 1.17135525e+00
8.16838562e-01 -3.37084174e-01 -2.78446287e-01 -3.24689567e-01
-6.20560825e-01 1.25864387e+00 -1.60449028e-01 -8.36834162e-02
7.98063457e-01 5.24192631e-01 4.00143772e-01 -1.97983176e-01
-4.98686761e-01 -4.18940425e-01 3.86659838e-02 8.62926126e-01
7.21445560e-01 2.99278229e-01 -1.30710995e+00 9.31327999e-01
-1.28859901e+00 -3.39246392e-01 6.74021304e-01 5.60367584e-01
-4.79470551e-01 -1.54378235e+00 -5.88383675e-01 1.94652945e-01
-9.80708241e-01 -3.40282053e-01 -2.85369009e-01 1.29058552e+00
5.11106253e-01 8.72010529e-01 3.25073034e-01 -1.48558691e-01
1.76152408e-01 1.43677667e-01 4.24289942e-01 -8.57914150e-01
-9.93831933e-01 5.52016616e-01 6.03399217e-01 8.02756771e-02
-3.12789291e-01 -9.83162224e-01 -1.50690842e+00 -4.69126105e-01
-3.93615812e-01 6.80395901e-01 9.88352299e-01 9.47843730e-01
-2.89932698e-01 2.14592889e-01 3.84650886e-01 -6.78793848e-01
-1.42720163e-01 -1.28903031e+00 -1.17551851e+00 -1.09767206e-01
-3.56586963e-01 -1.58963710e-01 -3.15921575e-01 3.72002274e-01] | [10.368390083312988, 10.21158504486084] |
19f60b90-9d5d-43e0-b96a-6ca8556ed1fb | gif-generative-interpretable-faces | 2009.00149 | null | https://arxiv.org/abs/2009.00149v2 | https://arxiv.org/pdf/2009.00149v2.pdf | GIF: Generative Interpretable Faces | Photo-realistic visualization and animation of expressive human faces have been a long standing challenge. 3D face modeling methods provide parametric control but generates unrealistic images, on the other hand, generative 2D models like GANs (Generative Adversarial Networks) output photo-realistic face images, but lack explicit control. Recent methods gain partial control, either by attempting to disentangle different factors in an unsupervised manner, or by adding control post hoc to a pre-trained model. Unconditional GANs, however, may entangle factors that are hard to undo later. We condition our generative model on pre-defined control parameters to encourage disentanglement in the generation process. Specifically, we condition StyleGAN2 on FLAME, a generative 3D face model. While conditioning on FLAME parameters yields unsatisfactory results, we find that conditioning on rendered FLAME geometry and photometric details works well. This gives us a generative 2D face model named GIF (Generative Interpretable Faces) that offers FLAME's parametric control. Here, interpretable refers to the semantic meaning of different parameters. Given FLAME parameters for shape, pose, expressions, parameters for appearance, lighting, and an additional style vector, GIF outputs photo-realistic face images. We perform an AMT based perceptual study to quantitatively and qualitatively evaluate how well GIF follows its conditioning. The code, data, and trained model are publicly available for research purposes at http://gif.is.tue.mpg.de. | ['Michael Black', 'Anurag Ranjan', 'Pravir Singh Gupta', 'Timo Bolkart', 'Roy Uziel', 'Partha Ghosh'] | 2020-08-31 | null | null | null | null | ['3d-face-modeling'] | ['computer-vision'] | [ 2.55306035e-01 3.00960183e-01 4.17527109e-01 -4.01705474e-01
-3.92139882e-01 -8.96050274e-01 9.55772936e-01 -6.24135315e-01
2.51984209e-01 4.65810120e-01 2.28981435e-01 -1.17900521e-01
3.69845390e-01 -7.95987785e-01 -7.13073850e-01 -7.58776724e-01
7.13664107e-03 5.50304532e-01 -4.11440194e-01 -2.26897389e-01
-1.18892655e-01 7.70615160e-01 -1.36315143e+00 1.56891674e-01
6.71326280e-01 6.55706286e-01 -2.84471184e-01 8.99940968e-01
9.86306891e-02 4.78953958e-01 -7.95655370e-01 -5.67285120e-01
4.97972190e-01 -7.98030734e-01 -4.36539412e-01 1.69792533e-01
5.90308726e-01 -4.31639284e-01 -5.67708015e-02 8.74440968e-01
4.93921638e-01 1.13332737e-02 9.51501846e-01 -1.71285892e+00
-7.72592902e-01 2.09101632e-01 -4.29187030e-01 -4.10581410e-01
4.92580891e-01 7.69183755e-01 5.38127780e-01 -6.36345506e-01
6.92690253e-01 1.62273502e+00 5.81983507e-01 8.97374570e-01
-1.65812480e+00 -8.36526990e-01 -2.03076661e-01 -4.38801110e-01
-1.40857112e+00 -7.19867110e-01 1.02151656e+00 -5.97375453e-01
4.92923677e-01 5.76088548e-01 7.57605731e-01 1.44530809e+00
6.62327036e-02 2.83541530e-01 1.30819786e+00 -3.71890396e-01
2.58127689e-01 9.05036181e-02 -6.75948203e-01 8.20467353e-01
-1.37212187e-01 4.94858354e-01 -2.32197523e-01 -1.95698455e-01
1.17803180e+00 -3.64896804e-01 -3.28262478e-01 -3.12153518e-01
-7.90612936e-01 8.49858999e-01 3.25491279e-01 -1.50342301e-01
-1.56277984e-01 4.50983971e-01 2.34441049e-02 4.63286430e-01
4.26334977e-01 7.06671953e-01 -1.63160309e-01 1.46299470e-02
-6.29248500e-01 3.93431067e-01 8.30438852e-01 9.02338147e-01
8.76505554e-01 4.98536617e-01 -2.47890487e-01 8.00506294e-01
2.81855822e-01 6.98891938e-01 -3.11035644e-02 -1.47595119e+00
-2.16912910e-01 3.21833134e-01 -5.33617549e-02 -1.04191411e+00
-1.98882520e-02 4.03776057e-02 -9.14110720e-01 9.99789476e-01
4.33634788e-01 -3.68603021e-01 -1.37657368e+00 2.01762080e+00
3.22576791e-01 2.12369591e-01 -8.58589411e-02 9.67968225e-01
7.06006587e-01 7.21603990e-01 2.54094929e-01 -1.06617875e-01
1.08897817e+00 -6.75688028e-01 -6.25943959e-01 -8.99048001e-02
1.10988602e-01 -9.72729802e-01 1.25186133e+00 3.04919571e-01
-1.31847668e+00 -4.51506853e-01 -8.54226828e-01 -2.15270072e-02
-1.81925133e-01 -5.83124794e-02 6.02091014e-01 1.00151610e+00
-1.40576923e+00 4.81201172e-01 -7.66335845e-01 -1.26579389e-01
4.27804768e-01 4.14568335e-01 -4.39043641e-01 -6.37476332e-03
-1.05638540e+00 7.37174511e-01 -1.28659591e-01 -2.63239983e-02
-1.24224985e+00 -7.31904924e-01 -1.00591636e+00 -4.37183008e-02
1.57833755e-01 -1.03865147e+00 1.12745917e+00 -1.30060732e+00
-1.96912956e+00 1.00370049e+00 5.71399517e-02 2.88271066e-02
8.01090479e-01 1.40333593e-01 -2.02796623e-01 1.04533114e-01
-3.18344057e-01 1.21345282e+00 1.29524481e+00 -1.72362590e+00
3.35514933e-01 -1.16015458e-02 3.45469296e-01 1.36218995e-01
-5.36138378e-02 2.59377863e-02 -4.84067023e-01 -7.03592479e-01
-4.88644451e-01 -1.02820015e+00 -2.36378431e-01 3.83941323e-01
-6.02648735e-01 2.14477599e-01 8.70618820e-01 -5.15305221e-01
5.56083798e-01 -2.13412261e+00 2.03393325e-01 2.32662886e-01
2.19426960e-01 1.06646165e-01 -4.17088926e-01 2.15902776e-01
-3.14397126e-01 6.36076689e-01 -3.33835989e-01 -6.63186371e-01
1.68325618e-01 3.09922338e-01 2.83431709e-02 3.10550660e-01
5.90880990e-01 7.80309319e-01 -6.03897095e-01 -4.44322288e-01
3.73035580e-01 1.00920320e+00 -9.86940324e-01 3.90033305e-01
-5.75872600e-01 9.01657343e-01 -2.39182293e-01 5.06906867e-01
7.78779030e-01 2.76967008e-02 7.16179982e-02 -3.19163710e-01
1.34682789e-01 -1.15308300e-01 -9.93295491e-01 1.45829618e+00
-4.69085902e-01 5.33666849e-01 1.72544643e-01 -3.87632251e-01
9.24100816e-01 3.51781934e-01 2.16359541e-01 -3.79489541e-01
4.16002572e-01 -1.00149080e-01 1.43076954e-02 -2.60517836e-01
2.57137448e-01 -4.92629915e-01 1.10274829e-01 3.19416404e-01
-7.27288797e-02 -9.23322856e-01 -2.03505475e-02 1.19262516e-01
9.40652788e-01 3.12108487e-01 -1.00479379e-01 -1.51955992e-01
1.90779448e-01 -3.66942406e-01 3.99526298e-01 3.43268603e-01
3.54663841e-02 1.10927463e+00 7.78549433e-01 -1.65245131e-01
-1.35761034e+00 -1.15282941e+00 1.41189620e-01 4.77818727e-01
-1.61623076e-01 -4.17128593e-01 -1.05833602e+00 -4.20089811e-01
-7.74121583e-02 8.82711351e-01 -9.45119262e-01 -2.04692572e-01
-5.15232205e-01 -4.09283519e-01 6.62966967e-01 2.78363883e-01
3.66730988e-01 -1.02777267e+00 -3.03472877e-01 -2.71418482e-01
2.02097818e-01 -8.45890284e-01 -6.42297447e-01 -2.38071039e-01
-3.53702217e-01 -8.41821432e-01 -7.39008904e-01 -3.98783088e-01
8.51871014e-01 -3.23531270e-01 1.29249430e+00 2.63888329e-01
-3.87423575e-01 3.57851893e-01 -3.24348249e-02 -3.52193534e-01
-9.05783355e-01 -2.77655691e-01 -1.47606423e-02 -7.39039630e-02
-3.99574786e-01 -8.41026306e-01 -6.54935062e-01 4.00886685e-01
-1.07669997e+00 4.61298496e-01 1.00825794e-01 5.88453472e-01
4.27227348e-01 -2.50247419e-01 1.58946827e-01 -1.02910972e+00
5.70321977e-01 -2.30859488e-01 -5.11734545e-01 2.26530600e-02
-3.65346879e-01 1.44397855e-01 6.59519911e-01 -6.98605359e-01
-1.14174879e+00 3.41699570e-02 -2.76115865e-01 -1.05549324e+00
-3.54046464e-01 -1.24556184e-01 -6.51333272e-01 -1.23578861e-01
7.60758102e-01 -1.78035825e-01 1.92626700e-01 -3.11698943e-01
5.98981619e-01 1.09741941e-01 4.62541223e-01 -9.73150074e-01
1.10660243e+00 3.86355460e-01 7.50759766e-02 -5.77967942e-01
-3.25208604e-01 5.74174106e-01 -3.14212590e-01 -3.31289947e-01
8.99879873e-01 -6.58237994e-01 -7.18672454e-01 5.10110855e-01
-1.10810328e+00 -8.41462433e-01 -4.30400342e-01 1.29447989e-02
-7.87518680e-01 -4.66025211e-02 -5.29975951e-01 -8.62084329e-01
-8.64638910e-02 -1.25421715e+00 1.07113194e+00 2.17981488e-01
-4.65482652e-01 -8.93325031e-01 -2.52149086e-02 1.40075296e-01
5.04298925e-01 9.63385701e-01 8.64008129e-01 9.68567329e-04
-4.96729106e-01 2.09078863e-01 -8.43244717e-02 4.40466046e-01
3.50997984e-01 5.48910141e-01 -1.13738847e+00 -2.31868684e-01
-2.47073621e-01 -2.76798695e-01 4.77665037e-01 2.24754468e-01
1.09775233e+00 -6.29903197e-01 -7.11225271e-02 8.96106064e-01
1.15967500e+00 2.58146971e-01 7.99960315e-01 -2.54547507e-01
8.58954966e-01 6.13993347e-01 -3.69923003e-02 3.35204333e-01
-6.78826272e-02 6.19439125e-01 5.01692712e-01 -3.09734881e-01
-4.59977597e-01 -5.21182716e-01 3.49244058e-01 2.49807239e-01
-2.85386890e-01 -5.34466445e-01 -7.80547857e-01 -2.81768646e-02
-1.12194383e+00 -9.41580296e-01 1.85343876e-01 1.88107789e+00
9.74280119e-01 1.79183055e-02 4.22928408e-02 -2.33278070e-02
6.71228290e-01 1.03114685e-02 -5.67809045e-01 -6.10277832e-01
-1.56951159e-01 4.54534680e-01 5.34690060e-02 8.00445378e-01
-6.26485229e-01 1.01763153e+00 6.19460773e+00 6.82711840e-01
-1.34123003e+00 -5.99321872e-02 1.06768012e+00 -2.13446707e-01
-8.23227286e-01 1.22934729e-01 -2.58819789e-01 4.85711515e-01
6.29677832e-01 -1.23455741e-01 7.52474189e-01 5.74004173e-01
3.79587889e-01 2.13675946e-01 -1.14990008e+00 8.59209001e-01
4.44669202e-02 -1.29474759e+00 3.67150217e-01 2.39983574e-01
7.84385264e-01 -5.61612010e-01 3.45468193e-01 1.45283997e-01
6.35925114e-01 -1.52643859e+00 8.60529840e-01 5.97949862e-01
1.37360668e+00 -8.49935472e-01 7.21001402e-02 -1.15933465e-02
-7.10193098e-01 4.58245486e-01 6.40887916e-02 5.45978807e-02
3.04498971e-01 3.20159167e-01 -7.39558399e-01 2.06316456e-01
4.40097809e-01 2.19447419e-01 -5.13132274e-01 5.46294630e-01
-4.02688086e-01 6.43366933e-01 -4.57610786e-01 3.38662058e-01
-3.79259549e-02 -3.47002476e-01 5.76866806e-01 9.96000409e-01
3.35232794e-01 3.24514300e-01 4.66401093e-02 1.31184554e+00
-1.53371274e-01 -1.45921722e-01 -7.38340437e-01 -2.67087758e-01
2.52849311e-01 1.15405941e+00 -5.39453685e-01 -1.04628474e-01
1.07364640e-01 1.03146374e+00 8.84763896e-03 4.95524526e-01
-9.11479652e-01 5.71120605e-02 1.04435980e+00 3.73071402e-01
-9.31539461e-02 -2.96160817e-01 -2.84689724e-01 -9.64465439e-01
-4.73312289e-01 -1.04148865e+00 -9.98590067e-02 -1.26448560e+00
-1.19438708e+00 7.46027172e-01 1.27465904e-01 -8.26896191e-01
-5.51420450e-01 -4.89889503e-01 -8.08503747e-01 1.00285149e+00
-1.08915246e+00 -1.34776020e+00 -2.81503826e-01 5.69312215e-01
3.50787371e-01 4.22842838e-02 1.00398111e+00 9.43167061e-02
-5.56753218e-01 8.13377261e-01 -4.70387846e-01 1.10728934e-01
7.95354843e-01 -1.10802078e+00 5.21008909e-01 6.01124227e-01
-7.03577325e-02 6.20804071e-01 1.09074295e+00 -5.26228964e-01
-1.37908888e+00 -9.40623701e-01 2.76715815e-01 -5.41922212e-01
1.82310089e-01 -4.60176647e-01 -7.08754480e-01 5.70232391e-01
5.30026495e-01 -4.91481693e-03 5.91352642e-01 -2.17265263e-01
-5.02011955e-01 1.76122159e-01 -1.49319530e+00 9.18577671e-01
1.11211586e+00 -3.56842548e-01 5.03141694e-02 1.53132752e-01
7.39515483e-01 -5.04973233e-01 -7.73280084e-01 2.84461021e-01
5.47181964e-01 -1.15983748e+00 1.03453887e+00 -4.71526742e-01
5.64148545e-01 -3.94939840e-01 -6.47839904e-02 -1.57610214e+00
-1.84692457e-01 -1.11348784e+00 1.80979311e-01 1.50236535e+00
2.35596269e-01 -5.47729969e-01 6.55670702e-01 8.72221053e-01
-7.96103254e-02 -5.13142467e-01 -4.06785131e-01 -6.49784327e-01
3.15244675e-01 -3.91380221e-01 9.26567197e-01 1.02469110e+00
-5.50792515e-01 3.17122012e-01 -4.18081522e-01 1.46870047e-01
5.06152391e-01 -7.06014112e-02 1.04140103e+00 -9.42553639e-01
-4.24648941e-01 -5.69821954e-01 -2.18170300e-01 -5.89191914e-01
3.39036226e-01 -5.94458699e-01 -1.06252946e-01 -1.03845465e+00
-1.07594401e-01 -7.43226230e-01 3.85072887e-01 8.23242605e-01
5.44459000e-03 6.24598145e-01 4.82286155e-01 3.22136283e-02
2.07111090e-01 5.70721805e-01 1.76916611e+00 -8.69308859e-02
-5.71270585e-02 -3.21178019e-01 -7.18978226e-01 6.57869518e-01
9.63863432e-01 -2.10808054e-01 -6.22002244e-01 -4.44805980e-01
4.79530310e-03 1.52104899e-01 6.77829444e-01 -8.17409813e-01
-3.03165525e-01 -3.80941242e-01 6.91002667e-01 2.35192627e-01
8.75900924e-01 -5.72537601e-01 8.70035112e-01 1.87238887e-01
-2.24706918e-01 -3.79980840e-02 3.98480535e-01 1.46852255e-01
1.06287554e-01 6.55456260e-02 1.01584888e+00 -1.57682061e-01
-1.17994420e-01 4.69171733e-01 -1.87665477e-01 1.48833424e-01
7.04803884e-01 -2.29963824e-01 -1.22990005e-01 -8.35291624e-01
-8.52268696e-01 -1.15893379e-01 1.18712080e+00 3.02786082e-01
4.35814142e-01 -1.34314132e+00 -7.91148841e-01 6.64424837e-01
-2.20613763e-01 8.07733759e-02 1.46118954e-01 2.31700078e-01
-6.40178263e-01 -4.08733517e-01 -3.27786207e-01 -4.62692201e-01
-1.17777169e+00 4.63402092e-01 4.47095066e-01 1.84467778e-01
-3.14568073e-01 8.82078826e-01 5.52415192e-01 -3.83839130e-01
-2.07940429e-01 -1.88148990e-02 1.39654979e-01 -2.23452076e-01
2.33733252e-01 -6.94604591e-02 -2.39964306e-01 -6.26070261e-01
-1.14147097e-01 5.14708102e-01 3.46500486e-01 -4.13349330e-01
1.13549888e+00 1.43096969e-01 9.89735723e-02 3.32530662e-02
1.21258473e+00 1.87715590e-01 -1.72078359e+00 5.13688028e-01
-8.59656036e-01 -5.88668644e-01 -9.86816809e-02 -9.64100063e-01
-1.48106766e+00 8.49271297e-01 4.98732209e-01 2.32909590e-01
1.25033963e+00 -1.33169383e-01 3.88641357e-01 -2.48945534e-01
1.10619426e-01 -3.89560133e-01 1.74069479e-01 2.71567225e-01
1.28222716e+00 -9.00299609e-01 -2.50496358e-01 -7.02660322e-01
-5.57416499e-01 8.88177991e-01 8.07884514e-01 -1.45547181e-01
6.02281988e-01 8.19825709e-01 3.18704844e-01 -1.45535078e-02
-7.27284074e-01 1.54596850e-01 2.19705060e-01 7.68132746e-01
4.49430078e-01 5.07143699e-02 1.72360137e-01 1.13249406e-01
-7.52526164e-01 -9.35274288e-02 3.71989220e-01 5.68381011e-01
1.35659680e-01 -1.24153638e+00 -5.32467246e-01 4.04986627e-02
-1.81877643e-01 6.76501468e-02 -5.53498924e-01 7.76838779e-01
2.70169616e-01 7.96541512e-01 1.80084422e-01 -4.48869318e-01
1.54076502e-01 8.23105723e-02 8.18270445e-01 -4.25811321e-01
-5.36135852e-01 8.44574645e-02 5.74229695e-02 -3.66039157e-01
-2.13505685e-01 -5.48884690e-01 -1.11151242e+00 -7.53646910e-01
-8.05231333e-02 1.43087860e-02 6.58587694e-01 5.56624770e-01
4.41350639e-01 4.22243953e-01 7.33207524e-01 -1.24706697e+00
-1.13487631e-01 -7.50976324e-01 -2.09835410e-01 5.82441390e-01
2.05046773e-01 -6.89159870e-01 -5.50904334e-01 4.53883082e-01] | [12.441621780395508, -0.330223023891449] |
f9807d11-d10c-4b7c-a273-58cd26f995c9 | shifts-2-0-extending-the-dataset-of-real | 2206.15407 | null | https://arxiv.org/abs/2206.15407v2 | https://arxiv.org/pdf/2206.15407v2.pdf | Shifts 2.0: Extending The Dataset of Real Distributional Shifts | Distributional shift, or the mismatch between training and deployment data, is a significant obstacle to the usage of machine learning in high-stakes industrial applications, such as autonomous driving and medicine. This creates a need to be able to assess how robustly ML models generalize as well as the quality of their uncertainty estimates. Standard ML baseline datasets do not allow these properties to be assessed, as the training, validation and test data are often identically distributed. Recently, a range of dedicated benchmarks have appeared, featuring both distributionally matched and shifted data. Among these benchmarks, the Shifts dataset stands out in terms of the diversity of tasks as well as the data modalities it features. While most of the benchmarks are heavily dominated by 2D image classification tasks, Shifts contains tabular weather forecasting, machine translation, and vehicle motion prediction tasks. This enables the robustness properties of models to be assessed on a diverse set of industrial-scale tasks and either universal or directly applicable task-specific conclusions to be reached. In this paper, we extend the Shifts Dataset with two datasets sourced from industrial, high-risk applications of high societal importance. Specifically, we consider the tasks of segmentation of white matter Multiple Sclerosis lesions in 3D magnetic resonance brain images and the estimation of power consumption in marine cargo vessels. Both tasks feature ubiquitous distributional shifts and a strict safety requirement due to the high cost of errors. These new datasets will allow researchers to further explore robust generalization and uncertainty estimation in new situations. In this work, we provide a description of the dataset and baseline results for both tasks. | ['Elena Volf', 'Efi Tsompopoulou', 'Vasileios Tsarsitalidis', 'Eli Sivena', 'Francesco La Rosa', 'Vatsal Raina', 'Antonis Nikitakis', 'Nataliia Molchanova', 'Po-Jui Lu', 'Konstantinos Kyriakopoulos', 'Nikolay Kartashev', 'Mara Graziani', 'Cristina Granziera', 'Mark J. F. Gales', 'Meritxell Bach Cuadra', 'Muhamed Barakovic', 'Andreas Athanasopoulos', 'Andrey Malinin'] | 2022-06-30 | null | null | null | null | ['weather-forecasting'] | ['miscellaneous'] | [ 2.36655623e-01 -4.67628315e-02 -3.05135041e-01 -4.28744107e-01
-8.48586261e-01 -7.55583525e-01 9.02822912e-01 1.09781899e-01
-5.08268416e-01 9.41964149e-01 -1.58883199e-01 -5.03995776e-01
-4.49423283e-01 -4.61400658e-01 -8.87798965e-01 -8.24770629e-01
-1.59660831e-01 4.84968394e-01 -1.02825963e-03 2.07019880e-01
1.55299082e-01 4.13230509e-01 -1.52793491e+00 8.34896509e-03
9.28225994e-01 1.19885647e+00 2.46573031e-01 2.47498736e-01
1.61961928e-01 2.89253771e-01 -5.38524032e-01 -5.28303802e-01
3.96862537e-01 -6.15278743e-02 -5.14996767e-01 -1.84768707e-01
4.51934814e-01 4.71095443e-02 -8.32895457e-04 1.03274238e+00
9.01602507e-01 1.72556296e-01 1.09616768e+00 -1.52421725e+00
-3.78165424e-01 5.40191054e-01 -5.00708282e-01 1.86680764e-01
1.92750543e-01 3.63719344e-01 7.80257761e-01 -4.82946485e-01
6.31336331e-01 1.13574600e+00 7.94944048e-01 3.65720958e-01
-1.20740211e+00 -6.79610014e-01 1.88899413e-01 1.44667119e-01
-1.13211489e+00 -2.47218549e-01 5.90637922e-01 -9.00359690e-01
7.75623620e-01 1.41873583e-01 2.10850447e-01 1.73457086e+00
6.41915679e-01 5.81131876e-01 1.49703526e+00 -1.58336297e-01
5.73081434e-01 3.77567351e-01 -4.14683111e-02 2.08260044e-01
4.40062791e-01 4.13126230e-01 -3.92133951e-01 -9.76174027e-02
1.33796290e-01 -2.61962652e-01 -9.88385975e-02 -6.65468574e-01
-1.25415671e+00 7.60513544e-01 1.49539560e-01 1.79884374e-01
-2.43356392e-01 9.42226201e-02 7.40908921e-01 2.55873263e-01
7.19584167e-01 4.44147646e-01 -8.14559937e-01 -1.80392370e-01
-8.26724708e-01 5.09096026e-01 7.32629716e-01 9.44102466e-01
3.90904695e-01 -1.19118914e-01 -1.64016411e-01 6.63839221e-01
2.15114549e-01 7.44604170e-01 4.13941056e-01 -9.08062458e-01
5.78800559e-01 1.51952013e-01 -6.79328740e-02 -9.72396255e-01
-7.19221652e-01 -5.14408886e-01 -6.98152423e-01 3.20088387e-01
5.31349540e-01 -3.33538532e-01 -7.63998449e-01 1.91478157e+00
2.95652241e-01 -2.34116614e-02 2.13821717e-02 6.42098844e-01
6.34680152e-01 1.81100994e-01 1.50454059e-01 -8.82852897e-02
1.26285553e+00 -3.91150862e-01 -4.91418689e-01 -3.57314318e-01
6.80706024e-01 -5.38616121e-01 1.23991454e+00 4.34775531e-01
-7.55988002e-01 -3.80350798e-01 -9.45026577e-01 2.26929635e-01
-5.67022979e-01 -1.93285197e-01 4.84802336e-01 8.87364686e-01
-3.51429492e-01 7.26671755e-01 -6.50448620e-01 -3.23973984e-01
6.01800978e-01 1.23079523e-01 -2.63332278e-01 -1.39739797e-01
-1.26154840e+00 1.32822049e+00 2.67276585e-01 -8.48712400e-02
-8.70419800e-01 -1.08798492e+00 -9.82061803e-01 -4.09766376e-01
3.63152266e-01 -6.67769432e-01 1.19194436e+00 -5.14513493e-01
-1.22002172e+00 9.06954646e-01 4.02633041e-01 -5.84858656e-01
9.95841682e-01 -4.50898930e-02 -3.91101897e-01 -4.61488187e-01
2.81224519e-01 5.09385109e-01 7.26474464e-01 -9.93196130e-01
-7.33067513e-01 -6.34053171e-01 -7.01190531e-02 -3.13227959e-02
1.83459669e-01 -2.29185820e-01 7.81450048e-02 -5.09277821e-01
-4.22479779e-01 -1.13045740e+00 -1.62206188e-01 -4.03474420e-01
-4.67735142e-01 -1.56879053e-01 5.56727111e-01 -4.40798819e-01
9.23087895e-01 -2.10141492e+00 1.20532088e-01 -1.61936358e-02
-6.80934563e-02 -2.74107873e-01 -3.37385722e-02 1.10253282e-01
-1.07805081e-01 2.19890893e-01 -5.02237856e-01 -2.86553264e-01
4.54666436e-01 2.05387995e-01 -1.74057737e-01 8.01322818e-01
3.03337038e-01 8.61635923e-01 -7.60724902e-01 -2.94336945e-01
3.26423228e-01 1.41538382e-01 -2.90304631e-01 -2.73333758e-01
-3.07047367e-01 7.04293251e-01 -3.79474312e-01 5.27761638e-01
6.38701916e-01 2.55254000e-01 -1.72068283e-01 -2.90111601e-01
-3.77934240e-02 1.88886419e-01 -1.03707397e+00 1.86565828e+00
-5.80661535e-01 6.28052294e-01 -1.11592084e-01 -1.11607075e+00
6.37592912e-01 4.52214219e-02 6.37846351e-01 -9.87168312e-01
2.64012516e-01 2.79264659e-01 1.47674605e-01 -6.89575613e-01
3.65242094e-01 -4.85643923e-01 -4.92589414e-01 1.25375420e-01
-3.22597623e-02 -5.98743320e-01 1.43866315e-01 -2.94146270e-01
9.17743504e-01 2.93099165e-01 3.44760977e-02 -6.15010083e-01
1.54520780e-01 3.09701543e-02 6.43730283e-01 4.89311427e-01
-3.00379425e-01 5.38463414e-01 5.58686495e-01 -2.05451474e-01
-1.03692901e+00 -1.19072044e+00 -8.57326090e-01 8.38223159e-01
-5.53230103e-03 5.49528096e-03 -6.56743586e-01 -8.93365502e-01
4.88564700e-01 1.17127323e+00 -6.69432819e-01 -3.78473073e-01
-2.17411920e-01 -1.09076107e+00 5.70266068e-01 4.80932862e-01
1.24960579e-01 -7.53105521e-01 -7.96643972e-01 -1.16875328e-01
2.53087655e-02 -1.21804869e+00 -8.50202888e-02 4.28146809e-01
-7.20188141e-01 -1.10588706e+00 -6.71896994e-01 -2.12194473e-01
1.22350864e-01 -3.69268239e-01 1.32187617e+00 -6.95127249e-01
-3.51007134e-01 4.22473013e-01 -5.75905442e-02 -1.06971180e+00
-3.56444180e-01 3.43531281e-01 3.23991984e-01 -2.77791768e-01
3.15817088e-01 -3.29463691e-01 -4.40033168e-01 4.94532168e-01
-9.74140465e-01 -4.61476445e-01 5.62852263e-01 6.29604220e-01
5.43442369e-01 2.13196408e-02 9.71280336e-01 -7.73772538e-01
7.07931817e-01 -8.74290168e-01 -5.64186275e-01 1.88368291e-01
-8.07022512e-01 1.34218901e-01 2.90152341e-01 -4.93159115e-01
-8.60345602e-01 -1.35099396e-01 -1.61992162e-02 -2.32770309e-01
-3.45848382e-01 5.03210783e-01 -2.31719896e-01 3.61146778e-01
8.73518944e-01 -3.40819329e-01 8.80124569e-02 -3.88016492e-01
3.51132065e-01 7.29521632e-01 5.55499375e-01 -9.49604034e-01
6.00424230e-01 1.46285251e-01 2.24038288e-01 -7.66758978e-01
-6.01266742e-01 -7.37643838e-02 -6.88854694e-01 -1.43345535e-01
9.36746716e-01 -7.90835679e-01 -4.16011214e-01 2.63477266e-01
-7.87522852e-01 -6.40791893e-01 -6.25241458e-01 7.57799029e-01
-8.47596288e-01 2.00720623e-01 -1.73957676e-01 -8.23339999e-01
3.02067697e-02 -1.59139919e+00 1.16644263e+00 -1.78776264e-01
-3.94967645e-01 -1.14614213e+00 -7.82496110e-02 3.29267174e-01
3.53648543e-01 7.67174184e-01 1.20120895e+00 -7.84786224e-01
-6.36151135e-02 -1.36824235e-01 6.79311231e-02 5.60841680e-01
5.40722832e-02 -2.02264190e-01 -1.14481413e+00 -3.41197073e-01
-2.27545612e-02 -3.91621202e-01 8.03793907e-01 6.76926970e-01
1.13893282e+00 2.10813791e-01 -3.21208060e-01 2.76404679e-01
9.45327759e-01 1.21104166e-01 3.48327130e-01 4.52250749e-01
3.17407936e-01 1.08007205e+00 7.90551126e-01 2.12191001e-01
2.83194244e-01 6.92566216e-01 5.53078353e-01 1.15412697e-01
6.89142272e-02 -1.55443829e-02 4.52371657e-01 4.25104976e-01
2.40044281e-01 -2.42799386e-01 -1.07278335e+00 4.02205974e-01
-1.69096220e+00 -7.66959131e-01 -1.87015191e-01 2.57073355e+00
6.21011794e-01 2.44376108e-01 1.81545481e-01 1.28391847e-01
4.88893837e-01 1.42488116e-03 -8.84761155e-01 -4.00732368e-01
-1.55831724e-02 -1.27130166e-01 7.52689540e-01 1.87769786e-01
-1.17932463e+00 2.73274481e-01 6.57689428e+00 9.98365641e-01
-9.96625245e-01 2.17230424e-01 9.29192781e-01 -3.75950336e-01
-4.36487496e-01 -4.63236153e-01 -6.53394222e-01 6.76013350e-01
1.27959263e+00 -4.25791927e-02 2.05255702e-01 6.66749001e-01
7.43889585e-02 -2.43947148e-01 -1.52469885e+00 9.43451345e-01
-3.30892317e-02 -7.70229518e-01 -3.93391341e-01 2.14086577e-01
6.40243053e-01 3.66944939e-01 4.54425663e-01 5.09750485e-01
1.88015759e-01 -1.27171504e+00 1.14270508e+00 4.78206009e-01
9.34484541e-01 -7.36722410e-01 8.66407454e-01 5.36500812e-01
-5.80920339e-01 -2.84687251e-01 -3.77934605e-01 9.26960856e-02
1.53666839e-01 9.60153639e-01 -5.78735530e-01 8.58695030e-01
9.31099534e-01 4.64613885e-01 -3.24222535e-01 8.50552499e-01
9.58261266e-02 3.86909723e-01 -4.60024953e-01 8.29968452e-02
1.38014555e-01 -1.77487478e-01 5.31318545e-01 1.16904294e+00
5.06974578e-01 -5.27080536e-01 3.73961329e-02 1.02507293e+00
1.86118521e-02 1.92463193e-02 -9.66728985e-01 4.61226702e-02
2.73721486e-01 1.18433070e+00 -4.88692909e-01 2.08574399e-01
-4.71337438e-01 4.93608713e-01 -6.09314740e-02 2.19273984e-01
-1.12199938e+00 -1.11266807e-01 9.26231682e-01 2.38747761e-01
-1.56209946e-01 -2.69547105e-01 -4.88210917e-01 -8.49836290e-01
6.63356632e-02 -8.13129008e-01 3.01301658e-01 -4.84101117e-01
-1.61661291e+00 3.81444991e-01 5.15340686e-01 -1.07286167e+00
-4.27652031e-01 -1.01294732e+00 -2.51184016e-01 8.33536088e-01
-1.61655724e+00 -9.75797772e-01 -8.26738700e-02 2.70371705e-01
4.72147763e-01 -2.78495312e-01 4.82205719e-01 4.07257617e-01
-7.55788684e-01 3.65874410e-01 2.25098997e-01 -3.31468344e-01
9.70572233e-01 -1.31009972e+00 1.84341863e-01 3.71319830e-01
-3.98237035e-02 2.56263673e-01 8.15668464e-01 -6.16572261e-01
-1.17503583e+00 -1.19852865e+00 3.57543558e-01 -9.82526660e-01
6.92504168e-01 -5.21256328e-01 -4.95026261e-01 5.61888874e-01
4.19273302e-02 4.60201642e-03 6.45342708e-01 1.17123283e-01
-1.16486721e-01 -1.79090425e-01 -1.35688055e+00 3.88486385e-01
1.07367015e+00 -4.70753998e-01 -4.94167805e-01 4.30328459e-01
5.71759999e-01 -4.51673836e-01 -1.14482093e+00 8.45843017e-01
5.44821441e-01 -8.53504658e-01 9.67832804e-01 -7.36567974e-01
4.24485326e-01 -5.88071346e-03 -4.79586422e-01 -1.49157476e+00
-3.10881566e-02 -1.55345619e-01 1.90718457e-01 1.43598318e+00
8.14946771e-01 -7.10557163e-01 4.95911777e-01 9.18407261e-01
-3.24188709e-01 -8.35393071e-01 -1.20230007e+00 -1.07228065e+00
7.87359953e-01 -9.44014907e-01 6.19016647e-01 8.36767554e-01
-2.31624112e-01 2.27123365e-01 -1.88220777e-02 4.02013697e-02
5.57944775e-01 -3.68327498e-02 6.28168166e-01 -1.47430384e+00
1.21566132e-02 -5.95329404e-01 -4.60417926e-01 -4.46395278e-01
4.70587224e-01 -1.14904451e+00 4.89096306e-02 -1.24273539e+00
5.81357405e-02 -6.45218849e-01 -2.10051760e-01 8.28575492e-02
1.25871077e-01 2.27082372e-02 8.34209472e-02 -5.39402477e-02
-2.64726162e-01 4.64964837e-01 1.02379918e+00 -3.90414566e-01
-2.04775799e-02 2.69981712e-01 -5.32065094e-01 7.64519155e-01
9.67335641e-01 -4.36792225e-01 -6.78175092e-01 -3.78051341e-01
3.17696154e-01 -2.48848453e-01 4.83309478e-01 -9.29763913e-01
-1.53566942e-01 -1.66019604e-01 3.71671736e-01 -3.18889081e-01
1.31141424e-01 -1.06584251e+00 2.21692353e-01 3.53806764e-01
-3.26865554e-01 2.52309859e-01 3.54041070e-01 6.19842350e-01
1.05330624e-01 -2.99563617e-01 7.33224690e-01 -4.67014909e-02
-5.02044737e-01 3.48034382e-01 -2.49565348e-01 4.39665079e-01
1.43831170e+00 -1.02806717e-01 -3.10836017e-01 -1.55702591e-01
-5.77269495e-01 2.45475203e-01 4.18567955e-01 6.67441726e-01
4.54917923e-02 -1.31863379e+00 -7.41840601e-01 6.79232404e-02
2.93314993e-01 2.13755995e-01 3.26932162e-01 1.16352606e+00
-4.66628559e-02 4.64299172e-01 -2.19846457e-01 -8.01112711e-01
-7.40914166e-01 7.45737493e-01 4.40873802e-01 -8.78329873e-02
-3.60499054e-01 4.12581921e-01 3.25944841e-01 -8.02061021e-01
2.18452603e-01 -8.76102269e-01 1.07642442e-01 3.63281310e-01
1.57297507e-01 5.85161507e-01 3.41242343e-01 -5.15355706e-01
-4.23667014e-01 4.42238927e-01 3.66311729e-01 3.06539107e-02
1.24548733e+00 -2.20821276e-01 2.18843952e-01 8.41333807e-01
1.11432624e+00 -1.03528753e-01 -1.40083563e+00 -2.09265742e-02
2.90441245e-01 -1.99478447e-01 1.31000904e-02 -1.01876295e+00
-9.67011094e-01 1.02039754e+00 8.10704410e-01 2.05841228e-01
8.17676425e-01 9.30375084e-02 4.03348446e-01 7.18889534e-02
6.11501575e-01 -1.25770533e+00 -1.86216965e-01 2.52232224e-01
1.05041599e+00 -1.37271619e+00 -4.19173799e-02 -6.41200393e-02
-7.17596292e-01 6.93185151e-01 3.37346137e-01 3.06254059e-01
6.81496263e-01 6.29270434e-01 -1.38756245e-01 -9.36395228e-02
-5.28756976e-01 -9.07529220e-02 4.42113966e-01 9.30605948e-01
3.07023674e-01 1.34703949e-01 -2.82657385e-01 9.39278603e-01
-4.16039973e-01 -1.79437414e-01 1.64417028e-01 6.90698862e-01
-2.85505243e-02 -9.83571768e-01 -1.70229450e-01 6.57341540e-01
-4.87679899e-01 2.83677161e-01 -1.88076928e-01 9.46663976e-01
3.46076787e-01 8.65692258e-01 2.88638156e-02 -4.54961330e-01
7.12729752e-01 2.60273516e-01 5.08877754e-01 -3.61288011e-01
-3.26985091e-01 -4.73335803e-01 2.40407556e-01 -4.74477440e-01
-3.71700883e-01 -9.46882010e-01 -1.01724052e+00 -3.18491012e-01
-3.35562289e-01 -8.66782218e-02 9.93188202e-01 8.42431009e-01
3.15688938e-01 7.29525030e-01 3.39503050e-01 -1.04707837e+00
-1.07395184e+00 -1.11991251e+00 -8.14625442e-01 5.40728867e-01
1.68004274e-01 -1.09978330e+00 -3.42335522e-01 -3.05609196e-01] | [8.2706937789917, 3.849998712539673] |
a30f526c-13a7-487c-9e16-13afc64a9b59 | tjudem-a-combination-classifier-for-aspect | null | null | https://aclanthology.org/S15-2131 | https://aclanthology.org/S15-2131.pdf | TJUdeM: A Combination Classifier for Aspect Category Detection and Sentiment Polarity Classification | null | ['Jian-Yun Nie', 'Zhifei Zhang', 'Hongling Wang'] | 2015-06-01 | null | null | null | semeval-2015-6 | ['aspect-category-detection'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.285002708435059, 3.6735422611236572] |
36f2e781-ecbe-4fec-bebc-b074598170eb | actionspotter-deep-reinforcement-learning | 2004.06971 | null | https://arxiv.org/abs/2004.06971v2 | https://arxiv.org/pdf/2004.06971v2.pdf | ActionSpotter: Deep Reinforcement Learning Framework for Temporal Action Spotting in Videos | Summarizing video content is an important task in many applications. This task can be defined as the computation of the ordered list of actions present in a video. Such a list could be extracted using action detection algorithms. However, it is not necessary to determine the temporal boundaries of actions to know their existence. Moreover, localizing precise boundaries usually requires dense video analysis to be effective. In this work, we propose to directly compute this ordered list by sparsely browsing the video and selecting one frame per action instance, task known as action spotting in literature. To do this, we propose ActionSpotter, a spotting algorithm that takes advantage of Deep Reinforcement Learning to efficiently spot actions while adapting its video browsing speed, without additional supervision. Experiments performed on datasets THUMOS14 and ActivityNet show that our framework outperforms state of the art detection methods. In particular, the spotting mean Average Precision on THUMOS14 is significantly improved from 59.7% to 65.6% while skipping 23% of video. | ['Adrien Chan-Hon-Tong', 'Guillaume Vaudaux-Ruth', 'Catherine Achard'] | 2020-04-15 | null | null | null | null | ['action-spotting'] | ['computer-vision'] | [ 4.65502709e-01 -3.15020591e-01 -4.69934553e-01 -7.00931028e-02
-5.54972589e-01 -5.23774028e-01 4.69522774e-01 8.58265758e-02
-7.00887561e-01 4.67916399e-01 2.26412550e-01 4.03400278e-03
-1.08692190e-02 -5.99877834e-01 -7.06559956e-01 -6.10308826e-01
-2.27206081e-01 2.10058540e-01 7.11151540e-01 3.61773074e-01
5.27446747e-01 4.48790401e-01 -1.50451863e+00 4.90426630e-01
6.04574263e-01 1.13092971e+00 4.35407549e-01 6.53259635e-01
3.38238217e-02 1.12229514e+00 -6.37379408e-01 5.19083217e-02
3.30459118e-01 -6.65758312e-01 -6.06826842e-01 5.25835752e-01
5.89777231e-01 -7.84327388e-01 -3.66717517e-01 1.01546335e+00
1.15418874e-01 3.57898921e-01 2.66450435e-01 -1.09041286e+00
7.60104656e-02 7.52424300e-01 -7.69019127e-01 5.56070924e-01
5.45236409e-01 7.40562305e-02 1.07777309e+00 -7.04360604e-01
8.20772171e-01 9.41165209e-01 2.92715967e-01 4.00011837e-01
-9.81913030e-01 -5.98700523e-01 2.77486771e-01 5.71999609e-01
-1.42612934e+00 -5.17032802e-01 5.25715411e-01 -3.88151526e-01
9.03037906e-01 1.82198197e-01 7.36102581e-01 8.78895998e-01
2.94522542e-05 1.22857749e+00 6.71693146e-01 -3.22013140e-01
4.10995692e-01 -4.59503531e-01 -1.63099766e-01 7.15763211e-01
1.27203569e-01 -4.33003843e-01 -7.56004810e-01 4.87971418e-02
8.39365542e-01 2.07817480e-01 -3.89823586e-01 -2.52642184e-01
-1.31336033e+00 4.44408685e-01 1.31165400e-01 2.14240924e-01
-5.46267033e-01 3.47545505e-01 6.63146853e-01 1.89713001e-01
3.33692521e-01 3.96546215e-01 -2.45522574e-01 -5.77147007e-01
-1.26809955e+00 1.67049989e-01 5.10044277e-01 8.10273230e-01
7.04677820e-01 -2.54003257e-01 -3.43183428e-01 6.15379632e-01
-2.45754421e-01 2.66304910e-01 3.07955593e-01 -1.40922117e+00
6.03028893e-01 6.45381331e-01 4.23172593e-01 -1.05510414e+00
-2.42072776e-01 7.28202686e-02 -5.99164963e-01 -1.32093176e-01
5.61765134e-01 -8.54851454e-02 -7.52841115e-01 1.35687196e+00
3.58095855e-01 6.91459119e-01 -4.56504703e-01 1.20136058e+00
2.62950093e-01 6.60355449e-01 1.62886605e-01 -4.72043037e-01
1.25305784e+00 -1.14388776e+00 -6.60938680e-01 -1.97500274e-01
7.01151788e-01 -5.79294264e-01 8.20815384e-01 6.43696845e-01
-9.08202171e-01 -4.64462072e-01 -9.85994637e-01 1.86389223e-01
8.24762881e-02 5.92887938e-01 4.01388913e-01 2.15825379e-01
-7.90341437e-01 7.29406536e-01 -1.30134439e+00 -3.57919455e-01
5.89547813e-01 1.78158835e-01 -3.24726909e-01 -6.51529431e-02
-9.29605246e-01 3.84286016e-01 6.10524118e-01 -1.81911707e-01
-1.07545745e+00 -5.22897363e-01 -6.74796939e-01 2.28408039e-01
1.33342826e+00 -3.00545305e-01 1.42661667e+00 -1.15091312e+00
-1.36538804e+00 5.78952491e-01 -2.14909703e-01 -1.00331724e+00
7.76905179e-01 -7.99689472e-01 -1.29370496e-01 7.95078456e-01
2.78687000e-01 4.97158229e-01 1.01769054e+00 -4.34861839e-01
-1.10820329e+00 -1.85865596e-01 3.01806837e-01 2.48502403e-01
-3.68263900e-01 2.00714141e-01 -1.06582785e+00 -5.63895524e-01
2.47528162e-02 -8.67302656e-01 -2.15350181e-01 7.50126364e-03
-3.84858280e-01 -4.24062967e-01 8.87871861e-01 -8.10022891e-01
1.51129329e+00 -2.08545613e+00 3.04127438e-03 -5.76049797e-02
2.37398282e-01 4.15410697e-01 2.09977552e-02 3.29857886e-01
2.84481883e-01 -5.92370853e-02 -7.90951923e-02 -1.05681352e-01
-1.47130653e-01 -6.72600567e-02 -2.63189971e-01 6.68032527e-01
1.42005265e-01 5.22700965e-01 -1.12037015e+00 -7.59945631e-01
4.64068234e-01 2.67256975e-01 -7.34450161e-01 1.47125825e-01
-4.38852906e-01 3.82350892e-01 -6.44842982e-01 5.74717402e-01
2.89008439e-01 -3.25483441e-01 4.53071028e-01 -1.06155626e-01
-2.11793438e-01 4.19506311e-01 -1.30777502e+00 1.74135852e+00
-2.84246832e-01 7.35764503e-01 -7.91838467e-02 -1.26722014e+00
4.30066407e-01 2.24370897e-01 1.01571071e+00 -6.90368235e-01
1.72059685e-02 -6.91755563e-02 -1.39750928e-01 -5.63272536e-01
4.76152062e-01 4.58615571e-01 -5.38201490e-03 5.68252921e-01
-1.09227009e-01 5.86907625e-01 8.55199695e-01 3.62524241e-01
1.59399056e+00 3.17152679e-01 5.68962157e-01 7.95486048e-02
5.80489874e-01 -1.03135616e-01 7.33414590e-01 8.57972443e-01
-3.02112222e-01 4.62895215e-01 7.10261405e-01 -5.37046969e-01
-6.99704528e-01 -7.36706257e-01 2.94639766e-01 1.25768924e+00
2.51272321e-01 -9.55566704e-01 -1.08323109e+00 -8.57763648e-01
-2.52620727e-01 4.09319282e-01 -5.32932997e-01 -8.53173062e-03
-8.55028629e-01 -3.04158121e-01 2.10072920e-01 5.89451134e-01
8.28746617e-01 -1.11419427e+00 -1.33412445e+00 3.50262672e-01
-4.05527771e-01 -1.49796653e+00 -6.29094183e-01 -2.27631211e-01
-7.58422136e-01 -1.43681800e+00 -6.82047486e-01 -3.42271954e-01
5.64805210e-01 5.21343470e-01 1.04414690e+00 1.21130362e-01
-3.27213168e-01 2.19334260e-01 -6.57832086e-01 1.87094718e-01
-1.83945149e-01 1.71590641e-01 -1.05016381e-01 3.94341826e-01
3.60118151e-01 -2.97949523e-01 -8.85168672e-01 5.41092336e-01
-8.01248848e-01 1.76481605e-01 6.44950449e-01 4.37587798e-01
7.80556858e-01 2.22522348e-01 2.30223790e-01 -8.30417693e-01
1.45039618e-01 -2.83390522e-01 -6.88566506e-01 2.75726527e-01
-1.95446149e-01 -9.71901938e-02 7.14905977e-01 -3.12568158e-01
-9.34958696e-01 4.55814987e-01 1.45120531e-01 -6.81373954e-01
-2.73769021e-01 3.13087583e-01 -9.72442478e-02 3.04491997e-01
4.32303488e-01 4.39203143e-01 -3.74914557e-01 -5.29357791e-01
2.27992430e-01 5.14940560e-01 4.60262835e-01 -2.35588804e-01
2.67822444e-01 7.63595581e-01 -1.90854698e-01 -1.05808604e+00
-9.18711424e-01 -8.81161034e-01 -7.65232742e-01 -4.21078533e-01
8.59975517e-01 -8.35260630e-01 -7.03371942e-01 2.75698155e-01
-1.00575352e+00 -4.89122570e-01 1.23232290e-01 4.31591004e-01
-6.59775615e-01 6.43147290e-01 -5.30524194e-01 -6.65310621e-01
-2.08658978e-01 -1.01090705e+00 1.29861319e+00 5.04652262e-02
-4.22871798e-01 -4.76308227e-01 -2.10339174e-01 4.93007153e-01
3.83943357e-02 1.27357721e-01 1.28730386e-01 -4.75389719e-01
-9.76384521e-01 -2.72946894e-01 -1.85726419e-01 2.29842380e-01
2.70341218e-01 2.95389779e-02 -4.98948157e-01 -1.65164068e-01
-3.26200992e-01 -1.00598775e-01 1.24877071e+00 5.44276476e-01
1.59751225e+00 -2.87439138e-01 -4.32347000e-01 3.45940351e-01
1.14840734e+00 3.27651709e-01 5.43607473e-01 3.42566371e-01
5.71768880e-01 2.98588604e-01 1.10631597e+00 8.99433434e-01
-4.96044084e-02 1.04516828e+00 4.17540908e-01 3.86834145e-01
-2.20382944e-01 -1.59000218e-01 5.71454287e-01 1.97594687e-01
-2.91465908e-01 -1.99443266e-01 -8.38209629e-01 4.44873422e-01
-2.08167386e+00 -1.36954176e+00 2.37149015e-01 2.30843258e+00
6.66463017e-01 4.43291962e-01 4.46293294e-01 2.04020128e-01
8.15137386e-01 4.48147088e-01 -6.13454700e-01 1.09525003e-01
3.50446761e-01 -6.43195659e-02 5.22957206e-01 2.00779095e-01
-1.57086170e+00 1.06158936e+00 5.04344749e+00 8.95750165e-01
-1.05752265e+00 6.76036580e-03 5.61930001e-01 -4.38915581e-01
6.02815747e-01 -6.10855743e-02 -8.35951507e-01 7.30492294e-01
7.17441022e-01 -7.28888959e-02 3.52385670e-01 9.08692956e-01
7.09776521e-01 -6.78727329e-01 -1.08029890e+00 1.09700310e+00
2.45249301e-01 -1.21401012e+00 -4.58757356e-02 -2.91985512e-01
6.11530185e-01 -1.83662400e-01 -4.87856239e-01 9.90124941e-02
9.10733081e-03 -5.67063630e-01 6.68231726e-01 4.16890949e-01
6.02675438e-01 -7.23074794e-01 3.44693094e-01 3.33488494e-01
-1.41581059e+00 -1.49252579e-01 -2.38222599e-01 -1.58317238e-01
3.46300781e-01 6.36171818e-01 -8.47357273e-01 1.34214357e-01
8.06844115e-01 1.22784770e+00 -4.54952359e-01 1.37948895e+00
-4.40532565e-01 8.18943381e-01 -2.82636136e-01 3.89080751e-03
2.61795640e-01 -1.72334239e-01 4.73487377e-01 1.17060268e+00
3.51041466e-01 9.19985548e-02 6.11919761e-01 3.68216068e-01
-1.98317960e-01 -5.24845347e-02 -3.76260966e-01 -2.07820907e-01
3.89932245e-01 1.17462420e+00 -1.25395882e+00 -6.11028075e-01
-4.13300753e-01 1.31594479e+00 1.01627611e-01 1.83736756e-01
-1.19014382e+00 -4.31914389e-01 6.46945059e-01 3.58206928e-01
5.79258263e-01 -3.29010755e-01 2.34089181e-01 -1.12232232e+00
1.89158648e-01 -7.21974432e-01 5.22520125e-01 -5.11981785e-01
-5.43990552e-01 1.83796600e-01 -1.11854792e-01 -1.51595545e+00
-2.54859775e-01 -3.84603292e-01 -3.80544960e-01 -8.00024439e-03
-9.67736363e-01 -5.32751381e-01 -5.36144078e-01 3.77289414e-01
1.15203393e+00 1.74625039e-01 1.17570423e-01 4.54412222e-01
-8.88272643e-01 3.25605512e-01 -1.88366413e-01 3.47838700e-01
6.57036185e-01 -1.06589663e+00 2.76699215e-01 1.01034486e+00
6.08759522e-01 3.45101029e-01 6.40614748e-01 -6.72269344e-01
-1.26371920e+00 -1.13034093e+00 7.44561613e-01 -1.20174207e-01
7.04559922e-01 -1.41922146e-01 -7.63981760e-01 5.87179840e-01
5.81690706e-02 2.41983179e-02 3.57171714e-01 -2.29914635e-01
-7.24463388e-02 -2.24801555e-01 -5.09076595e-01 7.84070075e-01
1.38653362e+00 -3.04361343e-01 -3.82791907e-01 5.55261850e-01
4.76736069e-01 -4.35195118e-01 -5.03121674e-01 1.37517840e-01
4.17251885e-01 -1.12078738e+00 9.70098853e-01 -3.96860123e-01
5.38853467e-01 -4.46985304e-01 1.53070465e-01 -8.74650836e-01
-2.28753351e-02 -7.13033557e-01 -5.13949394e-01 8.02325726e-01
8.38494673e-03 -4.82169352e-02 9.45112109e-01 4.14485902e-01
9.05986130e-03 -6.98868513e-01 -9.35561299e-01 -7.85762727e-01
-8.42197120e-01 -4.80654389e-01 1.92506924e-01 6.43508017e-01
5.89656197e-02 9.48946625e-02 -6.02568567e-01 8.21158960e-02
5.39910853e-01 2.70723850e-01 8.96186650e-01 -8.33123326e-01
-3.74181002e-01 -4.53303427e-01 -5.07787287e-01 -1.53538489e+00
5.91649264e-02 -4.56602454e-01 1.00016199e-01 -1.52006161e+00
2.82909721e-01 -1.50140878e-02 -4.37435985e-01 4.84540015e-01
-4.59870845e-02 1.83659628e-01 2.61561394e-01 2.95726746e-01
-1.47863770e+00 3.48578215e-01 1.01882577e+00 -9.53398198e-02
-2.36315101e-01 1.03059180e-01 -1.42258614e-01 9.68075991e-01
9.42307413e-01 -3.64810765e-01 -4.15047795e-01 -3.46554339e-01
1.24352291e-01 2.44793221e-01 4.01739210e-01 -1.50216627e+00
3.35657924e-01 -3.29112649e-01 1.71227992e-01 -6.77740037e-01
2.15976387e-01 -6.56510830e-01 -2.04764962e-01 5.56695938e-01
-4.11646038e-01 -1.45833120e-01 -1.16890296e-01 7.93676078e-01
-1.76447794e-01 -2.98511863e-01 5.94527483e-01 -1.86570168e-01
-1.40604401e+00 2.43961126e-01 -5.82856536e-01 6.45857751e-02
1.34307325e+00 -1.41344622e-01 5.35758920e-02 -2.62156010e-01
-7.43975997e-01 1.95548937e-01 3.62451613e-01 3.96505028e-01
6.43944144e-01 -1.09040129e+00 -3.28713208e-01 -1.23762958e-01
2.62126606e-02 -2.29170024e-01 2.49136552e-01 1.04992211e+00
-7.39054084e-01 4.53123659e-01 -1.46207780e-01 -5.76776624e-01
-1.47849119e+00 5.55573404e-01 1.89355537e-01 -2.23785624e-01
-8.13524187e-01 6.05903506e-01 1.13745995e-01 6.08574927e-01
4.11013335e-01 -4.58334476e-01 -3.52145314e-01 2.43428499e-01
9.93026197e-01 6.49408996e-01 -5.27120344e-02 -4.53008562e-01
-4.25574809e-01 3.90466392e-01 -8.29964802e-02 1.33180499e-01
1.10465121e+00 -1.43000543e-01 5.31402864e-02 1.16190143e-01
9.98456717e-01 -2.52105743e-01 -1.69711101e+00 -2.10553497e-01
3.81904274e-01 -7.30435193e-01 -3.31461281e-02 -4.94447410e-01
-1.24899125e+00 7.92909622e-01 2.78001577e-01 1.21832371e-01
1.41054475e+00 -6.55808449e-02 1.10277534e+00 5.64934611e-01
5.51599145e-01 -1.44281709e+00 3.91351610e-01 2.81009465e-01
7.41845727e-01 -1.21594727e+00 1.31136119e-01 -4.61326033e-01
-4.56065416e-01 1.02623856e+00 7.33225465e-01 -2.49147967e-01
4.01757836e-01 9.04321596e-02 -3.01964253e-01 -1.01953045e-01
-6.82637274e-01 -3.17556649e-01 1.18097454e-01 2.39129644e-03
1.98634878e-01 3.02415080e-02 -5.33629775e-01 2.31014460e-01
3.76374990e-01 2.50933379e-01 4.91520941e-01 9.51154888e-01
-6.63660109e-01 -7.45251179e-01 -1.21871196e-01 5.66424906e-01
-6.90319598e-01 1.14471436e-01 -3.63423377e-01 5.25219977e-01
-3.04241776e-02 8.37062180e-01 2.04915494e-01 -3.66958171e-01
1.73902705e-01 -8.45902637e-02 3.29561710e-01 -6.62770212e-01
-1.21847644e-01 2.02229500e-01 1.23503022e-01 -1.11210835e+00
-7.52600849e-01 -8.29047322e-01 -1.36341429e+00 -5.52583747e-02
-2.60540564e-02 -5.71756214e-02 1.24589033e-01 9.80480492e-01
4.11890090e-01 4.91110712e-01 4.50231940e-01 -9.00807381e-01
-4.21273589e-01 -6.50828063e-01 -4.66261327e-01 5.73687434e-01
1.36159912e-01 -8.29477370e-01 -1.65338933e-01 4.22538877e-01] | [8.400504112243652, 0.4452205300331116] |
48f32043-f8b5-4172-b327-e344222685ee | conceptdistil-model-agnostic-distillation-of | 2205.03601 | null | https://arxiv.org/abs/2205.03601v1 | https://arxiv.org/pdf/2205.03601v1.pdf | ConceptDistil: Model-Agnostic Distillation of Concept Explanations | Concept-based explanations aims to fill the model interpretability gap for non-technical humans-in-the-loop. Previous work has focused on providing concepts for specific models (eg, neural networks) or data types (eg, images), and by either trying to extract concepts from an already trained network or training self-explainable models through multi-task learning. In this work, we propose ConceptDistil, a method to bring concept explanations to any black-box classifier using knowledge distillation. ConceptDistil is decomposed into two components:(1) a concept model that predicts which domain concepts are present in a given instance, and (2) a distillation model that tries to mimic the predictions of a black-box model using the concept model predictions. We validate ConceptDistil in a real world use-case, showing that it is able to optimize both tasks, bringing concept-explainability to any black-box model. | ['Pedro Bizarro', 'Pedro Saleiro', 'Vladimir Balayan', 'Ricardo Moreira', 'João Bento Sousa'] | 2022-05-07 | null | null | null | null | ['explainable-models'] | ['computer-vision'] | [ 3.11931282e-01 1.08221960e+00 -1.63267732e-01 -4.21606302e-01
-2.69069165e-01 -3.22059780e-01 7.70373106e-01 4.04598027e-01
1.96833119e-01 5.58841825e-01 -6.27458021e-02 -7.36678541e-01
-2.97944605e-01 -7.62388527e-01 -7.19730973e-01 -1.04800232e-01
7.41434619e-02 9.62692499e-01 5.15995286e-02 -1.21174172e-01
3.15096915e-01 3.38829249e-01 -1.89528608e+00 8.53016734e-01
1.21162331e+00 7.57861733e-01 2.27480069e-01 6.11846626e-01
-5.48709512e-01 9.48008657e-01 -4.76558894e-01 -2.89359003e-01
-4.43054959e-02 -4.16355908e-01 -1.16289008e+00 4.83856797e-02
3.38593483e-01 7.91249052e-02 4.89824116e-01 7.10942388e-01
-4.12270308e-01 -1.99212916e-02 8.39560568e-01 -1.72737682e+00
-7.01635182e-01 7.25505769e-01 1.37644410e-01 -3.12109649e-01
3.53420883e-01 5.92804514e-02 9.51620162e-01 -8.05226624e-01
4.72284973e-01 1.59747612e+00 5.65654278e-01 1.08670712e+00
-1.67738748e+00 -5.03757834e-01 2.94694096e-01 4.03986990e-01
-1.05882096e+00 4.37714569e-02 6.12911284e-01 -7.27683425e-01
1.08734655e+00 4.08831000e-01 7.05302835e-01 9.91352439e-01
3.72538306e-02 7.59992838e-01 1.17326736e+00 -7.56890833e-01
5.86815715e-01 7.21072316e-01 3.13926756e-01 6.41254544e-01
5.27524114e-01 2.96245277e-01 -5.38323224e-01 -1.99668765e-01
7.57779539e-01 8.46789628e-02 -1.12756856e-01 -5.90469599e-01
-1.08406830e+00 8.97151828e-01 5.01942039e-01 4.59162503e-01
-3.18634629e-01 1.59783810e-01 -3.31233181e-02 1.84878141e-01
4.57911193e-01 1.16069281e+00 -1.07252073e+00 2.66838193e-01
-8.34965587e-01 3.78378749e-01 1.11352479e+00 9.76422310e-01
1.13603187e+00 -4.13754284e-02 1.25740334e-01 2.24648386e-01
3.00314248e-01 1.93543702e-01 7.92004943e-01 -7.94325113e-01
-3.93387303e-02 1.07558799e+00 -8.88161957e-02 -6.36174500e-01
-3.28722030e-01 -3.78709972e-01 -4.84096855e-01 5.99976838e-01
2.11673021e-01 -1.06380545e-01 -1.02879763e+00 1.46750259e+00
9.16896462e-02 4.14079100e-01 3.73399734e-01 7.22006857e-01
8.20310473e-01 4.49601412e-01 3.48361552e-01 1.06336325e-01
1.33891070e+00 -1.36244714e+00 -2.75914282e-01 -6.38080359e-01
8.29071999e-01 -2.35512927e-01 1.10769987e+00 5.28915107e-01
-8.50467741e-01 -9.66068268e-01 -1.11416686e+00 -5.93362823e-02
-8.00177276e-01 1.06813833e-01 7.52523839e-01 3.86850119e-01
-9.46025610e-01 8.90950441e-01 -5.82529664e-01 -4.43338394e-01
2.82039702e-01 4.41380471e-01 -3.70266050e-01 2.33239513e-02
-8.59504819e-01 1.27277350e+00 9.28756952e-01 -4.96922553e-01
-1.03935277e+00 -1.01908720e+00 -9.32479322e-01 5.71942687e-01
6.65322006e-01 -1.13959563e+00 1.33718956e+00 -1.40530598e+00
-9.96801436e-01 7.18221784e-01 -1.32651389e-01 -7.83727825e-01
2.85341889e-01 -3.85354668e-01 -4.96386111e-01 -2.67016310e-02
1.91210338e-03 1.06995654e+00 9.84762371e-01 -1.76442409e+00
-6.68624222e-01 -2.58920249e-02 4.66823965e-01 -1.39427274e-01
-1.20920762e-01 -4.56649065e-01 1.54592618e-01 -3.61647695e-01
2.26844311e-01 -8.16945851e-01 -4.64931279e-01 2.48296112e-01
-6.50452733e-01 -1.19316317e-01 1.04814470e+00 -3.72302949e-01
8.23026061e-01 -1.76303148e+00 9.00565013e-02 3.32951903e-01
5.95390320e-01 4.00390774e-01 -1.67527348e-01 6.42478317e-02
-7.51954794e-01 5.22727311e-01 -2.08878770e-01 -3.58147919e-01
-3.63202021e-02 5.00768840e-01 -2.81097770e-01 -4.81179208e-01
6.90568030e-01 8.63070369e-01 -8.55369031e-01 -4.33306992e-01
5.52484155e-01 3.35771799e-01 -7.33050704e-01 3.62588644e-01
-9.28827167e-01 3.93471599e-01 -3.52868259e-01 2.52072781e-01
5.08957207e-01 -2.65150398e-01 4.25331742e-01 1.91762391e-02
1.88445255e-01 1.91742182e-01 -1.14606285e+00 1.68275952e+00
-8.14917624e-01 6.08020425e-01 -5.71107864e-01 -1.09752262e+00
1.13193560e+00 3.56781006e-01 2.69528721e-02 -3.28434199e-01
-7.00504929e-02 2.83501863e-01 9.02677476e-02 -6.28569543e-01
2.87221670e-01 -5.65841973e-01 2.88072884e-01 5.17991781e-01
3.32955599e-01 -3.94786894e-01 1.02190934e-01 1.43204480e-01
8.12105954e-01 5.06654739e-01 7.08181441e-01 -5.00745535e-01
6.33094847e-01 3.38736057e-01 2.74276555e-01 7.10399270e-01
3.78572851e-01 7.72003949e-01 5.69459617e-01 -9.68332708e-01
-9.40886199e-01 -8.34874153e-01 2.17041209e-01 6.74499094e-01
-2.41446141e-02 -6.79206491e-01 -9.44320381e-01 -1.02280402e+00
1.06011979e-01 1.46961498e+00 -9.42424953e-01 -3.47307444e-01
-9.96048376e-02 -1.54821901e-02 1.14622347e-01 5.40096521e-01
2.84773171e-01 -1.01431322e+00 -1.08644795e+00 9.38733742e-02
-1.14401720e-01 -9.25723076e-01 2.68712789e-01 4.81900275e-01
-1.24865139e+00 -1.40799105e+00 -3.41997780e-02 -3.68334591e-01
1.03340137e+00 6.65745586e-02 1.53388965e+00 7.86658108e-01
-2.28330582e-01 3.67509961e-01 -4.26861674e-01 -9.77886319e-01
-6.74439609e-01 2.18551327e-02 -2.82471031e-01 -1.77527726e-01
6.17529094e-01 -5.86234272e-01 -1.97929844e-01 2.97356635e-01
-1.06678975e+00 7.54501879e-01 5.87274730e-01 7.96891868e-01
3.21820587e-01 6.33331668e-03 1.81400716e-01 -1.13353491e+00
6.23645067e-01 -3.02470893e-01 -2.70636171e-01 7.73870349e-01
-1.05312812e+00 7.15621233e-01 6.16166592e-01 -5.11526644e-01
-1.07316458e+00 3.49224716e-01 2.57857352e-01 -3.62895012e-01
-6.56243920e-01 5.50824463e-01 -1.56801775e-01 2.12574556e-01
1.18666947e+00 6.27620891e-02 -1.70218609e-02 -4.51494217e-01
6.98690176e-01 2.58843660e-01 4.79469806e-01 -8.96068037e-01
1.06417811e+00 1.86593175e-01 1.72343180e-01 -4.28549349e-01
-1.00551271e+00 -8.63592550e-02 -9.36136067e-01 2.43875328e-02
8.01270962e-01 -6.45215392e-01 -6.26000762e-01 -3.86445045e-01
-1.38377500e+00 -2.30401590e-01 -6.68429434e-01 3.04480851e-01
-7.26026773e-01 -2.19817042e-01 2.15993136e-01 -8.80721748e-01
-6.38629720e-02 -9.06132042e-01 9.35180128e-01 2.26899385e-01
-7.14145482e-01 -1.27276790e+00 -2.08762348e-01 2.43094116e-01
5.09449542e-01 2.70509094e-01 1.44860899e+00 -1.39690197e+00
-5.55117726e-01 -1.63997516e-01 -1.83112383e-01 3.43082041e-01
-5.89775704e-02 -3.90617810e-02 -1.34298742e+00 2.29749411e-01
-3.24823782e-02 -1.88575566e-01 6.45542502e-01 8.15770403e-02
1.45459557e+00 -5.14132857e-01 -6.53500259e-01 1.08273186e-01
1.32171965e+00 1.60802126e-01 7.22724319e-01 2.84564823e-01
3.71300578e-01 7.62546062e-01 6.15947008e-01 3.58919590e-03
2.16146141e-01 6.05135322e-01 6.25537217e-01 -4.55319405e-01
-1.11013345e-01 -5.35126746e-01 -2.99463540e-01 -1.31558493e-01
-1.99396208e-01 8.62649269e-03 -1.08184969e+00 4.82294470e-01
-2.31333685e+00 -7.73448646e-01 -2.87075996e-01 2.00257921e+00
4.75345105e-01 2.92796642e-01 -2.14363858e-02 1.77581072e-01
4.62094665e-01 -6.42567754e-01 -4.29292202e-01 -6.42827630e-01
1.42074108e-01 3.27474296e-01 -3.71143550e-01 6.06977701e-01
-7.80255735e-01 9.28247929e-01 6.40871906e+00 5.92262924e-01
-8.86142075e-01 2.98704933e-02 5.34314275e-01 1.53584406e-01
-5.94436824e-01 6.85558259e-01 -3.80011439e-01 -8.23770743e-03
8.74789357e-01 -4.50600386e-01 2.84784138e-01 1.40130424e+00
9.12627354e-02 -1.29421264e-01 -1.75993025e+00 6.42404318e-01
2.94544727e-01 -1.52698481e+00 6.29113019e-01 -1.02421977e-01
4.76543158e-01 -8.21396768e-01 -1.82074040e-01 5.93344748e-01
2.54652768e-01 -1.38625383e+00 7.50143230e-01 6.02495193e-01
2.56883532e-01 -4.09867585e-01 6.44018769e-01 6.23404562e-01
-9.39105093e-01 -3.62136930e-01 -2.33587727e-01 -3.97778481e-01
-2.51935720e-01 5.36049664e-01 -1.34911668e+00 7.15711296e-01
4.43599492e-01 5.75562298e-01 -8.52653444e-01 9.54200625e-01
-6.73942924e-01 4.01850998e-01 8.31758007e-02 9.80058014e-02
-5.12815602e-02 1.02134243e-01 2.18087271e-01 1.13206589e+00
3.78870279e-01 1.79153785e-01 1.24038197e-03 1.52772439e+00
4.48885202e-01 -2.26809412e-01 -4.55514371e-01 5.23360558e-02
6.08379692e-02 1.31910503e+00 -6.94537878e-01 -8.25432420e-01
-1.18902512e-01 8.68093252e-01 2.98729509e-01 2.73503602e-01
-5.59673786e-01 -2.37165108e-01 6.00090384e-01 3.19359750e-01
8.76814499e-02 3.78261022e-02 -6.29569352e-01 -1.08844209e+00
-2.83233196e-01 -8.74626815e-01 2.09787637e-01 -1.55863500e+00
-9.57310915e-01 8.02656054e-01 3.91709596e-01 -1.19773126e+00
-5.43472052e-01 -9.39642727e-01 -8.29488575e-01 9.06407654e-01
-1.44734621e+00 -1.45810461e+00 -5.93901932e-01 3.49452436e-01
6.77622080e-01 -1.38443917e-01 1.36448014e+00 -4.32791412e-01
1.74859315e-02 8.27805996e-02 -5.89840293e-01 -3.77959251e-01
2.29429096e-01 -1.57864869e+00 5.15737176e-01 5.67051232e-01
3.22166651e-01 8.91683340e-01 1.06835544e+00 -4.63732719e-01
-7.98074543e-01 -9.26181912e-01 1.15013182e+00 -6.18548393e-01
3.66427183e-01 -3.44243973e-01 -1.17107654e+00 8.08168709e-01
5.75837605e-02 -2.97647744e-01 7.61690140e-01 4.51092035e-01
-6.75170422e-01 1.26960009e-01 -1.34347987e+00 6.22576237e-01
8.60825777e-01 -4.25389022e-01 -1.14286554e+00 4.49694395e-01
8.47991526e-01 -2.35483989e-01 -4.36894774e-01 1.79091185e-01
5.53992450e-01 -1.18527699e+00 9.44510639e-01 -1.13880014e+00
9.59910631e-01 -3.95017028e-01 1.64865613e-01 -1.65783370e+00
-2.27854639e-01 -2.03061521e-01 -3.64567220e-01 8.57583106e-01
9.36907589e-01 -4.56413835e-01 7.61392057e-01 1.25157440e+00
-2.22734064e-01 -6.71352506e-01 -5.01770198e-01 -7.90574014e-01
-2.70197541e-03 -5.79129875e-01 9.80321050e-01 9.20318663e-01
3.84387434e-01 6.02609456e-01 -4.68421206e-02 1.27113923e-01
2.93779075e-01 2.20576391e-01 8.58614445e-01 -1.83460569e+00
-4.66734111e-01 -2.54334748e-01 -3.71020585e-01 -8.55818212e-01
1.99973211e-01 -8.11078489e-01 4.78483029e-02 -1.83597910e+00
2.71025121e-01 -2.86166519e-01 -4.28721979e-02 1.10318971e+00
-1.58475474e-01 -4.56069231e-01 5.13271570e-01 1.08144388e-01
-3.16315949e-01 2.10722253e-01 1.18424666e+00 -3.00430238e-01
-1.45039260e-01 5.95483594e-02 -1.17677605e+00 9.19220448e-01
6.73250794e-01 -7.58370459e-01 -7.29044378e-01 -2.16041401e-01
1.91680595e-01 3.71235795e-02 7.85082638e-01 -1.35148549e+00
1.33773968e-01 -3.51750314e-01 3.72562706e-01 -9.71413553e-02
4.19878274e-01 -1.25492597e+00 2.89677173e-01 7.28775978e-01
-6.02558911e-01 -9.82993096e-02 5.73340595e-01 2.67593890e-01
-1.95557788e-01 -6.10190272e-01 3.65142494e-01 -3.15516740e-01
-9.29430842e-01 -2.83650637e-01 -3.40793669e-01 -4.38748717e-01
1.08168888e+00 -5.73869824e-01 -4.89618301e-01 -3.08271617e-01
-1.06038022e+00 2.12113336e-01 1.61739916e-01 6.11353278e-01
7.23713994e-01 -1.07091773e+00 -5.81913590e-01 3.61823946e-01
4.73605663e-01 4.47996520e-02 -7.21585937e-03 1.29925027e-01
-3.02319080e-01 6.34689212e-01 -2.28506178e-01 -4.75507855e-01
-1.40811741e+00 9.22905326e-01 6.48902297e-01 -2.37425417e-01
-3.90189856e-01 6.34992540e-01 3.52376819e-01 -6.78454220e-01
-1.13766141e-01 -7.36791253e-01 -2.91270256e-01 -5.21153927e-01
4.92451042e-01 8.04696698e-03 -8.26786980e-02 -6.02856874e-02
-2.15012506e-01 4.99809533e-01 1.83773823e-02 -1.38215259e-01
1.20794702e+00 3.75203609e-01 1.14291184e-01 3.67406100e-01
6.07496560e-01 -7.00332463e-01 -1.06498158e+00 -6.24686144e-02
3.75957280e-01 -3.53230208e-01 -2.40974739e-01 -1.37556100e+00
-5.94914079e-01 1.20627785e+00 4.31460917e-01 4.20480311e-01
1.11557591e+00 1.32125482e-01 8.46525133e-02 7.31775939e-01
3.40518445e-01 -7.47107148e-01 4.36092943e-01 2.48105720e-01
1.17686987e+00 -1.11172307e+00 4.53630798e-02 -7.75228620e-01
-8.33689630e-01 1.52811599e+00 9.33387101e-01 8.42911750e-02
5.47065675e-01 -6.17607385e-02 -3.86067480e-02 -4.49832290e-01
-1.01905060e+00 -1.28199920e-01 5.82488835e-01 1.10619795e+00
3.17049086e-01 5.60866110e-02 1.12457849e-01 8.71743619e-01
-3.14992726e-01 3.89642030e-01 5.20353138e-01 7.59175181e-01
-5.65698445e-01 -1.32013547e+00 -4.06876624e-01 3.78547698e-01
2.43846208e-01 -1.36431456e-01 -7.31582224e-01 9.86263692e-01
7.44485497e-01 1.14162731e+00 3.08456793e-02 -5.55633247e-01
4.65700448e-01 3.66467267e-01 2.61845887e-01 -1.12290287e+00
-7.00082421e-01 -4.51335847e-01 2.20358059e-01 -4.79060829e-01
-4.02717590e-01 -3.18838418e-01 -1.39723897e+00 -1.00320190e-01
-3.99181396e-01 4.55634087e-01 9.34933424e-01 1.55923164e+00
4.28276569e-01 6.94714963e-01 3.75102386e-02 -7.65206516e-01
-2.25972295e-01 -8.87554586e-01 -3.13412577e-01 4.90802258e-01
1.30971819e-01 -6.13750577e-01 -2.98051775e-01 3.48582178e-01] | [8.88485050201416, 5.779441833496094] |
f30d5ff9-e9d2-495f-99cf-0ec95a8e7f63 | heuristic-ternary-error-correcting-output | 1303.2132 | null | http://arxiv.org/abs/1303.2132v2 | http://arxiv.org/pdf/1303.2132v2.pdf | Heuristic Ternary Error-Correcting Output Codes Via Weight Optimization and Layered Clustering-Based Approach | One important classifier ensemble for multiclass classification problems is
Error-Correcting Output Codes (ECOCs). It bridges multiclass problems and
binary-class classifiers by decomposing multiclass problems to a serial
binary-class problems. In this paper, we present a heuristic ternary code,
named Weight Optimization and Layered Clustering-based ECOC (WOLC-ECOC). It
starts with an arbitrary valid ECOC and iterates the following two steps until
the training risk converges. The first step, named Layered Clustering based
ECOC (LC-ECOC), constructs multiple strong classifiers on the most confusing
binary-class problem. The second step adds the new classifiers to ECOC by a
novel Optimized Weighted (OW) decoding algorithm, where the optimization
problem of the decoding is solved by the cutting plane algorithm. Technically,
LC-ECOC makes the heuristic training process not blocked by some difficult
binary-class problem. OW decoding guarantees the non-increase of the training
risk for ensuring a small code length. Results on 14 UCI datasets and a music
genre classification problem demonstrate the effectiveness of WOLC-ECOC. | ['Xiao-Lei Zhang'] | 2013-03-08 | null | null | null | null | ['genre-classification'] | ['computer-vision'] | [ 6.68657303e-01 4.77043800e-02 -2.82112390e-01 -1.14444651e-01
-7.87717402e-01 -3.84935886e-01 2.91006472e-02 3.17369044e-01
-3.98939550e-01 7.09833562e-01 -3.31287324e-01 -6.64809883e-01
-5.83280504e-01 -5.40771127e-01 -3.89473706e-01 -9.29075241e-01
-9.68342274e-02 4.32198524e-01 3.15206826e-01 2.16817856e-01
7.12988973e-01 3.81876618e-01 -1.94521940e+00 7.93271244e-01
1.21735942e+00 1.21921158e+00 1.68184549e-01 9.95612025e-01
-2.64077067e-01 4.25179422e-01 -4.59371001e-01 -6.22691810e-01
1.69600263e-01 -6.15203738e-01 -7.45213032e-01 9.87498909e-02
-6.41012341e-02 5.34992874e-01 5.34223020e-01 1.01830840e+00
2.40188673e-01 -9.14602429e-02 9.09301817e-01 -1.74674094e+00
-3.71672779e-01 7.38099277e-01 -5.92453599e-01 -3.07231456e-01
-6.29593655e-02 -7.55407453e-01 9.98001039e-01 -1.00634778e+00
3.29991430e-01 9.23463821e-01 9.84170854e-01 5.32157719e-01
-1.04163659e+00 -7.65002251e-01 2.92848349e-02 4.42050993e-01
-1.59144807e+00 1.00174263e-01 4.92462099e-01 -5.41785300e-01
7.64309883e-01 7.30540276e-01 6.90154016e-01 6.04990780e-01
4.92261909e-02 6.78582847e-01 1.11745107e+00 -9.59478915e-01
5.47582507e-01 1.79992035e-01 5.40903986e-01 8.69601786e-01
5.89660585e-01 7.82536119e-02 -1.38978243e-01 -3.80326003e-01
-2.68761486e-01 -9.43896845e-02 -3.44754785e-01 -1.48759469e-01
-7.76030540e-01 8.86010170e-01 3.89590440e-03 4.25148904e-01
2.54786819e-01 1.52268171e-01 2.88877487e-01 3.35952997e-01
1.71244621e-01 4.23079669e-01 -3.95988792e-01 -1.81965381e-01
-1.11236274e+00 -2.34507203e-01 8.29498887e-01 1.16454411e+00
3.72882426e-01 -2.06031486e-01 9.63151082e-02 1.35639870e+00
1.67447403e-01 3.56588215e-01 5.57256043e-01 -8.89013290e-01
6.01384580e-01 7.50375211e-01 -4.30999368e-01 -8.02113831e-01
-5.27249873e-01 -8.56109500e-01 -1.05639958e+00 2.04916432e-01
3.99719924e-01 -1.49386481e-03 -7.36422360e-01 1.25138342e+00
-2.02584993e-02 3.38686444e-02 1.42008290e-02 5.62370718e-01
4.18140024e-01 6.81301951e-01 -1.75398350e-01 -4.54670966e-01
1.22339606e+00 -1.06838989e+00 -5.98451138e-01 1.38456956e-01
8.19312215e-01 -6.96382105e-01 5.77049911e-01 9.30106163e-01
-7.85551965e-01 -4.81370240e-01 -1.52716458e+00 2.66033620e-01
-5.33653677e-01 6.84135318e-01 4.24733102e-01 1.00663865e+00
-5.93923569e-01 4.89024371e-01 -2.53785640e-01 -7.87285417e-02
3.28627735e-01 4.75082308e-01 -1.95694908e-01 -1.93843320e-01
-8.29785466e-01 6.28279686e-01 7.09150016e-01 2.31395569e-02
-2.78861761e-01 -3.24072927e-01 -6.15076065e-01 1.69017866e-01
2.83693343e-01 9.22683701e-02 6.97716355e-01 -1.10603678e+00
-1.26167488e+00 6.92371190e-01 -1.36872754e-01 -3.21063459e-01
3.06191504e-01 2.91178167e-01 -5.21949887e-01 3.82788368e-02
-1.05963685e-01 6.61186695e-01 1.01097703e+00 -1.59708178e+00
-1.07845795e+00 -1.44548580e-01 -3.99049103e-01 -1.83835715e-01
-3.89360815e-01 -1.16392218e-01 -4.57519740e-01 -1.07196212e+00
5.97981453e-01 -1.15353596e+00 6.79056197e-02 -9.13736224e-02
-3.38782996e-01 -3.90979737e-01 7.40075111e-01 -2.31259272e-01
1.82822120e+00 -2.32734179e+00 5.00280321e-01 6.14145100e-01
2.71687489e-02 1.56599358e-01 -9.38325748e-02 9.24097151e-02
-4.54801351e-01 2.59859055e-01 -8.71518791e-01 -4.15840894e-01
-2.49059364e-01 4.40420300e-01 -1.44845694e-01 3.52019697e-01
9.45565328e-02 4.48180526e-01 -6.14341438e-01 -7.83739626e-01
-4.04671192e-01 -1.50672361e-01 -8.30846012e-01 -1.80575058e-01
1.38350442e-01 -2.96569586e-01 1.13119297e-01 1.01659858e+00
9.34060752e-01 -1.28927097e-01 3.15329492e-01 1.10997304e-01
-1.59638435e-01 -3.62366408e-01 -1.53387094e+00 1.31331551e+00
-3.53282630e-01 4.23072517e-01 -1.36063740e-01 -1.43240035e+00
9.53388035e-01 1.69680387e-01 4.36492741e-01 -7.61421844e-02
2.34586686e-01 7.73591161e-01 1.03653565e-01 -3.64550769e-01
3.89182419e-01 -1.61423475e-01 -1.53630093e-01 3.56389195e-01
2.83743829e-01 -1.74020454e-01 2.58887142e-01 9.16272402e-02
7.35984862e-01 -7.53119215e-02 3.43192786e-01 -4.47710186e-01
1.00408220e+00 -3.65664437e-02 9.24127698e-01 8.63038540e-01
-7.56001845e-02 8.20613742e-01 5.38324654e-01 -1.66284457e-01
-8.57389331e-01 -6.08075082e-01 -7.16944516e-01 8.77339363e-01
1.18301392e-01 -8.17675889e-01 -8.94161224e-01 -8.85178149e-01
1.94270611e-01 6.99423313e-01 -5.45476496e-01 -2.22663477e-01
-2.07572296e-01 -9.09944057e-01 6.00993991e-01 3.63953918e-01
4.17292982e-01 -6.22829020e-01 -4.26775128e-01 2.97580570e-01
-3.10433716e-01 -9.40691888e-01 -2.83135235e-01 9.03057694e-01
-9.14233208e-01 -1.25558031e+00 -7.13839531e-01 -1.03743827e+00
7.77191699e-01 1.80798680e-01 5.70132375e-01 6.66631043e-01
-5.25383890e-01 3.07828486e-02 -8.57599199e-01 -4.63644654e-01
-2.94071823e-01 2.21000075e-01 -7.26022874e-04 2.09267616e-01
2.22968549e-01 -2.60690391e-01 -8.66845027e-02 5.08755803e-01
-5.52685857e-01 7.30186775e-02 4.25213277e-01 1.06290376e+00
4.62307036e-01 3.41378927e-01 4.61034387e-01 -6.83328569e-01
3.48608762e-01 -3.59491289e-01 -4.94085848e-01 6.90317214e-01
-8.75609756e-01 1.34870932e-01 7.47878969e-01 -6.35963023e-01
-4.91844803e-01 2.78044552e-01 7.13653350e-03 -4.29965146e-02
1.83378384e-01 3.61909837e-01 3.60810198e-02 -1.80455178e-01
6.47160470e-01 1.39004484e-01 -3.68538350e-01 -3.29869956e-01
-5.46551384e-02 1.39199519e+00 3.69478375e-01 -7.66403019e-01
6.69248879e-01 1.83888167e-01 1.67019576e-01 -6.53716564e-01
-6.45409584e-01 -5.28925776e-01 -6.82844818e-01 -5.65817535e-01
1.01727164e+00 -3.74089092e-01 -8.31763387e-01 4.93985474e-01
-1.08481240e+00 -2.16236543e-02 -1.76183388e-01 5.18113732e-01
-4.34343189e-01 4.65522230e-01 -2.75324821e-01 -1.20381165e+00
-7.40075707e-02 -1.29896259e+00 9.62055206e-01 -7.93212652e-02
-4.62769195e-02 -4.58480656e-01 -4.23300326e-01 1.35761514e-01
-1.24472514e-01 9.73663479e-03 1.32370400e+00 -5.92797875e-01
-1.80327147e-01 -3.95967305e-01 -2.36641124e-01 4.67171967e-01
-3.73477459e-01 8.87576863e-02 -8.46576571e-01 -1.96844429e-01
-2.58787066e-01 -2.72246599e-01 1.10073960e+00 1.08064443e-03
1.88716948e+00 -6.46783486e-02 -3.59653056e-01 7.88776934e-01
1.52126086e+00 6.37172580e-01 5.77722192e-01 4.81434256e-01
2.80314922e-01 3.58644009e-01 5.31412542e-01 3.94802600e-01
-4.27299440e-02 7.27749407e-01 2.88928509e-01 7.85617083e-02
7.05247298e-02 5.87491170e-02 3.46482284e-02 1.15137887e+00
-4.54539627e-01 -2.84255713e-01 -9.28071856e-01 1.93993095e-02
-1.90676522e+00 -9.06546772e-01 -6.35125995e-01 2.14619780e+00
7.87537873e-01 3.44601840e-01 -2.76521474e-01 1.22664285e+00
1.03043115e+00 -5.65916777e-01 -1.50667980e-01 -8.52378368e-01
-1.55411243e-01 3.12725931e-01 5.40925324e-01 5.80871582e-01
-1.06095695e+00 4.17211682e-01 5.41178274e+00 1.24698901e+00
-5.54517567e-01 1.03119031e-01 5.42268217e-01 1.25280425e-01
-3.26048471e-02 7.94172660e-02 -1.04250300e+00 6.95468783e-01
7.02975273e-01 2.75529623e-01 5.32814503e-01 9.23982918e-01
-4.78914469e-01 -3.15914124e-01 -8.70911658e-01 1.21671212e+00
4.68977660e-01 -1.10332036e+00 -4.65903543e-02 4.40156274e-02
7.03276515e-01 -7.30596602e-01 -2.16180652e-01 6.02523267e-01
1.74491629e-01 -5.60073316e-01 1.07142794e+00 2.83343226e-01
1.08825529e+00 -9.02769744e-01 7.24017441e-01 4.81100023e-01
-1.23773551e+00 -8.57456744e-01 -4.70318794e-01 2.81089693e-01
-2.17609137e-01 5.11109352e-01 1.28602251e-01 6.50508225e-01
8.61958206e-01 7.51996875e-01 -8.06053758e-01 1.28674018e+00
2.95173712e-02 5.83377004e-01 -1.98591918e-01 -1.33058667e-01
2.14413255e-01 -3.91842753e-01 4.56425816e-01 1.48279846e+00
6.99837625e-01 4.08309668e-01 1.31678078e-02 1.41433761e-01
6.41792780e-03 7.40911812e-02 -6.78622629e-03 3.79045397e-01
4.68194038e-01 1.02172053e+00 -1.10464478e+00 -4.37653065e-01
-1.34399563e-01 9.67968643e-01 2.79511005e-01 1.81198835e-01
-9.86457527e-01 -8.15749168e-01 5.54811247e-02 -3.10777724e-01
3.38330090e-01 -4.30514887e-02 -8.80452693e-01 -1.07557201e+00
-1.69208318e-01 -9.41031456e-01 7.77511299e-01 -6.48148417e-01
-8.89005303e-01 5.95224500e-01 -9.53409299e-02 -1.69753397e+00
4.67350662e-01 -9.08671319e-01 -3.56409818e-01 4.32077885e-01
-1.50249970e+00 -4.92992938e-01 -3.44674587e-01 4.45401639e-01
6.42830729e-01 -5.28993964e-01 8.83601546e-01 4.07751620e-01
-7.58751512e-01 9.82973039e-01 4.37314034e-01 -1.77722454e-01
6.09700859e-01 -1.28460181e+00 -6.76802635e-01 8.37323725e-01
-1.43037841e-01 2.81177402e-01 1.54237598e-01 -4.25051033e-01
-9.30657923e-01 -1.07310116e+00 1.36303818e+00 1.41990736e-01
3.78785610e-01 -5.29775441e-01 -7.08903372e-01 -4.02625911e-02
-2.62696266e-01 -1.49672225e-01 1.03363001e+00 9.19571668e-02
-4.95936811e-01 -2.82908350e-01 -1.06033719e+00 3.48293126e-01
1.13092101e+00 -2.63535678e-01 -3.98341566e-01 4.63971138e-01
4.22201961e-01 -1.89834625e-01 -6.72735691e-01 5.25857329e-01
7.07842708e-01 -7.15164721e-01 8.31381917e-01 -6.19201243e-01
7.98156083e-01 -4.37546670e-01 -4.32754368e-01 -1.02565551e+00
-4.81335908e-01 -2.27964491e-01 -9.44300368e-02 1.07868040e+00
8.62868011e-01 -4.47156519e-01 3.60775650e-01 -1.91699504e-03
-4.92168427e-01 -9.13955867e-01 -1.10022938e+00 -1.03350842e+00
1.14774533e-01 -9.51356292e-01 4.64353323e-01 1.00545990e+00
2.93215692e-01 -1.56033173e-01 -1.94334611e-01 6.66801184e-02
6.67926431e-01 3.40005130e-01 9.66544598e-02 -1.59774530e+00
-3.63952965e-01 -6.78693116e-01 -4.74146515e-01 -5.26166737e-01
8.89555886e-02 -1.58832228e+00 2.06154078e-01 -1.18758941e+00
7.69524351e-02 -9.10385549e-01 -3.01236957e-01 6.68013692e-01
-1.06581390e-01 4.26047593e-01 4.78657663e-01 3.01358491e-01
-5.62639832e-01 1.37742221e-01 7.71247685e-01 -4.55359787e-01
-9.32949632e-02 1.16558790e-01 -4.36980665e-01 5.67505836e-01
7.38906324e-01 -8.86972785e-01 -1.23509146e-01 -1.03513254e-02
2.85058200e-01 4.47788909e-02 2.13267915e-02 -1.34365213e+00
4.86199945e-01 9.06082243e-02 3.90615612e-01 -5.42489946e-01
5.50149679e-02 -1.20788169e+00 -3.32625397e-02 9.04208481e-01
-5.21457136e-01 -7.39830583e-02 2.06465255e-02 6.68877661e-01
-2.93598231e-02 -1.15415502e+00 1.11898792e+00 3.47603083e-01
-4.39187050e-01 -2.16685280e-01 -6.80535018e-01 -7.63833057e-03
1.26397920e+00 -5.82612455e-01 -8.70051980e-02 1.40152574e-01
-1.08282554e+00 1.97130889e-01 -1.06481753e-01 2.72564977e-01
6.11713231e-01 -1.45883811e+00 -5.59657037e-01 3.86496067e-01
5.04274666e-01 -4.97536629e-01 -2.42934838e-01 7.62013435e-01
-4.92106140e-01 3.48735809e-01 2.46464938e-01 -7.75645435e-01
-1.65125525e+00 5.80116928e-01 2.42354199e-01 -9.10188481e-02
-4.67994928e-01 9.64739919e-01 -5.94545066e-01 -3.60447735e-01
7.52643585e-01 -2.76684463e-01 -3.07556421e-01 3.73905152e-01
2.88903981e-01 5.66175163e-01 4.40071523e-01 -3.17603648e-01
-3.17873478e-01 9.08955157e-01 1.94793835e-01 1.83407329e-02
1.23269475e+00 1.48058310e-01 -2.66732037e-01 2.84579962e-01
1.34626079e+00 -3.07149202e-01 -7.28364170e-01 1.01128750e-01
4.19305593e-01 -3.51235539e-01 -2.66528316e-02 -9.41547573e-01
-8.89169574e-01 7.73844302e-01 6.35448575e-01 2.75150239e-01
1.37193167e+00 -4.77967650e-01 4.40477103e-01 4.17221606e-01
4.60921437e-01 -1.41212761e+00 -1.97866574e-01 7.86427557e-01
8.69721830e-01 -8.24507177e-01 4.94403914e-02 -6.31684363e-01
-6.64448380e-01 1.56942725e+00 3.73100281e-01 1.72863856e-01
1.07876837e+00 4.95535165e-01 -3.86620104e-01 3.97219986e-01
-7.12998390e-01 -1.91525191e-01 4.34331506e-01 3.88878882e-01
1.39707133e-01 8.66591260e-02 -1.08567345e+00 9.39310849e-01
-2.02260301e-01 1.02719618e-02 3.20055068e-01 8.60791087e-01
-5.83741844e-01 -1.29069734e+00 -6.28704071e-01 5.04954875e-01
-1.82776585e-01 -1.18086278e-01 -2.91653931e-01 5.53962052e-01
8.68014514e-01 1.20408022e+00 1.01577520e-01 -7.09457815e-01
2.09428683e-01 4.43491280e-01 2.52516806e-01 -3.98705333e-01
-6.65291667e-01 7.01874495e-02 -4.35107127e-02 -9.18228999e-02
-2.50259727e-01 -4.93243366e-01 -1.33518660e+00 -1.59850955e-01
-7.75288641e-01 6.33969367e-01 9.20145571e-01 8.41822028e-01
6.68229237e-02 5.12125134e-01 8.92737806e-01 -5.54643512e-01
-3.72834891e-01 -5.84153831e-01 -5.50929844e-01 1.49284914e-01
5.87408133e-02 -8.44406009e-01 -7.19821274e-01 1.06625177e-01] | [8.873046875, 4.119327068328857] |
cd8b2570-acab-44bd-8e09-07150ce1705d | temporal-modeling-matters-a-novel-temporal | 2211.08233 | null | https://arxiv.org/abs/2211.08233v2 | https://arxiv.org/pdf/2211.08233v2.pdf | Temporal Modeling Matters: A Novel Temporal Emotional Modeling Approach for Speech Emotion Recognition | Speech emotion recognition (SER) plays a vital role in improving the interactions between humans and machines by inferring human emotion and affective states from speech signals. Whereas recent works primarily focus on mining spatiotemporal information from hand-crafted features, we explore how to model the temporal patterns of speech emotions from dynamic temporal scales. Towards that goal, we introduce a novel temporal emotional modeling approach for SER, termed Temporal-aware bI-direction Multi-scale Network (TIM-Net), which learns multi-scale contextual affective representations from various time scales. Specifically, TIM-Net first employs temporal-aware blocks to learn temporal affective representation, then integrates complementary information from the past and the future to enrich contextual representations, and finally, fuses multiple time scale features for better adaptation to the emotional variation. Extensive experimental results on six benchmark SER datasets demonstrate the superior performance of TIM-Net, gaining 2.34% and 2.61% improvements of the average UAR and WAR over the second-best on each corpus. The source code is available at https://github.com/Jiaxin-Ye/TIM-Net_SER. | ['Xin-Cheng Wen', 'Hongming Shan', 'KunHong Liu', 'Yong Xu', 'Yujie Wei', 'Jiaxin Ye'] | 2022-11-14 | null | null | null | null | ['speech-emotion-recognition'] | ['speech'] | [-7.51981512e-02 -2.96007603e-01 -1.48453489e-01 -6.63394988e-01
-7.30566442e-01 -2.82799900e-01 4.65174884e-01 -1.14830382e-01
-2.84128577e-01 6.16908848e-01 4.50119406e-01 1.57794073e-01
8.11522827e-03 -3.07353735e-01 -2.65244573e-01 -6.33655787e-01
-3.60236764e-01 -1.28012523e-01 -2.10497659e-02 -4.82239455e-01
-2.42060751e-01 3.31968874e-01 -1.29996598e+00 4.08110797e-01
5.91966093e-01 1.46413732e+00 7.29721924e-03 6.68014348e-01
-7.53763542e-02 9.93550122e-01 -6.36695087e-01 -1.44695461e-01
-2.99510449e-01 -5.74853778e-01 -6.02019191e-01 -1.46832481e-01
-6.41935408e-01 1.88548639e-01 -5.61518550e-01 6.54832006e-01
5.21442890e-01 5.69585323e-01 2.63431340e-01 -1.24706161e+00
-4.73472536e-01 4.51506764e-01 -5.28556347e-01 5.65648019e-01
2.26799071e-01 3.59555632e-02 9.29503322e-01 -9.01655197e-01
2.53210247e-01 1.26744509e+00 4.86605436e-01 6.44687235e-01
-8.11514080e-01 -7.52120256e-01 6.10688746e-01 6.18204832e-01
-1.22446060e+00 -5.67187965e-01 1.25715053e+00 -1.48296922e-01
9.59046066e-01 2.91559935e-01 6.05692744e-01 1.51965034e+00
3.86642627e-02 1.25690389e+00 1.09896398e+00 -1.00296102e-01
1.56459972e-01 -1.27733693e-01 1.83000877e-01 4.04281676e-01
-9.93591309e-01 1.67759910e-01 -9.78479803e-01 2.70563867e-02
2.49909073e-01 -3.94966677e-02 -1.37364626e-01 3.66929561e-01
-1.15710807e+00 5.56964099e-01 4.28817302e-01 6.12118125e-01
-8.33519042e-01 3.03482707e-03 8.82096112e-01 4.18537587e-01
9.86751378e-01 1.11513250e-01 -8.13863456e-01 -8.44500005e-01
-7.16207862e-01 -3.80258620e-01 3.30636829e-01 4.95801240e-01
4.99969602e-01 2.84702778e-01 -2.92196404e-02 1.20022237e+00
4.57738675e-02 3.02955389e-01 6.85340643e-01 -5.80432475e-01
2.73697823e-01 3.33088279e-01 -7.62972757e-02 -9.49936688e-01
-6.32343769e-01 -3.27647716e-01 -9.27313805e-01 -4.01083261e-01
-1.63936004e-01 -6.01196945e-01 -6.19897187e-01 1.96966898e+00
4.03804868e-01 6.02549970e-01 2.96856701e-01 9.59789813e-01
7.43205369e-01 1.17628360e+00 3.18623364e-01 -6.58842444e-01
1.38920736e+00 -1.03119481e+00 -1.21577835e+00 -4.36877251e-01
2.04292402e-01 -5.44128001e-01 9.21769381e-01 4.63248581e-01
-9.67411697e-01 -5.54027379e-01 -9.10017788e-01 2.59535372e-01
-4.67291117e-01 2.72276044e-01 7.52779186e-01 1.37159571e-01
-7.92196274e-01 4.04748499e-01 -1.00374818e+00 -4.79172543e-02
1.91951737e-01 1.84890747e-01 -2.11330891e-01 4.94060040e-01
-1.88520551e+00 7.09717333e-01 2.15052247e-01 4.56196427e-01
-5.81261277e-01 -5.52258968e-01 -9.01464224e-01 -4.50158976e-02
3.81122231e-01 -8.54816660e-02 1.43741715e+00 -1.25347888e+00
-1.89457560e+00 2.72796839e-01 -6.13140285e-01 -4.30693567e-01
7.57120475e-02 -3.26455235e-01 -1.06395876e+00 3.41424465e-01
-2.91652560e-01 3.57735634e-01 8.24701965e-01 -9.15230513e-01
-5.40706396e-01 -3.44610721e-01 -2.79789209e-01 1.64127126e-01
-6.70347512e-01 3.90988380e-01 -5.70384681e-01 -9.31719005e-01
-1.44394577e-01 -8.56760085e-01 -2.03612119e-01 -4.56141889e-01
2.06390489e-02 -4.97661948e-01 1.02465832e+00 -8.90497327e-01
1.64671397e+00 -2.28945971e+00 3.12779993e-01 -1.00030139e-01
-1.12684593e-01 3.10239106e-01 -1.60400614e-01 5.11726260e-01
-2.36869976e-01 -2.15711638e-01 -7.64127970e-02 -5.36435544e-01
1.13720810e-02 2.08251387e-01 -4.19290602e-01 2.33792678e-01
3.62750918e-01 1.02252162e+00 -9.07457888e-01 -3.35713744e-01
2.56131977e-01 7.87776828e-01 -9.70294848e-02 3.10227185e-01
-1.42550141e-01 6.77794755e-01 -5.58731556e-01 6.36128604e-01
3.32595527e-01 -2.01812863e-01 1.86312079e-01 -7.85310790e-02
2.28092656e-04 4.41646367e-01 -7.90619016e-01 1.71435285e+00
-9.37484145e-01 7.74356902e-01 1.22444211e-02 -1.09175515e+00
1.08279574e+00 8.52817178e-01 8.44798148e-01 -1.04357743e+00
2.23909989e-01 -6.64523765e-02 -4.05571252e-01 -5.78196585e-01
3.81650567e-01 -3.56818467e-01 -4.75622296e-01 2.62031019e-01
9.17807519e-02 2.52014041e-01 -1.73967242e-01 -1.47226844e-02
9.41701770e-01 -1.04649521e-01 3.92419815e-01 2.03697115e-01
7.31150091e-01 -5.64669013e-01 1.10455167e+00 -3.46009918e-02
-6.10643387e-01 1.41936123e-01 5.71847200e-01 -5.56266129e-01
-5.06635070e-01 -8.59548092e-01 1.84520006e-01 1.42975986e+00
1.43075913e-01 -5.07021546e-01 -3.55808318e-01 -7.26708055e-01
-3.90500069e-01 6.13917470e-01 -8.76948595e-01 -3.39428037e-01
-5.21065295e-01 -6.87522233e-01 4.53097284e-01 7.99021780e-01
2.86140501e-01 -1.29481256e+00 -4.52493131e-01 3.32611173e-01
-6.82295203e-01 -1.21324110e+00 -4.77318108e-01 2.62704819e-01
-4.88184661e-01 -3.98508608e-01 -4.76398528e-01 -4.82344091e-01
4.27340269e-02 -7.61024579e-02 8.63766313e-01 -2.60273308e-01
-4.82313484e-02 3.54127705e-01 -7.15462565e-01 -4.19516146e-01
-6.53953776e-02 -1.15839966e-01 1.90105334e-01 5.13948381e-01
4.46495891e-01 -9.06453490e-01 -5.90497673e-01 3.64485502e-01
-6.77283585e-01 -8.12379941e-02 5.09648263e-01 9.46429670e-01
5.23727834e-01 8.32054913e-02 1.07308304e+00 -2.72230387e-01
6.00924551e-01 -8.34183931e-01 -5.98072000e-02 1.19187228e-01
-2.04868570e-01 -1.69531897e-01 7.07528472e-01 -8.42457592e-01
-1.43378329e+00 -7.04271626e-03 -2.35541493e-01 -7.59329021e-01
-1.27795354e-01 5.92575014e-01 -1.07127562e-01 5.06151199e-01
1.65879175e-01 3.85494441e-01 -2.56860316e-01 -1.60111874e-01
5.05805790e-01 8.39518368e-01 4.27682549e-01 -5.78097701e-01
3.31253558e-01 4.40573305e-01 -4.43095446e-01 -9.48870003e-01
-8.65081131e-01 -6.37942195e-01 -4.63653594e-01 -5.47287941e-01
9.52864587e-01 -1.02201724e+00 -7.05028296e-01 3.54987323e-01
-1.09594715e+00 -4.47043985e-01 -4.88735884e-02 5.57083130e-01
-5.96798539e-01 1.01467386e-01 -8.43253851e-01 -1.20214117e+00
-4.58417207e-01 -8.23158860e-01 1.11450231e+00 4.23804462e-01
-4.23188180e-01 -1.03880692e+00 4.87225838e-02 2.97223896e-01
3.34534883e-01 2.85777718e-01 3.83744031e-01 -5.33401072e-01
6.96730092e-02 -6.98156953e-02 1.10574365e-01 3.01309735e-01
2.46406570e-01 -1.29445300e-01 -1.14906693e+00 5.65615743e-02
2.70735305e-02 -4.42302763e-01 7.40460515e-01 2.95280457e-01
1.19453681e+00 -3.24226677e-01 -1.22868516e-01 3.76542211e-01
8.24964345e-01 7.98648596e-01 2.26444766e-01 1.78944059e-02
3.81316274e-01 7.45270908e-01 1.05205190e+00 8.20141435e-01
4.86022562e-01 7.88278461e-01 2.50725031e-01 5.60880117e-02
2.58336037e-01 -6.65091276e-02 6.91501141e-01 1.36075556e+00
1.24719664e-01 -2.88445186e-02 -7.12708831e-01 8.50882053e-01
-2.11117578e+00 -1.16380370e+00 4.70672339e-01 1.46118355e+00
8.29838872e-01 4.37660813e-02 2.98299402e-01 2.32908744e-02
6.23560071e-01 6.51790559e-01 -7.47963905e-01 -5.32321572e-01
-1.49687156e-01 1.91500813e-01 -2.55646795e-01 3.25610906e-01
-1.28782415e+00 1.13601589e+00 4.91989899e+00 9.43617165e-01
-1.46772730e+00 3.61882120e-01 8.65417898e-01 -5.04792392e-01
-3.14720310e-02 -4.11055595e-01 -3.09678644e-01 5.38363099e-01
1.44076014e+00 -3.94771904e-01 5.64739108e-01 7.00525522e-01
6.85932875e-01 3.35291356e-01 -6.03874624e-01 1.02119374e+00
-7.13792220e-02 -9.34057832e-01 -5.26812673e-01 -2.95626074e-01
5.20698965e-01 4.87713553e-02 1.52742118e-01 6.51840687e-01
2.36036107e-01 -7.09526777e-01 7.24378705e-01 7.52849817e-01
5.86060107e-01 -1.15680754e+00 5.61884463e-01 2.12703258e-01
-1.77061045e+00 -1.65504202e-01 8.28097761e-02 3.39985527e-02
4.13288981e-01 5.16368449e-01 -5.17266333e-01 6.86733067e-01
9.55886662e-01 1.06875348e+00 -4.03874815e-02 1.93571970e-01
-2.30611444e-01 7.82017946e-01 -9.55151543e-02 -2.65568882e-01
3.39538127e-01 -1.51862383e-01 5.34444392e-01 1.47770560e+00
2.40000442e-01 4.89284128e-01 -2.94920318e-02 3.02159995e-01
5.02272435e-02 5.16516380e-02 -3.90775561e-01 -4.43312943e-01
5.32063186e-01 1.50688255e+00 -3.67076218e-01 -2.16449723e-01
-4.52861071e-01 1.16915476e+00 3.31047237e-01 4.22549576e-01
-1.22748828e+00 -3.51846099e-01 1.22429895e+00 -6.90488875e-01
4.40053821e-01 -3.50216150e-01 -6.67090202e-03 -1.19338739e+00
1.12984642e-01 -8.86585891e-01 5.71660757e-01 -9.77736533e-01
-1.19355392e+00 1.05694258e+00 -3.53665084e-01 -1.22755075e+00
-4.01342481e-01 -2.94329792e-01 -7.00056434e-01 6.42296433e-01
-1.41293752e+00 -1.24602032e+00 -4.89528291e-02 6.71967626e-01
1.02024364e+00 -3.56888250e-02 6.57211483e-01 2.48672426e-01
-6.91460371e-01 5.59688866e-01 -1.42797396e-01 1.32676601e-01
6.62333548e-01 -1.07574785e+00 1.67501569e-01 8.07649016e-01
1.63634703e-01 4.93030757e-01 5.84414840e-01 -3.02076727e-01
-1.33189237e+00 -1.05063629e+00 8.36407900e-01 -9.98802260e-02
1.05108273e+00 -2.53422230e-01 -9.02418196e-01 5.64668894e-01
4.34457779e-01 1.63384780e-01 8.32602561e-01 3.11306596e-01
-3.46413255e-01 -2.72821724e-01 -7.42969871e-01 7.25924373e-01
8.03360164e-01 -9.58202243e-01 -4.33553427e-01 -1.77595497e-03
1.14288592e+00 -2.23259687e-01 -1.20723462e+00 5.60123384e-01
5.87015748e-01 -7.61128306e-01 8.76251757e-01 -6.27320468e-01
3.68150771e-01 -1.57284960e-01 -3.30436915e-01 -1.60983980e+00
-1.44438222e-01 -8.86523366e-01 -3.39710087e-01 1.29021573e+00
3.50857407e-01 -4.71292794e-01 3.19891572e-01 1.18778661e-01
-3.08803260e-01 -1.23036635e+00 -1.03977227e+00 -6.32369816e-01
-2.28807375e-01 -9.02660012e-01 7.25730956e-01 1.05867708e+00
4.65217292e-01 4.01148170e-01 -7.64172375e-01 3.90309632e-01
-4.62958589e-02 2.02203557e-01 3.68236601e-01 -8.02025735e-01
-1.69854864e-01 -5.52067995e-01 -2.16794789e-01 -9.66319799e-01
6.30214095e-01 -4.68004644e-01 1.82306409e-01 -1.10655284e+00
-2.08245620e-01 6.02621958e-02 -8.85064244e-01 5.43412566e-01
-3.50028962e-01 -9.69761536e-02 6.25135675e-02 -1.52967185e-01
-9.42358375e-01 1.19251764e+00 1.00941944e+00 -7.55258277e-02
-3.01006138e-01 5.27512915e-02 -5.67585707e-01 6.21438265e-01
1.11186087e+00 -1.21145181e-01 -5.73003173e-01 5.52121624e-02
2.84350123e-02 5.80304921e-01 -2.23212801e-02 -9.97868180e-01
2.11898863e-01 -4.30783421e-01 2.11142316e-01 -6.32045925e-01
8.69853079e-01 -7.72588789e-01 2.83094053e-03 2.60820091e-02
-4.40728962e-01 1.83365107e-01 5.36647558e-01 7.94827402e-01
-7.69084871e-01 4.58611190e-01 7.22209156e-01 1.40247226e-01
-1.07330751e+00 5.33483386e-01 -4.24149305e-01 -9.21271294e-02
1.06584311e+00 3.56376439e-01 -1.38380989e-01 -6.62689865e-01
-1.09802341e+00 3.72627556e-01 -2.59516150e-01 9.49147165e-01
6.91809595e-01 -1.59678781e+00 -3.30911487e-01 3.05599160e-02
1.86401114e-01 -5.71899354e-01 8.43574643e-01 1.03793144e+00
3.27165186e-01 4.29961771e-01 8.89415890e-02 -4.32794899e-01
-1.49853814e+00 4.76404190e-01 4.33271587e-01 -2.68228054e-01
-3.22059453e-01 9.26003575e-01 1.85919017e-01 -3.43322694e-01
1.27253443e-01 -2.27587000e-02 -3.52144122e-01 4.30700034e-01
6.17219687e-01 1.72740921e-01 7.59584038e-03 -8.48812461e-01
-5.87301970e-01 2.97908157e-01 9.44060013e-02 -4.25908327e-01
1.59804165e+00 -4.57998216e-01 1.43144876e-02 1.03705359e+00
1.36276352e+00 -9.94299576e-02 -1.47670245e+00 -6.87726438e-01
1.11877188e-01 1.88396163e-02 1.30408198e-01 -9.01260436e-01
-1.23437858e+00 9.80641365e-01 3.70531976e-01 1.30292654e-01
1.78275037e+00 2.57915258e-02 1.08866501e+00 1.16992742e-01
1.45799592e-01 -1.26028240e+00 5.23258388e-01 6.69612586e-01
1.07395661e+00 -1.14988577e+00 -5.19094706e-01 4.10261489e-02
-1.30960941e+00 9.89643753e-01 5.58991611e-01 2.81469285e-01
8.67984712e-01 2.44507656e-01 3.54423732e-01 -5.78174591e-02
-1.37222469e+00 -2.48357177e-01 3.57878119e-01 1.98236153e-01
4.07173991e-01 3.25787455e-01 -8.92995223e-02 1.11630738e+00
-5.02941087e-02 -1.25822827e-01 -1.34552494e-01 7.36260235e-01
-1.51698783e-01 -9.94648099e-01 -6.71374202e-02 2.06738450e-02
-4.71397370e-01 2.24726181e-02 -4.07134473e-01 4.90006685e-01
-4.19994965e-02 1.23785746e+00 6.23805029e-03 -9.06692028e-01
3.49161386e-01 3.64014059e-01 -1.00814819e-01 -4.46558557e-02
-5.58767796e-01 4.42476362e-01 2.83831656e-01 -8.60787868e-01
-4.87033963e-01 -9.36948478e-01 -1.39756835e+00 7.89813846e-02
9.58893821e-02 4.16330934e-01 6.42026901e-01 9.50655580e-01
5.99474669e-01 9.69335854e-01 1.16073573e+00 -7.75694728e-01
-3.95329706e-02 -1.12863946e+00 -4.13218081e-01 2.81429768e-01
6.36539757e-01 -7.03346908e-01 -4.83967006e-01 2.71923225e-02] | [13.449735641479492, 5.693144798278809] |
bd4866e5-8e4c-4e38-a0db-60fb43a49306 | delayed-feedback-in-kernel-bandits | 2302.00392 | null | https://arxiv.org/abs/2302.00392v1 | https://arxiv.org/pdf/2302.00392v1.pdf | Delayed Feedback in Kernel Bandits | Black box optimisation of an unknown function from expensive and noisy evaluations is a ubiquitous problem in machine learning, academic research and industrial production. An abstraction of the problem can be formulated as a kernel based bandit problem (also known as Bayesian optimisation), where a learner aims at optimising a kernelized function through sequential noisy observations. The existing work predominantly assumes feedback is immediately available; an assumption which fails in many real world situations, including recommendation systems, clinical trials and hyperparameter tuning. We consider a kernel bandit problem under stochastically delayed feedback, and propose an algorithm with $\tilde{\mathcal{O}}(\sqrt{\Gamma_k(T)T}+\mathbb{E}[\tau])$ regret, where $T$ is the number of time steps, $\Gamma_k(T)$ is the maximum information gain of the kernel with $T$ observations, and $\tau$ is the delay random variable. This represents a significant improvement over the state of the art regret bound of $\tilde{\mathcal{O}}(\Gamma_k(T)\sqrt{T}+\mathbb{E}[\tau]\Gamma_k(T))$ reported in Verma et al. (2022). In particular, for very non-smooth kernels, the information gain grows almost linearly in time, trivializing the existing results. We also validate our theoretical results with simulations. | ['Ciara Pike-Burke', 'Alberto Bernacchia', 'Danyal Ahmed', 'Sattar Vakili'] | 2023-02-01 | null | null | null | null | ['bayesian-optimisation'] | ['methodology'] | [ 1.15005866e-01 6.50967434e-02 -3.31715077e-01 -2.11639062e-01
-1.01683068e+00 -7.29445577e-01 -2.97098104e-02 2.90602982e-01
-9.04148519e-01 1.12753081e+00 -3.77952993e-01 -8.69003296e-01
-9.17106509e-01 -5.79165995e-01 -1.05020010e+00 -1.08223140e+00
-5.42213023e-01 2.70978361e-01 -1.98086482e-02 7.84463957e-02
1.41510859e-01 1.85531959e-01 -1.02057171e+00 -2.09805891e-01
1.04065681e+00 1.63375115e+00 -1.39674619e-01 8.59072804e-01
3.72504070e-02 5.86114168e-01 -5.18788695e-01 -6.60011709e-01
3.11020285e-01 -5.39090931e-01 -7.39930212e-01 -3.16050291e-01
-1.28639847e-01 -1.93390533e-01 -1.32362396e-01 1.23068440e+00
4.24754560e-01 4.63953674e-01 7.03338623e-01 -7.72843719e-01
-1.75796926e-01 4.17327046e-01 -5.68718553e-01 4.42081153e-01
-1.59979761e-01 3.24381404e-02 7.59034693e-01 -1.59504473e-01
-4.37403359e-02 7.68737674e-01 5.46349883e-01 1.94649756e-01
-1.15261805e+00 -7.94800520e-01 4.63840574e-01 1.24247797e-01
-1.19229341e+00 4.97839376e-02 2.19920710e-01 -4.67052460e-01
6.17535472e-01 6.77998066e-01 7.12064147e-01 4.33822036e-01
2.13542372e-01 8.82867515e-01 1.19413018e+00 -5.00997245e-01
7.16526568e-01 2.92171001e-01 2.82382011e-01 5.61818302e-01
8.95791575e-02 4.54524398e-01 -3.29918146e-01 -2.48975411e-01
9.22226965e-01 1.28331274e-01 -4.35829312e-01 5.65298870e-02
-8.59222829e-01 1.05758369e+00 2.85044283e-01 -1.33886814e-01
-4.31428492e-01 4.25086319e-01 2.56588280e-01 6.60402119e-01
8.68362546e-01 1.17263421e-01 -7.64572859e-01 -5.31865716e-01
-8.85950089e-01 2.64641613e-01 9.28687334e-01 9.19133961e-01
6.10302508e-01 -3.22374217e-02 -2.46267542e-01 6.78798258e-01
1.39830053e-01 7.45259941e-01 1.24192625e-01 -8.83429706e-01
5.49214661e-01 1.87998086e-01 5.87382615e-01 -1.97497770e-01
-3.40465605e-01 -7.29277551e-01 -8.79276156e-01 9.96123850e-02
8.97821724e-01 -6.30313993e-01 -6.74194157e-01 1.39824367e+00
3.88473809e-01 1.01118833e-01 -1.98856905e-01 9.38170612e-01
3.94819565e-02 5.99359095e-01 -3.08744133e-01 -8.36027741e-01
1.03811705e+00 -7.43202984e-01 -4.36033875e-01 -1.41461626e-01
5.95245957e-01 -7.30808377e-01 7.61801898e-01 8.05796266e-01
-1.15826833e+00 1.14008030e-02 -8.53228211e-01 7.28126943e-01
-1.52646482e-01 -2.21102774e-01 1.00497401e+00 1.30863225e+00
-8.71874511e-01 5.85304141e-01 -6.48744166e-01 3.38840604e-01
3.35649222e-01 6.63493812e-01 -1.11112511e-02 -2.53995091e-01
-1.02912164e+00 5.23404062e-01 2.79017746e-01 3.33358616e-01
-8.16642761e-01 -8.72938573e-01 -4.28127140e-01 -1.49827704e-01
8.14936280e-01 -3.82601082e-01 1.40227866e+00 -9.15999174e-01
-1.79919553e+00 2.48624429e-01 1.52168367e-02 -7.10346758e-01
6.80356264e-01 -3.47153366e-01 -1.31717622e-01 -1.60622075e-01
-3.00030500e-01 -9.89564434e-02 9.55053866e-01 -5.38683832e-01
-7.88744330e-01 -5.97667634e-01 1.99091077e-01 -3.51198157e-03
-1.24986894e-01 1.28525093e-01 -1.22339264e-01 -7.66052902e-01
-5.71699627e-02 -9.29900110e-01 -3.59757572e-01 -3.67083311e-01
-7.24492222e-02 -9.21034068e-02 2.26405039e-01 -6.56002820e-01
1.45274746e+00 -1.90178299e+00 -5.08070132e-03 5.31824708e-01
-6.55427054e-02 1.80736870e-01 5.20689428e-01 2.29808807e-01
-4.65367809e-02 4.54210229e-02 -2.47665659e-01 1.89241827e-01
1.55253643e-02 -7.63449892e-02 -2.73421437e-01 6.83090925e-01
-4.42589194e-01 7.28271484e-01 -8.19020212e-01 1.80437371e-01
1.21799752e-01 2.15424523e-01 -2.83361495e-01 -7.63646737e-02
-4.08901215e-01 4.32694465e-01 -6.45938754e-01 4.10626203e-01
4.47580963e-01 -4.06671852e-01 -8.24273974e-02 2.35694841e-01
-1.15267962e-01 -1.36829048e-01 -1.59561574e+00 1.21615279e+00
-2.57213384e-01 1.05792701e-01 2.53930688e-01 -1.52940142e+00
5.98081291e-01 2.80172348e-01 6.67502642e-01 -6.06345713e-01
2.41648987e-01 1.95236832e-01 -2.98197359e-01 -1.55151591e-01
-1.47510871e-01 -4.55740780e-01 1.89318173e-02 6.82802558e-01
-1.50252432e-01 -3.07681337e-02 -3.04988082e-02 -2.22289070e-01
1.30557334e+00 -1.57082051e-01 1.72462523e-01 -2.94775814e-01
2.73075849e-01 -1.62880167e-01 3.54070455e-01 1.15276527e+00
-1.37374684e-01 -4.99734990e-02 7.30424583e-01 -4.51094300e-01
-8.02364528e-01 -9.95008886e-01 -2.73487031e-01 1.39724588e+00
4.74845134e-02 9.56794806e-03 -8.76364052e-01 -5.73453963e-01
8.70409012e-02 7.48807430e-01 -7.71329403e-01 -1.62556663e-01
-1.94992781e-01 -1.27250767e+00 3.39269131e-01 4.28807139e-01
3.59786183e-01 -7.03115463e-01 -7.00112879e-01 4.25794780e-01
1.83054686e-01 -5.19835949e-01 -6.93687379e-01 6.13776445e-01
-1.01213789e+00 -7.51470506e-01 -9.56456959e-01 7.65380338e-02
5.51585138e-01 -9.37939212e-02 6.59844458e-01 -3.97182882e-01
-3.91469091e-01 4.72525448e-01 -4.35826898e-01 -9.76330638e-01
2.32600793e-01 -2.50553727e-01 -7.05017075e-02 1.33918211e-01
1.00624323e-01 -3.19559813e-01 -9.68005121e-01 4.26542222e-01
-1.03590930e+00 -2.76870966e-01 5.08405387e-01 9.77160990e-01
7.62418270e-01 1.21680722e-01 4.89454478e-01 -8.11249673e-01
6.19519353e-01 -4.25924510e-01 -1.24454963e+00 3.38837922e-01
-9.22048569e-01 2.97486514e-01 5.00502348e-01 -7.22354352e-01
-8.69142413e-01 -1.76760986e-01 2.42278904e-01 -3.84317189e-01
1.78938001e-01 6.27788961e-01 3.92914921e-01 -1.65424824e-01
6.33092165e-01 3.72813493e-01 -8.94964039e-02 -5.11396766e-01
3.70794982e-01 5.19643605e-01 2.53116399e-01 -7.74843454e-01
2.94189453e-01 3.93109858e-01 -5.51130809e-02 -3.90270084e-01
-1.02494097e+00 -4.46912944e-01 -2.89904457e-02 -1.89498842e-01
3.91395539e-01 -6.82609439e-01 -1.41142452e+00 1.35640860e-01
-5.40557146e-01 -6.85072184e-01 -4.86178964e-01 9.64654088e-01
-9.22294021e-01 7.38570541e-02 -3.18879813e-01 -1.59486651e+00
-5.53948164e-01 -9.01938915e-01 6.08692706e-01 4.54775572e-01
6.06501997e-02 -1.00030041e+00 -1.95848569e-01 5.32868445e-01
5.03217578e-01 7.23015666e-02 8.12379539e-01 -4.89425421e-01
-6.32575452e-01 -4.62159693e-01 -1.90480471e-01 6.67349756e-01
-1.47300199e-01 -5.88037193e-01 -4.96317774e-01 -5.33457339e-01
2.24553138e-01 -2.81365365e-01 7.23444045e-01 7.98178375e-01
1.17856920e+00 -7.43111372e-01 -1.21773362e-01 5.65400839e-01
1.31216216e+00 5.29268980e-01 4.17979836e-01 2.14539245e-01
-4.08846289e-02 8.79981071e-02 7.09544122e-01 8.94254386e-01
-2.46313829e-02 4.28949833e-01 4.34201121e-01 1.10746406e-01
6.70378923e-01 3.92534912e-01 4.49009985e-01 2.68581748e-01
-1.70977920e-01 5.74809983e-02 -6.19086742e-01 5.67073166e-01
-2.00912118e+00 -6.52333438e-01 8.75264332e-02 2.83790302e+00
1.06057048e+00 2.44349942e-01 2.63064533e-01 7.09939897e-02
6.46615267e-01 -3.38530242e-01 -7.96502709e-01 -6.78488672e-01
1.77893743e-01 9.00934398e-01 1.06286836e+00 5.12341678e-01
-8.58667374e-01 5.62202215e-01 5.07963610e+00 1.31021452e+00
-1.10308897e+00 2.71253347e-01 8.21857393e-01 -7.27200568e-01
1.24037683e-01 7.59319291e-02 -7.25720823e-01 8.62545550e-01
1.21523392e+00 -2.41226047e-01 8.85430276e-01 7.44998932e-01
1.43027738e-01 -7.90745258e-01 -1.02234674e+00 9.55079436e-01
-3.38016510e-01 -1.18915677e+00 -7.86009073e-01 2.00910956e-01
8.56380820e-01 -2.58072875e-02 2.35490024e-01 3.44314784e-01
8.22861493e-01 -1.23086965e+00 6.09381080e-01 5.58849454e-01
8.06444466e-01 -1.01508653e+00 7.19053388e-01 7.54396319e-01
-8.92081320e-01 -3.40995789e-01 -1.20825283e-01 -1.67852029e-01
3.95011492e-02 1.12304604e+00 -6.91093028e-01 8.17230821e-01
1.03836823e+00 -1.45945147e-01 1.93948165e-01 1.34058583e+00
-5.13346978e-02 8.20281625e-01 -7.68365026e-01 -3.87122750e-01
4.25573379e-01 -4.33383584e-01 3.00716788e-01 8.96338344e-01
3.01658332e-01 3.88600618e-01 2.33332887e-01 4.48882967e-01
1.52957216e-01 2.23759338e-01 7.23665878e-02 1.95797265e-01
2.05671892e-01 7.62112081e-01 -6.62208796e-01 -1.88427910e-01
-1.06157660e-01 7.79609263e-01 2.34630629e-01 5.43934882e-01
-9.18133020e-01 -5.13274431e-01 5.03941715e-01 -1.03097469e-01
6.09553516e-01 5.52391633e-03 -2.39670396e-01 -7.08619416e-01
3.63086313e-01 -4.30625021e-01 6.51319742e-01 -2.36537337e-01
-1.17186987e+00 -7.77459368e-02 3.29292089e-01 -8.30329299e-01
-2.55657852e-01 -5.82709670e-01 -5.41165322e-02 1.01891232e+00
-1.22893083e+00 -5.10600686e-01 3.17026079e-01 8.09533179e-01
9.97874066e-02 -1.47256218e-02 7.77827501e-01 8.38261023e-02
-4.15690571e-01 7.54577518e-01 1.10673559e+00 -1.43522218e-01
4.29540753e-01 -1.14484000e+00 -2.87016183e-01 3.25441152e-01
-3.58546227e-01 4.25611436e-01 7.59855032e-01 -4.90299463e-01
-1.29757583e+00 -9.28102195e-01 2.05124378e-01 -1.85756758e-01
7.92008936e-01 -9.24869403e-02 -5.13062000e-01 6.04338944e-01
-2.63785332e-01 2.88164794e-01 7.38207817e-01 6.44546524e-02
-1.66558117e-01 -5.06778240e-01 -1.30665720e+00 3.24574471e-01
5.87715626e-01 -1.11415073e-01 5.33434749e-02 4.97142196e-01
4.37951833e-01 -5.69517136e-01 -1.29572618e+00 5.25528967e-01
7.18433619e-01 -9.22811925e-01 6.21070981e-01 -5.62445521e-01
-4.05641884e-01 7.38198906e-02 4.18106504e-02 -1.08152831e+00
7.59191290e-02 -1.18758857e+00 -4.60142851e-01 5.58049202e-01
5.28795183e-01 -9.52988863e-01 8.58974993e-01 9.26156580e-01
2.11612970e-01 -1.01398730e+00 -1.62915218e+00 -8.69069695e-01
3.58406723e-01 -6.60567224e-01 2.10707128e-01 5.88177145e-01
6.36809855e-04 -2.57080436e-01 -4.03086990e-01 -5.74502833e-02
6.22013509e-01 -1.45293206e-01 4.02833462e-01 -9.40166354e-01
-7.45799303e-01 -3.73661041e-01 1.84351038e-02 -1.12428594e+00
-3.75862718e-01 -4.11564946e-01 -4.68868576e-02 -1.02812135e+00
2.05773562e-01 -7.61432290e-01 -6.37770057e-01 5.04680872e-01
-9.78864953e-02 -4.03215140e-01 -1.42426431e-01 -1.86739743e-01
-6.37404919e-01 2.91234314e-01 1.00993538e+00 -8.18359330e-02
-2.30117157e-01 6.10959411e-01 -6.42301321e-01 4.80609447e-01
6.16084814e-01 -5.38583338e-01 -4.95553851e-01 -2.58278161e-01
5.32250643e-01 6.59624279e-01 4.18869764e-01 -5.45626819e-01
3.37324977e-01 -5.70610523e-01 2.28459179e-01 -3.57797176e-01
3.73241216e-01 -6.19196117e-01 2.73280501e-01 3.66187274e-01
-1.28682107e-01 -2.09550828e-01 1.59138709e-01 9.63682294e-01
1.79772243e-01 -3.43592346e-01 6.80043995e-01 -1.87478468e-01
1.36710638e-02 2.84750044e-01 -1.69181749e-01 -6.99003786e-02
1.27588332e+00 -1.69992447e-01 -2.77308464e-01 -6.10345244e-01
-8.03322136e-01 4.75276291e-01 -8.19054618e-02 -1.42137073e-02
2.71442473e-01 -8.74672949e-01 -5.84083140e-01 -7.54291192e-03
-2.35553548e-01 3.19830865e-01 5.49396217e-01 1.24226236e+00
-1.44334614e-01 5.93967557e-01 5.33788800e-01 -5.14221072e-01
-7.98321128e-01 3.75031352e-01 3.83476853e-01 -5.56297898e-01
-1.77430302e-01 1.26571453e+00 -1.47544473e-01 6.61904290e-02
5.75038850e-01 -3.37192893e-01 5.54519713e-01 -1.20750852e-01
5.74597657e-01 8.13997328e-01 3.85262579e-01 2.31775314e-01
-8.39101896e-02 1.14838660e-01 -3.97163570e-01 -4.72701043e-01
1.20684958e+00 4.68272120e-02 -1.33333400e-01 4.42194939e-01
1.06625044e+00 -4.73848313e-01 -1.48573089e+00 -3.87549251e-01
-3.85161862e-02 -2.76249915e-01 8.84461999e-02 -1.14149523e+00
-8.00410509e-01 6.97199523e-01 9.50767636e-01 3.14326227e-01
1.09645462e+00 -3.50479595e-02 4.60866570e-01 4.17197585e-01
6.73238575e-01 -1.39331472e+00 -1.85087830e-01 5.97102106e-01
5.15143096e-01 -9.27285254e-01 1.03211235e-02 -5.33190891e-02
-4.49892163e-01 8.87981534e-01 6.23602942e-02 -1.95135467e-03
9.63057756e-01 1.26960635e-01 -2.66493082e-01 1.10659622e-01
-5.46086907e-01 -1.09557599e-01 2.40268305e-01 1.13570623e-01
2.12685376e-01 2.77554274e-01 -4.86967504e-01 1.01338911e+00
-1.89101130e-01 2.13963568e-01 2.29356512e-02 1.10415149e+00
-5.25019586e-01 -1.08419442e+00 -5.77485502e-01 8.56095850e-01
-8.41925681e-01 -8.71012062e-02 1.29850447e-01 3.96143198e-01
-8.81009642e-03 9.21541393e-01 -2.34574407e-01 1.40326470e-01
2.55424291e-01 2.59861350e-01 6.17217302e-01 -2.16368780e-01
-5.96250355e-01 4.00747508e-01 -1.77343920e-01 -3.82196754e-01
-3.27757597e-02 -4.79793340e-01 -1.06742942e+00 -2.89202332e-01
-5.08512378e-01 5.53019166e-01 8.54610801e-01 1.15233827e+00
1.22593939e-02 3.25653702e-01 6.85350776e-01 -4.86766666e-01
-9.26843226e-01 -8.98462474e-01 -9.33907270e-01 -4.19176444e-02
3.81314367e-01 -6.78063452e-01 -3.75940800e-01 -1.53035671e-01] | [4.654546737670898, 3.393979549407959] |
b6d9e75f-e6c1-4469-bf01-a3f4e916a954 | deep-identity-aware-transfer-of-facial | 1610.05586 | null | http://arxiv.org/abs/1610.05586v2 | http://arxiv.org/pdf/1610.05586v2.pdf | Deep Identity-aware Transfer of Facial Attributes | This paper presents a Deep convolutional network model for Identity-Aware
Transfer (DIAT) of facial attributes. Given the source input image and the
reference attribute, DIAT aims to generate a facial image that owns the
reference attribute as well as keeps the same or similar identity to the input
image. In general, our model consists of a mask network and an attribute
transform network which work in synergy to generate a photo-realistic facial
image with the reference attribute. Considering that the reference attribute
may be only related to some parts of the image, the mask network is introduced
to avoid the incorrect editing on attribute irrelevant region. Then the
estimated mask is adopted to combine the input and transformed image for
producing the transfer result. For joint training of transform network and mask
network, we incorporate the adversarial attribute loss, identity-aware adaptive
perceptual loss, and VGG-FACE based identity loss. Furthermore, a denoising
network is presented to serve for perceptual regularization to suppress the
artifacts in transfer result, while an attribute ratio regularization is
introduced to constrain the size of attribute relevant region. Our DIAT can
provide a unified solution for several representative facial attribute transfer
tasks, e.g., expression transfer, accessory removal, age progression, and
gender transfer, and can be extended for other face enhancement tasks such as
face hallucination. The experimental results validate the effectiveness of the
proposed method. Even for the identity-related attribute (e.g., gender), our
DIAT can obtain visually impressive results by changing the attribute while
retaining most identity-aware features. | ['WangMeng Zuo', 'David Zhang', 'Mu Li'] | 2016-10-18 | null | null | null | null | ['face-hallucination'] | ['computer-vision'] | [ 5.04422665e-01 4.43082780e-01 1.82846829e-01 -5.81779718e-01
-4.57470655e-01 -2.24709451e-01 3.13933104e-01 -4.23697352e-01
-2.46095479e-01 6.68170273e-01 -4.41034511e-02 3.26850027e-01
1.19867690e-01 -9.27704453e-01 -8.23688030e-01 -1.03719890e+00
4.51679915e-01 4.89107035e-02 -2.58654833e-01 -2.24973261e-01
-1.61797971e-01 5.33340931e-01 -1.49719095e+00 2.50259936e-01
1.09361672e+00 1.33717716e+00 7.01979697e-02 -1.11276641e-01
-4.77492716e-03 4.67099249e-01 -6.69427693e-01 -7.70947874e-01
3.02523673e-01 -6.96561515e-01 -4.80122447e-01 2.81757444e-01
6.11392796e-01 -5.41928113e-01 -3.18209916e-01 1.21057308e+00
7.16640949e-01 6.82002306e-02 7.89418399e-01 -1.62878764e+00
-9.58555162e-01 2.01628640e-01 -7.16440558e-01 -4.17360634e-01
1.23803213e-01 1.38119355e-01 3.95001829e-01 -1.08216965e+00
6.32415593e-01 1.49887168e+00 4.84790504e-01 8.50799859e-01
-1.22596049e+00 -1.21493971e+00 1.46419764e-01 1.75788254e-01
-1.71033168e+00 -6.18038833e-01 9.91980076e-01 -7.39871487e-02
2.54506599e-02 2.55257249e-01 5.54225445e-01 1.18635321e+00
8.79868791e-02 5.60412109e-01 1.19589937e+00 -1.59878984e-01
-1.99785799e-01 4.06507611e-01 -6.97152674e-01 7.47348011e-01
-1.73503980e-01 3.47250104e-02 -1.93326876e-01 -6.96635619e-03
9.29793954e-01 -1.44661278e-01 -3.12961668e-01 -1.21388741e-01
-9.12234187e-01 5.63484669e-01 6.20884895e-01 -7.86377490e-02
-5.11981249e-01 -1.27896577e-01 2.76800245e-01 3.38300437e-01
5.25437355e-01 1.41761139e-01 -1.10297747e-01 6.27127349e-01
-7.53311038e-01 1.49581075e-01 2.98981369e-01 1.09235227e+00
1.00184977e+00 3.80198509e-01 -6.01180971e-01 1.13598835e+00
1.10837519e-01 5.34020603e-01 1.94916472e-01 -1.23479712e+00
2.39561751e-01 6.11787558e-01 4.75225551e-03 -9.38705385e-01
-2.84656975e-03 -4.87080395e-01 -1.20215893e+00 5.38247645e-01
2.71413773e-01 -1.10318772e-01 -9.56491590e-01 2.31467605e+00
4.95009154e-01 2.69895434e-01 -6.44401982e-02 9.56466556e-01
1.04936707e+00 4.67038393e-01 3.35107088e-01 -2.70783603e-01
1.45420921e+00 -7.11760461e-01 -9.15568531e-01 -1.26202062e-01
1.10292330e-01 -8.86932194e-01 1.00566542e+00 1.10645860e-01
-1.19719231e+00 -7.47796237e-01 -9.90039706e-01 -8.07490051e-02
1.96251478e-02 4.91499513e-01 3.67734700e-01 4.98155266e-01
-1.03110850e+00 4.59221691e-01 -1.84297457e-01 -2.19003826e-01
7.89326608e-01 5.70237041e-01 -7.29305506e-01 -9.21803340e-02
-1.33090603e+00 6.87658727e-01 1.96073353e-01 2.33998761e-01
-7.85294831e-01 -8.42240810e-01 -9.07471001e-01 1.18091941e-01
1.44731596e-01 -9.91278887e-01 7.21021652e-01 -1.76657426e+00
-1.38154662e+00 1.18605959e+00 -1.07804306e-01 8.88346806e-02
6.43025517e-01 1.17019258e-01 -4.77807790e-01 1.36455700e-01
1.38441145e-01 9.44754124e-01 1.35475171e+00 -1.32814610e+00
-4.60338175e-01 -5.14332771e-01 -2.05782875e-01 3.82378131e-01
-6.03563666e-01 1.44278616e-01 -5.33166826e-01 -9.80060875e-01
5.78223867e-03 -8.34676683e-01 2.23936662e-01 6.04515553e-01
-4.40903217e-01 1.21844515e-01 9.05676365e-01 -9.92394507e-01
9.20629203e-01 -2.53390741e+00 1.64912522e-01 2.42256492e-01
1.17551856e-01 1.73853546e-01 -5.28706551e-01 -2.56897546e-02
-3.85930240e-01 -2.23504901e-02 -4.32049513e-01 -4.78743672e-01
-2.73810118e-01 1.45993978e-01 -1.85183659e-01 4.71185774e-01
5.14152646e-01 8.39108288e-01 -5.91967583e-01 -7.39856243e-01
-6.27174526e-02 8.78320277e-01 -4.54376310e-01 3.93832892e-01
2.76748147e-02 8.23203385e-01 -4.80345696e-01 5.86044729e-01
1.22228217e+00 2.52460003e-01 -1.65741295e-01 -6.52782440e-01
2.60092586e-01 -4.87339407e-01 -1.07922244e+00 1.37661135e+00
-4.80794013e-01 2.19231412e-01 6.68753862e-01 -7.56516099e-01
1.23933685e+00 3.23114514e-01 4.36127216e-01 -6.76782906e-01
2.73646116e-01 1.67740807e-01 2.97086742e-02 -3.88101429e-01
1.68087497e-01 -3.69868845e-01 2.83271998e-01 1.80191725e-01
1.06129728e-01 5.38517237e-02 -3.99151653e-01 -1.28999874e-01
3.76463383e-01 2.75685817e-01 -2.76569687e-02 -1.75951734e-01
9.64349270e-01 -5.55545568e-01 8.47890735e-01 1.09108463e-01
-2.18586579e-01 7.72240281e-01 3.58685225e-01 -2.63908654e-01
-1.17133975e+00 -1.16160965e+00 -1.30906895e-01 1.03480697e+00
3.28086972e-01 1.55990347e-01 -1.02299619e+00 -6.34713113e-01
-4.89701070e-02 5.38083017e-01 -8.23843360e-01 -7.26495445e-01
-5.07282913e-01 -5.72137594e-01 7.51023114e-01 3.54242772e-01
1.06384027e+00 -1.25705040e+00 3.25102359e-01 5.56053482e-02
-4.20527399e-01 -9.10261929e-01 -8.55323553e-01 -5.45252264e-01
-4.91070002e-01 -7.77466595e-01 -8.76672268e-01 -1.02836955e+00
1.19735348e+00 -3.18376124e-02 6.17456257e-01 1.97990000e-01
-9.49014574e-02 1.14447303e-01 -5.16887940e-02 -4.29744154e-01
-6.22964084e-01 -2.87185222e-01 1.93520412e-02 8.71775329e-01
8.43449775e-03 -8.38054955e-01 -9.25013661e-01 5.13587475e-01
-1.02894664e+00 2.52633821e-02 7.02486634e-01 1.01160038e+00
5.76371372e-01 -1.59575734e-02 8.23355675e-01 -8.56003940e-01
5.15753567e-01 -3.28019559e-01 -1.30308554e-01 1.56294718e-01
-4.96917218e-01 -1.98935658e-01 7.94814348e-01 -8.29230547e-01
-1.51425314e+00 5.32949641e-02 -3.37395817e-01 -8.05879831e-01
-1.07443258e-01 -3.85934338e-02 -9.35898840e-01 -2.74027556e-01
4.52240288e-01 4.75156814e-01 6.12850070e-01 -1.96165383e-01
1.95745066e-01 6.56659484e-01 7.12620080e-01 -5.65265059e-01
1.00184822e+00 6.91358328e-01 1.54456869e-01 -5.03339112e-01
-6.70130968e-01 2.26856366e-01 -4.30938274e-01 -2.71985263e-01
8.14128280e-01 -1.04491270e+00 -6.79399967e-01 8.56170356e-01
-1.10155535e+00 2.10194923e-02 -2.33929724e-01 2.49100700e-01
-6.20446861e-01 1.06550694e-01 -4.32953179e-01 -5.40864646e-01
-5.52722037e-01 -1.06030548e+00 1.01830590e+00 3.98825377e-01
-6.62245899e-02 -6.42837703e-01 -5.86163819e-01 3.42749238e-01
3.66870314e-01 4.29161191e-01 1.12572622e+00 -3.55431557e-01
-4.89416420e-01 -5.11485152e-02 -3.73610705e-01 8.09672356e-01
4.95236695e-01 -9.72780883e-02 -1.16802287e+00 -4.17913347e-01
1.19289253e-02 -2.71732122e-01 8.35313141e-01 2.48482764e-01
1.30611742e+00 -6.00915372e-01 -1.64707497e-01 1.03205764e+00
1.07822514e+00 2.90888131e-01 1.08057451e+00 7.56249996e-03
7.93181658e-01 8.94342244e-01 8.32364798e-01 2.90740162e-01
1.14485584e-01 6.20215476e-01 4.98680472e-01 -5.87548196e-01
-3.70567381e-01 -3.77333909e-01 2.07243979e-01 2.16408297e-01
-2.90321976e-01 3.26505564e-02 -9.00417194e-02 3.84399354e-01
-1.48779440e+00 -9.63353336e-01 2.37706602e-01 2.31972790e+00
8.83947492e-01 -3.53489310e-01 -1.06177591e-01 -9.43427682e-02
1.20623362e+00 -1.53736714e-02 -8.02967191e-01 -2.97579825e-01
-3.41251731e-01 2.74838120e-01 1.03234597e-01 1.81699768e-01
-1.09922397e+00 9.55250978e-01 4.92199850e+00 1.24218190e+00
-1.24371696e+00 1.64334118e-01 9.65520263e-01 1.43604591e-01
-4.14901048e-01 -3.06256175e-01 -4.61789906e-01 4.68611628e-01
2.11495087e-01 -2.57698417e-01 4.75673586e-01 7.18001962e-01
3.34494084e-01 4.05817091e-01 -7.79059410e-01 8.63332808e-01
1.35426104e-01 -8.88125598e-01 4.44516182e-01 4.64116149e-02
5.62953889e-01 -8.95183206e-01 5.44253826e-01 2.48502791e-01
-2.14736402e-01 -1.22953308e+00 6.74210906e-01 5.42228758e-01
1.52597821e+00 -1.06302369e+00 6.44583881e-01 8.57091099e-02
-1.11068821e+00 -6.33661076e-02 -3.89102042e-01 3.27592611e-01
-1.83375552e-01 1.79650411e-01 -5.25079727e-01 6.73458755e-01
5.86124659e-01 5.06622434e-01 -4.14798588e-01 6.37561142e-01
-3.72219503e-01 1.62384644e-01 -1.69025455e-02 6.41987026e-01
-1.48905054e-01 -6.19855642e-01 5.72509766e-01 6.31230652e-01
5.64904571e-01 1.38621911e-01 -2.12519333e-01 1.19699109e+00
-4.42258865e-01 4.05724376e-01 -5.47638893e-01 3.98654938e-01
6.58488214e-01 1.50545180e+00 -1.47579774e-01 -2.23980904e-01
-2.88282961e-01 1.39659011e+00 1.81665644e-01 6.00704610e-01
-9.08870280e-01 -4.32271212e-01 9.66359556e-01 3.28449517e-01
2.13134468e-01 5.71973383e-01 -1.40889034e-01 -6.81361258e-01
1.28391668e-01 -8.64906907e-01 2.40064159e-01 -9.35071349e-01
-1.25554037e+00 8.10802937e-01 -2.50485182e-01 -1.49451780e+00
2.28562765e-02 -1.78756148e-01 -9.14229631e-01 1.33288336e+00
-1.42218220e+00 -1.69454932e+00 -5.92798889e-01 8.52796137e-01
8.92735422e-02 -3.82477641e-01 7.47029424e-01 5.17252386e-01
-5.24469554e-01 1.16161966e+00 -1.81369543e-01 1.83472544e-01
1.15186691e+00 -7.04524338e-01 -1.09615117e-01 4.86865580e-01
-6.39260590e-01 6.39688075e-01 4.56801474e-01 -6.37516141e-01
-9.35996532e-01 -1.75825202e+00 5.57486773e-01 6.03997074e-02
1.04036734e-01 -2.25475147e-01 -1.09108388e+00 6.30141497e-01
1.26059100e-01 1.68723643e-01 4.43858922e-01 -4.95901257e-01
-3.80389929e-01 -5.98093867e-01 -1.59977496e+00 7.10917771e-01
1.08429313e+00 -3.43126059e-01 -8.73643681e-02 2.52081007e-02
6.30965948e-01 -3.05808395e-01 -1.04748929e+00 6.49163425e-01
6.60631537e-01 -7.61597097e-01 1.12198448e+00 -4.05864149e-01
6.34778142e-01 -3.43100876e-01 1.96365982e-01 -1.39988279e+00
-4.12975580e-01 -6.15904868e-01 4.78008479e-01 1.87585664e+00
1.18069500e-01 -5.89568794e-01 7.38916814e-01 5.48284054e-01
-2.35341385e-01 -6.61227226e-01 -1.01505756e+00 -8.41760755e-01
1.75025910e-01 2.62998104e-01 1.00945270e+00 1.07517338e+00
-7.20722497e-01 1.94836929e-01 -6.25260115e-01 2.40009695e-01
7.47906506e-01 3.36840600e-02 6.88486040e-01 -1.08220541e+00
1.75884858e-01 -2.89562285e-01 -3.22051644e-01 -5.62327147e-01
6.94085002e-01 -9.56439257e-01 -1.77737266e-01 -1.23732519e+00
1.69527546e-01 -4.74463791e-01 -3.00873607e-01 5.82664549e-01
-4.30057853e-01 5.81285059e-01 4.45499346e-02 1.08649112e-01
1.77775651e-01 9.99191523e-01 1.99295843e+00 -2.78399289e-01
-7.82493949e-02 1.61220416e-01 -9.69196498e-01 6.59887791e-01
8.06929231e-01 -2.87900507e-01 -5.04810035e-01 -9.24653858e-02
-2.82243550e-01 1.88386235e-02 5.62284350e-01 -6.58143699e-01
-2.50312164e-02 -2.53567815e-01 8.01212132e-01 -1.44457482e-02
6.57947004e-01 -9.15535808e-01 3.08067501e-01 2.75586575e-01
-1.21910416e-01 -2.86314577e-01 2.36394018e-01 3.21849197e-01
-3.89809012e-01 6.34875596e-02 1.22637916e+00 3.60788517e-02
-6.04463398e-01 7.37635434e-01 8.25124681e-02 -1.41393811e-01
1.14813411e+00 -5.76389432e-01 -2.40127683e-01 -6.38725579e-01
-7.62228549e-01 1.24747910e-01 5.80428123e-01 4.58039224e-01
8.93000245e-01 -1.71795058e+00 -1.17076719e+00 5.34261346e-01
2.58775234e-01 -2.08161294e-01 4.90946352e-01 7.33018219e-01
-1.05057359e-01 -3.41129929e-01 -7.33812392e-01 -2.13789925e-01
-1.43048263e+00 6.86899960e-01 4.81878400e-01 2.41628066e-01
-3.77228677e-01 8.52842629e-01 9.94160354e-01 -4.82081383e-01
-1.59393512e-02 4.89022702e-01 -3.66657913e-01 3.40690874e-02
4.95074630e-01 1.55404165e-01 -1.53987065e-01 -9.52158391e-01
-2.31839895e-01 6.92099094e-01 -4.81821708e-02 1.90763652e-01
1.05205798e+00 -2.30796471e-01 -4.59930629e-01 -2.37649858e-01
1.20128489e+00 8.58095214e-02 -1.43453717e+00 -2.25719422e-01
-8.41271281e-01 -5.20119548e-01 -2.01038122e-01 -7.10811019e-01
-1.54708552e+00 7.34368443e-01 7.57982850e-01 -5.36754489e-01
1.77155554e+00 -7.74223804e-02 6.44044697e-01 -2.88283706e-01
1.45212680e-01 -8.96250606e-01 1.87664162e-02 2.70138904e-02
1.27582157e+00 -1.04365611e+00 -2.28835806e-01 -8.70318472e-01
-6.80404723e-01 8.06184471e-01 1.10963094e+00 8.80798548e-02
3.32697958e-01 2.88672838e-02 1.66996509e-01 1.33728221e-01
-3.00587744e-01 -3.40279751e-02 3.80321354e-01 9.32834864e-01
1.27674133e-01 -1.34659678e-01 -3.03748280e-01 7.37504900e-01
-1.52431875e-01 -1.26655146e-01 2.82232404e-01 2.04117388e-01
-1.15255065e-01 -1.04424250e+00 -5.13994873e-01 4.48662728e-01
-4.23644722e-01 -1.25817969e-01 -3.78747582e-01 8.27006519e-01
5.81021190e-01 5.97572565e-01 2.23328322e-01 -4.16036814e-01
4.09002930e-01 1.03627659e-01 5.90745807e-01 -3.49935144e-01
-5.90424895e-01 1.29648775e-01 -4.19302210e-02 -4.83262330e-01
-2.57701457e-01 -3.59298408e-01 -1.15025711e+00 -4.09647018e-01
-6.43298998e-02 -1.43036753e-01 3.49665910e-01 5.63570678e-01
3.99508238e-01 4.91042197e-01 9.77377295e-01 -4.58175898e-01
-3.39381337e-01 -1.05834353e+00 -9.16095614e-01 9.27460790e-01
1.69354066e-01 -7.53697157e-01 -3.03714246e-01 3.24438483e-01] | [12.783480644226074, 0.05958179011940956] |
99220286-819f-47bf-91ae-306000991893 | automatic-spatial-context-sensitive | 1701.04256 | null | http://arxiv.org/abs/1701.04256v1 | http://arxiv.org/pdf/1701.04256v1.pdf | Automatic Spatial Context-Sensitive Cloud/Cloud-Shadow Detection in Multi-Source Multi-Spectral Earth Observation Images: AutoCloud+ | The proposed Earth observation (EO) based value adding system (EO VAS),
hereafter identified as AutoCloud+, consists of an innovative EO image
understanding system (EO IUS) design and implementation capable of automatic
spatial context sensitive cloud/cloud shadow detection in multi source multi
spectral (MS) EO imagery, whether or not radiometrically calibrated, acquired
by multiple platforms, either spaceborne or airborne, including unmanned aerial
vehicles (UAVs). It is worth mentioning that the same EO IUS architecture is
suitable for a large variety of EO based value adding products and services,
including: (i) low level image enhancement applications, such as automatic MS
image topographic correction, co registration, mosaicking and compositing, (ii)
high level MS image land cover (LC) and LC change (LCC) classification and
(iii) content based image storage/retrieval in massive multi source EO image
databases (big data mining). | ['Andrea Baraldi'] | 2017-01-16 | null | null | null | null | ['shadow-detection'] | ['computer-vision'] | [ 5.48157275e-01 -7.46999800e-01 1.98547244e-01 -9.25400387e-03
-2.55014002e-01 -8.92830789e-01 5.24942100e-01 2.96222448e-01
-2.88496912e-01 5.72912335e-01 -4.04660851e-01 -6.49538219e-01
-5.76615334e-01 -1.07326257e+00 -1.15718648e-01 -6.29669130e-01
-2.40163669e-01 1.24733873e-01 1.93273276e-01 -8.91042233e-01
2.17758462e-01 1.05432248e+00 -2.27243829e+00 -3.59510966e-02
1.01812661e+00 1.57873297e+00 8.08416724e-01 9.46400225e-01
-2.84098405e-02 5.45005262e-01 -3.00395004e-02 5.15486419e-01
6.52412832e-01 2.22496197e-01 -5.86935520e-01 1.05181322e-01
3.14513534e-01 9.84499082e-02 5.86109519e-01 1.36847675e+00
1.92269057e-01 -8.78896713e-02 4.66068804e-01 -1.51925695e+00
-4.73540008e-01 -3.54638040e-01 -7.11164773e-01 3.61256331e-01
-2.15932488e-01 2.81730276e-02 3.92286837e-01 -1.01980460e+00
7.64051080e-01 8.54562163e-01 5.21813869e-01 -6.17646337e-01
-5.73813856e-01 -5.99687994e-01 -3.99515092e-01 3.94259632e-01
-1.58817947e+00 -2.14645743e-01 2.63286352e-01 -5.47161400e-01
1.02192843e+00 9.73315120e-01 9.81423974e-01 -5.54532349e-01
6.12870038e-01 5.93025722e-02 1.56794178e+00 -7.16440916e-01
2.27280840e-01 2.77225941e-01 -9.66136828e-02 6.79547340e-02
5.41003287e-01 3.41695935e-01 -1.74191847e-01 -4.64474633e-02
5.09795189e-01 2.61059761e-01 -4.42478899e-03 -3.56663316e-02
-1.02477825e+00 6.57114208e-01 6.25310183e-01 4.17363018e-01
-9.95968640e-01 -4.80884388e-02 1.18104376e-01 9.33296740e-01
5.55310071e-01 1.78858310e-01 -8.73909891e-01 2.35004067e-01
-1.06565726e+00 2.46947825e-01 -9.38946158e-02 1.10904336e+00
1.24075377e+00 4.70597029e-01 4.75383222e-01 3.84575486e-01
6.26873434e-01 1.28469670e+00 2.76166320e-01 -9.06291127e-01
9.09378827e-02 7.93925881e-01 3.51788491e-01 -1.05881512e+00
-2.83876687e-01 -4.47299689e-01 -7.83317149e-01 4.46352750e-01
-9.26013172e-01 -1.13022253e-01 -9.38923419e-01 7.62667716e-01
4.72512901e-01 -1.23833403e-01 4.06003296e-01 7.41256475e-01
9.50173140e-01 6.78468645e-01 4.55219299e-01 -2.15588942e-01
1.52632773e+00 -3.98592532e-01 -7.30184913e-01 -4.00502175e-01
5.25629222e-01 -8.46110106e-01 3.59648526e-01 -1.40981346e-01
-5.27953148e-01 -5.25021851e-01 -9.87680078e-01 4.77562279e-01
-1.33887255e+00 9.51015279e-02 1.16824293e+00 4.87466902e-01
-1.44879615e+00 4.19988334e-01 -1.53180212e-01 -4.73546565e-01
3.06693614e-01 2.87049681e-01 -3.03689659e-01 3.98426317e-03
-1.28609204e+00 8.40534747e-01 7.13802218e-01 2.72281989e-02
-4.20603782e-01 -7.07108915e-01 -8.04418683e-01 2.91971918e-02
1.30936736e-02 -5.69793105e-01 2.63914198e-01 -1.63780820e+00
-5.73293924e-01 1.40098655e+00 1.43288206e-02 -5.35859764e-01
-3.69601436e-02 8.39621872e-02 -1.12941277e+00 3.41346949e-01
4.50913370e-01 5.38223267e-01 9.77820814e-01 -1.44186175e+00
-1.08920181e+00 -9.11136508e-01 -3.78882349e-01 4.50213224e-01
-1.06151380e-01 5.66889822e-01 4.80648398e-01 -6.31171644e-01
3.58845830e-01 -1.05372727e+00 -3.39288652e-01 1.84698567e-01
4.84758884e-01 4.79457021e-01 1.35940647e+00 -8.40561569e-01
1.00552464e+00 -2.23390532e+00 -4.28788424e-01 5.28015256e-01
-1.67684451e-01 6.03204966e-01 -1.69188365e-01 4.62537974e-01
-4.45884049e-01 2.59043515e-01 -5.76473355e-01 5.43246925e-01
-5.82702160e-01 2.90900409e-01 -1.96427837e-01 3.18274081e-01
6.43110201e-02 8.34272861e-01 -6.24600053e-01 -4.38860446e-01
7.12198198e-01 1.45671904e-01 3.30023944e-01 -1.08201616e-01
-7.24260509e-02 2.53078103e-01 -3.06680888e-01 1.29685307e+00
1.47647798e+00 5.11106104e-02 -5.98950237e-02 -7.31160119e-02
-9.26304936e-01 -8.25222850e-01 -1.66957092e+00 1.07159376e+00
-3.36070895e-01 8.21662486e-01 3.58720630e-01 -4.43611532e-01
9.64596212e-01 3.52698833e-01 7.23569274e-01 -7.81794429e-01
6.85465187e-02 4.67363775e-01 -3.33331048e-01 -6.57812238e-01
1.21018612e+00 -6.09621964e-02 6.95595503e-01 3.30066793e-02
-2.84623981e-01 -3.36328447e-02 -1.43045649e-01 -3.24363783e-02
1.97800130e-01 4.80671937e-04 7.29457676e-01 -6.02027655e-01
6.60060585e-01 8.89974415e-01 1.45237833e-01 1.66322574e-01
-3.27092677e-01 3.00752342e-01 -5.79740107e-01 -4.34871405e-01
-1.18956208e+00 -7.13519454e-01 -5.65821767e-01 8.37030292e-01
3.72220248e-01 2.20454380e-01 -2.51412511e-01 5.47468066e-01
3.51526171e-01 4.64997530e-01 -2.09851280e-01 4.51742083e-01
2.71427423e-01 -1.19504976e+00 2.97950720e-03 8.90071243e-02
1.10137093e+00 -1.10705650e+00 -1.07588935e+00 3.05911869e-01
-7.06627369e-02 -9.49411809e-01 4.52483416e-01 1.58034358e-02
-1.28467143e+00 -1.13972938e+00 -3.64910483e-01 -4.81920183e-01
4.46083874e-01 1.07630050e+00 1.10512042e+00 3.03199798e-01
-6.39567077e-01 6.03844106e-01 -6.84147179e-01 -9.69108164e-01
-1.53702006e-01 -5.96974075e-01 -6.21693768e-02 2.00173110e-01
3.16511422e-01 -5.38981199e-01 -6.42170012e-01 -2.48661619e-02
-1.34047151e+00 9.82596166e-03 7.70117998e-01 2.03861684e-01
8.32142889e-01 4.40384299e-01 2.86087424e-01 -6.25394821e-01
1.69893041e-01 -7.35266030e-01 -1.02879703e+00 4.67848778e-01
-1.09293973e+00 -6.02200747e-01 -2.01852754e-01 3.63046020e-01
-1.15805435e+00 1.48082405e-01 1.29130498e-01 -2.78335243e-01
-5.66055357e-01 8.85467291e-01 8.40605274e-02 -8.95621061e-01
6.14292979e-01 6.06905460e-01 -1.34120762e-01 -3.39421630e-01
2.04650491e-01 1.41635036e+00 6.39578104e-01 3.05178821e-01
1.03895164e+00 8.50578547e-01 3.96579266e-01 -1.26554048e+00
5.89703470e-02 -1.15622962e+00 -9.59620774e-01 -6.16420329e-01
9.70126748e-01 -1.53331888e+00 -1.73517764e-01 6.09263003e-01
-8.69536221e-01 6.83805943e-02 -4.60680798e-02 1.90272644e-01
-1.72499731e-01 1.82093412e-01 3.19007337e-01 -1.07226777e+00
-8.16441596e-01 -5.46845615e-01 9.68757212e-01 3.95727992e-01
5.79300165e-01 -7.60295510e-01 -6.98413923e-02 2.78511167e-01
9.69700396e-01 1.03403604e+00 5.48324347e-01 3.63379121e-01
-8.78896594e-01 -3.05928081e-01 -6.91092908e-01 4.49803114e-01
3.03229153e-01 2.88957626e-01 -8.37167859e-01 -2.47283205e-01
-2.74597406e-01 3.83260876e-01 6.11495733e-01 5.51420927e-01
6.78182483e-01 -1.20393492e-01 -2.25101024e-01 8.69839311e-01
2.52181888e+00 2.97861159e-01 1.10064435e+00 1.05038202e+00
3.08417380e-01 3.49320263e-01 1.55380428e+00 8.19783926e-01
2.09554136e-01 1.99123740e-01 1.19032419e+00 -4.86803561e-01
1.41195282e-01 6.39596224e-01 8.23536143e-02 2.42414236e-01
-6.91972196e-01 1.75881013e-01 -8.79907787e-01 8.16262841e-01
-1.67856121e+00 -1.07422948e+00 -9.88487959e-01 2.18900609e+00
1.54978544e-01 -6.32915080e-01 -3.18945765e-01 9.41392481e-02
7.68754423e-01 2.98431039e-01 -4.10578996e-01 -3.60355258e-01
-6.58702195e-01 5.00250459e-01 1.48934996e+00 4.23211791e-02
-1.24890614e+00 1.14058793e+00 5.08801413e+00 5.31721711e-01
-1.16720867e+00 4.80907053e-01 9.90442559e-02 6.41833365e-01
-5.27259052e-01 3.47951829e-01 -3.64064515e-01 2.06842810e-01
9.43966210e-01 -1.78654283e-01 4.03514832e-01 1.10828257e+00
2.98974782e-01 -6.13251209e-01 5.08336663e-01 1.02416515e+00
-3.57902884e-01 -1.65342021e+00 8.08174014e-02 1.42966568e-01
1.05403137e+00 7.92985618e-01 -2.52764046e-01 -5.30504286e-01
9.27517414e-02 -5.44793606e-01 7.67328620e-01 5.81825018e-01
1.28127491e+00 -7.51343071e-01 9.19052362e-01 7.06991032e-02
-1.81669724e+00 -1.97992399e-01 -4.75425929e-01 3.84777561e-02
-1.90979227e-01 3.89082670e-01 1.10696061e-02 1.19009650e+00
1.28827798e+00 6.35087252e-01 -8.67882848e-01 1.04103720e+00
3.18361938e-01 8.80019665e-02 -3.72291446e-01 5.97791433e-01
4.16314811e-01 -5.01888335e-01 8.36925328e-01 8.61091852e-01
5.91399372e-01 6.63884640e-01 -3.51921350e-01 3.81491780e-01
5.34803331e-01 3.01131248e-01 -7.48273432e-01 -1.65470049e-01
4.18726474e-01 1.70123196e+00 -6.12688541e-01 -5.68295240e-01
-2.23232895e-01 7.58114398e-01 -6.78474963e-01 1.58569321e-01
-1.91360518e-01 -5.35767019e-01 1.05240369e+00 1.13013089e-01
1.37864709e-01 -5.25735259e-01 -3.04982632e-01 -1.02626371e+00
-2.91578174e-01 -4.38710392e-01 3.48658115e-01 -1.54884005e+00
-4.56627280e-01 6.50340140e-01 -4.78464328e-02 -1.88142443e+00
-1.55443354e-02 -5.77104449e-01 -4.64690119e-01 1.01676524e+00
-2.39213300e+00 -1.45804358e+00 -1.09699094e+00 8.68334293e-01
1.75987801e-03 -3.62227947e-01 8.52774918e-01 1.78760633e-01
2.37594143e-01 -4.74390686e-01 7.96828747e-01 -5.77005088e-01
3.32296312e-01 -1.22562802e+00 -1.44288868e-01 8.93036842e-01
-3.01359266e-01 2.99008321e-02 5.36894977e-01 -8.71498108e-01
-1.46167600e+00 -1.76793611e+00 7.60923028e-01 1.10407256e-01
3.33173960e-01 6.14367306e-01 -6.63847744e-01 7.10370183e-01
7.26215392e-02 8.01711828e-02 5.80557108e-01 -6.35528266e-01
3.18694174e-01 -5.31149030e-01 -1.67415273e+00 2.59780645e-01
5.95817804e-01 -7.89262727e-02 1.74999107e-02 8.11297774e-01
5.44191122e-01 -1.55087113e-01 -1.28587377e+00 6.51628911e-01
5.76741278e-01 -9.94466782e-01 1.11342967e+00 -2.98221618e-01
2.33839169e-01 -8.49601686e-01 -6.53962731e-01 -8.90635788e-01
-6.32006526e-01 -1.42010152e-01 7.31790006e-01 9.64337826e-01
-1.63034722e-01 -9.25834835e-01 -1.62414145e-02 3.49126220e-01
-3.56503457e-01 2.00512111e-01 -8.51840317e-01 -9.72123623e-01
-5.74147522e-01 -5.39507687e-01 1.03893065e+00 1.16472495e+00
-8.86325657e-01 -5.83052158e-01 -1.95922256e-01 1.02177811e+00
7.62701571e-01 4.85012144e-01 6.25667751e-01 -1.92577326e+00
4.64496583e-01 9.48680714e-02 -5.44291496e-01 3.03778946e-01
-5.59275866e-01 -6.27351165e-01 -6.42292619e-01 -1.56525064e+00
-3.90314013e-02 -8.54558825e-01 -1.79121956e-01 4.16553974e-01
1.40101701e-01 6.31059170e-01 4.06979263e-01 8.10304761e-01
-6.25481978e-02 3.50004852e-01 9.05503213e-01 1.13836661e-01
2.44926978e-02 -1.20314650e-01 -3.97606373e-01 5.21685660e-01
6.72400832e-01 -6.37064695e-01 1.28010646e-01 -2.55040884e-01
6.87177241e-01 2.08611578e-01 7.97007203e-01 -1.16180551e+00
2.84339068e-03 -8.70672107e-01 2.98684746e-01 -1.20162666e+00
-1.23916231e-01 -1.39805818e+00 1.02992797e+00 6.95079744e-01
6.08105361e-01 2.82234132e-01 5.25939703e-01 2.28299439e-01
-6.17533803e-01 -3.48470390e-01 1.02035964e+00 -4.90419418e-01
-1.81211579e+00 5.83824396e-01 -5.50187767e-01 -8.36232960e-01
1.20497918e+00 -5.50805032e-01 -2.01908156e-01 -1.69212937e-01
-5.35962462e-01 2.63845891e-01 7.92365611e-01 3.41030329e-01
5.18607855e-01 -1.30898619e+00 -7.20295548e-01 3.95377874e-01
7.72725821e-01 -1.76666066e-01 6.82678044e-01 4.52299446e-01
-1.14986813e+00 5.57555020e-01 -5.89484751e-01 -5.22272229e-01
-1.69231629e+00 2.10551932e-01 3.31246912e-01 2.96706911e-02
-2.18260750e-01 2.77948171e-01 -5.53301692e-01 -4.09867942e-01
-9.21637952e-01 5.58650270e-02 -5.55499434e-01 1.50997385e-01
5.50833642e-01 3.72729331e-01 3.03341269e-01 -1.19970477e+00
-5.53486824e-01 8.88335228e-01 9.14734662e-01 7.84846619e-02
1.75333190e+00 -6.08184636e-01 -6.93216980e-01 2.44376853e-01
6.52954102e-01 -4.29199129e-01 -6.14989340e-01 -2.54596353e-01
1.28165290e-01 -8.68489325e-01 6.88748658e-01 -9.63209271e-01
-1.06827724e+00 5.07067025e-01 1.58621359e+00 1.45109117e-01
1.70091617e+00 -4.34880137e-01 5.65892100e-01 4.45348412e-01
4.73147899e-01 -1.40048122e+00 -8.22941601e-01 2.58438885e-01
1.15037274e+00 -1.74143100e+00 4.84038174e-01 -4.92725849e-01
-7.64918208e-01 1.20080960e+00 -2.47623492e-02 -1.10707417e-01
1.08618248e+00 1.34584993e-01 1.53372407e-01 -5.82831502e-01
-4.12326545e-01 -1.02113044e+00 9.92791057e-02 9.47612107e-01
-1.63643852e-01 6.40419185e-01 -8.05097297e-02 -1.68509528e-01
1.54030591e-01 4.29269373e-01 5.33397496e-01 1.07634199e+00
-9.35736418e-01 -6.67104602e-01 -1.03687215e+00 7.17127979e-01
-4.56237532e-02 -6.26096785e-01 3.34336376e-03 8.07399392e-01
3.68983299e-01 7.93215156e-01 2.44711772e-01 -1.79052457e-01
1.96185913e-02 -3.17024261e-01 -1.39217051e-02 -2.98523098e-01
-6.52493954e-01 -1.32635802e-01 -9.46658477e-02 -1.78967848e-01
-9.70021546e-01 -9.69040453e-01 -1.01774931e+00 -4.29642111e-01
-4.61109132e-01 -2.48113364e-01 1.47240460e+00 9.07457292e-01
5.81032872e-01 1.70928881e-01 9.58145380e-01 -7.24731147e-01
1.89128354e-01 -9.92300272e-01 -1.53392315e+00 1.29896358e-01
3.56898189e-01 -7.05091000e-01 -2.18068674e-01 1.33100793e-01] | [9.728398323059082, -1.7752231359481812] |
3fa89a66-02bb-4c95-a0e2-15f7bb4412f0 | artificial-counselor-system-for-stock-1 | 1903.00955 | null | https://arxiv.org/abs/1903.00955v1 | https://arxiv.org/pdf/1903.00955v1.pdf | Artificial Counselor System for Stock Investment | This paper proposes a novel trading system which plays the role of an artificial counselor for stock investment. In this paper, the stock future prices (technical features) are predicted using Support Vector Regression. Thereafter, the predicted prices are used to recommend which portions of the budget an investor should invest in different existing stocks to have an optimum expected profit considering their level of risk tolerance. Two different methods are used for suggesting best portions, which are Markowitz portfolio theory and fuzzy investment counselor. The first approach is an optimization-based method which considers merely technical features, while the second approach is based on Fuzzy Logic taking into account both technical and fundamental features of the stock market. The experimental results on New York Stock Exchange (NYSE) show the effectiveness of the proposed system. | ['Mark Crowley', 'Ali Saheb Pasand', 'Benyamin Ghojogh', 'Hadi NekoeiQachkanloo'] | 2019-03-03 | artificial-counselor-system-for-stock | https://www.aaai.org/ojs/index.php/AAAI/article/view/5016 | https://arxiv.org/pdf/1903.00955.pdf | proceedings-of-the-aaai-conference-on | ['stock-market-prediction', 'stock-price-prediction', 'stock-prediction'] | ['time-series', 'time-series', 'time-series'] | [-6.59546673e-01 -8.77588913e-02 -3.40063810e-01 -3.48977655e-01
3.41422111e-01 -6.23565972e-01 2.78195739e-01 -1.39050916e-01
-3.41578692e-01 9.13247883e-01 -1.60025477e-01 -5.24138331e-01
-5.04124701e-01 -1.16055357e+00 -3.35437544e-02 -4.32143778e-01
2.53632516e-01 4.09524918e-01 3.20291698e-01 -6.33626938e-01
1.19082427e+00 8.03350866e-01 -1.23901284e+00 6.16101176e-02
7.22120821e-01 1.40037847e+00 -2.77577676e-02 7.13159367e-02
-4.91279393e-01 1.16782105e+00 -5.94025791e-01 -5.03939629e-01
8.80870998e-01 -1.94868192e-01 3.41662802e-02 -3.42540741e-02
-6.54948175e-01 -6.26080513e-01 5.27648143e-02 1.32166898e+00
-1.71701819e-01 7.58719519e-02 7.14050055e-01 -1.05840194e+00
-3.25287580e-01 8.09829414e-01 -4.05220330e-01 4.99120116e-01
-2.36049131e-01 -2.43415788e-01 1.01580477e+00 -8.00678194e-01
6.53713895e-03 7.72004545e-01 3.02583367e-01 7.39841983e-02
-5.37610173e-01 -7.52351105e-01 1.13653049e-01 2.78970357e-02
-1.06703782e+00 7.04201460e-02 9.06133413e-01 -5.20501494e-01
8.12729895e-01 2.92704612e-01 8.93207550e-01 -4.47886407e-01
6.95340276e-01 2.73281068e-01 1.33073401e+00 -4.45621431e-01
5.56089938e-01 7.86518991e-01 3.58628958e-01 1.25750884e-01
7.30842292e-01 3.48359555e-01 1.58473641e-01 -4.30092752e-01
8.93150449e-01 2.65612930e-01 -2.23929867e-01 1.00733742e-01
-5.91363549e-01 1.07391441e+00 4.95758951e-02 6.82795465e-01
-5.64246178e-01 -3.21038008e-01 3.82891633e-02 4.81236517e-01
3.61556113e-01 5.35748303e-01 -4.33658391e-01 2.41762057e-01
-1.12544096e+00 1.73233345e-01 1.09577024e+00 5.13335645e-01
3.92978728e-01 4.72833902e-01 2.76691616e-01 1.92078233e-01
8.23428869e-01 3.71193588e-01 7.78130591e-01 -8.43284547e-01
3.42567027e-01 6.55283213e-01 4.86818761e-01 -1.24615896e+00
-1.40583500e-01 -4.08059508e-01 -2.25674435e-01 8.93527925e-01
2.38641411e-01 -4.25355673e-01 -4.22761589e-01 8.49807858e-01
-1.97323978e-01 2.38312602e-01 3.81753474e-01 6.06468022e-01
9.35716853e-02 8.15816462e-01 -5.84151268e-01 -6.25977635e-01
1.00568807e+00 -7.27188826e-01 -9.27796781e-01 4.62958455e-01
-1.88256413e-01 -7.47389317e-01 1.25409499e-01 8.97832870e-01
-1.10342407e+00 -3.11038345e-01 -1.36572230e+00 8.77478004e-01
-5.37832797e-01 2.22396806e-01 4.73855436e-01 7.16542959e-01
-8.88951838e-01 1.00781655e+00 -4.81494635e-01 3.59404534e-01
-2.56759435e-01 5.01020133e-01 5.07379651e-01 9.94731307e-01
-1.27937984e+00 1.11450934e+00 7.68741310e-01 4.28098172e-01
-1.24980407e-02 -2.51849204e-01 -3.21357191e-01 3.59011263e-01
2.62294173e-01 8.61270055e-02 1.10292137e+00 -1.03792882e+00
-1.66012752e+00 1.58712745e-01 5.12344837e-01 -8.92294824e-01
6.62948728e-01 2.88687766e-01 -5.98547161e-01 3.66026849e-01
-3.46655309e-01 -1.20045863e-01 5.43481350e-01 -8.47769678e-01
-1.06575620e+00 -9.74404141e-02 2.38584921e-01 1.59583405e-01
-2.40244240e-01 9.99196023e-02 3.14918756e-01 -9.60133195e-01
3.94718081e-01 -6.91797972e-01 -2.70281106e-01 -3.53841364e-01
4.38404046e-02 -5.29195182e-02 3.88268977e-01 -7.00385630e-01
1.32837641e+00 -1.55132294e+00 -5.52796304e-01 1.10116434e+00
-1.67490751e-01 3.59734118e-01 7.90654123e-01 4.50203717e-01
-1.71390846e-01 2.59501755e-01 5.61690293e-02 9.54486847e-01
3.07656694e-02 -2.04767242e-01 -6.14410281e-01 -1.47888828e-02
-1.88830253e-02 3.83967727e-01 -2.90191680e-01 -1.97089329e-01
2.22264692e-01 -2.46483609e-01 -2.77701557e-01 1.30852431e-01
-1.61301613e-01 -1.94443598e-01 -9.33500409e-01 7.21312463e-01
7.38140166e-01 7.55489096e-02 2.17011586e-01 -2.16071695e-01
-6.55986786e-01 -5.79255447e-02 -1.65004683e+00 2.20685363e-01
8.81298333e-02 1.52710557e-01 -1.86585531e-01 -9.67909276e-01
1.41018188e+00 4.38749164e-01 3.76335174e-01 -1.80671722e-01
3.61115277e-01 6.54381812e-01 5.34825847e-02 -2.10929483e-01
6.42071843e-01 -6.00146472e-01 2.76500940e-01 7.07682133e-01
-6.13202155e-01 1.84035220e-03 3.41522813e-01 -2.36568183e-01
3.64294678e-01 2.18212679e-02 6.37738049e-01 -6.30506396e-01
6.27071977e-01 4.23729159e-02 8.25012565e-01 1.82056010e-01
-1.83995768e-01 1.03636354e-01 8.01563561e-01 -5.79880178e-01
-6.74302816e-01 -5.11616468e-01 -9.58889201e-02 1.11239761e-01
1.72161624e-01 3.72193664e-01 -5.50899267e-01 -4.90493774e-01
3.44491035e-01 1.10767579e+00 -4.28099751e-01 2.14187548e-01
-2.54634917e-01 -7.50230789e-01 4.79029231e-02 4.85957384e-01
6.03947818e-01 -1.07323170e+00 -9.48276222e-01 4.33542162e-01
4.47266757e-01 -3.24030668e-01 -1.56939939e-01 -3.71126123e-02
-9.71762955e-01 -9.98671591e-01 -8.91694188e-01 -2.19007969e-01
6.36191547e-01 7.04649836e-02 5.43244123e-01 3.08758557e-01
3.30375463e-01 6.37883227e-03 -2.77904540e-01 -8.14406395e-01
-2.93380350e-01 -5.35195112e-01 -4.66622524e-02 1.81756496e-01
3.59383404e-01 -1.05497450e-01 -4.11532283e-01 3.35055798e-01
-7.74504244e-01 -4.55106080e-01 5.46071351e-01 6.76597655e-01
5.14332950e-01 9.61664021e-01 8.97106946e-01 -8.93168688e-01
7.65766978e-01 -4.38906312e-01 -1.35626817e+00 7.51635611e-01
-1.29734755e+00 1.26631811e-01 3.46558690e-01 -1.62289083e-01
-1.55800343e+00 -5.56280792e-01 2.72816330e-01 -3.57395798e-01
5.39912045e-01 8.22390556e-01 -1.03961185e-01 -1.74529016e-01
-1.39058635e-01 2.88486481e-01 1.54371366e-01 -4.52249944e-01
-4.01441157e-01 6.99119031e-01 1.66869372e-01 -2.02373683e-01
6.97590470e-01 1.53428808e-01 1.71281304e-02 -8.36804882e-02
-3.29590172e-01 3.08381510e-03 -4.53824282e-01 -2.47662544e-01
5.07840455e-01 -4.04744953e-01 -8.95098269e-01 4.72994894e-01
-7.19898283e-01 3.38007748e-01 -1.69471204e-01 9.16355073e-01
-4.80774909e-01 1.31209418e-02 -7.65627563e-01 -1.38560045e+00
-4.61040676e-01 -8.94582272e-01 -1.71676323e-01 5.54223955e-01
-1.59414746e-02 -1.04784596e+00 -1.00018889e-01 2.33884141e-01
3.56967211e-01 3.46855104e-01 9.88295496e-01 -1.18559158e+00
-7.88250804e-01 -4.18131858e-01 -7.05762580e-02 6.06548250e-01
2.53626376e-01 5.04352093e-01 -2.34182611e-01 -3.03502977e-02
8.28944266e-01 2.27818742e-01 4.24952388e-01 5.18281698e-01
4.24624324e-01 -4.55354840e-01 -9.66102630e-02 3.05603236e-01
1.82500923e+00 1.29789996e+00 6.62614763e-01 8.11262369e-01
-9.32179093e-02 8.69081378e-01 1.39480829e+00 9.16556478e-01
1.17668986e-01 2.28061259e-01 1.93278700e-01 4.68779474e-01
9.85114634e-01 3.36165167e-02 1.48743257e-01 7.27991879e-01
-2.99037993e-01 1.86355505e-02 -6.35831773e-01 1.21354222e-01
-1.78902626e+00 -1.28289831e+00 3.67238790e-01 2.22024202e+00
5.15609026e-01 7.90838301e-01 2.10788578e-01 3.88243556e-01
1.01986694e+00 -1.92602932e-01 -6.64618671e-01 -6.45391822e-01
-8.53118747e-02 -1.17651680e-02 8.05637240e-01 2.31136128e-01
-7.08826542e-01 2.81757712e-01 5.80046082e+00 7.31455624e-01
-1.32262933e+00 -7.15433836e-01 9.09514487e-01 6.71996921e-02
-6.78141117e-01 3.44489962e-01 -1.02543843e+00 8.55934799e-01
6.30373955e-01 -9.05207098e-01 2.46095568e-01 7.48653114e-01
2.80127525e-01 -1.99193940e-01 -4.53340590e-01 2.44191095e-01
-2.43028462e-01 -1.66698277e+00 -8.52053165e-02 1.62485987e-01
4.60281968e-01 -6.75358653e-01 2.60177646e-02 -2.06032917e-02
1.86663605e-02 -4.15761948e-01 1.15121305e+00 1.27255607e+00
-5.45178950e-02 -1.24656117e+00 1.23188651e+00 3.45499307e-01
-1.21988237e+00 -2.50488281e-01 -6.13138139e-01 -1.94415852e-01
1.49861932e-01 3.50593477e-01 -5.30933499e-01 6.12280488e-01
4.01550770e-01 2.33586490e-01 -2.10662596e-02 9.45206165e-01
8.59242603e-02 3.90462101e-01 -2.64023095e-01 -4.30879861e-01
3.36776793e-01 -1.00672090e+00 2.47078925e-01 4.56071824e-01
7.12311685e-01 7.32064545e-01 -2.15353593e-01 1.03980434e+00
3.97076517e-01 3.48842323e-01 -3.32482010e-01 -1.75908506e-01
6.96125984e-01 9.22570109e-01 -1.13631535e+00 -6.27541840e-01
-4.44889992e-01 2.16817051e-01 -4.13256943e-01 9.66586471e-02
-5.46026230e-01 -7.12440073e-01 -7.66563043e-02 3.35551977e-01
6.49891376e-01 2.54884928e-01 -7.74154544e-01 -9.55103397e-01
1.07016936e-01 -6.38828218e-01 4.33291048e-01 -5.57666779e-01
-1.15680325e+00 5.16826808e-01 1.59801081e-01 -1.53370464e+00
-5.27996063e-01 -7.60316432e-01 -1.10101891e+00 1.07356799e+00
-1.59351218e+00 -3.42051685e-01 5.59412003e-01 2.67805398e-01
2.26072550e-01 -9.98256147e-01 2.41719499e-01 -2.89030641e-01
-4.74451274e-01 2.48173937e-01 5.53874671e-01 8.09393600e-02
-2.07166635e-02 -1.28264177e+00 -2.12777212e-01 6.97150767e-01
-1.85134351e-01 6.01635993e-01 5.62360704e-01 -1.17111969e+00
-7.00521231e-01 -4.87661153e-01 1.02182782e+00 4.20011252e-01
9.55663860e-01 5.61952770e-01 -9.92339909e-01 5.22400558e-01
2.69211024e-01 -4.51127231e-01 7.07096040e-01 -6.74983680e-01
1.87684357e-01 -4.36024338e-01 -1.75508881e+00 3.44152093e-01
-2.83236355e-01 1.71656936e-01 -1.14513457e+00 -1.60231724e-01
4.16389316e-01 2.93051843e-02 -9.84349787e-01 4.62904602e-01
8.10114503e-01 -1.22325718e+00 5.98973751e-01 -8.61198530e-02
-1.95806265e-01 -2.88740009e-01 -1.67861301e-02 -1.00212252e+00
-2.98562080e-01 -5.72524011e-01 3.21675301e-01 1.23519266e+00
5.19182146e-01 -1.31595731e+00 8.52421105e-01 1.12495530e+00
2.87362516e-01 -9.51095879e-01 -6.80026829e-01 -7.60455906e-01
-1.39370978e-01 2.83646077e-01 1.01653349e+00 8.67516398e-01
1.77944705e-01 -3.46021414e-01 -1.65061042e-01 -1.54249743e-01
7.36015141e-01 6.29332721e-01 -2.01049857e-02 -1.14885747e+00
-4.53771532e-01 -8.11478555e-01 -1.89519823e-01 -3.33036900e-01
7.74184689e-02 -4.10033733e-01 -3.62245232e-01 -9.22324121e-01
-2.75507480e-01 -4.10031706e-01 -1.10265529e+00 1.06225178e-01
1.07234560e-01 -3.65366042e-01 5.37491918e-01 5.52560151e-01
3.43821973e-01 2.55417258e-01 7.46174693e-01 1.69012938e-02
-3.64497066e-01 7.17432141e-01 -8.08667362e-01 1.14027154e+00
1.06797028e+00 -2.15373427e-01 -5.35930395e-01 1.70286447e-01
1.64753601e-01 7.22686350e-01 -2.65826255e-01 -7.13022351e-01
2.93435574e-01 -8.28714848e-01 3.20478588e-01 -1.06192958e+00
5.53637482e-02 -1.09225309e+00 4.92728144e-01 1.05255461e+00
-3.86496037e-01 4.07150418e-01 -1.15666397e-01 3.52728367e-01
-5.00331283e-01 -1.01166189e+00 6.48616791e-01 -2.59743124e-01
-5.46879232e-01 -8.93601850e-02 -4.52335298e-01 -4.91505176e-01
1.37196422e+00 -5.86994350e-01 -9.77783799e-02 -3.14735651e-01
-5.27253866e-01 1.72267914e-01 2.57505298e-01 -3.86581719e-02
7.52816081e-01 -1.27320158e+00 -6.67991340e-01 4.73966412e-02
-3.40499252e-01 -7.02432752e-01 -1.13424510e-01 4.08510566e-01
-1.08875036e+00 6.44050539e-01 -5.99412799e-01 3.93929213e-01
-7.64162302e-01 4.65169698e-01 4.56023246e-01 -4.97132242e-01
-1.75473198e-01 5.93716204e-01 -1.27086461e-01 3.49152088e-01
-1.74837708e-01 -2.15227485e-01 -9.93514836e-01 4.97656137e-01
7.15017319e-01 7.70667195e-01 2.44867131e-02 -4.81975734e-01
-2.42937326e-01 6.61800146e-01 -8.03128555e-02 -4.33340907e-01
1.42387557e+00 1.51173193e-02 -4.64248449e-01 5.25323987e-01
4.54663515e-01 2.52430022e-01 -1.14519095e+00 -1.29121706e-01
7.23303735e-01 -6.87219143e-01 2.33506579e-02 -7.79147208e-01
-1.37032831e+00 3.75083864e-01 2.47478411e-01 5.94247043e-01
1.24544239e+00 -8.25630128e-01 7.58738518e-01 5.01400292e-01
6.06118798e-01 -1.51966727e+00 -5.44369161e-01 5.94796017e-02
1.03733230e+00 -8.63061547e-01 2.15369508e-01 -1.94445744e-01
-8.47091198e-01 1.62633264e+00 3.69912565e-01 -7.47602820e-01
1.42757332e+00 4.15904224e-01 1.87538415e-01 1.40891418e-01
-6.93825483e-01 2.74365932e-01 3.58112663e-01 4.30149259e-03
1.36608005e-01 1.31762967e-01 -1.10626614e+00 1.32797027e+00
-1.39691412e-01 2.71908820e-01 1.03023303e+00 1.18040383e+00
-9.69316423e-01 -9.03688073e-01 -6.49969399e-01 7.41809666e-01
-7.43942082e-01 -8.08952972e-02 -2.64285237e-01 7.31491029e-01
6.83138594e-02 8.88773441e-01 2.95464337e-01 -2.66766101e-01
4.70167488e-01 -2.34816223e-01 6.73154518e-02 -5.04072249e-01
-7.36112416e-01 4.63013053e-01 -2.86214277e-02 2.35902816e-02
-2.09828243e-01 -8.51850271e-01 -1.67789066e+00 -2.57444918e-01
-6.46080375e-01 7.49046683e-01 7.26246238e-01 9.14668143e-01
-2.42355078e-01 9.55967978e-02 1.09724724e+00 -4.84183192e-01
-1.32325339e+00 -6.37029409e-01 -1.42049801e+00 -3.54715556e-01
-1.99633643e-01 -7.39357293e-01 -5.11989594e-01 -2.63833642e-01] | [4.754644393920898, 4.02319860458374] |
a56d4ea7-0686-4875-8295-211b80773ce4 | can-sam-boost-video-super-resolution | 2305.06524 | null | https://arxiv.org/abs/2305.06524v2 | https://arxiv.org/pdf/2305.06524v2.pdf | Can SAM Boost Video Super-Resolution? | The primary challenge in video super-resolution (VSR) is to handle large motions in the input frames, which makes it difficult to accurately aggregate information from multiple frames. Existing works either adopt deformable convolutions or estimate optical flow as a prior to establish correspondences between frames for the effective alignment and fusion. However, they fail to take into account the valuable semantic information that can greatly enhance it; and flow-based methods heavily rely on the accuracy of a flow estimate model, which may not provide precise flows given two low-resolution frames. In this paper, we investigate a more robust and semantic-aware prior for enhanced VSR by utilizing the Segment Anything Model (SAM), a powerful foundational model that is less susceptible to image degradation. To use the SAM-based prior, we propose a simple yet effective module -- SAM-guidEd refinEment Module (SEEM), which can enhance both alignment and fusion procedures by the utilization of semantic information. This light-weight plug-in module is specifically designed to not only leverage the attention mechanism for the generation of semantic-aware feature but also be easily and seamlessly integrated into existing methods. Concretely, we apply our SEEM to two representative methods, EDVR and BasicVSR, resulting in consistently improved performance with minimal implementation effort, on three widely used VSR datasets: Vimeo-90K, REDS and Vid4. More importantly, we found that the proposed SEEM can advance the existing methods in an efficient tuning manner, providing increased flexibility in adjusting the balance between performance and the number of training parameters. Code will be open-source soon. | ['Xinchao Wang', 'Zhiwei Xiong', 'Jiawang Bai', 'Zeyu Xiao', 'Zhihe Lu'] | 2023-05-11 | null | null | null | null | ['video-super-resolution'] | ['computer-vision'] | [ 9.87014100e-02 -4.32933986e-01 -1.79699257e-01 -2.38816857e-01
-5.13904631e-01 -4.60381776e-01 4.61190104e-01 -3.02849710e-01
-4.06670749e-01 6.43102944e-01 4.38048661e-01 3.93760651e-02
5.73265702e-02 -8.13610494e-01 -4.91601110e-01 -5.05555809e-01
2.83627391e-01 -1.66833490e-01 7.14910746e-01 -3.79689246e-01
1.03235006e-01 4.37142253e-01 -1.62753630e+00 1.55232802e-01
1.08749020e+00 9.11338449e-01 5.22194088e-01 5.05860686e-01
-8.40130821e-02 8.33592176e-01 -2.91277796e-01 -2.04901844e-01
4.19658840e-01 -4.10280108e-01 -7.56616473e-01 1.07720077e-01
4.68505681e-01 -6.33366644e-01 -4.01692957e-01 8.31485689e-01
4.82869744e-01 4.25861776e-01 3.29934686e-01 -8.75201583e-01
-3.23175371e-01 3.06324452e-01 -7.22036123e-01 4.40057933e-01
3.64818633e-01 3.64229947e-01 8.03997576e-01 -7.93031871e-01
7.60257661e-01 1.30995011e+00 4.72119570e-01 5.83618999e-01
-1.03385782e+00 -6.76650345e-01 2.47296527e-01 3.86777371e-01
-1.23585355e+00 -7.52267241e-01 7.70637512e-01 -3.70643944e-01
6.35900438e-01 2.44000241e-01 6.54477596e-01 9.72870648e-01
-9.58143771e-02 6.22107208e-01 8.99725854e-01 -8.88586119e-02
1.07163675e-01 -4.99874875e-02 -2.19522174e-02 6.13903582e-01
2.18810990e-01 1.67103320e-01 -6.00206852e-01 2.03823507e-01
1.30008876e+00 -9.24426597e-03 -6.73328161e-01 -2.10673347e-01
-1.29997766e+00 5.54775119e-01 5.34047186e-01 3.51057261e-01
-3.41768086e-01 1.12047039e-01 1.80933282e-01 1.81332231e-02
5.09580374e-01 3.02355170e-01 -2.18925595e-01 -2.78600335e-01
-1.11466146e+00 2.66713575e-02 4.06501114e-01 6.67185366e-01
9.29541707e-01 2.41884753e-01 -3.42512250e-01 7.25415587e-01
2.65293539e-01 2.51897931e-01 3.17012966e-01 -1.48641360e+00
3.91978651e-01 4.56458777e-01 3.46448034e-01 -1.11019194e+00
-2.08952874e-01 -5.00557244e-01 -8.05842102e-01 2.93944478e-01
3.38555366e-01 3.15453745e-02 -8.81110787e-01 1.81318641e+00
5.19744575e-01 8.22731793e-01 1.03103876e-01 1.33607447e+00
8.42931092e-01 4.73565847e-01 -3.19923535e-02 -2.68852800e-01
1.20129919e+00 -1.13892210e+00 -6.88893080e-01 -9.08363536e-02
2.77591795e-01 -8.03294241e-01 9.98699427e-01 5.33169396e-02
-1.19490409e+00 -7.77144372e-01 -1.09044445e+00 -2.57184923e-01
3.17053571e-02 -1.36949524e-01 5.02862275e-01 2.22434759e-01
-1.07962656e+00 7.82397687e-01 -9.96291578e-01 -2.53236771e-01
3.95691335e-01 1.37223020e-01 -4.00653422e-01 -1.49147227e-01
-1.14226055e+00 8.31164718e-01 2.09138796e-01 1.49046868e-01
-6.88390255e-01 -9.23087239e-01 -8.98394346e-01 8.59004855e-02
4.89047855e-01 -1.07045984e+00 8.25730324e-01 -1.07432532e+00
-1.75220668e+00 3.91248256e-01 -4.66933072e-01 -3.37427646e-01
7.00028896e-01 -3.24726760e-01 -1.79567099e-01 4.36388612e-01
-9.21766162e-02 6.90232038e-01 9.38369572e-01 -1.26139534e+00
-6.23229444e-01 -6.07457869e-02 3.96562636e-01 3.74827534e-01
-3.21452677e-01 8.38453770e-02 -7.82315314e-01 -8.55567634e-01
-1.14043534e-01 -6.39581263e-01 -2.76463181e-01 9.87241194e-02
-4.25699838e-02 1.07611455e-01 8.58748376e-01 -8.19756210e-01
1.23388910e+00 -2.10834527e+00 2.83853561e-01 -1.06591895e-01
3.50939333e-01 6.86212838e-01 -1.80879459e-01 1.23201190e-02
2.70447612e-01 5.40468143e-03 -2.39808843e-01 -4.71073300e-01
-4.23779815e-01 2.78759181e-01 -1.74957812e-01 3.32251489e-01
2.62177914e-01 7.61383414e-01 -1.04958773e+00 -5.90809047e-01
6.84314191e-01 9.43837464e-01 -7.37071216e-01 2.97757059e-01
-7.29945451e-02 8.08097720e-01 -5.10752857e-01 2.78052241e-01
6.52174234e-01 -3.46206605e-01 -1.51473656e-01 -5.41917324e-01
-2.85421252e-01 1.77950948e-01 -1.50241601e+00 1.85392332e+00
-5.32742023e-01 5.06358564e-01 1.00794002e-01 -6.39143407e-01
8.42085481e-01 2.66808510e-01 6.39023900e-01 -5.94990551e-01
1.01456121e-01 5.51306866e-02 -1.93341196e-01 -4.02647078e-01
6.27156019e-01 2.89304227e-01 5.50431907e-01 1.29800335e-01
2.00464553e-03 1.80832922e-01 3.79575372e-01 2.34920591e-01
9.59494710e-01 5.14199972e-01 2.48583034e-01 -7.10244626e-02
8.10102224e-01 -2.50746459e-01 1.00400758e+00 4.73673046e-01
-1.80998266e-01 1.06523299e+00 4.50198650e-02 -3.38817745e-01
-9.69490588e-01 -9.02432084e-01 -8.59530494e-02 7.28722334e-01
6.15851343e-01 -5.77421904e-01 -6.49615467e-01 -4.88964319e-01
-2.82998174e-01 3.85203332e-01 -3.84418875e-01 -3.46370303e-04
-7.78248250e-01 -6.01085424e-01 8.36495385e-02 7.00466037e-01
7.89283931e-01 -7.96703398e-01 -7.33109713e-01 3.02461118e-01
-6.09188974e-01 -1.51505876e+00 -6.71860754e-01 -5.32505512e-01
-8.43494177e-01 -1.02334070e+00 -8.34385693e-01 -2.62930095e-01
5.03480434e-01 5.88353634e-01 8.52559566e-01 2.81446844e-01
-1.36585355e-01 3.27239245e-01 -5.36953747e-01 1.46308854e-01
-2.89088249e-01 -1.56214992e-02 -6.58252761e-02 3.20330054e-01
-1.29532769e-01 -6.73620880e-01 -9.00652647e-01 5.43235421e-01
-9.11688626e-01 5.27510345e-01 3.15490156e-01 6.67758703e-01
5.04237115e-01 -2.74915665e-01 4.59873766e-01 -5.45541644e-01
6.65336102e-02 -3.24596137e-01 -4.72076327e-01 8.36423412e-02
-4.83604819e-01 1.54643238e-01 6.64376915e-01 -2.80113399e-01
-1.27251112e+00 -3.21698189e-02 -2.30064020e-01 -7.81584263e-01
6.26444221e-02 1.99690998e-01 -6.59136027e-02 -2.68391788e-01
3.66137981e-01 1.87742844e-01 6.94985688e-02 -5.85583270e-01
4.35628533e-01 4.57095385e-01 7.88939595e-01 -5.23249269e-01
9.31588471e-01 8.17448258e-01 2.13029776e-02 -6.41609967e-01
-7.69753098e-01 -6.43496454e-01 -5.81979573e-01 -1.51478380e-01
8.95883441e-01 -1.11568844e+00 -4.85379875e-01 4.30980742e-01
-1.03671181e+00 -2.87594318e-01 -1.89830735e-01 5.94503999e-01
-5.03484547e-01 6.47907376e-01 -7.12150395e-01 -4.41481590e-01
-4.57272470e-01 -1.38030744e+00 1.00489366e+00 4.92015779e-01
1.53505756e-02 -9.90014076e-01 -9.77681577e-02 5.07089734e-01
7.22807348e-01 3.37242424e-01 2.46114984e-01 -1.78658992e-01
-9.14188147e-01 3.79393041e-01 -4.91586298e-01 5.02114773e-01
3.34245950e-01 1.38481304e-01 -9.44529235e-01 -4.25202549e-01
-2.00001568e-01 9.39070806e-02 1.00200367e+00 3.70139867e-01
9.05023277e-01 -6.65692836e-02 -8.57351720e-02 9.94297445e-01
1.53060138e+00 -3.73077877e-02 8.30759108e-01 4.66568708e-01
1.01016843e+00 3.84403348e-01 6.53610349e-01 4.47016299e-01
6.77795410e-01 1.04801905e+00 3.90543252e-01 -3.21062505e-01
-6.58991098e-01 1.65566634e-02 5.09437263e-01 8.90748024e-01
-7.61668861e-01 -8.27756748e-02 -4.60841388e-01 4.32076544e-01
-2.08392477e+00 -1.05557740e+00 2.92660184e-02 2.13278794e+00
8.26586127e-01 -2.34735742e-01 -6.17043301e-03 -8.50368850e-03
6.62757099e-01 3.75547469e-01 -3.73175830e-01 8.63215253e-02
-7.74810016e-02 2.23193273e-01 2.78484434e-01 7.00316846e-01
-9.58923161e-01 1.06807601e+00 5.35644817e+00 7.22097516e-01
-1.16811955e+00 1.64267004e-01 3.38822961e-01 -9.03382152e-02
-2.46925190e-01 1.33177653e-01 -7.44291067e-01 5.78053534e-01
6.80177033e-01 -1.32215545e-01 5.20617366e-01 5.43337584e-01
7.18308568e-01 -5.58554083e-02 -9.66858864e-01 9.91173446e-01
-1.49366990e-01 -1.58112156e+00 7.97361434e-02 -1.29130512e-01
6.82207584e-01 -6.16528355e-02 -1.66583896e-01 -1.26321211e-01
1.74130023e-01 -7.81506836e-01 7.11994410e-01 7.05793023e-01
7.55956411e-01 -5.93093872e-01 7.69716740e-01 -2.85204481e-02
-1.59263778e+00 1.31468922e-02 -2.45550931e-01 3.98231968e-02
5.60579002e-01 5.06661952e-01 -2.46136963e-01 9.64121699e-01
8.55079174e-01 1.17968071e+00 -4.94130731e-01 1.08487749e+00
-1.43663570e-01 4.17011917e-01 -2.93852925e-01 6.41274333e-01
6.04706332e-02 -2.18429685e-01 7.66812742e-01 1.13006413e+00
3.28646183e-01 1.38779610e-01 2.70057470e-01 8.22664022e-01
1.88061833e-01 4.97880094e-02 -1.18542373e-01 4.17842209e-01
4.77287769e-01 1.37248874e+00 -6.39115810e-01 -3.59021872e-01
-7.04387844e-01 8.87089670e-01 1.95066705e-01 5.23449719e-01
-1.01402724e+00 -8.57970268e-02 1.02621615e+00 2.12788284e-01
4.58028346e-01 -3.34854394e-01 -7.55669177e-02 -1.59824574e+00
1.60361070e-03 -7.83820748e-01 2.43119091e-01 -7.56868899e-01
-9.67852354e-01 6.77912593e-01 -2.44821180e-02 -1.31171918e+00
-2.85581648e-01 -1.72614247e-01 -4.75403965e-01 8.57652664e-01
-2.01049972e+00 -1.11209798e+00 -7.72189856e-01 8.59147787e-01
6.72343194e-01 1.37450308e-01 4.02146041e-01 5.25679827e-01
-7.84318984e-01 3.80098224e-01 -2.70744979e-01 1.53900906e-01
8.74987185e-01 -8.98167253e-01 2.67976373e-01 1.28559971e+00
-2.08344012e-02 5.62625945e-01 6.29023373e-01 -4.44740027e-01
-1.19958174e+00 -1.24163532e+00 3.46953839e-01 -3.76100272e-01
3.72333586e-01 -1.42270885e-02 -1.12635911e+00 5.55350721e-01
3.26793641e-02 4.40542608e-01 1.39007196e-01 -3.59889686e-01
-2.77297944e-01 -2.12763697e-01 -1.02839732e+00 5.88270843e-01
1.33537769e+00 -2.80182421e-01 -4.43549722e-01 -3.76859158e-01
1.01146913e+00 -5.65882266e-01 -9.77239132e-01 6.77493036e-01
4.24771160e-01 -1.22295153e+00 1.23873067e+00 -1.38698608e-01
3.96732181e-01 -7.45646179e-01 -1.09561391e-01 -1.17671335e+00
-3.96744967e-01 -7.44987011e-01 -4.49875355e-01 1.41865635e+00
-1.28304124e-01 -7.20190287e-01 4.83599067e-01 6.09572291e-01
-1.85604408e-01 -6.57014370e-01 -7.08353221e-01 -5.35267413e-01
-4.44354117e-01 -3.81308615e-01 6.69636190e-01 9.44426715e-01
-4.32570934e-01 1.34190336e-01 -5.02049327e-01 1.86832920e-01
6.38282061e-01 3.90228741e-02 8.42740178e-01 -1.07738388e+00
-3.45076144e-01 -3.67249608e-01 -5.22381544e-01 -1.15399599e+00
9.06529278e-03 -5.77101350e-01 -1.29557043e-01 -1.55671453e+00
1.92507863e-01 -5.10475278e-01 -3.27443540e-01 4.30146873e-01
-5.15809894e-01 4.10040408e-01 5.59651315e-01 3.30386877e-01
-6.39253974e-01 6.41665697e-01 1.51150584e+00 2.76235461e-01
-3.86635065e-01 -1.65959865e-01 -7.03630507e-01 7.65849471e-01
5.57024777e-01 -1.00319758e-01 -4.98848587e-01 -5.79236507e-01
-1.50736690e-01 1.28256693e-01 5.26361227e-01 -1.09203267e+00
1.71418488e-01 -2.41977036e-01 2.52561182e-01 -2.44848505e-01
2.96200186e-01 -6.94956720e-01 2.64540255e-01 1.12796485e-01
-2.19017230e-02 6.68338547e-03 2.40956694e-02 5.07261097e-01
-2.72762686e-01 5.99914119e-02 9.20552552e-01 -6.88650683e-02
-1.10712683e+00 6.02190077e-01 3.64812203e-02 8.13793913e-02
8.54959071e-01 -3.60874325e-01 -3.88193905e-01 -3.57318372e-01
-5.11824787e-01 2.20985696e-01 7.82692671e-01 6.11421585e-01
6.40631676e-01 -1.11761856e+00 -7.93462276e-01 8.16170573e-02
-2.09859610e-01 2.48304859e-01 3.98301482e-01 9.72308517e-01
-4.80879903e-01 2.11054534e-02 -2.69434005e-01 -6.82322204e-01
-1.06610084e+00 4.78546977e-01 3.70654225e-01 -1.78800106e-01
-9.83646929e-01 4.64774758e-01 3.60567659e-01 1.70763835e-01
6.03206381e-02 -3.25906962e-01 -4.55140382e-01 -1.32351235e-01
8.64790201e-01 5.44933677e-01 -1.67821318e-01 -9.33658063e-01
-3.45321715e-01 9.41103876e-01 1.06707916e-01 3.73544358e-02
1.29063392e+00 -5.48032641e-01 1.83475129e-02 2.08544619e-02
1.00370646e+00 6.27676621e-02 -1.88675189e+00 -4.07132387e-01
-3.47655594e-01 -7.99597859e-01 2.61206150e-01 -3.94755274e-01
-1.49500132e+00 6.53364182e-01 4.61269140e-01 -1.77997723e-01
1.32722712e+00 -1.86044782e-01 1.06016707e+00 -2.50289232e-01
5.15384078e-01 -7.48922348e-01 1.02483176e-01 2.49283448e-01
6.52872980e-01 -1.15842807e+00 4.95216995e-02 -7.44331241e-01
-5.50967395e-01 1.21105957e+00 7.56150961e-01 4.38238494e-03
3.07258576e-01 3.33243817e-01 7.13562444e-02 2.95861363e-01
-7.19151616e-01 -4.87944722e-01 4.38967586e-01 6.08978808e-01
3.57743323e-01 -3.94045144e-01 -2.13787690e-01 3.95937592e-01
5.96804544e-02 3.03632379e-01 5.77638149e-01 6.57777071e-01
-2.92597443e-01 -1.00157905e+00 -1.88421577e-01 4.56206873e-02
-5.18337131e-01 -6.40311018e-02 3.78477305e-01 6.00931704e-01
4.86362651e-02 9.63945925e-01 1.10737279e-01 -3.28832686e-01
1.87087119e-01 -4.74956900e-01 4.69267279e-01 -3.47620875e-01
-3.16775084e-01 5.86143434e-02 4.13474813e-02 -1.15991533e+00
-9.58305061e-01 -4.80811894e-01 -1.29677045e+00 -4.79356468e-01
-2.57120162e-01 -1.76187828e-01 2.82871068e-01 1.07380927e+00
6.52949512e-01 6.85875297e-01 5.63322067e-01 -1.15875471e+00
-1.39723808e-01 -6.14580154e-01 -1.22135513e-01 5.77845097e-01
4.96425211e-01 -9.20135021e-01 -3.58893603e-01 2.55731642e-01] | [10.904471397399902, -1.7194281816482544] |
cfd72d43-664d-4b86-9c86-022fe9007b99 | deep-learning-for-hand-gesture-recognition-on | null | null | https://hal-mines-paristech.archives-ouvertes.fr/hal-01737771/file/DeepLearning-HandSkeletalGestureRecognition_MINES-ParisTech_FG2018.pdf | https://hal-mines-paristech.archives-ouvertes.fr/hal-01737771/file/DeepLearning-HandSkeletalGestureRecognition_MINES-ParisTech_FG2018.pdf | Deep Learning for Hand Gesture Recognition on Skeletal Data | In this paper, we introduce a new 3D hand gesture recognition approach based on a deep learning model.
We introduce a new Convolutional Neural Network (CNN) where sequences of hand-skeletal joints’ positions are processed by parallel convolutions; we then investigate the performance of this model on hand gesture sequence classification tasks. Our model only uses hand-skeletal data and no depth image.
Experimental results show that our approach achieves a state-of-the-art performance on a challenging dataset (DHG dataset from the SHREC 2017 3D Shape Retrieval Contest), when compared to other published approaches. Our model achieves a 91.28% classification accuracy for the 14 gesture classes case and an 84.35% classification accuracy for the 28 gesture classes case. | ['Guillaume Devineau', 'Fabien Moutarde', 'Jie Yang', 'Wang Xi'] | 2018-05-15 | null | null | null | ieee-fg-2018-2018-5 | ['3d-shape-retrieval', 'temporal-information-extraction'] | ['computer-vision', 'natural-language-processing'] | [ 1.64098926e-02 -5.07207632e-01 -3.35697651e-01 -2.18005925e-01
-3.93486768e-01 -4.78106976e-01 8.95187259e-01 -6.62398756e-01
-8.67165685e-01 7.38102645e-02 1.34249687e-01 -1.86014295e-01
1.33432940e-01 -4.46257234e-01 -4.96775299e-01 -8.05128694e-01
-2.30058476e-01 7.78138041e-01 3.63364518e-01 9.02619734e-02
1.52386338e-01 1.13825095e+00 -1.22387636e+00 4.03903395e-01
-2.21569642e-01 1.21571016e+00 -1.49644434e-01 9.69062030e-01
-1.97630644e-01 6.53527200e-01 -6.41889989e-01 -2.21170038e-01
6.88242555e-01 8.12467858e-02 -1.09520006e+00 -2.07469482e-02
7.41947174e-01 -1.22384930e+00 -8.54338765e-01 4.08134907e-01
1.21625102e+00 7.34023005e-02 9.31275010e-01 -9.12892520e-01
-6.89859331e-01 2.06567258e-01 -4.91595656e-01 1.64992824e-01
3.20464134e-01 2.90590405e-01 8.87369394e-01 -1.11960828e+00
9.25735593e-01 1.23532367e+00 6.42116129e-01 1.04629207e+00
-7.03270972e-01 -4.61569428e-01 5.63032255e-02 4.05598074e-01
-1.41173553e+00 1.16334118e-01 5.33636928e-01 -3.80488098e-01
1.48930085e+00 3.65908854e-02 8.99538577e-01 1.41018355e+00
-2.74610490e-01 1.65377069e+00 6.94808781e-01 -4.04492110e-01
1.81645229e-01 -1.25661135e+00 2.84710288e-01 8.59375000e-01
-9.39854160e-02 3.04842263e-01 -6.12303078e-01 1.02585755e-01
1.25517714e+00 2.22910747e-01 1.96298677e-02 -3.39910239e-01
-1.19806743e+00 4.47893202e-01 6.19810164e-01 3.94243449e-01
-4.90033895e-01 3.21335733e-01 3.13223690e-01 3.21737140e-01
1.64199874e-01 -4.90022004e-01 -4.27295715e-01 -4.14178222e-01
-9.48230445e-01 7.25941360e-01 9.89842296e-01 9.46698427e-01
-1.76772714e-01 2.71474831e-02 -4.08189178e-01 8.03421199e-01
4.20662731e-01 6.35487258e-01 4.81577843e-01 -5.62543333e-01
5.40665507e-01 6.08331442e-01 -2.36303374e-01 -2.69698530e-01
-7.08873987e-01 4.30581979e-02 -9.15311515e-01 5.96329212e-01
8.96615326e-01 1.73693985e-01 -1.58195257e+00 1.24877870e+00
1.28986463e-01 3.39521058e-02 -2.06487283e-01 1.50821424e+00
1.43267059e+00 6.60487190e-02 -3.12181599e-02 4.61115956e-01
1.08738339e+00 -1.11596251e+00 -3.45561147e-01 4.62382942e-01
2.65266776e-01 -8.21464598e-01 8.77177477e-01 3.25464457e-01
-9.65868831e-01 -5.09767056e-01 -9.18845057e-01 -1.81065515e-01
-2.15766147e-01 5.06823838e-01 4.67047781e-01 7.81931996e-01
-1.02142799e+00 6.56845927e-01 -1.23569655e+00 -5.61116278e-01
7.34707355e-01 5.70479453e-01 -2.10131034e-01 -4.23536003e-02
-6.12155974e-01 8.67013812e-01 1.02457762e-01 1.85101926e-01
-8.43572319e-01 -3.06610763e-01 -3.89101416e-01 -2.10163325e-01
-1.78461112e-02 -5.59520841e-01 1.58450294e+00 -2.69800842e-01
-1.87668717e+00 1.19276559e+00 -4.59075607e-02 -2.67944574e-01
8.79905164e-01 -4.53231961e-01 1.22266077e-01 3.49911094e-01
-6.64994061e-01 8.15324545e-01 8.99410307e-01 -8.51596653e-01
-3.88176531e-01 -8.24413121e-01 -5.77369966e-02 1.08618312e-01
-1.98421013e-02 1.69590175e-01 -6.83236122e-01 -8.18402946e-01
4.79855329e-01 -1.12459421e+00 3.04518119e-02 3.09947044e-01
-2.96032667e-01 -7.70388246e-01 9.97419536e-01 -7.08396733e-01
8.37770581e-01 -1.70097923e+00 3.38357598e-01 3.03808212e-01
2.98354894e-01 7.56318390e-01 -4.26971138e-01 3.31767857e-01
1.37229085e-01 -1.24300912e-01 -1.38713941e-01 -5.26219726e-01
2.86540955e-01 3.99837077e-01 -1.97488070e-01 2.25903973e-01
1.87514752e-01 1.41532743e+00 -4.33784574e-01 -9.93750542e-02
4.01226193e-01 7.47698009e-01 -2.42772013e-01 4.30898041e-01
-8.71535763e-03 4.37396228e-01 -5.56944370e-01 1.30738163e+00
5.88306546e-01 -1.89588279e-01 1.22915484e-01 -7.24400207e-03
1.22126676e-01 1.47693351e-01 -9.12736714e-01 2.05457306e+00
1.03352234e-01 5.60997665e-01 -2.13304251e-01 -7.88964450e-01
9.88347530e-01 3.00897211e-01 6.40325665e-01 -5.08925557e-01
3.60787213e-01 4.16507304e-01 1.16013840e-01 -7.47632086e-01
2.31561363e-02 1.44665733e-01 4.20550048e-01 9.38774288e-01
2.65803039e-01 -8.46707895e-02 -1.74346521e-01 -3.62906337e-01
1.14031553e+00 4.30843025e-01 2.55192574e-02 7.19912127e-02
2.99969882e-01 -2.18386665e-01 -1.13061048e-01 8.99644792e-01
-5.59779704e-01 9.77014303e-01 2.26800129e-01 -1.20573723e+00
-1.11116064e+00 -1.07152426e+00 1.23818584e-01 9.57762241e-01
-2.49391869e-01 2.29793005e-02 -6.81832671e-01 -8.59182000e-01
1.75211877e-01 -7.40448460e-02 -6.97252870e-01 4.01349366e-01
-1.18021858e+00 -5.48021257e-01 1.22542286e+00 1.23287201e+00
8.60790312e-01 -1.42072022e+00 -1.00967085e+00 4.71469108e-03
2.90723383e-01 -1.10053372e+00 -5.16333401e-01 -5.11652604e-02
-1.14826417e+00 -1.40023577e+00 -1.49788177e+00 -1.06491923e+00
2.53580272e-01 -8.28036368e-02 8.18108559e-01 3.51002216e-01
-6.04775667e-01 6.27008617e-01 -6.79141700e-01 -4.59719688e-01
1.02900714e-01 5.15169084e-01 -1.04440063e-01 -3.58402669e-01
9.42645133e-01 -4.47166443e-01 -6.78391099e-01 1.68039709e-01
-8.71353567e-01 -4.25938576e-01 6.39756978e-01 1.02401412e+00
3.88475120e-01 -1.08359134e+00 2.24335417e-01 2.17896625e-01
6.27002716e-01 3.44914228e-01 -3.67151409e-01 3.58371139e-01
-3.74576986e-01 5.12457974e-02 -2.16039240e-01 -6.60072625e-01
-7.87421703e-01 6.67304337e-01 -7.46424973e-01 -8.68185580e-01
-5.25231540e-01 7.87577853e-02 1.55072063e-01 -2.97451317e-01
3.87353241e-01 5.46304584e-01 3.18190932e-01 -1.00309026e+00
2.57534057e-01 7.82858074e-01 4.00246650e-01 -6.90622985e-01
5.26270032e-01 5.36821067e-01 2.78444469e-01 -8.73123050e-01
-2.56347001e-01 -4.75331426e-01 -1.52219987e+00 -3.31166089e-01
9.19543743e-01 -4.37046617e-01 -1.20007551e+00 1.33391047e+00
-1.43286049e+00 -8.07188332e-01 -6.75296038e-02 5.19207776e-01
-7.19591677e-01 4.92874146e-01 -8.90411198e-01 -7.88224518e-01
-7.56315947e-01 -1.21086168e+00 1.52086854e+00 -7.34768286e-02
-2.78187424e-01 -7.02138424e-01 1.05133869e-01 -3.20361555e-02
3.04984957e-01 2.89646387e-01 7.13888943e-01 -8.45032930e-01
-6.96844578e-01 -2.30423659e-01 -4.09830064e-01 1.89466015e-01
-5.86234853e-02 -2.35352471e-01 -9.70666528e-01 -2.79277027e-01
-5.72797775e-01 -5.50341964e-01 1.12462997e+00 3.65107775e-01
1.30925632e+00 1.38884410e-01 -8.59830156e-02 5.45920670e-01
9.67444241e-01 4.02016938e-01 7.29608893e-01 2.87945092e-01
6.69934630e-01 2.41422027e-01 1.83313824e-02 3.99854064e-01
1.26178041e-01 8.35863709e-01 5.19252121e-01 9.06020328e-02
-5.94375432e-01 -4.93137315e-02 -1.09491989e-01 5.84636390e-01
-1.03391874e+00 -1.08683482e-01 -1.20475435e+00 4.30864245e-01
-1.80643451e+00 -8.69849384e-01 -1.13326609e-01 1.61962640e+00
5.15771866e-01 -3.16589117e-01 5.80784440e-01 4.14574414e-01
2.98644960e-01 2.16160178e-01 -5.77705920e-01 -6.15044981e-02
-1.25893220e-01 1.19803548e+00 2.47528121e-01 2.41148993e-01
-1.47089207e+00 9.70822573e-01 6.89445925e+00 5.08997500e-01
-1.15052903e+00 -3.16313319e-02 -1.06136072e-02 -3.31120938e-01
5.53554237e-01 -9.60569084e-01 -6.65500879e-01 -1.55287415e-01
2.27319628e-01 6.85383677e-01 3.63612443e-01 6.63258612e-01
-2.91233987e-01 4.33719069e-01 -1.13258326e+00 1.29008853e+00
2.66016573e-01 -1.40113795e+00 2.95173049e-01 1.40140519e-01
4.89862800e-01 3.55567187e-01 1.42042560e-03 5.36559075e-02
1.11590452e-01 -1.41232407e+00 7.96851933e-01 4.15984988e-01
9.39660072e-01 -4.76058453e-01 7.80249894e-01 2.55994231e-01
-1.31354117e+00 8.81119519e-02 -1.04754053e-01 -2.39520192e-01
1.02499612e-01 -2.02961788e-01 -8.59887540e-01 2.96158403e-01
9.49101448e-01 7.30173826e-01 -1.72821298e-01 1.03993285e+00
-4.53120768e-01 5.98854303e-01 -4.16211694e-01 -5.32001972e-01
5.16747594e-01 3.39284658e-01 5.76021492e-01 1.36733127e+00
1.96268260e-01 6.98484004e-01 2.92535909e-02 6.19280815e-01
-1.73107132e-01 -2.07234710e-01 -3.82930577e-01 -3.81278470e-02
7.00033009e-02 5.72242677e-01 -6.84971690e-01 -3.45599443e-01
-3.24063897e-01 1.21072221e+00 6.33689389e-02 4.90276963e-01
-1.88014567e-01 -2.75106043e-01 8.30684662e-01 -5.44207633e-01
7.22091377e-01 -7.24162102e-01 -6.05105400e-01 -1.07021403e+00
1.92750230e-01 -5.78593373e-01 5.04223466e-01 -4.55096126e-01
-1.54111981e+00 4.01674449e-01 -2.62318134e-01 -1.16448414e+00
-4.18245643e-01 -1.39479077e+00 -4.33630526e-01 7.82318473e-01
-1.47637427e+00 -1.43668365e+00 -4.93839681e-01 7.82125652e-01
5.84109485e-01 -5.53107262e-01 1.17936933e+00 3.01671207e-01
-8.29952117e-03 7.86721945e-01 -2.40492504e-02 1.00102270e+00
3.71142447e-01 -1.11710739e+00 9.62642133e-01 2.00337395e-01
2.67098218e-01 4.85503286e-01 -3.31998229e-01 -4.78906244e-01
-1.66477871e+00 -6.90398097e-01 9.72448707e-01 -6.48509979e-01
3.08710933e-01 2.40549110e-02 -6.06067300e-01 7.46710360e-01
5.29875048e-02 9.31242257e-02 6.21915817e-01 -1.39457494e-01
-6.65888071e-01 6.11175656e-01 -1.23659670e+00 3.69881421e-01
1.79076600e+00 -4.21201050e-01 -9.13329184e-01 -2.54962202e-02
1.41909942e-01 -9.80176926e-01 -8.78333509e-01 6.72578156e-01
1.57733107e+00 -6.08589113e-01 1.31710494e+00 -1.21339381e+00
6.12584412e-01 2.13106535e-02 -4.08235282e-01 -7.60321200e-01
-5.14522456e-02 -1.69153884e-02 -6.22548699e-01 2.35992402e-01
-1.70802534e-01 -9.82751623e-02 1.20643020e+00 4.48871076e-01
5.07246815e-02 -9.39038992e-01 -1.29090405e+00 -1.07131493e+00
3.66047323e-01 -5.95377445e-01 8.33383024e-01 7.18464434e-01
-3.99668902e-01 -1.24684669e-01 -1.52999416e-01 -2.48190016e-01
7.24136770e-01 3.69553924e-01 8.88029993e-01 -1.16194379e+00
2.85146083e-03 -1.05685127e+00 -7.21986651e-01 -1.49725759e+00
8.65313709e-02 -1.03738284e+00 -2.02781305e-01 -1.57813859e+00
2.21898004e-01 -5.21165095e-02 -1.86977461e-01 9.17417228e-01
2.87596554e-01 6.91485763e-01 6.59509182e-01 4.23301578e-01
-1.78473517e-01 3.89690042e-01 1.54498875e+00 -5.42085886e-01
-2.10247114e-01 2.55195975e-01 5.25293648e-01 4.49657917e-01
6.32778823e-01 3.92191410e-02 2.05006242e-01 -7.83158779e-01
-6.20364964e-01 -1.77297831e-01 7.89189756e-01 -6.71072960e-01
1.09506302e-01 1.34058163e-01 7.61464179e-01 -1.27993810e+00
3.90635878e-01 -6.29385948e-01 -3.25485766e-01 1.01750267e+00
-3.28542233e-01 -2.10918456e-01 1.70233354e-01 1.46955550e-01
-5.89210764e-02 -3.79881673e-02 3.52885604e-01 -2.84192357e-02
-8.10085535e-01 6.19651556e-01 -3.97372395e-01 -4.14148092e-01
5.40752411e-01 -2.72514671e-01 2.51901541e-02 -1.64525688e-01
-1.18165874e+00 -1.30759582e-01 -1.66710675e-01 7.68527508e-01
8.84077132e-01 -1.69117093e+00 -7.47030258e-01 4.04035002e-01
4.66362759e-02 -4.34250310e-02 -6.90089241e-02 4.99586493e-01
-7.80611336e-01 8.26398909e-01 -5.48972845e-01 -1.00600731e+00
-1.53904092e+00 -1.70166060e-01 3.82098377e-01 -2.09531695e-01
-9.68008816e-01 1.05297351e+00 -5.04146159e-01 -7.56370187e-01
8.77358854e-01 -5.70922852e-01 -6.98923618e-02 -7.37609342e-02
7.58760571e-01 5.77914596e-01 3.20989132e-01 -4.92607176e-01
-5.46811342e-01 8.12771201e-01 2.05805168e-01 -1.66843086e-01
1.66918182e+00 7.32084334e-01 1.82859629e-01 1.32275596e-01
1.28579950e+00 -9.08451319e-01 -1.40752017e+00 -3.88429552e-01
2.93990187e-02 -5.83970785e-01 -2.99240410e-01 -1.19659913e+00
-1.24639928e+00 1.42347217e+00 1.10767829e+00 -3.27353776e-01
9.21298087e-01 1.41186818e-01 1.08892858e+00 9.74884450e-01
3.86587441e-01 -8.24527800e-01 2.62093335e-01 1.26110172e+00
1.18110776e+00 -1.03777063e+00 -3.21848541e-01 -3.83482710e-03
-1.57051191e-01 1.79449284e+00 3.63918871e-01 -3.62584740e-01
8.79013538e-01 2.74172008e-01 2.51314789e-01 -3.26518446e-01
-1.04632117e-01 -6.45027697e-01 6.79121196e-01 7.46613264e-01
3.97416055e-01 2.48690873e-01 -5.13494194e-01 6.98532522e-01
-7.15078861e-02 3.04502457e-01 -2.19511926e-01 1.22644389e+00
-1.77791223e-01 -1.29443491e+00 -2.88531184e-01 1.59168482e-01
-2.44289830e-01 1.58082560e-01 -1.15025926e+00 1.03223550e+00
-2.86214441e-01 2.14544967e-01 6.60398453e-02 -6.77779317e-01
5.22545278e-01 7.50485778e-01 1.26090062e+00 -2.32388675e-01
-6.55258298e-01 9.47283581e-02 -2.74318725e-01 -6.10592961e-01
-8.31310928e-01 -6.98545873e-01 -1.40982985e+00 -2.51072556e-01
1.50171667e-01 -8.08221698e-01 8.59153926e-01 1.11675310e+00
8.86954740e-02 1.87145814e-01 9.47477743e-02 -1.38065588e+00
-1.05245423e+00 -1.19461119e+00 -4.87865359e-01 3.27783585e-01
2.53552884e-01 -8.94963026e-01 3.29880923e-01 3.43607925e-02] | [6.652824878692627, -0.4239647388458252] |
bccce757-a7fb-4055-bf7f-7fc53bf6150e | learning-an-adaptation-function-to-assess | 2206.01417 | null | https://arxiv.org/abs/2206.01417v1 | https://arxiv.org/pdf/2206.01417v1.pdf | Learning an Adaptation Function to Assess Image Visual Similarities | Human perception is routinely assessing the similarity between images, both for decision making and creative thinking. But the underlying cognitive process is not really well understood yet, hence difficult to be mimicked by computer vision systems. State-of-the-art approaches using deep architectures are often based on the comparison of images described as feature vectors learned for image categorization task. As a consequence, such features are powerful to compare semantically related images but not really efficient to compare images visually similar but semantically unrelated. Inspired by previous works on neural features adaptation to psycho-cognitive representations, we focus here on the specific task of learning visual image similarities when analogy matters. We propose to compare different supervised, semi-supervised and self-supervised networks, pre-trained on distinct scales and contents datasets (such as ImageNet-21k, ImageNet-1K or VGGFace2) to conclude which model may be the best to approximate the visual cortex and learn only an adaptation function corresponding to the approximation of the the primate IT cortex through the metric learning framework. Our experiments conducted on the Totally Looks Like image dataset highlight the interest of our method, by increasing the retrieval scores of the best model @1 by 2.25x. This research work was recently accepted for publication at the ICIP 2021 international conference [1]. In this new article, we expand on this previous work by using and comparing new pre-trained feature extractors on other datasets. | ['Nicolas Lomenie', 'Camille Kurtz', 'Hala Djeghim', 'Amine Marzouki', 'Olivier Risser-Maroix'] | 2022-06-03 | null | null | null | null | ['image-similarity-search', 'image-categorization'] | ['computer-vision', 'computer-vision'] | [ 2.02099428e-01 -6.45659268e-02 3.20467591e-01 -5.54297447e-01
-5.94758727e-02 -4.96699870e-01 8.76246512e-01 3.74138981e-01
-7.84340978e-01 3.25681746e-01 9.35319960e-02 -5.97288162e-02
-4.69613671e-01 -6.53134108e-01 -6.30717814e-01 -3.62911433e-01
-1.08002990e-01 3.48377705e-01 2.53939688e-01 -2.96451092e-01
6.50016546e-01 6.46188319e-01 -2.06313682e+00 7.09978402e-01
5.51989794e-01 1.22430205e+00 4.70238686e-01 6.05555236e-01
-2.10821658e-01 7.74623632e-01 -3.83052289e-01 -3.16260755e-01
4.42904592e-01 -5.88306785e-01 -1.12007093e+00 -1.06905408e-01
8.13389897e-01 6.07483461e-02 -6.72435760e-02 1.31888032e+00
3.07981700e-01 1.60542831e-01 9.66112852e-01 -1.19612694e+00
-1.08607626e+00 4.38528448e-01 -1.77840471e-01 5.26253939e-01
2.80094087e-01 -5.94976880e-02 8.71399522e-01 -9.38170969e-01
5.34555674e-01 1.26232636e+00 4.79603618e-01 4.26506251e-01
-1.20792890e+00 -5.22173405e-01 -1.30328134e-01 1.03176975e+00
-1.50578928e+00 -2.83024877e-01 7.41021812e-01 -4.82949615e-01
9.86846685e-01 2.45449007e-01 5.73175609e-01 9.26712453e-01
1.63372040e-01 4.60758567e-01 1.76065445e+00 -6.28939629e-01
3.43536735e-01 3.23548734e-01 -1.16058826e-01 5.22783101e-01
-6.45380989e-02 1.70177519e-01 -2.46656865e-01 2.79038757e-01
7.00590909e-01 2.01438740e-01 -2.25834668e-01 -4.04159695e-01
-1.31288743e+00 8.47286761e-01 1.06243837e+00 9.45045888e-01
-4.65877891e-01 -8.32276493e-02 2.94592261e-01 6.48329496e-01
6.35273457e-02 8.02852452e-01 -2.65388221e-01 2.97374368e-01
-1.04589283e+00 1.60221547e-01 5.77128649e-01 7.06676781e-01
9.02869523e-01 -2.41815463e-01 -7.79698044e-02 8.64102602e-01
1.29297674e-01 2.99028814e-01 9.39913094e-01 -8.44580889e-01
-1.39685690e-01 6.53684318e-01 -2.93158382e-01 -1.19156134e+00
-5.85391641e-01 -3.32820445e-01 -8.56104612e-01 5.49727917e-01
6.48095787e-01 4.96256948e-01 -8.40547144e-01 1.50538421e+00
-2.35013485e-01 5.89936115e-02 1.12399563e-01 1.02536500e+00
8.79341543e-01 3.42088908e-01 9.28137377e-02 1.28108561e-01
1.41776490e+00 -8.90602529e-01 -2.10664511e-01 -1.55015528e-01
3.12994093e-01 -9.71608698e-01 1.19333363e+00 4.88748133e-01
-9.49194193e-01 -1.14824092e+00 -1.15063488e+00 8.82162526e-03
-9.38279688e-01 6.84934556e-02 3.53546560e-01 3.84687841e-01
-1.38340247e+00 8.78440201e-01 -2.67596364e-01 -9.10552323e-01
4.80792046e-01 1.25250772e-01 -5.18691003e-01 -8.70045200e-02
-1.00798464e+00 1.33761454e+00 6.52004004e-01 -1.89838409e-01
-6.92193568e-01 -4.21354830e-01 -4.50816423e-01 1.88227385e-01
-1.29999220e-02 -5.18015981e-01 9.38029587e-01 -1.60629439e+00
-1.21003556e+00 1.33024120e+00 2.08896667e-01 -7.25885749e-01
2.59155482e-01 1.34194210e-01 -5.18915594e-01 3.36424738e-01
4.29760572e-03 1.05436003e+00 8.93463373e-01 -1.32829714e+00
-4.77908552e-01 -3.26828390e-01 1.91314220e-01 -1.03502367e-02
-5.06654263e-01 2.71678846e-02 5.21965288e-02 -5.58528185e-01
6.33675158e-02 -8.66884053e-01 -8.25235769e-02 2.51861662e-01
-2.03373954e-02 -2.44087636e-01 4.21377599e-01 -5.31030178e-01
6.30179167e-01 -2.05873966e+00 1.02553129e-01 2.29357243e-01
-1.85993910e-02 3.80463958e-01 -4.23188895e-01 2.80631185e-01
-3.84100109e-01 5.22513799e-02 -2.45722905e-01 4.98707891e-02
8.11235607e-02 1.68216437e-01 -1.73550308e-01 3.84809643e-01
1.75252587e-01 8.35897982e-01 -8.21698904e-01 -5.18230855e-01
3.96670282e-01 2.75122613e-01 -3.12853277e-01 2.58142889e-01
1.60183311e-01 2.92252421e-01 -9.21132267e-02 4.17156637e-01
7.18567431e-01 -2.17469595e-02 -9.59149078e-02 -5.04050374e-01
-1.66024402e-01 -1.53387621e-01 -9.12667572e-01 1.87092340e+00
-5.13698876e-01 7.95821309e-01 -5.17001987e-01 -1.58203781e+00
1.19757116e+00 2.01446041e-02 2.22828075e-01 -1.35891843e+00
2.87282258e-01 2.16600835e-01 3.35332394e-01 -5.44409871e-01
1.71457350e-01 -1.52954638e-01 2.82002658e-01 3.43063116e-01
4.89618152e-01 -2.27885246e-01 2.34630480e-01 -1.19173340e-01
7.32408345e-01 -9.47429836e-02 5.37736773e-01 -6.48235619e-01
9.19306040e-01 -1.25830010e-01 -6.83619305e-02 7.88223624e-01
-3.35343361e-01 8.40615273e-01 1.16600908e-01 -6.88814998e-01
-1.24237871e+00 -1.34847724e+00 -3.93646240e-01 1.02058625e+00
7.65336230e-02 -1.90236509e-01 -8.59596014e-01 -5.39521635e-01
-9.56646577e-02 6.46985292e-01 -8.94748807e-01 -1.91677049e-01
-2.20925093e-01 -3.78816724e-01 4.30559099e-01 4.18242723e-01
6.94749653e-01 -1.40816069e+00 -8.41266513e-01 2.29416359e-02
2.04929203e-01 -1.04873025e+00 6.34555817e-02 9.35321897e-02
-7.90010750e-01 -1.17230272e+00 -8.22279215e-01 -9.58846748e-01
5.54982603e-01 2.06337407e-01 1.29313576e+00 1.81873769e-01
-6.02161169e-01 5.38270116e-01 -5.11261940e-01 -5.45942724e-01
-4.78311747e-01 -1.87227562e-01 4.64806072e-02 1.32159099e-01
5.61061740e-01 -6.80136263e-01 -8.13627779e-01 3.25373888e-01
-1.05856395e+00 -6.50101900e-02 8.34988832e-01 7.24146843e-01
5.21580338e-01 -1.43513978e-01 3.37108493e-01 -4.66716230e-01
5.04515409e-01 -2.59044319e-01 -1.79642498e-01 4.53284264e-01
-6.01577997e-01 2.94756621e-01 8.19294095e-01 -4.34146464e-01
-7.95711994e-01 6.89323917e-02 2.02140361e-02 -4.76511270e-01
-5.56586742e-01 2.57298291e-01 1.04051523e-01 -5.05684733e-01
9.80771780e-01 2.66812801e-01 9.32030380e-03 -3.74959588e-01
4.88591343e-01 5.47846794e-01 6.58152342e-01 -3.52267891e-01
7.69727290e-01 6.40468717e-01 8.14443380e-02 -8.31382692e-01
-6.60601735e-01 -3.58848631e-01 -1.06516850e+00 -2.91756481e-01
1.05796194e+00 -4.36601758e-01 -6.34485722e-01 2.41701812e-01
-1.11857080e+00 6.52868748e-02 -3.48369241e-01 6.72279656e-01
-7.80488789e-01 4.13070798e-01 -8.22694227e-02 -3.53146225e-01
-2.27097645e-01 -9.27948415e-01 5.90338945e-01 3.09055537e-01
-1.55165598e-01 -9.73208308e-01 2.13087887e-01 2.05610737e-01
8.15071106e-01 9.71768424e-02 1.01215422e+00 -8.89949143e-01
-2.80138820e-01 4.76398617e-02 -6.40899241e-01 6.42860234e-01
7.34580755e-02 -2.06186101e-01 -1.24026680e+00 -1.70718372e-01
8.99101123e-02 -5.03415525e-01 1.12279570e+00 1.63915917e-01
1.44512558e+00 -4.89971507e-03 3.01446393e-02 5.62694132e-01
1.61300576e+00 1.00908674e-01 8.41037512e-01 6.16057396e-01
3.38601857e-01 9.14354324e-01 3.91849577e-01 1.07152335e-01
1.79625884e-01 6.15078926e-01 5.33456266e-01 -5.94782047e-02
-3.44916701e-01 1.14949957e-01 6.41643107e-02 7.29333580e-01
-3.56612891e-01 2.16493055e-01 -8.93616140e-01 4.89860266e-01
-1.58261704e+00 -1.07416725e+00 8.57665762e-02 2.24910402e+00
5.59005678e-01 1.69533864e-01 -4.54356149e-02 1.60692245e-01
5.02282381e-01 -9.48029384e-02 -2.05527201e-01 -7.17519879e-01
-2.55032599e-01 6.08295321e-01 2.37986937e-01 4.45532240e-02
-1.18223631e+00 6.91551507e-01 5.60833311e+00 7.43454874e-01
-1.39522767e+00 7.49629363e-02 5.21531224e-01 8.77862126e-02
2.24167649e-02 8.44285414e-02 -1.01649843e-01 2.98982054e-01
1.02448845e+00 -2.42895201e-01 5.62212288e-01 8.78295779e-01
-1.82604924e-01 -7.52487332e-02 -1.39988542e+00 1.31551683e+00
5.67014158e-01 -1.26467013e+00 3.58933836e-01 -2.26908028e-01
5.65331280e-01 5.65455630e-02 1.85794190e-01 2.88739830e-01
-2.95493543e-01 -1.44202852e+00 6.53213382e-01 1.00013864e+00
3.49829316e-01 -6.23237312e-01 7.37281680e-01 3.49423140e-02
-1.10114610e+00 -2.58507013e-01 -8.61041248e-01 -1.82433397e-01
-3.37537766e-01 9.62057114e-02 -7.83820570e-01 4.27838922e-01
1.12834215e+00 7.43886709e-01 -1.30500329e+00 1.29282963e+00
2.31775269e-02 1.75884515e-01 2.20813081e-02 -1.71963662e-01
4.07924294e-01 -1.36038065e-02 1.99348316e-01 1.29562032e+00
4.06393737e-01 -1.37131423e-01 -9.03511345e-02 1.07801652e+00
8.51121470e-02 4.29096192e-01 -7.11062670e-01 1.55052006e-01
8.15715790e-02 1.47209883e+00 -9.81940567e-01 -4.01857585e-01
-5.13420701e-01 1.10957015e+00 4.16922063e-01 1.63438484e-01
-5.58450162e-01 -4.63460386e-01 3.99582446e-01 7.34403729e-02
3.36176962e-01 -8.27293750e-03 -2.54070871e-02 -8.55176926e-01
-9.65572298e-02 -7.46926129e-01 2.93636799e-01 -1.02763128e+00
-1.72529745e+00 1.02907228e+00 1.41647771e-01 -1.34961653e+00
-2.42705002e-01 -1.11570382e+00 -7.16274083e-01 8.66730690e-01
-1.50202394e+00 -1.08198118e+00 -5.49338639e-01 7.87242651e-01
3.71943176e-01 -4.31544453e-01 1.03301263e+00 9.65856090e-02
1.82427838e-01 5.33038378e-01 2.01739267e-01 8.92826095e-02
8.75321686e-01 -1.34000444e+00 1.73881918e-01 4.57131535e-01
6.63529634e-01 5.69160938e-01 5.76499820e-01 1.82225928e-01
-8.92733037e-01 -7.03719020e-01 7.95899630e-01 -3.82273883e-01
6.61544383e-01 -2.20305264e-01 -1.24961793e+00 2.77643353e-02
5.94161153e-01 1.57218769e-01 6.20484352e-01 -3.86167198e-01
-6.23705029e-01 -3.78767014e-01 -1.19843304e+00 4.37343359e-01
9.33654785e-01 -7.88078308e-01 -1.04282761e+00 2.32070982e-01
2.42817342e-01 2.49320820e-01 -9.46105957e-01 3.00157070e-01
5.97084224e-01 -1.31290877e+00 1.29763532e+00 -6.99730992e-01
4.62508351e-01 -2.36373693e-01 -4.89905626e-01 -1.30416894e+00
-4.87420410e-01 2.14658871e-01 4.79972214e-01 9.62951541e-01
7.16886371e-02 -4.45537090e-01 4.20446515e-01 1.02537833e-01
1.24446422e-01 -5.21480322e-01 -9.90989029e-01 -9.43302810e-01
2.83108562e-01 -2.00104699e-01 3.06873322e-01 1.04420149e+00
-2.79420167e-01 2.09076762e-01 6.01766109e-02 -1.00286365e-01
6.91213310e-01 3.14367115e-01 4.44521487e-01 -1.50898266e+00
-1.93622768e-01 -9.94650960e-01 -1.09251130e+00 -2.04232544e-01
2.64162421e-01 -1.20116568e+00 -1.70716688e-01 -1.31080341e+00
2.79373705e-01 -6.49213940e-02 -7.32687116e-01 2.40604460e-01
2.00615972e-01 6.33456886e-01 5.51290095e-01 1.91597506e-01
-5.00512362e-01 2.95179188e-01 1.18580091e+00 -3.88002992e-01
2.73374379e-01 -2.36976087e-01 -4.71214175e-01 7.98068762e-01
8.94306600e-01 -2.80905873e-01 -4.59688604e-01 -2.02171534e-01
1.47581296e-02 -5.77285528e-01 8.06306839e-01 -1.50282741e+00
2.80374140e-01 -4.79298411e-03 8.11473489e-01 -1.30242765e-01
1.91236839e-01 -9.44203496e-01 -7.60515928e-02 5.77126265e-01
-4.83067453e-01 2.99094588e-01 2.59096086e-01 3.28120291e-01
-4.37572002e-01 -6.17397010e-01 9.38282847e-01 -4.79405880e-01
-1.35872078e+00 5.76536655e-02 -2.88216442e-01 4.58382778e-02
9.84174490e-01 -3.32743376e-01 -2.42819101e-01 -2.78890312e-01
-7.14385331e-01 -2.50148088e-01 4.32619780e-01 6.50512278e-01
8.51819456e-01 -1.30392575e+00 -8.03849339e-01 2.21965183e-02
5.66206813e-01 -7.86918402e-01 2.72330850e-01 5.73783576e-01
-6.27032936e-01 4.18471396e-01 -1.15648949e+00 -6.60091162e-01
-1.27758574e+00 1.04414225e+00 3.68635118e-01 1.83288679e-01
-2.25205526e-01 7.38191307e-01 2.15266466e-01 -3.99180293e-01
6.73673004e-02 -3.87395501e-01 -5.85378885e-01 1.14564024e-01
6.37213647e-01 1.25184834e-01 1.91730440e-01 -7.96339512e-01
-5.15058279e-01 9.13079798e-01 -9.23780575e-02 7.59313256e-02
1.24337244e+00 4.59342152e-02 -3.66768479e-01 4.88764614e-01
1.60561073e+00 -5.69979668e-01 -8.04948509e-01 -4.29528564e-01
2.75140345e-01 -4.19109464e-01 -4.13936563e-02 -7.90260732e-01
-9.64822054e-01 1.21026027e+00 1.30153227e+00 3.37853968e-01
1.32528579e+00 9.52697843e-02 1.23546124e-01 6.81590080e-01
3.50161046e-01 -1.13719320e+00 4.52308327e-01 3.97455156e-01
1.29233778e+00 -1.42267394e+00 3.44819874e-02 1.37595341e-01
-5.97152531e-01 1.47802448e+00 6.19052649e-01 -5.51696897e-01
7.05855072e-01 -2.65738010e-01 7.18956143e-02 -1.32222712e-01
-4.00099903e-01 -5.29279172e-01 7.76765347e-01 7.17523694e-01
4.82963562e-01 -1.11530460e-02 -3.50455880e-01 3.85034680e-01
-2.94636250e-01 -1.78378880e-01 2.82690823e-01 6.96465075e-01
-4.96093392e-01 -1.04263198e+00 -3.77441645e-01 4.11432475e-01
-1.44188657e-01 -1.92862600e-01 -5.93539476e-01 8.40950906e-01
4.29552197e-01 7.21433878e-01 3.69448334e-01 -4.67640072e-01
4.14209247e-01 5.34812585e-02 8.42930794e-01 -3.11591089e-01
-6.83985591e-01 -6.23656034e-01 -4.41354990e-01 -6.50743783e-01
-7.98329771e-01 -5.12644589e-01 -9.58927572e-01 -2.07218781e-01
1.65029913e-01 7.09347846e-03 8.11830759e-01 8.82151842e-01
1.10787362e-01 4.24340844e-01 5.34020185e-01 -1.16036665e+00
-5.20602405e-01 -1.12477160e+00 -4.24581915e-01 9.99625921e-01
-2.05184147e-02 -7.96036780e-01 -2.28777006e-01 1.25710875e-01] | [9.821967124938965, 2.204354763031006] |
21139517-6f7a-46bb-a78f-9972f6b53b04 | lexmae-lexicon-bottlenecked-pretraining-for | 2208.14754 | null | https://arxiv.org/abs/2208.14754v2 | https://arxiv.org/pdf/2208.14754v2.pdf | LexMAE: Lexicon-Bottlenecked Pretraining for Large-Scale Retrieval | In large-scale retrieval, the lexicon-weighting paradigm, learning weighted sparse representations in vocabulary space, has shown promising results with high quality and low latency. Despite it deeply exploiting the lexicon-representing capability of pre-trained language models, a crucial gap remains between language modeling and lexicon-weighting retrieval -- the former preferring certain or low-entropy words whereas the latter favoring pivot or high-entropy words -- becoming the main barrier to lexicon-weighting performance for large-scale retrieval. To bridge this gap, we propose a brand-new pre-training framework, lexicon-bottlenecked masked autoencoder (LexMAE), to learn importance-aware lexicon representations. Essentially, we present a lexicon-bottlenecked module between a normal language modeling encoder and a weakened decoder, where a continuous bag-of-words bottleneck is constructed to learn a lexicon-importance distribution in an unsupervised fashion. The pre-trained LexMAE is readily transferred to the lexicon-weighting retrieval via fine-tuning. On the ad-hoc retrieval benchmark, MS-Marco, it achieves 42.6% MRR@10 with 45.8 QPS for the passage dataset and 44.4% MRR@100 with 134.8 QPS for the document dataset, by a CPU machine. And LexMAE shows state-of-the-art zero-shot transfer capability on BEIR benchmark with 12 datasets. | ['Daxin Jiang', 'Linjun Yang', 'Binxing Jiao', 'Xiaolong Huang', 'Can Xu', 'Chongyang Tao', 'Xiubo Geng', 'Tao Shen'] | 2022-08-31 | null | null | null | null | ['passage-retrieval'] | ['natural-language-processing'] | [ 1.60980895e-01 -2.97488064e-01 -5.59774280e-01 -6.82086647e-02
-1.42279375e+00 -2.89739817e-01 5.87223709e-01 2.60032743e-01
-8.60849500e-01 4.76792872e-01 5.88896632e-01 -1.58247858e-01
-7.02124685e-02 -8.06556225e-01 -4.45945084e-01 -5.72165668e-01
-4.95478213e-02 7.70188630e-01 1.56380475e-01 -7.81341851e-01
2.46454537e-01 1.25284195e-01 -1.62340581e+00 4.50928837e-01
4.12082106e-01 1.05655682e+00 3.36848855e-01 5.55925965e-01
-3.09494644e-01 8.63302827e-01 -3.84528011e-01 -2.85117239e-01
1.60309345e-01 -1.95714384e-01 -5.18058419e-01 -3.92479241e-01
1.37120336e-01 -2.57740259e-01 -7.96625614e-01 7.95560598e-01
8.07965755e-01 3.08659524e-01 8.09095323e-01 -4.73689437e-01
-8.91391933e-01 7.73445308e-01 -5.20473838e-01 3.41158599e-01
2.30451763e-01 -1.61714733e-01 1.61458170e+00 -1.23316920e+00
5.12459576e-01 1.19306159e+00 1.94149211e-01 6.42594159e-01
-1.13782871e+00 -7.30575085e-01 4.78336066e-02 1.73395351e-01
-1.87204862e+00 -5.65220952e-01 4.21959758e-01 -1.60394058e-01
1.59365177e+00 1.35206774e-01 4.75281179e-01 1.00108314e+00
2.52481371e-01 9.40045416e-01 5.97880960e-01 -5.54665089e-01
1.41775995e-01 1.32333025e-01 4.21872348e-01 6.01404250e-01
2.42642209e-01 8.08314905e-02 -7.71186113e-01 -1.57376513e-01
4.24479663e-01 6.79237023e-02 -1.09843522e-01 -1.54683545e-01
-1.01839006e+00 1.14577568e+00 2.67295003e-01 3.72797579e-01
-2.93907046e-01 3.17559153e-01 7.13119268e-01 6.69855177e-01
7.46466637e-01 4.28906649e-01 -5.01013458e-01 -9.41440165e-02
-1.29734921e+00 4.75168154e-02 6.21234894e-01 9.25052047e-01
7.64898002e-01 2.08944961e-01 -4.47864652e-01 1.17769825e+00
4.23375458e-01 7.19413519e-01 1.08596277e+00 -1.63483545e-01
4.05239999e-01 4.51656848e-01 -2.89179534e-01 -7.61791348e-01
-3.20242971e-01 -5.94474971e-01 -9.14319217e-01 -5.04586995e-01
-2.59412855e-01 2.79783815e-01 -1.03544295e+00 1.68116200e+00
-2.30755985e-01 3.61850485e-02 2.59138048e-01 7.99487472e-01
9.22197819e-01 1.06720376e+00 1.17376588e-01 -3.09554517e-01
1.57585025e+00 -1.17130339e+00 -6.62665784e-01 -4.98563111e-01
8.41594338e-01 -7.91001141e-01 1.37904370e+00 1.99096367e-01
-1.25668633e+00 -4.68911320e-01 -1.16587532e+00 -5.38951993e-01
-5.74665964e-01 3.56901735e-02 5.28206110e-01 4.89827663e-01
-1.04497993e+00 2.14745745e-01 -4.73228544e-01 -2.60350466e-01
2.46628791e-01 2.18937442e-01 -2.01589182e-01 -1.83082640e-01
-1.71543193e+00 1.05955565e+00 5.63344598e-01 -3.94529015e-01
-1.04263723e+00 -8.45475376e-01 -1.07178152e+00 5.45099795e-01
3.56481850e-01 -4.83792603e-01 1.05276990e+00 -5.44149995e-01
-1.51520622e+00 1.00479007e+00 -2.16454968e-01 -7.03146756e-01
-1.22391865e-01 -2.87648052e-01 -6.62930787e-01 1.05964921e-01
-2.37977579e-02 7.84872949e-01 9.48694825e-01 -8.98337781e-01
-3.89100075e-01 3.80944759e-02 -2.53332593e-02 4.06672537e-01
-7.10910916e-01 1.62720650e-01 -9.09812987e-01 -1.13851476e+00
-1.90995052e-01 -6.37674451e-01 -1.68097280e-02 -4.48994488e-01
2.40300670e-02 -4.39109176e-01 2.08843932e-01 -5.21526814e-01
1.95688617e+00 -2.17765903e+00 3.11349720e-01 1.62098080e-01
8.69963989e-02 3.99809569e-01 -6.56028509e-01 6.12462461e-01
-4.74983789e-02 -2.60120660e-01 -1.10574715e-01 -5.10790765e-01
1.70760676e-01 1.08037174e-01 -8.00166965e-01 3.32499504e-01
2.39255950e-01 1.21331990e+00 -9.38860416e-01 -4.92811829e-01
1.09483436e-01 6.84585392e-01 -8.31569791e-01 2.13583484e-01
-1.50689930e-01 -3.46126050e-01 -2.16536596e-01 7.49295354e-01
3.85824382e-01 -4.57277834e-01 9.77521166e-02 -4.46746722e-02
4.02297646e-01 5.47683001e-01 -1.03872037e+00 2.21799040e+00
-9.11139846e-01 4.69094098e-01 -3.18825305e-01 -7.76301742e-01
8.65022361e-01 2.74983257e-01 2.31566265e-01 -1.01280057e+00
2.63199210e-01 3.27753961e-01 -2.54216164e-01 -7.32513964e-02
1.05272090e+00 -2.75973022e-01 -5.38336694e-01 6.64673984e-01
6.33553982e-01 -1.70482889e-01 4.04107392e-01 6.22243345e-01
1.00803590e+00 -3.97182614e-01 5.19594967e-01 -3.67460519e-01
6.96459651e-01 -1.91174597e-01 -1.11049758e-02 7.34400988e-01
2.02126786e-01 7.13741004e-01 6.57900721e-02 -1.83565632e-01
-8.29854846e-01 -1.00759542e+00 -1.89047709e-01 1.80012500e+00
8.77711922e-02 -8.47947359e-01 -3.65348577e-01 -3.57803345e-01
3.40377423e-03 8.55777621e-01 -4.45460498e-01 -7.78138161e-01
-4.88258868e-01 -7.62336254e-01 6.11900389e-01 3.91561419e-01
-3.51872817e-02 -1.16742456e+00 -3.57401192e-01 3.83601159e-01
-1.57774925e-01 -7.92838931e-01 -9.14098442e-01 5.14221728e-01
-6.36465669e-01 -5.88130713e-01 -8.75648737e-01 -8.34343791e-01
3.68088633e-01 4.89366859e-01 1.59255147e+00 1.92266881e-01
-3.46793056e-01 2.98374325e-01 -4.65088904e-01 -2.84939855e-01
-1.62850991e-01 4.85173732e-01 1.67289689e-01 -1.53277576e-01
8.05077791e-01 -3.59560460e-01 -6.52204275e-01 1.29347488e-01
-1.12408352e+00 -2.28804737e-01 7.12897003e-01 1.23091364e+00
7.46591389e-01 -2.36667857e-01 6.97823048e-01 -5.50228715e-01
8.77573371e-01 -6.96731031e-01 -5.45180678e-01 3.99420917e-01
-7.74229705e-01 3.01815778e-01 5.16495883e-01 -6.18903637e-01
-5.97407579e-01 -3.91436100e-01 -3.85430157e-01 -3.74242038e-01
5.01068950e-01 6.95215285e-01 8.92722085e-02 1.81558996e-01
6.60839736e-01 7.02318907e-01 -9.94061083e-02 -5.70548892e-01
8.04516137e-01 9.09419775e-01 1.85396731e-01 -5.66511691e-01
6.98729932e-01 2.28132367e-01 -5.31210065e-01 -6.69808626e-01
-9.98772681e-01 -9.67508256e-01 -1.97894394e-01 2.27972209e-01
5.74362993e-01 -1.59918511e+00 -2.61911809e-01 3.01879030e-02
-8.59901428e-01 -1.80018038e-01 -6.69381797e-01 3.36787641e-01
-4.61554706e-01 1.33150101e-01 -7.78291166e-01 -3.93469244e-01
-9.94854748e-01 -1.21054935e+00 1.43016946e+00 -1.26405463e-01
-3.09339762e-01 -7.65887618e-01 2.09896967e-01 1.99622050e-01
9.22231436e-01 -7.68232346e-01 1.06441760e+00 -9.38968778e-01
-3.10309708e-01 -5.75496793e-01 -3.97957832e-01 4.06394035e-01
-1.80056229e-01 -8.12411129e-01 -1.05498517e+00 -7.09100246e-01
-1.44650698e-01 -9.43307459e-01 1.27735603e+00 1.22001171e-01
1.23591876e+00 9.58291739e-02 -5.18367216e-02 4.18545753e-01
1.57444274e+00 2.94574481e-02 5.38826287e-01 4.77971025e-02
4.17984128e-01 3.07189804e-02 4.43494171e-01 3.27894509e-01
4.18950915e-01 8.80120814e-01 -7.37314895e-02 1.69019744e-01
-3.70023400e-01 -4.02098924e-01 5.80340862e-01 1.71623731e+00
3.15936059e-01 -2.84643620e-01 -7.51133442e-01 5.90983808e-01
-1.62826169e+00 -6.96822762e-01 5.23376048e-01 2.15237069e+00
1.33989346e+00 2.27868378e-01 -2.10276395e-01 -1.93938073e-02
2.08883464e-01 5.65253556e-01 -4.02562261e-01 -3.52124959e-01
-1.61936060e-01 7.38031924e-01 4.74009752e-01 8.09724152e-01
-9.48444426e-01 1.36585426e+00 6.09131527e+00 1.82718098e+00
-1.01355791e+00 5.01710594e-01 3.41122776e-01 -5.77953696e-01
-4.60125327e-01 -3.04838955e-01 -1.03920865e+00 3.34477961e-01
1.35549748e+00 -5.34499526e-01 6.61109626e-01 9.17864740e-01
-4.52438504e-01 1.20669484e-01 -9.80211973e-01 1.11493468e+00
4.59304452e-01 -1.33986783e+00 4.43076164e-01 -1.56218842e-01
5.48143208e-01 3.56029987e-01 5.81738204e-02 1.07003307e+00
1.30112439e-01 -1.10638833e+00 6.46198988e-01 3.45853388e-01
1.27600563e+00 -8.15077484e-01 7.57853627e-01 1.91159815e-01
-1.40662348e+00 1.16875336e-01 -8.55723977e-01 8.57559368e-02
2.10655645e-01 6.02491736e-01 -3.18043411e-01 3.36675316e-01
2.37169221e-01 6.25382543e-01 -4.10415232e-01 5.54809093e-01
-1.59669936e-01 5.52900672e-01 -2.29430720e-01 -1.72782078e-01
3.27602744e-01 9.39603373e-02 2.89127916e-01 1.61029696e+00
1.66271493e-01 -3.11621614e-02 1.58220585e-02 3.44041854e-01
-5.48430622e-01 4.13401961e-01 -4.18094486e-01 -1.61939457e-01
4.11388636e-01 1.19258237e+00 -3.35763276e-01 -6.79378510e-01
-4.84812558e-01 9.76628423e-01 3.86729211e-01 2.54398346e-01
-7.97803164e-01 -4.38687921e-01 7.58877635e-01 -2.08730236e-01
3.66612881e-01 -8.24554861e-02 8.37360602e-03 -1.59009552e+00
-2.59354323e-01 -1.12048316e+00 5.10132313e-01 -4.93166566e-01
-1.30556118e+00 7.20973849e-01 -2.83194128e-02 -1.22366130e+00
-1.55092761e-01 -5.50693750e-01 -3.00271641e-02 1.00931227e+00
-2.11498308e+00 -8.61419261e-01 1.43809423e-01 5.68462610e-01
9.96808290e-01 -6.96175635e-01 1.33689201e+00 7.37357974e-01
-2.75924385e-01 9.64684010e-01 3.73402327e-01 -4.98672724e-02
7.49074280e-01 -1.10294294e+00 4.24683988e-01 4.45685416e-01
5.48750877e-01 8.90238345e-01 1.01144679e-01 -2.81752706e-01
-1.77643991e+00 -1.04809928e+00 1.32411730e+00 -3.97672415e-01
9.76598740e-01 -7.07647860e-01 -1.02086353e+00 4.07002568e-01
3.26803774e-01 1.08419858e-01 9.02742088e-01 2.26304635e-01
-7.67432213e-01 -5.97443059e-02 -7.30163038e-01 6.41321898e-01
8.76783371e-01 -1.13102615e+00 -9.61627901e-01 5.00957072e-01
1.10247457e+00 -3.37044328e-01 -8.26741397e-01 2.81201839e-01
5.26922345e-01 -3.01122487e-01 1.44925845e+00 -6.12907648e-01
4.91849154e-01 -2.03984231e-02 -5.16142249e-01 -1.24646533e+00
-3.79006773e-01 -4.12196040e-01 -5.48170269e-01 7.76317537e-01
4.66201663e-01 -2.63885379e-01 5.25572956e-01 -9.71394554e-02
1.21841021e-01 -9.77251410e-01 -1.05188191e+00 -7.09837079e-01
4.25357372e-01 -6.65989637e-01 4.63574618e-01 5.99322796e-01
5.54911755e-02 8.30687106e-01 -5.96201479e-01 -3.12997073e-01
3.09736311e-01 2.64736079e-03 4.25546795e-01 -9.11845028e-01
-3.16197246e-01 -5.66230059e-01 -7.85379857e-02 -1.48445320e+00
3.53595406e-01 -1.39521992e+00 6.06986731e-02 -1.27298224e+00
3.60454798e-01 -1.73327401e-01 -8.21903586e-01 3.96436602e-01
-3.65134448e-01 3.80924433e-01 2.04424873e-01 2.84379005e-01
-1.00644147e+00 9.39441442e-01 8.63855958e-01 -4.53753740e-01
4.35946770e-02 -4.91274685e-01 -8.16878200e-01 3.20194781e-01
5.35349429e-01 -7.14036286e-01 -6.10305488e-01 -5.36985934e-01
4.58621681e-01 -1.08279169e-01 -3.19947660e-01 -8.06946278e-01
3.50816578e-01 3.45061332e-01 -1.43627346e-01 -6.14133418e-01
5.31362414e-01 -7.83894777e-01 -4.30279613e-01 4.60310251e-01
-6.34204745e-01 7.51831988e-03 2.98212081e-01 6.30910099e-01
-5.46141863e-01 -2.79634118e-01 6.51512325e-01 -1.08471051e-01
-8.80184472e-01 2.79994816e-01 -4.65709418e-01 5.08661330e-01
5.24784505e-01 2.71641344e-01 -1.17733017e-01 -4.00643170e-01
-4.77110386e-01 2.73890883e-01 -4.01765518e-02 7.08183289e-01
8.21529746e-01 -1.49918985e+00 -8.93493533e-01 5.00157297e-01
6.62060380e-01 -3.35689366e-01 2.72994190e-01 3.86740655e-01
-1.80118680e-01 8.43505621e-01 2.37496778e-01 -2.83820540e-01
-8.48162830e-01 6.25348032e-01 6.14112578e-02 -8.90117526e-01
-4.74496454e-01 1.13400352e+00 1.15011372e-01 -3.61595482e-01
4.74712878e-01 -2.38566324e-01 -2.26168782e-01 4.92572784e-01
9.27073598e-01 4.31808606e-02 3.92500788e-01 -6.46971285e-01
-3.20714265e-01 6.76754594e-01 -5.01702607e-01 -2.89592315e-02
1.17565823e+00 -8.91536251e-02 -1.66271061e-01 4.55840856e-01
1.56538177e+00 -6.33838028e-02 -5.61177135e-01 -8.28807473e-01
1.42763570e-01 -1.96811095e-01 5.97207844e-01 -8.24239910e-01
-9.17815447e-01 1.05348873e+00 6.00476921e-01 -1.01447158e-01
1.15773582e+00 -5.73066287e-02 1.11987710e+00 7.62958467e-01
6.02409720e-01 -1.10920119e+00 2.54063070e-01 1.04497659e+00
1.02752578e+00 -1.40574884e+00 -5.28184325e-03 1.39642105e-01
-7.64872670e-01 8.76738489e-01 1.61107689e-01 -2.58136958e-01
8.87156069e-01 1.92844138e-01 1.09685231e-02 -2.39830703e-01
-1.08402526e+00 -4.70687509e-01 8.03505242e-01 1.46728074e-02
5.52163661e-01 6.01976253e-02 -4.25438970e-01 8.75524998e-01
-1.72330737e-02 5.50832972e-02 -6.80020079e-02 7.01395512e-01
-5.20612717e-01 -1.05484152e+00 5.75713329e-02 5.91469347e-01
-3.99269581e-01 -9.71497416e-01 1.61685631e-01 4.79688376e-01
-3.51347536e-01 7.86518812e-01 2.24358350e-01 -4.05101359e-01
3.69447976e-01 3.95260721e-01 1.92551881e-01 -9.76784110e-01
-8.07374001e-01 2.94652939e-01 2.37849504e-02 -6.35174572e-01
-1.42869607e-01 -2.61869989e-02 -8.93254280e-01 -1.87560603e-01
-5.66252172e-01 3.54875505e-01 4.28851485e-01 7.43684828e-01
5.04045308e-01 6.43061757e-01 4.75105107e-01 -6.75860286e-01
-8.31387639e-01 -1.14324760e+00 -6.58024609e-01 2.69132584e-01
1.70336574e-01 -4.97352719e-01 -4.35568362e-01 -3.76918942e-01] | [11.458114624023438, 7.767665386199951] |
347caac6-934a-4a8a-b849-b1da859821bc | lessons-learned-in-multilingual-grounded | 1809.07615 | null | http://arxiv.org/abs/1809.07615v1 | http://arxiv.org/pdf/1809.07615v1.pdf | Lessons learned in multilingual grounded language learning | Recent work has shown how to learn better visual-semantic embeddings by
leveraging image descriptions in more than one language. Here, we investigate
in detail which conditions affect the performance of this type of grounded
language learning model. We show that multilingual training improves over
bilingual training, and that low-resource languages benefit from training with
higher-resource languages. We demonstrate that a multilingual model can be
trained equally well on either translations or comparable sentence pairs, and
that annotating the same set of images in multiple language enables further
improvements via an additional caption-caption ranking objective. | ['Grzegorz Chrupała', 'Marc-Alexandre Côté', 'Ákos Kádár', 'Desmond Elliott', 'Afra Alishahi'] | 2018-09-20 | lessons-learned-in-multilingual-grounded-1 | https://aclanthology.org/K18-1039 | https://aclanthology.org/K18-1039.pdf | conll-2018-10 | ['grounded-language-learning'] | ['natural-language-processing'] | [-1.88416958e-01 1.12042181e-01 -4.68705356e-01 -5.69354594e-01
-1.29914284e+00 -7.28531122e-01 9.92282033e-01 2.09410023e-02
-1.01771712e+00 7.81360209e-01 7.12938070e-01 -3.27758133e-01
5.43959022e-01 -3.95445198e-01 -9.61663008e-01 -1.86317056e-01
8.83529037e-02 5.44573188e-01 -8.34317505e-02 -2.24364772e-01
-2.28430942e-01 3.06988060e-01 -9.77492630e-01 5.56062043e-01
6.20997608e-01 3.85848552e-01 6.06408119e-01 4.48414832e-01
-1.70157209e-01 6.93251252e-01 -2.54502475e-01 -7.08806813e-01
3.56450647e-01 -3.34563017e-01 -7.52962470e-01 -3.49651501e-02
8.59207809e-01 -3.03675205e-01 -4.40852433e-01 8.99803460e-01
4.31689203e-01 -1.61349505e-01 6.15794897e-01 -1.19178522e+00
-1.60075128e+00 5.40360212e-01 -3.16113442e-01 2.52300799e-01
4.29809839e-01 1.43156394e-01 1.14477634e+00 -1.32460535e+00
1.19637430e+00 1.43060148e+00 5.62160552e-01 4.53332096e-01
-1.37890935e+00 -5.72183490e-01 3.30518395e-01 8.32736641e-02
-1.65642476e+00 -5.22206843e-01 4.85725403e-01 -4.82296199e-01
1.11354423e+00 -3.18773568e-01 4.43265408e-01 1.36949801e+00
3.44897881e-02 5.29652715e-01 1.29682314e+00 -5.34707308e-01
-4.07128364e-01 5.95584810e-01 -2.43421078e-01 9.00521696e-01
4.08236325e-01 -2.50686752e-03 -6.58263266e-01 1.73670366e-01
6.69285715e-01 -1.95100263e-01 -2.70465791e-01 -6.50907218e-01
-1.50831139e+00 1.02055073e+00 7.45783031e-01 5.48018754e-01
-3.33817042e-02 2.87456483e-01 4.80339110e-01 2.66175836e-01
6.48898005e-01 7.68427908e-01 -3.60932350e-01 1.84569046e-01
-8.89341652e-01 -2.02032149e-01 3.99897248e-01 1.25707734e+00
1.17060554e+00 5.31622171e-02 -1.26728058e-01 8.63714218e-01
2.07831055e-01 8.56602728e-01 4.21485007e-01 -8.06496441e-01
6.15585029e-01 1.99860767e-01 1.28582031e-01 -6.60948634e-01
3.72341760e-02 -2.15213746e-01 -3.59847307e-01 -6.13465868e-02
5.08767590e-02 -1.59838069e-02 -8.55710745e-01 2.05278063e+00
-2.25400716e-01 -6.11299202e-02 4.24920738e-01 1.03641605e+00
6.65721595e-01 6.07679486e-01 4.37753081e-01 1.83928281e-01
1.29362917e+00 -1.27276921e+00 -5.06873846e-01 -5.50425947e-01
7.68033266e-01 -7.90446997e-01 1.43933547e+00 -2.28977472e-01
-1.02380109e+00 -5.95454574e-01 -1.12437284e+00 -6.12496257e-01
-6.00844622e-01 2.48875663e-01 5.59364498e-01 2.72526264e-01
-1.60470533e+00 -8.80886521e-03 -7.20147073e-01 -8.15795124e-01
4.45013970e-01 -1.66153274e-02 -8.52530837e-01 -5.77854395e-01
-1.24503756e+00 1.46548069e+00 2.48437494e-01 -1.33415014e-01
-1.28606594e+00 -6.02070332e-01 -1.30169129e+00 -3.05798501e-01
-5.61347567e-02 -7.28398144e-01 9.12768900e-01 -1.29098725e+00
-7.53576398e-01 1.44845748e+00 -2.70102531e-01 -3.75833571e-01
3.63789231e-01 -1.82434723e-01 -3.80960345e-01 4.11399841e-01
5.70180118e-01 1.40608215e+00 5.66667855e-01 -1.51167297e+00
-2.83798277e-01 -1.78753868e-01 4.78733450e-01 5.26899278e-01
-6.28158808e-01 2.03362554e-01 -7.54091382e-01 -4.98837084e-01
-4.94187117e-01 -9.64309514e-01 -1.49222299e-01 3.64258975e-01
1.60523448e-02 -4.67140339e-02 3.35030735e-01 -8.74015987e-01
4.57087368e-01 -2.20769525e+00 1.46379590e-01 -2.97729731e-01
-3.60516757e-02 -9.16402265e-02 -7.61132777e-01 5.15934706e-01
1.07381098e-01 4.94036436e-01 1.49691775e-02 -4.79314506e-01
6.66799396e-02 4.72272217e-01 -2.37969965e-01 5.22477090e-01
5.67376316e-01 1.22108138e+00 -1.12839556e+00 -8.85232627e-01
5.22093251e-02 8.04413617e-01 -5.19562840e-01 1.18773945e-01
1.16210826e-01 4.80777085e-01 -1.54063240e-01 3.29620034e-01
4.21434969e-01 -3.14573407e-01 3.26006204e-01 -3.30058366e-01
-1.75540462e-01 3.42722118e-01 -3.48865747e-01 2.24847937e+00
-1.14502370e+00 1.13542879e+00 6.01514764e-02 -8.49252641e-01
5.72416008e-01 4.35175657e-01 1.93343371e-01 -9.92595077e-01
-3.23696166e-01 3.66795301e-01 -1.34658769e-01 -5.50915718e-01
4.22179580e-01 -4.21802759e-01 -1.78188622e-01 5.17620265e-01
5.52161336e-01 -2.79538959e-01 2.00005129e-01 4.78603989e-01
5.96659362e-01 2.87124544e-01 1.91380769e-01 -4.35899347e-01
2.16422930e-01 3.01174134e-01 3.62830281e-01 6.53150260e-01
-3.58513325e-01 5.16861737e-01 5.18351011e-02 -3.04135770e-01
-1.53019202e+00 -1.27171779e+00 -3.24807256e-01 1.38292861e+00
2.02732503e-01 -4.18585896e-01 -3.71110946e-01 -7.70181775e-01
-1.06996554e-03 7.27116644e-01 -5.73917031e-01 5.09398021e-02
-5.04819393e-01 -3.48147810e-01 6.78566813e-01 7.73347378e-01
2.75719851e-01 -8.83650541e-01 -2.61014342e-01 -2.22933650e-01
-1.49647087e-01 -1.59089136e+00 -5.28676629e-01 -7.42584094e-03
-6.07236147e-01 -7.77878344e-01 -1.08474278e+00 -1.36317289e+00
8.65036786e-01 4.06744957e-01 1.52267635e+00 6.58447370e-02
-5.63871525e-02 8.06293011e-01 -4.74301189e-01 -2.06267461e-01
-4.59796250e-01 1.73395142e-01 1.74750730e-01 -4.14469779e-01
4.33186144e-01 -1.13322258e-01 -2.68060952e-01 -4.59261350e-02
-7.56607234e-01 2.59944797e-01 6.06609643e-01 8.61529768e-01
4.15294528e-01 -7.40172803e-01 4.90104347e-01 -6.72530234e-01
5.08038342e-01 -3.40304971e-01 -2.84896702e-01 6.59726024e-01
-5.59992015e-01 3.57284904e-01 4.02473509e-01 -2.87331343e-01
-9.20885086e-01 4.75788675e-02 -3.44843082e-02 -5.47152519e-01
-4.06135945e-03 6.02029502e-01 4.38913843e-03 -1.90706208e-01
6.49296641e-01 9.89743844e-02 -2.07141057e-01 -2.10268602e-01
9.28081930e-01 5.91045678e-01 3.98103386e-01 -7.71211982e-01
1.01717639e+00 6.11221731e-01 -4.82405782e-01 -8.12456727e-01
-7.73106456e-01 -4.47326094e-01 -9.87738848e-01 2.88814562e-03
1.26842678e+00 -1.61757970e+00 1.56860366e-01 -3.20547640e-01
-1.42455471e+00 -3.48255575e-01 -2.08669007e-01 8.17551255e-01
-4.85008359e-01 9.78532434e-02 -5.35770118e-01 -3.27752590e-01
7.11626466e-03 -1.31511581e+00 1.37447536e+00 -2.30943337e-01
-9.92365927e-03 -1.39151418e+00 1.82562739e-01 2.78542399e-01
1.35718644e-01 -1.47305444e-01 7.70033896e-01 -5.95922589e-01
-6.96818531e-01 1.81701064e-01 -5.35262108e-01 4.21818137e-01
9.91689563e-02 -2.99428880e-01 -8.40033412e-01 -6.68640375e-01
-2.40683079e-01 -9.17334378e-01 8.86919141e-01 5.78742148e-03
5.09150267e-01 -1.44610211e-01 -2.95099646e-01 6.99251473e-01
1.89928567e+00 -2.78826743e-01 3.58579457e-01 4.41519320e-01
9.49568391e-01 5.82641482e-01 3.60001057e-01 -4.15466338e-01
7.02425897e-01 5.92996299e-01 -3.26040313e-02 -5.54861724e-01
-4.18198675e-01 -5.63959420e-01 6.22539580e-01 1.09224439e+00
2.10677832e-01 -7.04791099e-02 -1.14111412e+00 1.06702077e+00
-1.52223766e+00 -7.38233924e-01 3.51769269e-01 2.01144528e+00
9.88663912e-01 -3.15861434e-01 -2.56245792e-01 -8.14889848e-01
5.41888118e-01 2.97269106e-01 -3.81018937e-01 -4.89642769e-01
-5.10294616e-01 1.64499521e-01 7.23077655e-01 7.94690907e-01
-9.61733937e-01 1.44745469e+00 7.79852867e+00 3.55568975e-01
-1.15807009e+00 7.10095584e-01 3.96985561e-01 3.37151699e-02
-8.39817643e-01 1.06155910e-01 -5.64889371e-01 -6.04550987e-02
9.36331987e-01 -2.64923960e-01 3.95672560e-01 7.15247810e-01
8.34406838e-02 2.23158807e-01 -1.30751061e+00 1.01168180e+00
7.65293062e-01 -1.35194802e+00 5.23401558e-01 -1.70006361e-02
1.02307439e+00 5.30006289e-01 2.29599014e-01 3.64427358e-01
4.78637844e-01 -1.05043161e+00 8.82963121e-01 2.29258746e-01
1.15046275e+00 -3.33320349e-01 4.84838754e-01 -4.02190797e-02
-1.31479347e+00 2.49612838e-01 -4.10066396e-01 2.31057972e-01
3.47571850e-01 -1.02400020e-01 -7.01015413e-01 6.40095651e-01
6.32312119e-01 9.41982925e-01 -9.73290443e-01 6.62082791e-01
-5.32003641e-01 2.24879533e-01 1.27450287e-01 2.20257297e-01
5.13132691e-01 -1.42645851e-01 1.93362117e-01 1.35027003e+00
3.62904251e-01 -3.97872150e-01 4.21214372e-01 8.72158408e-01
-3.97127062e-01 3.45868826e-01 -1.19492936e+00 -4.56765145e-01
3.11640710e-01 1.04157066e+00 -3.38652581e-01 -4.81605917e-01
-1.13440228e+00 1.26952028e+00 7.89396822e-01 7.04428136e-01
-6.84219241e-01 1.09708093e-01 6.67594850e-01 7.47800842e-02
1.97562844e-01 -5.89300573e-01 2.46787677e-04 -1.64846563e+00
-1.80301309e-01 -7.60989845e-01 1.30204067e-01 -1.37250125e+00
-1.52661681e+00 6.92008913e-01 2.35209223e-02 -1.00414228e+00
-2.58466750e-01 -9.27375734e-01 -5.89638017e-02 1.13927543e+00
-1.86278093e+00 -1.68288219e+00 -1.80126633e-02 6.85173452e-01
5.49557030e-01 -2.00189903e-01 9.58650708e-01 5.08381903e-01
-2.48409986e-01 6.25734627e-01 -1.03737088e-02 4.70384508e-01
1.17077065e+00 -1.23703682e+00 3.51104051e-01 9.58758116e-01
9.48515177e-01 7.35085726e-01 4.76598650e-01 -5.09642899e-01
-1.20291495e+00 -1.13944697e+00 1.27718365e+00 -6.63154364e-01
8.55151772e-01 -6.76395357e-01 -6.32476032e-01 1.23787332e+00
9.23672259e-01 2.31788177e-02 7.66373396e-01 4.45011109e-01
-8.53326380e-01 1.60261571e-01 -7.46361494e-01 7.45190859e-01
1.11867595e+00 -1.57924891e+00 -7.41842806e-01 6.05193675e-01
8.45519781e-01 -1.22706361e-01 -8.71291041e-01 9.75612924e-02
2.96406955e-01 -4.36952382e-01 1.06050909e+00 -7.75174975e-01
4.79617000e-01 -2.12555900e-01 -3.83995593e-01 -1.38968444e+00
-2.33195439e-01 -1.33162364e-01 6.69898868e-01 8.89287770e-01
6.43577516e-01 -5.12531936e-01 2.30491042e-01 1.71287894e-01
-1.82807565e-01 -3.09268415e-01 -8.39542210e-01 -1.06545448e+00
6.89265430e-01 -2.32238382e-01 2.80914247e-01 1.13454282e+00
-1.23365551e-01 6.14443243e-01 -3.26196522e-01 2.63490558e-01
3.76379341e-01 6.44380525e-02 5.13445497e-01 -7.58407772e-01
1.14886403e-01 -1.35444522e-01 -4.36839700e-01 -1.06520653e+00
8.19771945e-01 -1.43215775e+00 1.95758864e-01 -1.89315355e+00
6.51454985e-01 -3.02130729e-01 -4.21401888e-01 5.51112831e-01
-2.58679181e-01 7.67943680e-01 4.04943198e-01 4.02182877e-01
-7.74946034e-01 5.49071610e-01 1.24707639e+00 -4.75450009e-01
3.89623970e-01 -1.07739019e+00 -7.76448846e-01 4.71262157e-01
4.08364028e-01 -4.93156821e-01 -5.03651977e-01 -1.21470821e+00
3.23083907e-01 -4.22980368e-01 5.04918635e-01 -6.66489482e-01
2.38417797e-02 -1.55764567e-02 3.87460768e-01 -2.06182674e-02
3.76178265e-01 -7.19191015e-01 -2.80532569e-01 1.45447835e-01
-6.20690763e-01 5.56664407e-01 3.55307281e-01 4.96945828e-01
-4.85136062e-01 -2.20412821e-01 6.64883196e-01 -2.95763582e-01
-9.73683298e-01 2.15469658e-01 -2.36607790e-01 3.58913213e-01
8.35897982e-01 -5.62005416e-02 -4.01374221e-01 -3.65441859e-01
-6.20782495e-01 1.25595912e-01 1.11087430e+00 7.72159398e-01
3.65250379e-01 -1.84127700e+00 -8.69958639e-01 -2.25824360e-02
6.00664496e-01 -6.23353601e-01 -2.36925021e-01 7.78856337e-01
-5.78519762e-01 7.08371639e-01 -4.54469621e-01 -5.68755865e-01
-8.69411349e-01 5.91343105e-01 3.95550221e-01 -1.81528293e-02
-2.99722970e-01 7.87168562e-01 4.07657027e-01 -6.26499653e-01
-1.24753036e-01 -4.93930019e-02 4.06544954e-02 -3.39870155e-02
1.69307396e-01 -3.39871049e-01 -2.36898631e-01 -1.33033621e+00
-6.86677575e-01 6.97524130e-01 -4.76132110e-02 -8.71053338e-01
1.16205394e+00 -1.84510827e-01 3.74636762e-02 6.29476547e-01
1.62505794e+00 7.24159926e-02 -1.21998298e+00 -2.19273940e-01
-1.70208722e-01 -3.85400623e-01 1.25044465e-01 -6.50659025e-01
-8.22876692e-01 1.06731498e+00 5.74642241e-01 -3.43821675e-01
8.06190789e-01 3.73801053e-01 5.06590724e-01 4.24311638e-01
6.31441891e-01 -9.94104505e-01 1.98404863e-01 4.20476466e-01
1.03239739e+00 -1.59276688e+00 4.50058728e-02 2.13802550e-02
-9.97025192e-01 6.28234863e-01 5.04836380e-01 -1.34675056e-01
4.60924953e-01 1.22591471e-02 3.67661774e-01 -1.33821756e-01
-5.68523526e-01 -5.24258196e-01 4.37104762e-01 8.75430107e-01
8.05514097e-01 1.30116371e-02 -2.51410753e-01 3.66186202e-02
-5.94083592e-02 -1.73046321e-01 1.84830695e-01 9.00658250e-01
5.76929599e-02 -1.30216062e+00 -8.93614665e-02 6.40885830e-02
-2.19793692e-01 -5.99181235e-01 -5.14863014e-01 1.03698647e+00
2.71943808e-01 8.27372193e-01 4.51400638e-01 -5.44154681e-02
-2.36168243e-02 2.03113824e-01 8.01999629e-01 -1.07870495e+00
-3.27849627e-01 -1.78606182e-01 1.83154270e-01 -5.83095610e-01
-8.91116560e-01 -5.83058536e-01 -1.00005651e+00 1.48483515e-01
5.19090295e-02 2.18986437e-01 6.84335887e-01 8.91591787e-01
3.76080841e-01 1.23894677e-01 3.20612282e-01 -6.95735574e-01
-1.78860471e-01 -9.01767135e-01 -2.22898990e-01 6.37778282e-01
3.07958812e-01 -5.89076877e-01 -3.33841622e-01 4.00198430e-01] | [11.226873397827148, 1.6178866624832153] |
da9a5f15-6514-41b5-baa6-eac3b48000fb | tablex-a-benchmark-dataset-for-structure-and | 2105.06400 | null | https://arxiv.org/abs/2105.06400v1 | https://arxiv.org/pdf/2105.06400v1.pdf | TabLeX: A Benchmark Dataset for Structure and Content Information Extraction from Scientific Tables | Information Extraction (IE) from the tables present in scientific articles is challenging due to complicated tabular representations and complex embedded text. This paper presents TabLeX, a large-scale benchmark dataset comprising table images generated from scientific articles. TabLeX consists of two subsets, one for table structure extraction and the other for table content extraction. Each table image is accompanied by its corresponding LATEX source code. To facilitate the development of robust table IE tools, TabLeX contains images in different aspect ratios and in a variety of fonts. Our analysis sheds light on the shortcomings of current state-of-the-art table extraction models and shows that they fail on even simple table images. Towards the end, we experiment with a transformer-based existing baseline to report performance scores. In contrast to the static benchmarks, we plan to augment this dataset with more complex and diverse tables at regular intervals. | ['Mayank Singh', 'Pratik Kayal', 'Harsh Desai'] | 2021-05-12 | null | null | null | null | ['table-extraction'] | ['miscellaneous'] | [ 3.89775455e-01 6.88470826e-02 -1.78226292e-01 -9.36418325e-02
-1.24514186e+00 -1.34603465e+00 7.13570654e-01 5.82170904e-01
-2.12887768e-02 8.01252782e-01 2.30804488e-01 -6.46460235e-01
8.37153718e-02 -7.35363483e-01 -7.68904328e-01 -1.66816667e-01
1.59133196e-01 6.80402517e-01 1.57583177e-01 -3.66401039e-02
5.59270501e-01 5.77319801e-01 -1.35144556e+00 8.13181460e-01
5.94899595e-01 1.13815892e+00 -1.79443106e-01 7.02553570e-01
-4.90637690e-01 9.85593915e-01 -8.09367061e-01 -9.91950512e-01
3.76514763e-01 -3.09168816e-01 -6.75329268e-01 1.99134648e-01
8.86708736e-01 -2.22575262e-01 -2.50249743e-01 9.43060398e-01
2.32811928e-01 -3.65386486e-01 6.68034732e-01 -1.24709916e+00
-3.95341337e-01 7.50788569e-01 -6.26610458e-01 3.43230397e-01
5.54480791e-01 9.92611274e-02 1.13557291e+00 -9.28284228e-01
1.38646984e+00 1.11734557e+00 5.05498111e-01 5.97411618e-02
-1.24828982e+00 -6.73493385e-01 -5.44241071e-02 7.85770938e-02
-9.80261147e-01 -7.04559624e-01 5.40412664e-01 -4.80682164e-01
1.13036513e+00 4.25037801e-01 3.75888020e-01 1.05988431e+00
3.45938653e-01 6.33733451e-01 1.35566795e+00 -3.58198464e-01
-1.38213756e-02 3.96035522e-01 1.38618842e-01 7.11200356e-01
8.22727680e-01 -6.96033657e-01 -6.34245813e-01 -6.43030480e-02
3.96031171e-01 -4.91828710e-01 -1.11347340e-01 -6.01041794e-01
-1.77555239e+00 4.38380659e-01 1.46023557e-02 1.43914167e-02
-1.56014279e-01 -3.06466788e-01 7.03670800e-01 1.33687764e-01
8.47351998e-02 8.73636723e-01 -4.25245583e-01 -4.30556476e-01
-8.58374894e-01 5.23359239e-01 1.14245784e+00 1.54426157e+00
4.52225089e-01 -4.02293652e-01 -3.26911330e-01 6.49364769e-01
-2.38424137e-01 4.21420991e-01 -2.47634679e-01 -1.00274086e+00
1.21833789e+00 7.63755858e-01 -9.50981025e-03 -1.31693435e+00
-2.15836480e-01 -2.91876137e-01 -7.48636901e-01 1.98173046e-01
7.94039786e-01 2.03146473e-01 -7.84205317e-01 8.12136054e-01
2.27407247e-01 -8.80122662e-01 -1.39614508e-01 3.03545505e-01
1.22117221e+00 6.41298115e-01 -2.63021767e-01 6.93154931e-02
1.90627074e+00 -6.87242866e-01 -1.07981622e+00 5.03747091e-02
4.09257442e-01 -1.21533501e+00 1.20028687e+00 8.92972291e-01
-1.28748643e+00 -2.68729031e-01 -1.33478820e+00 -6.46569788e-01
-1.01764703e+00 3.17545831e-01 5.06979346e-01 5.99351704e-01
-7.30144680e-01 4.30491298e-01 -6.01378560e-01 -3.28356028e-01
6.67800903e-01 7.72171915e-02 -5.64510047e-01 -6.61674216e-02
-5.91317296e-01 7.81012058e-01 1.12848684e-01 -1.68304071e-01
-1.53075784e-01 -9.69018221e-01 -9.25764441e-01 1.19133070e-01
8.47247720e-01 -3.36005896e-01 1.17269850e+00 -1.95941310e-02
-8.93731475e-01 1.22476053e+00 4.54862714e-02 -2.52537876e-01
6.24561787e-01 -5.02350368e-03 -2.27699220e-01 3.75914663e-01
2.78217822e-01 5.17728865e-01 4.33821678e-01 -1.34230173e+00
-4.97934371e-01 -3.97431105e-01 -1.43584371e-01 -4.73678485e-02
-1.35181949e-01 1.72128037e-01 -1.09086359e+00 -1.05826223e+00
-2.35223353e-01 -8.25991690e-01 2.07075864e-01 1.45024270e-01
-9.21781719e-01 2.95374513e-01 6.56773806e-01 -9.04761434e-01
1.52419686e+00 -1.84547102e+00 -9.76539403e-02 2.73896039e-01
4.67546195e-01 -2.29200155e-01 1.32531911e-01 5.39525688e-01
-4.72655371e-02 4.76505339e-01 -2.17294469e-01 -2.68888384e-01
1.58453017e-01 -2.03165814e-01 -4.36778784e-01 1.67446122e-01
2.88230598e-01 8.25208783e-01 -4.58674401e-01 -9.86959875e-01
-9.25142989e-02 3.15713048e-01 -4.23478454e-01 -1.22239105e-01
-3.45729083e-01 -9.64704156e-02 -2.33036846e-01 1.09053028e+00
8.00242960e-01 -4.40852076e-01 3.34501952e-01 -5.38212240e-01
-2.62079448e-01 6.82209492e-01 -1.43652332e+00 1.45606327e+00
-6.55312389e-02 8.76558661e-01 9.04118717e-02 -3.29977334e-01
7.68687427e-01 -5.51044345e-02 3.92163724e-01 -7.37070620e-01
5.93517870e-02 2.97990084e-01 -1.03354350e-01 -9.67781618e-02
6.52037859e-01 7.07329810e-01 -2.56696671e-01 4.95569468e-01
-9.64801162e-02 -3.71048570e-01 1.07873857e+00 6.89579427e-01
1.21460044e+00 3.48343790e-01 3.66981208e-01 -4.31484759e-01
5.70719957e-01 4.83154178e-01 8.66285861e-02 7.22005844e-01
9.84293520e-02 9.22075272e-01 1.17995989e+00 -5.23074567e-01
-1.31701016e+00 -1.00021684e+00 -4.49005991e-01 5.20772994e-01
-2.44557828e-01 -1.19376183e+00 -7.83588409e-01 -8.19868505e-01
8.42654258e-02 3.73116523e-01 -7.85745919e-01 4.38834012e-01
-7.04327106e-01 -7.58356631e-01 2.91622162e-01 4.53539103e-01
4.07589704e-01 -1.06487799e+00 -6.57615364e-01 1.03222780e-01
-2.54156828e-01 -1.47234750e+00 -5.68655074e-01 4.72942978e-01
-5.40589094e-01 -1.24180710e+00 -4.45602417e-01 -5.27322471e-01
6.39580607e-01 -3.66075970e-02 1.89703512e+00 -1.37969166e-01
-4.55619544e-01 1.70431331e-01 -1.99283451e-01 -7.60185778e-01
-5.34949720e-01 4.14674908e-01 -6.02673173e-01 -4.60363775e-01
3.04149747e-01 -1.40358895e-01 -3.84770185e-01 1.34690434e-01
-1.00526762e+00 2.19171301e-01 5.84281862e-01 6.33533299e-01
1.00878358e+00 6.02517873e-02 -1.07901037e-01 -1.44586003e+00
5.96385837e-01 -9.04991031e-02 -9.90395427e-01 4.04738426e-01
-7.30060637e-01 2.87007034e-01 8.39611828e-01 -3.86100821e-02
-7.18133509e-01 -5.78450672e-02 1.42168060e-01 1.29058301e-01
8.23457167e-02 4.57609564e-01 -5.40201306e-01 1.60875306e-01
4.80267376e-01 -2.38647461e-02 -3.99512909e-02 -7.00686097e-01
2.78406352e-01 5.40102184e-01 7.89057434e-01 -5.54180920e-01
9.53362405e-01 2.70750940e-01 1.87798560e-01 -5.00315249e-01
-6.84841335e-01 -7.88804069e-02 -6.15667105e-01 -1.26756830e-02
6.25745416e-01 -8.60967457e-01 -6.48981452e-01 2.20382214e-01
-8.41751099e-01 -2.40445569e-01 -1.44904912e-01 -6.30159453e-02
-2.81747520e-01 1.79266676e-01 -8.24197948e-01 -1.84320271e-01
-3.06742549e-01 -1.28994286e+00 1.27096975e+00 -1.39830858e-01
-4.86493558e-01 -5.47362864e-01 -1.79406747e-01 5.52721500e-01
1.02224812e-01 4.83787030e-01 1.14460099e+00 -6.15719616e-01
-1.07500243e+00 -1.36583313e-01 -4.74058241e-01 -3.57615739e-01
9.88674611e-02 6.67215526e-01 -7.05105484e-01 1.35492040e-02
-3.38080764e-01 -1.97097868e-01 8.10093105e-01 4.06581275e-02
1.48584497e+00 -4.39639330e-01 -4.38054740e-01 7.36784160e-01
1.39732814e+00 3.44884336e-01 8.57395828e-01 1.00895154e+00
8.07777166e-01 5.95579445e-01 5.96095383e-01 4.41671938e-01
5.09717941e-01 5.89226246e-01 3.68170217e-02 -2.12237164e-01
-2.57257164e-01 -1.15498267e-01 -3.50240059e-02 5.89560628e-01
4.16654974e-01 -5.68647981e-01 -1.06192923e+00 3.87970775e-01
-1.19033706e+00 -7.94599712e-01 -2.99083680e-01 2.03320646e+00
1.12526500e+00 6.55824840e-01 4.37921941e-01 5.59049785e-01
1.68256000e-01 6.97107166e-02 -3.87156427e-01 -4.79384184e-01
-4.64064687e-01 2.29047224e-01 6.90420806e-01 1.45858750e-01
-1.21991670e+00 5.63573837e-01 6.70858431e+00 6.90130293e-01
-7.98444927e-01 -5.99730194e-01 1.14647150e+00 -6.81024790e-02
-5.00296712e-01 -1.54417410e-01 -1.10281420e+00 3.97091448e-01
8.57321680e-01 -1.48485184e-01 3.01134259e-01 5.47496498e-01
-2.42010638e-01 -2.52223641e-01 -1.13601673e+00 1.06634307e+00
1.01983257e-01 -1.87780380e+00 2.85619646e-01 2.14644849e-01
3.47456038e-01 -4.48162407e-01 5.73173687e-02 1.18057050e-01
8.33480135e-02 -1.09838879e+00 8.22610438e-01 2.92260587e-01
1.15273905e+00 -7.00747311e-01 5.80170512e-01 -4.26629871e-01
-1.09493816e+00 3.50444824e-01 -2.68048979e-02 3.51377457e-01
-3.45662385e-01 5.35507143e-01 -8.11895013e-01 4.77119178e-01
1.08559358e+00 8.88055146e-01 -1.49346781e+00 7.38267362e-01
2.03494355e-01 3.88453543e-01 -3.35540116e-01 -1.36544764e-01
-1.70186013e-01 -3.51144522e-01 2.87094712e-01 1.54322493e+00
1.02684125e-01 -1.75322428e-01 -2.32773647e-01 8.91041875e-01
-5.95122874e-01 2.46502325e-01 -8.85343313e-01 -3.90772969e-01
5.50967634e-01 1.53729391e+00 -1.42101264e+00 -6.74668193e-01
-6.57059133e-01 5.57670653e-01 1.77443147e-01 5.46179079e-02
-6.50689363e-01 -9.05619323e-01 3.74599665e-01 3.04290324e-01
7.00843453e-01 -1.57551438e-01 -1.17044950e+00 -1.25203657e+00
5.17295897e-01 -1.60780394e+00 6.15185082e-01 -8.62399459e-01
-9.23969686e-01 8.80652964e-01 1.69143379e-01 -1.34163690e+00
-8.27859119e-02 -8.95305932e-01 5.14135323e-02 6.67897642e-01
-1.26006889e+00 -7.94316113e-01 -3.21322113e-01 3.25232387e-01
3.65824223e-01 -2.13486925e-01 7.34945774e-01 4.10251588e-01
-7.23655164e-01 7.64186263e-01 3.03119391e-01 4.64490026e-01
9.72771347e-01 -1.68004787e+00 9.59179282e-01 8.38501990e-01
2.51297832e-01 7.76550293e-01 8.44146550e-01 -6.46806777e-01
-1.92279255e+00 -7.50424981e-01 1.04495692e+00 -1.07938612e+00
6.79989040e-01 -1.04150510e+00 -9.84100521e-01 5.98976433e-01
4.95007604e-01 -1.97221234e-01 5.59441030e-01 -1.07250966e-01
-6.03178918e-01 -3.87516946e-01 -1.04144990e+00 9.20646727e-01
7.47390211e-01 -3.96453291e-01 -2.40341038e-01 3.93501192e-01
3.64648640e-01 -8.46072078e-01 -1.23932433e+00 2.46571302e-01
7.62627304e-01 -8.77593398e-01 1.01490295e+00 -2.00862482e-01
9.71087933e-01 -4.21575606e-01 8.35757703e-02 -6.60913229e-01
-7.34146014e-02 -7.38946259e-01 -2.32163772e-01 1.40231216e+00
7.91537702e-01 -2.21573934e-01 9.17379439e-01 6.04351163e-01
1.57428086e-01 -8.30352783e-01 -4.24937993e-01 -5.57342172e-01
1.15950182e-01 -1.14492230e-01 6.60415888e-01 6.62845433e-01
5.06889448e-02 5.65509021e-01 2.37373747e-02 -4.38405216e-01
5.49894333e-01 4.99695718e-01 1.05816662e+00 -1.14542186e+00
1.38047099e-01 -6.60996258e-01 -1.38091981e-01 -6.04529440e-01
-4.70423758e-01 -7.14159966e-01 -2.23402768e-01 -1.75514054e+00
3.72540832e-01 -1.37090042e-01 1.56559125e-01 5.27761102e-01
-2.86410004e-01 5.54389298e-01 4.46580350e-01 1.30757183e-01
-7.26256311e-01 -6.76173791e-02 1.27256918e+00 -3.76415670e-01
8.12157989e-02 -5.95035255e-01 -9.97088790e-01 2.67819375e-01
6.09366477e-01 -5.22637427e-01 -2.81804949e-01 -3.66236605e-02
5.82040548e-01 9.93309915e-02 -9.24867541e-02 -1.02096367e+00
2.82995086e-02 1.30431503e-02 8.42581034e-01 -1.22864425e+00
-4.85255793e-02 -6.07370794e-01 6.64997622e-02 5.52351475e-02
-3.29839975e-01 8.40155065e-01 6.07178152e-01 2.05657631e-02
-1.39914319e-01 2.42925346e-01 5.05863190e-01 -2.43804023e-01
-2.01859772e-01 3.79340164e-02 -4.72252548e-01 3.32099020e-01
5.91048896e-01 -3.18075150e-01 -7.12944090e-01 -2.22387210e-01
-1.61566541e-01 1.32553250e-01 8.99027169e-01 3.62391829e-01
4.11203146e-01 -1.10124338e+00 -6.10819519e-01 2.32787549e-01
3.98687154e-01 -2.76167765e-02 -2.66470194e-01 5.52673995e-01
-1.07205963e+00 5.92272282e-01 -4.72153485e-01 -2.69436002e-01
-1.41043758e+00 8.92969012e-01 -1.47151187e-01 -6.66565180e-01
-7.06993163e-01 3.37883025e-01 -1.26777822e-02 -2.49240279e-01
4.72765207e-01 -7.38716960e-01 -2.96051383e-01 1.77047595e-01
6.71255767e-01 2.48431101e-01 5.90327263e-01 -3.19609612e-01
-5.28777838e-01 3.79146338e-01 -3.41660440e-01 -1.72824994e-01
1.18577909e+00 -1.94903426e-02 -2.15315390e-02 4.21239913e-01
1.14665568e+00 7.67271936e-01 -8.55679095e-01 1.99273020e-01
2.43274406e-01 -4.41338748e-01 -3.45000923e-01 -1.29667902e+00
-9.88559186e-01 5.01972198e-01 1.17451362e-01 2.27998033e-01
1.22624159e+00 -2.69672036e-01 6.29250705e-01 5.16656280e-01
-9.80116427e-03 -9.53930020e-01 -1.70419320e-01 3.98820430e-01
9.50176358e-01 -1.41400230e+00 7.92590559e-01 -7.86972642e-01
-5.95002830e-01 1.41559923e+00 3.44429374e-01 2.38186687e-01
4.10136819e-01 9.69771385e-01 2.81744003e-01 -4.94077533e-01
-9.97236073e-01 1.53493464e-01 6.22500420e-01 4.20933574e-01
7.37607002e-01 -4.02835578e-01 -1.68257356e-01 3.41029793e-01
-6.83073163e-01 -1.07569009e-01 8.90827596e-01 1.23575473e+00
2.26251930e-01 -1.31415403e+00 -7.48515666e-01 8.21092844e-01
-8.42569113e-01 -2.59849042e-01 -8.85136247e-01 1.19538271e+00
-2.02525571e-01 8.26310575e-01 1.11715913e-01 -1.11853100e-01
5.10201931e-01 -1.20389909e-01 6.08151376e-01 -3.24648291e-01
-8.34455311e-01 1.99601606e-01 4.09404129e-01 -5.62701523e-01
-8.59308243e-03 -9.10229027e-01 -9.15240109e-01 -4.60555643e-01
4.27642643e-01 -8.19840282e-02 5.94842553e-01 3.29594404e-01
5.89228630e-01 6.78764760e-01 -2.50428859e-02 -2.88733840e-01
-1.15330063e-01 -7.64243186e-01 -2.80086964e-01 4.94600296e-01
3.21185410e-01 -5.40668905e-01 -2.49389336e-02 5.40485561e-01] | [11.688565254211426, 2.989166021347046] |
33f07143-faa8-4270-a974-5ba97836ecd1 | auditory-separation-of-a-conversation-from | 1905.10751 | null | https://arxiv.org/abs/1905.10751v1 | https://arxiv.org/pdf/1905.10751v1.pdf | Auditory Separation of a Conversation from Background via Attentional Gating | We present a model for separating a set of voices out of a sound mixture containing an unknown number of sources. Our Attentional Gating Network (AGN) uses a variable attentional context to specify which speakers in the mixture are of interest. The attentional context is specified by an embedding vector which modifies the processing of a neural network through an additive bias. Individual speaker embeddings are learned to separate a single speaker while superpositions of the individual speaker embeddings are used to separate sets of speakers. We first evaluate AGN on a traditional single speaker separation task and show an improvement of 9% with respect to comparable models. Then, we introduce a new task to separate an arbitrary subset of voices from a mixture of an unknown-sized set of voices, inspired by the human ability to separate a conversation of interest from background chatter at a cafeteria. We show that AGN is the only model capable of solving this task, performing only 7% worse than on the single speaker separation task. | ['Bruno Olshausen', 'Shariq Mobin'] | 2019-05-26 | null | null | null | null | ['speaker-separation'] | ['speech'] | [ 4.52593088e-01 2.24352390e-01 3.02458197e-01 -1.69809565e-01
-7.54950345e-01 -6.16728365e-01 6.31086290e-01 -1.57319948e-01
-4.11722094e-01 3.74124050e-01 2.71777302e-01 -7.92641938e-02
-6.14935113e-03 -1.55764788e-01 -3.54867220e-01 -9.51508224e-01
-7.30041936e-02 6.30652547e-01 3.16740334e-01 -2.42403626e-01
-1.95937939e-02 5.73876083e-01 -1.84043670e+00 4.55011964e-01
3.88218194e-01 4.87735242e-01 4.51958537e-01 1.20244813e+00
-2.90409833e-01 3.75692248e-01 -1.09759724e+00 3.70069921e-01
1.07388660e-01 -7.53037751e-01 -8.41752768e-01 6.98203668e-02
8.01405609e-01 2.13729635e-01 -6.00501634e-02 1.09759200e+00
6.56337261e-01 5.01389205e-01 5.84229827e-01 -1.27032578e+00
-5.41622043e-01 9.29987788e-01 -1.30245724e-04 8.67938757e-01
1.73010796e-01 -3.68768326e-03 1.18397701e+00 -8.90461028e-01
1.18168183e-01 1.43933380e+00 2.88953245e-01 1.08014798e+00
-1.75031340e+00 -8.11235785e-01 6.01493955e-01 -2.73336023e-01
-1.28113103e+00 -1.14600444e+00 7.78117657e-01 -3.50725681e-01
1.27055657e+00 7.16733932e-01 6.08203888e-01 1.11307645e+00
6.23061806e-02 5.73157847e-01 7.22164094e-01 -4.26842570e-01
4.13579673e-01 5.49641848e-01 3.53447348e-01 2.51992643e-01
1.06464894e-02 -7.63047338e-02 -8.80802274e-01 -3.39140505e-01
4.42428827e-01 -1.15168020e-01 -4.15380299e-01 -1.48162708e-01
-1.06849873e+00 8.12050283e-01 2.47910529e-01 7.86829174e-01
-3.55973482e-01 1.24570675e-01 -5.54649308e-02 3.52912396e-01
2.57776320e-01 7.56787062e-01 -4.91657823e-01 2.06002310e-01
-1.15436554e+00 1.74201131e-01 1.07720375e+00 6.52684391e-01
4.83174056e-01 4.31168586e-01 -1.17460199e-01 8.00763249e-01
4.46965873e-01 4.52390283e-01 7.98744261e-01 -9.92726207e-01
2.31652632e-01 8.55905786e-02 6.78936392e-02 -4.61181551e-01
-3.37300301e-01 -3.98072392e-01 -3.68019134e-01 4.58721101e-01
4.32260692e-01 -3.31565797e-01 -7.75056958e-01 2.11327147e+00
1.30269110e-01 3.56021196e-01 7.45805949e-02 8.15588772e-01
8.12269628e-01 7.09944308e-01 -5.86153530e-02 -3.56745124e-01
1.59435308e+00 -1.04085398e+00 -8.95852685e-01 -7.21272528e-01
-1.93741266e-02 -6.59221351e-01 8.57268989e-01 4.46957409e-01
-1.25968122e+00 -6.97454751e-01 -1.07872355e+00 2.23535091e-01
-4.09027457e-01 -2.69095868e-01 2.53070563e-01 9.32594597e-01
-1.30503368e+00 3.46930146e-01 -6.43589258e-01 -3.20576131e-01
1.75623104e-01 6.41464531e-01 -3.27950865e-01 5.39314866e-01
-1.13667941e+00 9.18344736e-01 3.57759446e-02 3.83006223e-03
-1.10330319e+00 -6.62395835e-01 -8.53739262e-01 5.00805140e-01
7.19756782e-02 -6.80671513e-01 1.19781113e+00 -1.40315259e+00
-1.68296063e+00 6.91099405e-01 -5.89943290e-01 -2.74730265e-01
-2.49683950e-02 -7.16650784e-02 -8.96692872e-01 1.20333605e-01
-3.35785747e-02 6.36297226e-01 1.51296186e+00 -1.25847697e+00
-6.34136081e-01 -2.07290530e-01 -1.17286295e-01 2.12432310e-01
-2.75597543e-01 4.12310123e-01 -3.81810926e-02 -4.69553828e-01
-3.26528512e-02 -8.64996850e-01 1.30796535e-02 -4.23719496e-01
-4.69539732e-01 -1.62668854e-01 7.70811856e-01 -3.91728312e-01
1.12806571e+00 -2.39954758e+00 5.66772878e-01 -2.19315253e-02
5.44085503e-01 7.26662129e-02 -5.24148405e-01 1.03278816e-01
-6.05521023e-01 3.05395007e-01 -2.24662587e-01 -7.81071305e-01
3.85020929e-03 1.08179934e-01 -3.79450977e-01 3.56339574e-01
4.92142916e-01 2.59089112e-01 -8.82740498e-01 -1.66907668e-01
-1.28459379e-01 5.97000659e-01 -7.62710810e-01 4.18279380e-01
1.54381897e-02 3.41527462e-01 1.59688085e-01 1.85717210e-01
3.93900961e-01 1.63415611e-01 1.82124555e-01 2.82807201e-01
-6.78377226e-02 6.10639095e-01 -1.40739465e+00 1.43761325e+00
-2.73218691e-01 9.89862442e-01 7.91030169e-01 -5.06310105e-01
6.15509272e-01 7.23493278e-01 -4.80752699e-02 1.59693316e-01
3.67564857e-01 -7.15118870e-02 8.33329260e-01 -3.47965419e-01
4.83781010e-01 -7.55560040e-01 -1.57380015e-01 7.15564966e-01
4.94266480e-01 -1.90151349e-01 4.57823016e-02 1.83498889e-01
1.17228389e+00 -6.58186138e-01 7.25107417e-02 -5.44639707e-01
4.76867914e-01 -6.71454072e-01 4.19904083e-01 7.47758806e-01
-7.49537468e-01 6.41934097e-01 6.25593305e-01 1.75625198e-02
-4.33624029e-01 -1.27671993e+00 1.31461024e-01 1.69331288e+00
-1.61709741e-01 -1.77684158e-01 -7.09177911e-01 -4.16004360e-01
-3.24913450e-02 1.17025876e+00 -8.42488885e-01 -3.64571095e-01
-5.57482600e-01 -6.04096711e-01 5.04928529e-01 3.57574850e-01
-4.02780950e-01 -1.34335971e+00 -4.23309594e-01 1.47769786e-02
-1.13513350e-01 -8.02646160e-01 -9.68542278e-01 8.09552670e-01
-4.93018389e-01 -6.86107099e-01 -3.37075263e-01 -9.24726903e-01
5.28607428e-01 2.10492194e-01 1.01545107e+00 -1.62815139e-01
-9.72463563e-02 3.47282052e-01 1.96015373e-01 -7.43685305e-01
-7.44760633e-01 1.69831127e-01 4.69452858e-01 3.63441586e-01
4.94125992e-01 -7.28187442e-01 -1.76759914e-01 8.41331109e-02
-8.67364883e-01 -4.46797192e-01 2.45281965e-01 5.81407309e-01
-1.24167852e-01 5.11127226e-02 6.64075553e-01 -6.47507906e-01
8.99212778e-01 -6.61586523e-01 -4.59783897e-02 -2.58911341e-01
6.70719817e-02 1.59676209e-01 4.69923109e-01 -9.58590329e-01
-8.78903866e-01 6.76204190e-02 1.49693027e-01 -5.39856613e-01
-5.28679550e-01 -1.52467415e-01 -3.80039036e-01 3.10526729e-01
6.00721061e-01 9.77888554e-02 1.02985166e-01 -4.28922594e-01
4.47569996e-01 6.18961632e-01 3.51367444e-01 -2.10121110e-01
6.23298109e-01 2.34842226e-01 -6.61575973e-01 -1.08787358e+00
-5.86132884e-01 -6.53733313e-01 -7.82641172e-01 -7.94009790e-02
9.77738321e-01 -6.34197950e-01 -6.19004667e-01 9.69739747e-04
-1.30794668e+00 -2.73087025e-01 -5.63933372e-01 6.24241948e-01
-3.73931140e-01 -6.42533600e-02 -3.34558010e-01 -1.08280301e+00
3.22030112e-02 -1.00765979e+00 8.85159433e-01 1.91932485e-01
-7.80313611e-01 -8.66717875e-01 3.37205291e-01 2.03476638e-01
4.27384436e-01 -5.54446340e-01 8.42692971e-01 -1.52654600e+00
-6.10074401e-02 -2.03338433e-02 3.90519619e-01 4.62920189e-01
5.08245945e-01 3.07001993e-02 -1.50923872e+00 -3.07275742e-01
5.11890113e-01 1.34344369e-01 1.11731005e+00 2.10016206e-01
6.36322796e-01 -1.30275950e-01 -3.37894112e-01 1.30047262e-01
8.14382195e-01 2.56256372e-01 2.20140591e-01 -2.72199035e-01
5.15473425e-01 6.99028075e-01 -3.85862589e-01 1.20831743e-01
-6.67099133e-02 2.24754840e-01 5.62234163e-01 1.90526307e-01
-2.40511075e-01 2.47443557e-01 7.09131658e-01 1.02446496e+00
-3.47526222e-02 -5.35142362e-01 -5.95977843e-01 9.09063935e-01
-1.26691902e+00 -1.24940479e+00 2.34102950e-01 2.29842186e+00
8.41523349e-01 2.63219595e-01 3.08341622e-01 1.75171152e-01
1.11319613e+00 3.96156073e-01 -4.34730440e-01 -6.88477695e-01
4.91760932e-02 4.27279800e-01 -7.30972812e-02 1.04495871e+00
-1.16955554e+00 8.97059679e-01 7.71119452e+00 2.88177580e-01
-1.18172598e+00 2.18443483e-01 -1.08830445e-01 -8.68593812e-01
-2.26377025e-01 -1.76520258e-01 -8.39699209e-01 4.45626289e-01
1.51766515e+00 1.77451279e-02 7.26200581e-01 3.34240913e-01
5.45805357e-02 2.40465149e-01 -1.61389697e+00 7.57179379e-01
6.34893537e-01 -8.67955446e-01 4.75972146e-02 9.72246155e-02
3.30583155e-01 -1.95100196e-02 1.20654486e-01 5.23938358e-01
4.87215757e-01 -1.27300727e+00 7.04025745e-01 4.47719067e-01
9.17748734e-02 -5.30945957e-01 3.85840803e-01 5.30678153e-01
-9.36800539e-01 -3.63617510e-01 -1.39765143e-01 -2.25112304e-01
8.14733952e-02 8.32951590e-02 -8.72275054e-01 -2.19653472e-01
6.42440856e-01 -1.23188850e-02 -2.85556853e-01 6.95238054e-01
-1.25765353e-01 7.32833624e-01 -2.74508804e-01 1.08675845e-01
1.15338460e-01 2.96714336e-01 1.13791645e+00 1.37355447e+00
1.95662737e-01 -1.37549490e-01 -7.56088942e-02 1.03964996e+00
-2.96911806e-01 -1.43715978e-01 -4.81756896e-01 -1.20052524e-01
6.66501045e-01 1.34942448e+00 -5.48839450e-01 -4.56287086e-01
-1.42162755e-01 9.59068954e-01 4.19169575e-01 3.92659873e-01
-7.10661829e-01 -6.03886902e-01 1.10965395e+00 -2.15360476e-03
8.18072081e-01 3.20070423e-02 8.14706534e-02 -9.07105446e-01
-3.45345050e-01 -8.38325679e-01 1.28245026e-01 -7.72620857e-01
-1.33068979e+00 9.68276322e-01 -2.77942747e-01 -1.01772499e+00
-1.82146579e-01 -4.56148535e-01 -9.42582667e-01 1.37199342e+00
-1.21151638e+00 -5.33827007e-01 5.28289601e-02 3.63649130e-01
7.76168883e-01 -4.39666182e-01 1.26264811e+00 -2.13109441e-02
-7.17249572e-01 3.08894098e-01 -3.33934009e-01 6.38263822e-02
9.46266711e-01 -1.46481121e+00 2.46391803e-01 7.37191856e-01
4.27890390e-01 9.47029948e-01 1.01552117e+00 -1.64954647e-01
-9.73311961e-01 -9.27326083e-01 1.32170451e+00 -9.14730191e-01
6.48625255e-01 -8.38233411e-01 -1.14172423e+00 9.19178069e-01
7.44793177e-01 3.04476935e-02 1.31236970e+00 1.93819135e-01
-4.38580513e-01 4.48777787e-02 -9.95953500e-01 6.43923223e-01
6.85422599e-01 -6.93773508e-01 -1.49413836e+00 -6.52214587e-02
9.42606628e-01 -8.77787266e-03 -3.67037952e-01 -4.10737753e-01
4.70181555e-01 -9.77894127e-01 9.05165374e-01 -1.02435529e+00
-3.06452271e-02 -3.81922752e-01 -3.16949636e-01 -1.60232544e+00
-7.08893776e-01 -7.90137589e-01 -1.04996733e-01 1.20183074e+00
3.88610005e-01 -9.10564840e-01 3.52840126e-01 4.38927382e-01
-3.59246060e-02 9.66035761e-03 -1.13356280e+00 -7.04505086e-01
1.74426422e-01 -4.12623107e-01 4.94459391e-01 8.96043837e-01
1.02739580e-01 9.15104389e-01 -1.28304511e-01 5.91343284e-01
2.68501788e-01 -9.74901617e-02 3.39765638e-01 -1.63877439e+00
-4.45215881e-01 -7.87995517e-01 -1.94318742e-01 -8.20688725e-01
4.75984246e-01 -1.08412254e+00 5.23522019e-01 -1.06603920e+00
-3.73476669e-02 3.32775302e-02 -7.30649173e-01 3.36392581e-01
-2.36126721e-01 1.56982824e-01 4.69633341e-01 3.94908711e-02
-3.73912513e-01 4.07838255e-01 7.54129350e-01 -3.11179042e-01
-4.73598570e-01 2.04172596e-01 -1.17677784e+00 9.59573686e-01
6.83464110e-01 -8.75156522e-01 -3.24740887e-01 -2.52022564e-01
-2.64501542e-01 -2.99395025e-02 2.97856152e-01 -9.95942414e-01
2.88255036e-01 1.19895086e-01 3.59474063e-01 -2.07524300e-01
7.18105495e-01 -8.83837581e-01 -1.72750667e-01 2.68114924e-01
-6.74493670e-01 -1.32312521e-01 5.35574317e-01 6.19319141e-01
-5.65966964e-02 -4.43827808e-01 7.47485995e-01 -4.07474302e-02
-1.42397866e-01 -1.69392437e-01 -1.04850578e+00 7.71941096e-02
7.66912818e-01 -1.06150709e-01 -8.26660395e-02 -3.22182715e-01
-1.40224159e+00 -8.52474123e-02 -1.15551174e-01 5.91610551e-01
3.37966919e-01 -1.21504724e+00 -7.44926512e-01 6.54778183e-01
-2.13373080e-01 -3.63940686e-01 1.44365221e-01 7.60302067e-01
1.50998592e-01 3.27427447e-01 -1.36983708e-01 -4.79306519e-01
-1.73797703e+00 6.16522014e-01 4.66147453e-01 -7.65939057e-03
-1.89786807e-01 1.23131180e+00 4.76832956e-01 -5.09572208e-01
2.08679482e-01 -6.38520777e-01 -3.96028817e-01 4.38558698e-01
7.81460643e-01 2.03801513e-01 -1.11460671e-01 -7.80622959e-01
-6.05960429e-01 1.03794150e-01 4.59098369e-02 -3.85027885e-01
1.10148847e+00 1.25216767e-02 -1.66244671e-01 1.18568838e+00
1.00914526e+00 6.78934336e-01 -9.96477723e-01 -1.04972601e-01
-3.30834657e-01 -1.90132886e-01 -8.06568842e-03 -6.35006249e-01
-7.77411103e-01 1.06820738e+00 4.54450309e-01 8.82111847e-01
7.08297670e-01 3.61504614e-01 2.94734210e-01 2.00922221e-01
-3.24876904e-01 -7.53730476e-01 5.68479411e-02 6.32392704e-01
1.14473772e+00 -9.66309786e-01 -3.79341006e-01 -2.11042255e-01
-6.35415614e-01 1.02233696e+00 3.50598991e-01 -3.95681351e-01
8.64058077e-01 5.40858090e-01 2.58681506e-01 -1.63945735e-01
-1.27492976e+00 -3.06189299e-01 3.83668542e-01 5.94831288e-01
5.27615309e-01 2.70100404e-02 5.34557462e-01 9.01555836e-01
-3.71307939e-01 -6.68286324e-01 4.25598115e-01 6.03290915e-01
-7.40161896e-01 -7.72208869e-01 -7.51226962e-01 1.78332984e-01
-4.39999372e-01 -2.40280777e-01 -5.71984589e-01 6.39204085e-01
2.77382851e-01 1.17038095e+00 6.08223081e-01 -1.76206321e-01
2.93145329e-01 8.61470282e-01 4.21159446e-01 -1.31454718e+00
-9.35449958e-01 4.01088387e-01 -1.34506286e-03 -5.46949022e-02
-3.21097702e-01 -8.11314225e-01 -1.10221863e+00 2.47640684e-01
-3.49061251e-01 4.14154798e-01 5.28239191e-01 9.51235831e-01
1.89276487e-01 1.15943587e+00 4.02560741e-01 -1.24503791e+00
-4.42367792e-01 -1.12240779e+00 -7.90338397e-01 6.70286715e-02
9.99943614e-01 -3.67476314e-01 -1.11751509e+00 1.58089280e-01] | [14.904731750488281, 5.969191074371338] |
7eb1df3d-c380-48c6-ab1c-9db632cff5e6 | binding-language-models-in-symbolic-languages | 2210.02875 | null | https://arxiv.org/abs/2210.02875v2 | https://arxiv.org/pdf/2210.02875v2.pdf | Binding Language Models in Symbolic Languages | Though end-to-end neural approaches have recently been dominating NLP tasks in both performance and ease-of-use, they lack interpretability and robustness. We propose Binder, a training-free neural-symbolic framework that maps the task input to a program, which (1) allows binding a unified API of language model (LM) functionalities to a programming language (e.g., SQL, Python) to extend its grammar coverage and thus tackle more diverse questions, (2) adopts an LM as both the program parser and the underlying model called by the API during execution, and (3) requires only a few in-context exemplar annotations. Specifically, we employ GPT-3 Codex as the LM. In the parsing stage, with only a few in-context exemplars, Codex is able to identify the part of the task input that cannot be answerable by the original programming language, correctly generate API calls to prompt Codex to solve the unanswerable part, and identify where to place the API calls while being compatible with the original grammar. In the execution stage, Codex can perform versatile functionalities (e.g., commonsense QA, information extraction) given proper prompts in the API calls. Binder achieves state-of-the-art results on WikiTableQuestions and TabFact datasets, with explicit output programs that benefit human debugging. Note that previous best systems are all finetuned on tens of thousands of task-specific samples, while Binder only uses dozens of annotations as in-context exemplars without any training. Our code is available at https://github.com/HKUNLP/Binder . | ['Tao Yu', 'Noah A. Smith', 'Luke Zettlemoyer', 'Mari Ostendorf', 'Dragomir Radev', 'Caiming Xiong', 'Yushi Hu', 'Rahul Nadkarni', 'Chengzu Li', 'Peng Shi', 'Tianbao Xie', 'Zhoujun Cheng'] | 2022-10-06 | null | null | null | null | ['table-based-fact-verification'] | ['natural-language-processing'] | [ 6.39307722e-02 4.39774454e-01 -1.54677108e-01 -5.57245791e-01
-1.06050706e+00 -1.12489212e+00 2.16372564e-01 4.68056835e-02
-1.67671204e-01 4.95583892e-01 -2.54583895e-01 -8.92478287e-01
9.77202281e-02 -8.63181233e-01 -1.23803878e+00 -3.66487540e-02
3.78204435e-01 7.65055418e-01 1.07698657e-01 -1.57921627e-01
-3.17458436e-02 -1.50431804e-02 -1.48047078e+00 6.80494487e-01
1.17095566e+00 7.40193605e-01 4.66651738e-01 7.43228912e-01
-5.22799790e-01 8.99651468e-01 -8.33244622e-01 -6.00623727e-01
-7.14890882e-02 -2.15424113e-02 -1.11113048e+00 -6.06581330e-01
3.97725582e-01 -3.17264020e-01 1.17115222e-01 1.07351398e+00
3.06956977e-01 -2.11291596e-01 1.01192631e-01 -1.28085995e+00
-7.04274058e-01 1.16120493e+00 7.25681111e-02 -1.27526850e-01
5.78688085e-01 7.13752925e-01 1.27598810e+00 -6.90255821e-01
6.67365074e-01 1.14293981e+00 5.04881024e-01 8.87800217e-01
-1.38933694e+00 -5.23004532e-01 -2.09209025e-02 -5.99756576e-02
-9.84353125e-01 -4.60062087e-01 3.97660017e-01 -4.01825309e-01
1.73870134e+00 5.93573928e-01 1.97674379e-01 1.22993362e+00
-6.15101047e-02 8.24768066e-01 7.43338823e-01 -5.08493066e-01
2.30466768e-01 6.92142844e-02 5.91960430e-01 8.73408318e-01
-1.31625205e-01 -2.21645981e-02 -2.35615030e-01 -3.00356209e-01
4.17955548e-01 -2.73250520e-01 -3.64521444e-01 1.80930793e-01
-1.12163305e+00 6.56991661e-01 3.37355435e-01 1.90208972e-01
-1.16551772e-01 3.70747954e-01 5.78299761e-01 3.17143172e-01
-9.95134339e-02 9.70821619e-01 -1.10708833e+00 -4.87971842e-01
-7.99437642e-01 3.53635758e-01 1.25870526e+00 1.19515157e+00
8.65313292e-01 -1.00852720e-01 -3.03243726e-01 6.85850322e-01
-3.07133570e-02 5.10323882e-01 4.85926807e-01 -1.12862122e+00
8.77523780e-01 1.04949570e+00 -5.94411343e-02 -4.26855922e-01
-3.04903358e-01 -3.79265994e-01 -2.95818925e-01 1.30185395e-01
6.25244260e-01 -2.32600272e-01 -6.73840761e-01 1.94881701e+00
1.62579000e-01 -1.13949649e-01 1.87724099e-01 6.40235364e-01
1.02320611e+00 7.08409011e-01 1.45978808e-01 2.21063644e-01
1.63851607e+00 -8.91262531e-01 -3.48481685e-01 -8.00254345e-01
1.06335163e+00 -2.56743073e-01 1.94752443e+00 4.84231174e-01
-1.18950832e+00 -5.85005462e-01 -7.57302880e-01 -4.63615000e-01
-3.51182997e-01 3.46138954e-01 6.72839701e-01 2.75818467e-01
-1.02474928e+00 5.12882531e-01 -8.66924882e-01 -1.51374832e-01
2.11914316e-01 3.06237251e-01 -2.96262234e-01 -5.25460057e-02
-1.22280979e+00 7.37154305e-01 8.63864124e-01 -9.39646736e-02
-8.16071033e-01 -9.12941813e-01 -1.05219293e+00 3.37387770e-01
7.20176637e-01 -8.55838895e-01 1.67112899e+00 -9.47228312e-01
-1.37451994e+00 8.68220627e-01 -1.30288556e-01 -3.12875688e-01
4.06725407e-01 -3.11844200e-01 -1.83729932e-01 -1.79422885e-01
2.00495794e-01 7.24258602e-01 3.78876269e-01 -7.67802060e-01
-4.42721814e-01 -2.97438741e-01 5.32633364e-01 -1.30587816e-01
7.46124238e-02 2.41309628e-01 -6.62368417e-01 -2.14501575e-01
-2.32351437e-01 -8.58730018e-01 1.16686605e-01 -3.13986838e-01
-8.87506723e-01 -3.46262813e-01 3.39028358e-01 -8.98682475e-01
1.32399690e+00 -2.13856792e+00 1.50301740e-01 -1.19826987e-01
2.72825032e-01 2.34552994e-01 -2.29982346e-01 2.19000474e-01
-2.34783173e-01 4.98836845e-01 -4.71496403e-01 -1.89445928e-01
4.55640554e-01 4.83582228e-01 -5.14764071e-01 -2.10459813e-01
5.09174049e-01 1.29128683e+00 -8.72362196e-01 -3.13949257e-01
-1.04720622e-01 -5.68241775e-02 -8.18912089e-01 5.89129090e-01
-1.01866174e+00 1.18607648e-01 -4.80674118e-01 7.59866536e-01
3.39094996e-01 -4.59565371e-01 1.65319666e-01 3.11736222e-02
2.05103476e-02 8.92095208e-01 -1.02545989e+00 1.77299988e+00
-7.74081230e-01 4.22718108e-01 8.37053433e-02 -7.19219685e-01
7.45891273e-01 1.97514474e-01 -5.20696640e-01 -5.09611130e-01
-2.33785927e-01 4.01434183e-01 6.80548102e-02 -8.64643931e-01
2.24300876e-01 3.46383333e-01 -4.62085664e-01 4.45626348e-01
2.77898282e-01 -1.24715649e-01 3.87823582e-01 9.37011912e-02
1.48478734e+00 4.98606235e-01 4.90075141e-01 1.13315605e-01
4.58111227e-01 2.60410994e-01 5.74434102e-01 9.03934777e-01
3.17457318e-01 4.28037763e-01 9.97594297e-01 -5.03591895e-01
-8.95454526e-01 -8.73252690e-01 1.04638040e-01 1.52253592e+00
-4.47194040e-01 -5.09514868e-01 -9.96847332e-01 -8.42159152e-01
-1.49373472e-01 1.43470860e+00 -4.36857611e-01 -2.56099924e-02
-8.36990654e-01 -3.24256212e-01 1.08319592e+00 5.71199596e-01
3.57362241e-01 -1.61919165e+00 -8.13388765e-01 2.36876741e-01
-3.45054686e-01 -1.02095175e+00 -3.47079664e-01 6.33287549e-01
-6.88662648e-01 -1.17268538e+00 1.50133997e-01 -6.23891532e-01
5.53986967e-01 -5.41612446e-01 1.69837141e+00 3.02920491e-01
-9.29778814e-02 6.98961839e-02 -2.13885412e-01 -1.16815053e-01
-8.31685483e-01 2.94836730e-01 -5.88726461e-01 -5.91699481e-01
4.72236484e-01 -6.01823568e-01 3.62544172e-02 7.50897527e-02
-8.21229458e-01 2.78812051e-01 4.29820061e-01 7.41048515e-01
5.58431208e-01 -3.76714110e-01 3.43132347e-01 -1.38701761e+00
6.64541662e-01 -5.45206785e-01 -8.67783427e-01 5.00206530e-01
-3.78007114e-01 3.17811430e-01 1.07394373e+00 -3.91457677e-01
-7.49482036e-01 8.16661492e-02 -3.01816076e-01 -2.70089507e-01
-5.94329953e-01 8.93766403e-01 -5.48941016e-01 4.71976191e-01
1.09282410e+00 3.35695028e-01 -1.51904032e-01 -5.29593050e-01
4.87619728e-01 4.79948878e-01 8.78998339e-01 -1.13811505e+00
6.01754844e-01 -4.64632809e-01 -4.99289125e-01 -1.20431170e-01
-6.63128734e-01 3.08120809e-02 -3.79752487e-01 2.99648702e-01
6.42424643e-01 -6.33929253e-01 -9.54366982e-01 1.77691579e-01
-1.57899928e+00 -8.62083614e-01 -2.16789037e-01 -2.13147908e-01
-5.14239430e-01 4.77945879e-02 -6.49502993e-01 -6.37922108e-01
-7.00011134e-01 -1.40011036e+00 1.04030704e+00 1.39551282e-01
-6.33126020e-01 -7.48821735e-01 -1.77438900e-01 3.72978300e-01
3.29261154e-01 1.53098151e-01 1.62103486e+00 -1.16425157e+00
-5.89326322e-01 -1.86291590e-01 -1.14922807e-01 4.53010291e-01
-4.08075571e-01 1.90079227e-01 -1.00009775e+00 2.34478921e-01
-1.16189979e-01 -5.58455706e-01 3.42493474e-01 5.91959544e-02
1.50054443e+00 -8.06885064e-01 -2.89576918e-01 8.84989321e-01
1.16603088e+00 -3.97274923e-03 6.90354526e-01 3.48560065e-01
5.85070908e-01 5.58238029e-01 2.42876217e-01 -1.27492677e-02
4.76112366e-01 5.02334118e-01 6.57594144e-01 2.30590105e-01
1.90886706e-01 -4.44673955e-01 6.47422254e-01 3.24643642e-01
4.23552752e-01 -1.69143274e-01 -1.34145832e+00 4.35224950e-01
-1.81103909e+00 -6.31610453e-01 -2.74928302e-01 2.03400922e+00
1.37925959e+00 2.49208480e-01 -1.62578776e-01 -1.52173489e-01
4.71136034e-01 -1.97654977e-01 -7.07458138e-01 -6.38410330e-01
8.18905309e-02 4.00975078e-01 5.11821359e-02 6.00085795e-01
-8.00486326e-01 1.18515587e+00 4.93177843e+00 7.93783367e-01
-1.08885133e+00 1.04163788e-01 4.58287060e-01 -7.85569474e-02
-4.65189815e-01 1.66282281e-01 -9.08133984e-01 5.43509424e-01
1.16795635e+00 1.72542464e-02 9.27221656e-01 1.25692689e+00
-1.28100500e-01 1.16371028e-01 -1.82743275e+00 6.81741357e-01
-3.37146282e-01 -1.24571609e+00 -1.18540637e-01 -2.97030717e-01
-7.62699172e-02 2.40292758e-01 -2.15427801e-01 1.05548584e+00
2.23739699e-01 -1.24103677e+00 1.01291704e+00 3.98937762e-01
1.08697391e+00 -4.09201950e-01 5.66614151e-01 8.75326514e-01
-6.62226439e-01 -1.60660043e-01 -1.69592261e-01 -1.64941326e-01
-1.64092213e-01 5.26348770e-01 -1.01875126e+00 3.35001558e-01
6.23277068e-01 1.96624622e-01 -8.60330462e-01 5.60923755e-01
-9.58336353e-01 7.73594320e-01 -4.49705154e-01 -1.57132268e-01
3.43555026e-02 1.16523825e-01 5.92205286e-01 1.33732927e+00
1.47328436e-01 1.19512784e-03 1.95191056e-01 1.69866621e+00
-2.47237444e-01 -3.49731110e-02 -4.13694233e-01 -2.00038284e-01
6.56813920e-01 1.35682583e+00 -4.09138203e-02 -4.69242811e-01
-2.56098211e-01 7.57020414e-01 8.67486298e-01 4.88468438e-01
-9.34226274e-01 -6.05185986e-01 4.51779991e-01 1.58751920e-01
7.28205740e-02 1.02420963e-01 -4.65042740e-01 -1.11256027e+00
5.26866674e-01 -1.30342674e+00 3.68909627e-01 -1.20126712e+00
-1.11134255e+00 7.00084686e-01 -1.23649202e-02 -5.00743985e-01
-6.72262907e-01 -7.88451433e-01 -7.99556077e-01 1.23007584e+00
-1.29264832e+00 -1.15875697e+00 -2.21035972e-01 2.90704042e-01
3.94825131e-01 -3.36565115e-02 1.24979413e+00 1.50253162e-01
-7.53289521e-01 7.07274318e-01 -5.01979113e-01 3.91576558e-01
4.70620722e-01 -1.61777437e+00 6.85287595e-01 8.75933290e-01
-1.02230512e-01 1.09630001e+00 5.68721354e-01 -5.71589231e-01
-1.60267603e+00 -1.26308954e+00 1.12513602e+00 -7.40613341e-01
7.07366884e-01 -6.21206105e-01 -1.25086105e+00 1.01410079e+00
-5.32715917e-02 2.83499770e-02 5.96264184e-01 3.12148750e-01
-6.70762300e-01 8.07078406e-02 -1.00374699e+00 5.29710650e-01
8.06201875e-01 -6.62176311e-01 -9.52427208e-01 5.11196434e-01
1.10345864e+00 -8.79512191e-01 -5.90612233e-01 5.99024706e-02
3.58684391e-01 -7.16343403e-01 6.07694864e-01 -8.78165007e-01
8.02985132e-01 -3.91070008e-01 -1.93416685e-01 -1.02567279e+00
-2.25590151e-02 -6.62471116e-01 -4.43564415e-01 1.37080324e+00
9.37725127e-01 -7.29366362e-01 4.52264041e-01 1.07914340e+00
-5.88406801e-01 -9.70341682e-01 -7.91543961e-01 -5.44890225e-01
4.90385443e-02 -8.53573442e-01 8.43827724e-01 8.55389953e-01
1.15803383e-01 4.74196315e-01 1.79507330e-01 2.39100829e-01
1.20800994e-01 3.56758475e-01 8.48003805e-01 -1.09684479e+00
-1.02625549e+00 -4.45077419e-01 3.37943435e-01 -9.85157132e-01
6.40493512e-01 -1.46016502e+00 1.90866441e-01 -1.38130295e+00
8.13919380e-02 -5.21042049e-01 1.87971741e-01 1.41958690e+00
-3.03300321e-01 -4.45208013e-01 6.27822056e-02 1.13554686e-01
-4.23956484e-01 -6.54192343e-02 6.62351787e-01 -1.83178112e-01
-2.97905833e-01 -7.29836598e-02 -8.84730101e-01 6.68995738e-01
8.12776566e-01 -6.25358343e-01 -1.05401948e-01 -9.32745874e-01
7.08588302e-01 2.86743641e-01 6.43261254e-01 -8.26084733e-01
2.77582735e-01 -1.72818795e-01 5.67037128e-02 -1.06445581e-01
-6.66326210e-02 -5.16672075e-01 1.53011754e-01 2.72016406e-01
-6.27558410e-01 2.00459316e-01 5.08413672e-01 2.35110559e-02
-5.77725051e-03 -7.87409604e-01 4.80665416e-01 -3.38344544e-01
-6.05557621e-01 -2.73980610e-02 -1.62884295e-02 4.44240719e-01
5.99776566e-01 2.88243517e-02 -7.35496402e-01 1.34151191e-01
-6.32875502e-01 4.41112787e-01 5.61892807e-01 2.60208279e-01
2.74060518e-01 -6.45996630e-01 -5.50341904e-01 1.31660596e-01
3.07359546e-01 4.37271059e-01 -2.32922863e-02 6.46734595e-01
-6.72347903e-01 3.83528858e-01 7.09086284e-02 -5.29049754e-01
-9.75808561e-01 6.09941185e-01 5.38537621e-01 -4.86031264e-01
-3.62290740e-01 9.89533305e-01 8.63105580e-02 -1.21717930e+00
1.52898863e-01 -8.06625187e-01 9.62225944e-02 -3.93584698e-01
6.33163750e-01 -1.03655197e-01 3.56066167e-01 1.60441443e-01
-5.05433679e-01 7.39159584e-02 4.19605859e-02 7.48824477e-02
1.32383096e+00 6.24125779e-01 -4.69403982e-01 3.70079100e-01
9.07731891e-01 -1.28866592e-02 -9.57086027e-01 -3.35650668e-02
8.94235075e-02 3.16409394e-02 -2.42987469e-01 -1.60605252e+00
-7.44272947e-01 1.01686716e+00 -1.68107197e-01 2.49842539e-01
1.05506265e+00 1.74771786e-01 5.42962193e-01 8.19673598e-01
3.16251040e-01 -6.54592335e-01 -3.12138647e-01 9.41205800e-01
1.08408117e+00 -1.00035131e+00 -5.08465588e-01 -3.38860184e-01
-4.63321060e-01 1.29292750e+00 9.84465003e-01 8.89021009e-02
-7.35892206e-02 5.70659578e-01 -2.23158047e-01 -9.19054821e-02
-1.19843352e+00 1.97998315e-01 9.79793891e-02 5.26150942e-01
5.37007570e-01 1.07578434e-01 1.15182295e-01 1.40173507e+00
-5.62178910e-01 8.77268240e-02 3.68851244e-01 7.01291859e-01
-2.61388928e-01 -1.07805359e+00 -2.06154943e-01 5.43959022e-01
-5.32339513e-01 -6.68110430e-01 -3.41065109e-01 8.69607925e-01
1.71683088e-01 6.73767090e-01 -1.52091593e-01 -2.35486403e-01
5.28818011e-01 5.85765302e-01 2.65458465e-01 -1.03203297e+00
-1.15016937e+00 -3.14636827e-01 4.52325612e-01 -8.36965382e-01
5.27208865e-01 -3.86584818e-01 -1.61614096e+00 -1.00768328e-01
-5.04733585e-02 6.19390383e-02 5.75107813e-01 9.53953803e-01
7.08618522e-01 5.77316821e-01 -1.14363089e-01 -3.69182169e-01
-8.31585944e-01 -1.02841926e+00 4.09117602e-02 2.27939129e-01
1.21767938e-01 7.24876663e-05 -3.06962401e-01 2.39190087e-01] | [9.626266479492188, 7.63092041015625] |
391c22c5-5eb8-4ac7-8856-00832894570c | heterogeneous-graph-transformer-for-graph-to | null | null | https://aclanthology.org/2020.acl-main.640 | https://aclanthology.org/2020.acl-main.640.pdf | Heterogeneous Graph Transformer for Graph-to-Sequence Learning | The graph-to-sequence (Graph2Seq) learning aims to transduce graph-structured representations to word sequences for text generation. Recent studies propose various models to encode graph structure. However, most previous works ignore the indirect relations between distance nodes, or treat indirect relations and direct relations in the same way. In this paper, we propose the Heterogeneous Graph Transformer to independently model the different relations in the individual subgraphs of the original graph, including direct relations, indirect relations and multiple possible relations between nodes. Experimental results show that our model strongly outperforms the state of the art on all four standard benchmarks of AMR-to-text generation and syntax-based neural machine translation. | ['Xiaojun Wan', 'Tianming Wang', 'Shaowei Yao'] | 2020-07-01 | null | null | null | acl-2020-6 | ['graph-to-sequence'] | ['natural-language-processing'] | [ 4.53244448e-01 7.01645613e-01 -3.78039241e-01 -2.75287569e-01
-4.17632580e-01 -6.90262616e-01 8.58924091e-01 1.47473723e-01
4.73508686e-02 1.12098265e+00 6.54089987e-01 -8.57577682e-01
2.62733698e-01 -1.43375564e+00 -7.16198623e-01 -3.17904860e-01
6.10387735e-02 8.59209538e-01 9.16526914e-02 -8.12030554e-01
-9.59968269e-02 1.06034026e-01 -7.20772147e-01 4.51992124e-01
9.76790726e-01 2.72883713e-01 -9.48538557e-02 8.78591061e-01
-6.56628966e-01 9.49554980e-01 -8.11445475e-01 -9.34856355e-01
1.14725074e-02 -1.09922528e+00 -9.42614377e-01 -2.85093188e-01
1.27604946e-01 -1.02728099e-01 -5.91761112e-01 1.06748796e+00
5.67162931e-01 2.45172456e-02 7.19717920e-01 -1.14880610e+00
-1.27038622e+00 1.42403591e+00 -3.74742836e-01 1.20509893e-01
7.39181995e-01 9.98999774e-02 1.49353850e+00 -6.29058957e-01
8.55706215e-01 1.62766004e+00 2.73583084e-01 7.42805004e-01
-1.16255975e+00 -3.79281938e-01 3.77416670e-01 8.23236108e-02
-1.26321650e+00 -2.26707116e-01 6.40064001e-01 -9.32087004e-02
1.56528747e+00 2.48736441e-01 6.11759543e-01 1.31220627e+00
4.11195993e-01 3.65415514e-01 6.74955904e-01 -5.64049959e-01
-1.39540702e-01 -4.29053128e-01 1.96511909e-01 9.77905929e-01
4.18097734e-01 4.16021571e-02 -2.90253639e-01 -2.00698704e-01
7.78559029e-01 -6.82253361e-01 -1.98338822e-01 1.07811235e-01
-1.38132226e+00 1.20423305e+00 6.34310007e-01 2.36682504e-01
-9.00169834e-02 3.66257995e-01 2.87726283e-01 4.33149755e-01
5.72030783e-01 5.72306693e-01 -3.16213250e-01 1.61402285e-01
-1.13083541e-01 2.18748122e-01 1.02420759e+00 1.30757594e+00
7.41798878e-01 4.86998051e-01 -6.14084482e-01 6.65647149e-01
3.62805754e-01 4.08209652e-01 4.71800119e-01 -3.19495469e-01
1.03474045e+00 7.37526834e-01 -3.91584277e-01 -1.09921873e+00
-1.99400678e-01 -2.30716571e-01 -1.16357088e+00 -6.00614965e-01
7.57473186e-02 -5.99149704e-01 -1.05316973e+00 1.69820678e+00
1.48044541e-01 1.33224621e-01 1.18087657e-01 6.13434792e-01
1.31247807e+00 1.04072905e+00 5.80813810e-02 -7.24350438e-02
9.49273407e-01 -1.04355335e+00 -8.35672557e-01 -5.34651816e-01
9.94386554e-01 -6.18539572e-01 8.87699008e-01 -2.11422384e-01
-1.17294073e+00 -4.63371664e-01 -7.16066241e-01 -2.76872627e-02
-4.22468543e-01 -2.73654103e-01 8.62354040e-01 2.97804505e-01
-1.35072553e+00 5.20165324e-01 -3.07667017e-01 -1.89599186e-01
-1.45387083e-01 3.64183366e-01 -2.33897761e-01 -7.37125203e-02
-2.04473972e+00 8.30384076e-01 7.19394922e-01 2.56808519e-01
-4.49662894e-01 -3.20763201e-01 -1.25538611e+00 -5.97902685e-02
2.53956676e-01 -1.28599465e+00 1.06201601e+00 -8.14380646e-01
-1.53651524e+00 6.19794011e-01 -3.37791562e-01 -5.45731723e-01
7.10875317e-02 1.10468447e-01 -3.52007866e-01 -1.31208271e-01
-1.83175340e-01 6.67553663e-01 4.54212844e-01 -9.11663949e-01
-1.04427986e-01 -2.15342594e-03 3.30789268e-01 4.04948115e-01
-1.01603884e-02 -2.24050530e-03 -3.64215344e-01 -8.44981968e-01
-2.07587361e-01 -9.64052320e-01 -4.95965719e-01 -7.70613968e-01
-9.93672848e-01 -5.67336500e-01 2.21642807e-01 -6.50417805e-01
1.35589778e+00 -1.30387223e+00 6.91486359e-01 9.56387073e-02
2.30805904e-01 3.23996514e-01 -6.84178174e-01 8.88184726e-01
-1.75659910e-01 5.48903525e-01 -1.13150135e-01 8.88122395e-02
-1.02993669e-02 4.49423254e-01 -3.07334453e-01 -1.82430133e-01
3.73151332e-01 1.75239408e+00 -1.25936687e+00 -6.04369640e-01
-5.81196621e-02 3.99134308e-01 -3.52610976e-01 4.17045087e-01
-6.64360940e-01 3.29513788e-01 -6.28780961e-01 1.21736586e-01
4.04202342e-01 -3.78548056e-01 6.53245926e-01 7.10428879e-02
5.33117354e-01 9.46229458e-01 -6.56126559e-01 1.71396458e+00
-4.01930839e-01 3.83280814e-01 -3.18595260e-01 -9.00674820e-01
1.03169155e+00 3.21731597e-01 -9.03704017e-02 -6.08690023e-01
3.46049964e-02 -4.15475555e-02 4.23983306e-01 -1.57779157e-01
5.57898581e-01 -9.68268067e-02 -6.67359829e-02 3.93273890e-01
1.77155063e-01 -4.13142592e-01 3.97105724e-01 6.96522951e-01
1.08833838e+00 2.35875741e-01 5.21962523e-01 7.16616958e-02
4.86311346e-01 -2.87946731e-01 2.27639169e-01 7.10082710e-01
3.85553688e-01 5.05780399e-01 9.36351776e-01 -2.11686492e-01
-7.40037739e-01 -1.15972364e+00 8.02660227e-01 9.73327994e-01
-1.53975770e-01 -7.69171178e-01 -6.85482681e-01 -1.00151992e+00
-3.38660926e-01 9.85703409e-01 -5.33033252e-01 -4.50809896e-01
-1.04342556e+00 -7.13713288e-01 8.51609766e-01 3.70182604e-01
7.38980174e-02 -1.26767361e+00 5.71145356e-01 2.58434564e-01
-5.21786094e-01 -1.24915016e+00 -7.68752694e-01 -2.16621429e-01
-9.98135269e-01 -7.51481593e-01 -4.48591679e-01 -9.99856412e-01
7.46875584e-01 4.37865919e-03 1.65224135e+00 3.20598483e-01
1.16145439e-01 -1.10814556e-01 -5.48968792e-01 -5.38965650e-02
-8.58661890e-01 5.34201682e-01 -4.57310706e-01 -2.16403201e-01
-4.98709865e-02 -4.76973772e-01 -2.26385906e-01 -4.66437824e-02
-7.00349450e-01 3.90096545e-01 5.49904346e-01 8.33300412e-01
4.50353682e-01 -3.84546041e-01 6.95550025e-01 -1.28736210e+00
1.13336813e+00 -5.71580350e-01 -3.55200976e-01 4.55761671e-01
-5.90125680e-01 3.94500852e-01 9.55293775e-01 -2.47273728e-01
-1.00961375e+00 -1.42530233e-01 -1.84826732e-01 3.35707213e-03
1.47115551e-02 9.01681125e-01 -4.02135044e-01 2.98230499e-01
5.76451004e-01 3.74595702e-01 -3.28509718e-01 2.61882506e-02
1.00523782e+00 2.37426117e-01 1.10953562e-01 -6.33308828e-01
9.60894644e-01 -2.19464019e-01 2.36937493e-01 -5.56693017e-01
-7.57537723e-01 5.98417670e-02 -6.71164930e-01 1.27104059e-01
9.33216095e-01 -7.94410050e-01 -1.64987698e-01 1.27195150e-01
-1.75646281e+00 -4.88112897e-01 -2.19098687e-01 2.17588425e-01
-2.10442290e-01 5.05743682e-01 -9.42600727e-01 -4.66356397e-01
-7.45909214e-01 -9.24274325e-01 1.08881402e+00 1.88570507e-02
-3.43671888e-01 -1.52256167e+00 2.74298221e-01 1.60809636e-01
3.39708149e-01 2.83348829e-01 1.36787868e+00 -5.97877741e-01
-6.24421179e-01 3.80868018e-02 -1.63528785e-01 1.65420607e-01
3.67377102e-01 -8.62926990e-02 -2.56013632e-01 -1.41123179e-02
-8.24826837e-01 -1.73535228e-01 9.15929139e-01 1.77620187e-01
7.13215888e-01 -7.24193752e-01 -3.17160487e-01 5.03445625e-01
9.70211685e-01 2.67473161e-02 9.65264738e-01 -2.26845652e-01
1.28828990e+00 5.87794781e-01 5.41443266e-02 -6.49583265e-02
5.30527174e-01 5.55603862e-01 4.46025729e-01 -2.12222695e-01
-3.87156606e-01 -7.68673956e-01 6.47311211e-01 1.43893862e+00
-6.48503155e-02 -1.02973509e+00 -7.01633334e-01 3.75627697e-01
-1.77512312e+00 -7.61440217e-01 -7.33497381e-01 1.69370258e+00
9.70131159e-01 1.14827462e-01 -1.73789412e-01 -3.70219827e-01
1.00980389e+00 6.04270160e-01 -6.37808666e-02 -9.09737885e-01
-3.79095525e-01 5.64111173e-01 3.75133485e-01 8.71845901e-01
-5.67850947e-01 1.50318241e+00 6.69436836e+00 6.30190372e-01
-6.00204945e-01 -2.83038616e-01 6.36425376e-01 2.18898132e-01
-9.37124133e-01 2.00754479e-01 -6.45271420e-01 9.09230560e-02
1.10239303e+00 -4.68010068e-01 7.37948835e-01 3.54049355e-01
-7.69085065e-03 6.00209534e-01 -1.15810323e+00 5.48603058e-01
1.54370189e-01 -1.37007189e+00 9.80367303e-01 -6.83544129e-02
9.37161803e-01 -5.17820604e-02 -2.16188699e-01 4.30905789e-01
1.02028477e+00 -1.36338711e+00 3.63629550e-01 4.35657173e-01
7.42480576e-01 -7.19767153e-01 7.39482880e-01 2.58405924e-01
-1.41528594e+00 5.27958393e-01 -3.72928858e-01 -3.94622087e-01
3.41087341e-01 6.87956393e-01 -1.18112195e+00 1.20177281e+00
-2.38978669e-01 6.65605307e-01 -6.13196969e-01 1.91857219e-01
-9.34866548e-01 8.67783308e-01 2.68770933e-01 -5.43335676e-01
3.19025964e-01 -5.94562292e-01 5.97312629e-01 1.28685391e+00
3.47871780e-01 1.01715356e-01 1.27315357e-01 9.57173765e-01
-5.17308652e-01 2.09353641e-01 -1.06075060e+00 -4.96609062e-01
2.57489204e-01 1.17834318e+00 -4.78638589e-01 -4.25199568e-01
-3.68256629e-01 1.17260551e+00 6.83406591e-01 5.19849658e-01
-7.15112567e-01 -5.26008546e-01 4.53725815e-01 5.29547259e-02
1.38128147e-01 -4.66019481e-01 5.42724766e-02 -1.32079923e+00
-2.04536960e-01 -1.16450143e+00 6.04644120e-01 -8.83962810e-01
-1.40960491e+00 8.30060363e-01 7.45229125e-02 -6.00691319e-01
-7.73415387e-01 -4.37564582e-01 -8.38610172e-01 1.18103707e+00
-1.28159547e+00 -1.19913483e+00 1.46959901e-01 5.42747259e-01
5.04117310e-01 -1.29924491e-01 1.10177565e+00 -1.12634450e-01
-3.75739098e-01 5.89804590e-01 -4.10148650e-01 6.08346343e-01
4.15728986e-01 -1.27595878e+00 1.38777232e+00 9.08695817e-01
5.81438601e-01 8.83682787e-01 3.56056213e-01 -1.00180769e+00
-1.66100788e+00 -1.43232369e+00 1.40983117e+00 -3.56314272e-01
7.86727965e-01 -6.06556058e-01 -7.43268073e-01 1.02372944e+00
6.52999878e-01 -1.23038009e-01 4.64706302e-01 4.49607447e-02
-4.69945312e-01 3.00113320e-01 -4.34036732e-01 9.73760247e-01
1.52149808e+00 -4.76852536e-01 -5.38771689e-01 5.63764095e-01
1.46225226e+00 -5.16260326e-01 -5.98500311e-01 2.28611887e-01
3.24211419e-02 -4.75749284e-01 8.79401505e-01 -1.15543723e+00
8.29275846e-01 -1.28269464e-01 5.72079904e-02 -1.92384064e+00
-4.66081858e-01 -9.69855785e-01 -4.00321633e-01 1.23018360e+00
1.01143885e+00 -7.83074081e-01 6.71224058e-01 -1.80147037e-01
-1.33294135e-01 -7.04713464e-01 -8.08527648e-01 -5.88462532e-01
3.30744117e-01 -3.36412415e-02 8.94366324e-01 1.05066252e+00
-1.14014838e-02 1.32894480e+00 -5.11339545e-01 -1.30059123e-01
3.27994198e-01 1.21675827e-01 6.75764084e-01 -1.05216956e+00
-7.03516781e-01 -3.84378672e-01 -1.84038758e-01 -1.11568725e+00
5.73739529e-01 -1.57431459e+00 -8.46573263e-02 -2.44430804e+00
-5.51434085e-02 -1.07949346e-01 -4.63430956e-03 3.47377002e-01
-6.60601795e-01 -1.47813454e-01 3.43550369e-02 -3.29967052e-01
-4.90363955e-01 7.69529521e-01 1.71133888e+00 -4.66737896e-01
-1.41490176e-01 -2.39154592e-01 -8.81472409e-01 3.02185327e-01
1.00235772e+00 -4.09005493e-01 -8.21647644e-01 -9.18063045e-01
6.52475119e-01 3.24462652e-01 9.54224076e-03 -2.44485319e-01
-9.31423753e-02 -4.21628118e-01 3.21644247e-02 -2.09900588e-01
6.67273924e-02 -1.12413935e-01 1.15698196e-01 4.86563951e-01
-7.83813357e-01 3.44760060e-01 -8.12511295e-02 6.10970616e-01
-1.54084086e-01 -7.09660351e-02 2.05924794e-01 -4.62237954e-01
-1.42738804e-01 4.63402420e-01 -2.73390710e-01 5.17219067e-01
5.53701222e-01 3.32277060e-01 -5.83595812e-01 -8.86889935e-01
-5.52657068e-01 9.05021802e-02 6.26675179e-03 7.46210873e-01
7.71089077e-01 -1.42752874e+00 -1.25944042e+00 -1.57224983e-01
-5.79628348e-02 7.69751593e-02 -3.66723061e-01 3.34612578e-01
-3.20294112e-01 5.13619840e-01 9.82355922e-02 -1.76550783e-02
-1.09319162e+00 5.57933450e-01 3.14586788e-01 -9.03625488e-01
-3.11091870e-01 9.28047955e-01 2.69459039e-01 -7.78237998e-01
-1.98486894e-01 -3.41605484e-01 -3.43424290e-01 -2.65808821e-01
2.19187036e-01 2.13174954e-01 -7.50732571e-02 -7.67910779e-01
-9.91097018e-02 3.24030727e-01 -5.60440868e-02 -1.17592186e-01
7.86192477e-01 1.98402286e-01 -5.72853744e-01 3.69467735e-01
1.08871126e+00 -2.48938100e-03 -3.06245446e-01 -3.03400427e-01
-2.50643250e-02 -6.58511668e-02 -4.08580244e-01 -6.06263518e-01
-1.00102997e+00 9.83219624e-01 -3.52838010e-01 2.92194635e-01
7.18088031e-01 2.82897223e-02 9.80325460e-01 5.07105112e-01
2.75953829e-01 -5.81883907e-01 1.73187599e-01 1.04691565e+00
1.02537894e+00 -7.80467510e-01 -6.20122515e-02 -8.37234378e-01
-7.69094706e-01 9.76723731e-01 7.10373223e-01 -7.29122758e-02
1.90251380e-01 1.69596463e-01 -1.68040156e-01 5.39475828e-02
-1.12268174e+00 -2.36450985e-01 6.43729627e-01 8.50585580e-01
9.00128961e-01 3.13775003e-01 -5.85044622e-01 2.72552669e-01
-3.95932645e-01 -4.20997590e-01 4.88802671e-01 4.68116045e-01
-1.10352762e-01 -1.65766168e+00 6.36720508e-02 4.93914992e-01
-3.27304721e-01 -8.53402257e-01 -1.20445478e+00 6.56815529e-01
-2.01130480e-01 1.24972153e+00 -1.11244224e-01 -6.24910295e-01
4.00606930e-01 1.41052842e-01 6.54983580e-01 -9.83400583e-01
-7.76975513e-01 -2.11332470e-01 7.27172852e-01 -2.94580936e-01
-5.00217415e-02 -1.72597870e-01 -1.50067878e+00 -2.82985300e-01
-4.14664000e-01 3.34920883e-01 2.29373157e-01 6.53718054e-01
5.23369610e-01 9.67609644e-01 5.42936444e-01 -4.30124104e-01
-4.48129714e-01 -1.20434475e+00 -1.74341336e-01 3.60566437e-01
-1.98925380e-02 -3.39884534e-02 -1.86931863e-01 1.74476914e-02] | [10.2379732131958, 8.319182395935059] |
7bba5a64-6211-4fe0-abd1-e0989611146e | mc-bert-efficient-language-pre-training-via-a | 2006.05744 | null | https://arxiv.org/abs/2006.05744v2 | https://arxiv.org/pdf/2006.05744v2.pdf | MC-BERT: Efficient Language Pre-Training via a Meta Controller | Pre-trained contextual representations (e.g., BERT) have become the foundation to achieve state-of-the-art results on many NLP tasks. However, large-scale pre-training is computationally expensive. ELECTRA, an early attempt to accelerate pre-training, trains a discriminative model that predicts whether each input token was replaced by a generator. Our studies reveal that ELECTRA's success is mainly due to its reduced complexity of the pre-training task: the binary classification (replaced token detection) is more efficient to learn than the generation task (masked language modeling). However, such a simplified task is less semantically informative. To achieve better efficiency and effectiveness, we propose a novel meta-learning framework, MC-BERT. The pre-training task is a multi-choice cloze test with a reject option, where a meta controller network provides training input and candidates. Results over GLUE natural language understanding benchmark demonstrate that our proposed method is both efficient and effective: it outperforms baselines on GLUE semantic tasks given the same computational budget. | ['Li-Wei Wang', 'Tie-Yan Liu', 'Shuxin Zheng', 'Zhenhui Xu', 'Linyuan Gong', 'Guolin Ke', 'Di He', 'Jiang Bian'] | 2020-06-10 | null | null | null | null | ['cloze-test'] | ['natural-language-processing'] | [ 3.02767068e-01 2.04532459e-01 -4.07396913e-01 -5.35067320e-01
-1.17977679e+00 -4.01282519e-01 7.23043859e-01 1.74088359e-01
-5.11225581e-01 6.44546032e-01 2.50020236e-01 -1.86465129e-01
5.18058956e-01 -1.01808059e+00 -1.15052760e+00 -4.73017186e-01
3.04043472e-01 7.18359768e-01 1.20306462e-01 -7.93442130e-02
2.39062443e-01 -2.31886744e-01 -1.71260929e+00 9.12342489e-01
1.15789413e+00 8.75758231e-01 6.75094664e-01 2.50414073e-01
-1.19023539e-01 6.08698189e-01 -7.21724331e-01 -5.23933411e-01
-3.77050266e-02 -5.29229879e-01 -9.85409439e-01 -3.36568117e-01
1.20761082e-01 -2.27265030e-01 1.54177889e-01 8.51657987e-01
5.02603829e-01 1.24394856e-01 6.73162460e-01 -1.16289282e+00
-8.92466724e-01 1.34882557e+00 -1.59716815e-01 -2.77568907e-01
2.53466606e-01 -6.52850047e-02 1.47070146e+00 -1.23178577e+00
5.63346446e-01 1.45322442e+00 4.63803291e-01 8.28160226e-01
-1.29880524e+00 -5.87025702e-01 2.65503526e-01 3.60035360e-01
-1.25501406e+00 -3.45581114e-01 6.01678193e-01 -4.27588373e-02
1.22302794e+00 -1.10892303e-01 2.84971178e-01 1.58957410e+00
-4.90080975e-02 1.23793232e+00 1.05783272e+00 -8.04741383e-01
4.50138241e-01 1.15354441e-01 8.78775865e-02 7.50032306e-01
1.50509238e-01 1.11633964e-01 -6.89772785e-01 -6.09208904e-02
3.52214247e-01 -3.04467499e-01 -3.11486870e-02 -1.94565192e-01
-1.00432146e+00 9.67556894e-01 4.04615164e-01 2.73902386e-01
-1.21949650e-01 3.15863729e-01 3.70337248e-01 1.84355468e-01
4.21445340e-01 6.87678933e-01 -4.86259639e-01 -2.32961625e-01
-9.94006395e-01 3.28731775e-01 7.05151916e-01 1.08025527e+00
7.95323789e-01 -6.69347569e-02 -2.11242929e-01 1.11723268e+00
1.70126364e-01 2.01690495e-01 7.16621101e-01 -7.68134892e-01
7.14112282e-01 6.43174410e-01 -9.21058506e-02 -5.12949288e-01
-1.89530388e-01 -4.96214181e-01 -5.98041236e-01 5.64583093e-02
3.81507725e-01 -3.95736238e-03 -9.37265754e-01 2.06789088e+00
7.76230618e-02 3.03170443e-01 2.48833790e-01 8.29338849e-01
6.81984961e-01 8.17911923e-01 4.66628462e-01 2.17480585e-01
1.42574561e+00 -1.45454466e+00 -4.26472485e-01 -7.10594475e-01
9.70915318e-01 -6.59878135e-01 1.35733533e+00 3.19143295e-01
-1.06478179e+00 -6.17900431e-01 -1.11115301e+00 -4.36429024e-01
-5.83038032e-01 5.34252286e-01 9.59697604e-01 5.60241759e-01
-7.99552083e-01 6.42870963e-01 -6.40132010e-01 -2.48589531e-01
4.61121917e-01 5.36798909e-02 -3.10777366e-01 -2.44673416e-01
-1.33659399e+00 8.38253140e-01 7.04486012e-01 -3.97251509e-02
-1.10910964e+00 -7.58628726e-01 -9.96280730e-01 4.18614030e-01
6.99864626e-01 -8.39231849e-01 1.23833454e+00 -9.73535001e-01
-1.64215541e+00 1.06108499e+00 -2.95120716e-01 -5.31811059e-01
4.59570527e-01 -6.55686796e-01 -7.33668730e-02 -1.46936283e-01
3.87308121e-01 1.07393813e+00 7.85795331e-01 -1.21315587e+00
-6.23001695e-01 6.55874088e-02 1.15538627e-01 1.57428622e-01
-4.25767004e-02 -4.53894474e-02 -3.92339945e-01 -8.98073435e-01
-1.02533221e-01 -8.15749645e-01 -3.62869166e-02 -5.47922611e-01
-4.03103918e-01 -7.27446496e-01 2.92042583e-01 -5.14063716e-01
1.15778327e+00 -2.10049963e+00 1.52726039e-01 -5.30415960e-02
-2.82896429e-01 3.00940394e-01 -4.93011326e-01 4.22160119e-01
-1.10229239e-01 3.47107142e-01 -2.39585817e-01 -7.20792055e-01
8.21804851e-02 1.18578166e-01 -5.80341518e-01 -1.70187324e-01
4.29600745e-01 9.23420310e-01 -9.39939857e-01 -3.63276094e-01
1.48146302e-01 2.06595868e-01 -9.70568955e-01 3.66976470e-01
-7.91936517e-01 1.88288927e-01 -1.40116960e-01 8.81146669e-01
4.14936721e-01 -3.37658077e-01 1.89232215e-01 -5.72739914e-02
1.71626523e-01 6.91458166e-01 -9.29191351e-01 2.20497274e+00
-8.69042337e-01 2.16704398e-01 -2.57615536e-01 -1.12883055e+00
7.75348067e-01 1.97790608e-01 -4.83058058e-02 -8.15160632e-01
-6.07313495e-03 3.95817995e-01 -1.40585080e-01 -4.10252959e-01
6.77402139e-01 1.55767739e-01 -4.18614388e-01 5.06619811e-01
2.48405740e-01 -6.43547997e-02 3.85104477e-01 3.64816308e-01
9.70147669e-01 7.73367584e-01 3.07193846e-01 -2.04015210e-01
2.97572881e-01 -2.84060501e-02 7.96063423e-01 1.03006709e+00
2.61785418e-01 6.22936845e-01 5.51385641e-01 -2.98647970e-01
-9.84639645e-01 -9.33351398e-01 1.80864796e-01 1.55898917e+00
1.96435347e-01 -8.34981143e-01 -1.01118875e+00 -9.96654689e-01
-1.32772461e-01 1.10557032e+00 -5.61996460e-01 -4.11487460e-01
-8.56094122e-01 -8.56750488e-01 5.12821317e-01 7.47914195e-01
5.05148649e-01 -1.41335607e+00 -6.76134586e-01 3.38320941e-01
-4.07971710e-01 -1.03765130e+00 -1.58750311e-01 3.64908218e-01
-7.28045940e-01 -8.78364742e-01 -3.55224103e-01 -1.00461483e+00
8.28174889e-01 6.01137765e-02 1.44151568e+00 1.13197647e-01
-3.46396834e-01 -1.81171045e-01 -6.59138262e-01 -2.59369791e-01
-4.20722723e-01 3.40247422e-01 -3.57406974e-01 -2.92151541e-01
4.16806012e-01 -4.26921219e-01 -5.16972363e-01 3.08761030e-01
-8.01905096e-01 5.04315376e-01 8.97064507e-01 1.22462833e+00
6.03686810e-01 -2.65054226e-01 7.22686648e-01 -1.18175077e+00
5.90891719e-01 -4.90351617e-01 -3.93889874e-01 6.60756588e-01
-5.87952316e-01 2.96271801e-01 7.42283642e-01 -3.50230604e-01
-1.29127979e+00 2.83886734e-02 -1.06672257e-01 4.32297867e-03
-1.05223224e-01 5.68311632e-01 -4.23404306e-01 4.92212027e-01
5.46812773e-01 3.89742516e-02 -4.86280859e-01 -6.34665489e-01
6.84771180e-01 4.13372368e-01 5.66488862e-01 -1.20614433e+00
2.71095306e-01 1.41399279e-01 -4.10226554e-01 -4.99064356e-01
-1.16491508e+00 -2.39820555e-01 -4.07252967e-01 2.54061818e-01
7.26375878e-01 -1.23224247e+00 -3.57171774e-01 3.49739611e-01
-1.31809199e+00 -4.84465331e-01 -1.33487999e-01 3.20199162e-01
-5.39630771e-01 4.52887639e-03 -5.33449113e-01 -6.09573722e-01
-3.36220324e-01 -1.17521322e+00 1.23298466e+00 1.02230661e-01
-4.67204779e-01 -7.18479633e-01 -1.12988435e-01 6.47826433e-01
3.27662498e-01 -7.20883235e-02 1.52843499e+00 -6.97971404e-01
-6.87444568e-01 5.83392829e-02 -1.74753487e-01 4.16237116e-01
-1.36771470e-01 -1.43956006e-01 -1.08175719e+00 -1.02143213e-01
-1.84384525e-01 -8.84688854e-01 1.30708051e+00 1.70795709e-01
1.40638471e+00 -1.55743122e-01 -4.14701313e-01 5.92249990e-01
1.34386349e+00 3.94609310e-02 5.51268458e-01 3.76520425e-01
5.00217795e-01 5.28634250e-01 7.65797198e-01 2.59930849e-01
5.24715424e-01 6.09344721e-01 4.07169104e-01 1.01529799e-01
-3.34299445e-01 -6.13383353e-01 6.26758695e-01 6.01147413e-01
3.19802016e-02 -4.07242328e-01 -7.96554804e-01 6.34603381e-01
-2.14741254e+00 -7.33899236e-01 3.84943217e-01 2.07735133e+00
1.10297084e+00 1.82371721e-01 -2.20769033e-01 4.72142845e-02
6.96977437e-01 7.88310021e-02 -4.49983090e-01 -2.94778466e-01
-2.77782619e-01 5.54879904e-01 -9.84550118e-02 5.02857566e-01
-1.17777789e+00 1.73004556e+00 5.61044550e+00 1.09425199e+00
-9.16667938e-01 2.98014373e-01 7.86450326e-01 -5.06191701e-02
-4.00243133e-01 2.06368148e-01 -9.92947280e-01 5.21705151e-01
9.24867094e-01 1.61605448e-01 3.34220231e-01 1.38890684e+00
-1.04874626e-01 -4.31071192e-01 -1.68649054e+00 8.38099241e-01
3.35755438e-01 -1.48371983e+00 3.84109706e-01 -3.64820659e-01
6.69442832e-01 -2.36032471e-01 5.57941198e-02 9.23294365e-01
4.44575071e-01 -1.09358335e+00 1.00923097e+00 1.64324135e-01
6.26056492e-01 -6.82289362e-01 7.20068753e-01 2.91770369e-01
-9.46577072e-01 -7.56587321e-03 -5.83812237e-01 -6.11790493e-02
2.17935368e-01 7.14036405e-01 -8.64441931e-01 4.61132616e-01
5.85681379e-01 4.81570929e-01 -5.39329946e-01 7.47002065e-01
-8.88445616e-01 6.72502279e-01 -1.62386581e-01 -1.61745146e-01
2.29012758e-01 4.01820615e-03 1.95551068e-01 1.39857936e+00
2.28415519e-01 -1.79157093e-01 2.53768176e-01 1.20492971e+00
-5.85777521e-01 8.41066614e-02 -2.79520124e-01 -1.53772041e-01
7.37691402e-01 1.12688422e+00 -6.01471603e-01 -4.75927293e-01
-2.86343664e-01 1.03475237e+00 8.15771401e-01 2.65266329e-01
-9.28278387e-01 -3.77713054e-01 4.80023295e-01 -1.28161341e-01
3.49408537e-01 1.71286948e-02 -4.76695061e-01 -1.31181610e+00
6.70329407e-02 -9.12128985e-01 4.10470098e-01 -8.17737103e-01
-1.13623881e+00 3.96824181e-01 -2.78522875e-02 -7.29786217e-01
-5.15971005e-01 -6.92210317e-01 -7.12330282e-01 6.87594652e-01
-1.50730813e+00 -1.44649708e+00 -2.80913431e-02 2.87561595e-01
8.32355201e-01 -1.56741619e-01 1.02306128e+00 2.15160310e-01
-5.43602169e-01 7.45665193e-01 -4.17165905e-01 2.52758086e-01
8.51060867e-01 -1.32330847e+00 3.87908101e-01 9.53030705e-01
3.56178164e-01 7.74217486e-01 3.91488254e-01 -7.17058659e-01
-1.20160699e+00 -1.03109384e+00 1.14722490e+00 -2.59302586e-01
5.29352307e-01 -7.74931312e-01 -6.92550182e-01 7.10700929e-01
2.26431876e-01 -1.13008983e-01 6.93951547e-01 3.19411933e-01
-5.86569786e-01 2.59193070e-02 -8.03470731e-01 5.84448457e-01
1.19180346e+00 -6.02134883e-01 -1.00777853e+00 2.81522036e-01
8.52855861e-01 -3.83315384e-01 -3.11586946e-01 3.99657130e-01
3.13058704e-01 -8.65367770e-01 8.47257078e-01 -6.21324122e-01
9.03166354e-01 -8.64681900e-02 -1.58173203e-01 -1.49305630e+00
-1.12889320e-01 -5.85116148e-01 -5.61789125e-02 1.52875984e+00
7.38011777e-01 -2.51566231e-01 8.45327020e-01 4.69780535e-01
-3.49773318e-01 -9.11525965e-01 -7.99465418e-01 -7.28487670e-01
9.58738551e-02 -5.81622481e-01 4.95119005e-01 7.24010706e-01
5.91024719e-02 6.28730178e-01 -2.70868629e-01 5.60331121e-02
3.91822278e-01 3.41643542e-01 5.99567711e-01 -1.01419187e+00
-4.40075934e-01 -2.98600256e-01 1.70839906e-01 -1.10501575e+00
6.64631426e-01 -1.18257666e+00 2.22368553e-01 -1.34176683e+00
2.39749566e-01 -7.06148505e-01 -2.85522491e-01 8.06762397e-01
-3.75213921e-01 -1.53829262e-01 7.50872418e-02 8.17910284e-02
-6.23048604e-01 6.69110060e-01 7.20402241e-01 -1.59823447e-01
-1.46621287e-01 -1.75266892e-01 -7.72007883e-01 7.87232578e-01
7.00818181e-01 -5.80729425e-01 -3.65103841e-01 -8.48896563e-01
5.22315025e-01 -2.75414735e-01 3.00320178e-01 -9.25698698e-01
1.68119773e-01 1.16526090e-01 1.24365598e-01 -4.98097301e-01
3.68954360e-01 -4.13813859e-01 -4.26779091e-01 1.93880245e-01
-6.54374480e-01 7.57444948e-02 1.73517223e-02 4.17306453e-01
-3.51748824e-01 -5.54950416e-01 6.58853590e-01 -3.72319907e-01
-9.20413256e-01 -3.21639217e-02 -5.50862662e-02 3.51080596e-01
8.55707705e-01 4.08260487e-02 -5.05682409e-01 -3.21858339e-02
-6.45299554e-01 1.38622537e-01 2.56180674e-01 6.19417012e-01
5.40447414e-01 -1.19519234e+00 -6.01996243e-01 1.23653956e-01
1.23172127e-01 3.75966251e-01 1.22252382e-01 3.21529597e-01
-3.73518288e-01 2.89222926e-01 2.66804677e-02 -4.20331597e-01
-1.11306667e+00 4.43901062e-01 -6.13048021e-03 -5.91141045e-01
-4.81626093e-01 1.22251260e+00 2.18236372e-01 -3.48077476e-01
4.10285056e-01 -4.62411940e-01 8.84133801e-02 9.75591242e-02
4.95901823e-01 2.04281688e-01 4.67669964e-02 -7.75806010e-02
-2.32867330e-01 2.16202244e-01 -3.71529639e-01 -1.97473109e-01
1.50076461e+00 1.43114820e-01 -2.78747022e-01 2.38644868e-01
8.29287648e-01 -1.30180538e-01 -1.10567200e+00 -2.70362377e-01
3.39350730e-01 -1.86020076e-01 -9.25517753e-02 -1.22310615e+00
-7.61414170e-01 1.09232640e+00 6.40750155e-02 -2.44965106e-01
6.93588138e-01 7.72174969e-02 8.04749608e-01 6.98147357e-01
5.70512235e-01 -1.51180041e+00 3.37597370e-01 6.74688697e-01
9.85142112e-01 -1.30833781e+00 -2.11275667e-01 -5.92395961e-01
-6.74408257e-01 9.89009500e-01 1.01030874e+00 -1.52633026e-01
2.94092119e-01 2.43085802e-01 -1.24207348e-01 2.45721057e-01
-1.03953350e+00 -3.14968824e-02 -7.92301670e-02 3.94474447e-01
5.81987619e-01 1.24915533e-01 -2.05349073e-01 1.03776324e+00
-4.62799132e-01 -1.38310432e-01 2.32027441e-01 8.19325030e-01
-3.17434669e-01 -1.35687566e+00 -3.73654105e-02 7.06720054e-02
-3.24037910e-01 -5.72596788e-01 -3.55358422e-01 7.22549200e-01
3.17270696e-01 9.60948646e-01 8.17565247e-03 -1.98400825e-01
3.06302775e-02 4.94592160e-01 4.64843035e-01 -1.04326749e+00
-7.35184789e-01 -4.40558307e-02 3.14899474e-01 -8.59465361e-01
-4.54925299e-01 -3.59692842e-01 -1.30202401e+00 5.14286458e-02
-2.80700505e-01 3.43662828e-01 6.65144682e-01 1.09551668e+00
5.74208796e-01 5.42584360e-01 2.26424634e-01 -7.35162735e-01
-6.00703120e-01 -1.03974819e+00 -1.61635384e-01 5.20090759e-01
-2.48712048e-01 -6.89519763e-01 -2.12464541e-01 1.23793945e-01] | [10.721213340759277, 8.565399169921875] |
09e4706b-bff1-40e4-84b5-d2aee0518ea0 | multi-step-entity-centric-information | 1909.07598 | null | https://arxiv.org/abs/1909.07598v1 | https://arxiv.org/pdf/1909.07598v1.pdf | Multi-step Entity-centric Information Retrieval for Multi-Hop Question Answering | Multi-hop question answering (QA) requires an information retrieval (IR) system that can find \emph{multiple} supporting evidence needed to answer the question, making the retrieval process very challenging. This paper introduces an IR technique that uses information of entities present in the initially retrieved evidence to learn to `\emph{hop}' to other relevant evidence. In a setting, with more than \textbf{5 million} Wikipedia paragraphs, our approach leads to significant boost in retrieval performance. The retrieved evidence also increased the performance of an existing QA model (without any training) on the \hotpot benchmark by \textbf{10.59} F1. | ['Abhishek Singhal', 'Dilip Kavarthapu', 'Ameya Godbole', 'Manzil Zaheer', 'Andrew McCallum', 'Zhiyu Gong', 'Xiaoxiao Guo', 'Rajarshi Das', 'Hamed Zamani', 'Mo Yu', 'Tian Gao'] | 2019-09-17 | multi-step-entity-centric-information-1 | https://aclanthology.org/D19-5816 | https://aclanthology.org/D19-5816.pdf | ws-2019-11 | ['multi-hop-question-answering'] | ['knowledge-base'] | [ 8.54423922e-03 5.71657658e-01 -2.20365852e-01 -1.59870133e-01
-2.10168052e+00 -1.02268219e+00 7.01954126e-01 6.61696196e-01
-8.44283104e-01 1.20411253e+00 2.12371632e-01 -5.33141494e-01
-6.96594596e-01 -8.11455488e-01 -1.07110071e+00 -7.82984123e-02
1.91191941e-01 1.02359366e+00 9.15490270e-01 -8.39299142e-01
6.07688963e-01 -7.38612711e-02 -1.54760015e+00 6.93634152e-01
1.18068957e+00 1.10821712e+00 2.51996398e-01 8.97184670e-01
-4.19164240e-01 1.07134604e+00 -9.96241152e-01 -7.82156348e-01
-1.23255044e-01 1.08393212e-03 -1.50565588e+00 -9.89972949e-01
1.06122637e+00 -1.26215115e-01 -2.48550013e-01 7.59892941e-01
4.91716117e-01 2.02824026e-01 6.90881371e-01 -6.15371704e-01
-8.16149771e-01 5.89490294e-01 -2.39109695e-01 1.01157343e+00
8.54314089e-01 -4.53035355e-01 1.52016330e+00 -1.02753234e+00
9.84461129e-01 9.86336827e-01 2.66572505e-01 3.04847986e-01
-5.17876387e-01 -3.89029443e-01 -3.75550836e-01 4.97570455e-01
-1.29269803e+00 -3.59806091e-01 2.19221875e-01 -2.70338934e-02
1.31542408e+00 4.91738379e-01 -1.55502662e-01 5.27046382e-01
-8.19566697e-02 6.81344926e-01 9.15097892e-01 -8.98517609e-01
-1.21126644e-01 2.64773518e-01 7.94560671e-01 6.92193270e-01
4.74789709e-01 -5.16422510e-01 -4.57566500e-01 -3.37381989e-01
-1.66312397e-01 -8.62968564e-01 -2.12707862e-01 5.83315790e-01
-7.51794577e-01 8.71241927e-01 3.77539843e-01 2.95413941e-01
-4.05178338e-01 1.68262482e-01 2.74101704e-01 5.16170144e-01
8.70741978e-02 1.15306914e+00 -6.46182716e-01 3.68423313e-02
-7.35296190e-01 5.33042550e-01 1.12766290e+00 7.39927053e-01
9.16714787e-01 -8.53504360e-01 -5.77993214e-01 8.81070316e-01
3.40255201e-01 8.78181040e-01 1.57860160e-01 -1.42688310e+00
1.24752676e+00 5.99676549e-01 4.68447149e-01 -9.98414457e-01
-1.83679149e-01 -3.02125514e-01 -1.09088300e-02 -6.25525475e-01
4.66982603e-01 -2.15239048e-01 -7.34352291e-01 1.44596696e+00
2.89996326e-01 -4.85136926e-01 3.53768766e-01 5.59605122e-01
1.33542871e+00 1.02801418e+00 1.23352334e-01 -3.67138465e-03
1.78820598e+00 -9.87909615e-01 -7.10130394e-01 -3.82727772e-01
8.09595942e-01 -1.01812816e+00 9.08713877e-01 1.88887373e-01
-1.23484921e+00 -2.48459995e-01 -9.71628845e-01 -4.41869169e-01
-6.02740049e-01 -1.34227052e-01 2.88333654e-01 4.31540668e-01
-1.20888305e+00 1.39236107e-01 1.93145350e-01 -1.44350380e-01
1.71545881e-03 1.62844181e-01 -1.49446413e-01 -7.42245615e-01
-1.86094308e+00 1.23627996e+00 4.63148654e-01 -2.54628152e-01
-4.98749048e-01 -6.86653674e-01 -7.19038010e-01 6.43142685e-02
6.82154357e-01 -1.02256680e+00 1.44083297e+00 -1.05435029e-01
-7.81754494e-01 6.20318711e-01 -2.63006032e-01 -3.74679506e-01
-3.55991423e-02 -5.57137430e-01 -5.21484256e-01 8.95840764e-01
5.74311912e-01 7.09812641e-01 4.83261228e-01 -1.25525010e+00
-6.45768940e-01 -1.26740605e-01 5.23388863e-01 4.05634284e-01
-1.70016900e-01 4.84914303e-01 -8.16777349e-01 -2.37509996e-01
-2.19160944e-01 -6.46814644e-01 1.95443973e-01 -6.58097982e-01
-2.35783026e-01 -9.17767823e-01 3.27537715e-01 -1.02318048e+00
1.57546628e+00 -1.52634192e+00 -1.97513670e-01 3.77947956e-01
-4.81642149e-02 3.61747295e-01 -2.97479033e-01 8.12856972e-01
5.09716809e-01 4.21228945e-01 8.84735733e-02 2.13283360e-01
1.09826803e-01 3.25475097e-01 -3.76851618e-01 -2.90026397e-01
7.90141672e-02 1.05497682e+00 -9.06174541e-01 -1.08203626e+00
-5.92057168e-01 1.65988445e-01 -3.08910906e-01 -3.71480780e-03
-6.41908765e-01 -4.60558645e-02 -7.56646931e-01 7.09179521e-01
1.57896001e-02 -6.04859531e-01 -2.24856913e-01 5.16746677e-02
2.69673467e-01 6.32453680e-01 -8.97596598e-01 1.49841750e+00
-3.97733212e-01 4.36962396e-01 -1.34183124e-01 -5.32334745e-01
6.67670310e-01 5.43609083e-01 2.14385614e-01 -1.40023780e+00
-1.75942004e-01 6.51897192e-01 -1.23957045e-01 -7.73350894e-01
7.92908609e-01 2.24356309e-01 -2.39816517e-01 5.85934222e-01
5.32581611e-03 -2.35766634e-01 7.72520185e-01 7.20516920e-01
1.48426104e+00 -3.49368781e-01 -2.35508651e-01 -1.64999187e-01
7.68366396e-01 6.00544810e-01 -6.80473000e-02 1.22884226e+00
3.48511010e-01 2.61100501e-01 1.45509869e-01 -1.71697717e-02
-8.77485096e-01 -9.32386398e-01 -1.71752602e-01 1.20554245e+00
1.78102672e-01 -6.25337422e-01 -5.65529227e-01 -8.28347564e-01
2.06413828e-02 9.51069951e-01 -4.02216762e-01 -1.78547502e-02
-9.32514846e-01 -4.60477144e-01 9.05544400e-01 2.84556806e-01
5.19787014e-01 -1.09591746e+00 -4.35379684e-01 2.23633662e-01
-8.73794556e-01 -1.07520032e+00 -1.21669121e-01 -2.16502026e-02
-6.52614236e-01 -1.06426597e+00 -8.63270700e-01 -8.21861088e-01
2.97491670e-01 1.18097506e-01 1.72916508e+00 4.73378062e-01
-2.11193681e-01 6.53044581e-01 -6.90753877e-01 -5.85065246e-01
-2.76326448e-01 4.20598179e-01 -5.19976139e-01 -9.28548753e-01
3.81750554e-01 1.39095038e-01 -8.25764954e-01 2.33093008e-01
-1.17753935e+00 -6.93682075e-01 6.32697046e-01 7.05141366e-01
4.55872536e-01 7.71059543e-02 1.15672696e+00 -7.98892140e-01
1.12718165e+00 -6.65514827e-01 -3.36361647e-01 9.66977715e-01
-9.30105150e-01 3.77926141e-01 3.96428317e-01 4.64871377e-02
-1.27459669e+00 -7.73081362e-01 -3.29072237e-01 2.47252345e-01
2.64947355e-01 1.05935872e+00 4.63533580e-01 1.07272282e-01
1.11785054e+00 -2.33370945e-01 -4.62642401e-01 -5.78170002e-01
5.67465305e-01 4.70651180e-01 4.19900298e-01 -8.91381502e-01
6.74811482e-01 -3.22759636e-02 -1.34612441e-01 -5.69412589e-01
-1.42375779e+00 -9.50075030e-01 -2.41558671e-01 -1.90574139e-01
7.19886601e-01 -6.44518077e-01 -5.46479285e-01 -4.19376373e-01
-1.16714585e+00 -1.65800434e-02 -2.18977183e-01 3.63214046e-01
-9.68613774e-02 5.26757896e-01 -7.17619479e-01 -5.76203525e-01
-7.92386651e-01 -5.04303515e-01 9.84370589e-01 2.85141259e-01
-2.31261104e-01 -7.70088792e-01 5.54446220e-01 1.23450530e+00
3.76659274e-01 -2.79101342e-01 1.30525625e+00 -8.74581218e-01
-5.87412655e-01 -5.26470125e-01 -3.42616618e-01 2.02050239e-01
-3.79874229e-01 -2.37988219e-01 -7.78844655e-01 -3.64890462e-03
-3.96186322e-01 -8.34359109e-01 9.73744512e-01 -1.17234036e-01
6.47374272e-01 -6.32962584e-01 -2.67770261e-01 -5.54086864e-01
1.40780139e+00 -3.38209346e-02 7.03792214e-01 6.99696600e-01
2.10338116e-01 8.52678001e-01 8.67742717e-01 -6.68039471e-02
7.02481627e-01 5.59573591e-01 2.12569863e-01 3.80046993e-01
-2.41134867e-01 -4.36314903e-02 1.32282168e-01 7.33188689e-01
-9.87421274e-02 -4.96397734e-01 -1.15550447e+00 7.58080125e-01
-1.55014360e+00 -9.96230543e-01 -2.22786039e-01 1.81848192e+00
1.27615941e+00 1.23818345e-01 -2.69843698e-01 -7.87440687e-02
4.19210315e-01 -1.81377336e-01 -2.38198027e-01 -4.13727105e-01
-3.06291550e-01 7.97173142e-01 3.35890591e-01 7.08382964e-01
-7.12747037e-01 8.63717020e-01 6.26201200e+00 9.93312597e-01
-2.54502207e-01 -6.13862984e-02 2.45578706e-01 3.28895479e-01
-5.83545923e-01 4.76063192e-02 -1.17884719e+00 1.27227515e-01
1.30497265e+00 -3.98456275e-01 1.07164726e-01 5.37289917e-01
-4.94903326e-01 -5.50323606e-01 -7.72363186e-01 2.75595367e-01
4.03503895e-01 -1.37643158e+00 4.44305837e-01 -2.28591159e-01
6.17850423e-01 1.62652895e-01 4.06773649e-02 6.99706256e-01
5.95499098e-01 -9.88335252e-01 2.56477624e-01 6.71086311e-01
6.10279560e-01 -6.55735493e-01 1.19863617e+00 6.62217081e-01
-8.31966400e-01 -1.22324340e-01 -4.47268337e-01 3.94331753e-01
2.79887557e-01 4.54228729e-01 -1.32140911e+00 9.90526319e-01
1.03803527e+00 -6.77741766e-02 -1.12276435e+00 9.62924659e-01
-4.98732716e-01 6.03253305e-01 -5.71048975e-01 -4.23980713e-01
4.84039307e-01 3.22010756e-01 4.95857626e-01 1.12752914e+00
2.87006974e-01 4.04390007e-01 -2.55503178e-01 1.60307705e-01
-6.19907141e-01 4.97061938e-01 -4.77922559e-01 6.05755858e-02
5.78667641e-01 1.01136231e+00 -3.06432664e-01 -5.57181001e-01
-2.06846520e-01 6.00425303e-01 4.01054233e-01 2.39174768e-01
-5.32736957e-01 -7.84776390e-01 -4.21016484e-01 -1.42451882e-01
3.80609930e-01 7.28815943e-02 3.20056796e-01 -7.87248313e-01
5.08722723e-01 -9.92124677e-01 1.21885455e+00 -1.06641018e+00
-1.43832779e+00 6.58479810e-01 2.94626862e-01 -5.41837335e-01
-5.44329286e-01 -4.37880099e-01 -1.31979391e-01 9.34202731e-01
-2.07835650e+00 -8.05738091e-01 1.99291054e-02 5.67151487e-01
3.46984684e-01 1.75341457e-01 8.52750301e-01 6.46521628e-01
-9.96758267e-02 5.69307923e-01 1.62883669e-01 1.92876220e-01
9.96959746e-01 -1.46148574e+00 -2.74715513e-01 5.48797369e-01
3.44696820e-01 9.92352962e-01 5.92187583e-01 -7.38542259e-01
-1.38779438e+00 -5.99209130e-01 1.51433384e+00 -1.14891744e+00
8.19190025e-01 3.29742312e-01 -1.05561447e+00 3.98352504e-01
7.42492914e-01 -4.89350736e-01 7.51012266e-01 2.08204508e-01
-6.52418554e-01 -2.23025694e-01 -1.06562388e+00 4.36731458e-01
5.16561747e-01 -6.05449319e-01 -1.49183166e+00 6.19611263e-01
9.47371960e-01 -2.75972486e-01 -1.14816785e+00 6.74406588e-01
2.34964222e-01 -1.20618947e-01 1.36046481e+00 -6.55958235e-01
4.52208102e-01 -3.14161986e-01 -2.16594696e-01 -7.94580758e-01
9.85481367e-02 -3.98986578e-01 -5.58600008e-01 1.09022427e+00
1.33194149e+00 -3.90968323e-01 3.68641943e-01 8.51348817e-01
-1.41253263e-01 -6.16406918e-01 -1.16384482e+00 -5.38873076e-01
4.24656481e-01 -1.15761422e-01 2.46371567e-01 5.90143442e-01
1.51493371e-01 7.92679191e-01 2.56984442e-01 2.25143522e-01
1.95956275e-01 1.19789042e-01 3.06848645e-01 -1.23433471e+00
-1.87372193e-01 -1.21788494e-01 2.96888888e-01 -1.28463995e+00
-6.66263849e-02 -1.04662359e+00 1.79921716e-01 -2.16595602e+00
1.64482430e-01 -7.14998484e-01 -3.56538057e-01 4.35469180e-01
-5.20045280e-01 3.02217543e-01 -1.87646076e-01 3.87108177e-01
-1.28019047e+00 6.02989979e-02 1.19733536e+00 -1.77893683e-01
3.47195983e-01 -3.33755553e-01 -8.72997046e-01 2.66043931e-01
6.13158405e-01 -6.29015744e-01 -3.93344969e-01 -7.85179853e-01
1.25848150e+00 2.91377693e-01 1.28494412e-01 -7.29349077e-01
7.77275681e-01 3.39954048e-01 1.07593477e-01 -7.87445545e-01
4.62689877e-01 -5.02955735e-01 -5.41773617e-01 2.01657325e-01
-6.61980212e-01 6.73699617e-01 2.77351260e-01 6.02216899e-01
-4.37848419e-01 -9.86487269e-01 2.89814658e-02 -3.61154348e-01
-5.44623137e-01 -1.93912536e-01 -3.37432116e-01 7.74912655e-01
5.38468778e-01 3.67893189e-01 -1.08543479e+00 -4.86950457e-01
-3.23539108e-01 6.28132939e-01 -8.41754749e-02 1.92267075e-01
5.67007661e-01 -8.61437023e-01 -9.09688771e-01 -8.97422731e-01
3.54850650e-01 1.09892681e-01 7.74761438e-02 6.15960479e-01
-5.99798918e-01 1.09180892e+00 3.84361953e-01 -2.77745724e-01
-1.19435251e+00 4.87162173e-01 -8.17857757e-02 -8.15351307e-01
-3.41023594e-01 1.00197506e+00 -6.21022463e-01 -4.33648795e-01
1.41849458e-01 2.05272079e-01 -7.75471866e-01 3.87634814e-01
6.82543993e-01 5.11183441e-01 5.61008394e-01 -2.80589521e-01
-1.34025171e-01 5.58465660e-01 -3.69721472e-01 -5.75490952e-01
1.05600357e+00 -1.35184214e-01 -3.72455865e-01 1.25059754e-01
1.15064204e+00 3.38916689e-01 -2.12400556e-01 -4.64126080e-01
7.09689081e-01 -1.34532422e-01 -1.06085032e-01 -1.42254913e+00
-2.81887144e-01 5.69088519e-01 1.42441303e-01 4.85374838e-01
7.15320289e-01 5.22211015e-01 1.04038966e+00 1.41091394e+00
3.87086660e-01 -1.39907479e+00 4.56215322e-01 8.52008462e-01
8.48792255e-01 -1.30332124e+00 7.46660009e-02 -2.42561281e-01
-4.66814935e-01 9.05042529e-01 4.03349847e-01 2.52495110e-01
3.60619277e-01 -3.55098426e-01 1.60515368e-01 -9.86734211e-01
-7.62845159e-01 -2.37362280e-01 7.74330437e-01 6.48470372e-02
2.05025628e-01 -5.75599134e-01 -6.21460080e-01 3.02113861e-01
-1.91969499e-01 -2.63488024e-01 1.46380067e-01 1.19079375e+00
-8.11283231e-01 -1.07381380e+00 -2.48782918e-01 5.62270582e-01
-9.59344268e-01 -5.47733545e-01 -5.80964267e-01 8.56278479e-01
-2.84213752e-01 1.54591405e+00 -3.28977019e-01 3.13823849e-01
3.33775431e-01 4.93683696e-01 4.28226173e-01 -7.21909463e-01
-8.69400203e-01 -3.90523046e-01 9.22505975e-01 -2.55012214e-01
-5.32204092e-01 -3.39438945e-01 -1.30510366e+00 6.78056329e-02
-5.60357213e-01 9.70331848e-01 5.47089636e-01 1.01418138e+00
4.04636651e-01 2.01686889e-01 3.85875493e-01 4.80422407e-01
-5.19088268e-01 -1.04353642e+00 8.22970197e-02 3.37361157e-01
7.59704486e-02 -3.79819751e-01 -3.48702580e-01 -2.53586650e-01] | [11.01725959777832, 7.883810520172119] |
941751a0-7534-44d9-8029-ae09ff2ae1e4 | hierarchical-dynamic-image-harmonization | 2211.08639 | null | https://arxiv.org/abs/2211.08639v3 | https://arxiv.org/pdf/2211.08639v3.pdf | Hierarchical Dynamic Image Harmonization | Image harmonization is a critical task in computer vision, which aims to adjust the foreground to make it compatible with the background. Recent works mainly focus on using global transformations (i.e., normalization and color curve rendering) to achieve visual consistency. However, these models ignore local visual consistency and their huge model sizes limit their harmonization ability on edge devices. In this paper, we propose a hierarchical dynamic network (HDNet) to adapt features from local to global view for better feature transformation in efficient image harmonization. Inspired by the success of various dynamic models, local dynamic (LD) module and mask-aware global dynamic (MGD) module are proposed in this paper. Specifically, LD matches local representations between the foreground and background regions based on semantic similarities, then adaptively adjust every foreground local representation according to the appearance of its $K$-nearest neighbor background regions. In this way, LD can produce more realistic images at a more fine-grained level, and simultaneously enjoy the characteristic of semantic alignment. The MGD effectively applies distinct convolution to the foreground and background, learning the representations of foreground and background regions as well as their correlations to the global harmonization, facilitating local visual consistency for the images much more efficiently. Experimental results demonstrate that the proposed HDNet significantly reduces the total model parameters by more than 80\% compared to previous methods, while still attaining state-of-the-art performance on the popular iHarmony4 dataset. Notably, the HDNet achieves a 4\% improvement in PSNR and a 19\% reduction in MSE compared to the prior state-of-the-art methods. | ['Huaxiong Li', 'Weiqiang Wang', 'Changhua Meng', 'Jun Lan', 'Yaohui Li', 'Zhangxuan Gu', 'Haoxing Chen'] | 2022-11-16 | null | null | null | null | ['image-harmonization'] | ['computer-vision'] | [ 1.03128783e-01 -5.50264657e-01 -8.11147615e-02 -2.01532707e-01
-1.62277207e-01 -5.87312765e-02 2.54943788e-01 -3.48059863e-01
-2.68456519e-01 2.70042360e-01 -5.65346330e-02 1.26624778e-01
5.94031513e-02 -8.71269584e-01 -4.80715960e-01 -9.78035688e-01
4.70399588e-01 -6.42538667e-02 6.80427730e-01 -2.70826340e-01
1.83878243e-01 5.44380367e-01 -1.45975852e+00 8.92925337e-02
1.11950779e+00 1.12428844e+00 4.34833795e-01 2.66305655e-01
-2.45082974e-01 6.01796269e-01 -4.50525016e-01 -3.96924704e-01
4.60789889e-01 -6.77882314e-01 -3.55743945e-01 3.64031911e-01
6.38976514e-01 -2.21553296e-01 -5.45957327e-01 1.42853189e+00
6.86462522e-01 1.06812522e-01 2.54190356e-01 -1.22515810e+00
-9.22658801e-01 3.16204846e-01 -1.02748716e+00 3.30059797e-01
-2.51145869e-01 3.24390680e-01 7.01159358e-01 -7.88720787e-01
5.63013375e-01 1.29241633e+00 6.27321124e-01 2.87134707e-01
-1.35547245e+00 -1.08678770e+00 3.26080769e-01 5.83257914e-01
-1.70718384e+00 -3.55720520e-01 1.04447007e+00 -4.71411720e-02
3.95305544e-01 3.71384084e-01 8.35377216e-01 5.27848482e-01
2.99158812e-01 6.82354808e-01 1.08090615e+00 -2.33928218e-01
-6.63570464e-02 -8.71563852e-02 -3.36390495e-01 6.79693520e-01
2.62200713e-01 -1.40604466e-01 -4.82282221e-01 2.45632842e-01
1.19424927e+00 -8.42429996e-02 -5.96942365e-01 -5.68378687e-01
-1.27503276e+00 5.52636206e-01 8.13477576e-01 4.85833436e-01
-3.13799590e-01 2.32060730e-01 2.61897296e-01 -3.21969762e-02
1.11160383e-01 1.27584696e-01 -2.19730094e-01 3.73397499e-01
-8.22976112e-01 1.15018204e-01 2.89733827e-01 8.89725089e-01
7.79598594e-01 3.42874736e-01 -3.97372544e-01 1.01604509e+00
2.43688494e-01 5.43414414e-01 4.66684997e-01 -1.09284139e+00
2.94720113e-01 7.76668429e-01 2.41625169e-03 -1.65608478e+00
-2.71403491e-01 -5.97935200e-01 -1.39848363e+00 2.78915137e-01
1.54884756e-01 3.31146806e-01 -8.24682117e-01 1.73430657e+00
5.44283628e-01 5.03951907e-01 -1.05411343e-01 8.93657207e-01
8.83904040e-01 8.12733173e-01 9.32330042e-02 -5.61670005e-01
1.16530943e+00 -1.10332739e+00 -9.54560459e-01 -1.70995951e-01
-1.22802362e-01 -1.11717975e+00 9.35868680e-01 1.56264275e-01
-1.26959586e+00 -1.09835851e+00 -1.03096771e+00 -6.54965788e-02
2.80729379e-03 -3.06937993e-02 4.35644567e-01 4.99096602e-01
-1.23081195e+00 2.99577475e-01 -6.23321056e-01 -1.65950850e-01
2.76422828e-01 2.31212005e-01 -2.21316181e-02 -6.81116506e-02
-1.05571997e+00 6.75452709e-01 5.08680761e-01 3.23878258e-01
-4.33736742e-01 -7.09482431e-01 -8.39048624e-01 -4.57187518e-02
2.68547654e-01 -8.37453246e-01 8.13272059e-01 -1.13044298e+00
-1.53608477e+00 8.54128182e-01 -1.81007147e-01 -2.37417549e-01
6.43230915e-01 2.13669360e-01 -6.51640654e-01 9.90251675e-02
-2.00706515e-02 7.41655529e-01 9.10140634e-01 -1.38752639e+00
-1.00593483e+00 -2.13170961e-01 -3.30274850e-01 3.46368283e-01
-4.91587013e-01 1.10518917e-01 -1.36627984e+00 -1.10379159e+00
4.23492253e-01 -7.68568277e-01 -2.63224572e-01 3.31660300e-01
-1.33598894e-01 1.09138303e-02 1.25676787e+00 -7.66231358e-01
1.47403014e+00 -2.16944838e+00 2.27855057e-01 3.08839113e-01
3.17673057e-01 5.82957268e-01 -1.64502978e-01 -2.80979633e-01
-5.11199273e-02 -1.62005171e-01 -3.29684943e-01 -1.22468300e-01
-1.56058282e-01 2.37519383e-01 1.79664455e-02 5.71526468e-01
-4.82299738e-02 1.07352471e+00 -5.79811633e-01 -7.87253141e-01
5.16323864e-01 6.88227952e-01 -5.08725226e-01 9.48815495e-02
1.92216158e-01 4.43633527e-01 -3.26818943e-01 5.83209753e-01
1.07822442e+00 -2.26649240e-01 2.14575216e-01 -7.08423018e-01
-1.96489424e-01 -5.22254586e-01 -1.60957932e+00 1.43754613e+00
-3.27637970e-01 5.72369993e-01 2.30417669e-01 -8.61252546e-01
1.12645280e+00 -6.05738871e-02 6.12194896e-01 -1.20860946e+00
2.04639509e-01 2.12883666e-01 5.69842570e-02 -1.46625146e-01
2.51758486e-01 2.12754875e-01 1.50168866e-01 -1.00113519e-01
-3.27815622e-01 -7.59297609e-02 8.96383375e-02 -1.64180145e-01
2.72757500e-01 1.61442067e-02 2.79536217e-01 -4.56485748e-01
8.62909615e-01 -3.57046366e-01 1.06908262e+00 5.15439272e-01
-2.91269660e-01 7.43229210e-01 -2.28828262e-03 -5.79975307e-01
-8.76791060e-01 -1.05042994e+00 -2.12033987e-01 7.63736844e-01
1.09779131e+00 -1.43797845e-01 -9.48742628e-01 -6.00980148e-02
-1.45828828e-01 5.56962073e-01 -5.16235352e-01 -3.60590339e-01
-9.81763721e-01 -8.39326978e-01 3.16467345e-01 5.22408426e-01
1.44309127e+00 -9.89312172e-01 -6.16637349e-01 2.97462612e-01
-4.26950395e-01 -1.14495528e+00 -8.15419078e-01 -3.07323605e-01
-6.50469244e-01 -8.30239058e-01 -7.90887952e-01 -1.10173905e+00
6.04725122e-01 5.50661027e-01 9.51888740e-01 2.25027800e-01
-4.53605920e-01 -6.02494814e-02 -7.89987221e-02 1.64119098e-02
-2.48189420e-01 -3.09595704e-01 -1.52924687e-01 4.38712806e-01
-3.21541354e-02 -5.20091712e-01 -1.04137993e+00 8.36023569e-01
-1.11150730e+00 3.18665534e-01 6.40291989e-01 8.21346641e-01
9.25020456e-01 3.73952329e-01 1.80029541e-01 -4.74908650e-01
2.32116282e-01 1.50158927e-01 -7.20823169e-01 4.73084092e-01
-7.29702890e-01 -3.28924030e-01 5.93244970e-01 -3.99018973e-01
-1.28640807e+00 -1.47225633e-01 9.75978225e-02 -6.75793946e-01
2.58018106e-01 -6.88236132e-02 -5.90805054e-01 -5.48486292e-01
3.34118485e-01 4.84567463e-01 -7.89860040e-02 -1.56214073e-01
4.13555413e-01 2.75745749e-01 9.39835727e-01 -3.87471259e-01
9.67268050e-01 5.95462501e-01 -2.85169259e-02 -4.52448070e-01
-4.84641761e-01 -3.55699390e-01 -5.89104414e-01 -3.65306705e-01
1.05423963e+00 -8.26010823e-01 -6.58830523e-01 9.58459973e-01
-8.32572281e-01 -4.17303890e-01 -2.92508572e-01 1.63650274e-01
-4.56412047e-01 5.02633274e-01 -5.57260275e-01 -2.84181595e-01
-4.14979339e-01 -1.30824053e+00 7.72113681e-01 7.68872261e-01
2.33644813e-01 -8.47009122e-01 -3.46786678e-01 1.10837705e-01
6.81800127e-01 3.59273762e-01 7.62375176e-01 2.06609234e-01
-7.70392835e-01 2.38232151e-01 -6.03469193e-01 2.90067732e-01
3.33473057e-01 1.77424982e-01 -6.64479434e-01 -3.84473383e-01
3.08928862e-02 3.07091683e-01 7.91242063e-01 6.59974992e-01
1.33270192e+00 -2.42393821e-01 -3.14771175e-01 1.14040911e+00
1.53098464e+00 5.23349404e-01 9.60505426e-01 5.39678454e-01
9.26358104e-01 3.61935139e-01 5.98905325e-01 3.52975368e-01
3.81236315e-01 1.11953878e+00 3.62904102e-01 -7.09496915e-01
-7.52666354e-01 3.21928002e-02 9.03840214e-02 7.74929762e-01
-3.40727903e-02 -1.44204274e-01 -6.66587174e-01 3.69605660e-01
-1.96863186e+00 -9.83884513e-01 -3.77099067e-02 2.07248592e+00
9.12153840e-01 -2.02873982e-02 -1.44538462e-01 -1.09195456e-01
1.19815791e+00 3.54004383e-01 -7.30887532e-01 -6.69092964e-03
-5.56414843e-01 3.41798849e-02 4.74614024e-01 2.96117544e-01
-1.13885939e+00 1.00705814e+00 5.40874529e+00 1.32501709e+00
-1.30276251e+00 7.38489702e-02 1.16209650e+00 7.39378035e-02
-1.06619276e-01 -2.60128736e-01 -7.21795619e-01 6.20228648e-01
1.26073420e-01 -2.30299175e-01 5.49260974e-01 6.21001244e-01
1.87032938e-01 1.13655232e-01 -4.67163742e-01 1.42958534e+00
1.67415276e-01 -1.46927404e+00 2.19299480e-01 -1.34095356e-01
1.21461165e+00 -3.60156268e-01 3.18439424e-01 -4.87375855e-02
-1.38054490e-02 -7.87306786e-01 8.83709252e-01 4.84315068e-01
7.88714170e-01 -1.10982025e+00 6.38170719e-01 6.01693476e-03
-1.76133168e+00 -5.93404323e-02 -5.86855054e-01 6.15652919e-01
1.60878092e-01 3.80522996e-01 2.24101637e-02 6.03171766e-01
1.23758256e+00 6.91152275e-01 -7.43913770e-01 9.42639768e-01
2.74083316e-02 1.28619820e-01 -1.29940316e-01 3.19370329e-01
1.51678711e-01 -6.16396070e-01 3.88952643e-01 1.14890981e+00
2.80517727e-01 1.86266273e-01 3.28653932e-01 8.97544503e-01
-3.63678634e-02 2.83021837e-01 6.25982732e-02 5.08141279e-01
4.32984710e-01 1.39523542e+00 -9.04710710e-01 -4.51702476e-01
-2.83443183e-01 1.26697159e+00 -1.78542379e-02 4.13475603e-01
-1.16956103e+00 -4.33483362e-01 7.63669431e-01 -9.48764682e-02
5.44426978e-01 -1.76157374e-02 -2.47082531e-01 -1.00576305e+00
7.15917051e-02 -9.97853339e-01 3.98925722e-01 -7.28980899e-01
-1.16862857e+00 6.24206662e-01 -2.10895956e-01 -1.33873069e+00
1.67886600e-01 -2.38545656e-01 -6.60926104e-01 8.83230627e-01
-1.57533157e+00 -1.20787287e+00 -9.05648589e-01 9.56485450e-01
4.53072757e-01 1.05186708e-01 2.58383811e-01 4.71578181e-01
-7.08323002e-01 7.74249673e-01 1.49620384e-01 1.38195708e-01
8.78555000e-01 -8.62426281e-01 3.34520161e-01 1.15283191e+00
8.41423962e-03 5.35196185e-01 5.31472564e-01 -5.31368971e-01
-1.21062076e+00 -1.36830115e+00 4.47725534e-01 1.37075394e-01
3.50736380e-01 1.24454752e-01 -1.13634562e+00 2.15457633e-01
2.79591650e-01 3.68089080e-01 2.09598750e-01 -6.55657232e-01
-2.41544068e-01 -6.75880551e-01 -1.10282946e+00 9.76229966e-01
1.25712144e+00 -2.17281505e-01 -1.03832453e-01 1.88130862e-03
7.75097549e-01 -6.19193196e-01 -7.00206637e-01 6.79763079e-01
4.36516732e-01 -1.27040708e+00 1.22914946e+00 9.41204056e-02
1.41150892e-01 -1.03644168e+00 -2.03174919e-01 -9.40665901e-01
-7.48819053e-01 -5.95183372e-01 1.15129448e-01 1.42750478e+00
-2.20965356e-01 -6.56154573e-01 3.36006403e-01 4.09508020e-01
-9.95047018e-02 -7.81458139e-01 -8.94803464e-01 -6.26913428e-01
-3.17843705e-01 -8.92644972e-02 5.91174543e-01 1.06289995e+00
-8.39236200e-01 -1.60539761e-01 -3.50360274e-01 2.81214237e-01
7.31084824e-01 4.93845403e-01 8.45281839e-01 -8.11553776e-01
-2.20529795e-01 -9.37931240e-01 -5.74308515e-01 -9.82947350e-01
2.91838520e-03 -7.04553127e-01 -2.74385847e-02 -1.47096312e+00
3.82851094e-01 -4.03195083e-01 -5.46392143e-01 2.48241425e-01
-5.82938790e-01 8.54770780e-01 4.62338328e-01 2.03819692e-01
-5.57837248e-01 6.94263160e-01 1.60059464e+00 -2.98108071e-01
-3.32715511e-01 -3.06135982e-01 -7.83981204e-01 8.02436829e-01
6.93188667e-01 -1.41060203e-02 -1.95612416e-01 -4.50497687e-01
-3.34586024e-01 -3.20307702e-01 4.06275451e-01 -1.17963576e+00
3.67821336e-01 -2.75151610e-01 8.79781604e-01 -4.48880970e-01
1.06152534e-01 -8.56168032e-01 6.07761264e-01 6.76862478e-01
-2.89994795e-02 1.40234843e-01 3.33999455e-01 5.73637962e-01
-4.73325729e-01 2.91239798e-01 1.39205372e+00 5.26829064e-02
-1.11556935e+00 5.90920329e-01 1.35375977e-01 -7.03103170e-02
1.27914751e+00 -4.28158075e-01 -2.65709639e-01 -1.59234568e-01
-3.15781891e-01 2.17809424e-01 7.08004296e-01 5.16624749e-01
6.66927874e-01 -1.49987626e+00 -7.02188730e-01 3.95192206e-01
-1.74636692e-02 -2.00600736e-02 6.55150771e-01 8.90028119e-01
-8.39813769e-01 -4.95953485e-02 -5.37707329e-01 -8.18164289e-01
-1.34179008e+00 5.30303001e-01 6.46605134e-01 -4.85313796e-02
-7.99950361e-01 8.23507726e-01 7.71095693e-01 -5.59362546e-02
1.79489315e-01 -1.20506883e-01 -1.69231638e-01 -1.00631699e-01
6.38337672e-01 4.25261945e-01 -1.62962496e-01 -8.72057498e-01
-3.10150981e-01 1.19631994e+00 1.20551717e-02 2.32726917e-01
9.65725958e-01 -4.22307968e-01 -4.30004209e-01 -6.23599254e-02
1.13821769e+00 1.41167819e-01 -1.52236176e+00 -4.80699956e-01
-4.22840774e-01 -8.93105745e-01 7.63936564e-02 -4.10001427e-01
-1.81042480e+00 6.03905976e-01 9.97620165e-01 -1.55320719e-01
1.73984289e+00 -1.06114909e-01 9.27539289e-01 -5.90713434e-02
1.78177893e-01 -1.10522711e+00 3.14548284e-01 1.26425087e-01
8.46818805e-01 -9.35879946e-01 1.11311980e-01 -5.54788411e-01
-6.84599578e-01 1.07480454e+00 9.69291151e-01 -1.11112848e-01
3.36097449e-01 1.22710161e-01 3.34915966e-01 8.32851604e-02
-1.09448686e-01 -7.83912539e-02 5.18602312e-01 4.92906332e-01
1.98197924e-02 -1.60399258e-01 -1.86833873e-01 1.77235574e-01
1.06260337e-01 -4.23124909e-01 -2.36306246e-03 4.34954941e-01
-4.33701038e-01 -8.34783614e-01 -4.78023112e-01 -4.97157387e-02
-2.34878957e-01 -6.06641523e-04 -1.96172167e-02 8.49314809e-01
5.59282601e-01 8.84446979e-01 3.28013361e-01 -2.97960103e-01
4.78347212e-01 -5.91098309e-01 3.19271088e-01 1.43835902e-01
-4.53202963e-01 5.12967765e-01 -5.73046803e-01 -9.32029903e-01
-6.20307922e-01 -4.38480198e-01 -1.25442743e+00 -4.11780506e-01
-2.32805386e-01 -3.61981332e-01 2.73195267e-01 5.20574570e-01
2.57449329e-01 9.08004224e-01 7.67868400e-01 -9.28805470e-01
-1.40099928e-01 -4.18307185e-01 -5.92867672e-01 5.37460089e-01
-4.33237329e-02 -5.48365533e-01 -1.42414123e-01 3.23286802e-01] | [11.180508613586426, -1.3152751922607422] |
550c0e15-2ea0-42f8-a5cf-ee4be176c94a | large-scale-unsupervised-audio-pre-training | 2306.15464 | null | https://arxiv.org/abs/2306.15464v1 | https://arxiv.org/pdf/2306.15464v1.pdf | Large-scale unsupervised audio pre-training for video-to-speech synthesis | Video-to-speech synthesis is the task of reconstructing the speech signal from a silent video of a speaker. Most established approaches to date involve a two-step process, whereby an intermediate representation from the video, such as a spectrogram, is extracted first and then passed to a vocoder to produce the raw audio. Some recent work has focused on end-to-end synthesis, whereby the generation of raw audio and any intermediate representations is performed jointly. All such approaches involve training on data from almost exclusively audio-visual datasets, i.e. every audio sample has a corresponding video sample. This precludes the use of abundant audio-only datasets which may not have a corresponding visual modality (e.g. audiobooks, radio podcasts, speech recognition datasets etc.), as well as audio-only architectures that have been developed by the audio machine learning community over the years. In this paper we propose to train encoder-decoder models on more than 3,500 hours of audio data at 24kHz, and then use the pre-trained decoders to initialize the audio decoders for the video-to-speech synthesis task. The pre-training step uses audio samples only and does not require labels or corresponding samples from other modalities (visual, text). We demonstrate that this pre-training step improves the reconstructed speech and that it is an unexplored way to improve the quality of the generator in a cross-modal task while only requiring samples from one of the modalities. We conduct experiments using both raw audio and mel spectrograms as target outputs and benchmark our models with existing work. | ['Maja Pantic', 'Yannis Panagakis', 'Triantafyllos Kefalas'] | 2023-06-27 | null | null | null | null | ['speech-synthesis'] | ['speech'] | [ 5.87888241e-01 1.76676303e-01 7.43070692e-02 -1.64273560e-01
-1.29830718e+00 -6.31869197e-01 8.53625953e-01 -1.50859624e-01
-1.00502558e-01 6.23096228e-01 4.03756291e-01 -2.34644681e-01
4.66856211e-01 -4.27838951e-01 -9.06422794e-01 -5.18943131e-01
8.72011185e-02 2.13548884e-01 2.09961772e-01 2.23927889e-02
-3.32797207e-02 2.18880892e-01 -2.06705117e+00 8.08476627e-01
2.76530921e-01 1.07220066e+00 2.96258807e-01 1.33134735e+00
-1.32546797e-01 8.50418389e-01 -7.84366012e-01 -1.18954152e-01
1.48953199e-01 -8.09912443e-01 -6.65603936e-01 3.71649384e-01
2.63961643e-01 -4.45868343e-01 -4.56009328e-01 7.81735539e-01
5.70879757e-01 1.93808272e-01 6.55937076e-01 -1.25984085e+00
-3.28886628e-01 5.65455198e-01 -1.06884703e-01 -1.11080974e-01
8.36879492e-01 1.76751107e-01 9.72864449e-01 -1.06499934e+00
6.08199179e-01 1.24859595e+00 2.51239151e-01 5.71013987e-01
-1.09739375e+00 -5.22077024e-01 -2.29854777e-01 1.80257767e-01
-1.37620330e+00 -1.12072623e+00 8.44776988e-01 -4.66101259e-01
9.73560691e-01 2.40167156e-01 6.00478172e-01 1.52379143e+00
-2.86704212e-01 7.19508231e-01 9.09698844e-01 -7.26765156e-01
2.53236115e-01 1.59311384e-01 -5.57713568e-01 4.05896544e-01
-5.25374770e-01 3.19298089e-01 -8.35902631e-01 -4.36837338e-02
6.36905372e-01 -4.58162010e-01 -4.97991472e-01 5.26524372e-02
-1.32405114e+00 6.23144865e-01 -1.13245554e-01 2.08816215e-01
-3.47254723e-01 9.71773639e-02 5.32882333e-01 6.69691741e-01
2.78952807e-01 -4.72392552e-02 -1.26369625e-01 -2.89493024e-01
-1.45291674e+00 1.35788694e-01 9.10244584e-01 9.00740981e-01
6.34758115e-01 4.52832341e-01 -5.17182387e-02 9.19922471e-01
4.84820634e-01 3.68030608e-01 6.07180536e-01 -9.71420288e-01
8.00784588e-01 -7.08306208e-02 3.57538834e-02 -4.57885563e-01
5.90041801e-02 1.07473936e-02 -5.36823273e-01 2.88773805e-01
3.92788768e-01 -2.38696367e-01 -1.07529294e+00 1.57640958e+00
3.14268500e-01 5.37499964e-01 2.88107395e-01 8.10010612e-01
1.02286232e+00 1.21293283e+00 -2.55499810e-01 -3.14501137e-01
1.15862775e+00 -1.03718221e+00 -7.33837605e-01 -1.14577606e-01
2.69196182e-01 -1.04961848e+00 1.12180841e+00 6.20242357e-01
-1.23172927e+00 -8.33201349e-01 -1.11800694e+00 -4.71412539e-02
-1.92947388e-01 3.51549834e-01 -1.38880521e-01 5.05310237e-01
-1.11967969e+00 4.71496761e-01 -7.98324943e-01 -1.63105056e-01
1.31017494e-03 1.31423727e-01 -5.02968311e-01 3.70751135e-02
-1.19771504e+00 4.87467796e-01 4.14601386e-01 -1.65359974e-01
-1.52692270e+00 -4.98049438e-01 -1.01862979e+00 9.92726386e-02
3.97339165e-01 -5.05749643e-01 1.57278061e+00 -1.47112727e+00
-2.11714554e+00 5.89255869e-01 -2.34531999e-01 -4.95724410e-01
4.44587618e-01 7.91718066e-02 -6.83690727e-01 4.85688955e-01
-2.12745458e-01 8.27247024e-01 1.50851190e+00 -1.15719700e+00
-6.95875287e-01 1.59459963e-01 1.51776358e-01 1.53129265e-01
-2.01190189e-01 1.42547369e-01 -4.60981250e-01 -7.49155998e-01
-2.32279003e-01 -1.01326108e+00 2.54176348e-01 -2.55535364e-01
-4.91194874e-01 7.36352755e-03 8.29164386e-01 -9.32754517e-01
1.04872143e+00 -2.41936660e+00 2.60696650e-01 3.67559604e-02
-2.68404484e-01 1.57872930e-01 -4.11908925e-01 7.51726806e-01
-3.70469600e-01 -8.51491988e-02 -1.99632511e-01 -7.33826399e-01
-9.17448779e-04 3.16638243e-03 -6.13484561e-01 4.14187163e-01
3.12949717e-01 4.01597172e-01 -7.64772594e-01 -4.77014422e-01
2.74308115e-01 8.46659660e-01 -5.64761043e-01 5.19042253e-01
-3.33071321e-01 6.54649079e-01 4.24420607e-04 5.35883367e-01
2.37825528e-01 7.48376325e-02 9.08787102e-02 -5.28657138e-02
-1.25822216e-01 7.61304915e-01 -1.32510757e+00 1.89651239e+00
-7.80958474e-01 8.78604472e-01 2.87682086e-01 -9.28614736e-01
7.87521362e-01 1.15398788e+00 3.40993762e-01 -4.31703180e-01
2.33073179e-02 4.21616733e-01 -2.21115425e-01 -5.92704654e-01
3.74140054e-01 -3.07518423e-01 2.26580799e-02 3.53253812e-01
4.94781405e-01 -3.08613896e-01 2.57661998e-01 -3.26949693e-02
9.67684150e-01 2.56292492e-01 1.13493808e-01 3.93999308e-01
7.12670386e-01 -1.67964071e-01 1.32961303e-01 4.34022188e-01
1.93855107e-01 1.04738712e+00 2.73184299e-01 2.04393297e-01
-1.35431004e+00 -1.04637873e+00 1.70807943e-01 1.15615594e+00
-4.30795521e-01 -7.31131256e-01 -8.43005240e-01 -4.74510431e-01
-4.53771740e-01 5.71062386e-01 -1.69825777e-01 -1.03746420e-02
-4.27327037e-01 3.90781313e-02 7.02294469e-01 3.44839275e-01
-3.98280360e-02 -1.15050852e+00 -3.27010930e-01 3.23803365e-01
-3.34967107e-01 -1.23648751e+00 -3.85857671e-01 1.43046409e-01
-6.91094339e-01 -6.90237105e-01 -8.72748792e-01 -7.25877047e-01
3.94355685e-01 5.12477243e-03 8.82401168e-01 -2.39830390e-01
7.69439787e-02 3.75111639e-01 -5.18343806e-01 -2.58262962e-01
-9.99983728e-01 -2.40996387e-03 6.06368296e-02 4.73724991e-01
-9.40317810e-02 -6.90981567e-01 -2.31780127e-01 9.10909697e-02
-1.02373111e+00 2.53855258e-01 4.79513049e-01 7.64443278e-01
6.08992994e-01 5.28382249e-02 7.92794704e-01 -4.70591098e-01
4.24788952e-01 -5.43029308e-01 -5.47140181e-01 -9.66492891e-02
1.24472633e-01 -3.00493427e-02 1.04013503e+00 -6.18031740e-01
-9.31272447e-01 4.88548756e-01 -4.37064826e-01 -8.94194126e-01
-4.38938409e-01 6.43736601e-01 -2.59685397e-01 3.64574850e-01
5.47107279e-01 3.47272992e-01 1.83113106e-02 -6.24323428e-01
2.91872948e-01 1.16356218e+00 7.20687628e-01 -3.54989976e-01
7.34379649e-01 2.55761623e-01 -2.48813659e-01 -1.26979744e+00
-3.65624994e-01 -3.46588850e-01 -3.40829462e-01 -2.80580163e-01
7.11848795e-01 -1.28684914e+00 -3.48340601e-01 2.89826781e-01
-1.21504009e+00 -3.96023184e-01 -3.18792641e-01 7.52505541e-01
-7.67678440e-01 2.47440457e-01 -5.07121265e-01 -8.74071896e-01
-6.83379620e-02 -1.32054377e+00 1.25198412e+00 -8.90297145e-02
-2.67088890e-01 -6.83093548e-01 1.03691176e-01 3.40705663e-01
1.34231210e-01 -2.66684126e-02 5.15598536e-01 -5.85655332e-01
-5.30972660e-01 -1.47775829e-01 2.60654092e-01 6.94384158e-01
1.14153408e-01 2.26631358e-01 -1.43911397e+00 -2.99552679e-01
-6.62434334e-03 -5.75293660e-01 7.96177268e-01 2.01847360e-01
9.90226209e-01 -3.32965851e-01 7.15608597e-02 4.61702436e-01
1.04099226e+00 3.12383175e-01 6.06969714e-01 -1.51427150e-01
5.32194376e-01 6.59972608e-01 4.65017051e-01 4.32945669e-01
3.24758321e-01 7.50574589e-01 3.68490458e-01 9.79827568e-02
-4.91846204e-01 -5.86238623e-01 1.01647472e+00 1.20111120e+00
3.72068491e-04 -3.85090888e-01 -6.79245353e-01 7.12869883e-01
-1.48674500e+00 -1.11463022e+00 1.48544550e-01 2.36996126e+00
1.01284349e+00 -9.61575843e-03 2.98305094e-01 6.51599884e-01
7.36663520e-01 1.97672546e-01 -1.94093555e-01 -2.92057574e-01
1.76752791e-01 4.29767489e-01 -1.94379780e-02 5.86652219e-01
-1.08946824e+00 6.01767898e-01 6.20803452e+00 7.79227614e-01
-1.44045091e+00 1.54966891e-01 3.27280790e-01 -3.69037718e-01
-2.78043956e-01 7.39824101e-02 -5.68489671e-01 5.49894214e-01
1.70628190e+00 2.65814587e-02 7.85198748e-01 6.70070112e-01
4.83614534e-01 4.85116243e-02 -1.41830277e+00 1.15330291e+00
5.96401431e-02 -1.22792065e+00 9.89553705e-02 -9.20558125e-02
5.51572382e-01 -5.79061806e-02 -6.97214231e-02 4.08145696e-01
-2.56091934e-02 -1.18629360e+00 1.24954021e+00 3.73499751e-01
1.17279470e+00 -6.93389535e-01 3.14287990e-01 4.61229205e-01
-1.42670190e+00 7.97331184e-02 1.54738659e-02 7.51328766e-02
4.25564647e-01 4.07428235e-01 -1.07643974e+00 6.20105147e-01
5.20697415e-01 5.53206921e-01 -2.23795354e-01 8.69603336e-01
-1.98973194e-01 9.46888208e-01 -4.37323719e-01 3.61754626e-01
1.89540803e-01 2.26591021e-01 5.99472463e-01 1.33013046e+00
7.44849265e-01 -2.84879029e-01 2.15296060e-01 5.17429233e-01
-1.79559782e-01 9.75119546e-02 -7.20338345e-01 -3.35284591e-01
5.11718452e-01 1.09617543e+00 -4.55398560e-01 -4.79245454e-01
-7.36232281e-01 8.45217526e-01 -2.44128972e-01 4.38177228e-01
-9.34805930e-01 -5.36468387e-01 4.19254601e-01 2.39516020e-01
4.02758777e-01 -1.67049125e-01 2.61456102e-01 -1.15595126e+00
3.83941419e-02 -1.24619174e+00 1.74367338e-01 -9.42573905e-01
-8.75455499e-01 7.75442779e-01 -9.68651846e-02 -1.60139585e+00
-1.04297316e+00 -3.26219440e-01 -4.90847558e-01 9.41438079e-01
-1.48865366e+00 -9.65009630e-01 -7.84221888e-02 8.15686047e-01
9.06823039e-01 -3.95805985e-01 7.31730103e-01 5.59398293e-01
-2.67905354e-01 3.03815901e-01 -1.18911766e-01 7.39500001e-02
7.63400555e-01 -9.77134585e-01 3.09751749e-01 7.71993041e-01
5.49115479e-01 1.59243703e-01 7.55892634e-01 -3.11690181e-01
-1.49408722e+00 -1.07585764e+00 9.22511041e-01 -1.58183903e-01
6.18961811e-01 -5.89522123e-01 -7.90023446e-01 5.93325257e-01
4.63241071e-01 6.21670559e-02 6.32308662e-01 -5.23869455e-01
-1.64797246e-01 -6.53434545e-02 -7.54426837e-01 3.42598736e-01
7.08290935e-01 -1.14709735e+00 -5.90673149e-01 4.92674708e-02
7.44581640e-01 -5.11235476e-01 -6.87170744e-01 4.18585353e-02
4.29931074e-01 -8.54544342e-01 9.61883247e-01 -3.74956608e-01
6.00991249e-01 -5.68761885e-01 -4.20746982e-01 -1.37850595e+00
2.56946981e-01 -9.29953635e-01 -2.87512213e-01 1.54130757e+00
5.37831545e-01 -1.89274803e-01 4.40359265e-01 -9.50270221e-02
-3.58358949e-01 -2.76914030e-01 -1.09730697e+00 -6.25531197e-01
-3.12102437e-01 -8.16678822e-01 5.40163755e-01 6.26967728e-01
-9.28682089e-02 5.10574520e-01 -6.12859428e-01 2.37510160e-01
2.28212252e-01 2.33675912e-02 9.19152558e-01 -9.40236747e-01
-5.93140364e-01 -6.30997727e-03 -2.03747645e-01 -1.06745183e+00
3.21534008e-01 -9.27502394e-01 2.82060146e-01 -1.32327747e+00
-3.99072200e-01 8.47969111e-03 2.04788707e-02 3.78384262e-01
2.87307590e-01 3.29235613e-01 3.07689190e-01 6.82316795e-02
-1.05862141e-01 5.51146507e-01 1.03466129e+00 -1.49268240e-01
-3.18017900e-01 3.61577421e-02 -1.92857519e-01 5.53521693e-01
5.14217377e-01 -5.76912463e-01 -6.93094075e-01 -2.35089839e-01
2.48490218e-02 7.29880571e-01 5.53949893e-01 -1.17760646e+00
6.61546811e-02 1.85431968e-02 2.09709361e-01 -4.37141508e-01
8.19985151e-01 -8.34526122e-01 4.65853393e-01 5.05493358e-02
-3.48887771e-01 -1.37001008e-01 4.12225584e-03 5.08193612e-01
-7.45131791e-01 -3.20397198e-01 6.37783170e-01 -9.02800411e-02
-4.04981017e-01 -1.72993150e-02 -6.40799403e-01 -3.39907557e-02
7.90591955e-01 -1.57467484e-01 2.36267317e-02 -7.77441919e-01
-1.01202965e+00 -3.21912140e-01 2.54802644e-01 4.57300693e-01
6.27063036e-01 -1.46632576e+00 -8.57414126e-01 4.00455892e-01
-2.29722317e-02 -1.01034962e-01 1.61503330e-01 7.60634542e-01
-3.21201771e-01 4.03694004e-01 -3.43671180e-02 -6.92936122e-01
-1.28873014e+00 4.42730457e-01 1.03092819e-01 1.09572425e-01
-5.51533878e-01 5.63581288e-01 6.56373426e-02 -8.53788629e-02
3.85070711e-01 -3.72575939e-01 -1.58035889e-01 3.17074239e-01
6.82118118e-01 1.39024153e-01 1.89355642e-01 -9.17777658e-01
-2.28556991e-01 2.37445235e-01 4.06415284e-01 -8.91262174e-01
1.28217304e+00 -6.53634667e-02 3.65218967e-01 7.89604008e-01
1.37076223e+00 2.02033266e-01 -1.37035775e+00 -5.04151583e-02
-2.46034876e-01 -3.31789821e-01 1.25293374e-01 -4.31605697e-01
-1.11522794e+00 1.16931295e+00 2.68194795e-01 4.41856056e-01
1.26351547e+00 -3.68456207e-02 7.38752723e-01 1.31614298e-01
3.15119952e-01 -1.11657596e+00 1.52916342e-01 4.32074934e-01
1.03938496e+00 -8.88210058e-01 -5.08803308e-01 -1.78266063e-01
-7.06557572e-01 1.29908049e+00 7.66129196e-02 5.94148748e-02
4.74219412e-01 4.88576531e-01 4.62852009e-02 3.35940570e-01
-1.11113286e+00 -2.54926831e-01 3.55104089e-01 5.13118088e-01
6.01989806e-01 -1.60483122e-01 1.41968757e-01 5.37742555e-01
-3.84126186e-01 2.22988039e-01 6.22702301e-01 9.25616324e-01
-3.11322093e-01 -1.23600566e+00 -6.54207766e-01 2.07788035e-01
-5.50344348e-01 -8.27425867e-02 -3.74501854e-01 3.60917330e-01
-1.44047113e-02 1.26655638e+00 1.86543819e-02 -5.08883715e-01
2.44945407e-01 5.34320116e-01 4.33762550e-01 -7.60915399e-01
-5.48243463e-01 5.28967917e-01 3.94518077e-01 -4.40212458e-01
-4.79364514e-01 -6.76385224e-01 -1.22399163e+00 -1.42227858e-01
-1.56159803e-01 1.87527031e-01 7.98938632e-01 8.03136110e-01
2.29868919e-01 6.76181138e-01 7.90172577e-01 -1.35690236e+00
-4.16690409e-01 -1.01084018e+00 -3.43217939e-01 2.23513827e-01
8.15144718e-01 -4.20774937e-01 -4.84492809e-01 7.47551322e-01] | [15.112496376037598, 5.127927303314209] |
76ab64f8-a61b-4d69-b573-721d3ab5baf4 | multi-directional-multi-level-dual-cross | 1401.5311 | null | https://arxiv.org/abs/1401.5311v2 | https://arxiv.org/pdf/1401.5311v2.pdf | Multi-Directional Multi-Level Dual-Cross Patterns for Robust Face Recognition | To perform unconstrained face recognition robust to variations in illumination, pose and expression, this paper presents a new scheme to extract "Multi-Directional Multi-Level Dual-Cross Patterns" (MDML-DCPs) from face images. Specifically, the MDMLDCPs scheme exploits the first derivative of Gaussian operator to reduce the impact of differences in illumination and then computes the DCP feature at both the holistic and component levels. DCP is a novel face image descriptor inspired by the unique textural structure of human faces. It is computationally efficient and only doubles the cost of computing local binary patterns, yet is extremely robust to pose and expression variations. MDML-DCPs comprehensively yet efficiently encodes the invariant characteristics of a face image from multiple levels into patterns that are highly discriminative of inter-personal differences but robust to intra-personal variations. Experimental results on the FERET, CAS-PERL-R1, FRGC 2.0, and LFW databases indicate that DCP outperforms the state-of-the-art local descriptors (e.g. LBP, LTP, LPQ, POEM, tLBP, and LGXP) for both face identification and face verification tasks. More impressively, the best performance is achieved on the challenging LFW and FRGC 2.0 databases by deploying MDML-DCPs in a simple recognition scheme. | ['DaCheng Tao', 'Larry S. Davis', 'Changxing Ding', 'Jonghyun Choi'] | 2014-01-21 | null | null | null | null | ['robust-face-recognition'] | ['computer-vision'] | [ 3.06117013e-02 -7.46693909e-01 -1.53351068e-01 -5.56089997e-01
-5.69295704e-01 -3.79305869e-01 5.63105524e-01 -2.94770926e-01
-1.41151631e-02 4.81836736e-01 -2.31804028e-01 2.03869551e-01
-4.64155674e-01 -6.79786682e-01 -4.15848881e-01 -1.16188693e+00
-3.27372164e-01 -8.44262689e-02 -3.72215696e-02 -2.00861946e-01
3.22253108e-01 1.36403561e+00 -2.09685564e+00 5.25708675e-01
2.53372043e-01 1.70942974e+00 -3.91378224e-01 1.67881995e-01
1.63709179e-01 3.33731622e-01 -4.61611331e-01 -5.40971160e-01
3.32967401e-01 -4.33904678e-02 -3.70805264e-01 7.30130449e-02
9.19071436e-01 6.85114190e-02 -1.51969552e-01 1.06273580e+00
8.80682707e-01 -1.29944071e-01 6.99872732e-01 -1.27254236e+00
-8.80378962e-01 -3.41390580e-01 -7.47866929e-01 2.51913577e-01
7.17279196e-01 5.44459410e-02 4.55518425e-01 -1.34358430e+00
5.31781852e-01 1.77060890e+00 9.54718530e-01 4.05161053e-01
-1.30787933e+00 -8.86667252e-01 -2.73620665e-01 4.71427798e-01
-1.88922632e+00 -6.08486533e-01 8.50815833e-01 -3.21450680e-01
7.74439514e-01 4.32329625e-01 1.11008577e-01 1.02953672e+00
4.61992413e-01 2.17557192e-01 1.57227039e+00 -6.25718176e-01
-1.14712700e-01 -1.89467594e-01 6.55489415e-02 1.16173744e+00
-2.37542689e-02 2.00165197e-01 -8.43413174e-01 -5.51406443e-01
5.25017560e-01 1.06384486e-01 -3.66073310e-01 -8.69896412e-02
-5.52977383e-01 4.97138977e-01 5.14830425e-02 5.35653651e-01
-3.11340362e-01 -3.26577991e-01 3.32155049e-01 4.16743100e-01
3.53939146e-01 -2.54932195e-01 -3.72846782e-01 1.69287220e-01
-8.46961915e-01 5.54342172e-04 6.40020490e-01 6.72043085e-01
9.67080474e-01 -8.88702571e-02 -4.37307924e-01 1.33625913e+00
3.53492409e-01 9.78147984e-01 5.14338493e-01 -7.09209442e-01
8.21473300e-02 5.10111511e-01 -2.21921489e-01 -1.79258573e+00
-2.15646580e-01 2.37635486e-02 -8.27901423e-01 1.19676590e-01
-4.76698689e-02 2.95233607e-01 -7.32941151e-01 1.46721244e+00
2.85065413e-01 1.51693255e-01 -2.17016846e-01 4.41289544e-01
9.52557981e-01 3.81487042e-01 -6.98640272e-02 -3.18556845e-01
1.58580410e+00 -4.36705500e-01 -8.15625429e-01 2.56746650e-01
-6.67538196e-02 -9.73613381e-01 5.54849625e-01 3.04033160e-01
-7.38805354e-01 -1.00844014e+00 -1.03312528e+00 2.57939875e-01
-6.35294080e-01 6.67992830e-01 3.14184219e-01 1.14873528e+00
-1.24865055e+00 5.00559032e-01 -4.05088574e-01 -3.17557514e-01
4.50333059e-01 6.72395766e-01 -8.92838836e-01 -3.54816616e-01
-7.89093494e-01 5.87674141e-01 7.85780624e-02 3.26201588e-01
-4.70858485e-01 -5.86513162e-01 -8.06470692e-01 -6.79472461e-02
1.02043301e-01 3.86234850e-01 3.71589363e-01 -8.65554035e-01
-1.65763187e+00 1.01113498e+00 -5.53784609e-01 1.81911379e-01
-7.18827620e-02 3.59799176e-01 -9.82386470e-01 4.99603152e-01
-5.20950109e-02 4.26879674e-01 1.25271523e+00 -1.08145702e+00
-1.08435988e-01 -8.27270746e-01 -5.88727951e-01 -4.08997446e-01
-3.36910754e-01 5.97847283e-01 -4.50469375e-01 -6.98348403e-01
4.57482375e-02 -8.22806418e-01 6.13998771e-01 2.63872087e-01
-2.71936879e-02 -5.56814671e-01 1.52813268e+00 -7.69253612e-01
9.46078956e-01 -2.48079681e+00 -3.17126334e-01 7.22321987e-01
-4.52106267e-01 7.21617460e-01 -4.71541017e-01 3.10009986e-01
-2.93339431e-01 -1.16990186e-01 -2.30365079e-02 -2.07740188e-01
3.92724276e-02 2.97641486e-01 -3.63375060e-02 7.45698750e-01
4.20074314e-01 6.68182552e-01 -3.86134595e-01 -6.68048918e-01
1.28496304e-01 9.68365669e-01 -3.60163391e-01 -7.65015781e-02
4.16380227e-01 2.02542737e-01 -2.05694586e-01 1.28679955e+00
1.34862447e+00 3.90341848e-01 2.41682827e-01 -5.50470650e-01
-2.66707130e-02 -4.54006046e-01 -1.23001337e+00 9.66060936e-01
-3.08084339e-01 5.34322739e-01 3.33959371e-01 -1.02387094e+00
1.35772836e+00 5.68701744e-01 4.90087867e-01 -8.60134363e-01
1.88016415e-01 3.03686321e-01 -4.99510825e-01 -4.78953332e-01
-7.79188424e-02 4.20774557e-02 2.15938702e-01 -1.97049543e-01
5.60561776e-01 4.94985163e-01 3.12307149e-01 -6.04308665e-01
5.31740487e-01 -3.09154298e-02 3.33997309e-01 -6.90583408e-01
1.44435322e+00 -1.02038133e+00 7.39569426e-01 3.56485009e-01
-4.45801675e-01 4.02452618e-01 4.01541054e-01 -4.37352389e-01
-1.86476678e-01 -1.13763118e+00 -7.27620542e-01 9.83586848e-01
-1.04718037e-01 -4.15611088e-01 -6.18283033e-01 -6.05777264e-01
4.03308809e-01 -4.82164025e-02 -6.61980569e-01 -6.14344850e-02
-5.29045284e-01 -8.49161685e-01 7.33251452e-01 2.91222990e-01
1.00607002e+00 -9.03500855e-01 -6.84451312e-02 -1.46304727e-01
1.88916485e-04 -1.28524208e+00 -5.63151538e-01 -3.56147051e-01
-3.57729107e-01 -1.25948572e+00 -4.53636259e-01 -1.04815829e+00
7.51116335e-01 5.56049310e-03 7.01485276e-01 1.15837835e-01
-1.00438130e+00 6.48519278e-01 -3.06851476e-01 -2.96391472e-02
1.12165488e-01 -8.02114606e-01 2.21008390e-01 9.00241375e-01
6.31357849e-01 -3.79026264e-01 -4.20290351e-01 6.95258498e-01
-5.28445601e-01 -1.07704496e+00 4.51959044e-01 1.09264302e+00
6.95271790e-01 3.14840019e-01 2.70446420e-01 -3.08130801e-01
5.03107071e-01 -9.00911242e-02 -4.83408391e-01 5.89515209e-01
-3.31879795e-01 -2.24773794e-01 5.76261520e-01 -3.29840541e-01
-1.26179016e+00 -4.33552489e-02 -2.12074667e-01 -4.12533641e-01
-3.47640216e-01 1.27139986e-01 -5.07661402e-01 -1.10860431e+00
4.70500559e-01 4.66284871e-01 1.39617890e-01 -6.25759363e-01
-1.40977427e-01 8.40109825e-01 6.25750899e-01 -9.40835238e-01
6.66837633e-01 5.45873523e-01 3.92486095e-01 -1.45913386e+00
-1.52390748e-01 -5.57366073e-01 -8.65944684e-01 -2.44233146e-01
4.99192178e-01 -7.00373590e-01 -1.16537046e+00 8.54755878e-01
-9.20807362e-01 3.67545426e-01 1.20696619e-01 1.42105773e-01
-2.34698966e-01 7.21556485e-01 -5.67449510e-01 -7.86674082e-01
-2.98957258e-01 -1.22560811e+00 1.26142633e+00 2.26688176e-01
3.85102868e-01 -7.59374857e-01 -2.11299032e-01 1.46928921e-01
5.55801094e-01 5.34962356e-01 6.69305384e-01 -1.12963878e-01
-1.46828875e-01 -3.87557924e-01 -3.12662035e-01 7.83337951e-01
4.15734887e-01 4.02382493e-01 -1.04278958e+00 -5.47892392e-01
6.38030022e-02 -2.72062063e-01 5.68549335e-01 -1.37532223e-02
1.37618685e+00 -4.96930987e-01 -1.93415731e-01 7.77529299e-01
1.55709720e+00 3.46161157e-01 7.38862753e-01 -1.75657436e-01
2.49903530e-01 6.38347149e-01 3.77763450e-01 4.61164773e-01
-1.04238763e-02 1.07826138e+00 -2.98052449e-02 8.16494673e-02
-2.89269835e-01 9.15279686e-02 5.39209366e-01 4.25876915e-01
-3.49140763e-01 2.98988253e-01 -5.81273735e-01 2.09553719e-01
-1.23350358e+00 -1.22802687e+00 1.92179039e-01 1.98733723e+00
7.15454459e-01 -6.07619703e-01 -1.30977377e-01 3.04698110e-01
7.70530164e-01 2.11434409e-01 3.98271903e-02 -6.66728497e-01
-5.95954359e-01 9.48948741e-01 3.30144733e-01 2.07136840e-01
-1.24914777e+00 6.35088503e-01 5.70194769e+00 1.22694480e+00
-1.43883920e+00 1.96942672e-01 6.40420914e-01 3.89111519e-01
5.70272148e-01 -5.17160058e-01 -1.02675104e+00 6.39851093e-01
7.92973280e-01 4.34953541e-01 3.97224814e-01 7.81599820e-01
-2.24460706e-01 9.74857807e-02 -8.41143906e-01 1.43745601e+00
5.67336023e-01 -8.46180439e-01 7.73291513e-02 7.01153427e-02
6.60979033e-01 -3.75796318e-01 4.99857128e-01 8.72274339e-02
-6.12605691e-01 -1.13080335e+00 3.34716022e-01 5.72583914e-01
9.68653917e-01 -8.38123500e-01 7.86995828e-01 -1.86651170e-01
-1.58387363e+00 -3.06609154e-01 -5.96108675e-01 4.69033629e-01
-5.88387191e-01 4.03328747e-01 -1.96142808e-01 7.54519045e-01
9.55209494e-01 5.91316462e-01 -8.42228293e-01 7.30489910e-01
-3.90081145e-02 2.41868779e-01 -3.67221057e-01 2.61131853e-01
-8.04400891e-02 -1.49210125e-01 2.84219921e-01 1.56109774e+00
5.94367564e-01 9.90384594e-02 1.32867455e-01 5.60599327e-01
-6.65695667e-02 2.91039199e-01 -3.51250142e-01 1.17822453e-01
5.22098899e-01 1.38686442e+00 -3.03199947e-01 4.86055668e-03
-4.04601872e-01 8.95800948e-01 -4.78504002e-02 3.13221753e-01
-4.91786778e-01 -4.96503741e-01 8.54745388e-01 -1.21252209e-01
7.71453738e-01 -1.62924677e-01 2.99606919e-01 -8.33363414e-01
2.66726494e-01 -1.08597839e+00 4.21635658e-01 -3.44353437e-01
-1.45568585e+00 7.25505531e-01 6.88420841e-03 -9.60582674e-01
8.54230002e-02 -1.29452956e+00 -6.29406691e-01 1.01466107e+00
-1.76601684e+00 -1.45517457e+00 -3.33794981e-01 1.22450519e+00
-1.12064546e-02 -5.07739186e-01 1.04666865e+00 5.48226953e-01
-6.21333838e-01 1.22190416e+00 1.65576071e-01 2.45753720e-01
7.19947755e-01 -5.78899741e-01 -2.94878602e-01 5.87254882e-01
-6.29234463e-02 8.81666481e-01 1.49227083e-01 -2.97379255e-01
-1.64653242e+00 -8.95904779e-01 8.85186970e-01 -6.28296286e-02
1.19036160e-01 -1.68616220e-01 -7.27851093e-01 1.92271501e-01
2.44931132e-03 5.96904159e-01 8.91519368e-01 -2.06650794e-01
-8.62674952e-01 -8.03500175e-01 -1.83449459e+00 1.22416466e-01
8.17541599e-01 -9.69191849e-01 -2.46790841e-01 3.04038882e-01
-2.79698163e-01 6.03412278e-02 -1.36834049e+00 7.06547976e-01
8.05668354e-01 -1.22845042e+00 1.28019822e+00 3.31255049e-02
-3.84769291e-01 -3.75661612e-01 -6.02495074e-01 -7.12307155e-01
-3.42133522e-01 -6.19814277e-01 9.84142125e-02 1.51021135e+00
-1.18844405e-01 -9.15569186e-01 3.29611212e-01 1.65909067e-01
2.30449632e-01 -7.54684210e-01 -1.42169380e+00 -1.06183887e+00
-3.99656445e-01 -8.84443074e-02 7.07308650e-01 7.75264382e-01
-1.30765110e-01 -5.50446332e-01 -2.53811508e-01 3.32095534e-01
8.36099148e-01 1.39091030e-01 4.18531388e-01 -1.22376812e+00
2.14785077e-02 -3.53838027e-01 -9.72340703e-01 -4.60042775e-01
4.78444099e-01 -7.63847768e-01 -1.46672621e-01 -5.28629601e-01
1.55692929e-02 -1.44009978e-01 -4.59193617e-01 7.22990274e-01
2.03657478e-01 8.72230887e-01 1.05903484e-01 1.05325907e-01
-6.85801134e-02 4.49673295e-01 9.43144441e-01 -2.36818343e-01
2.24152669e-01 -2.58292973e-01 -1.85548484e-01 5.82434952e-01
6.25337005e-01 1.19808773e-02 4.06250693e-02 1.55499220e-01
-7.30559170e-01 -3.68508250e-01 5.01135647e-01 -1.27398574e+00
2.28337064e-01 4.74698143e-03 8.07959199e-01 -5.56576848e-01
7.08232343e-01 -8.22821319e-01 4.55036983e-02 4.16861385e-01
2.86150903e-01 1.08598366e-01 3.68419677e-01 3.40346783e-01
-6.74575388e-01 5.69808260e-02 1.20609117e+00 1.67493597e-01
-7.81902254e-01 4.00915772e-01 1.32759067e-03 -4.74695563e-01
9.90132868e-01 -4.87470776e-01 -4.03248638e-01 2.74648756e-01
-4.69724417e-01 -3.76768202e-01 1.92826942e-01 4.26575273e-01
8.70060802e-01 -1.57536507e+00 -6.06341183e-01 1.11586797e+00
1.45024240e-01 -8.71564090e-01 4.42195892e-01 9.42615330e-01
-5.07569075e-01 7.54096091e-01 -6.92910850e-01 -6.34598672e-01
-1.84923172e+00 4.62668985e-01 3.82198960e-01 1.59249768e-01
-8.33687335e-02 9.88368154e-01 -1.34148121e-01 -4.04369950e-01
2.53286481e-01 1.37001336e-01 -2.36475095e-01 2.34773338e-01
7.17351496e-01 5.26533723e-01 4.65769440e-01 -1.49710035e+00
-8.10031056e-01 1.35028899e+00 9.31157246e-02 4.54469919e-01
1.10433578e+00 7.92991221e-02 -7.13849962e-01 -2.66560942e-01
1.82934952e+00 1.68962806e-01 -9.08355296e-01 -1.75937787e-01
3.23743746e-02 -9.14562523e-01 -1.87298849e-01 -6.70334876e-01
-1.30803239e+00 8.16776931e-01 1.15955138e+00 -1.92621559e-01
1.54792249e+00 -2.87145078e-01 4.11602527e-01 2.78061390e-01
6.32461309e-01 -1.07178605e+00 2.76251584e-01 4.18818921e-01
1.27234399e+00 -8.13346028e-01 1.67461094e-02 -6.24218583e-01
-2.32462287e-01 1.42053819e+00 3.28850567e-01 -3.40360031e-02
1.22258592e+00 2.08689839e-01 -1.63071185e-01 -1.44769728e-01
-3.02190959e-01 1.75004050e-01 7.09790647e-01 7.87649751e-01
3.65626633e-01 2.05186810e-02 -3.22371036e-01 1.38819411e-01
8.01113918e-02 -9.08128619e-02 -3.15569401e-01 1.01118624e+00
-2.39383727e-01 -1.30077481e+00 -7.99411535e-01 1.01670787e-01
-7.44934857e-01 3.22688162e-01 -3.43425363e-01 6.06374085e-01
7.42578149e-01 9.81549382e-01 -9.99196544e-02 -4.52920914e-01
3.38697702e-01 3.24724793e-01 7.32633948e-01 -6.96206540e-02
-5.10404289e-01 -3.27501856e-02 -1.59743637e-01 -9.47441459e-01
-7.58676708e-01 -6.96183383e-01 -7.25692630e-01 -4.41028416e-01
-1.04997970e-01 -7.31581450e-02 6.86826825e-01 6.23269439e-01
6.08344376e-01 -1.20830908e-01 1.00762784e+00 -1.04580641e+00
-5.59489906e-01 -8.04037154e-01 -8.25586319e-01 4.73451704e-01
2.98938602e-01 -1.06749809e+00 -3.22763026e-01 5.52239493e-02] | [12.944083213806152, 0.6121460795402527] |
22534386-dee3-4b34-8740-1b28b6645b7d | inductive-topic-variational-graph-auto | null | null | https://aclanthology.org/2021.naacl-main.333 | https://aclanthology.org/2021.naacl-main.333.pdf | Inductive Topic Variational Graph Auto-Encoder for Text Classification | Graph convolutional networks (GCNs) have been applied recently to text classification and produced an excellent performance. However, existing GCN-based methods do not assume an explicit latent semantic structure of documents, making learned representations less effective and difficult to interpret. They are also transductive in nature, thus cannot handle out-of-graph documents. To address these issues, we propose a novel model named inductive Topic Variational Graph Auto-Encoder (T-VGAE), which incorporates a topic model into variational graph-auto-encoder (VGAE) to capture the hidden semantic information between documents and words. T-VGAE inherits the interpretability of the topic model and the efficient information propagation mechanism of VGAE. It learns probabilistic representations of words and documents by jointly encoding and reconstructing the global word-level graph and bipartite graphs of documents, where each document is considered individually and decoupled from the global correlation graph so as to enable inductive learning. Our experiments on several benchmark datasets show that our method outperforms the existing competitive models on supervised and semi-supervised text classification, as well as unsupervised text representation learning. In addition, it has higher interpretability and is able to deal with unseen documents. | ['Jian-Yun Nie', 'Min Peng', 'Pan Du', 'Jimin Huang', 'Qianqian Xie'] | 2021-06-01 | null | null | null | naacl-2021-4 | ['semi-supervised-text-classification-1'] | ['natural-language-processing'] | [-8.66230801e-02 4.77525771e-01 -2.35011265e-01 -3.63625228e-01
-2.17556581e-01 -4.25805420e-01 8.83709967e-01 5.09336889e-01
6.89389408e-02 3.76713544e-01 4.72055316e-01 -3.30475241e-01
-1.92643330e-01 -1.12035358e+00 -7.27091610e-01 -7.31846631e-01
9.93121639e-02 8.82269025e-01 2.35662106e-02 -1.69181600e-01
-2.33364142e-02 -1.02915920e-01 -1.18108046e+00 2.53639728e-01
8.90509665e-01 5.69733202e-01 1.60240740e-01 5.33528388e-01
-6.58772051e-01 8.95695865e-01 -4.01322097e-01 -5.58176637e-01
-4.03125942e-01 -5.25839627e-01 -9.31830227e-01 3.50659639e-01
-4.57442775e-02 -1.24921829e-01 -8.63204062e-01 1.07906651e+00
1.14551419e-02 2.33616918e-01 1.11512172e+00 -1.38932562e+00
-1.42633164e+00 1.05834639e+00 -5.18488824e-01 -9.79672894e-02
4.22269739e-02 -3.65136534e-01 1.74065733e+00 -7.06641853e-01
5.31756163e-01 1.60314178e+00 4.10172820e-01 3.90291601e-01
-1.26898420e+00 -2.84663081e-01 6.15165174e-01 1.97575629e-01
-1.24689794e+00 1.73038572e-01 1.12784755e+00 -4.39410567e-01
9.26209211e-01 -1.37537587e-02 7.60932982e-01 1.37897646e+00
3.81155640e-01 1.08022869e+00 5.36115408e-01 -4.06824619e-01
2.78943419e-01 1.78853534e-02 5.51960230e-01 9.20796037e-01
4.51651484e-01 -2.76943386e-01 -4.48284030e-01 -2.02081591e-01
6.08002365e-01 3.21753442e-01 -4.45494205e-01 -6.65941298e-01
-9.50122476e-01 1.29623997e+00 7.89579630e-01 5.54887056e-01
-2.51590401e-01 3.80255789e-01 4.48410541e-01 7.48021752e-02
7.86547840e-01 -1.73325222e-02 -5.58239706e-02 4.97386664e-01
-5.97617984e-01 -1.05061382e-01 8.20026934e-01 9.80563223e-01
8.56256366e-01 1.58539042e-01 -2.45879814e-01 6.73694313e-01
1.00656402e+00 3.52491170e-01 5.98740399e-01 -2.34490529e-01
4.65973467e-01 1.02273726e+00 -5.80848277e-01 -1.37243807e+00
-3.43557626e-01 -5.69799423e-01 -1.15311432e+00 -4.14473474e-01
-1.48348495e-01 1.78808302e-01 -1.11006796e+00 1.57274115e+00
3.87364719e-03 2.78886519e-02 7.62435123e-02 7.18863547e-01
1.17360175e+00 1.01584470e+00 1.90484762e-01 -1.51068568e-02
1.34630895e+00 -9.16464627e-01 -1.01556766e+00 -3.55471224e-01
7.15232909e-01 -2.42766693e-01 9.56917584e-01 7.65002221e-02
-6.07167125e-01 -3.18438977e-01 -1.07205737e+00 -3.46787363e-01
-6.22660756e-01 -1.35921203e-02 7.68827438e-01 3.45575750e-01
-1.05098462e+00 2.12951854e-01 -9.51064944e-01 -3.06860298e-01
4.85526145e-01 3.64550769e-01 -2.70149350e-01 -4.94096205e-02
-1.42838717e+00 3.37166578e-01 7.60521173e-01 1.19053230e-01
-7.99397886e-01 -2.16439813e-01 -1.17129338e+00 4.52358872e-01
2.58122474e-01 -9.59840000e-01 7.73221076e-01 -9.40884888e-01
-1.43156588e+00 5.32613397e-01 -1.32697925e-01 -3.14146549e-01
6.13498576e-02 1.68054610e-01 -3.15161735e-01 2.49726042e-01
8.17493051e-02 5.04586637e-01 1.01174057e+00 -1.45513356e+00
-5.87962009e-02 -5.90261996e-01 3.63288485e-02 1.93137094e-01
-7.34481454e-01 -6.01322532e-01 -4.73639935e-01 -7.67308950e-01
2.08239123e-01 -7.50233233e-01 4.22068276e-02 -3.53644967e-01
-7.20873594e-01 -6.51368976e-01 1.19946504e+00 -5.06364644e-01
1.10833430e+00 -2.00044727e+00 5.01857638e-01 1.78292260e-01
6.79072440e-01 9.26769525e-02 -1.61082253e-01 6.81521416e-01
2.65817549e-02 1.58658311e-01 -3.11177880e-01 -4.91586804e-01
2.55594045e-01 6.34614944e-01 -4.59530801e-01 2.97527283e-01
1.53945446e-01 1.37878954e+00 -1.08390915e+00 -6.06178284e-01
1.59498915e-01 7.73017585e-01 -6.45452142e-01 1.32555038e-01
-6.80571616e-01 1.63748816e-01 -8.71965826e-01 2.83435285e-01
5.22119105e-01 -9.95930731e-01 5.79329371e-01 -2.74315700e-02
6.27972305e-01 2.77087599e-01 -8.24718118e-01 1.64811981e+00
-3.30069691e-01 7.09046960e-01 -2.50512928e-01 -1.43295264e+00
9.37867880e-01 4.03874725e-01 2.66714752e-01 -3.55759948e-01
2.66255796e-01 -1.05537839e-01 -2.82751262e-01 -2.13764235e-01
4.40890282e-01 -1.11540578e-01 -1.56418420e-02 6.15759969e-01
5.75555861e-01 -2.26076879e-02 9.49513242e-02 9.51585472e-01
7.43113577e-01 -5.37356548e-02 2.36156970e-01 -3.73877645e-01
5.56132555e-01 -3.57098460e-01 1.79616526e-01 4.95669574e-01
2.52506822e-01 3.94178063e-01 8.21041822e-01 -3.08935791e-01
-7.46068954e-01 -1.11374402e+00 2.08021134e-01 1.17410684e+00
2.27106497e-01 -8.48717391e-01 -7.01762736e-01 -9.18803930e-01
5.17517105e-02 8.94761145e-01 -8.68157685e-01 -4.64299589e-01
-7.62866512e-02 -7.70418823e-01 1.37580439e-01 6.18825555e-01
2.69040674e-01 -8.26198697e-01 5.22714674e-01 9.65002403e-02
-4.11422074e-01 -1.09333444e+00 -4.40475762e-01 -1.61792780e-03
-8.87593329e-01 -1.00255990e+00 -5.02893448e-01 -8.71696115e-01
9.95333850e-01 2.59285986e-01 1.14254355e+00 1.38405323e-01
-4.06213012e-03 6.06535971e-01 -5.56047440e-01 -3.46002460e-01
-4.55607504e-01 1.92646593e-01 -2.72800922e-01 2.94384331e-01
5.00472903e-01 -4.77189124e-01 -3.86938781e-01 1.14274092e-01
-1.26241481e+00 1.64947763e-01 4.10599649e-01 1.23534799e+00
4.34830159e-01 2.49104977e-01 4.54545796e-01 -1.23710978e+00
8.43178034e-01 -7.36189842e-01 -5.18419147e-01 3.67401868e-01
-8.41781259e-01 3.95594865e-01 8.04298401e-01 -1.59989387e-01
-1.03283703e+00 -3.53640229e-01 6.92654774e-02 -3.88354450e-01
1.64079458e-01 8.52087379e-01 -1.39270514e-01 3.68047386e-01
4.26089853e-01 3.76157433e-01 -7.49590099e-02 -3.46239179e-01
5.81344962e-01 6.49566293e-01 8.71322155e-02 -4.09156919e-01
7.18675017e-01 6.30297363e-01 6.81284163e-03 -8.75339389e-01
-1.04106438e+00 -4.52705979e-01 -6.84259713e-01 5.32612205e-03
1.02930963e+00 -8.42093170e-01 -4.09075558e-01 2.34217033e-01
-1.25040376e+00 -7.71112964e-02 -1.20803475e-01 4.53167737e-01
-3.58849525e-01 6.24799490e-01 -7.13069797e-01 -6.79253995e-01
-4.02423680e-01 -9.52808380e-01 1.30937409e+00 -5.25324140e-03
1.76185921e-01 -1.80285203e+00 -1.59971677e-02 3.70018870e-01
1.43582478e-01 9.25590917e-02 1.41649866e+00 -7.69105554e-01
-5.63816309e-01 -1.95717648e-01 -3.23181838e-01 2.75094837e-01
2.20924169e-01 -8.98298100e-02 -9.01046455e-01 -4.55821663e-01
-3.36612821e-01 -2.85594642e-01 1.22498453e+00 3.21271569e-01
1.11609888e+00 -4.71968591e-01 -5.58964312e-01 3.66840631e-01
1.35149634e+00 -2.80050069e-01 4.47595209e-01 -7.49399662e-02
1.29574120e+00 5.26958287e-01 -1.25768885e-01 2.77595133e-01
6.80009484e-01 4.33568865e-01 5.98962069e-01 8.04966688e-02
2.57963669e-02 -5.79854846e-01 3.85612547e-01 1.44945860e+00
-2.40999516e-02 -9.99623477e-01 -8.76231790e-01 4.76295322e-01
-2.19192243e+00 -5.79980135e-01 -3.88779730e-01 1.55500567e+00
4.29737717e-01 8.79913494e-02 -2.01935321e-01 1.37727097e-01
9.38845992e-01 4.30563658e-01 -2.82325804e-01 -2.31542632e-01
-1.66911587e-01 -7.59392753e-02 8.87299329e-02 5.03917515e-01
-9.14958596e-01 1.08279729e+00 5.99055004e+00 6.44127131e-01
-7.19531655e-01 6.28139898e-02 4.59809214e-01 4.52042013e-01
-9.08960581e-01 8.59453380e-02 -5.01850486e-01 2.93133467e-01
8.88167322e-01 -2.24708989e-01 1.81828082e-01 7.05069125e-01
-2.33359337e-01 5.43174088e-01 -1.06177354e+00 6.96427464e-01
2.72993207e-01 -1.32289314e+00 9.02519047e-01 1.27999470e-01
7.87031174e-01 -7.57715181e-02 -3.84119861e-02 4.42551732e-01
5.15727460e-01 -1.00111639e+00 3.99431348e-01 5.72579980e-01
3.53655726e-01 -7.80678689e-01 8.97485197e-01 4.13276225e-01
-1.20838058e+00 1.50196835e-01 -6.86786532e-01 1.53638214e-01
-1.21061184e-01 6.87596738e-01 -7.79976428e-01 9.17132616e-01
4.51035380e-01 1.08554709e+00 -4.96486604e-01 2.84508795e-01
-6.13929868e-01 7.92131722e-01 2.77002100e-02 -3.61285985e-01
5.30471683e-01 -3.70280325e-01 6.24124825e-01 1.17159426e+00
4.94648069e-02 -4.58029322e-02 2.01054201e-01 1.28616560e+00
-4.18686002e-01 2.88812011e-01 -6.33336127e-01 -6.96829557e-01
-2.76017543e-02 1.10883355e+00 -8.80437553e-01 -3.86971891e-01
-4.76678729e-01 1.04777849e+00 5.07152140e-01 6.27395451e-01
-5.31419456e-01 -3.67664993e-01 2.31022313e-01 -6.75588772e-02
4.94612217e-01 -3.49763095e-01 1.39004648e-01 -1.42335999e+00
-1.71338603e-01 -3.57515007e-01 6.43730402e-01 -7.17037320e-01
-1.54424250e+00 6.91117346e-01 8.26307572e-03 -8.61333668e-01
-4.08699483e-01 -8.80610526e-01 -5.62457502e-01 6.06169820e-01
-1.49738467e+00 -1.37561452e+00 -3.20917755e-01 7.85391092e-01
3.94067764e-01 -2.39077777e-01 8.61163557e-01 -1.04736842e-01
-5.07853210e-01 4.23533767e-01 3.91626656e-01 3.84001613e-01
1.74174681e-01 -1.57324111e+00 3.60791564e-01 5.42377591e-01
6.28729403e-01 6.68273151e-01 3.72715324e-01 -6.73624039e-01
-1.49050140e+00 -1.31513071e+00 8.02756727e-01 -4.11347181e-01
8.75211716e-01 -8.40817511e-01 -1.17753732e+00 1.04534101e+00
3.44326347e-01 6.76131397e-02 8.81119549e-01 3.50630254e-01
-4.92631346e-01 2.30897501e-01 -5.36186874e-01 4.57306862e-01
8.58881712e-01 -7.73723364e-01 -7.60283530e-01 6.24680340e-01
1.07929587e+00 -8.81928429e-02 -7.34005213e-01 -3.78753580e-02
4.46664654e-02 -7.00927079e-01 7.34596908e-01 -7.40198374e-01
5.59082031e-01 -7.24725500e-02 3.99081670e-02 -1.53056765e+00
-6.15939736e-01 -4.50374424e-01 -5.98629117e-01 1.26247203e+00
3.64502162e-01 -7.64328063e-01 6.04031384e-01 2.04201397e-02
-1.46612860e-02 -5.05913079e-01 -6.03007913e-01 -6.23963654e-01
6.71638995e-02 -3.76119226e-01 5.98988414e-01 1.18882596e+00
1.88876063e-01 9.98469949e-01 -2.55032837e-01 2.04693094e-01
7.87722647e-01 2.91153610e-01 4.22120839e-01 -1.58396482e+00
-3.58342737e-01 -5.23485839e-01 -6.73538744e-01 -1.04227650e+00
6.47365689e-01 -1.41707182e+00 -1.29518434e-01 -2.10119581e+00
3.06356162e-01 4.58954051e-02 -4.55952585e-01 4.76927668e-01
-3.40646774e-01 1.61938444e-02 -1.91391468e-01 3.29020500e-01
-7.08871186e-01 1.09828115e+00 1.39463353e+00 -5.36446512e-01
4.62608747e-02 -1.91480353e-01 -8.02140534e-01 5.79943776e-01
5.80869079e-01 -4.72967982e-01 -8.04652035e-01 -5.68455100e-01
4.90167528e-01 -2.73657292e-01 4.42204565e-01 -4.05919403e-01
3.21945012e-01 1.36549443e-01 2.47869238e-01 -5.45257330e-01
4.67679873e-02 -6.34600341e-01 -1.65139347e-01 2.40619212e-01
-3.48644376e-01 -3.55266094e-01 -6.23697378e-02 1.28889894e+00
-4.31845456e-01 -1.27510801e-01 3.81717384e-01 -3.81173305e-02
-5.46723127e-01 5.69510102e-01 -3.86770338e-01 6.36333078e-02
7.07365692e-01 -1.99199712e-04 -3.39247644e-01 -6.66730225e-01
-6.38678133e-01 2.97024041e-01 1.25082299e-01 7.33408093e-01
6.88714087e-01 -1.40736055e+00 -6.86498642e-01 2.43759826e-01
4.36445713e-01 1.84095278e-01 2.28660837e-01 4.23085123e-01
-3.50161254e-01 8.09897125e-01 1.91972569e-01 -7.52476335e-01
-7.23242462e-01 7.14618862e-01 8.81544203e-02 -4.83026952e-01
-9.95051563e-01 7.26412594e-01 7.22207308e-01 -5.20218432e-01
8.25966671e-02 -3.40457410e-01 -4.58067626e-01 3.46102454e-02
1.79410547e-01 1.59208681e-02 -7.92446956e-02 -7.07810760e-01
-1.76604733e-01 3.50406706e-01 -2.77219206e-01 2.12683871e-01
1.27679420e+00 -1.14176750e-01 -3.98237824e-01 6.48126066e-01
1.44186044e+00 -3.64844829e-01 -8.59016120e-01 -6.26140714e-01
2.59754248e-02 -1.72043264e-01 5.22087276e-01 -2.22797677e-01
-1.18123186e+00 1.10233665e+00 -9.78587791e-02 7.09005117e-01
7.47385025e-01 3.78165841e-01 6.76236391e-01 4.29853529e-01
2.16435082e-02 -1.00798213e+00 4.48437065e-01 6.01358414e-01
8.63836408e-01 -1.16033912e+00 2.69136373e-02 -6.03435755e-01
-6.03896439e-01 1.30324554e+00 3.24160188e-01 -1.28908560e-01
8.77474844e-01 -3.22386354e-01 -3.47801477e-01 -6.43483460e-01
-8.61513853e-01 -5.66318408e-02 7.36091733e-01 5.07870734e-01
5.20062268e-01 1.98136553e-01 -4.82379645e-02 6.74915969e-01
-3.76285650e-02 -4.19387311e-01 2.41925612e-01 6.80755675e-01
-2.03444675e-01 -9.66542661e-01 2.16997787e-02 4.78800118e-01
-1.81178868e-01 -1.90010816e-01 -7.13036835e-01 7.55805731e-01
-3.63802642e-01 8.51083457e-01 2.31368735e-01 -2.71845490e-01
-1.77191645e-01 1.11226946e-01 2.40772277e-01 -7.16555655e-01
-7.14861080e-02 9.14979130e-02 -1.84188589e-01 -1.29691482e-01
-4.00814652e-01 -2.66532511e-01 -1.43383360e+00 -1.96388960e-01
-6.44501626e-01 6.22076392e-01 5.98188162e-01 1.22618091e+00
4.14548039e-01 1.01503503e+00 4.51438487e-01 -4.77088332e-01
-3.06519896e-01 -1.05972838e+00 -9.59815919e-01 4.10740465e-01
2.34772176e-01 -7.52608716e-01 -5.77846348e-01 1.11855492e-01] | [10.142773628234863, 6.843301773071289] |
8807d9e6-a1a7-4302-a73f-30280ef774fe | exploring-the-adjugate-matrix-approach-to | 2205.09116 | null | https://arxiv.org/abs/2205.09116v1 | https://arxiv.org/pdf/2205.09116v1.pdf | Exploring the Adjugate Matrix Approach to Quaternion Pose Extraction | Quaternions are important for a wide variety of rotation-related problems in computer graphics, machine vision, and robotics. We study the nontrivial geometry of the relationship between quaternions and rotation matrices by exploiting the adjugate matrix of the characteristic equation of a related eigenvalue problem to obtain the manifold of the space of a quaternion eigenvector. We argue that quaternions parameterized by their corresponding rotation matrices cannot be expressed, for example, in machine learning tasks, as single-valued functions: the quaternion solution must instead be treated as a manifold, with different algebraic solutions for each of several single-valued sectors represented by the adjugate matrix. We conclude with novel constructions exploiting the quaternion adjugate variables to revisit several classic pose estimation applications: 2D point-cloud matching, 2D point-cloud-to-projection matching, 3D point-cloud matching, 3D orthographic point-cloud-to-projection matching, and 3D perspective point-cloud-to-projection matching. We find an exact solution to the 3D orthographic least squares pose extraction problem, and apply it successfully also to the perspective pose extraction problem with results that improve on existing methods. | ['Sonya M. Hanson', 'Andrew J. Hanson'] | 2022-05-17 | null | null | null | null | ['3d-point-cloud-matching'] | ['computer-vision'] | [ 6.25406951e-02 4.38186973e-02 -8.24346542e-02 5.99810742e-02
-1.78495407e-01 -8.09091568e-01 5.35481811e-01 1.14259191e-01
-4.74708945e-01 3.12460899e-01 -3.34946096e-01 -3.95626873e-01
-1.06733516e-01 -5.24674416e-01 -7.02037692e-01 -5.39406061e-01
-1.22831516e-01 8.74345243e-01 -2.81965099e-02 -7.14596450e-01
5.48710287e-01 9.32652593e-01 -1.02579522e+00 -4.36158359e-01
4.95958179e-01 5.82133472e-01 -3.91658157e-01 7.87052512e-01
2.76198417e-01 -1.74794286e-01 -3.60225946e-01 -3.28226656e-01
5.98920643e-01 -2.28005409e-01 -5.30658662e-01 3.65575850e-01
6.78543627e-01 2.06101555e-02 -7.11030141e-02 1.12882006e+00
1.60234839e-01 -2.40791366e-01 7.99702048e-01 -1.63772655e+00
-3.25886786e-01 -3.07770342e-01 -8.20626199e-01 -3.28721970e-01
8.29966724e-01 -6.36479974e-01 9.52359319e-01 -1.13136113e+00
1.07159257e+00 1.35741913e+00 6.51083291e-01 -4.43152636e-02
-1.24451208e+00 -1.56132668e-01 -3.67379278e-01 2.86849421e-02
-1.44000971e+00 6.70231953e-02 6.15753770e-01 -6.96178138e-01
8.53915513e-01 4.97425586e-01 9.37852323e-01 1.91995248e-01
4.28792119e-01 4.42741007e-01 8.65266442e-01 -5.93865693e-01
1.05894813e-02 -9.82196033e-02 -8.48784149e-02 9.54303861e-01
5.50508797e-01 -2.59712964e-01 -3.00692648e-01 -1.46936521e-01
1.52771378e+00 2.90189032e-02 -2.92090923e-01 -1.18435311e+00
-1.72651792e+00 8.75634789e-01 3.31307799e-01 -1.01068899e-01
-2.42990002e-01 -3.69682796e-02 -2.07607970e-01 4.12040204e-01
2.07575932e-01 8.32163215e-01 -2.72321075e-01 -2.30319612e-02
-3.39792520e-01 2.71977752e-01 1.15365648e+00 1.43685782e+00
1.19644761e+00 -4.03961027e-03 8.77225161e-01 2.22184524e-01
3.35341841e-01 9.99362409e-01 1.11799210e-01 -9.08276856e-01
5.36813855e-01 6.51639819e-01 2.29923114e-01 -1.37193537e+00
-8.74248147e-01 -2.11173698e-01 -8.34222913e-01 1.44145042e-01
6.12569392e-01 -8.46874639e-02 -3.10184509e-01 1.19004500e+00
3.90069097e-01 -3.72088589e-02 -1.65841177e-01 9.51400936e-01
8.66043419e-02 2.55481869e-01 -4.54670727e-01 -1.71593755e-01
1.55316114e+00 -3.01919490e-01 -2.20857814e-01 -1.44577131e-01
6.72496259e-01 -1.21428680e+00 5.65724552e-01 4.31983382e-01
-1.00451887e+00 -4.99717779e-02 -1.35597110e+00 -1.10752210e-01
-1.52381836e-02 3.14064711e-01 3.51022214e-01 5.23488462e-01
-7.32879758e-01 5.72319806e-01 -6.42772019e-01 -5.68234086e-01
-4.24759865e-01 5.39428353e-01 -8.55836511e-01 3.76426846e-01
-7.39847004e-01 1.29708481e+00 7.38753974e-02 9.90414917e-02
3.38294089e-01 -3.83977711e-01 -8.26253712e-01 -2.82735139e-01
3.68673772e-01 -8.29749167e-01 1.08930027e+00 -4.97153640e-01
-1.38019955e+00 1.19348848e+00 -1.42804354e-01 5.36155738e-02
4.96501386e-01 -1.49083197e-01 -1.64090469e-01 1.06244564e-01
8.52498598e-03 2.93011516e-01 1.00083363e+00 -1.10611498e+00
-4.46476370e-01 -6.93891644e-01 1.85923278e-01 5.26958704e-01
2.02263996e-01 -2.78244078e-01 -4.79858041e-01 -3.01666647e-01
1.26249480e+00 -1.61801004e+00 -2.78996736e-01 2.57073045e-02
-5.92004597e-01 5.62360063e-02 6.74949408e-01 -4.36016858e-01
8.66982639e-01 -1.83125269e+00 6.69850588e-01 4.82229441e-01
1.42108381e-01 2.72278264e-02 -1.15240499e-01 5.30331790e-01
-5.86306632e-01 1.19671440e-02 1.78744700e-02 8.84667262e-02
-1.43353745e-01 2.12792143e-01 -9.66541097e-02 1.18319821e+00
5.60878217e-01 5.22219837e-01 -1.02522850e+00 -3.29482645e-01
2.46811435e-01 3.03684443e-01 -4.05024856e-01 -1.76919416e-01
2.30621114e-01 1.96738571e-01 -6.15570307e-01 3.90432209e-01
9.35048163e-01 9.31191891e-02 1.53516859e-01 -6.04045093e-01
-2.19304115e-01 3.47411595e-02 -1.82634568e+00 1.28156269e+00
-4.26874101e-01 7.09109247e-01 6.33879825e-02 -6.52430117e-01
9.50759530e-01 2.02868402e-01 6.51497483e-01 -2.32802555e-01
1.62627026e-01 3.12271208e-01 -1.31906077e-01 -1.82819515e-01
9.17550504e-01 -2.97066182e-01 -1.57694831e-01 1.72198728e-01
-2.20033541e-01 -6.64791107e-01 2.64643114e-02 5.70275970e-02
6.48612440e-01 1.41812526e-02 8.29087675e-01 -2.64096707e-01
5.68492234e-01 2.62006857e-02 2.34953493e-01 -2.19900925e-02
1.69201985e-01 1.02230394e+00 9.06063735e-01 -1.97181448e-01
-1.35873973e+00 -1.12195611e+00 -3.13617706e-01 1.78063810e-01
3.44422132e-01 -4.55849260e-01 -3.99218887e-01 4.12187912e-02
3.07650566e-01 2.33983442e-01 -1.07534632e-01 -2.25762799e-02
-8.14486206e-01 -5.01463473e-01 1.68522447e-01 2.51919717e-01
1.24730222e-01 -7.85659254e-02 -7.96913028e-01 -4.36509699e-02
5.24499677e-02 -1.19927490e+00 -4.19953912e-01 4.47370037e-02
-1.06879735e+00 -1.33455062e+00 -6.57570839e-01 -6.08779907e-01
7.72031784e-01 5.31114995e-01 1.02872932e+00 -8.94598290e-02
8.30087662e-02 6.85287595e-01 5.79948649e-02 -4.63887304e-01
-3.26346278e-01 -1.16936922e-01 4.51995075e-01 -3.73155326e-02
1.89128861e-01 -4.43862587e-01 -4.71841604e-01 9.88156497e-01
-8.18440378e-01 -3.09857838e-02 3.11595976e-01 4.91504341e-01
6.61616147e-01 -3.99383575e-01 7.80945197e-02 -5.20089686e-01
3.83709490e-01 -3.14884961e-01 -8.55723679e-01 1.91066176e-01
-2.41273656e-01 2.43608043e-01 3.36287946e-01 -2.92527646e-01
-1.79874331e-01 6.33062243e-01 3.05021852e-01 -4.50440109e-01
3.74965459e-01 5.04738152e-01 -1.10935844e-01 -5.42190790e-01
6.26261890e-01 -2.02586651e-01 1.00913182e-01 -3.33553225e-01
7.28619277e-01 2.60677397e-01 4.98617351e-01 -2.61121839e-01
1.23335242e+00 6.12448692e-01 9.37662125e-01 -1.40512538e+00
-2.00347617e-01 -8.67002487e-01 -1.26582313e+00 -2.25800142e-01
8.49427342e-01 -9.03722703e-01 -8.71824503e-01 3.59562457e-01
-1.47375345e+00 5.05018055e-01 -1.57869786e-01 7.53644407e-01
-9.59179461e-01 7.11264491e-01 -1.58274606e-01 -6.37098253e-01
1.00957006e-01 -1.14384103e+00 1.49019802e+00 -7.52462223e-02
-3.94666404e-01 -8.73000860e-01 2.96326697e-01 -1.20703660e-01
-3.10110301e-01 4.34602767e-01 9.50314879e-01 -1.79370880e-01
-7.95602679e-01 -6.25606656e-01 -3.68170552e-02 -4.62014936e-02
-9.60962698e-02 3.29189926e-01 -1.97870597e-01 -2.90794551e-01
1.48143992e-01 3.29143137e-01 3.13477606e-01 -6.70265704e-02
3.96253690e-02 2.35517658e-02 -1.59360394e-01 8.11415374e-01
1.55592752e+00 -9.46862400e-02 4.27865058e-01 6.19160414e-01
8.95066202e-01 4.11011487e-01 6.17738366e-01 3.95513356e-01
3.79303157e-01 1.12500787e+00 6.63011789e-01 1.44740269e-01
3.88336390e-01 -2.87284642e-01 5.67740127e-02 1.12126291e+00
-3.88559908e-01 3.57584685e-01 -8.43334138e-01 1.20040461e-01
-1.71010697e+00 -4.67613727e-01 -7.60992825e-01 2.80218220e+00
2.97243774e-01 -1.68454230e-01 2.53604144e-01 2.28986174e-01
6.71700895e-01 -6.97743893e-02 -3.71090710e-01 -5.20108879e-01
-2.06363842e-01 1.43835828e-01 9.34156656e-01 6.07868314e-01
-1.00212133e+00 7.44766891e-01 6.09802198e+00 2.31173515e-01
-1.26928937e+00 -3.27405274e-01 -4.25022393e-01 5.71167350e-01
-3.75463307e-01 4.32323962e-01 -6.10670865e-01 -2.54563004e-01
3.96644473e-01 -3.14248085e-01 2.15490639e-01 8.78780365e-01
-2.38905281e-01 -2.49142781e-01 -1.37722933e+00 1.25553334e+00
1.90524384e-01 -1.00195861e+00 -1.42869512e-02 3.06076318e-01
7.33303726e-01 -2.08792403e-01 4.89441566e-02 -3.06160957e-01
-4.56322938e-01 -5.86983502e-01 6.71037555e-01 2.57072449e-01
7.52916336e-01 -4.21281427e-01 1.57603979e-01 2.38999590e-01
-1.27973044e+00 5.43985128e-01 -5.24824083e-01 -2.84619480e-01
2.15235546e-01 2.41421804e-01 -1.33580124e+00 8.73409271e-01
-4.87114228e-02 7.59757757e-01 -4.66234267e-01 1.30767810e+00
-1.10636555e-01 -7.52754956e-02 -7.35170424e-01 -1.02709517e-01
9.18277279e-02 -1.21868312e+00 1.00430143e+00 6.93283856e-01
6.11455739e-01 1.02695145e-01 -2.11159334e-01 3.01907420e-01
1.72618181e-01 3.46429676e-01 -7.42063820e-01 2.10310116e-01
1.51958644e-01 1.22645926e+00 -1.03054559e+00 9.80789959e-02
-2.94825286e-01 8.11423481e-01 -2.18895525e-01 3.77298981e-01
-3.44209880e-01 -4.06789780e-01 8.73374045e-01 2.24913731e-01
2.01401383e-01 -9.76529539e-01 -3.48464131e-01 -1.53164291e+00
6.30222037e-02 -7.39075720e-01 -1.21842422e-01 -1.04387963e+00
-8.02496731e-01 3.93501490e-01 2.24702165e-01 -1.83046830e+00
-6.29181027e-01 -1.11931646e+00 -3.66146147e-01 9.69707906e-01
-1.01336312e+00 -7.08386242e-01 1.08471200e-01 7.33986437e-01
-1.17543295e-01 9.16230381e-02 9.66399431e-01 3.41255814e-02
-1.94215164e-01 9.63138640e-02 2.93853730e-01 -1.35459453e-01
6.30380929e-01 -1.53129089e+00 5.94360352e-01 6.52356446e-01
3.89121532e-01 7.75761187e-01 9.28705871e-01 -4.78031427e-01
-2.26906705e+00 -5.07113874e-01 8.72876883e-01 -7.92263567e-01
8.84318292e-01 -1.45154133e-01 -5.06636858e-01 1.02182090e+00
-3.27422172e-01 -3.85801703e-01 7.87648708e-02 9.94521566e-03
-2.99076915e-01 1.17556073e-01 -8.67203712e-01 7.79882669e-01
5.90648234e-01 -5.23263156e-01 -5.21165431e-01 5.86523712e-01
2.73773313e-01 -9.14170682e-01 -9.50431585e-01 7.77077004e-02
7.99233615e-01 -5.93923807e-01 1.33425367e+00 -7.66393423e-01
1.41967610e-01 -5.31788945e-01 -1.98353931e-01 -1.29044771e+00
-1.29334480e-01 -8.68424594e-01 3.49847466e-01 4.60591137e-01
1.19862907e-01 -7.52350986e-01 9.72699881e-01 2.73491561e-01
1.40052557e-01 -6.37379766e-01 -1.26222944e+00 -7.43008494e-01
7.92042539e-02 -1.37722686e-01 3.29855978e-01 1.02280724e+00
1.62271827e-01 6.68623567e-01 -4.35343951e-01 3.98366928e-01
5.89235842e-01 1.05012424e-01 1.16396201e+00 -1.48223627e+00
-1.72638193e-01 -3.16858083e-01 -1.10979617e+00 -1.31483114e+00
-1.04570247e-01 -8.12231481e-01 -3.35963190e-01 -1.21979141e+00
-4.98333484e-01 -3.22041571e-01 4.75819528e-01 -1.42188028e-01
2.46109888e-01 2.25002378e-01 6.09794378e-01 4.50276166e-01
-1.45824820e-01 2.49455586e-01 1.44523931e+00 3.19347411e-01
-2.07795560e-01 3.94965321e-01 -2.31486231e-01 1.18847835e+00
3.49927783e-01 -3.46292675e-01 -1.43755242e-01 -3.70760769e-01
9.10041690e-01 4.24970895e-01 2.13923067e-01 -8.40845883e-01
2.91991621e-01 -2.65639752e-01 1.20616131e-01 -7.65667498e-01
7.90230334e-01 -9.31394517e-01 3.04790616e-01 4.60152924e-01
3.90216023e-01 7.45057046e-01 -2.35593408e-01 4.33509141e-01
-9.44823548e-02 -3.88849258e-01 5.15366018e-01 1.41525944e-03
-5.47137678e-01 1.97502747e-01 -2.63003588e-01 -1.64197937e-01
9.46477234e-01 -6.05098188e-01 -2.89271921e-01 -6.12505674e-01
-6.15054727e-01 -4.36992869e-02 9.74865913e-01 1.79362267e-01
6.81405902e-01 -1.45164382e+00 -5.32177091e-01 4.48399246e-01
1.93511277e-01 2.89721414e-02 -2.44318828e-01 1.03072655e+00
-1.15285671e+00 5.64861238e-01 -4.58987683e-01 -1.12928212e+00
-1.37930596e+00 3.09691876e-01 3.01613003e-01 2.12207735e-02
-1.34067759e-01 4.31781620e-01 7.04537928e-02 -6.84318960e-01
-2.74527282e-01 -6.03586555e-01 -1.07636629e-02 1.99367210e-01
-1.84735537e-01 6.04733467e-01 2.86857635e-01 -1.38129270e+00
-4.29532945e-01 1.43232882e+00 5.05964637e-01 -3.15186590e-01
1.05396509e+00 -6.16843849e-02 -4.24079448e-01 4.45607543e-01
1.56027627e+00 1.11611515e-01 -6.36080444e-01 -1.13749601e-01
-3.57466750e-02 -4.18249309e-01 -5.72788060e-01 6.39707269e-03
-6.29534066e-01 8.45794678e-01 1.93965465e-01 3.12953949e-01
8.07795227e-01 -2.06618682e-01 3.49028021e-01 8.70810688e-01
5.10332108e-01 -6.40216947e-01 -1.46851763e-01 8.02429020e-01
1.19420290e+00 -8.40691864e-01 4.93034482e-01 -9.62299705e-01
-3.87614131e-01 1.77262676e+00 -1.47295455e-02 -6.29513383e-01
9.31319475e-01 -1.82149276e-01 2.86311265e-02 8.28600451e-02
-1.07952766e-01 -4.99429256e-02 7.56032169e-01 5.11990070e-01
3.48965645e-01 3.63496929e-01 -4.48839992e-01 -2.33472034e-01
-6.25522971e-01 -3.17253947e-01 9.94586825e-01 9.54707444e-01
-4.33088034e-01 -1.44862211e+00 -9.58643615e-01 -4.45985273e-02
2.00891554e-01 2.40847901e-01 -5.12120843e-01 1.13176572e+00
-1.15267709e-01 5.69580197e-01 2.49902517e-01 -4.86202717e-01
9.12390053e-01 -1.88863963e-01 8.62830222e-01 -5.61902225e-01
-8.78444239e-02 1.90320626e-01 -9.56160668e-03 -4.25936580e-01
-4.76816148e-01 -8.49882126e-01 -1.09113765e+00 -3.66664141e-01
-3.19914252e-01 4.25624698e-02 1.25290978e+00 8.42993796e-01
-1.05649338e-03 -2.12802663e-01 6.49443328e-01 -1.10529792e+00
-8.79082620e-01 -5.07411420e-01 -9.80297983e-01 3.48184854e-01
4.70550627e-01 -6.68995559e-01 -2.99801707e-01 -1.90028816e-01] | [7.96394157409668, -2.3183233737945557] |
c14a9dda-f166-4886-94a5-5bce22dc33b5 | overcoming-the-domain-gap-in-contrastive | 2111.14595 | null | https://arxiv.org/abs/2111.14595v1 | https://arxiv.org/pdf/2111.14595v1.pdf | Overcoming the Domain Gap in Contrastive Learning of Neural Action Representations | A fundamental goal in neuroscience is to understand the relationship between neural activity and behavior. For example, the ability to extract behavioral intentions from neural data, or neural decoding, is critical for developing effective brain machine interfaces. Although simple linear models have been applied to this challenge, they cannot identify important non-linear relationships. Thus, a self-supervised means of identifying non-linear relationships between neural dynamics and behavior, in order to compute neural representations, remains an important open problem. To address this challenge, we generated a new multimodal dataset consisting of the spontaneous behaviors generated by fruit flies, Drosophila melanogaster -- a popular model organism in neuroscience research. The dataset includes 3D markerless motion capture data from six camera views of the animal generating spontaneous actions, as well as synchronously acquired two-photon microscope images capturing the activity of descending neuron populations that are thought to drive actions. Standard contrastive learning and unsupervised domain adaptation techniques struggle to learn neural action representations (embeddings computed from the neural data describing action labels) due to large inter-animal differences in both neural and behavioral modalities. To overcome this deficiency, we developed simple yet effective augmentations that close the inter-animal domain gap, allowing us to extract behaviorally relevant, yet domain agnostic, information from neural data. This multimodal dataset and our new set of augmentations promise to accelerate the application of self-supervised learning methods in neuroscience. | ['Pascal Fua', 'Pavan Ramdya', 'Sina Honari', 'Florian Aymanns', 'Semih Günel'] | 2021-11-29 | null | null | null | null | ['markerless-motion-capture'] | ['computer-vision'] | [ 5.62149942e-01 -2.63721973e-01 -2.32282132e-01 -4.72059578e-01
-1.64950699e-01 -7.74437785e-01 6.47790372e-01 2.48878356e-02
-9.05230522e-01 7.39850163e-01 1.70763418e-01 2.79665768e-01
-6.69788793e-02 -2.73178875e-01 -6.76153839e-01 -8.07616293e-01
-1.84571147e-01 3.98937792e-01 1.72208503e-01 -2.65725106e-01
4.69936222e-01 5.19951701e-01 -1.58420575e+00 3.33875865e-01
5.06670833e-01 7.77752995e-01 3.81479770e-01 7.05999911e-01
2.36172266e-02 5.86212516e-01 -2.98666686e-01 4.23647046e-01
-1.36644274e-01 -7.52304852e-01 -6.89981341e-01 -7.73098469e-02
3.91438812e-01 -2.09274665e-01 -1.50651678e-01 8.66866410e-01
3.30546379e-01 7.61832669e-02 1.01930976e+00 -1.38015020e+00
-4.37944859e-01 7.38048851e-02 -1.11893319e-01 6.06202364e-01
1.54060230e-01 2.28958786e-01 8.64614367e-01 -3.75757515e-01
8.40301871e-01 9.60415721e-01 2.92976409e-01 1.29759419e+00
-1.80931437e+00 -4.66549605e-01 -2.23121867e-01 1.29323557e-01
-7.36035466e-01 -5.22181988e-01 7.98153222e-01 -8.79704177e-01
8.04047763e-01 -1.98729336e-01 1.01442361e+00 1.81370187e+00
3.87244314e-01 5.81309080e-01 1.18574452e+00 -6.16256036e-02
5.88494480e-01 3.93866897e-02 2.94543207e-01 4.66605157e-01
1.12281702e-01 1.22109026e-01 -7.76494861e-01 -9.53887254e-02
1.02316105e+00 2.75447257e-02 -2.96412796e-01 -6.07351243e-01
-1.43112004e+00 7.61746705e-01 2.23218992e-01 5.80297172e-01
-2.11600542e-01 2.82423228e-01 2.91659743e-01 2.37861902e-01
1.09826304e-01 8.51069629e-01 -7.06108987e-01 -4.53244358e-01
-6.32862806e-01 2.64770240e-01 6.74343407e-01 2.48778120e-01
6.96673155e-01 1.43121243e-01 2.59878635e-01 7.49786377e-01
1.42408594e-01 3.13618600e-01 1.05815744e+00 -1.24297595e+00
-1.68278635e-01 1.05784142e+00 -1.70049638e-01 -8.45145702e-01
-7.45058894e-01 6.42917454e-02 -7.62365758e-01 4.30273533e-01
6.40562534e-01 8.14467110e-03 -7.51402259e-01 2.16156411e+00
1.35135025e-01 -6.01482466e-02 6.55589774e-02 9.37080204e-01
4.78622437e-01 5.54969847e-01 1.79905266e-01 -1.77901536e-01
1.18338871e+00 -2.31109887e-01 -5.86038172e-01 -5.05793333e-01
7.90421128e-01 -7.83514008e-02 1.01792431e+00 1.78533211e-01
-7.91975200e-01 -4.71569091e-01 -9.82961297e-01 5.20271109e-03
-7.57189989e-01 4.23282720e-02 6.01293743e-01 -7.16441795e-02
-8.22357178e-01 6.53604388e-01 -1.13472760e+00 -6.84432030e-01
6.78423345e-01 8.00488114e-01 -9.39090729e-01 3.94544333e-01
-7.91895270e-01 9.68541443e-01 3.51473749e-01 -4.05157179e-01
-1.06623447e+00 -6.04060531e-01 -7.47239053e-01 -2.24080428e-01
2.07575113e-02 -5.85461020e-01 1.00301242e+00 -1.05990553e+00
-1.49556732e+00 1.29732132e+00 -1.32898772e-02 -4.72854495e-01
-1.95518374e-01 -4.84065190e-02 -7.13109821e-02 3.29427898e-01
6.19883998e-04 1.19129765e+00 8.44354093e-01 -1.01627707e+00
-2.95003682e-01 -9.64312673e-01 -1.66435108e-01 9.10432562e-02
-7.43882239e-01 -2.51914799e-01 6.44952850e-03 -5.42693496e-01
-2.34094523e-02 -1.14553654e+00 -8.86490345e-02 3.61709058e-01
1.11039884e-01 2.13417653e-02 9.72762704e-01 -2.13682160e-01
6.76182032e-01 -2.24901605e+00 9.14701045e-01 -2.73457438e-01
4.25855011e-01 3.45914871e-01 -3.68973374e-01 2.22944871e-01
-1.92395627e-01 -1.92296043e-01 -5.44725001e-01 6.08857311e-02
-3.14526737e-01 4.39768821e-01 -2.04179749e-01 5.74114442e-01
5.76134741e-01 1.00949025e+00 -8.49672079e-01 -3.96914095e-01
1.07516736e-01 5.08723736e-01 -6.94120288e-01 5.77657163e-01
-4.22656655e-01 9.87043738e-01 -4.20899183e-01 4.70451951e-01
3.75052132e-02 -1.81044281e-01 9.32588950e-02 -2.49659583e-01
3.22584208e-04 -1.85584510e-03 -4.98142421e-01 1.92986357e+00
-1.65302932e-01 9.14887905e-01 9.45890844e-02 -1.67562664e+00
7.03147769e-01 1.70203120e-01 9.51644540e-01 -9.43320930e-01
3.91416818e-01 5.86826131e-02 3.43162626e-01 -7.48475611e-01
-2.19964162e-01 -1.52916491e-01 -1.08152114e-01 4.73631501e-01
7.24225283e-01 -3.42010498e-01 2.27042332e-01 -1.21791318e-01
1.28209543e+00 4.67945904e-01 4.49762732e-01 -3.03465247e-01
3.52717847e-01 4.11046565e-01 5.57806909e-01 1.40059575e-01
-4.24247205e-01 4.36623901e-01 6.46884680e-01 -5.41521370e-01
-9.92686510e-01 -1.04979336e+00 -1.12052627e-01 1.12283814e+00
-4.38939929e-02 -1.29751801e-01 -8.51969063e-01 -2.88996547e-01
-2.21289486e-01 2.82352984e-01 -1.04362357e+00 -6.34215951e-01
-5.75480282e-01 -8.07707727e-01 5.22912800e-01 5.32327592e-01
2.59584785e-01 -1.18985879e+00 -1.20358443e+00 2.65488595e-01
7.24239051e-02 -1.33813035e+00 3.91987674e-02 6.80506170e-01
-1.15504813e+00 -1.22538650e+00 -6.36443257e-01 -8.83058012e-01
8.51427376e-01 -3.76347750e-02 7.39670217e-01 -3.51278424e-01
-7.65781343e-01 6.04332983e-01 -1.13568097e-01 -3.07044804e-01
-3.20807815e-01 -2.57980347e-01 4.35176194e-01 -4.37674038e-02
3.57770026e-01 -8.94807100e-01 -4.38260108e-01 2.62539536e-01
-1.27226615e+00 -9.82236303e-03 5.75104117e-01 9.32617545e-01
6.10367596e-01 -4.13912565e-01 4.45862263e-01 -5.74070096e-01
6.15870416e-01 -3.74825031e-01 -5.02931476e-01 -1.53211385e-01
-6.32020608e-02 4.87853467e-01 8.94371748e-01 -9.91844893e-01
-6.47380054e-01 4.64098811e-01 3.50886844e-02 -2.10408047e-01
-6.86647773e-01 2.27510810e-01 4.85565960e-02 -1.46060839e-01
8.76503170e-01 4.29619431e-01 4.45204616e-01 -2.10908353e-01
1.90509588e-01 4.37753320e-01 7.80574799e-01 -4.23267782e-01
4.34920162e-01 6.66511536e-01 2.48514354e-01 -1.15831602e+00
-5.58025658e-01 -4.11893398e-01 -9.38638151e-01 -2.19491407e-01
1.19356489e+00 -6.30887687e-01 -8.19432735e-01 3.74226868e-01
-1.01947224e+00 -5.53938806e-01 -2.18859747e-01 7.04369724e-01
-1.08734632e+00 3.49733047e-02 -5.55079758e-01 -6.22168362e-01
-5.31722568e-02 -1.02882242e+00 1.04870427e+00 1.99181497e-01
-6.84092104e-01 -9.80737031e-01 7.30232775e-01 2.41875514e-01
2.53628284e-01 3.42934310e-01 1.21093643e+00 -7.85416126e-01
-3.47311124e-02 -1.10617653e-02 1.34762600e-01 3.97981763e-01
2.62931108e-01 1.50648691e-02 -8.51503789e-01 -3.41952518e-02
1.67544514e-01 -8.29090893e-01 6.55002415e-01 4.22085255e-01
9.73862469e-01 -5.97299188e-02 -3.23037297e-01 4.56623197e-01
1.08843946e+00 3.22362185e-01 3.78453732e-01 1.09537110e-01
5.10391831e-01 9.46228802e-01 1.79514334e-01 2.63433516e-01
1.04907311e-01 8.33405077e-01 4.23467726e-01 1.31658554e-01
1.06956542e-01 3.77567783e-02 6.83891118e-01 7.96758652e-01
-2.02005386e-01 1.55140191e-01 -8.97851408e-01 3.88014197e-01
-1.83788681e+00 -9.29185152e-01 1.35057956e-01 2.06266499e+00
9.55440164e-01 -2.42186952e-02 3.78528088e-01 1.09258465e-01
2.11753711e-01 -2.94341594e-01 -9.66888785e-01 -2.65282214e-01
-1.90497637e-01 1.23785749e-01 1.18012251e-02 6.61823824e-02
-1.04257262e+00 9.43671703e-01 6.57624626e+00 2.05477387e-01
-1.42225671e+00 -1.67885810e-01 3.17045569e-01 -1.93621308e-01
1.67077571e-01 -2.57821679e-01 -6.46026254e-01 5.61959803e-01
1.24239540e+00 1.11646719e-01 6.88787282e-01 6.98688269e-01
2.28642911e-01 -1.94844589e-01 -1.63747132e+00 1.22838128e+00
2.66558915e-01 -1.30099082e+00 -1.43258214e-01 2.20653921e-01
4.79948699e-01 4.65999842e-02 2.70633884e-02 1.54216856e-01
1.35339582e-02 -1.02279830e+00 1.92614302e-01 6.76227629e-01
4.20299768e-01 -1.89863816e-01 3.80973041e-01 7.03732431e-01
-7.93382823e-01 -2.26949736e-01 -2.11287066e-01 -1.35313675e-01
-1.88590854e-01 6.32972121e-02 -5.13909340e-01 -4.30401921e-01
7.84079969e-01 1.08814907e+00 -6.15196466e-01 6.18882000e-01
2.16813475e-01 4.28878725e-01 -2.48224929e-01 -2.00639382e-01
3.54750864e-02 -3.24778169e-01 4.91926223e-01 8.72647464e-01
4.53884900e-02 8.19891170e-02 -4.79587704e-01 1.24711430e+00
1.40870186e-02 -2.68592775e-01 -1.18370152e+00 -7.63380706e-01
-1.48862898e-01 1.25317550e+00 -7.30633497e-01 -9.82052833e-02
-2.61414170e-01 1.02875590e+00 6.62098527e-01 2.77075559e-01
-6.92693770e-01 7.46701062e-02 1.04726160e+00 2.60142349e-02
-7.91432559e-02 -6.72532141e-01 -1.48574471e-01 -1.25569355e+00
-2.99685746e-01 -8.08496594e-01 4.20046132e-03 -6.91850603e-01
-1.12547374e+00 3.14711034e-01 1.68911308e-01 -1.34384811e+00
-4.85563427e-01 -1.16342783e+00 -3.68964672e-01 1.26906529e-01
-1.00238085e+00 -7.42738366e-01 -1.91257223e-01 6.76021099e-01
5.75858593e-01 -2.40960181e-01 1.05847156e+00 1.01903573e-01
-5.85803151e-01 -4.62139249e-02 -1.10775474e-02 -2.89443992e-02
6.15760028e-01 -1.12257564e+00 -1.66417107e-01 2.39530653e-01
4.33331043e-01 6.74617887e-01 6.27222121e-01 -2.61998087e-01
-1.75066984e+00 -7.74416625e-01 3.66363078e-01 -7.20920801e-01
6.93611205e-01 -5.59859872e-01 -1.04110587e+00 4.67694938e-01
-1.51508665e-02 1.88262686e-01 1.06371081e+00 -4.53884929e-01
-1.80155486e-01 -7.48379230e-02 -8.93947363e-01 9.32978153e-01
9.94782805e-01 -6.15378559e-01 -1.02314043e+00 1.09478034e-01
2.46651083e-01 2.41974667e-01 -8.02941859e-01 4.28204268e-01
7.07727551e-01 -7.63742149e-01 9.43740427e-01 -1.02885282e+00
6.33137405e-01 -3.60588394e-02 -1.56899124e-01 -1.47915995e+00
-6.62990212e-02 -2.33766839e-01 -2.34338835e-01 9.22887683e-01
2.20480204e-01 -3.80817384e-01 8.62517238e-01 5.06364465e-01
3.98214422e-02 -6.13156021e-01 -7.99021959e-01 -5.82052827e-01
2.16497406e-01 -2.28828624e-01 -3.15325052e-01 9.15335119e-01
4.08913106e-01 6.10152304e-01 -1.47104427e-01 -3.35392118e-01
5.02462983e-01 -3.63222733e-02 7.24084914e-01 -1.42220867e+00
6.10529445e-03 -5.48562706e-01 -8.14698994e-01 -9.52954531e-01
5.17444968e-01 -8.47528875e-01 6.17742613e-02 -1.19227898e+00
3.01278263e-01 2.73105323e-01 -2.06274062e-01 6.06400967e-01
3.95469517e-01 2.51435995e-01 -2.41143152e-01 2.57181019e-01
-4.35262322e-01 6.58036828e-01 1.17789042e+00 -1.27897799e-01
-2.65624821e-01 -4.64639187e-01 -3.41217101e-01 9.95341539e-01
9.32143629e-01 -6.32605016e-01 -3.75458568e-01 -3.14473450e-01
5.35642579e-02 -5.11627905e-02 6.28290176e-01 -1.32108998e+00
2.30478004e-01 -2.47426644e-01 5.47478795e-01 -1.94190830e-01
7.10411429e-01 -9.49154615e-01 -3.40888053e-01 6.31846607e-01
-6.84656799e-01 -4.57579978e-02 3.20264876e-01 6.01769686e-01
-3.47072303e-01 3.29738036e-02 8.74724567e-01 -2.82457352e-01
-9.63166475e-01 1.59750357e-01 -9.76659775e-01 3.30529451e-01
1.17725348e+00 -3.62861961e-01 -2.71171898e-01 -4.92833555e-02
-7.69388974e-01 -9.05331969e-02 4.74728763e-01 4.65837538e-01
7.14235067e-01 -1.26297653e+00 -2.28348911e-01 5.27776361e-01
4.03629869e-01 -3.91176611e-01 1.15346342e-01 9.41049933e-01
-3.71889353e-01 4.87070203e-01 -1.14226210e+00 -7.91434288e-01
-1.10563731e+00 3.17645133e-01 1.81265742e-01 1.54676422e-01
-1.81256965e-01 5.78391731e-01 3.87985528e-01 -5.78851044e-01
1.94473732e-02 -4.69044626e-01 -6.21019781e-01 1.38875753e-01
3.46150130e-01 8.52912962e-02 -2.66700208e-01 -7.94562042e-01
-4.14561212e-01 8.05721045e-01 2.53412962e-01 -2.02310190e-01
1.58256412e+00 1.56323776e-01 -1.10462017e-01 9.25636113e-01
1.48502684e+00 -5.59313893e-01 -1.29328942e+00 1.03001937e-01
3.94504927e-02 -1.89025104e-02 -5.27424030e-02 -4.33236420e-01
-8.74584198e-01 1.36306489e+00 7.80678034e-01 3.85188311e-02
1.10763764e+00 7.62769952e-02 5.30104697e-01 7.54335225e-01
3.78783137e-01 -1.23160362e+00 6.28311992e-01 4.20469791e-01
7.42410779e-01 -1.58949518e+00 -2.29882032e-01 1.60593286e-01
-7.09072530e-01 1.32242239e+00 9.12776709e-01 -3.02518845e-01
4.73260462e-01 3.98889869e-01 -2.62136292e-02 -2.53039032e-01
-1.05597317e+00 -1.96575597e-01 2.11758673e-01 8.40804458e-01
4.39202338e-01 -3.02607268e-01 -1.61222085e-01 5.59811294e-01
1.20555110e-01 2.24335924e-01 3.19314778e-01 1.07011247e+00
-3.46120566e-01 -1.08925068e+00 -7.76381344e-02 4.37622964e-01
-3.30089092e-01 2.79161125e-01 -7.99103439e-01 6.76595926e-01
5.05509675e-02 5.46511531e-01 2.53781646e-01 -5.14549851e-01
8.65326598e-02 1.43748388e-01 6.71253622e-01 -7.12123811e-01
-1.18260734e-01 -3.33789252e-02 -3.15853894e-01 -7.77611196e-01
-7.03663051e-01 -4.93188441e-01 -1.51558149e+00 2.09160104e-01
2.10337222e-01 -3.05656105e-01 8.08890581e-01 1.04196358e+00
3.67051423e-01 4.02840644e-01 2.56311923e-01 -1.16739726e+00
-2.85431743e-01 -7.90366471e-01 -5.31451821e-01 8.01692963e-01
4.18176055e-01 -8.76584411e-01 -5.12555242e-01 6.55437768e-01] | [9.575722694396973, 2.5106472969055176] |
67e11188-64df-465c-ab9a-30cc61d593b8 | deepfn-towards-generalizable-facial-action | 2103.02484 | null | https://arxiv.org/abs/2103.02484v1 | https://arxiv.org/pdf/2103.02484v1.pdf | DeepFN: Towards Generalizable Facial Action Unit Recognition with Deep Face Normalization | Facial action unit recognition has many applications from market research to psychotherapy and from image captioning to entertainment. Despite its recent progress, deployment of these models has been impeded due to their limited generalization to unseen people and demographics. This work conducts an in-depth analysis of performance across several dimensions: individuals(40 subjects), genders (male and female), skin types (darker and lighter), and databases (BP4D and DISFA). To help suppress the variance in data, we use the notion of self-supervised denoising autoencoders to design a method for deep face normalization(DeepFN) that transfers facial expressions of different people onto a common facial template which is then used to train and evaluate facial action recognition models. We show that person-independent models yield significantly lower performance (55% average F1 and accuracy across 40 subjects) than person-dependent models (60.3%), leading to a generalization gap of 5.3%. However, normalizing the data with the newly introduced DeepFN significantly increased the performance of person-independent models (59.6%), effectively reducing the gap. Similarly, we observed generalization gaps when considering gender (2.4%), skin type (5.3%), and dataset (9.4%), which were significantly reduced with the use of DeepFN. These findings represent an important step towards the creation of more generalizable facial action unit recognition systems. | ['Mary Czerwinski', 'Alberto Fung', 'Rudovic', 'Ognjen', 'Daniel McDuff', 'Javier Hernandez'] | 2021-03-03 | null | null | null | null | ['facial-action-unit-detection'] | ['computer-vision'] | [ 3.08677942e-01 7.60822445e-02 -8.20683911e-02 -6.63980842e-01
-3.14991981e-01 -3.63525480e-01 5.61267316e-01 -4.89236981e-01
-6.06702745e-01 4.44906503e-01 3.39324236e-01 2.42226735e-01
2.76164263e-01 -6.89896643e-01 -4.20606494e-01 -8.34073544e-01
8.69486183e-02 -1.16458967e-01 -4.33765352e-01 -2.83598602e-01
-6.30711913e-02 7.74126709e-01 -1.94122028e+00 3.50972801e-01
4.47946459e-01 1.00523710e+00 -6.85847938e-01 2.71246254e-01
1.33219436e-01 4.08010334e-01 -8.87278795e-01 -7.34391212e-01
4.54396069e-01 -6.09366834e-01 -3.80694836e-01 2.37569556e-01
9.40653920e-01 -6.90510571e-01 -2.36403927e-01 9.31388855e-01
7.92677104e-01 2.86503762e-01 7.21983135e-01 -1.45055509e+00
-8.62703264e-01 5.15271612e-02 -7.18910098e-01 -9.81751680e-02
1.93291545e-01 2.15902016e-01 5.10848343e-01 -7.02249289e-01
5.66978335e-01 1.64852786e+00 6.55330777e-01 1.34758866e+00
-1.30830646e+00 -1.11122179e+00 -7.29226321e-02 2.20678113e-02
-1.39775479e+00 -7.95907795e-01 4.65656698e-01 -3.73764038e-01
7.98002064e-01 1.27576008e-01 5.22094190e-01 1.48050976e+00
-9.91967469e-02 3.96504700e-01 1.14479852e+00 -3.98312241e-01
1.31226808e-01 2.23402411e-01 -1.30052552e-01 4.12629515e-01
2.89320260e-01 1.34832552e-03 -5.11981666e-01 -1.64903738e-02
9.49901283e-01 -1.94729850e-01 7.67531693e-02 1.29059270e-01
-5.33533752e-01 7.94731617e-01 6.00381605e-02 3.59975666e-01
-4.14461970e-01 -9.92543846e-02 5.62234104e-01 2.71306127e-01
6.01555824e-01 2.10189566e-01 -3.06934208e-01 -2.02358007e-01
-7.66948879e-01 2.14416832e-01 7.01894104e-01 5.67496657e-01
6.13753021e-01 4.46970701e-01 -1.30232140e-01 1.11526060e+00
-6.29998446e-02 5.26632130e-01 5.65565765e-01 -1.20320308e+00
-5.49280411e-03 6.12543583e-01 -7.75131583e-02 -1.21761525e+00
-4.47554708e-01 -2.12290451e-01 -8.09596181e-01 4.58040327e-01
5.90960503e-01 -3.30080807e-01 -1.15642834e+00 2.28350878e+00
2.51126677e-01 -1.99749768e-01 8.20761994e-02 8.40827644e-01
7.65719652e-01 3.07457447e-01 5.70794880e-01 -1.01178281e-01
1.45507479e+00 -4.65217888e-01 -7.65579760e-01 -2.28499487e-01
6.34295881e-01 -6.19720280e-01 9.48323011e-01 3.27901185e-01
-9.48061883e-01 -6.02114201e-01 -8.82176757e-01 2.18091905e-02
-3.34582388e-01 1.71336249e-01 6.77777767e-01 1.28149664e+00
-1.25042081e+00 5.27509212e-01 -6.03114903e-01 -7.49308527e-01
8.29538107e-01 6.94191575e-01 -8.33888412e-01 -2.93770581e-02
-9.10944343e-01 8.88182759e-01 -3.51344086e-02 -7.24015236e-02
-6.61330938e-01 -6.41034722e-01 -8.13344181e-01 -6.81554303e-02
1.47793805e-02 -5.23812592e-01 9.00065362e-01 -1.70631611e+00
-1.38952732e+00 1.26182735e+00 -2.53633291e-01 -2.48361945e-01
4.06756431e-01 -1.04198352e-01 -6.05639100e-01 1.69954717e-01
-1.39864217e-02 1.00921953e+00 1.01186943e+00 -9.49749172e-01
-3.68612528e-01 -7.97952592e-01 -6.70535490e-02 1.07350312e-01
-8.38967800e-01 4.14555997e-01 -1.33656889e-01 -6.14939272e-01
-1.58204854e-01 -9.23990607e-01 3.36621493e-01 1.82901710e-01
1.32784814e-01 -2.83353359e-01 7.81224132e-01 -9.30851281e-01
9.40678060e-01 -2.47190142e+00 3.86582464e-02 9.72187743e-02
2.04607219e-01 4.27021801e-01 -4.52348232e-01 -9.94984410e-04
-5.50923347e-01 1.96672112e-01 4.42841500e-02 -5.34435451e-01
-1.53030708e-01 2.84364849e-01 3.61909926e-01 5.59241176e-01
4.13945287e-01 6.25602603e-01 -3.87871444e-01 -3.58124495e-01
9.82136950e-02 8.49599421e-01 -7.65968859e-01 -3.37464362e-02
4.27316755e-01 2.52979994e-01 -6.02823123e-02 9.59760487e-01
8.79446626e-01 4.64644253e-01 -3.83958290e-03 -3.85740340e-01
9.93321314e-02 -4.26148295e-01 -1.02866983e+00 1.12202847e+00
-2.90598571e-01 6.35016382e-01 3.09453756e-01 -8.20220053e-01
1.15103042e+00 2.92599976e-01 8.16122174e-01 -9.43568051e-01
5.45656562e-01 -5.59629574e-02 2.37287387e-01 -4.22883660e-01
1.94947377e-01 -5.53679585e-01 2.12840125e-01 3.10338259e-01
4.18006480e-01 3.54003787e-01 1.44544929e-01 -3.53107512e-01
8.23781192e-01 -1.12221256e-01 -8.66627842e-02 -2.47288108e-01
5.27307153e-01 -5.31895638e-01 5.81210613e-01 3.83469194e-01
-7.06568360e-01 6.15630388e-01 6.28610253e-01 -5.20019412e-01
-9.20444608e-01 -9.77204919e-01 -1.07932299e-01 1.33291578e+00
-4.42942709e-01 -1.99339330e-01 -1.28776884e+00 -4.22606766e-01
1.85991824e-01 6.69241726e-01 -1.07331848e+00 -6.23552084e-01
-2.88951397e-01 -9.01731670e-01 8.88065219e-01 6.86611176e-01
5.29667735e-01 -1.01891088e+00 -3.04076493e-01 -2.08382383e-01
-1.84201840e-02 -1.05491185e+00 -1.87509999e-01 -1.76722035e-01
-7.02066600e-01 -8.92272353e-01 -8.78658831e-01 -4.99291301e-01
8.01323056e-01 3.52899581e-02 7.86485732e-01 -4.30044532e-02
-3.22479397e-01 7.56153107e-01 -1.83223814e-01 -6.27908349e-01
-4.33482975e-01 -2.30755523e-01 6.05449378e-01 5.03007650e-01
8.09932172e-01 -5.62587321e-01 -7.07718313e-01 4.86304104e-01
-9.78305936e-01 -4.25829649e-01 5.36483049e-01 6.90944195e-01
2.03923881e-01 -2.26930231e-01 6.47398889e-01 -4.64484662e-01
6.41820550e-01 -2.19147190e-01 -8.73661712e-02 1.88613068e-02
-5.43297112e-01 -3.61765951e-01 2.75620997e-01 -7.03517854e-01
-1.27871335e+00 -7.32584372e-02 -1.61552206e-01 -4.78996783e-01
-4.31564063e-01 -5.27947918e-02 -2.54563302e-01 -2.87335694e-01
9.09062147e-01 -8.19453225e-02 7.29747355e-01 -3.66793662e-01
3.62880565e-02 7.95064747e-01 5.01095772e-01 -4.50987577e-01
5.67763150e-01 5.73214710e-01 5.14695607e-02 -1.00770426e+00
-5.62589705e-01 -5.04924990e-02 -7.27559984e-01 -3.96671802e-01
1.03077900e+00 -1.00251353e+00 -8.83309543e-01 7.58999944e-01
-8.05493474e-01 -1.44390598e-01 -2.02160731e-01 4.35491264e-01
-3.70968103e-01 2.01467797e-01 -4.28357959e-01 -8.09234500e-01
-4.35780197e-01 -8.98935676e-01 9.40533698e-01 3.53061050e-01
-6.63985908e-01 -7.25908041e-01 -1.87359914e-01 7.34952152e-01
4.13482875e-01 4.81935918e-01 7.12729990e-01 -6.27686739e-01
2.10833579e-01 -4.48653132e-01 -3.05362403e-01 1.02401543e+00
5.67075670e-01 1.57131329e-01 -1.24512446e+00 -2.70134538e-01
2.03955583e-02 -3.85456592e-01 4.98004198e-01 3.68036330e-01
1.11305749e+00 -1.86617628e-01 5.63181937e-02 5.74156880e-01
9.78420138e-01 3.65561903e-01 9.61327016e-01 2.47830778e-01
5.53401589e-01 9.18657899e-01 1.52904093e-01 4.49720949e-01
3.14483698e-03 5.31190693e-01 2.89235681e-01 -3.48815709e-01
-2.63935089e-01 -3.25378589e-02 5.76761186e-01 6.84466884e-02
-6.22858584e-01 5.19417636e-02 -5.70576191e-01 3.55154663e-01
-1.24226463e+00 -1.11941302e+00 2.39899918e-01 2.05370855e+00
6.23374820e-01 -2.89280534e-01 4.50707912e-01 7.92776002e-04
7.15781987e-01 -1.67178646e-01 -5.80481172e-01 -7.77830482e-01
-3.05398077e-01 4.80641812e-01 2.53980279e-01 4.34373654e-02
-1.02941549e+00 8.61221433e-01 6.61641550e+00 7.01376319e-01
-1.28750849e+00 -1.65557265e-01 9.02932644e-01 -2.53829807e-01
2.14882702e-01 -5.61484754e-01 -7.80340195e-01 2.97071666e-01
9.89697635e-01 -1.89739421e-01 4.59306091e-01 8.74437392e-01
2.90374666e-01 3.54505628e-02 -1.15160060e+00 1.33219564e+00
4.53663170e-01 -7.71290958e-01 1.49239838e-01 1.55540258e-01
7.12691545e-01 -4.84101206e-01 2.47296005e-01 4.03438866e-01
-2.72376865e-01 -1.26960897e+00 4.38915521e-01 3.62902403e-01
8.54203939e-01 -9.27964509e-01 7.66929328e-01 -1.73204601e-01
-6.33243680e-01 -2.31989428e-01 -3.17365378e-01 -2.51806200e-01
-2.82827258e-01 -1.08121829e-02 -5.46702206e-01 9.59977657e-02
9.09932733e-01 4.69541192e-01 -4.89906400e-01 2.38490731e-01
1.92216069e-01 5.04812837e-01 -2.82244861e-01 2.30921566e-01
-7.87179396e-02 -2.63424814e-01 1.82188660e-01 1.02299929e+00
2.49106556e-01 1.74103335e-01 -3.97166133e-01 7.78498292e-01
-2.66222298e-01 2.93306828e-01 -4.05313998e-01 -2.19821155e-01
1.12291001e-01 1.29766953e+00 -5.24434566e-01 -4.15468477e-02
-7.23358333e-01 1.05929172e+00 4.21677269e-02 3.80415559e-01
-7.59965003e-01 -1.55516878e-01 1.23011756e+00 3.16844374e-01
-7.56853446e-02 1.22828558e-01 -3.23431075e-01 -8.59787285e-01
-3.78703177e-02 -1.18371689e+00 3.63133788e-01 -5.56416154e-01
-1.37944829e+00 3.94279182e-01 9.81548429e-02 -9.35697377e-01
-2.16158152e-01 -8.67946804e-01 -3.88028413e-01 9.61433530e-01
-8.15942526e-01 -1.19485188e+00 -4.98624295e-01 6.81644320e-01
1.85455959e-02 -4.25840080e-01 1.07655835e+00 6.25494897e-01
-8.66194427e-01 1.25467479e+00 -2.06167266e-01 4.68913406e-01
9.71381724e-01 -6.42876804e-01 -2.69306600e-02 5.68123400e-01
-3.03465158e-01 8.15810084e-01 5.54740429e-01 -3.26245695e-01
-1.18089592e+00 -9.08341885e-01 6.40232563e-01 -3.87580395e-01
3.90701950e-01 -4.03004140e-01 -7.73510873e-01 6.07801855e-01
7.70095736e-02 -1.59462333e-01 1.07138860e+00 1.71400130e-01
-4.98183489e-01 -4.39758778e-01 -1.60251224e+00 1.00991678e+00
1.15983546e+00 -4.10122454e-01 -1.32659063e-01 -2.67518535e-02
1.69522524e-01 -9.36182737e-02 -1.10667109e+00 4.46179926e-01
1.00828719e+00 -1.21679544e+00 8.67054582e-01 -7.50448883e-01
3.84108484e-01 1.39688164e-01 -2.20689416e-01 -1.00462186e+00
-3.64748836e-01 -3.39898825e-01 2.23853856e-01 1.44314003e+00
1.55160710e-01 -7.37594128e-01 1.06955218e+00 1.22010076e+00
1.57471567e-01 -7.02585399e-01 -9.58235860e-01 -6.39499962e-01
2.38250375e-01 -2.52916276e-01 4.73754376e-01 9.47097182e-01
-3.57015342e-01 7.03214062e-03 -3.44904512e-01 -2.73920119e-01
4.21397299e-01 -5.37454784e-01 8.47140431e-01 -1.12786520e+00
2.11899132e-01 -5.36230147e-01 -8.92976284e-01 -2.84845591e-01
4.39039290e-01 -6.16558671e-01 -4.17781353e-01 -1.03006446e+00
3.25240672e-01 1.32350385e-01 -1.94649771e-01 1.01074696e+00
-9.16252262e-04 6.07297719e-01 1.50754884e-01 -1.72692046e-01
1.98353335e-01 5.71736693e-01 1.09682012e+00 2.52497718e-02
-1.80576622e-01 -2.90503148e-02 -1.03568327e+00 7.39083886e-01
9.40814257e-01 -1.71826586e-01 -3.47427577e-01 -2.24702299e-01
-3.20496321e-01 -5.49389601e-01 4.51449871e-01 -9.74797845e-01
-2.90065467e-01 -6.58836514e-02 9.56518114e-01 4.07098532e-02
5.50379455e-01 -7.53653169e-01 8.94821435e-02 3.55750382e-01
-8.18046462e-03 -8.92402083e-02 6.71514869e-01 -1.96981666e-04
-1.48311019e-01 1.14608482e-01 9.43734467e-01 7.27604106e-02
-6.17381871e-01 2.17620850e-01 -4.47605968e-01 -1.63651511e-01
9.96711314e-01 -7.25084543e-01 -8.87698382e-02 -6.85818732e-01
-9.15448070e-01 -2.31542408e-01 4.85327542e-01 5.81892431e-01
4.28346753e-01 -1.27630985e+00 -7.13592708e-01 4.08251286e-01
3.79757881e-02 -6.32107913e-01 5.44530571e-01 7.35452473e-01
-2.15938136e-01 4.31314856e-02 -9.22892213e-01 -3.95200491e-01
-1.57558918e+00 1.79832503e-01 5.65719485e-01 4.21051413e-01
-6.07791319e-02 9.69610035e-01 3.42684209e-01 -2.36609787e-01
2.99435973e-01 8.61394405e-02 -1.59585238e-01 4.82011229e-01
5.97659469e-01 7.04512537e-01 1.34685002e-02 -1.02706397e+00
-4.54240799e-01 5.96721828e-01 -1.25362143e-01 7.60755986e-02
1.41902864e+00 2.10734457e-01 -1.67596340e-01 -1.22373864e-01
1.40129137e+00 -1.05392829e-01 -1.22852886e+00 3.50950658e-01
-3.60707164e-01 -4.06373680e-01 -1.53847620e-01 -7.43757844e-01
-1.34929860e+00 6.02563977e-01 1.06403995e+00 -2.86499918e-01
1.39377081e+00 -1.43459573e-01 4.43027765e-01 -6.10607537e-03
1.07300848e-01 -1.20474160e+00 1.32305818e-02 2.39542112e-01
9.72321808e-01 -1.07704055e+00 1.79774761e-02 -3.31548035e-01
-6.48546875e-01 9.69179094e-01 9.35578704e-01 9.00685713e-02
4.31302547e-01 1.38084665e-01 3.07508886e-01 -1.20877624e-01
-2.65761077e-01 -9.46646929e-02 3.17530662e-01 9.20940816e-01
6.00392342e-01 -8.74727741e-02 -3.05660039e-01 8.59924734e-01
-3.01821470e-01 6.96113408e-02 2.64874369e-01 7.16581404e-01
5.33148088e-02 -1.21223795e+00 -4.48895931e-01 4.64557946e-01
-6.44384623e-01 2.68582672e-01 -6.39249921e-01 1.02588856e+00
4.40606773e-01 8.48203599e-01 2.91321665e-01 -5.09596348e-01
6.20004058e-01 3.90580237e-01 7.10674107e-01 -4.61868852e-01
-5.60744643e-01 -8.28320533e-02 4.59470786e-02 -6.56615078e-01
-5.83389580e-01 -8.02803874e-01 -9.06799734e-01 -6.08701408e-01
1.22549005e-01 -4.47322339e-01 5.89263856e-01 6.19988978e-01
5.11313796e-01 1.29230678e-01 4.91881073e-01 -7.66501606e-01
-4.44347173e-01 -1.10465527e+00 -6.41909122e-01 7.91541040e-01
-7.77557567e-02 -8.65062833e-01 -5.90145707e-01 2.22210839e-01] | [13.346035957336426, 1.4717254638671875] |
80371a28-a232-4113-905d-295671a9c6fd | real-time-variational-fisheye-stereo-without | 1909.07545 | null | https://arxiv.org/abs/1909.07545v1 | https://arxiv.org/pdf/1909.07545v1.pdf | Real-Time Variational Fisheye Stereo without Rectification and Undistortion | Dense 3D maps from wide-angle cameras is beneficial to robotics applications such as navigation and autonomous driving. In this work, we propose a real-time dense 3D mapping method for fisheye cameras without explicit rectification and undistortion. We extend the conventional variational stereo method by constraining the correspondence search along the epipolar curve using a trajectory field induced by camera motion. We also propose a fast way of generating the trajectory field without increasing the processing time compared to conventional rectified methods. With our implementation, we were able to achieve real-time processing using modern GPUs. Our results show the advantages of our non-rectified dense mapping approach compared to rectified variational methods and non-rectified discrete stereo matching methods. | ['Takeshi Oishi', 'Menandro Roxas'] | 2019-09-17 | null | null | null | null | ['stereo-matching'] | ['computer-vision'] | [ 2.17419311e-01 4.81407419e-02 4.44029689e-01 -3.64037365e-01
-2.30123937e-01 -5.18600583e-01 8.70014191e-01 -3.14910680e-01
-7.74968147e-01 7.65605152e-01 -5.91167472e-02 -1.89407229e-01
-2.83483695e-02 -9.54385221e-01 -9.20971155e-01 -3.16157103e-01
5.43340147e-01 9.07048166e-01 6.02912724e-01 -4.34539080e-01
6.78874731e-01 6.65361822e-01 -1.67610669e+00 -2.43799195e-01
8.81221056e-01 4.96937037e-01 5.72442055e-01 6.07671618e-01
3.13560143e-02 -5.60999252e-02 2.07935974e-01 -2.29258627e-01
6.60372257e-01 -1.71136498e-01 -5.64240575e-01 2.76000023e-01
8.47831964e-01 -5.56546628e-01 -4.85068202e-01 1.07429814e+00
1.65951416e-01 2.46665344e-01 4.51918423e-01 -1.02834451e+00
-1.03933051e-01 -1.62776381e-01 -7.32186317e-01 -7.84639195e-02
5.42406976e-01 -1.63328782e-01 5.46503246e-01 -9.05729651e-01
1.23887861e+00 1.19261360e+00 7.89814472e-01 4.15054679e-01
-1.26827979e+00 -4.00276989e-01 -2.57391542e-01 3.59928459e-01
-1.52754271e+00 -4.07541573e-01 6.14039183e-01 -4.73783463e-01
9.61693168e-01 -2.90938001e-02 8.83388460e-01 5.87146461e-01
3.99698198e-01 1.39445871e-01 1.03007102e+00 -5.04321694e-01
1.54900089e-01 -1.32285208e-01 -4.73729111e-02 5.81082642e-01
2.14735031e-01 4.34027195e-01 -2.46296644e-01 -3.63200046e-02
1.28025794e+00 2.98986793e-01 -3.41471612e-01 -9.49093223e-01
-1.41837347e+00 9.40373540e-01 2.57914156e-01 -1.19100481e-01
-2.98808396e-01 2.45872781e-01 -3.49778496e-02 1.76907808e-01
6.47834957e-01 2.34186679e-01 -5.89874461e-02 -2.01235041e-01
-9.39288139e-01 2.72508204e-01 7.66660571e-01 1.41227400e+00
1.23512888e+00 -1.54993683e-01 5.91874778e-01 4.93555218e-01
1.68555245e-01 7.32226968e-01 1.87753156e-01 -1.56471694e+00
2.83353001e-01 4.00335521e-01 2.92368621e-01 -8.23997557e-01
-3.62105310e-01 2.33189270e-01 -5.93482137e-01 6.08521760e-01
3.99903178e-01 1.12877727e-01 -6.99764550e-01 1.05663717e+00
3.82234931e-01 2.60874093e-01 1.65556103e-01 9.64676321e-01
3.99503410e-01 6.40081406e-01 -7.49349415e-01 -2.09024042e-01
9.31776106e-01 -7.08745718e-01 -5.75347483e-01 -5.61579503e-02
4.72080827e-01 -9.05415654e-01 6.42432928e-01 4.87698078e-01
-1.11159325e+00 -2.16793343e-01 -1.04444921e+00 -4.08694625e-01
4.44225781e-02 -3.05781692e-01 1.55013621e-01 3.14582646e-01
-1.15659857e+00 6.49144351e-01 -1.04964161e+00 -5.92671990e-01
-1.11714497e-01 5.03375292e-01 -7.96350539e-01 -1.24996543e-01
-6.64031744e-01 1.29457092e+00 4.94925082e-02 -1.16924860e-01
-1.98561594e-01 -5.95476627e-01 -1.08287394e+00 -2.03245699e-01
2.43141323e-01 -8.41499567e-01 1.10535157e+00 -3.77755195e-01
-1.73523664e+00 1.08957815e+00 -5.22931814e-01 -5.29728353e-01
8.17545295e-01 -3.46615493e-01 3.99824619e-01 2.18295559e-01
-6.27787188e-02 8.83685827e-01 4.70960885e-01 -1.06594050e+00
-5.58541596e-01 -4.55891460e-01 1.59098796e-04 4.73979205e-01
1.80906162e-01 -5.64649940e-01 -4.35700178e-01 -3.99391390e-02
5.50363243e-01 -1.32204890e+00 -4.57340688e-01 2.89807171e-01
3.22417952e-02 1.77564204e-01 1.24313748e+00 -2.98473954e-01
3.42794389e-01 -1.90563369e+00 4.15383190e-01 1.27166938e-02
8.97702798e-02 8.75869766e-02 1.34150535e-01 4.31097299e-01
3.82230282e-01 -4.05458480e-01 -2.83148319e-01 -5.61545610e-01
-1.54679850e-01 4.72262502e-01 -3.86038482e-01 1.06043923e+00
-1.56443045e-01 5.41242242e-01 -8.50196838e-01 -5.06537318e-01
7.78617501e-01 8.77428889e-01 -7.84113050e-01 4.11500335e-02
3.90654616e-02 6.44307673e-01 -3.22165340e-02 1.57029682e-03
1.04128480e+00 3.92537266e-01 -1.10087745e-01 -1.04417175e-01
-8.10577810e-01 2.47684911e-01 -1.31929338e+00 2.11392426e+00
-7.26849675e-01 9.24292982e-01 4.07774895e-01 -7.43050158e-01
1.15215111e+00 -4.19677123e-02 2.85779238e-01 -6.41621232e-01
9.63284820e-02 2.74483800e-01 -6.02179229e-01 -7.51963407e-02
1.05433702e+00 -2.97110587e-01 1.76365763e-01 2.65039712e-01
-2.06537053e-01 -9.30970073e-01 4.06172946e-02 1.21174268e-01
7.79669821e-01 5.50661325e-01 3.42525721e-01 -5.66241622e-01
4.27587360e-01 4.81790572e-01 5.01879036e-01 1.92727640e-01
3.18140626e-01 9.75470006e-01 4.69407300e-03 -5.66384733e-01
-1.56975102e+00 -1.08254027e+00 -5.11015475e-01 7.65755549e-02
8.35568190e-01 -3.32570851e-01 -5.73495328e-01 5.83538413e-02
6.03179298e-02 4.81886953e-01 -1.78929090e-01 2.77893364e-01
-8.79762173e-01 -1.75974905e-01 -1.19758556e-02 2.50618935e-01
7.09410906e-01 -5.11393011e-01 -9.68557656e-01 1.12619758e-01
-1.66079581e-01 -1.19071853e+00 -5.25636375e-01 -9.05810595e-02
-1.24276412e+00 -1.05140257e+00 -7.50261486e-01 -6.50973499e-01
7.78229594e-01 9.01612043e-01 7.93925464e-01 -1.72604784e-01
-1.62965819e-01 4.85666573e-01 -3.16156670e-02 1.84954286e-01
-2.80981719e-01 -2.31324673e-01 1.48306549e-01 -3.98195624e-01
2.89135784e-01 -7.21886456e-01 -5.39887547e-01 7.98771381e-01
-5.98647237e-01 2.78128505e-01 2.17271209e-01 7.51167953e-01
8.30107331e-01 -2.40430027e-01 -2.72158504e-01 -6.84205413e-01
1.25900134e-01 -1.71883166e-01 -1.46170616e+00 -2.95736283e-01
-6.38550103e-01 2.34670833e-01 2.77258843e-01 -1.50637358e-01
-1.07377923e+00 7.36763537e-01 -1.92636535e-01 -6.22799993e-01
2.72968947e-03 -9.18779373e-02 2.04240113e-01 -4.02760923e-01
4.56536412e-01 2.17460781e-01 3.92954379e-01 -3.97882640e-01
5.23137212e-01 5.38584113e-01 7.03997910e-01 -2.24335223e-01
8.14704478e-01 1.03508258e+00 4.08985972e-01 -1.18478370e+00
-1.21242711e-02 -7.57152498e-01 -1.12026322e+00 -2.24122584e-01
9.60118830e-01 -8.92174244e-01 -8.28361988e-01 3.65279377e-01
-1.53389776e+00 -2.47755185e-01 -2.58057535e-01 8.82676482e-01
-1.04934347e+00 7.92220891e-01 -3.13513994e-01 -5.65938532e-01
-7.78526068e-02 -1.38632774e+00 1.35386431e+00 -1.84047059e-03
-2.83072501e-01 -9.71086204e-01 4.31761682e-01 1.90660819e-01
2.59377837e-01 1.89777464e-01 6.91339150e-02 2.58827657e-01
-1.05013037e+00 -2.97367647e-02 -1.97220877e-01 -2.37864763e-01
-2.36596018e-01 1.56186633e-02 -7.00679481e-01 -1.87831521e-01
1.31347001e-01 4.97066155e-02 7.41157293e-01 3.95330012e-01
1.98439494e-01 1.12502322e-01 -4.44357187e-01 1.16467237e+00
1.63420522e+00 1.00015923e-01 9.33093786e-01 5.35885036e-01
5.39741039e-01 7.37439752e-01 8.48507285e-01 2.94418067e-01
4.94577289e-01 1.06623733e+00 6.26627743e-01 1.09757237e-01
6.58486262e-02 -3.56141925e-01 2.94759452e-01 5.81370950e-01
-2.98689455e-01 1.09645076e-01 -8.16407323e-01 6.23004615e-01
-1.94344783e+00 -9.54843283e-01 -8.48984122e-01 2.55807424e+00
4.35649633e-01 -2.02439398e-01 -2.33253404e-01 -8.34296923e-03
7.47308195e-01 -1.99780911e-01 -1.98382735e-01 -5.67510128e-01
1.14854351e-01 9.94163007e-02 1.00835216e+00 1.33748782e+00
-7.42701352e-01 1.02666163e+00 6.30038357e+00 2.55047023e-01
-1.04236460e+00 2.12758124e-01 -4.28105026e-01 -1.04305200e-01
-4.66079414e-01 4.04828161e-01 -1.08161676e+00 1.24672845e-01
6.41522527e-01 -1.45776182e-01 3.31195444e-01 7.73838103e-01
2.79648840e-01 -6.71039879e-01 -9.19293582e-01 1.10480678e+00
9.58715454e-02 -1.49080729e+00 -2.71190524e-01 6.25545740e-01
7.41012454e-01 2.77342170e-01 -4.86941546e-01 -4.29914027e-01
2.43437439e-01 -5.32712817e-01 5.88897109e-01 4.10274386e-01
7.00361550e-01 -5.47027707e-01 4.57742870e-01 6.47789955e-01
-1.07487357e+00 4.26784962e-01 -7.54714847e-01 -3.96962345e-01
6.41756296e-01 5.99645615e-01 -7.60742962e-01 2.85071403e-01
5.31213164e-01 8.54678810e-01 -7.94096477e-03 1.08575988e+00
-9.64427087e-03 -2.38812953e-01 -6.75142825e-01 1.60400242e-01
1.28104940e-01 -8.52121830e-01 7.93031394e-01 8.49891126e-01
7.22745657e-01 2.78549284e-01 -2.26048544e-01 7.57968485e-01
3.84596795e-01 -1.30112439e-01 -1.09004867e+00 7.15680599e-01
2.66693383e-01 9.90126014e-01 -7.72058487e-01 -1.90440759e-01
-4.52501118e-01 1.19061995e+00 1.35639207e-02 -7.74047971e-02
-5.13964176e-01 -4.74673033e-01 5.78607023e-01 5.53002894e-01
3.86452019e-01 -8.30851972e-01 -3.08292419e-01 -1.43440509e+00
-7.44015276e-02 8.76676664e-02 -1.69252291e-01 -1.09510589e+00
-6.32699668e-01 5.41621327e-01 4.04384732e-01 -1.45534933e+00
-4.97387677e-01 -5.32431424e-01 -3.98104668e-01 1.08579171e+00
-1.64710724e+00 -8.79239500e-01 -6.19588375e-01 8.66258442e-01
5.87720633e-01 3.29206675e-01 5.48121512e-01 -4.41558622e-02
2.86354840e-01 -2.83869892e-01 1.44244909e-01 -5.52473128e-01
8.05332482e-01 -1.16171885e+00 4.91603673e-01 9.06313002e-01
-1.15063302e-01 5.04958212e-01 9.18985486e-01 -6.60788417e-01
-1.61295414e+00 -6.95445001e-01 1.03385258e+00 -3.40811968e-01
4.55532312e-01 -3.52276087e-01 -7.93983400e-01 7.40875125e-01
3.41540337e-01 -1.06026828e-01 1.22161940e-01 -2.77200162e-01
-5.16101485e-03 -1.19014584e-01 -1.27661610e+00 3.55666846e-01
1.21702170e+00 -4.25006837e-01 -9.22668040e-01 2.00088605e-01
4.28797066e-01 -7.44075418e-01 -6.93174005e-01 1.57108098e-01
5.61941385e-01 -1.14478159e+00 9.51561868e-01 3.40651602e-01
3.42183709e-01 -3.37069064e-01 -9.27602276e-02 -1.29921091e+00
-1.38306305e-01 -8.97121131e-01 5.31117201e-01 6.68618917e-01
-1.27999067e-01 -8.75538588e-01 9.66716468e-01 5.41165411e-01
-5.14542937e-01 -9.10788327e-02 -1.14514410e+00 -7.89438963e-01
-8.95098075e-02 -2.46973902e-01 1.75848722e-01 6.71838224e-01
1.92147642e-01 1.42272592e-01 -5.28252184e-01 2.39009216e-01
9.92544055e-01 2.76964784e-01 1.03689492e+00 -1.41078055e+00
-6.62827045e-02 -4.40102406e-02 -7.90220439e-01 -1.59832239e+00
1.86159611e-01 -5.84589124e-01 2.30198771e-01 -1.47256875e+00
-5.21590933e-02 -3.58172894e-01 7.93837965e-01 -3.44512798e-02
4.27045137e-01 6.70684218e-01 6.28711581e-02 5.11609554e-01
-5.29466011e-02 5.07111609e-01 1.29003680e+00 2.74241239e-01
-3.47253561e-01 -1.08470075e-01 1.14532150e-01 8.27980816e-01
4.63055104e-01 -3.92018169e-01 -5.49052179e-01 -6.72849298e-01
2.81387985e-01 3.46891940e-01 4.97134626e-01 -8.12580287e-01
5.57896793e-01 -9.45297182e-02 -1.68116271e-01 -1.02437949e+00
8.56882811e-01 -9.63397980e-01 4.59719449e-01 5.12639284e-01
2.54345208e-01 3.20855618e-01 1.02542095e-01 5.36843300e-01
-1.96964338e-01 -5.02979457e-01 9.07489538e-01 -2.02054158e-01
-8.32593799e-01 1.31394520e-01 -5.35176635e-01 -2.28069440e-01
1.15501654e+00 -7.40382671e-01 -2.32483700e-01 -5.04520535e-01
-6.25564575e-01 1.19558014e-01 1.42994380e+00 9.79733914e-02
9.53228414e-01 -1.17259657e+00 -5.41657031e-01 5.11839569e-01
-1.12315670e-01 4.11994308e-01 1.04302205e-01 8.73677015e-01
-1.36998963e+00 6.28285050e-01 -5.56324184e-01 -1.11757910e+00
-1.41114879e+00 3.76129121e-01 2.11338744e-01 2.53324419e-01
-1.12059927e+00 3.02153379e-01 2.10290641e-01 -5.79484165e-01
-1.09929964e-01 -2.05092773e-01 5.65794893e-02 -3.61743897e-01
2.81620711e-01 6.68885171e-01 1.37153402e-01 -9.15432692e-01
-3.63568425e-01 1.33381712e+00 2.80472994e-01 -4.79787439e-01
1.46000624e+00 -4.12403584e-01 -1.38143048e-01 1.45498782e-01
1.19325554e+00 7.80508891e-02 -1.57355654e+00 3.95634659e-02
-3.86962682e-01 -7.99356341e-01 2.86862403e-01 1.89449593e-01
-8.72454941e-01 1.05496693e+00 3.59178811e-01 -1.44949675e-01
7.76942611e-01 -1.23599708e-01 8.40637028e-01 5.50988853e-01
8.32980752e-01 -7.90416241e-01 -6.05763674e-01 6.68880343e-01
8.11024427e-01 -1.12800717e+00 3.33159894e-01 -7.38754988e-01
-3.33673835e-01 1.46230054e+00 3.06818366e-01 -6.03930712e-01
6.70112431e-01 4.01889563e-01 -4.27693538e-02 6.46268995e-03
-5.00692785e-01 -1.74348339e-01 -1.22047298e-01 7.28179336e-01
7.55089223e-02 -4.59831804e-01 -6.40325248e-01 -6.24408126e-01
-2.54015684e-01 1.55819625e-01 1.15345740e+00 9.75206435e-01
-6.08763039e-01 -1.32984161e+00 -6.37946844e-01 -2.55270332e-01
2.58338124e-01 7.32033476e-02 -1.85035184e-01 9.40467894e-01
-1.58185720e-01 7.19287932e-01 7.15567470e-01 -2.96126753e-01
4.24066842e-01 -2.25424290e-01 8.37763429e-01 -5.07183075e-01
-5.34633249e-02 2.96857327e-01 5.04684635e-02 -7.43571341e-01
-7.07762241e-01 -8.09988618e-01 -1.30846071e+00 -5.10970771e-01
-3.20918351e-01 1.81225970e-01 1.09546936e+00 7.18916297e-01
3.49189550e-01 -3.84983957e-01 4.84138131e-01 -1.51791906e+00
-2.70253688e-01 -5.88578999e-01 -4.76936132e-01 2.83151746e-01
2.80849278e-01 -9.14888442e-01 -4.25696135e-01 2.33767778e-01] | [8.817781448364258, -2.567304849624634] |
3a49b367-405e-4b97-aec7-88b6dff4c643 | cae-mechanism-to-diminish-the-class | null | null | https://link.springer.com/chapter/10.1007/978-3-031-16210-7_12 | https://link.springer.com/content/pdf/10.1007/978-3-031-16210-7_12.pdf | CAE: Mechanism to Diminish the Class Imbalanced in SLU Slot Filling Task | Spoken Language Understanding (SLU) task is a wide application task in Natural Language Processing. In the success of the pre-trained BERT model, NLU is addressed by Intent Classification and Slot Filling task with significant improvement performance. However, classed imbalance problem in NLU has not been carefully investigated, while this problem in Semantic Parsing datasets is frequent. Therefore, this work focuses on diminishing this problem. We proposed a BERT-based architecture named JointBERT Classify Anonymous Entity (JointBERT-CAE) that improves the performance of the system on three Semantic Parsing datasets ATIS, Snips, ATIS Vietnamese, and a well-known Named Entity Recognize (NER) dataset CoNLL2003. In JointBERT-CAE architecture, we use multitask joint-learning to split conventional Slot Filling task into two sub-task, detect Anonymous Entity by Sequence tagging and Classify recognized anonymous entities tasks. The experimental results show the solid improvement of JointBERT-CAE when compared with BERT on all datasets, as well as the wide applicable capacity to other NLP tasks using the Sequence Tagging technique. | ['Nguyen Le Minh', 'Tung Le', 'Nguyen Minh Phuong'] | 2022-09-21 | null | null | null | advances-in-computational-collective | ['spoken-language-understanding', 'semantic-parsing', 'intent-detection', 'intent-classification', 'slot-filling', 'spoken-language-understanding'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'speech'] | [-6.49417564e-02 5.18423080e-01 -1.44546539e-01 -8.56714427e-01
-1.03942537e+00 -5.85221946e-01 3.89377087e-01 1.97655961e-01
-7.27859318e-01 1.27266264e+00 3.02037537e-01 -3.20923805e-01
4.26092446e-01 -8.31822336e-01 -7.13628829e-01 -2.22067654e-01
9.59654078e-02 9.80055928e-01 6.40518785e-01 -3.32359940e-01
3.83863389e-01 -1.00233138e-01 -1.24285507e+00 4.29193705e-01
1.45286810e+00 8.69083464e-01 3.49043518e-01 4.46201444e-01
-1.26683617e+00 8.46914351e-01 -7.26046741e-01 -4.86398280e-01
-9.81721506e-02 -3.87916595e-01 -1.37101889e+00 -2.16909125e-01
-1.04115196e-02 -1.39909342e-01 3.37467581e-01 8.98958623e-01
4.83758092e-01 1.28363073e-01 6.38322532e-01 -1.48410177e+00
-4.56270695e-01 8.07752609e-01 -3.32817763e-01 -3.52764651e-02
3.79395574e-01 -8.53912771e-01 1.20520568e+00 -5.70167601e-01
9.33239818e-01 1.66707098e+00 6.93007827e-01 7.89478302e-01
-5.81052601e-01 -7.57783771e-01 8.20792392e-02 3.15523654e-01
-1.09785831e+00 -1.86972171e-01 3.03498358e-01 -9.64464024e-02
1.27805018e+00 -1.02630734e-01 -2.19921201e-01 7.19303966e-01
4.18700129e-02 1.42597651e+00 1.19122934e+00 -6.24475002e-01
4.24195379e-01 3.36487859e-01 8.68101835e-01 7.94589281e-01
-7.64373690e-02 -4.17892039e-01 -5.71166575e-01 -2.09179461e-01
3.17976862e-01 -3.26857150e-01 1.95298746e-01 -2.03586787e-01
-8.50629151e-01 1.09610450e+00 2.30478823e-01 2.44837001e-01
-1.06597878e-01 -3.83291811e-01 7.41885543e-01 3.22964102e-01
7.60230362e-01 1.84949711e-01 -1.04828238e+00 -2.29233101e-01
-4.50858235e-01 8.25221613e-02 1.22764635e+00 1.16770709e+00
8.55517924e-01 -1.41897067e-01 2.15178747e-02 1.23002851e+00
2.70782292e-01 3.04581791e-01 6.11760557e-01 -6.58129632e-01
6.70394778e-01 7.59577692e-01 1.19704857e-01 -2.71251559e-01
-5.22328615e-01 1.41549036e-01 -3.61875504e-01 -1.67892143e-01
3.79636407e-01 -4.82263893e-01 -1.18021166e+00 1.69985342e+00
3.97411972e-01 2.70888824e-02 8.04065406e-01 6.89142823e-01
1.21907401e+00 9.42870677e-01 9.33716774e-01 -1.30219376e-02
1.63239086e+00 -1.12721181e+00 -1.06983483e+00 -3.67226362e-01
9.84414101e-01 -8.88691187e-01 6.32191658e-01 -3.34859230e-02
-5.03174067e-01 -4.47239518e-01 -7.86180735e-01 -2.81074971e-01
-7.70122886e-01 2.68762618e-01 1.02678084e+00 8.13009501e-01
-7.72089422e-01 2.26430371e-01 -6.47478878e-01 -1.02680683e+00
3.21762592e-01 4.16252434e-01 -6.95332527e-01 -8.61551017e-02
-1.56159258e+00 6.61174655e-01 9.75136280e-01 -2.38013133e-01
-3.29223633e-01 -1.42310649e-01 -1.07821929e+00 8.01406652e-02
5.65201581e-01 -2.71383166e-01 1.46063304e+00 -6.71382785e-01
-1.36285901e+00 1.12467337e+00 -5.70009589e-01 -6.38277769e-01
2.47793481e-01 -3.84890676e-01 -4.94809747e-01 -1.33261785e-01
5.55292308e-01 8.83502483e-01 1.74726248e-01 -1.06138682e+00
-9.69985366e-01 -5.05187094e-01 -2.36176446e-01 3.88232678e-01
-9.04145986e-02 2.59907484e-01 -7.49947205e-02 -2.75124818e-01
3.95887792e-01 -6.44347370e-01 1.85687691e-02 -4.98784542e-01
-2.28347227e-01 -9.86830473e-01 9.60420191e-01 -9.81851935e-01
1.02178717e+00 -2.12511635e+00 -4.64242935e-01 -1.98956519e-01
-2.88022727e-01 6.18233263e-01 -1.41472369e-01 6.66367829e-01
5.89515083e-02 1.24045491e-01 -3.22268039e-01 -4.47178364e-01
1.23524711e-01 7.26944506e-01 -3.09779435e-01 -2.02545449e-01
1.69901103e-01 7.74169922e-01 -6.92517579e-01 -8.68657589e-01
-2.26368643e-02 -2.98274279e-01 -4.23539937e-01 5.42196512e-01
-3.77744853e-01 6.48083687e-02 -8.76193941e-01 9.92299974e-01
8.89041662e-01 -2.78483797e-03 3.02983910e-01 9.12853628e-02
-2.26903439e-01 2.99781144e-01 -1.21451402e+00 1.84168971e+00
-3.17466646e-01 5.65847829e-02 1.71153516e-01 -1.36048377e+00
1.31954312e+00 4.72944140e-01 1.87821954e-01 -7.55485773e-01
8.75554755e-02 7.14659154e-01 -3.06336612e-01 -5.26625335e-01
6.27778769e-01 -1.27104387e-01 -5.56687772e-01 2.42449626e-01
4.56853807e-01 1.79767966e-01 1.33562177e-01 3.87114584e-01
9.01005387e-01 2.87770331e-01 6.50315940e-01 -4.53456402e-01
6.49152875e-01 4.52616364e-01 1.05988860e+00 7.34365761e-01
-6.66867614e-01 3.36101353e-01 6.48800969e-01 -2.66659290e-01
-7.77818680e-01 -1.02694297e+00 -2.59703219e-01 1.45589650e+00
3.47135186e-01 -4.64339964e-02 -1.02199006e+00 -1.23834264e+00
-1.17933387e-02 7.91537464e-01 -3.06821197e-01 3.29363346e-01
-4.23825085e-01 -1.01210558e+00 5.73974967e-01 4.71008927e-01
1.31025684e+00 -1.38553894e+00 -8.27917829e-03 4.90035862e-01
-4.52001363e-01 -1.43759167e+00 -2.10995257e-01 4.97212619e-01
-6.51655495e-01 -1.03159487e+00 -7.39334047e-01 -1.40489590e+00
2.47302249e-01 -9.46755186e-02 1.02009976e+00 -3.47958744e-01
-2.04285145e-01 1.10378221e-01 -9.04645205e-01 -7.16950595e-01
-2.97508895e-01 3.99157465e-01 -1.95510671e-01 -2.49057263e-01
9.93518114e-01 -1.34431366e-02 4.91959602e-02 4.58977848e-01
-6.14938319e-01 -8.44619051e-02 4.61374283e-01 1.11556172e+00
2.24452972e-01 -1.98768795e-01 1.23673153e+00 -1.41651773e+00
3.33247989e-01 -4.90638763e-01 -3.33213151e-01 7.08166897e-01
-3.86520445e-01 1.51274130e-01 3.45655203e-01 2.24143952e-01
-1.72790337e+00 9.72608402e-02 -6.04279518e-01 3.82204771e-01
-3.58239055e-01 2.76726276e-01 -6.43212497e-01 1.68840945e-01
3.54588151e-01 1.16962202e-01 -1.57167628e-01 -5.71479559e-01
4.61237021e-02 1.30627728e+00 4.25474495e-01 -6.81180656e-01
9.99552533e-02 -1.06732594e-02 -5.16423643e-01 -9.01678324e-01
-1.05419397e+00 -8.67518365e-01 -5.58351457e-01 3.89705360e-01
1.30266309e+00 -1.20859408e+00 -5.06626368e-01 7.68010497e-01
-1.32375002e+00 -2.11780131e-01 1.30919084e-01 3.56168240e-01
-2.95736939e-01 5.45331597e-01 -8.48525405e-01 -1.04582334e+00
-5.10761023e-01 -7.76628196e-01 1.26597440e+00 4.35389698e-01
-3.89357693e-02 -9.76683319e-01 -1.16006590e-01 6.93874180e-01
2.53724400e-03 -2.32242063e-01 1.12005818e+00 -1.53335559e+00
-4.02830094e-01 2.13128254e-02 -6.88333154e-01 2.88012743e-01
-1.28722176e-01 -7.19301522e-01 -1.04047298e+00 1.57164320e-01
-3.29721302e-01 -7.01415062e-01 8.63058031e-01 2.28736505e-01
5.93434513e-01 1.78184301e-01 -4.28329110e-01 1.93522319e-01
1.33880448e+00 7.68488288e-01 7.11442888e-01 5.84739566e-01
3.87109339e-01 9.08525825e-01 1.26987481e+00 1.84186533e-01
7.86862910e-01 2.66007066e-01 -7.84984455e-02 -4.50370423e-02
1.44925281e-01 -4.35791284e-01 8.30055922e-02 5.19989431e-01
3.69230151e-01 -4.71689463e-01 -9.40876484e-01 5.19805193e-01
-2.24904251e+00 -4.49773937e-01 -1.54760301e-01 1.66692853e+00
6.12678826e-01 -7.15465844e-02 -1.57300174e-01 -2.81250864e-01
1.01226306e+00 1.67891625e-02 -4.35025245e-01 -6.27195179e-01
-3.33676279e-01 1.52805805e-01 5.60001194e-01 5.13192713e-01
-1.38595521e+00 1.78737330e+00 5.27479315e+00 1.06551278e+00
-5.74140310e-01 2.53206164e-01 8.21990252e-01 1.02751577e+00
-6.41909614e-02 8.74843150e-02 -1.38843119e+00 3.11785817e-01
8.20471585e-01 -6.20121555e-03 -1.93628132e-01 1.11736894e+00
-2.72248030e-01 -5.27404308e-01 -5.47641456e-01 7.99055815e-01
4.78947274e-02 -9.12328601e-01 -5.54595776e-02 -2.47988582e-01
5.28593779e-01 -1.08189888e-01 -5.47876239e-01 9.78057861e-01
5.17150581e-01 -5.90602458e-01 3.03608954e-01 -1.70463145e-01
3.86741042e-01 -6.48239195e-01 1.43389559e+00 4.14433777e-01
-1.18665326e+00 8.85756314e-02 -7.56076217e-01 7.78160840e-02
5.25185645e-01 3.82280231e-01 -1.07264245e+00 7.25364745e-01
7.26938486e-01 4.30502921e-01 -9.78480652e-02 8.29133868e-01
-1.86963424e-01 7.02732086e-01 -3.08846354e-01 -2.85784423e-01
4.58120495e-01 -2.13190168e-01 1.79094374e-01 1.25017977e+00
2.72343814e-01 4.84131128e-01 5.95265269e-01 2.45841146e-01
-1.45964414e-01 6.73089921e-01 -2.64087647e-01 -2.66439527e-01
4.36850578e-01 9.95580614e-01 -1.02838421e+00 -5.92881143e-01
-4.58922029e-01 1.39973438e+00 4.74573404e-01 1.90330088e-01
-6.21108949e-01 -5.58706880e-01 3.10214818e-01 -4.32929158e-01
1.72332108e-01 1.00170113e-02 -3.09957653e-01 -1.20946276e+00
-1.55899465e-01 -3.34960312e-01 8.37029159e-01 -6.05954051e-01
-1.20661795e+00 7.14060068e-01 -1.28934056e-01 -7.56948113e-01
-2.37521335e-01 -7.38069057e-01 -6.09622300e-01 8.80591512e-01
-1.81302059e+00 -1.44054186e+00 5.28719462e-02 5.04485309e-01
1.00246131e+00 -5.25884569e-01 1.02080679e+00 4.07938123e-01
-6.50602520e-01 4.74952400e-01 1.56346206e-02 2.87655294e-01
1.01966798e+00 -1.49468386e+00 5.08379161e-01 5.82730234e-01
-1.84950992e-01 2.94943541e-01 5.45759678e-01 -9.98098671e-01
-6.91129088e-01 -8.95451248e-01 1.70373094e+00 -1.03417106e-01
4.41224307e-01 -4.98326629e-01 -1.00285900e+00 7.87232161e-01
1.33393615e-01 -2.48633474e-01 8.01818013e-01 4.16194588e-01
-2.70972639e-01 3.02474618e-01 -1.51028872e+00 -3.58325616e-02
1.09573817e+00 -3.01824719e-01 -1.10336113e+00 4.43842143e-01
9.21146154e-01 -3.97905588e-01 -5.09136796e-01 3.41316789e-01
1.75128266e-01 -8.70087206e-01 7.36252666e-01 -6.69897139e-01
8.80347192e-02 3.29909325e-02 -3.61628413e-01 -1.06857669e+00
2.10322931e-01 -1.04531385e-01 5.23320615e-01 1.89168000e+00
5.93085885e-01 -9.45628285e-01 1.12710416e+00 8.23372006e-01
-3.55465204e-01 -1.38811395e-01 -1.01527810e+00 -6.49801970e-01
2.64820233e-02 4.10168320e-02 4.54043090e-01 1.07370889e+00
1.21530145e-01 7.42405355e-01 -3.30281585e-01 1.27422616e-01
5.12784123e-01 1.35068387e-01 5.58840215e-01 -1.33927596e+00
1.11215189e-01 3.69369149e-01 -2.52870589e-01 -1.09988689e+00
6.94209874e-01 -7.83870518e-01 2.49215066e-01 -1.77820539e+00
8.63368884e-02 -6.82647765e-01 -1.91307776e-02 7.29416430e-01
-4.30294096e-01 -2.77654946e-01 9.85124633e-02 1.08468615e-01
-9.45237815e-01 7.70285606e-01 7.39683211e-01 1.19355373e-01
-2.61922628e-01 -2.14151796e-02 -4.03527141e-01 4.92471695e-01
7.65467703e-01 -3.92911345e-01 -2.29003698e-01 -2.48808578e-01
-5.11844456e-01 3.49680960e-01 -2.45345652e-01 -9.02370811e-01
1.18026644e-01 1.26927063e-01 5.20094782e-02 -6.76702142e-01
8.16807300e-02 -7.65982747e-01 -5.01865983e-01 4.16013896e-01
-1.93228811e-01 -1.68015942e-01 2.16628611e-01 4.74152863e-01
-6.11498177e-01 -5.95011175e-01 5.81235886e-01 -4.03117537e-01
-1.83205855e+00 1.02374606e-01 -3.06871772e-01 1.35550976e-01
1.04263330e+00 -1.58629134e-01 -4.53118175e-01 -7.31037185e-02
-8.49305689e-01 7.41535723e-01 -1.58050716e-01 4.01878268e-01
1.98042125e-01 -7.55924582e-01 -3.98461998e-01 1.47063762e-01
3.14326972e-01 -2.24679224e-02 7.00566471e-01 1.69533238e-01
-6.89396262e-01 6.57766640e-01 -3.30580711e-01 -3.58485579e-01
-1.25379372e+00 1.76553741e-01 -4.17410880e-02 -5.10568202e-01
-2.63735026e-01 1.09046423e+00 2.99422801e-01 -1.34542847e+00
3.19375366e-01 1.01813242e-01 -7.33659625e-01 1.88393161e-01
3.09092522e-01 1.62920892e-01 1.21161882e-02 -4.19529647e-01
-3.84382695e-01 2.09296629e-01 -2.99014539e-01 -1.08823972e-02
1.04278779e+00 -4.70652640e-01 -2.59430110e-01 2.03193694e-01
1.20496821e+00 -4.11918610e-01 -8.15112889e-01 -3.23329031e-01
5.22147417e-01 -3.77553739e-02 -3.49770039e-01 -1.08949471e+00
-4.52582836e-01 8.08821738e-01 7.69605041e-01 1.96449310e-02
6.73588753e-01 1.35875456e-02 1.32925296e+00 5.84507704e-01
8.39069784e-01 -1.42110789e+00 -3.55159789e-01 9.14899707e-01
2.73271322e-01 -1.75401640e+00 -6.03410006e-01 -1.05113459e+00
-1.11465883e+00 8.46443295e-01 1.04296732e+00 9.55602527e-02
6.46416843e-01 -3.68408449e-02 5.31074584e-01 1.86162814e-01
-3.21825802e-01 -5.03685296e-01 -3.09714556e-01 6.97215319e-01
5.43007195e-01 1.09450921e-01 -9.29750621e-01 9.39732313e-01
-9.68103707e-02 4.99328598e-02 2.74655551e-01 1.21959770e+00
-9.59848285e-01 -1.51084220e+00 -2.87557036e-01 3.45349193e-01
-5.04363656e-01 -1.40557125e-01 -3.13310504e-01 8.31514657e-01
8.96102563e-02 9.76784527e-01 1.38687208e-01 3.64464857e-02
2.83329576e-01 6.14742637e-01 -1.65829137e-01 -7.46677458e-01
-4.37825322e-01 -3.19698788e-02 6.99507952e-01 -3.92939150e-01
-4.37538743e-01 -6.49225712e-01 -1.76816833e+00 1.37963444e-01
-2.62767285e-01 1.07163107e+00 8.59103084e-01 1.16996992e+00
1.49084225e-01 2.77131855e-01 3.10745656e-01 -1.89705074e-01
-3.58117491e-01 -9.87690628e-01 -1.00721121e+00 3.62604618e-01
-2.85561085e-01 -6.94758773e-01 -1.15612134e-01 -2.07229540e-01] | [12.606319427490234, 7.445695877075195] |
4481f99a-1a0b-4158-af14-fefcba622126 | deep-learning-for-predicting-metastasis-on | 2303.05752 | null | https://arxiv.org/abs/2303.05752v1 | https://arxiv.org/pdf/2303.05752v1.pdf | Deep Learning for Predicting Metastasis on Melanoma WSIs | Northern Europe has the second highest mortality rate of melanoma globally. In 2020, the mortality rate of melanoma rose to 1.9 per 100 000 habitants. Melanoma prognosis is based on a pathologist's subjective visual analysis of the patient's tumor. This methodology is heavily time-consuming, and the prognosis variability among experts is notable, drastically jeopardizing its reproducibility. Thus, the need for faster and more reproducible methods arises. Machine learning has paved its way into digital pathology, but so far, most contributions are on localization, segmentation, and diagnostics, with little emphasis on prognostics. This paper presents a convolutional neural network (CNN) method based on VGG16 to predict melanoma prognosis as the presence of metastasis within five years. Patches are extracted from regions of interest from Whole Slide Images (WSIs) at different magnification levels used in model training and validation. Results infer that utilizing WSI patches at 20x magnification level has the best performance, with an F1 score of 0.7667 and an AUC of 0.81. | ['Kjersti Engan', 'Emiel A. M. Janssen', 'Helga Hardardottir', 'Saul Fuster', 'Christopher Andreassen'] | 2023-03-10 | null | null | null | null | ['whole-slide-images'] | ['computer-vision'] | [ 1.17686726e-01 -5.97659014e-02 -3.06379616e-01 1.41095728e-01
-8.20294917e-01 -4.97395217e-01 3.75470340e-01 6.60608351e-01
-8.10344458e-01 8.03157210e-01 -6.40234426e-02 -6.06522322e-01
3.25827450e-02 -6.91278338e-01 3.65484320e-02 -1.15479195e+00
8.26718286e-02 5.93731005e-04 2.32078031e-01 1.42195271e-02
2.80292809e-01 8.14399242e-01 -8.70490074e-01 1.33115565e-02
8.05441201e-01 9.77046549e-01 2.72470206e-01 9.52593029e-01
1.34158228e-02 4.12372977e-01 -7.30045617e-01 -2.93278366e-01
-1.49579495e-01 -2.89031595e-01 -7.71971643e-01 -6.52754977e-02
1.45663172e-01 -3.09021026e-03 -9.16225761e-02 8.55799735e-01
3.65597308e-01 -5.83413899e-01 9.61319625e-01 -6.88881755e-01
-3.41167033e-01 2.18409821e-02 -7.99996793e-01 3.87435615e-01
-1.58498734e-01 2.79607981e-01 6.09233856e-01 -4.02857482e-01
7.35699475e-01 3.59168202e-01 9.38873589e-01 5.03707588e-01
-1.08192658e+00 -3.57046872e-01 -4.97238785e-01 4.78420705e-02
-1.57231307e+00 7.94064254e-03 2.79745609e-01 -6.21248603e-01
7.77762592e-01 5.49491465e-01 9.90783691e-01 5.95541716e-01
8.19265127e-01 2.91773766e-01 1.33569086e+00 -5.00604451e-01
1.46693408e-01 2.33360976e-01 -2.58142024e-01 8.64415109e-01
4.30848718e-01 -1.26968950e-01 -1.10298194e-01 1.31430387e-01
7.99749136e-01 8.12802687e-02 -2.55676568e-01 1.83389857e-01
-9.59150970e-01 7.47696519e-01 6.25116169e-01 5.00449479e-01
-9.16670933e-02 1.05428182e-01 3.53558481e-01 2.14263096e-01
5.02738535e-01 2.52548754e-01 -1.54694244e-01 4.79113199e-02
-1.07692420e+00 -2.44318888e-01 2.79617727e-01 -1.47644803e-01
2.19837457e-01 -3.37813884e-01 6.53528944e-02 6.49727523e-01
3.27189803e-01 4.08991337e-01 6.78740621e-01 -2.57157803e-01
-4.04279351e-01 1.03415215e+00 -1.89112961e-01 -8.68611455e-01
-8.02276492e-01 -6.84910178e-01 -1.12864745e+00 6.20485663e-01
7.44475603e-01 -2.01693013e-01 -1.22124302e+00 1.13182878e+00
1.33534729e-01 -3.48157912e-01 -1.01947166e-01 8.29650283e-01
6.56064570e-01 2.49571040e-01 5.44490159e-01 5.10786735e-02
1.41551089e+00 -5.65582395e-01 -2.79166549e-01 5.04533648e-02
8.26530397e-01 -6.56368077e-01 8.21335912e-01 3.08152199e-01
-6.59062982e-01 -1.37847796e-01 -9.60894346e-01 1.77944630e-01
-5.35329580e-01 5.02708554e-01 5.69583118e-01 8.58923852e-01
-1.30882657e+00 4.45869833e-01 -9.40609097e-01 -1.06840456e+00
7.55580842e-01 4.48680490e-01 -7.66590118e-01 1.92715272e-01
-6.99140906e-01 1.01451981e+00 1.93120092e-01 7.15852007e-02
-7.85771966e-01 -6.81796670e-01 -4.41618562e-01 -3.43540490e-01
-2.56130159e-01 -5.58189631e-01 8.52308452e-01 -1.08184695e+00
-1.22691000e+00 1.21781075e+00 -1.72788888e-01 -3.86156231e-01
6.05174661e-01 5.75902581e-01 -5.65137625e-01 3.12191546e-01
-2.15956848e-02 6.97576761e-01 5.41726470e-01 -9.85913157e-01
-9.91020977e-01 -5.10865629e-01 -1.62268966e-01 3.68451774e-02
-4.29425150e-01 1.54852560e-02 -4.44634527e-01 -3.66818279e-01
-2.23378003e-01 -8.39380026e-01 -4.77893442e-01 4.58156496e-01
-3.24336559e-01 -5.92384562e-02 6.59678459e-01 -9.86313164e-01
1.10960078e+00 -2.15518832e+00 -2.61360973e-01 4.59130138e-01
4.51420128e-01 3.31625700e-01 1.51635617e-01 2.63777524e-01
1.27887890e-01 4.35444474e-01 -8.39734375e-02 1.77595764e-01
-3.97437930e-01 -1.75634339e-01 4.15532500e-01 7.65553474e-01
3.61357093e-01 8.86722445e-01 -7.15109885e-01 -6.78768635e-01
1.49317697e-01 7.37565517e-01 1.51182413e-01 -1.12478569e-01
2.11622834e-01 3.46309632e-01 -4.00221013e-02 9.64615166e-01
5.16178429e-01 -7.14261770e-01 2.45027021e-01 -5.15621006e-02
-1.97701469e-01 -4.90827411e-01 -4.53297049e-01 1.29304516e+00
-2.15376884e-01 1.10900152e+00 -2.13556853e-03 -4.44169164e-01
6.71956778e-01 2.56701738e-01 2.85889238e-01 -4.54356432e-01
2.35159189e-01 1.97596282e-01 1.90879941e-01 -7.81834304e-01
1.47611544e-01 -4.58743274e-01 2.77658254e-01 -9.04693678e-02
-3.14451337e-01 1.32480040e-01 2.22553238e-01 1.08024150e-01
1.20440590e+00 -4.87893969e-01 6.37435675e-01 -2.26984978e-01
4.73711520e-01 4.96905953e-01 1.52830005e-01 2.58456677e-01
-3.35945308e-01 5.39671421e-01 8.04327190e-01 -5.01199365e-01
-8.01140487e-01 -1.06452012e+00 -3.67514044e-01 4.02791589e-01
-5.17725758e-02 2.20113397e-01 -5.90397954e-01 -5.72830677e-01
-1.89205296e-02 7.44681805e-02 -1.08247769e+00 9.51899588e-02
-1.32025465e-01 -1.02110279e+00 6.64239228e-01 4.48251367e-01
4.13605928e-01 -7.94195533e-01 -8.63041937e-01 -3.33160311e-02
1.39772937e-01 -4.68873918e-01 -1.48506649e-02 1.60457700e-01
-8.24117661e-01 -1.40433288e+00 -1.33035576e+00 -8.56022477e-01
1.14142883e+00 -4.83571924e-02 7.19787478e-01 2.21576020e-01
-1.03523755e+00 2.86386255e-02 -9.69637111e-02 -5.23819208e-01
-4.26453918e-01 1.19903348e-01 -5.20607352e-01 -9.20457244e-02
4.43899155e-01 3.10119670e-02 -7.93708742e-01 -1.13572963e-02
-7.57639766e-01 1.61329061e-02 1.13193464e+00 8.00005376e-01
7.79498637e-01 2.14134082e-02 3.69895190e-01 -8.16294849e-01
5.78482151e-01 -3.57604176e-01 -3.53104144e-01 1.75017968e-01
-5.83014667e-01 -6.72747254e-01 5.14529645e-01 -2.24412844e-01
-8.09896052e-01 -2.76917089e-02 -1.86127648e-02 9.47611555e-02
-4.28298026e-01 6.69769406e-01 4.19279933e-01 -4.16772276e-01
8.70565176e-01 -3.87232304e-02 4.70502853e-01 1.16053373e-01
-4.41995591e-01 6.35601819e-01 3.04717779e-01 4.25824493e-01
4.28987652e-01 8.32255781e-01 4.19320077e-01 -1.08153677e+00
-5.14835775e-01 -6.32755160e-01 -3.81404102e-01 -6.62000299e-01
8.84988964e-01 -5.90992332e-01 -6.62411571e-01 7.96456575e-01
-5.63324451e-01 -4.64782000e-01 -8.82960111e-02 4.10263598e-01
1.57014150e-02 1.25243798e-01 -5.34536839e-01 -5.11540711e-01
-3.81315351e-01 -9.49900568e-01 8.31937790e-01 7.32297480e-01
-4.43117142e-01 -1.32229686e+00 1.37033761e-01 2.87936628e-01
6.66546106e-01 8.59260857e-01 9.06678379e-01 -2.99033523e-01
-1.09325655e-01 -7.37727702e-01 -5.67247570e-01 1.32624403e-01
3.19574058e-01 4.38727885e-01 -9.85541344e-01 -3.55436027e-01
-6.99646652e-01 -7.41607994e-02 8.77723455e-01 6.81088924e-01
1.05674696e+00 1.18781254e-01 -8.55429292e-01 5.86815774e-01
2.01970005e+00 2.75677145e-01 5.74715436e-01 4.06947255e-01
4.79378194e-01 6.12076879e-01 2.38619357e-01 9.77421775e-02
5.46633499e-03 1.07891701e-01 6.04478002e-01 -5.41288555e-01
-3.32663894e-01 2.48732902e-02 -1.98723122e-01 1.75890192e-01
-5.27242243e-01 -2.36658022e-01 -1.30126274e+00 7.75618136e-01
-9.59832311e-01 -5.72413862e-01 -3.90668929e-01 2.12128401e+00
5.88848531e-01 1.16959915e-01 4.28502448e-02 2.69410700e-01
6.81930184e-01 -1.67023763e-01 -3.61346364e-01 -2.71273226e-01
-1.63491994e-01 1.40808076e-01 6.67178273e-01 3.88713032e-01
-1.06098771e+00 5.31584442e-01 6.42064524e+00 8.55568230e-01
-1.69871974e+00 -1.24230549e-01 1.13818586e+00 -2.23324262e-02
1.67540729e-01 -2.41943330e-01 -5.35419881e-01 4.44857657e-01
8.86734307e-01 -1.06483817e-01 -2.39670977e-01 3.07563186e-01
3.19374949e-01 -7.15780139e-01 -4.41796511e-01 6.94407046e-01
1.03969447e-01 -1.45058405e+00 -8.02803412e-02 5.24556220e-01
4.93083119e-01 -7.83605352e-02 3.70457828e-01 -2.80340791e-01
-1.32679299e-01 -1.42798722e+00 3.20925042e-02 8.84825408e-01
1.42492294e+00 -7.76425898e-01 1.33038604e+00 -2.09805300e-03
-7.94746518e-01 7.42896572e-02 -1.43770427e-01 1.67493626e-01
-2.13532060e-01 5.47391295e-01 -1.47630823e+00 1.54018879e-01
5.48297405e-01 4.50919211e-01 -8.35051358e-01 1.31888485e+00
-3.74375992e-02 8.32540691e-01 -1.53529599e-01 -4.62220728e-01
2.97310680e-01 2.05976337e-01 2.47229397e-01 1.37398767e+00
4.29007560e-01 -2.03413248e-01 -3.72141778e-01 2.34123126e-01
2.42926627e-01 4.01353568e-01 -1.99106619e-01 -1.94941625e-01
5.71677685e-02 1.63909972e+00 -1.30183029e+00 2.46660993e-03
-3.31374139e-01 7.32704222e-01 8.39730054e-02 1.37056738e-01
-5.42341232e-01 -6.35610044e-01 3.55123043e-01 5.67925513e-01
-6.83639795e-02 2.60295749e-01 -4.67326224e-01 -4.59338367e-01
-3.75495553e-01 -5.44340909e-01 5.17662764e-01 -3.67660522e-01
-1.09431553e+00 5.18380582e-01 -6.16340697e-01 -1.16735661e+00
2.08697334e-01 -9.70223904e-01 -8.51650655e-01 1.02107954e+00
-1.53301156e+00 -1.29682887e+00 -6.62563026e-01 1.95200086e-01
2.60041744e-01 -1.38072327e-01 1.12435865e+00 -1.10590041e-01
-5.28303683e-01 7.29229569e-01 9.36321020e-02 2.68514544e-01
7.00704575e-01 -1.39660740e+00 -2.26964518e-01 4.01792616e-01
-6.05076492e-01 1.98777273e-01 3.41805428e-01 -4.75301474e-01
-8.77094805e-01 -1.06091058e+00 8.95524681e-01 -1.06913432e-01
8.86929631e-01 2.98775822e-01 -4.66248274e-01 1.63010925e-01
1.68194473e-01 -1.36817902e-01 1.26051342e+00 -1.74683183e-01
2.14777384e-02 -1.49977162e-01 -1.35038769e+00 6.93836927e-01
1.20256536e-01 -2.28155196e-01 2.07092673e-01 2.78577417e-01
-2.60334074e-01 -3.30114663e-01 -1.00167572e+00 3.12143266e-01
8.08682621e-01 -8.91948640e-01 5.95746994e-01 -2.30439037e-01
5.10354161e-01 -2.79179245e-01 3.12158704e-01 -1.15279078e+00
-4.28983599e-01 -7.15153385e-03 5.88175595e-01 6.78637266e-01
8.32163393e-01 -7.29325950e-01 1.21451020e+00 2.24594042e-01
1.59190789e-01 -1.35146773e+00 -8.87415707e-01 -3.51062059e-01
1.90852195e-01 2.04473242e-01 7.38255978e-02 7.24758148e-01
-3.32802832e-02 -1.17668055e-01 2.66912729e-01 1.08610682e-01
4.74058330e-01 -2.76287794e-01 4.27488208e-01 -1.31032407e+00
6.11336716e-02 -7.83159912e-01 -9.76506710e-01 -1.26359224e-01
-4.85091269e-01 -7.85644948e-01 -4.96639341e-01 -1.77797389e+00
3.54817569e-01 -2.06426874e-01 -4.13881183e-01 4.48123038e-01
-7.21427798e-02 8.94017279e-01 -2.22458661e-01 1.91100314e-01
-1.23977564e-01 -3.83339912e-01 1.45085037e+00 -4.40981060e-01
9.74043924e-03 -1.13255911e-01 -7.54143834e-01 7.20248461e-01
1.23708415e+00 -1.34360686e-01 -1.19281054e-01 -1.05506487e-01
9.70815197e-02 -1.31818689e-02 5.55476010e-01 -1.28539753e+00
3.72737646e-01 -2.58403838e-01 1.10995853e+00 -4.15927857e-01
2.35505074e-01 -6.38679564e-01 3.69548410e-01 9.35159743e-01
-1.33518934e-01 -1.95654407e-01 2.80775368e-01 5.19739747e-01
-2.85404384e-01 -2.14858159e-01 9.66010511e-01 -1.06066197e-01
-5.79207361e-01 2.51336604e-01 -6.18604481e-01 -4.09935683e-01
1.43377316e+00 -7.35469401e-01 -4.82762426e-01 8.77967998e-02
-9.30804431e-01 -9.64635015e-02 8.09940934e-01 -3.10626030e-01
5.54933727e-01 -7.46732712e-01 -8.14369678e-01 -1.20287715e-02
3.69496822e-01 -2.12167069e-01 5.47313750e-01 1.21716702e+00
-1.15983999e+00 3.24741483e-01 -4.34836626e-01 -5.61524272e-01
-1.58029485e+00 1.13155954e-01 5.69307446e-01 -4.43129718e-01
-3.76236469e-01 1.22196054e+00 -1.98855922e-01 2.05285490e-01
8.13564658e-02 -2.81211268e-02 -5.97682357e-01 6.43964112e-02
5.26877761e-01 4.61733162e-01 7.06601143e-02 -5.51976025e-01
-2.65269190e-01 6.20891929e-01 -3.77500802e-01 1.14909990e-03
9.21313107e-01 3.02860647e-01 -3.11292201e-01 4.89445567e-01
1.06037843e+00 1.20360054e-01 -1.00088120e+00 3.75409573e-01
-3.78725946e-01 -4.49976742e-01 1.48113936e-01 -1.26077926e+00
-1.00383675e+00 6.60118759e-01 9.86622512e-01 3.47750247e-01
1.20445967e+00 -1.76561791e-02 4.85797226e-01 -2.07292706e-01
1.11619577e-01 -9.01284337e-01 -2.56617516e-01 3.17438878e-02
4.73640412e-01 -1.31997097e+00 -7.09955320e-02 -3.18058044e-01
-4.25267249e-01 1.27394164e+00 3.37412030e-01 -2.50196368e-01
7.02384055e-01 3.29626679e-01 6.55948937e-01 -2.32814670e-01
-5.51187873e-01 -1.86366262e-03 1.50544360e-01 7.35005200e-01
6.73179448e-01 5.20484447e-01 -4.79834467e-01 1.74899101e-01
-4.72319834e-02 6.04474358e-02 3.53141457e-01 8.29463899e-01
-5.39873898e-01 -8.80221963e-01 -2.49129683e-01 1.01600504e+00
-7.00572908e-01 1.00808010e-01 -6.03942454e-01 1.14916468e+00
1.90049633e-01 7.68586338e-01 2.31835797e-01 -1.03435069e-01
-4.32717912e-02 -2.13255838e-01 3.27024698e-01 -1.73895553e-01
-6.20003939e-01 2.53913123e-02 -1.56430036e-01 -1.22591421e-01
-4.61152762e-01 -4.41751480e-01 -1.16986084e+00 -3.78218085e-01
-2.43509933e-01 1.18735507e-01 9.65491951e-01 5.74725688e-01
3.93945165e-03 6.22806549e-01 3.14014673e-01 -2.16341287e-01
2.03996301e-02 -9.84998822e-01 -8.79713833e-01 -1.13693140e-01
3.37444633e-01 -2.71756262e-01 -4.93608296e-01 1.98118061e-01] | [15.310314178466797, -3.024662971496582] |
0f6b356d-c71d-4a10-b21d-388e2694e7d6 | view-dialogue-in-2d-a-two-stream-model-in | null | null | https://aclanthology.org/2022.coling-1.531 | https://aclanthology.org/2022.coling-1.531.pdf | View Dialogue in 2D: A Two-stream Model in Time-speaker Perspective for Dialogue Summarization and beyond | Existing works on dialogue summarization often follow the common practice in document summarization and view the dialogue, which comprises utterances of different speakers, as a single utterance stream ordered by time. However, this single-stream approach without specific attention to the speaker-centered points has limitations in fully understanding the dialogue. To better capture the dialogue information, we propose a 2D view of dialogue based on a time-speaker perspective, where the time and speaker streams of dialogue can be obtained as strengthened input. Based on this 2D view, we present an effective two-stream model called ATM to combine the two streams. Extensive experiments on various summarization datasets demonstrate that ATM significantly surpasses other models regarding diverse metrics and beats the state-of-the-art models on the QMSum dataset in ROUGE scores. Besides, ATM achieves great improvements in summary faithfulness and human evaluation. Moreover, results on machine reading comprehension datasets show the generalization ability of the proposed methods and shed light on other dialogue-based tasks. Our code will be publicly available online. | ['Zhongfeng Wang', 'Siyuan Lu', 'Jiaxin Zhuang', 'Dongchen He', 'Keli Xie'] | null | null | null | null | coling-2022-10 | ['machine-reading-comprehension', 'document-summarization'] | ['natural-language-processing', 'natural-language-processing'] | [ 3.10141623e-01 3.41325164e-01 -3.38250399e-01 -5.11889756e-01
-1.00166500e+00 -6.18928254e-01 9.62708533e-01 5.17560482e-01
-1.08307935e-01 6.82095945e-01 1.17448449e+00 -2.10065618e-01
1.37652755e-01 -4.31832850e-01 -1.58250883e-01 -3.65706533e-01
1.74986809e-01 5.85241377e-01 2.51097649e-01 -6.81849599e-01
5.98950744e-01 -1.63662076e-01 -1.23745608e+00 7.07175314e-01
1.22529662e+00 5.23956597e-01 2.08529793e-02 1.05216599e+00
-4.77612972e-01 9.52559710e-01 -1.13363111e+00 -4.42470700e-01
-1.66155666e-01 -1.06925058e+00 -1.17134356e+00 2.42649272e-01
4.71929282e-01 -6.72543466e-01 -3.26494038e-01 7.55425334e-01
7.13395774e-01 2.86011219e-01 6.27873600e-01 -8.10658693e-01
-4.87348437e-01 1.00350261e+00 -4.81954843e-01 2.39944354e-01
1.13050544e+00 -2.26312280e-01 1.26969647e+00 -5.42196631e-01
4.14973617e-01 1.47005689e+00 3.10380757e-01 6.13937676e-01
-9.21068430e-01 -1.13522768e-01 4.55220222e-01 1.92116871e-01
-4.74538893e-01 -6.16072655e-01 8.35051477e-01 -8.09380487e-02
1.05989528e+00 6.54049754e-01 5.43924272e-01 9.82152164e-01
9.21662673e-02 1.37044227e+00 7.69855380e-01 -4.14297462e-01
8.09279680e-02 -1.33481607e-01 6.01141632e-01 2.62446761e-01
-1.26885593e-01 -7.44595289e-01 -9.76090729e-01 -1.61834031e-01
2.12062269e-01 -2.29347259e-01 -5.14226496e-01 1.09891929e-01
-1.22892869e+00 8.28887820e-01 -6.93112761e-02 1.92578748e-01
-3.97127330e-01 -3.83033216e-01 9.28275585e-01 4.29973900e-01
7.94462383e-01 4.85432327e-01 -1.42967492e-01 -5.18814683e-01
-1.05926836e+00 5.15184283e-01 1.24342120e+00 9.22035396e-01
2.25539982e-01 -1.85202956e-02 -7.17397511e-01 1.01617694e+00
5.94935156e-02 3.41783345e-01 5.44485211e-01 -1.05337167e+00
9.28848922e-01 6.32377744e-01 1.50135234e-01 -8.78084362e-01
-4.56948221e-01 -1.92571670e-01 -9.65748131e-01 -6.49340034e-01
2.42475286e-01 -1.28266856e-01 -4.67486769e-01 1.37732863e+00
1.43434644e-01 -1.78306505e-01 5.53985178e-01 7.48657227e-01
1.36515439e+00 1.03178799e+00 -2.63725489e-01 -7.34471381e-01
1.39115918e+00 -1.42763925e+00 -1.22793043e+00 -5.88709675e-02
5.55133343e-01 -7.24702954e-01 9.21997726e-01 5.18313348e-01
-1.41007280e+00 -5.41267633e-01 -1.08426499e+00 -2.95959473e-01
1.52053475e-01 4.40537110e-02 4.79308724e-01 5.18829107e-01
-9.38420236e-01 5.24575531e-01 -7.19771385e-01 -4.45121318e-01
-2.78028678e-02 -3.96645665e-02 -7.77304918e-02 1.07062355e-01
-1.30429983e+00 8.23494017e-01 2.61503369e-01 1.22980690e-02
-5.74579835e-01 -3.98842186e-01 -9.24777985e-01 1.69372231e-01
3.31825495e-01 -7.81771541e-01 1.95802402e+00 -5.51554084e-01
-2.11002827e+00 3.95649910e-01 -5.61255813e-01 -7.07028806e-01
6.19240582e-01 -5.37326038e-01 -4.38936166e-02 6.09135032e-01
1.22302743e-02 3.58707696e-01 5.16891241e-01 -1.13345015e+00
-7.44244814e-01 -2.13766024e-01 3.15697789e-01 7.88522005e-01
-3.12217534e-01 -8.18365812e-02 -2.90073335e-01 -2.87881076e-01
6.90783113e-02 -5.77248514e-01 6.39998466e-02 -9.13208127e-01
-7.24096239e-01 -6.05656385e-01 6.76817238e-01 -8.75335097e-01
1.64608788e+00 -1.68246591e+00 5.21167040e-01 -5.52044213e-01
2.35471547e-01 3.50457788e-01 -3.42957042e-02 1.21492064e+00
2.42125988e-01 4.23753681e-03 -2.94257939e-01 -7.04262674e-01
8.71460065e-02 -2.20696088e-02 -6.12705052e-01 1.61405087e-01
-2.08704378e-02 7.16950357e-01 -1.17321694e+00 -4.78569865e-01
1.92575321e-01 -5.58070876e-02 -3.27409089e-01 4.93112206e-01
-3.03699523e-01 5.56520045e-01 -5.15546322e-01 1.74313471e-01
4.39855516e-01 -8.05892199e-02 1.49651185e-01 7.06881061e-02
-1.30977228e-01 7.98146486e-01 -5.65699697e-01 2.05394936e+00
-3.03745627e-01 6.48919344e-01 -5.30684851e-02 -8.92812312e-01
8.70241165e-01 6.52429938e-01 2.94662297e-01 -5.88303030e-01
-5.80372848e-02 -5.03898636e-02 -1.08568653e-01 -5.42734206e-01
1.20676291e+00 4.24401835e-02 -4.08572704e-01 7.47000873e-01
1.02168456e-01 -4.31807518e-01 6.16498768e-01 7.75523663e-01
8.11595738e-01 -1.55992389e-01 6.65535271e-01 9.31747854e-02
6.69292569e-01 4.23121117e-02 2.63592064e-01 8.81793439e-01
-2.07131162e-01 6.87448502e-01 9.78005767e-01 3.12748514e-02
-8.54417384e-01 -6.97554886e-01 1.56823501e-01 1.37478137e+00
2.31371924e-01 -8.46924782e-01 -9.13870335e-01 -7.37165570e-01
-3.34862381e-01 1.17284083e+00 -4.30916458e-01 7.71061480e-02
-6.60427153e-01 -5.22308767e-01 6.13306999e-01 4.01206493e-01
6.13824368e-01 -7.69968629e-01 -6.28075838e-01 2.86725640e-01
-8.84740829e-01 -1.01666534e+00 -6.59780800e-01 -2.65501261e-01
-9.79884446e-01 -7.98277497e-01 -8.89309764e-01 -5.59867918e-01
2.15955913e-01 5.84388554e-01 1.04208565e+00 -1.66803762e-01
3.71442318e-01 5.28165638e-01 -7.77076006e-01 -5.99867880e-01
-8.24155688e-01 4.15049016e-01 -2.17622146e-01 -1.20573431e-01
1.01031318e-01 -3.82208765e-01 -5.06828725e-01 2.44607329e-02
-8.28366041e-01 4.53027368e-01 3.34214121e-01 9.19904351e-01
-5.04248850e-02 -3.37694496e-01 1.23560095e+00 -1.04527807e+00
1.36264598e+00 -4.65159625e-01 1.57750264e-01 3.52731973e-01
-3.13076437e-01 -6.49842853e-03 6.71964288e-01 -1.37162939e-01
-1.53287280e+00 -4.36017215e-01 -1.93521827e-01 2.63627082e-01
-6.17899634e-02 6.27609968e-01 -1.26202837e-01 8.46473396e-01
4.83367294e-01 4.75508124e-01 8.26975480e-02 -3.79367560e-01
6.39077485e-01 9.58761632e-01 6.10479414e-01 -5.17977655e-01
1.61348209e-01 2.11073637e-01 -5.12642860e-01 -1.14565623e+00
-1.03323317e+00 -8.99733007e-01 -6.48994505e-01 -2.98205346e-01
6.66697085e-01 -8.83248985e-01 -6.19680524e-01 4.96884465e-01
-1.61077523e+00 1.29236892e-01 -2.07602531e-01 2.43568599e-01
-7.34096706e-01 1.00546038e+00 -7.94552565e-01 -1.10322356e+00
-7.41898060e-01 -9.01097953e-01 1.25169051e+00 5.04062235e-01
-6.48731470e-01 -1.16734934e+00 1.69983268e-01 5.99556684e-01
2.92006016e-01 -2.80545782e-02 8.68807971e-01 -1.22600818e+00
-6.35713115e-02 -2.17829555e-01 1.42006591e-01 4.01677579e-01
2.98130155e-01 -1.41355291e-01 -8.76335680e-01 -1.47539690e-01
4.48506057e-01 -4.28036809e-01 1.10618567e+00 3.54792893e-01
6.80745184e-01 -5.59362888e-01 6.18518703e-02 -1.69002160e-01
7.29647517e-01 7.33825192e-02 5.33817589e-01 9.90541652e-02
3.56896192e-01 1.05451739e+00 7.40657806e-01 7.50547886e-01
8.51725996e-01 3.25951636e-01 2.45080575e-01 4.99127246e-02
1.00145796e-02 -4.35189545e-01 5.39911985e-01 1.60609829e+00
-1.29616708e-01 -7.65880883e-01 -5.73708057e-01 5.61290562e-01
-2.30844402e+00 -1.35161674e+00 -2.36801580e-01 1.98279369e+00
9.90788460e-01 1.00999393e-01 3.38530988e-01 1.52727440e-01
6.66790009e-01 7.67022908e-01 -4.18126076e-01 -8.17566812e-01
-2.32336909e-01 -3.54976445e-01 -1.10177062e-01 6.36160493e-01
-8.46306086e-01 7.79380620e-01 6.70537758e+00 5.84620953e-01
-8.43493581e-01 -9.29361284e-02 4.18146461e-01 -7.31595159e-02
-3.57801139e-01 -9.89709795e-02 -7.56484747e-01 2.06668690e-01
1.08619606e+00 -7.84573376e-01 2.14087497e-02 4.07454491e-01
4.80211914e-01 -4.25580740e-01 -1.37945175e+00 6.51612639e-01
4.94217992e-01 -1.17305446e+00 3.36192459e-01 -2.64941216e-01
7.00846314e-01 -3.83648694e-01 -2.07301170e-01 4.75907207e-01
9.79392976e-02 -8.61453652e-01 4.67555702e-01 5.23870111e-01
3.83868903e-01 -5.50349295e-01 9.20985341e-01 8.04120839e-01
-9.05857623e-01 2.10841954e-01 -2.60064334e-01 -2.62652874e-01
5.49791157e-01 1.75199807e-01 -1.21405923e+00 1.21044135e+00
4.51667607e-02 8.51656675e-01 -3.66201222e-01 6.45032525e-01
-1.93401664e-01 8.08812141e-01 2.97596194e-02 -2.81975448e-01
2.30033770e-01 -1.87080279e-01 8.68522882e-01 1.53085124e+00
4.80471700e-02 3.80799800e-01 2.90921718e-01 3.25060576e-01
-2.56089196e-02 3.13524693e-01 -4.92228210e-01 -1.06566064e-01
4.04375762e-01 1.06137419e+00 -3.92460763e-01 -6.66146517e-01
-3.36715698e-01 1.09195614e+00 -7.97547866e-03 3.72799039e-01
-5.67603827e-01 -4.72141415e-01 1.39643952e-01 -2.89174706e-01
1.20640740e-01 -1.79214031e-01 -2.38804042e-01 -1.30533266e+00
6.71937466e-02 -1.15433240e+00 3.45061302e-01 -4.01061893e-01
-1.10721087e+00 6.38878107e-01 3.28369588e-01 -1.31786883e+00
-7.38710523e-01 1.50847770e-02 -9.21081662e-01 7.36195087e-01
-1.42842340e+00 -8.70362222e-01 -2.21672535e-01 2.19120800e-01
1.38676703e+00 -6.25639930e-02 9.61519718e-01 -3.26144457e-01
-5.09940505e-01 3.84676158e-01 1.69622988e-01 -1.60687625e-01
9.74469721e-01 -1.48576176e+00 4.59270388e-01 6.76584840e-01
-7.21841231e-02 5.95710814e-01 1.00748920e+00 -5.50784767e-01
-1.25536990e+00 -5.99056780e-01 1.17800200e+00 -4.17749465e-01
4.63176131e-01 -3.41683440e-03 -1.11904538e+00 5.33866763e-01
1.06783736e+00 -1.14402854e+00 9.57202375e-01 2.60558844e-01
-1.31463101e-02 9.48397368e-02 -7.68297732e-01 5.25953770e-01
6.95739150e-01 -3.76880348e-01 -1.39213908e+00 4.22569454e-01
8.85625780e-01 -7.32584238e-01 -8.02920222e-01 2.30162561e-01
3.27378273e-01 -1.15648270e+00 5.40516436e-01 -6.69249892e-01
9.10715103e-01 4.98700328e-02 -3.62146758e-02 -1.63955307e+00
2.03511670e-01 -8.99155676e-01 -3.49250168e-01 1.36503017e+00
2.21032813e-01 -4.48600143e-01 4.56296086e-01 3.01251084e-01
-4.47874963e-01 -6.72113061e-01 -7.75179625e-01 -4.55906540e-01
1.27863660e-01 -6.16973601e-02 4.95837182e-01 5.76328516e-01
7.92890728e-01 1.13612556e+00 -6.31000698e-01 -2.45139718e-01
3.32711637e-01 2.30620176e-01 1.01924837e+00 -1.15554249e+00
-1.22241378e-01 -5.75415015e-01 1.56186000e-01 -1.82579780e+00
2.38226473e-01 -6.07838571e-01 2.37545043e-01 -2.12546945e+00
3.57446253e-01 3.87901336e-01 2.60439754e-01 -1.71672814e-02
-5.88665843e-01 -4.95361269e-01 3.30868363e-01 4.01908159e-01
-1.09233820e+00 1.01951969e+00 1.36657560e+00 -2.44348660e-01
-4.06299174e-01 2.44382054e-01 -9.68459845e-01 6.38206899e-01
6.90771461e-01 -7.50927031e-02 -6.22768283e-01 -4.53909636e-01
-2.64948338e-01 8.73971522e-01 -3.44884068e-01 -5.64693332e-01
6.04619801e-01 -6.78276867e-02 -6.06502779e-02 -1.14917612e+00
5.26002884e-01 -5.61843365e-02 -4.17051762e-01 2.04673037e-01
-9.62557733e-01 4.35472280e-02 -1.36722345e-02 7.65066624e-01
-6.27230167e-01 -3.48980665e-01 2.20493287e-01 -1.18563406e-01
-2.13540316e-01 -1.32304519e-01 -5.14274418e-01 2.34866470e-01
6.85936809e-01 -2.00128883e-01 -6.30447328e-01 -1.06955719e+00
-3.10492903e-01 6.87536180e-01 1.72291964e-01 4.62141126e-01
5.67208469e-01 -8.89717162e-01 -1.23095763e+00 -3.03379297e-01
8.18031281e-02 2.50732094e-01 5.44569612e-01 7.86341131e-01
-3.02830219e-01 9.21716213e-01 -1.07687954e-02 -7.55611598e-01
-1.40563273e+00 -8.14410741e-04 4.84355427e-02 -6.31900012e-01
-6.05629027e-01 5.64130008e-01 2.64903635e-01 -3.56554449e-01
4.10475969e-01 -2.02917382e-01 -7.60604739e-01 5.07038355e-01
9.09803331e-01 6.32182598e-01 -1.01806968e-02 -5.29372931e-01
-2.24650316e-02 1.48642600e-01 -4.96382952e-01 -4.39113498e-01
1.16090488e+00 -6.06297612e-01 -1.96802616e-01 1.05001414e+00
8.79188716e-01 2.27911100e-01 -1.05352175e+00 -4.04090822e-01
1.17849082e-01 -1.15744986e-01 -4.37599421e-01 -7.16593683e-01
-2.67747343e-01 9.84697580e-01 -3.02571505e-01 8.66030455e-01
1.09290886e+00 -9.18612927e-02 9.79587495e-01 5.25885701e-01
-7.71855935e-02 -1.10991323e+00 3.22849661e-01 9.09105361e-01
1.25344896e+00 -1.17934704e+00 1.52066857e-01 -3.58051449e-01
-1.19956255e+00 1.46175146e+00 5.25776446e-01 1.73335716e-01
-1.05532952e-01 -2.79476136e-01 1.56432316e-01 -2.12644972e-02
-1.20090389e+00 1.43317446e-01 3.27643007e-01 3.05367142e-01
8.14092934e-01 1.14706293e-01 -7.50226915e-01 7.33730733e-01
-5.36287546e-01 -3.25226277e-01 1.12012553e+00 8.06476176e-01
-6.16828024e-01 -1.03612065e+00 -2.47971490e-01 3.58523339e-01
-4.99798089e-01 3.94493826e-02 -8.05037975e-01 5.02632618e-01
-9.18117166e-01 1.50872076e+00 -7.86194354e-02 -2.34820291e-01
6.27410412e-01 2.25269347e-01 4.04880404e-01 -9.97986853e-01
-8.39504898e-01 1.04558781e-01 5.84109247e-01 -1.99285179e-01
-6.18157923e-01 -7.19730318e-01 -1.29919171e+00 -4.36334908e-01
-4.18678731e-01 6.02241933e-01 6.08433366e-01 9.72682834e-01
2.19932660e-01 7.69854784e-01 9.40617859e-01 -8.85711730e-01
-1.08415377e+00 -1.47172165e+00 -3.23297143e-01 3.40271324e-01
6.64964676e-01 -6.27920590e-03 -3.35315645e-01 1.06324917e-02] | [12.584975242614746, 9.178815841674805] |
2e4c8e18-4629-48fc-bbe5-2acf25e03271 | learning-non-autoregressive-models-from | null | null | https://openreview.net/forum?id=UNzc8gReN7m | https://openreview.net/pdf?id=UNzc8gReN7m | Learning Non-Autoregressive Models from Search for Unsupervised Sentence Summarization | Text summarization aims to generate a short summary for an input text. In this work, we propose a Non-Autoregressive Unsupervised Summarization (NAUS) approach, which does not require parallel data for training. Our NAUS first performs edit-based search towards a heuristically defined score, and generates a summary as pseudo-groundtruth. Then, we train an encoder-only non-autoregressive Transformer based on the search result. We also proposed a dynamic programming approach for length-control decoding, which is important for the summarization task. Experiments on the Gigaword headline generation and DUC2004 datasets show that NAUS achieves state-of-the-art performance for unsupervised summarization, yet largely improving inference efficiency. Further, our algorithm is able to perform length-transfer summary generation. | ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['headline-generation', 'abstractive-sentence-summarization', 'unsupervised-sentence-summarization'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [ 5.50255597e-01 4.87915456e-01 -1.34391636e-01 -4.02389020e-01
-1.55519617e+00 -4.77676034e-01 5.68660855e-01 4.07119840e-01
-3.32760125e-01 9.38919425e-01 7.99592257e-01 -1.41510606e-01
3.21422368e-01 -8.17853689e-01 -9.55872536e-01 -4.89328682e-01
2.64444113e-01 8.38287711e-01 8.76459554e-02 -7.01977685e-02
6.02019608e-01 -5.49569577e-02 -1.11758482e+00 3.77957791e-01
1.41033912e+00 4.43196714e-01 4.48308110e-01 1.31259227e+00
-7.84964859e-02 8.16534817e-01 -9.52515364e-01 -5.80941021e-01
-2.68716812e-01 -1.03664029e+00 -9.41825151e-01 1.19650573e-01
4.87189174e-01 -3.99282187e-01 -3.13915104e-01 7.43665099e-01
6.59604430e-01 3.17511111e-01 8.56209695e-01 -4.74399537e-01
-4.74694461e-01 1.18310153e+00 -3.04292381e-01 3.04613467e-02
5.00355661e-01 -9.55112278e-02 1.42743731e+00 -6.07867420e-01
6.00798011e-01 1.08257914e+00 3.93902123e-01 6.19916439e-01
-1.17871356e+00 -2.35144183e-01 5.41533045e-02 -5.14565110e-02
-8.76475036e-01 -5.21178484e-01 5.94551206e-01 -3.12394109e-02
1.20419848e+00 5.89367747e-01 6.75705671e-01 1.03715491e+00
3.31371337e-01 1.17324543e+00 3.43796730e-01 -5.68247437e-01
3.96315247e-01 -4.06480730e-01 3.81435484e-01 6.48260236e-01
3.06773543e-01 -5.43699622e-01 -6.06283128e-01 -2.35639304e-01
3.25632453e-01 -4.98988479e-01 -1.43434092e-01 2.54724771e-01
-1.20889914e+00 8.42403710e-01 -9.71626043e-02 5.47155514e-02
-7.63451457e-01 1.89775750e-01 7.22743571e-01 1.07172348e-01
7.88129210e-01 6.84565187e-01 -2.32654989e-01 -4.78000104e-01
-1.45439994e+00 3.11203659e-01 1.13832784e+00 1.02981222e+00
5.48989117e-01 2.87476748e-01 -8.06390762e-01 8.93778324e-01
-1.51417062e-01 4.81989533e-01 6.88673377e-01 -1.05782115e+00
6.93511307e-01 3.32054317e-01 5.82080670e-02 -6.92369759e-01
-1.01475880e-01 -3.99411380e-01 -1.06310594e+00 -5.14680505e-01
-6.44403175e-02 -4.54380095e-01 -8.00948083e-01 1.49546266e+00
-9.16026384e-02 1.37691781e-01 3.88565540e-01 4.50258166e-01
9.04022932e-01 1.02430141e+00 -1.90486044e-01 -6.63205445e-01
1.00870526e+00 -1.41480660e+00 -7.64123917e-01 -2.84617722e-01
7.18221068e-01 -4.51339394e-01 9.27942634e-01 3.98441374e-01
-1.56415260e+00 -3.89143646e-01 -1.03734422e+00 -3.67445081e-01
3.61834824e-01 4.87616748e-01 2.96405256e-01 2.97716051e-01
-1.08002079e+00 7.42228866e-01 -1.00930417e+00 -2.78412521e-01
1.29692286e-01 1.50498182e-01 -7.75262639e-02 2.86667734e-01
-8.31506073e-01 5.82102418e-01 8.54531765e-01 -1.53120533e-01
-6.58011198e-01 -4.75197345e-01 -1.00948524e+00 4.03304487e-01
3.78132850e-01 -1.14159107e+00 1.59907913e+00 -7.41048157e-01
-2.05371952e+00 4.84329402e-01 -7.04011977e-01 -1.03866196e+00
3.96154195e-01 -5.04249990e-01 8.88575688e-02 2.82994390e-01
1.04553938e-01 4.46986496e-01 6.38325155e-01 -1.02858496e+00
-5.56345880e-01 -1.23594195e-01 -2.76707470e-01 4.25332099e-01
-1.96293846e-01 -6.77961204e-03 -6.94280148e-01 -7.77375817e-01
-1.99246883e-01 -8.12693000e-01 -4.33862388e-01 -1.06061292e+00
-8.89862061e-01 -4.04241621e-01 2.91663557e-01 -1.16758716e+00
1.83024848e+00 -1.64671755e+00 5.46773016e-01 -8.69649276e-02
-9.83918607e-02 4.62758601e-01 -3.34110528e-01 7.74063230e-01
3.27918798e-01 8.31917077e-02 -6.81123555e-01 -8.09300005e-01
-7.03912228e-02 6.60747960e-02 -4.99670416e-01 -1.65602043e-01
2.47104406e-01 1.11284029e+00 -1.09442651e+00 -6.86323583e-01
-1.76286086e-01 1.88889783e-02 -8.59611928e-01 5.09815812e-01
-6.57221317e-01 2.63450205e-01 -4.16973889e-01 1.07812092e-01
2.17506409e-01 -5.12123555e-02 3.09438586e-01 7.33521255e-03
-1.93485275e-01 6.40457690e-01 -6.21984839e-01 1.97673416e+00
-5.72244763e-01 6.41945302e-01 -4.22947139e-01 -9.52201128e-01
9.59811330e-01 1.25646561e-01 1.21491194e-01 -3.14182639e-01
1.20681696e-01 1.60500079e-01 -2.91061401e-01 -2.32801303e-01
1.31048286e+00 3.45132917e-01 -2.23260731e-01 6.71852350e-01
1.34179115e-01 -5.21441281e-01 7.01123297e-01 5.36538303e-01
1.20801342e+00 1.26929238e-01 5.35549343e-01 -4.17472236e-03
6.01270258e-01 3.71467024e-02 5.46680629e-01 1.04076278e+00
4.97138560e-01 1.05156493e+00 8.43843222e-01 1.01547308e-01
-1.22834969e+00 -8.35996211e-01 4.12145019e-01 1.06937659e+00
-2.81850606e-01 -8.87599528e-01 -1.21647239e+00 -6.78641915e-01
-4.48886395e-01 1.44935131e+00 -2.99121290e-01 -1.94951057e-01
-8.52483034e-01 -6.50358140e-01 7.01975703e-01 4.90578532e-01
4.34050530e-01 -1.13474810e+00 -5.10037661e-01 6.05425358e-01
-5.45770824e-01 -8.74725163e-01 -8.62214446e-01 -2.86690053e-02
-1.01381266e+00 -6.80504262e-01 -8.65861356e-01 -8.35812032e-01
6.37900889e-01 -4.24872078e-02 1.08862817e+00 -1.97241157e-01
2.49677345e-01 1.33649498e-01 -3.63394022e-01 -4.52057481e-01
-9.56766963e-01 6.74421191e-01 -2.95237035e-01 -1.54375598e-01
-1.15328982e-01 -4.80888546e-01 -3.55151862e-01 -2.58217543e-01
-8.14642727e-01 3.17736268e-01 7.61685252e-01 9.27386284e-01
8.14111352e-01 -2.53533185e-01 8.50824833e-01 -1.14052439e+00
1.13126731e+00 -2.05972746e-01 -3.66960645e-01 3.58072996e-01
-4.35435981e-01 4.04409200e-01 9.71666992e-01 -1.40444964e-01
-1.27191985e+00 -4.82223816e-02 -3.40569943e-01 -1.20485742e-02
1.75023958e-01 7.66459107e-01 -2.18167920e-02 7.47291803e-01
6.09135449e-01 7.39375770e-01 -9.97220352e-02 -3.44805688e-01
4.78812754e-01 9.10042584e-01 8.57264102e-01 -4.05373037e-01
4.63955224e-01 5.63367568e-02 -2.83775389e-01 -1.09843898e+00
-9.85608578e-01 -4.41898555e-01 -5.44009030e-01 2.08469316e-01
7.94040859e-01 -8.76034260e-01 -2.34535038e-01 4.57830757e-01
-1.50314784e+00 -5.99127948e-01 -3.81298840e-01 3.78724128e-01
-9.81403589e-01 8.06558251e-01 -6.98148608e-01 -7.70943820e-01
-1.28161693e+00 -6.85746908e-01 1.17632329e+00 2.67738074e-01
-5.30653059e-01 -7.67523766e-01 4.21879113e-01 3.55076402e-01
1.56708524e-01 1.83693953e-02 7.54085958e-01 -1.09692979e+00
-4.52681243e-01 -2.48573884e-01 1.70320924e-02 6.20678246e-01
-1.65918484e-01 8.63956660e-02 -5.31523645e-01 -1.24191217e-01
-2.12590992e-01 -2.63741434e-01 1.21431553e+00 5.41161478e-01
1.22356045e+00 -7.98865139e-01 -7.04958737e-02 4.83900040e-01
1.00046170e+00 1.23672199e-03 8.01012695e-01 7.94242397e-02
6.34395242e-01 2.99441338e-01 7.52711177e-01 6.19872928e-01
4.72605854e-01 3.93595755e-01 -4.13984023e-02 2.80198961e-01
-4.56449725e-02 -5.90701044e-01 5.67097783e-01 1.38834953e+00
-8.70380253e-02 -7.47302115e-01 -5.52114427e-01 5.97019196e-01
-2.08719730e+00 -1.13431120e+00 -2.56800532e-01 2.21288610e+00
1.17069757e+00 1.33752227e-01 2.79812552e-02 -1.31665036e-01
6.40466332e-01 4.00237679e-01 -5.17363489e-01 -8.25818360e-01
-4.61550988e-02 3.70691001e-01 3.91008139e-01 7.12393403e-01
-9.03386176e-01 1.33159864e+00 5.94893217e+00 8.90092969e-01
-7.65940249e-01 -1.78327978e-01 5.15316486e-01 -7.78580904e-02
-4.59743351e-01 -6.38181493e-02 -7.56516337e-01 4.98980254e-01
1.32813430e+00 -7.72867441e-01 1.92194760e-01 8.24320316e-01
5.00542879e-01 -2.40656540e-01 -9.21277583e-01 6.71579897e-01
4.05100733e-01 -1.44142890e+00 4.27886784e-01 -2.59140491e-01
9.53353703e-01 -8.98777917e-02 -4.65953767e-01 4.90723640e-01
3.81800801e-01 -6.26722932e-01 4.79635239e-01 5.68681359e-01
6.81735575e-01 -1.00039113e+00 6.81388259e-01 6.37401879e-01
-7.51855791e-01 1.21094950e-01 -5.37361264e-01 1.13230385e-01
4.88952905e-01 6.57536983e-01 -9.93671775e-01 8.46124053e-01
-9.56864208e-02 7.37395525e-01 -4.77376640e-01 9.64137793e-01
-6.05399549e-01 1.04622722e+00 -8.85809734e-02 -2.72968918e-01
2.03551039e-01 -3.86802256e-01 8.67253900e-01 1.67701828e+00
4.42893982e-01 1.97288349e-01 8.40348303e-02 5.54540813e-01
-4.27401096e-01 2.77495980e-01 -2.46772677e-01 -2.65060306e-01
3.32306594e-01 1.03808260e+00 -4.00977105e-01 -8.18960130e-01
1.61389783e-01 1.55280805e+00 3.79292101e-01 2.80551940e-01
-7.80636847e-01 -7.14799643e-01 5.88332601e-02 -3.02016944e-01
3.48193258e-01 -2.81155914e-01 -3.23623747e-01 -1.42786455e+00
-4.00820374e-02 -9.16164160e-01 2.79645324e-01 -7.13742316e-01
-6.74003482e-01 6.36942625e-01 -7.88464174e-02 -8.64180148e-01
-9.32606995e-01 2.40390077e-01 -1.15647233e+00 6.41745806e-01
-1.18604410e+00 -9.14287329e-01 -1.11026570e-01 -1.65402293e-01
1.20041120e+00 -9.81967449e-02 8.74652863e-01 -2.14024812e-01
-7.02001929e-01 6.23491764e-01 4.04159635e-01 -8.12890846e-03
7.42561042e-01 -1.50530744e+00 9.50374067e-01 1.08266866e+00
1.06717981e-01 5.25100589e-01 1.00706542e+00 -9.04968262e-01
-1.33090639e+00 -1.31974971e+00 1.30462348e+00 -1.27756506e-01
4.92747873e-01 -1.05288498e-01 -9.06477153e-01 7.52590775e-01
5.65932512e-01 -1.01940751e+00 6.17263317e-01 -4.17631026e-03
7.38737211e-02 2.18049854e-01 -5.81303537e-01 6.83747828e-01
9.79942918e-01 -2.22825825e-01 -1.06816745e+00 3.89731199e-01
1.02652872e+00 -6.49230003e-01 -5.77018321e-01 9.62111428e-02
4.67392802e-01 -7.87544191e-01 6.01187110e-01 -6.39209688e-01
1.08073819e+00 1.33849710e-01 1.69655025e-01 -1.80270290e+00
-3.79818857e-01 -1.01929760e+00 -4.43910629e-01 1.50410342e+00
5.42208612e-01 -3.42573494e-01 8.13658476e-01 2.44291410e-01
-5.92622399e-01 -6.69802785e-01 -4.16798711e-01 -6.66058898e-01
-1.11149345e-02 -1.29918247e-01 3.84704709e-01 1.77756518e-01
1.45337600e-02 8.39672863e-01 -4.90765750e-01 -4.04373705e-02
6.40350997e-01 1.99910671e-01 1.09484971e+00 -8.17582309e-01
-4.07639474e-01 -4.24753964e-01 6.72479793e-02 -1.49954367e+00
4.89227355e-01 -9.10564899e-01 2.93983638e-01 -2.10990667e+00
4.08298463e-01 3.48668963e-01 1.91758364e-01 1.27881929e-01
-5.83972514e-01 -1.57563016e-01 1.49046376e-01 7.33429641e-02
-8.82386982e-01 9.50945854e-01 1.00053227e+00 -1.60089776e-01
-5.33667505e-01 2.55013287e-01 -8.20763528e-01 4.55858439e-01
9.61368084e-01 -3.71090084e-01 -3.37146789e-01 -4.16444629e-01
7.18377456e-02 4.27722633e-01 -1.90558553e-01 -9.08057451e-01
3.83415878e-01 5.83377071e-02 -4.92045730e-02 -1.21537697e+00
5.42363077e-02 7.60658309e-02 -2.00824901e-01 4.16925877e-01
-7.84962237e-01 -5.73652200e-02 -4.09142524e-02 6.58268332e-01
-2.87804395e-01 -5.90474844e-01 4.91518527e-01 1.13092298e-02
-1.59678951e-01 7.85876065e-02 -5.04757106e-01 3.47885042e-01
5.22518158e-01 3.18112858e-02 -2.10896552e-01 -5.65289199e-01
-3.86627555e-01 4.86932009e-01 4.11621958e-01 3.58571634e-02
4.78995949e-01 -8.48374963e-01 -1.39557672e+00 -1.32979393e-01
-1.57418013e-01 3.74073654e-01 1.81729481e-01 4.19688642e-01
-6.47748053e-01 5.89480937e-01 2.94998527e-01 -3.32613349e-01
-1.25080717e+00 1.91909477e-01 -2.14636490e-01 -7.98652351e-01
-6.94942594e-01 5.62536776e-01 -3.48115750e-02 -2.18924001e-01
6.37098700e-02 -3.04121524e-01 -1.52773768e-01 2.61284709e-01
7.35908031e-01 5.61885774e-01 1.25669330e-01 -2.65188098e-01
1.86144188e-01 1.87148616e-01 -3.13783675e-01 -3.48704308e-01
1.36569965e+00 -1.67479500e-01 -1.95179015e-01 4.41099674e-01
1.09407437e+00 2.61884600e-01 -1.01820624e+00 -5.54059930e-02
8.14319104e-02 6.23257533e-02 -1.38113156e-01 -6.20708883e-01
-5.02449334e-01 6.01260722e-01 -4.81591254e-01 1.35667115e-01
1.03883731e+00 -9.95529518e-02 1.23192477e+00 8.72991383e-01
-8.12064856e-02 -1.19002485e+00 6.70218244e-02 9.70184445e-01
1.05858362e+00 -7.65707493e-01 1.20083645e-01 -2.35048726e-01
-9.01167631e-01 1.16723037e+00 2.91265041e-01 -2.40695789e-01
-2.64632523e-01 -3.61725390e-02 -4.11244333e-01 1.66402921e-01
-1.12186599e+00 -7.07479864e-02 2.83050478e-01 2.90544778e-01
4.86110896e-01 4.72499207e-02 -7.66754746e-01 6.58430338e-01
-7.05134630e-01 -1.50546217e-02 9.06861842e-01 6.11859620e-01
-7.25070536e-01 -1.03694820e+00 -5.09699956e-02 7.42365241e-01
-4.75298136e-01 -3.67909044e-01 -5.14382303e-01 2.57317990e-01
-8.48939121e-01 1.00573933e+00 1.39103189e-01 -1.30345225e-01
3.67398590e-01 1.83485329e-01 3.78140301e-01 -9.32775676e-01
-7.61722565e-01 2.04641998e-01 5.26930034e-01 -2.90610790e-01
4.07646820e-02 -6.52764976e-01 -1.37614191e+00 -1.09935425e-01
-3.15499783e-01 5.26490033e-01 6.73536539e-01 7.27576494e-01
5.59230387e-01 7.12269306e-01 7.28591740e-01 -8.36950660e-01
-8.07291925e-01 -1.29475892e+00 -2.17902958e-01 1.54658183e-01
1.32467598e-01 3.42726588e-01 -2.64399350e-01 1.88995719e-01] | [12.464950561523438, 9.451364517211914] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.