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
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
d665a3b2-0d93-4a5b-a2ea-80cc9a2cd1d1
|
on-monoaural-speech-enhancement-for-automatic
|
2205.01751
| null |
https://arxiv.org/abs/2205.01751v2
|
https://arxiv.org/pdf/2205.01751v2.pdf
|
On monoaural speech enhancement for automatic recognition of real noisy speech using mixture invariant training
|
In this paper, we explore an improved framework to train a monoaural neural enhancement model for robust speech recognition. The designed training framework extends the existing mixture invariant training criterion to exploit both unpaired clean speech and real noisy data. It is found that the unpaired clean speech is crucial to improve quality of separated speech from real noisy speech. The proposed method also performs remixing of processed and unprocessed signals to alleviate the processing artifacts. Experiments on the single-channel CHiME-3 real test sets show that the proposed method improves significantly in terms of speech recognition performance over the enhancement system trained either on the mismatched simulated data in a supervised fashion or on the matched real data in an unsupervised fashion. Between 16% and 39% relative WER reduction has been achieved by the proposed system compared to the unprocessed signal using end-to-end and hybrid acoustic models without retraining on distorted data.
|
['Jon Barker', 'Rama Doddipatla', 'Catalin Zorila', 'Jisi Zhang']
|
2022-05-03
| null | null | null | null |
['robust-speech-recognition']
|
['speech']
|
[ 6.76061928e-01 -3.56558412e-02 8.39830339e-01 -4.15198684e-01
-1.48315632e+00 -2.79676616e-01 5.79128861e-01 -3.59927207e-01
-6.75743043e-01 6.18988156e-01 4.49699074e-01 -2.12708831e-01
-2.21478194e-01 -7.69339651e-02 -5.11986732e-01 -1.18963897e+00
1.10323653e-01 -3.50194395e-01 -6.87529370e-02 -2.08080083e-01
-9.59088206e-02 2.26334676e-01 -2.08566785e+00 1.09960563e-01
9.40840006e-01 1.06046855e+00 6.92697048e-01 9.84468758e-01
2.63662517e-01 1.94806248e-01 -9.67256904e-01 9.01715457e-03
4.60998356e-01 -5.83963990e-01 1.60554834e-02 2.38214701e-01
5.55742502e-01 1.06722079e-02 -4.46605772e-01 1.27252483e+00
1.23396552e+00 4.66566890e-01 5.04768193e-01 -4.89517301e-01
-1.57831848e-01 6.40708745e-01 -2.21482620e-01 2.24551439e-01
1.19164914e-01 1.08613737e-01 3.95000964e-01 -1.07112837e+00
-1.00204490e-01 7.15338230e-01 3.81067455e-01 3.32861125e-01
-9.37966943e-01 -7.15341508e-01 -4.50771123e-01 1.98771223e-01
-1.23878467e+00 -1.32204282e+00 8.18288505e-01 3.91933247e-02
9.66265559e-01 5.85791528e-01 9.48758423e-02 8.08066308e-01
-1.55091658e-01 3.40098470e-01 1.55350101e+00 -9.31583166e-01
1.29566401e-01 3.83931734e-02 2.64385700e-01 5.06203026e-02
-2.31121089e-02 5.68751752e-01 -6.28244638e-01 1.51591867e-01
1.25770107e-01 -8.03039014e-01 -6.29726648e-01 2.89646804e-01
-9.11693156e-01 1.10574022e-01 -7.56807998e-03 6.77392125e-01
-4.88151044e-01 -3.07997972e-01 2.09105045e-01 4.75483775e-01
6.30508602e-01 2.85369754e-01 -2.68384039e-01 -3.22158843e-01
-1.30658603e+00 -2.81827807e-01 5.09627938e-01 7.17773020e-01
2.21307710e-01 9.75308895e-01 3.31118256e-02 1.46836281e+00
4.48004305e-01 8.82995665e-01 6.33768618e-01 -6.00399256e-01
4.63400960e-01 -4.89455253e-01 5.40423533e-03 -3.92060459e-01
-5.60248382e-02 -1.07447124e+00 -7.00377464e-01 3.63594562e-01
9.44315195e-02 -3.96448940e-01 -1.48047304e+00 1.68271601e+00
2.47469768e-01 2.75264025e-01 5.78948438e-01 8.71398687e-01
6.75303400e-01 9.78648126e-01 -1.30008131e-01 -6.16836309e-01
7.98503280e-01 -9.51589823e-01 -1.26179266e+00 -1.29234239e-01
-1.25057310e-01 -1.42417872e+00 5.69381118e-01 8.54223490e-01
-1.39919448e+00 -1.10709178e+00 -1.41878843e+00 4.58759785e-01
-1.70632526e-01 2.94782460e-01 -3.33751708e-01 1.32401335e+00
-1.15531969e+00 3.33202988e-01 -4.85969812e-01 -3.21058370e-02
-2.33735397e-01 3.72020394e-01 -3.94270778e-01 -4.19769250e-02
-1.16327715e+00 7.52880752e-01 3.04370105e-01 4.43224072e-01
-8.93804610e-01 -4.34988409e-01 -9.23045874e-01 9.96248052e-02
-1.96455978e-03 -1.55531436e-01 1.34100497e+00 -8.57306361e-01
-1.88680184e+00 3.01993728e-01 -2.39405975e-01 -4.73411530e-01
1.60350710e-01 -2.13751853e-01 -1.27556455e+00 5.29133156e-03
-4.61765260e-01 -1.19865917e-01 1.31373346e+00 -1.24210393e+00
-5.46135485e-01 -1.96522355e-01 -8.10592115e-01 4.22368735e-01
-1.58601791e-01 1.92457497e-01 -1.65473461e-01 -9.05516505e-01
3.70992959e-01 -4.49242443e-01 -2.28916928e-02 -6.71278059e-01
-2.11471468e-02 3.85084033e-01 8.51243913e-01 -1.20792067e+00
9.81666803e-01 -2.45218945e+00 -1.68826610e-01 2.87389308e-01
-4.19181228e-01 9.21583712e-01 -2.08252341e-01 2.04658255e-01
-4.22145426e-01 -4.81709212e-01 -5.25615811e-01 -6.01120949e-01
2.56567933e-02 3.83468717e-02 -4.01670337e-02 4.97345477e-01
-3.27270851e-02 2.66133752e-02 -4.96775061e-01 -1.40842885e-01
4.70829010e-01 9.04337227e-01 -1.33453026e-01 5.95414579e-01
5.89219332e-01 4.90503490e-01 4.98442411e-01 4.40702111e-01
9.90245223e-01 8.58110189e-01 -7.64509384e-03 -1.18084483e-01
-1.88921228e-01 3.58680308e-01 -1.49045122e+00 1.54400432e+00
-6.97944582e-01 7.07471192e-01 8.60782027e-01 -8.98783863e-01
1.16592216e+00 9.36918378e-01 -8.71963520e-03 -8.28867316e-01
1.24042720e-01 5.59091449e-01 4.50963885e-01 -4.35343325e-01
5.21712542e-01 -5.15100837e-01 2.47325897e-01 8.25962946e-02
5.10193110e-01 -2.08554402e-01 -2.62851626e-01 -2.34078571e-01
6.54561639e-01 -1.27608806e-01 1.99859455e-01 -1.21701486e-01
8.61495912e-01 -7.72152901e-01 2.87649542e-01 6.88595176e-01
-2.58843869e-01 6.59236848e-01 -5.61372817e-01 6.77741170e-01
-9.94275153e-01 -1.35650861e+00 -3.50363761e-01 9.53398883e-01
-2.09166080e-01 1.15991384e-01 -9.26056445e-01 -9.15942341e-03
-5.35779893e-01 9.56284583e-01 3.51725109e-02 -5.77413253e-02
-4.25710052e-01 -7.06717789e-01 8.47119451e-01 2.41652802e-01
5.04952133e-01 -7.52865076e-01 1.71016201e-01 4.11268950e-01
-1.21089704e-01 -1.11443543e+00 -3.97087514e-01 5.57360351e-01
-4.78411615e-01 -2.75207847e-01 -7.79454470e-01 -9.61775720e-01
4.81864452e-01 2.91744828e-01 3.29860479e-01 -3.43525916e-01
3.76463570e-02 3.75613898e-01 -4.80177224e-01 -5.10822833e-01
-8.83710563e-01 -5.77587485e-01 4.46572870e-01 4.46134984e-01
-2.32435539e-02 -7.05459952e-01 -2.91607857e-01 4.38979268e-01
-8.75733435e-01 -3.48322630e-01 6.97203815e-01 9.40436661e-01
3.03198129e-01 4.82776135e-01 1.01129150e+00 -8.67750719e-02
7.50712574e-01 -1.55595735e-01 -4.42330539e-01 -1.34916216e-01
-6.65865421e-01 -1.43635318e-01 4.06486690e-01 -4.22777474e-01
-1.80448806e+00 -7.54536390e-02 -7.84309208e-01 -1.29313052e-01
-6.11142039e-01 3.35090101e-01 -6.74283147e-01 -2.21733361e-01
5.87667525e-01 6.17403090e-01 7.51026999e-03 -8.46129000e-01
1.87536448e-01 1.56092179e+00 1.15312719e+00 -3.32467765e-01
7.68573940e-01 -6.02340922e-02 -3.53004932e-01 -1.41958368e+00
-9.00035724e-03 -8.38060617e-01 -3.54790151e-01 -3.31295788e-01
4.79482621e-01 -9.21494603e-01 2.37181503e-02 9.30576563e-01
-9.22126830e-01 -1.09740384e-01 -6.19966984e-02 1.07162356e+00
-3.72993052e-01 6.76429093e-01 -4.92189199e-01 -1.41251767e+00
-5.17935872e-01 -1.17592978e+00 7.17957556e-01 1.46307945e-01
1.94405198e-01 -5.80304563e-01 -3.83278541e-02 6.22159004e-01
7.98177600e-01 -3.31251413e-01 3.87720495e-01 -7.24757493e-01
2.83301454e-02 -4.11880583e-01 3.27046394e-01 1.17607069e+00
3.83545518e-01 -2.93552577e-01 -1.67648268e+00 -3.23403955e-01
5.98927975e-01 1.46062180e-01 6.88577294e-01 4.21171814e-01
4.42795008e-01 7.91381225e-02 2.96488464e-01 4.60956186e-01
1.25456345e+00 7.84617364e-01 9.66992378e-01 -1.40546307e-01
9.94777009e-02 5.18506527e-01 5.38351715e-01 7.95197040e-02
-5.18456876e-01 5.09437621e-01 1.57651037e-01 -2.45193139e-01
-6.27036810e-01 5.04594762e-03 6.33155882e-01 1.59611344e+00
1.70912534e-01 -4.06325668e-01 -4.45569485e-01 7.35334754e-01
-1.15465915e+00 -9.65833306e-01 -2.59833068e-01 2.43331695e+00
7.92300761e-01 1.23195045e-01 -1.36959687e-01 8.72391939e-01
9.52829301e-01 7.99781084e-02 2.20024828e-02 -7.39241540e-01
-4.26787406e-01 9.35237110e-01 2.96858758e-01 9.71809387e-01
-8.83397937e-01 4.00311023e-01 5.90902185e+00 1.10027361e+00
-1.18776727e+00 4.32322919e-01 1.84117809e-01 -1.06511541e-01
1.90492630e-01 -4.24934894e-01 -2.74737239e-01 3.18885595e-01
1.57530737e+00 -8.09423998e-02 3.29313576e-01 4.66887414e-01
4.67262954e-01 -2.58002073e-01 -5.95592439e-01 1.00075400e+00
3.67409766e-01 -4.67589706e-01 -4.21424955e-01 -9.54416953e-03
5.73590159e-01 -1.52081817e-01 3.51806819e-01 2.39608467e-01
-2.94286102e-01 -8.67333651e-01 7.27950394e-01 5.49498260e-01
8.01666915e-01 -7.90900588e-01 7.99958348e-01 4.89231944e-01
-9.95427430e-01 1.87351942e-01 -3.80050912e-02 -3.46799940e-02
2.41445154e-01 5.76582134e-01 -1.08238518e+00 8.73981655e-01
3.49661171e-01 -1.66310549e-01 -1.96119905e-01 1.42922783e+00
-2.87985981e-01 1.21945298e+00 -3.22616786e-01 3.12166244e-01
8.79734755e-02 -5.66509515e-02 9.96378779e-01 1.50748873e+00
5.83736956e-01 -1.28142051e-02 -6.50348127e-01 2.33687416e-01
3.02635878e-01 7.68234804e-02 -4.34787422e-01 7.62814656e-02
1.56563342e-01 1.16037583e+00 -1.41845807e-01 -2.17019320e-01
-2.19295442e-01 8.73624146e-01 -5.81104159e-01 6.09892845e-01
-7.36013293e-01 -9.82175887e-01 4.02980983e-01 -1.72022417e-01
5.81443548e-01 -2.83394694e-01 -1.48246348e-01 -6.40608311e-01
5.62508963e-02 -1.03150046e+00 -1.28796965e-01 -1.02648079e+00
-9.23724174e-01 1.06657493e+00 -3.18527699e-01 -1.19274247e+00
-3.99787396e-01 -5.53609371e-01 -5.93469262e-01 1.46411943e+00
-1.27993631e+00 -7.95840621e-01 7.37146884e-02 5.51662087e-01
6.73776329e-01 -3.71784985e-01 7.57198930e-01 7.48636723e-01
-2.84120977e-01 7.80371785e-01 5.45024157e-01 -2.20346287e-01
8.01754057e-01 -1.10631621e+00 2.11633623e-01 1.55061579e+00
1.15637630e-01 4.68200505e-01 1.13175178e+00 -2.64154941e-01
-9.27038491e-01 -8.78499925e-01 6.46065295e-01 2.23699063e-01
1.37301564e-01 -2.77250260e-01 -1.01399148e+00 4.72468548e-02
7.17134774e-01 -2.51563698e-01 8.55262756e-01 -3.29028070e-01
-9.25742984e-02 -3.37384880e-01 -1.18748105e+00 2.40500987e-01
5.59136629e-01 -6.48248434e-01 -1.05818331e+00 -1.62274316e-01
6.48096323e-01 -2.88413286e-01 -7.85959005e-01 4.12293285e-01
4.17929024e-01 -8.65383148e-01 6.96930468e-01 -5.49857728e-02
-1.87474489e-01 -6.41039550e-01 -6.77228451e-01 -1.68840241e+00
7.18038976e-02 -1.06269252e+00 2.05855131e-01 1.68824220e+00
7.13054061e-01 -7.11077094e-01 1.99478790e-01 3.09085518e-01
-7.18607843e-01 -1.42806396e-01 -9.93700981e-01 -9.63189960e-01
-2.81733185e-01 -8.66058886e-01 2.10086778e-01 4.19537872e-01
-7.44962990e-02 1.62662819e-01 -5.98688781e-01 6.75270021e-01
8.62225235e-01 -5.10422170e-01 6.43648267e-01 -6.31903768e-01
-6.24753952e-01 -2.88475361e-02 -5.39097786e-01 -9.39435482e-01
-1.23800479e-01 -5.24369657e-01 5.68580985e-01 -1.26952767e+00
-5.32105744e-01 3.01114712e-02 -6.73296511e-01 -1.83396608e-01
-2.88980454e-01 3.53968680e-01 6.37433082e-02 -4.64248568e-01
-9.48996991e-02 5.88018477e-01 9.65173841e-01 -4.73813973e-02
-1.20198399e-01 3.65192145e-01 -5.31119645e-01 4.30359215e-01
6.26538157e-01 -3.59355509e-01 -3.30017895e-01 -1.66453809e-01
-6.85337722e-01 3.26028556e-01 1.81777999e-01 -1.55845773e+00
2.64096528e-01 3.91786546e-01 2.22604886e-01 -4.99479711e-01
7.74709225e-01 -8.85140002e-01 3.54368269e-01 4.80330400e-02
-3.48141462e-01 -6.30782306e-01 5.34871578e-01 6.09939992e-01
-6.03239179e-01 -2.88997531e-01 1.06780195e+00 3.66289347e-01
-4.68335122e-01 -4.61294889e-01 -4.90226269e-01 -4.84805048e-01
5.39138496e-01 -3.70838463e-01 -1.06109828e-01 -7.63947368e-01
-1.07996452e+00 -4.78152812e-01 -1.57271728e-01 9.26303416e-02
6.29244447e-01 -1.03190243e+00 -8.64972949e-01 4.12103951e-01
-2.04316631e-01 -5.02020359e-01 5.88751495e-01 9.09691513e-01
2.27757320e-02 4.19722021e-01 -1.34388238e-01 -3.88390094e-01
-1.70828009e+00 3.60007644e-01 5.53519428e-01 3.05794954e-01
-1.69506624e-01 9.10183549e-01 -1.92340299e-01 -1.66519061e-01
5.76412380e-01 -2.40399823e-01 -7.30544254e-02 -3.32349598e-01
6.15659058e-01 5.76052606e-01 6.90374434e-01 -9.54318464e-01
-6.80604801e-02 2.56739736e-01 1.16912484e-01 -8.02683532e-01
1.29033613e+00 -4.32437629e-01 1.21889114e-01 3.89639616e-01
1.29877841e+00 5.96950114e-01 -9.23461080e-01 -3.62217724e-01
-3.26228261e-01 -4.72913861e-01 5.44349670e-01 -1.24964452e+00
-7.03933358e-01 8.98858130e-01 1.14066637e+00 2.58225858e-01
1.68828833e+00 -4.35291201e-01 5.76621354e-01 1.33059546e-01
1.37109578e-01 -1.16366792e+00 -2.40166426e-01 2.22125202e-01
1.00496161e+00 -9.16403711e-01 -4.12552893e-01 -1.48501948e-01
-4.36253011e-01 9.11192834e-01 2.40908757e-01 2.33527377e-01
6.55582070e-01 5.50734699e-01 4.99485523e-01 3.42732549e-01
-3.23965520e-01 -5.42213976e-01 3.69783878e-01 9.43349004e-01
3.78868729e-01 3.89439724e-02 -3.33645046e-01 7.73219824e-01
-3.09623986e-01 -4.87245530e-01 4.41651911e-01 8.82358491e-01
-7.29587793e-01 -1.06188297e+00 -1.14634299e+00 1.85138747e-01
-6.76175356e-01 -2.82712936e-01 -1.26183957e-01 4.30798084e-01
1.05296157e-01 1.73849595e+00 -1.44667163e-01 -5.19276440e-01
6.42403007e-01 5.66887975e-01 4.05264378e-01 -3.54005188e-01
-7.13419080e-01 9.23492968e-01 3.73062849e-01 5.45387641e-02
-4.83476162e-01 -6.09251320e-01 -8.70705664e-01 2.41661832e-01
-7.20173061e-01 3.56665134e-01 1.23339248e+00 9.31134045e-01
1.69750959e-01 1.00302517e+00 9.52613294e-01 -8.27110827e-01
-6.42215014e-01 -1.38595855e+00 -8.89303923e-01 -1.47006509e-03
7.32078731e-01 -3.15213680e-01 -8.28620374e-01 2.30882004e-01]
|
[14.950444221496582, 5.841394901275635]
|
9b1a61ee-293e-441b-95f2-ec0d4d16e4b5
|
an-end-to-end-model-for-entity-level-relation
|
2102.05980
| null |
https://arxiv.org/abs/2102.05980v2
|
https://arxiv.org/pdf/2102.05980v2.pdf
|
An End-to-end Model for Entity-level Relation Extraction using Multi-instance Learning
|
We present a joint model for entity-level relation extraction from documents. In contrast to other approaches - which focus on local intra-sentence mention pairs and thus require annotations on mention level - our model operates on entity level. To do so, a multi-task approach is followed that builds upon coreference resolution and gathers relevant signals via multi-instance learning with multi-level representations combining global entity and local mention information. We achieve state-of-the-art relation extraction results on the DocRED dataset and report the first entity-level end-to-end relation extraction results for future reference. Finally, our experimental results suggest that a joint approach is on par with task-specific learning, though more efficient due to shared parameters and training steps.
|
['Adrian Ulges', 'Markus Eberts']
|
2021-02-11
| null |
https://aclanthology.org/2021.eacl-main.319
|
https://aclanthology.org/2021.eacl-main.319.pdf
|
eacl-2021-2
|
['document-level-relation-extraction', 'joint-entity-and-relation-extraction', 'nested-named-entity-recognition']
|
['natural-language-processing', 'natural-language-processing', 'natural-language-processing']
|
[ 1.47712797e-01 9.14586723e-01 -4.57882494e-01 -5.78285694e-01
-1.65926373e+00 -4.79946643e-01 7.16010451e-01 6.96467757e-01
-6.75948322e-01 1.00113654e+00 6.54931724e-01 -4.78211716e-02
-4.03232276e-01 -5.75213850e-01 -6.30431950e-01 -2.85697430e-01
-3.01063091e-01 9.95672703e-01 3.10449839e-01 -3.72324228e-01
-1.63935825e-01 1.58730105e-01 -8.87755156e-01 7.88095593e-01
4.07221317e-01 5.58199704e-01 -3.07748049e-01 5.11307240e-01
-1.37836337e-01 8.31212819e-01 -6.32967889e-01 -1.00412846e+00
-2.20363095e-01 -1.12526596e-01 -1.52880132e+00 -4.29394513e-01
4.72457975e-01 2.18830705e-01 -3.54399592e-01 7.00740993e-01
5.50307870e-01 1.46252990e-01 7.06610441e-01 -7.20718980e-01
-4.15380865e-01 1.17097735e+00 -7.27438509e-01 4.38715130e-01
5.02880812e-01 -5.79620659e-01 1.76713884e+00 -6.88646793e-01
9.97994244e-01 1.07520390e+00 8.27545643e-01 3.36787224e-01
-1.53391409e+00 -5.07350504e-01 2.49805853e-01 2.63154238e-01
-1.25584412e+00 -6.04715049e-01 5.95149100e-01 -1.12725385e-01
1.93126118e+00 1.52797893e-01 -1.86118688e-02 1.34846020e+00
1.24025844e-01 9.11620319e-01 8.82356644e-01 -7.19925046e-01
-3.09101909e-01 -7.77702481e-02 5.65832138e-01 4.32660699e-01
4.61263090e-01 -1.32244512e-01 -6.87273324e-01 -1.69168979e-01
2.40163848e-01 -7.37817168e-01 -2.92242974e-01 -2.12688446e-01
-1.07911670e+00 7.08575785e-01 4.01757956e-01 8.77251983e-01
-5.05422831e-01 -1.05636068e-01 8.07499409e-01 2.77224243e-01
6.43200397e-01 6.04897857e-01 -1.05339277e+00 -9.45755187e-03
-1.22404552e+00 3.05291116e-01 1.08260190e+00 8.50215733e-01
5.89259863e-01 -5.98063171e-01 -4.92599994e-01 9.47236121e-01
3.29920590e-01 -2.00494796e-01 2.44880632e-01 -5.92989326e-01
1.00268519e+00 4.98108506e-01 -1.23724811e-01 -7.58608401e-01
-9.36229289e-01 -5.18407822e-01 -5.30065536e-01 -3.34300011e-01
3.27083319e-01 -4.20403123e-01 -4.14979309e-01 1.75278127e+00
2.40187213e-01 2.63211399e-01 4.86830264e-01 6.59886956e-01
1.43418288e+00 2.03227103e-01 3.83603811e-01 -4.97583598e-01
1.84256101e+00 -9.74624038e-01 -1.11688137e+00 -3.39128703e-01
8.77038479e-01 -4.77992445e-01 1.35959446e-01 1.51353717e-01
-1.09855747e+00 -3.49423170e-01 -1.06105542e+00 -3.23508352e-01
-4.43778604e-01 1.94107339e-01 9.04145479e-01 4.01226610e-01
-7.41579294e-01 5.97970903e-01 -7.90343106e-01 -3.76278490e-01
1.52057320e-01 5.56988955e-01 -8.37168634e-01 4.41093057e-01
-1.59764576e+00 1.48504162e+00 8.60806108e-01 -2.96905227e-02
-1.39235750e-01 -6.20746315e-01 -1.20642996e+00 8.30674320e-02
5.24742842e-01 -6.45633876e-01 1.27665937e+00 -1.70006484e-01
-1.30704820e+00 1.20211005e+00 -2.80003637e-01 -7.84395993e-01
-2.84918491e-02 -5.46045601e-01 -5.70932984e-01 1.98620468e-01
9.22080353e-02 3.93122792e-01 6.62528872e-02 -1.19389081e+00
-5.58797061e-01 -3.02861542e-01 4.33365218e-02 1.10427529e-01
4.36738618e-02 5.78416109e-01 -5.18745244e-01 -3.96308839e-01
8.80033970e-02 -5.78572512e-01 -1.50321603e-01 -1.00166440e+00
-6.74665928e-01 -7.21247494e-01 4.85445589e-01 -7.92352855e-01
1.25098300e+00 -1.75354064e+00 4.10023242e-01 -2.67046094e-02
1.76526845e-01 3.25595200e-01 -1.76808149e-01 6.05558813e-01
-4.91881758e-01 1.44980662e-02 -7.04603568e-02 -7.77288318e-01
-1.53286085e-01 -8.14289898e-02 -8.78542438e-02 3.27697247e-01
6.07316375e-01 1.08568335e+00 -9.81498599e-01 -7.43848205e-01
-1.66301206e-01 5.45872867e-01 -3.43579441e-01 4.37651798e-02
-9.06670690e-02 3.14104378e-01 -5.74138463e-01 3.31098437e-01
3.97023618e-01 -1.51878476e-01 7.91220009e-01 -6.80824459e-01
9.15597081e-02 1.14467216e+00 -1.14242125e+00 1.91130626e+00
-5.38385093e-01 4.03128564e-01 1.82007238e-01 -1.28805959e+00
8.86418402e-01 7.80481219e-01 4.92704332e-01 -3.54474902e-01
1.60803989e-01 1.73803031e-01 3.54545750e-02 -4.09009844e-01
7.65355051e-01 -1.68101758e-01 -5.85396469e-01 2.30226874e-01
8.67555320e-01 6.59948587e-02 2.47802839e-01 5.26229084e-01
1.35716760e+00 6.50286615e-01 8.19013655e-01 -1.86550528e-01
5.40692985e-01 -1.72263771e-01 7.15958774e-01 4.29865628e-01
1.03394151e-01 4.18080360e-01 6.69823170e-01 -4.24599722e-02
-4.68727618e-01 -7.39332259e-01 -3.85522246e-01 9.12211359e-01
-2.38607913e-01 -8.01175117e-01 -3.93050909e-01 -1.06878948e+00
-1.38932839e-01 8.20993841e-01 -6.85905933e-01 2.63963699e-01
-9.99945402e-01 -9.51815426e-01 8.71629238e-01 5.90477288e-01
1.52836964e-01 -1.15221524e+00 -2.64294028e-01 6.31838799e-01
-4.42601562e-01 -1.67811680e+00 8.38292092e-02 7.73197174e-01
-6.23781621e-01 -1.12808251e+00 -3.69707406e-01 -7.27625787e-01
5.00662141e-02 -3.72220725e-01 1.49739230e+00 -2.93965966e-01
-6.89672828e-02 1.27021238e-01 -4.11963671e-01 -1.71957910e-01
-1.72375947e-01 6.26695275e-01 -1.51195228e-01 -3.62375915e-01
7.35655904e-01 -5.92963696e-01 -5.07721640e-02 -4.72211123e-01
-3.02090436e-01 -2.57392764e-01 7.85372853e-01 9.21661854e-01
5.32706022e-01 -2.29184583e-01 7.77034700e-01 -1.42410696e+00
5.97354650e-01 -4.89183336e-01 -1.37543365e-01 4.13646966e-01
-4.76902366e-01 2.47612193e-01 1.44247487e-01 6.14278717e-03
-1.35524786e+00 -4.26666550e-02 -2.07733989e-01 1.34875104e-01
-3.65158826e-01 7.83370912e-01 -3.12048852e-01 4.81773674e-01
6.75943851e-01 -3.64089131e-01 -4.41628695e-01 -6.20992124e-01
6.90045297e-01 6.34752095e-01 7.94656157e-01 -7.83187926e-01
5.20876944e-01 -1.97776146e-02 1.21058067e-02 -5.33941925e-01
-1.40024590e+00 -6.77453697e-01 -1.17473948e+00 3.88455272e-01
1.06503737e+00 -1.14656353e+00 -6.65174186e-01 1.31126363e-02
-1.49652016e+00 -1.07019171e-01 -1.88935086e-01 6.58495486e-01
-4.15485382e-01 3.50520521e-01 -1.14050508e+00 -7.25288570e-01
-5.23671508e-01 -6.10025585e-01 1.28888583e+00 1.07238017e-01
-5.75390160e-01 -1.14914060e+00 4.39478815e-01 6.15534961e-01
-1.49342582e-01 2.20205665e-01 6.91999733e-01 -1.29075694e+00
-3.52543265e-01 -2.75834322e-01 -4.23836619e-01 -2.09846780e-01
2.43623294e-02 -2.32834369e-01 -1.00338531e+00 -7.21100792e-02
-5.31312227e-01 -5.49491048e-01 1.16904068e+00 2.31170252e-01
3.41256142e-01 -9.60016847e-02 -9.50768709e-01 2.13138804e-01
1.27955711e+00 -2.66188830e-01 4.92458552e-01 5.66076159e-01
6.09771788e-01 7.53531516e-01 8.47732961e-01 8.22338760e-02
7.89148748e-01 1.10216570e+00 3.97118255e-02 -9.36884657e-02
-2.76205063e-01 3.82660665e-02 1.01152532e-01 6.78577363e-01
-1.97573736e-01 -2.41799861e-01 -7.76448488e-01 9.24097359e-01
-2.06784511e+00 -1.19238532e+00 -4.71504509e-01 1.76189542e+00
1.45902157e+00 3.97847503e-01 2.16172650e-01 1.04429014e-01
5.24506271e-01 2.41468981e-01 1.58847928e-01 -5.23494422e-01
-2.59637266e-01 7.03828275e-01 4.69792873e-01 7.27341831e-01
-1.55811656e+00 1.38150716e+00 6.31445217e+00 6.00541174e-01
-6.34461880e-01 4.06927168e-01 9.47888419e-02 8.29034522e-02
-1.79229110e-01 1.44098654e-01 -1.34367967e+00 -1.42538920e-01
1.21247137e+00 1.50341317e-01 -1.11261219e-01 3.66948634e-01
-3.46157908e-01 -6.84996620e-02 -1.19231379e+00 5.39798439e-01
-1.02555081e-02 -1.36489844e+00 -4.89771605e-01 -6.54811040e-02
3.52445513e-01 3.73237699e-01 -6.71700418e-01 6.06849194e-01
6.28636122e-01 -7.97008693e-01 2.97506720e-01 3.18056166e-01
5.79242826e-01 -6.94871008e-01 1.19393063e+00 2.33179495e-01
-1.37929475e+00 3.05511057e-01 -8.48639533e-02 1.00370333e-01
5.92907727e-01 8.07218075e-01 -7.84508228e-01 1.43279564e+00
4.48843956e-01 6.38734519e-01 -5.05318999e-01 7.22017288e-01
-6.50937319e-01 5.18222272e-01 -2.00199723e-01 1.95391789e-01
-4.29123566e-02 1.15200996e-01 5.65030336e-01 2.01099467e+00
-2.29509711e-01 2.67082244e-01 1.55103877e-01 6.11788392e-01
-2.51144975e-01 2.76191622e-01 -4.28020418e-01 3.00173402e-01
5.16399384e-01 1.70558608e+00 -3.99161160e-01 -2.33948559e-01
-6.79165959e-01 8.54045689e-01 1.06952870e+00 6.02905042e-02
-5.58622360e-01 -5.19688845e-01 3.72228533e-01 -4.97088999e-01
5.27579486e-01 -1.13074049e-01 -1.86931908e-01 -1.14981544e+00
-1.72777101e-01 -5.38735867e-01 8.13224554e-01 -8.63748342e-02
-1.28926325e+00 7.49996781e-01 2.30434865e-01 -6.42289102e-01
-6.66417181e-01 -4.05678481e-01 -5.09899378e-01 7.97876537e-01
-1.62313867e+00 -1.70126987e+00 4.14995074e-01 2.51177490e-01
1.74722329e-01 -4.14440930e-02 1.40314138e+00 3.40322673e-01
-6.47601008e-01 9.47854698e-01 -5.36401212e-01 7.19069004e-01
8.75453651e-01 -1.47684634e+00 4.66490299e-01 7.00316608e-01
7.23510265e-01 8.01011145e-01 6.58552647e-01 -6.46324635e-01
-7.96297550e-01 -8.35862637e-01 1.84858119e+00 -8.06832373e-01
7.77233899e-01 -3.25695425e-01 -1.01612830e+00 1.12345564e+00
6.68255746e-01 2.10949630e-02 1.03430462e+00 1.30913532e+00
-5.78092396e-01 1.91383630e-01 -1.11802113e+00 1.61880136e-01
1.05268061e+00 -6.56455815e-01 -1.31902993e+00 2.08667785e-01
6.69755459e-01 -5.96071541e-01 -1.42452300e+00 8.17152560e-01
1.54089093e-01 -6.72073185e-01 8.94809723e-01 -7.87827671e-01
1.82491347e-01 -7.09571317e-02 -1.31076589e-01 -1.17684186e+00
-5.76504469e-01 -5.28283358e-01 -5.28107643e-01 1.85551536e+00
1.04890323e+00 -3.31073344e-01 4.15422648e-01 3.79317284e-01
-1.72836497e-01 -7.06719458e-01 -1.13261354e+00 -6.10478580e-01
2.13183984e-01 -3.67803425e-01 1.57115355e-01 1.10641599e+00
7.10991383e-01 1.22397733e+00 -1.71865165e-01 2.96784610e-01
5.50172329e-01 3.45255047e-01 4.41630512e-01 -1.44093835e+00
-6.23405337e-01 -2.66328841e-01 -1.43974587e-01 -6.77849770e-01
9.97329295e-01 -1.08589780e+00 2.38661412e-02 -1.62177908e+00
3.46627474e-01 -2.22393408e-01 -5.10924399e-01 8.23606670e-01
-4.67871308e-01 1.52520254e-01 2.46882383e-02 -7.55245751e-03
-9.10517931e-01 2.82623202e-01 6.79662764e-01 -3.83450091e-02
-2.62138605e-01 -1.56688467e-01 -9.11538899e-01 4.36110646e-01
4.37274545e-01 -5.54153562e-01 5.99620603e-02 -2.02509835e-01
1.63890257e-01 2.32425645e-01 -5.37429936e-02 -5.52853048e-01
2.20589340e-01 3.09929252e-01 1.69427738e-01 -7.01210737e-01
5.50851166e-01 -4.08435911e-01 -2.79782087e-01 -3.41863744e-02
-4.88897681e-01 -4.72263277e-01 3.49550039e-01 3.57922733e-01
-4.58454728e-01 -2.96790272e-01 4.39347029e-01 1.09530957e-02
-4.47633296e-01 -5.93838505e-02 -5.96330985e-02 1.66317374e-01
7.40295231e-01 3.77049536e-01 -5.27373135e-01 -5.79656847e-02
-1.09182560e+00 1.96110189e-01 -2.22210690e-01 5.15951097e-01
8.63988549e-02 -1.09593236e+00 -1.17715371e+00 -3.05612773e-01
1.83865502e-01 -7.96236023e-02 -2.71408074e-03 9.54969704e-01
3.66411567e-01 8.79995465e-01 2.55060941e-01 -1.35476977e-01
-1.50693989e+00 5.74547589e-01 2.41357446e-01 -1.15921545e+00
-6.77774906e-01 1.09271145e+00 -1.73190236e-01 -5.86969435e-01
6.12381510e-02 -5.87914251e-02 -7.08616555e-01 6.35287344e-01
4.37430829e-01 -7.82558247e-02 4.83209997e-01 -8.55283618e-01
-9.03178751e-01 2.85346985e-01 -4.55473989e-01 -2.35538080e-01
1.56831539e+00 4.49086875e-02 -7.51887858e-02 4.56386119e-01
1.00356472e+00 3.55080664e-01 -6.19333446e-01 -4.66152906e-01
7.82854438e-01 1.51953593e-01 1.18858337e-01 -9.88324642e-01
-7.88546324e-01 5.93081713e-01 -7.80256614e-02 1.77941099e-02
6.27621293e-01 5.81927001e-01 6.21901810e-01 5.36470115e-01
5.09211719e-01 -9.99240398e-01 -4.15715843e-01 6.11422360e-01
8.71330678e-01 -1.15555096e+00 3.65473241e-01 -8.85491729e-01
-5.42130351e-01 8.88936698e-01 6.89596474e-01 -1.59576729e-01
5.92209756e-01 7.25472629e-01 -6.87033758e-02 -4.10748899e-01
-9.79683816e-01 -7.53732502e-01 4.98672694e-01 6.28549993e-01
1.27746737e+00 5.08577377e-02 -6.92188680e-01 1.25678730e+00
-7.41803199e-02 -2.49580637e-01 1.02885507e-01 5.74741483e-01
-5.99162132e-02 -1.77859974e+00 8.74759033e-02 2.67659724e-01
-1.02686417e+00 -3.73631090e-01 -5.20670831e-01 8.51135552e-01
-4.15471904e-02 1.18981326e+00 -1.33583263e-01 -3.00971031e-01
5.48910916e-01 3.89776826e-01 8.50030184e-01 -9.71511722e-01
-9.96344328e-01 1.05438836e-01 1.13880539e+00 -5.65168560e-01
-9.14079189e-01 -1.00767076e+00 -1.55026472e+00 2.51281470e-01
-7.08971918e-01 3.92772645e-01 1.55148134e-01 1.35340571e+00
4.22188908e-01 8.81884336e-01 1.32602170e-01 -8.06636989e-01
-2.54918665e-01 -1.42493761e+00 -4.12995338e-01 2.68219054e-01
1.73080429e-01 -7.29927778e-01 2.13208012e-02 -2.80100435e-01]
|
[9.456439018249512, 8.966801643371582]
|
074d3b2e-597b-4cbe-818f-68c2df5c74f3
|
learning-from-the-dictionary-heterogeneous
|
2210.10320
| null |
https://arxiv.org/abs/2210.10320v1
|
https://arxiv.org/pdf/2210.10320v1.pdf
|
Learning from the Dictionary: Heterogeneous Knowledge Guided Fine-tuning for Chinese Spell Checking
|
Chinese Spell Checking (CSC) aims to detect and correct Chinese spelling errors. Recent researches start from the pretrained knowledge of language models and take multimodal information into CSC models to improve the performance. However, they overlook the rich knowledge in the dictionary, the reference book where one can learn how one character should be pronounced, written, and used. In this paper, we propose the LEAD framework, which renders the CSC model to learn heterogeneous knowledge from the dictionary in terms of phonetics, vision, and meaning. LEAD first constructs positive and negative samples according to the knowledge of character phonetics, glyphs, and definitions in the dictionary. Then a unified contrastive learning-based training scheme is employed to refine the representations of the CSC models. Extensive experiments and detailed analyses on the SIGHAN benchmark datasets demonstrate the effectiveness of our proposed methods.
|
['Haitao Zheng', 'Yunbo Cao', 'Chao Li', 'Ruiyang Liu', 'Shulin Huang', 'Li Yangning', 'Zhongli Li', 'Qingyu Zhou', 'Shirong Ma', 'Yinghui Li']
|
2022-10-19
| null | null | null | null |
['chinese-spell-checking']
|
['natural-language-processing']
|
[ 2.29956493e-01 -7.08394408e-01 -1.88218281e-01 -2.64648348e-01
-6.42889380e-01 -6.71839297e-01 4.34160888e-01 7.48773888e-02
-6.34869337e-01 4.48389143e-01 4.01110083e-01 -4.17076081e-01
3.13087195e-01 -5.32263041e-01 -4.96497720e-01 -6.02661490e-01
7.15418577e-01 1.74891368e-01 6.63639158e-02 -2.47631267e-01
6.37410164e-01 9.66684222e-02 -1.14931917e+00 5.87587297e-01
1.43946636e+00 7.89293766e-01 7.15976119e-01 5.99885345e-01
-5.56020021e-01 9.22680557e-01 -8.03955615e-01 -5.03857672e-01
-2.69929111e-01 -5.86580694e-01 -5.76685905e-01 5.06098270e-02
1.43556967e-01 -1.43291831e-01 -2.99303472e-01 1.49814987e+00
4.62959677e-01 1.66739728e-02 4.69889551e-01 -6.63333416e-01
-1.28511941e+00 7.87729681e-01 -1.19818427e-01 -4.30091731e-02
4.29054916e-01 -1.32053778e-01 8.05036426e-01 -1.08834898e+00
3.58315587e-01 1.04912162e+00 4.36423779e-01 7.09185004e-01
-3.11420381e-01 -6.91834211e-01 4.30542082e-01 5.80082178e-01
-1.46468008e+00 -2.19733790e-01 6.34225726e-01 -1.52758673e-01
6.69425547e-01 3.04892004e-01 6.40333831e-01 8.91005039e-01
-2.44034559e-01 1.46270311e+00 1.22435379e+00 -8.34836841e-01
-7.84153566e-02 2.45234936e-01 4.79312360e-01 5.63065290e-01
1.94283023e-01 -5.58733083e-02 -5.16679108e-01 3.81984234e-01
6.07372284e-01 6.95929304e-02 -5.52101791e-01 1.66535243e-01
-1.17716110e+00 3.02732795e-01 7.75648141e-03 4.88763630e-01
1.13721594e-01 -2.61441678e-01 3.56597066e-01 7.39877447e-02
-2.06149459e-01 3.68032038e-01 -5.43578207e-01 -2.91990757e-01
-7.10407734e-01 -2.32851550e-01 5.47824681e-01 1.36257792e+00
5.88239551e-01 6.35389462e-02 -2.19412878e-01 8.81829381e-01
4.34820771e-01 8.84895623e-01 7.91608512e-01 -4.78436381e-01
6.32717848e-01 7.11471260e-01 1.57279477e-01 -1.03464878e+00
7.85586145e-03 -4.90597069e-01 -5.84650576e-01 -5.91599464e-01
5.02142049e-02 -1.86857283e-01 -8.22694361e-01 1.11159992e+00
-1.35206535e-01 4.69376057e-01 1.49274841e-01 8.69911432e-01
1.09220707e+00 6.39857650e-01 4.43052966e-03 -7.79231265e-02
9.68824983e-01 -1.19957137e+00 -1.03004766e+00 -2.58248001e-01
8.68894279e-01 -9.01048899e-01 1.42280793e+00 6.83810532e-01
-7.74861038e-01 -8.39690745e-01 -1.05671406e+00 -2.32602865e-01
-4.96416390e-01 8.88383269e-01 3.02312583e-01 8.79382193e-01
-6.59861684e-01 2.18272433e-01 -6.15263939e-01 1.14575125e-01
2.89853275e-01 1.00159928e-01 8.74360800e-02 -2.54189342e-01
-1.22904944e+00 7.51427710e-01 5.59964061e-01 4.40449893e-01
-5.98250687e-01 -2.98336565e-01 -9.13544059e-01 1.14043862e-01
1.88182592e-01 -1.57135919e-01 1.17826605e+00 -1.24200618e+00
-1.47510052e+00 6.51189566e-01 -3.37681174e-01 -5.89737110e-02
1.93299428e-01 -4.26659614e-01 -9.68731403e-01 7.02670068e-02
-1.45090863e-01 4.01929289e-01 5.66435218e-01 -1.28508210e+00
-1.08491814e+00 -1.81465283e-01 -3.31827663e-02 4.37751561e-01
-5.66546381e-01 2.16103435e-01 -1.13888025e+00 -7.76071548e-01
9.26956311e-02 -6.01782143e-01 -6.93014935e-02 -4.21291679e-01
-4.21004891e-01 1.51533969e-02 5.94580352e-01 -9.76215959e-01
1.85155332e+00 -2.26954150e+00 4.97480966e-02 3.77225488e-01
-2.13469669e-01 6.60031617e-01 -2.62615681e-01 2.21032530e-01
3.51496041e-01 1.43367767e-01 -9.34459940e-02 -2.74556458e-01
1.36227496e-02 3.30456197e-01 -2.69170105e-01 1.61325842e-01
-2.03699648e-01 8.13952148e-01 -1.03191435e+00 -5.58460832e-01
9.85631645e-02 3.68028462e-01 -4.35128570e-01 1.86263725e-01
-4.10931371e-03 3.38696867e-01 -4.39368755e-01 7.60914385e-01
8.39947820e-01 -1.94450915e-01 3.72018903e-01 -9.64841247e-02
-1.95148557e-01 4.66761380e-01 -1.36399651e+00 1.59467006e+00
-4.52219486e-01 4.54668701e-01 -2.61843979e-01 -7.51617789e-01
1.04357779e+00 8.36678222e-02 -2.34230652e-01 -1.11299825e+00
1.74944460e-01 3.66552413e-01 -7.47144446e-02 -7.01978505e-01
5.90023696e-01 2.73975372e-01 -5.38704544e-03 6.37677014e-02
-3.06635648e-01 4.37842309e-02 2.22013548e-01 1.86192375e-02
3.61007333e-01 2.25062370e-01 2.26097807e-01 9.14818607e-03
1.10708952e+00 1.26819968e-01 7.26174891e-01 6.22202575e-01
-1.27334356e-01 5.73844612e-01 1.19156480e-01 -1.95896566e-01
-3.93464535e-01 -7.59977102e-01 1.32291779e-01 1.07524228e+00
5.93076825e-01 -5.90966463e-01 -8.70582163e-01 -8.43288600e-01
-4.63110864e-01 8.04532766e-01 -2.58099616e-01 -2.85669059e-01
-7.30561733e-01 -4.06778514e-01 7.14227140e-01 7.94428051e-01
7.19649374e-01 -1.14861774e+00 -1.42940372e-01 -7.29370117e-02
-3.40170622e-01 -1.04658437e+00 -7.40942955e-01 -2.46489510e-01
-5.92951894e-01 -1.04665518e+00 -6.01450741e-01 -1.46686089e+00
8.28773439e-01 3.19371551e-01 7.19813645e-01 5.63427627e-01
2.28797451e-01 3.06878269e-01 -7.42920995e-01 -5.80250800e-01
-3.55287999e-01 -5.41746765e-02 -2.23584697e-01 1.21635154e-01
7.53966033e-01 1.56013891e-01 -3.48301291e-01 1.87952697e-01
-7.22607851e-01 1.57069415e-01 6.37813509e-01 6.91493213e-01
7.98481166e-01 -6.89150672e-03 2.22125456e-01 -9.15997744e-01
6.99684083e-01 -1.16930351e-01 -4.96987224e-01 7.46732354e-01
-5.69679022e-01 -1.55274034e-01 6.87205136e-01 -3.43002379e-01
-1.27407217e+00 7.49697462e-02 -2.21464962e-01 -2.00475425e-01
-2.54948735e-01 7.26302862e-01 -6.22709036e-01 -7.39307106e-02
1.92553639e-01 1.00265181e+00 -3.71901393e-01 -7.14457452e-01
2.08989009e-01 1.01870954e+00 7.27649808e-01 -6.36326909e-01
4.36086655e-01 2.37845644e-01 -7.28655338e-01 -7.93622255e-01
-7.89391339e-01 -5.39137900e-01 -7.65352130e-01 -1.32862106e-01
6.71555638e-01 -1.03297675e+00 -5.89575768e-01 8.36688280e-01
-1.06322348e+00 -9.69337821e-02 2.34288350e-01 5.86586118e-01
-1.26916081e-01 8.59868765e-01 -7.09302843e-01 -6.54805064e-01
-2.15510711e-01 -1.05356419e+00 9.13138866e-01 6.50773227e-01
1.51150987e-01 -1.04237461e+00 -1.31319359e-01 2.39298120e-01
8.88182223e-02 -5.85939944e-01 1.07335865e+00 -7.11039841e-01
-5.56951165e-01 -1.47266671e-01 -1.68076023e-01 7.48513997e-01
1.82483077e-01 6.18335977e-02 -7.03654528e-01 -3.59730911e-04
-1.32758275e-01 -8.29775445e-03 7.91292429e-01 3.43170762e-02
1.47476411e+00 -2.04023063e-01 -2.02786699e-01 7.97679722e-01
1.40678930e+00 6.45191252e-01 8.62479329e-01 3.11787456e-01
9.87724602e-01 2.30207279e-01 7.79350340e-01 4.23859864e-01
6.15915239e-01 3.13745409e-01 1.18518285e-01 2.12906033e-01
-1.17704764e-01 -5.35977602e-01 5.57907283e-01 1.50344932e+00
2.11333688e-02 -1.83012217e-01 -1.05824804e+00 5.46143115e-01
-1.59919047e+00 -7.33932555e-01 -2.70119697e-01 1.96220720e+00
1.00476325e+00 9.05637741e-02 -4.26921964e-01 1.88546523e-01
7.35690236e-01 -3.56633402e-02 -1.84500784e-01 -2.22917244e-01
-5.52545071e-01 7.84641355e-02 2.30429947e-01 5.22529781e-01
-1.09176421e+00 1.37220037e+00 6.11983204e+00 1.08283651e+00
-1.34748614e+00 -1.33329913e-01 2.23974496e-01 4.02275085e-01
-6.30668938e-01 -3.73085117e-04 -1.20188284e+00 7.47606754e-01
4.48854208e-01 8.56020823e-02 5.44671178e-01 6.57488227e-01
1.31529525e-01 3.71818095e-02 -7.52749324e-01 1.23727095e+00
4.83348459e-01 -1.44638216e+00 5.27029157e-01 -4.23838884e-01
9.67817605e-01 -2.19495028e-01 1.48083493e-01 4.17686999e-01
1.24755025e-01 -9.34509993e-01 9.73836124e-01 7.94318557e-01
5.71960926e-01 -7.22478092e-01 6.76615715e-01 5.99840105e-01
-1.14230824e+00 -8.84772092e-02 -4.32351530e-01 -8.09535757e-02
-1.68809101e-01 -8.05479214e-02 -4.48530734e-01 5.91836572e-01
4.81282234e-01 1.15363955e+00 -9.59189296e-01 1.20625162e+00
-7.25165546e-01 8.42861176e-01 2.66468346e-01 -5.07509112e-01
2.67356902e-01 -3.99547964e-01 3.76512595e-02 1.55013859e+00
3.50581408e-01 6.34471327e-02 2.81364471e-01 4.67942953e-01
8.23673531e-02 4.14728850e-01 -5.63594513e-02 -1.92569599e-01
7.14425087e-01 8.20658982e-01 -3.82378757e-01 -4.89885241e-01
-7.72093654e-01 1.03822172e+00 2.50740081e-01 6.44561827e-01
-6.57902896e-01 -5.10151267e-01 3.05435210e-01 -5.15350282e-01
5.56464612e-01 -3.14530373e-01 -6.28986239e-01 -1.41468215e+00
-5.25612906e-02 -1.34280860e+00 3.17986071e-01 -7.71835089e-01
-1.06648362e+00 3.16131234e-01 -4.56818700e-01 -1.47843850e+00
3.26953471e-01 -9.34254527e-01 -6.67832494e-01 8.95776391e-01
-1.79684484e+00 -1.09309590e+00 -2.66387999e-01 7.17819870e-01
7.86852777e-01 -3.42573255e-01 6.58946991e-01 4.86842066e-01
-8.54882061e-01 7.73890674e-01 3.26693267e-01 6.74525857e-01
6.00926518e-01 -1.11892903e+00 3.89079675e-02 1.07518554e+00
3.78354430e-01 1.03926039e+00 2.58723736e-01 -7.77264297e-01
-1.32053387e+00 -9.94178176e-01 1.20455456e+00 -2.87979692e-01
3.36799830e-01 -8.95632729e-02 -1.03033400e+00 4.99027848e-01
1.75327867e-01 -2.40337044e-01 8.28509569e-01 -7.65627846e-02
-2.64648318e-01 -7.41132349e-02 -4.74784940e-01 6.96075857e-01
8.21619749e-01 -7.83043563e-01 -7.71492660e-01 6.13500439e-02
4.77965087e-01 -5.28641343e-01 -2.40327269e-01 1.61308751e-01
4.96627808e-01 -6.51603997e-01 5.94897866e-01 -7.61578858e-01
2.67759591e-01 -5.78639567e-01 -3.62610489e-01 -1.24662673e+00
-2.12563679e-01 -2.85410762e-01 8.94880202e-03 1.27599001e+00
4.92674142e-01 -1.70775905e-01 4.25553083e-01 1.26110062e-01
-5.67678690e-01 -5.40098429e-01 -4.28021818e-01 -5.95435619e-01
4.28227149e-02 -5.35214782e-01 9.27821279e-01 1.05368423e+00
9.73666981e-02 1.63750932e-01 -3.20434541e-01 4.60047424e-01
-1.32850632e-02 1.62399635e-01 4.75130826e-01 -9.08904135e-01
-1.33642048e-01 -5.77437818e-01 1.02324538e-01 -1.64101923e+00
1.98872566e-01 -8.28686655e-01 1.13732159e-01 -1.48792279e+00
2.09624648e-01 -3.51612270e-01 -5.77084780e-01 2.09099039e-01
-7.00323582e-01 -1.38539031e-01 3.28277320e-01 9.75392312e-02
-9.84109700e-01 4.80537027e-01 1.47477889e+00 -3.08269262e-01
-1.29406765e-01 -6.23662136e-02 -7.69757569e-01 9.10376668e-01
7.76334107e-01 -1.43886462e-01 -3.09341609e-01 -1.05141115e+00
4.00041252e-01 -2.80362785e-01 1.20107837e-01 -9.35406744e-01
5.01460850e-01 -2.43947506e-01 4.69375163e-01 -7.76509523e-01
-1.87386479e-02 -7.78835833e-01 -5.65512538e-01 5.21187663e-01
-5.43638349e-01 3.32959622e-01 7.42560029e-02 4.30689901e-01
-3.87799799e-01 -5.45839965e-01 4.73986715e-01 -2.55254239e-01
-1.35531259e+00 3.22962776e-02 -3.41066450e-01 2.32632086e-01
5.46969712e-01 -3.07835102e-01 -1.82743385e-01 -2.87170321e-01
-5.00601232e-01 3.86323124e-01 3.74137998e-01 5.66478133e-01
9.21213925e-01 -1.38863134e+00 -4.98567879e-01 4.57249105e-01
1.86863631e-01 -3.24699521e-01 3.35164338e-01 6.51216865e-01
-7.92349041e-01 3.85915428e-01 -1.91840515e-01 -3.20022464e-01
-1.20781600e+00 5.62745035e-01 4.54212517e-01 4.51248102e-02
-1.33980051e-01 7.62830019e-01 1.09016551e-02 -5.02147555e-01
6.35347366e-01 -4.42140251e-01 -8.40281069e-01 -1.31582662e-01
8.90913427e-01 3.98699909e-01 -3.22874375e-02 -7.68654346e-01
-3.17968100e-01 7.36355543e-01 -1.92856714e-01 1.86217595e-02
7.23928332e-01 -4.35671598e-01 -1.37019426e-01 3.36067677e-01
9.50552881e-01 3.47591072e-01 -8.31172884e-01 -4.22467262e-01
2.09087834e-01 -5.01328170e-01 -1.88813552e-01 -1.01268256e+00
-9.06693637e-01 1.18603468e+00 4.75945890e-01 -4.99774516e-01
1.14675987e+00 -4.13780421e-01 9.30641472e-01 6.93556607e-01
1.99860483e-01 -1.58491504e+00 -3.95530313e-02 9.06543612e-01
5.20198166e-01 -1.16051960e+00 -4.47823495e-01 -4.28745389e-01
-1.08810079e+00 1.10994792e+00 8.61386180e-01 -1.08040739e-02
6.11603558e-01 1.23551652e-01 5.08793652e-01 2.50720352e-01
-3.40081245e-01 -2.60570139e-01 4.85288352e-01 7.92288125e-01
4.73006457e-01 1.54485062e-01 -5.62851369e-01 1.23888791e+00
-1.15284495e-01 -3.04601919e-02 4.61197078e-01 8.94500375e-01
-6.15787685e-01 -1.21565700e+00 -2.05393657e-01 9.26302448e-02
-1.84498489e-01 -5.36359072e-01 -5.44912279e-01 5.02989471e-01
5.61042249e-01 1.05423462e+00 -8.16287175e-02 -4.56686646e-01
2.87320226e-01 2.69365665e-02 2.57794917e-01 -7.59726465e-01
-5.69008648e-01 2.04554081e-01 -8.07162225e-02 -3.18331718e-01
-3.96925747e-01 -5.74927628e-01 -1.62415218e+00 1.36921480e-01
-2.86298662e-01 3.03881794e-01 3.91319841e-01 1.20675445e+00
2.41901711e-01 2.75359720e-01 5.08808374e-01 1.34612331e-02
-5.35364032e-01 -6.78163588e-01 -4.04337704e-01 3.72131824e-01
2.82861739e-01 -1.25844195e-01 1.09814145e-01 2.13320524e-01]
|
[10.940875053405762, 10.843379974365234]
|
a9e58268-0bac-4c2a-8815-0a7160e1f1c0
|
deep-transformer-based-data-augmentation-with
|
2007.06949
| null |
https://arxiv.org/abs/2007.06949v3
|
https://arxiv.org/pdf/2007.06949v3.pdf
|
Deep Transformer based Data Augmentation with Subword Units for Morphologically Rich Online ASR
|
Recently Deep Transformer models have proven to be particularly powerful in language modeling tasks for ASR. Their high complexity, however, makes them very difficult to apply in the first (single) pass of an online system. Recent studies showed that a considerable part of the knowledge of neural network Language Models (LM) can be transferred to traditional n-grams by using neural text generation based data augmentation. In our paper, we pre-train a GPT-2 Transformer LM on a general text corpus and fine-tune it on our Hungarian conversational call center ASR task. We show that although data augmentation with Transformer-generated text works well for isolating languages, it causes a vocabulary explosion in a morphologically rich language. Therefore, we propose a new method called subword-based neural text augmentation, where we retokenize the generated text into statistically derived subwords. We compare Morfessor and BPE statistical subword tokenizers and show that both methods can significantly improve the WER while greatly reducing vocabulary size and memory requirements. Finally, we also demonstrate that subword-based neural text augmentation outperforms the word-based approach not only in terms of overall WER but also in recognition of OOV words.
|
['Péter Mihajlik', 'Tibor Fegyó', 'György Szaszák', 'Balázs Tarján']
|
2020-07-14
| null | null | null | null |
['text-augmentation']
|
['natural-language-processing']
|
[ 4.33517367e-01 2.75812566e-01 1.83176905e-01 -3.76759142e-01
-1.12963474e+00 -5.24006009e-01 4.56822008e-01 8.57039541e-02
-7.06128001e-01 7.29615092e-01 4.19785202e-01 -8.33256721e-01
2.37092242e-01 -7.66310275e-01 -7.08138168e-01 -4.77392972e-01
2.87752122e-01 9.17488277e-01 -3.48687768e-02 -6.88962817e-01
-1.90654367e-01 3.94616783e-01 -1.09834862e+00 4.37121093e-01
8.61148298e-01 4.72959340e-01 4.21957403e-01 7.83474743e-01
-6.08103216e-01 6.18520319e-01 -8.86997461e-01 -4.44788009e-01
1.78851709e-01 -5.26127696e-01 -8.23509097e-01 -1.00995280e-01
2.88763344e-01 -9.93778780e-02 -3.77037108e-01 7.51344681e-01
6.33752823e-01 3.48511815e-01 4.64666307e-01 -4.36169207e-01
-5.87905169e-01 1.40202558e+00 -7.30894282e-02 3.19926471e-01
-8.20728019e-02 -2.50978112e-01 1.18566370e+00 -9.98559296e-01
2.33078629e-01 1.43277001e+00 5.26440561e-01 8.07224751e-01
-1.36298299e+00 -7.47217417e-01 1.92753330e-01 -1.20557420e-01
-1.34374774e+00 -7.21360743e-01 7.43404686e-01 -5.96339814e-02
1.62820065e+00 4.37394470e-01 3.65110606e-01 9.84700739e-01
-2.35517234e-01 7.63198256e-01 7.85072267e-01 -1.03881979e+00
-8.81443098e-02 2.22100869e-01 6.14930540e-02 5.62704265e-01
-2.03027520e-02 -2.01810881e-01 -4.80827153e-01 3.55552733e-02
6.04425907e-01 -4.69729424e-01 -5.93657903e-02 3.56167793e-01
-1.16430461e+00 1.11771822e+00 2.48362217e-02 7.37859726e-01
-3.05762738e-01 1.88528001e-01 5.41095555e-01 4.58674848e-01
7.61159778e-01 6.55316591e-01 -6.50104046e-01 -1.06012061e-01
-9.38937545e-01 -4.30746451e-02 6.70964718e-01 7.48930156e-01
4.23634857e-01 6.89355016e-01 -3.27670008e-01 1.46318769e+00
-3.55587937e-02 7.37809598e-01 1.04478216e+00 -3.56313080e-01
6.53070092e-01 3.56281817e-01 -3.66963714e-01 -4.39257979e-01
-2.69209683e-01 -4.95939970e-01 -6.79643869e-01 -3.43973577e-01
3.28730434e-01 -2.75455117e-01 -1.20320737e+00 1.91182160e+00
-7.65057057e-02 -2.44662330e-01 2.12077677e-01 2.79851049e-01
6.59568727e-01 1.03298783e+00 3.09895165e-02 -2.46066928e-01
1.23545027e+00 -7.98061550e-01 -9.12230313e-01 -5.13714612e-01
1.01275134e+00 -8.26735497e-01 1.34257877e+00 4.22108471e-01
-1.33959186e+00 -5.00926554e-01 -8.55946302e-01 -2.00699672e-01
-4.36312139e-01 2.64318496e-01 3.73048276e-01 1.05963564e+00
-1.14149690e+00 3.27300340e-01 -7.10997820e-01 -2.32638493e-01
-9.31635126e-03 6.68240607e-01 -2.41931096e-01 1.70358643e-01
-1.43845141e+00 1.00557458e+00 6.22205138e-01 -9.61614624e-02
-5.70633113e-01 -4.91043150e-01 -1.03168690e+00 2.53803819e-01
2.29843229e-01 -2.82852501e-01 1.41753280e+00 -7.49043345e-01
-1.81474495e+00 8.36929440e-01 -5.71313381e-01 -6.92309439e-01
9.77724418e-02 -1.30072013e-01 -3.95032078e-01 -2.55678296e-01
-4.57397372e-01 6.22981429e-01 6.72376573e-01 -8.21279287e-01
-7.11213529e-01 -1.57020628e-01 -1.79145217e-01 2.20232442e-01
-9.09726381e-01 2.42402449e-01 -2.30842143e-01 -8.77643347e-01
-1.37738302e-01 -8.31731200e-01 -1.55837357e-01 -1.01537740e+00
-4.24375087e-01 -6.07682645e-01 4.20942605e-01 -9.91194129e-01
1.47836781e+00 -1.88781404e+00 8.13571513e-02 3.98291558e-01
-6.34345040e-02 6.52288616e-01 -3.95278990e-01 3.48029405e-01
-1.43722281e-01 3.52848589e-01 -1.58162862e-01 -6.35793507e-01
-1.46344732e-02 5.08162200e-01 -6.28538311e-01 1.57944374e-02
3.32393259e-01 1.12503839e+00 -5.66731870e-01 -1.85018539e-01
4.57633525e-01 5.75923562e-01 -6.06642544e-01 3.84804234e-02
-2.20096841e-01 1.57806456e-01 4.96189035e-02 1.34523243e-01
1.64270207e-01 2.96630979e-01 2.06457883e-01 3.25834930e-01
3.29106413e-02 1.09054518e+00 -7.11827219e-01 1.34855938e+00
-1.12018955e+00 7.70777106e-01 -2.81174898e-01 -1.27064967e+00
1.09119511e+00 5.90289652e-01 6.96652159e-02 -6.15003228e-01
3.18041414e-01 4.27709997e-01 4.07299638e-01 -4.05004285e-02
6.32982492e-01 -5.40001988e-01 -8.02441984e-02 3.63412023e-01
2.99283296e-01 -4.26861107e-01 3.04151893e-01 -2.33027875e-01
9.68490243e-01 -3.70610774e-01 2.84409404e-01 -1.51080862e-01
6.27386510e-01 -6.68438226e-02 2.13765249e-01 8.43548715e-01
5.03534138e-01 6.14113688e-01 1.77247748e-01 -1.21411253e-02
-1.25949228e+00 -7.76246786e-01 1.44793885e-02 1.53107536e+00
-6.84517384e-01 -4.22788262e-01 -9.50360954e-01 -4.36041266e-01
-2.97143281e-01 1.31744683e+00 -2.20343307e-01 -2.90576786e-01
-1.26536798e+00 -8.25366378e-01 9.71006632e-01 6.16836786e-01
-3.09203640e-02 -1.31082880e+00 1.95110261e-01 5.85608244e-01
-4.66768831e-01 -1.37323833e+00 -6.28495753e-01 5.52136064e-01
-9.34903145e-01 -2.71620452e-01 -8.38109016e-01 -1.02477360e+00
4.46844429e-01 -3.62517498e-02 1.05027795e+00 -5.24336752e-03
7.46909482e-03 -2.93931738e-03 -4.08829570e-01 -7.92685866e-01
-9.94184732e-01 5.29189229e-01 1.89717069e-01 -1.36168242e-01
6.80944622e-01 -4.15338308e-01 1.93491951e-01 5.13176881e-02
-7.65132785e-01 -9.83715951e-02 5.33690751e-01 9.51151669e-01
4.38942999e-01 -8.55532438e-02 6.99709952e-01 -9.82317567e-01
9.28433001e-01 8.32830667e-02 -5.34233391e-01 1.69098645e-01
-3.80204082e-01 3.73951167e-01 9.39202189e-01 -6.92922771e-01
-1.02342737e+00 -4.75552604e-02 -7.27375329e-01 -1.22921400e-01
-1.87434163e-02 5.73340654e-01 -1.71626523e-01 1.51088879e-01
6.79058909e-01 4.23439473e-01 -1.73076570e-01 -6.34411633e-01
3.60195935e-01 9.40522730e-01 3.27040136e-01 -4.24891740e-01
9.95393276e-01 9.62768272e-02 -4.98167634e-01 -1.27665031e+00
-8.78215730e-01 -4.76391703e-01 -3.88604462e-01 1.89794078e-01
7.17530251e-01 -7.05510378e-01 -4.40042168e-01 1.88131765e-01
-1.37992799e+00 -5.12632430e-01 -6.17253065e-01 5.82964659e-01
-4.37517464e-01 1.60389230e-01 -6.63379371e-01 -9.00806367e-01
-6.11052215e-01 -8.77035499e-01 1.00392377e+00 -2.64345735e-01
-4.54228789e-01 -1.15692329e+00 7.24142268e-02 2.90634662e-01
5.88294923e-01 -5.60122430e-01 1.26548481e+00 -1.30974078e+00
6.43847808e-02 -3.23375612e-01 1.12975046e-01 9.16807830e-01
7.87578598e-02 -5.38907647e-01 -1.13450086e+00 -1.64576977e-01
1.44411186e-02 -2.15088964e-01 7.97398984e-01 4.29634392e-01
1.14151776e+00 -2.85626233e-01 4.41980213e-02 4.57424641e-01
8.79437983e-01 3.84374201e-01 5.71502566e-01 1.18814334e-01
7.92963743e-01 5.99043012e-01 2.21375868e-01 -1.18785789e-02
-1.85857080e-02 6.86203301e-01 -1.10568166e-01 -3.14732313e-01
-2.13018253e-01 -2.80518144e-01 6.51451409e-01 1.52746189e+00
6.54398724e-02 -6.74467206e-01 -1.03987384e+00 7.95060575e-01
-1.42116749e+00 -4.79171693e-01 -7.18846023e-02 2.22860193e+00
1.12638164e+00 3.38608027e-01 -5.92816174e-02 3.95851880e-01
7.49308825e-01 -8.76410306e-02 2.50267293e-02 -8.45434964e-01
-2.99503356e-01 9.18895245e-01 5.82598865e-01 8.90158236e-01
-6.55443132e-01 1.47962940e+00 6.19512081e+00 1.20696747e+00
-1.08013785e+00 3.65822464e-01 4.57933754e-01 -1.56287044e-01
-4.34390664e-01 -3.35800648e-01 -1.16401505e+00 7.09193619e-03
1.50773859e+00 -2.69525409e-01 5.80060363e-01 6.93480253e-01
2.69313663e-01 3.17458540e-01 -1.19586408e+00 1.01565456e+00
2.77497977e-01 -1.09176862e+00 5.38262367e-01 3.99684086e-02
5.13846099e-01 1.96546972e-01 -1.19389787e-01 7.24942565e-01
3.79105806e-01 -1.31719267e+00 4.84098345e-01 -1.27936348e-01
8.43078375e-01 -1.00026059e+00 8.48519564e-01 4.07112449e-01
-1.00626779e+00 1.61005184e-01 -5.60264170e-01 -1.76209137e-02
3.20494860e-01 4.63033497e-01 -1.22645283e+00 1.90714777e-01
6.41919225e-02 -2.93195695e-02 -3.01092237e-01 6.67973518e-01
-1.82404429e-01 1.13593233e+00 -5.20546257e-01 -2.75212556e-01
4.21071887e-01 -1.20333990e-03 7.02777207e-01 1.54699957e+00
4.19071555e-01 -2.71934643e-02 -6.77930042e-02 5.56466043e-01
-3.24062198e-01 5.68777800e-01 -5.03387034e-01 -4.93600726e-01
2.98026085e-01 9.46853518e-01 -5.91620684e-01 -5.26202500e-01
-2.15564102e-01 9.55416381e-01 3.96969885e-01 2.11529791e-01
-4.29157406e-01 -6.60978019e-01 4.68329549e-01 1.02502361e-01
3.03304493e-01 -4.27647293e-01 -2.21141979e-01 -9.90945756e-01
1.18815182e-02 -1.06743050e+00 5.45380171e-03 -4.44085628e-01
-1.08257210e+00 1.01775110e+00 -1.39596283e-01 -5.74747264e-01
-6.04274154e-01 -8.14157844e-01 -3.67162913e-01 1.23121154e+00
-1.52600062e+00 -1.05449271e+00 1.90484285e-01 5.04979730e-01
1.04476142e+00 -5.01514673e-01 1.14410138e+00 4.11840349e-01
-4.05383289e-01 9.50455964e-01 2.60404080e-01 3.77544552e-01
5.00072062e-01 -1.33599102e+00 8.27032149e-01 7.71310091e-01
6.19159281e-01 6.38004005e-01 6.31983459e-01 -5.15649259e-01
-1.04377806e+00 -1.15219820e+00 1.37127984e+00 -4.80731905e-01
7.53339231e-01 -9.56393838e-01 -1.09483600e+00 7.38575101e-01
8.24323520e-02 -4.50998813e-01 5.87484479e-01 3.60134125e-01
3.73267308e-02 -3.81595735e-03 -6.51584566e-01 7.75082827e-01
8.57797801e-01 -6.34778678e-01 -9.55367684e-01 4.93645400e-01
1.09618425e+00 -2.89697319e-01 -5.06243527e-01 3.39574993e-01
2.76644260e-01 -2.93273509e-01 8.47570002e-01 -8.27732563e-01
6.95003346e-02 2.25336924e-01 -1.68749616e-01 -1.51363206e+00
-5.21591194e-02 -8.13236892e-01 1.37465358e-01 1.38163853e+00
8.18748534e-01 -7.11440146e-01 6.37778401e-01 1.20365798e-01
-2.74566412e-01 -3.31184834e-01 -8.68487477e-01 -1.07776058e+00
4.34227437e-01 -8.89035106e-01 3.64590138e-01 7.32756317e-01
1.30095914e-01 6.94054246e-01 -2.95636088e-01 -2.22134426e-01
6.37562126e-02 -4.92476642e-01 5.21844804e-01 -1.08968234e+00
-4.25108969e-01 -5.01264811e-01 -1.86230510e-01 -1.12059367e+00
4.91103619e-01 -1.21860778e+00 3.38208020e-01 -1.27677727e+00
-1.12619445e-01 -4.65350449e-01 -3.60821933e-01 6.25435233e-01
-1.42639503e-01 3.43820959e-01 1.08026378e-01 -3.24227273e-01
7.96269849e-02 5.58115423e-01 7.54197955e-01 -6.06828630e-02
-4.93275911e-01 1.45662099e-01 -4.99998808e-01 6.44987762e-01
9.34674740e-01 -6.55107260e-01 -2.92451233e-01 -6.04105294e-01
1.50806382e-01 -8.05621818e-02 -2.59717971e-01 -7.53676355e-01
-6.41165003e-02 2.01599211e-01 2.14685872e-02 -5.83222330e-01
4.66145694e-01 -5.60900211e-01 -5.32297432e-01 3.94894958e-01
-6.10044003e-01 -1.39158843e-02 6.38761222e-01 9.44993123e-02
-2.53020912e-01 -5.61999977e-01 7.39963412e-01 -1.24195078e-02
-1.76707596e-01 -2.63140574e-02 -9.04673696e-01 2.63469607e-01
3.16966623e-01 -1.25458166e-01 2.15925306e-01 -5.18806159e-01
-6.60009325e-01 -1.23397060e-01 -2.11642310e-01 5.26515841e-01
4.68766332e-01 -9.75757718e-01 -9.67041135e-01 4.36731130e-01
-1.08445399e-01 -1.94427118e-01 -1.07941195e-01 5.82078815e-01
-3.01537573e-01 8.46986353e-01 2.99890488e-01 -3.02697808e-01
-1.53749418e+00 2.85164624e-01 3.67681712e-01 -5.37230670e-01
-6.11175358e-01 1.09882975e+00 4.38296407e-01 -6.28092110e-01
4.05455440e-01 -5.50620198e-01 -2.16664553e-01 -3.66449393e-02
5.68280518e-01 6.28880560e-02 5.58711886e-01 -5.53649962e-01
-1.50788978e-01 3.67453426e-01 -3.34341258e-01 -5.28957307e-01
1.27073109e+00 -6.43343255e-02 -2.95414608e-02 7.08170593e-01
1.00013876e+00 3.33290607e-01 -2.59864569e-01 -5.14384568e-01
2.58640170e-01 1.45889506e-01 2.92742133e-01 -5.89349210e-01
-6.99838817e-01 1.21621144e+00 2.10080668e-01 2.14499861e-01
1.00379479e+00 -6.65475875e-02 1.21402442e+00 9.38460827e-01
-8.05060342e-02 -1.30357480e+00 -2.36084059e-01 1.13274360e+00
7.88099647e-01 -1.05283964e+00 -6.24054372e-01 -3.56534779e-01
-4.58172977e-01 1.14120114e+00 2.22502857e-01 -6.24496751e-02
4.31178391e-01 4.68726903e-01 9.39010307e-02 1.38125494e-01
-7.43991315e-01 -3.71326774e-01 4.91915494e-01 3.93853724e-01
7.59462953e-01 7.56964087e-02 -2.70909011e-01 4.54645842e-01
-8.95934641e-01 -4.59930420e-01 4.38184053e-01 5.38325906e-01
-6.49935246e-01 -1.37889791e+00 -4.13193583e-01 4.89758700e-01
-7.62114823e-01 -7.98154175e-01 -4.77493584e-01 7.78378487e-01
-2.77497172e-01 9.91874635e-01 1.86714604e-01 -3.13794166e-01
3.08430374e-01 6.14128411e-01 4.12496835e-01 -1.20772922e+00
-7.31057286e-01 3.57564569e-01 4.83280629e-01 -8.23532268e-02
5.95105626e-02 -4.19718564e-01 -1.23861337e+00 -7.79720619e-02
-6.63050890e-01 3.84837449e-01 1.01131546e+00 1.29424667e+00
-2.55962282e-01 9.00775373e-01 3.42927903e-01 -6.02472603e-01
-6.47492707e-01 -1.40067220e+00 -3.82592499e-01 6.95702480e-03
2.20353216e-01 -1.20095007e-01 -4.32101727e-01 2.14184877e-02]
|
[14.329045295715332, 6.915004253387451]
|
fe9ab15d-98aa-4287-8b17-ae7a549f6ec7
|
combining-contrastive-and-supervised-learning
|
2205.10406
| null |
https://arxiv.org/abs/2205.10406v1
|
https://arxiv.org/pdf/2205.10406v1.pdf
|
Combining Contrastive and Supervised Learning for Video Super-Resolution Detection
|
Upscaled video detection is a helpful tool in multimedia forensics, but it is a challenging task that involves various upscaling and compression algorithms. There are many resolution-enhancement methods, including interpolation and deep-learning-based super-resolution, and they leave unique traces. In this work, we propose a new upscaled-resolution-detection method based on learning of visual representations using contrastive and cross-entropy losses. To explain how the method detects videos, we systematically review the major components of our framework - in particular, we show that most data-augmentation approaches hinder the learning of the method. Through extensive experiments on various datasets, we demonstrate that our method effectively detects upscaling even in compressed videos and outperforms the state-of-the-art alternatives. The code and models are publicly available at https://github.com/msu-video-group/SRDM
|
['Dmitriy Vatolin', 'Ivan Molodetskikh', 'Viacheslav Meshchaninov']
|
2022-05-20
| null | null | null | null |
['video-super-resolution']
|
['computer-vision']
|
[ 1.61517203e-01 -5.25258660e-01 -3.64750832e-01 6.22400455e-02
-1.23945117e+00 -2.97342539e-01 2.22937912e-01 -3.33514541e-01
-4.93805408e-02 5.32115042e-01 3.85936141e-01 -1.57930255e-01
2.61915743e-01 -6.48836970e-01 -6.65117860e-01 -5.27271450e-01
-3.21572810e-01 -2.28844643e-01 5.11263132e-01 -1.42376378e-01
5.53589106e-01 5.34590125e-01 -1.36984682e+00 7.94822395e-01
5.42693794e-01 1.01206362e+00 2.11022168e-01 7.42601216e-01
6.99678659e-02 1.15715766e+00 -4.52570409e-01 -5.82539260e-01
3.41230512e-01 -4.70575482e-01 -7.13541567e-01 -3.52063961e-02
5.36935270e-01 -1.10483778e+00 -1.10396850e+00 1.15559113e+00
6.68259382e-01 -3.47917415e-02 4.85241771e-01 -9.17208850e-01
-8.59528601e-01 4.48543638e-01 -1.08946157e+00 1.17089951e+00
4.82494384e-01 2.20089499e-02 6.59534693e-01 -1.12814879e+00
6.64483547e-01 1.42836666e+00 8.33885729e-01 4.88275737e-01
-1.01204383e+00 -8.88597548e-01 -3.72765869e-01 7.23949373e-01
-1.36646295e+00 -7.00571001e-01 1.02734125e+00 -3.44420671e-01
5.52423239e-01 8.78552198e-02 2.98745513e-01 1.30943537e+00
-3.83317135e-02 8.79741609e-01 1.19490349e+00 -2.19522551e-01
3.09028644e-02 -1.69485107e-01 -2.25051880e-01 7.40191877e-01
2.32556686e-01 1.76913410e-01 -8.87158632e-01 -1.30977005e-01
1.41057265e+00 1.43134943e-03 -3.94035667e-01 -7.37457871e-02
-7.69099891e-01 8.80212426e-01 1.95394814e-01 2.17041701e-01
-1.74298301e-01 4.80493589e-04 5.13031483e-01 2.22168729e-01
6.40019119e-01 -6.49476424e-02 -4.07182537e-02 -8.56096148e-02
-1.45145261e+00 1.67704523e-01 1.45421028e-01 6.85631931e-01
5.29435873e-01 1.80343896e-01 -1.09851345e-01 8.31515193e-01
-2.22731024e-01 1.69295505e-01 4.03297633e-01 -1.39328849e+00
5.25679708e-01 4.84368317e-02 -4.44208756e-02 -1.24169612e+00
-8.25269073e-02 -2.37466261e-01 -9.46944773e-01 2.68701017e-01
1.91125512e-01 1.96405858e-01 -4.19018894e-01 1.31590378e+00
2.96150863e-01 6.77159488e-01 -2.78724313e-01 9.63843703e-01
8.80229414e-01 4.47833419e-01 -2.09399700e-01 -5.78525543e-01
1.26711512e+00 -9.27282393e-01 -8.65288317e-01 8.96714255e-02
-7.81937167e-02 -7.76200831e-01 8.86510134e-01 5.29605687e-01
-1.39018321e+00 -7.98840761e-01 -1.01691711e+00 -4.37452406e-01
8.06838349e-02 -7.74925873e-02 3.60076338e-01 4.50467736e-01
-8.72060359e-01 9.87342119e-01 -6.30477130e-01 -1.29744142e-01
9.75651324e-01 -2.09050477e-01 -2.92019904e-01 -2.31594265e-01
-1.09188211e+00 4.96978998e-01 1.75137714e-01 -4.53695595e-01
-1.09482551e+00 -5.74362576e-01 -7.05285788e-01 -9.21448916e-02
3.75995338e-01 -3.78999531e-01 1.08423066e+00 -4.66048956e-01
-1.07503557e+00 9.45616126e-01 -2.68285006e-01 -5.67686319e-01
8.26347649e-01 -4.16423082e-01 -5.68966985e-01 8.65817904e-01
9.59251300e-02 2.89543211e-01 1.34159243e+00 -1.28801978e+00
-5.73645651e-01 -1.01301938e-01 -3.03474292e-02 -1.77907154e-01
-3.68771732e-01 7.01879442e-01 -6.67808890e-01 -9.17142630e-01
-5.76987490e-02 -2.55970269e-01 2.46878505e-01 2.07740694e-01
-2.62561083e-01 2.22628698e-01 9.06803250e-01 -1.35118759e+00
1.48872256e+00 -2.19209886e+00 -2.10337102e-01 -4.60718066e-01
4.78210211e-01 3.67216051e-01 -1.58653215e-01 1.48161471e-01
-1.57323077e-01 3.48429114e-01 -2.50461340e-01 -4.05744344e-01
-2.29776576e-01 -3.91219586e-01 -5.71654618e-01 7.27424741e-01
2.80027479e-01 6.01661801e-01 -8.66697729e-01 -8.74490499e-01
3.26601058e-01 1.00142455e+00 -3.59730542e-01 3.45255464e-01
2.90647298e-01 5.75936139e-01 -1.31134346e-01 1.00169313e+00
1.13178289e+00 -2.93774694e-01 7.24614784e-03 -5.66590667e-01
-1.06392711e-01 2.47013614e-01 -1.22619069e+00 1.69309938e+00
-1.19357280e-01 8.80145013e-01 3.61335129e-02 -9.42512691e-01
6.70930207e-01 1.27378717e-01 5.31690001e-01 -1.00341964e+00
4.02670540e-03 2.43628219e-01 -6.14826441e-01 -7.86418378e-01
6.81257188e-01 3.96469422e-02 4.34988827e-01 4.49399561e-01
8.58600289e-02 4.68420088e-01 2.81650901e-01 3.68128300e-01
1.30228615e+00 2.59896100e-01 4.06666517e-01 3.11129272e-01
5.55110157e-01 -4.77250785e-01 7.70773709e-01 6.56295300e-01
-5.66677511e-01 1.07475960e+00 6.26748621e-01 -3.81534219e-01
-1.33547533e+00 -9.72832203e-01 -2.39013419e-01 1.00554252e+00
2.93468088e-01 -5.14892280e-01 -8.43143463e-01 -5.32390952e-01
-3.06267112e-01 2.89669663e-01 -5.41633844e-01 2.29986951e-01
-8.69694531e-01 -6.87581241e-01 5.63813806e-01 4.57511395e-01
8.32818449e-01 -8.79643619e-01 -4.74757135e-01 -2.15543844e-02
-6.99028134e-01 -1.50632262e+00 -4.91441458e-01 -5.13385177e-01
-9.75295424e-01 -1.22742307e+00 -7.60810256e-01 -4.82811332e-01
1.96075544e-01 8.34666669e-01 1.16287160e+00 3.66163045e-01
-5.49698472e-01 1.96312323e-01 -4.23444688e-01 3.02274048e-01
-5.19980252e-01 -3.73717099e-01 1.04743196e-02 1.75140481e-02
1.82874054e-01 -7.95290232e-01 -9.60578382e-01 8.56219605e-02
-9.74167347e-01 -3.96913812e-02 5.22909641e-01 5.97069561e-01
8.31089497e-01 3.10240954e-01 2.83434808e-01 -7.05543041e-01
5.10396242e-01 -6.03332102e-01 -2.97908038e-01 2.43134946e-02
-4.06434417e-01 -2.49345288e-01 5.10899007e-01 -4.45195764e-01
-1.06771386e+00 -3.77393484e-01 4.89772391e-03 -1.03398609e+00
-9.97981951e-02 -1.02331735e-01 -5.13517112e-02 -2.30914846e-01
6.37292922e-01 4.20287549e-01 -2.20405698e-01 -9.53198433e-01
2.32779875e-01 5.99283993e-01 8.20692241e-01 -2.75257617e-01
9.66708958e-01 8.98395956e-01 -9.31514204e-02 -8.42895746e-01
-7.44885981e-01 -3.40325058e-01 -6.94691122e-01 -2.89408743e-01
5.89675486e-01 -1.17742383e+00 -3.21255982e-01 3.57397944e-01
-1.17730403e+00 -1.40418485e-01 -9.32565257e-02 1.69554114e-01
-6.33735299e-01 1.12566233e+00 -1.13937795e+00 -9.23948884e-01
-4.48308259e-01 -1.16499829e+00 1.00570536e+00 2.76474714e-01
2.78274357e-01 -6.79165542e-01 1.81492165e-01 7.56415784e-01
5.16538084e-01 2.94898897e-01 5.35452545e-01 -3.36701930e-01
-5.98136425e-01 2.32429788e-01 -6.43700957e-01 5.55429399e-01
-1.56084836e-01 -1.22940436e-01 -1.09152615e+00 -5.68120658e-01
1.50943190e-01 -3.15687478e-01 1.29371440e+00 3.02323699e-01
1.65819252e+00 -3.63287508e-01 5.07011414e-02 1.12503600e+00
1.55996323e+00 -2.50856280e-01 1.02987158e+00 6.17324471e-01
5.08690059e-01 3.42762768e-01 6.29905343e-01 6.39526784e-01
1.54560104e-01 6.79046631e-01 5.87706387e-01 -5.21288626e-02
-6.95073307e-01 -3.42687756e-01 5.52682281e-01 5.51569402e-01
-5.13252079e-01 2.28828825e-02 -3.76925051e-01 2.19093308e-01
-1.66518950e+00 -1.56733859e+00 -8.95803720e-02 2.19767165e+00
8.32592189e-01 2.25341376e-02 4.40137118e-01 2.48072937e-01
1.19524801e+00 5.72029591e-01 -5.74993074e-01 -7.14634210e-02
-4.24092382e-01 3.35030586e-01 3.64920855e-01 3.81388575e-01
-1.38477135e+00 8.60193014e-01 6.41547394e+00 1.26329005e+00
-9.22905922e-01 5.66600740e-01 8.05093110e-01 -1.79455325e-01
2.91560050e-02 -2.90305436e-01 -5.83972156e-01 7.90943384e-01
7.53594160e-01 1.46751240e-01 6.87368929e-01 6.90494418e-01
4.35235590e-01 9.58152264e-02 -6.66538119e-01 1.38412571e+00
4.07043457e-01 -1.60885179e+00 -5.08846864e-02 2.46987455e-02
5.75439394e-01 -1.14956811e-01 2.54637271e-01 1.34979635e-01
-3.27688485e-01 -7.91451752e-01 6.10686541e-01 3.37080806e-01
1.29992795e+00 -8.05019557e-01 4.65254664e-01 -1.10826418e-01
-1.32539296e+00 -2.88292378e-01 -7.15598583e-01 2.23164529e-01
1.51715636e-01 7.01636255e-01 -1.28267139e-01 5.37917078e-01
1.08824563e+00 7.91122913e-01 -6.35829866e-01 1.11640906e+00
-2.65003949e-01 6.10978961e-01 1.90912336e-01 8.85539234e-01
-4.92165148e-01 1.46800792e-02 6.49221122e-01 1.56348729e+00
5.87600112e-01 3.00267041e-02 -1.51294693e-01 9.82064068e-01
-3.46265405e-01 -1.05768979e-01 -3.15531313e-01 2.20214143e-01
7.31647253e-01 1.38808954e+00 -5.43930471e-01 -2.58520633e-01
-5.60312450e-01 1.16873693e+00 2.59719402e-01 3.11640322e-01
-1.16711998e+00 -1.57044977e-01 6.21708572e-01 4.69971031e-01
5.40701568e-01 -1.38030589e-01 -2.93358285e-02 -1.51045632e+00
7.15521947e-02 -1.02385747e+00 5.25233805e-01 -8.19757640e-01
-1.29885316e+00 4.34934348e-01 -1.79762930e-01 -1.51416659e+00
-3.87201980e-02 -3.85467529e-01 -6.62987590e-01 4.14276958e-01
-1.97781706e+00 -9.35486555e-01 -5.00845969e-01 8.06504369e-01
8.43495965e-01 -2.80719280e-01 2.75252759e-01 6.59813464e-01
-6.97454035e-01 6.45578980e-01 1.97150230e-01 4.01451826e-01
9.71927762e-01 -7.90619135e-01 2.69004315e-01 1.29640746e+00
-4.42365138e-03 3.20010632e-01 7.05978155e-01 -6.22956991e-01
-1.18013358e+00 -1.00195324e+00 2.66630888e-01 -2.88003739e-02
7.77324140e-01 -7.27850422e-02 -1.20443380e+00 4.40816104e-01
2.92945713e-01 3.05287331e-01 6.40371084e-01 -4.87881929e-01
-7.95449436e-01 -6.44195154e-02 -1.39411652e+00 1.97958454e-01
1.14696968e+00 -6.26854241e-01 -4.18336034e-01 3.19588125e-01
5.94350278e-01 -3.89383018e-01 -8.11601162e-01 3.60936522e-01
4.32620525e-01 -1.70018315e+00 1.55050051e+00 -3.73839825e-01
8.55311036e-01 -2.55637914e-01 -2.82619715e-01 -5.94482958e-01
-6.51739657e-01 -6.98388338e-01 -1.08237374e+00 1.31303036e+00
-3.28883827e-01 -2.10162938e-01 6.02580488e-01 2.46596262e-02
1.97563708e-01 -5.32842934e-01 -1.10141230e+00 -7.93484569e-01
7.45454878e-02 -3.60754311e-01 4.25421029e-01 1.00600672e+00
-2.12167442e-01 -3.75410840e-02 -8.62584531e-01 2.83069104e-01
1.32364750e+00 5.27342409e-02 5.25742650e-01 -7.64722347e-01
-4.93234783e-01 -4.70019132e-01 -3.06114107e-01 -8.88737500e-01
-1.80427104e-01 -4.33666795e-01 -5.82385242e-01 -1.33927715e+00
8.23273361e-01 2.04307973e-01 -3.58289808e-01 9.22811851e-02
-1.92719087e-01 6.70561492e-01 2.67680407e-01 6.33333921e-01
-8.71754110e-01 1.89227045e-01 1.04579651e+00 1.37411403e-02
1.22983985e-01 -3.57669860e-01 -8.02530169e-01 8.10819983e-01
1.08262646e+00 -3.56517881e-01 1.71503071e-02 -3.48486960e-01
4.82076183e-02 3.31245363e-01 6.33103311e-01 -1.14059007e+00
1.59609661e-01 -2.11429037e-02 7.32931077e-01 -6.89326704e-01
4.64121878e-01 -2.61252373e-01 -7.46240467e-02 3.55454445e-01
-2.91256577e-01 -2.41418779e-02 -7.61339441e-02 6.51589453e-01
-2.01583967e-01 -1.87342912e-01 1.24805772e+00 -3.22209507e-01
-8.79849553e-01 4.23552573e-01 -1.14809580e-01 1.94701567e-01
7.32128382e-01 -6.48823306e-02 -7.01109707e-01 -5.14940262e-01
-3.48196656e-01 -3.72149974e-01 6.57412410e-01 2.52173275e-01
1.13206923e+00 -1.30867243e+00 -1.02129257e+00 -1.73542798e-02
-2.88973242e-01 -4.00658011e-01 6.86045706e-01 8.48375559e-01
-6.84271991e-01 -6.36794865e-02 -2.89076060e-01 -2.81393588e-01
-1.38093555e+00 9.85271811e-01 1.68436170e-01 -2.37982094e-01
-9.07597899e-01 6.54990554e-01 1.61324382e-01 3.05335164e-01
2.21950382e-01 3.48894268e-01 -3.50213289e-01 -4.39163297e-02
1.23061538e+00 9.37973380e-01 -2.63967365e-01 -6.92633033e-01
-1.46126866e-01 4.89903718e-01 -1.50407284e-01 1.24855392e-01
1.56951261e+00 -4.59585667e-01 -2.91785877e-02 -2.71866620e-02
1.33169138e+00 9.55554247e-02 -1.48386478e+00 -3.69744629e-01
-4.73691016e-01 -1.16328228e+00 1.81142777e-01 -3.61096680e-01
-1.44514430e+00 1.14962542e+00 8.74768376e-01 1.64912105e-01
1.38566661e+00 -6.75407723e-02 1.08495593e+00 -2.21758470e-01
3.52715969e-01 -1.11295879e+00 5.20738184e-01 8.10430795e-02
8.55015278e-01 -1.40590155e+00 4.26502138e-01 -4.32485342e-01
-4.82230961e-01 1.32950819e+00 4.46297050e-01 -2.87976235e-01
3.31870496e-01 2.04762399e-01 -3.71642977e-01 1.07934713e-01
-4.96338993e-01 -2.00688660e-01 -1.19187266e-01 8.05879474e-01
3.94930959e-01 -3.53720039e-01 -1.97672293e-01 4.92504716e-01
9.36847553e-02 1.44670352e-01 7.85705507e-01 5.50940394e-01
-5.86531818e-01 -8.97376120e-01 -7.27743387e-01 1.64744109e-01
-8.88879716e-01 -2.02225447e-01 -6.06165081e-02 7.28557825e-01
1.31476641e-01 1.01101673e+00 -5.21873944e-02 -6.20886564e-01
-9.72106755e-02 -3.18440914e-01 5.82888544e-01 2.48157280e-03
-2.66224034e-02 2.85482615e-01 -1.65282845e-01 -1.01403069e+00
-5.12736738e-01 -8.60042334e-01 -8.93508673e-01 -7.57118523e-01
-2.96142558e-03 -3.15967709e-01 3.38346064e-01 4.65519100e-01
4.41120118e-01 3.94322753e-01 9.05944884e-01 -1.09913886e+00
-4.57838118e-01 -7.51683652e-01 -7.30652928e-01 5.01226246e-01
4.87740546e-01 -6.54851198e-01 -7.50310481e-01 2.53524154e-01]
|
[11.102272987365723, -2.0060501098632812]
|
a5f68d08-8b05-4b6e-ae88-2c2a2a0e33d0
|
bert-goes-shopping-comparing-distributional
|
2012.09807
| null |
https://arxiv.org/abs/2012.09807v2
|
https://arxiv.org/pdf/2012.09807v2.pdf
|
BERT Goes Shopping: Comparing Distributional Models for Product Representations
|
Word embeddings (e.g., word2vec) have been applied successfully to eCommerce products through~\textit{prod2vec}. Inspired by the recent performance improvements on several NLP tasks brought by contextualized embeddings, we propose to transfer BERT-like architectures to eCommerce: our model -- ~\textit{Prod2BERT} -- is trained to generate representations of products through masked session modeling. Through extensive experiments over multiple shops, different tasks, and a range of design choices, we systematically compare the accuracy of~\textit{Prod2BERT} and~\textit{prod2vec} embeddings: while~\textit{Prod2BERT} is found to be superior in several scenarios, we highlight the importance of resources and hyperparameters in the best performing models. Finally, we provide guidelines to practitioners for training embeddings under a variety of computational and data constraints.
|
['Jacopo Tagliabue', 'Bingqing Yu', 'Federico Bianchi']
|
2020-12-17
| null |
https://aclanthology.org/2021.ecnlp-1.1
|
https://aclanthology.org/2021.ecnlp-1.1.pdf
|
acl-ecnlp-2021-8
|
['product-recommendation']
|
['miscellaneous']
|
[-0.17203002 0.21157399 -0.24691662 -0.46110868 -0.3968241 -0.7317641
0.78310156 0.25900447 -0.48084182 0.37924308 0.5012901 -0.59287757
-0.36032808 -0.7069598 -0.47359687 -0.23117009 -0.05201464 0.25989267
-0.3647124 -0.5070938 0.25797856 0.10884058 -1.3814503 0.08411982
0.66734844 0.9617384 0.30924928 0.6699431 -0.3121246 0.4146094
-0.52324706 -1.1278193 0.29234204 -0.12492981 -0.7114521 0.24521735
0.30726358 -0.14740187 -0.7047155 0.69089216 0.30778354 0.3339685
0.854794 -1.1609416 -1.8387729 0.5433201 -0.0778142 0.27074188
0.17116353 0.15792976 1.7288482 -1.1592593 0.8534871 1.0149266
0.6061 0.2156592 -1.4603426 -0.25446305 0.5290437 0.20490788
-1.3359492 -0.01683511 0.4184278 -0.31887007 1.1344787 0.30275324
0.3898557 1.6242818 0.40213028 0.9951256 0.8113513 -0.47196332
0.1596321 0.8924654 0.4771831 0.25512442 0.46724954 0.09614085
-0.3877213 -0.09960809 0.9164843 0.05522515 0.04958855 -0.3146195
-0.88138914 1.4274191 0.4428727 0.43568775 -0.41800523 0.29206678
0.3128929 0.7125424 0.75987226 1.0811782 -0.75734425 -0.04220406
-0.37078184 0.591788 0.9248094 1.4853625 0.41292888 -0.05656907
-0.39030308 1.3025639 0.38107717 0.11114974 0.49695608 -0.66978616
0.49314368 0.1765953 0.43898553 -0.95112276 -0.19917184 -0.59141636
-0.11574002 -0.14995429 0.1394852 -0.47843614 -0.94966507 1.1686282
-0.21508436 -0.31451094 -0.12298851 0.8166675 0.5967489 0.6332928
0.3884539 0.23928593 1.369978 -1.1138241 -0.76448905 -0.41628975
0.89763 -0.8628802 1.0340275 0.30802065 -0.85240513 -0.82341933
-1.0301982 -0.24708016 -1.0574057 0.04079328 0.9485918 0.90919906
-0.8799758 0.5751945 -0.21667251 -0.84185237 0.17310539 0.306895
-0.33813545 -0.36756337 -1.070733 1.1460638 0.2969397 0.21909384
-0.6314306 -0.5569687 -1.0648103 0.2627273 -0.24131654 -0.5373905
1.2533991 -0.45048138 -1.3057245 0.36030558 0.13835114 -0.2970522
0.06116641 -0.44079417 -0.9155 -0.28574273 -0.08797181 0.72697407
0.47833562 -1.0535156 -0.6170527 -0.15168127 0.19736567 0.13218853
-0.82190377 0.16024296 -0.33139184 -0.61561954 -0.08935151 -0.8871588
-0.53967667 -0.28364715 -0.2068727 -0.5590861 0.34641737 -0.60286206
1.2291158 -2.3244822 0.00978189 0.09238487 0.04721054 0.2098577
-0.7968987 0.9077812 0.03973436 0.41905314 0.4132273 -0.39104488
0.701191 0.17355706 -0.09140642 0.10984773 0.5778398 1.2495364
-0.7179525 -0.02321687 0.4270064 0.60448325 -0.5780425 0.09381564
-0.33758473 -0.31920683 -0.64150935 0.47817618 0.50595236 -0.13400829
0.33747518 0.20809692 -0.04350685 0.20171611 -0.84756887 1.4843276
-0.87333655 0.75303966 -0.21719152 -0.6165513 1.0595771 0.2624516
0.3027037 -0.79606926 0.50245625 0.13306306 -0.09883992 -0.6786289
1.2121961 0.25419208 -0.29120353 0.2879612 0.7111073 -0.12413074
0.17578593 0.03558975 1.0723447 0.1793508 -0.28706574 -0.24487989
-0.20157708 0.12700717 -0.0528403 0.83723605 -0.11228055 0.74829537
0.42424572 -0.34392002 -1.0528003 -0.93513757 -0.39078036 1.6985338
0.1503287 -0.49349827 -0.3148946 -0.81794006 0.25935003 1.4000057
-0.7925273 -0.02527315 -0.15586893 -0.88803285 0.17453754 0.8017317
-0.1619368 -0.9889865 -0.04271171 0.76816964 0.42978704 -0.9284919
-0.49488536 0.474282 -0.6679578 -0.71292114 -0.6921353 -1.035996
0.36523882 0.30403852 1.1609389 -0.32874894 -0.30782026 0.54727596
-0.8925529 -0.7138688 0.15899828 0.28775263 -0.07859765 -0.13289341
1.0492723 -0.50598824 -0.60795027 0.46332502 -0.9841092 -0.651021
0.6407113 0.7857611 0.18024713 -0.01917175 0.60192317 -1.0783832
1.4355406 -0.79898775 -0.07297716 0.28653407 -1.0200039 -0.18899581
0.35010338 -0.71188927 -0.9444953 -0.5862829 -0.20936982 -0.24172537
-0.10972242 0.66250277 0.33585155 0.26320034 0.7412027 -0.00978304
-0.2645808 -0.8338967 1.0606242 0.8512965 -0.03494688 -0.3019545
0.61084574 -0.33887854 -0.8060194 -0.8914551 -0.6211941 -0.446167
-0.41790456 0.34901527 0.9715392 -0.7769101 -0.4884533 -0.24602197
-0.99155986 -0.09452436 -0.46937984 0.74218345 -0.15762669 -0.1056108
-0.794142 -0.6885034 0.04198766 -0.8560909 0.6866544 0.1764681
-0.5939338 -1.3675104 0.0336928 0.3930488 0.6634453 -0.15188003
1.1444316 -1.1411626 -0.33999383 -0.57767934 -0.36340436 0.591914
0.02091288 -0.042665 -0.9075234 -0.064897 -0.45355102 -0.35540307
0.7018508 0.3225 0.9221306 -0.07074551 -0.18937163 0.21220636
1.3050033 0.19951658 0.56240374 0.466311 0.45774406 0.53763425
0.66317695 0.42826083 0.41686872 0.587132 0.20245732 0.09072634
-0.01210439 -0.53932077 0.3339452 0.7329657 0.03638865 -0.5350364
-0.14899093 0.79087436 -1.6151023 -0.5990316 -0.20582739 1.7585291
0.2578387 0.10285752 0.01380356 -0.33874512 0.22197206 0.4829091
-0.30349064 -1.2591182 0.1961975 0.70639694 0.5928454 0.2674087
-1.1011877 0.9185245 7.1844487 0.55686873 -0.6420418 0.29251257
0.46574348 -0.18174349 -0.74199635 -0.24307495 -0.821801 0.44201684
1.2099588 -0.31027904 0.35385132 1.0648392 -0.03151307 0.25792935
-1.4531327 0.6896603 0.17232496 -1.3057292 -0.25983125 0.3898589
0.9188846 0.01420445 0.4969936 1.003209 0.80729425 -1.3906555
0.45753717 -0.09566104 0.63866496 -0.599284 0.8798744 -0.16869754
-0.8997864 -0.3882904 -0.66024566 -0.01979995 0.16380456 0.2827507
-0.77532136 0.9051661 0.4575998 0.6663524 -0.57236266 0.8474787
-0.10190403 0.40844622 0.1244325 -0.49435633 0.661249 -0.2580495
-0.2726198 1.3575877 0.5191815 -0.17508559 -0.04237612 0.8873908
-0.02922085 0.32294905 -0.75315565 -0.5989993 0.3588169 1.1921949
-0.15909341 0.13567516 -0.75905967 1.2585832 0.28324893 0.7147533
-0.71799755 -0.7749994 1.0458773 0.05233185 0.7782711 -0.44665563
-0.32456398 -0.909615 -0.1034273 -0.4966655 0.19448118 -0.737772
-1.7354764 0.83402085 -0.01927325 -0.96475637 -0.14535981 -1.0622531
-0.48449454 0.93205994 -1.4051801 -1.0906596 0.19917221 0.03792252
0.9044645 -0.28027254 1.0272359 0.15258133 -0.5855924 0.8124736
0.7109393 -0.09128838 0.48668042 -1.1430159 0.7764564 0.18895088
0.43084744 0.8525544 0.7879155 -0.30250135 -1.5185828 -1.1919833
1.4179721 -0.89153534 0.97308415 -0.7047195 -0.41088185 1.1650317
0.6905035 -0.28630415 0.9767926 0.97403175 -0.53215307 0.04230854
-1.1178243 0.73073775 1.0347121 -0.497119 -0.6454054 0.60941726
0.9715902 0.16397026 -1.3041309 -0.32401013 0.6433349 -0.48265505
0.8713008 -0.9017302 0.68283564 0.5167123 -0.336009 -1.7150385
-0.9707241 -0.3782674 0.02230675 1.2685562 0.84867054 -0.638274
0.6070735 1.0010818 -0.14048646 -0.75183713 -0.5894113 -0.8343176
0.3497263 -0.43740022 0.57442373 0.79503506 0.01481683 0.3710328
-0.2680012 -0.04312434 0.13304256 -0.2857427 0.5560605 -1.0935773
-0.26163378 -0.28582597 -0.30036455 -1.329225 -0.03552287 -0.8165525
-0.10021781 -1.6976441 -0.10572633 -0.44732556 -0.77270484 0.13207608
0.0563009 0.29291806 0.30793875 -0.2739123 -0.33869445 0.5977608
1.1532971 -0.13148755 -0.31162328 -0.21087448 -1.367974 0.0121424
0.5253002 -0.1642198 -0.63464785 -0.7848928 0.05360372 -0.33237383
-0.02344486 -0.43332037 -0.20105776 -0.16936669 0.575459 -0.05947413
0.55081594 -0.8948335 -0.08204706 -0.20591912 -0.68935335 0.23245068
0.15982589 0.74493605 -0.0530028 -0.62773114 -0.15002288 -0.12915494
-0.837537 0.222912 -0.62891614 -0.3200102 0.9470368 -0.23195781
-0.19113098 -0.3991378 -0.92619234 0.37842628 -0.03220892 1.1167722
0.3386512 -1.4561634 -0.574896 0.30149055 0.5544896 -0.6823826
0.09379435 0.4059777 -0.32523942 0.9102407 0.05438443 0.10613243
-0.88129634 0.56407416 -0.2280246 -0.32859042 -0.30824357 1.0674769
0.11937756 -0.528884 0.17964296 -0.29547033 -0.20943746 0.20910387
0.36973587 0.30731663 0.13844842 -0.17586493 0.06431817 -0.04863492
-0.7736431 -0.2171924 1.5461042 -0.03453531 0.50054365 0.22562663
1.4341453 -0.45613864 -1.1188883 -0.10961726 0.07656644 -0.7892986
0.1909572 -0.8957067 -0.8210317 0.6671092 0.5842132 0.5598956
0.5401372 -0.04953419 0.71538305 0.3586151 0.5900682 -1.374566
0.0675571 0.22530654 0.84136623 -1.0008899 -0.18731688 -0.3123684
-0.98484254 0.89117205 0.445173 -0.29214618 0.8158851 -0.41704732
0.16685605 -0.2623719 -0.806096 -0.17562824 -0.02040597 0.66675717
0.59739906 0.2515079 -0.66032743 0.7214692 -0.21694252 0.06845322
0.44854444 0.9225683 -0.21664424 -1.5575955 0.1794991 0.6439368
-0.18096109 -0.04907745 -0.35591263 1.0561417 0.19938646 1.398452
0.19461901 -0.6032817 0.64848346 0.2708391 0.46142045 -0.9172693
-0.9986946 -0.03035774 0.5623213 -0.28197408 0.14984651 -0.63118947
-0.36755508 -0.43859342 -0.4840737 0.17966554 0.88234174 0.54442567
0.58776134 0.56302285 0.9114708 -0.5891195 -0.8989403 -1.3662393
-1.190802 0.69173723 -0.24947518 -0.46444678 -0.23881242 -0.30142754]
|
[10.538130760192871, 8.406134605407715]
|
884da18b-b6bf-410f-9727-0cbaad00107a
|
svt-supertoken-video-transformer-for
|
2304.00325
| null |
https://arxiv.org/abs/2304.00325v2
|
https://arxiv.org/pdf/2304.00325v2.pdf
|
SVT: Supertoken Video Transformer for Efficient Video Understanding
|
Whether by processing videos with fixed resolution from start to end or incorporating pooling and down-scaling strategies, existing video transformers process the whole video content throughout the network without specially handling the large portions of redundant information. In this paper, we present a Supertoken Video Transformer (SVT) that incorporates a Semantic Pooling Module (SPM) to aggregate latent representations along the depth of visual transformer based on their semantics, and thus, reduces redundancy inherent in video inputs.~Qualitative results show that our method can effectively reduce redundancy by merging latent representations with similar semantics and thus increase the proportion of salient information for downstream tasks.~Quantitatively, our method improves the performance of both ViT and MViT while requiring significantly less computations on the Kinectics and Something-Something-V2 benchmarks.~More specifically, with our SPM, we improve the accuracy of MAE-pretrained ViT-B and ViT-L by 1.5% with 33% less GFLOPs and by 0.2% with 55% less FLOPs, respectively, on the Kinectics-400 benchmark, and improve the accuracy of MViTv2-B by 0.2% and 0.3% with 22% less GFLOPs on Kinectics-400 and Something-Something-V2, respectively.
|
['Madian Khabsa', 'Senem Velipasalar', 'Qifan Wang', 'Hanchao Yu', 'Rui Hou', 'Chenbin Pan']
|
2023-04-01
| null | null | null | null |
['video-understanding']
|
['computer-vision']
|
[-4.53439765e-02 -4.76410948e-02 -1.99736252e-01 -2.61588097e-01
-7.00367689e-01 -5.39791465e-01 1.95201337e-01 -6.58084005e-02
-7.99699426e-01 4.18516368e-01 4.79064584e-01 1.32689774e-01
1.97634429e-01 -7.79831290e-01 -9.66004729e-01 -5.08954048e-01
-7.49598965e-02 -1.34768456e-01 5.44228494e-01 -1.54580563e-01
1.34645700e-02 2.05104694e-01 -1.64827502e+00 7.76466548e-01
2.44436041e-01 1.34714448e+00 2.72804558e-01 6.60421968e-01
2.08971184e-02 1.12996006e+00 -4.01038766e-01 -4.01626766e-01
3.62775534e-01 -5.26738130e-02 -7.56520331e-01 -1.58955920e-02
6.38165116e-01 -5.18220603e-01 -9.51894104e-01 9.97025788e-01
4.20746773e-01 2.07533702e-01 4.13079560e-01 -1.27575529e+00
-1.64054558e-01 6.21767461e-01 -7.30970562e-01 5.80917180e-01
1.97429597e-01 4.45115000e-01 9.77066338e-01 -9.97663915e-01
6.79456055e-01 1.38608408e+00 6.00948155e-01 4.97256696e-01
-1.18311942e+00 -8.05867255e-01 2.77866870e-01 2.95339614e-01
-1.43039680e+00 -5.67895949e-01 3.98919404e-01 -2.68806607e-01
1.38221896e+00 4.21539694e-02 5.77861547e-01 1.12029421e+00
1.99127600e-01 8.10713232e-01 7.77687490e-01 4.04218286e-01
1.04244493e-01 -3.07843477e-01 3.29761356e-02 9.64608014e-01
7.14610443e-02 -4.85423356e-01 -1.05986845e+00 4.38671172e-01
1.10800529e+00 2.66170293e-01 -2.95187950e-01 -7.77663216e-02
-1.40824199e+00 5.69452822e-01 6.42334461e-01 2.30293386e-02
-4.00270492e-01 7.65804350e-01 7.36075759e-01 3.26911598e-01
6.27913415e-01 6.70632198e-02 -3.19577515e-01 -5.77954531e-01
-9.98683274e-01 1.47635356e-01 4.16158229e-01 1.18795586e+00
7.96967447e-01 5.08562997e-02 -1.01516001e-01 6.11444116e-01
1.05712585e-01 5.75273037e-01 6.54284894e-01 -1.30155098e+00
1.12179279e+00 6.56119823e-01 -7.05290288e-02 -1.00193226e+00
-3.52569640e-01 -8.59679878e-02 -8.29558372e-01 2.34693199e-01
1.08280107e-01 -1.12618402e-01 -1.19720852e+00 1.71516228e+00
-8.15533847e-02 2.21401036e-01 8.63330588e-02 9.53010857e-01
1.10677242e+00 7.35283256e-01 2.46788383e-01 -8.34222045e-03
1.49803185e+00 -1.21005535e+00 -5.77806711e-01 -3.74726653e-01
4.29470360e-01 -5.20309448e-01 1.18360937e+00 2.32940182e-01
-1.45410717e+00 -7.05425322e-01 -1.06262767e+00 -5.53779602e-01
-2.32769981e-01 -1.18600331e-01 4.95577008e-01 2.79683143e-01
-1.43186438e+00 9.56482887e-01 -1.23278368e+00 -1.55908689e-01
6.12756252e-01 5.98738015e-01 -7.27829039e-01 -1.70538142e-01
-9.43796337e-01 5.43725491e-01 2.50155419e-01 -2.02277660e-01
-1.02753246e+00 -8.79593253e-01 -1.06145263e+00 4.43543941e-01
4.45853263e-01 -7.10981667e-01 1.15433192e+00 -7.55730391e-01
-1.46202695e+00 5.47084391e-01 -3.11093062e-01 -4.93995041e-01
5.38526058e-01 -5.09830773e-01 1.69710312e-02 7.66084373e-01
2.43673041e-01 1.24333763e+00 7.66459405e-01 -6.88079238e-01
-7.08338559e-01 -3.04419994e-01 1.68374538e-01 5.68964660e-01
-3.71852994e-01 8.50829296e-03 -1.07066345e+00 -7.17294872e-01
1.97638497e-01 -7.50675857e-01 -1.02127671e-01 2.07104355e-01
-2.17432044e-02 -3.11352443e-02 7.38231838e-01 -7.90385187e-01
8.71315956e-01 -2.40482831e+00 5.88753164e-01 -8.16722140e-02
7.32702792e-01 -3.52759250e-02 -3.62916917e-01 1.13092370e-01
-9.07911640e-03 2.40949929e-01 2.09548119e-02 -7.03852832e-01
-2.59880692e-01 2.67744988e-01 -1.56073526e-01 4.92303848e-01
2.90680259e-01 1.09991574e+00 -7.91318953e-01 -3.78949910e-01
2.48510078e-01 5.26010811e-01 -8.48843634e-01 1.29449740e-02
3.07808723e-02 7.81102404e-02 -1.29420727e-01 7.02884316e-01
4.12062913e-01 -3.45814526e-01 5.18559888e-02 -3.31820130e-01
-3.71960141e-02 4.66408670e-01 -8.92286241e-01 2.06652427e+00
-3.97458941e-01 1.03127062e+00 2.26930588e-01 -9.11318302e-01
4.75247383e-01 4.16061163e-01 7.57442057e-01 -8.31384361e-01
1.30229428e-01 -8.59978721e-02 -4.56703514e-01 -4.88664806e-01
6.80792212e-01 -1.48293391e-01 -1.69429004e-01 2.13128012e-02
3.63878399e-01 2.40327358e-01 1.27658516e-01 4.13039744e-01
1.58623016e+00 1.19576015e-01 -2.31167600e-01 -1.28832877e-01
8.03404972e-02 3.15678455e-02 4.64498758e-01 4.75487828e-01
-2.56434202e-01 6.82969034e-01 7.17124224e-01 -2.18738779e-01
-1.04135060e+00 -1.33731413e+00 4.50422138e-01 1.14573908e+00
3.55230242e-01 -7.19297349e-01 -6.35494471e-01 -4.12723541e-01
4.55854423e-02 2.23975018e-01 -5.12328207e-01 -1.77208945e-01
-6.34030342e-01 -3.72084647e-01 7.34305680e-01 9.71943676e-01
8.07853043e-01 -8.58233631e-01 -7.46102393e-01 1.70399874e-01
-5.19654870e-01 -1.41853368e+00 -2.19195276e-01 1.69398904e-01
-1.25573957e+00 -8.31743956e-01 -8.55001330e-01 -4.78567183e-01
4.78310645e-01 5.64998925e-01 1.02324772e+00 -2.67842323e-01
-4.72418629e-02 1.38931736e-01 -3.28901112e-01 1.10583261e-01
2.35262021e-01 1.43897653e-01 1.01202510e-01 -2.84591079e-01
1.07342489e-01 -5.60761213e-01 -7.37583220e-01 3.59383434e-01
-8.18216622e-01 4.10122395e-01 4.03586268e-01 4.13402110e-01
4.33803856e-01 -1.00331701e-01 6.72668964e-02 -4.06727642e-01
1.95325300e-01 -5.88026166e-01 -3.77560735e-01 -2.65309632e-01
-1.09019242e-01 1.86997801e-01 5.62317491e-01 -4.03351247e-01
-7.32503414e-01 -3.60296704e-02 -4.62308638e-02 -1.26829934e+00
3.50501239e-01 2.93623745e-01 -1.20619675e-02 1.36186361e-01
4.78031039e-01 1.64890304e-01 9.16498378e-02 -6.32806182e-01
4.60100889e-01 3.26813012e-01 8.15702617e-01 -1.90846831e-01
7.38428712e-01 7.30926514e-01 -1.59531072e-01 -5.73219299e-01
-4.38928157e-01 -7.09789515e-01 -3.54944795e-01 -1.67293817e-01
1.10974967e+00 -1.52687716e+00 -1.02333128e+00 4.67142522e-01
-8.76193404e-01 -4.00014549e-01 -2.01735258e-01 4.65024024e-01
-6.02355957e-01 3.86044592e-01 -1.10017598e+00 -1.37237027e-01
-4.80024725e-01 -1.55095029e+00 1.23136973e+00 6.70596585e-02
-1.90406740e-01 -3.64456445e-01 -4.19771463e-01 5.41665137e-01
3.79530698e-01 4.86007705e-02 4.76458341e-01 5.84901264e-03
-7.79102445e-01 -8.58208258e-03 -6.50494814e-01 4.09022778e-01
6.21006973e-02 -3.42041612e-01 -1.05677366e+00 -3.29102665e-01
-1.79541498e-01 -2.12670624e-01 1.25204456e+00 2.18577325e-01
1.14041114e+00 -3.34605604e-01 -1.57836810e-01 9.37180102e-01
1.49512053e+00 -7.83441588e-02 6.89917564e-01 2.65015364e-01
9.80008185e-01 5.26269495e-01 4.48008448e-01 3.82115632e-01
2.95146555e-01 7.00171709e-01 6.62954211e-01 1.95695329e-02
-3.27582210e-01 -1.57223135e-01 8.33412230e-01 8.94137442e-01
-3.33092123e-01 -1.71191037e-01 -6.78977966e-01 6.64995730e-01
-1.98173153e+00 -9.24272239e-01 1.72698915e-01 1.84706843e+00
8.32296848e-01 1.70700148e-01 4.21141349e-02 1.24406807e-01
4.71664667e-01 3.86187375e-01 -3.79329920e-01 -1.99388713e-01
1.18749082e-01 3.97105604e-01 9.59349513e-01 3.72007191e-01
-1.07476783e+00 1.25665843e+00 5.70020342e+00 8.15989316e-01
-1.19925845e+00 2.03500018e-01 5.61875939e-01 -9.13948953e-01
3.23005170e-02 -4.51048821e-01 -7.06056356e-01 5.10075867e-01
1.10106468e+00 -4.51444425e-02 5.91490149e-01 1.00810874e+00
3.50876451e-01 -1.20433107e-01 -1.23236310e+00 1.29679453e+00
-9.52601507e-02 -1.62446320e+00 5.98788261e-02 8.25125817e-03
4.22602862e-01 4.53242034e-01 -6.59620911e-02 3.40714961e-01
1.16645649e-01 -1.01853406e+00 1.03983855e+00 3.52454394e-01
9.39501762e-01 -8.18242192e-01 7.90594816e-01 -8.21487755e-02
-1.58147001e+00 3.27449734e-03 -5.69908977e-01 -3.25843364e-01
3.00052464e-02 2.76296318e-01 -4.69005555e-01 3.14099371e-01
1.25014722e+00 9.86650348e-01 -4.33477700e-01 5.80803275e-01
-2.30091617e-01 3.87681037e-01 -4.56575245e-01 1.23208322e-01
5.33142805e-01 2.66573131e-01 2.69861102e-01 1.17169607e+00
1.07340574e-01 1.64805219e-01 -5.73784439e-03 4.56444472e-01
-6.39840007e-01 -2.02016637e-01 -4.40751463e-01 -3.46449912e-02
2.35623568e-01 1.03164732e+00 -6.13497734e-01 -7.82066703e-01
-3.95052284e-01 1.32936776e+00 2.29491517e-01 4.05110985e-01
-1.14721191e+00 -3.50663394e-01 1.32464921e+00 1.84603900e-01
5.88633716e-01 -1.50546700e-01 -3.26806456e-01 -1.41788661e+00
1.58074126e-01 -6.87772095e-01 2.24958867e-01 -9.66858983e-01
-8.63021672e-01 5.07856846e-01 1.07214861e-01 -1.15941358e+00
-1.59720615e-01 -6.64669156e-01 -1.96066901e-01 8.26392770e-01
-1.36127758e+00 -7.74439096e-01 -6.38515174e-01 8.80804062e-01
8.26704681e-01 1.73670158e-01 4.30732995e-01 5.26606500e-01
-5.60342133e-01 4.89197731e-01 -6.33099973e-02 4.01844740e-01
5.00977397e-01 -9.17172372e-01 7.49327719e-01 8.63929749e-01
-1.00588500e-01 7.18823791e-01 4.13562357e-01 -5.48918843e-01
-1.55248594e+00 -1.22123981e+00 6.43162847e-01 -2.44013727e-01
7.06738710e-01 -4.69369799e-01 -7.84665823e-01 8.33466589e-01
2.25791372e-02 3.00523132e-01 2.49200165e-01 -2.12479874e-01
-6.70165956e-01 -2.17328444e-01 -1.06045473e+00 7.85499215e-01
1.31258905e+00 -7.83692241e-01 -6.59080565e-01 1.42873302e-01
1.18917966e+00 -5.66306233e-01 -1.20329154e+00 2.29804352e-01
4.75277811e-01 -8.79332304e-01 1.22241032e+00 -3.66870105e-01
8.48323941e-01 -2.52723634e-01 -3.61177415e-01 -8.42570662e-01
-3.92612606e-01 -4.05734897e-01 -1.02395229e-01 8.48585784e-01
5.36732301e-02 -3.24385285e-01 9.97449875e-01 3.70048076e-01
-1.63700148e-01 -6.07239366e-01 -1.24284065e+00 -6.02663338e-01
-3.33812565e-01 -8.03291380e-01 1.39577970e-01 5.69886684e-01
2.11702824e-01 2.43694648e-01 -3.57801527e-01 1.88465491e-01
5.16826391e-01 -2.77223080e-01 6.57298565e-01 -5.09051919e-01
-3.24737579e-01 -4.06208426e-01 -7.31088877e-01 -1.28396642e+00
-1.99082419e-01 -6.89379454e-01 -9.51202363e-02 -1.40188324e+00
1.65507317e-01 1.55905724e-01 -4.74738061e-01 9.60901618e-01
-1.45113409e-01 7.17162192e-01 5.89398265e-01 3.73706341e-01
-7.24695444e-01 4.82469589e-01 9.58069801e-01 -2.21672684e-01
-3.69856395e-02 -5.42014182e-01 -5.36273599e-01 8.12087357e-01
5.36673844e-01 -5.04587114e-01 -5.79389751e-01 -8.38353932e-01
4.66253757e-02 1.26609966e-01 4.63344127e-01 -1.24976695e+00
7.91754872e-02 2.24649161e-01 5.33852756e-01 -5.02025545e-01
7.25015223e-01 -6.66134179e-01 -9.07594264e-02 5.71895540e-01
-1.87200814e-01 3.44242126e-01 6.53969288e-01 6.69621289e-01
-3.08100671e-01 2.03401625e-01 5.87717116e-01 -3.76304835e-01
-1.10961008e+00 2.38398224e-01 -6.04852498e-01 -3.36581990e-02
8.51821601e-01 -3.44654173e-01 -3.72291207e-01 -3.85955662e-01
-7.02429950e-01 2.14180693e-01 5.21681309e-01 5.40078104e-01
8.91711831e-01 -1.12399757e+00 -3.53700578e-01 1.31937310e-01
-2.41733193e-02 2.58566499e-01 3.27452004e-01 7.39112735e-01
-9.19043422e-01 3.41251642e-01 -5.24992704e-01 -6.72470987e-01
-1.53047180e+00 4.37997639e-01 1.35662824e-01 -2.54153371e-01
-9.51600432e-01 1.19276118e+00 4.73388493e-01 2.35897496e-01
4.17345315e-01 -9.05227721e-01 -4.86493744e-02 1.67896986e-01
5.19106567e-01 4.74456608e-01 -3.96717116e-02 -5.45160770e-01
-5.98125815e-01 5.70485473e-01 5.33684641e-02 -3.75786424e-01
1.53126967e+00 -1.80742089e-02 3.27965394e-02 2.25389972e-01
1.42267895e+00 -2.09481075e-01 -1.70537233e+00 -7.77619779e-02
-2.23981962e-01 -4.25588816e-01 1.94804072e-01 -3.53196234e-01
-1.52661693e+00 1.10445082e+00 4.41529423e-01 -2.83382714e-01
1.30704594e+00 6.60078926e-03 8.16140592e-01 2.86069661e-01
4.25364941e-01 -7.53101468e-01 3.53080541e-01 2.32083753e-01
7.54535079e-01 -8.95795226e-01 7.30533078e-02 -4.77130353e-01
-7.84799159e-01 9.78778481e-01 5.68203330e-01 -2.51214325e-01
1.56845391e-01 3.10831815e-01 -1.72968745e-01 -2.78639108e-01
-9.24979508e-01 -3.95215629e-03 1.50061622e-01 1.82361782e-01
1.49900556e-01 -7.27063939e-02 1.88487142e-01 4.71052140e-01
-9.22399685e-02 4.53281924e-02 4.50602472e-01 9.86226737e-01
-4.94087189e-01 -5.85147500e-01 -1.71238810e-01 3.63770396e-01
-5.66589594e-01 -4.45418149e-01 -5.44109643e-02 7.99823046e-01
4.43403376e-03 9.68712687e-01 4.22397673e-01 -7.35049963e-01
3.40096802e-01 -8.65938738e-02 2.73364246e-01 -6.17516816e-01
-7.88402200e-01 2.53794014e-01 1.33689657e-01 -1.25400627e+00
-4.70540822e-01 -3.75385582e-01 -1.44996786e+00 -7.88584411e-01
2.26332456e-01 -2.82742858e-01 5.46102941e-01 7.90480614e-01
5.66053927e-01 7.56307602e-01 2.04995960e-01 -1.11366570e+00
-2.62102544e-01 -8.49806726e-01 -2.59440988e-01 3.95194799e-01
1.38543844e-01 -6.48049355e-01 -2.20809445e-01 1.64978862e-01]
|
[9.077914237976074, 0.5086672902107239]
|
cddec456-1545-4f83-8fe1-4ab7ccbf5824
|
learning-when-to-trust-which-teacher-for
|
2306.12012
| null |
https://arxiv.org/abs/2306.12012v1
|
https://arxiv.org/pdf/2306.12012v1.pdf
|
Learning When to Trust Which Teacher for Weakly Supervised ASR
|
Automatic speech recognition (ASR) training can utilize multiple experts as teacher models, each trained on a specific domain or accent. Teacher models may be opaque in nature since their architecture may be not be known or their training cadence is different from that of the student ASR model. Still, the student models are updated incrementally using the pseudo-labels generated independently by the expert teachers. In this paper, we exploit supervision from multiple domain experts in training student ASR models. This training strategy is especially useful in scenarios where few or no human transcriptions are available. To that end, we propose a Smart-Weighter mechanism that selects an appropriate expert based on the input audio, and then trains the student model in an unsupervised setting. We show the efficacy of our approach using LibriSpeech and LibriLight benchmarks and find an improvement of 4 to 25\% over baselines that uniformly weight all the experts, use a single expert model, or combine experts using ROVER.
|
['Andreas Stolcke', 'Gopinath Chennupati', 'Anit Kumar Sahu', 'Milind Rao', 'Aakriti Agrawal']
|
2023-06-21
| null | null | null | null |
['automatic-speech-recognition']
|
['speech']
|
[ 3.40295553e-01 2.79762924e-01 -2.86630094e-01 -6.32650137e-01
-1.05726635e+00 -9.67892170e-01 5.20090699e-01 -8.98702145e-02
-5.63423872e-01 6.00829959e-01 2.37796456e-01 -3.99445027e-01
1.72133908e-01 -2.85588443e-01 -6.42673969e-01 -5.71956694e-01
4.70635831e-01 9.36111271e-01 3.60830694e-01 -3.04083049e-01
-1.24839313e-01 3.96941692e-01 -1.29909420e+00 2.10220665e-01
9.58099365e-01 8.52709591e-01 3.19125146e-01 9.32729006e-01
-1.31404176e-01 8.04826140e-01 -9.61900473e-01 -2.38684058e-01
9.69314598e-04 -3.05021167e-01 -8.09440017e-01 3.48048091e-01
4.13751572e-01 -1.14686467e-01 -4.30459142e-01 8.73702407e-01
5.40159166e-01 4.76207823e-01 5.57209849e-01 -7.06736147e-01
-3.95909131e-01 9.68264341e-01 -4.47900385e-01 2.40571752e-01
1.76843837e-01 -1.04344133e-02 8.53031635e-01 -8.39331448e-01
2.94894367e-01 1.09841025e+00 3.56350869e-01 8.80421877e-01
-1.36767101e+00 -7.04915404e-01 6.01375103e-01 1.66570365e-01
-1.29965246e+00 -1.01651239e+00 7.49910295e-01 -1.96069181e-01
7.06787646e-01 2.25285381e-01 1.02927998e-01 1.34063816e+00
-6.70415044e-01 9.63492572e-01 1.05953717e+00 -5.46740949e-01
2.47655168e-01 5.16031027e-01 3.40802610e-01 3.87518317e-01
-2.87553817e-01 -7.70777762e-02 -7.22688496e-01 -2.04924807e-01
3.87834013e-01 -2.86984712e-01 -4.52833861e-01 -3.85425910e-02
-9.77066040e-01 6.71140909e-01 -7.79787451e-02 2.62797922e-01
-2.75020570e-01 -1.20439127e-01 9.89252403e-02 6.23675823e-01
4.90424722e-01 5.01434803e-01 -7.77115107e-01 -2.75816262e-01
-1.15021753e+00 -2.04742819e-01 7.27932274e-01 8.66338968e-01
6.99138463e-01 3.92190486e-01 -2.54712719e-02 1.35300136e+00
3.08700532e-01 3.99519563e-01 8.04205537e-01 -1.00717998e+00
4.35066104e-01 2.00854182e-01 1.12984218e-01 -1.80481732e-01
-3.99344116e-02 -5.87309062e-01 -5.29700458e-01 7.24795908e-02
2.63830006e-01 -3.59072417e-01 -1.36501002e+00 1.60842800e+00
2.84666151e-01 6.51101232e-01 2.42108598e-01 7.97612965e-01
7.78577626e-01 7.51175523e-01 -8.42480455e-03 -1.57321930e-01
1.14092445e+00 -1.13623512e+00 -5.81009448e-01 -5.90826631e-01
3.36504430e-01 -8.52278888e-01 1.00653720e+00 8.28989863e-01
-1.01547766e+00 -6.41892791e-01 -9.40469205e-01 2.07114235e-01
-1.45419866e-01 5.30659914e-01 -4.54625450e-02 7.90924072e-01
-1.00015104e+00 2.91367292e-01 -8.82878065e-01 -1.43665105e-01
1.01991007e-02 4.09555197e-01 -1.22714959e-01 7.28402659e-02
-1.09545791e+00 8.54647398e-01 2.57280201e-01 -9.52821672e-02
-1.17216384e+00 -5.83068013e-01 -6.86911941e-01 1.09239385e-01
4.74683851e-01 -2.42937490e-01 1.78846073e+00 -1.32853162e+00
-2.14780068e+00 6.35064542e-01 -1.58774585e-01 -5.00439823e-01
3.47953379e-01 -1.75953001e-01 -4.05299693e-01 1.11236498e-01
-3.89165908e-01 3.76397669e-01 1.02583301e+00 -1.36203277e+00
-7.12040484e-01 -2.03329355e-01 2.06150711e-01 3.29772562e-01
-4.88177031e-01 1.88436300e-01 -4.93112743e-01 -7.47583330e-01
3.59335006e-03 -9.82743442e-01 -3.14688474e-01 -6.03779435e-01
-2.56272078e-01 -3.32477957e-01 8.23201060e-01 -6.03752494e-01
1.38521028e+00 -2.32793045e+00 1.02308959e-01 5.09662569e-01
1.73708111e-01 5.74160933e-01 -1.90218583e-01 1.69025332e-01
-1.91191182e-01 6.04463108e-02 -1.25314683e-01 -4.26522553e-01
-4.19956669e-02 2.88713902e-01 -4.94232982e-01 1.06705330e-01
1.07647441e-01 1.80337891e-01 -8.97160292e-01 -3.57323736e-01
5.66567518e-02 4.15542752e-01 -3.59816402e-01 6.46090269e-01
-2.13608652e-01 4.34482127e-01 -5.04969835e-01 2.62082785e-01
1.69574618e-01 -2.26869315e-01 4.00702834e-01 1.04964010e-01
5.69623411e-02 8.08418095e-01 -1.48168242e+00 1.46634769e+00
-8.67466033e-01 6.45477355e-01 5.19655764e-01 -1.15043592e+00
1.15441906e+00 8.55612636e-01 1.88157260e-01 -1.53058350e-01
-1.79503292e-01 2.66756743e-01 1.53860688e-01 -4.02747728e-02
3.70386720e-01 -1.13268204e-01 2.44433627e-01 6.10465944e-01
3.59771580e-01 -2.27679670e-01 -8.94803647e-03 1.07758120e-01
1.26737666e+00 -1.32520005e-01 1.07639410e-01 5.57221621e-02
5.55029213e-01 -2.47643143e-01 5.80874443e-01 8.49105477e-01
-7.52283558e-02 7.54821301e-01 1.14870280e-01 -1.80251732e-01
-6.07338309e-01 -1.00013471e+00 -3.50123681e-02 1.52315247e+00
-2.24028736e-01 -4.37262625e-01 -6.77346587e-01 -9.12737608e-01
-3.77957910e-01 8.64758015e-01 -3.17074507e-01 -1.57239720e-01
-8.41177523e-01 -3.59099358e-01 5.59291780e-01 4.94195163e-01
6.98717833e-02 -9.30472493e-01 -2.79628187e-01 3.57344270e-01
-7.12561160e-02 -1.15968823e+00 -6.94378912e-01 6.43597960e-01
-5.96709132e-01 -6.71634614e-01 -7.62602866e-01 -7.32087553e-01
6.61686540e-01 1.07922405e-01 1.15777004e+00 -4.52902615e-02
4.17637467e-01 4.96420890e-01 -4.53051925e-01 -3.05647552e-01
-7.39349067e-01 5.24171889e-01 2.64171183e-01 2.59103864e-01
3.06176305e-01 -6.32247746e-01 -2.77850747e-01 4.36284363e-01
-6.32692337e-01 -1.60193652e-01 4.88402903e-01 9.07748282e-01
3.01238030e-01 1.38951659e-01 8.29765499e-01 -1.16920435e+00
5.33158243e-01 -3.95487130e-01 -5.55932522e-01 4.37935680e-01
-5.48345745e-01 1.37748986e-01 8.52405548e-01 -8.39897096e-01
-1.22826815e+00 2.93628037e-01 -3.13136101e-01 -7.19062805e-01
-4.45421249e-01 5.12288988e-01 -1.72143787e-01 1.89818040e-01
7.39172280e-01 7.95842484e-02 -3.88018876e-01 -6.97895348e-01
2.48190150e-01 1.06405544e+00 4.23359364e-01 -9.68645334e-01
9.68205690e-01 -1.11408509e-01 -8.55476677e-01 -8.40168595e-01
-9.94104862e-01 -5.18813312e-01 -4.82742697e-01 -8.78203362e-02
3.80435824e-01 -1.11032712e+00 -3.11070204e-01 2.79344559e-01
-1.20724261e+00 -6.11605525e-01 -3.06805193e-01 5.83143413e-01
-2.68117934e-01 7.12945163e-02 -5.36942542e-01 -9.86309528e-01
-1.08497269e-01 -1.42537093e+00 8.45052004e-01 2.75474817e-01
-4.59197611e-01 -8.89050961e-01 6.18022531e-02 5.12704372e-01
3.91590416e-01 -5.92563510e-01 6.54555202e-01 -1.40741730e+00
-3.12069178e-01 -4.86995019e-02 2.89415270e-01 5.73688328e-01
2.32646540e-01 1.36894614e-01 -1.33917332e+00 -2.70908892e-01
-2.24773824e-01 -5.97012281e-01 7.25460291e-01 6.64777830e-02
1.27495801e+00 -2.55283743e-01 -1.53524250e-01 1.59892604e-01
7.87083566e-01 3.00219387e-01 2.28036821e-01 1.06328852e-01
5.70948541e-01 5.72521031e-01 2.40953535e-01 1.99662879e-01
2.93574840e-01 8.32534194e-01 -6.44808114e-02 -4.52199727e-02
-1.56010062e-01 -2.37722784e-01 8.02577496e-01 1.26323795e+00
1.11516610e-01 -3.24714780e-01 -9.98787105e-01 7.23994136e-01
-1.47570157e+00 -7.21758366e-01 3.44551504e-01 2.26325488e+00
1.38658345e+00 3.10724586e-01 2.70209908e-01 5.84001429e-02
7.18356252e-01 2.30803490e-01 -4.12122756e-01 -3.65236729e-01
9.85745266e-02 6.29864216e-01 3.66265893e-01 7.09023058e-01
-9.76075411e-01 9.63564992e-01 6.24257946e+00 9.13276970e-01
-1.29220426e+00 1.48890063e-01 6.15365624e-01 3.92163433e-02
-3.24327856e-01 1.66415628e-02 -9.27151263e-01 3.73206109e-01
1.31384540e+00 1.57681517e-02 6.31622136e-01 8.13024044e-01
2.53271937e-01 1.18002072e-01 -1.09160149e+00 7.22453177e-01
2.13425979e-02 -8.33356917e-01 -2.36475289e-01 -1.37085453e-01
7.08152533e-01 1.58930838e-01 1.53007984e-01 4.88025665e-01
9.58473563e-01 -9.31582391e-01 6.20226204e-01 1.66845128e-01
5.70213079e-01 -7.44678020e-01 4.30101395e-01 5.79103172e-01
-9.24685717e-01 1.29669067e-02 -7.99602345e-02 3.52104604e-01
-6.70458656e-03 3.38793844e-01 -1.26645458e+00 2.67710745e-01
5.50650835e-01 1.87050894e-01 -4.35880244e-01 8.73426497e-01
-4.84903514e-01 1.54267848e+00 -4.68108445e-01 1.28800914e-01
1.36105850e-01 -3.46437916e-02 5.59690475e-01 1.32320464e+00
2.30938718e-01 1.33162871e-01 5.88827312e-01 3.86391222e-01
-2.96431184e-01 2.31017321e-02 -2.78525174e-01 -1.92178875e-01
7.63686299e-01 1.21084058e+00 -4.10067916e-01 -4.77025270e-01
-5.04211187e-01 8.36828470e-01 3.24107885e-01 6.96151555e-01
-5.92665076e-01 -4.78769004e-01 5.08960426e-01 1.64195240e-01
4.13554788e-01 -1.66632116e-01 2.85111666e-02 -1.16752696e+00
-1.74778059e-01 -1.42218196e+00 3.13115001e-01 -6.84482276e-01
-1.16729176e+00 8.83987069e-01 -1.09393321e-01 -1.00898755e+00
-5.49247980e-01 -5.08233547e-01 -7.03626931e-01 8.76742840e-01
-1.50511622e+00 -7.87698805e-01 3.05159427e-02 4.92921948e-01
9.28303957e-01 -5.10459602e-01 8.74810159e-01 2.74568200e-01
-6.44999146e-01 8.79825830e-01 1.68043934e-02 3.64304572e-01
9.17986274e-01 -1.56297410e+00 2.72700548e-01 7.54499376e-01
6.66942596e-01 6.48358166e-01 7.86219180e-01 -2.00191244e-01
-9.82558906e-01 -8.66683483e-01 8.20954204e-01 -5.70579171e-01
8.03806603e-01 -3.26382518e-01 -1.15663373e+00 8.34485650e-01
3.50230098e-01 -5.92252165e-02 8.20328832e-01 4.69968081e-01
-4.47151035e-01 -3.31445545e-01 -8.04947674e-01 5.04745245e-01
5.19434690e-01 -6.94656968e-01 -7.18627453e-01 1.72199711e-01
7.61806011e-01 -6.17504120e-01 -7.34503269e-01 2.18487203e-01
2.93578744e-01 -4.64706600e-01 6.71365499e-01 -6.57414496e-01
-8.16000402e-02 -4.24833447e-01 -9.55006108e-02 -1.73198032e+00
5.51597774e-03 -7.97909260e-01 -2.27606848e-01 1.54736781e+00
7.77743816e-01 -4.81696010e-01 7.56465018e-01 5.60385942e-01
-3.09901923e-01 -4.00193244e-01 -8.92910182e-01 -6.63712442e-01
-1.87828764e-01 -4.09399450e-01 5.73268056e-01 1.04804754e+00
-2.87157774e-01 7.37541020e-01 -2.57768273e-01 4.33958173e-01
1.35869354e-01 -1.90947950e-01 7.59640038e-01 -1.19203448e+00
-7.27096736e-01 -2.60653496e-01 5.38950674e-02 -1.37899446e+00
5.95839679e-01 -7.23151922e-01 3.93004179e-01 -1.00507498e+00
-2.66922534e-01 -6.31308436e-01 -5.91086328e-01 7.81431496e-01
-2.51630753e-01 -1.14279725e-01 4.51789200e-02 -1.00313641e-01
-6.76775992e-01 4.48598593e-01 7.56660521e-01 -3.28005224e-01
-4.16213006e-01 5.00208080e-01 -6.69597268e-01 7.53843069e-01
7.83526778e-01 -5.81163704e-01 -7.28014171e-01 -5.32983243e-01
-2.47918889e-01 2.33700797e-02 -1.32493451e-02 -8.73335123e-01
4.89499003e-01 -2.28092462e-01 1.86187357e-01 -1.64368540e-01
4.71128881e-01 -9.33202446e-01 -7.06255734e-02 -2.20928758e-01
-6.14061415e-01 -1.13579519e-01 3.65031995e-02 3.80384028e-01
-3.51979226e-01 -6.52996659e-01 9.36340868e-01 1.02714757e-02
-4.11020517e-01 1.15573362e-01 -5.69438338e-01 1.54597670e-01
4.28471982e-01 -2.65336521e-02 -1.43888950e-01 -6.21435165e-01
-7.87711024e-01 2.74533510e-01 2.59215117e-01 4.26394284e-01
5.43203354e-01 -1.01545906e+00 -6.52463377e-01 8.30573291e-02
4.53765765e-02 1.80199340e-01 -8.40781331e-02 5.07528543e-01
1.00027211e-01 2.79016346e-02 4.20086414e-01 -4.06372666e-01
-1.29763401e+00 2.38586724e-01 4.11092222e-01 -3.63518447e-01
-1.89571872e-01 9.27929103e-01 1.75718039e-01 -8.31859410e-01
6.35457397e-01 -1.20477125e-01 -2.91112036e-01 6.93634525e-02
4.86150324e-01 1.55234590e-01 2.72420108e-01 -6.69711769e-01
-2.47614786e-01 2.92095691e-01 -3.88720632e-01 -4.75376248e-01
1.27608836e+00 5.69108017e-02 4.48432684e-01 7.41305232e-01
7.31715381e-01 5.09486020e-01 -1.22848821e+00 -6.00084901e-01
2.62518138e-01 -1.20906122e-01 9.41697136e-02 -9.99710560e-01
-1.01010418e+00 8.98285508e-01 4.31435585e-01 2.51624882e-01
1.20619977e+00 1.34643480e-01 6.22341514e-01 5.60394108e-01
6.22801743e-02 -1.36695313e+00 1.89282894e-01 6.54515088e-01
6.08789742e-01 -1.10030890e+00 -2.73365974e-01 -1.98935091e-01
-7.85174787e-01 1.01667821e+00 8.65234673e-01 1.42553180e-01
5.83988428e-01 2.88222730e-01 5.98762035e-01 2.15387225e-01
-1.12404346e+00 -2.89326966e-01 3.90048742e-01 5.09646475e-01
6.37167871e-01 1.21974505e-01 1.90642625e-01 8.42597067e-01
-2.55511850e-01 -3.21009129e-01 4.66152370e-01 7.32248247e-01
-5.46729207e-01 -1.51396453e+00 -4.36936080e-01 2.81838804e-01
-5.04918873e-01 -1.16776980e-01 -5.10439277e-01 3.16938728e-01
-1.17271692e-01 1.19764745e+00 -1.91624343e-01 -3.76401305e-01
4.65290874e-01 4.33310330e-01 2.21532404e-01 -1.04506826e+00
-7.74946749e-01 4.10045773e-01 2.87047744e-01 3.29933725e-02
-3.01948547e-01 -5.39129972e-01 -1.07017910e+00 2.10596859e-01
-3.96305770e-01 6.07991219e-01 4.87343132e-01 1.08352613e+00
2.24203318e-01 5.35702348e-01 9.64437246e-01 -4.07676935e-01
-8.31266105e-01 -1.06430924e+00 -3.39108825e-01 -9.45935249e-02
4.99284655e-01 -5.16402543e-01 -4.54386443e-01 3.59562457e-01]
|
[14.451835632324219, 6.628710746765137]
|
7569dd98-a948-4680-bcfc-f67bf499832a
|
limits-of-machine-learning-for-automatic
|
2306.17193
| null |
https://arxiv.org/abs/2306.17193v1
|
https://arxiv.org/pdf/2306.17193v1.pdf
|
Limits of Machine Learning for Automatic Vulnerability Detection
|
Recent results of machine learning for automatic vulnerability detection have been very promising indeed: Given only the source code of a function $f$, models trained by machine learning techniques can decide if $f$ contains a security flaw with up to 70% accuracy. But how do we know that these results are general and not specific to the datasets? To study this question, researchers proposed to amplify the testing set by injecting semantic preserving changes and found that the model's accuracy significantly drops. In other words, the model uses some unrelated features during classification. In order to increase the robustness of the model, researchers proposed to train on amplified training data, and indeed model accuracy increased to previous levels. In this paper, we replicate and continue this investigation, and provide an actionable model benchmarking methodology to help researchers better evaluate advances in machine learning for vulnerability detection. Specifically, we propose (i) a cross validation algorithm, where a semantic preserving transformation is applied during the amplification of either the training set or the testing set, and (ii) the amplification of the testing set with code snippets where the vulnerabilities are fixed. Using 11 transformations, 3 ML techniques, and 2 datasets, we find that the improved robustness only applies to the specific transformations used during training data amplification. In other words, the robustified models still rely on unrelated features for predicting the vulnerabilities in the testing data. Additionally, we find that the trained models are unable to generalize to the modified setting which requires to distinguish vulnerable functions from their patches.
|
['Marcel Böhme', 'Niklas Risse']
|
2023-06-28
| null | null | null | null |
['vulnerability-detection', 'benchmarking', 'benchmarking']
|
['miscellaneous', 'miscellaneous', 'robots']
|
[ 3.97723317e-01 1.48658678e-01 -1.36738479e-01 -2.19603881e-01
-8.41645181e-01 -1.09198201e+00 2.43169278e-01 3.45326602e-01
-1.01759166e-01 4.08112943e-01 -3.56834441e-01 -9.29987490e-01
-3.04711387e-02 -1.01008642e+00 -8.59044254e-01 -3.91400009e-01
-1.56215623e-01 -3.20114605e-02 3.21778148e-01 -4.38949287e-01
6.88271821e-01 3.20493937e-01 -1.57277548e+00 5.39425492e-01
7.14484632e-01 7.24069774e-01 -5.21866620e-01 7.19764054e-01
1.27672434e-01 6.53059602e-01 -8.52296472e-01 -5.93323946e-01
5.61970592e-01 -1.97548807e-01 -9.75689530e-01 -3.02143395e-01
4.05516654e-01 -1.05649203e-01 1.70306101e-01 1.19985747e+00
3.06847412e-02 -4.35820878e-01 2.95217514e-01 -1.43356001e+00
-4.74833071e-01 6.89002991e-01 -3.74354064e-01 1.10264875e-01
5.99477708e-01 2.34091669e-01 1.05477667e+00 -4.91403520e-01
4.65793043e-01 9.47033882e-01 8.99647236e-01 6.04048848e-01
-1.21300948e+00 -6.23641670e-01 9.72000733e-02 5.30305654e-02
-1.07216489e+00 -2.68379956e-01 7.11810291e-01 -5.58158994e-01
1.25302792e+00 4.49593782e-01 3.68749380e-01 9.85024154e-01
2.46414900e-01 8.62577856e-02 1.15438223e+00 -6.88714147e-01
3.34880501e-01 6.58382595e-01 3.91229331e-01 7.50947952e-01
5.01665592e-01 4.42512751e-01 1.09055907e-01 -6.30747616e-01
-3.93665805e-02 -2.41817772e-01 -9.49960276e-02 -1.89639479e-01
-6.29927516e-01 1.15383315e+00 1.20210424e-01 5.80018640e-01
1.39529154e-01 -2.31050313e-01 6.23519421e-01 6.80758655e-01
2.64265597e-01 9.64772761e-01 -9.21495616e-01 -5.97094484e-02
-8.34445238e-01 9.92998034e-02 8.45002830e-01 4.58515674e-01
9.46913600e-01 4.35846224e-02 5.21529496e-01 3.73789579e-01
5.49862124e-02 2.16018528e-01 4.09171373e-01 -7.22772598e-01
6.01497889e-01 1.03173637e+00 -2.36591041e-01 -1.03985822e+00
-2.34824672e-01 -4.19926286e-01 -1.34752497e-01 6.66270375e-01
5.83508909e-01 -1.91790506e-01 -6.73483431e-01 1.86774468e+00
6.92024902e-02 -9.26780328e-02 2.02128291e-01 1.26215667e-01
8.10444504e-02 3.22389781e-01 -5.05185574e-02 1.15005024e-01
9.41054523e-01 -4.71222341e-01 2.76935585e-02 -2.90089756e-01
1.40601599e+00 -5.96345663e-01 1.36697030e+00 5.47377408e-01
-8.75094891e-01 -5.99405050e-01 -1.49396968e+00 4.99732137e-01
-8.22609603e-01 -2.45444521e-01 4.68972176e-01 1.45145273e+00
-1.16941369e+00 7.47873783e-01 -5.08641720e-01 -2.85876095e-01
1.89042032e-01 4.79405910e-01 -5.16137719e-01 8.16945732e-03
-1.22581422e+00 9.00567472e-01 3.71442616e-01 -4.30229455e-01
-7.72203505e-01 -7.21038640e-01 -8.68478239e-01 1.10240012e-01
1.97896183e-01 -2.99554430e-02 7.12547004e-01 -1.27038670e+00
-9.35225904e-01 9.15251553e-01 7.69713745e-02 -4.12655175e-01
2.55980432e-01 1.80750415e-01 -5.89179575e-01 -7.68881366e-02
-2.26141233e-02 1.09532759e-01 9.09348369e-01 -1.20272052e+00
-5.49282253e-01 -4.14552391e-01 5.09044170e-01 -6.63042903e-01
-6.43100858e-01 3.25247139e-01 2.34683499e-01 -4.12422568e-01
-2.18675479e-01 -1.00966787e+00 1.15476258e-01 -5.12216747e-01
-3.95995140e-01 2.02795193e-01 9.01313841e-01 -8.53181243e-01
1.41726983e+00 -2.37075543e+00 -1.93169281e-01 7.30555356e-01
6.24542311e-02 5.69817305e-01 -3.62770975e-01 1.85761929e-01
-7.79142559e-01 7.46532142e-01 -7.07318306e-01 7.77519494e-02
-1.30666316e-01 -1.57016054e-01 -6.05171084e-01 3.22957754e-01
5.02955139e-01 5.17269492e-01 -5.27782023e-01 -1.40118465e-01
-1.20500967e-01 8.84626880e-02 -9.16713655e-01 4.75101359e-02
-1.08547479e-01 -6.22871071e-02 -2.86130518e-01 8.51329088e-01
6.15619719e-01 8.46497789e-02 1.25691175e-01 1.29463658e-01
2.31163041e-03 1.47023469e-01 -9.69272137e-01 9.58402216e-01
-4.73129839e-01 5.01306951e-01 -1.30388528e-01 -1.21766520e+00
9.88280833e-01 1.68155909e-01 6.52395263e-02 -5.98014235e-01
-8.28361958e-02 3.80203336e-01 4.28345531e-01 -4.94632453e-01
1.67097226e-02 4.03737426e-02 -2.47794211e-01 6.54397786e-01
-1.74940020e-01 -2.64609288e-02 4.33913730e-02 5.38689122e-02
1.57035065e+00 -5.01816766e-03 1.37992546e-01 -1.68606594e-01
7.89488196e-01 1.78531095e-01 5.03271818e-01 7.73421705e-01
-1.24859251e-01 3.15757483e-01 9.08618629e-01 -3.95517379e-01
-1.04216003e+00 -8.82853448e-01 -1.56299010e-01 9.21792448e-01
-3.96846026e-01 -4.65858638e-01 -1.15855777e+00 -1.26288676e+00
3.20292749e-02 8.55022430e-01 -9.21070814e-01 -7.37844884e-01
-6.72904015e-01 -9.06926632e-01 9.05018628e-01 4.40243900e-01
2.94482350e-01 -8.32220554e-01 -5.86601496e-01 -6.98368996e-02
8.41600001e-02 -6.56988382e-01 -1.04390547e-01 2.05514446e-01
-8.44926476e-01 -1.48308611e+00 1.83360353e-01 -6.12885714e-01
6.93390727e-01 -2.23491713e-01 8.31541121e-01 7.13060319e-01
-3.37370396e-01 2.90551066e-01 -4.94326323e-01 -1.54845297e-01
-9.11318481e-01 9.25095230e-02 -1.38611883e-01 -1.98211923e-01
2.79440284e-01 -6.04333460e-01 -6.66143000e-03 4.40762907e-01
-9.81310487e-01 -8.26248169e-01 5.42669296e-01 6.30952537e-01
-7.97407106e-02 4.12852734e-01 6.66895986e-01 -9.70584929e-01
5.42865455e-01 -6.64683580e-01 -5.98348916e-01 4.79017437e-01
-9.11773682e-01 2.61883646e-01 8.23727608e-01 -5.84535539e-01
-6.22728944e-01 1.95721090e-02 -3.46335351e-01 -1.24435900e-02
-7.60373250e-02 4.97876197e-01 -4.74502116e-01 -5.27405202e-01
1.06197214e+00 -4.69802804e-02 -1.67257302e-02 -2.94999957e-01
7.63807967e-02 3.91838461e-01 9.77353528e-02 -7.20869541e-01
1.12599158e+00 1.17383838e-01 -1.47558555e-01 -4.26195055e-01
-2.92117864e-01 1.45788625e-01 -5.04855096e-01 1.89148754e-01
4.77055103e-01 -2.22329333e-01 -4.96447533e-01 3.86425346e-01
-8.74688864e-01 -4.89636302e-01 -1.57188565e-01 9.93920416e-02
-2.17082709e-01 5.42043209e-01 -3.93032074e-01 -7.41887927e-01
-1.29148647e-01 -1.36171925e+00 6.08467281e-01 -1.32883802e-01
-3.27526361e-01 -1.02149355e+00 2.97397494e-01 2.95475245e-01
4.60373223e-01 3.72297496e-01 1.58803153e+00 -1.05592704e+00
-1.10607743e-01 -6.19698405e-01 1.09150335e-01 6.90261900e-01
3.32930088e-01 5.40289700e-01 -1.09806991e+00 -4.40897584e-01
2.44514897e-01 -3.19306135e-01 5.81826687e-01 -3.63287121e-01
1.01829696e+00 -4.27680045e-01 -2.63243377e-01 4.27641034e-01
1.34387136e+00 4.59387541e-01 8.20688844e-01 7.10739493e-01
3.84325743e-01 9.48719442e-01 3.88130277e-01 6.64921030e-02
-5.54185733e-02 5.72054327e-01 5.55061996e-01 1.54205607e-02
2.38588810e-01 -2.74750758e-02 7.33522356e-01 1.67566046e-01
1.97226495e-01 -1.32141151e-02 -1.02154100e+00 3.21595550e-01
-1.23199522e+00 -1.02907205e+00 8.73425305e-02 2.47537875e+00
8.29517126e-01 5.66958427e-01 2.86946744e-01 7.86677241e-01
6.01222575e-01 -1.13381192e-01 -3.67143482e-01 -7.94952929e-01
-1.02622770e-02 5.73346496e-01 2.60734081e-01 6.69575214e-01
-1.01460099e+00 6.65025055e-01 6.55149412e+00 5.07588089e-01
-1.26357961e+00 -7.35080056e-03 6.81815743e-01 1.88658938e-01
-4.56267208e-01 3.69479656e-01 -6.60981476e-01 3.86333674e-01
1.28449261e+00 9.36013684e-02 4.37842995e-01 9.22084391e-01
-2.31759354e-01 1.72519721e-02 -1.13036096e+00 2.14934081e-01
2.17434734e-01 -9.85827148e-01 -5.14147729e-02 6.58274442e-02
4.06140715e-01 -4.98032391e-01 3.70264798e-01 6.48500621e-01
1.26327693e-01 -1.09133744e+00 6.24342620e-01 4.12147462e-01
4.08775240e-01 -8.16477239e-01 8.75333488e-01 2.48064637e-01
-9.50260580e-01 -5.81701517e-01 -1.30987866e-02 -6.27789050e-02
-4.65097010e-01 2.84852862e-01 -1.01497555e+00 3.60991150e-01
7.37213492e-01 2.36747131e-01 -1.27569878e+00 6.09552324e-01
-1.18256502e-01 7.92894185e-01 -9.34537202e-02 2.54847586e-01
-1.52835278e-02 1.83785662e-01 3.72512102e-01 9.90213871e-01
2.80356646e-01 -4.61261749e-01 -1.91776186e-01 9.35144782e-01
3.14972013e-01 -1.01527348e-02 -7.00865149e-01 -1.44750103e-01
2.65813321e-01 1.00837684e+00 -5.61596036e-01 -1.07634291e-01
-4.97414917e-01 4.31925684e-01 3.13064218e-01 2.35086635e-01
-8.57661963e-01 -7.04658985e-01 5.74590445e-01 2.90901750e-01
2.08290040e-01 2.87844893e-02 -4.37778264e-01 -9.80393291e-01
3.33481222e-01 -1.23271585e+00 4.87984478e-01 -4.57906127e-01
-9.86532927e-01 6.69280291e-01 5.99591061e-02 -9.75105941e-01
-3.73028368e-01 -7.30387211e-01 -8.39989007e-01 9.02000427e-01
-1.18808341e+00 -9.07324910e-01 9.58522484e-02 5.09035945e-01
1.53408051e-02 -3.36924106e-01 8.86253774e-01 2.33306020e-01
-6.34598970e-01 1.15757787e+00 -2.55156875e-01 3.98312300e-01
6.89598739e-01 -1.08186233e+00 5.37020147e-01 1.19864023e+00
-3.72210215e-03 9.89808023e-01 6.72844470e-01 -8.07082832e-01
-1.09335434e+00 -1.00602794e+00 6.77669227e-01 -8.78794849e-01
7.86559701e-01 -4.57637459e-01 -1.29448438e+00 7.43547380e-01
-1.58343822e-01 -9.18800533e-02 7.55982757e-01 1.36383651e-02
-9.75927413e-01 -1.18411612e-02 -1.66752422e+00 3.03014159e-01
6.12227738e-01 -6.95685983e-01 -5.85506797e-01 8.29717442e-02
8.20786297e-01 8.52712616e-02 -8.34432960e-01 5.51088452e-01
5.50917923e-01 -1.14764643e+00 9.05121148e-01 -9.29211974e-01
3.67437840e-01 -2.53278106e-01 -4.72366244e-01 -9.73163307e-01
4.23832610e-03 -1.45912245e-01 3.40371802e-02 1.40006495e+00
1.01323140e+00 -1.11995888e+00 7.31218815e-01 7.62882471e-01
8.74595717e-03 -7.11408317e-01 -6.85730398e-01 -7.72742212e-01
5.32275915e-01 -5.12454391e-01 7.82970846e-01 1.10947299e+00
3.31114769e-01 -2.92647839e-01 1.44872531e-01 3.62404644e-01
3.86364430e-01 -8.75174329e-02 5.25778830e-01 -1.16394007e+00
-7.06437588e-01 -4.85680759e-01 -5.24766088e-01 7.61255473e-02
4.43585068e-01 -7.84974694e-01 -3.87581289e-01 -5.19487560e-01
1.20078608e-01 -6.88349783e-01 -2.88161814e-01 9.51448143e-01
-3.60182136e-01 3.07780504e-01 8.52339640e-02 2.56699082e-02
6.17705435e-02 -2.17077106e-01 3.32872808e-01 -2.78429180e-01
-1.28954619e-01 2.71822035e-01 -9.50144827e-01 5.93805254e-01
1.08816433e+00 -5.90321004e-01 -3.63479555e-01 1.68791190e-02
5.40076375e-01 -2.20391870e-01 5.27777910e-01 -8.82415473e-01
-2.47887462e-01 1.69105381e-02 1.82959288e-01 1.60433382e-01
-3.49401869e-02 -7.35220075e-01 -1.82571337e-02 7.67351747e-01
-5.00575006e-01 3.35085303e-01 5.69189966e-01 2.10662976e-01
2.98021957e-02 -7.03241706e-01 7.79552758e-01 -1.83008481e-02
-3.87721866e-01 -1.27170041e-01 -3.46045166e-01 3.85587029e-02
1.13316405e+00 -3.16384494e-01 -6.22565866e-01 -2.70062070e-02
-5.83742499e-01 -3.94740909e-01 9.45668876e-01 3.36256981e-01
4.58333969e-01 -7.86925435e-01 -5.30354142e-01 4.51607347e-01
8.35141018e-02 -7.38178015e-01 5.02212495e-02 6.03847086e-01
-2.23104104e-01 1.51804850e-01 -1.74429357e-01 -2.18179166e-01
-1.44924712e+00 1.05966413e+00 5.21487594e-01 -2.15076238e-01
-1.01901360e-01 5.60048342e-01 -1.66792348e-01 -6.56673789e-01
-3.45840044e-02 -2.70515442e-01 -1.83922350e-01 -1.65491804e-01
5.24623752e-01 2.88828969e-01 4.43106920e-01 -3.85606885e-01
-5.17305017e-01 5.99105477e-01 -1.88059941e-01 -6.47835582e-02
1.13731313e+00 5.12673318e-01 -2.52980947e-01 8.99929479e-02
1.47131586e+00 4.13984090e-01 -8.45381141e-01 2.12965295e-01
2.95657456e-01 -4.37770247e-01 -4.12212729e-01 -9.79772627e-01
-1.14463532e+00 8.89738560e-01 7.28191614e-01 6.55060947e-01
1.23981631e+00 -1.38788566e-01 3.08565319e-01 3.69495183e-01
3.85316521e-01 -6.22756958e-01 1.92118399e-02 3.26577991e-01
5.53808153e-01 -1.04283106e+00 -9.47588682e-02 -2.86558658e-01
-2.58071125e-01 1.28187299e+00 6.21664464e-01 -1.59354180e-01
4.77088392e-01 4.95821536e-01 -1.43989354e-01 3.04812565e-02
-6.02230787e-01 3.28634828e-01 -1.18997857e-01 7.66797066e-01
2.51338512e-01 -1.87996328e-01 -1.58530604e-02 5.75806797e-01
-3.72101605e-01 -4.83599603e-01 5.84204912e-01 1.15723312e+00
-4.12525952e-01 -1.57630432e+00 -7.18196034e-01 4.88347501e-01
-4.12170768e-01 -2.47989357e-01 -7.54970789e-01 9.61351275e-01
3.48423958e-01 1.21521318e+00 -4.35819209e-01 -9.25534487e-01
2.89184481e-01 3.07389796e-01 3.40557426e-01 -6.40325546e-01
-1.08831143e+00 -7.22252250e-01 9.39316377e-02 -5.57708144e-01
1.13651603e-01 -8.12468171e-01 -1.02928638e+00 -2.73431242e-01
-2.96339333e-01 4.35941875e-01 5.88291407e-01 9.44825649e-01
3.46438617e-01 2.61214435e-01 9.97414410e-01 -3.72126251e-01
-7.93082893e-01 -7.10512877e-01 -1.73260033e-01 3.52695256e-01
3.76883984e-01 -5.20680070e-01 -1.05632758e+00 -3.73738236e-03]
|
[7.075378894805908, 7.7659406661987305]
|
d8609545-b1da-4934-89ba-87d91a3eb892
|
effective-sampling-for-large-scale-automated
|
1412.5659
| null |
http://arxiv.org/abs/1412.5659v1
|
http://arxiv.org/pdf/1412.5659v1.pdf
|
Effective sampling for large-scale automated writing evaluation systems
|
Automated writing evaluation (AWE) has been shown to be an effective
mechanism for quickly providing feedback to students. It has already seen wide
adoption in enterprise-scale applications and is starting to be adopted in
large-scale contexts. Training an AWE model has historically required a single
batch of several hundred writing examples and human scores for each of them.
This requirement limits large-scale adoption of AWE since human-scoring essays
is costly. Here we evaluate algorithms for ensuring that AWE models are
consistently trained using the most informative essays. Our results show how to
minimize training set sizes while maximizing predictive performance, thereby
reducing cost without unduly sacrificing accuracy. We conclude with a
discussion of how to integrate this approach into large-scale AWE systems.
|
['Nicholas Dronen', 'Peter W. Foltz', 'Kyle Habermehl']
|
2014-12-17
| null | null | null | null |
['automated-writing-evaluation']
|
['natural-language-processing']
|
[-1.52641818e-01 -8.45627636e-02 -7.44979829e-02 -6.28673673e-01
-1.08983862e+00 -8.40821505e-01 3.32429618e-01 4.46228415e-01
-6.93949163e-01 8.48286569e-01 -1.92407206e-01 -7.00584352e-01
-2.54642487e-01 -7.54135489e-01 -2.68454194e-01 -2.66118459e-02
4.84899670e-01 5.82063854e-01 2.03341112e-01 -1.17187671e-01
8.99002969e-01 4.09541756e-01 -1.47479403e+00 2.40563244e-01
9.40567136e-01 4.16321784e-01 1.40615344e-01 1.28535986e+00
-3.92611772e-01 1.12949550e+00 -1.33073974e+00 -9.21593070e-01
1.29438490e-01 -4.69924629e-01 -8.37158620e-01 -8.13507140e-02
9.06920016e-01 -5.86497009e-01 3.23148891e-02 7.12240875e-01
4.52286661e-01 2.40054384e-01 4.94070113e-01 -1.06446707e+00
-8.12043548e-01 3.59260023e-01 -3.18793207e-01 2.59537786e-01
4.98595238e-01 5.88065237e-02 1.21680868e+00 -9.84500051e-01
5.93214273e-01 6.04280829e-01 7.19157040e-01 6.50754571e-01
-1.10233617e+00 -4.99149293e-01 -1.20219514e-01 1.99779317e-01
-8.61318350e-01 -5.00016272e-01 5.44809520e-01 -5.11821330e-01
8.92591715e-01 4.54514742e-01 7.32637882e-01 6.25658333e-01
6.17663637e-02 1.00575590e+00 1.23176587e+00 -6.97697461e-01
2.16302067e-01 4.65645641e-01 7.27606177e-01 7.61939108e-01
4.41915065e-01 -4.58690226e-01 -8.01033199e-01 -8.27074647e-02
4.00633544e-01 -2.43824705e-01 -2.71784868e-02 5.25521040e-02
-8.01528692e-01 9.40736473e-01 -1.40108883e-01 7.03996494e-02
-2.22482726e-01 8.97681117e-02 2.96820790e-01 9.02201653e-01
2.95854151e-01 1.02832305e+00 -3.05876821e-01 -8.31578910e-01
-1.26851368e+00 5.44304132e-01 1.06031692e+00 8.69937003e-01
3.40412050e-01 -1.53795198e-01 -2.26774484e-01 1.15447938e+00
1.59533873e-01 2.71840900e-01 6.32887483e-01 -1.10042334e+00
5.28974533e-01 8.36817324e-01 4.24693137e-01 -7.06946850e-01
1.41802341e-01 -2.50076354e-01 -1.03697151e-01 4.04398412e-01
7.01347530e-01 -2.27582991e-01 -4.07081366e-01 1.43276644e+00
-4.67894636e-02 -1.67151824e-01 -1.96562767e-01 5.28601229e-01
5.99032462e-01 5.15883982e-01 1.61474291e-02 -1.69452325e-01
1.07971931e+00 -1.08612633e+00 -6.71223879e-01 -3.68809879e-01
7.58687317e-01 -1.29490709e+00 1.51109719e+00 8.22908759e-01
-1.52058351e+00 -2.71724582e-01 -1.01448393e+00 -1.31696865e-01
-1.45013466e-01 8.68430808e-02 4.86502767e-01 1.15291917e+00
-1.03450167e+00 7.24474728e-01 -5.90482414e-01 -4.16352265e-02
4.56685185e-01 2.42553756e-01 -1.36901796e-01 -1.69765994e-01
-7.33540356e-01 1.13508081e+00 -2.02480197e-01 -2.66454905e-01
-3.64343375e-01 -7.13228583e-01 -3.46706927e-01 3.34503919e-01
1.14914849e-01 -3.02404851e-01 2.02723575e+00 -5.32796860e-01
-1.77615738e+00 7.68533170e-01 -4.86505851e-02 -1.34986684e-01
6.40277267e-01 -5.39484978e-01 -9.30026695e-02 -1.04256421e-01
-1.39471859e-01 2.36842230e-01 3.24455857e-01 -4.89935488e-01
-4.68299776e-01 -1.66959167e-01 -5.38753197e-02 3.05399567e-01
-1.19627094e+00 4.89510417e-01 -2.39219591e-01 -4.19705957e-01
6.08368660e-04 -8.21148396e-01 -4.08090204e-02 -9.98619497e-02
9.18202400e-02 -6.61688805e-01 5.11545837e-01 -6.19003713e-01
1.96408904e+00 -1.48897922e+00 -3.67446661e-01 1.93813324e-01
3.12002629e-01 5.45518458e-01 -1.38287947e-01 5.17298341e-01
5.32409072e-01 1.95126116e-01 1.49864748e-01 -3.48529577e-01
2.52594441e-01 -1.44607663e-01 -7.77118430e-02 -1.91350296e-01
-1.99340247e-02 8.94952536e-01 -1.03945518e+00 -5.72767973e-01
7.57966787e-02 4.17155847e-02 -5.89846790e-01 5.05366027e-01
-1.25846922e-01 -5.45728803e-01 -3.15835148e-01 4.67626631e-01
7.17844740e-02 -5.79940915e-01 5.39696775e-03 9.78028953e-01
-8.45044926e-02 7.00440407e-01 -1.31399512e+00 1.07976174e+00
-6.79003656e-01 1.08967423e+00 -1.61899835e-01 -5.35782278e-01
1.34611177e+00 2.81727225e-01 4.04851660e-02 -5.24221361e-01
-8.88424292e-02 4.82950807e-01 8.86194929e-02 -4.14064467e-01
7.80557394e-01 -4.72381674e-02 -8.02961737e-03 1.17267931e+00
2.92724445e-02 -4.40861315e-01 6.98644459e-01 5.82335532e-01
1.25498104e+00 -1.33925229e-01 3.86785448e-01 -5.50344363e-02
3.26620817e-01 1.10513203e-01 1.25292167e-01 9.29450154e-01
-2.60381013e-01 2.91209251e-01 2.47243315e-01 -4.03808057e-01
-1.21506655e+00 -6.38683558e-01 -5.29347220e-03 1.51168013e+00
-5.09298265e-01 -7.48207629e-01 -8.33275080e-01 -7.35398829e-01
3.39688249e-02 8.69892180e-01 -2.61851251e-01 9.66096297e-02
-6.57551289e-01 -5.97776949e-01 3.47907335e-01 6.04342818e-01
2.22565144e-01 -1.10766554e+00 -8.34066451e-01 4.98261184e-01
-3.90826836e-02 -5.75147390e-01 -4.80290145e-01 1.59150615e-01
-9.64185655e-01 -1.01210606e+00 -7.61700749e-01 -6.40633643e-01
7.41298378e-01 3.24971825e-01 1.51001501e+00 6.00071609e-01
-2.46879578e-01 5.65350592e-01 -1.21948242e-01 -8.54709387e-01
-6.62241638e-01 2.15820804e-01 -2.70651430e-01 -7.89778054e-01
1.16020930e+00 -2.81408221e-01 -5.04912376e-01 3.48045260e-01
-5.23578525e-01 1.26113698e-01 5.29341578e-01 8.81195545e-01
-5.14573120e-02 -5.78200102e-01 9.39211071e-01 -1.30351388e+00
1.46518743e+00 -1.15756862e-01 -6.74903691e-01 5.98760843e-01
-1.33236551e+00 -2.21480772e-01 7.19287753e-01 -4.62551951e-01
-9.57882226e-01 -3.78932476e-01 -9.62119922e-02 1.38511807e-01
4.74878065e-02 6.52829766e-01 5.88404953e-01 -1.79064408e-01
9.35474694e-01 -1.36292810e-02 -4.91724275e-02 -3.42421293e-01
-6.29390106e-02 9.15724099e-01 2.95247287e-01 -8.48015070e-01
4.94275302e-01 -6.68948591e-01 -2.82484204e-01 -5.61303556e-01
-1.07569385e+00 -5.86042404e-01 -5.17380476e-01 -5.35978138e-01
5.68140075e-02 -5.53843558e-01 -9.41771746e-01 1.89616963e-01
-8.77569139e-01 -5.70294619e-01 -4.02529836e-01 3.53096068e-01
-2.62997746e-01 1.02243550e-01 -9.62145865e-01 -8.76521111e-01
-4.53664035e-01 -9.21039462e-01 3.62770170e-01 6.35947645e-01
-8.92375171e-01 -1.02377403e+00 4.72659826e-01 8.40752125e-01
5.94527841e-01 -4.84555155e-01 6.95905089e-01 -8.47278655e-01
-4.91216511e-01 -7.40976274e-01 2.26952042e-02 4.34541851e-01
-4.71338958e-01 3.43939483e-01 -1.01621044e+00 -1.40625268e-01
-1.84554413e-01 -7.24260390e-01 4.42447126e-01 4.49783430e-02
1.15800071e+00 -2.21453205e-01 2.51201510e-01 2.18903851e-02
1.33279574e+00 1.20694205e-01 1.85515538e-01 6.00484431e-01
2.57012159e-01 4.54871178e-01 6.28931701e-01 4.31996256e-01
1.27892733e-01 4.07786638e-01 -3.75846803e-01 5.08847475e-01
-1.35996431e-01 -2.77328104e-01 5.56003571e-01 1.36314046e+00
-4.67572808e-02 -2.55278945e-01 -1.19121838e+00 6.47105038e-01
-1.59236753e+00 -9.82858479e-01 -1.96973354e-01 2.22016501e+00
1.05589437e+00 3.66080433e-01 1.14365108e-01 2.39062697e-01
3.43192309e-01 -6.00359924e-02 -3.77144098e-01 -9.47095573e-01
3.56431246e-01 7.33894169e-01 3.43812108e-02 7.41814017e-01
-5.31100214e-01 6.64453268e-01 7.26811266e+00 4.89416987e-01
-7.56640553e-01 4.25900482e-02 6.31325305e-01 -2.92006105e-01
-3.31431657e-01 -2.45707095e-01 -1.16463196e+00 3.96204770e-01
1.37127864e+00 -7.48761535e-01 1.78452954e-01 1.19343221e+00
7.14370562e-03 -1.75272942e-01 -1.06615031e+00 5.68576872e-01
7.94126540e-02 -1.24395919e+00 -2.24074394e-01 -3.70424278e-02
1.08588898e+00 -3.15246493e-01 -5.14631793e-02 6.00906134e-01
8.78835618e-01 -1.05154717e+00 3.92436534e-01 3.33451092e-01
6.15536571e-01 -6.17905080e-01 7.75027990e-01 6.42571270e-01
-6.68195009e-01 -2.21712202e-01 -5.58660805e-01 -4.79749978e-01
-2.56836742e-01 5.08085310e-01 -1.27159846e+00 -2.07074940e-01
2.72823304e-01 1.92520708e-01 -9.88930106e-01 1.13798773e+00
-5.05565286e-01 9.96486902e-01 -2.30188757e-01 -8.60909581e-01
2.54821241e-01 -1.62933305e-01 1.10896215e-01 1.32845771e+00
4.57549363e-01 1.59143507e-01 5.96028492e-02 6.87489450e-01
-3.81238997e-01 2.81931698e-01 -3.78421307e-01 -9.70886052e-02
9.08268154e-01 1.32601964e+00 -6.24730229e-01 -6.96788132e-01
-4.08217192e-01 8.17281663e-01 6.80651188e-01 -5.40869609e-02
-2.58550614e-01 -7.19099045e-01 2.49489322e-01 2.57141113e-01
-1.10443413e-01 -1.42715171e-01 -8.62225711e-01 -9.22089040e-01
-1.77255797e-03 -1.00752771e+00 4.09521401e-01 -8.43671918e-01
-1.17698121e+00 1.97350726e-01 -3.89765531e-01 -9.58065629e-01
-4.25060004e-01 -8.03148329e-01 -9.69164252e-01 1.07937682e+00
-1.28993607e+00 -3.48203212e-01 -4.89350259e-01 1.22339375e-01
8.06924045e-01 -3.98344010e-01 1.18353271e+00 2.73321867e-01
-6.47267580e-01 9.58054304e-01 2.15639949e-01 1.33309633e-01
1.24960661e+00 -1.75621259e+00 3.53178144e-01 8.02866638e-01
3.97123605e-01 1.11729634e+00 6.75771177e-01 -5.26984513e-01
-1.02573884e+00 -5.91482103e-01 1.59370351e+00 -9.82808948e-01
6.73478127e-01 1.89502258e-02 -1.01625466e+00 6.07178032e-01
2.24437013e-01 -2.60082334e-01 1.08696508e+00 6.48239315e-01
-2.79652953e-01 6.83769286e-02 -8.59581053e-01 6.65074289e-01
5.06250024e-01 -6.00382864e-01 -7.70995617e-01 5.05899787e-01
5.83080947e-02 -3.96409124e-01 -1.17770302e+00 -1.74333513e-01
7.45744586e-01 -9.61963654e-01 6.75472140e-01 -6.28743112e-01
1.12313640e+00 1.76947504e-01 3.58642966e-01 -1.28078830e+00
-2.70650685e-01 -7.32799768e-01 -4.56674010e-01 1.20001662e+00
5.82934320e-01 -2.93945551e-01 1.29130292e+00 1.29512727e+00
-5.15753888e-02 -1.10678208e+00 -1.59771293e-01 -7.58503854e-01
3.15977693e-01 -2.11920515e-01 5.28123856e-01 1.04071009e+00
4.80269998e-01 5.39450288e-01 -3.78660299e-02 -4.84113902e-01
4.53958243e-01 1.56317651e-01 1.07399857e+00 -1.51643610e+00
-2.96573162e-01 -9.36046422e-01 -2.46210881e-02 -1.11413109e+00
-3.18269134e-01 -7.25288928e-01 8.08615610e-03 -1.46077597e+00
4.92467284e-01 -4.54867780e-01 -3.46924752e-01 3.70144427e-01
-7.64427483e-01 4.24501657e-01 2.20072433e-01 1.28450125e-01
-8.48115802e-01 -1.16557613e-01 1.20541048e+00 3.12146366e-01
-7.36674741e-02 2.70031840e-01 -8.57238472e-01 6.19439840e-01
8.10513496e-01 -4.75454003e-01 -5.33465028e-01 -2.62693554e-01
5.12813687e-01 -1.10682677e-02 -1.66794226e-01 -9.19850349e-01
6.81266427e-01 -3.95317584e-01 6.36515975e-01 -3.32091838e-01
-1.28172353e-01 -3.51190329e-01 -4.13670689e-01 8.17898214e-02
-8.58160496e-01 4.97499257e-01 -1.99729297e-02 3.72625887e-02
-3.76797140e-01 -1.14379406e+00 7.56519914e-01 -2.80407310e-01
-4.63464528e-01 -1.14152022e-01 -4.34721410e-01 1.90145224e-01
1.03682220e+00 -2.91433543e-01 -4.42551166e-01 -3.96200895e-01
-2.67459691e-01 2.40876794e-01 4.51381564e-01 6.26542345e-02
5.04123688e-01 -1.06998038e+00 -6.59412086e-01 1.15719341e-01
1.59960896e-01 -1.51329696e-01 1.60034448e-02 3.45751941e-01
-7.36979604e-01 5.37577868e-01 -1.81758568e-01 -1.95739359e-01
-1.97493708e+00 1.19183203e-02 -5.63490316e-02 -7.72142410e-01
-5.45663178e-01 1.21492910e+00 -7.73340106e-01 -5.17013311e-01
3.95354450e-01 1.36497796e-01 -2.81489104e-01 -6.91087544e-02
9.14072692e-01 6.09904885e-01 4.17338789e-01 3.20114523e-01
2.87279159e-01 1.37424588e-01 -4.28454787e-01 -3.49125266e-01
1.46421182e+00 2.33655125e-01 2.14739561e-01 5.99963009e-01
6.78773820e-01 3.33836734e-01 -9.50607598e-01 -8.52122828e-02
4.52936918e-01 -6.34834111e-01 8.65164474e-02 -1.21671271e+00
-2.46943653e-01 1.16027832e+00 4.64479774e-02 3.70258570e-01
7.85101771e-01 -6.22051537e-01 7.70165443e-01 1.04002130e+00
1.44967541e-01 -1.55333769e+00 4.75074023e-01 5.15172720e-01
5.09739399e-01 -1.43692863e+00 4.11945671e-01 1.24561645e-01
-6.01794720e-01 1.40805531e+00 9.75183249e-01 -5.10044338e-05
1.76031619e-01 2.88817823e-01 2.39299655e-01 -2.53025830e-01
-1.26412475e+00 6.61332428e-01 4.50592428e-01 -2.68189982e-02
1.18541718e+00 3.34468096e-01 -5.45735061e-01 5.81201315e-01
-5.46039701e-01 1.85450703e-01 1.00127745e+00 1.08838570e+00
-9.10935640e-01 -1.49664164e+00 -5.28409779e-01 9.45589542e-01
-5.64810872e-01 -1.16911195e-01 -7.12534070e-01 6.73647285e-01
-5.72234809e-01 8.36816609e-01 -2.35170379e-01 -1.18105374e-01
3.26134056e-01 8.86352956e-01 5.70341349e-01 -9.29413557e-01
-1.10186338e+00 -3.24092478e-01 1.37661338e-01 -8.71764794e-02
8.22378322e-02 -6.27203345e-01 -7.18323946e-01 -9.67925608e-01
-5.62161803e-01 4.11485255e-01 6.94936335e-01 6.64384067e-01
1.57015070e-01 4.33792621e-01 4.93207246e-01 -1.25838742e-01
-1.22289824e+00 -1.18350494e+00 -3.81278545e-01 3.57657254e-01
-7.36170039e-02 -2.05580264e-01 -1.80219233e-01 2.82958262e-02]
|
[11.286768913269043, 9.333086967468262]
|
37644c8a-1257-4717-81a9-1ec535687d8f
|
vectorized-scenario-description-and-motion
|
2302.01161
| null |
https://arxiv.org/abs/2302.01161v1
|
https://arxiv.org/pdf/2302.01161v1.pdf
|
Vectorized Scenario Description and Motion Prediction for Scenario-Based Testing
|
Automated vehicles (AVs) are tested in diverse scenarios, typically specified by parameters such as velocities, distances, or curve radii. To describe scenarios uniformly independent of such parameters, this paper proposes a vectorized scenario description defined by the road geometry and vehicles' trajectories. Data of this form are generated for three scenarios, merged, and used to train the motion prediction model VectorNet, allowing to predict an AV's trajectory for unseen scenarios. Predicting scenario evaluation metrics, VectorNet partially achieves lower errors than regression models that separately process the three scenarios' data. However, for comprehensive generalization, sufficient variance in the training data must be ensured. Thus, contrary to existing methods, our proposed method can merge diverse scenarios' data and exploit spatial and temporal nuances in the vectorized scenario description. As a result, data from specified test scenarios and real-world scenarios can be compared and combined for (predictive) analyses and scenario selection.
|
['Steffen Müller', 'Constantin Vasconi', 'Max Winkelmann']
|
2023-02-02
| null | null | null | null |
['motion-prediction']
|
['computer-vision']
|
[ 8.27456117e-02 -2.79409051e-01 -3.64730269e-01 -5.97391665e-01
-4.76250261e-01 -7.84272194e-01 1.03961575e+00 3.70280206e-01
-4.78314459e-01 7.59646177e-01 1.24857113e-01 -6.42927885e-01
-4.69400704e-01 -1.01934576e+00 -3.79261106e-01 -7.67019212e-01
-2.03218952e-01 5.36382616e-01 5.55779278e-01 -1.97815165e-01
2.99145728e-01 1.10692143e+00 -2.00455809e+00 -4.10510227e-02
8.20262671e-01 6.23820305e-01 4.08288747e-01 7.43797541e-01
-2.28888810e-01 2.73901671e-01 -5.79194367e-01 -7.89637491e-02
1.09058827e-01 6.78295782e-03 -2.54263282e-01 1.75719690e-02
1.02798708e-01 -2.84408242e-01 -4.80305761e-01 6.06985271e-01
2.58680820e-01 6.50993824e-01 1.03138316e+00 -1.74087203e+00
1.73367932e-01 6.22493029e-01 -3.67848724e-02 -8.75889361e-02
2.21024334e-01 6.26299500e-01 5.47136247e-01 -5.91517091e-01
7.88225710e-01 1.15317857e+00 5.01339018e-01 4.04956251e-01
-1.18046677e+00 -5.47170699e-01 4.45983768e-01 5.17642736e-01
-1.60482597e+00 -3.81480306e-01 9.58623707e-01 -7.08182216e-01
8.51373136e-01 5.50594926e-01 6.54523015e-01 1.42792726e+00
9.50039774e-02 7.56141484e-01 3.98976922e-01 2.54044741e-01
4.06966507e-01 4.06985104e-01 3.45435329e-02 1.54345913e-03
3.35638076e-01 3.57117623e-01 7.03223199e-02 -5.28900661e-02
8.00805613e-02 -3.19315717e-02 -1.92998603e-01 -6.31536305e-01
-1.40731549e+00 7.10515916e-01 1.32666364e-01 1.48184195e-01
-5.32355368e-01 2.34676432e-03 5.22217989e-01 5.67909963e-02
1.19776867e-01 3.55887890e-01 -5.46799362e-01 -2.32897818e-01
-1.11003804e+00 7.15289354e-01 4.61905956e-01 1.17209542e+00
9.13012564e-01 3.38990599e-01 -5.61668813e-01 5.39950490e-01
3.00317019e-01 8.90451610e-01 1.63721040e-01 -6.79489911e-01
7.58922637e-01 6.52801573e-01 3.97700578e-01 -1.26460886e+00
-6.75258934e-01 -1.86646760e-01 -6.87077940e-01 4.09027189e-02
1.99908316e-01 -3.92058313e-01 -9.47037220e-01 1.69467843e+00
3.41812223e-01 3.64436537e-01 2.06038326e-01 7.55907595e-01
6.94834948e-01 8.10727358e-01 3.99667412e-01 -1.93240777e-01
5.84311306e-01 -3.22762012e-01 -4.65609521e-01 1.66704506e-01
8.26267660e-01 -2.67158210e-01 7.03306913e-01 -1.42597109e-02
-7.25751758e-01 -6.81038380e-01 -9.88075972e-01 7.80459285e-01
-8.75715613e-01 -4.71719839e-02 2.88280606e-01 6.26557946e-01
-8.67048085e-01 5.20596147e-01 -4.73014832e-01 -3.61319900e-01
2.06533805e-01 3.36157322e-01 -1.64001390e-01 1.85913835e-02
-1.38856196e+00 1.11988151e+00 5.99163353e-01 2.55117267e-01
-1.25836825e+00 -8.38379145e-01 -9.96556938e-01 -1.20530203e-01
1.34448977e-02 -3.99016798e-01 7.56355584e-01 -1.23281874e-01
-1.08739555e+00 1.44699469e-01 -7.14326277e-02 -3.68355453e-01
7.94570863e-01 3.13025564e-01 -9.69834685e-01 -1.07397407e-01
-4.06774692e-03 6.62701190e-01 4.42750812e-01 -1.65530920e+00
-7.82063246e-01 8.04236010e-02 -3.70384827e-02 1.20198108e-01
-2.17793137e-02 -1.77315161e-01 -4.90249217e-01 -2.85023868e-01
-2.30381086e-01 -8.83137703e-01 -7.07200229e-01 -4.95006800e-01
-6.66081369e-01 -8.04833248e-02 9.74169970e-01 -3.34454060e-01
1.34506941e+00 -2.03538728e+00 -3.44879040e-03 6.85169935e-01
-2.24333107e-02 3.70754600e-01 -3.77671331e-01 6.21814311e-01
2.49067679e-01 2.34806210e-01 -3.34962636e-01 -1.65548176e-02
5.19923903e-02 1.61819920e-01 -1.60528123e-01 5.87966621e-01
2.37115413e-01 7.43836522e-01 -9.90354478e-01 -5.31034589e-01
1.05071902e+00 5.06547868e-01 -1.30801946e-01 1.62279904e-01
-2.63673246e-01 6.26277149e-01 -8.18390369e-01 3.71304423e-01
8.63688588e-01 5.30815780e-01 5.04104942e-02 -1.28029585e-01
-2.83487469e-01 -2.86221325e-01 -1.24655390e+00 8.76728117e-01
-6.56883955e-01 9.67025459e-01 -5.12433887e-01 -8.96777689e-01
1.24602294e+00 1.89595386e-01 7.93704569e-01 -5.65767348e-01
2.41206512e-01 -3.90418656e-02 -1.30429894e-01 -7.25477636e-01
7.28259742e-01 4.53416944e-01 -3.35595816e-01 -7.06349164e-02
-3.04060131e-01 -1.93589896e-01 3.83633882e-01 -1.17619760e-01
8.91294062e-01 1.18851781e-01 5.11851311e-02 -1.75544177e-03
6.57182634e-01 4.70062941e-01 3.13566446e-01 7.34058142e-01
-1.22427978e-01 4.78606731e-01 2.86050498e-01 -4.76128519e-01
-1.26046562e+00 -1.22564352e+00 -3.64107668e-01 4.90396708e-01
6.54456377e-01 -1.38224140e-01 -2.31533900e-01 -7.52014339e-01
2.09860012e-01 1.38929939e+00 -5.90962291e-01 -2.45945588e-01
-6.84567690e-01 -5.40795445e-01 6.19280517e-01 4.33679372e-01
1.67342111e-01 -8.94733369e-01 -6.42563939e-01 3.49603355e-01
1.27227768e-01 -1.05827677e+00 -5.47153410e-03 -2.19199404e-01
-4.14342195e-01 -1.06949437e+00 -3.26611489e-01 -8.16175044e-02
4.83913004e-01 4.20512825e-01 7.72083044e-01 -2.68595546e-01
1.11746959e-01 2.36599624e-01 -4.02973741e-01 -4.04612929e-01
-5.79122841e-01 6.05570599e-02 6.73985258e-02 2.14233965e-01
2.88622648e-01 -5.39677262e-01 -6.07558489e-01 6.52234793e-01
-7.80394673e-01 -4.77342717e-02 3.63610297e-01 3.47415805e-01
4.46789205e-01 2.48241186e-01 5.87654591e-01 -5.30236125e-01
4.82118845e-01 -1.05869174e+00 -5.96407473e-01 3.68157804e-01
-6.47610664e-01 -5.23777194e-02 8.19583476e-01 -7.09754705e-01
-1.10393226e+00 -5.08107990e-03 -6.09022751e-03 -5.50179839e-01
-8.29738021e-01 4.32764977e-01 -4.57759887e-01 1.70217618e-01
6.12463236e-01 8.25134367e-02 -3.44526708e-01 -5.18795997e-02
5.70382595e-01 5.33591270e-01 3.86877358e-01 -1.51212066e-01
1.09344471e+00 2.97720730e-01 -2.29612179e-03 -9.56023216e-01
5.16483076e-02 -6.02673471e-01 -7.92469919e-01 -9.29002702e-01
8.39182258e-01 -6.34823799e-01 -4.98423487e-01 1.08386770e-01
-1.12805057e+00 -3.32912296e-01 -2.25713208e-01 9.42366600e-01
-5.35132766e-01 -2.93548312e-02 3.88372868e-01 -1.04615295e+00
3.88400733e-01 -1.34722888e+00 9.82753813e-01 9.83348265e-02
-5.40172696e-01 -1.29135346e+00 3.56664099e-02 -2.15031400e-01
5.72546422e-01 7.23817706e-01 7.17393696e-01 -8.92459273e-01
-5.38083196e-01 -5.63910484e-01 -1.49899036e-01 -9.56506357e-02
-2.19539879e-03 4.74861711e-01 -7.73706794e-01 -1.65613383e-01
-5.44405937e-01 3.24245006e-01 7.16855824e-01 3.97410035e-01
1.26936686e+00 -2.67203778e-01 -8.40356410e-01 6.52056038e-01
1.36606479e+00 5.20183802e-01 5.22961974e-01 3.28701466e-01
7.93451011e-01 9.87384737e-01 1.01838899e+00 4.90589142e-01
3.98997039e-01 9.31156456e-01 6.04531050e-01 6.77707344e-02
3.94620188e-02 -2.14912653e-01 2.01790795e-01 2.30447918e-01
8.44008997e-02 -6.84232116e-01 -1.14958262e+00 9.45489764e-01
-1.86746383e+00 -1.33727992e+00 -4.78776783e-01 2.26176071e+00
-3.90662178e-02 1.38823226e-01 2.56424427e-01 1.50768518e-01
8.32017481e-01 4.89857525e-01 -7.44728923e-01 -5.09546101e-01
-2.86812007e-01 -4.71672058e-01 1.06786942e+00 4.31376487e-01
-9.98781502e-01 7.75893211e-01 6.40242815e+00 9.01876807e-01
-1.25268590e+00 -3.08866173e-01 2.86135554e-01 -1.58569068e-01
-8.79597425e-01 -1.08156994e-01 -8.25443268e-01 6.10102713e-01
1.29822040e+00 -4.00384545e-01 4.97728325e-02 7.04515338e-01
8.79085958e-01 5.35200238e-02 -9.70574081e-01 6.10993922e-01
-2.19137877e-01 -1.43473399e+00 2.82505035e-01 7.95041174e-02
6.89660430e-01 9.56903398e-02 9.10207555e-02 4.75505352e-01
5.58596253e-01 -1.12296689e+00 5.71559608e-01 7.64682710e-01
7.98288882e-01 -9.20540035e-01 8.55631888e-01 3.43910784e-01
-1.27490687e+00 -3.07572782e-01 -2.34086648e-01 5.01977324e-01
5.70846498e-01 1.75829202e-01 -1.18514168e+00 9.39570308e-01
3.41962278e-01 6.08761370e-01 -6.05306327e-01 1.22625768e+00
2.90573537e-01 7.88236082e-01 -2.74632663e-01 -3.12962681e-01
5.59719265e-01 -2.24742100e-01 9.10963356e-01 1.38320577e+00
5.68638682e-01 -3.74812663e-01 1.42998323e-01 6.08885646e-01
6.61695361e-01 1.32185325e-01 -1.16690016e+00 3.19696754e-01
8.54225755e-01 9.65124190e-01 -4.90928113e-01 -1.61102042e-01
-2.87330955e-01 1.56755954e-01 -2.05195934e-01 9.09447908e-01
-1.23128998e+00 -2.94587851e-01 9.59260643e-01 1.03383876e-01
3.04652154e-01 -2.98290253e-01 -3.51023972e-01 -6.11774445e-01
-1.29410759e-01 -1.76329762e-01 8.08050111e-03 -5.82528830e-01
-9.17294443e-01 7.44280696e-01 9.07953441e-01 -1.84373128e+00
-7.46823430e-01 -5.00508070e-01 -9.91355240e-01 7.42615402e-01
-1.37515199e+00 -1.32305813e+00 -4.41388726e-01 5.27492702e-01
5.97562671e-01 -4.75626856e-01 3.61583024e-01 1.23557128e-01
-5.09390593e-01 3.36554497e-01 3.58580410e-01 -2.33997956e-01
2.11271808e-01 -9.03466165e-01 4.80782926e-01 7.44864345e-01
-1.98865086e-01 1.71179980e-01 1.12198675e+00 -6.38572037e-01
-1.07516563e+00 -1.90403211e+00 7.85710454e-01 -4.32812780e-01
6.84374034e-01 -1.64379135e-01 -8.22606623e-01 2.86172718e-01
-3.09839845e-01 -2.53279865e-01 5.19215822e-01 -2.62815922e-01
-4.64160889e-02 -1.61066175e-01 -1.24176788e+00 1.05143547e+00
9.66553211e-01 -1.03501186e-01 -1.62246197e-01 9.66746286e-02
6.77974761e-01 2.69609038e-02 -8.45544994e-01 7.55392671e-01
5.94790757e-01 -7.31575012e-01 1.05759394e+00 -7.51483500e-01
3.41699496e-02 -3.90310884e-01 -4.55592006e-01 -1.53511238e+00
-2.03100532e-01 -2.55858004e-01 4.17913757e-02 1.19922137e+00
7.80319691e-01 -6.08147979e-01 5.70156693e-01 5.97823203e-01
-2.61289805e-01 -6.47060335e-01 -9.03125048e-01 -9.08431590e-01
1.90342188e-01 -1.17407191e+00 1.44387603e+00 5.74151754e-01
-5.23228049e-01 -1.79260582e-01 -4.30105865e-01 5.48363805e-01
4.96235639e-01 1.83946602e-02 1.01282454e+00 -1.17324936e+00
4.53302830e-01 -8.42506707e-01 -7.97560036e-01 -7.56400108e-01
3.87491465e-01 -6.66535735e-01 8.10059458e-02 -1.59626901e+00
-3.01786214e-01 -7.63613105e-01 -2.43769232e-02 -1.03466369e-01
-4.98238690e-02 1.80448554e-02 1.35536909e-01 -1.16566373e-02
-3.77949744e-01 6.05840325e-01 9.02836740e-01 -3.36777717e-01
-4.49555933e-01 4.34178054e-01 -6.73990548e-02 4.17318821e-01
1.12869990e+00 -8.91871154e-02 -8.77662182e-01 -6.50760308e-02
-1.71258211e-01 3.65125686e-01 4.12414670e-01 -1.04386592e+00
2.01993734e-01 -9.93285656e-01 2.26161793e-01 -1.22090626e+00
2.16284364e-01 -9.65580463e-01 6.10146403e-01 2.49185087e-03
-5.45358360e-01 -5.55784181e-02 2.67853886e-01 5.30181706e-01
-1.72791138e-01 -2.51577590e-02 4.70767170e-01 4.23258513e-01
-1.01373708e+00 7.10410178e-01 -9.17561173e-01 -3.59327525e-01
1.60253847e+00 -8.74452889e-01 -1.07801437e-01 -4.09323782e-01
-6.19850874e-01 7.96742082e-01 5.04902124e-01 6.87893629e-01
8.01384807e-01 -1.38785720e+00 -9.01701093e-01 9.66415741e-03
4.70132083e-01 -1.18356217e-02 7.53023684e-01 4.73881811e-01
-3.12559932e-01 6.08625352e-01 -1.21124640e-01 -7.54470646e-01
-9.87976670e-01 8.75088692e-01 1.80721804e-01 6.01512007e-02
-4.12568569e-01 1.12295523e-01 9.04115383e-03 -8.21338952e-01
-4.70739119e-02 -1.14468887e-01 -6.74376428e-01 2.31570810e-01
3.21155965e-01 6.11466348e-01 -4.06949334e-02 -1.18825841e+00
-4.26487118e-01 4.89789039e-01 5.92376709e-01 -1.93375036e-01
1.11398149e+00 -3.79230589e-01 7.29724824e-01 5.82868099e-01
1.27961552e+00 -2.40888596e-02 -1.36868882e+00 2.24543571e-01
7.44550005e-02 -4.02282953e-01 -2.01943740e-01 -4.74052459e-01
-9.35363710e-01 6.89096808e-01 5.75395525e-01 2.54834831e-01
8.09136987e-01 -7.19331726e-02 4.82203066e-01 2.79294759e-01
3.74736249e-01 -8.98650289e-01 -6.10839844e-01 2.73122698e-01
8.92263353e-01 -1.14190090e+00 -3.56336623e-01 -2.02753842e-01
-9.46755171e-01 9.49279368e-01 6.52674973e-01 1.32075980e-01
7.67979085e-01 1.65588424e-01 -1.49793243e-02 -1.49021316e-02
-8.63735914e-01 -4.66116637e-01 4.22753960e-01 1.25342751e+00
-1.53185427e-01 4.12209034e-01 -5.21143451e-02 2.37518877e-01
-3.12699616e-01 -5.98958790e-01 4.18137401e-01 3.29298437e-01
-3.01669538e-01 -7.04316378e-01 -3.48849833e-01 6.16562724e-01
3.21437240e-01 3.59693497e-01 -3.42665846e-03 1.13253975e+00
2.86528736e-01 1.23006487e+00 3.84890616e-01 -7.83879697e-01
5.18456697e-01 -2.32176200e-01 -2.36098468e-01 -1.65288910e-01
-3.32629412e-01 -5.36791265e-01 4.62492436e-01 -5.86749971e-01
-1.41609430e-01 -1.01770866e+00 -1.03670144e+00 -4.84749943e-01
-1.39179498e-01 1.72169223e-01 9.51015174e-01 9.92506385e-01
2.34742448e-01 5.50587773e-01 1.09426641e+00 -1.11392093e+00
-2.73220718e-01 -7.16149986e-01 -2.97598481e-01 5.55876136e-01
3.63093674e-01 -9.50123906e-01 -4.69149411e-01 -2.60776669e-01]
|
[5.751068592071533, 1.1551727056503296]
|
660e5f34-077b-4fe5-a7fa-76004910faf3
|
pars-absa-a-manually-annotated-aspect-based
| null | null |
https://aclanthology.org/2022.lrec-1.763
|
https://aclanthology.org/2022.lrec-1.763.pdf
|
Pars-ABSA: a Manually Annotated Aspect-based Sentiment Analysis Benchmark on Farsi Product Reviews
|
Due to the increased availability of online reviews, sentiment analysis witnessed a thriving interest from researchers. Sentiment analysis is a computational treatment of sentiment used to extract and understand the opinions of authors. While many systems were built to predict the sentiment of a document or a sentence, many others provide the necessary detail on various aspects of the entity (i.e., aspect-based sentiment analysis). Most of the available data resources were tailored to English and the other popular European languages. Although Farsi is a language with more than 110 million speakers, to the best of our knowledge, there is a lack of proper public datasets on aspect-based sentiment analysis for Farsi. This paper provides a manually annotated Farsi dataset, Pars-ABSA, annotated and verified by three native Farsi speakers. The dataset consists of 5,114 positive, 3,061 negative and 1,827 neutral data samples from 5,602 unique reviews. Moreover, as a baseline, this paper reports the performance of some aspect-based sentiment analysis methods focusing on transfer learning on Pars-ABSA.
|
['Sauleh Eetemadi', 'Behrouz Minaei-Bidgoli', 'Soroush Javdan', 'Kamyar Darvishi', 'Taha Shangipour ataei']
| null | null | null | null |
lrec-2022-6
|
['aspect-based-sentiment-analysis']
|
['natural-language-processing']
|
[-2.43417457e-01 -9.90867149e-03 -4.39084649e-01 -7.65648007e-01
-7.40491629e-01 -6.96630955e-01 5.34782052e-01 3.41538101e-01
-3.85350674e-01 7.64735103e-01 3.21459174e-01 -1.76167116e-01
5.26615083e-01 -6.75994337e-01 -4.62527573e-01 -5.18316805e-01
5.03774524e-01 2.42316782e-01 -3.16191278e-02 -8.31942856e-01
6.47836626e-01 -2.03435384e-02 -1.31104076e+00 5.17077088e-01
7.33584583e-01 1.05698717e+00 -3.45439345e-01 5.62112510e-01
-4.93251711e-01 8.27298164e-01 -1.02533650e+00 -1.22575927e+00
-1.64852336e-01 -3.61286104e-01 -4.78288651e-01 -1.06174886e-01
1.68360144e-01 3.42158377e-01 4.16548789e-01 1.10743988e+00
3.84159207e-01 -3.03059161e-01 6.49996102e-01 -1.35626733e+00
-9.90910947e-01 8.61960351e-01 -6.73061013e-01 1.28720909e-01
4.04496700e-01 -4.37495917e-01 1.49589241e+00 -1.11423111e+00
8.04026902e-01 8.91082048e-01 5.56533635e-01 2.34725475e-01
-2.15644285e-01 -8.44096363e-01 3.01931381e-01 1.24203879e-02
-8.96700740e-01 -1.39892787e-01 9.83101070e-01 -4.60907996e-01
9.33054090e-01 1.07638650e-01 1.00121272e+00 1.16971517e+00
5.12066007e-01 9.05304015e-01 1.41928232e+00 -5.01343310e-01
2.95862824e-01 1.10172105e+00 5.82670093e-01 1.31315455e-01
3.18216681e-01 -6.63735867e-01 -9.95191813e-01 -2.04250321e-01
-3.36756378e-01 -2.90273428e-01 4.99232113e-02 -1.07072249e-01
-9.46889400e-01 1.06176972e+00 -1.20778059e-04 2.67611712e-01
-2.47628391e-01 -6.67493999e-01 8.79624724e-01 7.02682018e-01
9.49219644e-01 3.83792073e-01 -1.20932722e+00 -3.71993035e-01
-5.50264955e-01 1.99589476e-01 1.32245624e+00 9.66747046e-01
7.09359527e-01 6.71198368e-02 5.24337530e-01 8.18284869e-01
4.70132202e-01 1.18741834e+00 7.45550931e-01 -1.34339318e-01
5.24736822e-01 1.00566959e+00 -8.98995772e-02 -1.41907644e+00
-2.39968956e-01 -4.28633422e-01 -5.89968204e-01 -1.10964999e-01
-1.62592515e-01 -5.13217509e-01 -6.61094189e-01 1.33377397e+00
2.35343501e-01 -6.94547236e-01 6.15135550e-01 8.01143169e-01
1.25667989e+00 8.24712515e-01 -1.07341597e-03 -5.10375127e-02
1.62221694e+00 -9.87255335e-01 -8.35954547e-01 -4.92642760e-01
6.47312939e-01 -1.22921741e+00 1.15489876e+00 6.93369150e-01
-7.48037398e-01 -1.45092770e-01 -1.09996891e+00 1.80294782e-01
-1.20253420e+00 2.95244157e-01 8.17729771e-01 1.09918058e+00
-7.92210102e-01 -1.22620538e-01 -3.63519549e-01 -2.98302263e-01
3.37706923e-01 7.63198808e-02 -7.02699006e-01 2.56464154e-01
-1.42186379e+00 8.44552636e-01 -3.01521540e-01 -1.65286418e-02
-8.59714001e-02 -6.79507196e-01 -1.06054103e+00 -2.50649482e-01
1.94169611e-01 -1.22230589e-01 1.26544106e+00 -1.51724732e+00
-1.43596125e+00 1.20846128e+00 -4.17373031e-01 -1.15736000e-01
-1.16953447e-01 -2.12220669e-01 -9.39856589e-01 -1.08410873e-01
4.87365782e-01 7.75836408e-02 7.01961935e-01 -8.56517196e-01
-5.88944793e-01 -6.27717853e-01 5.94813451e-02 1.72893822e-01
-8.10101509e-01 5.98731875e-01 -2.12712809e-01 -6.32705390e-01
-3.05046529e-01 -1.08866978e+00 -1.17838748e-01 -7.72454441e-01
-6.03984416e-01 -2.78753161e-01 6.68349504e-01 -5.92866182e-01
1.09222090e+00 -2.06884718e+00 -2.24049568e-01 1.68551341e-01
-2.06471905e-01 2.41494045e-01 9.30733606e-02 6.77045763e-01
-1.93654254e-01 2.42246449e-01 -1.34934083e-01 -4.12410140e-01
1.53943628e-01 -1.21964306e-01 -5.94027698e-01 2.03053683e-01
1.03568301e-01 7.40070939e-01 -6.55422807e-01 -3.22724164e-01
-2.28322208e-01 5.91684759e-01 -4.91894573e-01 2.65642349e-02
6.94977939e-02 7.77486414e-02 -7.82953739e-01 9.08043385e-01
5.24189532e-01 -2.03058571e-01 1.81185436e-02 -1.57806531e-01
-1.40665293e-01 5.29761791e-01 -8.20842266e-01 1.19444919e+00
-6.59441829e-01 8.50140631e-01 -3.84813175e-02 -7.46940136e-01
1.14000630e+00 3.53414357e-01 2.98667043e-01 -4.91833389e-01
3.39354485e-01 4.08150703e-01 5.87763228e-02 -1.88602880e-01
9.80948031e-01 -3.66999090e-01 -7.64514804e-01 6.16537929e-01
-5.08651957e-02 -4.25375044e-01 6.24128759e-01 4.98726040e-01
7.36601055e-01 -2.70843148e-01 5.12303054e-01 -3.25794876e-01
1.01853311e+00 4.40856040e-01 5.70886075e-01 2.83598211e-02
-3.78390372e-01 4.65518564e-01 6.77302241e-01 -5.76397777e-01
-6.40467048e-01 -6.14013612e-01 -1.06019653e-01 1.10827327e+00
-1.17667854e-01 -8.49853516e-01 -7.00520694e-01 -7.35160112e-01
-2.18609810e-01 5.60721397e-01 -8.84935796e-01 5.12884296e-02
-9.72196013e-02 -8.74192595e-01 1.17909543e-01 3.50825876e-01
4.87289399e-01 -1.50192499e+00 -1.81596220e-01 -1.30830720e-01
-1.66965779e-02 -1.15441287e+00 -3.53735715e-01 1.46203339e-01
-2.61819184e-01 -1.04838991e+00 -3.62260371e-01 -8.45449507e-01
5.73402584e-01 4.71223705e-02 1.57168067e+00 -2.09865272e-01
1.16519310e-01 1.42997757e-01 -7.47495234e-01 -1.14148462e+00
-4.41348493e-01 6.26260042e-01 4.73126397e-02 -5.85735142e-02
1.28083026e+00 -3.61880548e-02 -3.35114568e-01 1.42157912e-01
-5.20580232e-01 -3.28630686e-01 5.90220034e-01 5.01810312e-01
3.92630637e-01 -6.72843084e-02 9.31424618e-01 -1.62278712e+00
8.36118937e-01 -4.82719660e-01 -1.55873865e-01 1.28336400e-01
-7.30892181e-01 -5.70788443e-01 9.32621360e-01 2.78177373e-02
-1.29135203e+00 -3.65682721e-01 -3.76263648e-01 4.57847923e-01
-2.68793494e-01 1.06786883e+00 -1.39457464e-01 2.82864898e-01
4.44260120e-01 1.45740777e-01 -2.19076276e-01 -1.44572139e-01
2.79497296e-01 1.40779400e+00 -7.72322714e-03 -1.66465595e-01
5.09156644e-01 3.36469024e-01 -7.82217920e-01 -1.06955004e+00
-1.62977493e+00 -6.98877692e-01 -4.50063348e-01 -2.86464632e-01
6.92357183e-01 -1.14364040e+00 -4.07412529e-01 7.50931144e-01
-6.99172735e-01 1.62166864e-01 -2.01764777e-01 2.41081446e-01
-8.05584639e-02 7.96725452e-02 -6.84658647e-01 -6.77610219e-01
-9.21555817e-01 -1.20351565e+00 9.74522650e-01 3.13820690e-01
-6.54828608e-01 -1.07076800e+00 4.70933735e-01 7.26401746e-01
4.70681101e-01 -2.01057360e-01 6.34958506e-01 -1.00523353e+00
2.81360239e-01 -7.84841239e-01 9.31839198e-02 5.81016541e-01
3.79744858e-01 4.31783050e-01 -9.84110951e-01 5.63478656e-03
2.24450618e-01 -5.32761157e-01 4.11431402e-01 2.52952248e-01
3.73478651e-01 -1.01855911e-01 6.59994185e-02 -4.70858589e-02
1.11670387e+00 2.79483914e-01 4.81739402e-01 8.31814587e-01
5.97629011e-01 7.05095112e-01 1.09105551e+00 5.39219260e-01
7.85782516e-01 -2.55143102e-02 1.38482779e-01 -1.55765628e-02
4.50903803e-01 3.74372257e-03 7.77823627e-01 1.71610570e+00
2.88870692e-01 -2.39545494e-01 -7.70804226e-01 6.98394120e-01
-1.27870989e+00 -5.57865143e-01 -2.46222243e-01 1.56214631e+00
8.72299671e-01 4.16710585e-01 -3.90218869e-02 2.47975200e-01
5.47843575e-01 3.94444615e-01 -5.08398890e-01 -1.04699898e+00
-4.95067358e-01 5.58349416e-02 1.38972372e-01 8.02435055e-02
-1.21946490e+00 1.15754843e+00 5.88185167e+00 4.88261402e-01
-1.11965525e+00 -2.27330223e-01 8.96925926e-01 2.06368014e-01
-5.50893784e-01 -1.23645350e-01 -1.21416295e+00 4.92184192e-01
1.10043263e+00 -3.52387249e-01 -2.33978480e-01 1.37174129e+00
-2.76185032e-02 -6.65563121e-02 -4.83933240e-01 5.74283481e-01
8.03833365e-01 -9.82067645e-01 8.42620805e-02 -1.23261265e-01
1.07615650e+00 2.11561054e-01 2.89370388e-01 4.93007004e-01
2.04741746e-01 -7.27243841e-01 7.66292155e-01 7.01992437e-02
3.30103129e-01 -1.15023375e+00 1.40620792e+00 1.71913460e-01
-7.92348206e-01 2.58219004e-01 -2.62842506e-01 -2.22006276e-01
1.03121109e-01 1.00604296e+00 -3.42156947e-01 4.65514064e-01
1.15520334e+00 1.44089794e+00 -6.37143850e-01 1.64868012e-01
-5.05244017e-01 8.18450570e-01 1.14908174e-01 -6.78954959e-01
2.20345497e-01 -5.85036576e-01 2.81409144e-01 1.23724663e+00
1.52217537e-01 -2.45612234e-01 -1.58269823e-01 6.41084313e-02
-2.65946060e-01 8.90800297e-01 -7.02516317e-01 -5.32817602e-01
1.64714400e-02 1.91860974e+00 -7.17140675e-01 -4.79075462e-01
-9.30980265e-01 6.08509243e-01 6.27035201e-02 6.31348267e-02
-4.12273496e-01 -5.58994889e-01 8.47381055e-01 -2.45853528e-01
2.85562336e-01 2.13068575e-01 -4.85016763e-01 -1.50156987e+00
8.64612833e-02 -1.39600980e+00 3.28405768e-01 -8.44317138e-01
-1.76522696e+00 1.03187966e+00 -6.35962248e-01 -1.08692861e+00
-1.97558463e-01 -1.11729503e+00 -6.02099895e-01 7.31483519e-01
-1.78342521e+00 -1.16008878e+00 2.28845440e-02 2.18339473e-01
6.37182117e-01 -8.08933616e-01 9.09229875e-01 1.69382766e-01
-5.85614204e-01 4.05055344e-01 -1.68522805e-01 1.87199280e-01
1.36993468e+00 -1.41995049e+00 4.13058847e-01 4.95670915e-01
-3.44675593e-02 8.75673592e-01 8.10251474e-01 -6.12988055e-01
-1.41258943e+00 -7.01874495e-01 1.57681704e+00 -6.82401776e-01
1.20427048e+00 -3.27150851e-01 -6.88035548e-01 8.41101289e-01
7.12635458e-01 -4.61872607e-01 1.42905986e+00 4.32726115e-01
-3.40315461e-01 -2.40783662e-01 -9.15219545e-01 7.03311741e-01
1.93978682e-01 -5.14707148e-01 -7.12075353e-01 1.65740237e-01
4.61037636e-01 -2.26038963e-01 -1.01984525e+00 1.77167356e-01
5.13202310e-01 -1.07387900e+00 6.38428986e-01 -5.04388869e-01
1.01106620e+00 -2.02302277e-01 -2.24736422e-01 -1.63420236e+00
3.01954120e-01 -8.24239030e-02 2.56766796e-01 1.45588720e+00
9.19182479e-01 -8.00565422e-01 9.59643245e-01 6.05243444e-01
-1.14260539e-01 -8.72651279e-01 -4.17503983e-01 -1.76069196e-02
3.00671577e-01 -6.21684372e-01 6.43207729e-01 1.13770628e+00
4.30873364e-01 8.07390273e-01 -1.73049197e-01 -1.88838005e-01
3.32220085e-02 6.29512906e-01 9.85719323e-01 -1.14397585e+00
1.23049572e-01 -3.04821730e-01 -3.46104562e-01 -4.64543521e-01
5.43270290e-01 -7.64550030e-01 -2.07376659e-01 -1.40364265e+00
2.40098700e-01 -2.21657619e-01 -2.43186116e-01 4.47066993e-01
-1.91277996e-01 5.54523647e-01 -3.05473298e-01 5.45285735e-03
-6.12522066e-01 5.39856732e-01 1.10857391e+00 -2.10399821e-01
8.23012218e-02 8.80589932e-02 -1.48937058e+00 1.10516524e+00
1.06828189e+00 -3.47520143e-01 -5.05513430e-01 4.88260165e-02
1.24386919e+00 -6.00200653e-01 -4.93382990e-01 -3.92181873e-01
7.71340653e-02 -5.45538329e-02 2.31462181e-01 -1.03794122e+00
2.52420694e-01 -6.98937654e-01 -3.52688402e-01 1.51886150e-01
-3.28298271e-01 3.95306587e-01 6.84287176e-02 2.15341642e-01
-9.23459828e-01 -3.37269068e-01 2.90847749e-01 -1.04762934e-01
-7.26905167e-01 7.91690871e-02 -7.23336995e-01 3.15907478e-01
8.89677405e-01 2.54207812e-02 -4.26641047e-01 -4.59076285e-01
-1.49323568e-01 -3.83287814e-04 4.98165190e-01 6.91047966e-01
5.19059122e-01 -9.12845194e-01 -6.53472602e-01 3.00087154e-01
5.61056972e-01 -3.71327609e-01 1.42317891e-01 6.42440736e-01
-2.71936297e-01 4.80485201e-01 -4.52799685e-02 -1.12789728e-01
-1.23615587e+00 2.96778172e-01 -1.33635849e-01 -3.57142866e-01
-3.52163374e-01 6.88903391e-01 9.09991376e-03 -9.57961380e-01
-3.58101577e-01 -2.18235198e-02 -8.55882883e-01 6.10856712e-01
5.86514473e-01 -8.34274366e-02 3.41867983e-01 -1.11634636e+00
-5.57977200e-01 6.78479493e-01 -4.92161036e-01 -1.50373429e-01
1.39593339e+00 -1.87726825e-01 -4.99798983e-01 9.89419341e-01
1.09298193e+00 6.19217217e-01 -2.25941613e-01 1.26943588e-02
-1.28309960e-02 -2.77247727e-01 -7.78346285e-02 -9.79885578e-01
-1.26914787e+00 7.16984093e-01 -4.62884940e-02 4.39898223e-01
9.40524161e-01 -1.15718320e-01 8.58025193e-01 5.79968214e-01
1.47691458e-01 -1.51121378e+00 -7.56916925e-02 9.05664980e-01
6.94728374e-01 -1.72262156e+00 -4.40394925e-03 -5.51909506e-01
-1.31863225e+00 8.41393292e-01 5.47406077e-01 5.63069386e-03
1.18657196e+00 4.55695957e-01 1.02940059e+00 -3.74776542e-01
-8.12613070e-01 3.26971173e-01 3.89075875e-02 4.19877827e-01
1.02744925e+00 2.46308353e-02 -3.94701034e-01 1.26871657e+00
-9.81455505e-01 -3.14826369e-01 8.88586998e-01 8.71590137e-01
-1.15605913e-01 -1.16155505e+00 -5.08561097e-02 6.04632318e-01
-1.03045833e+00 -3.07827026e-01 -7.25694895e-01 7.17042148e-01
-4.54329818e-01 1.21433258e+00 -1.63727641e-01 -1.12240553e-01
4.78371024e-01 -4.33473289e-02 -4.24132079e-01 -7.24815726e-01
-1.00223494e+00 -1.75850332e-01 4.02193278e-01 -1.62399728e-02
-8.13629210e-01 -8.10851634e-01 -1.10915077e+00 -2.40459576e-01
-1.98178232e-01 6.88298106e-01 1.06950843e+00 1.03515005e+00
2.60871768e-01 3.12457561e-01 7.91414976e-01 -2.38974944e-01
-1.02736004e-01 -9.60927069e-01 -9.73193526e-01 4.25208211e-01
6.17687590e-02 -2.07776636e-01 -6.81426764e-01 1.10236600e-01]
|
[11.226937294006348, 6.875585079193115]
|
f43edd6b-64cd-435a-9e92-f9759797a35d
|
attribute-consistent-knowledge-graph
|
2304.01563
| null |
https://arxiv.org/abs/2304.01563v1
|
https://arxiv.org/pdf/2304.01563v1.pdf
|
Attribute-Consistent Knowledge Graph Representation Learning for Multi-Modal Entity Alignment
|
The multi-modal entity alignment (MMEA) aims to find all equivalent entity pairs between multi-modal knowledge graphs (MMKGs). Rich attributes and neighboring entities are valuable for the alignment task, but existing works ignore contextual gap problems that the aligned entities have different numbers of attributes on specific modality when learning entity representations. In this paper, we propose a novel attribute-consistent knowledge graph representation learning framework for MMEA (ACK-MMEA) to compensate the contextual gaps through incorporating consistent alignment knowledge. Attribute-consistent KGs (ACKGs) are first constructed via multi-modal attribute uniformization with merge and generate operators so that each entity has one and only one uniform feature in each modality. The ACKGs are then fed into a relation-aware graph neural network with random dropouts, to obtain aggregated relation representations and robust entity representations. In order to evaluate the ACK-MMEA facilitated for entity alignment, we specially design a joint alignment loss for both entity and attribute evaluation. Extensive experiments conducted on two benchmark datasets show that our approach achieves excellent performance compared to its competitors.
|
['JianXin Li', 'Jiawei Sheng', 'Lihong Wang', 'Cheng Ji', 'Yangyifei Luo', 'Shu Guo', 'Qian Li']
|
2023-04-04
| null | null | null | null |
['multi-modal-entity-alignment', 'entity-alignment', 'entity-alignment']
|
['knowledge-base', 'knowledge-base', 'natural-language-processing']
|
[ 5.16246189e-04 4.66994822e-01 -4.96118516e-01 -5.92471182e-01
-8.81246448e-01 -2.53160566e-01 3.02394718e-01 5.27376711e-01
-1.55014321e-01 8.38921010e-01 2.17122048e-01 2.73443982e-02
-4.31014687e-01 -1.12491441e+00 -9.28865910e-01 -4.73886609e-01
-1.14356503e-01 7.79942513e-01 4.15106900e-02 -1.21966349e-02
-4.55553949e-01 2.89267190e-02 -1.08132946e+00 1.75958022e-01
1.28095853e+00 9.80762005e-01 -4.68194298e-02 1.67140979e-02
-2.47349426e-01 9.57047462e-01 -3.00347388e-01 -1.04437923e+00
1.59353420e-01 -1.68129697e-01 -1.20335078e+00 -1.95391312e-01
4.92246866e-01 3.15998904e-02 -4.93546307e-01 1.10324430e+00
7.41494715e-01 2.82025665e-01 8.80346179e-01 -1.72130561e+00
-1.23933506e+00 1.26945710e+00 -6.40899003e-01 -2.61772964e-02
3.14674854e-01 -3.07457626e-01 1.61061180e+00 -8.87139082e-01
6.91887021e-01 1.35079932e+00 7.72776544e-01 1.27641395e-01
-1.16387892e+00 -6.73768759e-01 5.27972281e-01 5.46456337e-01
-1.79122055e+00 -1.35518387e-01 7.19074368e-01 -8.83259550e-02
8.88356507e-01 1.52676329e-01 3.41773838e-01 1.04109418e+00
-1.59826756e-01 9.16499257e-01 5.64335346e-01 -2.30380818e-01
-1.71731859e-01 9.86595824e-02 3.99105251e-01 7.50521302e-01
5.83659589e-01 -4.55751389e-01 -4.01018202e-01 -1.07467994e-01
5.52902520e-01 -1.57516465e-01 -3.34512562e-01 -6.49715304e-01
-1.38775301e+00 6.23289943e-01 9.96919274e-01 1.68718457e-01
-6.25769377e-01 9.27433372e-02 5.18973053e-01 3.47369105e-01
2.27754995e-01 5.09595931e-01 -5.41337192e-01 6.22626781e-01
-2.59766787e-01 2.04329863e-01 4.86030251e-01 1.50471246e+00
1.02373254e+00 -1.81602880e-01 -5.44249237e-01 9.42581892e-01
2.92305171e-01 3.85476351e-01 2.92201012e-01 -5.06583512e-01
9.90102649e-01 1.18463230e+00 -2.53125429e-01 -1.29506433e+00
-6.36105299e-01 -6.92653000e-01 -1.20812321e+00 -6.21131957e-01
9.59318597e-03 -4.92795371e-02 -9.10880148e-01 2.15284491e+00
5.60910881e-01 4.49838519e-01 5.50021768e-01 6.68208241e-01
1.48157477e+00 5.10834038e-01 6.70865595e-01 -6.88844174e-02
1.45612121e+00 -9.84812856e-01 -8.18722486e-01 -2.15542942e-01
8.97231519e-01 -3.30389112e-01 9.68097389e-01 -4.53315467e-01
-8.49322796e-01 -3.52375060e-01 -8.91797960e-01 -2.06266940e-01
-5.99536419e-01 4.25757421e-03 8.36375058e-01 2.81425238e-01
-6.97504997e-01 4.15336043e-01 -4.37623203e-01 -1.76867247e-01
5.55570483e-01 4.42908108e-01 -8.27666283e-01 -1.98787808e-01
-1.89066184e+00 9.29116964e-01 9.23346162e-01 1.73880488e-01
-2.45526940e-01 -8.42254639e-01 -1.35134757e+00 1.96857631e-01
5.49381196e-01 -1.30350387e+00 7.28103101e-01 -7.10307002e-01
-6.04329944e-01 9.74283695e-01 -9.50917751e-02 -3.54900986e-01
1.07174411e-01 -1.93836123e-01 -8.41690421e-01 -9.89394784e-02
4.61551696e-01 5.71007133e-01 2.12541401e-01 -1.37704134e+00
-5.69249332e-01 -4.82913524e-01 1.75080642e-01 6.44921660e-01
-2.18492642e-01 -3.31653476e-01 -7.28434205e-01 -6.22571588e-01
4.01209205e-01 -7.37730265e-01 -1.77904308e-01 -6.64582491e-01
-1.03024268e+00 -4.59793091e-01 4.48045492e-01 -7.25807190e-01
1.31650817e+00 -1.75073826e+00 4.60681885e-01 4.19530958e-01
3.55331212e-01 -5.18074222e-02 -2.47444287e-01 2.60105103e-01
-2.68085241e-01 -5.59773929e-02 -1.76572815e-01 -2.93271840e-01
1.35927320e-01 2.67210275e-01 -8.34053457e-02 3.24630812e-02
2.70845234e-01 1.32250226e+00 -8.55583549e-01 -8.58384430e-01
-2.56702274e-01 4.08268034e-01 -2.92888671e-01 2.12736815e-01
-4.92096394e-02 2.13520378e-01 -6.92382812e-01 1.08680439e+00
7.77955770e-01 -6.10349536e-01 4.02946144e-01 -9.26165640e-01
6.85818195e-01 4.44223695e-02 -1.41717386e+00 1.68785310e+00
-3.47836912e-01 3.63172293e-02 -4.49614108e-01 -9.39586341e-01
9.45931256e-01 2.43927285e-01 5.50584137e-01 -6.74361289e-01
5.86159565e-02 2.65314639e-01 -2.28992358e-01 -3.90909165e-01
7.63869405e-01 1.50066093e-01 -3.45092237e-01 -1.28961941e-02
3.69032264e-01 5.22325575e-01 9.69510749e-02 5.96216261e-01
7.76195645e-01 -3.61592788e-03 3.71844769e-01 1.07694134e-01
5.82222342e-01 -2.62144148e-01 9.54572201e-01 4.96326894e-01
-6.03912957e-03 4.03848559e-01 4.75602359e-01 -3.90856564e-01
-7.01744437e-01 -1.21108913e+00 -4.82323356e-02 1.04305387e+00
6.57166302e-01 -4.92358983e-01 -3.44247669e-01 -1.04138005e+00
2.85901576e-01 6.17167830e-01 -6.75669730e-01 -5.64310253e-01
-4.43883061e-01 -9.84369278e-01 4.59769785e-01 9.26957548e-01
6.77790225e-01 -9.59001184e-01 3.64900112e-01 2.08000287e-01
-5.42904019e-01 -1.25329101e+00 -4.42724764e-01 1.02360032e-01
-4.27853912e-01 -1.18691957e+00 -5.22972167e-01 -1.14475179e+00
8.24747205e-01 -1.29531339e-01 1.34883916e+00 -1.63538992e-01
1.44899068e-02 2.60992318e-01 -2.83244789e-01 -1.15743183e-01
2.51912642e-02 6.09683692e-01 6.72624409e-02 1.41548216e-01
5.54653704e-01 -5.93128741e-01 -4.23557013e-01 2.65274376e-01
-5.81370234e-01 7.41100982e-02 9.42952693e-01 9.13693726e-01
1.09039795e+00 -9.78817195e-02 9.87902522e-01 -1.36392725e+00
5.67960262e-01 -8.11707377e-01 -1.62000686e-01 8.38699281e-01
-7.90297329e-01 2.63547629e-01 4.23472971e-01 -2.24222243e-01
-9.17830884e-01 1.51316682e-02 1.51588440e-01 -5.06757259e-01
1.79913267e-01 9.31396604e-01 -8.83165956e-01 2.03251347e-01
3.16121340e-01 3.34387906e-02 -5.71871400e-01 -2.86676556e-01
7.08053172e-01 3.93364191e-01 8.27896237e-01 -8.13860774e-01
6.91572070e-01 -1.42545938e-01 1.55510083e-01 -1.39421508e-01
-9.91306782e-01 -3.33355933e-01 -6.64801061e-01 1.64076447e-01
8.45915020e-01 -1.14003658e+00 -5.44488013e-01 5.02049565e-01
-8.61075580e-01 1.90182194e-01 -2.24505514e-01 6.24550343e-01
-3.29518318e-01 1.85098886e-01 -4.14718807e-01 -4.22691137e-01
-6.06519520e-01 -9.00704265e-01 9.64032650e-01 4.86225098e-01
-1.10833138e-01 -1.03624976e+00 -8.15164000e-02 3.47722590e-01
1.78891793e-01 3.56997848e-01 1.23932314e+00 -1.17167068e+00
-6.69211090e-01 -1.95137829e-01 -5.59248626e-01 -1.27188027e-01
2.36573741e-01 -2.45179787e-01 -6.92830980e-01 -1.76236838e-01
-8.79695833e-01 -4.53495055e-01 9.41148818e-01 2.62855887e-01
9.34335947e-01 -4.16979253e-01 -8.00816238e-01 8.01916957e-01
1.43876386e+00 -1.83729261e-01 5.44361234e-01 4.82154995e-01
1.34175372e+00 5.01288891e-01 8.33185852e-01 8.34317952e-02
1.23012662e+00 6.93038166e-01 4.52197641e-01 -8.91130939e-02
-6.98898286e-02 -4.69087511e-01 -1.57974124e-01 8.30355883e-01
-1.62649095e-01 -4.03738290e-01 -6.99114501e-01 7.50052631e-01
-2.01582408e+00 -8.37209165e-01 -2.92600274e-01 2.13552499e+00
9.96811688e-01 -2.02680111e-01 1.95339210e-02 -2.84017712e-01
1.10965681e+00 3.97201069e-02 -6.67015910e-01 1.94870666e-01
-5.78605950e-01 -1.87654927e-01 4.72068727e-01 2.48043552e-01
-1.49262488e+00 1.02193141e+00 4.72204924e+00 7.71823347e-01
-4.89832103e-01 -1.38794735e-01 5.45958698e-01 4.65503216e-01
-6.94192886e-01 -4.53133732e-02 -8.95671487e-01 3.75360608e-01
4.83262092e-01 -4.29926813e-01 -1.25182793e-01 8.54051948e-01
-7.00390160e-01 3.81455451e-01 -1.11906266e+00 9.86246347e-01
-1.68392524e-01 -1.23247886e+00 4.28129524e-01 -2.81562597e-01
7.75039375e-01 -2.37464562e-01 -3.82546224e-02 7.23126948e-01
5.55404782e-01 -9.85959530e-01 1.79029942e-01 7.16293752e-01
7.40458608e-01 -9.51314390e-01 1.09271550e+00 -3.02451462e-01
-1.86735380e+00 9.14686695e-02 -4.95148867e-01 7.51532614e-01
1.86944634e-01 3.98112983e-01 -6.76972568e-01 1.56409431e+00
6.12291515e-01 7.70918846e-01 -7.19753742e-01 9.71967220e-01
-6.72434866e-02 1.18702903e-01 -1.77164659e-01 3.33935589e-01
-3.77239473e-02 -1.49645835e-01 3.55669886e-01 9.85738933e-01
2.66603380e-01 5.32682762e-02 3.16451550e-01 7.41779208e-01
-6.51318967e-01 4.83418465e-01 -4.36331660e-01 -1.35913247e-03
1.13297558e+00 1.29669666e+00 -2.86172539e-01 -2.10359633e-01
-5.70113301e-01 9.96218562e-01 9.29575920e-01 3.81232977e-01
-7.93207288e-01 -5.60823798e-01 5.71546316e-01 -3.77586782e-01
1.90224573e-01 4.60122406e-01 -8.22184756e-02 -1.20613468e+00
1.29265949e-01 -5.80610454e-01 1.30747640e+00 -7.82559156e-01
-1.84217584e+00 7.64525831e-01 1.21685127e-02 -1.12501109e+00
-1.38215208e-02 -1.97246581e-01 -4.37549412e-01 9.22230721e-01
-1.65173757e+00 -1.93280196e+00 -4.74411428e-01 8.29529464e-01
-1.97560221e-01 -3.54346395e-01 8.76077592e-01 6.35028362e-01
-8.99203360e-01 1.24716640e+00 -2.95559913e-01 4.57521081e-01
8.37631881e-01 -1.53514981e+00 3.98324817e-01 6.13764048e-01
1.16891198e-01 7.15077937e-01 5.18110454e-01 -8.74654531e-01
-1.18869722e+00 -1.58988535e+00 9.80362773e-01 -3.54955614e-01
4.27788347e-01 1.05490923e-01 -1.33557034e+00 1.25983489e+00
1.34578437e-01 3.43923390e-01 8.57477367e-01 5.26892066e-01
-6.44314110e-01 -2.04935893e-01 -1.05168569e+00 4.87636179e-01
1.23025036e+00 -4.14242327e-01 -6.72332227e-01 3.21382642e-01
1.06722808e+00 -5.13131082e-01 -1.42645264e+00 8.75306666e-01
2.21689016e-01 -3.08962226e-01 1.14324093e+00 -1.09088349e+00
2.58199990e-01 -4.48716640e-01 -3.70237708e-01 -1.51622343e+00
-3.89497995e-01 -1.92212895e-01 -3.64469200e-01 1.88996649e+00
7.93499291e-01 -6.75957918e-01 6.50187075e-01 7.02842236e-01
-1.65590420e-01 -8.83564830e-01 -7.00331509e-01 -8.05853426e-01
-5.44468686e-02 9.47684050e-02 1.16708279e+00 1.66294086e+00
-5.71453534e-02 8.58599901e-01 -4.16840672e-01 8.26487005e-01
5.95758915e-01 6.27411723e-01 6.26270413e-01 -1.28229940e+00
-6.33931533e-02 -2.82190114e-01 -7.88829923e-01 -6.44962370e-01
4.54997182e-01 -1.36647403e+00 -4.78396833e-01 -1.89741921e+00
4.91292059e-01 -7.94590890e-01 -5.53704143e-01 7.44117022e-01
-8.91994774e-01 -1.31881714e-01 -1.93730235e-01 1.06820077e-01
-8.62846315e-01 8.83614421e-01 9.58345592e-01 -3.01513314e-01
-1.25624880e-01 -9.35898349e-02 -9.55532551e-01 4.28949714e-01
5.10185599e-01 -3.66929501e-01 -5.95690787e-01 -3.29521745e-01
3.29761535e-01 -5.71554489e-02 2.62211204e-01 -7.47870564e-01
4.56433803e-01 4.13575508e-02 4.34194088e-01 -5.86164832e-01
1.18779421e-01 -9.38525021e-01 4.59602714e-01 -8.81166384e-02
-5.34069896e-01 1.38720170e-01 -8.98953006e-02 8.51718545e-01
-4.81695622e-01 6.00237362e-02 4.18334424e-01 1.46071628e-01
-1.10001945e+00 7.54589736e-01 6.44756079e-01 4.64279830e-01
1.15064991e+00 4.17486113e-03 -6.55774653e-01 -1.78634405e-01
-1.03697073e+00 8.55313480e-01 1.98178574e-01 4.31828350e-01
5.13069868e-01 -1.93162632e+00 -9.44974065e-01 -1.24629870e-01
5.97922504e-01 4.51059669e-01 6.28503442e-01 7.53272891e-01
-7.87470937e-02 1.54952733e-02 -2.06066534e-01 -3.34372401e-01
-1.25274026e+00 6.81003749e-01 4.57426310e-01 -5.38595855e-01
-5.99688530e-01 1.16448581e+00 2.30061322e-01 -8.74770701e-01
2.91185856e-01 1.14525661e-01 -5.31463683e-01 2.26470262e-01
7.83197582e-02 1.67987213e-01 1.22857377e-01 -8.71896327e-01
-5.88281214e-01 3.78575534e-01 -4.84925479e-01 5.72894394e-01
1.17984736e+00 -2.00527087e-01 -1.77076772e-01 2.68682301e-01
1.12680447e+00 -3.47064920e-02 -6.53353512e-01 -7.13304579e-01
1.26811862e-01 -3.34401399e-01 -2.81728178e-01 -7.05547392e-01
-1.44243670e+00 2.75752187e-01 2.24758670e-01 -3.12624574e-02
1.13007390e+00 4.51631099e-01 8.23509991e-01 5.35670698e-01
3.16769928e-01 -8.66484046e-01 -3.38879973e-01 2.01037422e-01
7.94934869e-01 -1.29391646e+00 1.01192981e-01 -9.16495919e-01
-1.10998809e+00 5.82962573e-01 1.21353447e+00 2.94839829e-01
3.86993676e-01 -2.55601019e-01 -9.15474519e-02 -4.99090075e-01
-5.09037852e-01 -5.24250984e-01 5.67511559e-01 8.18891048e-01
2.96553105e-01 1.85592055e-01 -1.76710561e-01 1.18765736e+00
-9.05019715e-02 -4.49759454e-01 4.46764007e-02 5.98920405e-01
-7.76312947e-02 -9.49791312e-01 2.33189195e-01 5.89566827e-01
-1.79879218e-01 -2.48673126e-01 -5.33375919e-01 8.64524305e-01
7.28383614e-03 4.72182184e-01 -8.56641009e-02 -5.58839500e-01
6.38162851e-01 1.73868313e-01 3.19398850e-01 -3.57656300e-01
-4.06798363e-01 -5.68159461e-01 4.24649894e-01 -4.48565513e-01
-3.89016211e-01 -5.03566086e-01 -1.45179534e+00 -1.72262087e-01
-5.76566279e-01 2.55656451e-01 -1.17827229e-01 8.36155593e-01
6.98736131e-01 8.45382810e-01 5.69871366e-01 -1.27014235e-01
-2.17909783e-01 -8.57837558e-01 -8.55683506e-01 8.48647416e-01
-6.33715168e-02 -8.52062523e-01 1.03842784e-02 -2.30258211e-01]
|
[8.794084548950195, 8.026026725769043]
|
b186bddb-f6da-4220-bc24-241740f90042
|
green-portfolio-optimization-a-scenario
|
2305.16712
| null |
https://arxiv.org/abs/2305.16712v1
|
https://arxiv.org/pdf/2305.16712v1.pdf
|
Green portfolio optimization: A scenario analysis and stress testing based novel approach for sustainable investing in the paradigm Indian markets
|
In this article, we present a novel approach for the construction of an environment-friendly green portfolio using the ESG ratings, and application of the modern portfolio theory to present what we call as the ``green efficient frontier'' (wherein the environmental score is included as a third dimension to the traditional mean-variance framework). Based on the prevailing action levels and policies, as well as additional market information, scenario analyses and stress testing are conducted to anticipate the future performance of the green portfolio in varying circumstances. The performance of the green portfolio is evaluated against the market returns in order to highlight the importance of sustainable investing and recognizing climate risk as a significant risk factor in financial analysis.
|
['Siddhartha P. Chakrabarty', 'Rishabh Raj', 'Shashwat Mishra']
|
2023-05-26
| null | null | null | null |
['portfolio-optimization']
|
['time-series']
|
[-9.67115723e-03 -1.29343709e-03 2.26351544e-02 1.93996415e-01
-2.16789380e-01 -7.50516057e-01 5.99354863e-01 2.32538477e-01
-1.97072908e-01 4.43974882e-01 4.24283028e-01 -8.61183643e-01
-8.65172267e-01 -9.25001860e-01 -2.93352008e-01 -7.47465372e-01
1.15544543e-01 -4.57592398e-01 -2.86823183e-01 -2.17958599e-01
9.11991894e-01 6.17572248e-01 -1.35800922e+00 -3.94901037e-01
1.00192809e+00 1.17804003e+00 2.29024827e-01 3.43414962e-01
6.53041080e-02 3.60593051e-01 -4.53177989e-01 -5.76518118e-01
5.64176202e-01 -1.99938998e-01 -1.47858381e-01 -6.57554209e-01
-3.75181466e-01 -3.57564628e-01 6.01003230e-01 1.15819299e+00
6.21498346e-01 2.41962343e-01 4.35876966e-01 -8.68993223e-01
-5.71508169e-01 5.82411945e-01 -1.13383740e-01 2.63056487e-01
1.00999586e-01 3.71470183e-01 1.05069649e+00 -8.04916739e-01
1.92697719e-01 8.75964046e-01 3.10186177e-01 -2.08105519e-01
-9.14889395e-01 -6.27837062e-01 2.00754449e-01 1.31714195e-02
-9.68837380e-01 -2.20228791e-01 7.00858057e-01 -5.99053144e-01
1.18165553e+00 5.08042157e-01 1.07903111e+00 3.28827024e-01
7.41051495e-01 -2.86551207e-01 1.47035742e+00 -7.52835512e-01
6.48373127e-01 -5.15445322e-02 -3.33469033e-01 -2.97631770e-01
9.43181098e-01 8.04947555e-01 -2.39340112e-01 -1.33581564e-01
1.40402257e-01 -2.19875813e-01 8.52858275e-02 -7.27743357e-02
-7.32555985e-01 5.60863733e-01 2.24708229e-01 1.75654501e-01
-7.14961290e-01 1.99234664e-01 2.40469668e-02 6.28188401e-02
5.83419085e-01 6.16910577e-01 -2.04210728e-01 5.44569008e-02
-7.21081197e-01 1.86182261e-01 5.90080500e-01 4.01423991e-01
3.61514568e-01 3.10611516e-01 -3.10593486e-01 3.98014039e-02
9.75289524e-01 9.23295081e-01 -1.88722983e-01 -1.47006714e+00
3.16989660e-01 5.99457443e-01 6.70172870e-01 -7.65513062e-01
-1.28331602e-01 -8.66725445e-01 -1.61787216e-02 6.13486946e-01
-6.10697716e-02 -2.94560850e-01 -4.77150559e-01 1.32208669e+00
1.33422479e-01 -2.21088469e-01 1.16244636e-01 1.70436680e-01
2.69380342e-02 6.72256947e-01 4.20517504e-01 -4.47352141e-01
8.41021717e-01 -3.28198105e-01 -7.25594699e-01 -3.38247828e-02
2.93635368e-01 -5.08391678e-01 5.75747073e-01 5.23322463e-01
-9.55530822e-01 -1.23917177e-01 -8.43184531e-01 9.92475569e-01
-6.36121809e-01 -4.41543639e-01 6.51299536e-01 1.10097957e+00
-9.86856699e-01 7.18087375e-01 -3.01117688e-01 -1.23533539e-01
-4.76909876e-02 -1.39151335e-01 3.31195086e-01 3.42899472e-01
-1.16712248e+00 1.42048347e+00 4.83554989e-01 3.39486986e-01
-6.72600508e-01 -6.97806954e-01 -5.03582895e-01 3.63248378e-01
6.77929446e-02 -2.01189682e-01 8.68285060e-01 -4.31452572e-01
-1.16911519e+00 2.20042974e-01 6.67723477e-01 -3.37107062e-01
7.52286613e-01 -4.91507173e-01 -8.08305681e-01 6.86250627e-02
-3.22674238e-03 -2.95287281e-01 1.05812348e-01 -1.07971692e+00
-9.39318955e-01 -3.32272649e-01 -9.24886204e-03 1.19263247e-01
-7.07536414e-02 4.63526338e-01 7.12482631e-01 -6.72858477e-01
-2.22784385e-01 -9.81775939e-01 -4.28235233e-01 -6.42809391e-01
-5.09817377e-02 3.62451114e-02 1.77448958e-01 -8.34592223e-01
1.58464062e+00 -1.87518835e+00 -6.05200887e-01 7.26450503e-01
-4.97649699e-01 8.69354010e-02 3.47604752e-01 9.39053118e-01
-2.33951628e-01 3.43221307e-01 -1.70423523e-01 6.10612690e-01
5.07499337e-01 -2.19060406e-01 -4.63166386e-01 3.17545980e-01
2.33771503e-01 4.24228907e-01 -9.10605907e-01 2.26448685e-01
5.17544448e-01 4.07489613e-02 -1.56199977e-01 2.04765320e-01
-3.25823463e-02 3.04671377e-01 -3.81058484e-01 8.04228663e-01
6.61808372e-01 3.39350581e-01 2.39912286e-01 -2.91118212e-02
-9.27067518e-01 1.93459406e-01 -1.11467731e+00 8.49846244e-01
-2.81564683e-01 1.33807361e-01 -2.45926693e-01 -6.17776930e-01
1.16129160e+00 2.47036561e-01 3.60571504e-01 -1.05437458e+00
-2.22090632e-01 6.46078289e-01 1.46545574e-01 -4.99259025e-01
5.23473918e-01 -3.55249405e-01 -8.15302879e-02 4.53004092e-01
-3.06040972e-01 3.94645520e-02 1.68422669e-01 -3.57065827e-01
7.18606830e-01 6.67026758e-01 3.55862021e-01 -7.93716550e-01
3.35200787e-01 -2.62473553e-01 6.81704462e-01 1.54469445e-01
-4.01498318e-01 -2.41396129e-01 4.48881388e-01 2.96974424e-02
-8.28387380e-01 -1.03069127e+00 -1.17621139e-01 8.70752037e-01
-9.54738334e-02 1.44071996e-01 -4.92690384e-01 -1.96441427e-01
4.04109836e-01 1.66799068e+00 -4.65486884e-01 -2.44631097e-01
-2.17143491e-01 -1.11993253e+00 9.04905573e-02 4.45125133e-01
3.45227271e-01 -6.49930894e-01 -1.42979240e+00 4.21412349e-01
5.62133014e-01 -8.68302733e-02 2.03884900e-01 5.74651361e-01
-6.17480814e-01 -1.13671637e+00 -7.60224044e-01 2.50854194e-01
3.67149144e-01 1.24079973e-01 1.02774394e+00 -3.97801548e-01
1.07126772e-01 3.97637367e-01 -2.98702598e-01 -9.81576383e-01
-3.14658791e-01 -6.68753624e-01 -1.90504670e-01 -3.77662033e-01
1.84823379e-01 -2.88205802e-01 -9.58653390e-01 9.87529159e-02
-6.63161039e-01 -4.36915487e-01 5.78197360e-01 1.64917931e-01
4.22172874e-01 4.39869404e-01 1.23616147e+00 -5.17029166e-01
8.70065033e-01 -6.06688142e-01 -1.04450762e+00 5.49639225e-01
-1.56250834e+00 -1.94526702e-01 2.96597164e-02 1.94346666e-01
-1.71625292e+00 -5.85621238e-01 1.51732489e-01 1.32711649e-01
1.47046775e-01 9.42120433e-01 -3.12602192e-01 -1.37690395e-01
1.72616854e-01 -4.60226089e-01 -2.99597621e-01 -6.21987641e-01
3.85147482e-01 2.83769101e-01 4.37728614e-01 -6.68961525e-01
1.05026042e+00 7.56508261e-02 2.54390180e-01 1.27127066e-01
-4.57673758e-01 -1.74471408e-01 -3.41199577e-01 -9.17337239e-01
5.13074160e-01 -7.04867005e-01 -4.88896370e-01 2.01110721e-01
-5.34212708e-01 -2.74811566e-01 -5.40185273e-01 8.29909980e-01
-5.23457229e-01 -1.48579746e-01 2.12301880e-01 -1.49995410e+00
-4.76194412e-01 -8.55708361e-01 -3.48533206e-02 3.84957284e-01
-6.51688054e-02 -1.06402636e+00 4.09255445e-01 1.26567751e-01
9.53318655e-01 1.08634162e+00 8.96630466e-01 -4.46081907e-01
-6.13969028e-01 -2.02705711e-01 6.11877954e-03 5.96007943e-01
1.53075099e-01 5.45575380e-01 -7.18738556e-01 -1.34691969e-01
2.67695040e-01 4.85885292e-01 4.84771729e-01 2.65229821e-01
4.65255529e-01 -4.09274548e-01 5.50345965e-02 3.36246312e-01
2.04757833e+00 8.89213502e-01 6.67390108e-01 1.05147254e+00
-2.93787615e-03 1.36438227e+00 1.13600767e+00 7.67128289e-01
9.03716907e-02 1.38764560e-01 7.79610276e-01 4.39776391e-01
3.10013831e-01 2.74663940e-02 4.01690066e-01 6.71999395e-01
-4.34017211e-01 -5.44635020e-02 -1.06954050e+00 3.02518785e-01
-1.32037556e+00 -1.13227272e+00 3.79088372e-02 2.48336530e+00
2.24087521e-01 4.58371550e-01 1.62276432e-01 1.11022815e-02
7.40915835e-01 1.83722317e-01 -4.32825744e-01 -8.18365991e-01
-1.52887613e-01 2.46017426e-01 1.05019021e+00 -4.04398106e-02
-6.61532223e-01 9.13569480e-02 7.77103138e+00 3.30162555e-01
-8.13356638e-01 -3.17377746e-01 7.14538217e-01 1.98722966e-02
-1.13916218e+00 4.71130580e-01 -5.10979116e-01 6.07738495e-01
1.44468951e+00 -9.98580813e-01 1.30853698e-01 3.83240163e-01
7.61449158e-01 -5.65931380e-01 -4.74069059e-01 -3.34995426e-03
-4.50093389e-01 -1.12817693e+00 -2.85587460e-01 4.59384173e-01
6.26256466e-01 -2.67586917e-01 3.13438803e-01 4.38709185e-02
4.33099478e-01 -6.49568737e-01 1.24438107e+00 1.04354715e+00
4.88966197e-01 -1.19415367e+00 8.50439429e-01 -4.01343741e-02
-1.24571419e+00 -5.93494058e-01 -1.11125380e-01 -1.63708776e-01
8.26230496e-02 6.13951683e-01 -6.96096644e-02 8.94174695e-01
7.73783386e-01 5.54770343e-02 -4.80128229e-01 1.10082984e+00
-4.75584008e-02 7.04437137e-01 -6.72163740e-02 2.92722821e-01
2.23412424e-01 -8.34427893e-01 3.97911251e-01 1.01990366e+00
1.02246225e+00 -4.91481125e-02 -6.97199583e-01 1.20454681e+00
4.43447679e-01 2.22584918e-01 -6.31732762e-01 -3.38602930e-01
7.77770340e-01 1.05052412e+00 -6.13102674e-01 2.36496076e-01
-5.94425201e-01 -1.10438382e-02 -5.06019354e-01 3.20730448e-01
-4.94962692e-01 -4.53495651e-01 3.15904617e-01 -2.08666250e-01
1.69647634e-01 5.22189774e-02 -6.06238782e-01 -2.59809494e-01
-1.44312188e-01 -3.45371634e-01 2.68387109e-01 -7.04859674e-01
-9.03254330e-01 -2.56137252e-01 4.38565731e-01 -1.21529055e+00
-1.78232640e-02 -4.85851258e-01 -1.23383439e+00 1.45664573e+00
-1.72330236e+00 -7.27577746e-01 -8.35723877e-02 -2.03338429e-01
-3.85233462e-01 -1.82747304e-01 6.78125918e-01 -6.37205392e-02
-5.74221313e-01 1.25218585e-01 7.86937416e-01 -6.31953299e-01
3.50165606e-01 -1.44912374e+00 2.69840062e-01 1.02181304e+00
-8.20094347e-01 4.52699989e-01 7.54372835e-01 -1.13063669e+00
-1.06663764e+00 -7.98592865e-01 5.84166288e-01 -1.21138617e-01
1.02617133e+00 3.77757668e-01 -5.34417629e-01 3.28027904e-01
3.82651031e-01 -8.26187551e-01 1.31812465e+00 -1.46641791e-01
1.27210379e-01 -4.56907272e-01 -1.51958394e+00 4.28950042e-01
6.41395867e-01 -3.79913330e-01 -5.92872143e-01 -2.50683367e-01
6.44226193e-01 2.88906872e-01 -1.41252291e+00 4.66145426e-01
9.27915275e-01 -1.00675511e+00 7.28147089e-01 -2.31710553e-01
-1.21676931e-02 -1.93093196e-01 -5.55591047e-01 -1.21467113e+00
-5.41772962e-01 -8.11926961e-01 1.88571572e-01 1.56778324e+00
3.78685385e-01 -9.71357584e-01 3.12312216e-01 1.09240437e+00
-3.03507090e-01 -6.87678158e-01 -1.20520091e+00 -1.07676625e+00
2.10080013e-01 -2.60480046e-01 1.03383803e+00 6.22893512e-01
-6.24699146e-02 -7.40604997e-01 -7.36057088e-02 3.11008021e-02
1.02916336e+00 7.06514269e-02 2.18343645e-01 -1.21262169e+00
1.26463711e-01 -7.43366361e-01 4.03992012e-02 6.61885560e-01
-3.30141187e-01 -5.26676655e-01 -3.20476890e-01 -1.51731420e+00
4.23712395e-02 -7.93654993e-02 -1.06243491e+00 -7.41546154e-02
-2.92477965e-01 -3.53113204e-01 7.77364552e-01 1.05940260e-01
2.28666067e-01 6.16789877e-01 7.81949043e-01 1.10567175e-01
3.07651912e-03 3.04711938e-01 -1.24285805e+00 6.92903459e-01
9.24433410e-01 -3.00324142e-01 -2.68234521e-01 2.09604055e-01
9.68818486e-01 -6.07394911e-02 3.16884637e-01 -9.74862516e-01
-3.48074108e-01 -1.06830335e+00 1.14322171e-01 -8.65664363e-01
-3.93931597e-01 -1.06138134e+00 1.17018378e+00 8.45897257e-01
-2.70650268e-01 2.81944931e-01 1.38323799e-01 4.64813441e-01
1.00102104e-01 -4.73606974e-01 5.16176045e-01 -1.58563033e-02
-5.46552360e-01 -2.57857412e-01 -2.26645365e-01 -4.22859818e-01
1.47432649e+00 -6.37826264e-01 -7.53734231e-01 1.69006526e-01
-6.28923655e-01 3.79352987e-01 7.74628103e-01 8.64348412e-02
1.46625966e-01 -1.38141632e+00 -6.52570963e-01 -4.93587136e-01
-7.90545121e-02 -5.69859803e-01 3.92416924e-01 4.93521273e-01
-7.57537663e-01 6.06162548e-01 -5.62659919e-01 4.59532320e-01
-3.59152883e-01 5.56861877e-01 5.44995308e-01 -2.44609475e-01
-3.74481708e-01 2.62326419e-01 3.31699699e-01 1.55040160e-01
-9.69706178e-02 -3.55826728e-02 -3.62574130e-01 4.16905403e-01
4.07761246e-01 1.18283820e+00 -1.23409024e-02 -4.92746145e-01
-5.84311426e-01 2.98371166e-01 7.88450181e-01 -3.84842098e-01
1.57855916e+00 -4.58332866e-01 -2.35069245e-01 7.45670557e-01
4.24272746e-01 3.04377645e-01 -1.46754491e+00 2.75804043e-01
6.83602095e-01 -6.80516005e-01 -1.06156524e-02 -1.31652701e+00
-9.15008605e-01 5.47533453e-01 9.73813415e-01 2.92495936e-01
1.29174197e+00 -6.40184343e-01 1.03033938e-01 1.21501677e-01
2.07018763e-01 -1.93758428e+00 -2.50026196e-01 -1.58582240e-01
1.14096200e+00 -4.80701953e-01 3.35777760e-01 -7.67751457e-03
-2.79577732e-01 9.83613074e-01 1.96605280e-01 -6.97322413e-02
8.38303566e-01 1.86531860e-02 -2.48810589e-01 8.62366259e-02
-4.82716084e-01 -1.27042204e-01 1.13916166e-01 6.42612338e-01
2.64071494e-01 6.01521492e-01 -7.05537379e-01 7.19038486e-01
6.65881187e-02 -1.09569944e-01 6.59525454e-01 9.57416356e-01
-8.23850989e-01 -8.93813193e-01 -6.49439216e-01 5.58737338e-01
-8.46769273e-01 -7.40751475e-02 -2.33438611e-01 5.86882770e-01
3.52681011e-01 9.96452928e-01 -6.20899051e-02 -3.03379148e-01
8.05274844e-01 6.96191341e-02 -3.71629139e-03 -3.40102047e-01
-8.49945843e-01 3.87477249e-01 3.76116097e-01 -4.23095196e-01
-7.51667261e-01 -9.21177387e-01 -8.70193422e-01 -3.31462979e-01
-4.65420961e-01 2.89738923e-01 1.06681204e+00 9.46963072e-01
9.89293829e-02 6.33945405e-01 1.08319688e+00 -6.68278575e-01
-1.11036050e+00 -7.70847201e-01 -1.04703200e+00 -1.34577304e-01
-1.42160594e-01 -5.85807502e-01 -6.66098058e-01 -4.55144376e-01]
|
[5.439409255981445, 3.9798524379730225]
|
7a0863bc-f210-459d-bd41-aa2929b93009
|
aligning-instruction-tasks-unlocks-large
|
2305.11159
| null |
https://arxiv.org/abs/2305.11159v1
|
https://arxiv.org/pdf/2305.11159v1.pdf
|
Aligning Instruction Tasks Unlocks Large Language Models as Zero-Shot Relation Extractors
|
Recent work has shown that fine-tuning large language models (LLMs) on large-scale instruction-following datasets substantially improves their performance on a wide range of NLP tasks, especially in the zero-shot setting. However, even advanced instruction-tuned LLMs still fail to outperform small LMs on relation extraction (RE), a fundamental information extraction task. We hypothesize that instruction-tuning has been unable to elicit strong RE capabilities in LLMs due to RE's low incidence in instruction-tuning datasets, making up less than 1% of all tasks (Wang et al., 2022). To address this limitation, we propose QA4RE, a framework that aligns RE with question answering (QA), a predominant task in instruction-tuning datasets. Comprehensive zero-shot RE experiments over four datasets with two series of instruction-tuned LLMs (six LLMs in total) demonstrate that our QA4RE framework consistently improves LLM performance, strongly verifying our hypothesis and enabling LLMs to outperform strong zero-shot baselines by a large margin. Additionally, we provide thorough experiments and discussions to show the robustness, few-shot effectiveness, and strong transferability of our QA4RE framework. This work illustrates a promising way of adapting LLMs to challenging and underrepresented tasks by aligning these tasks with more common instruction-tuning tasks like QA.
|
['Yu Su', 'Bernal Jiménez Gutiérrez', 'Kai Zhang']
|
2023-05-18
| null | null | null | null |
['instruction-following', 'relation-extraction']
|
['natural-language-processing', 'natural-language-processing']
|
[ 1.72071487e-01 2.59301633e-01 -6.23288572e-01 -2.66055644e-01
-1.46815920e+00 -4.71807182e-01 6.17840827e-01 1.23555310e-01
-4.41603720e-01 7.45304108e-01 5.26975632e-01 -8.35712910e-01
-4.46379930e-02 -6.78775072e-01 -1.01271129e+00 9.81499851e-02
1.89174771e-01 3.98597747e-01 5.04901707e-01 -6.95927858e-01
2.89850563e-01 7.42223859e-02 -1.63967240e+00 4.33433741e-01
1.16406512e+00 5.32285571e-01 2.46474549e-01 6.16671801e-01
-3.24447960e-01 1.02969527e+00 -8.03765297e-01 -3.65860581e-01
-1.36950642e-01 -4.54783859e-03 -1.08777499e+00 -6.91252768e-01
9.87970948e-01 -1.93185344e-01 -3.71694326e-01 6.67181611e-01
5.40995419e-01 4.47434634e-01 4.10193592e-01 -1.10932398e+00
-6.99081421e-01 9.64379430e-01 -3.85998696e-01 4.27302003e-01
5.33442855e-01 3.11558396e-01 1.31879222e+00 -8.43111336e-01
6.01798415e-01 1.18477452e+00 6.62268281e-01 4.09157217e-01
-1.22537887e+00 -8.11908603e-01 -1.43897563e-01 1.60332099e-01
-1.32567716e+00 -8.01968575e-01 3.14504623e-01 -2.94960469e-01
1.79928029e+00 2.63745457e-01 -2.20326975e-01 1.14844263e+00
6.88824236e-01 9.80265021e-01 1.18332183e+00 -5.97747326e-01
-6.63532615e-02 -1.52780920e-01 9.21459496e-01 8.14133108e-01
2.43876308e-01 -3.17540602e-03 -9.33593750e-01 -2.86874563e-01
8.50795023e-03 -4.57601935e-01 -2.70037621e-01 -4.30970751e-02
-1.27756238e+00 7.60416150e-01 8.69041160e-02 4.79907840e-01
1.03061656e-02 2.21744820e-01 4.65196997e-01 4.93340999e-01
1.42465904e-01 1.00718951e+00 -9.87419188e-01 -5.28551579e-01
-8.09384346e-01 -5.59556438e-03 1.13350534e+00 1.16403079e+00
8.34236860e-01 -8.80943518e-03 -7.73727834e-01 6.96456254e-01
-1.17710322e-01 4.33444709e-01 8.72611761e-01 -8.20389926e-01
7.85585463e-01 6.69507265e-01 -2.54163772e-01 -4.99827981e-01
-4.46751475e-01 -5.56563079e-01 -3.83311719e-01 -3.82477850e-01
2.78195918e-01 3.01107652e-02 -7.03133881e-01 1.87146616e+00
-9.98314843e-02 1.06891938e-01 9.09249634e-02 2.50999808e-01
1.33524501e+00 5.52488625e-01 4.18487608e-01 2.48888005e-02
1.66880429e+00 -1.07891023e+00 -9.17400777e-01 -7.25180387e-01
1.26211977e+00 -8.20548236e-01 1.90709245e+00 2.66633064e-01
-7.44597435e-01 -7.51969874e-01 -1.20200872e+00 -5.09683490e-01
-5.39889574e-01 5.08530065e-02 8.59967828e-01 6.47152960e-01
-7.34452128e-01 4.64759409e-01 -6.97587371e-01 -3.15956652e-01
2.38262221e-01 1.96791589e-01 -2.15460420e-01 -1.76157117e-01
-1.43373609e+00 1.05091035e+00 2.47909099e-01 -5.20078659e-01
-9.25553322e-01 -1.24614847e+00 -1.11170542e+00 2.33768031e-01
9.31956649e-01 -5.62174618e-01 1.58713782e+00 1.52963370e-01
-1.41017389e+00 6.95671916e-01 -5.80613673e-01 -4.25054967e-01
-2.01121330e-01 -7.34479666e-01 -5.98415613e-01 -3.93658459e-01
1.44396752e-01 3.32318515e-01 4.01463270e-01 -9.45787132e-01
-2.51382738e-01 -2.40924694e-02 1.96716890e-01 -1.33347541e-01
-3.44626456e-01 -2.49718241e-02 -4.36169475e-01 -2.96724588e-01
-5.29692233e-01 -5.28263032e-01 6.71108365e-02 -1.18811834e+00
-5.39967597e-01 -6.90858960e-01 5.43125868e-01 -4.70756888e-01
1.95268726e+00 -1.96038175e+00 -2.39023656e-01 -2.73828894e-01
3.04156870e-01 5.04433930e-01 -7.33007908e-01 5.77660143e-01
-8.53081942e-02 1.18032321e-01 -2.89287031e-01 -3.12193334e-01
4.22950953e-01 2.63318509e-01 -6.49297953e-01 -5.28304186e-03
3.85599852e-01 1.64018047e+00 -1.01860857e+00 -5.27327240e-01
2.74241772e-02 4.92618196e-02 -5.09178579e-01 3.97909701e-01
-5.19248843e-01 -1.41946189e-02 -3.24895591e-01 7.90288985e-01
3.54561210e-03 -4.88641471e-01 -1.53718010e-01 -2.87876397e-01
5.01914732e-02 9.87972736e-01 -6.81300402e-01 1.94872820e+00
-8.58000994e-01 7.04704285e-01 -2.79044509e-01 -5.77116609e-01
6.46277964e-01 1.12520672e-01 2.24877506e-01 -1.11421895e+00
-1.38557374e-01 1.23780392e-01 4.96296771e-02 -6.82727158e-01
9.37731683e-01 1.44851401e-01 -5.10165215e-01 6.70880795e-01
2.90301293e-01 -3.91067773e-01 3.69818449e-01 5.47371745e-01
1.58778262e+00 1.78445041e-01 5.25624216e-01 -3.35733920e-01
5.10846257e-01 2.05502927e-01 2.74551332e-01 9.06837523e-01
-3.12374145e-01 2.62850598e-02 3.02524209e-01 1.07902020e-01
-3.17878962e-01 -8.31481695e-01 -3.23560774e-01 1.86063027e+00
-2.88444996e-01 -9.80517149e-01 -7.61875749e-01 -1.12264931e+00
2.81474385e-02 1.44306910e+00 -4.66800034e-01 -5.17922819e-01
-7.52384305e-01 -8.58853281e-01 7.62724638e-01 3.84004921e-01
3.31624031e-01 -9.42033887e-01 -4.32193995e-01 2.16523573e-01
-3.76234144e-01 -1.29268909e+00 -6.40611768e-01 5.25184453e-01
-6.54943645e-01 -1.20096040e+00 -1.07408114e-01 -4.64530498e-01
6.38351887e-02 3.71194005e-01 2.10107422e+00 -4.51409863e-03
-1.77425951e-01 4.86795992e-01 -1.86884269e-01 -3.01681072e-01
-5.65244377e-01 5.73755264e-01 -5.74512631e-02 -9.00254130e-01
1.07454252e+00 -2.07555607e-01 -2.82507151e-01 3.92242163e-01
-8.15284729e-01 -4.49407518e-01 7.49597609e-01 7.73460269e-01
5.20522475e-01 -3.23234409e-01 9.07271564e-01 -1.28547204e+00
9.30894434e-01 -6.05621219e-01 -3.80943626e-01 6.25195205e-01
-8.35195541e-01 4.58017975e-01 4.89351809e-01 -4.29214031e-01
-8.75541866e-01 -4.04226631e-01 -2.43698537e-01 5.38717806e-02
-1.66591436e-01 5.35110891e-01 -1.59504697e-01 -1.30367070e-01
1.03040338e+00 3.90192717e-02 -2.97640353e-01 -4.99744534e-01
7.56522119e-01 5.32685220e-01 6.54960752e-01 -9.64105129e-01
7.10193515e-01 -2.21124604e-01 -1.46636739e-01 -6.28232837e-01
-1.31040061e+00 -6.15815878e-01 -4.80048776e-01 4.42906439e-01
7.56270051e-01 -9.13102686e-01 -7.71894932e-01 -1.04462773e-01
-8.11232805e-01 -7.76937127e-01 -3.37799668e-01 2.65517354e-01
-4.68119234e-01 3.95981461e-01 -8.21703017e-01 -2.54282475e-01
-5.17099917e-01 -1.48929596e+00 1.33237422e+00 -5.52087091e-03
-7.46175230e-01 -1.07529497e+00 2.41036981e-01 6.87742114e-01
5.00835001e-01 -3.18777680e-01 1.35257840e+00 -9.17370915e-01
-4.43276435e-01 1.54884949e-01 -1.23814173e-01 1.41729817e-01
1.09615333e-01 -1.91600353e-01 -1.22630322e+00 -1.86446443e-01
-6.47262409e-02 -8.33343089e-01 9.57912803e-01 4.07962874e-02
1.22920632e+00 6.68214459e-04 -4.00422812e-01 4.26139295e-01
1.14318645e+00 -2.27010876e-01 4.43863750e-01 3.56358260e-01
6.56999171e-01 4.49500442e-01 8.87036383e-01 -1.40902892e-01
6.61517084e-01 6.18429542e-01 1.72505349e-01 1.41668513e-01
-5.90266764e-01 -4.14006889e-01 8.20818007e-01 1.32504046e+00
5.81565380e-01 -2.23873705e-01 -1.16272748e+00 4.72904265e-01
-1.47227097e+00 -3.81134421e-01 -4.34580147e-01 2.11506891e+00
1.28684330e+00 1.40981749e-01 -3.82640779e-01 -2.65051812e-01
-4.56317365e-02 1.30459532e-01 -4.07271504e-01 -5.51901519e-01
-1.63482986e-02 8.58825922e-01 4.49120641e-01 5.93942404e-01
-9.41829741e-01 1.22725093e+00 6.67413902e+00 1.33028543e+00
-6.21134043e-01 3.14366519e-01 3.18552405e-01 9.64006260e-02
-5.42627394e-01 -3.83401215e-02 -1.47345054e+00 1.33914933e-01
1.66383111e+00 -4.00573850e-01 2.36314699e-01 6.41869426e-01
-3.61761004e-01 -1.73477694e-01 -1.51623130e+00 5.72771549e-01
2.12975904e-01 -1.23908997e+00 3.82522568e-02 -1.63971528e-01
9.28268731e-01 3.54874253e-01 1.10523254e-02 1.35067582e+00
8.12072992e-01 -1.45801508e+00 1.54329091e-01 4.15532768e-01
9.40981448e-01 -5.55758774e-01 7.23356366e-01 5.55276453e-01
-1.16418743e+00 1.10175550e-01 -2.73240298e-01 2.20086873e-02
6.86547756e-02 4.45481747e-01 -8.39931071e-01 5.69568098e-01
5.44603407e-01 3.99318159e-01 -9.72223461e-01 4.75292861e-01
-4.11051363e-01 9.34058726e-01 -1.48373157e-01 9.64784548e-02
1.56773821e-01 3.23231310e-01 3.00662071e-01 1.37476230e+00
1.06692784e-01 1.50129721e-01 1.14616618e-01 8.05738330e-01
-3.75959486e-01 2.34696586e-02 -5.04535377e-01 -3.14485639e-01
6.51963174e-01 1.27674294e+00 8.47318769e-02 -5.31170428e-01
-7.07030714e-01 5.45385718e-01 5.00997841e-01 3.55507046e-01
-6.73548877e-01 -5.11071861e-01 8.34632695e-01 -1.77800491e-01
-8.93822163e-02 -2.41160080e-01 -2.04870626e-01 -1.30078661e+00
-3.22685748e-01 -1.37451828e+00 5.32491207e-01 -5.86753070e-01
-1.37443757e+00 4.78132159e-01 7.44128153e-02 -8.40994895e-01
-5.86839080e-01 -6.35879934e-01 -4.64996070e-01 9.51790869e-01
-1.78934526e+00 -9.32081282e-01 -1.28093287e-01 5.40307403e-01
7.42734075e-01 -1.12020709e-01 9.97096062e-01 3.58902037e-01
-7.68790841e-01 1.09242189e+00 -1.38610765e-01 -1.32500678e-01
1.37540531e+00 -1.35205805e+00 6.23472810e-01 7.67094970e-01
2.53657401e-01 1.16249430e+00 5.57278514e-01 -7.28233337e-01
-1.95849586e+00 -1.08404696e+00 1.22349966e+00 -1.26191652e+00
1.23783636e+00 -4.36368644e-01 -1.24723291e+00 9.39108312e-01
3.17722261e-01 -1.26679212e-01 9.25200403e-01 5.35934269e-01
-5.36645830e-01 1.72833547e-01 -5.94908535e-01 4.54727978e-01
1.00530231e+00 -9.18220043e-01 -1.23076832e+00 3.96696836e-01
1.27832317e+00 -5.42560279e-01 -1.35385525e+00 7.63779104e-01
2.44673993e-02 -5.20326793e-01 1.06749821e+00 -9.23736155e-01
5.62447309e-01 5.10309525e-02 -3.76429200e-01 -1.30664194e+00
-1.72399521e-01 -6.20343566e-01 -8.15080464e-01 1.42717397e+00
4.65485841e-01 -5.55727601e-01 4.20505077e-01 4.28613901e-01
-4.33719456e-01 -8.56457829e-01 -6.26024783e-01 -8.73299718e-01
5.77619493e-01 -7.16331363e-01 6.88045859e-01 9.39262271e-01
-8.18324685e-02 9.07438397e-01 1.34635255e-01 -1.09253554e-02
3.72516006e-01 2.16234267e-01 1.13323331e+00 -9.56088424e-01
-4.98416752e-01 -4.28945869e-01 3.33346993e-01 -1.38477695e+00
6.32747114e-01 -1.09543848e+00 1.74318105e-01 -1.27517438e+00
3.63701403e-01 -4.84669119e-01 -3.57450664e-01 6.33500874e-01
-6.70276940e-01 2.04889979e-02 -1.20081007e-01 9.81062800e-02
-8.69624496e-01 5.20758271e-01 1.28252423e+00 -8.29138607e-02
5.11041060e-02 -2.61305720e-01 -8.08404624e-01 5.25605142e-01
5.19591928e-01 -2.34986320e-01 -4.23078448e-01 -4.78343159e-01
3.21522653e-01 -5.81459664e-02 -1.77903131e-01 -7.15103626e-01
2.39923432e-01 -8.25646818e-02 -1.66851357e-01 -4.50930744e-01
7.36163035e-02 -3.05621684e-01 -5.73959053e-01 3.08008760e-01
-4.22369540e-01 -3.34617607e-02 7.31487155e-01 3.47152323e-01
-3.28310519e-01 -3.28496665e-01 1.96928769e-01 9.30985212e-02
-1.24337888e+00 8.47455561e-02 -1.67943150e-01 6.94292307e-01
5.75997412e-01 1.89442828e-01 -1.03544474e+00 2.48152409e-02
-1.83661506e-01 5.07653892e-01 1.51200950e-01 7.26513028e-01
2.39850208e-01 -1.04141569e+00 -6.13991320e-01 1.84721828e-01
5.76501727e-01 -1.81150138e-01 -5.83813619e-03 8.12163293e-01
1.47464290e-01 1.00558877e+00 3.41238022e-01 -5.20893455e-01
-1.05578876e+00 5.22662044e-01 5.19494899e-02 -8.36544037e-01
-2.55976737e-01 9.81820762e-01 3.67287308e-01 -8.24561417e-01
1.79843307e-01 -7.08593667e-01 -2.68458545e-01 -9.23016593e-02
5.21947145e-01 3.53555322e-01 4.44673747e-01 -2.15058476e-01
-2.63484389e-01 4.14550990e-01 -1.07581235e-01 3.12774092e-01
1.08624780e+00 7.24645099e-03 -1.52640775e-01 6.06833160e-01
1.23455513e+00 5.10372519e-01 -7.89592624e-01 -5.85901856e-01
3.82556319e-01 7.98717588e-02 1.25841692e-01 -1.09882498e+00
-5.34844100e-01 9.01174963e-01 9.70325097e-02 -2.81238347e-01
9.19908643e-01 1.29817843e-01 1.21876240e+00 7.61969149e-01
6.21307731e-01 -7.65169561e-01 1.93714634e-01 1.11023510e+00
5.89696288e-01 -1.61330748e+00 -7.45811909e-02 -3.70745927e-01
-2.43356541e-01 9.06639814e-01 9.31827486e-01 3.65367144e-01
3.82917762e-01 5.12428999e-01 -1.67813003e-01 -2.82988012e-01
-1.15858793e+00 -3.47900778e-01 8.06840360e-01 3.12600583e-01
1.04201221e+00 9.73845571e-02 -2.78611600e-01 9.21407759e-01
-6.12510562e-01 6.97247535e-02 2.77586073e-01 9.74000573e-01
-4.76475507e-01 -1.25054908e+00 -2.23036736e-01 7.17889249e-01
-4.88853872e-01 -6.84540927e-01 -2.00787693e-01 9.82309878e-01
-2.51834065e-01 1.23949957e+00 -7.85650909e-02 -4.90581393e-01
4.70216841e-01 4.52065796e-01 4.75843847e-01 -1.06334782e+00
-9.60824430e-01 -5.62704086e-01 4.21494126e-01 -1.18093967e+00
2.03778014e-01 -1.63264662e-01 -1.26788306e+00 -3.44535381e-01
-3.31256986e-01 2.34453738e-01 2.83526570e-01 1.02601755e+00
5.41748166e-01 1.00489998e+00 -5.80278933e-02 1.86617747e-02
-1.01647437e+00 -1.19978774e+00 1.03047788e-01 3.28018546e-01
2.00568900e-01 -5.93689084e-01 -3.44148457e-01 -2.14925200e-01]
|
[10.57347297668457, 8.443232536315918]
|
d13e2e67-88e0-46b3-a1a0-e984277c6cc6
|
read-attend-and-comment-a-deep-architecture-1
| null | null |
https://arxiv.org/abs/1909.11974
|
https://arxiv.org/pdf/1909.11974.pdf
|
Read, Attend and Comment: A Deep Architecture for Automatic NewsComment Generation
|
Automatic news comment generation is a new testbed for techniques of natural language generation. In this paper, we propose a “read-attend-comment” procedure for news comment generation and formalize the procedure with a reading network and a generation network. The reading network comprehends a news article and distills some important points from it, then the generation network creates a comment by attending to the extracted discrete points and the news title. We optimize the model in an end-to-end manner by maximizing a variational lower bound of the true objective using the back-propagation algorithm. Experimental results on two datasets indicate that our model can significantly outperform existing methods in terms of both automatic evaluation and human judgment.
|
['Zhoujun Li', 'Ze Yang', 'Wei Wu', 'Can Xu']
|
2019-10-01
| null | null | null |
emnlp2019-2019-10
|
['comment-generation']
|
['natural-language-processing']
|
[ 2.38924146e-01 9.82817054e-01 -3.27985972e-01 -5.38589120e-01
-1.14672506e+00 -5.96142411e-01 1.06566179e+00 3.83555740e-01
-1.84872970e-01 1.02846301e+00 9.43161786e-01 -2.31041834e-01
1.40810460e-01 -7.04559147e-01 -6.12672389e-01 -4.85254973e-01
1.10279001e-01 7.45104671e-01 -5.65646552e-02 -3.45483243e-01
6.51908100e-01 -4.36397612e-01 -1.33840716e+00 5.06496310e-01
1.03662860e+00 6.82192683e-01 -3.23515316e-03 9.94426370e-01
-2.20170408e-01 1.10202312e+00 -7.61220455e-01 -7.87559271e-01
6.74347878e-02 -9.10805345e-01 -1.01360977e+00 7.08103105e-02
2.19847813e-01 -1.22313000e-01 3.42467993e-01 9.25964475e-01
4.81450677e-01 5.65072358e-01 9.95924175e-01 -1.04688311e+00
-9.86812890e-01 1.30149138e+00 -3.71906281e-01 1.54685125e-01
7.97190189e-01 -2.88886189e-01 1.64505661e+00 -1.25012994e+00
8.61671388e-01 1.11122823e+00 5.10731757e-01 5.75800061e-01
-1.35708201e+00 -1.34039121e-02 4.67443466e-01 -2.23200843e-01
-8.24682117e-01 -4.54685748e-01 7.37112343e-01 -5.37545264e-01
8.95270586e-01 3.02444071e-01 7.02313423e-01 1.05983734e+00
3.24813366e-01 9.44290578e-01 5.44732630e-01 -5.29800713e-01
3.19365948e-01 3.51765156e-01 2.16706291e-01 6.03088617e-01
-2.11166576e-01 1.03943087e-02 -8.71968925e-01 -3.83115768e-01
2.29175657e-01 -3.37179244e-01 -3.89745802e-01 9.75352973e-02
-1.24929988e+00 1.22060108e+00 3.30326051e-01 -2.10682884e-01
-5.61476111e-01 6.95441961e-02 2.07635432e-01 4.01516557e-01
1.36074650e+00 7.77971983e-01 -1.31324932e-01 -1.42270699e-01
-1.20992589e+00 7.25917935e-01 1.39152300e+00 7.96417594e-01
5.21847606e-01 -2.18782976e-01 -6.84295595e-01 7.46925414e-01
5.28062940e-01 3.83702427e-01 2.67931700e-01 -6.98940337e-01
5.30797720e-01 1.85696945e-01 5.10186970e-01 -1.09373605e+00
-2.61260778e-01 -4.47758734e-01 -7.35406101e-01 1.51935369e-01
2.95982696e-02 -7.47556150e-01 -6.88252449e-01 1.48888218e+00
2.55802661e-01 -7.59555623e-02 1.87020466e-01 7.12811589e-01
9.14706886e-01 1.23626244e+00 -1.61612779e-01 -5.80048978e-01
9.84003186e-01 -1.41672623e+00 -7.17410743e-01 -2.47292295e-02
3.73288870e-01 -7.90602565e-01 9.10734117e-01 3.04081142e-01
-1.58120799e+00 -3.78487170e-01 -6.15305960e-01 -1.03658706e-01
-1.89992681e-03 2.73817986e-01 7.53160119e-02 -2.86441464e-02
-1.33333850e+00 4.93408263e-01 -4.19237465e-01 -2.41867945e-01
4.40192856e-02 -1.75465383e-02 2.19128400e-01 4.44923580e-01
-1.18245649e+00 7.65698493e-01 3.72306645e-01 3.19972001e-02
-7.03546345e-01 -8.61722052e-01 -7.41182268e-01 8.69572461e-02
3.06396544e-01 -1.34636092e+00 1.87646830e+00 -1.01062512e+00
-1.97018409e+00 7.47725427e-01 -4.89905268e-01 -6.55492961e-01
6.67810202e-01 -3.23465586e-01 -1.98342443e-01 -2.47589946e-02
3.29784870e-01 8.04102302e-01 9.75750148e-01 -1.38354468e+00
-8.18668187e-01 2.48853832e-01 3.12688291e-01 5.45712888e-01
8.76541436e-02 4.62544970e-02 -9.33122188e-02 -8.51413727e-01
-3.35271537e-01 -8.75740111e-01 -2.23604620e-01 -2.45192543e-01
-9.26550567e-01 -5.16856253e-01 2.58634180e-01 -5.49153090e-01
1.29631412e+00 -1.67668688e+00 4.64465201e-01 3.03497761e-01
5.34407675e-01 -2.47967809e-01 -5.11452034e-02 7.87184000e-01
5.77580705e-02 2.40163296e-01 -1.31736055e-01 -7.96256244e-01
2.01736480e-01 -2.26466671e-01 -8.24082077e-01 1.09036699e-01
-1.22440420e-01 9.30055916e-01 -1.21404243e+00 -2.96777070e-01
-4.38503176e-01 7.41811097e-02 -8.31722081e-01 4.20407534e-01
-9.44053292e-01 1.60327002e-01 -3.76701385e-01 -3.36510427e-02
1.62312210e-01 -5.76816022e-01 -1.50314331e-01 1.81529909e-01
-3.13896179e-01 7.00173378e-01 -7.90580571e-01 1.47996998e+00
-1.85826883e-01 7.14652002e-01 -2.53196806e-01 -5.87774456e-01
9.30609047e-01 3.51340950e-01 1.22807264e-01 -2.40800545e-01
-3.74059267e-02 -1.41425565e-01 -4.79511619e-01 -4.33160037e-01
1.00295401e+00 -2.20868036e-01 -1.65460840e-01 1.25110126e+00
-1.09561414e-01 -2.89980829e-01 5.06381691e-01 7.80856788e-01
5.61486542e-01 6.57169595e-02 3.92445147e-01 -3.43036890e-01
2.14542821e-01 1.11151524e-01 4.91570681e-02 1.13116050e+00
3.72011989e-01 7.17683554e-01 7.90370464e-01 -3.09581339e-01
-1.07865512e+00 -1.13542926e+00 4.74543899e-01 1.57468808e+00
-1.53443128e-01 -7.82399654e-01 -7.19051600e-01 -7.04724371e-01
-1.84014246e-01 1.33992445e+00 -9.55753803e-01 9.01974365e-02
-2.07064688e-01 -4.91673410e-01 2.00138256e-01 3.57628644e-01
8.85980427e-02 -1.07555521e+00 -2.87908256e-01 2.38295168e-01
-5.62338769e-01 -5.15269101e-01 -9.17878389e-01 -2.59129912e-01
-4.44402337e-01 -6.05290830e-01 -1.02127957e+00 -9.26351964e-01
8.24189126e-01 -6.04129918e-02 1.45355046e+00 3.01482733e-02
4.40624475e-01 1.24066897e-01 -5.43708980e-01 -7.07756400e-01
-7.57506847e-01 4.14355993e-01 -1.34062812e-01 9.34923142e-02
-8.67589265e-02 -3.78290534e-01 -3.90448183e-01 -1.86571032e-01
-6.41709566e-01 6.49862289e-01 3.05114806e-01 8.62250984e-01
4.66366172e-01 -3.28769505e-01 8.40114951e-01 -1.23679769e+00
1.55694604e+00 -7.95280933e-01 -4.46674049e-01 2.34887764e-01
-7.60861099e-01 2.39759952e-01 6.16038322e-01 -8.13492835e-02
-1.38214350e+00 -4.48253781e-01 -1.09225482e-01 3.76910567e-01
1.81604713e-01 1.18484282e+00 5.19158006e-01 6.78167582e-01
9.69205141e-01 3.42191488e-01 -1.19735144e-01 -1.58644289e-01
7.21613705e-01 4.27304536e-01 4.75173086e-01 -4.39327031e-01
7.89199710e-01 2.64660299e-01 -6.28349662e-01 -5.22706747e-01
-1.57914567e+00 -3.51301849e-01 -3.79693657e-01 -4.42006320e-01
8.46225321e-01 -7.79401183e-01 -3.85403901e-01 2.76903123e-01
-1.43725193e+00 -2.69149423e-01 -7.24677324e-01 3.10290694e-01
-5.75590312e-01 4.84468266e-02 -6.55249536e-01 -7.25016654e-01
-7.52056539e-01 -5.40315270e-01 9.86118019e-01 4.24442917e-01
-5.85492492e-01 -1.48080397e+00 5.65438986e-01 4.47437316e-02
5.33872068e-01 2.95184642e-01 6.76923215e-01 -8.46584558e-01
-3.51319760e-01 -8.69811550e-02 -1.34786323e-01 1.87072679e-01
-2.76597768e-01 2.70068020e-01 -6.99851036e-01 4.43093032e-02
-2.63441771e-01 -3.64790648e-01 9.15887415e-01 5.05236983e-01
6.50638521e-01 -9.15767908e-01 -2.07066879e-01 2.40842700e-01
1.09039545e+00 -1.65485352e-01 3.67172897e-01 -1.54959243e-02
6.05180919e-01 5.75592935e-01 3.31932217e-01 7.98768461e-01
6.88705802e-01 2.73691833e-01 3.86920758e-02 -5.00411019e-02
1.59621611e-01 -7.35138178e-01 5.25870025e-01 1.09321523e+00
-3.22659798e-02 -1.01738310e+00 -7.09253848e-01 6.06874228e-01
-2.19539571e+00 -1.34337330e+00 -2.60589778e-01 1.67216539e+00
1.10798168e+00 2.17188194e-01 2.46982247e-01 -2.71725059e-01
5.86480975e-01 5.40037870e-01 -3.88423324e-01 -6.56732857e-01
-2.49371259e-03 -9.36105326e-02 -4.05922811e-03 1.02007616e+00
-6.44114912e-01 8.71182859e-01 7.63914967e+00 4.54923660e-01
-7.20732987e-01 -1.89984292e-01 5.73857367e-01 -5.75259924e-02
-9.73232031e-01 6.36422113e-02 -9.20275867e-01 3.35449994e-01
8.39407086e-01 -8.92659009e-01 2.88629055e-01 7.43483186e-01
5.61166108e-01 -2.05351397e-01 -1.26142740e+00 4.22268897e-01
5.97253501e-01 -1.94367397e+00 4.39901501e-01 -2.33635560e-01
1.27222192e+00 -1.63288012e-01 -2.07040366e-02 9.72031504e-02
8.29577565e-01 -9.59870398e-01 9.92375195e-01 1.02062905e+00
4.23872828e-01 -6.50932789e-01 5.43604434e-01 8.15784454e-01
-8.35170746e-01 2.35489666e-01 -1.27573490e-01 -4.07561153e-01
7.22310245e-01 8.73624086e-01 -1.09050524e+00 3.75090152e-01
1.52624503e-01 9.53420937e-01 -2.38216981e-01 9.64533865e-01
-7.88827479e-01 8.00169826e-01 -9.84525830e-02 -6.42060757e-01
3.03792715e-01 -1.92197040e-01 9.94445384e-01 1.26992106e+00
1.64574623e-01 2.18474001e-01 4.40566152e-01 1.16827083e+00
-4.20827359e-01 2.57603765e-01 -2.03495711e-01 -5.72751462e-02
2.25612640e-01 1.12718344e+00 -5.44564426e-01 -7.05340624e-01
8.41251090e-02 9.68972385e-01 5.03762901e-01 6.16611600e-01
-7.00273156e-01 -4.71131533e-01 1.80348217e-01 5.24452589e-02
1.81208000e-01 1.42164141e-01 -4.24885273e-01 -1.35510027e+00
-2.64037419e-02 -7.76758552e-01 2.89488077e-01 -1.16983914e+00
-1.32276320e+00 7.17555523e-01 1.25797525e-01 -9.05841053e-01
-7.61591792e-01 3.91047373e-02 -1.13338947e+00 1.02667582e+00
-1.23919618e+00 -7.64332831e-01 -1.52080551e-01 2.55354971e-01
7.88403273e-01 -3.54952440e-02 8.94807577e-01 -2.73827970e-01
-1.52672961e-01 4.71723765e-01 1.02871768e-01 3.25854979e-02
5.72007418e-01 -1.63709295e+00 6.72111213e-01 9.03667688e-01
2.72850633e-01 5.39602578e-01 1.10835695e+00 -6.34517312e-01
-5.26054502e-01 -9.40328240e-01 1.55764842e+00 -4.73755211e-01
6.10290170e-01 -3.31851929e-01 -5.98152637e-01 8.49024594e-01
8.97285759e-01 -9.07713532e-01 8.61426413e-01 3.59709948e-01
-9.83439386e-02 3.58958393e-01 -7.84334183e-01 8.28163624e-01
7.73852587e-01 -3.93014222e-01 -1.01034343e+00 7.94562399e-01
9.29667950e-01 -6.87606454e-01 -5.54063559e-01 -3.72494221e-01
3.78498524e-01 -8.50926876e-01 4.86279041e-01 -6.39730453e-01
1.16101992e+00 -2.14551553e-01 4.22347158e-01 -2.11832070e+00
-4.76489156e-01 -1.15679884e+00 -1.45529196e-01 1.13449526e+00
1.25994968e+00 -4.89974439e-01 6.00652099e-01 5.76718450e-01
-3.23237687e-01 -8.28686714e-01 -1.95727825e-01 -4.36417088e-02
-1.11178115e-01 -2.83490449e-01 3.74383897e-01 6.81666017e-01
4.49936658e-01 8.77697587e-01 -6.19611919e-01 -2.95140833e-01
4.47747886e-01 2.52812117e-01 8.69446456e-01 -1.32192028e+00
-4.66043383e-01 -5.09091079e-01 3.46042931e-01 -1.59880579e+00
1.95437223e-01 -9.62653756e-01 4.44507241e-01 -2.14078665e+00
2.94641882e-01 -6.91194981e-02 2.49745756e-01 1.30411461e-01
-4.00696248e-01 -3.97494808e-02 9.37779248e-02 2.52462208e-01
-8.61171126e-01 6.14279985e-01 1.24693561e+00 3.35330889e-02
-2.81629205e-01 4.15192902e-01 -1.14565551e+00 7.31567621e-01
5.88760376e-01 -6.03068411e-01 -5.87564290e-01 -5.22037268e-01
1.05524695e+00 2.51048863e-01 1.33474693e-01 -4.57936317e-01
6.30872428e-01 -1.96263209e-01 9.20929238e-02 -9.00648832e-01
5.28561473e-02 3.65290008e-02 -9.77438390e-02 -3.15103424e-03
-1.29070783e+00 3.08307081e-01 -3.68852466e-01 8.58393908e-01
-3.07408512e-01 -3.16427946e-01 4.30373192e-01 -4.08339500e-02
-1.50732666e-01 9.56361070e-02 -5.55718601e-01 5.78827143e-01
8.23614299e-01 1.10439491e-02 -5.25556684e-01 -1.13237286e+00
-9.01137471e-01 4.53360975e-01 2.68275827e-01 4.04124260e-01
6.12230718e-01 -1.18340945e+00 -1.31758177e+00 2.14585923e-02
5.12036681e-02 1.59338653e-01 -2.51315951e-01 6.33527577e-01
-3.77381742e-01 1.74173445e-01 3.81541491e-01 -2.62289286e-01
-9.36808169e-01 1.39378205e-01 2.42586911e-01 -6.27026141e-01
-5.73791325e-01 1.19494343e+00 2.03603096e-02 -3.24059010e-01
2.09551096e-01 -5.71761847e-01 -4.26160157e-01 4.23456430e-01
8.03740323e-01 2.96103835e-01 -3.68292123e-01 -4.53052223e-01
2.14854583e-01 7.23489001e-02 -2.91353762e-01 -6.93603575e-01
1.20528340e+00 -3.44762236e-01 -1.05129495e-01 7.79955089e-01
1.06354094e+00 3.71952146e-01 -1.01646626e+00 -4.02383119e-01
2.79613826e-02 -7.65942931e-02 -1.72292009e-01 -8.40788901e-01
-5.97356081e-01 2.81186372e-01 -4.13982689e-01 8.90448928e-01
5.82104683e-01 1.76331639e-01 9.16322052e-01 4.72120553e-01
-2.77364612e-01 -1.43011224e+00 1.65606603e-01 7.98847437e-01
1.22856784e+00 -1.12463009e+00 -1.16456971e-02 -1.31210297e-01
-1.05289364e+00 1.10677135e+00 1.89248592e-01 -4.21260357e-01
8.83422196e-01 9.94490311e-02 2.72338033e-01 -3.22625726e-01
-1.28971493e+00 2.08795011e-01 7.81634331e-01 3.63823652e-01
6.92192137e-01 3.03213894e-02 -5.08753955e-01 4.66452390e-01
-7.56890833e-01 -1.61934849e-02 7.79436946e-01 6.43482566e-01
-6.93390846e-01 -7.21418202e-01 -5.29962927e-02 5.74338019e-01
-3.67005348e-01 -4.39139962e-01 -7.63552487e-01 1.72267556e-01
-3.23302329e-01 1.12956727e+00 1.56218141e-01 -2.44629815e-01
1.94536462e-01 1.14053831e-01 -9.54138264e-02 -9.57944453e-01
-7.93995500e-01 -9.51255579e-03 5.11008203e-01 -1.77166075e-01
-5.50025821e-01 -6.11695945e-01 -1.24868929e+00 -4.32588577e-01
-3.97544682e-01 6.90437078e-01 4.84698623e-01 1.03802693e+00
5.19639254e-01 4.83331650e-01 6.53190136e-01 -8.68849337e-01
-8.42215180e-01 -1.04243255e+00 -3.07324320e-01 3.27785790e-01
4.86341417e-01 -4.40879678e-03 -5.62472582e-01 5.96182466e-01]
|
[12.32429313659668, 9.287490844726562]
|
6617924e-0894-4f14-96bf-6d72b40d08b2
|
features-fusion-framework-for-multimodal
|
2209.01728
| null |
https://arxiv.org/abs/2209.01728v1
|
https://arxiv.org/pdf/2209.01728v1.pdf
|
Features Fusion Framework for Multimodal Irregular Time-series Events
|
Some data from multiple sources can be modeled as multimodal time-series events which have different sampling frequencies, data compositions, temporal relations and characteristics. Different types of events have complex nonlinear relationships, and the time of each event is irregular. Neither the classical Recurrent Neural Network (RNN) model nor the current state-of-the-art Transformer model can deal with these features well. In this paper, a features fusion framework for multimodal irregular time-series events is proposed based on the Long Short-Term Memory networks (LSTM). Firstly, the complex features are extracted according to the irregular patterns of different events. Secondly, the nonlinear correlation and complex temporal dependencies relationship between complex features are captured and fused into a tensor. Finally, a feature gate are used to control the access frequency of different tensors. Extensive experiments on MIMIC-III dataset demonstrate that the proposed framework significantly outperforms to the existing methods in terms of AUC (the area under Receiver Operating Characteristic curve) and AP (Average Precision).
|
['Xianchao Zhang', 'Peiwang Tang']
|
2022-09-05
| null | null | null | null |
['irregular-time-series']
|
['time-series']
|
[-2.36730441e-01 -8.64247739e-01 2.90480368e-02 -3.62589151e-01
-3.58356714e-01 -3.65853667e-01 7.70434737e-01 2.64987707e-01
-3.43201607e-01 3.49700361e-01 4.44644153e-01 -1.20829204e-02
-4.68617171e-01 -7.55147219e-01 -4.06074703e-01 -7.03790426e-01
-4.45161641e-01 1.70676917e-01 3.05538535e-01 -2.80075043e-01
2.14332044e-01 5.81975102e-01 -1.50382662e+00 4.73489106e-01
6.07843161e-01 1.40135455e+00 6.71328083e-02 3.47307473e-01
-2.83295184e-01 1.18317044e+00 -6.21119976e-01 -2.95210689e-01
-2.68627942e-01 5.29141873e-02 -4.26902205e-01 -1.17976837e-01
-3.48083913e-01 -4.24246974e-02 -6.98588848e-01 7.60358870e-01
4.61011052e-01 2.09556356e-01 4.34641659e-01 -1.22034359e+00
-3.97120595e-01 6.29744351e-01 -4.01739627e-01 4.59945291e-01
2.32606500e-01 -1.79478358e-02 7.33944416e-01 -1.05916202e+00
4.09733176e-01 1.07636595e+00 7.25594759e-01 -2.32387736e-01
-8.44880164e-01 -4.86211926e-01 1.14614360e-01 6.73097014e-01
-1.32344234e+00 -1.01983242e-01 1.03632498e+00 -3.03430557e-01
9.15736079e-01 2.04511493e-01 6.46118104e-01 1.23815262e+00
5.24052978e-01 6.91735148e-01 8.78180861e-01 1.76528171e-02
-3.80599976e-01 -1.74139217e-01 1.22091539e-01 5.74044287e-01
-2.54681557e-01 6.26074895e-02 -7.84362257e-01 -1.83380321e-01
5.56883216e-01 7.52454758e-01 -1.35729387e-01 7.72856474e-02
-1.88229001e+00 5.77571630e-01 3.58351350e-01 7.70793915e-01
-7.72404194e-01 -1.74962431e-01 9.08336401e-01 4.69825089e-01
3.22913289e-01 -5.26254363e-02 -6.49020851e-01 -1.65186808e-01
-6.01432383e-01 1.70300726e-03 6.94375455e-01 8.27605605e-01
3.89670193e-01 2.14192316e-01 -4.77013469e-01 7.61948407e-01
2.53747046e-01 6.05224073e-01 9.81746316e-01 -3.91799837e-01
6.78523421e-01 7.79223323e-01 -3.12714688e-02 -1.66476130e+00
-7.27027714e-01 -5.44208765e-01 -1.28349400e+00 -8.21996391e-01
3.07366967e-01 -7.93711543e-02 -4.86867458e-01 1.58187318e+00
2.74661839e-01 3.60780954e-01 -4.77025807e-02 6.95399046e-01
9.76339877e-01 1.10115647e+00 3.05457693e-02 -7.44056165e-01
1.57959592e+00 -4.53913838e-01 -1.19930124e+00 3.28129321e-01
1.82841986e-01 -1.02197039e+00 7.47338772e-01 3.90037626e-01
-1.04774082e+00 -5.35320461e-01 -9.21010137e-01 -4.58533615e-02
-6.50354922e-01 2.98123688e-01 5.02870739e-01 3.73320915e-02
-4.74816889e-01 5.47620952e-01 -8.63041520e-01 -8.39303806e-02
-8.02368000e-02 2.27305189e-01 -1.99345022e-01 2.66347736e-01
-1.62023377e+00 6.45709991e-01 3.99224639e-01 7.47553468e-01
-5.76110005e-01 -6.57826543e-01 -6.66288853e-01 -1.15877293e-01
2.67566502e-01 -2.57406384e-01 9.27547157e-01 -6.02600396e-01
-1.38049650e+00 2.74462521e-01 -2.01174796e-01 -1.81479678e-01
2.08236873e-01 1.23858498e-02 -1.02618778e+00 5.41145802e-02
-7.60981068e-02 -1.12189390e-01 7.71261275e-01 -6.25221193e-01
-7.16713727e-01 -3.50310564e-01 -2.97552198e-01 1.04214154e-01
-5.89743197e-01 3.59383702e-01 -3.08193833e-01 -9.16642845e-01
2.95151681e-01 -7.21427560e-01 -1.35199977e-02 -5.13728321e-01
-3.38371515e-01 -4.80654597e-01 8.12987983e-01 -6.38285875e-01
1.60230267e+00 -2.18867850e+00 3.10408086e-01 3.62880468e-01
1.36420026e-01 -8.80298838e-02 -8.63473117e-02 7.59987772e-01
-1.76468521e-01 -1.72175825e-01 2.49910742e-01 -2.68064048e-02
2.32445840e-02 1.01100989e-01 -6.99329078e-01 3.55641872e-01
8.40798914e-02 6.97097540e-01 -8.01712334e-01 -4.91643995e-01
2.23450109e-01 6.78618193e-01 2.86248356e-01 4.05908436e-01
1.08617857e-01 2.93591470e-01 -5.07913947e-01 7.12010026e-01
3.60152483e-01 -4.13614303e-01 8.07588324e-02 -8.64801466e-01
-2.94175416e-01 3.06941003e-01 -1.05934453e+00 1.54116547e+00
-3.61644894e-01 3.37814689e-01 -6.11189544e-01 -9.09536123e-01
1.00345373e+00 7.88021863e-01 8.19009066e-01 -7.19778299e-01
3.96276325e-01 2.40967020e-01 -1.31598294e-01 -8.73663485e-01
4.97077346e-01 2.85623241e-02 -1.51345834e-01 3.42265785e-01
9.59272757e-02 5.11960268e-01 3.06373000e-01 -1.36269033e-01
8.90151262e-01 -1.37094006e-01 -1.37262456e-02 1.02431707e-01
7.10841954e-01 -4.62455362e-01 6.70950532e-01 3.90074700e-01
1.78220704e-01 2.98306078e-01 4.33317423e-01 -8.14528048e-01
-8.44448447e-01 -8.68104756e-01 -2.76752770e-01 1.02319598e+00
6.90028295e-02 -4.54603583e-01 -3.58310454e-02 -3.33557278e-01
-2.84359962e-01 2.83831358e-01 -5.99749207e-01 -1.50038719e-01
-5.03169477e-01 -9.72344577e-01 6.36567712e-01 3.78167033e-01
4.19326544e-01 -1.25647855e+00 -5.23935914e-01 5.46195745e-01
-7.24879861e-01 -1.18160582e+00 -5.90960085e-01 -4.43307832e-02
-8.66019428e-01 -8.67519379e-01 -5.29958367e-01 -6.76502228e-01
1.97022215e-01 -6.85674092e-03 9.95566428e-01 -2.27665767e-01
3.48130949e-02 2.03111872e-01 -5.36585093e-01 -3.07151467e-01
6.89423531e-02 5.30181406e-03 -1.31250434e-02 7.11944342e-01
4.44492429e-01 -7.31405973e-01 -5.39140224e-01 4.04028445e-01
-1.23559606e+00 -1.14453010e-01 5.45585155e-01 9.31601584e-01
7.05312908e-01 4.34719861e-01 4.29577827e-01 -3.09141964e-01
7.50967383e-01 -7.98945308e-01 -3.79194796e-01 4.77816612e-01
-3.40396643e-01 -1.27797216e-01 8.45197499e-01 -1.05118728e+00
-1.02848375e+00 -4.30705518e-01 3.47133249e-01 -6.98649168e-01
1.62690461e-01 1.13120377e+00 1.83247449e-03 2.49012992e-01
2.31245726e-01 6.39991939e-01 -2.82840043e-01 -3.49891961e-01
-1.19152658e-01 3.56315106e-01 3.53167892e-01 -6.07988119e-01
5.30113816e-01 3.49800766e-01 1.18858121e-01 -5.07318437e-01
-7.93760002e-01 -1.24940656e-01 -4.88196284e-01 -3.12543541e-01
6.39466643e-01 -8.99249732e-01 -1.00712657e+00 9.70319390e-01
-1.32168424e+00 2.75677174e-01 -4.71534543e-02 8.57808828e-01
-2.51486689e-01 1.82503507e-01 -1.09943604e+00 -6.05446041e-01
-4.78222579e-01 -9.62647676e-01 9.05219495e-01 2.82932192e-01
7.87541866e-02 -9.73495781e-01 -1.40092120e-01 -1.03644431e-01
5.53920031e-01 5.09785056e-01 8.64522576e-01 -5.40575027e-01
-4.41330642e-01 -5.32174587e-01 -1.75402492e-01 -1.51223436e-01
1.94791943e-01 2.78987318e-01 -4.39087719e-01 -5.26416972e-02
4.26338434e-01 1.96662191e-02 4.54715014e-01 3.64625156e-01
1.29596007e+00 -5.09541750e-01 -7.70238116e-02 5.52696764e-01
1.18664455e+00 5.57139635e-01 6.76941633e-01 1.94147795e-01
8.49434078e-01 6.81792140e-01 5.39587200e-01 6.72901690e-01
7.76159048e-01 4.23323900e-01 2.39246070e-01 1.85846508e-01
5.29673874e-01 -1.45516381e-01 4.38041180e-01 1.47322726e+00
-3.21152776e-01 3.06947678e-02 -8.41560423e-01 4.15858448e-01
-1.99413311e+00 -1.35334468e+00 -4.34276044e-01 2.15727997e+00
8.32597375e-01 1.52496129e-01 4.73594852e-02 2.57198989e-01
8.11044991e-01 3.53422552e-01 -4.50665444e-01 7.83455092e-03
-3.37656021e-01 -2.91394293e-01 2.82249570e-01 -1.45890489e-01
-1.13517690e+00 3.95293206e-01 5.80417967e+00 8.70480776e-01
-1.51350939e+00 1.17622480e-01 4.60870981e-01 -1.12215385e-01
-1.77683681e-01 -3.03934872e-01 -4.37635928e-01 8.39285493e-01
1.23439598e+00 -3.62529516e-01 5.53062618e-01 1.69416711e-01
2.58653820e-01 4.02901769e-01 -7.21497059e-01 1.23634291e+00
9.29168165e-02 -1.08287442e+00 1.66667074e-01 -1.89939380e-01
2.44085386e-01 1.71284720e-01 3.04636151e-01 3.44469905e-01
-1.54552106e-02 -7.38182664e-01 6.43694997e-01 1.38223422e+00
4.67058897e-01 -8.58991086e-01 9.14883256e-01 7.33765215e-02
-1.77325940e+00 -1.25845209e-01 -1.28369138e-01 1.79680064e-01
2.96935916e-01 9.10856426e-01 -2.35858962e-01 9.61197078e-01
9.10298944e-01 1.22450864e+00 -5.36324799e-01 7.88708329e-01
2.54661769e-01 4.71667945e-01 -3.29138935e-01 -1.77322462e-01
1.13957174e-01 -2.76037514e-01 6.01976395e-01 1.07560265e+00
5.77245235e-01 5.81695400e-02 1.01253062e-01 5.90604067e-01
1.09901980e-01 2.90052921e-01 -3.78557116e-01 -2.55901992e-01
5.53130329e-01 1.34881747e+00 -5.03658056e-01 -4.45387453e-01
-5.39187431e-01 3.92062426e-01 5.76877482e-02 4.99956518e-01
-9.85400617e-01 -4.55656290e-01 3.13849077e-02 -2.95856714e-01
2.34091759e-01 -1.74958184e-01 6.15082942e-02 -1.61099672e+00
4.14766818e-01 -9.22864556e-01 6.85283840e-01 -9.28634107e-01
-1.89864492e+00 9.69501317e-01 -3.83277424e-02 -1.84551287e+00
-9.37601477e-02 -1.93573490e-01 -5.59534252e-01 6.85752094e-01
-1.24379086e+00 -1.33593106e+00 -1.80595204e-01 1.14798939e+00
2.87815124e-01 -4.18734282e-01 7.05556870e-01 5.93947172e-01
-7.73635685e-01 2.88867563e-01 1.46628050e-02 2.76544362e-01
5.77832401e-01 -8.46857786e-01 -3.52058649e-01 7.40722179e-01
-5.93778789e-02 7.44138181e-01 5.49262404e-01 -4.41790998e-01
-1.76803923e+00 -1.14622021e+00 9.95471478e-01 4.15285304e-02
1.18892515e+00 3.99873853e-02 -9.95490551e-01 6.69756532e-01
1.44771427e-01 2.49843135e-01 7.84985304e-01 1.99860796e-01
-4.97270554e-01 -6.27616107e-01 -7.77663410e-01 4.85581994e-01
4.64995891e-01 -7.92847753e-01 -6.35192037e-01 3.43938231e-01
7.04675436e-01 -4.52939481e-01 -1.51630795e+00 6.77058578e-01
6.72228336e-01 -7.74338663e-01 9.41113353e-01 -5.82510173e-01
3.88263315e-01 -4.69221771e-01 -3.20230216e-01 -1.03680301e+00
-3.42915714e-01 -5.44252038e-01 -4.48633730e-01 1.31166530e+00
1.74217150e-01 -8.15229893e-01 -2.39202734e-02 2.08745569e-01
6.13167956e-02 -8.30064833e-01 -9.77993906e-01 -5.11908472e-01
-5.11003792e-01 -4.40749705e-01 9.81760085e-01 1.21487188e+00
-2.22991761e-02 4.06920284e-01 -6.57578230e-01 1.66685790e-01
3.65335643e-01 3.07257503e-01 2.62901008e-01 -1.13979959e+00
-1.23705275e-01 -4.39270854e-01 -4.11229014e-01 -5.69335341e-01
-1.43670574e-01 -6.51501894e-01 -3.82456601e-01 -1.08909082e+00
1.36283994e-01 -2.11083680e-01 -8.78241301e-01 3.84087473e-01
-9.71450731e-02 -2.33764380e-01 -3.38228904e-02 4.72605050e-01
-6.22740984e-01 8.59774351e-01 1.11274028e+00 -1.87625229e-01
-2.68657565e-01 -7.60203451e-02 -1.23751573e-02 6.15365863e-01
6.12918913e-01 -2.78960884e-01 -3.58469158e-01 -3.90412688e-01
4.95184332e-01 7.17633665e-01 2.68453777e-01 -8.55636358e-01
4.15557265e-01 -2.13917986e-01 4.82311696e-01 -1.07747817e+00
4.09311146e-01 -1.05807209e+00 4.90598798e-01 2.57681310e-01
-2.58832544e-01 7.75197744e-01 7.18527585e-02 5.31589031e-01
-5.95843852e-01 2.96524584e-01 1.78187713e-01 9.22008827e-02
-2.01143339e-01 7.41204977e-01 -4.74301159e-01 -2.14213297e-01
7.08896458e-01 2.19721839e-01 -4.37388927e-01 -3.18269491e-01
-7.07426012e-01 1.07411966e-01 -2.41574839e-01 6.73748016e-01
7.30680704e-01 -1.85059047e+00 -6.34401381e-01 2.08068997e-01
1.66429713e-01 -2.03772321e-01 6.44288838e-01 1.27958584e+00
-1.58435762e-01 1.82169154e-01 -1.56551719e-01 -6.79411173e-01
-1.06235182e+00 7.35997796e-01 4.85486567e-01 -6.83093250e-01
-4.03572917e-01 2.67735779e-01 -1.53089672e-01 -4.65918362e-01
1.47477999e-01 -3.52714777e-01 -7.11404562e-01 5.05911648e-01
6.62396610e-01 3.44220340e-01 4.41822149e-02 -9.50438261e-01
-1.80727497e-01 5.77641726e-01 -1.08587541e-01 7.38384426e-02
1.58381641e+00 -1.75777242e-01 -6.05130672e-01 1.16628933e+00
1.22793496e+00 -1.75848633e-01 -5.40913701e-01 -6.81868196e-01
5.28167486e-02 -2.15349734e-01 -1.43975481e-01 -5.61127841e-01
-1.27170908e+00 7.96294093e-01 6.32526159e-01 8.07490110e-01
1.42272913e+00 -4.62262660e-01 1.09190977e+00 4.05343056e-01
3.07782412e-01 -8.75209987e-01 6.60074130e-02 8.20671082e-01
1.02311254e+00 -7.68952966e-01 -2.07641482e-01 -1.29830400e-02
-5.97454071e-01 1.37405705e+00 2.29883656e-01 -1.47036374e-01
1.02357793e+00 1.32627741e-01 -1.18334122e-01 -4.12303329e-01
-1.02926540e+00 5.37143238e-02 6.28923476e-01 2.83242762e-03
3.88756782e-01 6.47299271e-03 -4.06447560e-01 6.85149312e-01
-1.81929931e-01 9.01213661e-03 3.07720333e-01 6.38324916e-01
9.49575305e-02 -8.88139486e-01 -6.22166932e-01 6.44378126e-01
-6.92502618e-01 1.75813753e-02 3.12670827e-01 3.96894276e-01
1.38409391e-01 1.12907851e+00 1.18981607e-01 -8.05241406e-01
4.65482146e-01 -3.62138040e-02 1.73290551e-01 5.22626517e-03
-7.40167260e-01 3.60340089e-01 -1.84565350e-01 -6.08555019e-01
-6.89450681e-01 -7.65329599e-01 -1.14250112e+00 -4.60521311e-01
-1.44111753e-01 1.39575014e-02 4.85321164e-01 1.07599103e+00
4.20765460e-01 8.10679615e-01 1.06426263e+00 -7.99586773e-01
-2.48934135e-01 -1.13372004e+00 -6.81112826e-01 6.26736164e-01
3.96066785e-01 -7.87407100e-01 -3.13114524e-01 1.72873348e-01]
|
[6.930964469909668, 2.9453487396240234]
|
31279dcf-30dd-4610-8890-ff5634120670
|
applying-feature-underspecified-lexicon
|
2204.07228
| null |
https://arxiv.org/abs/2204.07228v1
|
https://arxiv.org/pdf/2204.07228v1.pdf
|
Applying Feature Underspecified Lexicon Phonological Features in Multilingual Text-to-Speech
|
This study investigates whether the phonological features derived from the Featurally Underspecified Lexicon model can be applied in text-to-speech systems to generate native and non-native speech in English and Mandarin. We present a mapping of ARPABET/pinyin to SAMPA/SAMPA-SC and then to phonological features. This mapping was tested for whether it could lead to the successful generation of native, non-native, and code-switched speech in the two languages. We ran two experiments, one with a small dataset and one with a larger dataset. The results supported that phonological features could be used as a feasible input system for languages in or not in the train data, although further investigation is needed to improve model performance. The results lend support to FUL by presenting successfully synthesised output, and by having the output carrying a source-language accent when synthesising a language not in the training data. The TTS process stimulated human second language acquisition process and thus also confirm FUL's ability to account for acquisition.
|
['Jiewen Zheng', 'Huang Liu', 'Huinan Zeng', 'Cong Zhang']
|
2022-04-14
| null | null | null | null |
['language-acquisition']
|
['natural-language-processing']
|
[ 2.78257191e-01 4.40875292e-01 -1.42214634e-03 -3.74718159e-01
-6.14070117e-01 -5.94135761e-01 8.14714372e-01 -1.63023293e-01
-4.25616324e-01 8.32227468e-01 2.89454967e-01 -8.83598089e-01
-3.08533125e-02 -4.65692341e-01 -7.30469346e-01 -3.73290360e-01
2.86680937e-01 7.11889148e-01 3.62297535e-01 -5.13002217e-01
2.16981977e-01 4.26006854e-01 -1.82946372e+00 5.82958937e-01
8.52881670e-01 2.30030552e-01 1.04005110e+00 7.84667850e-01
-2.85566300e-01 8.05792987e-01 -9.59205151e-01 -1.32605106e-01
1.50913283e-01 -9.32105839e-01 -7.79758751e-01 6.66416511e-02
3.48434955e-01 -7.23316669e-02 1.57528281e-01 6.54931486e-01
3.29200715e-01 2.02647179e-01 6.63165689e-01 -8.43935370e-01
-8.56985629e-01 1.18094718e+00 2.23910466e-01 3.88322860e-01
7.80669093e-01 2.25488022e-01 6.73965693e-01 -8.67128491e-01
7.13820636e-01 1.42573893e+00 2.87133902e-01 8.27429473e-01
-1.41760969e+00 -6.92651510e-01 -1.18587650e-01 -2.08643869e-01
-1.26299000e+00 -1.05462790e+00 5.71885824e-01 -5.29745579e-01
1.27017665e+00 3.46912891e-01 9.19774532e-01 9.78529990e-01
-1.14776976e-01 6.27269864e-01 1.73434293e+00 -1.03558671e+00
-7.49878287e-02 6.22020423e-01 -2.57128507e-01 4.20326114e-01
1.15698591e-01 7.63256967e-01 -9.45112109e-01 5.03258228e-01
8.38654160e-01 -1.08571744e+00 -3.05169493e-01 4.45318460e-01
-1.33759058e+00 8.22113693e-01 -6.03934601e-02 9.42981124e-01
-2.70403296e-01 -3.42690498e-01 2.74108738e-01 7.40600944e-01
1.28662229e-01 7.84861445e-01 -7.28099644e-01 -4.18425649e-01
-1.01653969e+00 2.10669190e-01 6.79354846e-01 9.82334435e-01
4.09396887e-01 6.78230882e-01 1.25982761e-01 1.03215802e+00
4.25414443e-01 8.03424001e-01 1.08893609e+00 -6.32806361e-01
4.39269722e-01 1.99151367e-01 -2.24144563e-01 -3.94935995e-01
-1.35884717e-01 2.83492267e-01 9.36224759e-02 1.54668704e-01
6.63952827e-01 -2.31402695e-01 -1.26732349e+00 1.83999693e+00
-1.78051367e-02 -4.07846034e-01 4.91811007e-01 5.89826047e-01
7.99259067e-01 9.04403329e-01 1.90532029e-01 -4.02363271e-01
1.15317404e+00 -6.93259656e-01 -7.80965567e-01 -3.44637990e-01
7.86147118e-01 -1.13186193e+00 1.49542797e+00 3.81167024e-01
-1.35730910e+00 -1.13966107e+00 -9.29280579e-01 1.79804638e-01
-3.97702157e-01 1.10253461e-01 5.85516989e-01 1.13375902e+00
-1.42412937e+00 4.77464676e-01 -5.91332495e-01 -5.92526019e-01
-4.86853927e-01 5.09691596e-01 -5.39296210e-01 3.40894818e-01
-1.35026670e+00 1.15663636e+00 8.05473924e-01 -1.28299922e-01
-6.06737971e-01 -1.40011758e-01 -9.97221053e-01 -3.55503291e-01
-2.19380870e-01 -3.67621481e-02 1.34080076e+00 -1.28448677e+00
-1.71566665e+00 9.85047936e-01 -7.69145563e-02 -2.29110464e-01
2.11674526e-01 6.42524511e-02 -9.06589150e-01 -8.23846310e-02
2.48709470e-01 7.95732617e-01 6.50639296e-01 -1.22058225e+00
-8.49565446e-01 -9.61109325e-02 -4.27524477e-01 2.64557660e-01
1.40937701e-01 5.36512017e-01 7.02182278e-02 -9.02186215e-01
-4.43829084e-03 -9.53643382e-01 3.38526040e-01 -8.12615812e-01
-3.45518559e-01 -2.72470921e-01 2.36416593e-01 -1.03691983e+00
1.21790850e+00 -2.01321554e+00 2.43891124e-02 4.20372635e-01
-5.99156082e-01 3.99732441e-01 -2.01025069e-01 5.00577450e-01
-1.85464233e-01 6.22352660e-01 1.29539659e-02 7.08013847e-02
-4.82299291e-02 3.83321881e-01 -8.32259059e-02 -3.10496800e-02
5.55459082e-01 7.00231791e-01 -6.47075176e-01 -3.10029685e-01
3.02292645e-01 3.24271619e-01 -2.94050932e-01 3.21582615e-01
-6.62369430e-02 6.40671730e-01 5.16227484e-02 2.83648342e-01
-6.07963912e-02 7.26210237e-01 1.00973502e-01 5.03705680e-01
-5.95100164e-01 8.27541530e-01 -1.04466665e+00 1.17025256e+00
-6.39980435e-01 6.65693939e-01 3.17183137e-02 -5.50678968e-01
1.10382187e+00 7.15432465e-01 -4.22581434e-01 -7.96509147e-01
3.08295429e-01 8.60950947e-01 9.90461111e-01 -5.51511228e-01
5.28612077e-01 -6.54953778e-01 8.05550814e-02 2.39864886e-01
9.42608565e-02 -6.04128659e-01 3.09545338e-01 -4.57842559e-01
3.91306758e-01 2.50327468e-01 1.70205623e-01 -7.70705819e-01
6.27279639e-01 9.95704457e-02 2.47102439e-01 6.58674836e-01
1.37477726e-01 4.82580096e-01 6.95863292e-02 2.80094948e-02
-8.54540944e-01 -9.82897162e-01 -2.30298474e-01 1.31820285e+00
-5.43044508e-01 -9.87508148e-02 -9.28189516e-01 -2.44261935e-01
-3.78269106e-01 1.49436116e+00 -2.67879844e-01 -1.22753913e-02
-8.72915447e-01 -6.70344010e-02 7.00939417e-01 3.27552319e-01
-5.72907031e-02 -1.81368625e+00 -3.96786779e-01 5.39316416e-01
-1.39709696e-01 -7.74932146e-01 -5.16764581e-01 4.48785186e-01
-5.90564072e-01 -7.21789956e-01 -6.67438686e-01 -1.31816542e+00
4.54956919e-01 -3.29084963e-01 5.64270854e-01 2.44214624e-01
2.92404145e-01 2.01102078e-01 -4.81129050e-01 -6.93510056e-01
-1.49785006e+00 1.47854596e-01 5.21099746e-01 -3.98954988e-01
4.26815778e-01 -1.69459879e-01 3.42504919e-01 2.24965259e-01
-9.21802580e-01 1.71106488e-01 3.43782425e-01 6.23420000e-01
3.03166270e-01 -1.46150291e-01 7.97429860e-01 -8.26516032e-01
7.54309893e-01 -2.08975658e-01 -3.31969559e-01 8.43083933e-02
-4.02628422e-01 2.91450143e-01 7.27648437e-01 -7.47196674e-01
-1.14817560e+00 5.14031807e-03 -7.38749444e-01 1.25960246e-01
-5.00784695e-01 5.74931622e-01 -4.71199512e-01 8.93098041e-02
8.12143743e-01 3.47898066e-01 1.47771686e-01 -4.58948761e-01
1.76863581e-01 1.03090107e+00 4.49759305e-01 -8.59813690e-01
7.65911639e-01 -5.72196186e-01 -5.91922224e-01 -1.42904878e+00
-2.50928868e-02 -7.26944506e-02 -6.92806423e-01 -1.41755521e-01
9.22379076e-01 -8.21999848e-01 -6.41048178e-02 6.32312179e-01
-1.17492652e+00 -7.78115034e-01 -5.27540445e-01 8.04704189e-01
-6.89066887e-01 -6.39832169e-02 -5.88923633e-01 -8.53406191e-01
9.85185727e-02 -1.34343469e+00 6.62637234e-01 1.01126283e-01
-6.14473045e-01 -1.03369713e+00 1.16117269e-01 1.05424121e-01
4.42008853e-01 -2.45287254e-01 1.18380523e+00 -1.00876498e+00
-3.36660221e-02 2.53773808e-01 3.67570877e-01 6.53583288e-01
4.23300713e-01 2.71926522e-01 -8.52499545e-01 -2.08047077e-01
1.15727164e-01 -3.51037681e-01 2.00238571e-01 2.02990592e-01
-6.82209432e-02 -3.09338540e-01 1.67350709e-01 2.48537973e-01
1.01675665e+00 6.87133133e-01 3.70641083e-01 1.85599983e-01
3.67645383e-01 9.74587858e-01 4.06405300e-01 -2.36591533e-01
2.49470457e-01 6.89586163e-01 -5.68959296e-01 7.95929432e-02
-6.30997419e-01 -5.12379766e-01 1.03060293e+00 1.61494350e+00
-8.12899396e-02 -2.13256761e-01 -1.07436752e+00 9.02697265e-01
-1.09976053e+00 -7.55461335e-01 -4.73807335e-01 2.26915097e+00
1.19024754e+00 3.55599642e-01 2.24574327e-01 3.98698896e-01
8.59509289e-01 -9.53052267e-02 1.88369766e-01 -1.22987294e+00
-3.08784157e-01 6.72876656e-01 1.50422171e-01 9.63627338e-01
-4.70995575e-01 1.42144167e+00 6.86575985e+00 5.83247364e-01
-1.23497832e+00 -1.34346545e-01 3.67952168e-01 3.53681207e-01
-5.46788692e-01 -9.20520276e-02 -1.05748188e+00 3.97023022e-01
1.81919897e+00 -2.88236678e-01 6.88272595e-01 3.52077872e-01
4.00613189e-01 -6.14616461e-02 -1.16073728e+00 4.60718513e-01
1.02271006e-01 -8.61661196e-01 2.22306162e-01 -8.69731456e-02
4.59071696e-01 -9.41111445e-02 -2.55862415e-01 5.04129171e-01
3.29553097e-01 -1.05280471e+00 1.30650866e+00 2.02143475e-01
1.11940122e+00 -5.16658366e-01 4.64634538e-01 4.55622792e-01
-1.09105814e+00 3.76985699e-01 -1.22402467e-01 -2.42956698e-01
2.52523839e-01 -3.63888979e-01 -1.24176979e+00 2.42460817e-02
1.85022548e-01 -5.85880652e-02 -8.57025027e-01 6.09770834e-01
-5.59111357e-01 1.25294638e+00 -4.92956728e-01 -2.16791764e-01
1.28897622e-01 8.02721307e-02 6.38933957e-01 1.36025155e+00
6.57804251e-01 -1.55238762e-01 1.75201818e-01 5.72718740e-01
6.51277542e-01 6.62474871e-01 -7.52103984e-01 -4.46483791e-01
5.74444771e-01 3.81578803e-01 -8.33233476e-01 -1.55272782e-01
-4.84604239e-01 9.68126893e-01 9.01239961e-02 2.59852022e-01
-3.02332908e-01 -4.48332429e-01 3.03771913e-01 5.59832275e-01
1.66892990e-01 -1.62664384e-01 2.08867360e-02 -6.70005202e-01
-2.36813948e-01 -1.27119112e+00 -2.77113765e-01 -6.60692155e-01
-1.14963222e+00 1.15721631e+00 2.35295534e-01 -5.24181008e-01
-8.57955158e-01 -8.94358099e-01 -4.05227035e-01 1.56251538e+00
-1.18766510e+00 -1.21829879e+00 4.64007318e-01 5.38531899e-01
7.73487449e-01 -5.99122643e-01 1.06368577e+00 -9.32792798e-02
-2.12481990e-01 5.81031740e-01 -2.63572484e-01 1.49125293e-01
5.87132573e-01 -1.38232660e+00 7.17082322e-01 9.68395114e-01
3.38206738e-01 7.57209897e-01 6.87524140e-01 -8.61236572e-01
-6.66654170e-01 -6.50404930e-01 1.44859529e+00 -3.84398550e-01
6.43804729e-01 -4.21819001e-01 -8.63223255e-01 6.15112364e-01
5.52873731e-01 -6.61665678e-01 6.54836833e-01 -1.40555590e-01
2.60794759e-01 3.64348918e-01 -1.04747200e+00 5.99777460e-01
8.92744064e-01 -5.88578045e-01 -1.19781387e+00 3.32300104e-02
8.78071249e-01 -2.54218906e-01 -7.70647526e-01 -1.87392756e-02
3.42067361e-01 -7.51274824e-01 2.93914020e-01 -5.91672599e-01
2.39399925e-01 -2.35824257e-01 -1.64855734e-01 -1.58040011e+00
-3.31845731e-01 -8.30921471e-01 5.55392265e-01 1.51468134e+00
9.00368512e-01 -6.82303548e-01 1.94557428e-01 4.75109309e-01
-4.70112413e-01 -5.61377183e-02 -1.01107895e+00 -9.51727033e-01
4.30776417e-01 -6.20137751e-01 6.90884411e-01 7.74579346e-01
2.75880415e-02 4.75982726e-01 -4.60944735e-02 -8.16993266e-02
-1.33560196e-01 -6.39001966e-01 4.36791241e-01 -1.31347251e+00
-4.74176317e-01 -5.44964135e-01 -2.91443139e-01 -4.50806111e-01
3.80636334e-01 -1.20592141e+00 5.49676836e-01 -1.22102845e+00
-6.65543795e-01 -7.27468371e-01 -1.20684141e-02 5.84419608e-01
-2.40383327e-01 1.74677745e-02 6.19199097e-01 6.67107180e-02
6.40311480e-01 1.81137994e-01 1.18222034e+00 3.78375173e-01
-6.96841300e-01 2.17004642e-01 -7.19905436e-01 5.14104843e-01
9.59872365e-01 -5.00451267e-01 -6.08532250e-01 -3.75092804e-01
-1.18544787e-01 1.79965153e-01 -2.53698766e-01 -1.14562511e+00
-2.11342618e-01 -3.03263783e-01 2.62295932e-01 1.05564065e-01
-1.35230506e-02 -7.19900608e-01 3.07009012e-01 3.68355423e-01
-2.93885916e-01 5.07562757e-01 7.88916171e-01 -3.41371208e-01
-1.95878044e-01 -7.05118835e-01 7.34962404e-01 -2.08365992e-01
-7.90645540e-01 -4.28304195e-01 -9.52995658e-01 2.44971246e-01
9.46112096e-01 -6.90726459e-01 -9.23196375e-02 -1.93105668e-01
-8.56857240e-01 -3.16110790e-01 5.76304913e-01 8.34795654e-01
4.01196718e-01 -1.27975011e+00 -9.12288249e-01 1.02790332e+00
1.13516292e-02 -6.13073766e-01 -4.41388011e-01 3.08769047e-01
-9.37464952e-01 6.86332524e-01 -4.41071868e-01 -1.24364808e-01
-1.10164642e+00 4.27309036e-01 2.62779057e-01 2.51029849e-01
-1.89671293e-01 9.11799073e-01 -1.00059263e-01 -7.79428720e-01
2.83724293e-02 -3.67716849e-01 -2.60557145e-01 -1.18222043e-01
3.83330941e-01 7.64524983e-03 -1.13367088e-01 -1.30333483e+00
-2.51218885e-01 1.47554532e-01 -8.40711668e-02 -8.55787635e-01
9.16170478e-01 -1.50768086e-02 -4.55475645e-03 1.05905211e+00
6.81136549e-01 6.87553167e-01 -6.55934811e-01 2.61154115e-01
1.44596800e-01 -1.69441774e-01 -7.88575858e-02 -9.73742664e-01
-5.92129827e-01 7.72865176e-01 4.38845098e-01 2.55993962e-01
7.26027489e-01 -7.87254572e-02 4.86007512e-01 6.77672550e-02
3.31451863e-01 -1.24264741e+00 -4.80833292e-01 7.54883409e-01
1.06578863e+00 -8.95182908e-01 -6.33355975e-01 -3.45115513e-01
-9.83026683e-01 1.10038781e+00 6.72680020e-01 2.01049641e-01
5.14029443e-01 2.94379890e-01 4.55063015e-01 7.89403170e-02
-7.38073528e-01 -3.58564466e-01 3.83822441e-01 8.93519044e-01
1.14302421e+00 3.51671427e-01 -6.64767981e-01 5.37058532e-01
-1.12401605e+00 -2.56314158e-01 5.92189968e-01 8.65976632e-01
-5.79364657e-01 -1.59193623e+00 -4.57479298e-01 5.74492216e-01
-3.01309884e-01 -4.29712236e-01 -6.41039848e-01 1.28134549e+00
5.73858500e-01 1.18615711e+00 1.43083051e-01 -3.60040188e-01
4.87519920e-01 7.18056977e-01 4.70179677e-01 -1.21551847e+00
-1.04365289e+00 3.00934613e-01 4.16428328e-01 8.98998380e-02
-2.27878630e-01 -1.04501927e+00 -1.33678794e+00 -6.72340989e-02
-5.49069643e-01 3.58866125e-01 5.62764287e-01 9.82887864e-01
-3.39247346e-01 4.27769065e-01 2.30564341e-01 -6.01615191e-01
-1.33573726e-01 -1.30031073e+00 -6.75979733e-01 1.78031653e-01
1.36240602e-01 -2.91483343e-01 -4.96107936e-01 2.89099127e-01]
|
[14.262474060058594, 7.05139684677124]
|
ac4fca6f-d19b-429c-9908-c6f73dc32886
|
cross-dataset-propensity-estimation-for
|
2212.13892
| null |
https://arxiv.org/abs/2212.13892v1
|
https://arxiv.org/pdf/2212.13892v1.pdf
|
Cross-Dataset Propensity Estimation for Debiasing Recommender Systems
|
Datasets for training recommender systems are often subject to distribution shift induced by users' and recommenders' selection biases. In this paper, we study the impact of selection bias on datasets with different quantization. We then leverage two differently quantized datasets from different source distributions to mitigate distribution shift by applying the inverse probability scoring method from causal inference. Empirically, our approach gains significant performance improvement over single-dataset methods and alternative ways of combining two datasets.
|
['Sarah Dean', 'Fengyu Li']
|
2022-12-22
| null | null | null | null |
['selection-bias']
|
['natural-language-processing']
|
[ 3.28828067e-01 -1.53220266e-01 -7.16300130e-01 -7.70469129e-01
-1.00646162e+00 -7.38877714e-01 7.27465391e-01 2.49143422e-01
-3.61715049e-01 1.12965286e+00 7.34944940e-01 -3.84548455e-01
-4.58915532e-01 -1.03103936e+00 -8.15235138e-01 -5.58528483e-01
-6.01768456e-02 4.11731988e-01 1.19858377e-01 -9.28345397e-02
7.07903802e-01 -2.87492108e-02 -1.69120824e+00 3.80259067e-01
8.16098809e-01 4.88924325e-01 -1.98943675e-01 5.22611260e-01
1.53519750e-01 7.83648431e-01 -1.03129613e+00 -3.31815898e-01
3.96010578e-01 -3.12443912e-01 -3.90151709e-01 -4.70840067e-01
8.82514060e-01 -3.54340434e-01 -3.88767153e-01 1.08503652e+00
8.65232229e-01 3.18577766e-01 1.00664222e+00 -1.27121174e+00
-1.10987961e+00 1.15141308e+00 -5.70651233e-01 6.57835901e-01
4.51160789e-01 3.48871239e-02 1.19291914e+00 -5.60158908e-01
5.86148143e-01 1.53993750e+00 8.10901225e-01 1.37288347e-01
-1.74100769e+00 -1.17498922e+00 1.41910300e-01 1.45688519e-01
-1.45548248e+00 -4.93357211e-01 4.21271920e-01 -5.94511390e-01
6.60263062e-01 2.22017467e-01 2.16586262e-01 1.44212329e+00
1.96343958e-01 4.71911788e-01 1.26809680e+00 -7.41846487e-02
5.63805759e-01 5.10852933e-01 7.32346326e-02 -3.98820527e-02
6.10956609e-01 2.83646077e-01 -7.95102179e-01 -7.80992687e-01
3.34239125e-01 -3.37689854e-02 -4.72879916e-01 -2.18947738e-01
-1.03719676e+00 1.34548044e+00 2.04528585e-01 -1.81686983e-01
-4.23689544e-01 3.29228282e-01 1.17792279e-01 3.85158271e-01
4.70996290e-01 5.70818365e-01 -7.85988271e-01 -1.87165484e-01
-1.05699360e+00 6.77575171e-01 8.75669837e-01 1.13176835e+00
3.55141997e-01 -1.76704168e-01 -4.88203645e-01 7.76179254e-01
4.71609205e-01 9.46505487e-01 6.00387990e-01 -8.34810674e-01
2.99775958e-01 1.51617587e-01 3.12163740e-01 -1.14314699e+00
-1.99909687e-01 -2.03859091e-01 -5.85477293e-01 -2.60526478e-01
5.39022624e-01 -4.40131158e-01 -8.05824637e-01 1.78328252e+00
3.82048428e-01 3.70856017e-01 4.14458551e-02 8.53644133e-01
6.89685822e-01 4.83114600e-01 2.60185748e-01 -1.29847229e-02
1.09129977e+00 -1.52276158e-01 -8.11903179e-01 4.50492293e-01
5.95183790e-01 -6.31553829e-01 1.45840228e+00 6.38657510e-01
-6.77134454e-01 -1.75603285e-01 -9.89369988e-01 1.95036590e-01
-3.13974082e-01 -1.50592670e-01 4.62505698e-01 1.04988492e+00
-7.67579794e-01 8.53769898e-01 -2.26762995e-01 -1.80357933e-01
7.85969496e-01 3.83690238e-01 2.72432357e-01 -1.13451809e-01
-1.58316302e+00 4.56289083e-01 2.34323163e-02 -7.52329588e-01
-7.05348432e-01 -1.43798149e+00 -2.83744693e-01 6.66052103e-02
2.08987668e-01 -5.82846820e-01 1.13189840e+00 -3.49105805e-01
-1.38120723e+00 -8.59539881e-02 7.85853341e-02 -5.81991792e-01
2.29731083e-01 -4.23023403e-01 -4.33424771e-01 -4.24376756e-01
8.69506523e-02 2.13530958e-01 7.44399667e-01 -1.20934618e+00
-7.54539669e-01 -3.85953397e-01 -6.91374838e-02 1.20664619e-01
-4.44351554e-01 -1.93897471e-01 8.68372247e-02 -9.56455052e-01
-5.10206342e-01 -9.62749541e-01 -4.25165653e-01 -5.85828364e-01
-4.05343086e-01 -3.28976721e-01 5.06703317e-01 -1.52099714e-01
1.59840667e+00 -2.11274505e+00 -2.31143475e-01 4.80280608e-01
7.17084929e-02 -1.52448460e-01 -1.27440855e-01 3.50212693e-01
2.09972903e-01 5.43001413e-01 2.07841307e-01 9.54830647e-03
1.16032690e-01 3.62711906e-01 -7.28899777e-01 4.99212772e-01
-7.23346695e-02 2.48474836e-01 -1.22839677e+00 -2.78285325e-01
-3.84587273e-02 5.66427648e-01 -1.10439444e+00 3.61917615e-02
-2.89739162e-01 2.97820777e-01 -2.71432132e-01 3.23383212e-01
6.04431868e-01 -2.59516299e-01 4.40182030e-01 -2.12674052e-01
9.13305581e-02 7.23308623e-01 -1.27294207e+00 1.44302988e+00
-4.10102934e-01 6.06210768e-01 -5.13122737e-01 -4.22496825e-01
7.20852315e-01 3.94323580e-02 4.16255772e-01 -6.46456718e-01
5.67826778e-02 -1.25247240e-01 2.20797285e-01 -1.07402086e-01
6.65222526e-01 -2.14071319e-01 -2.82000721e-01 6.51228249e-01
2.74357628e-02 5.44314608e-02 -1.50699943e-01 4.22474265e-01
1.17995083e+00 -2.72310287e-01 5.35780527e-02 -5.94742775e-01
-2.18936011e-01 -1.31951198e-01 5.83327293e-01 1.31339419e+00
-2.13636905e-02 4.71867859e-01 4.52926189e-01 -1.68832049e-01
-9.28308845e-01 -1.13150990e+00 -5.97272813e-01 1.28882420e+00
2.94515435e-02 -7.00109601e-01 -2.20244452e-01 -9.36228812e-01
6.50844038e-01 1.33559024e+00 -9.28256214e-01 -2.98092335e-01
-6.12079501e-02 -1.23065102e+00 7.41215944e-01 4.81633872e-01
-1.52204454e-01 -3.30538601e-01 -4.09780800e-01 2.14457382e-02
7.00569293e-03 -6.87606931e-01 -4.43473309e-01 2.87142485e-01
-7.19161987e-01 -9.43307340e-01 -2.54620343e-01 2.78830111e-01
2.14057431e-01 2.29857102e-01 1.41481602e+00 -3.02156657e-01
2.58553416e-01 1.72738612e-01 -3.77666652e-01 -7.81805813e-01
-3.18649024e-01 2.81098247e-01 5.04179835e-01 -2.68231094e-01
4.69259620e-01 -7.06024349e-01 -8.55202615e-01 2.84890145e-01
-8.62306952e-01 -7.83488870e-01 3.85465503e-01 7.17437327e-01
4.69979405e-01 4.02368665e-01 8.49699616e-01 -1.61941874e+00
1.08069539e+00 -1.12671065e+00 -4.65333760e-01 -1.34357572e-01
-1.34480274e+00 2.13751167e-01 5.72199225e-01 -5.69329262e-01
-1.04287326e+00 -3.73225421e-01 2.67336011e-01 -4.55807984e-01
-2.49515042e-01 2.80692190e-01 -3.27435970e-01 3.41715693e-01
1.10909891e+00 -5.34892380e-01 -3.75769019e-01 -4.61515993e-01
8.38245392e-01 9.53943074e-01 2.03286558e-01 -5.64271331e-01
5.47731876e-01 4.23529506e-01 -3.12678427e-01 -3.74573231e-01
-1.04083085e+00 -2.66444266e-01 -5.02663076e-01 2.08490670e-01
4.21514958e-01 -1.01278532e+00 -2.43863806e-01 -1.65817276e-01
-7.06128657e-01 -4.42723006e-01 -5.95523596e-01 6.78721786e-01
-2.57445476e-03 -1.86539188e-01 -1.44126192e-01 -6.33828998e-01
-1.18491791e-01 -9.38445807e-01 1.01344156e+00 1.19781882e-01
-3.13189119e-01 -1.00149643e+00 4.28294152e-01 -7.54967853e-02
4.34517682e-01 1.03437761e-02 6.02928579e-01 -1.00157559e+00
-2.25437462e-01 4.40024994e-02 -2.28569254e-01 -1.57756160e-03
3.82341653e-01 1.98487148e-01 -1.03428400e+00 -3.37750912e-01
-4.66543794e-01 -1.23983711e-01 8.11584473e-01 5.09041607e-01
1.20219696e+00 -6.15210116e-01 -3.98383707e-01 4.15207863e-01
1.44616210e+00 6.87074736e-02 4.99964207e-01 1.79390848e-01
6.96547568e-01 5.00436246e-01 8.31154048e-01 9.41326320e-01
6.49201870e-01 5.70534766e-01 7.33390823e-02 1.86441988e-01
-6.37294725e-02 -5.32921910e-01 2.99745470e-01 4.62710828e-01
3.03696275e-01 -6.03557765e-01 -8.23557913e-01 6.23137057e-01
-1.64668810e+00 -1.11718607e+00 -3.86365592e-01 2.48897076e+00
1.01505709e+00 1.15782671e-01 3.35661680e-01 -3.00878696e-02
4.90221500e-01 -1.69883668e-01 -6.10581636e-01 -3.44595969e-01
-1.27458423e-01 1.56763136e-01 1.20998812e+00 3.84922683e-01
-9.01778460e-01 5.74392319e-01 7.78797340e+00 8.05182517e-01
-7.88082838e-01 1.80531144e-01 3.48182350e-01 -6.43920243e-01
-8.18976045e-01 1.63240060e-02 -1.00699306e+00 8.89184058e-01
1.56310129e+00 -4.43667114e-01 2.17624366e-01 4.69424158e-01
1.21943235e-01 -9.25967991e-02 -1.09547961e+00 6.67899072e-01
-8.35570246e-02 -1.26379633e+00 2.72289157e-01 4.39797014e-01
1.10857654e+00 2.53838629e-01 4.08402950e-01 3.44065428e-01
1.18884611e+00 -9.24596429e-01 4.12395835e-01 6.77665651e-01
7.50368834e-01 -8.59649420e-01 7.37603009e-01 4.34692064e-03
-6.34415567e-01 -2.48868451e-01 -4.63974684e-01 -1.77389726e-01
-9.36600268e-02 1.12778997e+00 -1.00521922e+00 2.02939078e-01
9.39743936e-01 5.78091860e-01 -5.74359953e-01 7.54930079e-01
-2.71271691e-02 1.36924756e+00 -4.00486648e-01 -1.98236123e-01
-3.40884268e-01 1.13719702e-01 3.96158099e-01 1.29239380e+00
2.75456727e-01 1.63908973e-01 -2.87252665e-02 5.98278284e-01
-2.89412498e-01 1.91239025e-02 -8.45204532e-01 2.37520918e-01
1.18618059e+00 7.40992367e-01 -3.77838522e-01 -6.39306605e-01
-3.35695148e-01 4.13349062e-01 4.31640306e-03 3.40899438e-01
-8.35711241e-01 -2.82056928e-01 8.41355741e-01 7.88341314e-02
3.62045646e-01 2.81936735e-01 -4.17842895e-01 -1.13235509e+00
-6.97127521e-01 -7.94917107e-01 6.18231535e-01 -4.54355806e-01
-1.76895142e+00 1.41896717e-02 3.19535762e-01 -9.60211515e-01
3.45127098e-02 -1.26738131e-01 -2.60887802e-01 7.98514605e-01
-1.39632750e+00 -4.59707439e-01 1.23778880e-02 5.03475428e-01
1.98348090e-01 -1.50317296e-01 7.66217828e-01 6.44781590e-01
-4.24294442e-01 1.03481340e+00 5.44656217e-01 -3.13676029e-01
1.31232715e+00 -1.75250435e+00 2.31677964e-01 2.67503083e-01
2.24979624e-01 9.36756909e-01 8.54463339e-01 -7.17301488e-01
-1.23868930e+00 -1.38445711e+00 4.29043472e-01 -1.02243185e+00
6.55317187e-01 -3.34998727e-01 -8.74274671e-01 5.36060691e-01
2.24521115e-01 -3.26692194e-01 1.49996948e+00 6.60636604e-01
-5.17280161e-01 -6.10800013e-02 -1.42127597e+00 3.73526931e-01
9.01976049e-01 -3.66710156e-01 -5.42174459e-01 2.34588280e-01
8.82110417e-01 -1.64291188e-01 -1.55426168e+00 3.20097655e-01
7.88626075e-01 -6.35572076e-01 8.40574503e-01 -8.11618268e-01
5.23921371e-01 -2.57332742e-01 -6.30831480e-01 -1.73235929e+00
-6.00897074e-01 -3.89496982e-01 -3.96302305e-02 1.51384211e+00
5.39069295e-01 -5.06436348e-01 6.73065722e-01 4.68424112e-01
4.24791247e-01 -4.46197808e-01 -5.38826168e-01 -4.02446300e-01
3.44948649e-01 -3.99914294e-01 1.00042927e+00 1.23979163e+00
-1.18085176e-01 7.31625021e-01 -2.92385131e-01 4.18080032e-01
8.51784468e-01 1.73666611e-01 9.63273823e-01 -1.26977134e+00
-3.13082337e-01 -2.94291437e-01 -9.98729914e-02 -8.91488850e-01
-9.39732790e-02 -6.76143408e-01 5.17919883e-02 -1.00320780e+00
3.08116190e-02 -8.78046095e-01 -7.09175169e-01 3.12791407e-01
-3.29016179e-01 3.72513413e-01 -8.75808224e-02 3.39794904e-01
-6.15375757e-01 6.43744588e-01 7.35140979e-01 8.13317746e-02
-2.30037436e-01 -1.03040308e-01 -1.11017752e+00 6.15075707e-01
9.43635762e-01 -8.39741170e-01 -8.89594793e-01 -2.59481519e-01
4.92048919e-01 -3.68392348e-01 1.49024397e-01 -6.50438726e-01
-7.12652365e-03 -2.27636561e-01 3.93351823e-01 -6.02477014e-01
-2.71185189e-01 -5.28549790e-01 5.61441407e-02 1.26196714e-02
-8.47133815e-01 1.00155741e-01 1.94437116e-01 9.41594303e-01
1.23723097e-01 2.11990640e-01 5.66748142e-01 1.86386973e-01
-1.51956722e-01 1.09555805e-02 -5.93228459e-01 3.18521947e-01
5.11842728e-01 3.62656504e-01 -4.27184612e-01 -3.01464677e-01
-3.15052330e-01 2.22424075e-01 4.18522924e-01 4.99149114e-01
4.94168103e-01 -1.37807202e+00 -7.04755247e-01 -1.10706814e-01
1.02181576e-01 -2.42638677e-01 -8.36318880e-02 7.10627317e-01
2.64805764e-01 2.58653134e-01 1.56000152e-01 -4.36018109e-01
-1.00019598e+00 3.98584634e-01 3.03261708e-02 5.36811799e-02
-1.29113346e-01 9.70658064e-01 3.81437340e-03 -7.27114737e-01
2.00336128e-01 -5.00470281e-01 -2.83231139e-01 2.53514171e-01
5.14475465e-01 6.35028839e-01 4.74768542e-02 -2.93694675e-01
-3.11794639e-01 -3.45401070e-03 -1.47534147e-01 -2.23243147e-01
1.21466112e+00 -2.67736942e-01 3.61244589e-01 7.07094371e-01
9.30238128e-01 5.05378246e-01 -1.02129543e+00 -2.00021699e-01
-1.26659825e-01 -9.63814795e-01 3.33721697e-01 -1.03611326e+00
-9.54840839e-01 5.39048970e-01 8.79375994e-01 4.39098626e-01
9.63911474e-01 -2.41787910e-01 5.27525187e-01 4.55211289e-02
4.20344472e-01 -8.60277712e-01 -2.56771296e-01 2.58966368e-02
5.44359267e-01 -1.21544158e+00 4.80152965e-01 -1.39665470e-01
-4.50727969e-01 5.33433676e-01 3.91627431e-01 -3.31742495e-01
1.17158186e+00 3.86748910e-01 1.67636592e-02 -2.47469291e-01
-1.19770002e+00 -5.14650904e-03 3.21707517e-01 7.72204578e-01
7.46640384e-01 2.04190731e-01 -4.01999116e-01 8.74363124e-01
-6.64192438e-01 1.41820088e-01 7.81902730e-01 6.87140405e-01
-3.95431370e-02 -1.20777130e+00 -3.38552892e-01 1.10717285e+00
-5.73846459e-01 -4.11956966e-01 -4.61476684e-01 6.19990587e-01
4.44639027e-01 1.14026916e+00 1.34307846e-01 -7.19228506e-01
6.13938510e-01 -3.32224876e-01 2.91175872e-01 -6.57755852e-01
-5.87241828e-01 -6.15317933e-02 1.07443936e-01 -5.77635527e-01
-4.78536338e-01 -1.20443988e+00 -1.05875826e+00 -4.82045174e-01
-6.68582380e-01 3.07629466e-01 4.48245317e-01 5.36894143e-01
8.01143825e-01 6.96011961e-01 7.20683336e-01 -5.19534826e-01
-8.70277762e-01 -9.69350278e-01 -6.06481373e-01 6.11013591e-01
1.80711418e-01 -1.06901491e+00 -5.72539866e-01 -5.06106727e-02]
|
[9.705270767211914, 5.476047039031982]
|
8d534311-5658-4151-b20d-06b290ed0a00
|
enhancing-the-open-domain-dialogue-evaluation
| null | null |
https://aclanthology.org/2021.findings-acl.432
|
https://aclanthology.org/2021.findings-acl.432.pdf
|
Enhancing the Open-Domain Dialogue Evaluation in Latent Space
| null |
['Rui Yan', 'Shuming Shi', 'Dongyan Zhao', 'Haisong Zhang', 'Juntao Li', 'Lemao Liu', 'Zhangming Chan']
| null | null | null | null |
findings-acl-2021-8
|
['dialogue-evaluation']
|
['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]
|
[-7.375680923461914, 3.5948355197906494]
|
1b07b9a2-b083-4947-937b-b5145cfe2136
|
abnormal-chest-x-ray-identification-with
|
1903.02040
| null |
http://arxiv.org/abs/1903.02040v1
|
http://arxiv.org/pdf/1903.02040v1.pdf
|
Abnormal Chest X-ray Identification With Generative Adversarial One-Class Classifier
|
Being one of the most common diagnostic imaging tests, chest radiography
requires timely reporting of potential findings in the images. In this paper,
we propose an end-to-end architecture for abnormal chest X-ray identification
using generative adversarial one-class learning. Unlike previous approaches,
our method takes only normal chest X-ray images as input. The architecture is
composed of three deep neural networks, each of which learned by competing
while collaborating among them to model the underlying content structure of the
normal chest X-rays. Given a chest X-ray image in the testing phase, if it is
normal, the learned architecture can well model and reconstruct the content; if
it is abnormal, since the content is unseen in the training phase, the model
would perform poorly in its reconstruction. It thus enables distinguishing
abnormal chest X-rays from normal ones. Quantitative and qualitative
experiments demonstrate the effectiveness and efficiency of our approach, where
an AUC of 0.841 is achieved on the challenging NIH Chest X-ray dataset in a
one-class learning setting, with the potential in reducing the workload for
radiologists.
|
['Yu-Xing Tang', 'You-Bao Tang', 'Jing Xiao', 'Ronald M. Summers', 'Mei Han']
|
2019-03-05
| null | null | null | null |
['one-class-classifier']
|
['methodology']
|
[ 5.72059035e-01 5.59000254e-01 2.40560900e-02 -5.27185977e-01
-1.19961929e+00 -5.08326530e-01 2.24812955e-01 4.62941043e-02
-2.73054630e-01 5.31608820e-01 9.33169872e-02 -7.87857413e-01
-6.25250190e-02 -7.15258837e-01 -1.04282749e+00 -7.49587655e-01
2.33451519e-02 7.76428223e-01 1.01352736e-01 3.60658944e-01
-3.92476290e-01 2.70965517e-01 -7.78898537e-01 7.10752249e-01
3.71371031e-01 1.04346168e+00 1.73694149e-01 1.42864108e+00
4.13655251e-01 1.42751038e+00 -6.53763592e-01 -5.28949082e-01
3.91307205e-01 -8.80021513e-01 -9.65390563e-01 2.17158496e-01
4.03288960e-01 -9.02156711e-01 -8.46838534e-01 7.92966247e-01
7.10623324e-01 -1.84071913e-01 7.00376987e-01 -7.24717915e-01
-4.58586603e-01 4.70635265e-01 -3.23328972e-01 4.25509691e-01
1.80445492e-01 4.34303284e-01 8.50113809e-01 -6.96752369e-01
5.52573383e-01 6.43254757e-01 6.06768548e-01 8.40880036e-01
-9.46286023e-01 -5.81865847e-01 -2.16502935e-01 -1.32885724e-01
-8.06229532e-01 8.15740898e-02 5.19150913e-01 -3.34900916e-01
3.69525164e-01 5.91789544e-01 7.62948692e-01 1.39324367e+00
7.15357721e-01 7.86821067e-01 9.56886649e-01 -2.50235468e-01
1.81148425e-01 -1.70686916e-01 -3.43048163e-02 7.10882604e-01
8.67038369e-02 9.45345312e-02 6.28554076e-02 -4.00102645e-01
8.15516770e-01 5.37674546e-01 -3.23924601e-01 -3.54099393e-01
-1.37303841e+00 7.76072562e-01 6.22690737e-01 3.08322340e-01
-7.61880577e-01 1.16975583e-01 3.53520364e-01 2.80192435e-01
1.86336532e-01 2.39511803e-01 -9.74681675e-02 2.37069488e-01
-9.61072266e-01 9.81380790e-02 5.77136815e-01 3.57477665e-01
-2.25684583e-01 -1.53558314e-01 -3.34854871e-01 5.33016801e-01
1.45268723e-01 4.83099818e-01 8.64768386e-01 -6.39341593e-01
3.11300695e-01 5.59519053e-01 -1.67904541e-01 -2.40211055e-01
-5.17046511e-01 -6.42972112e-01 -1.09680617e+00 3.12262923e-01
4.41286266e-01 -1.86469033e-01 -1.54504037e+00 1.46473598e+00
2.70364523e-01 2.30182886e-01 1.44976124e-01 1.06325042e+00
9.67421889e-01 4.00495559e-01 1.35967210e-01 -4.74659614e-02
1.35825527e+00 -7.08496630e-01 -2.99290389e-01 -3.15450132e-01
4.11097258e-01 -8.30699980e-01 1.00112569e+00 3.71535748e-01
-1.53015196e+00 -6.86429143e-01 -1.07044554e+00 2.86499977e-01
2.54194200e-01 1.90141007e-01 2.97122329e-01 4.55058336e-01
-9.01257575e-01 2.97150552e-01 -1.07755136e+00 2.03042012e-03
6.18094444e-01 3.71517509e-01 -2.80872434e-01 -3.97873610e-01
-1.01614845e+00 7.48258412e-01 3.57918859e-01 -4.08430435e-02
-1.25934589e+00 -7.88386166e-01 -3.12280148e-01 1.85256988e-01
4.31880563e-01 -9.64807987e-01 1.58991337e+00 -8.77302170e-01
-9.29452002e-01 1.11292160e+00 2.80238181e-01 -5.54618597e-01
1.08910680e+00 4.75630350e-02 -3.86410177e-01 6.45395219e-01
1.42351881e-01 5.54531932e-01 9.09753144e-01 -1.27177799e+00
-4.88345623e-01 -3.02866757e-01 1.09487191e-01 -1.15903929e-01
1.55074507e-01 -2.95751125e-01 -3.95350337e-01 -8.92620802e-01
3.78220648e-01 -1.15352535e+00 -3.21628928e-01 2.23690152e-01
-5.04282892e-01 3.02173793e-01 6.04362190e-01 -7.98632383e-01
8.85777473e-01 -2.18251801e+00 -3.67034346e-01 4.11063552e-01
6.88766062e-01 1.69075221e-01 2.40317181e-01 -8.49040300e-02
-6.18195713e-01 1.17194444e-01 -4.42191362e-01 4.06086817e-02
-2.77808249e-01 2.94599891e-01 -3.28780323e-01 5.01140118e-01
4.24740463e-01 1.22702312e+00 -7.62753367e-01 -4.92655158e-01
7.58027434e-02 4.36196864e-01 -3.63878071e-01 7.56476283e-01
-4.48243655e-02 7.83685863e-01 -4.75879431e-01 6.62756622e-01
3.78232747e-01 -8.38660061e-01 1.33922756e-01 1.14602251e-02
8.45168710e-01 9.28873420e-02 -7.01415598e-01 1.26914513e+00
-4.89668757e-01 4.41080749e-01 -5.31368218e-02 -8.27282608e-01
4.87555534e-01 7.39388287e-01 6.24010444e-01 -8.46715987e-01
2.30386332e-01 2.38150418e-01 5.57802975e-01 -7.52650201e-01
-1.14964016e-01 -5.41931629e-01 1.40168637e-01 1.07636249e+00
-1.23359188e-01 -1.06147259e-01 -1.62476525e-01 3.45802009e-01
1.61229861e+00 -3.36431384e-01 1.26840755e-01 3.21329683e-01
3.77372712e-01 -2.35497981e-01 -2.68176273e-02 1.15576661e+00
-1.27067283e-01 1.22199154e+00 4.87970978e-01 -6.48401916e-01
-1.21537459e+00 -1.41688561e+00 -2.43578568e-01 8.32691669e-01
-2.77550727e-01 2.46405825e-01 -6.64334416e-01 -1.19906187e+00
-2.76607722e-01 6.25823617e-01 -1.03147435e+00 -3.21323156e-01
-8.90412509e-01 -7.95318007e-01 5.56327403e-01 7.50458121e-01
2.37249255e-01 -1.14467061e+00 -1.07100189e+00 1.69025853e-01
-2.45875791e-01 -8.16721499e-01 -2.63977557e-01 2.67806411e-01
-8.19336534e-01 -1.51828730e+00 -1.07815659e+00 -6.30280256e-01
9.17755961e-01 -3.96458656e-02 1.42604041e+00 5.02461553e-01
-6.69520199e-01 5.42646229e-01 -1.99778959e-01 -4.66744363e-01
-1.02794588e+00 -1.92558393e-01 -4.51367646e-01 3.00894817e-03
-3.26600373e-02 -3.15079004e-01 -1.09293056e+00 1.38924718e-01
-1.22228479e+00 -4.25014421e-02 1.01845300e+00 1.13103974e+00
9.02856886e-01 -1.62891112e-02 3.37551266e-01 -1.47353435e+00
4.76974577e-01 -7.04314768e-01 6.83020353e-02 3.05284739e-01
-4.38302815e-01 -3.76867175e-01 7.40183353e-01 -5.13763964e-01
-8.12075257e-01 1.50287062e-01 -3.51935059e-01 -5.95972121e-01
-2.44417384e-01 2.45980352e-01 5.46507835e-01 1.27511635e-01
8.43379796e-01 5.30022621e-01 -1.00457154e-01 -1.82304367e-01
-2.15959430e-01 2.98936039e-01 9.05299544e-01 -1.31880298e-01
6.77000463e-01 5.73158205e-01 1.30734459e-01 -2.75698632e-01
-1.12016499e+00 -3.34830433e-01 -5.36084771e-01 -3.86135727e-01
1.06234014e+00 -7.57441938e-01 -3.52147400e-01 -2.34468449e-02
-8.21960270e-01 -1.75904810e-01 -6.15705550e-01 6.41515672e-01
-5.04013181e-01 2.01696455e-01 -5.47125518e-01 -4.98940438e-01
-4.99625981e-01 -1.35671532e+00 1.02296007e+00 -1.55306607e-01
-2.87824035e-01 -9.68834281e-01 1.48310840e-01 4.20625687e-01
2.79617637e-01 6.20060086e-01 1.40132940e+00 -1.07659686e+00
-5.18461347e-01 -8.46115530e-01 -4.84847687e-02 5.46900690e-01
-6.95524663e-02 -2.89360315e-01 -1.00303590e+00 -5.00496686e-01
5.32181501e-01 -5.23915410e-01 6.79906428e-01 5.17062962e-01
1.57637656e+00 -1.23736866e-01 9.81517322e-03 6.37204945e-01
1.10351562e+00 3.61454397e-01 5.28710902e-01 -9.57083423e-03
4.97516453e-01 2.05192536e-01 2.94297069e-01 2.03633517e-01
4.37832810e-02 -1.84724629e-01 8.48817527e-01 -7.75388837e-01
-2.19299465e-01 -1.21791653e-01 -1.44686446e-01 3.40871125e-01
1.01518683e-01 -4.05417264e-01 -1.20545936e+00 2.23061606e-01
-1.38635111e+00 -6.87040150e-01 -1.67076454e-01 1.90086293e+00
5.89932323e-01 3.76742154e-01 1.16351480e-02 1.60156265e-01
4.67907369e-01 7.44504109e-02 -8.84033561e-01 6.83835074e-02
1.42929450e-01 7.37336099e-01 1.12793520e-01 8.75453278e-03
-1.11595011e+00 -1.13439083e-01 6.60200548e+00 2.01617867e-01
-1.34116971e+00 2.58187592e-01 1.28409791e+00 -2.47635946e-01
-1.53362259e-01 -4.04797018e-01 -2.37949882e-02 4.23713893e-01
8.30946267e-01 1.58944726e-01 -1.99016050e-01 9.81649339e-01
-1.43777668e-01 -4.16063629e-02 -1.39265466e+00 7.50738263e-01
1.70619547e-01 -1.27583933e+00 4.94300053e-02 5.81286736e-02
6.49852633e-01 9.97575596e-02 2.86451191e-01 2.26197809e-01
8.67241398e-02 -1.24634504e+00 6.25757575e-01 4.88312781e-01
1.10778427e+00 -5.73682606e-01 1.08178985e+00 4.78858560e-01
-4.96503323e-01 1.97882075e-02 -4.07475419e-02 5.17897069e-01
-6.70747682e-02 1.55903503e-01 -1.34695506e+00 3.49008650e-01
5.64225376e-01 7.58693367e-02 -7.29815423e-01 8.86513114e-01
-1.77785188e-01 1.05588889e+00 -4.91421297e-02 3.52734894e-01
4.12136912e-01 2.59641349e-01 3.63184720e-01 1.24906111e+00
1.76173970e-01 5.13092816e-01 3.44012320e-01 6.65899694e-01
-3.87849361e-01 -6.76696673e-02 -5.88943183e-01 1.72879428e-01
-1.76574156e-01 1.01254988e+00 -8.67808223e-01 -7.09265649e-01
-3.66055369e-01 7.59765506e-01 -3.02674651e-01 1.93359032e-01
-1.04801893e+00 1.39169574e-01 -4.33257669e-01 5.31479836e-01
1.59832403e-01 5.04463613e-01 -3.97723317e-01 -9.31743562e-01
2.13191465e-01 -1.05064929e+00 5.92553735e-01 -9.83752489e-01
-1.24411476e+00 8.89962316e-01 -3.08277816e-01 -1.41348815e+00
-7.36511111e-01 -5.18722236e-01 -8.09206426e-01 8.56952012e-01
-1.21032715e+00 -1.25528681e+00 -5.50991833e-01 6.85663998e-01
5.63275456e-01 -1.97920233e-01 9.66409147e-01 1.66003048e-01
-7.90161490e-02 7.02916861e-01 -4.33898419e-02 4.47572827e-01
4.64033306e-01 -1.46073949e+00 2.63357341e-01 9.41788554e-01
1.98591389e-02 3.92900378e-01 3.49291831e-01 -5.71445882e-01
-1.05097580e+00 -1.29004407e+00 3.65886420e-01 -4.36042130e-01
4.41298127e-01 -1.79601684e-02 -1.31581879e+00 6.64756954e-01
-3.11122537e-02 4.15087014e-01 9.97679889e-01 -5.95234454e-01
-2.47387171e-01 2.70385891e-01 -1.28822041e+00 2.38036767e-01
4.99240160e-01 -4.43314701e-01 -6.58693194e-01 6.82702661e-01
3.35946888e-01 -8.56115818e-01 -9.89835083e-01 4.84400064e-01
6.58291340e-01 -1.03403628e+00 1.10432494e+00 -9.31966722e-01
9.15626466e-01 2.34304056e-01 -4.19503972e-02 -9.00085509e-01
-2.15405717e-01 -1.29856318e-01 -7.85697475e-02 1.79652885e-01
3.91821533e-01 -4.19563681e-01 1.03520894e+00 4.27961856e-01
-1.17667139e-01 -1.18831277e+00 -8.38108957e-01 -1.43027455e-01
2.54043993e-02 -4.40118551e-01 3.67264152e-01 7.63831139e-01
-7.79083550e-01 2.68463969e-01 -7.48581514e-02 8.39791894e-02
4.77122277e-01 1.14456132e-01 5.61206162e-01 -8.99784148e-01
-9.43627059e-01 -2.85390943e-01 -2.50183612e-01 -7.00606048e-01
-6.31439984e-02 -9.40744400e-01 -5.90853356e-02 -1.44713211e+00
5.11678696e-01 -3.40085745e-01 -4.94606316e-01 4.12964016e-01
-5.26841879e-01 6.12196863e-01 1.49622917e-01 4.25758541e-01
-3.58549684e-01 -1.68120474e-01 1.59347653e+00 -2.65930504e-01
1.58428520e-01 6.96825564e-01 -5.00884652e-01 8.87730062e-01
6.25721872e-01 -7.53103197e-01 -4.43592817e-01 -3.24626625e-01
3.77427191e-02 6.69038177e-01 7.49632120e-01 -9.31356847e-01
-1.29300859e-02 2.34426335e-01 9.91002023e-01 -7.62307167e-01
3.40393484e-01 -9.99283731e-01 1.55213907e-01 1.06714654e+00
-6.26173615e-01 2.32918024e-01 -7.70940185e-02 7.52165079e-01
-2.49904990e-01 -3.03617120e-01 1.12512124e+00 -6.43782437e-01
6.12139404e-02 4.25757438e-01 -1.98582351e-01 2.66023666e-01
9.92588401e-01 -2.04322278e-01 4.45165038e-02 -4.44309324e-01
-1.06106186e+00 -1.02132738e-01 1.50415167e-01 1.43610863e-02
9.81263161e-01 -9.08861816e-01 -1.11647284e+00 4.34879571e-01
-1.87707003e-02 3.51343572e-01 3.22301656e-01 8.56694698e-01
-8.15294504e-01 1.76117256e-01 4.58784029e-02 -9.68266428e-01
-1.26924992e+00 4.48344171e-01 5.85664332e-01 -7.56292999e-01
-8.23692083e-01 8.14623952e-01 6.30264640e-01 -1.08660735e-01
1.86209559e-01 -2.46234089e-01 2.85974026e-01 -4.48787600e-01
5.11345804e-01 -1.24551944e-01 4.56634372e-01 -4.44429308e-01
3.04209050e-02 6.79330602e-02 -4.99610633e-01 4.24470678e-02
1.45038116e+00 4.31766659e-01 2.01491550e-01 4.14894402e-01
1.24218392e+00 -2.62107193e-01 -9.27239001e-01 -2.14201316e-01
-4.53928918e-01 -3.89752924e-01 7.52004236e-02 -1.11878991e+00
-1.13044882e+00 9.67869282e-01 9.79628205e-01 3.51762086e-01
1.29799962e+00 2.60874152e-01 8.02561045e-01 4.42734629e-01
-2.45634913e-01 -3.87495995e-01 5.17294228e-01 -4.66497801e-02
7.63430893e-01 -1.35826766e+00 3.44976485e-02 -3.29034626e-02
-7.65326202e-01 1.13151264e+00 4.27413315e-01 -2.96123594e-01
4.06374335e-01 2.23663673e-01 2.40038916e-01 -4.20138031e-01
-7.89407194e-01 3.30975920e-01 2.67537951e-01 3.45721036e-01
4.19519901e-01 2.96975315e-01 2.22508922e-01 2.79254228e-01
-1.67052537e-01 -2.17387617e-01 3.85070324e-01 1.07949769e+00
-2.87877947e-01 -8.61271441e-01 -5.26035428e-01 1.01910353e+00
-1.05110371e+00 -5.17059769e-03 -4.58234757e-01 1.10775745e+00
6.19085692e-02 3.78923118e-01 8.08792487e-02 -2.40940109e-01
5.78220963e-01 1.87803835e-01 4.02882844e-01 -7.29415178e-01
-1.03057635e+00 2.38199979e-01 -4.21630949e-01 -4.05520171e-01
-1.50758073e-01 -4.11969900e-01 -1.24400795e+00 1.05604462e-01
-1.43354297e-01 -1.39849871e-01 4.55151975e-01 7.06649065e-01
-1.12003006e-01 1.06488740e+00 9.45077777e-01 -2.23684520e-01
-9.13775504e-01 -8.67790103e-01 -2.52020985e-01 8.27835560e-01
7.40211606e-01 -1.34387389e-01 -2.44672000e-01 3.43128026e-01]
|
[15.134420394897461, -1.9662021398544312]
|
a3364be9-5317-4c21-990b-6768f5f66d18
|
exemplars-guided-empathetic-response
|
2106.11791
| null |
https://arxiv.org/abs/2106.11791v3
|
https://arxiv.org/pdf/2106.11791v3.pdf
|
Exemplars-guided Empathetic Response Generation Controlled by the Elements of Human Communication
|
The majority of existing methods for empathetic response generation rely on the emotion of the context to generate empathetic responses. However, empathy is much more than generating responses with an appropriate emotion. It also often entails subtle expressions of understanding and personal resonance with the situation of the other interlocutor. Unfortunately, such qualities are difficult to quantify and the datasets lack the relevant annotations. To address this issue, in this paper we propose an approach that relies on exemplars to cue the generative model on fine stylistic properties that signal empathy to the interlocutor. To this end, we employ dense passage retrieval to extract relevant exemplary responses from the training set. Three elements of human communication -- emotional presence, interpretation, and exploration, and sentiment are additionally introduced using synthetic labels to guide the generation towards empathy. The human evaluation is also extended by these elements of human communication. We empirically show that these approaches yield significant improvements in empathetic response quality in terms of both automated and human-evaluated metrics. The implementation is available at https://github.com/declare-lab/exemplary-empathy.
|
['Soujanya Poria', 'Rada Mihalcea', 'Alexander Gelbukh', 'Devamanyu Hazarika', 'Deepanway Ghosal', 'Navonil Majumder']
|
2021-06-22
| null | null | null | null |
['empathetic-response-generation']
|
['natural-language-processing']
|
[-1.80277318e-01 1.91580862e-01 1.09689906e-01 -4.52773333e-01
-8.66247058e-01 -6.22188747e-01 7.26924896e-01 1.64979011e-01
-2.53904015e-01 9.46570873e-01 8.08716118e-01 4.00177449e-01
7.93373659e-02 -4.75111961e-01 -9.44800302e-02 -5.30869365e-01
5.51694036e-01 4.67244655e-01 -6.13989472e-01 -7.43199766e-01
5.63585043e-01 3.72918218e-01 -1.12924480e+00 7.53250122e-01
8.26851845e-01 7.56998479e-01 -3.52767199e-01 6.90221906e-01
-1.11180313e-01 1.44778645e+00 -9.51557279e-01 -8.80893409e-01
2.44593229e-02 -1.08985126e+00 -1.11616063e+00 -2.09128767e-01
-2.07171887e-01 -2.29186669e-01 1.97148323e-01 7.64413416e-01
7.98151374e-01 4.76441324e-01 7.60187984e-01 -1.16934311e+00
-8.27253044e-01 8.94432366e-01 -2.34253913e-01 -1.07096314e-01
1.03687835e+00 2.12089643e-01 9.90293086e-01 -8.43464971e-01
7.77489781e-01 1.26355684e+00 6.69224203e-01 9.93243515e-01
-9.07191932e-01 -5.72830141e-01 -4.89483565e-01 4.25825417e-02
-7.89299846e-01 -5.69620788e-01 1.16267228e+00 -4.98112768e-01
4.19400305e-01 3.50311249e-01 6.47259414e-01 1.66687012e+00
-4.35116768e-01 7.90770710e-01 1.34859169e+00 -4.11347121e-01
1.23106018e-01 6.46639466e-01 1.81816891e-01 1.38730317e-01
-4.57637936e-01 1.38487056e-01 -8.04481387e-01 -3.62433434e-01
4.99945402e-01 -3.68352205e-01 -4.07306969e-01 1.34851053e-01
-1.08558917e+00 1.22849083e+00 5.72186053e-01 3.37159634e-01
-7.59926498e-01 4.29714397e-02 6.49189532e-01 3.87126863e-01
4.01451886e-01 1.13503361e+00 2.39273891e-01 -6.05229318e-01
-7.52407193e-01 3.28125060e-01 1.16436195e+00 7.09990621e-01
4.37704891e-01 -2.24802896e-01 -4.34287816e-01 1.14166605e+00
-1.36897311e-01 2.63723016e-01 3.95192653e-01 -1.35364151e+00
7.80038834e-02 6.49978340e-01 6.45345151e-01 -1.11692715e+00
-3.73236269e-01 -1.88332513e-01 -3.13657433e-01 1.21462956e-01
4.92228568e-01 -3.56297940e-01 1.68668658e-01 1.80951798e+00
4.90864128e-01 -3.78222853e-01 3.50271791e-01 1.40296447e+00
1.07923234e+00 5.96599162e-01 2.54664063e-01 -5.84204271e-02
1.27616179e+00 -1.24907947e+00 -7.38070488e-01 -3.64741609e-02
7.97558486e-01 -1.11225724e+00 1.51849139e+00 3.77079964e-01
-1.40582430e+00 -2.93196350e-01 -6.93533480e-01 -1.48807019e-01
9.98052768e-03 2.08255574e-01 6.29308045e-01 2.48081610e-01
-7.05292821e-01 6.07580066e-01 1.07662000e-01 -4.99778569e-01
1.27373606e-01 -5.61611652e-02 -3.45578402e-01 2.15717316e-01
-1.48297334e+00 1.10694325e+00 -1.14634104e-01 -8.49945173e-02
-1.23401351e-01 -6.17690086e-01 -5.75271904e-01 -2.76352823e-01
-2.94237137e-01 -8.78312111e-01 1.58162379e+00 -1.61223733e+00
-1.91046739e+00 1.20447123e+00 2.04979151e-01 -2.60877460e-01
8.65049660e-01 -4.48263586e-01 -1.90154389e-01 5.55519402e-01
7.65497014e-02 8.89100432e-01 6.39163852e-01 -1.38148963e+00
-1.76375672e-01 1.17852226e-01 1.95234224e-01 2.35332593e-01
-2.31603220e-01 6.80890679e-01 3.60289574e-01 -6.77131295e-01
-5.24751961e-01 -9.64093626e-01 3.62330601e-02 -1.29100427e-01
-2.95904785e-01 -2.75962085e-01 3.66315037e-01 -4.86735046e-01
1.01282001e+00 -2.21266413e+00 -1.53974807e-02 1.30110336e-02
2.87838161e-01 8.67940113e-02 -1.48964971e-01 1.12699223e+00
-5.87385371e-02 -1.70851395e-01 -4.77004871e-02 -4.53561008e-01
3.21036369e-01 -1.31797761e-01 -5.84761143e-01 2.27366045e-01
1.18487738e-01 1.03758729e+00 -1.11999714e+00 -6.25749707e-01
-5.93088493e-02 5.72722435e-01 -5.50654054e-01 5.96514821e-01
1.11467831e-01 8.90813470e-01 -5.07217288e-01 2.19544277e-01
2.68818408e-01 -1.59219190e-01 -2.02399150e-01 8.85198489e-02
1.08808503e-01 3.81511688e-01 -4.54519749e-01 1.40363038e+00
-6.88753068e-01 5.34873605e-01 -2.38906555e-02 -3.83927196e-01
1.24387932e+00 4.88721788e-01 2.42338762e-01 -5.24803162e-01
5.15070438e-01 2.76647121e-01 -9.24283788e-02 -7.55581260e-01
7.87772655e-01 -5.58693290e-01 -5.81633091e-01 1.16370583e+00
-3.08922619e-01 -4.99293119e-01 1.13980837e-01 4.50163513e-01
8.33515584e-01 1.69002369e-01 4.40209329e-01 5.24465777e-02
4.93550837e-01 1.54028982e-01 2.31161043e-01 4.80677217e-01
-4.07711893e-01 5.31836808e-01 6.55712485e-01 -3.43014777e-01
-9.02555346e-01 -6.80328310e-01 1.65427923e-01 1.17273438e+00
7.15092570e-02 -2.42286921e-01 -9.84768510e-01 -2.51678497e-01
-2.41974518e-01 1.12546492e+00 -8.41344476e-01 -2.33396873e-01
-4.57220554e-01 -1.78464487e-01 8.11847985e-01 4.68402624e-01
8.21909457e-02 -1.74913192e+00 -1.02647042e+00 1.42675817e-01
-7.12108314e-01 -1.09437644e+00 -3.95943493e-01 -3.04811984e-01
-4.15250272e-01 -8.11025918e-01 -5.67407906e-01 -4.00629491e-01
5.51436365e-01 4.52526994e-02 1.14747787e+00 2.67178595e-01
-1.44444983e-02 4.17469025e-01 -9.41218376e-01 -3.17505628e-01
-8.57139468e-01 -3.29865158e-01 -1.77224532e-01 4.04863395e-02
4.68657911e-01 -6.20662034e-01 -7.33974576e-01 3.18834215e-01
-6.42638624e-01 1.33524537e-01 2.26790592e-01 9.69879508e-01
6.54371157e-02 -9.15042877e-01 9.45181012e-01 -9.59424198e-01
1.45107210e+00 -7.53057539e-01 4.52784210e-01 -1.32299230e-01
7.44551001e-03 -2.78387964e-01 8.24006379e-01 -6.31765246e-01
-1.03705668e+00 -1.81821823e-01 -6.86882436e-02 -2.83033580e-01
-3.42568725e-01 2.21046403e-01 2.62177080e-01 4.91662361e-02
1.17510295e+00 -4.47006553e-01 -8.46749693e-02 -1.32615060e-01
6.01292133e-01 8.42312634e-01 6.52250588e-01 -1.01379335e+00
3.16576421e-01 4.00694370e-01 -4.26699221e-01 -3.26533228e-01
-1.21608639e+00 -4.02989060e-01 -3.35822433e-01 -6.70359731e-01
6.08889759e-01 -6.49034619e-01 -8.01313341e-01 1.61579713e-01
-1.51058304e+00 -5.40452242e-01 -5.43436229e-01 5.39670050e-01
-1.01661611e+00 2.69302905e-01 -9.80562568e-01 -8.96676958e-01
-7.40135610e-01 -7.85053253e-01 9.36788023e-01 2.84646660e-01
-1.40368867e+00 -9.52807367e-01 3.55047166e-01 7.68956602e-01
4.25074369e-01 5.92515111e-01 6.60327256e-01 -7.97629237e-01
3.75641763e-01 -5.01312435e-01 -5.18709347e-02 1.79245844e-01
-1.61818489e-01 5.03697619e-02 -1.06491411e+00 3.36444825e-01
2.69041449e-01 -9.64861214e-01 2.54639655e-01 -1.99316189e-01
6.07281148e-01 -5.93187273e-01 2.47036427e-01 1.92265406e-01
7.66817212e-01 -2.11938381e-01 6.09033167e-01 2.19350234e-01
3.15037251e-01 1.41257989e+00 1.05264175e+00 9.40879285e-01
4.03380543e-01 5.20399928e-01 1.42082885e-01 1.74081530e-02
1.11330450e-02 -4.19911683e-01 4.58371639e-01 7.51466095e-01
4.61065993e-02 -1.20228469e-01 -7.22226143e-01 4.86618668e-01
-1.82142234e+00 -1.39081597e+00 -4.46307153e-01 1.67513537e+00
1.08164513e+00 -4.38547373e-01 3.90591145e-01 2.25028228e-02
7.09529936e-01 -7.79378507e-03 -1.34917527e-01 -1.11841393e+00
-2.04797193e-01 1.68201581e-01 -4.41691667e-01 6.57155991e-01
-5.20545304e-01 1.10155272e+00 5.30470037e+00 4.38012451e-01
-1.08013129e+00 1.06377497e-01 4.71527249e-01 -4.17890817e-01
-4.74612415e-01 -8.93452242e-02 -2.14821249e-01 3.60828191e-01
6.27986550e-01 -8.73663276e-02 2.91011661e-01 7.61676669e-01
5.22156775e-01 -6.93329126e-02 -1.23938692e+00 1.08878541e+00
2.22927406e-01 -7.92187333e-01 -1.89771578e-01 -4.01653856e-01
6.35296702e-01 -7.40995109e-01 9.68408957e-02 1.82570353e-01
4.52712923e-01 -1.11106431e+00 8.53940070e-01 8.14902008e-01
5.01425564e-01 -7.61717796e-01 7.80716479e-01 3.17613095e-01
-3.23135227e-01 -5.31484792e-03 -1.34762913e-01 -3.31526309e-01
4.69484538e-01 2.13037342e-01 -7.77520955e-01 8.80131349e-02
3.17231357e-01 4.18954879e-01 -1.69355229e-01 6.52058065e-01
-8.82713437e-01 4.25975531e-01 3.95764187e-02 -3.48197997e-01
3.19280207e-01 -2.79835254e-01 6.02323532e-01 1.36030066e+00
2.83001453e-01 4.02175158e-01 -2.38173410e-01 1.13156009e+00
-1.18675373e-01 6.38153732e-01 -5.28867424e-01 -1.46377087e-01
6.70062363e-01 1.61886322e+00 -2.20802411e-01 -2.01341122e-01
1.71227440e-01 1.06379974e+00 6.43412292e-01 1.75764114e-01
-8.98374677e-01 -2.05027416e-01 3.71979088e-01 -1.16026931e-01
-3.48073035e-01 3.93013030e-01 -4.11157578e-01 -7.69191921e-01
-8.00509229e-02 -1.13980103e+00 3.02931160e-01 -1.26507318e+00
-1.67890751e+00 7.26371646e-01 -3.74732018e-01 -1.20406854e+00
-6.45377219e-01 -2.04130307e-01 -9.71776307e-01 8.57849240e-01
-1.05342269e+00 -1.22616708e+00 -6.48095906e-01 4.64718491e-01
2.22991839e-01 2.21914664e-01 1.05227959e+00 9.75870714e-02
-2.82509774e-01 6.93504691e-01 -5.93684971e-01 1.12648807e-01
1.41913915e+00 -9.68419075e-01 -2.06682310e-01 2.76415646e-01
-2.36617967e-01 7.13678479e-01 1.08338606e+00 -1.68681189e-01
-7.62933552e-01 -5.57898343e-01 1.18863320e+00 -6.57626987e-01
9.82475400e-01 1.18238777e-02 -7.54595637e-01 2.66287565e-01
5.36207616e-01 -3.97834241e-01 1.37046421e+00 -7.02676326e-02
-5.26267767e-01 4.51424450e-01 -1.33709991e+00 9.10912931e-01
8.39072883e-01 -6.18113160e-01 -8.15917313e-01 3.17205697e-01
4.06306952e-01 -2.56543010e-01 -1.02791882e+00 -2.56162230e-02
5.33569276e-01 -1.31655347e+00 6.08653188e-01 -8.50797117e-01
1.12787449e+00 1.00509919e-01 -5.24101891e-02 -1.46328080e+00
-9.30886157e-03 -1.17617416e+00 1.83675736e-01 1.39006889e+00
2.85233736e-01 -4.91851240e-01 4.93060827e-01 8.98977697e-01
-1.84766099e-01 -7.87280381e-01 -3.90716553e-01 -2.37272099e-01
3.18321556e-01 -2.46102735e-01 6.66443169e-01 1.28881073e+00
8.36966813e-01 6.55101717e-01 -6.34077668e-01 -4.90583360e-01
2.28296116e-01 4.49859411e-01 1.21086586e+00 -1.00179958e+00
-3.20394248e-01 -8.34758878e-01 -2.80527445e-03 -7.56499946e-01
4.74423945e-01 -7.47344971e-01 3.58547159e-02 -1.37392807e+00
1.21284761e-01 -3.51375997e-01 1.37396559e-01 1.34744391e-01
-2.46610776e-01 2.97614157e-01 3.17670673e-01 3.77981037e-01
-5.96200883e-01 5.78241527e-01 1.43687344e+00 4.13080186e-01
-2.74684131e-01 -1.98708773e-01 -1.01090360e+00 6.81354046e-01
1.32389581e+00 -5.48603117e-01 -2.34261170e-01 -4.71179113e-02
5.21821856e-01 4.89268869e-01 6.66387856e-01 -6.46166921e-01
1.58097774e-01 -1.43865928e-01 -2.21229680e-02 -1.25904549e-02
7.34784663e-01 -3.49644184e-01 6.44560307e-02 1.63060218e-01
-9.26759064e-01 6.74298033e-02 -2.22138166e-01 1.10198550e-01
-3.83470118e-01 -5.55912137e-01 1.02898669e+00 -8.46676975e-02
-2.30290204e-01 -3.02621901e-01 -2.49876648e-01 2.99090236e-01
9.15957332e-01 -2.68908441e-01 -5.26323676e-01 -1.13741815e+00
-5.84251702e-01 8.86200890e-02 8.84696662e-01 2.77595818e-01
5.66244006e-01 -1.45721078e+00 -1.02437150e+00 -4.35138196e-01
3.80557448e-01 -6.05631173e-01 3.89394015e-01 1.22661865e+00
-4.34879243e-01 -2.61983909e-02 -4.46179658e-01 3.01735532e-02
-1.10183847e+00 4.39526558e-01 3.51055145e-01 -1.47012338e-01
-4.78525966e-01 9.42674160e-01 1.78038493e-01 -3.93946558e-01
-7.41312280e-02 3.61628771e-01 -4.77839530e-01 3.66507441e-01
5.98091304e-01 4.08414781e-01 -4.56702560e-01 -1.05421078e+00
5.22631407e-02 3.26131135e-01 2.23555461e-01 -4.85608310e-01
1.06702864e+00 -8.00822452e-02 -2.65939742e-01 5.22681594e-01
1.05844784e+00 4.76971120e-01 -8.96306992e-01 -7.20788538e-03
-2.12902837e-02 -4.72186655e-01 -7.05995917e-01 -9.59419906e-01
-5.23833334e-01 1.08488154e+00 -2.90114731e-01 1.83446676e-01
8.72520745e-01 -8.65050708e-04 1.14758575e+00 1.98034525e-01
2.48231038e-01 -1.33519268e+00 5.49572647e-01 3.52151424e-01
1.37176907e+00 -1.03292191e+00 -3.01661491e-01 -1.74033165e-01
-1.42866540e+00 1.09619772e+00 5.88891327e-01 -2.98858404e-01
-7.49053285e-02 8.37209597e-02 7.52817988e-01 -3.13777983e-01
-8.37868154e-01 1.22001566e-01 3.54127921e-02 4.55445766e-01
8.01251531e-01 1.72674969e-01 -6.20437503e-01 1.00575268e+00
-9.40246582e-01 -1.92762613e-01 6.82577193e-01 5.07036269e-01
-2.84611076e-01 -9.97912526e-01 -3.86414230e-01 -8.45649764e-02
-5.12755394e-01 -3.42246108e-02 -1.35222113e+00 6.02921188e-01
-4.16391313e-01 1.31176054e+00 -3.53425175e-01 -3.09763610e-01
3.45310688e-01 3.72557305e-02 4.51221257e-01 -4.81536329e-01
-1.33750474e+00 -3.20972919e-01 6.69912040e-01 -4.94056433e-01
-5.07607579e-01 -6.14686668e-01 -1.53253651e+00 -6.63464904e-01
2.34321523e-02 5.69109380e-01 4.26535279e-01 7.30405509e-01
2.30323866e-01 -2.70331334e-02 8.19640756e-01 -8.08920324e-01
-7.97326088e-01 -1.08814836e+00 -2.43605465e-01 1.09767652e+00
-6.52561039e-02 -4.81373101e-01 -6.35037720e-01 -1.67119190e-01]
|
[13.172581672668457, 7.6189985275268555]
|
dbea21e5-5b94-49de-9d85-80fc114874b3
|
on-the-efficacy-of-3d-point-cloud
|
2306.06799
| null |
https://arxiv.org/abs/2306.06799v1
|
https://arxiv.org/pdf/2306.06799v1.pdf
|
On the Efficacy of 3D Point Cloud Reinforcement Learning
|
Recent studies on visual reinforcement learning (visual RL) have explored the use of 3D visual representations. However, none of these work has systematically compared the efficacy of 3D representations with 2D representations across different tasks, nor have they analyzed 3D representations from the perspective of agent-object / object-object relationship reasoning. In this work, we seek answers to the question of when and how do 3D neural networks that learn features in the 3D-native space provide a beneficial inductive bias for visual RL. We specifically focus on 3D point clouds, one of the most common forms of 3D representations. We systematically investigate design choices for 3D point cloud RL, leading to the development of a robust algorithm for various robotic manipulation and control tasks. Furthermore, through comparisons between 2D image vs 3D point cloud RL methods on both minimalist synthetic tasks and complex robotic manipulation tasks, we find that 3D point cloud RL can significantly outperform the 2D counterpart when agent-object / object-object relationship encoding is a key factor.
|
['Hao Su', 'Xuanlin Li', 'Yunchao Yao', 'Zhan Ling']
|
2023-06-11
| null | null | null | null |
['3d-point-cloud-reinforcement-learning', 'robot-manipulation']
|
['computer-vision', 'robots']
|
[-1.92747712e-01 1.54929131e-01 -5.11329532e-01 2.41069347e-02
-2.35393882e-01 -6.80615425e-01 9.75755155e-01 1.93243876e-01
-3.09448183e-01 2.99786001e-01 1.00972466e-01 -4.20194417e-01
-3.65424722e-01 -5.13919175e-01 -9.29569244e-01 -5.22333145e-01
-2.09018514e-01 6.57647073e-01 -1.61747009e-01 -4.17142779e-01
6.31038368e-01 1.00441098e+00 -1.56724787e+00 2.35525612e-02
2.86095440e-01 8.31719518e-01 3.29797268e-01 4.85445559e-01
-7.52413943e-02 8.16268921e-01 -7.04892337e-01 3.60525578e-01
6.72951996e-01 -2.47884840e-01 -6.47963107e-01 1.98829755e-01
4.20161277e-01 -3.77106965e-01 -3.43725711e-01 7.14736998e-01
5.33718944e-01 2.49281690e-01 9.74553764e-01 -1.74075460e+00
-1.05952954e+00 3.18325549e-01 -7.05010653e-01 2.08874464e-01
5.01082480e-01 7.71281540e-01 9.16329980e-01 -7.50635564e-01
8.87778819e-01 1.73228467e+00 3.91796619e-01 5.48462987e-01
-1.33139110e+00 -3.50791305e-01 3.46664041e-01 -6.63019940e-02
-8.11766863e-01 -9.36370566e-02 8.90634418e-01 -7.63087511e-01
1.32137215e+00 -1.99527100e-01 9.41489637e-01 1.34016907e+00
2.44866505e-01 9.67650473e-01 1.42035675e+00 -4.54464525e-01
1.13278367e-01 -3.78703400e-02 -5.50753288e-02 7.05381334e-01
2.89177150e-01 8.24479699e-01 -5.49456000e-01 -1.27403021e-01
1.31917036e+00 -1.18481994e-01 -1.16164997e-01 -1.24747598e+00
-1.32958138e+00 9.51317310e-01 1.02807975e+00 2.09177330e-01
-4.67605919e-01 7.31538534e-01 2.39306092e-01 4.92485225e-01
1.81205675e-01 1.08061218e+00 -3.43892306e-01 2.23420635e-01
-1.92773476e-01 7.39229321e-01 4.37111586e-01 1.17100573e+00
5.84859908e-01 5.56913912e-01 -1.56288326e-01 4.05177861e-01
4.84488606e-01 4.43779945e-01 1.69101804e-01 -1.37925470e+00
3.49281490e-01 6.01451516e-01 1.48446053e-01 -9.67258751e-01
-5.86523592e-01 -2.28627399e-01 -2.88123965e-01 1.20073271e+00
3.36503148e-01 1.11669786e-01 -9.25352693e-01 1.72206903e+00
1.98266685e-01 -4.64538187e-01 3.72209489e-01 1.23121810e+00
9.88836765e-01 4.44184840e-01 2.72256762e-01 2.32727200e-01
9.47108686e-01 -8.27803195e-01 -3.77479881e-01 -3.50769341e-01
6.44421875e-01 -2.57720321e-01 1.18177843e+00 -9.71505642e-02
-1.25390947e+00 -6.39281631e-01 -1.12592292e+00 -3.58222038e-01
-6.22787416e-01 8.54612440e-02 8.38276267e-01 7.05066770e-02
-9.83510077e-01 4.20905948e-01 -4.93412405e-01 -5.10075450e-01
4.76162791e-01 4.01281863e-01 -5.12420595e-01 4.09798287e-02
-7.40619421e-01 1.76040339e+00 3.10084075e-01 -1.88171640e-01
-1.42321491e+00 -5.93929291e-01 -1.10012662e+00 -2.28030443e-01
2.62282014e-01 -9.82880175e-01 1.36189783e+00 -8.69760513e-01
-1.22459674e+00 1.12619758e+00 2.07870454e-01 -4.21108186e-01
2.82010943e-01 -1.23685338e-01 5.06674051e-01 2.41593540e-01
1.52549952e-01 1.20773578e+00 1.03603435e+00 -1.87771165e+00
-1.19562641e-01 -5.30794263e-01 5.14404535e-01 5.79002559e-01
5.79787135e-01 -4.89867479e-01 3.17741781e-01 -6.11007154e-01
7.25593120e-02 -1.04528093e+00 -1.52663186e-01 5.62925100e-01
6.75453469e-02 -6.30877256e-01 8.50843072e-01 -1.43997833e-01
7.00967386e-02 -2.23212099e+00 5.65451682e-01 -1.22359596e-01
3.36370736e-01 1.03207722e-01 -2.12857574e-01 5.95382631e-01
-2.53155529e-01 1.70997545e-01 7.87804425e-02 1.17861390e-01
2.89610535e-01 3.74412805e-01 -4.84605193e-01 5.36716640e-01
7.11813807e-01 1.23800719e+00 -1.08855426e+00 -2.97747225e-01
4.14451331e-01 4.52258706e-01 -6.12693369e-01 2.30762079e-01
-5.31146824e-01 5.58065832e-01 -7.10779667e-01 5.49857914e-01
2.71562785e-01 -2.12694541e-01 -1.78881466e-01 1.03796966e-01
-2.78422236e-01 2.68848598e-01 -3.69124651e-01 1.87180781e+00
-3.13585937e-01 7.70578682e-01 -7.44601041e-02 -9.11363244e-01
1.09132385e+00 2.46990517e-01 3.60439032e-01 -8.63045812e-01
1.04731828e-01 9.33018848e-02 1.73182979e-01 -6.76683903e-01
3.49663764e-01 -2.72776216e-01 -7.66827315e-02 3.63506913e-01
2.52547145e-01 -1.03276742e+00 -1.58448726e-01 3.56125273e-03
8.48756850e-01 9.02628064e-01 4.36219215e-01 -1.25977099e-01
-1.87164977e-01 6.08139932e-01 -8.39178562e-02 9.57813323e-01
-3.88651073e-01 4.81735826e-01 7.05442011e-01 -5.66655159e-01
-1.23892403e+00 -1.32045078e+00 1.25534311e-01 8.11971903e-01
3.83812428e-01 -2.82033179e-02 -2.99335923e-02 -6.08143330e-01
5.66833138e-01 9.92519557e-01 -9.21386003e-01 -3.64911169e-01
-4.57500398e-01 -6.51456416e-02 3.18903506e-01 5.65567970e-01
1.28059104e-01 -1.47503376e+00 -1.44038427e+00 -1.62308797e-01
3.53700072e-01 -7.07630217e-01 1.93869457e-01 4.64661300e-01
-1.02420747e+00 -1.16842556e+00 -7.47847497e-01 -7.42199838e-01
5.98562717e-01 5.86797774e-01 1.23257077e+00 7.92621821e-02
-2.12015182e-01 1.09993768e+00 -5.13043940e-01 -6.77542269e-01
-4.50805217e-01 -2.99407780e-01 7.25369751e-02 -8.49318385e-01
1.52894631e-01 -4.65966403e-01 -3.13540488e-01 1.46228157e-03
-5.98549128e-01 -4.19943873e-03 6.57990038e-01 6.51996791e-01
2.85613745e-01 -5.25267005e-01 6.23493075e-01 -3.26185495e-01
8.45038235e-01 -4.78491902e-01 -5.40466547e-01 -9.14381351e-03
-3.55704665e-01 2.82031149e-01 2.28263903e-02 -5.85974157e-01
-6.20196342e-01 1.09028906e-01 3.28924358e-01 -9.08718228e-01
-1.85510337e-01 2.26617694e-01 4.24639165e-01 -6.27429187e-02
9.71245527e-01 8.81204009e-03 4.06354100e-01 -1.38811529e-01
7.24506021e-01 7.99304992e-02 9.49211121e-02 -8.02120805e-01
7.48336017e-01 3.91630143e-01 2.37475887e-01 -7.15129375e-01
-3.84700358e-01 -3.82258371e-02 -6.91689730e-01 -3.54297400e-01
1.04923129e+00 -8.94533396e-01 -1.03744686e+00 -5.20391995e-03
-1.20754683e+00 -7.01093614e-01 -6.18896186e-01 5.00812531e-01
-1.47049844e+00 1.31333312e-02 -3.35825801e-01 -7.96767652e-01
1.00142293e-01 -1.38553047e+00 1.19264913e+00 -5.57726761e-03
-4.84918684e-01 -6.87542140e-01 -1.76387370e-01 -9.72102433e-02
2.54264474e-01 5.42348862e-01 1.53150845e+00 -2.66495258e-01
-7.34828949e-01 9.71856192e-02 -3.39802325e-01 -2.32784376e-01
-1.52762868e-02 -2.32311249e-01 -8.06805015e-01 -3.00942987e-01
-1.24609068e-01 -9.07692373e-01 7.08630443e-01 4.83695477e-01
8.25102508e-01 8.95582885e-02 -2.38624379e-01 3.41451317e-01
1.14438438e+00 3.22339177e-01 2.48581737e-01 4.52356875e-01
5.35926223e-01 8.37174058e-01 8.10148776e-01 2.45377228e-01
3.55973154e-01 6.63165212e-01 1.03190637e+00 1.26049332e-02
-3.38311523e-01 -4.46593106e-01 2.48601630e-01 6.42341003e-02
-7.07919002e-02 -1.75838873e-01 -8.35255623e-01 4.17568445e-01
-1.81626379e+00 -7.41606772e-01 2.46929482e-01 1.75701129e+00
4.68518049e-01 8.21488127e-02 1.69654682e-01 -9.32335407e-02
3.69813085e-01 3.86690348e-01 -8.82717311e-01 -4.45841879e-01
-3.80165502e-03 -6.39877766e-02 2.72311568e-01 1.53389961e-01
-8.04353714e-01 1.05288327e+00 6.65191746e+00 1.01758994e-01
-9.83368814e-01 -2.65061766e-01 -2.24073287e-02 -1.78070500e-01
-1.20675229e-01 3.73114571e-02 -3.33583057e-01 -1.94523588e-01
1.22523174e-01 -2.08380795e-03 4.77790147e-01 1.09915352e+00
8.39419961e-02 -1.54159710e-01 -1.64959633e+00 1.10716283e+00
1.86209902e-02 -1.38309574e+00 2.68822432e-01 1.21128447e-01
4.32337850e-01 -1.22413889e-01 3.86556536e-01 5.82957566e-01
5.82669318e-01 -1.32233143e+00 1.16367614e+00 4.23983753e-01
5.20604908e-01 -3.31128925e-01 4.51666303e-02 4.36804801e-01
-6.58312619e-01 -2.48325676e-01 -2.39067569e-01 -1.31009221e-01
-1.13752924e-01 -3.94639999e-01 -1.00951517e+00 2.70900607e-01
9.34682190e-01 8.75149608e-01 -2.55215943e-01 8.04464996e-01
-2.34899417e-01 -1.46720544e-01 -4.18181792e-02 -2.57895350e-01
6.09789133e-01 7.62484297e-02 7.45684743e-01 7.05210686e-01
2.55795214e-02 -8.51845294e-02 2.19525531e-01 1.27293229e+00
2.73294061e-01 -4.21912760e-01 -1.49032414e+00 -2.71373361e-01
3.61324281e-01 6.61238372e-01 -5.31011879e-01 -4.99438196e-02
-1.09738044e-01 4.06563878e-01 7.24606097e-01 5.38991570e-01
-5.08132696e-01 3.44289057e-02 8.09642732e-01 5.64183034e-02
4.45790380e-01 -7.93098092e-01 -1.50092855e-01 -5.09165108e-01
-2.40047187e-01 -9.95923936e-01 5.03526591e-02 -1.60548711e+00
-1.35181534e+00 2.05481395e-01 3.85606766e-01 -1.24602020e+00
-4.95248020e-01 -1.07773077e+00 -2.58468181e-01 7.19883800e-01
-1.46745420e+00 -1.09923458e+00 -1.48241401e-01 5.95053256e-01
5.10863006e-01 -2.02944875e-01 8.62325072e-01 -5.72165370e-01
2.32902065e-01 -7.52840713e-02 -4.08785522e-01 -7.31055364e-02
4.62611258e-01 -1.15146935e+00 3.48555505e-01 9.69134364e-03
1.80272341e-01 6.31819963e-01 5.14996529e-01 -5.49690366e-01
-1.85256910e+00 -5.53895772e-01 1.98439006e-02 -1.00351644e+00
2.38310605e-01 -2.79120326e-01 -5.61172843e-01 1.01523900e+00
3.10075939e-01 -1.96071900e-02 5.46078458e-02 -8.66669714e-02
-6.20358706e-01 4.29066449e-01 -1.21192014e+00 8.91339004e-01
1.28946400e+00 -5.26760519e-01 -1.02910030e+00 2.62928128e-01
9.67490315e-01 -5.47913432e-01 -8.16768348e-01 5.84543288e-01
3.98720801e-01 -8.21471453e-01 1.33147490e+00 -9.66838956e-01
6.85124874e-01 -4.24511909e-01 -1.89378396e-01 -1.69885468e+00
-4.44447666e-01 -1.93164989e-01 -1.25964299e-01 4.67682958e-01
1.53278828e-01 -4.73911047e-01 4.58203465e-01 -2.94206105e-02
-2.86622703e-01 -4.96719182e-01 -7.31945157e-01 -6.13410592e-01
3.52764428e-01 -1.12007417e-01 2.64231056e-01 8.92888129e-01
-1.50580734e-01 2.60132432e-01 4.59999181e-02 4.22333032e-02
4.41920519e-01 4.14684594e-01 9.98637199e-01 -1.09377217e+00
-1.17937364e-02 -7.52244115e-01 -5.18418908e-01 -1.06661117e+00
5.48844934e-01 -1.19406509e+00 1.05223931e-01 -1.81488621e+00
-1.87514201e-01 -6.16458118e-01 3.07817280e-01 5.38397312e-01
3.38660508e-01 -1.94875583e-01 7.29138196e-01 2.64381260e-01
-1.87148497e-01 8.72622550e-01 1.89581966e+00 -3.62872034e-01
-8.89888108e-02 -3.11735392e-01 -5.91171801e-01 6.00363672e-01
8.60204577e-01 -2.49961913e-01 -6.49784565e-01 -8.48494828e-01
1.04419127e-01 2.40220606e-01 9.32683945e-01 -7.10699797e-01
-1.76714763e-01 -3.32585722e-01 7.71268547e-01 -5.76210737e-01
8.37763667e-01 -1.01075506e+00 -3.19720268e-01 5.35344601e-01
-7.17767656e-01 4.13056374e-01 3.65738571e-01 7.25595117e-01
1.84970677e-01 -2.58604616e-01 6.28859818e-01 -6.84282064e-01
-8.76650751e-01 1.25695124e-01 -5.46428025e-01 3.85946184e-02
1.15725911e+00 -4.06249285e-01 -4.34138983e-01 -3.42460841e-01
-6.48075223e-01 2.66804099e-01 6.59150898e-01 8.78547907e-01
8.26215386e-01 -1.36036503e+00 -5.08112192e-01 1.65613219e-01
2.88969874e-01 1.46576256e-01 -1.81267709e-01 3.80153418e-01
-5.12954772e-01 4.21143234e-01 -7.56144941e-01 -9.55268025e-01
-8.21531773e-01 9.94009197e-01 4.92201805e-01 2.92608857e-01
-8.59026909e-01 6.50784314e-01 4.63549376e-01 -7.78022885e-01
3.90722394e-01 -5.26352346e-01 -1.34663507e-01 -2.60060996e-01
-6.51485100e-02 -5.58838286e-02 -4.66178060e-01 -6.76220298e-01
-1.69733673e-01 7.94807196e-01 2.39110336e-01 -1.42551541e-01
1.33125675e+00 2.82270730e-01 2.59476423e-01 7.38980889e-01
8.67645681e-01 -5.79976737e-01 -1.68483734e+00 -6.28422573e-03
-7.61291459e-02 -4.70903784e-01 -1.60889149e-01 -6.26219630e-01
-9.10128355e-01 1.08236182e+00 5.15850484e-01 1.61668509e-01
5.07418334e-01 3.59915197e-01 -1.41497985e-01 6.05138302e-01
6.10125959e-01 -6.54947281e-01 6.59023762e-01 6.26879156e-01
1.68057251e+00 -1.34497070e+00 2.01488659e-01 -3.22355516e-02
-9.60112333e-01 9.17114913e-01 8.77915204e-01 -6.31162047e-01
4.14850265e-01 -8.92083198e-02 3.11146658e-02 -6.48157716e-01
-7.36187339e-01 -4.94283587e-01 8.02258030e-02 1.16737616e+00
1.32322282e-01 -7.85071701e-02 2.69833595e-01 -1.88119352e-01
-2.71253645e-01 -2.20956221e-01 2.63135433e-01 1.40286970e+00
-3.68926048e-01 -5.74062467e-01 -2.53703684e-01 6.00499325e-02
2.07576156e-01 2.46735245e-01 -6.91347241e-01 1.41668987e+00
-1.43846929e-01 7.13388085e-01 2.43468434e-01 -1.25348046e-01
4.56023961e-01 6.43050745e-02 1.29508793e+00 -7.07249999e-01
-7.97929108e-01 -3.34808081e-01 -1.86465725e-01 -4.23857212e-01
-6.85177863e-01 -5.38375258e-01 -1.55667543e+00 2.22768281e-02
-2.00953782e-02 -2.17921317e-01 9.12106097e-01 6.75238729e-01
4.21461344e-01 4.33457881e-01 2.31744021e-01 -1.56390703e+00
-6.97072685e-01 -7.50718474e-01 -3.89633536e-01 5.14118493e-01
6.36395991e-01 -1.48304498e+00 -2.19420955e-01 -3.25574070e-01]
|
[4.685592174530029, 0.6537683606147766]
|
b5d0b0c8-fda5-495a-bfcd-53435370c20d
|
phagocytosis-unveiled-a-scalable-and
|
2304.13764
| null |
https://arxiv.org/abs/2304.13764v1
|
https://arxiv.org/pdf/2304.13764v1.pdf
|
Phagocytosis Unveiled: A Scalable and Interpretable Deep learning Framework for Neurodegenerative Disease Analysis
|
Quantifying the phagocytosis of dynamic, unstained cells is essential for evaluating neurodegenerative diseases. However, measuring rapid cell interactions and distinguishing cells from backgrounds make this task challenging when processing time-lapse phase-contrast video microscopy. In this study, we introduce a fully automated, scalable, and versatile realtime framework for quantifying and analyzing phagocytic activity. Our proposed pipeline can process large data-sets and includes a data quality verification module to counteract potential perturbations such as microscope movements and frame blurring. We also propose an explainable cell segmentation module to improve the interpretability of deep learning methods compared to black-box algorithms. This includes two interpretable deep learning capabilities: visual explanation and model simplification. We demonstrate that interpretability in deep learning is not the opposite of high performance, but rather provides essential deep learning algorithm optimization insights and solutions. Incorporating interpretable modules results in an efficient architecture design and optimized execution time. We apply this pipeline to quantify and analyze microglial cell phagocytosis in frontotemporal dementia (FTD) and obtain statistically reliable results showing that FTD mutant cells are larger and more aggressive than control cells. To stimulate translational approaches and future research, we release an open-source pipeline and a unique microglial cells phagocytosis dataset for immune system characterization in neurodegenerative diseases research. This pipeline and dataset will consistently crystallize future advances in this field, promoting the development of efficient and effective interpretable algorithms dedicated to this critical domain. https://github.com/ounissimehdi/PhagoStat
|
['Daniel Racoceanu', 'Morwena Latouche', 'Mehdi Ounissi']
|
2023-04-26
| null | null | null | null |
['cell-segmentation']
|
['medical']
|
[ 7.66023546e-02 -6.17777586e-01 3.30747336e-01 -2.47769803e-01
-4.85588759e-01 -4.23708826e-01 2.54095107e-01 1.67755559e-01
-8.71999085e-01 9.02487695e-01 4.06885855e-02 -1.44188926e-01
1.30268604e-01 -3.75843465e-01 -5.31845093e-01 -9.86103833e-01
-1.74141124e-01 9.87655103e-01 3.00534457e-01 2.84214377e-01
2.20412359e-01 8.04462731e-01 -1.49216938e+00 7.98730195e-01
7.10934460e-01 6.42743587e-01 4.79359180e-01 1.06171381e+00
1.93029165e-01 6.49144530e-01 -3.91513526e-01 3.36181551e-01
-1.96021553e-02 4.76313159e-02 -6.84356868e-01 7.13766217e-02
1.70828685e-01 -6.29030108e-01 1.26368970e-01 6.80466950e-01
8.19214880e-01 -2.93783218e-01 6.33039296e-01 -1.15320313e+00
-5.29046237e-01 -1.45603672e-01 -3.49210471e-01 8.96006167e-01
-1.83430210e-01 8.62118840e-01 4.72626626e-01 -6.42989457e-01
7.21858263e-01 1.28915644e+00 6.06253743e-01 6.76995575e-01
-1.47131145e+00 -1.55284390e-01 -1.42834201e-01 6.10087395e-01
-8.02177429e-01 -6.10521495e-01 3.11644059e-02 -9.18635726e-01
1.16381025e+00 2.17714667e-01 1.10316086e+00 1.24424720e+00
3.92821103e-01 7.09532678e-01 1.52633405e+00 5.02872504e-02
5.34642875e-01 -6.92246497e-01 3.87203425e-01 7.61121988e-01
6.20686352e-01 -1.58034831e-01 -1.36633441e-01 -1.38104334e-01
9.45154190e-01 4.36680347e-01 -3.43313783e-01 -2.46703420e-02
-1.48661578e+00 5.25621235e-01 9.41285715e-02 1.27377808e-01
-3.40950698e-01 5.36452055e-01 4.71777141e-01 -7.41710439e-02
2.68581241e-01 2.19804600e-01 -6.27326310e-01 -1.26201719e-01
-7.62385607e-01 5.13249516e-01 3.21400762e-01 1.67635828e-01
3.95881236e-01 -6.00972436e-02 -1.88694283e-01 5.94111025e-01
1.78597599e-01 7.62138069e-01 3.60837489e-01 -1.55011833e+00
-2.70455092e-01 8.17677796e-01 2.49035299e-01 -3.68241072e-01
-8.28766525e-01 -1.63317055e-01 -6.98504746e-01 9.59692538e-01
8.73541832e-01 6.91006286e-03 -7.53349781e-01 1.33101141e+00
4.13129598e-01 -1.13100886e-01 -4.40884620e-01 1.19139743e+00
3.61329347e-01 -9.01460648e-02 1.28586516e-01 -1.41952485e-01
1.87170005e+00 -7.10401952e-01 -4.50277895e-01 6.41028106e-04
8.12314868e-01 -5.82187772e-01 1.35693336e+00 4.14359003e-01
-1.02798820e+00 -3.12144253e-02 -7.40949869e-01 -6.32394195e-01
-4.12419528e-01 2.15718418e-01 5.54625332e-01 2.32066497e-01
-1.26295912e+00 4.67140734e-01 -1.49257076e+00 -5.66175401e-01
1.27479780e+00 4.96844202e-01 -2.69046903e-01 -2.87104342e-02
-4.15536821e-01 7.08517671e-01 -9.20799524e-02 -3.72233763e-02
-9.54757035e-01 -8.93392205e-01 -9.03063565e-02 -1.35810852e-01
-2.69187152e-01 -1.43513834e+00 1.09931278e+00 -4.39965189e-01
-1.07055664e+00 1.18864048e+00 -6.28612339e-01 -6.82598829e-01
7.16589391e-01 -3.45047176e-01 3.00353765e-01 6.91511512e-01
1.00616150e-01 8.78870785e-01 5.78690648e-01 -9.62398529e-01
-6.58658147e-01 -6.82599068e-01 -1.70754761e-01 -2.61517137e-01
1.10426456e-01 1.60520241e-01 8.99809301e-02 -3.33218575e-01
-3.28354925e-01 -8.37377012e-01 -3.28230076e-02 7.69937575e-01
-1.75107047e-01 1.61722720e-01 1.08622456e+00 -6.63285434e-01
3.86401862e-01 -1.63733232e+00 -1.48820534e-01 -4.59238529e-01
8.99900854e-01 4.40080911e-01 9.55872089e-02 -1.10520467e-01
2.12110355e-02 2.83429652e-01 -1.76225439e-01 -4.33582127e-01
-1.19361967e-01 1.02700636e-01 -3.80721502e-02 8.88107836e-01
4.21776056e-01 1.23548126e+00 -7.84028411e-01 -4.42653745e-01
2.93761998e-01 9.72881258e-01 -6.31097317e-01 -1.72791690e-01
-4.81895924e-01 1.03312433e+00 -2.41008922e-01 8.45014036e-01
5.05950093e-01 -6.99550211e-01 1.70833096e-02 -4.18059051e-01
-2.62722254e-01 -7.82511085e-02 -6.02534413e-01 9.02507901e-01
-5.53787649e-02 8.55234683e-01 2.75897741e-01 -7.10787654e-01
3.15622300e-01 -2.01688670e-02 6.00448966e-01 -6.93733633e-01
5.80624640e-01 2.39208505e-01 4.65921201e-02 -6.93208158e-01
-3.38640511e-01 -2.84791198e-02 7.55737782e-01 7.06958234e-01
-5.32826662e-01 4.48452145e-01 5.04970312e-01 -1.96791459e-02
1.38613629e+00 -1.18188225e-02 2.04509795e-01 -3.88232470e-01
4.30436134e-01 2.19919086e-01 3.52319658e-01 5.60267091e-01
-7.52474368e-01 6.80349827e-01 7.06422985e-01 -8.38930666e-01
-1.52814162e+00 -1.13540983e+00 -2.48266265e-01 8.15143347e-01
-2.04300135e-01 1.97825328e-01 -8.11108530e-01 -2.18013480e-01
1.09876499e-01 3.21054049e-02 -7.25302398e-01 3.99911672e-01
-6.53699696e-01 -1.16612887e+00 5.57472706e-01 5.26098013e-01
5.10862947e-01 -8.07045162e-01 -1.06231391e+00 8.42472091e-02
-3.08978528e-01 -1.09857666e+00 -4.40883860e-02 -4.94940840e-02
-9.27237749e-01 -1.63730669e+00 -6.57172918e-01 -6.20829761e-01
7.68890977e-01 6.29718125e-01 7.47479141e-01 5.45576453e-01
-8.86545300e-01 2.32250348e-01 -2.40995243e-01 -4.48338360e-01
-1.86173439e-01 -4.47653383e-01 1.54560238e-01 -3.71357977e-01
4.95266140e-01 -6.76331043e-01 -1.27935171e+00 2.31504664e-01
-8.89303088e-01 1.99075535e-01 4.97258842e-01 7.34146595e-01
1.32337785e+00 -2.06083849e-01 1.22454979e-01 -4.86407787e-01
4.81988370e-01 -2.04572544e-01 -5.21953940e-01 -1.61474094e-01
-3.49137247e-01 6.69120923e-02 7.05605865e-01 -3.07468086e-01
-8.21511149e-01 -1.84596647e-02 -1.92311425e-02 -2.14414135e-01
-3.85585606e-01 -1.18873850e-01 1.16350256e-01 -4.69225198e-02
6.11486554e-01 8.53994861e-02 7.19077408e-01 -2.64202088e-01
9.56626683e-02 5.30744553e-01 4.00378644e-01 -3.68433625e-01
1.57139823e-01 1.66231155e+00 1.76483184e-01 -9.83072340e-01
-5.69478512e-01 -6.43384159e-01 -4.48057383e-01 -2.67907083e-01
1.19514990e+00 -7.62459338e-01 -1.17311585e+00 8.88512313e-01
-1.28314829e+00 -9.10923302e-01 -8.83531868e-02 4.77522045e-01
-6.92348719e-01 4.91446704e-01 -8.59098136e-01 -6.11259460e-01
-7.52000034e-01 -1.50775981e+00 1.40820372e+00 -2.03087870e-02
-4.47112739e-01 -1.03042471e+00 2.58383095e-01 1.00340581e+00
2.84967214e-01 3.77335846e-01 9.52252388e-01 -5.46366751e-01
-8.60411704e-01 3.46917272e-01 -6.38275266e-01 5.84958643e-02
-9.49349627e-02 2.76501954e-01 -9.53608871e-01 -2.69584894e-01
1.30259573e-01 -2.53301173e-01 9.63733852e-01 6.40855253e-01
1.04866445e+00 -2.77308464e-01 -4.01359081e-01 6.57782316e-01
1.22505164e+00 -2.17999756e-01 8.47839832e-01 7.82070220e-01
5.93387902e-01 5.45996010e-01 1.37058467e-01 4.37871695e-01
2.77325004e-01 4.85588312e-01 5.50030708e-01 -2.18063205e-01
-4.34211701e-01 8.94987106e-01 4.07710969e-01 4.59435284e-01
-3.09785068e-01 -1.93622544e-01 -1.05229509e+00 4.79328632e-01
-1.74064207e+00 -1.20977068e+00 -8.75839889e-01 1.75482535e+00
6.46861434e-01 -1.26710251e-01 3.27841520e-01 5.29683791e-02
6.74101949e-01 -3.70245159e-01 -8.64611149e-01 3.86271402e-02
-4.98573393e-01 2.53276825e-01 2.71227270e-01 5.84972501e-01
-8.08562160e-01 6.91907406e-01 6.02023220e+00 3.51256520e-01
-1.20063424e+00 4.49567944e-01 6.86667085e-01 -7.14768946e-01
-5.29197082e-02 -2.89512813e-01 -5.98575830e-01 6.23356283e-01
6.33174956e-01 3.87877285e-01 6.32854223e-01 4.53553647e-01
1.10135841e+00 -2.89717913e-01 -1.00105345e+00 9.25796330e-01
-4.06221181e-01 -1.64911318e+00 9.94949117e-02 4.46609557e-01
2.85475969e-01 6.12405062e-01 3.38804685e-02 -4.08631921e-01
2.07539916e-01 -9.86512423e-01 5.24369776e-01 8.86821389e-01
4.62384105e-01 -2.70577908e-01 8.54542434e-01 -5.99620380e-02
-7.96152830e-01 -5.56403436e-02 -3.62438083e-01 -1.69826195e-01
2.62191951e-01 9.36658382e-01 -1.04690790e+00 -3.45409930e-01
9.14320111e-01 5.02904296e-01 -6.96460545e-01 1.19467235e+00
3.69033739e-02 2.75520891e-01 -2.12416515e-01 6.61743060e-02
-2.03852922e-01 -1.46212667e-01 6.74439609e-01 1.25615931e+00
1.41163483e-01 -8.28374699e-02 -2.29864165e-01 1.16831660e+00
3.01042020e-01 -1.73159570e-01 -1.34248594e-02 -3.34214717e-01
2.88055092e-01 1.44271791e+00 -1.19774187e+00 -2.39114240e-01
-1.82335079e-01 8.32041979e-01 5.84688902e-01 3.06775182e-01
-8.35472465e-01 3.71568739e-01 1.04671347e+00 5.02500594e-01
2.35795528e-01 -6.03666127e-01 -7.34014928e-01 -9.92801666e-01
-1.04595624e-01 -6.84224367e-01 2.84680188e-01 -1.07228410e+00
-1.26963603e+00 8.92775878e-02 -4.43970323e-01 -7.69288123e-01
3.96316618e-01 -1.15032732e+00 -5.00524461e-01 4.15990561e-01
-1.45304811e+00 -1.14267778e+00 -5.96224546e-01 4.41355139e-01
3.48107368e-01 -3.02377641e-02 5.16287148e-01 1.64184600e-01
-8.16058576e-01 -1.28128558e-01 2.88564742e-01 -1.59776717e-01
5.79625726e-01 -1.25959337e+00 1.68526471e-01 6.86133981e-01
-4.68826145e-01 8.62703741e-01 8.31729233e-01 -7.19296396e-01
-1.12962675e+00 -1.32460880e+00 4.51779008e-01 -7.84432292e-01
9.31772768e-01 -1.13388225e-01 -8.98380518e-01 4.65247124e-01
-2.32073769e-01 3.05103771e-02 9.43973660e-01 -4.73989010e-01
-3.94820087e-02 -5.11768684e-02 -1.19554663e+00 8.24336767e-01
9.65309381e-01 -4.13596600e-01 -4.50202376e-01 7.77680516e-01
3.65464956e-01 -2.41255928e-02 -8.22897732e-01 8.92277807e-02
8.02983880e-01 -1.35020876e+00 1.11538363e+00 -5.53399861e-01
2.82529205e-01 -6.08097494e-01 7.80907571e-02 -5.66521704e-01
-2.47488394e-01 -2.76223451e-01 -3.38118643e-01 5.76224744e-01
-9.00418088e-02 -7.95403421e-01 9.96895134e-01 3.91344190e-01
-3.28492910e-01 -9.88018990e-01 -9.31943357e-01 -5.47665119e-01
8.87582004e-02 -2.46942148e-01 2.32224822e-01 4.56967503e-01
-3.59380454e-01 -8.16352218e-02 3.81251723e-01 1.70830145e-01
8.95802736e-01 -1.75047725e-01 6.49918199e-01 -1.22517335e+00
-1.68773085e-01 -5.57861805e-01 -5.68710804e-01 -3.43277305e-01
-1.08727328e-02 -8.00182760e-01 -4.13972080e-01 -1.82777953e+00
5.54352939e-01 -1.03727572e-01 -1.24487497e-01 4.23926920e-01
-7.71504864e-02 4.69939291e-01 -1.70020148e-01 7.60948479e-01
-8.24844539e-01 2.36910120e-01 1.51243389e+00 -2.31656119e-01
-2.27833409e-02 -4.56746936e-01 -5.68661392e-01 9.06609237e-01
1.20834625e+00 -3.51768970e-01 -2.58934647e-01 -7.25875914e-01
8.61210600e-02 -7.46044576e-01 1.37926841e+00 -1.26044154e+00
1.05817676e-01 -1.68771576e-02 6.90366149e-01 -5.10965586e-01
2.93830574e-01 -4.26952600e-01 2.03116208e-01 8.82867873e-01
8.02833289e-02 -1.15404725e-02 1.33009646e-02 5.16815186e-01
4.85605985e-01 2.31501862e-01 1.20769608e+00 -4.73508716e-01
-4.53172445e-01 2.91926622e-01 -1.05357933e+00 1.69688597e-01
9.62323368e-01 -6.26669884e-01 -1.24130368e+00 2.73841769e-01
-8.82787704e-01 1.21492177e-01 1.03906870e+00 -2.62072384e-01
5.66738844e-01 -8.46291006e-01 -6.37983978e-01 -3.75983976e-02
-6.48827255e-02 -3.81699800e-02 4.98652697e-01 1.54745531e+00
-1.31072736e+00 4.28127766e-01 -8.26372504e-01 -9.73357737e-01
-1.44511330e+00 2.25858331e-01 5.28731167e-01 -1.70324430e-01
-5.74325383e-01 5.38770795e-01 1.45203516e-01 8.09963122e-02
-2.87103087e-01 -5.70295453e-01 -1.06605619e-01 -7.43666887e-02
9.14638221e-01 9.36191380e-01 1.01909362e-01 -4.03398782e-01
-3.46021920e-01 1.77042693e-01 -2.82491237e-01 8.95129144e-02
1.45474076e+00 -4.56982493e-01 -5.91644168e-01 2.35554263e-01
9.18203533e-01 -2.87377954e-01 -1.55050850e+00 4.94235098e-01
-3.25830430e-01 -1.18037120e-01 -3.03933863e-02 -6.86293006e-01
-9.50882733e-01 9.68841732e-01 1.11240673e+00 7.95867890e-02
7.09368885e-01 2.66611520e-02 1.12216520e+00 4.83329147e-01
3.95480961e-01 -1.00968552e+00 2.31181204e-01 3.93067420e-01
7.53222167e-01 -1.00732541e+00 2.31182918e-01 -2.24721044e-01
-1.01571448e-01 1.07282245e+00 5.55738270e-01 -8.36331248e-02
1.95942387e-01 4.84645426e-01 2.02421576e-01 -5.27608633e-01
-9.14922655e-01 -4.04309809e-01 -2.82075077e-01 1.11361909e+00
2.27852196e-01 -2.10025266e-01 -3.77350062e-01 4.90179241e-01
3.80146243e-02 2.47652665e-01 6.13359690e-01 8.41464281e-01
-8.39222491e-01 -9.90302563e-01 -4.03190792e-01 6.93063438e-01
-5.05683184e-01 -1.71273928e-02 -4.67326850e-01 7.78710783e-01
2.86001772e-01 6.51903510e-01 2.64222831e-01 2.12875187e-01
-1.47816256e-01 -9.40195397e-02 6.41903102e-01 -3.80811632e-01
-1.87267587e-01 2.01020524e-01 9.83082950e-02 -7.63681471e-01
-7.14956880e-01 -8.17615628e-01 -1.78292537e+00 -5.37928462e-01
2.79892474e-01 -5.68817019e-01 3.29437107e-01 1.24141073e+00
8.21134150e-01 5.16613841e-01 -2.37850532e-01 -1.09422445e+00
9.24371108e-02 -5.31249404e-01 -4.14158642e-01 4.81576502e-01
3.48988086e-01 -8.77367496e-01 -4.94376600e-01 7.29073882e-01]
|
[14.319708824157715, -3.0240187644958496]
|
41e1f77f-090e-412d-b5bb-53684b9a01fe
|
3d-shape-classification-using-collaborative
|
1711.04875
| null |
http://arxiv.org/abs/1711.04875v2
|
http://arxiv.org/pdf/1711.04875v2.pdf
|
3D Shape Classification Using Collaborative Representation based Projections
|
A novel 3D shape classification scheme, based on collaborative representation
learning, is investigated in this work. A data-driven feature-extraction
procedure, taking the form of a simple projection operator, is in the core of
our methodology. Provided a shape database, a graph encapsulating the
structural relationships among all the available shapes, is first constructed
and then employed in defining low-dimensional sparse projections. The recently
introduced method of CRPs (collaborative representation based projections),
which is based on L2-Graph, is the first variant that is included towards this
end. A second algorithm, that particularizes the CRPs to shape descriptors that
are inherently nonnegative, is also introduced as potential alternative. In
both cases, the weights in the graph reflecting the database structure are
calculated so as to approximate each shape as a sparse linear combination of
the remaining dataset objects. By way of solving a generalized eigenanalysis
problem, a linear matrix operator is designed that will act as the feature
extractor. Two popular, inherently high dimensional descriptors, namely
ShapeDNA and Global Point Signature (GPS), are employed in our experimentations
with SHREC10, SHREC11 and SCHREC 15 datasets, where shape recognition is cast
as a multi-class classification problem that is tackled by means of an SVM
(support vector machine) acting within the reduced dimensional space of the
crafted projections. The results are very promising and outperform state of the
art methods, providing evidence about the highly discriminative nature of the
introduced 3D shape representations.
|
['G. Economou', 'A. Papathanasiou', 'S. Oikonomou', 'F. Fotopoulou', 'S. Fotopoulos']
|
2017-11-13
| null | null | null | null |
['3d-shape-retrieval']
|
['computer-vision']
|
[ 1.81991041e-01 -2.94894222e-02 4.54162955e-02 -2.58731872e-01
-2.58082598e-01 -5.82447708e-01 9.18310106e-01 3.04102123e-01
-2.03563049e-01 3.53499174e-01 2.93960989e-01 -8.97872597e-02
-8.59290540e-01 -1.03853583e+00 -2.62181848e-01 -8.66211116e-01
-1.31939054e-01 8.73807967e-01 1.35265838e-03 -3.16451132e-01
3.81827384e-01 1.21650279e+00 -1.61095846e+00 2.17667088e-01
5.12048900e-01 9.10542488e-01 9.61810201e-02 2.97852963e-01
-4.02580410e-01 2.82597810e-01 -1.69349313e-01 -3.03427100e-01
3.73065561e-01 -7.33999684e-02 -3.78967226e-01 4.63720977e-01
1.12902306e-01 4.59839672e-01 1.03643999e-01 6.82036638e-01
3.89188647e-01 2.55790114e-01 1.18902194e+00 -8.48431706e-01
-3.78566176e-01 3.41326892e-01 -2.93441504e-01 -2.59291768e-01
6.28129005e-01 -2.00072080e-01 1.07615757e+00 -1.45770121e+00
9.18857098e-01 9.77944255e-01 5.56643546e-01 5.68357557e-02
-1.58824146e+00 -1.37997463e-01 -2.73075402e-01 5.42932265e-02
-1.55763662e+00 -2.56430626e-01 1.27984715e+00 -6.62939608e-01
5.93100250e-01 6.02961063e-01 8.29342544e-01 8.65788221e-01
-3.58454138e-02 5.03368258e-01 1.07772851e+00 -4.20993626e-01
5.68158984e-01 3.97972882e-01 1.05395585e-01 3.83058369e-01
2.99401820e-01 -1.13348246e-01 -1.54482096e-01 -5.90507984e-01
4.10714865e-01 1.77834511e-01 -1.56011224e-01 -1.28580213e+00
-5.95764875e-01 8.26697886e-01 4.05572981e-01 8.02212775e-01
-6.49510860e-01 -5.36530495e-01 2.60418862e-01 -8.21709633e-03
4.15041476e-01 5.42639419e-02 -1.05597945e-02 1.74659401e-01
-9.37271535e-01 2.50052899e-01 1.11416197e+00 6.63817167e-01
8.04943085e-01 1.78377360e-01 -3.52669135e-02 7.41398275e-01
5.78618228e-01 4.90942031e-01 4.69268382e-01 -2.92427801e-02
4.97055113e-01 1.36541843e+00 -1.21950604e-01 -1.54875576e+00
-4.39036071e-01 -5.37540376e-01 -9.33925807e-01 4.47289020e-01
4.40254100e-02 1.16867296e-01 -4.11072433e-01 1.28586912e+00
6.48962080e-01 1.01292819e-01 2.32851982e-01 6.46204591e-01
7.00659215e-01 3.84134501e-01 -1.43647373e-01 -2.54952878e-01
1.02343810e+00 -2.08336666e-01 -2.15672880e-01 4.80210274e-01
2.75566041e-01 -7.46210873e-01 5.97142994e-01 3.33531469e-01
-8.07328224e-01 -5.45657516e-01 -1.09011638e+00 4.03122723e-01
-5.74074447e-01 3.56582612e-01 3.54051262e-01 7.94568300e-01
-9.16735291e-01 6.29162788e-01 -4.82820064e-01 -3.06107581e-01
2.67506480e-01 2.92607844e-01 -5.74572921e-01 -3.89089324e-02
-5.19784510e-01 8.50108743e-01 3.79525632e-01 2.29625419e-01
-3.12319458e-01 -4.16046470e-01 -6.89392090e-01 9.32964310e-02
2.09583297e-01 -5.03824830e-01 9.93601084e-02 -7.96649754e-01
-1.47281325e+00 9.27199125e-01 1.25529677e-01 -2.91286200e-01
5.59134781e-01 3.95680755e-01 -2.97764093e-01 1.33103579e-01
-3.24570417e-01 -1.98830292e-02 1.27680135e+00 -1.67033327e+00
5.75254075e-02 -7.34276652e-01 -3.66521925e-01 1.50831863e-01
-4.98050302e-01 -2.72222877e-01 -2.06718385e-01 -5.49903572e-01
4.97820169e-01 -9.89780664e-01 -2.50184506e-01 -2.87467211e-01
-2.43034914e-01 -2.19947860e-01 9.72410917e-01 -5.25671363e-01
1.13506007e+00 -2.17516899e+00 7.37426996e-01 1.17744696e+00
1.18477777e-01 4.45282996e-01 -5.14860265e-02 8.23896587e-01
-2.64916480e-01 -4.01211709e-01 -6.89715385e-01 -5.34100473e-01
1.52263008e-02 2.67565697e-01 -2.48785228e-01 6.54608190e-01
1.98320925e-01 5.42273104e-01 -5.51357865e-01 -3.20444107e-01
4.78241622e-01 6.02867782e-01 -4.04448062e-01 9.89980698e-02
-1.02984592e-01 3.26771379e-01 -6.10241175e-01 3.63227397e-01
8.96849394e-01 2.09001154e-01 3.92363101e-01 -3.18306565e-01
-3.24570894e-01 -4.75371718e-01 -1.57865012e+00 1.51798153e+00
-5.00263751e-01 -8.95970762e-02 -1.26364648e-01 -1.42797220e+00
1.75376534e+00 3.17985028e-01 9.60235596e-01 -2.25043640e-01
6.45520985e-02 3.99597645e-01 -2.12558389e-01 -2.96695501e-01
3.16777915e-01 9.10138115e-02 2.36287281e-01 3.85968417e-01
1.00237273e-01 -2.35617861e-01 1.66543722e-01 -9.01790112e-02
8.06004167e-01 2.97562450e-01 6.91577911e-01 -3.40642452e-01
1.17688835e+00 -1.94332048e-01 3.08441259e-02 1.62355557e-01
3.12778443e-01 5.09686112e-01 4.99441534e-01 -6.01470113e-01
-1.06220341e+00 -8.41162145e-01 -2.36757189e-01 4.68163043e-01
-4.54588085e-01 -2.66083121e-01 -4.62838948e-01 -7.34283686e-01
3.45918000e-01 3.11226904e-01 -4.86247241e-01 6.83249384e-02
-4.74801421e-01 -4.70445901e-01 1.63897455e-01 1.98799577e-02
1.33775501e-02 -1.03346908e+00 -6.86513066e-01 2.69241631e-01
5.64002156e-01 -7.28146076e-01 1.53386906e-01 1.66580662e-01
-1.16305959e+00 -1.17540169e+00 -7.93283641e-01 -4.32278067e-01
7.82442212e-01 3.14796977e-02 6.46152854e-01 -7.98858032e-02
-2.71915048e-01 7.23144710e-01 -5.83018720e-01 -2.12692514e-01
-3.17243397e-01 -9.49995443e-02 8.26201141e-02 9.04127121e-01
2.10864663e-01 -1.09366620e+00 -1.49536952e-01 -3.17986757e-02
-9.57999349e-01 -3.87330115e-01 8.66363704e-01 8.32869411e-01
6.82536066e-01 -1.60263613e-01 4.04244035e-01 -9.78253901e-01
5.30527830e-01 -5.40085554e-01 -7.36236453e-01 1.90999061e-01
-3.55029732e-01 1.37334809e-01 1.00374341e+00 -2.14857548e-01
-9.11450922e-01 6.18352354e-01 -2.09138408e-01 -6.52594030e-01
-2.21016839e-01 5.80547690e-01 -2.93024391e-01 -3.94419611e-01
7.30561733e-01 7.24437177e-01 1.79316118e-01 -7.37008274e-01
4.52959269e-01 5.28069019e-01 7.88514093e-02 -5.28160572e-01
1.14385319e+00 4.17238444e-01 6.23957813e-01 -1.28970110e+00
-2.02833638e-01 -6.47841990e-01 -8.87582719e-01 -3.07440698e-01
4.95223194e-01 -4.13474947e-01 -5.31973541e-01 4.41968739e-02
-9.61478353e-01 5.18623292e-01 -4.85249609e-01 4.31225300e-01
-7.83979595e-01 4.78761792e-01 1.57149479e-01 -1.23282266e+00
-2.84222037e-01 -8.81584823e-01 8.92523050e-01 -7.79152438e-02
-8.27050209e-02 -9.37487900e-01 3.50361019e-01 2.22099453e-01
2.41433561e-01 5.34919322e-01 1.19818604e+00 -1.11417699e+00
-3.79568577e-01 -4.16846573e-01 3.83883342e-02 4.93086696e-01
4.56104614e-02 -4.87545021e-02 -6.74560428e-01 -3.37980956e-01
3.35378319e-01 4.51109186e-02 5.43344319e-01 8.59365761e-02
7.73763001e-01 -3.24549600e-02 -1.74302682e-01 4.55411792e-01
1.78637731e+00 1.38151452e-01 4.17888165e-01 -4.51994203e-02
5.54874599e-01 6.51985943e-01 3.93437117e-01 7.50990808e-01
-7.57271936e-03 9.77171004e-01 3.25488597e-01 1.79048419e-01
7.40128895e-03 -1.61988422e-01 2.48490244e-01 9.65456545e-01
-5.21440804e-01 1.43024489e-01 -9.22543347e-01 9.10561010e-02
-1.85045373e+00 -8.57987344e-01 -2.69240111e-01 2.56030130e+00
8.55213031e-02 7.30176121e-02 2.70313084e-01 7.24802494e-01
5.41243374e-01 2.20159367e-01 -5.26555292e-02 -4.55361068e-01
-2.00970665e-01 5.08666694e-01 5.75532205e-02 3.70043159e-01
-9.77817178e-01 4.24664229e-01 4.61907291e+00 6.94396734e-01
-1.04211545e+00 -2.37567618e-01 9.34923440e-02 5.33175111e-01
-3.85898054e-01 1.09094847e-02 -6.05367184e-01 3.30000490e-01
6.54640973e-01 3.99097875e-02 2.19887242e-01 1.03009439e+00
1.67293400e-01 9.39220935e-02 -8.28756392e-01 1.03032982e+00
4.30294722e-01 -1.06032777e+00 3.88259768e-01 1.76029265e-01
7.01854050e-01 -3.60474706e-01 1.64685160e-01 -5.62847331e-02
-3.48046958e-01 -6.78041935e-01 6.01652443e-01 1.00507224e+00
4.71383601e-01 -8.94141376e-01 6.80061579e-01 4.35171008e-01
-1.55690169e+00 -1.96570233e-01 -4.57858205e-01 2.20473692e-01
8.48541707e-02 7.14747429e-01 -7.54528046e-01 1.11465740e+00
-2.93923505e-02 9.15841222e-01 -4.66625631e-01 1.15110385e+00
7.70063102e-02 3.22068602e-01 -4.09663349e-01 -2.30283320e-01
2.34106630e-01 -8.65589678e-01 1.07040906e+00 1.07829332e+00
5.22470534e-01 -7.32649490e-02 2.70024240e-01 9.43145931e-01
2.66640931e-01 8.59704554e-01 -1.15411210e+00 1.25902891e-01
1.34175807e-01 1.44758391e+00 -9.30727422e-01 -9.30176154e-02
-2.66619653e-01 5.79405248e-01 1.81071877e-01 6.06757551e-02
-3.79869819e-01 -8.27449001e-03 2.90828139e-01 -6.28117705e-03
6.43285990e-01 -3.56699169e-01 -2.37661853e-01 -9.21007395e-01
1.40762955e-01 -6.76307917e-01 2.25512266e-01 -3.10001463e-01
-1.25862980e+00 5.55098772e-01 -4.47770096e-02 -1.70492125e+00
-2.72535920e-01 -6.85254753e-01 -5.64407229e-01 9.46704566e-01
-1.02939248e+00 -1.33397138e+00 -1.28966168e-01 8.54521692e-01
2.51175284e-01 -5.80999732e-01 1.03555071e+00 3.33763212e-01
-2.57049501e-01 1.58270091e-01 2.64378756e-01 -2.01812357e-01
-2.29982510e-02 -1.16181648e+00 -2.39382967e-01 5.73169708e-01
7.04293430e-01 3.89823198e-01 5.13994634e-01 -6.27541065e-01
-1.91989970e+00 -7.91322649e-01 1.01583016e+00 -2.55964100e-01
5.82140386e-01 -3.31614256e-01 -8.36462080e-01 2.18827024e-01
-4.28336859e-01 2.71141291e-01 7.39379525e-01 -6.59992099e-02
-1.73496649e-01 -1.40248448e-01 -1.09853625e+00 1.78270549e-01
9.11603034e-01 -5.24656892e-01 -7.60667026e-01 1.62744761e-01
2.33006151e-03 -7.52748549e-02 -1.18491590e+00 9.93186086e-02
6.31505489e-01 -1.06016147e+00 1.20424581e+00 -5.16784191e-01
9.92477015e-02 -2.51197726e-01 -3.42340469e-01 -1.25085533e+00
-1.32327914e-01 -2.25034192e-01 -1.62720099e-01 1.22291756e+00
2.01173410e-01 -5.62002242e-01 1.05086029e+00 1.70238569e-01
1.14517540e-01 -8.93203259e-01 -1.02524841e+00 -6.73315287e-01
-4.10154134e-01 -3.65613103e-01 5.11018336e-01 8.13130081e-01
-2.13992789e-01 3.09774190e-01 -1.89008206e-01 -9.14202817e-03
7.08975852e-01 5.09455740e-01 8.16425323e-01 -1.50602758e+00
-3.84605646e-01 -4.73052740e-01 -1.06173849e+00 -4.44792181e-01
2.39574641e-01 -1.26504493e+00 -5.31820118e-01 -9.53095675e-01
-2.44590580e-01 -6.55335367e-01 -2.45825231e-01 1.07720204e-01
4.40130174e-01 1.67567953e-01 5.78002632e-01 1.91691339e-01
2.29318310e-02 7.40671754e-01 9.15612638e-01 -1.26550391e-01
-3.90985757e-01 5.32454729e-01 -2.22784206e-01 6.23962164e-01
2.79193282e-01 -2.86401749e-01 -5.44940293e-01 2.59344012e-01
6.69925734e-02 2.06495330e-01 1.13108441e-01 -1.22105980e+00
2.87314087e-01 1.07049793e-01 3.44787568e-01 -7.92227387e-01
5.62286854e-01 -1.36704338e+00 6.09621227e-01 5.80944955e-01
-4.99467291e-02 3.72913256e-02 -2.22951755e-01 6.55133963e-01
-3.93853098e-01 -4.68251467e-01 6.25038683e-01 -5.20267226e-02
-4.73066866e-01 2.88360894e-01 -3.64965796e-02 -3.98748755e-01
1.09233844e+00 -5.32718778e-01 6.20753586e-01 -6.75519481e-02
-1.11133373e+00 -4.79028821e-01 1.44031316e-01 6.03064969e-02
7.98753977e-01 -1.48387575e+00 -7.43487000e-01 3.61194491e-01
2.57826388e-01 -3.11003566e-01 4.16452110e-01 7.61209249e-01
-2.11492538e-01 3.52394670e-01 -4.10968721e-01 -6.43788218e-01
-1.35953712e+00 6.12213969e-01 4.76084417e-04 -4.25577819e-01
-7.14008033e-01 3.24661016e-01 -5.46223044e-01 -4.94303584e-01
1.11014477e-03 -1.22655958e-01 -9.19481277e-01 6.60456777e-01
9.58161131e-02 4.73710120e-01 4.83575732e-01 -1.18562567e+00
-3.85701448e-01 9.59298730e-01 6.26425683e-01 8.55590701e-02
1.85887647e+00 2.65964150e-01 -3.33467096e-01 3.58151019e-01
1.32156599e+00 4.10571069e-01 -8.13415051e-01 -2.19810992e-01
3.84604514e-01 -4.29341793e-01 -1.85484618e-01 -3.78159493e-01
-7.78846264e-01 4.99729693e-01 6.06662571e-01 4.06378746e-01
1.03146386e+00 -2.82818794e-01 1.10173203e-01 5.12795746e-01
4.86390024e-01 -7.18394697e-01 -2.54030794e-01 3.75562519e-01
1.40716934e+00 -9.12755728e-01 1.69016615e-01 -5.42397618e-01
-4.21056569e-01 1.52763820e+00 -4.77036163e-02 -5.81839383e-01
8.00220609e-01 6.20525740e-02 -5.09026051e-01 -2.85072565e-01
-2.64722109e-01 -3.85170579e-01 6.02401316e-01 5.12340665e-01
8.10821280e-02 3.62432837e-01 -8.09354246e-01 4.87213731e-01
-1.32381320e-01 -9.85582843e-02 1.18095614e-01 5.67294717e-01
-3.03141564e-01 -1.46092308e+00 -6.21066391e-01 3.51224691e-01
3.24248940e-01 3.43402267e-01 -5.28746545e-01 7.29172885e-01
3.01137865e-01 3.78912210e-01 -3.14199239e-01 -4.36509579e-01
6.07054532e-01 2.55681992e-01 3.69217604e-01 -5.53731322e-01
-7.23280430e-01 -2.64151931e-01 -1.38937593e-01 -3.99826139e-01
-4.93719548e-01 -8.44138384e-01 -8.25997889e-01 1.48736477e-01
-1.01036072e-01 1.95821032e-01 1.02549303e+00 9.06590939e-01
1.84143618e-01 4.84430930e-03 1.10795271e+00 -1.24289572e+00
-8.38719666e-01 -5.86738229e-01 -8.49889100e-01 7.55498469e-01
-1.62346184e-01 -1.02100313e+00 -1.56036079e-01 -3.51772338e-01]
|
[8.048518180847168, 3.964935064315796]
|
c39b5b22-f08e-452e-9c21-c0422bbc80ee
|
crowd-source-scene-change-detection-and-local
|
2203.05205
| null |
https://arxiv.org/abs/2203.05205v1
|
https://arxiv.org/pdf/2203.05205v1.pdf
|
Crowd Source Scene Change Detection and Local Map Update
|
As scene changes with time map descriptors become outdated, affecting VPS localization accuracy. In this work, we propose an approach to detect structural and texture scene changes to be followed by map update. In our method - map includes 3D points with descriptors generated either via LiDAR or SFM. Common approaches suffer from shortcomings: 1) Direct comparison of the two point-clouds for change detection is slow due to the need to build new point-cloud every time we want to compare; 2) Image based comparison requires to keep the map images adding substantial storage overhead. To circumvent this problems, we propose an approach based on point-clouds descriptors comparison: 1) Based on VPS poses select close query and map images pairs, 2) Registration of query images to map image descriptors, 3) Use segmentation to filter out dynamic or short term temporal changes, 4) Compare the descriptors between corresponding segments.
|
['Ofer Kruzel', 'Feng Wensen', 'Omri Asraf', 'Firas Shama', 'Lin Manqing', 'Nati Daniel', 'Itzik Wilf']
|
2022-03-10
| null | null | null | null |
['scene-change-detection']
|
['computer-vision']
|
[ 3.19091618e-01 -7.51709580e-01 1.85321718e-01 -4.58946019e-01
-7.14522481e-01 -8.17563832e-01 7.10663080e-01 7.60556400e-01
-5.19158065e-01 4.65231538e-01 -4.36516732e-01 1.20862067e-01
-1.59478948e-01 -1.13760984e+00 -7.08825946e-01 -2.03527465e-01
-7.58810788e-02 1.06854153e+00 1.31902051e+00 -2.73891091e-01
7.59101808e-01 1.15232909e+00 -1.90344107e+00 -1.87749386e-01
8.46053243e-01 1.07005692e+00 5.82850754e-01 5.47628760e-01
-6.08494759e-01 -5.79822920e-02 -4.46047455e-01 3.17146540e-01
7.01827526e-01 2.92289872e-02 -8.10475647e-01 -1.48373246e-01
7.28666842e-01 -2.88389057e-01 1.26365982e-02 1.12793422e+00
3.10119867e-01 3.97137970e-01 5.13680398e-01 -1.41340232e+00
7.21482784e-02 -2.27536842e-01 -7.28623211e-01 3.59732687e-01
6.96714699e-01 6.91608191e-02 1.66229010e-01 -7.59990692e-01
1.07727289e+00 1.09565973e+00 1.00235486e+00 -1.45403504e-01
-1.40526092e+00 -6.18876934e-01 4.34096828e-02 5.45122027e-01
-1.90056813e+00 -3.45478535e-01 7.21712232e-01 -6.18595123e-01
8.36824179e-01 4.92400110e-01 8.45243752e-01 2.35810921e-01
-3.67601365e-02 -1.05435356e-01 1.02319872e+00 -2.85224140e-01
3.11587632e-01 2.35320032e-01 3.03210113e-02 4.06774282e-01
9.45181027e-02 -3.71190421e-02 -3.94728601e-01 -3.92409950e-01
6.12301886e-01 7.43838549e-02 -1.36005595e-01 -6.12066925e-01
-1.19088089e+00 5.46582103e-01 4.62844938e-01 4.03437823e-01
-6.62522316e-01 1.20153472e-01 2.44012237e-01 4.60462362e-01
3.66598010e-01 2.11568207e-01 -4.14368212e-01 -2.89977163e-01
-1.20749760e+00 3.88708115e-01 3.29949766e-01 1.11660087e+00
1.46346521e+00 -5.84469795e-01 3.16921234e-01 4.66738105e-01
2.95984596e-01 7.48223603e-01 3.57810259e-01 -6.32258415e-01
3.94155860e-01 7.47447848e-01 2.09366471e-01 -1.44500637e+00
-2.41518438e-01 3.41427952e-01 -3.23713988e-01 2.39952862e-01
5.18121244e-03 7.29636252e-01 -8.28164816e-01 8.42707157e-01
5.99041700e-01 3.42993081e-01 -2.70875067e-01 4.69320804e-01
7.80930400e-01 7.61420667e-01 -1.71936736e-01 -1.45876169e-01
9.28171396e-01 -2.00092226e-01 -4.15349126e-01 9.54393446e-02
4.02805150e-01 -1.03925478e+00 7.12047160e-01 -3.21057849e-02
-9.18510675e-01 -7.86234915e-01 -1.00522864e+00 1.04750186e-01
-8.77472758e-01 -4.43482041e-01 -2.39891429e-02 2.11333066e-01
-1.39682353e+00 8.02361488e-01 -8.45289052e-01 -8.84046972e-01
-3.86822689e-03 7.55569100e-01 -2.14106143e-01 1.91779777e-01
-7.60333478e-01 9.30363417e-01 6.84617281e-01 -2.22306356e-01
-3.54506403e-01 -6.66365325e-01 -6.29138768e-01 -2.96419442e-01
5.69253676e-02 -6.10935628e-01 7.78383195e-01 -7.32969761e-01
-1.22958243e+00 1.06477320e+00 -5.49340606e-01 -2.17705905e-01
5.22218466e-01 1.31283760e-01 -2.88143545e-01 1.07335106e-01
6.11227214e-01 6.65381432e-01 5.95088959e-01 -1.44262779e+00
-9.75947917e-01 -6.12893760e-01 -3.16359490e-01 4.56544936e-01
2.79439420e-01 -3.26033756e-02 -7.27592111e-01 -5.44085028e-03
9.91990745e-01 -8.52969706e-01 -2.15798646e-01 3.05156521e-02
-1.45245776e-01 -1.57241881e-01 1.47501135e+00 -4.32628810e-01
9.16687191e-01 -2.03112006e+00 -4.53035563e-01 6.76519394e-01
-9.45144892e-02 1.94857270e-01 2.69524977e-02 5.74314177e-01
1.73133746e-01 2.20257901e-02 -8.32479447e-02 -1.98012128e-01
-2.32915759e-01 2.05968738e-01 -3.00831676e-01 5.01222789e-01
-1.99234992e-01 3.76294285e-01 -7.68918872e-01 -8.17549169e-01
7.54779696e-01 2.88767099e-01 -2.76398044e-02 -9.05927345e-02
1.14699071e-02 6.18322432e-01 -5.83418608e-01 6.93972111e-01
1.30303192e+00 3.72608423e-01 -5.65466940e-01 -2.64503717e-01
-7.40668118e-01 3.56321931e-01 -1.56347322e+00 1.64301622e+00
-5.83110787e-02 6.93253994e-01 -1.38054669e-01 -7.22559988e-01
1.29597652e+00 -8.69352669e-02 8.05959940e-01 -6.93330348e-01
-1.42433628e-01 4.85973537e-01 -8.33412647e-01 -3.61757189e-01
9.39355254e-01 3.47527534e-01 2.25369155e-01 2.63003614e-02
-5.07240176e-01 -8.06176901e-01 1.15308866e-01 -1.10243507e-01
8.97421300e-01 1.33423358e-01 1.73023984e-01 -1.64746031e-01
7.95603037e-01 6.66937828e-01 4.00725722e-01 6.91200972e-01
-1.40898585e-01 5.66983879e-01 -1.49161592e-01 -6.65525258e-01
-8.78828526e-01 -1.09113169e+00 -4.55166131e-01 4.23499763e-01
8.78708303e-01 -1.88414782e-01 -3.14072758e-01 -2.74411857e-01
2.45459706e-01 2.32075870e-01 -1.07325852e-01 4.20192659e-01
-7.76813984e-01 -3.76175165e-01 7.71750808e-02 1.37077287e-01
7.81465828e-01 -5.80532253e-01 -9.22599316e-01 3.55688483e-01
-1.32107005e-01 -8.98653090e-01 -2.50297427e-01 -1.19016849e-01
-1.20429730e+00 -9.96205270e-01 -3.45856875e-01 -7.28090465e-01
6.54775858e-01 6.34384632e-01 9.94254947e-01 -4.14313423e-03
-2.45899320e-01 6.69307828e-01 -3.09135616e-01 -3.84943634e-01
-3.30248177e-01 -6.47841841e-02 8.63915868e-03 -2.17015132e-01
4.03555185e-01 -7.76019216e-01 -4.88907248e-01 6.47625804e-01
-6.69448376e-01 -1.11724325e-01 5.37810810e-02 1.61226004e-01
1.27928829e+00 2.35566661e-01 -1.30793482e-01 -3.86401772e-01
2.63028026e-01 -2.61984199e-01 -1.35225189e+00 1.92695513e-01
-7.36178935e-01 -2.57190257e-01 -2.42142342e-02 -1.23338416e-01
-5.91610670e-01 5.26076078e-01 6.98375255e-02 -4.05196816e-01
-3.42379153e-01 1.95218876e-01 2.69330651e-01 -6.47490263e-01
7.01426446e-01 5.10055482e-01 -6.29289746e-02 -3.70063156e-01
2.44045094e-01 7.26605475e-01 6.90295339e-01 -1.95708305e-01
1.15796793e+00 8.30267906e-01 1.85337260e-01 -9.73372936e-01
2.65821725e-01 -1.12862384e+00 -1.33276248e+00 -4.31193471e-01
7.85204113e-01 -7.95489311e-01 -5.70009291e-01 2.57989168e-01
-1.54738891e+00 6.40404969e-02 -2.80526131e-01 5.00080526e-01
-5.93616784e-01 5.04826069e-01 -1.37414470e-01 -6.05555177e-01
-2.53236085e-01 -1.07283187e+00 1.25436735e+00 3.98035049e-01
3.10475472e-02 -7.28213966e-01 5.34073293e-01 -4.46575023e-02
4.69653398e-01 6.67743146e-01 4.45689172e-01 -2.31211126e-01
-1.12031543e+00 -3.61950547e-01 -3.31679106e-01 -3.71069223e-01
1.79306909e-01 2.38409355e-01 -7.54125953e-01 -2.48843268e-01
-1.01511583e-01 5.10255098e-01 3.45080882e-01 3.93223792e-01
6.67370558e-01 3.14108014e-01 -9.93437350e-01 6.59066021e-01
1.87620246e+00 6.74064577e-01 6.08234525e-01 7.55496681e-01
5.01814425e-01 4.93880689e-01 1.19216597e+00 3.36774856e-01
5.74919879e-01 1.06680226e+00 4.03964847e-01 -7.66878873e-02
6.99885860e-02 -2.44045883e-01 7.51761124e-02 7.75446594e-01
-1.46994233e-01 2.38043919e-01 -1.32306588e+00 5.27689874e-01
-1.83858562e+00 -7.76461840e-01 -8.07556570e-01 2.72305274e+00
2.79496580e-01 2.33229585e-02 -4.45754007e-02 -1.62594333e-01
9.38915968e-01 9.94712114e-03 -4.03784186e-01 -7.99249187e-02
7.63713121e-02 1.31528690e-01 1.00089526e+00 6.13547921e-01
-1.09898388e+00 1.04987049e+00 6.06245470e+00 4.27400649e-01
-1.39163542e+00 1.58740580e-01 -8.57917592e-02 2.18790144e-01
-4.14904319e-02 3.82102430e-01 -8.37334216e-01 6.96275592e-01
6.41595900e-01 -2.09991306e-01 1.20372348e-01 7.40320861e-01
1.98355407e-01 -7.21411347e-01 -7.90160120e-01 1.43714106e+00
-5.13224769e-03 -1.20048189e+00 -1.60613447e-01 3.47456336e-02
6.34665370e-01 5.32249808e-01 -4.34517115e-01 -6.95095211e-02
-1.78922579e-01 -4.23161715e-01 7.67008245e-01 7.41274834e-01
7.84617305e-01 -4.89328951e-01 6.49992466e-01 2.01670110e-01
-1.82932210e+00 4.96441245e-01 -7.05024540e-01 2.44456887e-01
2.61561394e-01 3.08928639e-01 -1.13168252e+00 6.78436100e-01
8.95320296e-01 4.75016922e-01 -8.25263679e-01 1.42472947e+00
2.94052511e-01 -7.29057044e-02 -8.30867469e-01 1.41518772e-01
1.30760625e-01 -5.51344037e-01 8.04962516e-01 1.00972641e+00
5.78652978e-01 -1.54526606e-01 4.32797670e-01 9.40307558e-01
6.12832963e-01 2.59258091e-01 -9.63354468e-01 4.81110036e-01
8.43660891e-01 8.52301776e-01 -1.15333056e+00 -4.18276072e-01
-7.43089020e-02 1.21231794e+00 -3.19586724e-01 1.78050488e-01
-5.55950880e-01 -6.05298579e-01 3.80001813e-01 5.79870164e-01
1.45860001e-01 -6.13281846e-01 -1.26738455e-02 -6.84118211e-01
2.05141410e-01 3.67271081e-02 1.06875680e-01 -8.80442142e-01
-7.46721566e-01 6.51326656e-01 2.98911065e-01 -1.51276946e+00
-2.65615851e-01 1.16837941e-01 -5.43730915e-01 1.07195020e+00
-1.64748931e+00 -9.86895323e-01 -8.59587610e-01 1.00312304e+00
4.51996148e-01 2.82649517e-01 5.61584949e-01 5.15132308e-01
7.21132681e-02 -1.37938216e-01 3.73330057e-01 -3.05549234e-01
6.86712682e-01 -1.06436110e+00 6.66635394e-01 6.84619069e-01
9.53197777e-02 1.09448858e-01 6.78777218e-01 -1.05004096e+00
-1.06768346e+00 -9.11349058e-01 1.00809693e+00 -6.30300343e-01
3.07250649e-01 -2.30024263e-01 -9.54589367e-01 5.30255079e-01
-3.02936137e-01 -2.60932799e-02 1.43648386e-01 -1.56563297e-01
1.29642144e-01 -5.11855543e-01 -1.36057687e+00 8.72423351e-02
9.04862821e-01 -5.61619461e-01 -4.59680468e-01 4.67034996e-01
4.12632346e-01 -6.67444587e-01 -8.05189550e-01 4.34160262e-01
3.47446918e-01 -1.01939201e+00 9.02989566e-01 3.25708807e-01
-6.17550015e-01 -1.02280343e+00 -2.54851252e-01 -9.59035814e-01
-2.00198814e-01 -4.64393526e-01 7.15073466e-01 1.29239130e+00
-4.13297080e-02 -8.51186335e-01 7.45222211e-01 5.57143807e-01
-6.99191839e-02 -6.28152415e-02 -1.32686067e+00 -9.87572968e-01
-6.83096051e-01 -2.51497298e-01 9.88035619e-01 1.03864479e+00
-4.70991522e-01 -5.22372276e-02 3.64702731e-01 6.62695587e-01
3.76456529e-01 2.18633011e-01 1.19275153e+00 -1.75642490e+00
3.05697143e-01 -3.96378070e-01 -1.14953828e+00 -9.34990406e-01
-4.32838559e-01 -6.17973506e-01 1.45816922e-01 -1.68152297e+00
-1.39466852e-01 -1.16549540e+00 1.71118304e-01 5.50311916e-02
3.19063306e-01 2.76186019e-01 1.36230081e-01 8.45726252e-01
-5.55157959e-01 2.50958681e-01 4.79934514e-01 6.97168782e-02
-5.80139816e-01 3.59398983e-02 4.04691845e-01 4.86611307e-01
5.60458481e-01 -7.02256620e-01 -3.74706894e-01 -2.86912441e-01
1.01716623e-01 8.28417763e-03 3.53383332e-01 -1.47654414e+00
4.57066149e-01 -2.97838748e-01 2.11753860e-01 -1.50381398e+00
5.68815887e-01 -1.14321816e+00 8.07975054e-01 6.15272582e-01
4.37301725e-01 5.30344665e-01 2.70805389e-01 4.99972224e-01
-4.24229234e-01 -5.83963275e-01 8.47514808e-01 -1.75250486e-01
-1.01964378e+00 4.83957708e-01 -2.30843246e-01 -6.66682780e-01
1.45800817e+00 -9.10534501e-01 -1.91442650e-02 -1.80104673e-01
-4.27774489e-01 2.19450280e-01 1.03515613e+00 2.23857358e-01
4.90413100e-01 -1.35173905e+00 -3.53100657e-01 1.79193377e-01
2.35866487e-01 3.70801508e-01 2.01642707e-01 8.13260436e-01
-1.28431594e+00 3.94164354e-01 -1.43253103e-01 -1.24073052e+00
-1.69626105e+00 5.28129816e-01 2.54623562e-01 1.17210954e-01
-5.91072261e-01 6.48485780e-01 -2.37920120e-01 -4.59522218e-01
-3.62857580e-02 -4.95890498e-01 -1.65114373e-01 2.74017930e-01
1.73872188e-01 6.09781146e-01 2.46794015e-01 -1.04893136e+00
-6.58246815e-01 1.37453270e+00 1.96928039e-01 -2.28587806e-01
1.09229732e+00 -5.29785156e-01 -2.08104759e-01 5.53038180e-01
1.25976372e+00 -7.87957311e-02 -8.36239159e-01 -3.38814765e-01
3.14621001e-01 -1.03667033e+00 1.99274160e-02 -1.93897083e-01
-5.89730263e-01 5.66895366e-01 1.50093210e+00 2.88410425e-01
1.08709848e+00 1.03341497e-01 7.70035386e-01 3.57578665e-01
8.66970003e-01 -1.24607575e+00 -5.28655648e-01 4.28293377e-01
6.91198885e-01 -1.25149536e+00 1.92025051e-01 -5.28813481e-01
-1.31684616e-01 1.26427162e+00 3.07592809e-01 -5.53552844e-02
6.90429270e-01 -1.80749267e-01 -2.49198768e-02 -5.45365453e-01
-1.24682235e-02 -3.76661092e-01 1.34328514e-01 7.76825368e-01
-3.43701065e-01 -4.44695801e-02 -3.01952302e-01 -5.65147340e-01
-3.13786656e-01 -7.04766884e-02 1.91451281e-01 1.11126554e+00
-7.93234646e-01 -1.17405558e+00 -8.35959256e-01 3.14296126e-01
3.75489682e-01 1.14864245e-01 -2.44950980e-01 8.76409948e-01
4.88282949e-01 6.50490820e-01 6.59260690e-01 -2.85347760e-01
7.08099365e-01 -1.82391137e-01 4.25198376e-01 -3.62053812e-01
-4.92309809e-01 4.34649456e-03 -3.87380272e-01 -7.31482923e-01
-5.79884648e-01 -1.00708330e+00 -1.20334792e+00 -3.69151413e-01
-4.75144655e-01 5.56127653e-02 1.36279738e+00 4.91363734e-01
6.63534582e-01 -2.44171992e-01 7.94096470e-01 -1.02143729e+00
2.19562929e-02 -7.01594770e-01 -5.67351878e-01 4.50063050e-01
1.44389242e-01 -7.49116659e-01 -1.26874685e-01 1.27773032e-01]
|
[7.759767532348633, -2.4599556922912598]
|
470784a9-746a-4da1-987c-29d6f6be55d2
|
pixel-level-kernel-estimation-for-blind-super
| null | null |
https://ieeexplore.ieee.org/document/9615068
|
https://ieeexplore.ieee.org/document/9615068
|
Pixel-Level Kernel Estimation for Blind Super-Resolution
|
Throughout the past several years, deep learning-based models have achieved success in super-resolution (SR). The majority of these works assume that low-resolution (LR) images are ‘uniformly’ degraded from their corresponding high-resolution (HR) images using predefined blur kernels — all regions of an image undergoing an identical degradation process. Furthermore, based on this assumption, there have been attempts to estimate the blur kernel of a given LR image, since correct kernel priors are known to be helpful in super-resolution. Although it has been known that blur kernels of real images are non-uniform (spatially varying), current kernel estimation algorithms are mostly done at image-level, estimating one kernel per image. These algorithms inevitably become sub-optimal in handling scenarios where an image is degraded non-uniformly. A divide-and-conquer form of approach, dividing an image into several patches for individual kernel estimation and SR can be a simple solution for this matter. Nevertheless, this approach fails in practice. In this paper, we address this issue by pixel-level kernel estimation. The three main components for training a SR framework based on pixel-level kernel estimation are as follows: Kernel Collage — a method for synthesizing non-uniformly degraded LR images, designed considering the coherency of kernels at neighboring regions while abruptly changing at times, the indirect loss — a novel loss for training the kernel estimator, based on the reconstruction loss, and an additional optimization — a scheme to robustify the SR network to minor errors in kernel estimations. Extensive experiments show the superiority of pixel-level kernel estimation in blind SR, surpassing state-of-the-art methods in terms of quantitative and qualitative results.
|
['Jae-Pil Heo', 'Euiyeon Kim', 'Jaihyun Lew']
|
2021-11-15
| null | null | null |
ieee-access-2021-11
|
['super-resolution']
|
['computer-vision']
|
[ 2.18370661e-01 -2.09890142e-01 -6.07698299e-02 -1.83745086e-01
-9.60953176e-01 -3.19429457e-01 2.67126471e-01 -4.28797334e-01
-2.95728952e-01 9.86750543e-01 1.15344413e-01 1.33900940e-01
-3.27606201e-01 -4.81407583e-01 -7.07967222e-01 -1.05119979e+00
-1.19284630e-01 -6.68630823e-02 3.99818033e-01 6.02043234e-02
1.80201083e-01 8.02908838e-01 -1.63612401e+00 2.91479081e-01
1.25241613e+00 9.19414580e-01 4.28924948e-01 7.93208599e-01
3.48209292e-01 9.73848939e-01 -5.38004518e-01 -7.48493448e-02
1.87345698e-01 -4.17650759e-01 -6.23110116e-01 1.86797336e-01
7.67382145e-01 -5.16154230e-01 -5.81876159e-01 1.23503482e+00
5.15178025e-01 5.66284359e-02 6.20062470e-01 -5.10513902e-01
-9.81789827e-01 4.74714041e-01 -7.48156607e-01 3.12412918e-01
1.39920011e-01 3.73112857e-02 5.84469914e-01 -8.53973329e-01
3.60176653e-01 1.01333284e+00 7.86414862e-01 3.97134066e-01
-1.62339318e+00 -1.56918526e-01 -3.31250995e-01 4.08934295e-01
-1.47002161e+00 -5.99804401e-01 7.43484318e-01 -3.09407592e-01
4.95021641e-01 3.68227810e-01 2.53514230e-01 9.25594926e-01
3.49642128e-01 5.64705908e-01 1.73747516e+00 -3.57628435e-01
2.97167391e-01 4.40827876e-01 1.20738126e-01 2.00079232e-01
1.98548168e-01 3.34491968e-01 -2.74493814e-01 -1.36376321e-01
1.32120967e+00 -2.77909517e-01 -1.00488639e+00 -4.52335358e-01
-1.09997261e+00 4.54171151e-01 5.79855680e-01 5.61703205e-01
-4.61848050e-01 -1.38222501e-01 7.77053311e-02 3.62478226e-01
6.31912947e-01 3.81356061e-01 -3.01240534e-01 1.92282364e-01
-1.42755854e+00 2.78022617e-01 5.55884123e-01 4.16793823e-01
6.98688090e-01 9.86735597e-02 -3.40217978e-01 1.02077723e+00
-1.51820257e-01 2.17086300e-01 3.93526852e-01 -9.27752018e-01
-9.11844149e-02 4.11147587e-02 6.16831362e-01 -8.45699012e-01
-9.52030867e-02 -6.99241281e-01 -1.23836958e+00 6.48138881e-01
7.33749986e-01 2.34424382e-01 -9.15429294e-01 1.62441278e+00
2.14100868e-01 5.10705590e-01 1.57675713e-01 1.36170983e+00
3.93836558e-01 4.48737234e-01 -2.85589278e-01 -4.36037302e-01
1.27642000e+00 -7.59461343e-01 -7.86024630e-01 -7.84569532e-02
-2.83839941e-01 -8.42065454e-01 1.08159387e+00 6.13305867e-01
-1.24207270e+00 -9.75548863e-01 -9.93349016e-01 1.57628674e-03
-1.52528435e-01 6.72427297e-01 2.69588649e-01 5.40596604e-01
-1.42461181e+00 8.57676446e-01 -4.46971804e-01 -1.97925717e-01
4.04579490e-01 6.33453578e-02 -3.70200813e-01 -2.66377449e-01
-1.28782618e+00 1.28898525e+00 2.64777988e-01 3.60710353e-01
-8.44739974e-01 -8.32835734e-01 -5.87819159e-01 1.05932243e-01
2.21664652e-01 -5.77243507e-01 9.35799181e-01 -1.25370824e+00
-1.54137433e+00 8.04554939e-01 -1.15535155e-01 -5.18217325e-01
6.61391377e-01 -3.32018763e-01 -5.24440110e-01 3.20121020e-01
-2.49430537e-01 6.60297275e-02 1.65867758e+00 -1.63685191e+00
-3.64456236e-01 -2.99127698e-01 1.35506943e-01 1.89398780e-01
6.94633797e-02 1.64868385e-01 -1.50263026e-01 -8.41528356e-01
-2.70555299e-02 -4.63611156e-01 -1.55754045e-01 -1.64517201e-02
-1.67290241e-01 2.11757734e-01 6.62417352e-01 -1.16607821e+00
1.07438684e+00 -2.25408387e+00 3.25973392e-01 -3.70745778e-01
1.81413695e-01 3.22185278e-01 4.62602265e-02 1.31032206e-02
-2.13034227e-01 -3.61153096e-01 -5.13327122e-01 -2.98727751e-01
-2.97136664e-01 -2.71382630e-01 -5.73435903e-01 8.36277187e-01
2.73206383e-01 6.24217510e-01 -8.26424181e-01 -3.54819328e-01
4.33107793e-01 9.54421699e-01 1.15536712e-01 4.02292579e-01
9.92960334e-02 4.83872980e-01 1.02004468e-01 4.71670926e-01
1.17333126e+00 -1.13909252e-01 -2.23225340e-01 -7.52629519e-01
-2.54366010e-01 -3.20341378e-01 -1.23139358e+00 1.36862254e+00
-5.75248182e-01 6.75661862e-01 4.82493013e-01 -9.19875979e-01
6.77536666e-01 3.77380818e-01 2.49705762e-01 -2.92895615e-01
-2.12191835e-01 3.44190925e-01 -2.81867355e-01 -2.84143150e-01
5.10409892e-01 -2.07712159e-01 4.47641969e-01 2.04462260e-01
-8.11326355e-02 -1.20179892e-01 -1.95508912e-01 -1.17578521e-01
9.12165046e-01 2.62732804e-01 3.06122839e-01 -2.05933154e-01
8.72766435e-01 -2.97816873e-01 2.86707342e-01 8.85154605e-01
-1.87161639e-01 1.09079134e+00 1.50964960e-01 -2.19314292e-01
-1.19461739e+00 -1.41266692e+00 -5.44740856e-01 3.47503096e-01
1.73323408e-01 3.23667526e-01 -1.13096142e+00 -3.78540039e-01
-2.13272944e-01 7.49610841e-01 -6.06137097e-01 -1.22259147e-01
-4.04315203e-01 -9.78208840e-01 3.86618823e-01 1.84965685e-01
8.75833392e-01 -9.13146973e-01 -6.93527281e-01 1.21255092e-01
-9.57223400e-02 -1.34436762e+00 -3.07564259e-01 1.28409797e-02
-7.87753701e-01 -9.96729195e-01 -1.33124936e+00 -4.73995835e-01
7.01373160e-01 5.32592773e-01 1.04121304e+00 -1.83130860e-01
-5.60602307e-01 2.35973164e-01 -2.06995219e-01 2.93714195e-01
-5.03031671e-01 -4.06455845e-01 1.27618507e-01 5.08774817e-01
-1.89135984e-01 -5.68636894e-01 -7.34088063e-01 3.57638001e-01
-1.05101359e+00 -1.45709310e-02 1.01538050e+00 9.69747901e-01
4.29722458e-01 6.40852869e-01 5.02531648e-01 -4.59764034e-01
4.85992700e-01 -1.53736398e-01 -6.68495357e-01 2.92676657e-01
-6.19152129e-01 -1.01995282e-01 8.58215868e-01 -6.24563754e-01
-1.47319365e+00 -9.96625125e-02 2.27698147e-01 -7.33322263e-01
-4.28150326e-01 8.12576860e-02 -1.69909656e-01 -4.60863918e-01
7.65227377e-01 6.71444535e-01 -1.62124392e-02 -5.54304481e-01
5.29781342e-01 7.30306745e-01 7.51751781e-01 -3.06633949e-01
1.15808368e+00 7.45179236e-01 -2.26554886e-01 -9.92399156e-01
-9.90627646e-01 -4.54494059e-01 -6.86153650e-01 -2.44004562e-01
9.05441523e-01 -9.73996341e-01 -3.03873032e-01 9.30123270e-01
-9.19321477e-01 -4.68680382e-01 -4.32677954e-01 4.77574646e-01
-7.45256007e-01 6.34389937e-01 -7.07431376e-01 -9.22749400e-01
-7.07870126e-02 -1.11128294e+00 8.86844039e-01 5.76060653e-01
2.57958323e-01 -9.40403521e-01 -1.47107616e-01 2.94789076e-01
9.66014564e-01 -5.45368791e-02 6.25949621e-01 1.05632395e-01
-7.22625673e-01 4.46271412e-02 -7.89076269e-01 9.11897242e-01
5.15006363e-01 -3.25541943e-01 -1.36399448e+00 -4.62343395e-01
5.26367605e-01 -1.97071880e-01 9.91874158e-01 7.85630822e-01
1.17321301e+00 -1.80845648e-01 9.26745776e-03 6.94498420e-01
1.81286764e+00 -4.29037869e-01 9.52376544e-01 3.21558863e-01
5.84470570e-01 5.40931880e-01 6.11004531e-01 -1.21719316e-02
-9.25982744e-02 1.04771173e+00 3.19046706e-01 -3.88239652e-01
-5.79663932e-01 2.71858335e-01 5.38075864e-01 5.55826537e-02
-2.84025401e-01 2.11898625e-01 -2.54553914e-01 5.67407250e-01
-1.67238605e+00 -1.17385185e+00 -2.49779224e-02 2.59728789e+00
1.01201248e+00 -5.20723080e-03 -1.07079245e-01 1.26815774e-02
7.62199581e-01 3.41238827e-01 -6.37026966e-01 2.34978925e-02
-5.76257348e-01 8.81806090e-02 6.18158340e-01 6.60583019e-01
-1.17515481e+00 8.54394495e-01 5.99450493e+00 1.04393911e+00
-1.24165130e+00 1.27893731e-01 7.73966074e-01 3.65941316e-01
-6.21747933e-02 -4.44826856e-02 -5.68217278e-01 5.76976418e-01
8.72425139e-01 9.32507068e-02 8.86655211e-01 6.55815482e-01
4.35133070e-01 -5.18101811e-01 -6.86243534e-01 1.04583228e+00
2.03386873e-01 -1.11577749e+00 -2.51780868e-01 -1.11758344e-01
6.38512552e-01 -1.83023408e-01 3.45097840e-01 -5.53059429e-02
9.49551724e-03 -1.10133159e+00 6.44279778e-01 1.05236113e+00
9.69679296e-01 -5.80866754e-01 8.05785418e-01 2.58394688e-01
-9.93952692e-01 -8.00610185e-02 -6.74336553e-01 2.55674154e-01
1.12892218e-01 1.16174352e+00 -3.29866022e-01 8.54911149e-01
9.80874479e-01 4.40149963e-01 -6.57619715e-01 1.15504122e+00
-1.80833995e-01 4.91888285e-01 7.83942640e-02 7.59004891e-01
-2.07149759e-01 -5.15297830e-01 7.59089589e-01 1.17740834e+00
3.48008096e-01 -7.53191411e-02 -2.87151337e-01 1.30783939e+00
3.59895527e-01 -3.22793543e-01 -3.04302752e-01 4.63579178e-01
1.46682829e-01 1.45901644e+00 -6.44632697e-01 -2.69064844e-01
-4.35118794e-01 1.60089374e+00 3.04712862e-01 8.13378155e-01
-9.65456843e-01 -1.95107594e-01 7.00744867e-01 1.41478553e-01
4.80246931e-01 -1.66028455e-01 -2.09829390e-01 -1.26979482e+00
1.41114041e-01 -8.64043593e-01 -6.06193282e-02 -1.09484315e+00
-1.48797834e+00 6.60767734e-01 3.06216180e-02 -1.25182545e+00
1.13941438e-01 -4.66535687e-01 -3.84530336e-01 1.26297045e+00
-2.04145050e+00 -1.15522218e+00 -4.74696964e-01 4.22161102e-01
5.04956901e-01 1.53009996e-01 5.13451695e-01 2.19009444e-01
-3.05311948e-01 3.04053307e-01 3.69113445e-01 -1.01194777e-01
9.86326098e-01 -1.52640986e+00 9.74667445e-03 1.16686261e+00
-5.87406754e-02 5.42928755e-01 9.92486477e-01 -4.05070484e-01
-1.25481844e+00 -1.01579916e+00 4.78188664e-01 -2.96466053e-01
6.09350801e-01 -1.70925424e-01 -1.33928931e+00 2.86786079e-01
1.05997548e-01 3.95465106e-01 2.35107020e-02 -4.17651206e-01
-2.46958479e-01 -4.09404159e-01 -1.33608985e+00 5.41622519e-01
5.61120212e-01 -6.87854886e-01 -4.14920360e-01 1.51706621e-01
5.07148683e-01 -3.98957402e-01 -1.02532446e+00 4.43810552e-01
3.53813648e-01 -1.48832035e+00 1.36866331e+00 5.32439863e-03
2.90624768e-01 -7.44191527e-01 6.37768582e-02 -1.44993639e+00
-4.32689309e-01 -4.40267056e-01 -3.99990827e-01 1.08916855e+00
7.12208301e-02 -6.78905368e-01 4.65153962e-01 2.22834334e-01
-5.94899133e-02 -5.09858787e-01 -9.15066123e-01 -8.94679189e-01
-2.69096851e-01 2.39295177e-02 2.24373400e-01 9.22986627e-01
-6.15808010e-01 -2.62536258e-02 -6.59643114e-01 6.42802775e-01
1.23833072e+00 7.76449293e-02 4.97908026e-01 -1.02009296e+00
-4.16258216e-01 -4.06114757e-01 -2.19248295e-01 -8.68753254e-01
8.90383944e-02 -2.72684157e-01 2.07418576e-01 -1.47043943e+00
2.60271668e-01 -3.11342567e-01 -3.00128341e-01 -1.69668142e-02
-4.23350871e-01 3.23426545e-01 -1.63785487e-01 4.29889202e-01
-1.00848049e-01 3.50975782e-01 1.27971840e+00 -7.02469004e-03
-5.10132499e-02 1.02856874e-01 -3.34742934e-01 5.02634943e-01
6.14939332e-01 -1.22551154e-02 -3.30440283e-01 -6.62846938e-02
-1.06974736e-01 9.84872058e-02 8.41642678e-01 -1.21644390e+00
1.09634124e-01 -1.08958129e-02 5.97115755e-01 -3.13931167e-01
5.00458896e-01 -8.84839475e-01 3.63846898e-01 1.11910783e-01
-2.22536206e-01 -5.35792232e-01 1.08577587e-01 6.53794885e-01
-3.68241012e-01 -2.23810822e-01 1.45192337e+00 -2.01996565e-01
-7.60748684e-01 1.48126274e-01 -1.69496268e-01 -4.24909681e-01
8.38124275e-01 -4.88600135e-01 -3.94649595e-01 -2.99552143e-01
-7.20403850e-01 -5.28962553e-01 9.06616926e-01 1.13388419e-01
7.25747049e-01 -9.86842036e-01 -9.37125444e-01 6.42623827e-02
-2.04346344e-01 -5.94873913e-02 7.79637992e-01 1.27949667e+00
-5.19037604e-01 1.10866867e-01 -1.72821403e-01 -4.63015974e-01
-1.14776766e+00 7.07501113e-01 7.26367474e-01 -3.11524540e-01
-7.62389183e-01 7.39648581e-01 3.13407898e-01 -1.03141842e-02
1.15926422e-01 -2.53629416e-01 -3.21800113e-01 -2.42611811e-01
8.13718915e-01 3.39447141e-01 2.86601717e-03 -6.88186765e-01
4.30874638e-02 7.76209176e-01 -7.18485042e-02 -4.26408416e-03
1.19902897e+00 -4.22073334e-01 -4.26880330e-01 5.24071634e-01
8.12360585e-01 1.08429670e-01 -1.59546983e+00 -3.36444616e-01
-8.64126384e-02 -7.89437890e-01 4.01586771e-01 -9.46325004e-01
-9.81577158e-01 6.03283346e-01 8.40116203e-01 3.23572129e-01
1.51158428e+00 -2.63258461e-02 4.73913819e-01 -3.29370826e-01
3.44791412e-01 -9.03484106e-01 -9.62113366e-02 -4.74442402e-03
9.60423589e-01 -1.18741608e+00 1.89752027e-01 -4.63355690e-01
-3.90654773e-01 1.21586072e+00 3.50299925e-01 -1.95492938e-01
3.93735290e-01 1.45732239e-01 8.90154690e-02 1.99445426e-01
-2.63408929e-01 -1.78110197e-01 3.96716863e-01 6.77098036e-01
2.94160664e-01 -1.38360798e-01 -1.07799783e-01 1.32029682e-01
2.60496974e-01 1.73139155e-01 7.10285544e-01 3.64088714e-01
-4.37006831e-01 -6.65368259e-01 -7.80489266e-01 1.96017459e-01
-4.72530186e-01 -1.42901868e-01 8.43687505e-02 5.59550762e-01
-7.49550089e-02 7.85784423e-01 -2.31384411e-01 4.14864682e-02
1.60348937e-01 -3.13283414e-01 7.15450525e-01 -2.63511717e-01
-2.19987482e-01 6.84773456e-03 -1.90499887e-01 -6.54228270e-01
-5.27446091e-01 -5.67543089e-01 -7.35018730e-01 6.74122944e-02
-3.20069522e-01 1.17335506e-01 5.06096065e-01 6.61257625e-01
1.19182251e-01 4.98797178e-01 7.39746332e-01 -1.19978213e+00
-8.74901772e-01 -1.01844931e+00 -1.10549831e+00 2.51623958e-01
7.50473499e-01 -5.60097456e-01 -9.70248640e-01 1.76552862e-01]
|
[11.509187698364258, -2.4688827991485596]
|
41f8514a-674f-4a95-bcae-9bafa2eb7ef8
|
understanding-abuse-a-typology-of-abusive
|
1705.09899
| null |
http://arxiv.org/abs/1705.09899v2
|
http://arxiv.org/pdf/1705.09899v2.pdf
|
Understanding Abuse: A Typology of Abusive Language Detection Subtasks
|
As the body of research on abusive language detection and analysis grows,
there is a need for critical consideration of the relationships between
different subtasks that have been grouped under this label. Based on work on
hate speech, cyberbullying, and online abuse we propose a typology that
captures central similarities and differences between subtasks and we discuss
its implications for data annotation and feature construction. We emphasize the
practical actions that can be taken by researchers to best approach their
abusive language detection subtask of interest.
|
['Thomas Davidson', 'Ingmar Weber', 'Zeerak Waseem', 'Dana Warmsley']
|
2017-05-28
|
understanding-abuse-a-typology-of-abusive-1
|
https://aclanthology.org/W17-3012
|
https://aclanthology.org/W17-3012.pdf
|
ws-2017-8
|
['abuse-detection']
|
['natural-language-processing']
|
[ 4.05698903e-02 -1.07276618e-01 -3.81398529e-01 -4.75338310e-01
-4.65088099e-01 -7.30139792e-01 5.90806842e-01 7.77216494e-01
-5.54103792e-01 2.99784452e-01 8.65379095e-01 -5.03647029e-01
-9.42778885e-02 -5.20463921e-02 1.48038805e-01 -1.95134029e-01
1.52348265e-01 -1.11144155e-01 2.35548895e-02 -2.90729702e-01
6.99863195e-01 6.30895734e-01 -9.08950806e-01 3.63853723e-01
7.29212284e-01 3.03500772e-01 -5.06135941e-01 3.80694181e-01
-3.53385329e-01 1.30431473e+00 -9.60619509e-01 -9.66699839e-01
-1.90253809e-01 -4.58472729e-01 -1.38954794e+00 2.36680001e-01
5.32896101e-01 -7.04392791e-01 -3.94763768e-01 1.18176222e+00
2.95650452e-01 2.49279916e-01 5.95624387e-01 -1.25574291e+00
-6.41695678e-01 6.36516988e-01 -4.00461018e-01 1.15453124e+00
6.40775442e-01 2.57337689e-01 7.84097075e-01 -5.73402822e-01
8.04522932e-01 1.33198678e+00 6.54909313e-01 6.23822033e-01
-1.08838725e+00 -7.84408212e-01 3.74399871e-01 4.93671373e-02
-1.17266440e+00 -8.18983495e-01 7.01979399e-01 -9.70655143e-01
1.13436258e+00 2.04146817e-01 8.19098294e-01 1.43104041e+00
-6.62362352e-02 4.94360656e-01 8.74069154e-01 -3.60021830e-01
-1.81895152e-01 2.54545271e-01 8.39569449e-01 4.98275876e-01
5.74127555e-01 -5.59594154e-01 -6.75988793e-01 -8.64826322e-01
3.12440008e-01 -2.93201089e-01 8.23237076e-02 -1.44324303e-02
-3.22353214e-01 1.29934525e+00 -1.88180376e-02 9.07076776e-01
-2.23840214e-02 -7.69414231e-02 1.17073774e+00 2.49777481e-01
7.96633959e-01 7.98606217e-01 2.76496857e-01 -7.91717947e-01
-7.75979280e-01 4.19348508e-01 7.57935405e-01 3.59665811e-01
2.21088558e-01 1.18464537e-01 8.34052265e-02 1.33158886e+00
1.85820565e-01 -1.57606915e-01 2.92265952e-01 -8.00899923e-01
5.13568580e-01 4.48084772e-01 -9.66192456e-04 -1.26247334e+00
-4.42092299e-01 7.25047439e-02 2.26405337e-01 -3.42009753e-01
3.45704198e-01 -1.11599669e-01 -5.44242203e-01 1.48528445e+00
2.95081764e-01 -2.38323510e-01 -6.13176465e-01 5.28922796e-01
4.60305035e-01 3.26825976e-02 8.22944582e-01 -3.47361356e-01
1.13232076e+00 -2.12510899e-01 -9.85604763e-01 -6.18048012e-01
1.35073924e+00 -8.75357985e-01 9.23514068e-01 2.21958771e-01
-9.49578822e-01 3.09650540e-01 -9.40982282e-01 -5.09383261e-01
-3.91174614e-01 -8.39773238e-01 7.00678349e-01 1.31789744e+00
-6.63109481e-01 4.38057363e-01 -8.52105141e-01 -7.50752151e-01
2.91861117e-01 -2.37887010e-01 -3.33013564e-01 1.30138904e-01
-1.04545605e+00 1.35089123e+00 2.63860732e-01 1.04048088e-01
-6.33243024e-01 -3.26276362e-01 -9.32627797e-01 -4.85825151e-01
4.18358743e-01 5.20141460e-02 1.26252520e+00 -8.66623819e-01
-6.72586858e-01 1.35284555e+00 1.65940542e-02 1.65379047e-01
6.68353364e-02 -4.26102877e-01 -5.88116169e-01 1.96994618e-01
4.32887316e-01 -1.61353067e-01 7.25798965e-01 -9.21819210e-01
-5.06640077e-01 -8.10744047e-01 2.72337765e-01 2.11087897e-01
-1.01268899e+00 1.43809140e+00 3.91458333e-01 -7.13436961e-01
-1.59883127e-01 -7.36860991e-01 3.01131338e-01 -3.11924189e-01
-5.12857676e-01 -5.71812212e-01 1.07172561e+00 -9.83599544e-01
1.88340652e+00 -2.48365617e+00 8.44368711e-02 7.15056583e-02
7.25415587e-01 2.68559694e-01 3.07648212e-01 8.69369388e-01
-4.75476775e-03 8.21857989e-01 -6.45166710e-02 -4.60280478e-01
-1.07643999e-01 1.88323393e-01 -3.68881196e-01 9.26788747e-01
-9.69563946e-02 4.94547129e-01 -9.77769077e-01 -5.82493246e-01
-5.81107177e-02 3.34992558e-01 -2.49179333e-01 1.60885945e-01
3.01470101e-01 2.14021966e-01 -6.16634011e-01 7.52029002e-01
3.28014404e-01 2.30571300e-01 8.87268484e-02 4.71495569e-01
-3.99673790e-01 1.11225796e+00 -4.92607415e-01 1.07129502e+00
-8.63291919e-02 8.73936176e-01 6.48064077e-01 -6.94982409e-01
7.32025504e-01 2.20829472e-01 4.00163919e-01 -1.16451457e-01
2.46188283e-01 2.19779938e-01 3.88484985e-01 -8.93279016e-01
5.61498821e-01 -4.58745480e-01 -2.21713796e-01 7.95845270e-01
2.66678315e-02 -5.67824729e-02 1.18843116e-01 4.05946672e-01
1.18137956e+00 -4.43073213e-01 6.16585612e-01 -2.11200789e-01
2.74922937e-01 1.02365121e-01 4.37303782e-01 6.25239849e-01
-8.78346503e-01 -7.45346695e-02 9.02177453e-01 -3.25640231e-01
-8.67447615e-01 -6.04671240e-01 -4.16639388e-01 1.58380008e+00
-2.58543819e-01 -7.20356762e-01 -7.85330296e-01 -8.55707049e-01
-1.08563025e-02 9.56039727e-01 -7.39282131e-01 -3.17288339e-01
-6.83901310e-01 -6.60508096e-01 1.17822146e+00 2.34062701e-01
-1.10277213e-01 -9.44964051e-01 -7.96384573e-01 1.73335969e-01
-1.68439537e-01 -1.32804620e+00 -1.61423877e-01 -1.93287075e-01
-6.15855217e-01 -1.17942035e+00 -1.18174657e-01 -2.45521069e-01
1.77935228e-01 6.94020331e-01 7.54386961e-01 6.75279915e-01
-2.76492655e-01 6.68155730e-01 -8.00180852e-01 -2.13179827e-01
-8.65403116e-01 1.43791884e-01 -2.89410371e-02 -2.40908265e-01
8.63633156e-01 -5.14413357e-01 -6.58157542e-02 -3.10084790e-01
-9.07688022e-01 -4.26192075e-01 -2.95531094e-01 3.79706860e-01
-4.60530907e-01 -3.92391056e-01 3.23146790e-01 -1.05548406e+00
1.23659611e+00 -1.16196024e+00 2.00985134e-01 -1.06721930e-01
-5.06015480e-01 -8.00693572e-01 2.56811380e-01 -2.48878777e-01
-8.99507523e-01 -7.97293723e-01 -3.56641054e-01 -1.37067184e-01
-4.87706810e-01 4.45918173e-01 2.16930553e-01 -2.47176006e-01
7.93701768e-01 -3.00164878e-01 2.46618435e-01 -6.72293484e-01
1.53231978e-01 9.51355040e-01 -1.96828127e-01 -5.06745636e-01
4.43138897e-01 2.92013645e-01 -5.43131769e-01 -1.42833781e+00
-9.78995621e-01 -7.07182825e-01 -6.55866325e-01 -3.39236438e-01
8.11472356e-01 -4.30704743e-01 -5.57640791e-01 3.37122798e-01
-1.17628503e+00 -1.89878404e-01 3.29950005e-01 1.06374815e-01
-2.42088422e-01 9.20661867e-01 -9.24443364e-01 -1.19209349e+00
-3.35148931e-01 -1.04671454e+00 6.49996877e-01 -3.15247327e-01
-1.13834894e+00 -1.11582339e+00 2.91158199e-01 6.89909279e-01
2.36042246e-01 4.12769318e-01 1.18367302e+00 -1.36657655e+00
4.62617218e-01 -1.49411738e-01 7.39596710e-02 2.32998028e-01
2.87326843e-01 2.87013650e-01 -7.75067687e-01 -7.35842064e-02
2.81752139e-01 -8.58831167e-01 3.25774103e-01 -1.13911323e-01
6.79950476e-01 -6.36974692e-01 -2.17506379e-01 2.66817156e-02
1.04449880e+00 3.24985027e-01 2.00597242e-01 5.70343256e-01
8.73881340e-01 1.04944694e+00 3.84678632e-01 4.76233780e-01
1.73286334e-01 5.96131086e-01 5.44584654e-02 5.92314601e-01
2.05305934e-01 -2.36397177e-01 4.93070811e-01 4.29805785e-01
2.98708588e-01 7.49196857e-02 -1.34415150e+00 7.04129100e-01
-1.53300035e+00 -1.21877849e+00 -2.87252545e-01 1.88557184e+00
6.34523392e-01 2.65043676e-01 8.40522528e-01 1.42813012e-01
6.06484532e-01 6.74443781e-01 -2.19367579e-01 -1.21388197e+00
4.33114916e-01 -1.83800951e-01 -2.61772368e-02 8.49606574e-01
-9.62213635e-01 1.02620804e+00 7.73373365e+00 4.74849433e-01
-1.14581418e+00 5.48785925e-01 5.39409220e-01 -3.82496357e-01
-9.29140523e-02 -4.75764200e-02 -6.75603867e-01 5.56762636e-01
1.00020766e+00 -3.17454994e-01 3.01656693e-01 8.06123376e-01
2.28178710e-01 -2.51050204e-01 -1.00514090e+00 4.76037771e-01
5.08432388e-01 -7.05864191e-01 -3.16614240e-01 7.71586746e-02
-7.12034553e-02 -4.78886403e-02 -2.62392443e-02 1.31001428e-01
3.60648304e-01 -9.63038743e-01 7.17780411e-01 -2.26495057e-01
3.82437408e-01 -5.57049155e-01 3.75465602e-01 2.32076496e-01
-3.92951995e-01 -4.47035551e-01 3.51234414e-02 -5.97532392e-01
3.29926044e-01 2.09359884e-01 -7.20618844e-01 -2.05657378e-01
6.46670222e-01 6.24692023e-01 -5.37774324e-01 6.93427205e-01
-3.30636322e-01 8.96478355e-01 -2.68596765e-02 -2.97314924e-04
4.40681785e-01 -1.54551223e-01 7.54406929e-01 1.52951312e+00
-2.57190406e-01 2.05882877e-01 2.14731276e-01 5.16375184e-01
3.34765136e-01 1.67328879e-01 -1.18055773e+00 -7.10283160e-01
8.68291259e-01 1.25152111e+00 -8.29399645e-01 -1.52730659e-01
-6.06552660e-01 6.80313408e-01 8.35224688e-01 -6.61009748e-04
-4.49133277e-01 -4.46318574e-02 1.09568429e+00 4.83170420e-01
-4.43729460e-01 -5.62139094e-01 -6.16987526e-01 -8.94579530e-01
-2.05588758e-01 -1.15197897e+00 7.11357951e-01 -3.30001384e-01
-1.37072098e+00 3.27687681e-01 5.98165214e-01 -3.40822726e-01
-5.65811515e-01 -3.69543582e-01 -4.09985304e-01 5.88046849e-01
-4.85127777e-01 -1.02863812e+00 5.34019098e-02 1.78487167e-01
4.15090919e-01 4.69888002e-01 7.74437904e-01 1.68417558e-01
-1.04056561e+00 3.70256066e-01 -5.24393022e-01 3.35687727e-01
5.03298223e-01 -6.11706436e-01 4.25428420e-01 1.01000357e+00
-1.53905019e-01 8.50990653e-01 9.96521056e-01 -1.05938184e+00
-8.20545077e-01 -4.31171536e-01 1.10845351e+00 -8.24093103e-01
1.57642102e+00 -5.80647349e-01 -1.24817038e+00 1.14790869e+00
3.03390682e-01 -6.82302535e-01 1.26649642e+00 5.33066034e-01
-7.82747567e-01 5.70619166e-01 -1.26215315e+00 6.60492182e-01
1.36109269e+00 -8.15813363e-01 -7.57434666e-01 7.06888855e-01
2.31919155e-01 -2.68567502e-01 -8.09885800e-01 -3.29080939e-01
4.18204069e-01 -9.62736130e-01 5.64068258e-01 -1.05530214e+00
3.74019653e-01 7.11876869e-01 2.87195593e-01 -9.62298810e-01
-5.57533145e-01 -6.97802961e-01 6.57304078e-02 1.25873041e+00
-8.34342092e-02 -6.00226462e-01 2.49449834e-01 1.36494696e+00
-1.24288024e-02 -5.16326964e-01 -1.15739369e+00 -4.69939291e-01
5.56473792e-01 -5.12411475e-01 4.55217175e-02 1.60821891e+00
9.44155633e-01 3.39695185e-01 -3.00760120e-01 -2.73052216e-01
2.09648579e-01 -6.42400205e-01 5.97204387e-01 -1.13985646e+00
2.27104709e-01 -6.28488719e-01 -4.99697268e-01 -4.52628762e-01
3.98275107e-01 -5.90460300e-01 -4.53883708e-01 -1.12807846e+00
4.96958911e-01 -1.07242428e-01 1.80689067e-01 6.35985434e-01
-1.09501407e-01 2.51755983e-01 4.92629498e-01 2.97242075e-01
-5.78354120e-01 1.25671685e-01 4.74964797e-01 2.13315055e-01
-3.12355101e-01 -5.50570011e-01 -1.27541888e+00 9.57880437e-01
7.43056893e-01 -7.14383006e-01 -1.76031724e-01 -3.55008245e-01
4.00234789e-01 -1.12636290e-01 3.02641064e-01 -4.41907823e-01
-2.66807705e-01 -6.95041418e-01 -2.75183350e-01 1.33389324e-01
4.65854824e-01 -5.31190813e-01 -2.62368500e-01 3.43106747e-01
-5.38527846e-01 2.33773977e-01 1.36424243e-01 6.83367848e-02
4.48198579e-02 -9.28213179e-01 7.79685140e-01 -1.41829178e-01
-3.94739270e-01 -9.74410549e-02 -1.18188584e+00 3.55022788e-01
1.25507367e+00 -3.08277249e-01 -3.83115649e-01 -7.42293775e-01
-7.50961423e-01 -1.00942262e-01 4.64653522e-01 6.00530446e-01
4.06450957e-01 -9.59708035e-01 -5.49628019e-01 -2.49917552e-01
2.33753204e-01 -8.90799880e-01 9.80269164e-02 9.15027499e-01
-9.71199125e-02 3.13133419e-01 -1.19191252e-01 2.31633708e-01
-1.50850821e+00 5.40781021e-01 3.37147087e-01 -5.22117987e-02
-6.04885757e-01 5.71589291e-01 -2.13326111e-01 1.96032584e-01
-3.08430865e-02 4.70786780e-01 -5.02083182e-01 3.98130804e-01
7.58490384e-01 8.90522540e-01 -9.55402032e-02 -1.29961538e+00
-4.36983049e-01 -2.96717197e-01 -7.79870391e-01 -2.83813298e-01
1.22301030e+00 -2.91288793e-01 -6.79351091e-01 8.44572425e-01
1.25850761e+00 2.69370943e-01 -5.04804790e-01 1.73120406e-02
5.51347077e-01 -9.42136884e-01 -1.62515879e-01 -4.99532282e-01
-3.57302666e-01 8.11261833e-01 -9.89598557e-02 1.12097216e+00
6.01359904e-01 2.13248134e-01 6.08812690e-01 -1.29412487e-01
2.05851898e-01 -1.18089283e+00 1.62994981e-01 6.37598515e-01
1.20947003e+00 -7.39773512e-01 2.98346102e-01 -7.95389056e-01
-6.69023216e-01 1.11451364e+00 9.44246471e-01 -1.26095206e-01
6.81339920e-01 2.62935340e-01 1.26340389e-01 -7.35287130e-01
-3.77352864e-01 1.84262142e-01 -2.26466075e-01 5.46578228e-01
1.01579750e+00 -9.97228101e-02 -1.13250947e+00 3.91267151e-01
-1.91358939e-01 -4.38583314e-01 6.85809433e-01 1.20242453e+00
-5.97892642e-01 -9.99067605e-01 -5.20009458e-01 6.59094512e-01
-1.16142976e+00 -2.68549137e-02 -1.30508769e+00 1.08081496e+00
-1.23037718e-01 1.23167169e+00 2.91116815e-02 -3.96627337e-01
5.16577018e-03 4.61357832e-01 2.14016289e-01 -1.10110795e+00
-1.13279796e+00 -1.48769841e-01 8.34178746e-01 -7.03815341e-01
-1.65973902e-01 -9.91402507e-01 -8.11460495e-01 -6.01485014e-01
-2.11572737e-01 -5.13317473e-02 4.28856671e-01 1.35267437e+00
3.52673650e-01 -2.06577331e-01 2.54269063e-01 -2.51347542e-01
-6.11933053e-01 -1.29725015e+00 -5.03962994e-01 7.62819767e-01
5.31044126e-01 -8.40778232e-01 -5.38675427e-01 -5.53783357e-01]
|
[8.649980545043945, 10.41579532623291]
|
bd082050-7a02-4642-87b2-f41ed9756a3f
|
automated-reconstruction-of-3d-open-surfaces
|
2210.15059
| null |
https://arxiv.org/abs/2210.15059v2
|
https://arxiv.org/pdf/2210.15059v2.pdf
|
Automated Reconstruction of 3D Open Surfaces from Sparse Point Clouds
|
Real-world 3D data may contain intricate details defined by salient surface gaps. Automated reconstruction of these open surfaces (e.g., non-watertight meshes) is a challenging problem for environment synthesis in mixed reality applications. Current learning-based implicit techniques can achieve high fidelity on closed-surface reconstruction. However, their dependence on the distinction between the inside and outside of a surface makes them incapable of reconstructing open surfaces. Recently, a new class of implicit functions have shown promise in reconstructing open surfaces by regressing an unsigned distance field. Yet, these methods rely on a discretized representation of the raw data, which loses important surface details and can lead to outliers in the reconstruction. We propose IPVNet, a learning-based implicit model that predicts the unsigned distance between a surface and a query point in 3D space by leveraging both raw point cloud data and its discretized voxel counterpart. Experiments on synthetic and real-world public datasets demonstrates that IPVNet outperforms the state of the art while producing far fewer outliers in the reconstruction.
|
['William J. Beksi', 'Mohammad Samiul Arshad']
|
2022-10-26
| null | null | null | null |
['mixed-reality']
|
['computer-vision']
|
[ 3.69564295e-01 2.27044627e-01 4.68574911e-01 -3.78688008e-01
-9.68918502e-01 -3.25046003e-01 4.81143981e-01 3.94351363e-01
1.42136499e-01 5.24736047e-01 -1.83337957e-01 -9.31732729e-02
-5.02009802e-02 -1.15138972e+00 -1.21104157e+00 -3.82528275e-01
-2.25705385e-01 9.41778779e-01 5.92424214e-01 -3.15466911e-01
1.85633957e-01 8.69083941e-01 -1.85862386e+00 3.88129383e-01
1.11805403e+00 1.02911973e+00 1.36866659e-01 3.03616464e-01
-6.62412286e-01 -1.91718880e-02 -1.98419049e-01 4.37871218e-02
4.78603244e-01 2.72458822e-01 -3.14079225e-01 -1.93594977e-01
9.64386106e-01 -1.78201109e-01 -1.06520183e-01 9.06168163e-01
3.22465926e-01 -3.41337104e-03 7.98654616e-01 -9.20107722e-01
-1.89641550e-01 -3.32737863e-01 -4.15244281e-01 -4.97378170e-01
4.37206328e-01 -4.66342755e-02 6.45855010e-01 -1.42887866e+00
8.08551848e-01 1.19184113e+00 1.37224174e+00 4.51090746e-02
-1.61462569e+00 -6.58216476e-01 1.27990514e-01 -3.10539573e-01
-1.55869961e+00 -3.94682139e-01 1.14051199e+00 -7.27788091e-01
8.52566898e-01 7.18309999e-01 7.12071538e-01 7.22333312e-01
3.59696120e-01 1.59426734e-01 1.13594770e+00 -1.67050343e-02
3.85815620e-01 -2.30850369e-01 -2.97356606e-01 4.45315510e-01
3.29410791e-01 2.81846911e-01 -6.09099507e-01 -6.06788158e-01
1.08666515e+00 8.79367813e-02 -3.57827038e-01 -1.03436494e+00
-1.11991584e+00 5.85852146e-01 6.41221642e-01 -2.74893135e-01
-3.73225689e-01 7.13881627e-02 5.42396866e-02 2.59745538e-01
1.18091595e+00 4.09231961e-01 -6.19607210e-01 2.98440736e-02
-1.04172862e+00 6.07013524e-01 8.13710928e-01 9.01301324e-01
1.33233094e+00 -5.50696626e-02 4.71445262e-01 7.21237540e-01
3.18222255e-01 7.00719178e-01 -4.44714308e-01 -1.05574453e+00
2.73699194e-01 7.83611000e-01 2.98871875e-01 -1.24875236e+00
-3.95760208e-01 -7.50298575e-02 -7.70880103e-01 8.41940820e-01
3.23575020e-01 3.34018558e-01 -1.12066555e+00 1.07651865e+00
8.33411396e-01 6.40421450e-01 -3.81191105e-01 8.75233710e-01
1.00601542e+00 5.43655515e-01 -5.10151863e-01 6.89702854e-02
5.40199339e-01 -4.19024944e-01 -4.55028564e-01 -1.64466575e-01
3.95160913e-01 -7.21212149e-01 1.06221497e+00 4.99492735e-01
-1.04054201e+00 -3.03794801e-01 -8.74951601e-01 -2.28850946e-01
-1.62735656e-01 -6.32368147e-01 5.67544103e-01 1.07865781e-01
-8.23694289e-01 1.01420212e+00 -9.88501847e-01 4.74772602e-02
4.29470748e-01 3.08725327e-01 -4.45300221e-01 -2.65488535e-01
-5.79149663e-01 4.58862513e-01 -2.63163090e-01 1.22172438e-01
-5.92850149e-01 -1.35074520e+00 -1.00956070e+00 -3.53659332e-01
4.60376769e-01 -6.14408910e-01 8.56231093e-01 -7.61372924e-01
-1.30223513e+00 8.37066770e-01 -1.03038825e-01 -1.38597824e-02
9.09605563e-01 -1.75766379e-01 -2.00981110e-01 -9.77563038e-02
1.02137603e-01 3.05293381e-01 8.73570919e-01 -1.78345799e+00
-2.25729555e-01 -5.78162193e-01 -3.67904156e-01 -1.22747058e-02
4.12275523e-01 -7.46100962e-01 -9.26153213e-02 -4.94436294e-01
9.79360223e-01 -6.87246799e-01 -5.04319072e-01 8.27841878e-01
-3.71598482e-01 1.30087167e-01 9.78844941e-01 -8.39865208e-01
6.11444116e-01 -2.15453577e+00 -1.20040867e-02 6.53081298e-01
3.37520748e-01 -1.91886693e-01 3.02327075e-03 3.28598529e-01
2.34523103e-01 3.99007387e-02 -7.40538418e-01 -4.52956647e-01
-1.58305436e-01 5.20988405e-01 -3.41344506e-01 9.20792639e-01
1.17541969e-01 4.39212859e-01 -9.68880713e-01 -2.79716641e-01
5.00832379e-01 8.25421214e-01 -8.20377946e-01 2.78282285e-01
-6.14306927e-01 8.70903254e-01 -4.23431545e-01 8.35618258e-01
1.22227156e+00 1.93920851e-01 -1.92346379e-01 -1.28417760e-01
-4.69929755e-01 3.90866846e-01 -1.42864394e+00 2.02079749e+00
-4.39689964e-01 3.26552570e-01 5.08638084e-01 -4.80036676e-01
1.27887046e+00 1.41875505e-01 6.75742626e-01 -7.35059023e-01
-3.07863444e-01 7.22593725e-01 -4.25129145e-01 -2.94898123e-01
5.44762909e-01 -3.12780499e-01 2.13212833e-01 -1.10036626e-01
-5.91659963e-01 -9.45146143e-01 -8.01264703e-01 -2.18169779e-01
1.14699543e+00 5.74680865e-01 -1.94906816e-01 -3.56083781e-01
1.56255752e-01 1.54917195e-01 7.61458099e-01 4.91899192e-01
3.58873248e-01 1.23647523e+00 2.16957722e-02 -7.81489193e-01
-1.36623275e+00 -1.41620409e+00 -5.75148165e-01 3.10597271e-01
4.96036112e-01 -2.58160532e-01 -4.22836632e-01 -1.05777957e-01
7.38722563e-01 6.00536942e-01 -6.29449725e-01 3.94733734e-02
-8.25638890e-01 -1.30972654e-01 -1.53780505e-02 2.22416475e-01
-5.73768131e-02 -6.78461611e-01 -5.64661980e-01 2.98197061e-01
6.24539740e-02 -9.83700871e-01 -3.21722515e-02 -5.11948392e-02
-1.27301395e+00 -1.13355875e+00 -2.07216382e-01 -4.47102189e-01
7.79958189e-01 2.89838403e-01 1.48016560e+00 2.29735300e-01
-2.29843393e-01 3.19780439e-01 -3.84736657e-02 -4.01044905e-01
-4.45414364e-01 -2.46735066e-01 -1.13108657e-01 1.06458813e-01
-2.47066706e-01 -1.11658001e+00 -4.42342430e-01 4.54218954e-01
-6.74743354e-01 3.96601200e-01 1.94132105e-01 6.04956210e-01
1.50256228e+00 -3.58715087e-01 2.25544721e-01 -9.84048247e-01
-9.05395448e-02 -7.57415533e-01 -6.30620182e-01 -2.76248068e-01
-2.32630178e-01 -3.35873589e-02 3.06482792e-01 -3.51018429e-01
-9.14867163e-01 2.93287247e-01 -2.62189478e-01 -8.06942582e-01
-1.28281146e-01 3.13891560e-01 -1.06766246e-01 -4.62236047e-01
6.78353906e-01 3.81358340e-02 2.03319147e-01 -1.04780829e+00
3.63513604e-02 4.31153655e-01 5.57820678e-01 -7.49412656e-01
9.31058288e-01 1.12990260e+00 3.48680884e-01 -1.29157913e+00
-6.73792303e-01 -4.85473603e-01 -7.90538549e-01 -1.69818208e-01
3.53164226e-01 -8.77420843e-01 -2.94255197e-01 3.27402085e-01
-1.18396962e+00 -5.58930874e-01 -6.78445220e-01 7.13994540e-03
-7.49080598e-01 2.84493655e-01 -2.66657740e-01 -7.60798991e-01
-1.76147148e-01 -9.32977915e-01 1.54972684e+00 -3.81465197e-01
-3.18835318e-01 -7.32974291e-01 3.85027438e-01 1.12345301e-01
2.88846195e-01 1.01283574e+00 8.88394594e-01 2.53601149e-02
-9.26226497e-01 -1.69908971e-01 7.35235512e-02 5.06962324e-03
2.08815843e-01 1.91447034e-01 -9.66873825e-01 -2.36961752e-01
4.63372841e-02 -8.26032925e-03 5.23029685e-01 2.56561816e-01
1.27361941e+00 -1.46357909e-01 -3.85035187e-01 1.24305880e+00
1.44498432e+00 -4.86556619e-01 6.72244608e-01 1.90307513e-01
9.16639447e-01 6.48249805e-01 7.69716740e-01 4.81129915e-01
3.53015423e-01 6.84280455e-01 1.04081428e+00 -3.20840210e-01
-1.53936893e-01 -4.48088527e-01 -2.30333749e-02 7.71603465e-01
-1.82868063e-01 2.77549148e-01 -1.10741389e+00 5.27035892e-01
-1.74566102e+00 -5.14322042e-01 -8.86139214e-01 2.58018327e+00
7.06295371e-01 1.27507031e-01 -4.36258286e-01 1.23716526e-01
2.88089603e-01 3.24193239e-02 -7.34744847e-01 -2.75960386e-01
-2.02959731e-01 5.37157118e-01 5.08781612e-01 7.27273047e-01
-7.63820291e-01 7.25539148e-01 5.71443605e+00 4.89090830e-01
-1.23555458e+00 1.77142978e-01 1.25048503e-01 8.74727145e-02
-8.47813308e-01 -2.59585716e-02 -4.24552411e-01 3.33913982e-01
7.89874136e-01 7.20084310e-02 2.33164415e-01 7.98762918e-01
2.15917379e-01 -1.90896884e-01 -1.11030018e+00 1.07785499e+00
-1.17023855e-01 -1.53726554e+00 1.65372133e-01 8.44282135e-02
8.83207262e-01 4.36402291e-01 -2.12303802e-01 -7.34688416e-02
1.43850353e-02 -1.20889211e+00 1.09250987e+00 8.44588757e-01
1.14452398e+00 -5.41124523e-01 3.48349005e-01 7.39544868e-01
-1.15749872e+00 4.64970440e-01 -4.41725910e-01 -2.84065455e-01
1.76296011e-01 8.30995202e-01 -7.55181551e-01 5.45504868e-01
8.22827876e-01 8.14283550e-01 -2.84483165e-01 1.30646944e+00
1.58696875e-01 3.33623558e-01 -8.37290049e-01 6.48497164e-01
-5.29338792e-02 -5.10755718e-01 9.88661706e-01 7.51500487e-01
4.32471067e-01 1.30784750e-01 4.28519994e-01 1.17484760e+00
6.00305609e-02 1.52911872e-01 -9.19585109e-01 6.20800197e-01
3.82103771e-01 8.86196554e-01 -5.07011890e-01 4.10232730e-02
-3.74395490e-01 6.71385884e-01 3.60006869e-01 2.61682689e-01
-4.42822009e-01 1.17314726e-01 1.07747388e+00 1.08358943e+00
1.27628624e-01 -5.61654568e-01 -7.90028691e-01 -8.91690612e-01
2.30109856e-01 -5.26934505e-01 -1.44198239e-01 -7.29762435e-01
-1.25571692e+00 3.29697520e-01 -2.03281313e-01 -1.40856659e+00
3.30406934e-01 -3.29125404e-01 -3.37377071e-01 9.63025570e-01
-1.67038918e+00 -1.05056787e+00 -7.27819264e-01 4.14222419e-01
4.37813252e-01 6.09708726e-01 9.50953305e-01 2.62724578e-01
2.24284500e-01 4.42254171e-03 3.66899639e-01 -3.78741235e-01
4.65686321e-01 -1.10187232e+00 7.94816256e-01 3.53169143e-01
-9.37415883e-02 1.12767346e-01 8.44890058e-01 -1.14591134e+00
-1.74165583e+00 -1.23831165e+00 2.68068403e-01 -5.68178236e-01
1.16128594e-01 -6.42848194e-01 -1.58921659e+00 6.33468568e-01
-7.35225618e-01 7.24100709e-01 2.00437784e-01 -5.62514812e-02
-2.91339606e-01 3.92723605e-02 -1.53245223e+00 4.14818853e-01
1.32969689e+00 -4.04361576e-01 -4.80954558e-01 1.54389396e-01
7.87141621e-01 -1.05729282e+00 -9.12403107e-01 9.79175806e-01
4.86920953e-01 -1.09432304e+00 1.31147277e+00 -2.94520706e-01
3.32539618e-01 -3.68197709e-01 -3.71386230e-01 -1.26802468e+00
-1.21536009e-01 -6.49856448e-01 -3.92927885e-01 6.78343594e-01
6.33001253e-02 -6.48803115e-01 9.94350195e-01 6.50321662e-01
-7.29705691e-01 -1.04096830e+00 -1.36880028e+00 -7.03638852e-01
3.51708531e-01 -7.27371097e-01 7.98510432e-01 9.76506829e-01
-6.78815961e-01 -3.26622039e-01 4.85574491e-02 6.00808799e-01
1.01267445e+00 2.72505730e-01 1.05770719e+00 -1.89621496e+00
2.59395927e-01 5.72263710e-02 -4.94086355e-01 -9.69025493e-01
-1.21546797e-02 -7.70038247e-01 2.22083688e-01 -1.56950498e+00
-5.60339570e-01 -1.47853768e+00 1.75972730e-01 2.97632795e-02
2.35015377e-01 3.26512009e-01 -3.98762226e-01 2.69550562e-01
-3.86394784e-02 8.41662049e-01 1.32394874e+00 -9.30591747e-02
-4.71634418e-01 -1.22411326e-02 5.26046604e-02 1.16362083e+00
4.99307245e-01 -5.28258264e-01 -1.34078395e-02 -6.79256082e-01
3.75455499e-01 1.57928526e-01 5.08833706e-01 -1.02379823e+00
9.60474685e-02 -4.08482611e-01 2.60482460e-01 -1.01020598e+00
8.13826442e-01 -1.16018021e+00 9.20779467e-01 1.17014639e-01
1.93178251e-01 -1.65103182e-01 1.78327337e-01 6.91297889e-01
4.58088815e-02 1.73050851e-01 7.26313531e-01 -2.34593838e-01
-4.04857159e-01 7.10065544e-01 3.04295927e-01 1.27174333e-01
8.49150062e-01 -5.94833434e-01 2.74126589e-01 1.15962498e-01
-6.11462712e-01 1.92790125e-02 1.34666932e+00 2.62541384e-01
1.12338829e+00 -1.24350214e+00 -7.30382085e-01 6.28163040e-01
1.37891442e-01 1.05851197e+00 3.08415174e-01 5.13936341e-01
-9.04892862e-01 -2.72861153e-01 1.39113411e-01 -1.06211388e+00
-1.04851437e+00 7.34033063e-02 6.38230503e-01 1.93726316e-01
-1.44444120e+00 7.63239861e-01 2.06001177e-01 -1.04469562e+00
8.34933072e-02 -6.85297906e-01 4.32238698e-01 -3.34103048e-01
1.28213763e-01 4.67084348e-01 4.68128085e-01 -8.08384240e-01
-1.52516887e-01 7.85581768e-01 4.61185306e-01 2.98870504e-01
1.71303439e+00 1.59007862e-01 -2.75419772e-01 8.50629985e-01
1.06367373e+00 3.30053210e-01 -1.60694778e+00 -3.70196074e-01
-1.85073286e-01 -1.04690111e+00 2.04137176e-01 -3.36445153e-01
-8.66251826e-01 8.05323005e-01 3.63516212e-01 -1.38808146e-01
3.94515932e-01 -8.63294080e-02 9.86021936e-01 2.31762886e-01
1.08915138e+00 -8.37269068e-01 -2.47787639e-01 3.53054106e-01
1.23054838e+00 -1.20205998e+00 2.42080912e-01 -1.07801247e+00
1.30846336e-01 1.04228330e+00 4.41035837e-01 -4.36722249e-01
8.97142351e-01 4.69312102e-01 -8.38535652e-02 -5.89233518e-01
-3.85113537e-01 4.04649466e-01 1.77854344e-01 6.49094522e-01
-4.33127210e-03 3.16066235e-01 3.62871349e-01 2.72746444e-01
-4.12544221e-01 -2.18543664e-01 3.94239664e-01 1.14160550e+00
-4.69706088e-01 -8.03166986e-01 -7.84250259e-01 7.67592907e-01
5.55150136e-02 1.66187733e-01 -8.32742900e-02 5.80187678e-01
2.74792284e-01 2.63577670e-01 4.74501133e-01 -1.22814693e-01
7.74244428e-01 -2.80817479e-01 4.70105767e-01 -8.60154092e-01
-3.59775215e-01 -1.77030608e-01 5.50465956e-02 -9.81402099e-01
-6.40009344e-02 -8.55028868e-01 -1.43560922e+00 -3.04678917e-01
-1.62425458e-01 5.58153875e-02 7.92728603e-01 5.85748672e-01
5.08469343e-01 1.96959525e-01 5.19271076e-01 -1.64046884e+00
-4.43597406e-01 -5.93145669e-01 -6.50160134e-01 5.86757243e-01
6.91125751e-01 -1.08000398e+00 -5.34706414e-01 -2.28813156e-01]
|
[8.615457534790039, -3.4812443256378174]
|
8f229f5f-89b9-4566-8ba1-eeaa66d65512
|
learning-causally-disentangled
|
2306.01213
| null |
https://arxiv.org/abs/2306.01213v1
|
https://arxiv.org/pdf/2306.01213v1.pdf
|
Learning Causally Disentangled Representations via the Principle of Independent Causal Mechanisms
|
Learning disentangled causal representations is a challenging problem that has gained significant attention recently due to its implications for extracting meaningful information for downstream tasks. In this work, we define a new notion of causal disentanglement from the perspective of independent causal mechanisms. We propose ICM-VAE, a framework for learning causally disentangled representations supervised by causally related observed labels. We model causal mechanisms using learnable flow-based diffeomorphic functions to map noise variables to latent causal variables. Further, to promote the disentanglement of causal factors, we propose a causal disentanglement prior that utilizes the known causal structure to encourage learning a causally factorized distribution in the latent space. Under relatively mild conditions, we provide theoretical results showing the identifiability of causal factors and mechanisms up to permutation and elementwise reparameterization. We empirically demonstrate that our framework induces highly disentangled causal factors, improves interventional robustness, and is compatible with counterfactual generation.
|
['Xintao Wu', 'Feng Chen', 'Yongkai Wu', 'Aneesh Komanduri']
|
2023-06-02
| null | null | null | null |
['disentanglement']
|
['methodology']
|
[ 5.20818710e-01 3.50138754e-01 -7.32094765e-01 -2.80404299e-01
-5.17976105e-01 -8.22523415e-01 1.05244577e+00 -2.00729638e-01
7.16899186e-02 1.27425444e+00 1.12593925e+00 -4.54977065e-01
-8.12507749e-01 -7.63876438e-01 -1.07307410e+00 -6.87884331e-01
-4.99639094e-01 2.98983395e-01 -5.99697709e-01 2.24576250e-01
7.89588094e-02 4.50908452e-01 -1.14751863e+00 1.42738312e-01
9.62672412e-01 -9.78884175e-02 -3.73216510e-01 6.39460206e-01
4.28946227e-01 6.99126959e-01 -1.35170981e-01 -2.63249874e-01
2.40796953e-01 -7.56299973e-01 -7.30506599e-01 -3.56792361e-01
3.56414288e-01 -1.93010718e-01 -5.50512135e-01 6.64285660e-01
6.50478676e-02 -8.92240852e-02 1.10330570e+00 -1.59659445e+00
-7.92782247e-01 9.80750024e-01 -8.54801476e-01 2.74386734e-01
6.12256899e-02 5.79894595e-02 1.53052557e+00 -6.52422965e-01
6.93162262e-01 1.42683852e+00 1.60018906e-01 6.05238616e-01
-1.95702946e+00 -1.11800301e+00 1.96199045e-01 7.33924285e-02
-8.63896966e-01 -2.25703463e-01 6.61602914e-01 -7.72493601e-01
2.52939761e-01 3.29192340e-01 4.24635142e-01 1.54866576e+00
2.79905915e-01 5.46882987e-01 1.17749166e+00 -3.80001903e-01
2.82736838e-01 -6.32585526e-01 4.43790033e-02 7.61205554e-01
7.74179280e-01 5.90031207e-01 -7.86402047e-01 -5.62274814e-01
1.09246778e+00 7.74577074e-03 -5.00306606e-01 -7.51960456e-01
-1.60147452e+00 1.41204154e+00 5.46391547e-01 -6.87535331e-02
-3.94151241e-01 6.60466611e-01 1.37108564e-01 2.12783590e-02
3.13991457e-01 8.47331583e-01 -4.71223235e-01 2.11840063e-01
-4.91393209e-01 3.21219593e-01 4.97632295e-01 5.24376631e-01
4.44636673e-01 -1.42944857e-01 -2.94168442e-01 2.39213526e-01
4.15534288e-01 5.09301662e-01 -9.73135382e-02 -1.08953977e+00
3.20998937e-01 4.75728214e-01 1.90491542e-01 -7.85723805e-01
-2.96642482e-01 -1.37049541e-01 -9.80058670e-01 1.76186368e-01
6.84517086e-01 -4.24058616e-01 -8.34829330e-01 2.42948604e+00
5.72476424e-02 9.30656552e-01 -1.74277857e-01 7.54712582e-01
1.45147502e-01 4.11730498e-01 3.63675863e-01 -4.72271472e-01
1.22024632e+00 -3.42372715e-01 -7.80344546e-01 1.39586002e-01
1.19740278e-01 -3.73639971e-01 9.68838692e-01 8.98530260e-02
-8.18473697e-01 4.08112817e-02 -9.48855698e-01 -2.06787628e-03
3.38831954e-02 -9.74138454e-02 1.22539032e+00 5.09215355e-01
-5.84574997e-01 6.29869521e-01 -8.98833454e-01 9.81395096e-02
7.89358497e-01 2.04197243e-01 -6.01341367e-01 -6.59643859e-02
-1.35949028e+00 5.47699749e-01 2.81100832e-02 -1.35298297e-01
-1.47733605e+00 -1.34173858e+00 -7.22902000e-01 2.66648501e-01
5.00063717e-01 -1.32824278e+00 8.24350715e-01 -2.48997092e-01
-1.05214214e+00 3.72161567e-01 -1.98646322e-01 -2.25971937e-01
5.22003591e-01 -4.22188997e-01 -3.56170796e-02 6.62083998e-02
2.57771403e-01 3.12458724e-01 9.30482626e-01 -1.21990144e+00
-2.12177947e-01 -3.50755960e-01 9.42677632e-02 8.93106218e-03
-2.11246219e-02 -2.28061691e-01 3.71157914e-01 -7.45891035e-01
-5.70620485e-02 -8.39663267e-01 -3.75845760e-01 1.80161428e-02
-7.62935638e-01 8.96953046e-02 3.32675099e-01 -3.51840824e-01
9.27162826e-01 -1.85918701e+00 7.30705976e-01 1.29650190e-01
8.48717034e-01 -6.23139381e-01 -2.08538234e-01 3.30756187e-01
-8.28409791e-01 5.14218807e-01 -4.14104223e-01 1.60268813e-01
-1.03405729e-01 1.48181498e-01 -7.42876053e-01 7.28703082e-01
4.07079548e-01 1.09167314e+00 -1.20506811e+00 -8.72930419e-03
-4.06590514e-02 4.01300490e-01 -8.15302730e-01 3.44551176e-01
-1.97568536e-01 8.46591294e-01 -5.65553248e-01 -1.45789340e-01
3.80759150e-01 -2.81339169e-01 5.08716643e-01 -1.62799746e-01
4.56295684e-02 4.46416050e-01 -1.01743686e+00 1.62575662e+00
-5.12659550e-01 6.01855159e-01 -3.33367914e-01 -8.63339603e-01
4.71704483e-01 6.06791556e-01 5.39256692e-01 8.14276859e-02
9.16603133e-02 -9.12852064e-02 1.47375748e-01 -4.47337776e-01
3.74458320e-02 -5.05121529e-01 -2.04818785e-01 7.82071590e-01
1.02998771e-01 2.81620473e-01 -1.27920806e-01 5.57972074e-01
1.25545824e+00 2.97737330e-01 7.04561472e-01 -4.84896630e-01
-1.11082874e-01 -2.34608009e-01 6.86700284e-01 6.76834583e-01
1.94888979e-01 4.55509216e-01 1.13020658e+00 -1.78813115e-01
-1.20185757e+00 -1.76205373e+00 -1.40011400e-01 6.66696370e-01
-8.65122750e-02 -2.77779669e-01 -3.94580394e-01 -7.65590250e-01
1.45229787e-01 9.37490880e-01 -1.16477454e+00 -4.34424400e-01
-7.28838503e-01 -1.17063677e+00 5.72689712e-01 6.41696513e-01
-1.49149373e-01 -5.14194906e-01 -3.77076089e-01 -3.18294503e-02
-2.07281038e-01 -4.61107343e-01 -4.67579722e-01 1.64661765e-01
-8.66483927e-01 -1.54497874e+00 -6.32213235e-01 -3.84672149e-03
7.68843174e-01 2.11979046e-01 8.07205439e-01 -4.88807440e-01
-3.68530720e-01 1.49562657e-01 2.28410199e-01 -1.99939143e-02
-3.80183995e-01 -2.72438824e-01 2.02437535e-01 2.56915372e-02
1.02924183e-02 -9.44510400e-01 -6.89377069e-01 2.08259448e-01
-6.71421707e-01 3.56815130e-01 3.86506826e-01 9.91746128e-01
1.81512728e-01 -2.12536842e-01 8.31597328e-01 -1.04113972e+00
6.99756384e-01 -8.04340124e-01 -5.97963393e-01 1.11714445e-01
-5.71434319e-01 7.86633193e-01 4.26469415e-01 -4.40962791e-01
-1.48314142e+00 -2.44412333e-01 5.44689476e-01 -3.44548941e-01
-2.39238292e-01 3.41461033e-01 -5.63057959e-01 6.08904183e-01
9.01584089e-01 -4.82578248e-01 -2.16902152e-01 -3.17967445e-01
1.35176575e+00 -1.24957427e-01 4.13684338e-01 -9.45529521e-01
9.03608620e-01 7.27045238e-01 4.30235863e-01 -1.35237053e-01
-7.50233412e-01 -2.15025768e-02 -6.42432570e-01 1.04111679e-01
1.07369375e+00 -8.83790910e-01 -1.02659237e+00 -2.56946832e-01
-1.17340362e+00 -1.06894016e-01 -4.51820135e-01 7.86104620e-01
-7.76089907e-01 -1.28957778e-01 -4.66270953e-01 -6.26857877e-01
2.77900934e-01 -9.10442889e-01 8.25248659e-01 -2.04507187e-02
-6.28221810e-01 -1.10740829e+00 5.51873684e-01 1.01763800e-01
-1.32965147e-01 6.66773379e-01 1.45086086e+00 -2.62048066e-01
-7.54373074e-01 1.88494414e-01 -4.10818964e-01 -4.45219427e-01
2.93659270e-01 -8.53629559e-02 -1.04664469e+00 1.09514594e-01
-4.15918350e-01 5.91089251e-03 1.21965361e+00 8.25916708e-01
9.27243173e-01 -6.23104393e-01 -6.59470141e-01 3.39506060e-01
1.00296378e+00 -5.69775924e-02 5.15374839e-01 -2.95921743e-01
9.52676117e-01 8.00353706e-01 -1.02858722e-01 2.70015091e-01
1.02825224e-01 3.71725798e-01 3.04527998e-01 -5.39060915e-03
-1.09830424e-01 -9.43284214e-01 4.45040800e-02 2.81883895e-01
-3.08216602e-01 -2.34109670e-01 -4.84193116e-01 5.60273290e-01
-1.88604093e+00 -1.19036698e+00 -4.52507377e-01 2.12011528e+00
9.82688725e-01 -2.21186385e-01 9.76781398e-02 -1.02912858e-01
7.98938155e-01 1.75054930e-02 -6.19803965e-01 -7.78155448e-03
7.19974283e-03 4.06170338e-01 5.66007733e-01 8.33320498e-01
-7.85480082e-01 5.25515139e-01 6.53077269e+00 4.00424272e-01
-5.24491429e-01 2.91822553e-01 6.26439571e-01 -2.54592210e-01
-1.09349263e+00 3.68776560e-01 -3.43729883e-01 1.98518530e-01
6.46629751e-01 -6.14541531e-01 3.85259271e-01 3.97059977e-01
7.38468409e-01 1.99998006e-01 -1.54440963e+00 5.43831348e-01
-4.84818935e-01 -1.64022756e+00 1.64500609e-01 3.98813218e-01
1.09053302e+00 -4.96406049e-01 7.31451362e-02 -2.65751719e-01
1.22132134e+00 -1.32737803e+00 3.83033454e-01 6.38338625e-01
9.66530323e-01 -6.54142201e-01 1.43621027e-01 -3.01462878e-02
-7.06941485e-01 -2.35833507e-02 -1.00360662e-01 -2.93634206e-01
2.74716169e-01 8.74059677e-01 -9.87001300e-01 6.12791598e-01
-1.81267243e-02 7.66068399e-01 1.28789648e-01 7.47437358e-01
-1.04566216e+00 1.04280555e+00 4.17459756e-02 2.44840249e-01
-2.89895594e-01 -1.24795556e-01 4.89543676e-01 7.79748201e-01
1.08916856e-01 1.77457035e-01 -3.24792176e-01 1.53761089e+00
-2.57649004e-01 -1.79412648e-01 -7.80776143e-01 -9.89277139e-02
4.63461936e-01 1.00573647e+00 -5.88802576e-01 -7.98179880e-02
-1.60671592e-01 7.96049654e-01 3.05768162e-01 6.59591138e-01
-8.10332656e-01 1.47763759e-01 1.11489427e+00 -7.98310712e-02
-1.66912496e-01 -2.05943391e-01 -6.69747829e-01 -1.57634819e+00
-4.55321342e-01 -3.56499016e-01 5.02655029e-01 -6.67257845e-01
-1.53816426e+00 8.48451555e-02 4.88760710e-01 -7.66059577e-01
-3.44605207e-01 -3.95950824e-01 -8.19116056e-01 1.16760504e+00
-1.12674487e+00 -1.05088234e+00 1.09652013e-01 3.41641217e-01
2.79265255e-01 1.59562141e-01 9.03097391e-01 -3.87039706e-02
-5.76758921e-01 1.95920557e-01 -1.38510033e-01 -1.45993069e-01
6.62949145e-01 -1.41513777e+00 3.24767232e-01 9.93286133e-01
2.20436558e-01 1.07979000e+00 9.06560838e-01 -8.89885247e-01
-1.17318535e+00 -1.01018226e+00 6.77674115e-01 -6.26198113e-01
1.01832330e+00 -5.46159923e-01 -6.89276636e-01 1.00025380e+00
2.00400501e-01 -1.24284886e-01 9.23495173e-01 5.91498613e-01
-8.72939765e-01 2.92060733e-01 -8.59542489e-01 9.03221846e-01
1.40919590e+00 -4.18438792e-01 -7.39826918e-01 1.27408683e-01
1.04845524e+00 2.45358139e-01 -6.36517882e-01 2.77149260e-01
6.86905921e-01 -4.97592956e-01 1.07156396e+00 -1.19434738e+00
1.00425971e+00 -2.56708264e-01 2.42885444e-02 -1.76377153e+00
-6.15329802e-01 -7.86964118e-01 -3.40305679e-02 1.02774143e+00
5.59985042e-01 -6.70298874e-01 6.11716092e-01 4.81049418e-01
3.60878438e-01 -3.59886289e-01 -8.59312177e-01 -4.37737584e-01
4.43260103e-01 -4.39203322e-01 5.52475512e-01 1.50614095e+00
2.33673781e-01 6.65763021e-01 -6.05081320e-01 4.09006536e-01
8.85763168e-01 3.54893893e-01 5.06019533e-01 -1.28866220e+00
-6.43266857e-01 -3.93868208e-01 -1.54578900e-02 -6.87026620e-01
3.59674543e-01 -1.06625462e+00 -2.60993987e-01 -1.26959383e+00
7.49936640e-01 -4.10006732e-01 -1.08337462e-01 5.02063096e-01
-5.51958084e-01 -6.15480803e-02 4.02603261e-02 4.17029597e-02
2.04342559e-01 6.87940359e-01 1.33864975e+00 1.00363023e-03
7.67134968e-03 -2.20002979e-01 -1.05439317e+00 6.68446541e-01
6.30452812e-01 -7.32670128e-01 -7.38596082e-01 -1.29447967e-01
5.03698170e-01 3.30849379e-01 8.81255329e-01 -5.34254238e-02
-3.30167234e-01 -6.89397275e-01 2.93798000e-01 4.11516316e-02
-5.33435382e-02 -3.51930737e-01 4.01921421e-01 3.85706723e-01
-9.95270491e-01 -2.29346082e-01 2.19544545e-02 9.01159942e-01
2.90116251e-01 7.72180269e-03 4.63763356e-01 1.28746152e-01
-2.15444863e-01 2.64927685e-01 -3.61675292e-01 2.02367768e-01
9.04413402e-01 5.36062896e-01 -6.27416492e-01 -3.05115700e-01
-8.24659407e-01 2.52538454e-02 -1.97907109e-02 4.80266064e-01
6.08658373e-01 -1.43655276e+00 -9.31028605e-01 8.32138509e-02
-3.95940877e-02 -3.94277781e-01 2.30209619e-01 5.84710479e-01
8.56561288e-02 2.74417549e-01 -1.73475340e-01 -1.54823720e-01
-8.38808298e-01 7.30501711e-01 5.27335703e-02 -2.38944203e-01
-4.99483168e-01 6.87702060e-01 1.10670924e+00 -2.64014989e-01
-3.88173699e-01 -2.02647567e-01 -6.87890798e-02 1.53315654e-02
4.77562904e-01 6.44720018e-01 -5.54512382e-01 -5.85567802e-02
-1.16657689e-01 2.10247084e-01 9.41242427e-02 -5.06188691e-01
1.35940409e+00 -7.99879152e-03 -9.89023075e-02 4.12304699e-01
1.14956033e+00 3.97140503e-01 -1.53459978e+00 1.47078604e-01
-3.33461054e-02 -6.08504415e-01 -1.01871252e-01 -8.81151974e-01
-8.06835592e-01 8.91537786e-01 2.11439654e-01 6.21730573e-02
7.31077194e-01 2.95020401e-01 -3.17775719e-02 -3.15927774e-01
1.58379167e-01 -6.41508866e-03 1.43759295e-01 -1.62724435e-01
1.09434736e+00 -6.54970467e-01 8.86454433e-02 -8.26005101e-01
-2.30718225e-01 9.02592361e-01 3.05380195e-01 -3.94510776e-01
5.50235093e-01 3.72806251e-01 -3.87075156e-01 -4.14801598e-01
-9.48693514e-01 1.38844669e-01 4.07025218e-01 6.23888552e-01
5.48046410e-01 6.39896572e-01 -5.12292266e-01 5.55507779e-01
-2.19463095e-01 -5.52076586e-02 8.34578454e-01 1.70763001e-01
8.87019485e-02 -1.07898605e+00 -3.97542745e-01 4.09835964e-01
-2.58968174e-01 -2.47111216e-01 -1.85328037e-01 8.55078161e-01
-3.80957313e-02 7.84867883e-01 1.15653090e-01 8.50107968e-02
8.11700448e-02 -5.41766025e-02 6.26494110e-01 -6.31519437e-01
2.08642319e-01 1.37286991e-01 -3.92504083e-03 -6.13558054e-01
-3.54202360e-01 -8.99910450e-01 -1.16397190e+00 -2.81882524e-01
-2.48761848e-01 8.18897262e-02 3.20421159e-01 1.04363763e+00
5.42139649e-01 8.74806464e-01 5.53315282e-01 -4.45553660e-01
-3.70693117e-01 -7.32874572e-01 -3.75620157e-01 4.43121403e-01
4.51172978e-01 -1.10775959e+00 -4.55301911e-01 3.12148631e-01]
|
[8.032697677612305, 5.368592262268066]
|
b56d5474-c6a9-4011-ad75-71587ebaf83a
|
human-head-pose-estimation-by-facial-features
|
1510.02774
| null |
http://arxiv.org/abs/1510.02774v1
|
http://arxiv.org/pdf/1510.02774v1.pdf
|
Human Head Pose Estimation by Facial Features Location
|
We describe a method for estimating human head pose in a color image that
contains enough of information to locate the head silhouette and detect
non-trivial color edges of individual facial features. The method works by
spotting the human head on an arbitrary background, extracting the head
outline, and locating facial features necessary to describe the head
orientation in the 3D space. It is robust enough to work with both color and
gray-level images featuring quasi-frontal views of a human head under variable
lighting conditions.
|
['Eugene Borovikov']
|
2015-10-09
| null | null | null | null |
['head-pose-estimation']
|
['computer-vision']
|
[-1.47152573e-01 1.17937200e-01 5.21132909e-02 -3.88002783e-01
-4.63947564e-01 -5.49760163e-01 1.34262145e-01 -2.57245183e-01
-5.58535993e-01 1.73753768e-01 -7.48791993e-02 1.34658039e-01
6.07651293e-01 -1.82451278e-01 -2.78122544e-01 -6.59276426e-01
-3.72425228e-01 4.67024565e-01 3.12545031e-01 3.81966829e-02
1.33249283e-01 1.31025124e+00 -1.66357744e+00 -4.47809905e-01
-1.41915813e-01 6.02297604e-01 -2.27433905e-01 7.61131823e-01
3.93445551e-01 2.44999006e-01 -5.66866815e-01 -3.68573904e-01
1.78290620e-01 -4.43850964e-01 -5.43435991e-01 9.06941712e-01
7.07109869e-01 -5.08011222e-01 -1.20657220e-01 1.06648636e+00
6.36978507e-01 -1.98149160e-01 6.66790426e-01 -1.27131629e+00
-5.78129515e-02 -5.90134263e-01 -1.10998237e+00 1.56836078e-01
1.10688055e+00 -7.57087469e-02 4.27503824e-01 -1.09743738e+00
7.71262884e-01 1.42256141e+00 6.16504729e-01 8.69713664e-01
-9.57538068e-01 -6.60134673e-01 6.81654885e-02 -6.09256094e-03
-1.81927812e+00 -7.56218791e-01 7.64830589e-01 -3.24452937e-01
5.01886487e-01 2.68326253e-01 1.02879477e+00 2.22786576e-01
3.04419696e-01 6.72225654e-01 1.37130237e+00 -7.16509104e-01
1.94609553e-01 -1.41659632e-01 7.20494241e-02 1.44364154e+00
4.09611970e-01 -1.62328817e-02 -4.94087577e-01 -4.00348037e-01
7.69071400e-01 -1.56079859e-01 -3.31713557e-01 -4.26354825e-01
-5.21489024e-01 3.29611212e-01 1.39359877e-01 2.54163086e-01
-3.16879630e-01 -7.60376081e-02 -4.08571213e-02 -3.93590450e-01
3.21216136e-01 -3.06234479e-01 5.63907623e-02 1.34542093e-01
-1.09945130e+00 7.35987201e-02 7.31794775e-01 1.09998608e+00
7.65836537e-01 6.62273169e-02 4.61122155e-01 2.83061385e-01
5.66515207e-01 8.03501248e-01 1.28928274e-01 -1.02776325e+00
-1.99706241e-01 4.05886233e-01 1.14863738e-01 -6.23550415e-01
-9.09816682e-01 5.21367550e-01 -4.13029864e-02 7.76704729e-01
6.87214255e-01 -4.36904579e-01 -1.13519859e+00 1.35652733e+00
9.84438121e-01 -3.17143112e-01 -4.31195527e-01 9.11578476e-01
8.34583640e-01 1.12639733e-01 -9.71146822e-02 -3.13136786e-01
1.87530041e+00 -2.91893065e-01 -8.39939773e-01 -6.38800085e-01
1.21862300e-01 -6.71719253e-01 4.92539197e-01 1.96445525e-01
-1.31416547e+00 -1.76910400e-01 -9.04199302e-01 -5.49085736e-02
-4.08009768e-01 1.46886781e-01 1.05799302e-01 1.05840719e+00
-1.41861224e+00 4.83716764e-02 -9.08837795e-01 -4.72585440e-01
-1.10249817e-01 7.34970808e-01 -9.80286837e-01 -8.97833332e-02
-5.78792930e-01 1.10432243e+00 -6.44753948e-02 4.25044060e-01
-2.54920483e-01 3.23311329e-01 -1.24158120e+00 -2.99740344e-01
-8.25942308e-02 -3.42019528e-01 1.23734450e+00 -9.18664157e-01
-1.54602194e+00 1.56363893e+00 -7.16629446e-01 4.04977411e-01
4.60775733e-01 7.55663812e-02 -4.79821593e-01 5.69479764e-01
-9.05290470e-02 5.38530529e-01 1.25867379e+00 -1.20522320e+00
-5.30258298e-01 -1.26723158e+00 -6.31260753e-01 2.76374251e-01
3.65281850e-01 8.91903341e-01 -8.74402165e-01 -1.70657501e-01
5.33073008e-01 -1.08675408e+00 9.22093540e-02 1.71598747e-01
-4.78011489e-01 -2.61792034e-01 1.14299178e+00 -7.87758589e-01
8.35541368e-01 -2.14706945e+00 -3.54528308e-01 4.64811593e-01
2.69242525e-01 -7.82569945e-02 1.35336712e-01 -2.42714554e-01
-1.86500415e-01 -3.68163764e-01 1.25139549e-01 -3.83257627e-01
-7.18467236e-02 1.28292441e-01 6.43951058e-01 1.38681471e+00
-8.94361734e-02 6.04983330e-01 -7.21438527e-01 -1.05639267e+00
-3.70687172e-02 8.31927717e-01 -1.22008547e-01 2.31147215e-01
7.18633711e-01 -5.15906401e-02 -3.89198244e-01 1.08042014e+00
9.17030573e-01 2.70661354e-01 2.35833481e-01 7.20677823e-02
-1.21916428e-01 -1.26863658e-01 -1.32947147e+00 9.02424216e-01
2.02229157e-01 7.55020142e-01 8.89541328e-01 -5.58815189e-02
6.59347415e-01 4.36246544e-01 2.78973728e-01 -2.24273086e-01
3.74187201e-01 3.06665022e-02 -3.08238536e-01 -4.52803373e-01
1.26562864e-01 -6.01168454e-01 1.09336369e-01 7.07744241e-01
-9.52923149e-02 -1.72269806e-01 -5.55236032e-03 -2.01912031e-01
4.84132141e-01 -5.01263849e-02 4.14486408e-01 -2.32638791e-01
5.10099292e-01 -6.22994065e-01 2.79324710e-01 -1.17540350e-02
-5.78666925e-01 7.66545415e-01 2.69986778e-01 -3.82222325e-01
-6.33768320e-01 -8.79828751e-01 -2.34759584e-01 1.22397768e+00
1.30347759e-02 2.27717683e-02 -1.28398645e+00 -5.99447012e-01
6.34052381e-02 1.55595347e-01 -8.56071472e-01 4.29234654e-01
-6.03560567e-01 -4.72016811e-01 2.14611918e-01 4.96596843e-01
1.60540324e-02 -7.86726892e-01 -1.02968907e+00 -2.79462934e-01
2.02253327e-01 -1.03031600e+00 -9.55969155e-01 1.52874202e-01
-5.85385621e-01 -1.28018653e+00 -1.07237840e+00 -1.08414054e+00
1.46935713e+00 2.64354527e-01 9.15819466e-01 1.52666405e-01
-6.69204652e-01 7.99535871e-01 1.10133015e-01 -4.07691777e-01
4.85831685e-02 -6.08445525e-01 2.98234254e-01 9.02721286e-02
5.86219847e-01 5.27072214e-02 -5.92714012e-01 4.05784965e-01
-4.04428452e-01 -3.91018629e-01 1.66609455e-02 2.64797330e-01
2.94974536e-01 -8.64362195e-02 -2.80324131e-01 -4.03155506e-01
3.74719977e-01 1.73467010e-01 -5.49332023e-01 4.12616104e-01
-1.74113646e-01 -3.26779485e-01 8.02692324e-02 -1.00086987e-01
-7.79142380e-01 5.91136396e-01 3.00990772e-02 -1.16661593e-01
-5.34514487e-01 -3.51618141e-01 -2.60016114e-01 -3.46591681e-01
3.41848940e-01 8.69603455e-02 2.77626723e-01 -3.63690168e-01
3.16664547e-01 5.66405237e-01 7.27229238e-01 -2.04654619e-01
8.03973615e-01 8.50303531e-01 2.10528538e-01 -1.08834541e+00
-3.85170579e-01 -6.86254859e-01 -1.58604181e+00 -4.83553469e-01
8.75864148e-01 -5.73419273e-01 -9.08551872e-01 4.44029123e-01
-1.10491145e+00 1.16331264e-01 3.11632335e-01 3.72709841e-01
-4.08145934e-01 4.32194471e-01 -5.20804107e-01 -1.20201349e+00
-3.95575523e-01 -9.46390092e-01 1.52571857e+00 3.71441960e-01
-1.31058425e-01 -1.12356985e+00 -9.42212716e-02 -1.53356344e-01
-2.26652652e-01 3.40887427e-01 5.23233116e-01 -1.05967909e-01
4.60561439e-02 -9.63347733e-01 3.58901657e-02 -2.24377766e-01
2.05664903e-01 4.78664070e-01 -1.20293105e+00 -3.62451434e-01
-1.11195460e-01 -2.34377477e-02 2.50062674e-01 6.22345984e-01
5.10132611e-01 -1.95379689e-01 -5.24542093e-01 3.64821643e-01
9.75601077e-01 4.31583822e-01 3.33212376e-01 2.11923242e-01
3.97703528e-01 1.01365733e+00 1.08540110e-01 4.13719386e-01
5.14050364e-01 5.49446821e-01 5.61676994e-02 -6.82205319e-01
-7.55986124e-02 -1.78941245e-05 4.54750061e-01 1.64525285e-01
-1.58979997e-01 4.17606294e-01 -8.66436660e-01 2.88501382e-01
-9.81573164e-01 -6.62325382e-01 7.96789899e-02 2.23332667e+00
5.93762279e-01 -1.33820906e-01 9.05627549e-01 2.75383025e-01
1.13750720e+00 -2.03343585e-01 -3.98309499e-01 -3.97937894e-01
1.76692575e-01 2.81239524e-02 4.98320878e-01 5.90852141e-01
-1.01118040e+00 7.22332656e-01 8.64984035e+00 -1.28530830e-01
-1.23125899e+00 -2.34843910e-01 4.95785385e-01 -3.36484760e-02
1.76844627e-01 -3.75411838e-01 -1.14836872e+00 1.09108314e-01
5.26875854e-01 -1.88604861e-01 -7.80768096e-02 8.93415749e-01
7.08096176e-02 -4.60542321e-01 -9.38016593e-01 1.00117576e+00
5.38066685e-01 -3.34295690e-01 -4.72110122e-01 1.52772173e-01
1.30301312e-01 -3.23082179e-01 2.87286006e-02 -3.23494852e-01
3.89888883e-02 -1.01721084e+00 7.68421888e-01 1.97032928e-01
1.06720877e+00 -6.64548993e-01 4.18665767e-01 5.73132075e-02
-1.35267889e+00 2.79437989e-01 -4.66041155e-02 5.94930090e-02
6.02122620e-02 -2.23827749e-01 -1.05484927e+00 -3.47145289e-01
6.42987609e-01 9.13012251e-02 -6.56139314e-01 1.16735053e+00
-3.21605027e-01 -3.38624790e-02 -6.64997756e-01 1.05256744e-01
1.70649022e-01 -1.23661399e-01 2.37305984e-01 1.48223639e+00
1.31120965e-01 4.50123161e-01 2.26984508e-02 3.29931915e-01
2.09111378e-01 3.16145450e-01 -7.08223343e-01 4.85977739e-01
1.09758914e-01 1.39231980e+00 -1.11274779e+00 -9.13334340e-02
-6.73099399e-01 1.04313397e+00 -1.17861636e-01 4.85249579e-01
-2.87820399e-01 -5.07495165e-01 4.00970578e-01 5.15567541e-01
7.81734437e-02 -2.34990686e-01 1.26525715e-01 -7.48151600e-01
-4.34554517e-02 -3.72891635e-01 4.93514121e-01 -1.00843072e+00
-5.64926028e-01 7.77472794e-01 9.63256359e-02 -9.55832899e-01
-4.89675641e-01 -9.80711162e-01 -7.24539757e-01 9.50771153e-01
-1.20054030e+00 -1.06980813e+00 -4.37738806e-01 8.47981572e-01
1.61834538e-01 3.02139312e-01 8.04631531e-01 -2.21136257e-01
-7.28497624e-01 6.87755167e-01 -3.16417813e-01 3.71605963e-01
5.71902275e-01 -1.34497690e+00 4.48032409e-01 6.96718454e-01
-2.42990628e-01 5.27219236e-01 7.79305220e-01 -4.17380095e-01
-1.48636830e+00 -3.99900496e-01 1.15858400e+00 -5.47206044e-01
2.14099109e-01 -3.81976157e-01 -5.76843560e-01 7.78405964e-01
1.51408210e-01 4.29336399e-01 6.73268020e-01 -5.28386161e-02
-2.99690366e-01 3.13832089e-02 -1.40145373e+00 3.64564925e-01
4.30965304e-01 -4.86987889e-01 -6.15781486e-01 2.87077308e-01
-5.89883208e-01 -6.79407418e-01 -5.62689006e-01 -5.91842877e-03
1.15279520e+00 -8.85489225e-01 8.15265536e-01 -3.36750209e-01
-6.81488156e-01 -3.54638696e-01 3.42848808e-01 -8.83245826e-01
-2.17663288e-01 -7.23268390e-01 3.60921383e-01 5.83505034e-01
2.20022142e-01 -4.99297947e-01 9.23052430e-01 1.23767316e+00
3.75415236e-01 -2.99814939e-01 -1.15291667e+00 -5.18275976e-01
-4.18708950e-01 1.43235013e-01 1.76448643e-01 3.24274898e-01
3.80886346e-01 1.01255644e-02 9.46100652e-02 1.80724636e-01
9.50523376e-01 5.67693263e-02 6.40111029e-01 -1.26594675e+00
5.38812280e-01 -5.35319746e-01 -8.57477844e-01 -5.60788274e-01
3.62547129e-01 -2.04059198e-01 6.55949637e-02 -1.24851382e+00
3.03603172e-01 3.04092497e-01 -3.86766084e-02 7.05869019e-01
-1.03219561e-01 4.66311783e-01 5.44054545e-02 -9.72176716e-02
-2.67548352e-01 -2.25130722e-01 1.05195546e+00 1.80210665e-01
-1.70956925e-02 2.16526166e-01 -2.84762025e-01 1.24344349e+00
4.04580444e-01 -3.27120811e-01 1.16202429e-01 6.05973639e-02
-5.18549323e-01 1.71894968e-01 1.52446523e-01 -7.53576040e-01
3.07758152e-01 -7.75528476e-02 1.16655588e+00 -7.89168596e-01
5.92632055e-01 -9.02304590e-01 -6.84004948e-02 5.50360799e-01
1.15688056e-01 6.65309489e-01 1.99607149e-01 1.18362613e-01
5.73076792e-02 -3.23788583e-01 1.33594155e+00 -4.25847918e-01
-6.79245412e-01 1.18527360e-01 -5.75209618e-01 -2.09382579e-01
1.09727335e+00 -6.07986569e-01 1.84409469e-02 -5.29370427e-01
-9.26651895e-01 -9.68428552e-02 1.08771098e+00 -4.48117182e-02
8.61478388e-01 -1.04640317e+00 -4.74317968e-01 9.53799784e-01
-1.63851827e-02 -6.07058644e-01 -1.54582635e-01 9.00334239e-01
-8.96898150e-01 3.07943702e-01 -3.14409435e-01 -6.42710805e-01
-2.05182028e+00 5.59769928e-01 6.96234822e-01 6.38852596e-01
-7.33513296e-01 9.52555954e-01 1.51793137e-01 9.26561281e-02
4.58895445e-01 -1.99897755e-02 -2.72332072e-01 1.55563876e-01
9.60911155e-01 3.46931815e-01 1.31210610e-01 -1.58511734e+00
-7.94022918e-01 1.17652273e+00 2.07859859e-01 -3.67992193e-01
7.51407146e-01 -3.53351802e-01 -2.14389086e-01 1.93938628e-01
1.52659798e+00 2.68992454e-01 -1.13192821e+00 -1.59757975e-02
1.29149007e-02 -6.73533738e-01 1.06428683e-01 -3.30071867e-01
-1.17864561e+00 1.08874822e+00 9.18517411e-01 6.00320064e-02
1.22351539e+00 1.52976170e-01 3.63177180e-01 -4.88901585e-02
5.00811696e-01 -1.08605063e+00 -1.32462978e-01 1.03088103e-01
6.41705692e-01 -9.95070159e-01 3.01666260e-01 -4.67206925e-01
-6.33016348e-01 1.40941703e+00 4.57567096e-01 1.77164182e-01
9.50736880e-01 5.93612254e-01 5.34619272e-01 -4.91360873e-01
-2.64845788e-01 -7.65901208e-01 6.99356198e-01 7.16828227e-01
6.72887385e-01 1.77771926e-01 5.96673824e-02 -3.65811169e-01
-1.23381384e-01 -2.88870007e-01 4.03094232e-01 1.30801535e+00
-7.94800222e-01 -6.03311360e-01 -9.55321848e-01 -9.01811868e-02
-6.81690753e-01 3.75958413e-01 -6.73321247e-01 1.03571379e+00
-2.77278684e-02 7.89556086e-01 1.61221951e-01 -5.93848564e-02
5.85842371e-01 4.36383575e-01 9.42110658e-01 -4.94957775e-01
-6.23377264e-02 8.65014672e-01 -5.19797578e-02 -5.10405779e-01
-3.76388341e-01 -9.13648963e-01 -1.32903016e+00 -2.03977302e-01
-1.85014129e-01 1.06088839e-01 7.89702058e-01 6.75493717e-01
-2.92559296e-01 -4.64414924e-01 7.75357306e-01 -1.42549729e+00
1.28602460e-01 -6.41246974e-01 -1.42634833e+00 2.22555861e-01
7.18069553e-01 -8.44475627e-01 -6.52850628e-01 1.65734082e-01]
|
[13.570521354675293, 0.24519406259059906]
|
d9e4771c-bad8-4605-956b-8cd7a9c054c7
|
deep-bayesian-active-learning-for-multiple
|
1912.01119
| null |
https://arxiv.org/abs/1912.01119v2
|
https://arxiv.org/pdf/1912.01119v2.pdf
|
Deep Bayesian Active Learning for Multiple Correct Outputs
|
Typical active learning strategies are designed for tasks, such as classification, with the assumption that the output space is mutually exclusive. The assumption that these tasks always have exactly one correct answer has resulted in the creation of numerous uncertainty-based measurements, such as entropy and least confidence, which operate over a model's outputs. Unfortunately, many real-world vision tasks, like visual question answering and image captioning, have multiple correct answers, causing these measurements to overestimate uncertainty and sometimes perform worse than a random sampling baseline. In this paper, we propose a new paradigm that estimates uncertainty in the model's internal hidden space instead of the model's output space. We specifically study a manifestation of this problem for visual question answer generation (VQA), where the aim is not to classify the correct answer but to produce a natural language answer, given an image and a question. Our method overcomes the paraphrastic nature of language. It requires a semantic space that structures the model's output concepts and that enables the usage of techniques like dropout-based Bayesian uncertainty. We build a visual-semantic space that embeds paraphrases close together for any existing VQA model. We empirically show state-of-art active learning results on the task of VQA on two datasets, being 5 times more cost-efficient on Visual Genome and 3 times more cost-efficient on VQA 2.0.
|
['Li Fei-Fei', 'Michael Bernstein', 'Khaled Jedoui', 'Ranjay Krishna']
|
2019-12-02
| null | null | null | null |
['question-answer-generation']
|
['natural-language-processing']
|
[ 1.55986413e-01 6.71476364e-01 -1.29947811e-01 -5.04776001e-01
-1.01743722e+00 -4.92060930e-01 8.43463361e-01 2.35686854e-01
-5.19520819e-01 7.66778946e-01 1.22078590e-01 -4.79690880e-01
1.72999993e-01 -8.76151264e-01 -1.07634461e+00 -4.89586294e-01
4.80308533e-01 7.18600988e-01 3.63925248e-01 1.82816207e-01
2.78746963e-01 -1.53284678e-02 -1.74503672e+00 4.74293441e-01
9.89182174e-01 1.10931098e+00 2.16258153e-01 7.73942173e-01
-8.57714772e-01 1.31930482e+00 -7.84481704e-01 -7.63657689e-01
-1.45670667e-01 -6.55233800e-01 -1.12854481e+00 6.64711073e-02
7.83281446e-01 -2.14798763e-01 -3.05087939e-02 1.07585895e+00
8.84668604e-02 1.38437524e-01 9.36354637e-01 -1.60597110e+00
-1.03223324e+00 5.64949214e-01 -5.79839759e-02 -7.05058575e-02
4.02667403e-01 4.47470278e-01 1.23876584e+00 -1.00455809e+00
6.38062298e-01 1.51061952e+00 3.62519145e-01 8.17856610e-01
-1.63170648e+00 -3.82267147e-01 1.88853279e-01 7.21597135e-01
-1.00873375e+00 -4.05461192e-01 6.98187113e-01 -5.47943771e-01
8.36406112e-01 2.91717649e-01 4.34600085e-01 1.46641707e+00
-1.17137400e-03 1.11209583e+00 1.18150520e+00 -6.20172203e-01
8.94038141e-01 6.20496750e-01 1.74861699e-01 5.83995223e-01
-6.68129548e-02 -2.09653061e-02 -6.04175627e-01 -2.62871593e-01
4.08532381e-01 -2.73659825e-01 -2.12836742e-01 -8.12545180e-01
-8.31652761e-01 1.17100275e+00 6.43912315e-01 5.89012429e-02
-1.23561732e-02 2.87246108e-01 5.83632439e-02 4.74721342e-01
4.92419809e-01 5.74793875e-01 -2.27838814e-01 1.05157889e-01
-9.50856924e-01 1.04226470e-01 8.55203092e-01 8.70517254e-01
8.58359635e-01 -9.14668366e-02 -4.54248339e-01 5.90882301e-01
7.07586169e-01 3.48926753e-01 3.04514766e-01 -1.33918011e+00
2.77503580e-01 7.04826772e-01 2.24193007e-01 -7.30246842e-01
1.56515419e-01 -8.14440697e-02 -5.93141556e-01 7.62736678e-01
6.75457895e-01 2.22576395e-01 -1.25337613e+00 1.86531365e+00
1.38064906e-01 -7.25345984e-02 9.04376134e-02 8.25898707e-01
1.00406444e+00 6.83355868e-01 1.64534420e-01 -7.53799453e-02
8.99995983e-01 -1.02882969e+00 -7.04259694e-01 -4.48382825e-01
2.28823319e-01 -3.16408604e-01 1.50975835e+00 2.96545863e-01
-1.06873357e+00 -7.07669616e-01 -1.05776596e+00 -3.86913449e-01
-5.06288767e-01 -2.86190152e-01 2.56897241e-01 6.34945869e-01
-1.14228797e+00 4.58658993e-01 -5.37040412e-01 5.16678765e-02
7.25926340e-01 1.19987866e-02 -2.79811114e-01 2.87082531e-02
-1.08727217e+00 1.23057890e+00 4.23663050e-01 -2.02877641e-01
-1.13020086e+00 -6.05700076e-01 -1.06454802e+00 1.72985420e-01
5.61527431e-01 -9.45650101e-01 1.39079773e+00 -1.39025903e+00
-1.44160581e+00 1.00667107e+00 -2.12901279e-01 -7.23053932e-01
6.95012033e-01 -8.43771398e-02 1.77261680e-01 1.67798787e-01
-6.15876839e-02 1.15092421e+00 1.37069678e+00 -1.54361784e+00
-1.24285601e-01 -3.57412994e-01 3.66433084e-01 1.31826058e-01
8.46452266e-02 -4.12000030e-01 -2.11930260e-01 -2.26833135e-01
-1.03722610e-01 -6.45188987e-01 -1.24024764e-01 6.45906508e-01
-3.04388255e-01 -4.40207750e-01 6.04418933e-01 -4.62345064e-01
7.44345546e-01 -1.98263359e+00 4.91240531e-01 -5.76056279e-02
4.45862412e-01 9.20981765e-02 -1.22554563e-01 1.30902156e-01
2.48893630e-02 1.52790517e-01 -6.79149032e-01 -8.35604191e-01
1.79666996e-01 4.90313470e-01 -7.31006324e-01 2.15567321e-01
5.36788940e-01 1.27957976e+00 -9.50596035e-01 -9.01955485e-01
3.30580741e-01 1.84872732e-01 -3.51009101e-01 5.43631971e-01
-8.48744869e-01 1.79273590e-01 1.42028285e-02 3.46000731e-01
5.21542013e-01 -5.42234957e-01 -2.16093674e-01 1.24128893e-01
1.21942468e-01 1.68022811e-01 -8.93664062e-01 1.84006250e+00
-5.26535273e-01 7.44168878e-01 -1.99298531e-01 -9.34281051e-01
9.68812108e-01 2.62521833e-01 -1.68475747e-01 -6.79363847e-01
2.27073673e-03 7.07324967e-02 -2.21039012e-01 -4.98186290e-01
2.05267400e-01 3.84037234e-02 1.70061767e-01 4.25200284e-01
5.33476591e-01 -5.88436306e-01 3.58512364e-02 6.86920583e-01
8.82934332e-01 3.61224979e-01 3.23387265e-01 4.66099270e-02
4.72959340e-01 6.57057315e-02 2.30560880e-02 1.10809278e+00
-2.84830213e-01 9.55911994e-01 7.61431515e-01 -8.12893957e-02
-1.25364506e+00 -1.56455171e+00 4.03829925e-02 7.57592857e-01
6.15471862e-02 -1.52815655e-01 -6.92858756e-01 -1.11649096e+00
4.08653691e-02 1.22485578e+00 -8.88197124e-01 -3.38836402e-01
6.61956239e-03 -2.55908757e-01 2.64142036e-01 3.89786988e-01
4.30323422e-01 -1.23453987e+00 -7.19980896e-01 -6.59588799e-02
-1.98805884e-01 -7.39585876e-01 -1.14331834e-01 1.36266321e-01
-8.52964759e-01 -1.05704534e+00 -7.95346439e-01 -4.33245659e-01
5.88167608e-01 -3.09359759e-01 1.66750169e+00 -3.36302780e-02
-2.22577587e-01 7.53459394e-01 -2.77134269e-01 -6.07431591e-01
-6.65153325e-01 -2.22410679e-01 -4.08614427e-01 5.06113432e-02
3.87804836e-01 -3.02920282e-01 -4.12731379e-01 -1.93338647e-01
-1.00454032e+00 5.72801493e-02 4.61689800e-01 9.34062779e-01
4.90309000e-01 -6.72905624e-01 6.08372867e-01 -7.93845654e-01
5.72866678e-01 -5.22336304e-01 -5.70981145e-01 6.12369955e-01
-5.79066396e-01 6.04917824e-01 4.30772096e-01 -5.58455884e-01
-1.18935263e+00 1.66196823e-01 -9.75761861e-02 -6.07487619e-01
-1.24401331e-01 9.71742943e-02 -1.39700130e-01 1.97943479e-01
1.01947510e+00 1.60917252e-01 1.94025099e-01 -2.42695197e-01
8.08284998e-01 6.26522541e-01 4.81145948e-01 -2.71619350e-01
7.58500457e-01 4.90255117e-01 -2.60959983e-01 -5.49893975e-01
-1.08290148e+00 -3.93069714e-01 -3.82842541e-01 -2.85267353e-01
9.09763277e-01 -5.91503859e-01 -6.33605123e-01 2.79029638e-01
-1.58861899e+00 -1.44086257e-01 -8.54628861e-01 1.75497532e-01
-9.04296458e-01 4.06126171e-01 -1.43533498e-01 -1.08919883e+00
-2.57939458e-01 -1.15374351e+00 1.01787901e+00 3.13468158e-01
-3.67998540e-01 -9.83423471e-01 4.00340289e-01 5.80684066e-01
5.10488093e-01 -1.58282667e-01 1.17322373e+00 -5.46588957e-01
-8.99403274e-01 4.14358437e-01 -3.57027709e-01 6.29723668e-01
-1.69758156e-01 -1.82358518e-01 -1.40814877e+00 4.54193093e-02
2.47802466e-01 -9.39726293e-01 1.28715241e+00 1.86422512e-01
1.26470435e+00 -3.61013591e-01 5.81679419e-02 1.83048576e-01
1.38272595e+00 -3.02534364e-02 9.61531281e-01 1.47244865e-02
3.78423542e-01 8.85423601e-01 2.66778588e-01 1.64487790e-02
2.19239250e-01 5.63094854e-01 8.98523867e-01 5.12478640e-04
-4.07040864e-01 -3.81201208e-01 2.52774745e-01 4.54613477e-01
4.84802902e-01 -3.46441209e-01 -8.29136193e-01 6.78941131e-01
-1.92773914e+00 -1.02130127e+00 -2.65868809e-02 2.42723846e+00
1.30903542e+00 9.92895737e-02 -2.54672319e-01 8.85345265e-02
3.61939490e-01 1.36883050e-01 -7.61720896e-01 -5.49862087e-01
-1.53903529e-01 3.67405295e-01 -1.79117601e-02 8.75814915e-01
-1.05773413e+00 5.95548630e-01 6.20720625e+00 8.21731448e-01
-6.64692223e-01 2.42525443e-01 7.83630311e-01 2.74427421e-02
-6.56773448e-01 1.56198934e-01 -4.55567062e-01 5.08416057e-01
7.34123886e-01 -3.45568396e-02 1.54600576e-01 9.78094280e-01
-3.75850827e-01 -4.88810003e-01 -1.52288842e+00 1.16794932e+00
2.81262636e-01 -1.48201478e+00 2.36212075e-01 -2.95366764e-01
6.98465347e-01 -2.91548878e-01 1.98267713e-01 4.93544847e-01
2.72289425e-01 -1.29959691e+00 9.39813852e-01 8.60978663e-01
6.92751586e-01 -4.01187629e-01 5.26692927e-01 5.47241807e-01
-5.19986212e-01 -4.08057906e-02 -4.36105192e-01 1.89519376e-02
3.60645168e-02 4.69674021e-01 -8.61691952e-01 1.27309456e-01
6.62523806e-01 2.69679070e-01 -8.60468924e-01 1.10440040e+00
-4.78215992e-01 5.49576342e-01 -1.41735584e-01 -1.64877579e-01
2.00725853e-01 -6.27711937e-02 5.36906540e-01 8.53460312e-01
-9.46674868e-03 -3.39588314e-01 -1.75274536e-01 1.60540712e+00
-1.45089835e-01 -8.78454372e-02 -7.17803717e-01 3.00884619e-02
3.32207561e-01 7.57936358e-01 -3.73905152e-01 -5.25975883e-01
-3.33605170e-01 1.25477540e+00 5.17044842e-01 3.74173373e-01
-8.84568512e-01 -2.95823008e-01 2.63292164e-01 -1.43365860e-01
6.00023419e-02 -1.06111751e-03 -3.46532047e-01 -1.18867683e+00
1.05234861e-01 -8.46937358e-01 1.38518885e-01 -1.22933710e+00
-1.44530821e+00 4.01498586e-01 1.14352390e-01 -8.44696879e-01
-5.90281427e-01 -7.02372372e-01 -5.20409405e-01 9.98941123e-01
-1.28957272e+00 -9.21293259e-01 -3.30533475e-01 5.33428788e-01
7.56344140e-01 -9.23402309e-02 9.80097711e-01 -1.12039246e-01
5.37113957e-02 4.00451422e-01 5.58259189e-02 3.17458585e-02
6.23213887e-01 -1.71281815e+00 4.70107615e-01 6.69101417e-01
7.49682486e-01 2.92205155e-01 8.49971116e-01 -4.89594191e-01
-9.79561627e-01 -6.94852889e-01 9.72278774e-01 -1.07620561e+00
5.79647362e-01 -6.31720304e-01 -1.13211703e+00 5.58886349e-01
3.71422946e-01 -5.87555580e-02 1.83435902e-01 -5.57930842e-02
-7.52706528e-01 1.46804750e-01 -1.16402221e+00 6.18573666e-01
7.73564756e-01 -9.58034456e-01 -1.10748458e+00 2.71027535e-01
1.03518140e+00 -1.67880774e-01 -8.41212869e-02 2.29348093e-01
2.43367910e-01 -1.23299444e+00 9.96527016e-01 -8.69916916e-01
4.96567309e-01 -1.37282640e-01 5.25121950e-03 -1.31914413e+00
-6.46211160e-03 -9.60559472e-02 -5.12020290e-01 9.84357178e-01
6.36612654e-01 -4.10307050e-01 8.43726575e-01 5.18151164e-01
2.67938618e-02 -5.03175855e-01 -1.27529132e+00 -5.63980222e-01
6.11732118e-02 -4.83349234e-01 1.77000910e-01 6.27370894e-01
-4.45229053e-01 6.52247548e-01 -1.22855768e-01 -2.99686909e-01
9.07999039e-01 -6.78885207e-02 4.58923131e-01 -1.24530625e+00
-4.93651479e-01 -2.61025637e-01 -4.03447092e-01 -8.69883299e-01
3.27783167e-01 -6.95804417e-01 2.18215331e-01 -1.64340103e+00
2.92209893e-01 -2.54297286e-01 8.91801063e-03 3.64339322e-01
-1.35793462e-01 1.75557792e-01 2.70734906e-01 7.13527277e-02
-7.50312150e-01 7.12644637e-01 1.00929773e+00 -6.50615156e-01
-2.11982131e-02 -1.09475508e-01 -3.45158696e-01 6.52883470e-01
5.27608633e-01 -5.92464626e-01 -8.28126729e-01 -4.69656885e-01
7.39160299e-01 -1.49043292e-01 8.90940487e-01 -8.10614407e-01
2.77612358e-01 -5.89153590e-03 4.10711795e-01 -3.86244506e-01
6.82446659e-01 -8.13704789e-01 -6.80553615e-02 2.56388962e-01
-7.69786656e-01 -3.28912705e-01 4.89357673e-02 7.32266843e-01
-3.37189198e-01 -6.98755682e-01 8.29559624e-01 -3.96632731e-01
-7.69706130e-01 1.09076621e-02 -2.00853586e-01 2.97514528e-01
9.51099455e-01 -1.76489294e-01 -5.44490099e-01 -6.26774669e-01
-9.36276913e-01 2.15220660e-01 4.50774670e-01 4.87928569e-01
9.12146211e-01 -1.15500009e+00 -7.63959646e-01 -1.28897399e-01
3.94490361e-01 1.31772235e-02 2.37452924e-01 5.09013057e-01
-4.11697954e-01 3.42619449e-01 -5.35451919e-02 -8.83212745e-01
-1.16211283e+00 7.26632535e-01 5.65100670e-01 -9.62652713e-02
-3.80164087e-01 1.10085404e+00 3.70021075e-01 -4.68826830e-01
4.55399513e-01 -2.12449491e-01 -2.01613247e-01 2.11921826e-01
5.02166033e-01 1.51303098e-01 -2.35640705e-02 -1.83606580e-01
-1.84956759e-01 2.38840237e-01 2.42720265e-02 -4.20765579e-01
7.33296037e-01 -1.18466079e-01 -9.02546383e-03 9.86188412e-01
1.13420570e+00 -3.62743706e-01 -1.38814473e+00 -3.67686033e-01
1.40477747e-01 -4.65382427e-01 -7.08219633e-02 -9.72244501e-01
-4.30210352e-01 1.33800864e+00 8.16500187e-01 3.31545323e-01
6.53696001e-01 4.81990218e-01 8.39740559e-02 5.01155436e-01
2.97416806e-01 -1.07648122e+00 5.15726328e-01 3.70283961e-01
1.20397174e+00 -1.69010091e+00 -2.42215231e-01 9.29961167e-03
-7.99584985e-01 8.01210284e-01 6.17912531e-01 2.71860342e-02
2.36198366e-01 -2.84728296e-02 1.24477670e-01 7.92649444e-05
-1.06213462e+00 -3.08667243e-01 4.96317536e-01 7.53370821e-01
1.10839501e-01 -1.14535913e-01 1.25951469e-01 2.04863206e-01
5.43465093e-02 -9.59377363e-03 3.93133521e-01 6.77071810e-01
-5.87375581e-01 -9.46079731e-01 -4.02630121e-01 2.64478832e-01
2.74356958e-02 -1.42850906e-01 -7.49705553e-01 5.45900702e-01
1.69354498e-01 1.03104138e+00 3.43398005e-01 5.24603091e-02
-4.16537561e-03 5.59880257e-01 8.17325652e-01 -7.88031042e-01
-2.85335511e-01 -7.72490919e-01 1.19762691e-02 -6.76438034e-01
-3.09844971e-01 -2.08098680e-01 -1.01878130e+00 6.47944063e-02
-2.78845102e-01 2.44043082e-01 7.67407358e-01 9.92039204e-01
1.91293642e-01 2.11879209e-01 3.14297944e-01 -2.95069784e-01
-1.06010771e+00 -9.64161396e-01 -2.35847741e-01 5.59883118e-01
4.12809342e-01 -6.09789968e-01 -7.41396427e-01 1.22822776e-01]
|
[10.796609878540039, 1.756005883216858]
|
9d199f13-775a-43df-a07c-b58ada41629d
|
ape-argument-pair-extraction-from-peer-review
| null | null |
https://aclanthology.org/2020.emnlp-main.569
|
https://aclanthology.org/2020.emnlp-main.569.pdf
|
APE: Argument Pair Extraction from Peer Review and Rebuttal via Multi-task Learning
|
Peer review and rebuttal, with rich interactions and argumentative discussions in between, are naturally a good resource to mine arguments. However, few works study both of them simultaneously. In this paper, we introduce a new argument pair extraction (APE) task on peer review and rebuttal in order to study the contents, the structure and the connections between them. We prepare a challenging dataset that contains 4,764 fully annotated review-rebuttal passage pairs from an open review platform to facilitate the study of this task. To automatically detect argumentative propositions and extract argument pairs from this corpus, we cast it as the combination of a sequence labeling task and a text relation classification task. Thus, we propose a multitask learning framework based on hierarchical LSTM networks. Extensive experiments and analysis demonstrate the effectiveness of our multi-task framework, and also show the challenges of the new task as well as motivate future research directions.
|
['Luo Si', 'Wei Lu', 'Qian Yu', 'Lidong Bing', 'Liying Cheng']
| null | null | null | null |
emnlp-2020-11
|
['argument-pair-extraction-ape']
|
['natural-language-processing']
|
[ 2.63135344e-01 4.07095850e-01 -5.96098721e-01 -3.67216557e-01
-1.05479491e+00 -5.57631135e-01 9.53722537e-01 5.39349556e-01
-3.46550345e-01 1.02556646e+00 4.46468562e-01 -8.80741060e-01
-1.48959786e-01 -5.91196835e-01 -9.60212529e-01 -1.90370351e-01
2.55696714e-01 3.35036784e-01 1.22927703e-01 -2.84115970e-01
6.12018108e-01 -3.80679190e-01 -1.17776525e+00 9.65292156e-01
1.20463669e+00 6.89976454e-01 3.25172804e-02 5.77505529e-01
-5.93615413e-01 1.40660512e+00 -8.38694811e-01 -9.33774650e-01
-3.00132662e-01 -4.65044260e-01 -1.34925890e+00 -3.33711535e-01
3.19153577e-01 -1.51853830e-01 2.74900962e-02 7.86693156e-01
3.63749206e-01 -3.69574726e-02 6.48612499e-01 -1.17337203e+00
-7.79960334e-01 1.57688808e+00 -6.03652239e-01 3.74025375e-01
3.84504884e-01 -5.62421203e-01 1.64333522e+00 -9.24573123e-01
5.35777628e-01 1.33223593e+00 7.25722432e-01 1.49416938e-01
-7.54599452e-01 -6.48126721e-01 3.66233498e-01 5.77988327e-01
-4.30464566e-01 -4.63376135e-01 1.06213057e+00 -4.49620426e-01
9.34590280e-01 -1.25138804e-01 5.91484666e-01 1.28204989e+00
5.61077744e-02 1.25526381e+00 1.07067680e+00 -7.74007618e-01
-1.88797399e-01 -1.19348727e-01 8.16880226e-01 5.39114892e-01
2.49128044e-01 -5.57217062e-01 -4.52037334e-01 -2.92587996e-01
1.11173190e-01 -4.05988365e-01 -9.51425880e-02 -1.71359964e-02
-1.10194194e+00 8.81563962e-01 4.00730610e-01 4.14765835e-01
-3.14193517e-01 2.58571953e-01 8.26755583e-01 4.69417810e-01
7.21627593e-01 3.31840634e-01 -6.09141111e-01 -1.59952432e-01
-3.49477023e-01 3.18536341e-01 1.19832468e+00 7.01060355e-01
4.48905796e-01 -6.42412961e-01 -2.57897079e-01 1.29795420e+00
4.20381576e-01 1.58300698e-01 3.71424645e-01 -6.57574296e-01
1.13472545e+00 7.47816503e-01 -1.25972345e-01 -1.09411407e+00
-3.23833615e-01 -8.62338692e-02 -4.57603008e-01 -5.50941885e-01
2.98203737e-01 -4.98017222e-01 -1.08192243e-01 1.47304583e+00
2.74641097e-01 4.54307124e-02 3.61734420e-01 4.32748824e-01
1.39825571e+00 6.40253842e-01 1.47591919e-01 -2.33662337e-01
1.47451484e+00 -1.44584131e+00 -1.05029237e+00 -1.62433192e-01
1.14046240e+00 -8.62473190e-01 8.47252369e-01 5.96826635e-02
-1.28044224e+00 -3.81600469e-01 -1.20118833e+00 -4.78888899e-01
-2.56858587e-01 6.01440966e-01 5.84813774e-01 9.15829465e-02
-3.36516082e-01 4.71824735e-01 -3.59876603e-01 1.40697360e-01
5.89719176e-01 -1.51242584e-01 4.98443060e-02 1.75425306e-01
-1.65982473e+00 1.06969333e+00 3.28028202e-01 3.77750158e-01
-2.88756788e-01 -3.75925362e-01 -1.03345251e+00 1.96443930e-01
7.50222385e-01 -4.56333071e-01 1.53722656e+00 -7.65306652e-01
-1.38438249e+00 9.79405999e-01 -1.74935922e-01 -6.04969025e-01
3.34486663e-01 -7.35616624e-01 -2.65796125e-01 8.13333318e-02
4.72415507e-01 1.70828268e-01 6.32040381e-01 -1.08191621e+00
-6.69564128e-01 -6.48702160e-02 3.66636157e-01 2.82484412e-01
-3.10420245e-01 4.82896686e-01 3.94120328e-02 -4.74598050e-01
-3.67787451e-01 -7.04057753e-01 9.35100168e-02 -5.95433235e-01
-8.07628393e-01 -1.09991467e+00 5.96146584e-01 -6.76653504e-01
1.27455187e+00 -1.70763838e+00 -3.86955626e-02 -2.10184783e-01
4.45178777e-01 5.40116727e-01 -2.43354693e-01 6.83696151e-01
-1.21713169e-01 4.48343843e-01 -3.01910847e-01 -3.28121215e-01
-9.11273248e-03 1.06074639e-01 -6.54658616e-01 7.83404261e-02
2.72254407e-01 1.42492318e+00 -9.23113406e-01 -6.71407282e-01
-3.69581670e-01 2.37884792e-03 2.57179644e-02 1.43510014e-01
-4.44203675e-01 2.87525244e-02 -7.55056739e-01 2.45400146e-01
3.85109991e-01 -4.57523257e-01 4.17744458e-01 -4.40609157e-02
-7.19900876e-02 1.31336808e+00 -4.84824568e-01 1.38278103e+00
-5.58986247e-01 8.46573293e-01 8.41082111e-02 -1.44759429e+00
1.01803327e+00 4.81545538e-01 4.87466976e-02 -7.17719615e-01
2.73799539e-01 4.46719974e-01 4.73942280e-01 -8.71485770e-01
3.99863958e-01 1.23748453e-02 -1.60342485e-01 1.20556366e+00
-3.81742626e-01 5.17209880e-02 5.84547579e-01 5.05775511e-01
9.68506277e-01 1.64895847e-01 5.11038601e-01 -1.47212312e-01
9.19429302e-01 -1.34808987e-01 5.01848817e-01 8.44706595e-01
-5.26814722e-02 -9.16144103e-02 1.09764230e+00 -4.33746248e-01
-8.49524498e-01 -4.36078429e-01 -8.74242410e-02 8.92733872e-01
1.40383244e-01 -5.80384672e-01 -5.41263103e-01 -1.38007796e+00
-2.34221704e-02 5.26356697e-01 -7.34093547e-01 2.18887895e-01
-1.22365272e+00 -6.18683279e-01 6.70659304e-01 4.87529278e-01
6.56788588e-01 -1.34877992e+00 -4.21072870e-01 2.96973020e-01
-6.92710221e-01 -1.36010170e+00 -1.04019977e-01 6.88943639e-02
-4.48948056e-01 -1.74952888e+00 -2.63072222e-01 -9.99617577e-01
1.42296165e-01 2.67416656e-01 1.40643632e+00 5.32345533e-01
1.80434495e-01 6.70237243e-02 -6.90044999e-01 -7.09062994e-01
-5.68533540e-01 3.73328269e-01 -4.10954565e-01 -3.28283846e-01
4.95768517e-01 -3.78795952e-01 3.54478578e-03 -3.66746970e-02
-4.41281736e-01 2.17173472e-01 5.09921610e-01 1.13210213e+00
9.10534933e-02 -4.45296019e-01 1.36904204e+00 -1.38044524e+00
1.46782136e+00 -7.18327761e-01 -1.83794603e-01 6.92884266e-01
-3.91569436e-01 -4.26436588e-02 4.87352252e-01 -2.31466815e-01
-1.45009017e+00 -8.36604655e-01 -1.11395217e-01 4.29507226e-01
2.10477322e-01 9.73805547e-01 1.12801343e-02 5.42047977e-01
4.17507261e-01 -2.40552977e-01 -1.90468337e-02 -4.10944998e-01
6.09222293e-01 9.07766938e-01 2.18397141e-01 -9.75962222e-01
5.56724370e-01 1.06263712e-01 -3.93006504e-01 -6.28674746e-01
-1.77985299e+00 -9.97087583e-02 -6.31046772e-01 1.90426614e-02
6.80705845e-01 -7.24674404e-01 -7.72580087e-01 1.65102094e-01
-1.73226440e+00 -3.90607834e-01 -4.43363003e-02 3.46303910e-01
-1.71551779e-01 7.05307662e-01 -1.05489814e+00 -6.92397654e-01
-7.04100370e-01 -8.93322945e-01 7.19971001e-01 1.03689112e-01
-4.09214586e-01 -1.20095146e+00 2.44186789e-01 8.36485684e-01
-3.71235907e-02 -1.26318946e-01 1.25028932e+00 -1.06362271e+00
-2.10967943e-01 1.31286383e-01 -5.95687389e-01 2.97523528e-01
1.71221301e-01 1.04055062e-01 -6.04843318e-01 2.63373166e-01
1.80680513e-01 -8.40256333e-01 1.20145309e+00 3.49884242e-01
9.68009233e-01 -4.69442248e-01 -4.68919069e-01 -7.97623843e-02
4.86339599e-01 1.01611726e-01 3.80611420e-01 5.81296325e-01
7.92306483e-01 1.13151157e+00 7.83894062e-01 2.24884525e-02
8.91696870e-01 4.42049384e-01 3.60577740e-02 4.56775818e-03
-5.70502691e-02 -3.47824991e-01 2.32629284e-01 1.29893494e+00
2.90506154e-01 -3.55427146e-01 -9.43719268e-01 6.09696984e-01
-2.19907498e+00 -9.47404325e-01 -6.89765692e-01 1.40260851e+00
1.08547425e+00 1.90971822e-01 -8.61810222e-02 2.66598552e-01
9.60290134e-01 3.46853465e-01 -2.03216210e-01 -7.85210311e-01
-2.16432810e-01 9.93249565e-02 -3.94810200e-01 5.43363154e-01
-1.20449436e+00 9.76122260e-01 5.74037600e+00 7.02198923e-01
-6.71867549e-01 6.61498159e-02 6.69947743e-01 3.45109731e-01
-4.49631214e-01 3.05460453e-01 -8.72409105e-01 5.23667097e-01
6.75628483e-01 -2.18728945e-01 -1.28601745e-01 6.95936143e-01
1.50210569e-02 2.30876990e-02 -1.07677925e+00 5.27041554e-01
3.15932572e-01 -1.60887825e+00 -1.18078232e-01 -9.18555483e-02
6.46619737e-01 1.10321186e-01 -4.47711289e-01 5.70146501e-01
4.44759518e-01 -9.13063943e-01 4.98278975e-01 -2.51513347e-02
1.34527951e-01 -3.56373250e-01 8.76355529e-01 4.58560854e-01
-9.20481265e-01 -1.40592709e-01 -1.76633880e-01 -6.36893034e-01
5.25875330e-01 7.34924614e-01 -6.28920138e-01 7.33616352e-01
3.97422135e-01 1.38799989e+00 -4.01952177e-01 6.02927268e-01
-1.22078693e+00 8.16582680e-01 -1.66032482e-02 -4.08521742e-01
2.88870186e-01 -1.82264149e-01 3.31305355e-01 1.13945448e+00
-1.19847111e-01 9.76785272e-02 1.86162069e-01 8.14145505e-01
-7.48492479e-01 3.68830860e-01 -5.00282943e-01 -2.88666815e-01
7.04392195e-01 1.52378976e+00 -5.61210215e-01 -6.21014118e-01
-3.89308840e-01 2.97113478e-01 9.39151347e-01 1.79914474e-01
-7.27361321e-01 -3.97398919e-01 1.53660536e-01 -4.19995010e-01
2.80198693e-01 8.44301507e-02 -4.22541022e-01 -1.42489100e+00
3.84395510e-01 -9.26534474e-01 5.30827463e-01 -5.68543136e-01
-1.69762695e+00 5.21489918e-01 -1.03279591e-01 -8.77133012e-01
-2.60798424e-01 -4.27484155e-01 -1.12138808e+00 5.87281346e-01
-1.93233359e+00 -1.38576317e+00 5.63780703e-02 -9.03280154e-02
7.03715861e-01 -3.77678461e-02 4.78503048e-01 3.36539358e-01
-7.43458927e-01 3.02571148e-01 -2.79666930e-01 4.47957337e-01
6.91184342e-01 -1.10990381e+00 5.91742575e-01 6.89477861e-01
3.50256741e-01 7.23273039e-01 9.83057320e-02 -8.04125845e-01
-9.06705678e-01 -5.35099089e-01 1.55981576e+00 -5.76250255e-01
1.01351893e+00 -2.60353506e-01 -1.18791819e+00 7.62643993e-01
6.76770568e-01 -4.47457492e-01 9.38599765e-01 7.16620445e-01
-4.50725079e-01 2.90027261e-01 -4.36333686e-01 4.86312330e-01
8.64915967e-01 -6.44130409e-01 -1.41663730e+00 4.63008106e-01
9.07337904e-01 -2.53234714e-01 -6.87045157e-01 4.99299705e-01
3.23925465e-01 -7.82808125e-01 8.19516778e-01 -7.52627492e-01
1.33058369e+00 1.11149669e-01 4.37767625e-01 -1.19504559e+00
1.00426592e-01 -5.76071322e-01 -3.92712414e-01 1.68667507e+00
6.90932095e-01 -5.87699413e-01 3.73073101e-01 2.75973797e-01
-3.62714410e-01 -1.20220518e+00 -6.25383198e-01 -3.54332954e-01
3.93398046e-01 -4.25245315e-01 4.31001067e-01 1.06993604e+00
6.36781573e-01 1.40694571e+00 -1.90080076e-01 -4.74460363e-01
3.48389477e-01 7.15780854e-01 7.55073309e-01 -1.37892151e+00
-2.21856788e-01 -5.27671635e-01 6.91308677e-01 -1.32703364e+00
9.45734441e-01 -1.06532753e+00 4.50536497e-02 -1.99515831e+00
3.10352743e-01 -6.62960887e-01 -2.23758369e-04 4.38501805e-01
-6.98652208e-01 -2.81765968e-01 -8.40425268e-02 4.61615950e-01
-7.20023215e-01 8.58428061e-01 1.21070838e+00 -1.33080468e-01
-1.49730384e-01 -7.81625956e-02 -1.01151991e+00 9.40300643e-01
6.96835458e-01 -4.92291540e-01 -3.44460219e-01 -7.80102849e-01
7.74129510e-01 5.42920008e-02 1.00009918e-01 -1.01058381e-02
2.98270315e-01 1.19456582e-01 -6.89074844e-02 -8.65908682e-01
1.80425704e-01 -1.03441663e-01 -9.96618569e-01 2.36088246e-01
-9.94251847e-01 2.11099327e-01 -9.71406996e-02 5.74306071e-01
-4.57686782e-01 -6.25220060e-01 1.32854655e-01 -2.02024169e-02
-1.66780323e-01 -2.87968218e-01 -3.52577537e-01 5.40034115e-01
7.47508526e-01 4.17379707e-01 -1.00604308e+00 -3.57121736e-01
-1.67951152e-01 7.13163972e-01 -2.05647737e-01 5.38599789e-01
5.39883971e-01 -1.02098930e+00 -1.14710319e+00 -2.61305779e-01
-2.57630441e-02 7.72171840e-02 1.93303321e-02 8.23399901e-01
-7.41885751e-02 5.69477975e-01 2.79524624e-01 -1.32211491e-01
-1.44601429e+00 3.42448622e-01 -1.63635910e-01 -7.97041237e-01
-7.42221415e-01 9.16908085e-01 -3.89824132e-03 -6.69338226e-01
8.77810922e-03 -2.52410918e-01 -1.01599622e+00 5.00488877e-01
4.13929909e-01 1.72206268e-01 -1.28580019e-01 -2.80323088e-01
-1.27560243e-01 1.62421241e-01 -5.60010076e-01 8.55528042e-02
1.47679138e+00 -2.41834700e-01 -8.62154841e-01 5.61874568e-01
9.69233096e-01 1.36231750e-01 -7.02624202e-01 -4.71742600e-01
5.34507751e-01 2.09505018e-02 -3.87304187e-01 -6.67570174e-01
-3.91574383e-01 1.12186539e+00 -6.22912884e-01 6.37400389e-01
3.25559199e-01 2.74500906e-01 1.18059421e+00 8.45176399e-01
-3.94407064e-01 -1.14562762e+00 2.24494278e-01 1.09677386e+00
8.82460415e-01 -1.23991573e+00 1.63746059e-01 -7.60611355e-01
-4.92081165e-01 1.10284233e+00 7.03736544e-01 -1.04770921e-01
5.38229942e-01 1.93614915e-01 2.02096090e-01 -6.37088180e-01
-1.20891929e+00 2.16377843e-02 1.92483827e-01 1.72580570e-01
9.38795686e-01 -2.16789618e-01 -9.62944686e-01 9.18099165e-01
-1.86868787e-01 -1.74874261e-01 6.43313408e-01 7.96906173e-01
-1.11940727e-01 -1.49350166e+00 1.65320471e-01 5.54428518e-01
-7.21628726e-01 -4.31990921e-01 -7.88867235e-01 4.54464525e-01
-2.72246450e-01 1.17756796e+00 -1.10082120e-01 2.65263626e-03
-1.71511378e-02 2.98971627e-02 2.22833022e-01 -7.53751397e-01
-9.86324966e-01 -2.56634444e-01 9.90727067e-01 6.18988313e-02
-8.29310000e-01 -3.99447411e-01 -1.17411578e+00 5.74973188e-02
-5.55045307e-01 7.68484056e-01 6.73792422e-01 1.62897861e+00
3.72037351e-01 7.35974312e-01 4.36426044e-01 -4.04592276e-01
-6.60012007e-01 -1.17530227e+00 4.70630266e-02 4.36426938e-01
8.52427408e-02 -6.88217402e-01 -2.38673091e-01 -2.13621870e-01]
|
[9.856151580810547, 9.292535781860352]
|
400ddef6-3a92-4049-a855-27e206e42ab9
|
learning-what-and-where-to-attend-with-humans
| null | null |
https://openreview.net/forum?id=BJgLg3R9KQ
|
https://openreview.net/pdf?id=BJgLg3R9KQ
|
Learning what and where to attend with humans in the loop
|
Most recent gains in visual recognition have originated from the inclusion of attention mechanisms in deep convolutional networks (DCNs). Because these networks are optimized for object recognition, they learn where to attend using only a weak form of supervision derived from image class labels. Here, we demonstrate the benefit of using stronger supervisory signals by teaching DCNs to attend to image regions that humans deem important for object recognition. We first describe a large-scale online experiment (ClickMe) used to supplement ImageNet with nearly half a million human-derived "top-down" attention maps. Using human psychophysics, we confirm that the identified top-down features from ClickMe are more diagnostic than "bottom-up" saliency features for rapid image categorization. As a proof of concept, we extend a state-of-the-art attention network and demonstrate that adding ClickMe supervision significantly improves its accuracy and yields visual features that are more interpretable and more similar to those used by human observers.
|
['Dan Shiebler', 'Drew Linsley', 'Sven Eberhardt', 'Thomas Serre']
|
2019-05-01
| null | null | null |
iclr-2019-5
|
['image-categorization']
|
['computer-vision']
|
[ 3.00594747e-01 -3.21947709e-02 -2.10998638e-04 -7.00394154e-01
-3.35908055e-01 -4.38628078e-01 5.58651984e-01 2.67778814e-01
-7.74769843e-01 3.03299397e-01 8.26551616e-02 -1.91878498e-01
-1.09937392e-01 -3.91329378e-01 -1.02237415e+00 -2.31036335e-01
-1.01091653e-01 1.31681964e-01 5.23205757e-01 -2.95085430e-01
6.83053970e-01 6.92200780e-01 -2.06246662e+00 6.94588006e-01
4.52382535e-01 1.09313190e+00 4.67723310e-01 6.68988168e-01
1.79285839e-01 6.06303453e-01 -6.20538056e-01 -6.51396811e-02
3.32215816e-01 -3.49587768e-01 -9.57165778e-01 -8.99533853e-02
1.10921872e+00 -4.42340314e-01 -1.69531003e-01 1.01895547e+00
4.38756078e-01 1.09611884e-01 5.74365735e-01 -1.33962536e+00
-1.29401565e+00 1.79088607e-01 -4.21096027e-01 7.62752116e-01
1.73406079e-01 4.70110744e-01 1.27902234e+00 -1.01976073e+00
5.45010507e-01 1.36013269e+00 3.82146001e-01 6.06079400e-01
-1.56755364e+00 -6.17130101e-01 2.95523703e-01 3.81014347e-01
-1.01491737e+00 -5.19140780e-01 4.40832704e-01 -5.63843131e-01
1.14716041e+00 2.35827744e-01 6.82739079e-01 1.17500949e+00
1.59990445e-01 7.83202946e-01 1.21478713e+00 -4.54714656e-01
1.99447945e-01 4.06449050e-01 1.40710577e-01 5.09230256e-01
3.26741278e-01 3.16779971e-01 -6.41222179e-01 1.78206652e-01
9.53457296e-01 1.69275001e-01 -1.68807194e-01 -5.55421293e-01
-1.29397655e+00 7.31566131e-01 1.22711563e+00 1.91313431e-01
-5.08713365e-01 3.19048107e-01 3.96738574e-02 2.26200715e-01
1.61179557e-01 9.87305284e-01 -4.19922173e-01 1.37067318e-01
-6.86826468e-01 -1.62117369e-02 3.82370770e-01 6.86601758e-01
8.55807662e-01 8.05712938e-02 -4.47624683e-01 6.40990198e-01
2.12915376e-01 4.94761527e-01 5.42306662e-01 -1.22730899e+00
2.63018925e-02 5.70732713e-01 3.12151641e-01 -9.76654053e-01
-3.66262883e-01 -6.63883567e-01 -1.93147913e-01 8.74067962e-01
3.80757213e-01 1.59972966e-01 -1.20652390e+00 1.73280323e+00
-1.09546095e-01 -4.62822884e-01 -3.13360751e-01 1.25954068e+00
5.24768651e-01 1.41794726e-01 4.89108473e-01 6.02912366e-01
1.70473313e+00 -8.55839729e-01 -2.44363129e-01 -5.77340424e-01
3.03661019e-01 -4.69702512e-01 1.42349398e+00 3.02481532e-01
-8.30943406e-01 -8.86230946e-01 -1.03047931e+00 -1.87940225e-01
-6.65289819e-01 9.55969095e-02 7.37905860e-01 2.51913577e-01
-1.21795082e+00 6.16771758e-01 -3.68220866e-01 -6.23859465e-01
9.41167951e-01 3.75974894e-01 -2.76641667e-01 2.72556961e-01
-1.04038835e+00 1.20037282e+00 9.65887308e-02 -2.26691559e-01
-1.46233308e+00 -8.30690145e-01 -6.43987894e-01 5.21215141e-01
2.48223737e-01 -5.10405242e-01 1.37395453e+00 -1.49727345e+00
-1.01275802e+00 1.22218525e+00 -1.68008402e-01 -5.36963403e-01
8.85327458e-02 -3.00655603e-01 -2.45257452e-01 6.68702543e-01
2.99884826e-01 1.51366782e+00 1.07472491e+00 -1.31967902e+00
-7.27997661e-01 -1.62866458e-01 2.31601000e-01 -8.42273757e-02
-4.22723711e-01 1.80903941e-01 6.59338534e-02 -5.14003098e-01
-2.96926022e-01 -7.01763034e-01 -5.57548590e-02 5.76655388e-01
-2.15662256e-01 -3.19754392e-01 7.98180938e-01 -4.39226985e-01
5.75203955e-01 -2.21585393e+00 -1.75334275e-01 7.94962123e-02
6.18658602e-01 4.34201717e-01 -4.16951239e-01 1.90886661e-01
-3.80123824e-01 3.36604685e-01 -1.17820799e-02 6.61112145e-02
1.33042336e-01 -1.58793166e-01 -4.21837866e-01 2.19848588e-01
6.50241554e-01 1.16265333e+00 -9.23535585e-01 -2.62298465e-01
2.18780696e-01 2.50988334e-01 -6.43805742e-01 4.10730653e-02
-1.16678230e-01 1.65109619e-01 -2.14762732e-01 5.04305720e-01
3.20428997e-01 -6.37876213e-01 -3.02712888e-01 -3.44752789e-01
-1.33480981e-01 1.56743363e-01 -4.83344853e-01 1.46348095e+00
-2.50380129e-01 1.11251140e+00 -1.23726716e-02 -8.49327087e-01
6.24744833e-01 -4.04057540e-02 -2.38892436e-02 -9.77830648e-01
2.49883041e-01 -7.93544501e-02 6.14200711e-01 -4.26741123e-01
2.92354584e-01 1.80317923e-01 3.65907758e-01 4.90940779e-01
3.76072019e-01 -1.82986781e-02 7.37388283e-02 3.87732118e-01
9.01804626e-01 -5.58584332e-02 1.56603739e-01 -5.78295827e-01
1.83027089e-01 3.31619270e-02 1.59265265e-01 1.15982497e+00
-6.74222052e-01 6.64529800e-01 4.70736355e-01 -4.45500463e-01
-1.22292650e+00 -9.98985767e-01 -8.60755593e-02 1.75035679e+00
1.69762373e-01 -1.77532613e-01 -7.61445463e-01 -5.96991360e-01
3.13127011e-01 7.42112279e-01 -1.15471745e+00 -3.09753001e-01
2.86700558e-02 -1.75573528e-01 3.69441867e-01 9.92559433e-01
4.49794352e-01 -1.40914702e+00 -1.10194731e+00 -1.66776210e-01
3.09495211e-01 -8.75969470e-01 -5.20393908e-01 3.17872375e-01
-4.94264930e-01 -1.07859707e+00 -8.03375781e-01 -7.41536617e-01
9.66841400e-01 5.73201358e-01 1.10472548e+00 2.43561074e-01
-8.47085536e-01 5.83767235e-01 -1.83092535e-01 -7.01044858e-01
-1.82823315e-02 4.48102169e-02 7.26749599e-02 -1.32607017e-02
6.45207584e-01 -3.22288662e-01 -8.41060817e-01 4.95114803e-01
-7.85019875e-01 4.60578389e-02 1.04832339e+00 7.05907345e-01
7.42639378e-02 -6.40439034e-01 7.54829228e-01 -6.09746635e-01
6.69832766e-01 -1.80585802e-01 -7.39696443e-01 1.39030695e-01
-7.01608956e-01 2.76651710e-01 4.42765743e-01 -4.21247810e-01
-8.20747316e-01 1.30314857e-01 2.97576189e-01 -5.07539332e-01
-4.59219545e-01 8.66165757e-02 3.65027905e-01 -4.65785801e-01
1.11469460e+00 4.62542623e-02 8.16225260e-02 -2.75503397e-01
3.92631769e-01 4.54498529e-01 4.05921906e-01 -3.45014989e-01
7.28313506e-01 4.77707446e-01 -2.08086878e-01 -6.43904150e-01
-1.19832230e+00 -3.81160200e-01 -6.63151622e-01 -2.25088045e-01
1.15450382e+00 -6.34417951e-01 -1.33433950e+00 2.18456417e-01
-1.00277853e+00 -4.69710350e-01 -2.79752851e-01 4.34892356e-01
-4.61099863e-01 -6.18702024e-02 -4.37420577e-01 -5.38618743e-01
-8.06770101e-02 -9.72457349e-01 1.05288887e+00 4.36405361e-01
-3.01254630e-01 -5.96570015e-01 -3.84870231e-01 9.69733968e-02
7.87497699e-01 -2.91848570e-01 8.11391294e-01 -7.79343128e-01
-7.05986619e-01 -2.01062788e-03 -8.83902490e-01 4.01940852e-01
3.46721672e-02 -1.35651287e-02 -1.39657748e+00 -2.63232231e-01
-4.19541240e-01 -7.10876226e-01 1.11014903e+00 3.38214636e-01
1.56538987e+00 -5.81401959e-02 -2.82720238e-01 5.35305858e-01
1.16943300e+00 3.13446322e-03 4.78132695e-01 4.30282682e-01
4.06795233e-01 6.90261841e-01 2.53447473e-01 6.47391230e-02
6.30366802e-02 4.94399756e-01 5.00363529e-01 -4.87783253e-01
-2.60364443e-01 -2.29399905e-01 1.38675705e-01 -6.87780753e-02
-6.87029660e-02 1.98266492e-03 -8.51784587e-01 7.59473562e-01
-1.58216584e+00 -9.72001612e-01 1.77726626e-01 2.18819737e+00
8.16692650e-01 4.59135354e-01 2.13175584e-02 -2.57420987e-01
7.80861259e-01 -1.27611458e-01 -8.11390579e-01 -4.31784809e-01
-1.31988237e-02 2.37901703e-01 6.04388356e-01 2.62714237e-01
-1.07736325e+00 8.92368734e-01 7.58027172e+00 3.85516375e-01
-1.28259635e+00 -1.59495831e-01 6.06323302e-01 1.55226784e-02
-8.78961757e-02 1.46281868e-02 -6.84964716e-01 2.34655321e-01
8.51599634e-01 -5.98995127e-02 4.02526379e-01 1.07723439e+00
2.77945437e-02 -2.99498498e-01 -1.56519115e+00 8.32354784e-01
1.62457883e-01 -1.36925066e+00 1.30843967e-01 6.50748536e-02
5.97558975e-01 1.75360903e-01 5.30535042e-01 2.31773838e-01
4.32096690e-01 -1.24553275e+00 7.91891515e-01 5.04539311e-01
7.97082305e-01 -1.74441785e-01 4.53563124e-01 6.23690002e-02
-6.90554202e-01 -3.28195333e-01 -5.30932724e-01 -2.45378017e-01
-2.80190617e-01 1.24302894e-01 -9.80403066e-01 -3.32556456e-01
9.86415923e-01 6.73677444e-01 -1.24794912e+00 1.19636071e+00
-4.74645048e-01 4.69423890e-01 2.44857185e-02 -1.95589855e-01
3.13805848e-01 5.14284074e-01 3.10370594e-01 1.19429564e+00
-1.11716561e-01 -2.03656685e-02 -1.37141407e-01 1.33953905e+00
-2.75030166e-01 -3.54160130e-01 -5.27226746e-01 -1.80909991e-01
4.07338947e-01 1.41118646e+00 -7.30718732e-01 -6.35068238e-01
-2.91493684e-01 1.01646411e+00 4.41999674e-01 5.89877427e-01
-6.54935300e-01 -9.25740659e-01 6.20576084e-01 8.29130113e-02
6.58464193e-01 -9.84905213e-02 -1.24325685e-01 -9.42639351e-01
-2.52355397e-01 -7.54359543e-01 -4.59656380e-02 -1.38528323e+00
-1.49553645e+00 4.62765396e-01 -1.81931540e-01 -1.05190349e+00
1.07751869e-01 -1.14256668e+00 -5.52295864e-01 1.07396996e+00
-1.45212376e+00 -8.30904901e-01 -6.80366039e-01 4.57374960e-01
4.93940413e-01 7.12639764e-02 7.45391607e-01 1.84370372e-02
-1.23750106e-01 5.38715243e-01 -3.24491471e-01 3.57829213e-01
9.60774302e-01 -1.33070731e+00 4.04538870e-01 6.68952167e-01
2.09197685e-01 9.58350122e-01 6.53851926e-01 -3.33266288e-01
-1.00367320e+00 -6.75725460e-01 5.02910733e-01 -8.18469286e-01
6.15676284e-01 -5.95847189e-01 -8.77664864e-01 8.17789972e-01
4.45164144e-01 2.38719761e-01 3.88575703e-01 2.34029382e-01
-6.18344605e-01 -3.07463855e-01 -1.11645186e+00 4.26196814e-01
1.04599333e+00 -8.09016049e-01 -9.53978717e-01 1.83160976e-01
6.95825517e-01 1.12500198e-01 -4.58847046e-01 9.03722420e-02
6.62369251e-01 -9.20605421e-01 1.02955794e+00 -1.09002781e+00
5.27319074e-01 -3.14863622e-01 5.93829677e-02 -1.36949360e+00
-7.61312187e-01 -1.57043710e-01 3.95655960e-01 8.39947760e-01
6.01097882e-01 -7.75406659e-01 4.88016218e-01 4.86441404e-01
-1.98688507e-01 -3.11406255e-01 -5.71239173e-01 -6.14682853e-01
-2.42024466e-01 -2.99835354e-02 5.90657368e-02 7.95630395e-01
-1.03569992e-01 3.79575789e-01 -6.69651479e-02 7.38541335e-02
6.86980724e-01 2.30110921e-02 4.97574776e-01 -1.36165237e+00
-5.19429892e-02 -6.65481269e-01 -6.91548526e-01 -9.90465283e-01
-1.13229252e-01 -8.22576225e-01 3.59779835e-01 -1.30971730e+00
3.45829129e-01 -4.86499183e-02 -7.86074936e-01 8.65023553e-01
-1.52454659e-01 4.78314102e-01 3.98685724e-01 1.55561373e-01
-8.11949730e-01 3.58440131e-01 1.23618937e+00 -1.82835653e-01
3.45928758e-01 -3.90537769e-01 -1.16011298e+00 6.71686828e-01
6.01268172e-01 -3.80584419e-01 -2.70886809e-01 -6.17523670e-01
-8.26845039e-03 -6.08520687e-01 8.93079042e-01 -1.17913783e+00
3.38799685e-01 -9.32533443e-02 1.03371501e+00 -2.58230686e-01
2.78291881e-01 -7.14023650e-01 -6.47110701e-01 5.41340828e-01
-9.36590552e-01 1.48538304e-02 5.01196444e-01 5.20812690e-01
-5.10085225e-02 -7.80783370e-02 9.61416543e-01 -2.51602799e-01
-1.15448332e+00 2.18878314e-02 -5.26979029e-01 1.15220351e-02
8.92619073e-01 -1.48405865e-01 -7.92241037e-01 -4.99454319e-01
-7.75933444e-01 2.14424595e-01 3.83710653e-01 4.98960167e-01
7.12821722e-01 -1.09918749e+00 -5.31972706e-01 2.46860653e-01
3.72405082e-01 -6.17295563e-01 2.05941703e-02 8.08686554e-01
-3.54594141e-01 6.51339471e-01 -7.62751400e-01 -7.51520455e-01
-9.31278288e-01 8.47516596e-01 3.14099908e-01 6.66124582e-01
-3.82687151e-01 1.14580822e+00 6.90454364e-01 -2.29837578e-02
2.90935457e-01 -4.91218835e-01 -6.76334277e-03 -1.21077172e-01
5.98317504e-01 -2.28535123e-02 -1.22219220e-01 -2.02910647e-01
-5.23892283e-01 3.91978383e-01 -2.52633899e-01 -9.46776047e-02
1.27016675e+00 6.85139373e-02 1.41026765e-01 3.22224230e-01
9.12349522e-01 -1.25927180e-01 -1.71278155e+00 -8.62792209e-02
-4.02321406e-02 -6.94304824e-01 -1.70319416e-02 -1.21066594e+00
-8.52087140e-01 1.30334938e+00 8.58871639e-01 3.42376679e-01
8.97800028e-01 2.23363549e-01 2.55328387e-01 7.02450156e-01
1.34133816e-01 -1.18445778e+00 5.26789486e-01 2.93474227e-01
1.12873566e+00 -1.59194112e+00 -1.93777263e-01 -1.17425025e-02
-6.95610881e-01 8.76458824e-01 1.10546410e+00 -3.87799621e-01
3.85634691e-01 -2.69955635e-01 9.12199356e-03 -4.89545375e-01
-8.91078770e-01 -4.92772102e-01 4.68682915e-01 7.52629459e-01
3.11785847e-01 -2.04656348e-01 2.58457869e-01 2.67090738e-01
5.19249626e-02 5.94601361e-03 4.33473557e-01 8.36401105e-01
-8.93925667e-01 -4.79048669e-01 -2.47813106e-01 6.28517807e-01
-2.06109658e-01 -3.38903874e-01 -5.32379329e-01 8.17799449e-01
7.42472766e-04 7.23434389e-01 2.25195080e-01 -2.16357335e-01
1.65380895e-01 2.45415028e-02 4.96574700e-01 -8.26604664e-01
-4.67208385e-01 -4.80369568e-01 -2.01364532e-01 -7.06118584e-01
-2.59773284e-01 -2.59727746e-01 -1.10614109e+00 4.66627367e-02
-1.88682765e-01 -4.93431203e-02 7.30908334e-01 6.75390244e-01
7.72854030e-01 7.81780303e-01 3.71630132e-01 -1.16084862e+00
-6.24768555e-01 -1.01888597e+00 -4.22463000e-01 7.84253538e-01
5.56295812e-01 -8.69092703e-01 -5.25447667e-01 2.31074035e-01]
|
[9.988508224487305, 1.8669776916503906]
|
8dc38937-f653-4be9-a83f-8a2af547e75b
|
spot-keywords-from-very-noisy-and-mixed
|
2305.17706
| null |
https://arxiv.org/abs/2305.17706v1
|
https://arxiv.org/pdf/2305.17706v1.pdf
|
Spot keywords from very noisy and mixed speech
|
Most existing keyword spotting research focuses on conditions with slight or moderate noise. In this paper, we try to tackle a more challenging task: detecting keywords buried under strong interfering speech (10 times higher than the keyword in amplitude), and even worse, mixed with other keywords. We propose a novel Mix Training (MT) strategy that encourages the model to discover low-energy keywords from noisy and mixed speech. Experiments were conducted with a vanilla CNN and two EfficientNet (B0/B2) architectures. The results evaluated with the Google Speech Command dataset demonstrated that the proposed mix training approach is highly effective and outperforms standard data augmentation and mixup training.
|
['Shi Yin', 'Jiqing Han', 'Lantian Li', 'Dong Wang', 'Ying Shi']
|
2023-05-28
| null | null | null | null |
['keyword-spotting']
|
['speech']
|
[ 3.12581748e-01 -8.95995125e-02 -3.14251855e-02 -1.40513536e-02
-1.11444092e+00 -2.84768820e-01 6.92900956e-01 -1.84174821e-01
-5.12300193e-01 6.19427681e-01 4.74497080e-01 -5.55649817e-01
1.58185497e-01 -2.28712305e-01 -8.27735722e-01 -7.96890497e-01
-1.86088637e-01 -1.16623715e-01 2.40260616e-01 -3.77183646e-01
-9.38102305e-02 3.49090338e-01 -1.69815087e+00 4.12308753e-01
6.14058077e-01 1.10582697e+00 6.83679461e-01 1.06091750e+00
-7.46896416e-02 8.31368446e-01 -1.33707511e+00 -3.75742689e-02
1.24047361e-01 -1.09887294e-01 -4.17429268e-01 -3.42211366e-01
3.96944612e-01 -2.80602455e-01 -6.73119724e-01 1.00469351e+00
1.18304348e+00 3.29872310e-01 4.34471071e-02 -1.24653149e+00
-1.89390808e-01 1.25934923e+00 -3.22593451e-01 8.32767665e-01
4.68440056e-01 2.12742165e-01 8.22879374e-01 -1.03252327e+00
-1.09600089e-02 1.22660065e+00 7.42190003e-01 4.76936996e-01
-8.23812604e-01 -1.01063097e+00 3.05496514e-01 5.57761669e-01
-1.46880209e+00 -7.90031075e-01 9.66241717e-01 8.99660811e-02
1.52120447e+00 5.29073179e-01 3.81022245e-01 1.87999773e+00
-3.05866629e-01 9.75737929e-01 9.33121383e-01 -3.01924914e-01
5.63772470e-02 1.98359758e-01 -3.38250771e-02 1.05235942e-01
-2.83721417e-01 2.70029426e-01 -8.73201072e-01 -3.33841860e-01
-1.78850189e-01 -6.52444720e-01 -7.97461152e-01 6.38288021e-01
-1.13186371e+00 4.00083661e-01 1.93263978e-01 6.31049037e-01
-5.03996909e-01 3.59478652e-01 3.70919466e-01 2.67698705e-01
5.82911015e-01 4.96294409e-01 -5.09012878e-01 -4.89864886e-01
-1.32436562e+00 8.50900710e-02 7.43684709e-01 7.73858786e-01
2.33980089e-01 8.09353471e-01 -2.34376222e-01 1.10890961e+00
2.46874183e-01 6.75572217e-01 6.21325910e-01 -4.04952556e-01
5.61685801e-01 -2.26301357e-01 -6.09156527e-02 -8.24702084e-01
-4.50633615e-01 -9.71599698e-01 -7.14665830e-01 -3.79026949e-01
-3.28926474e-01 -4.29399759e-01 -1.30568540e+00 1.60245407e+00
2.40873080e-02 6.22038484e-01 1.51275665e-01 7.82401502e-01
1.45457983e+00 1.22528720e+00 -1.41310528e-01 -3.30690533e-01
1.17330956e+00 -9.69157994e-01 -1.35311735e+00 -5.83158195e-01
1.04181528e-01 -8.31555665e-01 1.10845578e+00 6.63193107e-01
-1.02444255e+00 -6.10173643e-01 -1.17624962e+00 1.42614931e-01
-4.89579737e-01 1.83626711e-01 2.96793520e-01 7.34371603e-01
-1.28467512e+00 1.67777181e-01 -3.56297761e-01 4.54074368e-02
2.39046976e-01 3.09199005e-01 -2.30936315e-02 1.81716710e-01
-1.60145521e+00 7.91678429e-01 3.02977562e-01 3.62633675e-01
-1.36908805e+00 -7.91885674e-01 -8.27038288e-01 1.60791069e-01
7.25700796e-01 6.41324073e-02 1.49612403e+00 -6.21110380e-01
-1.50062430e+00 3.41560543e-01 -8.90635327e-02 -7.56430745e-01
2.46065065e-01 -4.38701630e-01 -1.08354056e+00 2.11971119e-01
-8.29782262e-02 4.76695240e-01 1.02056253e+00 -1.55460703e+00
-4.68917549e-01 9.47222114e-02 3.19937291e-03 1.62445754e-01
-4.25838202e-01 1.73952714e-01 -4.34802592e-01 -1.02829325e+00
-5.88275343e-02 -2.88901776e-01 9.73261073e-02 -5.32822371e-01
-9.71993804e-01 -7.07288682e-02 1.09535253e+00 -8.41358364e-01
1.46458614e+00 -2.21920180e+00 -1.47829100e-01 1.21852934e-01
9.25303176e-02 7.33017683e-01 -9.66429412e-02 2.82608181e-01
-1.77046880e-01 2.29676872e-01 -9.81634855e-02 -5.25917470e-01
-1.38398215e-01 1.13794036e-01 -5.04297197e-01 2.82896817e-01
6.48211613e-02 6.32103920e-01 -7.50530124e-01 -3.96934003e-01
-9.88442171e-03 6.45520985e-01 -2.76994884e-01 2.84772873e-01
-2.31065050e-01 1.47952691e-01 -2.07948219e-02 8.65176678e-01
8.00306261e-01 2.15991229e-01 -1.31845057e-01 -4.71565753e-01
-1.57984510e-01 7.16564178e-01 -1.16300511e+00 1.22721577e+00
-5.60861528e-01 9.59862530e-01 7.24089444e-01 -1.05078423e+00
7.22715616e-01 6.46248102e-01 2.38302469e-01 -7.94762850e-01
3.36791813e-01 3.79074335e-01 5.48175797e-02 -7.45554507e-01
7.12817132e-01 1.97286949e-01 2.01259941e-01 -2.51543727e-02
1.80993676e-01 -3.45009387e-01 -2.71186709e-01 9.92253274e-02
1.18365943e+00 -3.87810349e-01 -7.38035515e-02 -6.19263686e-02
4.95340288e-01 -4.62996006e-01 4.04627413e-01 1.13281226e+00
-1.89314693e-01 4.01881516e-01 2.69909322e-01 2.61459917e-01
-5.11604428e-01 -9.23134506e-01 -5.24358265e-02 1.05579114e+00
1.22621626e-01 -4.91475374e-01 -5.86672246e-01 -3.12274128e-01
-2.75955766e-01 9.28007901e-01 -2.32338145e-01 -3.11888665e-01
-6.03744984e-01 -6.32388353e-01 1.15273952e+00 4.10951257e-01
8.20729434e-01 -9.90530968e-01 -1.11122221e-01 3.08455050e-01
-5.38729548e-01 -1.61178935e+00 -5.33551574e-01 7.87285089e-01
-4.42803912e-02 -4.26058084e-01 -7.43900478e-01 -7.96847939e-01
-2.05453541e-02 4.43846464e-01 7.95190930e-01 7.83508569e-02
-7.51507506e-02 3.31338704e-01 -5.26254594e-01 -7.58191407e-01
-4.80362624e-01 -3.30450647e-02 2.09567532e-01 5.48198745e-02
6.24879971e-02 -6.03275836e-01 -3.17426860e-01 2.06905171e-01
-8.80507052e-01 -3.29366922e-01 5.53107440e-01 8.24138403e-01
4.30876426e-02 1.89877853e-01 5.33006370e-01 -1.19639533e-02
1.03047073e+00 -6.06115997e-01 -2.00832307e-01 -1.03512555e-01
-3.05147380e-01 -4.42504793e-01 5.34942627e-01 -9.53889728e-01
-8.29161048e-01 -3.73009533e-01 -8.24866176e-01 -5.67495704e-01
1.16874129e-02 4.13096577e-01 -5.90462029e-01 -9.57703963e-02
5.13901532e-01 5.81613183e-01 -2.83289790e-01 -4.96623397e-01
1.27059221e-01 1.25363898e+00 6.93769872e-01 -1.19674988e-01
7.28487194e-01 2.43496135e-01 -6.24010026e-01 -1.47606063e+00
-4.14194673e-01 -5.97675204e-01 9.11897644e-02 -3.58647734e-01
7.45466888e-01 -1.08321500e+00 -4.71708536e-01 6.87200308e-01
-1.24895203e+00 -4.57176954e-01 2.48546340e-02 6.36168838e-01
-7.82348961e-02 4.24477100e-01 -3.97341907e-01 -1.15284848e+00
-4.10843432e-01 -1.29697633e+00 1.08747137e+00 6.18911833e-02
5.59897441e-03 -4.89346147e-01 -2.39270508e-01 2.87933975e-01
8.29538882e-01 -2.83687502e-01 4.25664097e-01 -1.07675242e+00
-2.11866185e-01 -5.65914512e-02 1.81276962e-01 6.28674448e-01
-7.86607191e-02 -7.01154396e-02 -1.53609824e+00 1.10832034e-02
2.15039656e-01 -2.72917897e-01 1.12959170e+00 3.84801865e-01
1.39252007e+00 -5.29956758e-01 -3.75674605e-01 5.20362973e-01
7.10840642e-01 4.93071020e-01 5.28403163e-01 1.16479412e-01
5.44368327e-01 1.18374057e-01 4.15253490e-01 3.76964718e-01
5.02850153e-02 8.26767862e-01 6.06076360e-01 -4.40994889e-01
-2.87921876e-01 -7.30764344e-02 4.53234613e-01 8.93388927e-01
5.41678846e-01 -9.23927248e-01 -9.33595359e-01 7.98179865e-01
-1.20670545e+00 -9.86514211e-01 3.97905558e-02 1.63710546e+00
1.06003916e+00 4.97602463e-01 -1.48284316e-01 6.95788682e-01
7.59193361e-01 5.80013394e-01 -2.71606326e-01 -4.21320647e-01
-5.49228847e-01 4.78245616e-01 5.05323529e-01 6.56301498e-01
-1.22570801e+00 7.43339658e-01 7.16250992e+00 1.07413125e+00
-1.36241019e+00 4.32422310e-01 5.32363534e-01 -3.09049547e-01
-5.28350472e-01 -6.46171987e-01 -6.64836168e-01 5.25250435e-01
1.22337651e+00 1.00474417e-01 4.39877331e-01 6.86326981e-01
4.19251293e-01 -1.26592398e-01 -7.03882158e-01 1.09806442e+00
2.16864541e-01 -1.15248740e+00 -1.90809682e-01 -7.18348622e-02
4.21291023e-01 3.81473929e-01 1.50301903e-01 6.68847859e-01
-4.27129753e-02 -1.04863310e+00 9.44659948e-01 1.35387108e-01
6.28828347e-01 -5.16695619e-01 6.65971160e-01 4.02011782e-01
-1.26541507e+00 -2.58478731e-01 5.17521054e-02 2.70308435e-01
1.58796787e-01 6.26236260e-01 -1.09655559e+00 2.69429952e-01
1.00112915e+00 1.64097980e-01 -1.14857048e-01 8.78302515e-01
-1.01143368e-01 1.07064974e+00 -6.59043729e-01 -3.64237666e-01
4.25631821e-01 5.61830342e-01 1.08800626e+00 1.58400393e+00
2.73750961e-01 4.31472622e-02 7.41804391e-02 6.52635992e-01
-4.43077266e-01 -9.48148146e-02 -6.29359245e-01 -2.36335352e-01
6.96507156e-01 1.11274076e+00 -4.05138403e-01 -4.44427103e-01
-2.46146441e-01 7.83284485e-01 -3.77049327e-01 5.35736620e-01
-1.13540328e+00 -3.10700327e-01 6.53980017e-01 -1.56277701e-01
4.50838238e-01 -3.11071068e-01 5.64844627e-03 -6.46679640e-01
2.55196333e-01 -1.17381918e+00 -1.41867563e-01 -1.11467385e+00
-7.47612953e-01 9.70632553e-01 -3.24371876e-03 -9.55492437e-01
5.47185652e-02 -4.15762484e-01 -7.91328013e-01 8.43029976e-01
-1.53022659e+00 -6.72910511e-01 -3.68876070e-01 5.15684545e-01
9.28384900e-01 -2.23494440e-01 5.40054977e-01 7.62896478e-01
-5.44914484e-01 8.35309029e-01 -9.03285071e-02 -1.51229545e-01
2.51774758e-01 -1.11220431e+00 5.04810512e-01 1.04930925e+00
9.03651863e-02 3.60719949e-01 9.77404833e-01 -5.81462502e-01
-1.43911171e+00 -8.45006168e-01 6.59833014e-01 1.60269737e-01
7.44787574e-01 -7.72457123e-01 -9.61138129e-01 1.01285361e-01
6.83214962e-01 -1.54719070e-01 3.41444731e-01 -4.40932423e-01
-1.62510440e-01 -2.03224897e-01 -1.12214971e+00 5.37217736e-01
8.92454445e-01 -8.34912360e-01 -4.22570318e-01 4.46704566e-01
1.35779583e+00 -4.84785795e-01 -4.24678832e-01 6.82860911e-01
8.01426768e-02 -5.38982153e-01 1.14023077e+00 -2.35269442e-01
-8.36026520e-02 -1.43542841e-01 -5.17141461e-01 -1.50177419e+00
3.52630109e-01 -1.18755257e+00 -4.31448042e-01 1.20983934e+00
6.74983203e-01 -3.81300360e-01 2.84900129e-01 -3.32738101e-01
-6.39781415e-01 -7.69970775e-01 -1.30498147e+00 -8.31281185e-01
-4.85996455e-01 -1.07193935e+00 4.46212530e-01 6.64807796e-01
-1.96315438e-01 1.64662272e-01 -6.96403801e-01 6.44611895e-01
1.02082253e-01 -7.13650286e-01 4.62957710e-01 -5.57458818e-01
-2.13559195e-01 -4.75541472e-01 -1.31866887e-01 -1.32717788e+00
7.09846839e-02 -5.35441220e-01 6.58255875e-01 -1.41714501e+00
-5.20235419e-01 -1.03432074e-01 -2.91116506e-01 3.65366757e-01
-3.02327305e-01 2.24971995e-01 -1.95686147e-01 -2.53456801e-01
-2.29230016e-01 7.29151189e-01 7.19680369e-01 -6.42217278e-01
-8.63619149e-03 1.96886510e-01 -6.37377381e-01 5.10105908e-01
7.75468409e-01 -3.85150969e-01 -3.52225751e-01 -1.27344489e-01
1.06606901e-01 -9.96040329e-02 4.07553643e-01 -1.14384222e+00
2.57904500e-01 2.51532167e-01 -1.08971387e-01 -1.13047361e+00
8.69348645e-01 -6.73976481e-01 -2.34478712e-01 1.84460580e-01
-3.70600373e-01 -1.64977595e-01 7.54022360e-01 5.22489190e-01
-2.06118152e-01 -2.54936367e-01 6.44324362e-01 6.91443980e-02
-7.14596629e-01 -1.25346169e-01 -8.20149779e-01 1.85828637e-02
7.38439739e-01 -3.93923335e-02 -3.21294367e-01 -8.47723782e-01
-7.49770939e-01 2.02145308e-01 -5.31213999e-01 6.79054677e-01
1.02750421e+00 -1.10883427e+00 -5.87452769e-01 8.95264745e-02
-8.17797780e-02 -2.98909068e-01 9.67897400e-02 9.28139567e-01
-5.94744124e-02 3.54942232e-01 5.89936078e-01 -7.39460051e-01
-1.33683205e+00 3.63367319e-01 8.71338665e-01 3.07814181e-01
-5.77092230e-01 1.26267016e+00 -3.39431852e-01 -3.09529185e-01
1.13186741e+00 -5.40034711e-01 -2.79861093e-01 1.89131632e-01
5.79562008e-01 3.18811119e-01 4.83271420e-01 -6.61741138e-01
-5.71213543e-01 1.18782766e-01 8.48632008e-02 -2.77689010e-01
1.17345333e+00 2.63971537e-02 2.46780917e-01 5.41772604e-01
1.30071580e+00 2.55250603e-01 -6.71276033e-01 -4.09092933e-01
-8.82579535e-02 -1.38884082e-01 6.55236363e-01 -1.04269326e+00
-1.01908207e+00 6.98156118e-01 8.84944618e-01 6.30639553e-01
1.30042183e+00 1.30445004e-01 8.68056476e-01 5.05463958e-01
-1.18451387e-01 -1.33610904e+00 2.38824338e-01 5.20959139e-01
1.15634727e+00 -1.12756252e+00 -2.05487296e-01 -1.36452124e-01
-4.22135949e-01 8.60614657e-01 5.91762304e-01 5.07045209e-01
9.05839860e-01 9.06439960e-01 2.51745403e-01 -2.64972299e-01
-7.07806528e-01 -4.90839213e-01 2.26109639e-01 4.97080535e-01
9.38154608e-02 -2.37343639e-01 -5.17796911e-03 5.74169993e-01
-3.08965355e-01 -5.10532022e-01 5.36096036e-01 9.53365028e-01
-5.60071707e-01 -5.61552465e-01 -6.68164432e-01 5.75175226e-01
-8.02967012e-01 -5.63682377e-01 -4.70196038e-01 5.91173530e-01
1.83911920e-01 1.59782219e+00 1.45531259e-02 -8.68044853e-01
5.37251651e-01 1.44883560e-03 -1.53468937e-01 -3.26700270e-01
-8.45454514e-01 6.23998642e-01 4.89762396e-01 -5.78683078e-01
-4.56006616e-01 -2.63973325e-01 -1.12884927e+00 5.09118959e-02
-8.27855527e-01 2.12815672e-01 1.19417763e+00 8.94339144e-01
1.91015407e-01 1.33118117e+00 8.08102846e-01 -9.08871472e-01
-4.58848208e-01 -1.54474020e+00 -3.42021286e-01 -2.80011296e-01
8.59595656e-01 -5.55428207e-01 -8.80316079e-01 -3.73585135e-01]
|
[14.69292163848877, 6.120051383972168]
|
4858202d-9c8a-4f70-988a-b1d6b822e5dd
|
tsam-a-two-stream-attention-model-for-causal
|
2203.00819
| null |
https://arxiv.org/abs/2203.00819v2
|
https://arxiv.org/pdf/2203.00819v2.pdf
|
TSAM: A Two-Stream Attention Model for Causal Emotion Entailment
|
Causal Emotion Entailment (CEE) aims to discover the potential causes behind an emotion in a conversational utterance. Previous works formalize CEE as independent utterance pair classification problems, with emotion and speaker information neglected. From a new perspective, this paper considers CEE in a joint framework. We classify multiple utterances synchronously to capture the correlations between utterances in a global view and propose a Two-Stream Attention Model (TSAM) to effectively model the speaker's emotional influences in the conversational history. Specifically, the TSAM comprises three modules: Emotion Attention Network (EAN), Speaker Attention Network (SAN), and interaction module. The EAN and SAN incorporate emotion and speaker information in parallel, and the subsequent interaction module effectively interchanges relevant information between the EAN and SAN via a mutual BiAffine transformation. Extensive experimental results demonstrate that our model achieves new State-Of-The-Art (SOTA) performance and outperforms baselines remarkably.
|
['Jie zhou', 'Xiuyi Chen', 'Fandong Meng', 'Zhen Yang', 'Duzhen Zhang']
|
2022-03-02
| null |
https://aclanthology.org/2022.coling-1.588
|
https://aclanthology.org/2022.coling-1.588.pdf
|
coling-2022-10
|
['causal-emotion-entailment']
|
['natural-language-processing']
|
[ 1.04071319e-01 3.75706047e-01 1.00889981e-01 -9.25461233e-01
-7.02895105e-01 -1.65279090e-01 7.31053293e-01 -1.15176901e-01
-1.39308736e-01 3.33576411e-01 7.49510646e-01 5.71258627e-02
2.12356433e-01 -9.54382643e-02 -4.73433673e-01 -7.07619250e-01
-3.10965121e-01 4.05127943e-01 -4.95715946e-01 -3.93113315e-01
-8.34884048e-02 -1.18877076e-01 -1.32652676e+00 2.74133652e-01
6.97427928e-01 1.01830709e+00 -2.02796251e-01 9.75952983e-01
-1.60372406e-01 1.67424822e+00 -6.27182662e-01 -5.22981942e-01
-4.40151364e-01 -8.52673948e-01 -1.17089355e+00 1.63121074e-01
-3.74988496e-01 -1.31589249e-01 -3.40995282e-01 8.15075219e-01
4.59667504e-01 3.83113295e-01 5.46030045e-01 -1.94371593e+00
-5.22848606e-01 9.34400976e-01 -4.22537535e-01 6.75773248e-02
5.27077317e-01 -1.62443951e-01 1.26120842e+00 -1.08361113e+00
1.30683124e-01 1.71212184e+00 4.00360972e-01 7.06416070e-01
-8.68143380e-01 -6.09713495e-01 7.49628723e-01 4.89928424e-01
-8.52256894e-01 -6.71413779e-01 1.19560575e+00 -1.65260479e-01
1.07122147e+00 4.68472421e-01 4.48926181e-01 1.44233739e+00
8.05809535e-03 1.43467426e+00 7.88776875e-01 -2.80539393e-01
1.75028086e-01 1.47455260e-01 5.24064004e-01 2.59257287e-01
-1.01100791e+00 -2.55740285e-01 -9.10587072e-01 -2.35548541e-01
1.37127608e-01 -1.65190712e-01 -3.05957943e-01 2.28995383e-01
-1.05498052e+00 9.63936150e-01 2.03373209e-01 2.55882680e-01
-8.26521814e-01 5.73111102e-02 6.49711728e-01 6.90340936e-01
7.79207230e-01 1.30231187e-01 -5.50827146e-01 -4.95074540e-01
-1.65446997e-01 -1.13108397e-01 1.09577107e+00 8.35960925e-01
4.28254426e-01 -1.72599241e-01 -1.17926039e-01 9.48279619e-01
4.35761124e-01 1.15667962e-01 3.97940218e-01 -8.14097226e-01
9.08099860e-02 3.56533468e-01 -3.69645730e-02 -8.66303265e-01
-2.93918043e-01 -1.14427261e-01 -8.54867339e-01 -5.87118983e-01
-2.10015297e-01 -6.22279942e-01 -3.36197257e-01 2.12746167e+00
5.49105287e-01 5.76342344e-01 3.69513631e-01 9.49271619e-01
9.66176331e-01 9.48941648e-01 3.31522495e-01 -6.01461768e-01
1.33934510e+00 -1.38536084e+00 -1.32185078e+00 -4.66990501e-01
3.96598220e-01 -6.57595754e-01 8.11346412e-01 2.16670260e-01
-1.12341225e+00 -3.61144036e-01 -7.42775798e-01 -1.22933239e-01
-8.69779140e-02 -2.27344498e-01 5.98948002e-01 4.05417234e-02
-9.45028007e-01 1.66046750e-02 -5.80286860e-01 -2.01603115e-01
-8.46436098e-02 2.82544345e-01 -1.54471910e-02 1.82060510e-01
-1.90733957e+00 8.85830939e-01 -8.18825737e-02 4.62093681e-01
-8.56797576e-01 -4.42313939e-01 -1.02118158e+00 3.29667360e-01
4.73890424e-01 -3.92021686e-01 1.87391949e+00 -1.40459514e+00
-2.24832916e+00 4.49921638e-01 -7.49034107e-01 -2.05655262e-01
-4.14575972e-02 -2.90931970e-01 -6.48178637e-01 -5.87610640e-02
-2.16138437e-01 4.17812526e-01 7.36455202e-01 -1.20958173e+00
-6.54584467e-01 -1.75343677e-01 -1.41095519e-01 5.34299731e-01
-2.99519062e-01 6.57976449e-01 -5.84505320e-01 -2.56240547e-01
-1.83678731e-01 -7.51615465e-01 -5.49688153e-02 -7.35382259e-01
-3.03139508e-01 -8.65112543e-01 8.85069370e-01 -5.80277443e-01
1.34835231e+00 -2.21272159e+00 5.92159331e-01 2.12593377e-02
4.06856835e-01 -1.42244861e-01 -3.29092354e-01 3.92418921e-01
-5.00272095e-01 1.80920458e-03 -4.63193655e-02 -9.89505887e-01
2.80600339e-01 2.14949191e-01 -4.06658947e-01 2.75785685e-01
5.82640409e-01 1.10610676e+00 -1.08613443e+00 -3.96415114e-01
6.92227855e-02 6.27916574e-01 -3.80535662e-01 9.62250233e-01
-1.67758137e-01 5.25359213e-01 -4.13955599e-01 3.79615396e-01
2.33687028e-01 -1.99850112e-01 2.94704974e-01 7.88597614e-02
1.28966365e-02 6.20579720e-01 -7.72791445e-01 1.36805010e+00
-6.79249287e-01 7.19264567e-01 5.39683938e-01 -8.95402789e-01
8.51512313e-01 1.01827335e+00 4.56294864e-01 -5.42333663e-01
5.17721951e-01 -2.19899461e-01 1.45951748e-01 -7.50743389e-01
3.61564726e-01 -3.96711618e-01 -4.96531010e-01 6.59907162e-01
2.83911228e-01 2.66038299e-01 -3.86721671e-01 4.45781291e-01
8.32846463e-01 -3.57634157e-01 2.69319266e-01 6.31232783e-02
6.77777648e-01 -6.93269372e-01 9.30859983e-01 2.79563129e-01
-6.73883975e-01 4.41455096e-02 9.29551840e-01 -2.17180535e-01
-4.75949228e-01 -5.38467348e-01 3.51479679e-01 1.63515496e+00
1.57088712e-01 -3.19351345e-01 -8.01306307e-01 -6.42921269e-01
-4.53825742e-01 9.67344642e-01 -1.02926600e+00 -2.00747326e-01
-3.95767927e-01 -6.23866796e-01 2.84586310e-01 5.41928649e-01
4.39461648e-01 -1.35937238e+00 -2.87034303e-01 3.71150911e-01
-7.56064713e-01 -1.01278329e+00 -8.07078898e-01 2.83972502e-01
-6.66945502e-02 -6.26090467e-01 -3.70222151e-01 -5.80678761e-01
2.17659071e-01 1.07046843e-01 1.18577611e+00 -2.61148214e-01
1.76289916e-01 5.61559081e-01 -4.51853663e-01 -6.45523369e-01
-5.91684699e-01 -2.77369231e-01 1.19117133e-01 7.68383920e-01
7.64850080e-01 -5.97226620e-01 -2.11433247e-01 3.17374766e-01
-5.65478146e-01 6.22916929e-02 2.94559389e-01 9.63221312e-01
-3.89832929e-02 -4.74186838e-02 9.66926277e-01 -6.86682224e-01
8.58290672e-01 -1.05410802e+00 6.10222220e-02 3.14193547e-01
-2.69406885e-01 -1.21204115e-01 4.01363522e-01 -4.90366966e-01
-1.63622701e+00 -2.06506565e-01 -2.29812548e-01 -6.03807271e-01
-4.50915843e-01 6.73237324e-01 -6.56541228e-01 6.53105497e-01
-1.68494195e-01 8.49278048e-02 -1.07232472e-02 -1.44268841e-01
4.08303708e-01 1.13097250e+00 5.80235302e-01 -4.05834377e-01
1.24275289e-01 -1.06720580e-02 -6.36079073e-01 -7.58343995e-01
-1.00073278e+00 -7.75824428e-01 -4.26894277e-01 -7.49498904e-01
9.48507309e-01 -1.01743019e+00 -1.35135889e+00 5.90321481e-01
-1.57121336e+00 -2.98178911e-01 1.14684962e-01 5.54037213e-01
-3.82244706e-01 2.78016776e-02 -9.62598920e-01 -1.47399151e+00
-3.72515321e-01 -9.68539059e-01 1.13816202e+00 3.78937930e-01
-6.96136653e-01 -1.09756064e+00 2.98339706e-02 4.07419086e-01
2.67746508e-01 5.36522754e-02 6.09274983e-01 -1.00402844e+00
-1.38590401e-02 -1.99281827e-01 7.45273381e-02 2.32700139e-01
-6.70791119e-02 1.42701641e-02 -1.42126679e+00 4.47383188e-02
5.79534054e-01 -5.43255568e-01 5.32506227e-01 1.74911812e-01
8.15934777e-01 -3.66361260e-01 -2.33133193e-02 1.05106972e-01
6.37076139e-01 6.65665209e-01 4.77320343e-01 -3.02562267e-01
5.59617817e-01 1.09733975e+00 4.74732727e-01 4.10127580e-01
8.56464863e-01 4.40025300e-01 4.36870307e-01 -3.02965552e-01
4.78023738e-01 -1.85788900e-03 7.99295068e-01 1.60272729e+00
2.24010840e-01 -6.49395108e-01 -6.34106576e-01 5.98604500e-01
-2.19501400e+00 -1.06853247e+00 -1.39545321e-01 1.64863849e+00
7.09638596e-01 -2.21054509e-01 6.02516569e-02 -6.19113147e-02
8.23913753e-01 2.98364073e-01 -6.77341938e-01 -7.35891342e-01
-7.68465847e-02 -2.72560716e-01 -3.84061575e-01 8.17456186e-01
-1.08876765e+00 7.62830675e-01 5.90106106e+00 4.58910465e-01
-9.27548289e-01 3.42220664e-01 8.72690678e-01 -1.49924174e-01
-2.89350331e-01 -2.97837436e-01 -3.60944867e-01 2.85823643e-01
1.34251714e+00 -2.16404855e-01 4.01943237e-01 7.46422231e-01
1.54288560e-01 2.63972461e-01 -1.25315797e+00 8.04814517e-01
2.95024395e-01 -6.58202291e-01 -5.86355984e-01 -1.75542936e-01
4.39297795e-01 5.50843254e-02 -1.28398582e-01 7.19051719e-01
4.43030566e-01 -8.15202355e-01 6.53082609e-01 5.94059765e-01
3.18018883e-01 -9.97612536e-01 9.05726492e-01 2.78659225e-01
-1.19711435e+00 -2.18329299e-02 3.76478225e-01 -3.67036551e-01
4.81133491e-01 1.30647257e-01 -8.55647326e-01 5.38485706e-01
6.33950353e-01 8.04597855e-01 1.62610233e-01 1.37069598e-01
-4.78175938e-01 8.70475411e-01 -1.18174199e-02 -4.23399508e-01
3.49808216e-01 -8.93752873e-02 7.29514480e-01 1.28150105e+00
-2.55360425e-01 5.70122838e-01 3.20036039e-02 7.02928603e-01
-2.29562685e-01 4.50157747e-02 -3.03395987e-02 -1.44666255e-01
5.63634872e-01 1.37582982e+00 -2.74962485e-01 -5.28518379e-01
-5.18118501e-01 1.32293212e+00 4.50915694e-01 4.45588171e-01
-1.02986658e+00 -3.26734990e-01 1.08776855e+00 -9.31550026e-01
9.08619985e-02 3.97566259e-01 2.12640271e-01 -1.27368510e+00
-1.13870740e-01 -1.00276268e+00 4.06713486e-01 -8.39185536e-01
-1.61009598e+00 8.77859831e-01 -3.53621930e-01 -8.15991998e-01
-4.77295876e-01 -1.07701071e-01 -1.15591812e+00 9.06885803e-01
-1.32403469e+00 -9.38064039e-01 -5.49543835e-03 5.33167899e-01
8.81035686e-01 2.16808021e-01 1.16391456e+00 1.55939952e-01
-1.16911650e+00 5.54641545e-01 -2.38954753e-01 2.67459959e-01
6.85762763e-01 -1.40611637e+00 3.11148584e-01 6.64980471e-01
-2.39373714e-01 6.36726677e-01 6.89145744e-01 -3.72907043e-01
-1.34323704e+00 -9.12604749e-01 1.43500173e+00 -4.15510297e-01
7.98414350e-01 -6.35873973e-01 -1.09330022e+00 9.52427804e-01
8.54720831e-01 -5.36858737e-01 1.05892920e+00 6.90466225e-01
-3.81833315e-01 1.99436605e-01 -5.97052157e-01 5.64657211e-01
5.54130375e-01 -8.30844045e-01 -8.82051289e-01 1.00730807e-01
1.22585964e+00 -2.41090626e-01 -9.43403363e-01 2.12668076e-01
3.95287454e-01 -8.63500834e-01 5.48695862e-01 -8.42918754e-01
6.66546881e-01 1.46972135e-01 -7.92657956e-02 -1.58774149e+00
-1.98790491e-01 -1.25676286e+00 -2.51754731e-01 1.52174067e+00
3.72222573e-01 -6.02583170e-01 1.39922723e-01 1.09377360e+00
-2.86478221e-01 -7.99306870e-01 -9.04310942e-01 -1.45878941e-01
-2.89493024e-01 -6.07978642e-01 8.52299511e-01 1.35566878e+00
7.75767386e-01 1.25392711e+00 -7.66720712e-01 3.82388413e-01
2.48798355e-01 1.09050117e-01 5.45669675e-01 -1.07070088e+00
-2.54171163e-01 -6.72835469e-01 1.09538585e-01 -1.04462039e+00
7.64687121e-01 -4.72053289e-01 5.85301936e-01 -1.06390834e+00
3.20866048e-01 1.70451194e-01 -3.75804245e-01 3.80144119e-01
-7.55716205e-01 -5.30850828e-01 5.89069724e-02 -6.33784160e-02
-9.71928060e-01 1.22553074e+00 7.63993859e-01 3.20119374e-02
-2.81776249e-01 -9.61373821e-02 -6.45645976e-01 8.38317871e-01
5.61221421e-01 -2.98251867e-01 -3.21827292e-01 -1.50920361e-01
-6.58592880e-02 4.97910380e-01 9.78908613e-02 -2.01301187e-01
5.53412795e-01 -1.43799884e-02 -1.88246369e-01 -4.27840710e-01
4.88316983e-01 -6.16107225e-01 -1.87740654e-01 -6.14350699e-02
-8.33738983e-01 -1.55974776e-01 3.79798710e-02 6.87001526e-01
-7.05532670e-01 2.93855876e-01 3.79956901e-01 2.93263704e-01
-5.25035739e-01 3.39223146e-01 -6.96843386e-01 -1.43753558e-01
8.01409900e-01 5.53426683e-01 -4.32584696e-02 -1.10230768e+00
-7.19894171e-01 9.04335976e-01 -4.21700060e-01 7.94331610e-01
4.95597839e-01 -1.35700119e+00 -7.83318222e-01 1.37560010e-01
1.34439752e-01 -6.92497566e-02 6.83875918e-01 9.70034063e-01
6.77464426e-01 3.46558958e-01 3.80279630e-01 -3.02591711e-01
-1.53535342e+00 5.31606913e-01 3.93617928e-01 -2.36934721e-01
-8.81548389e-04 1.38146472e+00 4.79423642e-01 -6.45793796e-01
6.17058933e-01 -2.38716714e-02 -3.96469891e-01 3.80545288e-01
7.22020090e-01 1.44581392e-01 -3.18840206e-01 -9.28324759e-01
-3.74085337e-01 -2.13917680e-02 -1.48750857e-01 -2.71833092e-01
1.41268182e+00 -6.65068507e-01 -4.57090139e-01 1.04199076e+00
1.54831040e+00 -2.51122475e-01 -1.09729731e+00 -6.40643418e-01
-3.83274853e-02 2.06565663e-01 2.77409047e-01 -7.57382751e-01
-8.34876597e-01 9.29543972e-01 -5.46583273e-02 4.78368700e-01
1.34218216e+00 3.29020500e-01 8.32977712e-01 1.40478343e-01
-2.91100144e-01 -1.04666305e+00 1.87451959e-01 8.97465765e-01
1.16721225e+00 -1.31612909e+00 -7.83930242e-01 -3.43935370e-01
-1.20781600e+00 8.91174078e-01 7.36703396e-01 3.43810827e-01
6.98853016e-01 4.03155595e-01 3.49858373e-01 -3.89831007e-01
-1.55459845e+00 -1.34307116e-01 2.74860650e-01 2.07332652e-02
6.90948308e-01 2.23500505e-01 8.98358375e-02 1.35947490e+00
4.31951359e-02 -3.21608156e-01 2.14153126e-01 6.84839129e-01
5.16560674e-02 -8.38805735e-01 -1.91354737e-01 -1.88036375e-02
-3.14437956e-01 -2.43255049e-01 -7.79165268e-01 4.01659191e-01
-1.94687396e-01 1.45464277e+00 2.99524635e-01 -7.94421077e-01
2.56361663e-01 6.54476821e-01 -2.42346272e-01 -2.76568532e-01
-1.00494218e+00 6.00919545e-01 3.52498651e-01 -6.81032240e-01
-5.94881892e-01 -8.01846683e-01 -1.28341651e+00 -1.63436592e-01
-3.84448320e-01 5.63050866e-01 5.43836415e-01 1.23762000e+00
5.13510525e-01 1.09319806e+00 1.33580053e+00 -7.28580713e-01
-2.63598055e-01 -1.34818041e+00 -4.54403460e-01 5.03461957e-01
5.45761883e-01 -3.34247231e-01 -7.09065199e-01 8.21064115e-02]
|
[13.099064826965332, 6.019595623016357]
|
7b952281-6732-4914-81b3-5b05c69e0ea0
|
hrca-advanced-multiple-choice-machine-reading
| null | null |
https://aclanthology.org/2022.lrec-1.651
|
https://aclanthology.org/2022.lrec-1.651.pdf
|
HRCA+: Advanced Multiple-choice Machine Reading Comprehension Method
|
Multiple-choice question answering (MCQA) for machine reading comprehension (MRC) is challenging. It requires a model to select a correct answer from several candidate options related to text passages or dialogue. To select the correct answer, such models must have the ability to understand natural languages, comprehend textual representations, and infer the relationship between candidate options, questions, and passages. Previous models calculated representations between passages and question-option pairs separately, thereby ignoring the effect of other relation-pairs. In this study, we propose a human reading comprehension attention (HRCA) model and a passage-question-option (PQO) matrix-guided HRCA model called HRCA+ to increase accuracy. The HRCA model updates the information learned from the previous relation-pair to the next relation-pair. HRCA+ utilizes the textual information and the interior relationship between every two parts in a passage, a question, and the corresponding candidate options. Our proposed method outperforms other state-of-the-art methods. On the Semeval-2018 Task 11 dataset, our proposed method improved accuracy levels from 95.8% to 97.2%, and on the DREAM dataset, it improved accuracy levels from 90.4% to 91.6% without extra training data, from 91.8% to 92.6% with extra training data.
|
['Hayato Yamana', 'Yuxiang Zhang']
| null | null | null | null |
lrec-2022-6
|
['multiple-choice-qa', 'machine-reading-comprehension']
|
['natural-language-processing', 'natural-language-processing']
|
[ 4.42370921e-01 1.95869058e-01 1.07956283e-01 -4.60067481e-01
-1.16138816e+00 -4.34363604e-01 3.65751088e-01 6.71650469e-01
-5.59968174e-01 7.19936907e-01 5.61793149e-01 -5.78305185e-01
-1.81727663e-01 -8.38301003e-01 -5.97896218e-01 6.29889546e-04
3.33206743e-01 5.32639503e-01 4.74576056e-01 -4.65229124e-01
4.79097098e-01 -4.00860548e-01 -1.33845115e+00 6.35814130e-01
1.48129821e+00 1.10001612e+00 6.03959322e-01 7.37507999e-01
-6.03870213e-01 1.12612402e+00 -4.76007998e-01 -4.16118920e-01
-2.30419561e-01 -1.01362193e+00 -1.35880685e+00 -3.66367131e-01
5.57728231e-01 -3.52158993e-01 -1.97850093e-01 7.51001120e-01
1.83864266e-01 5.77768087e-01 6.30704165e-01 -6.74452901e-01
-8.47076774e-01 6.37937903e-01 -3.10368925e-01 3.62133980e-01
8.59516442e-01 -9.13135335e-02 1.18849659e+00 -1.03778493e+00
2.64293194e-01 1.28450644e+00 1.09150797e-01 6.40708923e-01
-9.89490330e-01 -3.64485919e-01 3.96455944e-01 8.64327252e-01
-9.29087758e-01 -1.02139212e-01 5.41604280e-01 -2.11152717e-01
1.01461577e+00 5.53757250e-01 4.86967593e-01 8.22968185e-01
1.34046718e-01 9.23943043e-01 1.10839140e+00 -6.34788573e-01
9.34608579e-02 -2.83720344e-01 8.63136113e-01 6.34430826e-01
-3.77070516e-01 -5.21711051e-01 -5.81350863e-01 -7.18022287e-02
3.15653235e-01 -2.48278126e-01 -5.62818646e-01 4.21265244e-01
-1.25972140e+00 9.06181633e-01 4.86207753e-01 1.25415996e-01
-4.72129554e-01 -4.74873215e-01 1.05469907e-02 4.86486226e-01
2.64461394e-02 7.53386915e-01 -6.83146775e-01 -1.72193781e-01
-5.58424950e-01 2.91808665e-01 8.87842178e-01 7.20556140e-01
6.68958008e-01 -5.47586441e-01 -8.93445790e-01 1.01931167e+00
3.17853540e-01 3.74889404e-01 4.07585025e-01 -9.20785964e-01
1.08380437e+00 7.22899616e-01 1.31765559e-01 -1.03820240e+00
-4.51156437e-01 -4.13482100e-01 -8.01081955e-01 -6.40930116e-01
4.53413457e-01 -1.74579531e-01 -6.65715218e-01 1.75487030e+00
3.05181473e-01 -1.48421288e-01 1.22842766e-01 8.35517645e-01
1.16752839e+00 1.18461585e+00 1.38883695e-01 -2.90622503e-01
1.64123952e+00 -1.33030963e+00 -9.42095816e-01 -5.05276024e-01
3.89921248e-01 -7.61930883e-01 1.53062379e+00 3.16707075e-01
-1.06730986e+00 -7.58208871e-01 -8.31416368e-01 -5.88823378e-01
2.23089471e-01 1.97394192e-01 -2.22966038e-02 -2.00676501e-01
-7.05281973e-01 2.53023952e-01 -2.19342813e-01 -7.87840560e-02
7.75560215e-02 9.56474629e-04 -8.62955227e-02 -5.79094946e-01
-1.58603120e+00 9.70441997e-01 2.36170024e-01 -3.69644687e-02
-4.87424284e-01 -6.75594568e-01 -8.11949134e-01 5.13089538e-01
5.84551990e-01 -8.27456892e-01 1.53154731e+00 -8.33823204e-01
-1.52411830e+00 3.63718510e-01 -7.05372930e-01 -4.50196564e-01
3.67147196e-03 -6.41646922e-01 -3.53107691e-01 4.99924988e-01
1.42045826e-01 6.87348008e-01 5.70697188e-01 -7.44704902e-01
-5.20304680e-01 -1.97302386e-01 1.97466433e-01 5.76809049e-01
1.61681816e-01 -2.63414621e-01 -4.39243436e-01 -4.02303427e-01
4.07088667e-01 -6.07104897e-01 3.37734469e-03 -2.83980638e-01
-2.82803625e-01 -7.42235124e-01 2.19019189e-01 -1.40504146e+00
1.47321904e+00 -1.79119217e+00 4.03619200e-01 -1.84265867e-01
1.31840035e-01 1.76901385e-01 -5.10987282e-01 6.10311389e-01
3.03586155e-01 3.42326723e-02 -2.99715489e-01 -1.38228044e-01
-7.85285532e-02 -1.55697599e-01 -3.64340246e-01 -1.33340001e-01
3.22872669e-01 8.64035964e-01 -8.66522849e-01 -3.41625422e-01
-6.19586632e-02 4.54264320e-02 -4.68678772e-01 6.91581905e-01
-7.73653924e-01 4.56381142e-01 -4.38189149e-01 2.83614606e-01
5.16109765e-01 -4.84360427e-01 -4.95550968e-02 -1.48366634e-02
2.82775611e-01 9.52302277e-01 -7.57088423e-01 1.63408029e+00
-6.45674646e-01 6.59801424e-01 -2.83149749e-01 -5.35970569e-01
1.13228285e+00 1.06980957e-01 -2.67019659e-01 -1.28939092e+00
6.18314296e-02 1.40450476e-02 2.39999607e-01 -9.27590787e-01
5.89628875e-01 1.54256508e-01 1.75110286e-03 3.96606594e-01
2.11803745e-02 1.48182094e-01 2.33306319e-01 6.16057575e-01
9.52525258e-01 -1.80145100e-01 4.80436891e-01 -1.05452873e-01
9.20478284e-01 -1.28030524e-01 2.29936346e-01 8.56677771e-01
1.49510458e-01 6.53644621e-01 5.89488804e-01 -2.62900293e-02
-4.48138416e-01 -8.81936133e-01 2.37560377e-01 1.23706162e+00
1.60508305e-01 -5.87616742e-01 -7.34014213e-01 -6.23006105e-01
-3.71766359e-01 1.36750340e+00 -6.25609100e-01 -2.45001227e-01
-7.10710287e-01 -9.72385425e-03 3.62632051e-02 3.80696028e-01
8.73291671e-01 -9.87681508e-01 -3.29748869e-01 3.13185126e-01
-1.03866696e+00 -9.93043482e-01 -6.64818704e-01 -2.28791714e-01
-7.26180017e-01 -1.21728671e+00 -3.44473243e-01 -9.43475544e-01
4.54597741e-01 2.00801983e-01 1.35493600e+00 4.71434832e-01
4.67318803e-01 3.51229280e-01 -7.86280334e-01 -8.01417530e-02
-2.38452986e-01 2.61348903e-01 -6.02882385e-01 2.18756959e-01
3.28202814e-01 -7.41887018e-02 -6.51383162e-01 2.82290459e-01
-6.88389361e-01 4.08714324e-01 5.28172195e-01 9.77683187e-01
6.82605267e-01 -5.68986297e-01 9.19760287e-01 -6.52244091e-01
9.89231586e-01 -8.63866508e-01 -2.12178692e-01 6.23868942e-01
-3.12514067e-01 1.98220536e-01 7.64039576e-01 -2.52896339e-01
-1.14108264e+00 -4.12600815e-01 -5.15539765e-01 3.61136459e-02
-4.34454083e-02 9.58963752e-01 -2.84742743e-01 7.37441540e-01
5.63623726e-01 3.24823916e-01 6.29190216e-03 -6.50333524e-01
2.77046055e-01 6.25365853e-01 5.33419549e-01 -5.54668128e-01
3.05214345e-01 -2.96584576e-01 -3.08673322e-01 -5.57759762e-01
-1.53493810e+00 -3.85721505e-01 -4.67388093e-01 -1.32446289e-01
9.70143914e-01 -7.43999422e-01 -5.17826736e-01 3.10900956e-01
-1.34469318e+00 -2.00148806e-01 4.39031236e-02 4.31634694e-01
-1.10831276e-01 4.37407196e-01 -4.57766026e-01 -6.25340521e-01
-5.20942032e-01 -8.55455875e-01 5.80763817e-01 6.30389810e-01
-4.86998558e-01 -8.09476435e-01 -2.26941660e-01 9.51143920e-01
5.63080370e-01 -2.02608317e-01 1.35466671e+00 -9.43564296e-01
-5.34091532e-01 1.73305050e-02 -1.46349430e-01 1.72935739e-01
1.02262497e-02 -3.93711895e-01 -6.01688266e-01 -1.63791850e-01
2.42040709e-01 -4.97904927e-01 1.07715428e+00 9.54334717e-03
1.37714887e+00 -5.88015020e-01 -5.49338059e-03 -1.01363376e-01
8.21788251e-01 1.19661748e-01 7.33061075e-01 1.93385199e-01
3.85474175e-01 7.62964547e-01 8.32390726e-01 1.24499135e-01
1.03952491e+00 4.65398729e-01 3.99558544e-01 4.71286535e-01
-1.54258817e-01 -5.30745625e-01 1.86333343e-01 1.20862436e+00
4.40726846e-01 -4.93477851e-01 -9.35664892e-01 6.09213173e-01
-1.65224242e+00 -8.36802304e-01 -5.49602211e-01 2.18284702e+00
1.12481987e+00 3.35261077e-02 -3.23089689e-01 4.73312996e-02
5.68269193e-01 2.58996803e-02 -6.37014985e-01 -2.94939399e-01
-3.99412401e-02 4.37780142e-01 -4.05606776e-01 9.46242690e-01
-7.86856115e-01 7.22500384e-01 4.47277355e+00 8.23244691e-01
-5.76118588e-01 -5.61334677e-02 6.96562469e-01 9.27120373e-02
-6.38924837e-01 1.27943844e-01 -7.27188528e-01 2.70054162e-01
9.34142709e-01 -1.33935168e-01 5.46008229e-01 2.91721106e-01
7.17155337e-02 -6.39664650e-01 -1.14778745e+00 7.24515140e-01
4.73239690e-01 -1.19130278e+00 3.17139953e-01 -6.62862420e-01
4.14934307e-01 -3.38204741e-01 -1.26184940e-01 6.72938168e-01
-2.52081722e-01 -1.36213577e+00 4.37287509e-01 9.23109412e-01
4.39491928e-01 -6.00798070e-01 7.78572023e-01 7.75164604e-01
-1.01873839e+00 -1.30770117e-01 -2.74963647e-01 -4.29127097e-01
1.19130857e-01 2.78478563e-01 -6.56302631e-01 6.18852556e-01
5.23613334e-01 5.47441781e-01 -9.62123513e-01 9.02673483e-01
-7.40250170e-01 1.13158119e+00 -1.08662330e-01 -5.19958556e-01
1.91838682e-01 -1.72546744e-01 4.25830960e-01 7.93194175e-01
1.72853246e-01 6.59763455e-01 -6.08608052e-02 7.38424659e-01
-2.42317453e-01 4.57795650e-01 2.63417125e-01 2.18076810e-01
7.54885435e-01 8.85453343e-01 -7.41676018e-02 -2.86133140e-01
-4.71190184e-01 9.81233597e-01 7.92240858e-01 3.34143788e-01
-5.72121263e-01 -5.54426372e-01 1.08761810e-01 4.59279157e-02
2.53195226e-01 -1.41749084e-01 -2.21778154e-01 -1.21619523e+00
2.18050122e-01 -1.19114518e+00 7.20181406e-01 -8.87999713e-01
-1.45917761e+00 6.59845531e-01 -6.73532933e-02 -9.86853778e-01
-1.52376771e-01 -2.27175504e-01 -7.71049619e-01 1.20367837e+00
-1.40169799e+00 -6.34657502e-01 -4.24020529e-01 4.60657477e-01
1.00954485e+00 7.59521276e-02 6.91070259e-01 -8.22593197e-02
-3.26981068e-01 6.04878306e-01 -4.84283753e-02 1.79789841e-01
5.77585936e-01 -1.17045772e+00 2.34159097e-01 7.41520703e-01
1.68307960e-01 7.11339414e-01 3.36979985e-01 -4.31179911e-01
-1.17137098e+00 -8.43615532e-01 1.48692739e+00 -3.91346604e-01
2.09250838e-01 -3.72394286e-02 -1.65333855e+00 4.65750724e-01
7.42870510e-01 -6.74620330e-01 8.28393042e-01 2.08829701e-01
-3.45619172e-01 -1.86860608e-03 -7.68649817e-01 6.72765017e-01
5.17742515e-01 -5.94465315e-01 -1.32955623e+00 9.58720446e-02
1.14249575e+00 -6.05503798e-01 -5.87073684e-01 2.56597757e-01
1.80732921e-01 -6.75666630e-01 6.04284465e-01 -6.73641205e-01
7.30157852e-01 -2.27688134e-01 -2.62845635e-01 -1.30764341e+00
-3.06724042e-01 -2.57836998e-01 -3.74475926e-01 1.11545813e+00
7.63482332e-01 -4.35451508e-01 -5.72367944e-02 6.68367326e-01
-2.68716007e-01 -1.11040747e+00 -1.08705008e+00 -5.27254157e-02
4.10889626e-01 -2.33004346e-01 6.39850676e-01 7.37607121e-01
4.45597135e-02 1.04450977e+00 -1.88378736e-01 1.70341462e-01
7.21997470e-02 3.47666621e-01 4.63133872e-01 -1.02156222e+00
-3.65009278e-01 -2.60663629e-01 4.99762326e-01 -1.81660509e+00
1.24610476e-01 -8.57388258e-01 2.01786235e-02 -2.18654990e+00
2.10013941e-01 1.90533083e-02 -1.51721984e-01 2.72286057e-01
-9.87948120e-01 -5.53795516e-01 4.33420271e-01 -1.71111003e-02
-7.88439095e-01 9.60972071e-01 1.47996926e+00 -1.86414897e-01
-1.84146404e-01 9.27636996e-02 -7.79258132e-01 4.04293090e-01
1.10893381e+00 -3.12094420e-01 -5.58310509e-01 -7.17799962e-01
3.99913043e-01 6.37877464e-01 3.13591152e-01 -8.11223090e-01
4.42966551e-01 -1.56569198e-01 3.82785678e-01 -8.60691428e-01
3.66532624e-01 -2.85843581e-01 -4.33088273e-01 2.60486454e-01
-9.99405622e-01 2.29181841e-01 1.53098211e-01 6.37687445e-01
-3.30129385e-01 -4.89035755e-01 4.14288670e-01 -8.25749114e-02
-5.46381772e-01 -1.01092108e-01 -5.32962799e-01 6.49439037e-01
5.36051512e-01 2.96994984e-01 -5.67049623e-01 -8.02435040e-01
-6.53763771e-01 9.01225626e-01 -7.62076676e-02 7.06859291e-01
9.47575986e-01 -1.08737755e+00 -1.01721668e+00 -4.55617383e-02
6.68411702e-02 2.56607831e-01 5.58804870e-01 6.17930591e-01
-2.45722532e-01 4.56896037e-01 1.41218640e-02 -3.61366957e-01
-1.27780044e+00 2.36519337e-01 4.03054267e-01 -3.96343499e-01
-3.09465617e-01 9.60826814e-01 -8.20422471e-02 -5.22444546e-01
2.41991550e-01 -4.62250143e-01 -8.94969523e-01 9.68965441e-02
8.14789474e-01 3.61188382e-01 1.54294372e-02 -3.58209938e-01
-7.04618171e-02 5.05909026e-01 -3.72888297e-01 -1.12925552e-01
7.63661742e-01 -4.48136866e-01 -2.54968047e-01 5.21429479e-01
1.00906229e+00 -5.67607060e-02 -8.90964985e-01 -4.77143914e-01
4.05270457e-02 -3.85771394e-01 -1.21689461e-01 -1.36466753e+00
-5.81886232e-01 1.06831002e+00 8.33708569e-02 3.08381915e-02
1.16343045e+00 2.20512241e-01 9.82149899e-01 6.50454342e-01
1.55072547e-02 -7.91065812e-01 4.30566162e-01 1.07497036e+00
1.37601745e+00 -1.17906678e+00 -2.33034670e-01 -3.81774068e-01
-8.90476108e-01 1.09419799e+00 1.25653481e+00 2.11710215e-01
2.51955837e-01 -8.18118930e-01 2.49488163e-03 -6.53952286e-02
-1.20884776e+00 -4.31333960e-04 8.16419661e-01 8.65463987e-02
5.21990120e-01 1.60429664e-02 -5.49511790e-01 1.10539043e+00
-2.87827820e-01 -3.53160232e-01 5.87226212e-01 6.63153768e-01
-7.17434943e-01 -8.50519478e-01 -2.61581808e-01 8.51383686e-01
-7.08098412e-02 -4.37294602e-01 -3.96441638e-01 3.59370619e-01
-1.13674901e-01 1.55856967e+00 2.10594147e-01 -2.61096090e-01
3.69068980e-01 3.50342542e-01 2.62671143e-01 -6.18178844e-01
-7.79613674e-01 -3.77580315e-01 2.99137414e-01 -3.44872087e-01
-1.84763119e-01 -6.78756475e-01 -1.49781787e+00 -1.09557211e-01
-3.63205820e-01 5.15740752e-01 1.00773618e-01 1.28150713e+00
6.24793470e-01 6.00063205e-01 5.19469976e-01 1.09842278e-01
-6.90338016e-01 -1.23504281e+00 3.76541652e-02 2.23758116e-01
4.16310787e-01 -2.82701552e-01 -3.25980097e-01 -2.48736218e-01]
|
[11.393810272216797, 8.063148498535156]
|
fc84e0ae-193e-46da-90dd-b96c950a0593
|
dialog-policy-learning-for-joint
|
2006.05456
| null |
https://arxiv.org/abs/2006.05456v3
|
https://arxiv.org/pdf/2006.05456v3.pdf
|
Dialog Policy Learning for Joint Clarification and Active Learning Queries
|
Intelligent systems need to be able to recover from mistakes, resolve uncertainty, and adapt to novel concepts not seen during training. Dialog interaction can enable this by the use of clarifications for correction and resolving uncertainty, and active learning queries to learn new concepts encountered during operation. Prior work on dialog systems has either focused on exclusively learning how to perform clarification/ information seeking, or to perform active learning. In this work, we train a hierarchical dialog policy to jointly perform both clarification and active learning in the context of an interactive language-based image retrieval task motivated by an online shopping application, and demonstrate that jointly learning dialog policies for clarification and active learning is more effective than the use of static dialog policies for one or both of these functions.
|
['Raymond J. Mooney', 'Aishwarya Padmakumar']
|
2020-06-09
| null | null | null | null |
['novel-concepts']
|
['reasoning']
|
[ 1.98794857e-01 6.17101848e-01 -2.43478477e-01 -9.16630030e-01
-8.59103799e-01 -8.82482171e-01 6.06472552e-01 7.52489626e-01
-5.86536169e-01 6.53041065e-01 2.93917537e-01 -5.48466623e-01
-3.90472233e-01 -4.32192892e-01 -2.37062529e-01 -2.19363555e-01
7.31383637e-03 1.00727391e+00 3.41513187e-01 -3.23933452e-01
5.54405034e-01 5.39806187e-01 -1.29287541e+00 2.78438956e-01
9.83699739e-01 4.77079779e-01 4.44142967e-01 7.29594648e-01
-7.64987290e-01 1.16438150e+00 -7.79607773e-01 3.03814257e-03
4.11627907e-03 2.25689784e-02 -1.16046488e+00 4.38332915e-01
-2.24330544e-01 -1.01498330e+00 3.80990505e-01 6.14911318e-01
1.02469318e-01 7.26850212e-01 3.55533004e-01 -1.20990014e+00
-2.53863692e-01 7.91074753e-01 1.72422379e-01 2.67518401e-01
5.82031786e-01 3.44655514e-01 8.08424413e-01 -5.45588791e-01
3.52028459e-01 1.38384831e+00 1.53853238e-01 7.41612434e-01
-1.17610765e+00 -5.33957422e-01 5.59192479e-01 -6.36264533e-02
-8.41537893e-01 -6.48969531e-01 2.46393889e-01 -5.01992524e-01
1.15241838e+00 6.16638251e-02 2.80286521e-01 7.60596097e-01
-3.17241699e-02 1.02281916e+00 3.66853267e-01 -6.25059128e-01
3.79932195e-01 8.52813840e-01 6.32337213e-01 4.86733794e-01
-1.26844108e-01 1.19937696e-01 -4.70957994e-01 -4.47170824e-01
5.67898452e-01 5.68235386e-03 -1.24551088e-01 -3.44895869e-01
-7.37118661e-01 8.41410816e-01 -4.66542765e-02 1.26988485e-01
-2.55267054e-01 -2.32204542e-01 2.13665664e-01 6.97142005e-01
1.41001716e-01 1.09380412e+00 -8.06160033e-01 -3.22634161e-01
-3.20685506e-01 9.43287686e-02 1.25414455e+00 9.20854509e-01
9.20209527e-01 -3.12309358e-02 -1.33784488e-01 7.08674848e-01
5.65203071e-01 6.99596247e-03 2.65992284e-01 -1.40350509e+00
2.67233495e-02 8.02969992e-01 5.98406553e-01 -1.52161077e-01
-2.83354670e-01 1.75180212e-01 3.63928258e-01 5.06540835e-01
3.35038155e-01 -3.77324194e-01 -8.32661927e-01 1.46353304e+00
1.92264259e-01 -3.00752610e-01 3.40089142e-01 5.06924450e-01
5.32972872e-01 3.76810193e-01 4.00896072e-01 -5.97315073e-01
7.77964890e-01 -6.88415647e-01 -8.88710916e-01 -2.82508492e-01
7.33069241e-01 -7.87940502e-01 9.82381463e-01 4.73639935e-01
-1.18406165e+00 -5.51646948e-01 -9.23220217e-01 -3.63502540e-02
-3.47510517e-01 -4.16257262e-01 3.05025578e-01 2.87704617e-01
-8.87507915e-01 2.55183905e-01 -1.03800786e+00 -3.10511470e-01
-1.12045020e-01 5.89254737e-01 2.17846498e-01 1.64310664e-01
-1.09217548e+00 9.49285805e-01 4.82402235e-01 -2.90341020e-01
-6.90784991e-01 -5.94880700e-01 -8.43807220e-01 4.17107224e-01
9.81579781e-01 -3.25317115e-01 2.14696503e+00 -9.99980271e-01
-1.54011822e+00 2.52248764e-01 2.52551168e-01 -5.18718839e-01
2.32806370e-01 -6.59436047e-01 -1.86492279e-02 1.20734431e-01
-2.61957258e-01 9.00697649e-01 6.22992873e-01 -1.50262582e+00
-7.90244222e-01 -3.34003687e-01 5.06849587e-01 7.81727195e-01
-2.54790753e-01 -4.75469716e-02 -4.37524557e-01 -7.25579401e-03
-3.74847911e-02 -7.85170317e-01 -2.45444134e-01 -1.11316748e-01
1.03243016e-01 -3.33828568e-01 1.36059177e+00 -5.25489509e-01
9.29346502e-01 -2.07521772e+00 -1.76125616e-01 2.81951606e-01
2.70097535e-02 2.97209501e-01 8.34866539e-02 4.85223800e-01
2.60685921e-01 -2.76262723e-02 8.75973776e-02 -4.45557952e-01
1.98095348e-02 2.92080373e-01 -5.37385523e-01 -5.23617864e-01
8.20046440e-02 4.01274621e-01 -8.41885626e-01 -4.38362688e-01
4.62954521e-01 2.90605240e-04 -5.56164026e-01 6.84938967e-01
-9.11057055e-01 7.29244649e-01 -4.21083450e-01 3.92318338e-01
7.84377083e-02 -3.72520119e-01 6.66647017e-01 3.58105421e-01
-1.09636560e-01 3.51557553e-01 -9.71364856e-01 1.42653441e+00
-7.40079165e-01 3.13405722e-01 4.67711538e-01 -6.82274044e-01
8.30619276e-01 3.89055818e-01 4.67939049e-01 -5.65412164e-01
1.71017587e-01 -1.89866245e-01 -3.11657526e-02 -7.01687694e-01
5.93030035e-01 2.69733459e-01 -3.82156856e-02 7.75525331e-01
1.34835109e-01 -2.85840958e-01 2.14233100e-02 5.90144336e-01
7.48667598e-01 -1.27180502e-01 3.14526260e-01 -2.27727666e-02
2.75777340e-01 2.32399598e-01 2.62032270e-01 1.11921084e+00
-2.15678453e-01 1.99482609e-02 1.59574479e-01 -5.46632633e-02
-3.97437781e-01 -9.66772437e-01 2.83239275e-01 1.54510558e+00
2.67259151e-01 -3.56842816e-01 -3.31480801e-01 -9.29075539e-01
9.08552706e-02 1.50212991e+00 2.01847106e-01 -2.23602906e-01
-4.19165522e-01 -1.80434152e-01 -2.89491802e-01 5.27023196e-01
7.12626696e-01 -1.01465213e+00 -7.60441184e-01 2.22684279e-01
9.51610282e-02 -8.44492793e-01 -4.89778906e-01 3.37546915e-01
-1.02484214e+00 -1.00358558e+00 -2.11294629e-02 -6.60285294e-01
4.99571979e-01 -2.80227717e-02 9.78067398e-01 5.91603518e-01
-6.04200624e-02 1.28198040e+00 -2.29200289e-01 -5.05523264e-01
-8.39003921e-01 -1.30567804e-01 -2.42098942e-01 -5.20867884e-01
3.83500546e-01 -3.00901651e-01 -3.88144284e-01 5.19851029e-01
-8.84860218e-01 -4.16600471e-03 4.71248031e-01 9.58510935e-01
-7.96497315e-02 4.62616868e-02 6.36402905e-01 -1.19237316e+00
1.08730471e+00 -2.86842495e-01 -8.37894142e-01 7.24358797e-01
-1.07632828e+00 1.88513026e-01 9.88707617e-02 -4.78984177e-01
-1.83681417e+00 6.05316758e-01 -1.48649618e-01 -1.87829986e-01
-4.30104822e-01 5.64871609e-01 -5.70907220e-02 1.50341615e-01
6.48212254e-01 -3.04540187e-01 9.41767991e-02 -2.85136461e-01
3.79579931e-01 8.55922222e-01 3.44291091e-01 -7.45457292e-01
3.49240273e-01 -1.01454593e-01 -8.79285097e-01 -5.50994635e-01
-7.28389382e-01 -7.06424177e-01 -5.75639009e-01 -1.91831023e-01
7.00274527e-01 -7.40653634e-01 -9.82274771e-01 1.05452426e-01
-7.38824487e-01 -7.89440095e-01 -4.18689132e-01 5.07290423e-01
-5.44360042e-01 3.47726583e-01 -4.69608307e-01 -1.15736043e+00
-1.88200057e-01 -1.17787361e+00 5.40227473e-01 6.17398679e-01
-6.47057235e-01 -1.13169658e+00 -1.55424118e-01 5.46067476e-01
6.33602321e-01 -5.55862963e-01 1.19104028e+00 -1.26475906e+00
-9.32802916e-01 -1.96528777e-01 2.76297301e-01 8.90862793e-02
3.45091671e-01 -1.54559150e-01 -7.67299473e-01 -3.75011235e-01
2.34839454e-01 -7.56482303e-01 4.26782429e-01 9.54943746e-02
8.55941474e-01 -4.79376227e-01 -2.46625036e-01 -3.64522576e-01
8.51199031e-01 9.25006092e-01 1.57822967e-01 1.60479546e-01
-1.16937511e-01 7.75023818e-01 9.20012891e-01 4.61023092e-01
3.53645623e-01 5.46088338e-01 1.39482901e-01 2.38979384e-01
3.24489772e-01 -5.99533319e-02 -3.37600894e-02 2.73245722e-02
6.72599435e-01 -1.97984487e-01 -1.09007859e+00 3.47478628e-01
-2.12720156e+00 -5.68606555e-01 7.71859407e-01 2.24510813e+00
9.22366142e-01 6.46527052e-01 -7.65610263e-02 -2.48933539e-01
5.86634517e-01 -3.10981661e-01 -1.06522894e+00 -3.28101099e-01
5.19846439e-01 -1.19557902e-01 1.97573170e-01 1.13906109e+00
-7.90024996e-01 8.76481950e-01 6.18588400e+00 6.36950731e-02
-9.82746303e-01 -1.43543929e-01 4.51257318e-01 2.30149813e-02
-3.23482215e-01 4.23270315e-01 -6.67252302e-01 -1.51923835e-01
6.68753624e-01 -1.10677265e-01 3.99489939e-01 9.19183254e-01
7.19468668e-02 -7.13465512e-01 -1.26687372e+00 5.51979065e-01
-2.20276743e-01 -1.17656422e+00 9.02103558e-02 -3.87014270e-01
2.78925419e-01 -5.79313755e-01 8.61068256e-03 7.48787999e-01
8.33196998e-01 -7.06173480e-01 1.37654349e-01 5.44049263e-01
8.67238492e-02 -5.70461810e-01 3.57212692e-01 1.01533866e+00
-4.79606837e-01 -6.03906214e-01 2.84360856e-01 -6.63154200e-02
1.75125360e-01 -4.06276643e-01 -1.68626368e+00 3.24272104e-02
4.68253076e-01 5.63029498e-02 -2.56569862e-01 9.51100469e-01
-1.89249650e-01 4.54749703e-01 -2.39069492e-01 -1.57917753e-01
-3.13844770e-01 1.52843788e-01 5.39757967e-01 7.80035794e-01
-2.18939558e-01 6.15939975e-01 7.02602327e-01 6.62419021e-01
2.64768571e-01 -1.24470785e-01 -1.58635616e-01 -1.86755165e-01
7.57951975e-01 8.92905056e-01 -5.53047478e-01 -4.75903630e-01
-3.02318096e-01 7.72194922e-01 9.60534737e-02 5.03455102e-01
-6.65494874e-02 -2.95775682e-01 3.14933836e-01 1.20954767e-01
1.25318661e-01 -4.58411098e-01 -1.70300052e-01 -8.64310324e-01
-5.87549984e-01 -8.17294896e-01 7.40456462e-01 -8.86887729e-01
-1.00164211e+00 2.48377994e-01 4.73429471e-01 -5.35816729e-01
-8.73891711e-01 -2.67451674e-01 -7.60755718e-01 7.09808409e-01
-1.20942318e+00 -2.97648370e-01 -2.43550166e-01 5.53634346e-01
1.18340337e+00 -2.98087120e-01 1.13943100e+00 -9.91261080e-02
-1.81759700e-01 2.64256775e-01 -1.82563633e-01 -3.32912616e-02
9.72751439e-01 -1.37730396e+00 -1.07135966e-01 3.06110471e-01
-1.41345143e-01 9.43342507e-01 9.96744156e-01 -9.62670565e-01
-1.18703985e+00 -3.74863982e-01 3.85005176e-01 -3.67546171e-01
2.21496180e-01 -2.28326365e-01 -1.09421158e+00 1.14603722e+00
4.66325939e-01 -9.04746711e-01 6.60959840e-01 3.50221813e-01
-1.10426042e-02 5.28539605e-02 -1.27196121e+00 4.81234491e-01
3.17838609e-01 -4.16751117e-01 -9.70966101e-01 4.83369946e-01
9.35648978e-01 -6.91118062e-01 -5.62883556e-01 5.45236170e-01
2.59576470e-01 -6.10806227e-01 8.11714053e-01 -5.95270753e-01
-2.74262249e-01 4.77448925e-02 1.66520521e-01 -1.08901608e+00
7.41540790e-02 -6.77981853e-01 -2.85342559e-02 1.15666306e+00
5.35517573e-01 -5.00082374e-01 7.20615029e-01 1.47118568e+00
-2.55169630e-01 -3.56975615e-01 -2.75084108e-01 -1.17483899e-01
-2.40626320e-01 -1.28753558e-01 2.91449040e-01 8.62810135e-01
4.95839387e-01 6.70027673e-01 -4.44879159e-02 4.60138947e-01
2.23389134e-01 7.58467987e-02 6.65118694e-01 -1.27144885e+00
-3.72312069e-01 -4.45021354e-02 2.76494950e-01 -1.45831335e+00
8.83506015e-02 -1.15888573e-01 4.64740247e-01 -1.52008283e+00
-1.62963226e-01 -5.06386936e-01 8.37084949e-02 5.56204855e-01
-7.00349808e-02 -8.84612024e-01 1.62962109e-01 4.89975572e-01
-8.08751881e-01 3.13196272e-01 7.96079636e-01 -8.00849348e-02
-9.08685684e-01 5.29910862e-01 -7.10038006e-01 6.95998967e-01
6.93791807e-01 3.72678638e-02 -1.02798986e+00 -2.64521956e-01
-6.52907342e-02 7.17062771e-01 -5.94568513e-02 -6.04893386e-01
9.11525488e-01 -2.69538105e-01 3.44880581e-01 -5.42883217e-01
6.26670301e-01 -1.06567061e+00 -1.89707026e-01 4.50950623e-01
-1.10934627e+00 -1.12130329e-01 4.26771045e-01 5.31185091e-01
-1.42796144e-01 -5.41101456e-01 5.61490357e-01 -4.92795825e-01
-1.06672525e+00 -2.25473076e-01 -9.35944021e-01 -1.60567045e-01
9.93079424e-01 -4.54973206e-02 -4.67684627e-01 -9.62898374e-01
-1.47874653e+00 9.42728460e-01 2.47216538e-01 5.88489175e-01
6.47750795e-01 -3.78400207e-01 -6.48534745e-02 1.72571108e-01
2.99695935e-02 1.36472136e-01 -1.11937495e-02 2.93988287e-01
-9.23274979e-02 3.63724262e-01 -9.17225033e-02 -6.94978535e-01
-1.56115496e+00 3.87895882e-01 5.09023190e-01 -2.01722607e-01
-2.58652329e-01 6.86802387e-01 1.14143573e-01 -6.81540072e-01
1.12044835e+00 -9.02559534e-02 -4.47950393e-01 -7.48656616e-02
4.01356518e-01 -7.63875768e-02 6.24367706e-02 4.07330066e-01
6.85209632e-02 -1.16447426e-01 -7.13839114e-01 -4.26493794e-01
8.89858902e-01 -7.80897498e-01 4.69464630e-01 7.10546255e-01
5.24711788e-01 -1.25518218e-01 -1.31639326e+00 -6.07050061e-01
4.84423012e-01 -3.74687493e-01 -8.11546296e-02 -1.53645408e+00
-3.51091653e-01 4.91455674e-01 7.04224169e-01 3.93282682e-01
9.53669608e-01 1.27550252e-02 3.54667395e-01 1.25018895e+00
9.00658295e-02 -1.26994359e+00 7.73022413e-01 6.05179667e-01
7.36445010e-01 -1.41471136e+00 2.29681358e-02 -4.74563807e-01
-9.01990652e-01 1.36610687e+00 1.22812116e+00 5.25715292e-01
7.53105342e-01 3.65703136e-01 4.74784255e-01 -2.78572649e-01
-1.22282445e+00 8.72457847e-02 7.06316950e-03 2.26533666e-01
2.36658171e-01 -4.84884083e-01 1.19889759e-01 1.83401376e-01
3.51373523e-01 -5.59433922e-02 5.37636399e-01 1.66440010e+00
-9.05981064e-01 -1.03533423e+00 -2.55318850e-01 3.85694414e-01
6.48200065e-02 1.14650875e-01 -8.36479843e-01 7.17561364e-01
-3.35266829e-01 1.27605951e+00 3.25945914e-01 -8.93486664e-03
3.08332205e-01 5.66205680e-01 2.85771161e-01 -1.19696188e+00
-8.52655828e-01 2.00703055e-01 3.06989223e-01 -2.48063937e-01
-1.82815135e-01 -6.69311941e-01 -1.63184249e+00 2.26816714e-01
-6.39503241e-01 6.22293830e-01 5.51645160e-01 1.15513504e+00
1.80959836e-01 2.53246754e-01 5.16263247e-01 -1.24124959e-01
-1.13667154e+00 -8.05299401e-01 -8.47004727e-02 1.76177174e-01
3.32573593e-01 -4.50268090e-01 -3.80635262e-01 -8.92558321e-02]
|
[12.898334503173828, 7.966475486755371]
|
a0bdef05-c556-4a41-b255-8180ae91a2f9
|
clustering-based-aggregations-for-prediction
|
2210.09738
| null |
https://arxiv.org/abs/2210.09738v1
|
https://arxiv.org/pdf/2210.09738v1.pdf
|
Clustering-based Aggregations for Prediction in Event Streams
|
Predicting the behaviour of shoppers provides valuable information for retailers, such as the expected spend of a shopper or the total turnover of a supermarket. The ability to make predictions on an individual level is useful, as it allows supermarkets to accurately perform targeted marketing. However, given the expected number of shoppers and their diverse behaviours, making accurate predictions on an individual level is difficult. This problem does not only arise in shopper behaviour, but also in various business processes, such as predicting when an invoice will be paid. In this paper we present CAPiES, a framework that focuses on this trade-off in an online setting. By making predictions on a larger number of entities at a time, we improve the predictive accuracy but at the potential cost of usefulness since we can say less about the individual entities. CAPiES is developed in an online setting, where we continuously update the prediction model and make new predictions over time. We show the existence of the trade-off in an experimental evaluation in two real-world scenarios: a supermarket with over 160 000 shoppers and a paint factory with over 171 000 invoices.
|
['Boudewijn F. van Dongen', 'Marwan Hassani', 'Yorick Spenrath']
|
2022-10-18
| null | null | null | null |
['marketing']
|
['miscellaneous']
|
[-2.93576866e-02 3.48976910e-01 -5.20276368e-01 -5.15855968e-01
-4.45248574e-01 -4.38644290e-01 2.81620294e-01 7.40907669e-01
-4.86654162e-01 5.79764903e-01 -1.30679786e-01 -7.71739557e-02
-2.73424238e-01 -1.23674810e+00 -8.64885688e-01 -4.50256974e-01
-2.95570046e-01 1.02133739e+00 4.38035935e-01 -3.07772845e-01
1.22262180e-01 2.73919046e-01 -1.37924230e+00 2.21315891e-01
6.38464689e-01 1.24182093e+00 4.00218189e-01 4.32514399e-01
-1.35370065e-02 7.43495166e-01 -1.01553842e-01 -7.82460332e-01
5.26851177e-01 1.67528316e-01 -4.79560852e-01 3.15013438e-01
-1.48945510e-01 -3.37687463e-01 7.15996623e-02 7.62529373e-01
-2.50025153e-01 9.93827954e-02 4.66787040e-01 -1.46849513e+00
-2.21941605e-01 1.02399886e+00 -2.81229913e-01 4.27623279e-02
2.58235276e-01 -1.08695686e-01 1.51552701e+00 -2.01704592e-01
5.17473459e-01 9.73136902e-01 6.57851160e-01 -1.99012861e-01
-1.39974105e+00 -2.84573436e-01 4.87081707e-01 1.77938938e-01
-8.78909230e-01 5.79524599e-02 4.14598525e-01 -2.35699862e-01
5.30171692e-01 2.52003908e-01 5.81073701e-01 7.36084878e-01
3.32546622e-01 9.99263346e-01 9.18544233e-01 8.88565630e-02
4.70225513e-01 5.31511068e-01 2.92280316e-02 -2.17982292e-01
2.21861452e-01 -1.19009510e-01 -3.10261816e-01 -1.31863147e-01
5.21603882e-01 5.15459239e-01 2.62987703e-01 -4.54487890e-01
-9.43373084e-01 1.12434304e+00 4.54727709e-01 2.11444125e-01
-8.39824498e-01 -8.08177814e-02 1.70781732e-01 6.64579213e-01
3.21606189e-01 5.59917986e-01 -9.45592105e-01 -4.48193878e-01
-5.94525695e-01 5.12891889e-01 1.43708670e+00 1.17278075e+00
8.31556439e-01 -5.89045405e-01 2.20246702e-01 6.26688004e-01
2.02715397e-03 1.69644043e-01 3.30565572e-01 -9.35156465e-01
5.80594540e-01 7.58857489e-01 7.30171800e-01 -1.07552958e+00
-4.91492599e-01 -3.91289204e-01 -4.89497602e-01 -4.43993568e-01
7.24822819e-01 7.42877126e-02 -5.59825003e-01 1.46142542e+00
3.36743474e-01 -2.02920884e-01 -4.30975288e-01 8.49173427e-01
-2.17319101e-01 5.56814671e-01 2.06165671e-01 -4.12109613e-01
1.22735059e+00 -9.04439747e-01 -5.96274912e-01 -4.45116431e-01
7.65443981e-01 -7.13038623e-01 7.32200086e-01 6.47751749e-01
-9.45231736e-01 -3.28322470e-01 -3.19822490e-01 3.88810992e-01
-4.78593886e-01 -3.75560015e-01 8.39946628e-01 3.80910873e-01
-4.81663197e-01 1.10188985e+00 -9.84323978e-01 -4.25266445e-01
9.71241202e-03 6.25257671e-01 -1.50986314e-01 -1.13535188e-01
-1.20265961e+00 9.60044086e-01 2.99925983e-01 -1.70595586e-01
-2.67185777e-01 -6.03240550e-01 -5.38212299e-01 4.06913787e-01
9.06602144e-01 -2.66363233e-01 1.56933939e+00 -1.03736246e+00
-9.09335196e-01 2.29833484e-01 7.29371682e-02 -9.07305241e-01
8.49591196e-01 7.35498890e-02 -4.61511612e-01 -4.29675043e-01
2.51708895e-01 1.09760903e-01 4.55038160e-01 -1.04186285e+00
-1.30303800e+00 -7.30505586e-01 2.92507827e-01 1.84249490e-01
8.29980895e-02 -4.32108521e-01 -2.01979637e-01 -1.27154037e-01
2.69724727e-01 -1.23175466e+00 -6.41768634e-01 -4.90899712e-01
-1.04746163e-01 -3.07948053e-01 3.58487159e-01 -5.88414133e-01
1.17902792e+00 -2.05245638e+00 -2.13073209e-01 4.63510960e-01
1.14912182e-01 -1.57045200e-01 1.54638678e-01 9.29926276e-01
3.95653516e-01 1.00958668e-01 1.77823290e-01 -1.09152384e-01
3.58152181e-01 4.68714207e-01 1.42005971e-02 1.18098915e-01
1.01794012e-01 7.61895657e-01 -8.14824998e-01 -1.65039167e-01
1.18182898e-01 -1.35875821e-01 -4.91581172e-01 4.52640615e-02
-4.08366412e-01 1.51964754e-01 -8.45921338e-01 6.31784260e-01
5.58692455e-01 -4.68404859e-01 4.92488116e-01 2.34421313e-01
-9.42348465e-02 2.39732221e-01 -1.37781894e+00 8.48150313e-01
-8.19052339e-01 -1.32156145e-02 3.20549339e-01 -1.10390699e+00
8.77750814e-01 -9.89340246e-02 5.43594897e-01 -8.70620310e-01
-5.84887601e-02 4.56882268e-01 3.63788307e-02 -1.34473115e-01
7.94174969e-01 -1.89698681e-01 -4.69227761e-01 3.69451731e-01
-3.46785545e-01 3.12770277e-01 5.96402645e-01 5.63552491e-02
1.23686516e+00 -1.66668907e-01 4.18696433e-01 -1.03240035e-01
5.94237335e-02 1.73469022e-01 6.01782024e-01 8.68319452e-01
-2.18683213e-01 -7.94039760e-03 7.20297515e-01 -6.77077711e-01
-1.21514499e+00 -8.82329285e-01 -2.15812549e-01 1.28017914e+00
3.18029851e-01 -2.86233723e-01 -3.58212650e-01 -6.32150769e-01
6.04944050e-01 8.72440100e-01 -6.48811102e-01 6.80501759e-02
-3.40489358e-01 -6.10929668e-01 -6.32443130e-01 4.16209310e-01
3.45312893e-01 -7.78382421e-01 -3.24202627e-01 8.25334370e-01
-2.38064080e-01 -1.34352791e+00 -3.17318082e-01 2.44588032e-01
-8.20620298e-01 -8.58137608e-01 -1.62479550e-01 -4.68197852e-01
3.37700337e-01 7.74013773e-02 1.21224749e+00 -2.05400348e-01
2.11929753e-01 2.22374320e-01 -4.56501514e-01 -5.51913679e-01
-6.26915514e-01 5.12588441e-01 5.33223562e-02 2.21756265e-01
7.77406216e-01 -6.66804075e-01 -7.79897749e-01 9.36945558e-01
-7.11803615e-01 -1.79561973e-01 6.75936759e-01 6.25890136e-01
5.36313057e-01 5.15899539e-01 8.91944468e-01 -1.27721572e+00
3.66874963e-01 -7.89505780e-01 -7.07288206e-01 1.88311309e-01
-9.85232115e-01 1.22867137e-01 9.69532251e-01 -5.16252756e-01
-7.96698451e-01 1.55520827e-01 -4.45232876e-02 1.69969365e-01
-1.17722906e-01 5.76794088e-01 1.56004637e-01 3.77892852e-01
1.21290155e-01 1.09930538e-01 -1.07796542e-01 -6.77594900e-01
1.19596168e-01 7.66954660e-01 2.02410430e-01 -1.35963112e-01
6.56688213e-01 2.10989058e-01 -5.83472513e-02 -5.28476655e-01
-6.52792811e-01 -9.05162096e-01 -5.71924269e-01 -8.03472847e-02
1.93233520e-01 -7.83686280e-01 -1.31365335e+00 5.56516126e-02
-5.65716326e-01 -2.84377784e-01 -4.15171653e-01 3.40622127e-01
-7.85711229e-01 -1.28115416e-01 -5.17913640e-01 -8.82133365e-01
1.95710406e-01 -8.86424899e-01 9.26524878e-01 2.07152992e-01
-3.38400126e-01 -1.09629214e+00 -3.49977583e-01 6.73119307e-01
3.77512187e-01 1.34564072e-01 7.54427791e-01 -1.36846042e+00
-9.90436554e-01 -6.69574976e-01 4.83157746e-02 5.34817316e-02
1.03148594e-01 -4.34475988e-01 -3.54283273e-01 -1.76456288e-01
-1.61941722e-01 8.02903250e-02 4.63767260e-01 2.84715325e-01
8.07654679e-01 -4.52205420e-01 -3.63279670e-01 -1.84663638e-01
1.38203037e+00 3.19912225e-01 2.73153275e-01 5.14800191e-01
1.60277635e-01 1.06307006e+00 1.28009772e+00 8.12732279e-01
7.61989594e-01 9.45785701e-01 6.23185098e-01 4.27136630e-01
6.87557101e-01 -5.94602466e-01 2.05330655e-01 4.65490788e-01
-1.41763553e-01 2.77061085e-03 -7.30825663e-01 7.07369685e-01
-2.26018167e+00 -7.91599810e-01 -1.45757854e-01 2.43904138e+00
5.67615926e-01 3.59648347e-01 7.30854213e-01 1.05617262e-01
5.72143793e-01 -2.41434753e-01 -8.18268239e-01 -5.90691507e-01
3.18061382e-01 -3.84228081e-01 1.11957467e+00 2.04973042e-01
-8.40986311e-01 6.40998721e-01 5.55599117e+00 7.92459428e-01
-6.11921728e-01 -8.87681916e-02 9.41218913e-01 -6.53692037e-02
-1.89140126e-01 1.36859030e-01 -1.16510522e+00 8.47979963e-01
1.32775879e+00 -1.90709904e-01 7.10764110e-01 1.13458538e+00
3.77236575e-01 -4.06175792e-01 -1.34572291e+00 6.16320133e-01
-5.68296075e-01 -1.06412685e+00 -4.42279607e-01 5.81394792e-01
5.08913040e-01 3.55890281e-02 -1.21589839e-01 5.89248896e-01
4.89626050e-01 -7.03919351e-01 5.05401373e-01 4.87860054e-01
-2.98296303e-01 -1.02039146e+00 9.49408650e-01 9.24677908e-01
-1.07053745e+00 -6.73052430e-01 -1.60361752e-01 -2.40588948e-01
4.49342847e-01 6.42716527e-01 -1.20475817e+00 2.73814738e-01
5.53462207e-01 4.06812012e-01 -2.71514207e-01 8.09743941e-01
2.77494818e-01 3.79655719e-01 -7.22807109e-01 -3.99175376e-01
4.07160461e-01 -5.70796907e-01 3.70014422e-02 6.14796042e-01
3.54068100e-01 -9.02349725e-02 3.31028074e-01 6.90159619e-01
-1.38188139e-01 3.41569245e-01 -3.70013535e-01 -1.79605678e-01
3.89355242e-01 1.31861162e+00 -6.93828046e-01 4.67749052e-02
-5.28981626e-01 6.30757868e-01 1.27169326e-01 8.23969170e-02
-6.32197440e-01 -2.80173472e-03 4.37038839e-01 9.48874772e-01
5.78831255e-01 -4.46288325e-02 -9.90366638e-02 -8.78643513e-01
1.76720366e-01 -6.41712189e-01 1.89331979e-01 -2.65979499e-01
-1.31141210e+00 7.83464685e-03 -1.67348236e-01 -1.03445578e+00
-5.09426057e-01 -6.42528951e-01 -1.17621258e-01 3.73725295e-01
-1.42408097e+00 -9.42876995e-01 1.85531914e-01 8.20565522e-02
6.18811250e-01 3.77395034e-01 3.35301489e-01 1.44904003e-01
-2.00079456e-01 1.28745884e-01 4.35228884e-01 -1.26245469e-01
3.47957164e-01 -1.46504498e+00 4.04855043e-01 1.11472495e-01
-5.58316708e-04 4.17785943e-01 1.04871488e+00 -6.61543369e-01
-1.39594984e+00 -9.56038773e-01 1.36570358e+00 -5.86685956e-01
1.11117530e+00 -4.27428156e-01 -8.09205890e-01 8.80243957e-01
-3.13287735e-01 -2.39757732e-01 4.88953888e-01 6.27330363e-01
2.96856374e-01 -4.15001303e-01 -1.14643443e+00 3.74133557e-01
8.10043573e-01 -1.17611729e-01 -2.30390906e-01 5.67683756e-01
5.08927941e-01 -1.66943282e-01 -1.56968713e+00 1.11414447e-01
7.48528361e-01 -1.04722238e+00 6.76699758e-01 -3.63869965e-01
3.33140880e-01 3.21460456e-01 -1.80234686e-01 -1.30177069e+00
-2.22651675e-01 -5.26160955e-01 -1.04427323e-01 1.06045997e+00
6.79628551e-01 -8.23863924e-01 1.05938351e+00 1.38892996e+00
4.94659483e-01 -8.94866526e-01 -9.11048830e-01 -8.68572593e-01
-9.40857455e-02 -2.73927659e-01 8.71157408e-01 5.95939159e-01
1.62157305e-02 2.76724607e-01 -4.57782924e-01 -1.20347049e-02
5.31985700e-01 3.84425581e-01 8.49895954e-01 -1.47011197e+00
-5.48915029e-01 1.03683159e-01 -5.11634588e-01 -1.13062501e+00
-2.76677310e-01 -4.36446100e-01 -2.32014403e-01 -1.26815403e+00
1.74591526e-01 -6.30232394e-01 -3.14001620e-01 5.81893250e-02
1.95728317e-01 -1.36992946e-01 4.13070977e-01 4.05989975e-01
-8.81518841e-01 -3.93010414e-04 1.23325837e+00 1.61378428e-01
-4.11887348e-01 8.62923801e-01 -6.84380472e-01 6.18116915e-01
7.74778545e-01 -3.93370599e-01 -1.93382904e-01 2.49493957e-01
8.71606469e-01 3.90298754e-01 1.60031468e-01 -5.19929290e-01
9.27577987e-02 -3.60540986e-01 -1.99772697e-02 -5.89351773e-01
3.36125642e-01 -1.22753727e+00 4.64027196e-01 3.58703613e-01
-5.43934107e-01 2.45592301e-03 -4.46882576e-01 1.02822340e+00
-2.46797174e-01 -2.89020240e-01 4.13126260e-01 -2.76146650e-01
-6.04542851e-01 4.89528894e-01 -3.73448968e-01 -3.36280406e-01
1.27489781e+00 -3.31727207e-01 -8.47155601e-02 -7.14651167e-01
-1.25113261e+00 7.40075350e-01 5.11773944e-01 4.72317368e-01
-1.22416943e-01 -1.14190781e+00 -4.17082280e-01 1.66299596e-01
3.06574643e-01 -1.08176246e-01 3.61054450e-01 1.04088283e+00
-9.02365148e-02 4.68946844e-01 -4.18189131e-02 -4.38891143e-01
-8.04121375e-01 6.88432693e-01 -3.13076600e-02 -8.91924620e-01
-8.60243961e-02 2.38550410e-01 1.92420006e-01 -4.16476935e-01
-7.78847262e-02 -3.81493479e-01 -2.61706114e-01 3.81943941e-01
1.29739270e-01 4.18592453e-01 1.46617889e-01 -6.45068228e-01
4.83303244e-04 -1.39708351e-02 -5.13386369e-01 3.06948006e-01
1.59030271e+00 -6.98144972e-01 2.56169409e-01 6.81589007e-01
9.98613298e-01 -2.27528051e-01 -1.43197036e+00 -4.43925977e-01
4.00339544e-01 -6.53107166e-01 -7.40543008e-02 -8.27972889e-01
-8.61184359e-01 4.31654543e-01 1.78775609e-01 1.17337143e+00
9.38429654e-01 1.55564189e-01 1.15892565e+00 2.53931135e-01
8.81705046e-01 -1.59855998e+00 -3.31633985e-01 1.05784632e-01
3.85747194e-01 -1.48759854e+00 -2.81336159e-01 -4.45184469e-01
-9.75814044e-01 7.63829291e-01 1.83570713e-01 -1.60047859e-01
7.37148523e-01 7.42243081e-02 -2.85904229e-01 1.14118151e-01
-1.05465865e+00 2.81323474e-02 -2.77370065e-01 3.61298740e-01
1.31472662e-01 6.51584387e-01 -4.72895503e-01 1.08317423e+00
-3.33999902e-01 -1.65181868e-02 7.27393210e-01 8.80034745e-01
-5.37003696e-01 -1.46101379e+00 -5.81525303e-02 1.03679729e+00
-6.09986246e-01 2.67490566e-01 -2.02825859e-01 9.18102026e-01
8.92171077e-03 1.02485871e+00 1.82296365e-01 -2.80675441e-01
7.99478412e-01 -5.02287038e-02 -8.09269585e-03 -4.77948636e-01
-6.53857172e-01 2.00702012e-01 4.42430615e-01 -6.67316556e-01
-2.55514085e-01 -1.03289652e+00 -9.16719258e-01 -6.93346381e-01
-5.43739855e-01 3.92823100e-01 8.14368427e-01 8.86882842e-01
3.37784380e-01 7.57882074e-02 1.17222381e+00 -4.91818994e-01
-1.16874242e+00 -8.10607910e-01 -1.37474477e+00 7.40977883e-01
1.41312703e-01 -6.49189472e-01 -4.57620740e-01 -1.24284633e-01]
|
[9.267672538757324, 5.822889804840088]
|
7ec2282c-4a51-410d-adf6-7f3adeea4368
|
neural-program-meta-induction
|
1710.04157
| null |
http://arxiv.org/abs/1710.04157v1
|
http://arxiv.org/pdf/1710.04157v1.pdf
|
Neural Program Meta-Induction
|
Most recently proposed methods for Neural Program Induction work under the
assumption of having a large set of input/output (I/O) examples for learning
any underlying input-output mapping. This paper aims to address the problem of
data and computation efficiency of program induction by leveraging information
from related tasks. Specifically, we propose two approaches for cross-task
knowledge transfer to improve program induction in limited-data scenarios. In
our first proposal, portfolio adaptation, a set of induction models is
pretrained on a set of related tasks, and the best model is adapted towards the
new task using transfer learning. In our second approach, meta program
induction, a $k$-shot learning approach is used to make a model generalize to
new tasks without additional training. To test the efficacy of our methods, we
constructed a new benchmark of programs written in the Karel programming
language. Using an extensive experimental evaluation on the Karel benchmark, we
demonstrate that our proposals dramatically outperform the baseline induction
method that does not use knowledge transfer. We also analyze the relative
performance of the two approaches and study conditions in which they perform
best. In particular, meta induction outperforms all existing approaches under
extreme data sparsity (when a very small number of examples are available),
i.e., fewer than ten. As the number of available I/O examples increase (i.e. a
thousand or more), portfolio adapted program induction becomes the best
approach. For intermediate data sizes, we demonstrate that the combined method
of adapted meta program induction has the strongest performance.
|
['Pushmeet Kohli', 'Rudy Bunel', 'Rishabh Singh', 'Jacob Devlin', 'Matthew Hausknecht']
|
2017-10-11
|
neural-program-meta-induction-1
|
http://papers.nips.cc/paper/6803-neural-program-meta-induction
|
http://papers.nips.cc/paper/6803-neural-program-meta-induction.pdf
|
neurips-2017-12
|
['program-induction']
|
['computer-code']
|
[ 2.67535269e-01 -2.44552642e-01 -9.24722552e-01 -4.51150954e-01
-7.58128285e-01 -4.53675896e-01 2.64258295e-01 3.37374330e-01
-5.31194210e-01 6.83456898e-01 -3.78788292e-01 -6.17576420e-01
-3.94451013e-03 -9.07088578e-01 -1.22036278e+00 -3.85779053e-01
-2.01227710e-01 2.99934715e-01 2.05558956e-01 -1.02340467e-01
4.94732201e-01 1.37228161e-01 -1.65326369e+00 4.76593167e-01
1.14179385e+00 6.57907963e-01 3.51432383e-01 6.97489858e-01
-2.91372269e-01 1.04501116e+00 -4.83082712e-01 -2.80258387e-01
1.53452173e-01 -2.16351241e-01 -1.12963212e+00 -2.67738611e-01
3.29367012e-01 -2.61776596e-01 9.53445211e-04 9.70324278e-01
3.08516353e-01 2.05547541e-01 4.05710697e-01 -1.21425784e+00
-4.97149736e-01 7.57783294e-01 -7.53283501e-01 2.20268667e-01
2.33268917e-01 1.31139070e-01 8.42827737e-01 -1.04634893e+00
5.44868410e-01 8.19892108e-01 8.28383327e-01 4.51226652e-01
-1.76669919e+00 -7.34736860e-01 5.90291917e-02 1.47613343e-02
-1.21956611e+00 -3.22127014e-01 7.78631151e-01 -6.20359302e-01
1.17162228e+00 -7.55691305e-02 1.29774183e-01 5.93819976e-01
-1.95490807e-01 9.07210648e-01 1.21947944e+00 -8.85383129e-01
2.85343528e-01 6.78677559e-01 6.52382016e-01 1.00579476e+00
2.09198549e-01 -7.75584485e-03 -4.85433668e-01 -5.16565859e-01
4.31730300e-01 -1.06202140e-01 7.70688849e-03 -5.78205764e-01
-1.13782954e+00 9.03631866e-01 3.12653601e-01 4.26465303e-01
2.29703635e-03 2.24903911e-01 8.38500798e-01 7.44248807e-01
5.23104370e-01 5.27945161e-01 -6.67193532e-01 -5.63940667e-02
-9.67207313e-01 2.58803368e-01 9.79029953e-01 1.21518576e+00
1.22295105e+00 3.67964469e-02 -2.21738234e-01 9.28058863e-01
-2.44444802e-01 1.40949354e-01 5.62520981e-01 -7.33570397e-01
8.56983900e-01 6.32844985e-01 -4.76690792e-02 -5.09669244e-01
6.88992962e-02 -1.86996982e-01 -5.37244797e-01 2.86772579e-01
3.69978130e-01 -4.24698114e-01 -8.54606688e-01 1.76665938e+00
4.26498801e-02 6.31280988e-02 9.84951258e-02 5.76934516e-02
5.99539220e-01 7.39842832e-01 1.92577347e-01 -2.06747547e-01
1.05955923e+00 -1.21470046e+00 -2.38328546e-01 -2.98354119e-01
1.17394221e+00 -5.23315549e-01 1.38272250e+00 2.59296954e-01
-9.02978539e-01 -8.12912047e-01 -9.86121356e-01 3.18993807e-01
-5.13290346e-01 9.55998674e-02 8.68318558e-01 7.01699853e-01
-1.02839828e+00 5.83236456e-01 -5.03703535e-01 -2.97941387e-01
4.21302527e-01 5.46036720e-01 -1.24906145e-01 -2.23599911e-01
-7.57423341e-01 6.43687904e-01 7.93656826e-01 -5.53611577e-01
-9.16747570e-01 -8.36647630e-01 -7.50425160e-01 1.28066570e-01
1.85957968e-01 -5.68676949e-01 1.31881607e+00 -1.31110537e+00
-1.33905947e+00 9.00115490e-01 -2.50980794e-01 -4.06983554e-01
9.20061991e-02 -1.48098633e-01 4.29033786e-02 -2.18036667e-01
-6.42021000e-02 4.75962698e-01 6.12012506e-01 -1.49869108e+00
-6.49177253e-01 -1.87130049e-01 3.19157183e-01 -5.34890890e-02
-7.03694820e-01 2.82548130e-01 -3.59465718e-01 -4.46661234e-01
-5.21436095e-01 -9.84559298e-01 -2.21318528e-01 -4.62349266e-01
-7.30626360e-02 -4.86467838e-01 7.21162856e-01 -3.08516860e-01
1.47482777e+00 -2.20340133e+00 1.29095986e-01 3.17088187e-01
3.82262841e-02 4.67445552e-01 -2.34040603e-01 3.42831850e-01
-3.03337544e-01 2.24691257e-01 -2.85831183e-01 -1.09590359e-01
-1.51728451e-01 1.92702159e-01 -2.75684953e-01 -4.26011533e-02
-9.33279004e-03 8.40847254e-01 -7.75456309e-01 -6.32911205e-01
-2.30839983e-01 -1.85435370e-01 -9.68103468e-01 3.70320588e-01
-4.88884836e-01 1.65700436e-01 -5.67817390e-01 5.44109225e-01
3.67463440e-01 -4.24617499e-01 -3.90149257e-03 1.49955094e-01
-2.30045781e-01 1.60425738e-01 -9.15996134e-01 1.86524475e+00
-1.04613364e+00 5.27573586e-01 -2.03369990e-01 -1.47428560e+00
1.03150129e+00 2.38023818e-01 2.78772950e-01 -3.62926215e-01
-2.98272908e-01 2.16192335e-01 1.10621549e-01 -6.59859300e-01
1.57511175e-01 5.26501238e-02 -2.96413571e-01 5.91649234e-01
2.66340673e-01 -1.12586290e-01 4.10929471e-01 2.21427139e-02
1.30308282e+00 1.97066396e-01 3.88003021e-01 -3.74676883e-01
5.23250401e-01 4.43393171e-01 4.31446224e-01 1.08635008e+00
1.33979976e-01 -4.72948775e-02 6.16159618e-01 -5.71346998e-01
-1.07436800e+00 -7.03102708e-01 -3.99836935e-02 1.87786174e+00
-2.71086633e-01 -2.99183041e-01 -8.30590129e-01 -9.29194629e-01
1.01174608e-01 8.71003509e-01 -6.47838771e-01 -4.33583446e-02
-7.58075178e-01 -9.73770201e-01 6.20199800e-01 6.84236705e-01
6.35082841e-01 -1.22528565e+00 -5.06418347e-01 9.69494730e-02
2.81807035e-02 -7.32968092e-01 -3.47332060e-01 5.80146194e-01
-1.12990165e+00 -8.34072173e-01 -6.21307015e-01 -1.30397522e+00
9.04900491e-01 9.74323079e-02 1.31795335e+00 1.60573378e-01
-5.10961339e-02 3.02081317e-01 -3.22348684e-01 -3.54302227e-01
-5.32342851e-01 4.20974642e-01 -2.18190521e-01 -3.43933642e-01
4.77465004e-01 -6.46250069e-01 -1.04085669e-01 -4.66047749e-02
-8.09327722e-01 6.02094047e-02 8.86703014e-01 1.14581740e+00
3.12550634e-01 9.11199376e-02 7.59162128e-01 -1.44670415e+00
6.61251247e-01 -6.40927017e-01 -7.09737301e-01 4.59244817e-01
-8.38344216e-01 4.46057647e-01 8.32892299e-01 -7.21393764e-01
-1.26087856e+00 2.67804891e-01 8.51824880e-02 -3.70065540e-01
-1.98617339e-01 8.88159335e-01 1.12986237e-01 -3.49361122e-01
9.53246176e-01 2.88232058e-01 -1.77155957e-01 -3.81243318e-01
1.55335248e-01 6.41010463e-01 2.21116260e-01 -1.12345231e+00
7.78304815e-01 -9.60435439e-03 -4.24462140e-01 -5.57319939e-01
-5.93250394e-01 -4.67303902e-01 -7.66430795e-01 1.96582362e-01
5.43126523e-01 -7.29063749e-01 -4.51076299e-01 4.35161442e-01
-1.16699910e+00 -9.57159996e-01 -2.20888615e-01 2.99057215e-01
-7.00329602e-01 9.98078510e-02 -6.17655754e-01 -4.80186343e-01
-3.90608549e-01 -1.34939611e+00 5.16972661e-01 -4.69212569e-02
-2.12633207e-01 -1.12980509e+00 3.56775850e-01 3.91293392e-02
3.59630138e-01 1.46099860e-02 1.55786586e+00 -9.96803284e-01
-4.76839930e-01 -1.57744139e-01 -2.08996549e-01 4.52950060e-01
2.25884356e-02 -1.82127342e-01 -8.32640529e-01 -2.67623395e-01
-5.47886379e-02 -7.34654963e-01 7.82290697e-01 2.64108106e-02
1.62550175e+00 -3.25057596e-01 -4.94997412e-01 6.59469903e-01
1.68882811e+00 4.33308959e-01 5.05042136e-01 3.85559767e-01
6.34581983e-01 4.09584373e-01 5.73242128e-01 2.18220517e-01
2.66768306e-01 5.88905334e-01 -5.11489660e-02 -1.06846415e-01
1.05512314e-01 -2.09579274e-01 5.02710700e-01 9.43984449e-01
-2.21228838e-01 1.72827139e-01 -1.16123688e+00 7.41357863e-01
-1.70217907e+00 -8.28375578e-01 4.24972743e-01 2.48343754e+00
1.26520658e+00 2.64084250e-01 1.88416258e-01 -1.07761696e-01
5.65570772e-01 -1.74431890e-01 -5.98318875e-01 -7.01459765e-01
4.84588385e-01 7.63303399e-01 4.75591213e-01 2.59498835e-01
-1.07247961e+00 8.33494008e-01 6.31408501e+00 7.43353128e-01
-9.69067276e-01 2.79716879e-01 6.34653032e-01 1.66644767e-01
-1.97170347e-01 1.14087433e-01 -9.68416274e-01 3.82721275e-01
1.09825504e+00 -4.15924460e-01 5.65951169e-01 1.33575964e+00
-4.15165484e-01 -1.72922090e-01 -1.50510025e+00 5.28923988e-01
1.16811521e-01 -1.22884381e+00 -1.72895774e-01 -1.61815569e-01
1.00349164e+00 2.70826444e-02 4.51132432e-02 9.32093203e-01
6.51403785e-01 -7.68088281e-01 2.03537911e-01 2.65803814e-01
8.10373485e-01 -8.17063630e-01 6.13817871e-01 7.02534974e-01
-1.10415494e+00 -2.21177220e-01 -7.96329007e-02 1.70867529e-03
-6.08466327e-01 3.63157332e-01 -7.71420240e-01 2.41702765e-01
6.91225767e-01 4.27141130e-01 -6.27510428e-01 8.52856159e-01
-4.99606878e-02 7.70053208e-01 -1.55107766e-01 -1.47943690e-01
5.09846071e-03 1.65641531e-01 5.86946718e-02 1.53815210e+00
3.23328435e-01 5.14930636e-02 4.17176694e-01 1.00686550e+00
-2.26595953e-01 2.33796164e-01 -1.02002764e+00 1.91399857e-01
4.58392322e-01 9.67687130e-01 -4.55781907e-01 -7.43370891e-01
-8.22369874e-01 7.09153533e-01 7.59001315e-01 4.46966320e-01
-6.61200702e-01 -7.81943262e-01 1.30168959e-01 -1.39408574e-01
3.49771529e-01 -1.14806086e-01 -3.86367708e-01 -1.06543887e+00
2.12290846e-02 -1.08104849e+00 4.86703575e-01 -3.22431356e-01
-9.83839989e-01 4.37500834e-01 4.07343388e-01 -8.60048890e-01
-5.26179969e-01 -5.27321517e-01 -9.23608422e-01 9.49675024e-01
-1.46047819e+00 -9.35786128e-01 -2.32818529e-01 4.90463257e-01
5.99474907e-01 -3.07175606e-01 9.36426938e-01 2.96225637e-01
-5.98975956e-01 9.90914285e-01 1.37735084e-01 3.14891815e-01
7.46065140e-01 -1.17241836e+00 2.58036017e-01 7.13729203e-01
-1.03281841e-01 7.79574275e-01 5.04810274e-01 -5.50376475e-01
-1.50606692e+00 -1.25942063e+00 6.33822262e-01 -3.06309372e-01
7.04887629e-01 -3.22880894e-01 -1.15247917e+00 1.00993347e+00
1.05615623e-01 2.25212023e-01 8.79863739e-01 4.08940256e-01
-5.58961928e-01 -2.62269020e-01 -9.25231636e-01 3.46060216e-01
7.53954232e-01 -5.36228716e-01 -7.13419735e-01 3.80280793e-01
6.77881598e-01 -2.22733632e-01 -1.09839773e+00 5.79281569e-01
3.55281830e-01 -7.32918680e-01 8.59528303e-01 -6.90191627e-01
8.12062442e-01 1.10618686e-02 -9.04865637e-02 -1.33817220e+00
-4.10251543e-02 -2.43622392e-01 -1.83874652e-01 1.27450752e+00
6.26297772e-01 -5.51780403e-01 6.75687611e-01 4.68472749e-01
-4.44792584e-02 -8.09929967e-01 -4.10142571e-01 -8.85269523e-01
3.35905045e-01 -2.53296494e-01 3.93135935e-01 1.07648897e+00
4.61162448e-01 4.13626939e-01 -1.46397769e-01 -1.31941214e-01
5.92012584e-01 5.56203127e-01 1.07961106e+00 -9.92092669e-01
-8.54941010e-01 -3.57033789e-01 6.78016469e-02 -9.38799024e-01
5.42463601e-01 -1.18576670e+00 1.24596037e-01 -9.08400059e-01
6.25040352e-01 -9.94518578e-01 -4.12143856e-01 6.99590981e-01
-3.68414313e-01 -1.95923075e-01 6.17700815e-02 2.99455971e-01
-5.11304617e-01 6.14690892e-02 7.42904186e-01 -2.00261980e-01
-6.46877229e-01 3.76823872e-01 -5.97324550e-01 8.76281321e-01
9.36821461e-01 -5.52175522e-01 -3.84318411e-01 -4.14008170e-01
1.39549702e-01 2.45590851e-01 1.41937211e-01 -1.01924813e+00
3.60784411e-01 -1.18883416e-01 5.21201603e-02 -1.71356544e-01
-1.33379161e-01 -5.95313430e-01 -1.57588914e-01 5.80692172e-01
-6.74901187e-01 1.08026609e-01 4.47396457e-01 3.79635185e-01
-2.00752288e-01 -8.59842777e-01 7.88866460e-01 -3.64431441e-01
-8.97160411e-01 1.61853746e-01 -2.18085572e-01 7.29143992e-02
1.09030163e+00 1.40914731e-02 -1.85806230e-01 4.96769920e-02
-5.38609862e-01 9.47552696e-02 4.49328184e-01 1.93557322e-01
3.74242455e-01 -1.20052361e+00 -5.45225263e-01 2.38401011e-01
3.90042782e-01 -8.48850608e-03 -2.10064024e-01 7.01965213e-01
-2.77215034e-01 3.80690694e-01 -1.19173110e-01 -5.77232718e-01
-1.33981919e+00 9.30942297e-01 4.73825522e-02 -7.35423028e-01
-2.32557461e-01 6.22356415e-01 2.12081015e-01 -6.52480304e-01
2.97341675e-01 -3.48698199e-01 -4.63657118e-02 -1.78670138e-01
4.70569164e-01 1.53760731e-01 8.43182802e-02 2.20394377e-02
9.15157795e-02 3.65875930e-01 -3.92349690e-01 1.48416325e-01
1.44683564e+00 5.37328899e-01 -3.64371985e-01 7.52574086e-01
1.28921723e+00 -6.56275079e-02 -9.69803512e-01 -6.32226706e-01
4.05543119e-01 -3.70002359e-01 -1.57061264e-01 -5.52323520e-01
-9.11854386e-01 9.24618125e-01 4.89864051e-01 1.45012178e-02
1.25840056e+00 -4.59440351e-02 3.37154865e-01 8.53604674e-01
5.87981999e-01 -9.39906716e-01 2.87666857e-01 6.44910991e-01
4.26165849e-01 -1.46430957e+00 -2.15681177e-02 -2.67793119e-01
-3.69732827e-01 1.02953899e+00 9.89629567e-01 -3.06925103e-02
6.11612439e-01 5.67569017e-01 -4.58052248e-01 -9.12150964e-02
-8.92353058e-01 -4.21912819e-02 7.58100227e-02 4.95901197e-01
6.75852001e-01 -6.58003688e-02 -4.30297293e-02 4.85987157e-01
1.63549513e-01 3.92381340e-01 2.85394937e-01 1.25599372e+00
-5.08494258e-01 -1.25019455e+00 -2.73644716e-01 7.99455106e-01
-4.04272169e-01 -3.39983374e-01 1.23594478e-01 9.58507597e-01
4.61895429e-02 4.89106804e-01 -1.49034709e-02 -3.56485367e-01
2.81475753e-01 4.34885830e-01 4.42852050e-01 -1.09833753e+00
-6.86061323e-01 -3.84309441e-01 -8.00990090e-02 -2.07663149e-01
-5.60899377e-01 -4.16140437e-01 -9.61277068e-01 -1.62537917e-01
-3.79093021e-01 2.26381183e-01 4.59625632e-01 7.97296405e-01
1.85161322e-01 4.79362160e-01 6.97682559e-01 -6.79937005e-01
-7.55130291e-01 -8.37755263e-01 -2.70879060e-01 3.81030411e-01
2.65299529e-01 -3.97644788e-01 -2.40063816e-01 4.91810560e-01]
|
[8.01919937133789, 7.6302032470703125]
|
0600965e-b6a3-498c-ab1e-c194eee8ef0c
|
medical-literature-mining-and-retrieval-in-a
|
2108.01436
| null |
https://arxiv.org/abs/2108.01436v1
|
https://arxiv.org/pdf/2108.01436v1.pdf
|
Medical Literature Mining and Retrieval in a Conversational Setting
|
The Covid-19 pandemic has caused a spur in the medical research literature. With new research advances in understanding the virus, there is a need for robust text mining tools which can process, extract and present answers from the literature in a concise and consumable way. With a DialoGPT based multi-turn conversation generation module, and BM-25 \& neural embeddings based ensemble information retrieval module, in this paper we present a conversational system, which can retrieve and answer coronavirus-related queries from the rich medical literature, and present it in a conversational setting with the user. We further perform experiments to compare neural embedding-based document retrieval and the traditional BM25 retrieval algorithm and report the results.
|
['Rohini K. Srihari', 'Sougata Saha', 'Souvik Das']
|
2021-07-23
| null | null | null | null |
['literature-mining']
|
['natural-language-processing']
|
[ 3.21049877e-02 1.05092861e-01 9.25200135e-02 -2.23702431e-01
-8.29498827e-01 -3.20429087e-01 5.30426800e-01 5.31580389e-01
-5.46160161e-01 8.31911445e-01 7.81202435e-01 -5.09499133e-01
-4.23688024e-01 -8.51331115e-01 1.44121006e-01 -4.10939485e-01
-1.31295947e-02 9.98751760e-01 -2.65608937e-01 -5.32795072e-01
2.67635137e-01 2.49019414e-01 -8.70866299e-01 4.65105921e-01
8.18384051e-01 5.15650928e-01 2.33285755e-01 1.32331991e+00
-5.95040381e-01 7.90857315e-01 -8.81452084e-01 -5.81450462e-01
-3.46484780e-01 -2.15549186e-01 -1.08237624e+00 -6.45351946e-01
-3.10906202e-01 -3.09745371e-01 -3.75404507e-01 3.05425704e-01
1.17461920e+00 2.21577719e-01 9.92473781e-01 -9.02909815e-01
-8.07305396e-01 2.72142351e-01 -2.99103446e-02 4.67707187e-01
6.94030166e-01 -2.72450745e-01 6.60093963e-01 -8.77121925e-01
1.06641591e+00 1.35568357e+00 4.41376358e-01 9.48703349e-01
-9.29431379e-01 -4.93496865e-01 -4.85049635e-01 3.73054028e-01
-9.45249438e-01 -4.86216247e-02 7.60102987e-01 -5.44524193e-01
1.41626859e+00 3.04906070e-01 6.12781525e-01 1.42332363e+00
6.63487136e-01 5.76300144e-01 4.13827091e-01 -3.75456423e-01
5.60162552e-02 4.25620407e-01 4.83872741e-01 6.26783788e-01
1.02219932e-01 -2.29835585e-01 -3.17040652e-01 -8.61385763e-01
5.22899255e-02 5.58280826e-01 -2.36397162e-01 4.17073667e-01
-1.02127206e+00 1.35811913e+00 1.45492420e-01 4.90984946e-01
-6.10658526e-01 -1.24891162e-01 7.43534088e-01 6.40007794e-01
7.61258841e-01 9.15371239e-01 -5.16959429e-01 -2.04502083e-02
-6.20457590e-01 3.22923809e-01 1.31608307e+00 4.90527481e-01
2.94840962e-01 -4.32106167e-01 -5.71683526e-01 9.72465694e-01
3.40806246e-01 5.98827720e-01 6.00351214e-01 -5.05924165e-01
1.12998083e-01 7.31283784e-01 -1.46438023e-02 -1.41930079e+00
-6.17885590e-01 8.24561268e-02 -8.01843822e-01 -6.89943850e-01
-4.15152967e-01 -6.17395759e-01 -4.70052063e-01 1.33152330e+00
4.92853552e-01 2.44616754e-02 5.15982687e-01 3.63067091e-01
1.60571909e+00 9.78527606e-01 1.05682157e-01 -3.12033862e-01
1.65215957e+00 -8.46307516e-01 -1.19943988e+00 2.39859834e-01
7.60426879e-01 -9.89017844e-01 4.51945603e-01 -6.86627552e-02
-6.79187238e-01 -3.20252419e-01 -8.56967211e-01 -1.85426325e-01
-8.74189138e-01 -3.63109469e-01 5.74612856e-01 3.60402733e-01
-1.35164273e+00 2.34624207e-01 -3.58697385e-01 -6.89928949e-01
1.37721807e-01 2.81970322e-01 -2.85476476e-01 -1.62619784e-01
-1.43889976e+00 1.27954364e+00 3.26522708e-01 -5.91611266e-02
-4.23814476e-01 -7.92916238e-01 -6.36753023e-01 -1.82675928e-01
-1.77278444e-02 -1.26794243e+00 1.06878424e+00 -1.27620623e-01
-1.30696225e+00 7.77947485e-01 -3.07026327e-01 -2.59006798e-01
-1.39523566e-01 -1.31332591e-01 -6.20300412e-01 5.75766802e-01
-2.02583984e-01 8.91211689e-01 4.82234120e-01 -8.46681297e-01
-1.84607863e-01 -4.18160886e-01 -3.56697515e-02 1.07311130e-01
-3.25684547e-01 3.11004847e-01 -1.60232797e-01 -5.75885773e-01
-8.67092967e-01 -9.42671359e-01 -4.36350673e-01 -2.99895823e-01
-1.43639490e-01 -8.89679670e-01 9.11694527e-01 -8.09353411e-01
1.33134878e+00 -1.79028487e+00 -3.45904101e-03 2.47521163e-03
5.53262413e-01 5.92278123e-01 -4.05059874e-01 1.22257924e+00
2.21306726e-01 2.02478692e-01 2.75179595e-02 -1.84685498e-01
-2.52234548e-01 3.53440225e-01 -4.86103684e-01 -1.77008599e-01
3.64445865e-01 1.21071959e+00 -9.66761231e-01 -6.90640092e-01
9.98977721e-02 8.46806228e-01 -7.14856565e-01 7.17789114e-01
-4.06571120e-01 1.62943512e-01 -9.01677072e-01 3.54612142e-01
2.64078230e-01 -3.94774646e-01 1.84183091e-01 1.81856632e-01
3.31248105e-01 -5.73379286e-02 -2.89884120e-01 1.55925298e+00
-6.71082795e-01 9.40505326e-01 -1.14505477e-01 -8.59319627e-01
9.40752983e-01 7.92973220e-01 3.92765164e-01 -4.32064325e-01
3.81189018e-01 -1.46673337e-01 -1.45712078e-01 -1.22353184e+00
5.75794280e-01 4.40528356e-02 6.46870658e-02 7.95438409e-01
-4.81540896e-02 3.78028378e-02 5.42707704e-02 5.14983356e-01
1.37650728e+00 -5.58162987e-01 2.49981329e-01 -1.02632433e-01
4.48178262e-01 1.54744312e-01 -1.00131519e-01 7.03521311e-01
-5.18956073e-02 8.91776606e-02 1.37277633e-01 -6.75433040e-01
-5.55848479e-01 -8.20801318e-01 -6.94232658e-02 1.15007007e+00
-2.47417420e-01 -5.44792652e-01 -6.29286468e-01 -4.57121313e-01
1.98961645e-02 6.07025743e-01 -7.62313426e-01 -2.42590785e-01
-4.64228123e-01 -8.07201266e-01 5.47891378e-01 3.51466238e-01
5.70825301e-02 -1.51076651e+00 -6.66216254e-01 6.46644533e-01
-4.18773144e-01 -6.11829877e-01 -4.31985080e-01 -1.16426118e-01
-8.38326991e-01 -1.06232417e+00 -9.46674287e-01 -1.11869633e+00
4.16364312e-01 1.64145604e-01 1.19419003e+00 2.49318376e-01
-9.45482075e-01 7.46616662e-01 -4.97751743e-01 -8.95341456e-01
-5.80747128e-01 9.73152742e-02 -5.95083237e-02 -4.21410739e-01
9.90998566e-01 -3.14992517e-01 -1.06505823e+00 -2.72236586e-01
-1.21831119e+00 -2.81400681e-01 4.15487677e-01 8.64817739e-01
-4.53220494e-02 -5.70000052e-01 1.26351857e+00 -7.74619281e-01
1.57834101e+00 -1.03258491e+00 8.08537453e-02 2.77116388e-01
-7.61954904e-01 2.62736440e-01 2.06478015e-01 -3.33221912e-01
-8.85944366e-01 -5.71952045e-01 -4.22434419e-01 -7.98554868e-02
1.62158124e-02 8.21832359e-01 4.53785181e-01 4.75086987e-01
8.83067131e-01 1.27249554e-01 3.96199584e-01 -3.93681556e-01
6.66793644e-01 1.49445450e+00 -7.96080679e-02 -6.60658404e-02
3.17709148e-01 2.43935063e-01 -5.81130326e-01 -1.16004694e+00
-4.41184312e-01 -8.54382277e-01 -1.70013271e-02 -2.83893555e-01
1.29240668e+00 -4.93918568e-01 -1.06620288e+00 -3.35021585e-01
-1.81686592e+00 2.67717570e-01 1.92939326e-01 4.05746222e-01
-1.84748784e-01 2.47475952e-01 -7.54949808e-01 -8.09152961e-01
-1.29744184e+00 -9.01549637e-01 9.23743844e-01 2.44521618e-01
-9.13665533e-01 -1.12553251e+00 9.50632989e-01 5.44241190e-01
9.19634402e-01 1.51019171e-01 1.34526443e+00 -1.28765285e+00
-2.24533081e-01 -4.08152550e-01 -2.11262062e-01 3.87322977e-02
1.35076821e-01 -2.40530208e-01 -9.45805669e-01 -7.95627683e-02
-5.17800450e-02 -1.56475574e-01 8.12362790e-01 2.56381482e-01
7.41689384e-01 -5.13333023e-01 -7.02910423e-01 -4.25065532e-02
1.14993227e+00 6.86541975e-01 4.78010178e-01 -4.31904197e-02
1.39301807e-01 8.69736791e-01 -1.04159519e-01 2.85299718e-01
5.17629266e-01 1.55004561e-01 -2.63801068e-01 -1.97704732e-02
2.92876601e-01 7.24864677e-02 -5.54858632e-02 1.38274240e+00
2.46321321e-01 -5.66826463e-01 -8.01029742e-01 6.19503021e-01
-1.81280053e+00 -1.05041075e+00 4.12138104e-01 1.44494581e+00
1.07064962e+00 -5.30399144e-01 -1.33979738e-01 -4.03447509e-01
4.09191549e-01 4.61467467e-02 -3.59541237e-01 -1.07740605e+00
3.59708250e-01 4.29732054e-01 -2.87511468e-01 6.31356657e-01
-8.06791723e-01 6.35343015e-01 7.33626890e+00 5.02973139e-01
-9.30823267e-01 3.28066908e-02 4.50546205e-01 1.27429292e-02
-3.72123659e-01 -6.75868750e-01 -3.41340750e-01 2.60616273e-01
1.38564610e+00 -6.11623704e-01 1.53383613e-01 7.03227520e-01
1.18064240e-01 3.15580249e-01 -1.07886147e+00 1.14220572e+00
3.94545376e-01 -1.76552987e+00 3.43687207e-01 -7.83862844e-02
3.60519201e-01 2.02126756e-01 -1.56146929e-01 5.69474638e-01
3.46771568e-01 -1.10092854e+00 -6.36157453e-01 9.71852958e-01
4.19619948e-01 -6.84807003e-01 1.12711799e+00 2.90763527e-01
-6.83218122e-01 4.14618142e-02 -3.37725610e-01 2.38025427e-01
2.10036740e-01 4.21765357e-01 -1.48537338e+00 6.02357745e-01
5.29393733e-01 3.53435159e-01 -9.20633823e-02 8.69477510e-01
2.96952426e-01 2.69758970e-01 -1.34196198e-02 -9.28128958e-01
1.29645258e-01 -1.44069046e-01 5.01841962e-01 1.79989386e+00
4.92701493e-02 4.55760181e-01 4.18573320e-02 5.89810371e-01
-1.85057685e-01 6.19579911e-01 -1.02479124e+00 -6.16892338e-01
2.62691051e-01 1.33143163e+00 -4.10890043e-01 -5.73352218e-01
-1.70932800e-01 8.86601865e-01 9.14179832e-02 2.96202749e-01
-2.90474415e-01 -7.28255212e-01 6.58598244e-01 -2.76495337e-01
9.15586278e-02 7.57400393e-02 4.23307866e-01 -9.19661522e-01
-2.80388474e-01 -1.09448898e+00 5.85398972e-01 -6.68285966e-01
-1.68897784e+00 1.05606556e+00 1.31720947e-02 -7.39138961e-01
-7.31181860e-01 -4.49055284e-01 -7.41736054e-01 7.06264257e-01
-1.22869825e+00 -6.70467257e-01 1.24308489e-01 4.26180422e-01
6.84751630e-01 -3.93543661e-01 1.63642216e+00 3.39037180e-01
-3.33190441e-01 2.19049647e-01 2.77556956e-01 1.12215765e-01
7.14966059e-01 -9.88461912e-01 -1.03660904e-01 -3.25228065e-01
-1.10748932e-01 1.20119941e+00 7.02186346e-01 -6.68919683e-01
-1.50690663e+00 -1.00830483e+00 1.37515199e+00 -8.81106496e-01
5.43431103e-01 -1.94959998e-01 -8.32477033e-01 2.02570990e-01
9.74342167e-01 -8.91829491e-01 1.50556672e+00 1.43505096e-01
-1.54117286e-01 6.55606464e-02 -1.23085225e+00 8.43334734e-01
5.27788043e-01 -7.17931271e-01 -1.34918952e+00 7.15849221e-01
1.27750981e+00 1.96583524e-01 -8.51540089e-01 3.37872535e-01
6.02478802e-01 -2.23087907e-01 9.78434503e-01 -1.14522719e+00
4.53781813e-01 1.38608381e-01 9.59631130e-02 -1.46308005e+00
-5.69531024e-02 -8.40679049e-01 1.01204617e-02 6.07008636e-01
6.90911472e-01 -6.99757099e-01 4.36781973e-01 5.42021453e-01
2.13071451e-01 -9.88747001e-01 -7.00441718e-01 -1.38721704e-01
4.31615040e-02 -2.49243885e-01 3.30713123e-01 8.60021532e-01
5.73808491e-01 9.25550759e-01 -2.89446920e-01 -3.24269861e-01
2.34835166e-02 1.70767754e-01 5.86592853e-01 -1.49007559e+00
-1.60400704e-01 -3.33845973e-01 -2.29540706e-01 -7.15027928e-01
1.39517859e-01 -1.13279927e+00 -2.58605406e-02 -1.93374825e+00
4.79461461e-01 2.33228460e-01 -3.87822509e-01 -5.52443005e-02
-1.94916651e-01 -1.73736259e-01 -1.45147875e-01 -8.88219550e-02
-5.73131561e-01 5.05237818e-01 9.15388525e-01 -4.07403678e-01
-3.50193322e-01 -3.33038807e-01 -7.31660664e-01 5.03937304e-01
7.40569949e-01 -6.85881913e-01 -5.54245293e-01 -2.34024853e-01
5.57125926e-01 2.99710035e-01 -2.29970604e-01 -2.10001051e-01
3.12537014e-01 2.06665248e-01 2.71600515e-01 -7.84657478e-01
4.19679761e-01 -6.56449795e-01 -1.13119461e-01 5.59935272e-01
-7.20607698e-01 5.60799420e-01 3.63118976e-01 7.91397512e-01
-6.60680085e-02 -1.91519082e-01 7.96255991e-02 -8.27752501e-02
-6.89074025e-02 3.00270051e-01 -1.02233636e+00 3.43796387e-02
7.06863046e-01 2.54358619e-01 -4.19774503e-01 -6.55129194e-01
-8.10773611e-01 5.95265150e-01 -3.62069190e-01 8.79107296e-01
1.08078051e+00 -9.35828567e-01 -1.01336944e+00 5.47308065e-02
2.90579200e-01 -5.54818094e-01 3.66094947e-01 4.27616924e-01
-5.87078929e-01 8.27276289e-01 1.54125899e-01 -2.71635592e-01
-1.41334331e+00 7.42342532e-01 3.91528346e-02 -3.84853005e-01
-4.96286929e-01 5.84265888e-01 -8.37546587e-02 -6.56490862e-01
3.41672689e-01 -5.83533496e-02 -8.47156763e-01 4.88120645e-01
1.17314184e+00 3.43476743e-01 -2.00509965e-01 -6.07863888e-02
-4.96808141e-01 2.87147433e-01 -2.76331872e-01 -1.51598170e-01
1.37634408e+00 1.57337561e-01 -3.65158945e-01 3.93475503e-01
1.55691314e+00 -1.07043289e-01 2.14445546e-01 -2.49965057e-01
2.08624512e-01 1.65579736e-01 -3.83295231e-02 -8.22973430e-01
-5.10543406e-01 8.02669823e-01 8.32422137e-01 4.19341654e-01
6.87931418e-01 2.37061247e-01 1.25439048e+00 1.21262467e+00
-3.40094239e-01 -1.04949331e+00 2.41183728e-01 5.64602315e-01
1.11801553e+00 -1.24729204e+00 -3.79589260e-01 3.05747569e-01
-5.71742654e-01 1.16762280e+00 2.13301107e-02 5.85477613e-02
1.12803280e+00 2.30148107e-01 5.93506634e-01 -6.61755145e-01
-1.35667539e+00 -1.73324291e-02 3.23397875e-01 6.08237326e-01
6.59894526e-01 -5.32167442e-02 -4.75072592e-01 5.81042349e-01
-5.79261295e-02 2.03879908e-01 1.18726462e-01 8.19249272e-01
-5.24373949e-01 -1.15929866e+00 1.93611868e-02 6.41054034e-01
-6.47197843e-01 -4.32189792e-01 -7.31965899e-01 3.64078343e-01
-3.39975685e-01 1.46558988e+00 -5.45685887e-02 -4.00504380e-01
2.39000648e-01 5.81222117e-01 -8.00318196e-02 -7.48533130e-01
-7.30942667e-01 -1.68926775e-01 4.61538315e-01 -2.28269294e-01
-4.55338150e-01 -2.41295602e-02 -1.18688226e+00 -2.86055237e-01
-3.31329614e-01 6.11124396e-01 6.69999003e-01 9.34365094e-01
9.91458595e-01 3.47441226e-01 4.82401609e-01 -3.60978663e-01
-2.47046977e-01 -1.26502275e+00 9.57202613e-02 1.31924376e-01
3.42505932e-01 -1.13921069e-01 -2.08374187e-01 -1.67318895e-01]
|
[8.644817352294922, 8.673627853393555]
|
abb1ae53-d712-4f0b-a3f9-a50ddafba85e
|
mpc-multi-view-probabilistic-clustering
| null | null |
http://openaccess.thecvf.com//content/CVPR2022/html/Liu_MPC_Multi-View_Probabilistic_Clustering_CVPR_2022_paper.html
|
http://openaccess.thecvf.com//content/CVPR2022/papers/Liu_MPC_Multi-View_Probabilistic_Clustering_CVPR_2022_paper.pdf
|
MPC: Multi-View Probabilistic Clustering
|
Despite the promising progress having been made, the two challenges of multi-view clustering (MVC) are still waiting for better solutions: i) Most existing methods are either not qualified or require additional steps for incomplete multi-view clustering and ii) noise or outliers might significantly degrade the overall clustering performance. In this paper, we propose a novel unified framework for incomplete and complete MVC named multi-view probabilistic clustering (MPC). MPC equivalently transforms multi-view pairwise posterior matching probability into composition of each view's individual distribution, which tolerates data missing and might extend to any number of views. Then graph-context-aware refinement with path propagation and co-neighbor propagation is used to refine pairwise probability, which alleviates the impact of noise and outliers. Finally, MPC also equivalently transforms probabilistic clustering's objective to avoid complete pairwise computation and adjusts clustering assignments by maximizing joint probability iteratively. Extensive experiments on multiple benchmarks for incomplete and complete MVC show that MPC significantly outperforms previous state-of-the-art methods in both effectiveness and efficiency.
|
['Jianqiang Huang', 'Chen Shen', 'Yaowu Chen', 'Boxuan Gu', 'Xiang Tian', 'Rongxin Jiang', 'Shaotian Yan', 'Junlong Liu', 'Junjie Liu']
|
2022-01-01
| null | null | null |
cvpr-2022-1
|
['incomplete-multi-view-clustering']
|
['computer-vision']
|
[-1.46888420e-01 -2.19503373e-01 -1.82467680e-02 -3.97119135e-01
-1.05573332e+00 -7.62822926e-01 3.23100746e-01 3.17741424e-01
8.68323892e-02 2.84187317e-01 3.31184089e-01 1.49762139e-01
-2.86005735e-01 -5.30096531e-01 -5.57080150e-01 -8.67227733e-01
-5.80771640e-02 9.47545409e-01 6.46116793e-01 4.18703109e-01
1.81519970e-01 6.16313294e-02 -1.45725870e+00 4.29336965e-01
8.56943905e-01 4.48459685e-01 2.24917680e-01 5.38442791e-01
3.12793069e-02 3.11621457e-01 -4.28847611e-01 -4.48867083e-01
1.97575390e-01 -2.82982111e-01 -8.05350602e-01 5.30836046e-01
1.48631573e-01 1.53965622e-01 1.29991546e-01 1.18646908e+00
4.93586987e-01 2.46493239e-02 6.20988071e-01 -1.65890527e+00
-4.84728903e-01 8.17161858e-01 -1.30032444e+00 -2.43923992e-01
2.78269738e-01 -3.03387176e-02 8.62537026e-01 -9.02630508e-01
6.50734782e-01 1.26707375e+00 6.56181335e-01 1.75223276e-01
-1.43310106e+00 -2.82350212e-01 5.54974496e-01 2.74690539e-01
-1.81684732e+00 -1.42842323e-01 6.59082174e-01 -4.00069833e-01
5.27528882e-01 3.41254801e-01 3.63827318e-01 6.73858702e-01
-2.08043039e-01 6.72177851e-01 1.10854733e+00 -1.13058779e-02
2.90182769e-01 7.93619305e-02 -3.13399471e-02 5.08639157e-01
2.54543543e-01 -4.34391320e-01 -3.80516112e-01 -4.99811381e-01
2.84908384e-01 2.26884693e-01 -4.24762338e-01 -9.19760585e-01
-1.53892589e+00 6.07187450e-01 1.15023270e-01 3.44068371e-02
-2.18087509e-01 -1.01684108e-01 3.50022525e-01 -1.72821373e-01
1.30234659e-01 -1.63963698e-02 -4.99825388e-01 5.30014224e-02
-1.07155931e+00 1.19866803e-01 6.94185913e-01 1.33499634e+00
8.80727649e-01 -4.95597839e-01 -1.25946440e-02 8.64122331e-01
3.14588010e-01 4.96962994e-01 -7.96138421e-02 -1.30964589e+00
4.68450934e-01 9.98568714e-01 -7.10298866e-02 -1.26033139e+00
-3.00826877e-01 -3.48585010e-01 -1.16841757e+00 -9.85155907e-03
2.56922811e-01 1.04057781e-01 -7.90322185e-01 1.51002169e+00
7.14502752e-01 2.05141604e-01 -1.80053100e-01 9.29937959e-01
6.93316340e-01 5.51154554e-01 -3.82147014e-01 -4.59327519e-01
1.28520775e+00 -9.74005222e-01 -6.04855239e-01 2.06943184e-01
1.90869600e-01 -9.11883175e-01 7.72128940e-01 5.34903824e-01
-1.09169710e+00 -3.83638591e-01 -8.97453725e-01 2.96623856e-01
-1.78881064e-01 -2.18993172e-01 2.97790736e-01 6.50524616e-01
-1.08993757e+00 3.45706612e-01 -9.65025008e-01 -3.18740278e-01
2.90073246e-01 4.17171061e-01 -4.46392715e-01 -5.47763705e-01
-4.78190392e-01 2.38777041e-01 4.55563009e-01 -3.90163995e-02
-6.94125056e-01 -8.70649099e-01 -7.33655334e-01 8.08555782e-02
8.38827908e-01 -8.57757568e-01 5.05329609e-01 -3.69743019e-01
-1.03174067e+00 5.69190681e-01 -4.52752084e-01 1.15975350e-01
3.39372158e-01 -2.04193071e-02 -4.69383448e-01 1.99505493e-01
3.91284436e-01 6.67935908e-01 6.13192201e-01 -2.14870572e+00
-7.33557582e-01 -6.75686717e-01 -3.76771957e-01 3.78810138e-01
1.61862466e-02 -1.55792609e-01 -1.57200539e+00 -5.35024107e-01
6.37031972e-01 -1.08897531e+00 -5.90918422e-01 -3.79751951e-01
-7.47135639e-01 -4.34208848e-02 1.04665995e+00 -3.39068145e-01
1.37478852e+00 -2.08073425e+00 6.06041193e-01 5.70351124e-01
3.93470556e-01 -3.02523226e-01 3.26350145e-02 5.60984433e-01
1.05371006e-01 2.79725194e-01 -6.38698816e-01 -7.65972018e-01
-6.14183471e-02 4.04421598e-01 1.73337087e-01 7.26747274e-01
-3.63853216e-01 4.29307818e-01 -8.25100422e-01 -8.05773914e-01
3.08970749e-01 4.25315887e-01 -7.66704559e-01 8.78931880e-02
-6.84368461e-02 6.88207150e-01 -5.61864153e-02 8.93714488e-01
1.23792827e+00 -5.15282869e-01 6.04268849e-01 -3.73734534e-01
4.56225909e-02 -5.27324915e-01 -1.78988326e+00 1.89050210e+00
8.05549920e-02 -1.39582515e-01 4.18979079e-01 -7.23774016e-01
5.60354531e-01 1.96577400e-01 7.64369726e-01 1.19068243e-01
-1.89711943e-01 -1.04021132e-01 -3.48419249e-01 -1.36498258e-01
4.40971136e-01 2.02727661e-01 -1.83380917e-01 5.02398193e-01
-9.02545601e-02 3.28108445e-02 1.79303259e-01 5.87211013e-01
9.53141868e-01 1.13076523e-01 1.20432176e-01 -2.95512497e-01
7.46775806e-01 5.94729669e-02 1.06047964e+00 4.63957816e-01
-2.63358682e-01 1.29567587e+00 5.25132060e-01 3.02886334e-03
-7.14414060e-01 -1.42442095e+00 2.63374120e-01 8.06573749e-01
5.01007795e-01 -9.02953386e-01 -7.60792077e-01 -9.88345146e-01
-2.57994264e-01 4.31390047e-01 -3.78764629e-01 2.22870409e-01
-3.43284190e-01 -9.61887300e-01 -2.56607961e-02 4.49336946e-01
2.02986047e-01 -3.06032538e-01 1.06320091e-01 7.34186918e-02
-6.11730456e-01 -1.32833803e+00 -8.07783604e-01 -1.58510759e-01
-8.66070747e-01 -1.37299514e+00 -5.73484838e-01 -5.84925294e-01
9.31481838e-01 7.39024520e-01 1.15632677e+00 1.33796021e-01
-2.06544041e-03 6.56093657e-01 -4.42865133e-01 2.39111885e-01
-2.02590838e-01 1.33445878e-02 -1.09705096e-02 2.17835009e-01
2.71697670e-01 -8.98548901e-01 -8.04261267e-01 7.30097353e-01
-7.96740413e-01 -9.61169600e-02 4.29374397e-01 5.25327027e-01
1.16384518e+00 5.16685188e-01 1.78258881e-01 -1.24145830e+00
1.06455669e-01 -6.40919983e-01 -5.45527458e-01 5.18149853e-01
-7.03631699e-01 -1.57835767e-01 4.93097514e-01 8.59455485e-03
-1.06125402e+00 2.97108591e-01 2.38455236e-01 -7.70168960e-01
-3.93537551e-01 3.16603869e-01 -6.95691526e-01 3.21123153e-01
-6.40181871e-03 2.10464448e-01 -2.64202774e-01 -5.34120142e-01
6.28615081e-01 2.74841636e-01 7.31445312e-01 -6.03686750e-01
9.38272357e-01 1.01974678e+00 -3.64338011e-02 -4.71915632e-01
-6.57467723e-01 -9.63700831e-01 -9.95810032e-01 -2.81437039e-01
1.04017210e+00 -1.06802320e+00 -8.65886807e-01 4.30568725e-01
-8.49899590e-01 1.54262185e-01 2.84954846e-01 3.87094468e-01
-3.59549969e-01 8.79613459e-01 -4.26936537e-01 -4.87481743e-01
-3.07860691e-02 -1.41071892e+00 1.15692806e+00 -1.05417771e-02
3.39478590e-02 -1.02094901e+00 6.04926609e-02 6.55417323e-01
-4.38287668e-02 4.21087742e-01 8.18604529e-01 -4.31757212e-01
-8.15289199e-01 -9.32041183e-02 -1.95434049e-01 1.50631711e-01
5.54447621e-02 3.76795292e-01 -6.89015090e-01 -5.20285726e-01
-7.78295696e-02 1.86279014e-01 6.69488251e-01 4.16041911e-01
1.24609935e+00 -1.42517820e-01 -7.02373564e-01 7.69447446e-01
1.71006858e+00 -9.20609534e-02 4.17014927e-01 -1.35900661e-01
1.07892597e+00 6.49327636e-01 6.50396824e-01 6.85062468e-01
9.75980818e-01 5.36756039e-01 7.08828092e-01 1.12261340e-01
1.72616802e-02 -1.99558109e-01 1.32764488e-01 1.30361652e+00
-1.34113222e-01 -3.48680705e-01 -8.82597744e-01 7.72266448e-01
-2.10711241e+00 -9.43574846e-01 -6.33887589e-01 2.27063107e+00
3.46654713e-01 -4.75847907e-02 2.82892466e-01 9.53767002e-02
1.22867918e+00 2.28699848e-01 -3.01147282e-01 7.62225837e-02
-1.23052061e-01 -4.19671774e-01 5.32020867e-01 4.50066984e-01
-1.14574242e+00 6.34634197e-01 5.93076563e+00 8.17880750e-01
-2.63122052e-01 2.89821804e-01 6.20874643e-01 -8.01773071e-02
-5.47202766e-01 3.21595132e-01 -7.02526867e-01 4.69206363e-01
3.95404756e-01 3.41075540e-01 4.84085143e-01 7.29543746e-01
-3.41799222e-02 -3.35515827e-01 -9.97668087e-01 1.09020054e+00
2.18849942e-01 -1.15586638e+00 1.10420004e-01 1.82824388e-01
1.24956810e+00 -5.39710373e-02 -1.41073376e-01 3.53547819e-02
6.12100124e-01 -5.51694095e-01 5.02942443e-01 5.33196330e-01
4.54594404e-01 -1.14661622e+00 7.00414538e-01 4.97449338e-01
-1.58349931e+00 1.72418743e-01 -3.26003194e-01 5.21903992e-01
3.29327494e-01 8.34953427e-01 -5.56623638e-01 1.09573615e+00
9.86942947e-01 7.69932151e-01 -7.00453639e-01 9.27641213e-01
4.26727301e-03 5.16012251e-01 -4.15083289e-01 5.16950428e-01
5.00471778e-02 -4.91819501e-01 7.84891069e-01 1.08726764e+00
2.25035936e-01 1.58006735e-02 6.84979320e-01 6.27183437e-01
1.12436898e-01 -8.60393792e-02 -2.94693232e-01 5.80479085e-01
9.31710243e-01 1.40881991e+00 -1.23709798e+00 -3.29688281e-01
-6.15029693e-01 1.20376682e+00 2.73496956e-01 4.28055674e-01
-9.12852883e-01 1.77242890e-01 5.52764118e-01 9.23468247e-02
6.05815411e-01 -2.39596680e-01 -5.37174165e-01 -1.16712070e+00
1.94687620e-02 -7.37118244e-01 9.02322114e-01 -5.10896564e-01
-1.42930174e+00 3.05156291e-01 8.90064985e-02 -1.24809659e+00
2.86908656e-01 2.13805307e-02 -5.88868916e-01 4.30966586e-01
-1.07506359e+00 -1.31009173e+00 -3.02259386e-01 8.99854243e-01
5.47688484e-01 1.17153555e-01 6.34832323e-01 4.13570106e-01
-6.29404187e-01 4.29804355e-01 1.22183710e-01 -2.62945652e-01
9.60470796e-01 -1.51603675e+00 9.97748505e-03 1.10525596e+00
-3.48953381e-02 5.88855565e-01 4.99887496e-01 -6.63860261e-01
-1.47005546e+00 -1.24304569e+00 4.82255816e-01 -6.57964289e-01
2.94297874e-01 -3.56711149e-01 -8.91179562e-01 5.79390645e-01
3.57162684e-01 7.31369257e-02 9.23630893e-01 1.62590131e-01
-5.18981516e-01 -2.52913445e-01 -1.07985473e+00 5.95553219e-01
1.04096615e+00 -2.27486849e-01 -2.17266873e-01 2.39977956e-01
7.56665766e-01 -1.85593054e-01 -1.22397196e+00 5.61629713e-01
1.01255290e-01 -1.46809661e+00 1.09166598e+00 -4.90139164e-02
3.14998440e-02 -9.76481318e-01 -5.30853987e-01 -1.17113936e+00
-4.56816405e-01 -6.94435775e-01 -1.59904405e-01 1.82739747e+00
3.38671237e-01 -2.84055412e-01 7.74439394e-01 4.42276865e-01
-3.00208002e-01 -6.63517296e-01 -7.07958579e-01 -5.33466101e-01
-3.05512667e-01 -4.58650053e-01 7.74486601e-01 1.20278943e+00
-2.00513989e-01 2.16759130e-01 -4.52690482e-01 8.48311961e-01
1.17298841e+00 5.21006107e-01 1.01142323e+00 -1.05323398e+00
-3.84600580e-01 -1.67709023e-01 -1.59245223e-01 -8.12886596e-01
-1.47402614e-01 -7.92837620e-01 2.30719782e-02 -1.82740581e+00
7.97030747e-01 -4.06800508e-01 -7.40443766e-02 3.04434914e-02
-6.77997828e-01 2.03534305e-01 1.88205734e-01 3.69085819e-01
-1.26771522e+00 4.45104212e-01 1.13219464e+00 1.09823905e-01
-9.39392298e-02 1.14159279e-01 -7.55197585e-01 8.65888298e-01
5.30772567e-01 -6.29117548e-01 -4.66077566e-01 -2.68968463e-01
1.38831049e-01 3.19251865e-01 1.93606049e-01 -9.65338945e-01
6.06690466e-01 -1.88292235e-01 2.21498489e-01 -1.40171587e+00
2.52552241e-01 -1.12670684e+00 4.96922344e-01 1.76056027e-01
2.63867795e-01 3.69262397e-01 -2.66885102e-01 1.23599911e+00
-3.94978255e-01 1.84539422e-01 7.96730399e-01 -2.59649426e-01
-6.40841842e-01 5.39922774e-01 -3.07439446e-01 2.55403727e-01
1.21871138e+00 -2.12064922e-01 -1.70504481e-01 -3.58568549e-01
-9.89034832e-01 8.07262421e-01 9.66852307e-01 2.69102126e-01
5.81057012e-01 -1.31682372e+00 -6.38018489e-01 -8.58369321e-02
2.04723254e-01 3.24727446e-01 7.60869145e-01 1.05446982e+00
-3.33753526e-01 -7.15699270e-02 4.42621201e-01 -1.10903525e+00
-1.49611998e+00 1.02312291e+00 -9.88743827e-02 -3.20108384e-01
-5.09051263e-01 7.06552625e-01 2.20254019e-01 -8.14621329e-01
2.35724136e-01 1.40539303e-01 2.82799471e-02 -1.10806212e-01
1.60243094e-01 7.78988123e-01 -2.00489327e-01 -7.62094676e-01
-6.34581506e-01 7.74933100e-01 3.95424962e-02 -7.62826726e-02
1.31810331e+00 -7.28178620e-01 -5.10617435e-01 4.66270596e-01
1.14619672e+00 2.40020782e-01 -1.20130718e+00 -1.41288340e-01
5.61701581e-02 -4.74540055e-01 -3.21679026e-01 -6.32106721e-01
-1.38406801e+00 6.17499888e-01 2.08755851e-01 -1.61719639e-02
1.20574856e+00 2.27853075e-01 7.41646230e-01 -7.31844008e-02
3.51313740e-01 -1.09703612e+00 -1.44720748e-01 1.80388793e-01
4.91829216e-01 -1.18897307e+00 3.36577684e-01 -9.61507380e-01
-1.01386404e+00 7.08324015e-01 7.04491138e-01 6.26045167e-02
1.00026643e+00 3.18802536e-01 -6.85472339e-02 -4.62143749e-01
-8.07463169e-01 -3.76551114e-02 1.95092008e-01 7.25423574e-01
1.04510002e-01 1.46366477e-01 -1.60764884e-02 5.76430976e-01
2.51921862e-01 -5.47850132e-01 4.01088893e-01 7.46382952e-01
-2.33839825e-01 -1.06278694e+00 -7.15456903e-01 2.41722643e-01
-4.89034861e-01 6.27570152e-02 -3.37246656e-01 6.74337566e-01
3.25287938e-01 1.25993574e+00 -1.10204667e-01 -3.83043826e-01
2.25809202e-01 -2.02804148e-01 2.17246681e-01 -6.29085958e-01
-4.57592309e-01 6.48650646e-01 -1.36477292e-01 -7.70735264e-01
-8.36847961e-01 -9.71506417e-01 -1.15598547e+00 -4.36214507e-01
-4.94306386e-01 2.55910903e-01 3.54200602e-01 6.20401502e-01
6.73340619e-01 4.45553064e-01 7.07903862e-01 -6.38312697e-01
-1.68006107e-01 -6.22657418e-01 -6.10060930e-01 4.93909359e-01
-1.41418576e-01 -5.28697312e-01 -3.52817059e-01 2.34119177e-01]
|
[8.237527847290039, 4.633541584014893]
|
7ae4d95d-4a20-4687-aafc-fb6799893b16
|
regularizing-face-verification-nets-for-pain
|
1702.06925
| null |
http://arxiv.org/abs/1702.06925v3
|
http://arxiv.org/pdf/1702.06925v3.pdf
|
Regularizing Face Verification Nets For Pain Intensity Regression
|
Limited labeled data are available for the research of estimating facial
expression intensities. For instance, the ability to train deep networks for
automated pain assessment is limited by small datasets with labels of
patient-reported pain intensities. Fortunately, fine-tuning from a
data-extensive pre-trained domain, such as face verification, can alleviate
this problem. In this paper, we propose a network that fine-tunes a
state-of-the-art face verification network using a regularized regression loss
and additional data with expression labels. In this way, the expression
intensity regression task can benefit from the rich feature representations
trained on a huge amount of data for face verification. The proposed
regularized deep regressor is applied to estimate the pain expression intensity
and verified on the widely-used UNBC-McMaster Shoulder-Pain dataset, achieving
the state-of-the-art performance. A weighted evaluation metric is also proposed
to address the imbalance issue of different pain intensities.
|
['Trac. D. Tran', 'Gregory D. Hager', 'Feng Wang', 'Xiang Xiang', 'Jian Cheng', 'Chang Liu', 'Harry Quon', 'Austin Reiter', 'Alan L. Yuille']
|
2017-02-22
| null | null | null | null |
['pain-intensity-regression']
|
['medical']
|
[ 1.00565232e-01 1.33177331e-02 -7.99657404e-01 -9.73219454e-01
-1.16337287e+00 6.04617484e-02 4.00674194e-02 -1.23884760e-01
-6.18128240e-01 8.71110439e-01 2.30807349e-01 3.15792322e-01
-4.09944616e-02 -4.06442642e-01 -4.17887360e-01 -8.55856121e-01
-5.76755367e-02 2.55458802e-01 -9.47261035e-01 -2.44733706e-01
-6.70872554e-02 6.34923816e-01 -1.47675514e+00 4.14824694e-01
6.30276024e-01 1.84465063e+00 -3.55456144e-01 -9.63372513e-05
3.29928324e-02 8.45811605e-01 -5.66110373e-01 -4.33503479e-01
1.21190302e-01 -3.63175541e-01 -6.90761387e-01 -1.27344355e-01
3.74932885e-01 -6.14318967e-01 2.84630898e-02 7.10143507e-01
7.10734665e-01 -2.10277960e-01 5.18463135e-01 -1.49265516e+00
-3.46624762e-01 2.17856750e-01 -8.99031401e-01 -1.90238789e-01
2.21727446e-01 4.25635977e-03 7.23586380e-01 -7.58021832e-01
5.83832026e-01 1.03394639e+00 6.58342361e-01 9.41328526e-01
-1.04775679e+00 -1.31578171e+00 -1.89894259e-01 2.25402132e-01
-1.50045240e+00 -4.02334929e-01 1.14533198e+00 -3.42916161e-01
6.05351150e-01 -5.08319912e-03 5.99311769e-01 1.55776072e+00
7.90331513e-02 4.63773161e-01 1.62375104e+00 -1.40171245e-01
3.08486402e-01 1.17917418e-01 -2.56917238e-01 8.19194555e-01
-2.20437869e-01 2.19336495e-01 -8.50539207e-01 -1.54819369e-01
3.13205332e-01 -6.94612563e-02 -1.09310873e-01 -9.53302383e-02
-3.65223080e-01 1.05810702e+00 6.87726676e-01 2.00383827e-01
-4.25260872e-01 1.70012400e-01 7.89359808e-01 3.10017258e-01
9.15731132e-01 3.54874700e-01 -6.78693295e-01 -2.13575318e-01
-1.22731078e+00 6.87274262e-02 5.68717957e-01 1.74826816e-01
8.13616097e-01 -1.33856013e-03 -3.36692244e-01 1.18427098e+00
2.04444632e-01 2.54587173e-01 5.09277642e-01 -1.01359499e+00
1.42326102e-01 6.30831063e-01 -2.45853499e-01 -9.08043087e-01
-6.37934268e-01 -4.13616449e-01 -1.16478539e+00 8.00829589e-01
2.73500890e-01 -4.32994694e-01 -9.95176792e-01 2.42929149e+00
3.46839637e-01 2.06706952e-02 -2.03437373e-01 1.08391154e+00
8.69937420e-01 1.90877005e-01 2.64177680e-01 -5.18529356e-01
1.41998875e+00 -8.43495071e-01 -7.68150985e-01 -2.29380969e-02
6.81673288e-01 -3.95739585e-01 1.09042728e+00 6.95529640e-01
-8.66790712e-01 -5.10991037e-01 -8.84311855e-01 3.89474886e-03
-3.13739419e-01 3.03082138e-01 9.31327999e-01 5.54974496e-01
-8.52226198e-01 6.47426546e-01 -5.93095899e-01 -1.62811801e-02
1.08766580e+00 6.78594172e-01 -8.47476423e-01 -1.70356318e-01
-1.23639524e+00 1.16961515e+00 -5.97803481e-02 4.52037781e-01
-1.01805651e+00 -8.05133045e-01 -7.96866715e-01 1.16882823e-01
2.27832511e-01 -5.07105291e-01 1.09983301e+00 -1.50457573e+00
-1.67236388e+00 1.34781730e+00 2.08932627e-02 -8.36102292e-02
4.53177929e-01 -2.70050317e-01 -2.52843618e-01 2.12837145e-01
-4.49346043e-02 9.03374434e-01 1.00498414e+00 -8.74239385e-01
2.18427498e-02 -8.30956936e-01 -2.98034579e-01 9.81165003e-03
-5.15947938e-01 2.81353682e-01 -2.40805581e-01 -4.85702842e-01
-5.33262908e-01 -9.25048113e-01 -1.82559386e-01 4.19611126e-01
-2.09002510e-01 -3.34446520e-01 7.26775944e-01 -7.09982336e-01
9.20266807e-01 -2.16050315e+00 1.98996067e-01 3.11402112e-01
6.53022304e-02 -5.83866835e-02 -3.84345680e-01 -2.48740357e-03
-4.34531599e-01 -7.93613773e-03 -1.58532962e-01 -8.33083630e-01
-1.93503037e-01 1.90662503e-01 8.17300677e-02 6.28073335e-01
6.29430354e-01 6.84540451e-01 -4.59184170e-01 -6.49720669e-01
-9.49401930e-02 7.02494323e-01 -5.45533717e-01 4.11946535e-01
7.78267672e-03 8.47550392e-01 -4.04713780e-01 9.28035915e-01
8.13081205e-01 -9.62869897e-02 -1.10755801e-01 -3.68466049e-01
3.23042035e-01 -1.76899001e-01 -6.40558422e-01 2.29071832e+00
-7.15392649e-01 2.94753164e-01 5.26810646e-01 -1.01282156e+00
9.08015013e-01 3.66610676e-01 8.88725758e-01 -8.25931549e-01
5.97174704e-01 2.47872174e-01 -9.41091999e-02 -4.29850250e-01
-8.53246450e-02 -7.52495468e-01 -1.92702606e-01 3.58072728e-01
3.41876149e-01 -1.01560891e-01 -2.12976217e-01 -2.13958338e-01
1.00740099e+00 -1.53108267e-02 9.44333375e-02 4.42359261e-02
4.65159297e-01 -4.00340706e-01 7.19301999e-01 1.96785331e-01
-5.15610158e-01 4.29815054e-01 6.64882958e-01 -3.11589807e-01
-6.02213323e-01 -6.13164186e-01 -4.89578843e-01 1.32991707e+00
-4.11759943e-01 -3.11726779e-01 -8.64769876e-01 -7.69680500e-01
1.63176864e-01 7.84630775e-02 -1.30551970e+00 -4.40349966e-01
-9.34929773e-02 -7.57313907e-01 7.34218419e-01 6.30497634e-01
3.35526377e-01 -1.32748461e+00 -3.89395028e-01 -1.39592394e-01
-9.71817002e-02 -7.75280654e-01 -2.35667918e-02 4.69555289e-01
-5.00496924e-01 -9.25315201e-01 -8.33661079e-01 -5.60913265e-01
5.66724598e-01 -6.60670161e-01 9.48614120e-01 3.60745750e-02
-5.78419387e-01 1.28420502e-01 -1.40624344e-01 -5.31613410e-01
-4.56498973e-02 3.32420389e-03 1.67302743e-01 1.51128992e-01
4.03221756e-01 -5.06672800e-01 -8.88928413e-01 1.78803746e-02
-8.89511049e-01 -2.17800871e-01 7.71755159e-01 1.14447320e+00
6.22368157e-01 -2.41220906e-01 8.78362358e-01 -6.01911724e-01
7.22027302e-01 -3.67135257e-01 -2.72911787e-01 6.08822666e-02
-6.53198123e-01 5.93030043e-02 4.27246809e-01 -4.83955950e-01
-1.21246707e+00 1.68987945e-01 -4.33346510e-01 -7.01389134e-01
-1.37817860e-01 8.05921257e-01 1.31617987e-03 -4.47215319e-01
5.97709835e-01 -4.81737822e-01 3.75791788e-01 -2.70772368e-01
4.18868810e-01 6.71015263e-01 6.04055643e-01 -5.90411246e-01
4.07852888e-01 4.04405117e-01 3.94184828e-01 -1.79854095e-01
-1.18597710e+00 -2.24392459e-01 -3.68876308e-01 -2.34431937e-01
8.11606467e-01 -1.08806849e+00 -8.58475983e-01 6.36224449e-01
-9.23573792e-01 -3.55923414e-01 -2.70432889e-01 3.42516661e-01
-6.49122655e-01 6.75352216e-02 -8.97170007e-01 -7.44014442e-01
-8.44119370e-01 -1.12759495e+00 1.36074197e+00 5.89281358e-02
-4.15250272e-01 -4.80794936e-01 4.66479689e-01 6.59765005e-01
6.10615313e-01 6.52304888e-01 7.32750654e-01 -2.04269961e-01
1.30517066e-01 -4.24506992e-01 -2.86802292e-01 7.32297182e-01
2.82975584e-01 -8.83236751e-02 -1.49752617e+00 -1.63073078e-01
2.76255369e-01 -1.49734354e+00 8.54409516e-01 4.94668156e-01
1.90248060e+00 -1.52530253e-01 -2.55043730e-02 7.67347097e-01
1.17401254e+00 -1.65759191e-01 5.84205985e-01 -2.78410059e-03
3.87215227e-01 8.02027524e-01 5.94031632e-01 7.12974966e-01
-1.95031837e-02 7.65544474e-01 4.98821855e-01 -6.38091266e-01
9.06253830e-02 8.16741511e-02 -1.28914611e-02 3.46065402e-01
-1.68791950e-01 3.18698257e-01 -5.30888975e-01 4.60609905e-02
-1.49646330e+00 -7.06184089e-01 5.31332850e-01 1.75877583e+00
1.28387892e+00 -1.33523643e-01 9.63404775e-02 1.99066505e-01
2.55696088e-01 1.14264406e-01 -6.75441921e-01 -6.24008834e-01
1.00657204e-02 8.04351807e-01 5.22964410e-02 2.18098626e-01
-1.08659267e+00 6.87969863e-01 5.98731613e+00 1.12517333e+00
-1.68692875e+00 2.83371121e-01 1.22261679e+00 -4.49826926e-01
2.16491684e-01 -5.83277643e-01 -3.16983610e-01 4.24449354e-01
9.76718605e-01 2.94705719e-01 4.47797328e-01 1.06255651e+00
3.27930093e-01 -2.21941277e-01 -1.16800654e+00 1.43546546e+00
2.50289679e-01 -8.97286713e-01 -4.33032393e-01 -7.91184511e-03
4.30784941e-01 -5.25264442e-02 4.04944688e-01 5.73621392e-01
-1.89065054e-01 -1.66185701e+00 1.59991518e-01 3.64016384e-01
1.33971393e+00 -8.61510813e-01 9.23230648e-01 1.75404083e-02
-6.33689523e-01 -1.84566006e-01 -9.74605903e-02 -4.57478575e-02
-1.36234080e-02 7.27473021e-01 -5.01807868e-01 2.16301978e-01
7.98286676e-01 4.26158160e-01 -4.28414106e-01 4.53592062e-01
-1.87939808e-01 4.87512082e-01 -3.33653331e-01 3.10287327e-01
1.49020962e-02 -1.32892638e-01 -4.19934899e-01 1.02727687e+00
1.15786500e-01 -5.42979091e-02 -1.05165221e-01 7.30090797e-01
-5.87333858e-01 3.12155217e-01 -5.06447613e-01 6.19338863e-02
-1.96111873e-01 1.79675102e+00 -1.37990028e-01 -6.97631463e-02
-3.37381274e-01 8.51394832e-01 7.44743645e-01 8.59144926e-02
-8.85289848e-01 6.26188293e-02 7.11508274e-01 -1.23478949e-01
-2.04323173e-01 4.06665087e-01 -2.85755366e-01 -9.08253670e-01
-1.16818182e-01 -9.41179216e-01 3.57979745e-01 -8.82625103e-01
-1.69348192e+00 6.04097426e-01 -2.01565757e-01 -8.45978439e-01
-3.34374189e-01 -7.99761474e-01 -4.88452435e-01 8.63811314e-01
-1.67297399e+00 -1.32201290e+00 -6.38686359e-01 8.67106736e-01
-3.01435217e-02 -7.39733353e-02 1.23953617e+00 5.98093212e-01
-7.40582645e-01 9.89865005e-01 -4.69680578e-01 2.72699237e-01
9.82973039e-01 -7.67152607e-01 -8.17380130e-01 5.04123792e-03
-1.93689704e-01 3.18560719e-01 3.51509005e-01 -2.80055851e-01
-1.24856520e+00 -9.69372749e-01 4.10986423e-01 -1.09934121e-01
4.94230777e-01 -4.06170011e-01 -8.53552997e-01 2.92568713e-01
1.53091401e-01 5.25412083e-01 1.15505159e+00 4.85996008e-01
-6.18050039e-01 -6.69426918e-01 -1.60338843e+00 2.35389784e-01
7.40229368e-01 -7.87021637e-01 -1.61129639e-01 4.92554992e-01
1.23236828e-01 -4.43801552e-01 -1.07701099e+00 7.56505787e-01
8.62849534e-01 -7.51358330e-01 6.93237305e-01 -9.80023563e-01
9.43746746e-01 3.07126880e-01 9.43453331e-03 -1.26343977e+00
-9.70725045e-02 -3.06048661e-01 -9.39264894e-04 1.12304211e+00
2.61337399e-01 -1.23805813e-01 1.16843188e+00 7.97612667e-01
2.50763744e-01 -1.41453314e+00 -1.31365871e+00 -4.05614376e-01
1.71084717e-01 -4.36425209e-01 2.74391323e-01 9.81071770e-01
2.74027318e-01 4.15754467e-01 -6.72676384e-01 -4.35607433e-01
3.31460565e-01 3.16267200e-02 5.88449299e-01 -1.17298794e+00
-2.30778217e-01 -3.49852562e-01 -4.98623252e-01 -3.46000083e-02
8.84884536e-01 -7.83635437e-01 7.38854781e-02 -9.85523283e-01
6.42853856e-01 -3.22041243e-01 -8.80578041e-01 9.53572929e-01
8.33615363e-02 4.92310911e-01 -6.35045841e-02 -2.55352765e-01
-3.27629179e-01 9.81882274e-01 1.17966986e+00 -1.71203449e-01
4.13787998e-02 -2.99089164e-01 -8.07436407e-01 6.31458104e-01
8.81170928e-01 -5.58239758e-01 -4.30792898e-01 -1.12794517e-02
1.11889571e-01 2.65332967e-01 2.78047204e-01 -8.38580787e-01
-8.90916213e-02 -1.58162907e-01 7.43539333e-01 -2.47727230e-01
9.64467704e-01 -8.70698154e-01 -1.93959400e-02 4.04001474e-01
-5.04056871e-01 -2.30151132e-01 3.83326769e-01 2.55077630e-01
-4.65576857e-01 7.89863765e-02 9.90326881e-01 -1.37079796e-02
-3.40409636e-01 5.41428745e-01 4.50132377e-02 -7.63214901e-02
9.64533329e-01 2.50471860e-01 -2.21370623e-01 -5.68616092e-01
-7.40541875e-01 3.08158044e-02 1.82397887e-01 1.99017644e-01
5.55254519e-01 -1.56100118e+00 -6.77813470e-01 1.15272440e-01
2.94468284e-01 -9.60729495e-02 4.31273609e-01 1.12164009e+00
-6.68933019e-02 3.97948688e-03 -6.64791286e-01 -5.20839870e-01
-1.32843673e+00 5.03128111e-01 4.25289422e-01 -6.47091448e-01
-7.29787946e-02 9.70302403e-01 1.22951888e-01 -4.73605961e-01
3.21453154e-01 -2.53392369e-01 -2.92889271e-02 2.95053095e-01
4.70314890e-01 -9.00619105e-02 3.02462339e-01 -5.10849476e-01
-5.67337155e-01 6.48238659e-01 5.62213995e-02 -3.29434909e-02
1.44651043e+00 3.71849746e-01 -2.20040843e-01 4.31699455e-02
1.66465855e+00 -2.18538210e-01 -1.11985540e+00 1.03245221e-01
-2.54791051e-01 -1.94974184e-01 2.44577035e-01 -1.07449925e+00
-1.67961180e+00 8.76598656e-01 1.09030867e+00 -7.22777784e-01
1.69922888e+00 -8.95793065e-02 6.41244531e-01 3.61149400e-01
3.40128899e-01 -1.21347570e+00 4.53318805e-01 -5.41945957e-02
1.14232993e+00 -1.62838328e+00 -4.78397794e-02 -2.13600948e-01
-7.75680482e-01 9.09810126e-01 8.51757348e-01 1.24217840e-02
8.11337292e-01 5.23980796e-01 3.74233007e-01 -4.33067888e-01
-6.13118172e-01 3.39434668e-02 2.24914908e-01 3.96449357e-01
6.02042735e-01 2.97745168e-02 -4.76074547e-01 8.77112865e-01
-2.00235561e-01 6.88309491e-01 -1.02377415e-01 6.77568853e-01
1.39631271e-01 -1.02089202e+00 -1.30040154e-01 6.56855643e-01
-8.60634863e-01 1.38914600e-01 -5.89293480e-01 6.89591467e-01
2.81977534e-01 8.27035904e-01 -1.38950124e-01 -3.24245244e-01
7.53085092e-02 2.24858686e-01 6.27804458e-01 -3.42614919e-01
-5.87263048e-01 -1.64279342e-01 1.94299638e-01 -8.91707361e-01
-6.07290983e-01 -1.59577116e-01 -1.04429293e+00 -4.85906191e-02
-8.19197372e-02 -5.56358397e-02 8.34485471e-01 8.37163746e-01
3.58480483e-01 2.38751456e-01 8.40393722e-01 -8.79103363e-01
-6.32199585e-01 -1.23791754e+00 -8.65308106e-01 8.84610951e-01
3.76769155e-01 -1.02166986e+00 -3.10621649e-01 -4.22499269e-01]
|
[13.61054515838623, 1.69781494140625]
|
8f004ab1-7d93-4366-a3a0-e74ba7086fc8
|
text-summarization-with-pretrained-encoders
|
1908.08345
| null |
https://arxiv.org/abs/1908.08345v2
|
https://arxiv.org/pdf/1908.08345v2.pdf
|
Text Summarization with Pretrained Encoders
|
Bidirectional Encoder Representations from Transformers (BERT) represents the latest incarnation of pretrained language models which have recently advanced a wide range of natural language processing tasks. In this paper, we showcase how BERT can be usefully applied in text summarization and propose a general framework for both extractive and abstractive models. We introduce a novel document-level encoder based on BERT which is able to express the semantics of a document and obtain representations for its sentences. Our extractive model is built on top of this encoder by stacking several inter-sentence Transformer layers. For abstractive summarization, we propose a new fine-tuning schedule which adopts different optimizers for the encoder and the decoder as a means of alleviating the mismatch between the two (the former is pretrained while the latter is not). We also demonstrate that a two-staged fine-tuning approach can further boost the quality of the generated summaries. Experiments on three datasets show that our model achieves state-of-the-art results across the board in both extractive and abstractive settings. Our code is available at https://github.com/nlpyang/PreSumm
|
['Yang Liu', 'Mirella Lapata']
|
2019-08-22
|
text-summarization-with-pretrained-encoders-1
|
https://aclanthology.org/D19-1387
|
https://aclanthology.org/D19-1387.pdf
|
ijcnlp-2019-11
|
['extractive-document-summarization']
|
['natural-language-processing']
|
[ 3.88410330e-01 4.34781313e-01 -1.58227041e-01 -4.56076801e-01
-1.05775917e+00 -4.63509142e-01 7.83638418e-01 2.21467942e-01
-3.27931613e-01 6.79879069e-01 1.04461658e+00 -2.29488671e-01
2.33493865e-01 -6.99119568e-01 -9.55244720e-01 -3.96933675e-01
3.06599379e-01 3.56245458e-01 7.39063025e-02 -4.71749455e-01
2.99544632e-01 9.15801972e-02 -1.18645787e+00 7.89589942e-01
1.05622828e+00 7.21945703e-01 3.63525420e-01 8.65845263e-01
-2.88837731e-01 1.06806898e+00 -8.33525598e-01 -7.39068747e-01
-6.86111599e-02 -5.83487272e-01 -9.73598599e-01 -3.00571304e-02
3.45840126e-01 -5.57046771e-01 -5.74959457e-01 7.35481501e-01
5.77934146e-01 -3.97448204e-02 5.66214204e-01 -7.88595319e-01
-7.97599792e-01 1.30916190e+00 -3.96651119e-01 2.39961073e-01
2.52296448e-01 -6.12873025e-02 1.44838226e+00 -9.01376903e-01
4.09318119e-01 1.25582445e+00 3.32070380e-01 5.87736607e-01
-1.02801633e+00 -4.10030574e-01 3.48616064e-01 1.69433981e-01
-8.83904934e-01 -8.20520461e-01 7.61902511e-01 -5.81048280e-02
1.37888086e+00 3.22898209e-01 5.61212420e-01 1.15293312e+00
4.43938971e-01 1.33269155e+00 3.79723042e-01 -2.86602944e-01
-5.49362786e-02 -3.87403518e-02 1.71730593e-01 6.86605811e-01
2.27702022e-01 -4.61579561e-01 -8.50148797e-01 3.59070003e-02
2.87876934e-01 -1.27356872e-01 -4.68259275e-01 -7.82949179e-02
-1.12335348e+00 8.39905024e-01 5.73290706e-01 2.64773428e-01
-4.17526126e-01 3.06434959e-01 7.01918900e-01 2.85990179e-01
6.38333261e-01 5.34327924e-01 -3.14684033e-01 -2.40215272e-01
-1.20376170e+00 1.50372252e-01 8.90107870e-01 9.82891023e-01
3.99185687e-01 1.17274247e-01 -7.35280693e-01 9.97792006e-01
6.74575493e-02 2.21777067e-01 6.95964515e-01 -6.35237157e-01
8.39592397e-01 6.58307970e-01 -6.40225261e-02 -5.13135314e-01
-7.49622285e-02 -6.57425165e-01 -9.38373625e-01 -4.91365194e-01
-1.85608372e-01 -1.01425610e-01 -7.88802564e-01 1.57675612e+00
-8.84266198e-02 7.79134333e-02 2.04559430e-01 5.60860455e-01
1.02717614e+00 1.06917584e+00 -2.47817054e-01 -1.58152685e-01
1.29106736e+00 -1.39449120e+00 -7.86123097e-01 -5.14996350e-01
5.60425818e-01 -5.90894043e-01 1.02354276e+00 1.50271103e-01
-1.64278913e+00 -2.57921964e-01 -1.27077472e+00 -7.00379729e-01
-2.39176042e-02 4.70781863e-01 2.83745527e-01 4.37426083e-02
-1.25765371e+00 7.00716376e-01 -1.09654021e+00 -2.00146303e-01
5.52688062e-01 1.95546493e-01 -2.15721905e-01 1.89313173e-01
-1.05270851e+00 9.69143510e-01 5.19052088e-01 2.15502024e-01
-8.23820055e-01 -5.77939689e-01 -9.22060668e-01 5.81609547e-01
3.03655148e-01 -1.22868586e+00 1.61902654e+00 -6.59851849e-01
-1.84915733e+00 6.74545825e-01 -3.98594499e-01 -7.76667714e-01
3.57394785e-01 -6.40570581e-01 -3.78738012e-04 2.20609635e-01
1.54770032e-01 5.01514673e-01 7.26192772e-01 -8.55469286e-01
-5.44065654e-01 -2.01229736e-01 7.77970105e-02 3.27089906e-01
-5.61843693e-01 9.22856033e-02 -5.32235444e-01 -8.43267143e-01
-2.86370069e-01 -5.64919174e-01 2.40161978e-02 -2.96879709e-01
-8.61487865e-01 -3.01795989e-01 4.60161477e-01 -9.01742578e-01
1.55866027e+00 -1.98973119e+00 6.54498458e-01 -3.82670313e-01
1.89342052e-01 4.38952118e-01 -2.38421962e-01 9.68705654e-01
7.70705119e-02 1.20963119e-01 -5.94816923e-01 -8.17936540e-01
2.30710730e-01 1.74979389e-01 -7.88228452e-01 1.39784589e-01
4.85829324e-01 1.10377324e+00 -9.14634168e-01 -2.44266734e-01
-1.00493640e-01 4.51126099e-01 -6.69045806e-01 4.50763911e-01
-2.77481586e-01 1.44996777e-01 -3.89821082e-01 7.29718730e-02
4.31508213e-01 -2.67239690e-01 1.36518776e-01 -1.56479895e-01
-7.48137683e-02 1.15815628e+00 -6.62202239e-01 2.00943518e+00
-7.06092536e-01 7.09503293e-01 7.14197084e-02 -1.08347058e+00
7.26143003e-01 3.74281794e-01 -1.12421429e-02 -5.34572721e-01
1.61650240e-01 2.79484302e-01 -2.24499688e-01 -2.98168778e-01
9.70864713e-01 -1.99139863e-02 -1.87092394e-01 6.46987021e-01
4.14964199e-01 -2.74347723e-01 5.34429014e-01 6.00299776e-01
1.17331529e+00 1.04601746e-02 4.71471041e-01 -1.65377349e-01
6.85771883e-01 -3.42260063e-01 3.94867718e-01 7.30838478e-01
3.69032383e-01 6.50113344e-01 8.27888787e-01 -1.15600906e-01
-1.06999362e+00 -9.85483050e-01 2.35771060e-01 1.14203060e+00
-2.01780006e-01 -9.72648144e-01 -7.31741548e-01 -6.50272310e-01
-1.76808938e-01 1.14018464e+00 -4.95142609e-01 -5.38767159e-01
-8.46371412e-01 -4.95566785e-01 6.37406945e-01 7.31710792e-01
4.85935003e-01 -1.00077391e+00 -6.42480135e-01 2.74792820e-01
-4.84264582e-01 -1.09412634e+00 -6.72726512e-01 1.71045691e-01
-9.18440878e-01 -5.48532546e-01 -6.02993011e-01 -7.12721348e-01
4.40989941e-01 1.76028758e-01 1.26333797e+00 1.65996417e-01
2.57381588e-01 1.59858704e-01 -4.23936039e-01 -4.88276660e-01
-7.80474961e-01 6.73856318e-01 -3.79391551e-01 1.52846575e-02
-9.22999978e-02 -6.85266018e-01 -4.40236926e-01 -4.01370734e-01
-1.18310225e+00 5.41732550e-01 6.97592974e-01 7.96939731e-01
2.26454422e-01 -4.44945961e-01 6.16366386e-01 -8.97094488e-01
1.07254517e+00 -3.25119019e-01 -2.74757594e-01 3.58102947e-01
-2.76819468e-01 7.07627416e-01 8.82333815e-01 8.16968158e-02
-1.00800121e+00 -3.86058033e-01 -5.16794860e-01 -1.88492350e-02
4.09218669e-01 7.85212755e-01 -1.06462076e-01 7.03239441e-01
2.42750376e-01 5.33945501e-01 -1.59934565e-01 -6.80762291e-01
4.16233510e-01 9.39795732e-01 6.48555756e-01 -4.73085374e-01
5.95002115e-01 2.71982431e-01 -4.52773571e-01 -6.40741050e-01
-1.08655441e+00 -2.98212498e-01 -4.32044834e-01 1.63313359e-01
6.40573919e-01 -1.10027933e+00 -2.53034383e-01 3.43583167e-01
-1.49640191e+00 -3.61818016e-01 -5.46241343e-01 1.35996029e-01
-5.48480988e-01 3.19438130e-01 -7.11769104e-01 -3.53475869e-01
-1.07260573e+00 -1.13249648e+00 1.45947063e+00 1.33737609e-01
-1.74534500e-01 -9.08629119e-01 8.72945115e-02 2.19440207e-01
4.76893514e-01 -6.45457804e-02 9.25297499e-01 -7.43535042e-01
-3.99263859e-01 5.31444652e-03 -9.18296054e-02 5.93349576e-01
7.25575611e-02 -2.29193494e-02 -8.44358921e-01 -3.35695505e-01
1.20111920e-01 -3.49069178e-01 1.51517272e+00 2.67615318e-01
1.12740314e+00 -8.24755788e-01 -1.18104964e-01 5.85753143e-01
1.17070889e+00 -3.59698206e-01 7.56877422e-01 2.74703562e-01
6.46895170e-01 2.84208238e-01 1.20564789e-01 5.18652022e-01
6.99254036e-01 6.62084520e-01 2.99840450e-01 9.51439589e-02
-3.23440492e-01 -5.17731369e-01 8.16797674e-01 1.33755517e+00
1.01589203e-01 -5.71290135e-01 -6.41442657e-01 6.78303957e-01
-2.01941442e+00 -9.62916851e-01 2.12992698e-01 1.93259418e+00
1.17729592e+00 1.87685013e-01 -2.42393706e-02 1.48545569e-02
4.79473889e-01 5.25336742e-01 -4.94067132e-01 -9.35784161e-01
4.62203380e-03 3.80183369e-01 2.26184711e-01 6.06984258e-01
-8.72266948e-01 9.85758483e-01 5.53943205e+00 6.39188528e-01
-1.19441080e+00 -3.14758271e-02 3.57752502e-01 -3.37027669e-01
-6.07788384e-01 -6.92947656e-02 -7.40534544e-01 4.52012122e-01
1.15692806e+00 -7.05656767e-01 3.00362498e-01 5.32879710e-01
2.02261925e-01 2.26663604e-01 -1.37565291e+00 5.33925533e-01
2.32073233e-01 -1.66313255e+00 4.95721579e-01 -2.54833937e-01
5.69621563e-01 2.27712184e-01 -2.09919155e-01 4.38707232e-01
2.08336234e-01 -7.24952638e-01 1.01263475e+00 4.57978010e-01
6.99277163e-01 -6.18232250e-01 6.08655751e-01 5.22512019e-01
-1.01908660e+00 -1.58921614e-01 -3.43682230e-01 -4.52945679e-02
4.14976418e-01 5.56203127e-01 -9.01138246e-01 9.18108046e-01
3.66532743e-01 1.12627470e+00 -6.17185891e-01 7.42672145e-01
-6.61672771e-01 6.24047995e-01 -8.18666369e-02 -2.36227587e-02
3.17366362e-01 -5.62269986e-02 8.53842676e-01 1.50310159e+00
3.53904098e-01 -2.25563005e-01 -1.43197328e-01 8.50324512e-01
-5.59209406e-01 -4.64790314e-02 -3.71044755e-01 -1.88099176e-01
4.28587466e-01 1.08351433e+00 -2.01090112e-01 -6.04271591e-01
-2.16325894e-01 1.23336613e+00 7.42284834e-01 2.25288853e-01
-8.89626086e-01 -4.70810354e-01 4.08035129e-01 -1.15407109e-02
6.62484884e-01 -1.59291729e-01 -2.60001123e-01 -1.67474174e+00
2.54027575e-01 -9.35422659e-01 2.79241979e-01 -7.67349482e-01
-8.58719885e-01 7.35312998e-01 -3.86381634e-02 -9.48890626e-01
-5.09217024e-01 -2.37634510e-01 -8.58851135e-01 7.19235659e-01
-1.70110190e+00 -1.10483062e+00 -2.83271559e-02 1.92924634e-01
9.19716120e-01 4.38844264e-02 7.98527062e-01 9.59686562e-02
-7.69648075e-01 5.06005764e-01 2.31201231e-01 9.89653990e-02
5.84816635e-01 -1.36348784e+00 7.31093049e-01 1.15296924e+00
1.23903044e-01 7.21300423e-01 7.38437891e-01 -2.45711073e-01
-1.41873276e+00 -1.23223364e+00 1.31370366e+00 -1.72354296e-01
6.66122317e-01 -5.76050460e-01 -8.32043290e-01 9.87335980e-01
8.53207886e-01 -5.37956119e-01 5.77010393e-01 -2.11576652e-02
-3.36684406e-01 -2.10968599e-01 -5.68210781e-01 6.82533920e-01
8.81042361e-01 -4.62428600e-01 -9.02364194e-01 2.46541455e-01
9.87973273e-01 -5.70704520e-01 -6.38361156e-01 2.81142533e-01
4.20063436e-01 -9.96874511e-01 7.38962233e-01 -5.51669002e-01
1.11593473e+00 -1.13388780e-03 1.32918963e-02 -1.69718516e+00
-2.79796928e-01 -7.05646455e-01 -6.06532156e-01 1.32215834e+00
4.37841147e-01 -5.61232328e-01 2.73762256e-01 1.28154173e-01
-6.71455979e-01 -1.03776252e+00 -7.42087662e-01 -6.09429717e-01
2.88709968e-01 -5.00319116e-02 7.07728386e-01 2.70855397e-01
2.17978045e-01 1.02821505e+00 -3.34798008e-01 -2.14196458e-01
2.58546233e-01 2.72078097e-01 7.40747511e-01 -8.05732727e-01
-3.18653435e-01 -7.43385434e-01 3.58832628e-02 -1.54257321e+00
3.19708228e-01 -1.30245578e+00 7.66800195e-02 -2.18347478e+00
4.52084810e-01 2.75072247e-01 -1.20734200e-01 4.98191983e-01
-3.98232788e-01 -1.96841508e-01 3.80029082e-01 1.03960760e-01
-5.75705469e-01 1.06704676e+00 1.09419513e+00 -3.89617026e-01
-1.06718868e-01 -3.69307622e-02 -1.26670420e+00 3.17268968e-01
9.17323351e-01 -3.96878839e-01 -3.89666080e-01 -1.03353739e+00
3.34427655e-01 4.22336487e-03 1.67436138e-01 -8.53056550e-01
3.46437126e-01 3.28712106e-01 -3.25161591e-02 -6.01171374e-01
2.91869849e-01 -3.53789777e-01 -3.19378644e-01 4.06159729e-01
-8.15264404e-01 1.80510312e-01 2.54590392e-01 4.29635137e-01
-4.31110322e-01 -2.09729552e-01 5.78581333e-01 -1.02438301e-01
-1.38598323e-01 6.73479885e-02 -2.79241830e-01 2.27039084e-01
4.19328600e-01 2.71714538e-01 -4.96840000e-01 -6.32902503e-01
-2.74720222e-01 4.03444231e-01 3.00551087e-01 4.31035519e-01
6.23156309e-01 -1.10716379e+00 -1.20917261e+00 5.89289032e-02
-9.39517915e-02 6.40247464e-02 3.86641696e-02 8.40279520e-01
-3.43935072e-01 6.90096498e-01 -3.52083109e-02 -3.60755563e-01
-1.01530039e+00 2.26893798e-01 1.72352746e-01 -9.24278855e-01
-7.13210762e-01 7.89238632e-01 2.30312452e-01 -3.37082744e-01
9.18453634e-02 -6.18762374e-01 -1.77377537e-01 3.25979590e-02
6.76044941e-01 3.73916686e-01 3.15639049e-01 -5.08636177e-01
-3.12789261e-01 1.43461406e-01 -5.87472796e-01 -8.48011002e-02
1.74943483e+00 -1.43488571e-01 -3.85652989e-01 5.41113317e-01
1.33227575e+00 9.72621888e-03 -1.09544396e+00 -2.77237862e-01
-7.15847984e-02 4.38585831e-03 9.84041542e-02 -6.64797902e-01
-9.16005731e-01 1.05730236e+00 -3.38733673e-01 2.23807797e-01
1.30483234e+00 -3.30088437e-02 1.17131615e+00 4.96196359e-01
-1.04027554e-01 -9.87028837e-01 8.29009488e-02 7.99625993e-01
1.36933160e+00 -8.23990285e-01 1.34666517e-01 -2.58146554e-01
-7.29244173e-01 1.21488023e+00 1.86835423e-01 -2.73487300e-01
1.16089597e-01 2.88582981e-01 -4.50682998e-01 -9.30249467e-02
-1.34161854e+00 2.76869349e-02 3.20181251e-01 -2.87309680e-02
7.65969634e-01 -1.25294160e-02 -4.75344479e-01 7.79956639e-01
-5.28223217e-01 3.11985593e-02 8.51310730e-01 9.81507599e-01
-4.75137770e-01 -1.16356802e+00 1.09366514e-01 5.67639410e-01
-5.96353829e-01 -4.02311593e-01 -4.64445144e-01 3.10946316e-01
-5.67664087e-01 8.83796453e-01 8.88550058e-02 -2.15814009e-01
4.79741037e-01 8.21511820e-02 5.12652636e-01 -8.50808918e-01
-8.47616673e-01 -7.39598051e-02 3.88430655e-01 -4.97719318e-01
-2.19103158e-01 -4.91026253e-01 -1.11695373e+00 -2.50081897e-01
-1.77741319e-01 2.77215034e-01 4.33782458e-01 8.68879378e-01
7.81471014e-01 9.06995416e-01 6.02113247e-01 -8.29846680e-01
-8.86988282e-01 -1.22952783e+00 -2.42664084e-01 1.09198734e-01
6.59160554e-01 -1.12558924e-01 -1.14561178e-01 1.82166412e-01]
|
[12.150152206420898, 9.204319953918457]
|
37426fea-f4aa-4f1f-9826-2d7408168c59
|
lcdctcnn-lung-cancer-diagnosis-of-ct-scan
|
2304.04814
| null |
https://arxiv.org/abs/2304.04814v1
|
https://arxiv.org/pdf/2304.04814v1.pdf
|
LCDctCNN: Lung Cancer Diagnosis of CT scan Images Using CNN Based Model
|
The most deadly and life-threatening disease in the world is lung cancer. Though early diagnosis and accurate treatment are necessary for lowering the lung cancer mortality rate. A computerized tomography (CT) scan-based image is one of the most effective imaging techniques for lung cancer detection using deep learning models. In this article, we proposed a deep learning model-based Convolutional Neural Network (CNN) framework for the early detection of lung cancer using CT scan images. We also have analyzed other models for instance Inception V3, Xception, and ResNet-50 models to compare with our proposed model. We compared our models with each other considering the metrics of accuracy, Area Under Curve (AUC), recall, and loss. After evaluating the model's performance, we observed that CNN outperformed other models and has been shown to be promising compared to traditional methods. It achieved an accuracy of 92%, AUC of 98.21%, recall of 91.72%, and loss of 0.328.
|
['Ahmed Abdelgawad', 'Mahabuba Meherin', 'Md Ishtyaq Mahmud', 'Muntasir Mamun']
|
2023-04-10
| null | null | null | null |
['lung-cancer-diagnosis']
|
['medical']
|
[-2.64071077e-01 -1.51192203e-01 -4.81576622e-01 9.80895981e-02
-7.00510621e-01 2.99495133e-03 4.77361739e-01 2.80184537e-01
-8.39806974e-01 6.20428920e-01 9.66757536e-02 -6.69668078e-01
-1.29062325e-01 -1.06473756e+00 -2.05567300e-01 -6.44187689e-01
-1.96390852e-01 5.89759767e-01 6.19637966e-01 2.35585064e-01
-1.84305519e-01 8.45190585e-01 -9.78705287e-01 3.89776766e-01
5.18685520e-01 1.19348037e+00 7.31254593e-02 7.91150272e-01
-2.89328825e-02 1.18467641e+00 -1.46899179e-01 -1.96869716e-01
2.19825044e-01 -4.15220000e-02 -9.35190618e-01 -4.39082295e-01
2.07466364e-01 -6.70493960e-01 -6.94299042e-01 5.40824831e-01
6.81786835e-01 -3.56720597e-01 8.11819434e-01 -8.03336978e-01
-1.58435211e-01 1.87148944e-01 -4.63525623e-01 6.04066074e-01
-2.15759024e-01 1.20949380e-01 5.07306039e-01 -1.07841921e+00
2.72225529e-01 7.48208046e-01 1.21719825e+00 8.23065162e-01
-4.92652714e-01 -8.56057346e-01 -7.49442339e-01 1.71619654e-01
-1.26842356e+00 1.45331159e-01 -5.89274429e-02 -3.76121044e-01
9.18994665e-01 3.60611737e-01 7.37195969e-01 8.71038973e-01
8.53001297e-01 5.52243650e-01 6.79722786e-01 -3.62132341e-01
-3.57096121e-02 1.95813701e-01 -4.23026979e-02 1.18674076e+00
6.42702639e-01 4.69188720e-01 2.41818011e-01 -3.12269211e-01
8.45566690e-01 6.10967636e-01 -1.31299958e-01 2.13627350e-02
-1.08628321e+00 8.57546151e-01 1.04891586e+00 7.54608810e-01
-4.72976387e-01 4.85583872e-01 5.41026890e-01 -2.53796011e-01
3.01846802e-01 9.46066156e-02 -1.55984819e-01 3.07299584e-01
-1.01019287e+00 -3.61095071e-02 6.35400116e-01 3.56018275e-01
2.35430952e-02 -1.17802415e-02 -7.30814934e-01 7.31765985e-01
1.77616641e-01 3.13821644e-01 9.26131308e-01 -4.28883106e-01
2.48610973e-01 6.25466108e-01 -1.02865793e-01 -4.56079125e-01
-7.42164791e-01 -7.27876544e-01 -1.55191529e+00 -3.10055893e-02
1.56825185e-01 -3.34651098e-02 -1.21638453e+00 1.17908096e+00
-1.64166018e-01 1.66145101e-01 -2.07348019e-01 5.05990803e-01
1.17291558e+00 4.28511977e-01 5.42948484e-01 -6.63704053e-02
1.26279986e+00 -1.03300011e+00 -3.12702656e-01 6.55125901e-02
8.05208087e-01 -3.82555217e-01 6.66558743e-01 -2.16608830e-02
-9.03158665e-01 -4.08763349e-01 -7.88128614e-01 9.12914947e-02
-1.21796623e-01 3.43225926e-01 4.81286407e-01 7.25652337e-01
-1.19929445e+00 6.70600772e-01 -1.00925672e+00 -7.21976221e-01
7.69960821e-01 6.19830787e-01 -3.31599206e-01 -8.40921700e-02
-1.03302741e+00 9.97400403e-01 4.16411102e-01 -3.62195790e-01
-1.31053674e+00 -7.49142408e-01 -4.34391387e-02 3.21906030e-01
1.58423379e-01 -1.15368497e+00 1.52645028e+00 -3.01494122e-01
-8.86660993e-01 8.80668461e-01 1.33056283e-01 -8.53520513e-01
7.25379825e-01 -1.95953369e-01 -6.96435943e-02 3.14756542e-01
-9.70482230e-02 7.63402581e-01 1.94702446e-01 -8.45074475e-01
-9.73655283e-01 -1.81152821e-01 -3.03088427e-01 -2.09877938e-02
-4.90788192e-01 -4.89922836e-02 -4.06724215e-01 -4.02706236e-01
6.37907162e-03 -8.64715695e-01 -5.94910026e-01 3.60176384e-01
-2.32185528e-01 -3.92187029e-01 8.07200968e-01 -6.74098194e-01
1.20680547e+00 -1.71789086e+00 -7.75908172e-01 5.14655039e-02
7.13633835e-01 7.07079232e-01 2.81722665e-01 1.94905907e-01
-1.80184886e-01 6.82281196e-01 -1.27310947e-01 8.33196789e-02
-7.53284872e-01 2.03277078e-02 3.46940875e-01 3.62555206e-01
-4.81862947e-02 1.33827913e+00 -6.36633694e-01 -7.02683747e-01
4.24937248e-01 7.48092532e-01 -2.41234377e-01 4.48513508e-01
1.84734404e-01 3.35918218e-01 -5.34264684e-01 7.98086643e-01
3.94157439e-01 -4.58522528e-01 -1.51765317e-01 3.47341709e-02
1.07261248e-01 -6.68664789e-03 -2.86408871e-01 9.21876848e-01
-5.99672914e-01 5.72470725e-01 -4.89052117e-01 -5.33860683e-01
5.40524662e-01 8.54418218e-01 9.27180767e-01 -6.62797153e-01
3.74072999e-01 3.25156301e-01 7.33211637e-02 -6.74863517e-01
-7.64618590e-02 -3.46742272e-01 4.64673847e-01 2.07217649e-01
-1.67080164e-01 9.44861844e-02 -3.14700454e-01 5.98037578e-02
1.55552948e+00 -8.15457404e-01 8.47408414e-01 -2.39981785e-01
5.19959629e-01 5.34314327e-02 6.64992556e-02 8.56345713e-01
-3.76128793e-01 7.43419647e-01 1.72704786e-01 -8.08568716e-01
-1.00555682e+00 -1.01607072e+00 -2.27885991e-01 3.53287995e-01
-4.80945110e-01 -6.66377768e-02 -6.43560052e-01 -8.49236608e-01
-9.38467830e-02 4.49989974e-01 -7.17088878e-01 -2.21290037e-01
-7.02778041e-01 -9.17773962e-01 8.91011834e-01 8.99197996e-01
1.07197368e+00 -9.25883949e-01 -8.09940815e-01 1.82498649e-01
-1.77383497e-01 -8.03786039e-01 -4.07384560e-02 2.78557897e-01
-1.40773177e+00 -1.29122257e+00 -1.10437822e+00 -7.65524745e-01
5.47564864e-01 3.57655525e-01 1.33456135e+00 7.05929816e-01
-6.55459642e-01 3.29225898e-01 -1.24261901e-02 -6.18752062e-01
-4.06708509e-01 2.50945687e-01 -2.14992881e-01 -6.10421956e-01
2.79610962e-01 -1.54074952e-01 -1.02069330e+00 1.74029567e-03
-9.76632833e-01 -1.63718194e-01 1.03942764e+00 7.88353503e-01
6.24462128e-01 2.47880667e-01 3.51032227e-01 -1.00132382e+00
6.29683614e-01 -7.01268792e-01 -2.02361435e-01 1.31184578e-01
-9.09621537e-01 -4.15174901e-01 5.87065518e-01 -7.65926242e-02
-6.69427633e-01 1.85922757e-01 -3.62368554e-01 -3.72604638e-01
-1.50992423e-01 3.96166086e-01 5.32848477e-01 -3.00001740e-01
8.21062684e-01 1.34955361e-01 -1.77685887e-01 -3.18322927e-01
-4.45272744e-01 6.16682231e-01 4.34173316e-01 1.26591802e-01
7.98691034e-01 5.23361087e-01 3.85212898e-01 -7.04615176e-01
-8.00867379e-01 -7.90720046e-01 -6.13793194e-01 -2.69475043e-01
9.81611907e-01 -1.05687165e+00 -5.90955138e-01 4.73188311e-01
-9.50227678e-01 -9.45758447e-02 -1.98705390e-01 7.01292157e-01
-8.94087255e-02 2.25218251e-01 -6.53739989e-01 -5.64131737e-01
-9.10174906e-01 -7.58162200e-01 6.34297013e-01 1.59196988e-01
-6.92275763e-02 -1.13245058e+00 1.28271088e-01 1.19004168e-01
9.84903038e-01 3.37201416e-01 1.00553870e+00 -9.39541876e-01
-4.57267046e-01 -8.72205019e-01 -6.47036672e-01 3.27786446e-01
1.69891581e-01 -2.27240976e-02 -9.08694327e-01 -4.16324139e-01
2.49345735e-01 -1.47011682e-01 1.21893084e+00 6.34420395e-01
1.81914270e+00 -2.56925374e-01 -8.60080719e-01 5.78332901e-01
1.75134277e+00 4.74605650e-01 8.20748389e-01 3.86055171e-01
5.67879140e-01 -5.14694974e-02 4.27579656e-02 1.53477699e-01
2.76927073e-02 5.62372617e-02 7.40045547e-01 -3.77848953e-01
-3.64775747e-01 -5.97564504e-02 -2.33244300e-01 5.17083645e-01
-2.97999114e-01 -6.81723475e-01 -1.52560377e+00 5.42293429e-01
-1.32991588e+00 -8.95199060e-01 -4.31782186e-01 2.12710142e+00
2.90386170e-01 2.85215318e-01 2.73218798e-03 1.32696629e-01
5.72649479e-01 -2.45144844e-01 -2.81793654e-01 -8.32911953e-03
4.83573854e-01 5.95286369e-01 7.93304145e-01 -4.19335486e-03
-1.31401098e+00 4.53617573e-01 6.89304638e+00 8.57290506e-01
-1.32949841e+00 3.65320086e-01 1.01787138e+00 2.45736048e-01
1.78432792e-01 -4.54685181e-01 -5.20948231e-01 1.91564053e-01
1.12484014e+00 -2.52936393e-01 -1.99948475e-01 8.46708119e-01
1.76441848e-01 -3.17556977e-01 -9.48572814e-01 6.15322530e-01
-2.02350870e-01 -1.34642410e+00 1.95956260e-01 1.44586012e-01
5.79506218e-01 5.53021967e-01 3.82197388e-02 4.14346308e-01
8.19877610e-02 -1.59979963e+00 -1.23153076e-01 4.04197484e-01
1.12384534e+00 -7.56028414e-01 1.43468916e+00 6.10976696e-01
-1.25121963e+00 -1.67326018e-01 -3.23944539e-01 2.57186472e-01
-2.94343829e-01 6.90722048e-01 -1.50319231e+00 3.90652567e-01
8.88486087e-01 3.84865522e-01 -7.31851280e-01 1.49253261e+00
4.21714596e-02 1.02254331e+00 -4.10566688e-01 -3.46063823e-01
3.97956103e-01 5.66958606e-01 7.26012290e-02 1.26228607e+00
5.78542292e-01 2.50734597e-01 1.05063446e-01 5.06330490e-01
-3.95401895e-01 1.09867878e-01 -7.22830176e-01 2.94972032e-01
2.78501540e-01 1.15469670e+00 -8.15268219e-01 -4.61382955e-01
-4.16839957e-01 5.18462002e-01 -3.86685878e-02 -2.13301569e-01
-1.09337711e+00 -4.14687470e-02 -2.56460130e-01 5.39025366e-01
-1.22223794e-01 2.78615654e-01 -3.47800940e-01 -7.16097355e-01
-4.25182372e-01 -4.66852397e-01 6.00886464e-01 -6.10059023e-01
-1.05199754e+00 7.66921997e-01 -3.33251320e-02 -1.36773705e+00
1.17433779e-02 -6.71250224e-01 -9.86030221e-01 6.05356932e-01
-1.50864470e+00 -9.53235209e-01 -7.18788028e-01 4.77771223e-01
5.63414037e-01 -2.82582909e-01 8.78847897e-01 2.83934981e-01
-5.48311591e-01 4.37987030e-01 1.63090050e-01 3.97512287e-01
3.35780144e-01 -1.15861118e+00 2.36211151e-01 4.87693965e-01
-3.91368330e-01 4.49676849e-02 2.23415848e-02 -6.37777090e-01
-7.82156289e-01 -1.64783812e+00 9.11445141e-01 -2.87982196e-01
2.16830730e-01 3.77979785e-01 -8.29990745e-01 5.30526221e-01
4.39998619e-02 2.04102084e-01 5.25156260e-01 -6.45842969e-01
2.96555855e-03 -8.60590413e-02 -1.37046027e+00 3.46620142e-01
5.62356472e-01 -1.09947830e-01 -1.12122923e-01 4.40514565e-01
4.36313361e-01 -2.70938098e-01 -8.60290289e-01 9.33731914e-01
6.71159625e-01 -1.24905992e+00 1.30550718e+00 -3.15186650e-01
5.52461088e-01 2.24268988e-01 1.35731578e-01 -7.63209641e-01
-9.34374571e-01 3.96324575e-01 -6.25389488e-03 4.36781794e-01
4.95548666e-01 -5.31626225e-01 1.19593740e+00 4.92074579e-01
4.77665029e-02 -1.32080245e+00 -9.63600397e-01 -7.69526660e-01
4.83725935e-01 -2.75694013e-01 2.09585235e-01 5.25513113e-01
-6.34432912e-01 -6.99981004e-02 -1.16200276e-01 -1.24634430e-01
3.94608319e-01 -6.16784215e-01 3.69082421e-01 -1.40595937e+00
1.60696596e-01 -4.86008167e-01 -2.87427366e-01 -2.65060067e-01
-3.96101147e-01 -9.71172512e-01 -2.59399444e-01 -1.97813284e+00
1.03899026e+00 -1.56482115e-01 -7.96541333e-01 5.79555213e-01
-1.63560852e-01 2.15073556e-01 7.21822828e-02 5.43837428e-01
-1.92599639e-01 -1.39420209e-02 1.31682992e+00 -2.21686199e-01
2.93387622e-01 4.49176103e-01 -3.79145563e-01 7.96714842e-01
1.31610167e+00 -8.82531464e-01 -4.65083830e-02 -3.30728620e-01
-2.07479864e-01 2.68029124e-01 5.74621916e-01 -1.65006518e+00
3.69302005e-01 -1.08467989e-01 8.63076866e-01 -8.75964046e-01
1.07434355e-02 -1.07405400e+00 2.06734896e-01 1.65231156e+00
-3.14967930e-01 -9.24358070e-02 2.25878268e-01 5.17025769e-01
-1.34343803e-01 -3.31334919e-01 1.16731465e+00 -6.34615421e-01
-4.33504134e-01 7.83390343e-01 -5.97834051e-01 -2.05361232e-01
1.24278069e+00 -2.21448272e-01 -1.37376249e-01 -3.51499647e-01
-2.86964774e-01 3.10004503e-03 -2.23875744e-03 5.47257476e-02
8.34099352e-01 -1.30873668e+00 -8.08823168e-01 -8.25276598e-04
4.16689329e-02 5.56329228e-02 2.79169269e-02 9.30163443e-01
-1.13362753e+00 9.29706335e-01 -1.64721981e-01 -7.27394521e-01
-1.39558852e+00 3.80470484e-01 9.20904040e-01 -1.04631209e+00
-5.60439944e-01 1.00461996e+00 1.92472175e-01 -2.35290393e-01
4.99248922e-01 -4.91262227e-01 -3.69177252e-01 -4.80449319e-01
2.57190734e-01 6.82959735e-01 1.98081419e-01 -1.01487450e-01
-5.55479467e-01 4.70545143e-01 -1.31217211e-01 2.86433518e-01
1.24162138e+00 4.28817034e-01 -1.14515074e-01 7.60984123e-02
1.19707322e+00 -4.18428302e-01 -3.77980441e-01 -1.77485347e-02
1.29327938e-01 -2.23112673e-01 3.80056322e-01 -1.19851518e+00
-1.25482547e+00 1.10693610e+00 1.17970371e+00 9.85507667e-02
1.19815898e+00 -1.91606998e-01 1.11099601e+00 5.43424904e-01
1.30955040e-01 -4.96974409e-01 4.11494106e-01 5.07299185e-01
6.07159436e-01 -1.52023602e+00 1.40276164e-01 -2.40354508e-01
-1.74418747e-01 1.39410651e+00 8.49765956e-01 -1.89112619e-01
1.01054764e+00 1.82471454e-01 -3.86782400e-02 -2.77307361e-01
-9.62148726e-01 -6.96787313e-02 3.14240068e-01 4.20165211e-01
8.09399545e-01 1.28440768e-01 -3.71880740e-01 4.39071655e-01
2.05955505e-01 4.92980450e-01 2.17126459e-01 8.51488948e-01
-8.89251709e-01 -6.66620612e-01 -3.12657267e-01 9.72840846e-01
-1.19697225e+00 -9.57089514e-02 -5.75008690e-01 1.27103150e+00
1.00033469e-01 6.81524634e-01 -4.26901244e-02 -4.32612211e-01
1.38020590e-01 -3.34646031e-02 4.47317027e-02 -6.96319997e-01
-8.86080980e-01 -2.36005872e-01 -2.07880601e-01 -2.34220922e-01
-3.18665177e-01 9.48041677e-03 -1.30189526e+00 -4.60538805e-01
-3.83954376e-01 -2.29287464e-02 6.24724686e-01 6.90865338e-01
-3.23187523e-02 8.77325833e-01 6.74643636e-01 -3.71584773e-01
-4.30152714e-01 -1.02015948e+00 -2.37655967e-01 9.70776938e-03
3.25480640e-01 -2.45215595e-01 -2.92056352e-01 -3.27606708e-01]
|
[15.33697509765625, -2.292292833328247]
|
83e47b17-efeb-4473-83a3-e3a1e2834bdf
|
autolearn-automated-feature-generation-and
| null | null |
https://ieeexplore.ieee.org/abstract/document/8215494
|
http://web2py.iiit.ac.in/research_centres/publications/download/inproceedings.pdf.88535e0ea3a74e72.4943444d2d20323031372e706466.pdf
|
AutoLearn - Automated Feature Generation and Selection
|
In recent years, the importance of feature engineering has been confirmed by the exceptional performance of deep learning techniques, that automate this task for some applications. For others, feature engineering requires substantial manual effort in designing and selecting features and is often tedious and non-scalable. We present AutoLearn, a regression-based feature learning algorithm. Being data-driven, it requires no domain knowledge and is hence generic. Such a representation is learnt by mining pairwise feature associations, identifying the linear or non-linear relationship between each pair, applying regression and selecting those relationships that are stable and improve the prediction performance. Our experimental evaluation on 18 UC Irvine and 7 Gene expression datasets, across different domains, provides evidence that the features learnt through our model can improve the overall prediction accuracy by 13.28%, compared to original feature space and 5.87% over other top performing models, across 8 different classifiers without using any domain knowledge.
|
['Ambika Kaul', 'Saket Maheshwary', 'Vikram Pudi']
|
2017-11-17
| null | null | null |
ieee-ieee-international-conference-on-data
|
['automated-feature-engineering']
|
['methodology']
|
[ 1.10861167e-01 -1.94014326e-01 -3.00130427e-01 -7.10945606e-01
-5.86887836e-01 -5.89401484e-01 3.83025289e-01 3.13883215e-01
-3.16263437e-01 9.65960443e-01 4.50981926e-04 -9.84312594e-02
-7.08272994e-01 -7.71968961e-01 -5.65532804e-01 -7.71054566e-01
-5.76504111e-01 5.93860745e-01 -8.32629576e-02 -2.26259410e-01
3.42480779e-01 6.02895319e-01 -1.77620804e+00 3.47856671e-01
7.00044632e-01 1.22867775e+00 -2.43967593e-01 4.76669073e-01
-1.09410979e-01 3.52523893e-01 -4.01013732e-01 -5.48019409e-02
3.06111693e-01 -3.56082231e-01 -8.39456737e-01 -2.03018129e-01
3.02419245e-01 1.08267620e-01 -2.60200910e-02 5.67272723e-01
3.78082514e-01 2.72737350e-02 6.68992162e-01 -1.38686872e+00
-5.77607572e-01 5.84422052e-01 -5.97369850e-01 1.48520663e-01
3.38756293e-01 -1.73181757e-01 1.41668975e+00 -1.13955367e+00
7.19475091e-01 8.72148216e-01 1.05270207e+00 3.74920607e-01
-1.47376859e+00 -8.89379978e-01 -1.99386597e-01 2.70157218e-01
-1.59288871e+00 -2.90741324e-01 5.46239018e-01 -5.49228966e-01
1.30120540e+00 3.33409280e-01 7.58090377e-01 7.87050962e-01
3.79449904e-01 3.81173640e-01 9.99738276e-01 -5.47841668e-01
1.74986973e-01 1.68419912e-01 3.06707829e-01 7.26147234e-01
3.32885265e-01 1.58649147e-01 -6.72546327e-01 -4.58633333e-01
3.25938821e-01 1.85328186e-01 4.60559689e-02 -5.45337617e-01
-9.92712319e-01 1.03287590e+00 3.10197085e-01 5.75413644e-01
-3.72319341e-01 -2.00245559e-01 3.49010825e-01 7.49379933e-01
4.09966797e-01 8.58401835e-01 -1.09634805e+00 -1.05032526e-01
-8.89328659e-01 1.39769167e-01 8.36697698e-01 7.75644064e-01
1.03942931e+00 -1.62679940e-01 1.20517261e-01 8.76307845e-01
-3.56735050e-04 1.60285473e-01 8.39387894e-01 -6.29953623e-01
-1.73124507e-01 1.10575736e+00 -8.54650661e-02 -1.00509918e+00
-9.19914007e-01 -4.15198714e-01 -8.13887358e-01 7.35188648e-02
3.32033336e-01 -2.44500086e-01 -7.53300905e-01 1.61421764e+00
2.81002223e-01 -5.05349375e-02 -2.96978951e-02 5.04337549e-01
8.51305187e-01 3.00424188e-01 4.04781103e-02 -2.46626198e-01
1.12246621e+00 -4.36367750e-01 -4.61540937e-01 5.53943552e-02
8.52175355e-01 -6.21588111e-01 7.70774662e-01 6.01440370e-01
-5.07656872e-01 -4.47738975e-01 -1.11872971e+00 1.10408641e-01
-6.64446592e-01 7.94760063e-02 1.17503405e+00 5.02148271e-01
-8.84227455e-01 8.69975746e-01 -5.66876888e-01 -5.18095613e-01
4.67525929e-01 8.93744946e-01 -9.34622109e-01 -1.23088196e-01
-1.23658466e+00 8.39582801e-01 4.21183378e-01 -1.26576722e-01
-4.21756476e-01 -8.72045755e-01 -7.75638461e-01 2.16715649e-01
1.80309832e-01 -7.46565402e-01 7.92382419e-01 -1.19029081e+00
-1.10257518e+00 6.57643855e-01 -2.01186448e-01 -4.89545316e-01
9.80448052e-02 -4.02519554e-01 -6.75182402e-01 -1.86523080e-01
-5.86113520e-02 4.66404915e-01 6.78046048e-01 -7.64387012e-01
-7.22509146e-01 -3.91166687e-01 -4.43389088e-01 -5.45453541e-02
-5.90461075e-01 4.08054292e-02 -4.64535467e-02 -3.43027771e-01
3.68465073e-02 -1.01196432e+00 -3.44211012e-01 -2.20682532e-01
-2.28373960e-01 -4.27362621e-01 8.61713767e-01 -2.92774945e-01
1.11797595e+00 -2.15400910e+00 -1.50872674e-02 6.52085721e-01
2.19457328e-01 1.69762418e-01 -1.87068745e-01 5.86588264e-01
-4.57891762e-01 1.09693609e-01 -1.79562405e-01 3.77707541e-01
-2.40572900e-01 1.47031382e-01 -4.70988825e-02 4.14006680e-01
5.31539738e-01 6.57160521e-01 -7.03203440e-01 -3.09591740e-01
1.21765144e-01 5.15881181e-01 -6.18594289e-01 1.60492435e-01
-5.68133779e-03 1.70196980e-01 -5.10589957e-01 7.01024771e-01
5.89254618e-01 -4.15877640e-01 5.04458070e-01 -1.28152803e-01
-1.20121231e-02 6.18794449e-02 -1.04362619e+00 1.50682747e+00
-4.28769648e-01 7.14311719e-01 -5.14918625e-01 -1.08699656e+00
1.31333232e+00 1.06061414e-01 9.16527987e-01 -5.50928771e-01
1.04778349e-01 2.50193238e-01 2.96275645e-01 -5.70228100e-01
1.26662508e-01 -9.05240793e-03 -2.09049419e-01 4.25649107e-01
4.09495234e-01 2.07061604e-01 1.55744031e-01 -8.16662014e-02
1.41969883e+00 -1.03389621e-02 8.60946476e-01 -3.33325237e-01
4.61483032e-01 1.30553409e-01 8.19820046e-01 5.95714211e-01
1.39737919e-01 4.38994557e-01 4.12862450e-01 -9.67213333e-01
-7.99468458e-01 -5.85879505e-01 -4.22453225e-01 1.34144092e+00
-3.09123844e-01 -5.37920117e-01 -3.53525579e-01 -7.86735833e-01
4.58831578e-01 5.32803774e-01 -9.59664583e-01 -5.48066378e-01
-3.79731417e-01 -9.37571287e-01 3.55150402e-01 5.52670777e-01
1.71798110e-01 -8.92695189e-01 -2.74446338e-01 3.40166807e-01
2.76793242e-01 -7.26002038e-01 6.20989464e-02 7.23788798e-01
-9.63161826e-01 -1.36195838e+00 -6.01831339e-02 -6.44068420e-01
6.39829397e-01 -5.93716055e-02 1.27712417e+00 3.78717750e-01
-6.63360000e-01 -1.00406133e-01 -5.59170127e-01 -4.59915966e-01
1.31822638e-02 1.86160520e-01 2.11510628e-01 -4.04274650e-02
9.83894765e-01 -6.96622014e-01 -3.38174671e-01 3.11827838e-01
-6.03633046e-01 -2.94604182e-01 9.43927169e-01 1.38785768e+00
5.94355226e-01 1.45393729e-01 9.68250811e-01 -1.26310599e+00
4.23243284e-01 -6.68140352e-01 -2.55559295e-01 2.22456798e-01
-9.97485578e-01 1.53673515e-01 6.65079832e-01 -1.68076292e-01
-6.79515183e-01 6.74059093e-01 -9.71484836e-03 -7.97572881e-02
-3.29224467e-01 6.14899874e-01 -3.25237773e-02 -1.85595661e-01
1.00933969e+00 3.15972954e-01 8.49058926e-02 -3.88103932e-01
1.97771698e-01 6.31239951e-01 2.35217474e-02 -3.21100265e-01
7.45285511e-01 2.25925684e-01 2.93901682e-01 -6.57663286e-01
-8.32671702e-01 -5.02432883e-01 -1.12737894e+00 1.81794956e-01
4.02829379e-01 -6.49519444e-01 -6.97336137e-01 8.01655799e-02
-6.37655139e-01 1.46815732e-01 -1.28785044e-01 4.68394399e-01
-4.20333207e-01 -1.01448253e-01 -1.13459736e-01 -4.49266613e-01
-5.37528217e-01 -8.58090222e-01 7.07419276e-01 1.47262529e-01
-6.53097987e-01 -8.60873401e-01 1.09973013e-01 7.15156598e-03
4.18958873e-01 4.51697767e-01 1.20244730e+00 -1.24450862e+00
-1.14929276e-02 -5.44354737e-01 2.12997198e-03 8.95511210e-02
5.41653454e-01 2.66897857e-01 -1.01562440e+00 -2.23384574e-01
-6.36038184e-01 -3.11126441e-01 7.95765519e-01 2.46085465e-01
1.30960739e+00 -7.58998170e-02 -5.77632487e-01 5.74138582e-01
1.25297213e+00 2.62125850e-01 2.40615398e-01 4.48657066e-01
3.44776124e-01 5.40798485e-01 8.95521402e-01 6.36407614e-01
1.87805101e-01 6.20861173e-01 1.82101443e-01 -1.47680119e-01
2.07705975e-01 5.03979437e-02 -4.20179404e-03 4.20749873e-01
-1.63424805e-01 2.18779355e-01 -9.53109920e-01 4.33549255e-01
-1.96499455e+00 -7.96099126e-01 -1.09588623e-01 1.98837245e+00
9.97825563e-01 -9.85714123e-02 3.42439592e-01 3.28445494e-01
3.81700307e-01 -5.18141568e-01 -7.96054721e-01 -6.06657028e-01
-1.74104348e-01 4.72357452e-01 2.94023722e-01 2.29091290e-02
-1.17797720e+00 8.79949927e-01 7.30938959e+00 4.64088529e-01
-1.10101461e+00 -3.91551435e-01 5.95360100e-01 -2.40555212e-01
-6.95010126e-02 -1.60646766e-01 -6.90963566e-01 6.48145974e-02
9.99514937e-01 -3.46931130e-01 2.30519488e-01 1.12591422e+00
-9.72035341e-03 3.97818275e-02 -1.42227495e+00 7.92554021e-01
-1.23388551e-01 -1.36940455e+00 -8.20240006e-02 -3.63328941e-02
5.34988642e-01 -9.65654477e-02 2.61444636e-02 4.19289023e-01
5.71930289e-01 -1.33873391e+00 1.84778757e-02 5.64624190e-01
5.57167292e-01 -1.20584786e+00 1.02823889e+00 8.07226002e-02
-8.57698917e-01 -3.62435192e-01 -4.98132885e-01 -9.57174599e-02
-4.51368302e-01 9.65436578e-01 -1.19004309e+00 6.02445781e-01
9.96266007e-01 9.34812665e-01 -7.67710209e-01 1.06131315e+00
3.76919471e-02 6.26673102e-01 -1.29121423e-01 -1.55197501e-01
5.17089441e-02 9.90962163e-02 1.77372783e-01 1.38685656e+00
4.37910289e-01 -1.32031247e-01 7.14959204e-02 3.48695308e-01
1.48835117e-02 4.21813965e-01 -8.32121015e-01 -8.45879316e-02
5.49523950e-01 1.54587579e+00 -5.61396718e-01 -7.43935034e-02
-4.96195614e-01 6.30659699e-01 3.83393645e-01 -2.51430739e-02
-5.10690033e-01 -8.55644286e-01 8.81947100e-01 -1.57897905e-01
4.62708086e-01 5.31051718e-02 -3.61888200e-01 -7.49144733e-01
-2.90128529e-01 -1.03366673e+00 7.88468361e-01 -3.69154751e-01
-1.60355926e+00 7.13216543e-01 -3.64950806e-01 -1.32013619e+00
-5.39471805e-01 -7.91752219e-01 -3.21514785e-01 7.05374360e-01
-1.30426240e+00 -9.96294022e-01 -1.91856712e-01 5.33150792e-01
1.28096357e-01 -5.40826976e-01 1.33921480e+00 2.72893518e-01
-4.01720077e-01 7.58913875e-01 2.77613759e-01 6.61627054e-02
8.64224076e-01 -1.27410805e+00 8.07367712e-02 2.48700231e-01
3.07331294e-01 9.40154612e-01 6.31130219e-01 -3.04352254e-01
-1.67246354e+00 -1.03856862e+00 1.10453141e+00 -4.55584019e-01
5.87610126e-01 -3.98981363e-01 -9.25411820e-01 5.64730406e-01
-3.64524983e-02 2.86639750e-01 1.52748001e+00 6.81967497e-01
-6.59822941e-01 -3.79576266e-01 -1.36758685e+00 2.30392978e-01
9.31737661e-01 -1.21835805e-01 -4.79110599e-01 2.69577205e-01
4.29003596e-01 7.69875199e-02 -1.23858356e+00 3.93795222e-01
8.40394378e-01 -9.49269176e-01 9.06459212e-01 -1.13153207e+00
2.39444688e-01 -2.65783936e-01 -1.68709904e-01 -1.55800831e+00
-7.71670878e-01 -2.72297561e-01 -1.15247086e-01 1.11181629e+00
8.32509816e-01 -5.90758860e-01 7.73750842e-01 6.54250205e-01
3.38782780e-02 -8.84807229e-01 -7.95202971e-01 -8.22209895e-01
-7.89921060e-02 -1.36863425e-01 8.98480356e-01 1.37945819e+00
2.30773121e-01 4.17399585e-01 -3.92135054e-01 -7.94526637e-02
2.01037183e-01 5.47581315e-01 8.35108519e-01 -1.75801051e+00
-2.81471014e-01 -3.94632131e-01 -8.10641825e-01 -1.34561405e-01
2.63119400e-01 -1.05749822e+00 -2.11290538e-01 -1.11052942e+00
2.52367079e-01 -4.70617443e-01 -6.62842989e-01 9.81326222e-01
-5.77864088e-02 2.87889779e-01 -3.87635916e-01 -4.58666869e-02
-4.73992586e-01 2.61766195e-01 7.76819646e-01 -3.59994322e-02
-1.82738453e-01 -3.86139192e-02 -1.19791424e+00 6.54008150e-01
8.69430363e-01 -6.26911998e-01 -1.68014139e-01 5.58160394e-02
2.96889096e-01 -4.09395397e-01 -1.26427442e-01 -8.93119276e-01
1.14632875e-01 -2.88672924e-01 1.08109653e+00 -3.40740174e-01
1.51357606e-01 -1.05568945e+00 2.50943512e-01 5.53637147e-01
-3.56955796e-01 9.25845504e-02 2.28502005e-01 5.12959957e-01
-2.81807780e-01 -1.84960693e-01 5.92319667e-01 9.49805602e-02
-9.85133767e-01 2.77239382e-01 -1.97521999e-01 -3.26154202e-01
1.16971374e+00 -1.21225625e-01 6.31207079e-02 -1.83820650e-01
-9.20998096e-01 -2.15094090e-02 9.20612216e-02 4.75723118e-01
5.78606009e-01 -1.39184582e+00 -6.34786487e-01 5.07238150e-01
4.72624958e-01 -3.87552440e-01 -6.46317750e-02 6.95936501e-01
-5.61819263e-02 3.67452204e-01 -3.76535684e-01 -4.97197896e-01
-1.63944900e+00 5.26626408e-01 1.38819322e-01 -1.62453562e-01
-5.74196517e-01 7.60092795e-01 -2.75347352e-01 -5.06810725e-01
-1.34018078e-01 3.30641381e-02 -4.95228887e-01 2.30896905e-01
5.82299232e-01 3.19148123e-01 4.49931324e-01 -4.30024087e-01
-6.63742483e-01 6.96015358e-01 -3.61448109e-01 4.01195318e-01
1.91769052e+00 3.00423205e-01 -2.28584573e-01 3.63666981e-01
1.25112009e+00 -1.86734617e-01 -7.65692592e-01 -3.24738652e-01
5.39823651e-01 -5.39034009e-01 1.20819472e-02 -1.03432226e+00
-1.12673259e+00 6.25224352e-01 6.34419799e-01 2.40610600e-01
1.21338332e+00 -1.11149997e-01 3.01851183e-01 9.11985397e-01
3.51592183e-01 -1.09615552e+00 -1.08311512e-01 7.16922462e-01
7.82410443e-01 -1.33490109e+00 2.27648824e-01 -3.55032027e-01
-5.19869864e-01 1.47686923e+00 7.58029819e-01 -2.60713995e-01
9.36868370e-01 2.93812841e-01 -6.11380525e-02 -3.17891449e-01
-1.08397138e+00 -6.78908378e-02 4.45792675e-01 7.07417130e-01
9.29992020e-01 6.42688200e-02 -3.84499520e-01 7.30235755e-01
-3.15010458e-01 9.48609635e-02 1.13814302e-01 9.15007889e-01
-4.61384654e-01 -1.39213431e+00 3.76491658e-02 9.30309713e-01
-4.12053078e-01 -5.49339205e-02 -7.52039552e-01 9.61404204e-01
8.83992165e-02 8.20501566e-01 8.91788453e-02 -7.48438954e-01
2.94579268e-01 4.52964246e-01 2.95909703e-01 -5.36133230e-01
-8.38302612e-01 -1.74659669e-01 2.86417097e-01 -5.36036074e-01
-2.20424697e-01 -8.63587081e-01 -1.29658723e+00 -1.93832695e-01
-4.23440784e-01 2.84790128e-01 5.27049720e-01 9.20913041e-01
8.87155056e-01 4.29978549e-01 9.57887650e-01 -3.43447953e-01
-4.16044950e-01 -1.02858281e+00 -6.30646646e-01 4.62035656e-01
2.09396243e-01 -9.11023259e-01 -1.86294198e-01 -8.29072669e-02]
|
[8.090826034545898, 4.498520374298096]
|
5e00827f-ed9e-42ab-a742-befd0ea5abf5
|
demixing-sines-and-spikes-using-multiple
|
2004.00259
| null |
https://arxiv.org/abs/2004.00259v2
|
https://arxiv.org/pdf/2004.00259v2.pdf
|
Demixing Sines and Spikes Using Multiple Measurement Vectors
|
In this paper, we address the line spectral estimation problem with multiple measurement corrupted vectors. Such scenarios appear in many practical applications such as radar, optics, and seismic imaging in which the signal of interest can be modeled as the sum of a spectrally sparse and a blocksparse signal known as outlier. Our aim is to demix the two components and for that, we design a convex problem whose objective function promotes both of the structures. Using positive trigonometric polynomials (PTP) theory, we reformulate the dual problem as a semi-definite program (SDP). Our theoretical results states that for a fixed number of measurements N and constant number of outliers, up to O(N) spectral lines can be recovered using our SDP problem as long as a minimum frequency separation condition is satisfied. Our simulation results also show that increasing the number of samples per measurement vectors, reduces the minimum required frequency separation for successful recovery.
|
['Mohammad Hossein Kahaei', 'Sajad Daei', 'Hoomaan Maskan']
|
2020-04-01
| null | null | null | null |
['seismic-imaging']
|
['miscellaneous']
|
[ 6.68187737e-01 -4.61060852e-02 2.31200799e-01 1.21963814e-01
-9.48621154e-01 -4.75983858e-01 2.05453066e-03 -4.08389373e-03
-2.66862035e-01 7.10043490e-01 2.06686985e-02 -4.45335731e-02
-4.58245516e-01 -2.70392478e-01 -7.48834074e-01 -9.46667612e-01
-2.19462976e-01 2.25374788e-01 -1.61434084e-01 7.80884847e-02
3.90121311e-01 6.42983854e-01 -1.09025931e+00 -3.79612803e-01
1.15134895e+00 9.43300605e-01 -2.10727146e-03 6.49534404e-01
4.09247100e-01 3.85954350e-01 -4.93943483e-01 8.66472498e-02
5.98804832e-01 -2.59441197e-01 -3.64251882e-01 5.14700234e-01
3.35772783e-01 -1.96919531e-01 -2.75782347e-01 1.53352475e+00
5.69214702e-01 2.12247476e-01 5.65709591e-01 -1.18101382e+00
-2.13478908e-01 4.11831141e-01 -1.25369823e+00 1.81972943e-02
3.68266195e-01 -2.59297431e-01 5.80490768e-01 -1.21354377e+00
-1.14036910e-02 7.61653483e-01 8.55631053e-01 -1.94471434e-01
-1.59089851e+00 -3.71236086e-01 -3.46644104e-01 1.04649164e-01
-1.54281521e+00 -8.19651008e-01 9.99723792e-01 -6.15853548e-01
5.12682199e-01 4.39966381e-01 5.26598990e-01 4.92257267e-01
-9.24817026e-02 4.30466175e-01 7.13448644e-01 -5.79994619e-01
1.47580251e-01 -1.66122228e-01 1.71744570e-01 2.40195915e-01
7.96703994e-01 4.30846959e-03 -5.72347999e-01 -6.82982028e-01
5.86633205e-01 -5.81984706e-02 -9.07528341e-01 -3.54110360e-01
-1.25626242e+00 8.13768089e-01 -1.71343192e-01 2.08890900e-01
-6.27069592e-01 9.03309062e-02 6.34287074e-02 3.77347291e-01
4.11128402e-01 5.47096312e-01 2.88907122e-02 1.63892388e-01
-1.17052472e+00 6.73759803e-02 8.50505829e-01 7.67981648e-01
6.56175435e-01 6.48851693e-01 2.31554836e-01 7.31326163e-01
2.95965850e-01 1.12873733e+00 1.94502681e-01 -9.66817021e-01
5.92705786e-01 -1.06715553e-01 5.36792815e-01 -1.22980464e+00
-3.29365253e-01 -4.88720655e-01 -9.22821879e-01 -1.61430538e-01
4.94475573e-01 -3.38100672e-01 -3.74364793e-01 1.45098579e+00
1.06236711e-01 7.79853284e-01 1.00645155e-01 9.60475504e-01
1.63756609e-01 8.71550143e-01 -8.36526573e-01 -8.24564099e-01
8.53430748e-01 -4.90260988e-01 -8.61228108e-01 -3.98818821e-01
2.60273963e-01 -1.13184524e+00 4.85852122e-01 6.52890682e-01
-1.20512629e+00 -7.90704563e-02 -1.38885200e+00 4.99520749e-01
5.96750617e-01 1.60835534e-01 1.33051559e-01 6.04526460e-01
-6.85413241e-01 5.44485748e-01 -7.23491311e-01 1.90546662e-01
-3.61625195e-01 3.26231658e-01 -3.39972854e-01 -1.11969650e-01
-7.36770868e-01 6.29047692e-01 1.37140840e-01 5.92990458e-01
-4.32303935e-01 -6.75020397e-01 -8.26329172e-01 4.63230573e-02
3.08290750e-01 -3.07970643e-01 8.28968406e-01 -8.40976000e-01
-1.13999212e+00 5.58843374e-01 -2.99174786e-01 -4.01517898e-01
2.84722000e-01 -2.69435287e-01 -6.18053257e-01 4.14942443e-01
2.37096027e-01 -4.61382002e-01 1.37719977e+00 -1.29386616e+00
-3.23894382e-01 -3.74559492e-01 -5.79809964e-01 -5.46592623e-02
1.73028801e-02 -1.49217084e-01 -9.69827846e-02 -5.44132531e-01
7.18755722e-01 -1.01558888e+00 -4.00448203e-01 -3.39725256e-01
-5.29906631e-01 5.18328965e-01 6.27978981e-01 -8.35256398e-01
1.00170755e+00 -2.49688721e+00 3.83081079e-01 7.75338650e-01
1.88122556e-01 -6.62400275e-02 -7.18671083e-02 7.08756089e-01
-4.42210406e-01 -3.17922860e-01 -6.06664836e-01 -2.73952961e-01
-2.98489243e-01 -4.45555709e-02 -5.37455916e-01 1.34696865e+00
-1.43811211e-01 4.36528623e-02 -8.08907151e-01 1.04603484e-01
1.97965324e-01 1.52322114e-01 -5.30676603e-01 1.19340703e-01
4.02179867e-01 4.24733132e-01 -3.54962111e-01 3.84709269e-01
1.12096953e+00 -1.37594625e-01 1.08450375e-01 -3.87954682e-01
-2.99118310e-01 -3.66554260e-01 -1.86155105e+00 1.25036132e+00
-2.51959950e-01 7.56328225e-01 6.24169648e-01 -1.31046045e+00
7.82471955e-01 4.23691422e-01 9.95126188e-01 -2.87378579e-01
-1.29211163e-02 5.15538514e-01 -4.68696654e-02 -5.57297349e-01
4.65450644e-01 -1.81585625e-01 9.17918161e-02 2.58433282e-01
-2.79103190e-01 -1.45211697e-01 1.03638060e-01 9.94335711e-02
1.12628317e+00 -6.15264714e-01 5.20354807e-01 -2.81229317e-01
5.44803739e-01 -2.46221855e-01 7.53448188e-01 6.41806841e-01
2.76835766e-02 5.92061996e-01 4.22684342e-01 2.13936180e-01
-1.14869344e+00 -8.12369585e-01 -3.20769906e-01 1.64585471e-01
1.58864021e-01 6.99941069e-02 -3.20219845e-01 2.30156004e-01
7.54166096e-02 4.70532954e-01 -3.88545319e-02 -1.01851888e-01
-7.36813724e-01 -7.81211972e-01 2.87318796e-01 -9.52476040e-02
3.19784403e-01 -2.52053827e-01 -3.65095735e-01 2.80351311e-01
-2.89549261e-01 -1.31450057e+00 -6.83864415e-01 1.76382363e-01
-7.52933621e-01 -1.02400851e+00 -7.50943899e-01 -6.55646801e-01
8.03725779e-01 7.89850175e-01 6.05982840e-01 -2.68741161e-01
4.94804641e-04 7.10014820e-01 -3.34715903e-01 -2.07890868e-01
-2.03711152e-01 -5.88744402e-01 3.57123554e-01 5.65465510e-01
-1.79376021e-01 -8.46626639e-01 -3.32878798e-01 1.43608168e-01
-9.61919606e-01 -2.61977524e-01 4.64319319e-01 9.64266062e-01
4.76680666e-01 2.92455226e-01 5.18457949e-01 -5.86718738e-01
6.78083718e-01 -5.47751248e-01 -9.27216172e-01 1.54729888e-01
-3.54812831e-01 7.74581656e-02 7.20361829e-01 -4.46667910e-01
-5.63487351e-01 1.86373487e-01 2.46063858e-01 -5.54447174e-01
4.66004312e-01 9.28550184e-01 -1.33896217e-01 -6.40323639e-01
6.60380304e-01 4.75306779e-01 -1.03130072e-01 -2.54743427e-01
1.06070966e-01 6.36028111e-01 7.49302030e-01 -4.85494494e-01
1.20461619e+00 6.23068035e-01 4.10371661e-01 -1.57463038e+00
-6.93361759e-01 -1.03474486e+00 -3.56518477e-01 -2.77620137e-01
1.74310490e-01 -6.53564632e-01 -8.02948713e-01 3.21830630e-01
-1.15137768e+00 2.33043894e-01 -2.17744708e-01 8.38857234e-01
-5.89915454e-01 9.94906962e-01 -1.29527882e-01 -1.23583674e+00
-1.54995278e-01 -9.35959101e-01 9.49468136e-01 -1.05255708e-01
5.50353415e-02 -8.14715803e-01 2.83699214e-01 6.63159490e-02
7.37974569e-02 4.19897079e-01 4.86707360e-01 -3.28132004e-01
-5.66159010e-01 -4.41553950e-01 -7.56448954e-02 4.53107208e-01
5.40232882e-02 -2.04483151e-01 -6.31208837e-01 -6.63197100e-01
8.07416379e-01 1.10697597e-01 5.50792217e-01 5.58935523e-01
7.92791009e-01 -5.09312689e-01 -1.60589471e-01 9.61604834e-01
1.62240958e+00 3.14184070e-01 5.03569424e-01 6.56862706e-02
5.17928481e-01 3.74325424e-01 5.07701814e-01 8.52340937e-01
-4.16811943e-01 5.32029390e-01 3.40565503e-01 8.00558180e-02
4.81247485e-01 1.83066338e-01 3.02661031e-01 1.12265944e+00
1.47345275e-01 -2.90642470e-01 -7.96308398e-01 6.85521722e-01
-2.07663608e+00 -1.20759177e+00 -5.15715301e-01 2.74985909e+00
4.70609039e-01 -3.57427925e-01 -2.10253075e-01 6.14157140e-01
7.34758496e-01 1.91830605e-01 -6.44928098e-01 -9.86162107e-03
-2.52728045e-01 1.07561752e-01 1.02801788e+00 6.13320470e-01
-8.36254239e-01 3.13167386e-02 6.11144447e+00 6.99192643e-01
-1.18422437e+00 -1.68216392e-01 -2.26463437e-01 5.79683334e-02
-3.71232599e-01 2.12196127e-01 -2.64199972e-01 5.25518239e-01
9.16587770e-01 -5.05142033e-01 6.23122036e-01 2.55762368e-01
4.27212238e-01 -2.96773583e-01 -7.79253960e-01 1.47051942e+00
2.49593362e-01 -9.77356911e-01 -4.27191347e-01 1.87609911e-01
7.50431001e-01 -2.35910103e-01 1.14958303e-03 -4.30252552e-01
-1.14682570e-01 -7.26874113e-01 6.68395996e-01 5.50101340e-01
5.25812805e-01 -7.60508060e-01 4.79892015e-01 5.06538630e-01
-1.05626321e+00 5.31090144e-03 -3.73083711e-01 7.94433802e-02
6.50548935e-01 1.22009754e+00 -6.78387225e-01 8.29677820e-01
9.84338596e-02 6.70537949e-01 2.93631107e-01 1.58452392e+00
-6.91666305e-02 7.59753048e-01 -6.30234063e-01 4.97919083e-01
7.58504942e-02 -8.79419804e-01 1.08447003e+00 8.26511323e-01
7.89528608e-01 4.22213227e-01 4.64433730e-01 3.62455934e-01
2.62796074e-01 8.14545453e-02 -4.88380015e-01 -1.35124046e-02
7.82229841e-01 8.41361225e-01 -4.80640650e-01 8.62451866e-02
-5.46577334e-01 9.84819353e-01 -1.61826581e-01 7.00059533e-01
-6.06588900e-01 -4.53636140e-01 4.57135260e-01 1.16754815e-01
1.31869540e-01 -5.85758090e-01 -3.28287572e-01 -1.37906933e+00
2.62671143e-01 -8.16180706e-01 1.20111816e-02 -6.23870850e-01
-1.10580885e+00 1.28905252e-01 -1.74099401e-01 -1.55118668e+00
-1.78060949e-01 -2.34067947e-01 -4.72184300e-01 8.66486430e-01
-1.15171003e+00 -7.71473348e-01 -2.12873742e-02 5.52999973e-01
1.50509477e-01 -5.95737547e-02 6.37962699e-01 5.34392059e-01
-6.44831419e-01 1.30958825e-01 7.44318366e-01 -1.16599984e-01
6.03771627e-01 -1.07885623e+00 -1.57995179e-01 1.32091939e+00
3.89567278e-02 5.12874424e-01 1.42017090e+00 -6.67795002e-01
-1.54571486e+00 -6.89549088e-01 4.69054073e-01 3.76358777e-01
9.59227383e-01 1.36828169e-01 -7.82388270e-01 7.04526186e-01
1.33809564e-03 1.23856999e-01 6.66808248e-01 -2.06175834e-01
-5.28543852e-02 -8.59162211e-02 -1.12093663e+00 3.18246454e-01
5.66246152e-01 -4.39062029e-01 -3.86132807e-01 6.18807793e-01
2.72885531e-01 -5.27516961e-01 -8.31546724e-01 4.56541210e-01
2.34803900e-01 -7.98925042e-01 1.12270713e+00 -1.79035410e-01
-3.62000652e-02 -5.26886463e-01 -5.20406306e-01 -1.50089788e+00
-2.51373917e-01 -1.08678758e+00 -2.89895803e-01 7.01963186e-01
1.19748205e-01 -8.36407125e-01 6.21198297e-01 2.28284091e-01
-2.99337208e-01 -3.12847495e-01 -1.05571663e+00 -1.06600106e+00
-5.33110619e-01 -2.84377217e-01 5.25686052e-03 1.05419803e+00
1.06988914e-01 3.57522041e-01 -9.42314625e-01 8.84190500e-01
1.12986422e+00 2.18559414e-01 6.81003690e-01 -1.26381338e+00
-5.96909940e-01 4.67123464e-02 -3.91545922e-01 -1.27064359e+00
6.16432726e-02 -5.74820757e-01 1.94689006e-01 -1.09473014e+00
4.25572470e-02 -4.18055475e-01 2.94235528e-01 -1.65365070e-01
-6.55803457e-03 -1.22334570e-01 1.58276156e-01 3.49790931e-01
7.57599920e-02 5.07926345e-01 8.26114357e-01 -1.28031313e-01
-1.55227467e-01 4.03524309e-01 -4.28804040e-01 8.17789972e-01
3.33157778e-01 -4.79005218e-01 -4.78837252e-01 -4.77430344e-01
3.90716791e-01 8.90064657e-01 1.38367787e-01 -1.12742674e+00
4.87678885e-01 -2.26712257e-01 2.41498332e-02 -6.89310431e-01
6.59945190e-01 -1.07679653e+00 4.82158303e-01 3.56376350e-01
1.55153230e-01 -1.17504351e-01 -7.64470920e-02 6.86358750e-01
-4.15973604e-01 -7.35379875e-01 9.05209601e-01 4.02561545e-01
-3.14534694e-01 2.06590056e-01 -3.21730912e-01 -1.01336718e-01
1.05535662e+00 -3.96800250e-01 -2.76067585e-01 -6.96642399e-01
-6.64199769e-01 2.22345039e-01 1.69447765e-01 -4.29058880e-01
7.62743115e-01 -1.13904691e+00 -7.91423440e-01 3.05521280e-01
-1.78529263e-01 -5.56165762e-02 5.05584776e-01 1.22686791e+00
-5.78019142e-01 4.28707182e-01 2.74237663e-01 -6.26603901e-01
-1.08457327e+00 3.50379050e-01 3.37028474e-01 -2.47477815e-02
-4.67744589e-01 6.45892441e-01 -2.73655325e-01 -7.12581351e-02
5.74512221e-03 -1.13060929e-01 2.10546121e-01 -1.76902171e-02
5.12804389e-01 7.61063933e-01 1.31516933e-01 -7.15368569e-01
-8.24566931e-02 7.42870092e-01 3.13383877e-01 -3.47615540e-01
1.33593142e+00 -1.26764104e-01 -3.43398631e-01 2.67922640e-01
1.33274198e+00 8.43138039e-01 -1.04420304e+00 -3.43468785e-01
-4.93744314e-02 -7.12003231e-01 1.29368365e-01 1.40284523e-01
-8.43542814e-01 3.79008830e-01 3.15214783e-01 5.78450799e-01
1.29091930e+00 -4.64870512e-01 7.36665785e-01 4.52603668e-01
4.45055872e-01 -6.79980576e-01 -1.87463257e-02 4.94351476e-01
9.34729278e-01 -7.98529148e-01 2.64519632e-01 -7.30486393e-01
-1.16492443e-01 1.12106979e+00 -1.72007419e-02 -5.62144697e-01
6.10493779e-01 2.87313074e-01 -4.81081992e-01 -5.71139418e-02
-1.19251676e-01 -2.62404848e-02 3.38355899e-01 4.05090362e-01
5.10926284e-02 1.37455627e-01 -3.34195703e-01 2.08211690e-01
-2.12154165e-01 -3.60181510e-01 1.04026628e+00 7.64310777e-01
-8.30950558e-01 -8.27691734e-01 -1.06224489e+00 5.57505727e-01
-2.99627036e-01 -2.52826884e-02 1.16136767e-01 3.27672094e-01
-3.87079209e-01 1.00723326e+00 -1.09528050e-01 -4.43108454e-02
5.13182580e-01 -2.94752538e-01 4.89542007e-01 -5.50751388e-01
1.32940561e-01 4.29632008e-01 5.87324500e-02 -3.10071439e-01
-5.76614380e-01 -7.95650780e-01 -1.00234818e+00 -4.49399620e-01
-5.53090394e-01 5.09015739e-01 5.12606680e-01 8.94099772e-01
-8.25176686e-02 3.38423431e-01 9.75955069e-01 -7.80462980e-01
-9.21378553e-01 -7.68015087e-01 -1.15857542e+00 -7.44234920e-02
7.70781517e-01 -3.67910266e-01 -7.93047547e-01 -3.27705480e-02]
|
[6.575272083282471, 1.5165603160858154]
|
21bc4800-776e-4a96-b651-b427d4fb6fb5
|
uncertainty-regularized-policy-learning-for
| null | null |
https://openreview.net/forum?id=rwSWaS_tGgG
|
https://openreview.net/pdf?id=rwSWaS_tGgG
|
Uncertainty Regularized Policy Learning for Offline Reinforcement Learning
|
Recent studies show the promising results of using online RL methods in the offline setting.
However, such a learning diagram may suffer from an overtraining issue, that is, the performance of the policy degrades significantly as the training process continues when the dataset is not sufficiently large and diverse.
In this work, we propose an alternative approach to alleviate and avoid the overtraining issue: we explicitly take the learning stability into account in the policy learning objective, and adaptively select a good policy before the overtraining issue happens.
To do so, we develop an Uncertainty Regularized Policy Learning (URPL) method.
URPL adds an uncertainty regularization term in the policy learning objective to enforce to learn a more stable policy under the offline setting.
Moreover, we further use the uncertainty regularization term as a surrogate metric indicating the potential performance of a policy.
Based on the low-valued region of the uncertainty term, we can select a good policy with considerable good performance and low computation requirements.
On standard offline RL benchmark D4RL, URPL achieves much better final performance over existing state-of-the-art baselines.
|
['Chengqi Zhang', 'Xuan Song', 'Guodong Long', 'Pengfei Wei', 'Jing Jiang', 'Han Zheng']
|
2021-09-29
| null | null | null | null |
['d4rl']
|
['robots']
|
[-2.32659608e-01 -1.46208927e-02 -6.89674020e-01 -2.36955479e-01
-1.03001738e+00 -6.82753682e-01 5.70365906e-01 1.40990019e-01
-7.32116044e-01 1.10664284e+00 3.08495998e-01 -3.58079106e-01
-1.06380343e-01 -3.89785230e-01 -7.71638930e-01 -9.36098635e-01
2.03232974e-01 3.53817612e-01 1.31507874e-01 2.60541979e-02
1.85977429e-01 4.10666257e-01 -1.27385449e+00 -2.06038699e-01
1.10801303e+00 1.26113963e+00 2.36029401e-01 8.90697241e-02
-2.98644491e-02 7.83862174e-01 -6.55739427e-01 5.18362820e-02
5.34336269e-01 -2.13383108e-01 -6.12101495e-01 -9.17590037e-02
1.65168509e-01 -4.25147474e-01 -4.02826667e-02 1.09180462e+00
7.02659070e-01 5.50790370e-01 4.41464692e-01 -1.02305996e+00
-2.54054397e-01 6.51395023e-01 -6.32494271e-01 4.82823253e-02
4.56496738e-02 2.16801375e-01 1.13839710e+00 -7.48416305e-01
4.78359431e-01 1.41544616e+00 2.06925556e-01 5.27740777e-01
-1.26565015e+00 -7.02001572e-01 7.75585473e-01 -9.56832692e-02
-1.07997406e+00 -4.22621429e-01 7.01321065e-01 -2.58966058e-01
4.57792461e-01 -1.21961460e-01 4.02102947e-01 1.14425194e+00
-5.29135875e-02 1.10468614e+00 1.24485505e+00 -2.66500413e-01
6.03060126e-01 6.91039488e-02 -1.57793596e-01 5.58059096e-01
1.53413758e-01 4.03772622e-01 -3.12202126e-01 -2.14665383e-01
7.00013161e-01 -2.45873004e-01 -1.94302723e-01 -4.41386729e-01
-9.81350780e-01 8.08955491e-01 3.35391998e-01 1.62972093e-01
-3.31207812e-01 2.45027110e-01 5.13035715e-01 3.64798874e-01
6.13537490e-01 6.91240847e-01 -5.42712867e-01 -4.13655072e-01
-8.76498401e-01 3.35868210e-01 4.14707989e-01 4.76711184e-01
6.12420619e-01 4.48416881e-02 -7.55776763e-01 8.88139665e-01
2.45475844e-01 3.13124001e-01 3.99790227e-01 -1.16358733e+00
6.48454845e-01 4.00810033e-01 6.01552784e-01 -5.06825507e-01
-1.74139813e-01 -5.81166387e-01 -5.40324926e-01 4.69751656e-01
5.52092493e-01 -4.62031305e-01 -7.30876267e-01 2.04921699e+00
5.39368689e-01 1.50254577e-01 1.44177780e-01 1.05609190e+00
1.59651041e-01 7.29112506e-01 4.53121476e-02 -6.75389707e-01
8.72124374e-01 -1.04686797e+00 -8.58932376e-01 -2.47287735e-01
5.42873502e-01 -4.82169122e-01 1.42465115e+00 6.00386798e-01
-9.68990684e-01 -3.79174083e-01 -1.07869518e+00 2.96968877e-01
1.33317992e-01 5.26537061e-01 6.02520645e-01 2.39646912e-01
-7.24309087e-01 7.71972358e-01 -8.46250892e-01 8.94082785e-02
2.09905729e-01 3.17105025e-01 8.11941847e-02 1.55273125e-01
-1.21660244e+00 8.67650688e-01 6.56582236e-01 -8.61254113e-04
-9.03366923e-01 -6.63853705e-01 -5.03374934e-01 -1.27974197e-01
1.16476524e+00 -2.66349465e-01 1.59478021e+00 -1.00307739e+00
-2.18740582e+00 2.55148888e-01 7.81728849e-02 -5.23776174e-01
9.09941137e-01 -6.12666607e-01 -1.02667473e-01 -1.29999071e-01
-1.04132138e-01 3.62549365e-01 1.10607243e+00 -1.41625321e+00
-6.29037857e-01 -1.44399196e-01 2.16866687e-01 5.23956537e-01
-1.55979425e-01 -1.90381274e-01 -4.80262816e-01 -6.98601067e-01
-3.44550341e-01 -1.02543890e+00 -3.84458750e-01 -9.10197943e-02
-1.05126604e-01 -4.40032333e-01 6.66780055e-01 -3.12097251e-01
1.51465499e+00 -2.11087990e+00 1.36899844e-01 1.70270652e-01
-1.79391474e-01 5.49307823e-01 -1.54585674e-01 9.97469798e-02
3.92643034e-01 7.46713728e-02 -1.53791800e-01 -2.76065767e-01
3.50866541e-02 4.24000442e-01 -5.95434725e-01 3.97137970e-01
3.02754808e-02 6.26624167e-01 -1.20767021e+00 -2.97212034e-01
9.55118313e-02 4.35892045e-02 -6.55365944e-01 3.85962158e-01
-7.29241490e-01 8.79272401e-01 -8.77705157e-01 3.60766351e-01
3.63306701e-01 -1.81412458e-01 2.04702020e-01 8.47843513e-02
-2.67882168e-01 3.57145816e-01 -1.26857436e+00 1.69397950e+00
-6.70024931e-01 2.64463812e-01 2.31678590e-01 -9.18098867e-01
7.99565196e-01 2.08014235e-01 7.04966426e-01 -5.78754008e-01
1.28094211e-01 2.28389218e-01 -1.38522079e-02 -1.11770548e-01
4.02809143e-01 9.00372937e-02 4.07078955e-03 4.86024380e-01
-3.70067745e-01 8.18190947e-02 6.05323054e-02 -6.38747439e-02
8.03793609e-01 6.09938741e-01 4.04042453e-01 -3.18017095e-01
5.42239487e-01 -2.92057514e-01 7.49927580e-01 7.74727643e-01
-3.14713418e-01 1.74717546e-01 6.39548302e-01 -1.73182398e-01
-6.39307857e-01 -6.96446955e-01 -2.15300787e-02 1.32568586e+00
1.42638475e-01 -2.25309923e-01 -4.26770955e-01 -1.10958195e+00
2.66878396e-01 7.91722953e-01 -4.20677215e-01 -3.26452792e-01
-6.23631597e-01 -7.33199239e-01 2.92011976e-01 4.52655315e-01
5.39599121e-01 -8.78447056e-01 -5.87227941e-01 4.02483135e-01
-5.92964031e-02 -1.04525745e+00 -6.87133670e-01 1.42323285e-01
-7.79821396e-01 -6.95079327e-01 -7.02399135e-01 -1.98611051e-01
4.86481279e-01 -4.58268896e-02 8.23890567e-01 -2.11376041e-01
3.65086287e-01 5.37747256e-02 -3.01785946e-01 -2.93007851e-01
-2.95512885e-01 -2.00301828e-03 3.67105693e-01 2.62730587e-02
-1.83713213e-01 -2.45970130e-01 -7.25633085e-01 4.63618606e-01
-5.59449732e-01 -2.87024468e-01 5.58707893e-01 9.44399416e-01
8.18019867e-01 -6.08415604e-02 9.12629187e-01 -6.32514834e-01
9.46586609e-01 -3.25745851e-01 -1.01088905e+00 3.89971673e-01
-1.10674560e+00 6.96158469e-01 9.52639163e-01 -8.20029676e-01
-1.22728944e+00 9.68162494e-04 3.04924715e-02 -6.95355594e-01
4.92030233e-01 3.66894186e-01 -1.82652727e-01 3.89625207e-02
5.34844756e-01 5.72981732e-03 -4.28122059e-02 -5.94012737e-01
5.46630442e-01 5.47698855e-01 2.61120230e-01 -1.12027419e+00
6.46579683e-01 2.67167300e-01 -1.44134566e-01 -3.11589003e-01
-1.34143698e+00 -4.31201667e-01 -2.31246248e-01 -2.09360391e-01
3.97028238e-01 -9.12295043e-01 -8.11030805e-01 8.15819502e-02
-7.67982841e-01 -6.49417818e-01 -4.50201988e-01 6.82555079e-01
-7.09904790e-01 2.69053042e-01 -1.64146423e-01 -1.09631908e+00
-3.85503352e-01 -1.30062950e+00 8.92307580e-01 2.17819542e-01
1.07807994e-01 -8.30470502e-01 1.72671214e-01 -3.89935002e-02
2.48259813e-01 2.41234645e-01 7.32450485e-01 -5.82000434e-01
-3.27469349e-01 1.94831938e-01 -1.68688908e-01 5.41883945e-01
1.41886726e-01 -6.22635707e-02 -7.01323390e-01 -7.07070649e-01
1.22592142e-02 -6.74086928e-01 9.05740857e-01 3.38143975e-01
1.47641683e+00 -4.32882518e-01 -9.95592922e-02 5.09842038e-01
1.29197097e+00 3.13710719e-01 2.47831479e-01 4.01800573e-01
4.64696139e-01 3.16802651e-01 1.21296299e+00 7.38002121e-01
1.28242582e-01 7.90540159e-01 3.37620288e-01 1.74594134e-01
1.41130716e-01 -6.29955590e-01 7.57686496e-01 5.38831890e-01
2.34988125e-04 -2.97475040e-01 -6.01891696e-01 1.00322284e-01
-2.28273129e+00 -6.26270294e-01 6.44978821e-01 2.60428619e+00
1.31917274e+00 3.42133194e-01 1.50483087e-01 -3.02266270e-01
5.01942039e-01 5.27914524e-01 -1.05031228e+00 -3.64262193e-01
1.55999973e-01 -1.31041333e-01 6.35075510e-01 5.01531661e-01
-1.03957093e+00 1.15303767e+00 5.81642151e+00 1.19032097e+00
-1.39197004e+00 1.03037348e-02 6.20807588e-01 -2.43823037e-01
-7.41594434e-02 6.72092512e-02 -9.56453443e-01 5.64956725e-01
7.14838862e-01 -9.33964700e-02 7.20571518e-01 9.12687421e-01
5.84194064e-01 -1.75106049e-01 -9.71009672e-01 8.39676976e-01
-4.59463447e-01 -1.02438486e+00 -1.80304185e-01 7.33623654e-02
8.11264157e-01 7.44226575e-02 8.30099732e-02 6.66457653e-01
6.17669523e-01 -8.64533246e-01 7.86240101e-01 5.12155831e-01
7.56635308e-01 -9.20239985e-01 3.50389689e-01 6.32656574e-01
-1.03992927e+00 -2.91663736e-01 -4.90944713e-01 1.70816228e-01
3.14226113e-02 4.39166963e-01 -6.84401214e-01 4.77868974e-01
2.12628126e-01 5.90756059e-01 -1.64014414e-01 8.75863671e-01
-6.22642219e-01 6.25781536e-01 -3.46577317e-01 -1.53601337e-02
6.27008319e-01 -4.08827335e-01 6.41922653e-01 7.23844469e-01
1.90838426e-01 -1.83249563e-02 7.12527692e-01 6.92976832e-01
-2.34671235e-01 3.17367852e-01 -2.84204513e-01 -2.20149487e-01
5.05399525e-01 1.01442099e+00 -3.24949265e-01 -1.06008887e-01
-1.13266632e-01 7.40818441e-01 6.91409111e-01 4.65132713e-01
-8.26564014e-01 1.13370992e-01 5.60383379e-01 -2.08756760e-01
1.80111960e-01 -2.28419542e-01 1.13382125e-02 -1.20934761e+00
1.11930780e-01 -9.75562274e-01 3.21924925e-01 -2.57981598e-01
-1.09106278e+00 3.28676373e-01 9.98898670e-02 -1.38599503e+00
-4.00590569e-01 -3.29976380e-01 -4.19859409e-01 6.05360866e-01
-1.68609238e+00 -5.36490917e-01 2.93295145e-01 4.06322807e-01
5.70504010e-01 -1.74664840e-01 4.02660668e-01 1.47864178e-01
-7.67263830e-01 7.62922168e-01 4.71176654e-01 -3.57850224e-01
9.94470000e-01 -1.35024369e+00 -1.99435204e-01 6.51097119e-01
-3.17865536e-02 5.55242479e-01 6.05501056e-01 -6.36125028e-01
-1.25038707e+00 -1.13424993e+00 1.90779418e-01 -1.90196276e-01
7.22822845e-01 -1.98295161e-01 -8.82306635e-01 4.23047841e-01
-2.37027884e-01 1.63685873e-01 1.74883097e-01 1.17289834e-01
-1.86243549e-01 -2.46314242e-01 -1.03133559e+00 7.07260787e-01
8.08051407e-01 -2.32610121e-01 -3.90586644e-01 3.89799118e-01
9.39263642e-01 -5.05868912e-01 -8.73402715e-01 6.01716816e-01
4.95933950e-01 -5.35641730e-01 7.68257439e-01 -5.59062600e-01
5.42838126e-03 -2.39278391e-01 5.47782443e-02 -1.49043405e+00
-7.44472146e-02 -1.14054036e+00 -6.00122035e-01 1.07608593e+00
4.74203020e-01 -6.28025591e-01 5.99965215e-01 5.07300138e-01
9.65583175e-02 -1.28407395e+00 -9.35986936e-01 -1.07603633e+00
2.50464052e-01 -2.05396786e-01 5.24582028e-01 5.44179440e-01
-1.31282821e-01 2.05608621e-01 -6.95757866e-01 -5.30931130e-02
5.06280065e-01 1.87896028e-01 6.70133471e-01 -1.07896674e+00
-4.72194344e-01 -4.28449959e-01 4.25821930e-01 -1.53432953e+00
5.43749750e-01 -5.59472620e-01 3.25828850e-01 -1.22843087e+00
-1.71309501e-01 -7.42596090e-01 -6.72727168e-01 5.00969291e-01
-3.67460668e-01 -6.02886200e-01 2.66647905e-01 4.96441573e-01
-8.57194006e-01 1.00401461e+00 1.53339875e+00 1.50671497e-01
-6.96085572e-01 3.48140717e-01 -6.49149776e-01 7.16260731e-01
9.95764077e-01 -4.11858231e-01 -7.49490619e-01 -2.32747421e-01
1.51255265e-01 2.35458642e-01 -1.27577692e-01 -5.64680994e-01
6.03797659e-02 -6.21080816e-01 5.08912019e-02 -3.77135694e-01
2.06299394e-01 -7.89697230e-01 -3.78374040e-01 4.54783440e-01
-6.07152402e-01 -1.78581789e-01 9.16963145e-02 8.71958256e-01
-6.89266846e-02 -1.77541837e-01 1.03814995e+00 1.97622683e-02
-6.33069694e-01 5.32854736e-01 -2.21103039e-02 3.12884003e-01
9.35020983e-01 2.71887124e-01 -1.49658784e-01 -3.35023731e-01
-4.65744108e-01 7.44310558e-01 4.32161242e-01 5.98738253e-01
3.41678292e-01 -1.26863825e+00 -4.12968308e-01 -5.37444949e-02
2.77038012e-02 2.01047901e-02 -3.32118452e-01 7.69841850e-01
5.57990298e-02 3.71809810e-01 1.85223475e-01 -3.70189399e-01
-7.09148943e-01 5.03400445e-01 3.37958843e-01 -7.91000247e-01
-6.16100013e-01 6.43778384e-01 -2.55441461e-02 -2.74912864e-01
7.16601253e-01 -4.15799409e-01 -2.63460129e-01 1.51434571e-01
3.51647705e-01 3.41983527e-01 -9.99567062e-02 -2.86161840e-01
-3.07601452e-01 3.27583790e-01 -1.45966798e-01 -3.36629450e-01
9.96694982e-01 -1.51487142e-01 3.82562429e-01 5.36139727e-01
9.78856146e-01 1.24546610e-01 -2.04040098e+00 -5.57256818e-01
2.72788584e-01 -4.87641782e-01 3.59064579e-01 -1.02769530e+00
-9.43082690e-01 3.80015045e-01 5.06476462e-01 -1.62982121e-01
1.01701093e+00 -3.25947553e-01 7.12973893e-01 5.80857635e-01
4.67702657e-01 -1.61626029e+00 1.90282583e-01 7.31055319e-01
9.44388807e-01 -1.52890265e+00 3.25470358e-01 1.18535995e-01
-9.83608127e-01 8.89030039e-01 6.11465871e-01 -1.83141366e-01
4.69858468e-01 2.19968595e-02 -1.08270971e-02 2.24390969e-01
-1.00739551e+00 -3.10827792e-01 4.48730528e-01 4.32100054e-03
1.83435231e-01 1.21934310e-01 -6.10904157e-01 4.80247498e-01
5.60565367e-02 -2.10473575e-02 5.10865450e-02 9.10638154e-01
-5.57815969e-01 -1.28109443e+00 -9.88684073e-02 2.62323946e-01
-6.54536664e-01 2.04452723e-01 -2.57493138e-01 6.61925972e-01
-1.71850204e-01 8.52572322e-01 -3.28706920e-01 -1.20983064e-01
3.31944317e-01 -7.08098263e-02 4.16092247e-01 -4.72396642e-01
-6.09764397e-01 3.18678170e-01 7.12315738e-02 -8.53933811e-01
-4.42974240e-01 -4.99197870e-01 -1.50447190e+00 1.45340547e-01
-4.50996250e-01 3.34877938e-01 5.01997411e-01 1.05033469e+00
3.58304441e-01 3.62827808e-01 9.78276730e-01 -6.20518267e-01
-1.44748271e+00 -7.53129423e-01 -4.31428105e-01 2.09081337e-01
4.97839838e-01 -1.01074755e+00 -3.80711287e-01 -6.03360653e-01]
|
[4.091881275177002, 2.204296827316284]
|
6921e664-cb54-4d55-83c2-c9bf43f5e60f
|
measuring-commonality-in-recommendation-of
|
2208.01696
| null |
https://arxiv.org/abs/2208.01696v1
|
https://arxiv.org/pdf/2208.01696v1.pdf
|
Measuring Commonality in Recommendation of Cultural Content: Recommender Systems to Enhance Cultural Citizenship
|
Recommender systems have become the dominant means of curating cultural content, significantly influencing the nature of individual cultural experience. While the majority of research on recommender systems optimizes for personalized user experience, this paradigm does not capture the ways that recommender systems impact cultural experience in the aggregate, across populations of users. Although existing novelty, diversity, and fairness studies probe how systems relate to the broader social role of cultural content, they do not adequately center culture as a core concept and challenge. In this work, we introduce commonality as a new measure that reflects the degree to which recommendations familiarize a given user population with specified categories of cultural content. Our proposed commonality metric responds to a set of arguments developed through an interdisciplinary dialogue between researchers in computer science and the social sciences and humanities. With reference to principles underpinning non-profit, public service media systems in democratic societies, we identify universality of address and content diversity in the service of strengthening cultural citizenship as particularly relevant goals for recommender systems delivering cultural content. Taking diversity in movie recommendation as a case study in enhancing pluralistic cultural experience, we empirically compare systems' performance using commonality and existing utility, diversity, and fairness metrics. Our results demonstrate that commonality captures a property of system behavior complementary to existing metrics and suggest the need for alternative, non-personalized interventions in recommender systems oriented to strengthening cultural citizenship across populations of users. In this way, commonality contributes to a growing body of scholarship developing 'public good' rationales for digital media and ML systems.
|
['Georgina Born', 'Fernando Diaz', 'Gustavo Ferreira', 'Andres Ferraro']
|
2022-08-02
| null | null | null | null |
['movie-recommendation']
|
['miscellaneous']
|
[-3.40786427e-01 -7.62375891e-02 -3.55308235e-01 6.14146516e-02
-1.24971896e-01 -8.92178118e-01 7.28891551e-01 3.16658616e-01
-3.91835093e-01 1.92795515e-01 1.18629682e+00 -1.76027775e-01
-2.68700927e-01 -6.84724391e-01 -1.83166385e-01 -5.37824571e-01
4.24689710e-01 -3.54283899e-01 -4.49008316e-01 -8.75019670e-01
6.77977085e-01 6.54530227e-02 -1.69998658e+00 5.14994621e-01
1.08131015e+00 5.42385757e-01 5.04087657e-02 3.99461716e-01
-1.66765839e-01 7.55519867e-01 -3.03258240e-01 -9.09786046e-01
7.68746138e-02 -8.46311271e-01 -4.34769481e-01 -2.82761455e-01
3.72339219e-01 -3.13581467e-01 1.06243603e-01 1.00298345e+00
8.15805137e-01 1.85920164e-01 7.12371230e-01 -9.12501216e-01
-1.65999770e+00 9.70670700e-01 1.73873007e-02 1.55324697e-01
6.14293337e-01 -1.74157679e-01 1.18117547e+00 -5.88842750e-01
7.73867905e-01 9.40591812e-01 8.87832880e-01 3.39462847e-01
-1.11445177e+00 -3.86871576e-01 3.03478509e-01 -5.07740855e-01
-1.29270756e+00 -3.57228875e-01 2.28647038e-01 -9.01430607e-01
5.03445685e-01 8.03629637e-01 1.21120036e+00 9.46871758e-01
1.32691368e-01 1.76785618e-01 1.13323045e+00 -4.04312909e-01
3.52599561e-01 6.64427996e-01 -4.59793061e-02 -6.22449368e-02
6.27588630e-01 -2.87918568e-01 -3.78869414e-01 -3.63962203e-01
5.93127728e-01 2.82103390e-01 -4.21508402e-01 -9.44259688e-02
-8.61129940e-01 9.19875622e-01 -7.33650401e-02 7.28456855e-01
-5.11739194e-01 -1.39219418e-01 3.33555102e-01 5.25308728e-01
7.04189360e-01 8.91308725e-01 3.69474292e-02 -6.34099662e-01
-6.15447044e-01 1.93767130e-01 1.11676514e+00 6.35608912e-01
1.11830398e-01 -5.99684566e-02 -2.36331701e-01 1.06158662e+00
4.09559757e-01 6.63686156e-01 3.24246615e-01 -1.16913188e+00
-5.01947463e-01 5.70941567e-01 2.73810476e-01 -1.26331997e+00
-1.29581288e-01 -7.74737656e-01 -1.38824999e-01 -8.61080177e-03
3.40494722e-01 -4.91191357e-01 1.48551300e-01 1.62872171e+00
1.55547345e-02 -2.70419329e-01 -3.69747914e-02 1.17103326e+00
8.40038776e-01 4.72668529e-01 1.73554957e-01 -2.15626150e-01
1.07890546e+00 -3.78976822e-01 -4.34240371e-01 3.21877658e-01
6.86097264e-01 -1.12021983e+00 1.33521235e+00 5.13990939e-01
-1.02908039e+00 -1.04134627e-01 -7.20796049e-01 8.88493285e-02
-4.17053014e-01 -3.70243609e-01 6.94900930e-01 1.52116275e+00
-1.26400244e+00 3.85611355e-01 5.01531735e-02 -1.05140352e+00
-6.93750009e-03 -2.37463444e-01 1.49167657e-01 -8.10958818e-03
-8.72274518e-01 9.47696924e-01 -5.17835200e-01 -2.80102283e-01
-2.80412257e-01 -8.33339632e-01 -1.24574080e-01 -1.17289290e-01
5.42830583e-03 -8.71280432e-01 8.73554647e-01 -1.77840817e+00
-1.43409789e+00 5.96527100e-01 6.14617884e-01 1.61552697e-01
2.87856311e-01 -4.66931015e-01 -9.76864398e-01 -2.93127179e-01
2.10387371e-02 -4.04160470e-02 1.04502156e-01 -1.57749474e+00
-8.30972493e-01 2.77148385e-04 4.80922431e-01 4.32631701e-01
-8.47489834e-01 3.95902187e-01 1.59110557e-02 -4.33054179e-01
-2.81271070e-01 -6.87478900e-01 -1.50942624e-01 -3.81673604e-01
2.93228358e-01 4.03180152e-01 2.93959677e-01 -5.46828151e-01
1.70665574e+00 -2.37959266e+00 -7.85855353e-02 4.97460395e-01
2.33562708e-01 -8.83095413e-02 -1.23860039e-01 1.08979642e+00
6.40851438e-01 5.38206041e-01 5.45196056e-01 2.03629628e-01
2.46016368e-01 -6.77116215e-02 -1.89820424e-01 4.67516035e-01
-4.90182579e-01 2.58632749e-01 -1.18825638e+00 2.06957832e-01
-2.27788553e-01 9.29388583e-01 -1.16613805e+00 -4.78899837e-01
1.33801386e-01 1.98842481e-01 -3.07667434e-01 4.01325047e-01
3.42463553e-01 -1.13202386e-01 6.25711143e-01 3.54530215e-01
-6.99198365e-01 1.03853047e-01 -1.02221143e+00 1.31850052e+00
-4.74529266e-01 5.38671553e-01 -1.36805460e-01 -1.42188713e-01
1.02392125e+00 1.56456992e-01 6.47080123e-01 -1.02042687e+00
3.40134323e-01 1.66367829e-01 3.15462589e-01 -5.71320415e-01
1.38823998e+00 -1.91392154e-01 -9.77950692e-02 7.88996279e-01
-3.93151462e-01 1.84424669e-01 -1.47324473e-01 3.82755876e-01
5.86606979e-01 1.65924847e-01 3.26695085e-01 -8.41594279e-01
2.81858165e-03 -5.13050035e-02 3.36494654e-01 5.96123636e-01
-1.54189125e-01 5.66109300e-01 4.37901348e-01 -8.34432989e-02
-1.25870740e+00 -6.42201960e-01 -1.22058809e-01 1.59587669e+00
4.27398086e-01 -7.21256137e-01 -6.77160859e-01 -7.95166939e-02
2.86286086e-01 1.03520894e+00 -6.89073503e-01 -7.71900490e-02
1.66166410e-01 -5.43026686e-01 3.72749716e-01 -4.08271700e-02
-9.90045145e-02 -5.39707422e-01 -7.22312629e-01 1.46549180e-01
-1.01703905e-01 -6.74809515e-01 -6.48325205e-01 -6.61322176e-01
-5.86140096e-01 -9.47274089e-01 -8.24427307e-01 -4.18651074e-01
2.37933874e-01 8.18006873e-01 1.12683499e+00 3.65071803e-01
2.36394599e-01 1.14110482e+00 -8.02099884e-01 -2.57142365e-01
-4.45186049e-01 -1.53805763e-01 1.66186884e-01 -5.95829636e-02
5.99542499e-01 -6.53665364e-01 -8.11389446e-01 3.47466052e-01
-7.63687849e-01 -2.77187943e-01 9.97507423e-02 2.13467434e-01
5.02870679e-02 -3.94916326e-01 8.75508845e-01 -9.89274502e-01
1.20423698e+00 -1.24576402e+00 1.99965343e-01 1.01377659e-01
-9.85078394e-01 -6.80629075e-01 5.19373715e-01 -4.26020622e-01
-1.14482021e+00 -9.09406126e-01 2.05860198e-01 3.45277816e-01
3.21232229e-01 7.38876462e-01 2.03149527e-01 -2.13222094e-02
1.04198408e+00 -1.62520424e-01 2.87372500e-01 -2.20127821e-01
6.79297507e-01 7.53584862e-01 -4.45786566e-02 -7.51742065e-01
1.18575528e-01 5.61263740e-01 -6.97418153e-01 -9.78757083e-01
-2.46659711e-01 -5.08796036e-01 -5.62274382e-02 -8.82881999e-01
7.36030281e-01 -1.06127238e+00 -8.01650107e-01 -2.31512904e-01
-2.76304752e-01 -1.83235794e-01 -3.14666182e-01 7.82639503e-01
-2.47303173e-01 3.63321900e-01 -4.72090930e-01 -1.18106902e+00
-3.55122060e-01 -6.73253834e-01 3.08448106e-01 4.17171180e-01
-5.11894882e-01 -1.02500451e+00 3.93368661e-01 4.63655740e-01
8.35910439e-01 2.81025052e-01 6.19633615e-01 -3.67538393e-01
-1.64668888e-01 -3.82411741e-02 9.42972377e-02 2.41601706e-01
-6.87879026e-02 4.10637259e-01 -6.65623963e-01 -1.69153690e-01
-3.15858155e-01 4.23449539e-02 1.45302430e-01 2.25824803e-01
-3.49375680e-02 -3.59608442e-01 2.80792624e-01 8.53320286e-02
1.74507844e+00 3.99228066e-01 7.85204887e-01 6.17500901e-01
3.66009533e-01 6.64276123e-01 5.44383764e-01 1.05426943e+00
8.14255476e-01 3.83094460e-01 2.41265632e-02 1.59479722e-01
3.66095565e-02 -2.49453396e-01 6.69032574e-01 1.05992019e+00
-7.41311133e-01 6.10273294e-02 -6.90843821e-01 3.77106845e-01
-1.61813533e+00 -1.35589027e+00 -3.57815892e-01 2.35590029e+00
2.52474278e-01 -3.10374767e-01 7.15055346e-01 -3.76893282e-01
5.72236419e-01 -1.21230945e-01 8.00866634e-02 -9.32686388e-01
-2.67616749e-01 -3.44201267e-01 4.04179424e-01 3.82646680e-01
4.19349000e-02 7.10551977e-01 6.34383202e+00 1.84200555e-01
-1.07080412e+00 1.43419772e-01 5.16969144e-01 -5.35565794e-01
-1.29272008e+00 8.32240954e-02 -4.01599318e-01 3.39928448e-01
8.56097817e-01 -6.11724913e-01 6.61794186e-01 8.34204853e-01
6.98821068e-01 -1.25004232e-01 -6.48378670e-01 6.10560596e-01
4.71267700e-01 -1.28215873e+00 -2.10767120e-01 4.46435541e-01
1.26548004e+00 -2.58914411e-01 5.81773818e-01 2.53282905e-01
4.87121969e-01 -7.07549751e-01 1.19738972e+00 8.29036057e-01
3.07921231e-01 -7.54497051e-01 3.64584416e-01 -2.08046660e-01
-5.81132293e-01 -2.20857531e-01 -3.13664973e-01 -7.66161382e-01
1.44325763e-01 3.53941172e-01 4.90258560e-02 1.16392493e-01
6.22465253e-01 4.26603824e-01 -2.98710674e-01 1.08790016e+00
5.74978411e-01 6.07386470e-01 2.37916201e-01 -3.43392551e-01
2.30195776e-01 -6.88477159e-01 3.41579527e-01 1.28627169e+00
6.56090975e-01 3.08318526e-01 -4.15034831e-01 5.68732738e-01
2.74417132e-01 8.58978689e-01 -7.20695972e-01 -4.31816638e-01
8.69783998e-01 1.09399176e+00 -5.79891443e-01 8.73635150e-03
-7.00050175e-01 4.51062530e-01 -8.55883360e-02 2.99530208e-01
-5.96565604e-01 2.30254486e-01 9.15790558e-01 3.65450233e-01
3.42886001e-02 -1.58474982e-01 -6.25921071e-01 -1.00482142e+00
-4.74100024e-01 -1.30946612e+00 2.62957215e-01 -4.63111967e-01
-1.25966763e+00 1.54313281e-01 -5.23753881e-01 -1.02734399e+00
2.89241046e-01 -9.01725888e-02 -3.84002715e-01 7.55985141e-01
-9.06953871e-01 -1.13611376e+00 -2.49445960e-01 3.11791152e-01
-1.74662367e-01 -1.35031417e-01 8.53829622e-01 5.68060040e-01
-3.87768358e-01 5.70053697e-01 8.55624318e-01 -7.17762113e-01
9.26109076e-01 -7.74771333e-01 -5.78068532e-02 7.82237291e-01
-1.23862624e-01 1.02801657e+00 8.20912719e-01 -6.41012609e-01
-1.54325426e+00 -4.84203935e-01 7.64400005e-01 -5.43790936e-01
6.09559953e-01 -1.05201960e-01 -3.60559046e-01 2.98642963e-01
4.18701768e-01 -9.50230718e-01 1.81850147e+00 8.76188278e-01
-5.28807759e-01 9.14028734e-02 -1.16865361e+00 1.02412271e+00
1.36600459e+00 -6.15089118e-01 5.13963588e-02 -1.06179453e-01
6.08693779e-01 2.27692500e-01 -1.62718856e+00 -5.21796823e-01
1.22472847e+00 -1.20512652e+00 5.91361105e-01 -4.79327977e-01
8.38237166e-01 1.03750855e-01 -8.45751822e-01 -1.34738970e+00
-9.52388406e-01 -5.95500648e-01 4.46787983e-01 1.45039785e+00
4.81596202e-01 -7.80855060e-01 4.06627148e-01 1.17729437e+00
-4.60245728e-01 -4.82539415e-01 -4.70205545e-02 -2.23396152e-01
4.68778044e-01 -2.59652823e-01 8.78992558e-01 1.69993293e+00
5.25378287e-01 7.63646215e-02 -4.83139783e-01 -1.20151110e-01
2.13035315e-01 -2.34177947e-01 9.93711412e-01 -1.29064643e+00
-3.20176840e-01 -1.01924348e+00 -3.00354898e-01 -1.95220217e-01
-5.15763521e-01 -7.04288125e-01 -6.88904703e-01 -1.62781000e+00
2.97099948e-01 -5.01851618e-01 -2.09902540e-01 -3.97565305e-01
2.00966433e-01 3.10093373e-01 5.31893551e-01 3.79381955e-01
-6.81618571e-01 6.65665716e-02 1.37449646e+00 5.74223816e-01
-5.31219423e-01 -4.86411422e-01 -2.10350680e+00 3.71074796e-01
6.61773026e-01 4.11949828e-02 -5.73167205e-01 -2.00294599e-01
1.24108553e+00 -3.18160206e-01 -3.25746238e-02 -6.64417565e-01
7.54907355e-02 -7.16014802e-01 -5.12392586e-03 3.36427808e-01
9.98465158e-03 -8.10554147e-01 7.29014218e-01 3.14517111e-01
-3.43414158e-01 1.36555910e-01 -1.75914034e-01 2.50316799e-01
1.67008713e-01 -4.20403779e-02 4.40384418e-01 4.16365452e-02
-6.29560709e-01 -2.01373175e-01 -6.19137824e-01 1.79017112e-01
9.32434678e-01 -6.91930294e-01 -7.02536345e-01 -7.91387379e-01
-6.50181055e-01 -2.52304077e-01 1.22766495e+00 6.11366212e-01
8.77192467e-02 -1.18211651e+00 -7.82698095e-01 -4.38854069e-01
3.21455926e-01 -1.10700619e+00 6.12291932e-01 8.78951788e-01
-4.70310330e-01 3.44751142e-02 -5.40421903e-01 1.32779345e-01
-9.60359573e-01 3.18979919e-01 2.92187363e-01 8.50607395e-01
-4.47745353e-01 5.54351628e-01 -1.11707551e-02 -1.34528697e-01
-1.56973705e-01 9.74869728e-02 -5.28696239e-01 4.99015719e-01
5.45753777e-01 7.08328009e-01 -3.79882216e-01 -9.18236971e-01
1.22783326e-01 1.88440174e-01 1.04336455e-01 -2.45969251e-01
1.22930789e+00 -5.00099540e-01 -1.53091386e-01 5.33366501e-01
8.41805220e-01 7.06879854e-01 -7.70779908e-01 7.40969628e-02
-3.53627145e-01 -1.04433453e+00 1.19825043e-02 -9.83050823e-01
-9.73483920e-01 1.65995419e-01 6.48446560e-01 6.01854563e-01
8.92977536e-01 -3.72841567e-01 3.35222512e-01 -7.36567825e-02
3.31738740e-01 -1.63232803e+00 -4.58435789e-02 1.64728314e-01
7.98822939e-01 -4.74917740e-01 -2.08316613e-02 -1.66933760e-01
-1.27450752e+00 7.41548419e-01 4.11708146e-01 -2.93003358e-02
8.67825210e-01 -6.63056970e-02 3.20267677e-01 1.25481114e-02
-6.23843670e-01 -2.32401416e-01 7.91090280e-02 6.77355647e-01
1.32003438e+00 5.46603501e-01 -1.31752849e+00 1.04142630e+00
-5.89843035e-01 1.25915945e-01 1.02232707e+00 6.05385125e-01
-6.68253720e-01 -9.64702189e-01 -3.54913771e-01 5.64200461e-01
-7.53014445e-01 -2.54137903e-01 -5.59813857e-01 7.52178907e-01
3.09250265e-01 1.06212926e+00 -5.71226235e-03 -5.99991322e-01
2.17647821e-01 -3.14962655e-01 2.29324296e-01 -3.99627596e-01
-1.36989462e+00 1.27446905e-01 6.84508562e-01 -2.99555987e-01
-3.30818474e-01 -1.12876260e+00 -8.49009335e-01 -1.03557205e+00
-1.30583927e-01 2.28208482e-01 8.36133122e-01 4.81521964e-01
7.62903988e-01 1.75724462e-01 5.64622045e-01 -1.25868395e-01
-2.20830813e-01 -4.89685297e-01 -1.04389596e+00 5.99277675e-01
-4.23938215e-01 -4.23494577e-01 -2.54650004e-02 -1.79499939e-01]
|
[9.73599624633789, 5.8445329666137695]
|
098f0381-8ed9-469c-b097-71a3b6907380
|
evaluating-token-level-and-passage-level
|
2203.11163
| null |
https://arxiv.org/abs/2203.11163v2
|
https://arxiv.org/pdf/2203.11163v2.pdf
|
Evaluating Token-Level and Passage-Level Dense Retrieval Models for Math Information Retrieval
|
With the recent success of dense retrieval methods based on bi-encoders, studies have applied this approach to various interesting downstream retrieval tasks with good efficiency and in-domain effectiveness. Recently, we have also seen the presence of dense retrieval models in Math Information Retrieval (MIR) tasks, but the most effective systems remain classic retrieval methods that consider hand-crafted structure features. In this work, we try to combine the best of both worlds:\ a well-defined structure search method for effective formula search and efficient bi-encoder dense retrieval models to capture contextual similarities. Specifically, we have evaluated two representative bi-encoder models for token-level and passage-level dense retrieval on recent MIR tasks. Our results show that bi-encoder models are highly complementary to existing structure search methods, and we are able to advance the state-of-the-art on MIR datasets.
|
['Jimmy Lin', 'Yuqing Xie', 'Jheng-Hong Yang', 'Wei Zhong']
|
2022-03-21
| null | null | null | null |
['math-information-retrieval']
|
['natural-language-processing']
|
[-8.36611018e-02 -2.77428269e-01 -3.86335731e-01 -8.92649740e-02
-1.52456546e+00 -5.10209501e-01 1.04095542e+00 4.86646801e-01
-6.14084959e-01 5.37446439e-01 8.05421650e-01 -2.30042953e-02
-7.41304755e-01 -8.61180365e-01 -6.23279035e-01 -2.87807614e-01
1.53378740e-01 9.46284413e-01 3.54282379e-01 -7.08733559e-01
5.88948250e-01 2.42837295e-01 -1.77204168e+00 6.66410923e-01
6.79682016e-01 7.42768645e-01 2.92434037e-01 6.30349219e-01
-1.34032726e-01 9.88029897e-01 -4.80323404e-01 -5.47790706e-01
-3.56934179e-04 -2.59163529e-01 -1.27403367e+00 -6.92090929e-01
5.62573731e-01 -3.67615819e-01 -8.33184898e-01 6.53736174e-01
7.41865098e-01 2.87189037e-01 9.06763971e-01 -4.06799823e-01
-9.10271108e-01 9.42166805e-01 -2.52114803e-01 5.28293550e-01
8.12915981e-01 -2.31170058e-01 1.57608414e+00 -8.61599743e-01
1.28161693e+00 1.34984899e+00 4.91158187e-01 2.26831645e-01
-1.12453294e+00 -3.39895546e-01 -3.09759915e-01 7.17116714e-01
-1.59079790e+00 -3.68078202e-01 4.43400502e-01 -1.49849311e-01
1.47133446e+00 1.95068955e-01 3.87840748e-01 1.10615742e+00
1.44560292e-01 1.24617839e+00 7.24679649e-01 -8.21519911e-01
-2.58721381e-01 -6.01969101e-02 3.85963410e-01 5.77818871e-01
-4.75166552e-02 2.75491983e-01 -6.89591706e-01 -8.85200202e-02
3.78918678e-01 6.52207956e-02 -2.63633877e-01 -7.12805390e-02
-1.16500056e+00 9.21725690e-01 3.77700806e-01 9.16961432e-01
-7.41247833e-02 2.37976983e-01 6.17860377e-01 7.18171179e-01
6.17097497e-01 9.83423591e-01 -3.66944611e-01 -4.57697451e-01
-1.45102382e+00 6.72765613e-01 7.85607576e-01 1.05578661e+00
6.53169334e-01 -6.47209883e-01 -7.60971844e-01 9.87161756e-01
2.30360687e-01 2.79005766e-01 7.71848500e-01 -7.86961555e-01
5.48976183e-01 4.32695627e-01 -2.50412732e-01 -9.87376332e-01
3.62184308e-02 -4.93073314e-01 -6.39240444e-01 -7.94570982e-01
1.20545037e-01 5.49711168e-01 -7.36346662e-01 1.36867678e+00
-1.60011142e-01 6.63710572e-03 -5.62582277e-02 6.10763013e-01
9.15785372e-01 7.60788143e-01 -1.01126686e-01 6.04798682e-02
1.15962267e+00 -1.03559303e+00 -8.38644922e-01 2.89636612e-01
1.02261519e+00 -1.06696141e+00 9.40303624e-01 3.27950686e-01
-1.61134791e+00 -6.33972228e-01 -9.75269735e-01 -5.49103200e-01
-7.25726962e-01 2.02137008e-01 6.34373307e-01 2.90938139e-01
-1.23083031e+00 9.01111901e-01 -5.94249070e-01 -2.45141417e-01
5.56260236e-02 1.70997128e-01 -2.18538404e-01 -6.91368043e-01
-1.52285993e+00 1.04284728e+00 2.45279729e-01 -7.66183808e-02
-9.84056830e-01 -8.42512727e-01 -7.41787553e-01 4.07430440e-01
2.09109202e-01 -8.69921505e-01 1.33685160e+00 -4.23729926e-01
-1.34292829e+00 9.16530132e-01 -1.60980120e-01 -5.25177717e-01
3.81561927e-02 -5.87661684e-01 -1.61856294e-01 5.14523149e-01
-1.37441114e-01 5.96269727e-01 4.89402682e-01 -7.83199430e-01
-2.05210239e-01 -1.95279360e-01 1.22199655e-01 2.10481897e-01
-5.45402408e-01 3.89307320e-01 -6.08166158e-01 -6.81954443e-01
-1.91272289e-01 -5.59466422e-01 -2.81692762e-02 -6.11528814e-01
-1.64357975e-01 -6.72847688e-01 3.20385873e-01 -4.23060358e-01
1.78920698e+00 -1.91446042e+00 5.18266678e-01 6.47302866e-02
1.18340917e-01 4.90742058e-01 -4.99890089e-01 1.16361785e+00
1.93378046e-01 2.56451845e-01 2.46299058e-01 -4.15126950e-01
3.29770237e-01 -1.42169759e-01 -5.61161220e-01 1.69456273e-01
1.10313289e-01 1.32311118e+00 -9.67681885e-01 -6.81784570e-01
1.06261432e-01 3.96897435e-01 -6.41380966e-01 3.84282142e-01
-2.64293402e-01 -1.44025728e-01 -5.87219417e-01 5.61662734e-01
1.39196992e-01 -3.89988184e-01 -6.52006119e-02 1.50570244e-01
3.46592069e-02 8.02695215e-01 -8.90978098e-01 2.09017420e+00
-5.73316514e-01 7.18874156e-01 -1.87898502e-01 -9.34808433e-01
5.24909973e-01 4.84550506e-01 5.05028009e-01 -1.19616354e+00
-4.41727117e-02 4.35055673e-01 -3.17042172e-01 -3.58535647e-01
1.08662057e+00 1.81096226e-01 -2.12560967e-03 4.98741925e-01
3.63926083e-01 -3.77787352e-01 5.30449450e-01 6.82024121e-01
1.51661277e+00 1.10400788e-01 4.33914959e-02 -2.45501980e-01
5.97755849e-01 -1.11811168e-01 -2.82407731e-01 1.09925508e+00
3.76904130e-01 7.44943917e-01 -2.50493120e-02 -7.79219642e-02
-1.01272154e+00 -8.76801074e-01 -3.85243118e-01 1.31304288e+00
2.16334220e-02 -1.10042131e+00 -3.23484182e-01 -2.46026248e-01
-3.00261993e-02 3.21018517e-01 -3.23723912e-01 -3.24576318e-01
-7.51785100e-01 -2.59273142e-01 7.49702752e-01 3.72520328e-01
2.01097354e-01 -1.12255061e+00 -1.04624443e-01 3.05957288e-01
-1.91166505e-01 -9.29075778e-01 -2.40325928e-01 2.17712119e-01
-7.77519763e-01 -7.54528046e-01 -1.13628960e+00 -7.67001748e-01
4.78670411e-02 3.00815463e-01 1.86742043e+00 3.89696211e-01
-7.46630847e-01 7.14522898e-01 -8.63583744e-01 -1.44923255e-01
-2.56874263e-01 6.59844279e-01 -3.59081388e-01 -7.37648666e-01
4.55250680e-01 -3.57840985e-01 -6.77101433e-01 5.34192249e-02
-1.23526311e+00 -3.83717209e-01 6.97781086e-01 9.81243908e-01
4.84062999e-01 -2.24374890e-01 6.69912755e-01 -7.36003101e-01
1.03690660e+00 -4.69410092e-01 -3.94746691e-01 5.76421440e-01
-8.44743788e-01 4.07047927e-01 2.79817551e-01 -1.55383348e-01
-1.00212491e+00 -5.35187185e-01 -4.70088124e-01 -4.04272556e-01
-2.08729804e-02 1.01229286e+00 4.54983681e-01 7.73711205e-02
7.09180057e-01 4.06583875e-01 -4.44540679e-01 -7.71271765e-01
7.86574006e-01 7.29562163e-01 5.97768575e-02 -1.09397495e+00
4.56107914e-01 -5.80797568e-02 -2.07371153e-02 -7.08082914e-01
-8.90825152e-01 -1.13131881e+00 -4.70701188e-01 2.10658893e-01
6.80949390e-01 -1.09267223e+00 -2.90896833e-01 3.77602465e-02
-1.07388628e+00 -7.27139562e-02 -3.82268578e-01 3.67157847e-01
-5.20367146e-01 7.07694829e-01 -1.28335726e+00 -5.20695984e-01
-5.45862556e-01 -1.23102105e+00 1.67787814e+00 -5.38764782e-02
-1.00519255e-01 -1.03517628e+00 6.67331636e-01 3.96978557e-01
6.95478857e-01 -5.47026277e-01 1.06488538e+00 -6.93039536e-01
-1.06797040e+00 -4.22525316e-01 -2.73996174e-01 2.20703989e-01
-4.02929842e-01 -2.13646308e-01 -9.19481099e-01 -2.19152465e-01
-5.55195153e-01 -9.76387858e-01 1.43841052e+00 2.40138620e-01
1.19614148e+00 7.71978870e-03 -4.26948160e-01 3.37249398e-01
1.47301793e+00 -2.73464233e-01 8.57699096e-01 2.55891174e-01
1.23089373e-01 3.99293453e-01 8.25504720e-01 1.80724815e-01
2.49460116e-01 9.19051349e-01 -3.36508602e-02 2.11250737e-01
-4.66578841e-01 -4.22616720e-01 1.83030173e-01 1.28177702e+00
-1.86230272e-01 -3.42757851e-01 -7.70448148e-01 6.96886122e-01
-1.95111740e+00 -1.13618529e+00 1.13732696e-01 1.89619112e+00
1.28164053e+00 -9.26304609e-02 -2.47228548e-01 -1.30251721e-01
2.99804974e-02 1.79008976e-01 2.53125429e-01 -3.89806271e-01
-3.87329124e-02 1.01070905e+00 1.17875889e-01 4.17286754e-01
-9.32058275e-01 1.14791548e+00 7.18086815e+00 1.44138551e+00
-4.34247553e-01 -1.83186606e-02 2.95622557e-01 -9.32318792e-02
-5.32284319e-01 -3.60285267e-02 -1.14302611e+00 5.60087040e-02
1.20295620e+00 -2.42402077e-01 2.53295302e-01 7.54271865e-01
-3.72364491e-01 -6.17799014e-02 -1.43510818e+00 8.67069066e-01
1.18599981e-01 -1.72556806e+00 1.53862327e-01 4.02434543e-02
7.37968862e-01 2.30583444e-01 7.46957511e-02 9.00546312e-01
2.74739087e-01 -1.27564168e+00 3.48744154e-01 6.91819847e-01
7.40378320e-01 -5.40390611e-01 8.92300725e-01 4.05283689e-01
-1.31512427e+00 7.05223083e-02 -5.23262858e-01 9.88056585e-02
-1.06892943e-01 4.70129639e-01 -6.82205379e-01 1.00630927e+00
3.96367043e-01 1.05859959e+00 -5.94901443e-01 1.14942074e+00
-1.18094258e-01 3.80594552e-01 -2.02450618e-01 -3.64764035e-01
4.97115493e-01 4.51251119e-02 3.23637873e-01 1.61537063e+00
1.41629487e-01 -2.00773463e-01 8.22925791e-02 7.96643376e-01
-1.71152472e-01 3.09208095e-01 -7.77644455e-01 -3.55184972e-01
2.73243546e-01 1.03080022e+00 -1.07711121e-01 -4.70075846e-01
-2.88473815e-01 8.00200880e-01 4.79518145e-01 2.97798187e-01
-4.17478621e-01 -3.50350618e-01 4.03740734e-01 9.28843319e-02
4.35211599e-01 -2.34561294e-01 2.08327949e-01 -1.32570171e+00
-2.20395356e-01 -1.13581610e+00 7.26577222e-01 -6.95558131e-01
-1.51656520e+00 4.51157987e-01 2.74575353e-01 -9.74888325e-01
-6.79962218e-01 -4.29862082e-01 4.02931161e-02 8.14292967e-01
-2.02595472e+00 -1.18334961e+00 3.46598089e-01 7.07481682e-01
8.11064959e-01 -2.09696338e-01 1.17882419e+00 7.29979277e-01
-2.43540909e-02 7.02357650e-01 5.16883612e-01 9.87298042e-02
9.66171086e-01 -1.22698438e+00 -8.78136605e-02 4.07121420e-01
8.06041539e-01 1.24288130e+00 2.33903810e-01 -4.59593952e-01
-1.69889116e+00 -4.98453677e-01 1.31479955e+00 -7.14863896e-01
7.73758948e-01 -1.74689338e-01 -7.26343036e-01 4.35869902e-01
5.43129385e-01 -2.71749735e-01 5.49625099e-01 7.44361460e-01
-6.09489501e-01 -2.87319720e-03 -5.33141792e-01 5.57274103e-01
1.04154456e+00 -9.03459132e-01 -1.08474433e+00 5.68079829e-01
8.18383157e-01 -4.02540833e-01 -1.13171315e+00 5.54766655e-01
4.77130651e-01 -7.05339491e-01 1.45037198e+00 -7.13763595e-01
9.56444502e-01 2.40033701e-01 -1.68352649e-01 -1.14597416e+00
-3.43149304e-01 -4.77889657e-01 -4.42459971e-01 1.06958532e+00
4.34148163e-01 -7.04543851e-03 4.46154892e-01 1.85337439e-01
-2.36120582e-01 -8.63236308e-01 -8.66073072e-01 -7.95608878e-01
5.04635334e-01 -4.45719630e-01 2.22920179e-01 6.07229829e-01
4.17551726e-01 6.62177801e-01 -1.18340209e-01 -5.32828689e-01
-2.06302069e-02 4.83156353e-01 5.96827686e-01 -1.05840635e+00
-5.64434052e-01 -7.45825052e-01 -3.18454534e-01 -1.52377307e+00
3.76993477e-01 -1.21351361e+00 -8.31610244e-03 -1.59786940e+00
6.87848628e-01 -3.28989804e-01 -4.93565649e-01 2.47917652e-01
-2.86443561e-01 7.32123181e-02 3.75797637e-02 3.18385243e-01
-1.20545721e+00 6.85057580e-01 1.29342759e+00 -2.85172969e-01
1.62998930e-01 -2.65331626e-01 -3.96910578e-01 -3.08963843e-02
1.93284348e-01 -3.18204612e-01 -5.97179711e-01 -6.13495648e-01
7.00689912e-01 8.99891481e-02 2.34395340e-02 -8.31030309e-01
3.94332975e-01 3.52172464e-01 1.16113253e-01 -7.37675786e-01
4.45364922e-01 -7.14009345e-01 -4.35973853e-01 8.70673656e-02
-9.23601210e-01 4.45199572e-02 7.41252974e-02 7.13005483e-01
-7.54590213e-01 -5.44332385e-01 1.57199457e-01 -4.78574514e-01
-6.09046876e-01 3.40677708e-01 -3.08038086e-01 3.44317406e-01
2.54679620e-01 2.49791726e-01 -4.71424043e-01 -4.86091495e-01
-2.90193081e-01 7.92935118e-02 5.50145619e-02 4.67892677e-01
7.55887151e-01 -1.14423990e+00 -8.21774483e-01 -8.31466764e-02
3.74812663e-01 -1.82351023e-01 -3.69705223e-02 6.29471958e-01
-3.08589101e-01 1.20965135e+00 2.45527834e-01 -4.25777763e-01
-1.34169912e+00 6.15898788e-01 -6.83753705e-03 -1.12638915e+00
-6.64351046e-01 9.05519128e-01 -2.59633631e-01 -1.82826459e-01
4.73348856e-01 -3.20379555e-01 -2.57497907e-01 1.79548785e-01
7.45833337e-01 2.04930454e-01 3.93219173e-01 -1.64465323e-01
-7.42764249e-02 5.15159130e-01 -4.50943768e-01 -3.18247527e-01
1.53154969e+00 2.76921652e-02 -3.28915685e-01 1.32489175e-01
1.48202085e+00 -7.40218163e-02 -2.12776393e-01 -5.83012640e-01
4.78501052e-01 -4.95763093e-01 3.36309344e-01 -9.37744975e-01
-5.96559346e-01 1.08608997e+00 3.12557727e-01 1.82202548e-01
1.05016077e+00 1.62344202e-01 1.00034988e+00 9.57612038e-01
4.68953192e-01 -1.19610560e+00 5.14038086e-01 9.70560193e-01
7.38551617e-01 -1.08913732e+00 1.53513461e-01 -2.14170530e-01
-2.27343023e-01 1.19398451e+00 1.16039179e-01 -1.42774075e-01
7.70157397e-01 1.48272023e-01 -5.22154093e-01 -5.87041676e-01
-1.07511139e+00 -6.94609702e-01 8.27802420e-01 2.10681766e-01
1.04873610e+00 -4.06849414e-01 -6.52200103e-01 3.64429802e-01
1.96842756e-02 1.85601190e-01 -3.87014300e-02 1.00459850e+00
-4.38418537e-01 -1.56413245e+00 8.73931870e-02 3.50081503e-01
-6.53305411e-01 -6.45183384e-01 -5.36730468e-01 7.70168483e-01
-4.46996570e-01 8.90242755e-01 -1.90891460e-01 -2.26094902e-01
1.97344899e-01 4.13991153e-01 9.51604724e-01 -8.75931740e-01
-1.06998467e+00 1.17455207e-01 3.28831702e-01 -7.28631377e-01
-3.58276218e-01 -4.09149736e-01 -6.37214184e-01 -1.62624121e-01
-4.69571978e-01 3.03189218e-01 5.13453543e-01 9.61836278e-01
4.46960032e-01 5.19317508e-01 4.16313261e-01 -5.35727262e-01
-9.12741423e-01 -1.28805685e+00 -4.90641236e-01 5.55251837e-01
1.33494079e-01 -3.80093813e-01 -1.61483586e-01 -1.98643744e-01]
|
[11.467082977294922, 7.644372463226318]
|
3dd396dc-23bf-4d6a-8d28-c6a2edac9527
|
debiasing-multilingual-word-embeddings-a-case
|
2107.10181
| null |
https://arxiv.org/abs/2107.10181v2
|
https://arxiv.org/pdf/2107.10181v2.pdf
|
Debiasing Multilingual Word Embeddings: A Case Study of Three Indian Languages
|
In this paper, we advance the current state-of-the-art method for debiasing monolingual word embeddings so as to generalize well in a multilingual setting. We consider different methods to quantify bias and different debiasing approaches for monolingual as well as multilingual settings. We demonstrate the significance of our bias-mitigation approach on downstream NLP applications. Our proposed methods establish the state-of-the-art performance for debiasing multilingual embeddings for three Indian languages - Hindi, Bengali, and Telugu in addition to English. We believe that our work will open up new opportunities in building unbiased downstream NLP applications that are inherently dependent on the quality of the word embeddings used.
|
['Animesh Mukherjee', 'Ayush Suhane', 'Vishal Garimella', 'Srijan Bansal']
|
2021-07-21
| null | null | null | null |
['multilingual-word-embeddings']
|
['methodology']
|
[-4.82216299e-01 -2.08870173e-01 -4.23720270e-01 -3.92659038e-01
-1.30547202e+00 -1.22738898e+00 6.12574518e-01 3.38574350e-01
-7.93353677e-01 8.25429857e-01 6.76870942e-01 -8.63153040e-01
1.34096757e-01 -5.34811616e-01 -6.35135531e-01 -4.92117167e-01
2.42219567e-01 4.54491615e-01 -1.85765438e-02 -5.98554850e-01
3.48608613e-01 4.59662080e-01 -8.62645686e-01 -1.89662993e-01
1.17831159e+00 2.07261816e-01 -2.31275678e-01 4.63363856e-01
1.84099451e-01 3.25197220e-01 -4.74976808e-01 -7.94345558e-01
1.30541131e-01 4.02822420e-02 -8.58251095e-01 -8.17703366e-01
7.30169892e-01 -3.02162111e-01 -4.91236717e-01 1.31096554e+00
7.84565747e-01 -1.89508826e-01 6.75659180e-01 -1.00628352e+00
-1.51552129e+00 1.06967247e+00 -2.78694481e-01 5.66957355e-01
5.09069860e-02 1.29228204e-01 1.41239226e+00 -1.10613155e+00
8.11249137e-01 1.25919914e+00 8.46422911e-01 4.55478966e-01
-1.07090700e+00 -7.80241966e-01 2.88614511e-01 4.90635157e-01
-1.36802626e+00 -4.45652038e-01 3.10977280e-01 -4.10467267e-01
1.04899359e+00 -6.93780929e-02 2.14084134e-01 1.29721892e+00
3.93875122e-01 8.35749447e-01 1.28794229e+00 -5.06648302e-01
5.12400568e-02 3.11692744e-01 7.45259523e-01 3.20471197e-01
5.47909439e-01 2.29047164e-01 -6.05628371e-01 -2.57854789e-01
5.55751547e-02 -6.51131570e-01 -2.73206025e-01 -9.64537710e-02
-1.25319028e+00 1.20471168e+00 9.51773971e-02 5.48933327e-01
-7.72892758e-02 2.01646343e-01 6.39903665e-01 4.04947758e-01
1.01459324e+00 6.15465105e-01 -8.10948312e-01 -3.18565845e-01
-1.11548185e+00 2.15016469e-01 6.16020918e-01 8.40115070e-01
6.36926770e-01 -2.55139396e-02 -1.25945508e-01 9.05578375e-01
2.63612717e-01 7.71209478e-01 5.80117166e-01 -5.67566156e-01
5.75630784e-01 -8.52693990e-02 1.52018115e-01 -4.81564224e-01
5.19329980e-02 -1.34550735e-01 -2.20731124e-01 -3.12916517e-01
4.58852321e-01 -2.83780962e-01 -7.23815262e-01 1.91715515e+00
2.28917003e-01 -2.25558624e-01 2.81767964e-01 7.98969448e-01
2.74570227e-01 7.25788653e-01 1.39295980e-01 1.14918098e-01
1.48804283e+00 -9.91834104e-01 -7.73915052e-01 -2.81587839e-01
9.88539279e-01 -1.16336906e+00 1.48235035e+00 2.21754611e-01
-7.54719257e-01 -1.57904312e-01 -9.85969484e-01 -5.07099807e-01
-6.68151915e-01 -9.40866545e-02 4.57672864e-01 1.01810825e+00
-9.86865044e-01 3.77095342e-01 -6.46234989e-01 -5.58591723e-01
5.92425317e-02 -4.95795161e-02 -4.18678403e-01 -2.89612710e-01
-1.84795594e+00 1.48095930e+00 -1.44033064e-03 -8.57355166e-03
-9.75120068e-01 -1.23001158e+00 -9.11376834e-01 -1.94962949e-01
-2.80321360e-01 -7.81127959e-02 1.13113403e+00 -3.52389663e-01
-1.20628989e+00 9.67744827e-01 -1.28423125e-01 -1.87036455e-01
3.73424053e-01 -6.32972062e-01 -5.32598555e-01 -4.15669262e-01
4.54837471e-01 4.01435703e-01 4.85366821e-01 -8.27978134e-01
-5.46391129e-01 -4.94023710e-01 -1.74903005e-01 1.04808971e-01
-7.82184124e-01 3.17759871e-01 -1.02247894e-01 -9.05270934e-01
-4.77871925e-01 -1.02124250e+00 -4.64090891e-02 -3.44940424e-01
-1.89943984e-01 -3.22240829e-01 6.69715583e-01 -1.05661821e+00
1.42309344e+00 -2.19082379e+00 2.67551571e-01 -1.87427491e-01
-3.53826523e-01 4.35208499e-01 -3.38326275e-01 7.08425879e-01
-4.86604050e-02 4.68848974e-01 -4.54078689e-02 -3.61674070e-01
3.94157618e-01 2.00741425e-01 -7.33214080e-01 8.49589825e-01
9.25156921e-02 8.59680772e-01 -1.08638585e+00 -2.30584979e-01
1.41432896e-01 6.05564892e-01 -4.89211202e-01 9.04586762e-02
1.89873576e-01 2.55003244e-01 -1.95054186e-03 6.25554264e-01
7.88525760e-01 6.69802666e-01 2.00194657e-01 -1.89327374e-01
-5.31784713e-01 8.19216907e-01 -6.24455988e-01 1.62074506e+00
-7.20982730e-01 8.04901123e-01 4.41315733e-02 -5.65350413e-01
6.76717818e-01 2.40422159e-01 -2.38997415e-01 -3.67984891e-01
-1.72519162e-02 5.72631419e-01 -7.80596137e-02 -2.83261359e-01
9.27285671e-01 -4.09573108e-01 -5.76565981e-01 6.97176933e-01
2.02423707e-01 -4.10413653e-01 2.01583803e-01 2.05569148e-01
6.90281451e-01 -1.06407672e-01 3.34098995e-01 -9.52298939e-01
3.61910671e-01 -4.49492782e-02 4.09327656e-01 7.15297341e-01
-5.76242983e-01 3.14354628e-01 3.98115873e-01 -1.17761865e-01
-1.36587119e+00 -1.25541449e+00 -5.03360569e-01 1.51648986e+00
-1.47716686e-01 -1.47245243e-01 -6.31911814e-01 -1.11129141e+00
2.88999617e-01 1.11260557e+00 -4.94400173e-01 -7.92044178e-02
-7.97299802e-01 -1.09179580e+00 1.19338858e+00 4.14662629e-01
-1.81398720e-01 -7.90926218e-01 4.14447784e-01 9.21327248e-02
-2.48050451e-01 -9.11961854e-01 -1.05076087e+00 2.69962192e-01
-6.65334165e-01 -7.70636737e-01 -6.81987464e-01 -9.71807539e-01
2.52094865e-01 1.42463455e-02 8.24652314e-01 -5.48563540e-01
1.13072544e-01 2.21555725e-01 -3.40286136e-01 -2.06122383e-01
-6.47905707e-01 6.32684171e-01 4.17019516e-01 -5.03449500e-01
7.58443058e-01 -3.40805203e-01 -2.87453473e-01 -4.07989770e-02
-9.79929626e-01 -8.66101921e-01 2.76530981e-01 7.12098300e-01
6.98056221e-02 -4.41241890e-01 6.90467596e-01 -9.83314574e-01
1.10034847e+00 -7.63428569e-01 -5.41911244e-01 3.31867188e-01
-1.00424993e+00 4.66830879e-01 5.60823977e-01 -5.18976569e-01
-9.67389941e-01 -6.66263461e-01 -4.99001563e-01 1.25595734e-01
-6.23323917e-02 4.84421939e-01 -1.18392497e-01 -4.19541374e-02
6.02273762e-01 -7.83980563e-02 -5.63305080e-01 -7.38072217e-01
1.10529709e+00 1.13141251e+00 3.63476455e-01 -7.64675915e-01
7.10595191e-01 2.41136640e-01 -6.18949413e-01 -7.25197852e-01
-7.37406313e-01 -4.82494235e-01 -7.94561923e-01 2.00937107e-01
7.98729122e-01 -1.02509308e+00 -3.64209786e-02 7.52602756e-01
-1.32861769e+00 -2.09736973e-01 -1.08316801e-01 5.75717330e-01
-7.77861977e-04 6.89305723e-01 -1.08104098e+00 -3.84110123e-01
-5.40149927e-01 -1.20848966e+00 1.00577927e+00 -3.05218309e-01
-4.89040673e-01 -1.49653149e+00 8.33827913e-01 3.71210515e-01
4.80084330e-01 -5.58759689e-01 1.21517777e+00 -8.17710817e-01
-4.47113765e-03 4.90704253e-02 -2.21805915e-01 7.22602844e-01
1.26484573e-01 3.35970987e-03 -9.14958477e-01 -5.62072217e-01
-2.58595109e-01 -4.32178319e-01 8.70848536e-01 1.39601931e-01
3.52252215e-01 -2.82975078e-01 -7.31366798e-02 3.65360379e-01
1.38846183e+00 -2.70948321e-01 5.95181882e-01 5.54479063e-01
6.93366885e-01 5.14484823e-01 6.72644258e-01 4.23085690e-02
6.27768695e-01 3.96598607e-01 9.09620598e-02 1.46572456e-01
-2.41615564e-01 -4.15662616e-01 8.10229719e-01 1.54536974e+00
6.47840440e-01 -3.00347209e-01 -1.07077801e+00 1.42799592e+00
-1.49236643e+00 -5.45192778e-01 -2.03037068e-01 2.07278848e+00
1.36773837e+00 -3.28361630e-01 -2.77959853e-01 -1.21494599e-01
7.51438022e-01 4.41681951e-01 -1.79609507e-01 -9.59796965e-01
-8.88967738e-02 5.58467865e-01 9.81842279e-01 9.74156678e-01
-1.01565361e+00 1.56399953e+00 7.11373520e+00 8.60071301e-01
-1.08885837e+00 5.76531351e-01 2.44969308e-01 1.41775683e-01
-6.99470043e-01 1.22191660e-01 -1.28568530e+00 4.15396601e-01
1.05516005e+00 -2.52908230e-01 5.06574452e-01 7.67167330e-01
1.31282642e-01 2.43541330e-01 -1.11939275e+00 5.27815223e-01
2.19293743e-01 -9.80908394e-01 -3.67300585e-03 1.89874973e-02
8.86058629e-01 5.63822448e-01 3.80185395e-01 5.00470996e-01
7.29519367e-01 -9.31003749e-01 8.83250296e-01 -1.56196624e-01
8.43051910e-01 -9.04848695e-01 7.88595259e-01 4.89834435e-02
-7.93092847e-01 1.93867594e-01 -3.85523289e-01 8.55712295e-02
4.79833305e-01 8.86154592e-01 -6.44243360e-01 3.20861846e-01
6.57609463e-01 7.20075727e-01 -4.66677219e-01 4.31483537e-01
-7.47521937e-01 1.02633929e+00 -1.51583761e-01 -6.86633140e-02
4.48977083e-01 7.73244724e-03 5.26643276e-01 1.66251993e+00
4.17215079e-01 -5.90746999e-01 -3.38616103e-01 4.77878124e-01
-2.15726793e-01 3.89656693e-01 -6.46227479e-01 -2.18747869e-01
8.37223172e-01 9.68633294e-01 -1.02941930e-01 -1.94927454e-01
-3.13082308e-01 1.07661927e+00 7.59435534e-01 3.18777740e-01
-7.82523453e-01 -6.34696305e-01 1.28977513e+00 -2.45034322e-01
4.10112679e-01 -5.56880236e-01 -2.60045022e-01 -1.56359327e+00
-3.75195704e-02 -1.12517333e+00 3.63807321e-01 -3.65275025e-01
-1.92031670e+00 4.00099725e-01 7.74335675e-03 -6.03109658e-01
-2.20116116e-02 -7.87272513e-01 -2.73966193e-01 1.00807858e+00
-1.95199168e+00 -1.24407327e+00 3.85064691e-01 1.06743313e-01
4.42631006e-01 1.18024107e-02 7.99401879e-01 5.68911433e-01
-4.94600803e-01 7.78898478e-01 5.25117993e-01 2.12553024e-01
1.48529410e+00 -1.13632858e+00 8.04945707e-01 1.10454202e+00
2.09555656e-01 9.11699176e-01 6.97102964e-01 -6.16069019e-01
-1.16285169e+00 -1.08684874e+00 1.66969979e+00 -7.27395535e-01
1.46497715e+00 -4.32012975e-01 -7.22807348e-01 1.05635476e+00
5.95482111e-01 -2.34419748e-01 7.89903104e-01 6.72118425e-01
-8.30486119e-01 -3.38509865e-02 -1.09070766e+00 6.74724162e-01
6.77002430e-01 -9.63460863e-01 -1.00765419e+00 4.93707508e-02
8.02093029e-01 -3.64769585e-02 -9.80867386e-01 1.53348967e-01
5.19760251e-01 -4.69185442e-01 7.71825731e-01 -7.13346481e-01
1.75709605e-01 -5.50739802e-02 -4.09092158e-01 -1.83433664e+00
-2.88735688e-01 -4.36363548e-01 3.68944049e-01 1.53532946e+00
5.24986506e-01 -9.03682530e-01 3.15202057e-01 1.75844103e-01
-6.32155463e-02 -1.91234663e-01 -1.31044674e+00 -9.34628069e-01
1.20195675e+00 -5.81654191e-01 4.91926849e-01 1.45450759e+00
2.80974239e-01 4.84164804e-01 -3.88811111e-01 3.49872708e-01
6.42582119e-01 -3.57655406e-01 3.77324969e-01 -8.10447931e-01
1.25462502e-01 -1.99521348e-01 -1.85905397e-01 -1.09442115e+00
7.39423573e-01 -1.16957104e+00 5.57807907e-02 -1.16240907e+00
1.93585262e-01 -6.86695099e-01 -3.14818591e-01 3.92212152e-01
-4.56991583e-01 4.56001222e-01 2.91365236e-02 1.98796336e-02
-1.21086739e-01 4.67273831e-01 5.62482655e-01 -2.24111468e-01
1.03720412e-01 -6.71890616e-01 -8.99582207e-01 3.55106443e-01
8.58563840e-01 -9.07557428e-01 -1.18942067e-01 -9.34031963e-01
3.68128955e-01 -6.99422896e-01 8.38018768e-03 -3.00114632e-01
-6.41806126e-02 -9.13281366e-02 -3.36771607e-01 -2.60964960e-01
-1.66105285e-01 -4.08065468e-01 -4.16819960e-01 2.40961537e-01
-3.98269236e-01 3.21852863e-01 3.35371912e-01 2.14568853e-01
-1.44650012e-01 -4.28656161e-01 8.32244277e-01 6.87868297e-02
-6.14839137e-01 4.02614772e-02 -7.29589522e-01 6.80777490e-01
6.99056447e-01 5.55333495e-01 -7.43455470e-01 -1.80947080e-01
-4.15660679e-01 1.68426648e-01 5.48177183e-01 7.65482724e-01
1.63447589e-01 -1.40870380e+00 -1.01392293e+00 1.53459653e-01
2.26812273e-01 -9.00719345e-01 -1.30264103e-01 7.46799111e-01
-5.54461539e-01 6.82965636e-01 -3.10597382e-02 -9.07259881e-02
-1.11433542e+00 6.23789668e-01 1.68236494e-02 -4.39437896e-01
3.77106853e-02 7.80700028e-01 -1.91824175e-02 -1.05098355e+00
5.46479672e-02 -3.99236917e-01 2.23085865e-01 3.52764428e-01
2.63236523e-01 5.96933722e-01 2.77109146e-01 -7.37870097e-01
-6.20109141e-01 4.54632670e-01 -2.99817413e-01 -5.39294839e-01
1.11732709e+00 -2.79439718e-01 -4.70755100e-01 5.39960742e-01
1.36018944e+00 7.75829315e-01 -4.13195580e-01 -5.61360344e-02
1.80102602e-01 -2.92470217e-01 2.57828325e-01 -7.80277014e-01
-5.93392074e-01 1.03879988e+00 4.68018323e-01 -2.30564088e-01
5.31605065e-01 -6.53406093e-03 1.03690696e+00 1.00448318e-01
4.55613822e-01 -1.27682281e+00 -3.14778507e-01 8.67280602e-01
5.26289940e-01 -1.01134944e+00 -4.75229323e-02 2.57278122e-02
-6.34351194e-01 8.51801038e-01 1.39236078e-01 -5.69006503e-02
7.69045889e-01 3.71345371e-01 6.15033507e-01 1.26766950e-01
-5.12826979e-01 2.93967221e-02 3.50473970e-02 7.02987373e-01
8.27767253e-01 1.26817554e-01 -8.15250278e-01 5.00358820e-01
-3.39939713e-01 -3.27446103e-01 5.52518010e-01 7.11485147e-01
-2.07691461e-01 -1.64219010e+00 -4.10215616e-01 -8.62476975e-02
-7.81175613e-01 -9.43974376e-01 -3.09993058e-01 9.10684288e-01
3.13812532e-02 9.07986343e-01 -9.06007588e-02 -1.16550270e-02
1.93475217e-01 3.92000586e-01 3.92319947e-01 -5.66173375e-01
-7.30311692e-01 -2.78406352e-01 4.09456730e-01 -2.30061263e-01
-2.04593521e-02 -8.06742907e-01 -7.34195888e-01 -6.72869146e-01
-4.01981533e-01 3.37226331e-01 6.43835306e-01 7.57227957e-01
3.70984346e-01 -1.28066838e-01 3.33427668e-01 -6.11327171e-01
-8.71846080e-01 -1.22626626e+00 -4.68596995e-01 2.41311803e-01
4.46893930e-01 -2.44247496e-01 -8.04251015e-01 -1.54689893e-01]
|
[10.916136741638184, 9.999638557434082]
|
f9f07a7a-a1b7-4557-839e-4b5170a2415e
|
medical-image-segmentation-via-cascaded
| null | null |
https://openaccess.thecvf.com/content/WACV2023/html/Rahman_Medical_Image_Segmentation_via_Cascaded_Attention_Decoding_WACV_2023_paper.html
|
https://openaccess.thecvf.com/content/WACV2023/papers/Rahman_Medical_Image_Segmentation_via_Cascaded_Attention_Decoding_WACV_2023_paper.pdf
|
Medical Image Segmentation via Cascaded Attention Decoding
|
Transformers have shown great promise in medical image segmentation due to their ability to capture long-range dependencies through self-attention. However, they lack the ability to learn the local (contextual) relations among pixels. Previous works try to overcome this problem by embedding convolutional layers either in the encoder or decoder modules of transformers thus ending up sometimes with inconsistent features. To address this issue, we propose a novel attention-based decoder, namely CASCaded Attention DEcoder (CASCADE), which leverages the multiscale features of hierarchical vision transformers. CASCADE consists of i) an attention gate which fuses features with skip connections and ii) a convolutional attention module that enhances the long-range and local context by suppressing background information. We use a multi-stage feature and loss aggregation framework due to their faster convergence and better performance. Our experiments demonstrate that transformers with CASCADE significantly outperform state-of-the-art CNN- and transformer-based approaches, obtaining up to 5.07% and 6.16% improvements in DICE and mIoU scores, respectively. CASCADE opens new ways of designing better attention-based decoders.
|
['Radu Marculescu', 'Md Mostafijur Rahman']
|
2023-01-03
| null | null | null |
proceedings-of-the-ieee-cvf-winter-conference-3
|
['polyp-segmentation']
|
['computer-vision']
|
[ 2.78722197e-01 1.46872044e-01 3.79307084e-02 -4.87789482e-01
-8.97995472e-01 -1.93095282e-01 4.25842017e-01 2.04678103e-01
-5.50987542e-01 4.05464560e-01 3.99788290e-01 -1.87699229e-01
2.41001785e-01 -7.33832300e-01 -7.66171575e-01 -6.63280785e-01
2.27349758e-01 1.40002012e-01 5.99833846e-01 -1.20885663e-01
5.72413355e-02 1.63781598e-01 -1.30835164e+00 6.21216118e-01
1.06503522e+00 1.10409701e+00 2.72213399e-01 6.91986203e-01
-9.33723450e-02 1.26960278e+00 -3.00323188e-01 -7.35680997e-01
-5.95554709e-03 -5.61121821e-01 -7.31474459e-01 -1.26177305e-02
5.02889872e-01 -4.28652734e-01 -4.15389746e-01 9.04319465e-01
5.51676631e-01 -3.25989604e-01 3.86635423e-01 -7.90369809e-01
-8.49223435e-01 5.78219056e-01 -6.69109106e-01 3.83566946e-01
-3.88745517e-02 2.69619554e-01 1.35054648e+00 -7.69317091e-01
2.47776121e-01 1.20843410e+00 8.77506316e-01 4.84654069e-01
-1.17162371e+00 -3.86699289e-01 1.02981739e-01 4.16395336e-01
-1.12375295e+00 -1.85359299e-01 6.65412843e-01 -3.15151721e-01
1.19641662e+00 1.33162066e-01 5.95400512e-01 1.02296555e+00
5.14067471e-01 1.00470257e+00 1.01312220e+00 -2.37831771e-01
-7.31683075e-02 -6.64278939e-02 5.68463951e-02 9.33121502e-01
-4.18585017e-02 -1.70860156e-01 -3.02690476e-01 2.88552761e-01
9.20114934e-01 6.08738745e-03 -1.04098298e-01 4.83641289e-02
-1.18391883e+00 8.77550066e-01 1.03937364e+00 3.25241715e-01
-4.83597159e-01 3.58130723e-01 4.58204716e-01 1.14433013e-01
5.31431198e-01 3.83708358e-01 -3.50896418e-01 1.65473536e-01
-8.86029661e-01 -1.92777619e-01 4.39982980e-01 6.57986403e-01
6.38569295e-01 -2.97300994e-01 -6.58448398e-01 8.17923725e-01
1.83875486e-01 9.95441526e-02 4.91107136e-01 -8.17819595e-01
4.24764603e-01 8.38880002e-01 -3.59103620e-01 -8.12439024e-01
-4.19872522e-01 -8.20716977e-01 -9.19605076e-01 1.01254046e-01
2.74845958e-01 4.54655811e-02 -1.21976006e+00 1.68739128e+00
1.73025191e-01 2.90734231e-01 -1.55657336e-01 8.34132135e-01
8.48864615e-01 4.26989585e-01 2.61310607e-01 2.00217068e-01
1.43773472e+00 -1.50102711e+00 -6.73755765e-01 -2.54108816e-01
4.66701239e-01 -6.86111510e-01 1.16808748e+00 1.73283681e-01
-1.34036148e+00 -6.50851905e-01 -9.30382252e-01 -6.23925269e-01
-3.11440468e-01 1.48379385e-01 4.24612939e-01 4.42312747e-01
-1.17624688e+00 5.91459811e-01 -1.04267287e+00 -1.48342177e-01
9.84480202e-01 5.28191209e-01 -1.07723273e-01 -3.17196846e-02
-9.67521727e-01 8.73827338e-01 -1.64068237e-01 9.57020745e-02
-8.41747165e-01 -7.85810351e-01 -9.42709446e-01 3.04464698e-01
1.46161035e-01 -9.98923182e-01 1.18312919e+00 -1.23648632e+00
-1.51807666e+00 7.83037603e-01 -1.31135374e-01 -6.27106607e-01
6.67189062e-01 -3.89474452e-01 9.76904854e-02 2.67492086e-01
1.09359778e-01 9.61991608e-01 7.48195291e-01 -7.65574396e-01
-6.26826704e-01 -3.14269871e-01 2.12357312e-01 1.23571746e-01
-5.41601777e-01 -8.37360024e-02 -8.00462484e-01 -7.68440127e-01
-1.39845729e-01 -7.74508536e-01 -4.72933769e-01 2.86101401e-01
-6.10335290e-01 3.82467639e-03 6.89132631e-01 -7.18832672e-01
1.21794343e+00 -2.04742622e+00 1.28993914e-01 -2.05641001e-01
4.47129041e-01 4.43203956e-01 -2.18157262e-01 9.70287248e-02
1.43726423e-01 5.68883941e-02 -4.71543819e-01 -6.47288918e-01
-3.90031815e-01 3.36585909e-01 9.11263078e-02 2.48207763e-01
6.75364852e-01 1.21682560e+00 -1.00929308e+00 -6.04435861e-01
4.14210588e-01 8.63405526e-01 -8.43451977e-01 5.73350266e-02
-6.05767295e-02 6.25722885e-01 -4.27101851e-01 5.78671753e-01
5.04222214e-01 -5.76844275e-01 -5.87251745e-02 -3.24940205e-01
-7.67630637e-02 4.77223337e-01 -4.20508206e-01 1.79453242e+00
-6.18482172e-01 7.28153408e-01 9.22540855e-03 -1.05079198e+00
5.78666091e-01 2.60917008e-01 3.84942383e-01 -8.58555079e-01
2.18207136e-01 1.35572478e-01 1.22332036e-01 -4.89113569e-01
2.44327679e-01 -1.30569831e-01 1.51585698e-01 5.61622810e-03
2.10614488e-01 2.24818736e-01 1.75819043e-02 8.52230787e-02
1.31780040e+00 -5.40365949e-02 5.03511801e-02 -1.88205466e-01
7.20640957e-01 -4.61290538e-01 5.23613036e-01 7.11375296e-01
-2.12744251e-01 1.03749287e+00 6.83108628e-01 -3.80961359e-01
-1.00052786e+00 -9.94193077e-01 -1.87590113e-03 9.78445411e-01
8.90600160e-02 -4.24557239e-01 -8.53338480e-01 -8.75105619e-01
-2.04846337e-01 2.80477732e-01 -9.98876393e-01 -1.91938967e-01
-5.91498375e-01 -8.09518814e-01 4.99706984e-01 9.83118117e-01
7.80673265e-01 -9.71734464e-01 -6.74559534e-01 4.18717682e-01
-3.91383827e-01 -1.24718130e+00 -6.15952969e-01 2.38611177e-01
-9.73332703e-01 -8.53628695e-01 -9.66138899e-01 -6.62405372e-01
8.23193192e-01 3.03782728e-02 1.19158971e+00 2.04422995e-01
-3.63912523e-01 8.40697531e-03 -3.52486074e-01 -2.50765324e-01
-2.30172440e-01 3.85113120e-01 -7.13118494e-01 1.47889361e-01
1.24537572e-01 -5.97318769e-01 -1.01233339e+00 2.91952431e-01
-9.64255512e-01 2.83883393e-01 1.05197370e+00 1.16402590e+00
5.77097893e-01 -3.78176183e-01 3.72311890e-01 -9.23308909e-01
3.61453921e-01 -2.86024272e-01 -3.84587795e-01 3.22058558e-01
-3.23929995e-01 3.40434074e-01 5.71842790e-01 -1.70771778e-01
-9.51787829e-01 2.41287634e-01 -6.12265885e-01 -3.47956806e-01
5.16029671e-02 2.30707705e-01 -1.07789986e-01 -1.71910957e-01
3.42429608e-01 -1.67391337e-02 -9.21179056e-02 -5.22126138e-01
2.75483459e-01 5.73311687e-01 6.20341539e-01 -2.13996544e-01
3.61452460e-01 5.79880714e-01 -1.92167476e-01 -4.84451294e-01
-1.17242992e+00 -4.76770937e-01 -6.49245203e-01 1.33803869e-02
1.33008039e+00 -1.00925887e+00 -4.20047522e-01 5.86169600e-01
-1.09622347e+00 -3.45700294e-01 -3.37001741e-01 3.06162834e-01
-4.53066051e-01 2.01677516e-01 -1.11156023e+00 -3.16134721e-01
-5.88681042e-01 -1.49242413e+00 1.21568859e+00 2.71902472e-01
7.75970798e-03 -9.19782102e-01 -1.75523475e-01 5.21383822e-01
5.69499373e-01 2.43655160e-01 8.70390058e-01 -2.86441684e-01
-6.97849333e-01 9.62353423e-02 -5.74919045e-01 5.25480092e-01
1.44936278e-01 -6.74299747e-02 -1.17019331e+00 -1.71623468e-01
-2.17522457e-01 -1.82664067e-01 1.42723334e+00 6.33787632e-01
1.33633244e+00 -1.63861230e-01 -3.06895733e-01 7.68730760e-01
1.38655114e+00 -8.77736956e-02 9.54458773e-01 1.67969853e-01
8.59686852e-01 3.57708275e-01 3.07292305e-02 2.18250617e-01
5.30122459e-01 6.87958717e-01 6.47644162e-01 -6.32644057e-01
-5.56674898e-01 -1.34066597e-01 2.15012863e-01 8.42914820e-01
-5.42834252e-02 -2.18723699e-01 -7.60779321e-01 7.42869437e-01
-1.83974516e+00 -6.16626918e-01 -2.26943448e-01 1.89883578e+00
9.02007937e-01 3.31121802e-01 -5.42854890e-03 1.48414820e-01
5.06734788e-01 5.16195074e-02 -5.85046232e-01 -5.47315776e-01
-7.48033961e-03 3.22137028e-01 5.25716245e-01 5.23489416e-01
-1.21454799e+00 9.03807104e-01 5.90638304e+00 6.93417490e-01
-1.19379735e+00 2.44176269e-01 1.05441725e+00 1.40606374e-01
-2.10713208e-01 -3.10322583e-01 -5.75054348e-01 4.37221140e-01
7.33635008e-01 4.92390871e-01 -6.88023940e-02 4.61707890e-01
-4.52906825e-02 -1.94004215e-02 -9.10572171e-01 7.44886279e-01
-2.39170436e-02 -1.33246648e+00 1.46428064e-01 4.48891670e-02
8.54479551e-01 3.23941410e-01 2.63618529e-01 4.13214192e-02
2.28447944e-01 -1.08967376e+00 6.99367344e-01 4.37936634e-01
7.77446926e-01 -6.74869478e-01 1.00534201e+00 -7.18511268e-02
-1.15291417e+00 -1.74147367e-01 -1.74008161e-01 -6.33717552e-02
8.64039436e-02 7.33582497e-01 -6.37410045e-01 4.79577512e-01
8.22351635e-01 1.01114166e+00 -6.71870530e-01 1.19821703e+00
-4.39971954e-01 6.52293205e-01 -1.92460045e-01 2.34524861e-01
6.37799144e-01 2.26894245e-01 3.00036967e-01 1.45312870e+00
1.33045495e-01 -9.15089250e-02 -1.94025069e-01 9.14791703e-01
-2.91186064e-01 -1.09960355e-01 -1.96923539e-01 2.82999158e-01
-8.19561556e-02 1.26522410e+00 -8.17366242e-01 -3.78785461e-01
-8.05834353e-01 1.29319715e+00 3.80672365e-01 1.55365556e-01
-1.08965659e+00 -3.86521280e-01 7.70993590e-01 2.04831138e-01
8.39909077e-01 -5.39738033e-03 -6.44330859e-01 -1.05424690e+00
-1.02537088e-02 -6.30878866e-01 3.93427998e-01 -6.12967491e-01
-1.31547558e+00 7.77936518e-01 -4.57921088e-01 -9.59081948e-01
1.83109820e-01 -5.84591568e-01 -5.76612711e-01 6.62181675e-01
-1.87166429e+00 -1.44169235e+00 -3.67963433e-01 5.89880466e-01
6.76053703e-01 2.47989252e-01 6.38715208e-01 4.98680413e-01
-6.30218983e-01 7.98519492e-01 8.67843106e-02 2.58517802e-01
6.34793222e-01 -1.38370013e+00 5.49796700e-01 6.38568401e-01
2.03434825e-01 3.51313263e-01 2.94731736e-01 -2.29980215e-01
-9.14025784e-01 -1.24426901e+00 9.91864264e-01 -4.31429565e-01
4.94324595e-01 -3.33947986e-01 -9.64538157e-01 6.01846814e-01
3.99176747e-01 3.55298907e-01 4.81658787e-01 -5.27005829e-02
-4.42915469e-01 -2.69417435e-01 -1.03725517e+00 4.94668186e-01
9.32471514e-01 -4.91090626e-01 -3.77043456e-01 -2.87511088e-02
7.49839246e-01 -4.20605749e-01 -6.66661143e-01 4.02607739e-01
5.33113599e-01 -1.24821925e+00 1.03509331e+00 -1.85966909e-01
9.04520690e-01 -1.15509033e-01 1.86867237e-01 -1.05739236e+00
-5.11969805e-01 -4.67171133e-01 1.09146461e-01 1.08492684e+00
5.64639270e-01 -4.52318251e-01 6.89808786e-01 3.46301317e-01
-3.77156675e-01 -1.15065718e+00 -9.23244536e-01 -3.83533597e-01
2.07794622e-01 -2.53582269e-01 3.13993216e-01 5.88812232e-01
-1.75519705e-01 5.09060264e-01 -3.09868693e-01 -6.69003930e-03
4.23563123e-01 -3.52035314e-02 1.27182975e-01 -9.65005457e-01
-2.31007904e-01 -5.51730692e-01 -5.17205417e-01 -1.27790964e+00
-1.53437227e-01 -7.07999408e-01 7.34154880e-02 -1.61151803e+00
4.74515945e-01 -2.83359617e-01 -7.07419574e-01 6.85543001e-01
-4.12119776e-01 6.42313659e-01 6.80532008e-02 -1.87245328e-02
-8.66828859e-01 5.37860036e-01 1.44996274e+00 -2.00744078e-01
-5.65428287e-02 -2.13287175e-01 -8.28878701e-01 6.93304002e-01
6.89679384e-01 -4.30161834e-01 -2.35539913e-01 -1.01461792e+00
-6.85237572e-02 -2.37679586e-01 5.66852331e-01 -1.19009852e+00
3.29177439e-01 4.31341112e-01 4.91192281e-01 -4.32608068e-01
2.37585381e-01 -6.49523079e-01 -3.31837445e-01 5.64378262e-01
-5.55617273e-01 9.81217623e-02 2.60527611e-01 6.01546288e-01
-3.80833507e-01 1.56737626e-01 8.64242613e-01 -1.46820068e-01
-5.37419021e-01 3.54802340e-01 -4.10843343e-01 6.06092177e-02
8.10187697e-01 -2.07601860e-02 -1.48963392e-01 -3.55278373e-01
-7.08658636e-01 2.56807327e-01 3.24204654e-01 3.72474223e-01
6.13033116e-01 -1.07651162e+00 -7.62913346e-01 2.88721353e-01
-4.85771522e-02 8.33218694e-02 2.74596989e-01 1.21979201e+00
-5.26685417e-01 3.62033427e-01 -2.92141438e-01 -7.87859142e-01
-1.22500730e+00 2.98825681e-01 4.71001327e-01 -6.98060215e-01
-8.02774668e-01 1.24909425e+00 5.67109823e-01 -1.94956064e-02
1.91200629e-01 -8.27432871e-01 -8.91328305e-02 -1.12938493e-01
5.06890476e-01 1.03731371e-01 1.95979714e-01 -5.99618435e-01
-4.63868052e-01 7.03409195e-01 -3.05605173e-01 7.36658946e-02
1.35593855e+00 -1.14902996e-01 -3.65922675e-02 9.18277055e-02
1.35363352e+00 -3.09049278e-01 -1.63887584e+00 -2.48681024e-01
-9.22758952e-02 -2.93144405e-01 3.95919114e-01 -1.00122881e+00
-1.47529888e+00 1.29266059e+00 5.93317449e-01 -6.17760420e-02
1.46889102e+00 1.19508319e-01 1.21641600e+00 -9.22625586e-02
-1.43832684e-01 -6.68093920e-01 3.04435283e-01 5.03007889e-01
6.85648859e-01 -1.37137127e+00 -2.40991712e-01 -4.15386587e-01
-5.64191163e-01 1.01803219e+00 5.62846959e-01 -1.70830175e-01
5.86433291e-01 5.02460539e-01 8.63082111e-02 -1.29398867e-01
-8.55650187e-01 -4.21866238e-01 4.85109299e-01 4.43981498e-01
7.24272490e-01 -1.95753410e-01 -2.35791340e-01 5.26177049e-01
1.62879705e-01 3.12838778e-02 1.18814446e-01 6.74339652e-01
-2.11139634e-01 -1.07157397e+00 6.85818642e-02 4.65426266e-01
-8.85956347e-01 -3.60285103e-01 -2.79791445e-01 5.01066387e-01
3.08279335e-01 8.64354253e-01 2.33812928e-01 -3.41160983e-01
2.74630815e-01 -3.97894770e-01 4.82197940e-01 -4.09875512e-01
-9.73587632e-01 1.67879134e-01 -1.07455052e-01 -7.71903753e-01
-4.83211815e-01 -4.39762026e-01 -1.13223684e+00 2.96530183e-02
-3.82404387e-01 -1.19246557e-01 3.14593107e-01 9.18408036e-01
4.07288939e-01 9.87144232e-01 3.76932055e-01 -6.36847079e-01
-2.90990740e-01 -8.12135518e-01 -9.73891690e-02 3.83184880e-01
6.37195408e-01 -3.92377406e-01 -1.47604821e-02 8.18400085e-02]
|
[14.543108940124512, -2.5945799350738525]
|
2dc9c508-448b-4287-8db6-822ac17b4fa2
|
sctn-sparse-convolution-transformer-network
|
2105.04447
| null |
https://arxiv.org/abs/2105.04447v4
|
https://arxiv.org/pdf/2105.04447v4.pdf
|
SCTN: Sparse Convolution-Transformer Network for Scene Flow Estimation
|
We propose a novel scene flow estimation approach to capture and infer 3D motions from point clouds. Estimating 3D motions for point clouds is challenging, since a point cloud is unordered and its density is significantly non-uniform. Such unstructured data poses difficulties in matching corresponding points between point clouds, leading to inaccurate flow estimation. We propose a novel architecture named Sparse Convolution-Transformer Network (SCTN) that equips the sparse convolution with the transformer. Specifically, by leveraging the sparse convolution, SCTN transfers irregular point cloud into locally consistent flow features for estimating continuous and consistent motions within an object/local object part. We further propose to explicitly learn point relations using a point transformer module, different from exiting methods. We show that the learned relation-based contextual information is rich and helpful for matching corresponding points, benefiting scene flow estimation. In addition, a novel loss function is proposed to adaptively encourage flow consistency according to feature similarity. Extensive experiments demonstrate that our proposed approach achieves a new state of the art in scene flow estimation. Our approach achieves an error of 0.038 and 0.037 (EPE3D) on FlyingThings3D and KITTI Scene Flow respectively, which significantly outperforms previous methods by large margins.
|
['Bernard Ghanem', 'Silvio Giancola', 'Cheng Zheng', 'Bing Li']
|
2021-05-10
| null | null | null | null |
['scene-flow-estimation']
|
['computer-vision']
|
[-2.69121826e-01 -5.13208687e-01 -1.19438343e-01 -1.29649103e-01
-3.22160482e-01 -5.96045852e-01 4.68919456e-01 -1.41241342e-01
-1.23559557e-01 5.06480038e-01 2.20401421e-01 -2.90046316e-02
-1.60811409e-01 -8.04943800e-01 -1.00470567e+00 -4.38839614e-01
-2.11598620e-01 4.19144779e-01 4.52148288e-01 -4.32305038e-02
3.38009119e-01 9.88105476e-01 -1.40119052e+00 7.64301717e-02
1.02750981e+00 1.14566123e+00 3.00599545e-01 7.89055824e-01
-4.30344343e-01 9.80582476e-01 -3.13243330e-01 -1.33590698e-01
5.84889352e-01 -1.19816162e-01 -9.38451767e-01 3.06338251e-01
1.00283384e+00 -6.69789076e-01 -6.39131784e-01 7.90972769e-01
1.06953129e-01 4.14207339e-01 6.38910174e-01 -1.36353230e+00
-4.09945786e-01 -1.03263810e-01 -8.36163640e-01 3.11568528e-01
2.76482016e-01 4.22762096e-01 8.40024650e-01 -8.86757433e-01
5.53016305e-01 1.43799531e+00 7.45076716e-01 2.07278833e-01
-9.05266643e-01 -7.00879574e-01 4.40940231e-01 1.19324490e-01
-1.26457179e+00 -3.28413159e-01 9.38196659e-01 -6.83811665e-01
8.38147581e-01 8.42293128e-02 8.97779465e-01 6.05875790e-01
-1.90231651e-01 7.06720591e-01 3.42071414e-01 1.35411605e-01
1.02391064e-01 -2.83423305e-01 -2.47282401e-01 8.23223114e-01
2.44067118e-01 1.08578801e-01 -5.00053883e-01 -4.24669730e-03
1.42781579e+00 3.48605275e-01 -4.83700007e-01 -5.72727263e-01
-1.60741746e+00 6.55608892e-01 1.03348315e+00 3.62019688e-02
-4.40943509e-01 5.45697033e-01 3.61453086e-01 -5.37203588e-02
4.79469568e-01 1.01100393e-01 -3.35374266e-01 -2.12324440e-01
-8.21782351e-01 4.15532380e-01 6.40022635e-01 1.38544667e+00
1.04068756e+00 8.19421485e-02 -1.30280003e-01 4.83115971e-01
4.35204118e-01 6.87665343e-01 -2.58545205e-02 -1.49097848e+00
7.88635612e-01 6.97134316e-01 3.53580624e-01 -1.53951669e+00
-2.58234870e-02 -4.13200051e-01 -1.16079617e+00 9.23798531e-02
3.60882282e-01 6.08970262e-02 -5.69623768e-01 1.37151015e+00
7.06129134e-01 1.13613689e+00 -2.52517760e-01 1.23725462e+00
7.38174379e-01 9.44833815e-01 -1.23405322e-01 2.60298029e-02
9.55902159e-01 -1.11125588e+00 -3.83482784e-01 -6.46089464e-02
5.37844539e-01 -6.61245525e-01 8.87884200e-01 -4.18718345e-02
-1.12773418e+00 -6.17354155e-01 -7.25170612e-01 -2.63986915e-01
2.16616988e-01 -9.98787433e-02 8.61722648e-01 -5.06346375e-02
-8.30117106e-01 6.99799061e-01 -1.06730700e+00 -1.16217442e-01
8.93880904e-01 2.51048863e-01 -3.78541917e-01 -2.30282798e-01
-6.19793713e-01 3.32513869e-01 4.83899899e-02 3.43403488e-01
-6.54239476e-01 -1.33657897e+00 -1.02052927e+00 1.82804018e-01
6.65167421e-02 -1.32393074e+00 9.86819804e-01 -4.82395202e-01
-1.41570330e+00 4.87957686e-01 -4.42785829e-01 -3.18435580e-01
6.93140507e-01 -3.86808246e-01 6.47448823e-02 4.49324340e-01
3.12746942e-01 7.62731433e-01 7.40397692e-01 -1.23432875e+00
-7.87418842e-01 -2.23285198e-01 5.85913770e-02 3.13750118e-01
2.01194501e-03 -2.98756927e-01 -4.88295972e-01 -5.21012008e-01
1.48648560e-01 -7.41645992e-01 -3.69461924e-01 7.23945379e-01
-2.98816711e-01 -1.41808718e-01 1.10062146e+00 -1.96744218e-01
8.42941105e-01 -2.02624822e+00 2.79518198e-02 -4.57823090e-03
5.37404954e-01 1.80949450e-01 3.32607552e-02 1.21627338e-01
2.87524581e-01 -1.61428317e-01 -5.51294863e-01 -4.48831618e-01
-1.45929500e-01 2.97450870e-01 -5.85904539e-01 7.63848126e-01
5.18824220e-01 9.54641402e-01 -1.20430970e+00 -4.67765301e-01
8.58263195e-01 7.50659704e-01 -1.00008631e+00 3.33948433e-01
-6.50139302e-02 7.12602735e-01 -7.38688886e-01 7.06417918e-01
1.20440018e+00 -6.33742452e-01 -6.40016735e-01 -3.26767832e-01
-4.28002656e-01 1.30740941e-01 -1.28320706e+00 2.17281222e+00
-5.83223999e-01 6.26610160e-01 2.73991413e-02 -8.90466213e-01
8.54073226e-01 -3.34821232e-02 7.88763463e-01 -5.10245003e-02
-1.23965116e-02 1.64129734e-01 -3.67515534e-01 -3.57581466e-01
5.66512644e-01 3.42107490e-02 2.83732206e-01 -2.16863826e-02
-4.01352271e-02 -4.81362820e-01 -4.74778078e-02 2.71966368e-01
1.13865900e+00 2.61648655e-01 3.93601768e-02 -2.10104689e-01
7.80488133e-01 2.19153449e-01 7.16435790e-01 4.26516175e-01
-3.91475439e-01 7.49812245e-01 2.27348968e-01 -6.70494139e-01
-9.42316175e-01 -9.98301208e-01 -1.00445978e-01 2.53601700e-01
7.72452891e-01 -2.82203853e-01 -2.01475620e-01 -6.36079073e-01
3.21153998e-01 9.17454883e-02 -3.73968154e-01 2.67139282e-02
-9.57584083e-01 -1.09784633e-01 1.66734353e-01 7.02574432e-01
7.95637667e-01 -7.86735117e-01 -6.28089011e-01 2.06611827e-01
-2.30290905e-01 -1.47703815e+00 -8.54108632e-01 -3.93505841e-01
-1.28462684e+00 -1.07767367e+00 -8.26780379e-01 -7.18365312e-01
7.35220373e-01 7.69197047e-01 1.26046252e+00 1.15104780e-01
-5.04837558e-02 2.92198569e-01 -2.32176170e-01 9.81039554e-02
5.72889820e-02 1.77537296e-02 -8.13680589e-02 5.10389991e-02
2.38013431e-01 -8.79750907e-01 -1.02569187e+00 3.60352635e-01
-7.69017577e-01 3.08074374e-02 3.41861129e-01 6.85944915e-01
5.74787259e-01 -1.96814671e-01 4.04695347e-02 -4.98200387e-01
6.26940653e-02 -5.93867540e-01 -8.09442639e-01 -1.53592810e-01
-6.94840448e-03 -8.79726410e-02 7.07942486e-01 -3.19012254e-01
-1.05099595e+00 4.08130288e-01 -1.09996076e-03 -1.38873947e+00
-2.03288227e-01 -3.14916708e-02 7.03732446e-02 -4.89165604e-01
3.41127664e-01 1.23764522e-01 -1.46615535e-01 -2.86826909e-01
4.07242149e-01 1.00280426e-01 7.25075185e-01 -5.88080883e-01
1.20921075e+00 9.06607568e-01 3.39283913e-01 -7.36551046e-01
-6.84759796e-01 -8.65981996e-01 -8.36745858e-01 -3.11569482e-01
7.89304793e-01 -1.13884747e+00 -1.17393827e+00 4.73738343e-01
-1.49296153e+00 -2.38481656e-01 -3.80130053e-01 7.22380579e-01
-7.03969181e-01 4.70752120e-01 -6.90943301e-01 -4.83498842e-01
-2.67902315e-01 -1.20541203e+00 1.47609496e+00 1.38010100e-01
5.46995215e-02 -1.16367328e+00 1.14247315e-01 1.75939128e-01
3.38266671e-01 5.26648998e-01 2.81905472e-01 1.70539141e-01
-1.28400505e+00 1.32392094e-01 -5.67305624e-01 8.82730037e-02
3.15003663e-01 1.20765284e-01 -7.71015406e-01 -1.30914375e-01
1.32989127e-03 -1.45624638e-01 7.74304628e-01 5.68702877e-01
1.37153935e+00 -2.68700093e-01 -3.73028398e-01 1.23966968e+00
1.43902516e+00 -1.18249632e-01 4.00335789e-01 3.09454557e-02
1.33587277e+00 4.45255220e-01 6.77550077e-01 6.57700896e-01
4.99512374e-01 5.92865705e-01 6.77573979e-01 -3.32835056e-02
-1.66081071e-01 -4.90404278e-01 7.66164660e-02 9.56787109e-01
-8.62851366e-02 -6.54901788e-02 -7.73944855e-01 6.33407772e-01
-1.95976400e+00 -8.68436992e-01 -3.80845785e-01 1.94047844e+00
4.52070504e-01 -1.67075440e-01 -9.81701240e-02 -8.48017409e-02
6.46842778e-01 3.72611314e-01 -6.30397618e-01 1.74026623e-01
1.45233735e-01 7.86030069e-02 5.85940897e-01 6.83984220e-01
-1.08977711e+00 1.05023324e+00 5.26959944e+00 5.86670101e-01
-1.10881448e+00 -1.68316498e-01 3.53795648e-01 -1.07773013e-01
-3.83447081e-01 5.52468151e-02 -7.46189058e-01 6.12319887e-01
3.03711623e-01 -1.74381509e-01 1.45008862e-01 7.21905470e-01
3.82954627e-01 1.99575320e-01 -9.34179604e-01 1.29948521e+00
-2.08033964e-01 -1.86141181e+00 3.06876123e-01 6.04744628e-02
8.64759028e-01 1.08771063e-01 -1.68336406e-01 -1.32371085e-02
3.76013190e-01 -7.59752274e-01 6.81626379e-01 4.98688221e-01
6.47862434e-01 -7.78227448e-01 3.20241213e-01 2.93830812e-01
-1.67488456e+00 2.53241032e-01 -6.23366296e-01 -1.10454790e-01
5.57144344e-01 6.76819444e-01 -5.43469429e-01 6.96576118e-01
8.23426902e-01 1.64775074e+00 -4.67238165e-02 1.33650267e+00
1.28411472e-01 3.45562935e-01 -4.78361696e-01 3.06441545e-01
5.43229401e-01 -4.16265547e-01 8.65711331e-01 9.98694658e-01
5.00831127e-01 2.75793582e-01 2.77584583e-01 1.24341965e+00
-1.94047064e-01 -6.16213121e-02 -7.90386081e-01 3.77896070e-01
5.28281927e-01 1.21227956e+00 -5.99401236e-01 -3.25097859e-01
-4.25372124e-01 9.86677408e-01 3.38591486e-01 3.68978143e-01
-7.42099226e-01 -3.34372669e-01 1.18822002e+00 9.44602937e-02
4.04346883e-01 -4.98236299e-01 -1.01902083e-01 -1.51856875e+00
1.37856469e-01 -1.09244451e-01 4.43978459e-02 -7.33858168e-01
-1.41274989e+00 5.33111334e-01 -1.00246489e-01 -1.89061701e+00
-4.67484519e-02 -4.46645081e-01 -7.44966030e-01 9.05981898e-01
-1.87194502e+00 -9.96942222e-01 -9.57506657e-01 8.28154981e-01
6.31231904e-01 2.23645940e-01 1.27756447e-01 4.42785919e-01
-3.88518184e-01 1.31968573e-01 -3.00851971e-01 1.20588779e-01
4.74388629e-01 -1.10089016e+00 7.18157887e-01 8.54786754e-01
4.43601906e-02 4.90337253e-01 2.94258624e-01 -5.63365579e-01
-1.51145315e+00 -1.52419960e+00 5.42816579e-01 -4.90049392e-01
5.53440154e-01 -2.09356755e-01 -1.09655106e+00 6.50946259e-01
-8.83606523e-02 8.65326405e-01 1.30648941e-01 -3.96917552e-01
-1.71963915e-01 -1.95154712e-01 -1.06797278e+00 3.77036542e-01
1.56169534e+00 -3.40653688e-01 -2.80164033e-01 3.53117287e-01
1.18968940e+00 -7.34905481e-01 -9.20008361e-01 5.77399552e-01
1.55120328e-01 -9.50130761e-01 1.29788864e+00 -4.02692139e-01
6.65941119e-01 -6.82055354e-01 -1.40981404e-02 -1.16598761e+00
-3.88021976e-01 -7.30868876e-01 -5.34052014e-01 8.71586502e-01
-2.11455449e-01 -4.47781146e-01 1.10631108e+00 4.56123710e-01
-4.67870712e-01 -5.59238434e-01 -7.69444406e-01 -7.81730413e-01
1.78674795e-02 -4.84143376e-01 7.59789169e-01 1.20097005e+00
-5.13650119e-01 1.40734419e-01 -3.08529168e-01 4.06036198e-01
8.23563635e-01 3.20958495e-01 1.05082333e+00 -1.13413358e+00
-3.33793685e-02 -4.75441247e-01 -7.84116685e-01 -1.97634590e+00
4.73723024e-01 -8.30129504e-01 4.16232878e-03 -1.44730377e+00
-3.09347183e-01 -7.91219115e-01 1.18409075e-01 5.50396070e-02
-2.39565700e-01 5.68918921e-02 2.53851056e-01 4.53258246e-01
-5.55244207e-01 9.62620795e-01 1.73860836e+00 -6.00976683e-02
-3.46228927e-01 1.55091882e-01 -2.06076607e-01 6.38316989e-01
4.48806196e-01 -2.01416358e-01 -6.11081064e-01 -8.36238921e-01
-2.19964072e-01 2.57717431e-01 7.43643820e-01 -1.23108983e+00
4.92469817e-01 -3.77833277e-01 3.98352087e-01 -1.00674355e+00
3.92177254e-01 -1.05481100e+00 7.68897980e-02 3.27788949e-01
-8.84438455e-02 1.18731111e-01 2.75068611e-01 9.70674813e-01
-4.55710411e-01 2.45214432e-01 5.80411196e-01 -2.57038832e-01
-8.08090746e-01 1.14491475e+00 2.46962652e-01 2.05284879e-01
9.74226713e-01 -2.48977199e-01 -1.23631217e-01 -2.17653140e-01
-2.15656653e-01 4.16296750e-01 4.86344516e-01 3.71047556e-01
8.49854708e-01 -1.62087286e+00 -6.31753623e-01 3.41421068e-01
1.16238773e-01 1.03421235e+00 3.81572515e-01 8.20245087e-01
-1.06550074e+00 3.81522864e-01 -1.02475487e-01 -1.27095127e+00
-7.72801757e-01 3.20236266e-01 3.13664645e-01 1.51865155e-01
-1.09965956e+00 9.32672501e-01 5.79802215e-01 -4.11658853e-01
9.45628956e-02 -8.07003200e-01 9.33418721e-02 -4.76501256e-01
4.71401930e-01 5.58844984e-01 -1.55286059e-01 -7.66501486e-01
-4.40968245e-01 1.03554535e+00 1.91085100e-01 2.65758246e-01
1.21988249e+00 -7.24187568e-02 -8.03349242e-02 2.63577282e-01
1.54048216e+00 -7.70218074e-02 -1.82986712e+00 -2.80249476e-01
-3.92560184e-01 -1.21758139e+00 8.76293331e-02 7.37273544e-02
-1.45801723e+00 1.00940883e+00 9.68235806e-02 -2.14565769e-01
1.01178837e+00 -4.91507165e-02 1.14338064e+00 3.13681424e-01
3.61938804e-01 -1.97090760e-01 -3.36987078e-02 6.39170349e-01
7.26758838e-01 -1.35617459e+00 -1.06343493e-01 -8.89866829e-01
-3.87080580e-01 1.07688868e+00 7.71737993e-01 -6.42534733e-01
7.61804521e-01 7.40036666e-02 -2.73253798e-01 -2.42936179e-01
-6.89064503e-01 -2.71257199e-02 3.56738865e-01 5.00912964e-01
5.97549751e-02 -2.05204129e-01 3.75786662e-01 -1.93353891e-01
-5.78572340e-02 2.45796457e-01 2.63578326e-01 8.53103459e-01
-2.92663634e-01 -6.41491592e-01 -2.64806002e-01 2.97461897e-01
8.37995857e-02 2.91855466e-02 7.25560561e-02 6.94191635e-01
-1.05106905e-01 5.58215797e-01 5.74777842e-01 -2.86874771e-01
4.72850055e-01 -6.70298696e-01 3.40909183e-01 -3.73022020e-01
-3.81131351e-01 -1.11179702e-01 -4.55153137e-01 -1.01739264e+00
-8.18873227e-01 -5.86480021e-01 -1.23639011e+00 -6.60835803e-01
2.16564140e-03 1.41845942e-01 2.32828110e-01 6.12093627e-01
6.11350596e-01 3.74845594e-01 8.83863330e-01 -1.14175856e+00
-1.21719621e-01 -4.81443524e-01 -3.88368338e-01 5.40317953e-01
7.36250460e-01 -8.30978096e-01 -6.39112651e-01 5.90546839e-02]
|
[8.551650047302246, -2.078713893890381]
|
bc360c43-7ee6-4a04-94d5-58225dcace46
|
everybody-sign-now-translating-spoken
|
2011.09846
| null |
https://arxiv.org/abs/2011.09846v4
|
https://arxiv.org/pdf/2011.09846v4.pdf
|
Everybody Sign Now: Translating Spoken Language to Photo Realistic Sign Language Video
|
To be truly understandable and accepted by Deaf communities, an automatic Sign Language Production (SLP) system must generate a photo-realistic signer. Prior approaches based on graphical avatars have proven unpopular, whereas recent neural SLP works that produce skeleton pose sequences have been shown to be not understandable to Deaf viewers. In this paper, we propose SignGAN, the first SLP model to produce photo-realistic continuous sign language videos directly from spoken language. We employ a transformer architecture with a Mixture Density Network (MDN) formulation to handle the translation from spoken language to skeletal pose. A pose-conditioned human synthesis model is then introduced to generate a photo-realistic sign language video from the skeletal pose sequence. This allows the photo-realistic production of sign videos directly translated from written text. We further propose a novel keypoint-based loss function, which significantly improves the quality of synthesized hand images, operating in the keypoint space to avoid issues caused by motion blur. In addition, we introduce a method for controllable video generation, enabling training on large, diverse sign language datasets and providing the ability to control the signer appearance at inference. Using a dataset of eight different sign language interpreters extracted from broadcast footage, we show that SignGAN significantly outperforms all baseline methods for quantitative metrics and human perceptual studies.
|
['Richard Bowden', 'Necati Cihan Camgoz', 'Ben Saunders']
|
2020-11-19
| null | null | null | null |
['sign-language-production']
|
['natural-language-processing']
|
[ 4.04203355e-01 2.17564702e-01 5.87208942e-02 -3.13016355e-01
-1.04852641e+00 -5.92790782e-01 7.76495159e-01 -1.18509531e+00
-1.86189383e-01 6.60878837e-01 6.12448215e-01 -4.93875444e-02
2.49790609e-01 -4.41061825e-01 -1.03223324e+00 -6.81594312e-01
2.74237931e-01 4.81586456e-01 8.42492878e-02 -1.24422237e-01
-6.13198243e-03 4.55946803e-01 -1.79648721e+00 3.38536382e-01
9.10506725e-01 5.39892554e-01 2.84228921e-01 1.13786209e+00
2.23314762e-01 1.04154348e+00 -7.84684777e-01 -3.59731317e-01
4.55590010e-01 -9.72783566e-01 -4.60628808e-01 1.57777458e-01
1.25658858e+00 -9.26275671e-01 -5.33175528e-01 7.68062115e-01
1.00110447e+00 -1.21188439e-01 7.15347469e-01 -1.46369159e+00
-7.75258660e-01 5.49460948e-01 -2.36715779e-01 -7.62460232e-01
6.64552331e-01 6.99606776e-01 9.23581362e-01 -7.73634672e-01
1.24620140e+00 1.56832600e+00 4.01707143e-01 1.06850398e+00
-1.25640690e+00 -8.00914586e-01 -2.84272265e-02 1.83373377e-01
-1.18791342e+00 -5.92215478e-01 6.58242583e-01 -3.18203658e-01
4.97754335e-01 2.57069916e-01 1.09201908e+00 1.62998068e+00
-4.36030835e-01 1.25791252e+00 1.10690451e+00 -6.01970553e-01
8.54851976e-02 -4.40278322e-01 -5.99091470e-01 6.77086830e-01
-6.77592307e-02 3.49476695e-01 -9.35033083e-01 7.08035007e-02
1.13195896e+00 -6.23324752e-01 -8.73676002e-01 -6.21841967e-01
-1.31688058e+00 5.38574398e-01 1.52866185e-01 -1.44691020e-01
-4.18224156e-01 6.50557697e-01 -7.29772775e-03 3.27230692e-01
-1.93604589e-01 1.72559768e-01 -1.24368578e-01 -4.02984202e-01
-9.63066697e-01 6.03738606e-01 8.11050057e-01 1.02538037e+00
-2.40467802e-01 4.30754781e-01 -4.75917041e-01 7.42061794e-01
7.23825574e-01 1.11004853e+00 2.69696146e-01 -1.29633629e+00
4.41512883e-01 1.50591373e-01 1.92633331e-01 -4.44434851e-01
3.66399288e-02 -1.51776344e-01 -4.85981494e-01 9.90564406e-01
8.49222541e-01 -1.70474395e-01 -1.61823142e+00 1.87135053e+00
1.16711058e-01 1.22686334e-01 1.18949227e-01 1.34309220e+00
7.08962798e-01 6.03626668e-01 5.24809994e-02 2.33327255e-01
1.07703614e+00 -9.98822153e-01 -6.07023716e-01 6.23773411e-02
2.41479874e-01 -6.81954741e-01 1.43204296e+00 6.47723734e-01
-1.32845879e+00 -2.02651665e-01 -8.42505395e-01 -2.47104883e-01
4.08277541e-01 1.32831022e-01 2.55824506e-01 5.07777870e-01
-1.22432148e+00 1.71471015e-01 -9.18217182e-01 -3.35332543e-01
3.33470851e-01 7.28682280e-02 -5.10953665e-01 -2.43577436e-01
-8.48995268e-01 9.14478540e-01 6.34322762e-02 1.08006448e-01
-1.02206254e+00 -7.33812690e-01 -9.89959359e-01 -3.31788361e-01
-9.17921141e-02 -1.07008767e+00 1.50829411e+00 -1.25166118e+00
-2.07790732e+00 9.06190872e-01 -9.94707420e-02 -3.32028270e-01
1.35688031e+00 -2.24108949e-01 -6.87928498e-02 3.65293831e-01
-3.01097572e-01 1.12203777e+00 1.14808440e+00 -1.63860583e+00
-4.88901705e-01 1.10213272e-01 -2.29176357e-01 4.43207979e-01
2.06079930e-01 1.41302854e-01 -6.43855989e-01 -8.87102425e-01
1.76192373e-01 -9.57015812e-01 1.36807978e-01 7.04029620e-01
-4.01522666e-01 2.36329928e-01 8.42079699e-01 -1.26817465e+00
6.20959640e-01 -1.86176240e+00 5.10081410e-01 2.35211626e-01
5.86618818e-02 3.73642325e-01 -5.49764991e-01 3.22848022e-01
2.72454351e-01 -3.50166351e-01 -4.54542547e-01 -4.28308696e-01
2.83156902e-01 3.28726828e-01 -4.01506484e-01 3.42397302e-01
-7.62979267e-03 1.03022671e+00 -9.15506065e-01 -4.12966102e-01
2.26612836e-01 1.09193373e+00 -6.95392132e-01 2.65930116e-01
-4.43099618e-01 8.21849406e-01 1.12255365e-01 8.49573851e-01
4.70619172e-01 1.51346967e-01 2.20656738e-01 -8.47944245e-02
1.47801459e-01 -2.10228059e-02 -1.08357322e+00 1.78952813e+00
-6.26674354e-01 9.27541554e-01 3.12227100e-01 -2.48825982e-01
6.50556684e-01 5.10355890e-01 9.61939320e-02 -6.55064821e-01
1.07373819e-01 4.99455005e-01 6.51279092e-02 -5.36291122e-01
1.85003072e-01 -2.39471942e-01 2.50085235e-01 3.60772371e-01
-1.90835018e-02 -7.06613898e-01 1.22534856e-01 -7.52494857e-02
9.48946834e-01 6.24483347e-01 -3.09887409e-01 3.76909405e-01
1.69933900e-01 -2.52062887e-01 1.77171960e-01 6.79301620e-01
1.98820345e-02 1.25155246e+00 2.55145103e-01 -6.09801663e-03
-1.27122748e+00 -1.57395697e+00 2.98855513e-01 6.89581335e-01
-5.63651212e-02 7.35497028e-02 -8.14962029e-01 -3.84593189e-01
-1.03301771e-01 8.85552824e-01 -1.02961577e-01 2.48581380e-01
-8.97926748e-01 4.62514050e-02 8.86443555e-01 6.39337361e-01
5.07009268e-01 -1.36814141e+00 -6.70201182e-01 4.70426343e-02
-3.92720670e-01 -1.12913251e+00 -8.54240119e-01 -6.49611890e-01
-3.51798803e-01 -8.86992097e-01 -1.65765655e+00 -1.15080214e+00
8.20250213e-01 -3.30022663e-01 7.17246950e-01 -2.07924321e-01
-2.91148216e-01 5.00771761e-01 -3.97739410e-01 -1.96327940e-01
-9.65044081e-01 -4.94480282e-01 1.21942259e-01 8.95878300e-02
-1.93033502e-01 -7.48321772e-01 -6.53598547e-01 2.35418379e-01
-9.61486042e-01 6.48872316e-01 6.10712886e-01 9.63185132e-01
3.02898347e-01 -8.84491742e-01 1.74659252e-01 2.75922045e-02
5.31481326e-01 3.70521724e-01 -6.98165059e-01 2.77788311e-01
-1.69348419e-01 2.33150408e-01 3.47953945e-01 -8.01932573e-01
-1.04268825e+00 2.11107612e-01 -2.19782799e-01 -4.78300065e-01
5.51276356e-02 -1.22015432e-01 -4.00022388e-01 -2.01359630e-01
6.16878033e-01 5.52752197e-01 3.41883719e-01 -3.28007191e-01
7.75004327e-01 8.64737928e-01 1.06405854e+00 -6.00040495e-01
9.52755153e-01 5.93998253e-01 -9.32520777e-02 -9.51662004e-01
-2.95421239e-02 1.26995951e-01 -3.59699100e-01 -5.06096363e-01
8.09803426e-01 -9.25637782e-01 -8.79361272e-01 1.15347219e+00
-1.23023665e+00 -9.07841861e-01 -2.85891652e-01 6.66753471e-01
-1.02887368e+00 5.11009932e-01 -5.66186488e-01 -7.86434352e-01
-2.28057966e-01 -1.18201268e+00 1.40774679e+00 1.52154654e-01
-4.59196180e-01 -2.90184498e-01 1.16651900e-01 4.89345908e-01
2.81413972e-01 3.70988637e-01 4.72353250e-01 2.69789070e-01
-9.83853340e-01 -1.29104882e-01 -2.08103195e-01 6.41365051e-01
-8.03817734e-02 1.23648576e-01 -8.66960466e-01 -3.58477265e-01
-8.00642669e-01 -5.03576577e-01 7.57860422e-01 5.11918426e-01
4.36408162e-01 -4.40713197e-01 1.20354056e-01 8.22297931e-01
9.48917747e-01 1.18824370e-01 8.63589287e-01 -8.93922225e-02
8.32135379e-01 5.20429611e-01 3.00034672e-01 1.73201278e-01
4.92104232e-01 8.87164474e-01 5.33430278e-02 -1.68697521e-01
-1.06514740e+00 -9.84855771e-01 7.89683998e-01 5.20161629e-01
-4.91092265e-01 -4.39911008e-01 -8.04296911e-01 5.88521957e-01
-1.66829336e+00 -1.00281596e+00 6.71170279e-02 1.96536839e+00
9.66714978e-01 -2.55966961e-01 2.15241611e-01 1.52796462e-01
5.23952007e-01 -1.12449519e-01 -4.93757635e-01 -1.12691574e-01
-5.05929470e-01 4.11358088e-01 4.17946100e-01 8.11289251e-01
-6.58766031e-01 1.14486432e+00 5.81833839e+00 4.37024981e-01
-1.30132449e+00 -1.58341214e-01 -9.91416201e-02 -3.73808235e-01
-4.73302841e-01 -1.08711272e-01 -4.29165483e-01 3.40788275e-01
3.85298818e-01 2.54269340e-03 3.79505485e-01 6.20497882e-01
7.21242666e-01 8.20140615e-02 -1.00375354e+00 1.23514187e+00
3.62647086e-01 -1.11738992e+00 4.21540380e-01 4.07764167e-02
8.92561615e-01 -4.94856574e-02 1.03633940e-01 -3.64464931e-02
5.54444015e-01 -1.08749020e+00 1.37351763e+00 8.20633769e-01
1.33388019e+00 -3.55236739e-01 1.67772651e-01 1.57782659e-01
-9.01512384e-01 7.66769722e-02 4.40216303e-01 1.91095456e-01
9.14211273e-01 -2.02385977e-01 -9.62662578e-01 -1.33091095e-03
3.39481741e-01 3.06058079e-01 -1.01972148e-01 1.22139049e+00
-8.07348907e-01 6.03203475e-01 -4.76835459e-01 -6.64491653e-02
-3.23501229e-02 1.08740874e-01 9.27794576e-01 9.90946233e-01
5.77681422e-01 -1.33381411e-01 -1.23661965e-01 8.32033038e-01
1.09358177e-01 6.32401779e-02 -5.94592214e-01 -9.47221816e-02
2.43163377e-01 4.29621309e-01 -2.39983603e-01 -2.26272553e-01
-8.85172188e-02 1.58241379e+00 -2.53941596e-01 6.64343596e-01
-6.64682031e-01 -3.94124061e-01 8.06058526e-01 2.74018645e-01
3.26165438e-01 -4.60247457e-01 -2.55754460e-02 -1.29408228e+00
4.49894488e-01 -1.17172945e+00 -1.97999910e-01 -1.21560907e+00
-9.11378443e-01 5.04999280e-01 -1.87544391e-01 -1.47948503e+00
-6.92069471e-01 -7.22191453e-01 -3.64282817e-01 8.26927483e-01
-1.27240849e+00 -1.70617294e+00 -3.70305657e-01 5.35839677e-01
4.68926102e-01 -3.24424468e-02 6.96609318e-01 4.03209955e-01
3.01271737e-01 9.25930321e-01 -1.86382815e-01 3.14798266e-01
7.20784545e-01 -1.00634181e+00 5.69564641e-01 9.42726314e-01
3.06179732e-01 2.49458313e-01 9.00233448e-01 -5.29236794e-01
-1.20403147e+00 -7.18348682e-01 8.05902839e-01 -4.67702955e-01
2.27523193e-01 -3.26402575e-01 -3.95623296e-01 5.45181811e-01
-1.62175819e-02 -2.16214657e-01 -1.57464612e-02 -7.35337198e-01
-4.95065242e-01 1.39632598e-01 -1.07366848e+00 1.09712815e+00
1.52113414e+00 -5.49091578e-01 -4.59554851e-01 1.35913923e-01
4.84891295e-01 -6.42722547e-01 -3.37635070e-01 3.47525775e-01
1.26519573e+00 -7.48332798e-01 9.21070516e-01 -5.30534506e-01
4.33155060e-01 -6.06164634e-01 -1.83314860e-01 -1.26004326e+00
3.86738122e-01 -1.01964986e+00 -2.44931161e-01 1.02966464e+00
3.71857762e-01 -3.77136528e-01 8.37061405e-01 5.26469290e-01
-3.88566703e-02 -2.12958857e-01 -9.20445740e-01 -1.03512609e+00
-1.45297330e-02 -5.07955313e-01 3.12299252e-01 4.33975518e-01
-1.47272050e-01 -8.57976899e-02 -9.24928069e-01 8.96521509e-02
9.22494948e-01 -3.30139659e-02 1.21853411e+00 -7.72312939e-01
-5.27940989e-01 -4.99265522e-01 -6.58806920e-01 -1.47604859e+00
1.08814284e-01 -7.59491146e-01 5.07865071e-01 -1.92057312e+00
-2.15694666e-01 9.42889675e-02 5.48518240e-01 3.17720115e-01
2.72564560e-01 4.75651562e-01 5.86215317e-01 1.12230219e-01
7.93917477e-02 6.10307813e-01 1.97698426e+00 -1.01990312e-01
-3.03177208e-01 4.75978218e-02 -1.56822518e-01 7.30731249e-01
4.41527933e-01 -6.23265505e-02 -4.02100950e-01 -6.35535538e-01
-1.07724726e-01 2.23210946e-01 8.53744805e-01 -1.01810026e+00
5.02656288e-02 2.34836973e-02 9.04262140e-02 -3.80429894e-01
6.44596934e-01 -4.81455982e-01 2.38797501e-01 6.48043394e-01
-3.54649454e-01 -3.76417458e-01 -1.78930238e-01 2.99939454e-01
-1.73084453e-01 2.64879525e-01 9.40877914e-01 6.55145049e-02
-6.94590271e-01 2.22431809e-01 -3.17191243e-01 8.52504969e-02
7.04209864e-01 -3.96271616e-01 7.32444227e-02 -1.14485526e+00
-7.21459925e-01 1.56442210e-01 6.02335334e-01 5.04855156e-01
8.59851480e-01 -1.46424842e+00 -1.16136336e+00 3.53551626e-01
1.12681650e-01 -4.84085418e-02 2.95785964e-01 4.93486136e-01
-1.24717271e+00 3.13073099e-02 -2.66296595e-01 -6.24979377e-01
-1.54232788e+00 -1.79704770e-01 3.72869343e-01 3.45339000e-01
-1.16224229e+00 1.02110863e+00 -3.96464616e-02 -5.50339222e-01
6.28230631e-01 -5.33962011e-01 4.49335188e-01 -3.84075910e-01
6.20313168e-01 1.40937150e-01 -5.66697717e-01 -8.31960797e-01
-7.60113671e-02 7.46329367e-01 4.26848203e-01 -9.95140553e-01
1.09782350e+00 1.21354342e-01 4.20584023e-01 3.60584515e-03
8.88289988e-01 2.42048785e-01 -1.76686966e+00 1.58030575e-03
-4.46063936e-01 -5.50676286e-01 -3.42146784e-01 -1.27535474e+00
-8.62540424e-01 7.63209820e-01 5.82919419e-01 -7.13355184e-01
1.00548220e+00 -3.95893119e-02 1.08769238e+00 2.75234967e-01
4.67180192e-01 -1.00846183e+00 6.45714775e-02 3.95226479e-01
1.45173275e+00 -9.34968054e-01 -6.90602481e-01 -2.13095948e-01
-8.38808179e-01 1.00766814e+00 4.36893135e-01 9.12489593e-02
2.26115137e-01 4.17085201e-01 7.56648719e-01 1.47654787e-01
-2.41438851e-01 -2.21177340e-01 4.21147913e-01 1.06597149e+00
1.62441149e-01 1.05662644e-01 -2.39514694e-01 1.11457467e-01
-6.40855670e-01 2.73283213e-01 4.58326995e-01 7.72958815e-01
-7.43708611e-02 -1.16362667e+00 -5.30836642e-01 -1.66701704e-01
1.62582338e-01 -3.04184910e-02 -5.53090155e-01 7.61492729e-01
9.03277695e-02 6.54702365e-01 -1.21541247e-01 -6.48388937e-02
4.27622378e-01 2.33689785e-01 1.12712872e+00 -2.81066895e-01
2.15156619e-02 -1.07097238e-01 2.37779304e-01 -4.82872158e-01
-2.83701360e-01 -6.69444144e-01 -1.32215750e+00 -1.59917668e-01
1.28186315e-01 -4.45695579e-01 7.67091990e-01 6.98427498e-01
8.21662843e-02 2.32368082e-01 8.80354419e-02 -1.23645985e+00
-5.84644735e-01 -9.79610085e-01 -5.89697957e-01 6.37464583e-01
4.80273604e-01 -3.88326943e-01 -3.86236608e-01 4.00206834e-01]
|
[9.216070175170898, -6.527680397033691]
|
1f9223c7-f222-4913-a660-a87d4eba201f
|
constrained-labeled-data-generation-for-low
| null | null |
https://aclanthology.org/2021.findings-acl.396
|
https://aclanthology.org/2021.findings-acl.396.pdf
|
Constrained Labeled Data Generation for Low-Resource Named Entity Recognition
| null |
['Dan Roth', 'Ruohao Guo']
| null | null | null | null |
findings-acl-2021-8
|
['low-resource-named-entity-recognition']
|
['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]
|
[-7.385695457458496, 3.6895151138305664]
|
018df2b8-20fa-4e8a-9078-f6f91a7d3247
|
interactive-sketch-fill-multiclass-sketch-to
|
1909.11081
| null |
https://arxiv.org/abs/1909.11081v2
|
https://arxiv.org/pdf/1909.11081v2.pdf
|
Interactive Sketch & Fill: Multiclass Sketch-to-Image Translation
|
We propose an interactive GAN-based sketch-to-image translation method that helps novice users create images of simple objects. As the user starts to draw a sketch of a desired object type, the network interactively recommends plausible completions, and shows a corresponding synthesized image to the user. This enables a feedback loop, where the user can edit their sketch based on the network's recommendations, visualizing both the completed shape and final rendered image while they draw. In order to use a single trained model across a wide array of object classes, we introduce a gating-based approach for class conditioning, which allows us to generate distinct classes without feature mixing, from a single generator network. Video available at our website: https://arnabgho.github.io/iSketchNFill/.
|
['Arnab Ghosh', 'Puneet K. Dokania', 'Philip H. S. Torr', 'Alexei A. Efros', 'Richard Zhang', 'Eli Shechtman', 'Oliver Wang']
|
2019-09-24
|
interactive-sketch-fill-multiclass-sketch-to-1
|
http://openaccess.thecvf.com/content_ICCV_2019/html/Ghosh_Interactive_Sketch__Fill_Multiclass_Sketch-to-Image_Translation_ICCV_2019_paper.html
|
http://openaccess.thecvf.com/content_ICCV_2019/papers/Ghosh_Interactive_Sketch__Fill_Multiclass_Sketch-to-Image_Translation_ICCV_2019_paper.pdf
|
iccv-2019-10
|
['sketch-to-image-translation']
|
['computer-vision']
|
[ 4.28440064e-01 1.35036632e-01 -2.04022214e-01 -2.89874494e-01
-6.71081901e-01 -9.38442528e-01 7.78307378e-01 -3.90306771e-01
1.35227844e-01 5.87931037e-01 -3.53959692e-03 -1.82712704e-01
4.91521627e-01 -9.75551903e-01 -8.01314592e-01 -3.03053707e-01
4.65938032e-01 5.75560212e-01 -7.02266619e-02 9.67013538e-02
2.23261565e-01 5.58719516e-01 -1.56810558e+00 6.60715759e-01
6.96488917e-01 8.51571381e-01 5.40654480e-01 8.98023367e-01
-2.18898907e-01 6.32021248e-01 -5.62527835e-01 -5.01060903e-01
3.63509625e-01 -6.65242910e-01 -3.86171967e-01 3.02166045e-01
6.61139131e-01 -6.28543079e-01 -1.38921037e-01 8.06755841e-01
2.07358897e-01 9.62440521e-02 8.36991310e-01 -1.23074985e+00
-8.06410193e-01 6.03260040e-01 -1.87767699e-01 -6.18517816e-01
4.76960331e-01 5.40097952e-01 9.95584369e-01 -1.20982683e+00
1.06092358e+00 1.05437398e+00 1.61762074e-01 1.07995212e+00
-1.47055221e+00 -9.43889797e-01 2.12701097e-01 -1.98484227e-01
-1.10037661e+00 -3.25010061e-01 9.86240625e-01 -5.52606225e-01
4.53872532e-01 5.26296556e-01 1.21828604e+00 1.29807270e+00
-9.94527563e-02 9.61226285e-01 9.07637358e-01 -3.83511245e-01
3.41783017e-01 1.65357158e-01 -4.59011614e-01 7.61472762e-01
-1.33260071e-01 5.10213450e-02 -5.33806562e-01 -3.91481034e-02
1.39268434e+00 3.23826790e-01 -2.05176517e-01 -7.18663812e-01
-1.05273092e+00 8.13604951e-01 4.41272348e-01 1.44418016e-01
-3.40136230e-01 5.60692966e-01 -1.39588103e-01 2.30395064e-01
2.79305160e-01 6.36770070e-01 -2.62012124e-01 1.28473714e-02
-1.07651842e+00 4.28789347e-01 6.94338620e-01 1.11942327e+00
6.65059328e-01 1.18241444e-01 -3.69001597e-01 6.52778149e-01
2.57095158e-01 4.81842220e-01 -4.64085042e-02 -1.13166678e+00
1.72771260e-01 4.81594950e-01 1.70399800e-01 -5.80773473e-01
2.84788817e-01 -2.12902457e-01 -7.40315974e-01 9.70855355e-01
3.08599055e-01 8.99046585e-02 -1.26491773e+00 1.63727403e+00
1.50898620e-01 6.44566044e-02 -3.87683153e-01 8.70686948e-01
8.93014610e-01 6.11674190e-01 -1.85968149e-02 1.89094245e-01
1.23046899e+00 -1.01072800e+00 -5.05326807e-01 -1.21281408e-01
2.68754233e-02 -7.80671477e-01 1.57814002e+00 5.41224360e-01
-1.43803751e+00 -6.84061110e-01 -1.00719571e+00 -1.06979683e-01
-1.61552668e-01 4.44353968e-01 6.15361810e-01 3.80760163e-01
-1.06042075e+00 9.60624635e-01 -7.95286059e-01 -2.37417400e-01
8.82426023e-01 1.84260353e-01 -2.05452606e-01 -1.04715325e-01
-5.95704436e-01 5.18069208e-01 6.91970214e-02 -9.85079929e-02
-1.25185859e+00 -8.07728410e-01 -6.51390731e-01 1.86671689e-01
1.91399381e-01 -1.19762862e+00 1.42430627e+00 -1.27613127e+00
-1.96339846e+00 7.34829724e-01 -9.27400738e-02 7.56844506e-02
8.76365721e-01 -2.07714245e-01 1.03948861e-01 1.61169544e-01
-1.18716732e-01 1.16673124e+00 1.37883675e+00 -1.64230371e+00
-2.89805323e-01 2.83638760e-02 1.90152556e-01 1.71895742e-01
1.93420172e-01 -2.24742040e-01 -5.62621593e-01 -1.02623963e+00
-8.25883299e-02 -9.12662745e-01 -4.83487546e-02 6.14389300e-01
-5.87505460e-01 -5.41870594e-02 7.57345855e-01 -4.23791260e-01
9.39608872e-01 -2.06395531e+00 3.01796824e-01 3.35482985e-01
3.17521125e-01 8.20207223e-02 -3.66484940e-01 6.35606050e-01
-2.39560127e-01 2.05328271e-01 -3.85924548e-01 -7.95357645e-01
-5.29341623e-02 2.21488550e-02 -4.77954596e-01 -7.46858865e-02
3.02983105e-01 1.08586550e+00 -9.94757295e-01 -1.86169937e-01
3.92584592e-01 5.71274281e-01 -7.08990037e-01 6.35625780e-01
-6.50184214e-01 8.27140868e-01 -3.23237598e-01 4.40635830e-01
6.28850102e-01 -3.56066585e-01 3.35542053e-01 -2.52478242e-01
5.45405187e-02 2.60401011e-01 -1.13229227e+00 1.99786150e+00
-6.66718543e-01 6.17103875e-01 -1.04149580e-01 -5.48752487e-01
8.99455190e-01 3.52565587e-01 -5.24737500e-02 -2.06113458e-01
8.07499215e-02 1.57628745e-01 -3.56422007e-01 -9.94314700e-02
1.74179614e-01 -1.28444523e-01 2.23348051e-01 9.19640183e-01
6.53108954e-02 -7.11030781e-01 1.53867856e-01 3.69797558e-01
8.32306445e-01 7.92797744e-01 -3.86189409e-02 5.85150644e-02
1.63481906e-02 -4.32632089e-01 -8.42309222e-02 7.94229209e-01
6.89979017e-01 1.19299686e+00 4.99399394e-01 -6.13994062e-01
-1.32684457e+00 -1.39190090e+00 2.89886713e-01 1.06497598e+00
-1.93512633e-01 -3.80910873e-01 -7.83774793e-01 -6.03802919e-01
-1.27724797e-01 8.05743396e-01 -7.87566960e-01 7.65159279e-02
-5.40198505e-01 1.93057999e-01 6.74860552e-03 4.06058282e-01
8.49711001e-02 -1.52004528e+00 -7.08386064e-01 -3.82871814e-02
2.89158393e-02 -4.46209520e-01 -8.15740883e-01 -9.44327042e-02
-1.01631844e+00 -8.48688662e-01 -8.86663556e-01 -8.65877807e-01
1.20644879e+00 -7.39114359e-02 1.21453750e+00 3.96402448e-01
-3.30444664e-01 3.25760305e-01 -9.53978896e-02 -7.34404698e-02
-6.53344810e-01 6.13560341e-02 -4.45404291e-01 7.61122033e-02
-5.89975595e-01 -8.61930728e-01 -8.60441804e-01 6.71320260e-02
-9.56963897e-01 8.75474691e-01 4.78568375e-01 7.38255560e-01
6.80418670e-01 -6.21765256e-01 2.58962721e-01 -1.02411664e+00
6.81062043e-01 -3.85220647e-02 -6.78099394e-01 2.00001806e-01
-2.84279346e-01 1.14367552e-01 6.60151064e-01 -6.95506334e-01
-8.51231813e-01 3.24149132e-01 -1.12986550e-01 -8.64315152e-01
-2.98929840e-01 1.59050316e-01 -2.13677928e-01 1.82872534e-01
5.73975921e-01 1.64297178e-01 -4.13071066e-02 -4.69556630e-01
9.64085519e-01 1.71391696e-01 5.56848288e-01 -6.44750118e-01
8.90844464e-01 2.38934189e-01 -3.49030614e-01 -2.76937872e-01
-3.88885975e-01 3.26566190e-01 -6.14343822e-01 -3.76381040e-01
9.32702899e-01 -5.73915958e-01 -6.81007624e-01 3.85034472e-01
-1.32639062e+00 -9.45738435e-01 -6.10986233e-01 4.74520661e-02
-7.72624373e-01 -2.08968416e-01 -5.22590220e-01 -5.30110240e-01
-3.35522622e-01 -1.15013194e+00 1.04052925e+00 2.89711267e-01
-5.78294694e-01 -8.07348430e-01 -1.75310120e-01 -1.26803860e-01
3.02169144e-01 3.41184944e-01 8.48868072e-01 -1.79504067e-01
-1.16965294e+00 -2.09579602e-01 -9.74042043e-02 1.90711692e-01
2.89815247e-01 3.45370680e-01 -1.05219913e+00 -2.48048678e-01
-4.54348892e-01 -3.89809340e-01 6.01710081e-01 3.01411241e-01
1.55123603e+00 -5.26898503e-01 -3.29101205e-01 6.74716771e-01
1.22138882e+00 1.73393771e-01 8.92929912e-01 -2.06270829e-01
8.60493064e-01 1.45903066e-01 1.16068617e-01 3.81847113e-01
-3.76602858e-02 6.20463789e-01 3.00170124e-01 -2.63508320e-01
-5.25495470e-01 -7.98347056e-01 3.82732484e-03 5.26878595e-01
-1.82034060e-01 -3.84753376e-01 -3.86975169e-01 1.37258425e-01
-1.56918681e+00 -1.01647675e+00 3.25624734e-01 2.38229036e+00
8.75305176e-01 1.61045507e-01 9.95767042e-02 -1.60849661e-01
5.92487156e-01 1.40008703e-01 -6.58880532e-01 -2.94151902e-01
3.35212529e-01 7.08984256e-01 5.65842986e-02 7.41319656e-01
-6.25268161e-01 1.05859292e+00 5.93382788e+00 9.16886866e-01
-1.43222713e+00 -8.60984251e-02 8.50041449e-01 -2.35573456e-01
-9.38889742e-01 2.47705132e-01 -2.87158489e-01 3.68115366e-01
2.64606208e-01 -1.25531167e-01 8.92299592e-01 7.25354910e-01
2.11365163e-01 -4.42337543e-02 -1.29290783e+00 9.37016129e-01
-7.28320107e-02 -1.74836314e+00 4.47624117e-01 -8.64598602e-02
8.01533461e-01 -4.78894502e-01 2.85330623e-01 1.50236143e-02
4.06394035e-01 -1.11145365e+00 1.05484796e+00 8.89278710e-01
1.44886994e+00 -6.39349461e-01 -1.96000308e-01 2.40801677e-01
-1.11983311e+00 2.82154024e-01 1.89192705e-02 -4.53400277e-02
1.69573918e-01 3.27595532e-01 -9.09777164e-01 2.53852934e-01
2.48351589e-01 5.38252234e-01 -4.07669634e-01 7.70291746e-01
-7.89295614e-01 5.10880709e-01 -1.59071207e-01 -9.10487585e-03
-2.22193673e-01 -3.84299815e-01 2.46796787e-01 1.02889538e+00
5.22816956e-01 1.69963360e-01 6.34823963e-02 1.50438070e+00
-3.40516061e-01 -3.70865278e-02 -6.70494795e-01 -1.11127034e-01
5.40181935e-01 1.50445712e+00 -9.47566986e-01 -6.11379027e-01
1.81856565e-02 1.43587947e+00 3.50638360e-01 5.12186646e-01
-7.21270084e-01 -5.03606319e-01 3.84202302e-01 3.77072543e-01
4.95664567e-01 -2.15425268e-01 -4.67073739e-01 -1.05056286e+00
1.81874111e-02 -9.68259156e-01 -1.05126342e-02 -1.34110129e+00
-8.78169954e-01 7.21805990e-01 -1.33828223e-01 -1.38135421e+00
-4.18088555e-01 -3.31810832e-01 -1.16430962e+00 8.56838644e-01
-7.18889713e-01 -1.39940572e+00 -4.92439687e-01 3.31780046e-01
6.51622057e-01 -6.83800578e-02 8.90982330e-01 1.22718647e-01
-1.33059546e-01 6.30554020e-01 -4.07344013e-01 7.84300864e-02
5.77642679e-01 -1.13120914e+00 8.18295002e-01 5.17817318e-01
4.62061375e-01 7.48328030e-01 6.65933967e-01 -7.87790775e-01
-1.21698678e+00 -9.03039157e-01 6.04801655e-01 -5.43387771e-01
2.21874446e-01 -6.89424872e-01 -7.55446017e-01 5.99177063e-01
4.80495185e-01 -5.39662465e-02 3.75849545e-01 -3.98937374e-01
-3.20447206e-01 1.11876681e-01 -1.01449418e+00 9.88408148e-01
1.16338611e+00 -6.67111695e-01 -6.07727095e-02 2.43105665e-01
4.48615372e-01 -5.33600688e-01 -4.86608982e-01 1.65219456e-02
9.31605577e-01 -8.89312565e-01 7.88637280e-01 -4.81132925e-01
8.04922044e-01 -4.62645888e-01 2.00090423e-01 -1.45163238e+00
-3.52489203e-01 -8.06777656e-01 -4.31311782e-03 1.00211775e+00
6.17266774e-01 -3.88621837e-01 8.95839095e-01 6.01952255e-01
-6.60866573e-02 -9.04842317e-01 -4.04630989e-01 -4.09772247e-01
-7.27774277e-02 -4.51356471e-01 8.21491599e-01 6.79089844e-01
-5.85189685e-02 2.27336466e-01 -4.65497464e-01 -2.60716677e-01
3.69002044e-01 4.39108819e-01 1.11850989e+00 -1.06511176e+00
-5.57587981e-01 -7.08911717e-01 1.79801017e-01 -1.23722899e+00
-1.34455428e-01 -1.15844667e+00 -6.70766681e-02 -1.76309180e+00
1.29914507e-01 -5.66548824e-01 1.38249129e-01 5.46210945e-01
-1.27436683e-01 5.59887350e-01 6.54818475e-01 1.81129321e-01
-3.39340419e-01 5.05227208e-01 1.78784680e+00 1.70561988e-02
-2.96166301e-01 1.38904616e-01 -6.80724800e-01 4.84380364e-01
7.98958778e-01 -4.97046649e-01 -4.48639989e-01 -4.36906189e-01
2.22176149e-01 3.23110819e-01 5.97618520e-01 -9.27558005e-01
-1.13905244e-01 -2.51646966e-01 8.04985940e-01 -4.01174754e-01
5.29884994e-01 -4.75931287e-01 7.29513347e-01 5.41933119e-01
-6.29805565e-01 2.27273162e-03 1.19412728e-01 1.40532240e-01
2.04983428e-01 -2.26888165e-01 7.42126942e-01 -3.54945302e-01
-4.43126783e-02 6.30412340e-01 -2.89843678e-01 -2.20784500e-01
6.07250452e-01 -1.91395715e-01 -7.17792064e-02 -8.55124354e-01
-1.12596846e+00 -2.08108932e-01 8.16450000e-01 3.58250946e-01
7.82402098e-01 -1.67897284e+00 -6.19709253e-01 4.70194399e-01
-4.02484387e-02 1.01845399e-01 8.91581550e-02 1.28317475e-01
-4.59051609e-01 -1.37911946e-01 -2.69294739e-01 -4.04023081e-01
-9.36530471e-01 3.66031796e-01 2.59404510e-01 -9.74767283e-02
-6.82429552e-01 8.99028420e-01 6.11180782e-01 -3.82600486e-01
-5.43286512e-03 -4.97597426e-01 1.87213525e-01 -1.53012186e-01
4.38493639e-01 -9.94846895e-02 -2.06914127e-01 1.72407925e-01
1.44422993e-01 4.49113786e-01 1.70384660e-01 -6.28714979e-01
1.28065431e+00 3.21622431e-01 2.16951430e-01 4.94249016e-01
9.91351068e-01 2.15501919e-01 -1.92158461e+00 2.29218826e-01
-6.44587636e-01 -7.49921501e-01 -4.57359046e-01 -9.32574272e-01
-1.18228841e+00 9.46454108e-01 5.07496476e-01 -2.29950175e-02
9.16455805e-01 1.38201922e-01 4.53576833e-01 8.26034993e-02
2.82334656e-01 -6.99191034e-01 6.74454153e-01 2.44882986e-01
1.49712706e+00 -6.87619030e-01 -1.49618844e-02 -2.80407012e-01
-7.81433284e-01 1.26156223e+00 4.53062296e-01 -4.96003449e-01
4.29780543e-01 1.94231346e-01 1.68989703e-01 -2.74725165e-02
-7.97585785e-01 7.46147335e-02 3.98920685e-01 6.79039598e-01
5.01542926e-01 2.56018192e-01 -1.65298451e-02 2.93689966e-01
-4.61070389e-01 2.91259766e-01 4.97142345e-01 7.17928231e-01
-2.22776219e-01 -1.53519154e+00 -5.86563461e-02 5.66678703e-01
-1.81707107e-02 -2.36199066e-01 -2.57995516e-01 5.00574172e-01
1.50949836e-01 2.66867429e-01 2.85606056e-01 -2.36473247e-01
1.88562468e-01 9.09923241e-02 7.98580289e-01 -8.77803385e-01
-3.61343443e-01 2.13058427e-01 -1.39612243e-01 -5.23776531e-01
-1.44596407e-02 -5.48485339e-01 -1.03663409e+00 -3.31796184e-02
-3.65293920e-02 -3.20976675e-02 6.23499155e-01 4.54340994e-01
3.86140138e-01 4.36289370e-01 7.86995888e-01 -1.45902765e+00
2.55566109e-02 -9.36690271e-01 -2.61881322e-01 5.08407414e-01
1.47009894e-01 -4.26880717e-01 -1.56623438e-01 4.35021967e-01]
|
[11.555671691894531, -0.37150782346725464]
|
3ab50c30-1104-47d2-b478-4314b595cdf7
|
complementary-time-frequency-domain-networks
|
2012.11974
| null |
https://arxiv.org/abs/2012.11974v2
|
https://arxiv.org/pdf/2012.11974v2.pdf
|
Complementary Time-Frequency Domain Networks for Dynamic Parallel MR Image Reconstruction
|
Purpose: To introduce a novel deep learning based approach for fast and high-quality dynamic multi-coil MR reconstruction by learning a complementary time-frequency domain network that exploits spatio-temporal correlations simultaneously from complementary domains. Theory and Methods: Dynamic parallel MR image reconstruction is formulated as a multi-variable minimisation problem, where the data is regularised in combined temporal Fourier and spatial (x-f) domain as well as in spatio-temporal image (x-t) domain. An iterative algorithm based on variable splitting technique is derived, which alternates among signal de-aliasing steps in x-f and x-t spaces, a closed-form point-wise data consistency step and a weighted coupling step. The iterative model is embedded into a deep recurrent neural network which learns to recover the image via exploiting spatio-temporal redundancies in complementary domains. Results: Experiments were performed on two datasets of highly undersampled multi-coil short-axis cardiac cine MRI scans. Results demonstrate that our proposed method outperforms the current state-of-the-art approaches both quantitatively and qualitatively. The proposed model can also generalise well to data acquired from a different scanner and data with pathologies that were not seen in the training set. Conclusion: The work shows the benefit of reconstructing dynamic parallel MRI in complementary time-frequency domains with deep neural networks. The method can effectively and robustly reconstruct high-quality images from highly undersampled dynamic multi-coil data ($16 \times$ and $24 \times$ yielding 15s and 10s scan times respectively) with fast reconstruction speed (2.8s). This could potentially facilitate achieving fast single-breath-hold clinical 2D cardiac cine imaging.
|
['Daniel Rueckert', 'Joseph V. Hajnal', 'Anthony N. Price', 'Claudia Prieto', 'René Botnar', 'Thomas Küstner', 'Jo Schlemper', 'Kerstin Hammernik', 'Jinming Duan', 'Chen Qin']
|
2020-12-22
| null | null | null | null |
['de-aliasing']
|
['computer-vision']
|
[ 5.35787225e-01 -1.56614766e-01 1.04377903e-01 -4.27596390e-01
-1.07613420e+00 -1.57184169e-01 1.87376902e-01 -2.39022300e-01
-6.08833611e-01 7.40283072e-01 9.16601717e-02 -2.34578878e-01
-8.85918558e-01 -3.20415407e-01 -5.07370651e-01 -1.03955758e+00
-7.95141816e-01 5.94180763e-01 1.07970148e-01 5.28592290e-03
-1.59699932e-01 5.78435302e-01 -9.77137685e-01 2.67134339e-01
3.89883935e-01 7.46808290e-01 6.62302732e-01 8.75697792e-01
5.02835035e-01 1.12188244e+00 -2.28243060e-02 4.30227488e-01
3.11085463e-01 -7.80082345e-01 -1.15281641e+00 2.64982320e-02
5.78294657e-02 -4.08333749e-01 -4.88495678e-01 6.49226427e-01
1.06924582e+00 4.48602796e-01 2.51082391e-01 -4.11061406e-01
-1.23928972e-01 4.76971537e-01 -8.24651957e-01 9.09234524e-01
-9.78835076e-02 2.28407294e-01 2.86615968e-01 -5.72273254e-01
9.47412789e-01 4.23478007e-01 8.47302854e-01 3.86910528e-01
-1.34150791e+00 -4.18430328e-01 -3.92092258e-01 2.21305072e-01
-8.20798099e-01 -2.69653350e-01 8.67500842e-01 -4.90672827e-01
1.23561692e+00 1.62365690e-01 6.37515783e-01 7.62871206e-01
7.76970625e-01 3.79598856e-01 1.67585444e+00 -2.89881140e-01
-1.88523099e-01 -5.27073383e-01 6.68832287e-02 7.00508773e-01
-4.75078046e-01 5.41711986e-01 -3.35686654e-01 -5.17885424e-02
1.07046843e+00 7.01029003e-02 -5.32125175e-01 -4.11404401e-01
-1.76202679e+00 7.29239643e-01 5.19556403e-01 9.97863233e-01
-9.33549821e-01 1.40704408e-01 8.93494010e-01 4.52892482e-01
3.62019986e-01 5.24624407e-01 -3.32246751e-01 -6.82095112e-03
-1.42042756e+00 5.68909496e-02 2.84412116e-01 4.04919773e-01
1.64033055e-01 4.09723938e-01 -7.02968612e-03 7.73648202e-01
1.25250474e-01 4.50569361e-01 1.07187498e+00 -1.04961228e+00
3.06383055e-02 -3.92684937e-01 -2.20600635e-01 -7.95268357e-01
-9.39280093e-01 -7.51921296e-01 -1.23178017e+00 -2.59212796e-02
2.30925173e-01 -2.20054477e-01 -8.38448286e-01 1.68090355e+00
4.89657640e-01 3.15343231e-01 -1.33169234e-01 1.44657052e+00
7.24935770e-01 3.72735023e-01 -4.40264791e-02 -8.94528031e-01
1.33825171e+00 -7.59551644e-01 -1.08474350e+00 1.52156278e-01
6.32942379e-01 -5.97403109e-01 4.23926115e-01 3.52618724e-01
-1.64603734e+00 -5.44467688e-01 -1.05840671e+00 2.34127522e-01
3.24351788e-01 -2.61282116e-01 4.18711305e-01 1.84137568e-01
-1.16616440e+00 1.03940332e+00 -1.27290297e+00 4.51534837e-01
2.30986208e-01 6.01778388e-01 -5.58277190e-01 -2.63813019e-01
-1.45011961e+00 1.09886742e+00 2.68580079e-01 2.59146392e-01
-9.52974081e-01 -1.04187524e+00 -6.61864758e-01 -6.33636296e-01
3.56709301e-01 -7.78596103e-01 1.14451742e+00 -9.79477942e-01
-1.38917756e+00 9.34168220e-01 -5.45358583e-02 -6.61116421e-01
4.94127989e-01 9.18802917e-02 -6.58060431e-01 6.86561942e-01
2.89415091e-01 9.77405980e-02 1.02270830e+00 -8.97130072e-01
5.79058081e-02 -5.09716451e-01 -3.88758808e-01 1.84674323e-01
5.96886218e-01 1.56140670e-01 1.66443467e-01 -7.18081832e-01
6.04626179e-01 -9.56651688e-01 -4.72717315e-01 -2.31417730e-01
-6.13037571e-02 5.77888429e-01 6.44136488e-01 -1.32896507e+00
1.05968416e+00 -1.72085118e+00 5.11897147e-01 1.04645155e-01
7.38406241e-01 1.91066280e-01 4.41889837e-02 1.37294875e-02
-9.00722206e-01 -4.72447127e-01 -6.45829558e-01 9.43248197e-02
-7.18444228e-01 9.32437256e-02 1.12456188e-01 1.01292861e+00
-1.83827654e-01 9.84969139e-01 -1.08245063e+00 -3.31804693e-01
4.15646702e-01 7.64192820e-01 -1.39058799e-01 3.37170571e-01
4.07354593e-01 1.20821762e+00 -1.64317384e-01 2.45091081e-01
6.75852358e-01 -3.75290304e-01 5.13243318e-01 -5.50820470e-01
-3.08413059e-01 8.01939890e-02 -8.78885865e-01 2.24820662e+00
-6.02807164e-01 5.05274713e-01 4.35729086e-01 -1.68101048e+00
5.63519120e-01 8.25010598e-01 1.28050315e+00 -1.28630126e+00
1.81665286e-01 4.79615957e-01 3.20524663e-01 -9.40336466e-01
5.85984811e-02 -1.22787070e+00 2.77348369e-01 7.85113871e-01
2.22427830e-01 -1.83950603e-01 -8.16079974e-02 -6.65673614e-02
1.12354445e+00 6.86061010e-02 6.64636493e-02 -5.49492776e-01
6.04259670e-01 -2.40627930e-01 3.70464146e-01 7.48019099e-01
-4.42902505e-01 5.87081850e-01 3.77528369e-02 -7.87173986e-01
-1.21910608e+00 -9.84334230e-01 -5.56383610e-01 5.46109855e-01
-2.23872542e-01 1.39459699e-01 -4.17098969e-01 -4.44769889e-01
-6.27875686e-01 2.33545825e-01 -6.12875342e-01 9.90685523e-02
-1.39857292e+00 -1.03362000e+00 1.80169284e-01 4.22239065e-01
3.83284003e-01 -1.10191405e+00 -1.28042400e+00 8.41171026e-01
-5.37617743e-01 -9.79832411e-01 -3.53996038e-01 7.09314704e-01
-1.33083797e+00 -9.64290142e-01 -1.14421332e+00 -6.15501940e-01
3.08106750e-01 1.76462829e-01 1.28309345e+00 -1.24154046e-01
-5.57696223e-01 3.00289810e-01 -2.07373574e-01 2.64466256e-01
-2.48721093e-01 -3.00663561e-01 1.07051596e-01 -2.79432058e-01
-3.21661890e-01 -9.60886180e-01 -8.00055563e-01 1.44643500e-01
-1.00932336e+00 2.49056041e-01 5.12422085e-01 1.39806199e+00
9.73474503e-01 -2.21244749e-02 5.09429216e-01 -7.28685856e-01
3.21367502e-01 -6.54689252e-01 -1.91878334e-01 -6.07528770e-03
-4.43262309e-01 3.42218637e-01 6.29129112e-01 -4.20494944e-01
-7.92508781e-01 1.99853890e-02 -2.99618453e-01 -5.69104016e-01
6.82555363e-02 5.72337687e-01 5.45469701e-01 -2.72537023e-01
6.43330932e-01 7.12397039e-01 4.63751525e-01 -2.30717957e-01
5.59264086e-02 2.12299973e-02 6.79508150e-01 -3.97688985e-01
2.44185671e-01 6.27687454e-01 4.01145011e-01 -7.02672958e-01
-4.45457876e-01 -3.78542632e-01 -9.95247066e-01 -5.07617474e-01
1.02611697e+00 -7.34380186e-01 -4.46020097e-01 4.68230188e-01
-7.27486849e-01 -3.76531929e-01 -4.63007033e-01 9.77580547e-01
-8.54196072e-01 5.34449816e-01 -1.03300762e+00 -2.28625655e-01
-6.90196931e-01 -1.53397429e+00 8.52652073e-01 -3.29209149e-01
-5.52954338e-02 -1.23422539e+00 2.94104904e-01 3.54099572e-01
8.16037357e-01 7.61579216e-01 8.36186826e-01 -1.43565744e-01
-3.92359376e-01 1.86195180e-01 1.12393193e-01 2.35040113e-01
-2.08639745e-02 -9.18562710e-01 -5.91528296e-01 -5.62794387e-01
9.19635355e-01 -4.25648153e-01 7.32703865e-01 1.21484947e+00
1.04488432e+00 9.19024870e-02 6.00571930e-02 9.53831613e-01
1.53384018e+00 3.16464007e-01 4.94590908e-01 1.69708014e-01
4.98675466e-01 5.24930835e-01 2.16470197e-01 2.66038686e-01
1.31249368e-01 7.29662418e-01 1.16071194e-01 -3.56654882e-01
-2.78384149e-01 3.35461318e-01 -9.04077888e-02 1.30029523e+00
-2.40848839e-01 5.53360820e-01 -1.00816941e+00 6.70343399e-01
-1.47986317e+00 -1.08283079e+00 -3.39140773e-01 2.02407002e+00
8.65114450e-01 -3.28973442e-01 1.54113844e-01 4.34893191e-01
5.84598660e-01 2.92981654e-01 -6.82996511e-01 -2.02287257e-01
-2.44386420e-02 6.30499780e-01 4.51378465e-01 5.77920258e-01
-9.71303523e-01 5.37023172e-02 6.01252031e+00 4.72502947e-01
-1.73162484e+00 8.38644028e-01 7.12005556e-01 -4.59228903e-01
-7.56072104e-02 -3.89930427e-01 1.40580073e-01 2.91690171e-01
1.34473836e+00 5.77548109e-02 5.54353237e-01 2.11711779e-01
3.90962034e-01 -4.35405038e-02 -7.83795238e-01 1.11406469e+00
-5.48281558e-02 -1.41786015e+00 -7.19420791e-01 -9.10880119e-02
5.44037044e-01 4.47291970e-01 1.79288723e-02 -2.62850579e-02
-3.10555637e-01 -1.16376209e+00 4.70893204e-01 6.65975213e-01
1.28229690e+00 -6.20162904e-01 5.28740346e-01 4.03091401e-01
-1.04767287e+00 1.78302571e-01 2.72782266e-01 3.12877059e-01
6.31984949e-01 8.22953463e-01 -4.93052751e-01 9.81094778e-01
6.41182661e-01 9.34478343e-01 9.60496068e-02 5.60107708e-01
2.64613211e-01 2.88425982e-01 -2.11551011e-01 7.13805616e-01
3.91939759e-01 -1.23344310e-01 6.44637406e-01 1.10022843e+00
2.32148379e-01 5.79439938e-01 -2.90827882e-02 6.97499514e-01
3.69724452e-01 -3.13159764e-01 -4.64118004e-01 4.33098346e-01
-1.71820521e-01 1.17480505e+00 -8.32284451e-01 -4.00024414e-01
-1.80145815e-01 8.94161820e-01 -1.78561397e-02 3.10003966e-01
-8.42425883e-01 2.16914341e-02 -2.82030374e-01 4.99550879e-01
1.98729619e-01 -3.33543122e-01 -6.37264326e-02 -1.11511600e+00
9.63445082e-02 -9.50282156e-01 3.44474584e-01 -7.03747690e-01
-1.09440768e+00 1.08345032e+00 2.32035115e-01 -1.06602168e+00
-6.84253573e-01 -1.64338678e-01 -1.28441185e-01 1.15908909e+00
-1.59237516e+00 -9.14233327e-01 8.48882794e-02 8.33256125e-01
3.26944917e-01 5.53501658e-02 8.49209905e-01 5.95745921e-01
-5.88776506e-02 2.55854949e-02 2.19380677e-01 -1.25894457e-01
3.88900012e-01 -1.17323232e+00 -1.52120054e-01 6.78866088e-01
-3.62715870e-01 6.00633919e-01 6.78718090e-01 -5.49376607e-01
-1.44007468e+00 -7.33810365e-01 6.92866564e-01 1.99216008e-01
5.14467895e-01 1.35449424e-01 -1.03957129e+00 4.07233596e-01
2.51288265e-01 7.63401031e-01 6.11170471e-01 -4.64826167e-01
2.45247126e-01 -2.76755169e-03 -1.50601315e+00 -2.47022271e-01
7.25111544e-01 -7.95322955e-01 -5.76861858e-01 5.35784304e-01
3.49630564e-01 -9.54907119e-01 -1.49922025e+00 4.38235611e-01
6.40624523e-01 -1.07676339e+00 1.15577519e+00 -3.33085984e-01
5.61560988e-01 -6.95257857e-02 5.75918034e-02 -1.33976042e+00
-4.26917970e-01 -6.99199259e-01 -2.67696291e-01 3.54542695e-02
-1.18364073e-01 -5.85857332e-01 3.25921297e-01 3.14183682e-01
-4.02090192e-01 -1.03759611e+00 -1.56955481e+00 -4.92557317e-01
2.51153022e-01 -4.41636592e-01 -1.26600964e-02 1.24410903e+00
-2.11179063e-01 1.11332528e-01 -6.08221889e-01 -8.07522312e-02
8.03683341e-01 1.51378378e-01 -3.12418610e-01 -6.76874995e-01
-7.24751472e-01 -1.06200926e-01 -1.57687843e-01 -8.16146791e-01
1.73833475e-01 -1.07425296e+00 -6.72756881e-02 -1.11453962e+00
2.19352663e-01 -5.40252328e-01 -4.94504064e-01 1.55093282e-01
1.14806108e-01 1.89629987e-01 4.03742902e-02 3.57935041e-01
-1.96871877e-01 2.57700443e-01 1.81642759e+00 1.31619394e-01
3.28285284e-02 -2.67042071e-01 -1.24256767e-01 2.39868879e-01
4.17958528e-01 -6.45029485e-01 -4.59033251e-01 -4.71435070e-01
-1.05998784e-01 1.29066634e+00 4.83896762e-01 -1.01786554e+00
7.41881132e-02 2.13781133e-01 3.95632744e-01 -3.99758905e-01
1.84188172e-01 -7.48028636e-01 6.11804008e-01 7.59788752e-01
-4.09700483e-01 4.38492477e-01 2.00875953e-01 1.44272268e-01
-3.06534886e-01 -1.01332217e-01 1.25961244e+00 -6.67157769e-01
-2.39676654e-01 4.30834353e-01 -3.28742415e-01 9.71926898e-02
8.18145990e-01 -5.22035100e-02 3.77691954e-01 -8.85039642e-02
-1.21256173e+00 -2.78983593e-01 -1.62787095e-01 7.34623149e-02
8.62279058e-01 -1.27958977e+00 -8.30250561e-01 4.74783540e-01
-6.33950710e-01 -1.40806288e-01 1.16453695e+00 1.70388532e+00
-5.76211751e-01 4.80299890e-01 -5.51247180e-01 -1.00497019e+00
-8.54669094e-01 4.91047353e-01 1.08871591e+00 -9.60190117e-01
-1.22990310e+00 7.36576557e-01 -2.48517230e-01 -4.79552418e-01
-3.39997858e-01 -1.90158814e-01 -8.74340907e-02 -1.68593779e-01
6.30397618e-01 6.09856658e-02 4.39391553e-01 -1.02519715e+00
-4.02605325e-01 7.12007523e-01 -6.21140599e-02 -3.90445083e-01
1.80788052e+00 -1.63342625e-01 -1.47114962e-01 5.51887751e-01
1.59384799e+00 -6.40638471e-01 -1.21035886e+00 -4.53812033e-01
-2.66307175e-01 -2.77336359e-01 6.89381242e-01 -1.02408910e+00
-1.49876082e+00 1.06472719e+00 1.22225201e+00 -8.92582014e-02
1.44490910e+00 -2.49489918e-01 1.10190988e+00 -2.52751887e-01
3.67575198e-01 -8.87164831e-01 3.18970494e-02 3.79272550e-01
1.11049974e+00 -1.19212365e+00 2.52699614e-01 2.74069905e-01
-6.25347376e-01 1.19268143e+00 -1.14323631e-01 -2.90174931e-01
7.42386460e-01 5.07234871e-01 2.63856631e-02 -6.40118539e-01
-5.11367500e-01 2.85976410e-01 2.09105909e-01 5.74064493e-01
8.79810750e-01 1.24987118e-01 -4.05041844e-01 2.98464984e-01
1.70841634e-01 2.02803716e-01 3.84391904e-01 9.85077322e-01
1.37624979e-01 -9.77467537e-01 -3.88763815e-01 5.06119072e-01
-7.62539327e-01 -4.23559733e-02 6.63570821e-01 7.59447813e-01
1.01318836e-01 4.16863173e-01 -2.94998705e-01 -3.84304039e-02
6.14360571e-02 4.76854406e-02 8.88805091e-01 -1.39065087e-01
-1.01387882e+00 4.90576386e-01 -9.13941562e-02 -8.27679098e-01
-9.49386895e-01 -8.82007897e-01 -1.50782776e+00 1.34294778e-02
-8.45755786e-02 4.72072791e-03 7.07116902e-01 9.15340126e-01
1.58356309e-01 1.08660066e+00 7.48181760e-01 -1.10173559e+00
-5.69329023e-01 -9.27551508e-01 -8.07558298e-01 4.93076146e-01
8.15400481e-01 -5.21403790e-01 -1.85940459e-01 3.11521022e-03]
|
[13.526352882385254, -2.4019837379455566]
|
337185fc-8848-4957-a7fd-a5b3983fa647
|
sc-block-supervised-contrastive-blocking
|
2303.03132
| null |
https://arxiv.org/abs/2303.03132v2
|
https://arxiv.org/pdf/2303.03132v2.pdf
|
SC-Block: Supervised Contrastive Blocking within Entity Resolution Pipelines
|
The goal of entity resolution is to identify records in multiple datasets that represent the same real-world entity. However, comparing all records across datasets can be computationally intensive, leading to long runtimes. To reduce these runtimes, entity resolution pipelines are constructed of two parts: a blocker that applies a computationally cheap method to select candidate record pairs, and a matcher that afterwards identifies matching pairs from this set using more expensive methods. This paper presents SC-Block, a blocking method that utilizes supervised contrastive learning for positioning records in the embedding space, and nearest neighbour search for candidate set building. We benchmark SC-Block against eight state-of-the-art blocking methods. In order to relate the training time of SC-Block to the reduction of the overall runtime of the entity resolution pipeline, we combine SC-Block with four matching methods into complete pipelines. For measuring the overall runtime, we determine candidate sets with 99.5% pair completeness and pass them to the matcher. The results show that SC-Block is able to create smaller candidate sets and pipelines with SC-Block execute 1.5 to 2 times faster compared to pipelines with other blockers, without sacrificing F1 score. Blockers are often evaluated using relatively small datasets which might lead to runtime effects resulting from a large vocabulary size being overlooked. In order to measure runtimes in a more challenging setting, we introduce a new benchmark dataset that requires large numbers of product offers to be blocked. On this large-scale benchmark dataset, pipelines utilizing SC-Block and the best-performing matcher execute 8 times faster than pipelines utilizing another blocker with the same matcher reducing the runtime from 2.5 hours to 18 minutes, clearly compensating for the 5 minutes required for training SC-Block.
|
['Christian Bizer', 'Roee Shraga', 'Alexander Brinkmann']
|
2023-03-06
| null | null | null | null |
['blocking', 'entity-resolution']
|
['natural-language-processing', 'natural-language-processing']
|
[ 1.03839867e-01 -1.68278646e-02 9.01136994e-02 -4.22570795e-01
-1.32114363e+00 -8.92656744e-01 4.50536281e-01 8.90230238e-01
-7.21366107e-01 4.62490201e-01 8.03877413e-03 -1.15439668e-01
-2.70144999e-01 -1.10591710e+00 -9.85050976e-01 -1.37078986e-01
-2.49174237e-01 7.78766870e-01 5.67561507e-01 7.75676817e-02
1.39623120e-01 4.58130300e-01 -1.54849553e+00 4.32019502e-01
5.87027192e-01 5.89859545e-01 8.45490098e-02 4.68860418e-01
-3.51873726e-01 3.48562062e-01 -6.66085064e-01 -6.52566254e-01
5.26247084e-01 -1.71953645e-02 -8.12864840e-01 -7.04108715e-01
9.29557681e-01 -1.16914406e-01 -2.32023988e-02 7.20048726e-01
5.68521202e-01 -1.03282601e-01 3.27921242e-01 -1.02342844e+00
8.49290192e-02 9.88305569e-01 -2.84931928e-01 1.18503004e-01
5.25543809e-01 -1.73417069e-02 1.17045081e+00 -8.52876186e-01
1.00847280e+00 9.67177689e-01 1.01046526e+00 9.06182602e-02
-1.73200667e+00 -9.05467451e-01 -2.70382404e-01 -4.06900942e-02
-1.77913737e+00 -4.92487997e-01 8.13623518e-02 -4.46466386e-01
1.32495093e+00 6.15739048e-01 2.76750773e-01 4.94739592e-01
-4.10017103e-01 4.15764928e-01 6.12797916e-01 -3.77290219e-01
2.85101712e-01 2.76106358e-01 3.17071974e-01 3.50970536e-01
6.65660024e-01 6.33713906e-04 -3.74286205e-01 -4.20985729e-01
1.87206835e-01 -1.24554731e-01 -7.55300000e-02 -4.52256858e-01
-1.31475627e+00 5.79198122e-01 4.74382818e-01 3.51913929e-01
-4.44149017e-01 -5.17398454e-02 4.82800066e-01 2.92345762e-01
-1.09195095e-02 1.07529414e+00 -6.96708977e-01 -1.74405620e-01
-1.31695330e+00 5.08848250e-01 1.09447551e+00 1.14560103e+00
1.04354274e+00 -7.53261387e-01 -2.74144381e-01 4.97512788e-01
-1.61578786e-02 2.40743563e-01 4.25626002e-02 -6.82363331e-01
6.47953808e-01 1.00474501e+00 2.03424498e-01 -8.10859382e-01
-6.42914176e-01 -4.17681009e-01 -2.34918773e-01 -3.38551290e-02
5.76323092e-01 9.41358060e-02 -7.68208861e-01 1.58352995e+00
5.44088423e-01 2.91175723e-01 7.62117654e-03 6.18468642e-01
8.16704273e-01 4.97871429e-01 1.08036771e-01 5.25332056e-02
1.64777517e+00 -6.87675476e-01 -2.55781204e-01 -1.00529008e-01
1.17952096e+00 -8.65182519e-01 9.18367028e-01 1.16335221e-01
-1.18861532e+00 -4.00076270e-01 -1.27256799e+00 -2.00868234e-01
-6.28779233e-01 -1.05684819e-02 3.76395255e-01 4.80748683e-01
-8.19328129e-01 8.78024220e-01 -7.29447365e-01 -5.22194207e-01
2.98787773e-01 6.96024418e-01 -6.52632058e-01 4.39973138e-02
-1.16184580e+00 1.00050330e+00 4.89464223e-01 -3.97889346e-01
-2.58944184e-01 -1.31671929e+00 -9.35133338e-01 4.24206614e-01
3.60510111e-01 -6.16305828e-01 9.10455287e-01 -1.99781284e-01
-6.11146152e-01 8.68471682e-01 -2.50065893e-01 -5.50840855e-01
4.52258617e-01 -3.41258854e-01 -3.31180304e-01 -2.42679819e-01
3.92645776e-01 4.75107849e-01 2.49155521e-01 -9.03453112e-01
-8.95508587e-01 -1.53034493e-01 -7.59524480e-02 -2.09337771e-01
-1.92070290e-01 8.01073462e-02 -8.45716476e-01 -3.23233187e-01
-8.88880491e-02 -9.04570341e-01 -2.28738040e-01 -4.62087601e-01
-1.08133323e-01 -2.49796018e-01 2.86075294e-01 -4.16579217e-01
1.64059734e+00 -2.19036412e+00 -4.91053760e-02 4.41413760e-01
1.92134932e-01 4.56906319e-01 -2.16397688e-01 6.21424079e-01
-7.87807480e-02 2.68073499e-01 -1.40604809e-01 -4.94583428e-01
6.40843660e-02 -9.22774971e-02 -1.30016014e-01 2.76315451e-01
2.77661383e-01 7.86779523e-01 -8.12112689e-01 -4.90169942e-01
1.07500620e-01 4.20617402e-01 -5.42704523e-01 1.42242670e-01
-8.18073936e-03 -9.98683795e-02 -9.33640003e-02 4.49426174e-01
7.52182364e-01 -2.32608765e-01 2.81390905e-01 -3.64571065e-01
-2.86354095e-01 7.58481741e-01 -1.65844476e+00 1.68191910e+00
-4.66725200e-01 5.23939610e-01 -2.12645769e-01 -6.01819992e-01
9.37708676e-01 4.66940738e-02 4.10844326e-01 -7.74731338e-01
-2.83326685e-01 5.95396101e-01 -1.97694600e-01 -3.10063779e-01
6.64278150e-01 4.04740185e-01 -2.91222066e-01 1.93850294e-01
-6.29541948e-02 1.70475289e-01 6.05120659e-01 2.65353262e-01
1.71098816e+00 4.11358625e-02 1.61869094e-01 -1.49477512e-01
4.06497210e-01 3.98101538e-01 6.88330352e-01 8.11069846e-01
2.82810688e-01 5.05514443e-01 3.94862384e-01 -6.05492294e-01
-1.22875071e+00 -8.89720619e-01 -2.24016026e-01 9.03626323e-01
2.31235519e-01 -1.09206688e+00 -7.06282914e-01 -5.68465233e-01
3.67739201e-01 7.48309731e-01 -3.99755955e-01 5.52849099e-02
-9.03476715e-01 -5.86026311e-01 8.92935991e-01 4.35719609e-01
1.46507651e-01 -7.87184536e-01 -8.21746409e-01 3.86652529e-01
4.48742546e-02 -1.07352936e+00 -2.62875795e-01 2.64685124e-01
-5.18884182e-01 -1.10077775e+00 -2.21726999e-01 -6.30775571e-01
5.98832846e-01 -6.61174431e-02 1.51794004e+00 2.64309734e-01
-4.97810900e-01 -1.00829586e-01 -1.63965493e-01 -3.28370839e-01
-3.07764918e-01 6.74370050e-01 -1.78759247e-01 -4.42653209e-01
1.00646532e+00 -3.58646035e-01 -5.57282507e-01 4.19012636e-01
-7.83146024e-01 -1.48051204e-02 8.06601763e-01 6.62766516e-01
7.82876253e-01 -7.01803491e-02 2.48178035e-01 -1.15636134e+00
1.50186464e-01 -6.32020593e-01 -1.14830458e+00 3.78410906e-01
-7.33389318e-01 3.14677596e-01 5.22516012e-01 -3.62203032e-01
-4.89718795e-01 4.66133893e-01 -8.51386506e-03 -8.63238573e-02
-6.29074574e-02 5.28047025e-01 -1.16684757e-01 2.69257516e-01
8.01015437e-01 -1.20529369e-01 -2.03305751e-01 -7.12199569e-01
1.95673868e-01 5.77878892e-01 6.23689771e-01 -4.67302144e-01
1.00360823e+00 2.93423146e-01 -4.06709462e-02 -2.33991280e-01
-4.79680657e-01 -8.14721942e-01 -6.31818473e-01 3.97287786e-01
6.56675577e-01 -1.00699604e+00 -6.49384320e-01 -1.28035218e-01
-1.02978504e+00 -1.24286495e-01 -1.99573755e-01 5.08247256e-01
-6.97955415e-02 -1.51656028e-02 -4.40534174e-01 -2.56514400e-01
-5.12011349e-01 -1.11685002e+00 1.11215401e+00 2.11431012e-01
-6.49542093e-01 -4.07870710e-01 3.26217622e-01 3.10681731e-01
4.10642654e-01 2.26109192e-01 6.92074776e-01 -1.20026517e+00
-8.00113797e-01 -5.27055681e-01 -2.74218559e-01 -3.07282060e-01
-9.15695056e-02 5.07055670e-02 -8.45025420e-01 -3.49858612e-01
-5.92515528e-01 2.25016639e-01 7.38560081e-01 -6.05511591e-02
5.83306551e-01 3.28401849e-02 -8.13977063e-01 7.59977281e-01
1.63846231e+00 -2.39363890e-02 6.58341229e-01 9.78658438e-01
4.99647349e-01 6.64856017e-01 1.05868578e+00 1.03976622e-01
4.42552060e-01 1.01143992e+00 1.42934352e-01 -3.59255642e-01
1.77066009e-02 -2.64526635e-01 1.80480093e-01 3.42810005e-01
3.19651812e-01 -3.26725654e-02 -1.17977846e+00 9.32570100e-01
-1.77579188e+00 -9.80975926e-01 -3.70316952e-01 2.59111047e+00
1.00646913e+00 2.78667897e-01 1.96430892e-01 1.32809848e-01
6.53724074e-01 -2.75190800e-01 -2.12169766e-01 -2.46401265e-01
1.42190814e-01 4.09731120e-01 9.90847886e-01 4.92311537e-01
-1.14073801e+00 8.43071640e-01 5.50833797e+00 5.96058190e-01
-1.14741755e+00 -1.65127084e-01 1.28226638e-01 -5.06550372e-01
-2.30449453e-01 4.56671655e-01 -1.31469202e+00 6.63251102e-01
1.28010535e+00 -8.83997455e-02 3.05203289e-01 8.57999980e-01
-1.00617953e-01 -1.26401186e-01 -1.54474878e+00 8.33265543e-01
-3.92214894e-01 -1.60135877e+00 -2.96904862e-01 1.85194328e-01
4.54750419e-01 4.74526405e-01 -3.79071772e-01 4.48535472e-01
2.03394085e-01 -1.02680957e+00 5.50622642e-01 2.48644352e-01
6.71271563e-01 -7.59740949e-01 9.24112022e-01 1.59252703e-01
-1.41312897e+00 1.75915554e-01 -4.23171699e-01 2.50978082e-01
1.90895393e-01 7.79615223e-01 -1.09388185e+00 5.79868853e-01
9.31077480e-01 3.00946772e-01 -8.54417384e-01 1.23661661e+00
2.08458260e-01 3.71878713e-01 -9.07433987e-01 1.24382958e-01
2.64197849e-02 1.28846005e-01 5.35978675e-01 1.73594582e+00
3.53347063e-01 -2.76325673e-01 -9.59256440e-02 8.65933836e-01
-2.86053747e-01 2.41274267e-01 -3.53495985e-01 2.09157869e-01
1.24351943e+00 1.50589371e+00 -6.01966143e-01 -4.16407883e-01
-5.56127310e-01 6.91340089e-01 5.49711883e-01 -1.45157546e-01
-1.11858499e+00 -8.28992188e-01 7.19785750e-01 4.57023680e-01
6.73017263e-01 1.40363723e-01 -3.68281543e-01 -8.36473048e-01
2.72442520e-01 -9.03832912e-01 5.21706879e-01 -2.14494497e-01
-8.24968517e-01 5.46242595e-01 8.75893310e-02 -1.11265552e+00
-8.77587125e-02 -1.61433890e-01 -3.16989303e-01 9.85114336e-01
-1.45919728e+00 -9.16016638e-01 -4.73091304e-01 1.15303516e-01
1.47563651e-01 1.83271959e-01 9.86987293e-01 9.45502758e-01
-5.90792120e-01 8.81651640e-01 9.25654471e-02 2.82934874e-01
1.16385829e+00 -1.39554322e+00 1.04867685e+00 8.52246583e-01
3.53573203e-01 1.17583847e+00 6.67045474e-01 -5.79865158e-01
-1.35172236e+00 -1.10898519e+00 1.31678951e+00 -9.25678551e-01
4.99409169e-01 -6.39951110e-01 -1.17024028e+00 6.03013933e-01
-1.79916024e-01 8.34232718e-02 7.34842718e-01 4.14848685e-01
-6.30658150e-01 -3.10025901e-01 -1.15287232e+00 3.38248730e-01
9.84189034e-01 -3.93900245e-01 -6.05507016e-01 -1.51979439e-02
5.30003428e-01 -5.46703279e-01 -1.49917138e+00 4.43079323e-01
7.02434301e-01 -7.78904974e-01 1.05816114e+00 -2.55816251e-01
1.09024078e-01 -8.00794482e-01 8.42873566e-03 -8.10774028e-01
-2.48882234e-01 -5.79301596e-01 1.59943607e-02 1.68276036e+00
8.30171824e-01 -5.31397402e-01 5.98587394e-01 8.07755411e-01
2.23576263e-01 -6.53878987e-01 -7.60053694e-01 -8.17024589e-01
-1.32274687e-01 -1.38250872e-01 1.15118825e+00 1.06440878e+00
-3.82554904e-02 3.78832132e-01 1.55602887e-01 4.76961553e-01
3.86280030e-01 1.42480955e-01 1.19493592e+00 -1.33679867e+00
-2.64467001e-01 -3.40245098e-01 -4.68266815e-01 -7.94551730e-01
-2.92049855e-01 -8.34801137e-01 -2.49295756e-02 -1.46093202e+00
1.47997350e-01 -1.03880799e+00 -3.27202022e-01 5.83830059e-01
-3.95565480e-01 2.45716631e-01 1.51318878e-01 3.52884591e-01
-5.98464608e-01 -3.36293221e-01 8.55169296e-02 -5.36868051e-02
-4.92025703e-01 -3.27209651e-01 -5.64098597e-01 3.91270190e-01
4.24011022e-01 -8.89392316e-01 -1.84510574e-01 -4.37454551e-01
6.89868212e-01 -3.75141412e-01 5.64869791e-02 -1.09485221e+00
4.70062435e-01 2.57703841e-01 3.18636417e-01 -7.48538196e-01
1.29708126e-01 -8.36575568e-01 7.60433376e-01 3.44348758e-01
-3.09107780e-01 1.69339389e-01 3.66325438e-01 2.28024796e-01
-1.47363707e-01 -2.92860121e-01 5.35742104e-01 1.35878369e-01
-5.85452497e-01 -8.26603547e-02 1.29109040e-01 -7.32505322e-02
1.09365284e+00 -3.29526544e-01 -4.32551682e-01 3.30766469e-01
-4.30110484e-01 4.06549126e-01 8.35782588e-01 3.60614836e-01
4.95117810e-03 -9.97836888e-01 -8.02595973e-01 1.26884207e-01
2.94372141e-01 3.71647388e-01 -2.54210413e-01 7.94352591e-01
-8.49311113e-01 3.45236957e-01 1.86948135e-01 -6.13521814e-01
-1.68466079e+00 6.21726990e-01 1.51287019e-01 -6.68357730e-01
-6.93512499e-01 8.22294176e-01 -1.35920838e-01 -4.33307052e-01
1.68517753e-01 -2.67844439e-01 -9.54198986e-02 2.42207557e-01
6.94616318e-01 5.07568538e-01 5.16734064e-01 -4.49047446e-01
-7.88801014e-01 6.18684471e-01 -4.08824950e-01 1.15352474e-01
1.40531886e+00 1.38271391e-01 -1.62690401e-01 3.02698705e-02
1.22464538e+00 4.78736311e-01 -7.24980116e-01 -5.87379523e-02
7.51213729e-01 -4.28805262e-01 -1.56089932e-01 -5.95482886e-01
-7.85669684e-01 1.93502679e-01 6.62363648e-01 1.82526439e-01
9.80381191e-01 2.63736248e-02 8.11132669e-01 2.61466920e-01
5.13026655e-01 -9.18976843e-01 -8.43554914e-01 1.26687303e-01
3.53121966e-01 -1.22793829e+00 4.09503400e-01 -5.70250928e-01
-1.94608361e-01 8.13956022e-01 5.52166641e-01 -5.94027378e-02
3.27597290e-01 4.55916256e-01 1.18294396e-02 -3.22842598e-01
-7.58262515e-01 -1.39918223e-01 2.66667575e-01 1.54519707e-01
4.47644085e-01 -1.12121776e-01 -4.93444532e-01 5.41279137e-01
-3.89381647e-01 -8.93230364e-02 1.92811117e-01 9.38344657e-01
5.98010793e-03 -1.39570844e+00 -2.31829628e-01 4.63187605e-01
-6.15862846e-01 -1.90144360e-01 -3.08732539e-01 8.11927915e-01
2.88623810e-01 7.36325562e-01 3.62969309e-01 -3.84073079e-01
6.99005485e-01 -3.48096155e-02 1.88057005e-01 -7.27277756e-01
-1.21566844e+00 -2.20951870e-01 4.64061975e-01 -8.56277943e-01
-1.99902713e-01 -6.99975669e-01 -1.45882452e+00 -5.01083195e-01
-6.65137231e-01 3.52929533e-01 7.17665970e-01 6.11024320e-01
9.04670060e-01 2.52750188e-01 1.71675488e-01 -4.69315588e-01
-3.17854404e-01 -8.03387165e-01 4.54420410e-03 6.33123994e-01
1.56742007e-01 -4.51130092e-01 -1.61843628e-01 -1.59002915e-01]
|
[9.426286697387695, 8.43648910522461]
|
bd5a2ed7-5d82-4804-b765-6782f82885f8
|
building-an-effective-email-spam
|
2303.08792
| null |
https://arxiv.org/abs/2303.08792v1
|
https://arxiv.org/pdf/2303.08792v1.pdf
|
Building an Effective Email Spam Classification Model with spaCy
|
Today, people use email services such as Gmail, Outlook, AOL Mail, etc. to communicate with each other as quickly as possible to send information and official letters. Spam or junk mail is a major challenge to this type of communication, usually sent by botnets with the aim of advertising, harming and stealing information in bulk to different people. Receiving unwanted spam emails on a daily basis fills up the inbox folder. Therefore, spam detection is a fundamental challenge, so far many works have been done to detect spam using clustering and text categorisation methods. In this article, the author has used the spaCy natural language processing library and 3 machine learning (ML) algorithms Naive Bayes (NB), Decision Tree C45 and Multilayer Perceptron (MLP) in the Python programming language to detect spam emails collected from the Gmail service. Observations show the accuracy rate (96%) of the Multilayer Perceptron (MLP) algorithm in spam detection.
|
['Kazem Taghandiki']
|
2023-03-15
| null | null | null | null |
['spam-detection']
|
['natural-language-processing']
|
[-2.94018179e-01 -2.87021190e-01 3.03853691e-01 -4.55889523e-01
2.04009518e-01 -6.44988239e-01 6.86831534e-01 5.40174305e-01
-4.08542812e-01 5.54127574e-01 1.68850690e-01 -8.79919052e-01
5.75357527e-02 -1.06027055e+00 -1.89250819e-02 -3.93115550e-01
1.90517440e-01 8.31585824e-01 6.34756446e-01 -3.09115499e-01
7.51820445e-01 8.87352049e-01 -1.51810873e+00 7.94227183e-01
5.80498755e-01 9.03266430e-01 4.49416135e-03 8.34986567e-01
-8.89684141e-01 1.07113528e+00 -8.27950358e-01 -5.05811751e-01
2.36129880e-01 -9.05823857e-02 -1.08713257e+00 7.74640813e-02
-2.66713873e-02 -4.09598857e-01 -1.39273003e-01 1.05719578e+00
2.17721492e-01 1.35243788e-01 5.73443413e-01 -1.14334762e+00
-7.32329860e-02 5.56889355e-01 -1.98076487e-01 1.48853123e-01
4.04639930e-01 3.96830589e-03 4.35352743e-01 -4.29956704e-01
2.55965769e-01 1.88901198e+00 5.56479156e-01 3.80846679e-01
-1.14344621e+00 -6.58078134e-01 -2.67198682e-01 3.00765574e-01
-7.11805165e-01 -1.16667077e-02 4.04085159e-01 -5.35810411e-01
7.81359196e-01 5.44995666e-01 1.91640541e-01 8.96622241e-01
3.85399908e-01 6.78417206e-01 1.48584902e+00 -3.33484143e-01
3.97022605e-01 9.19174135e-01 1.03405070e+00 2.90376008e-01
3.01436931e-02 -2.90510617e-02 -1.86980665e-01 -8.91524792e-01
8.29105303e-02 3.41174781e-01 4.21478450e-01 4.20098126e-01
-5.77002347e-01 1.18426573e+00 4.33691144e-01 4.88305926e-01
-2.47643843e-01 -2.06321910e-01 7.13086605e-01 9.81934488e-01
3.55821013e-01 3.10018599e-01 -4.87320453e-01 -5.72097488e-03
-3.70481700e-01 1.90373898e-01 1.47422004e+00 5.81240535e-01
7.01095343e-01 -2.85897970e-01 9.22407880e-02 8.03327501e-01
6.34876907e-01 2.31012866e-01 8.13147247e-01 -4.09337610e-01
1.82881001e-02 1.30098057e+00 1.67123765e-01 -1.24275756e+00
-2.63987720e-01 -1.26220226e-01 -8.46009970e-01 2.14849368e-01
6.14627063e-01 -1.24729671e-01 -4.59074497e-01 4.94915307e-01
2.87957430e-01 -5.02484202e-01 -3.17575365e-01 6.23479545e-01
8.05443645e-01 6.74012303e-01 2.74253577e-01 7.49355853e-02
1.50499237e+00 -6.22088194e-01 -3.63496035e-01 -1.25478223e-01
5.28337002e-01 -1.30648327e+00 7.41477311e-01 4.63324696e-01
-4.21787411e-01 -4.02349889e-01 -2.83228904e-01 3.64062577e-01
-9.05243397e-01 -2.72597283e-01 5.02066970e-01 1.05028403e+00
-6.88470662e-01 9.19646680e-01 -4.45035875e-01 -9.42938328e-01
3.13684106e-01 5.20138323e-01 1.86620466e-02 2.68792152e-01
-1.11683643e+00 1.02177644e+00 3.42041701e-01 -3.72600853e-01
-2.65175968e-01 -2.48165309e-01 -1.58721842e-02 -8.39529783e-02
-1.88595951e-01 -1.69498309e-01 1.35220301e+00 -1.05620587e+00
-1.27480984e+00 1.12061143e+00 -4.61444706e-02 -7.79920936e-01
7.12253034e-01 3.61918770e-02 -4.69104975e-01 5.18717710e-03
3.16911493e-03 2.20431119e-01 1.10376251e+00 -9.78282630e-01
-6.59257591e-01 -6.96093857e-01 -3.68445516e-01 -3.72637659e-01
-3.56298655e-01 7.78778970e-01 5.59691906e-01 -3.63767087e-01
2.45727345e-01 -5.92884898e-01 -8.56680721e-02 -3.44775021e-01
-3.80595058e-01 -6.16809368e-01 1.39672959e+00 -9.97820616e-01
1.13052833e+00 -1.78805065e+00 -6.90450907e-01 5.19722700e-01
1.19347125e-01 5.78602612e-01 6.19416296e-01 7.75723457e-01
4.47395518e-02 1.87155791e-02 5.22782356e-02 2.28448078e-01
1.13754041e-01 2.60015011e-01 -4.94143933e-01 3.30952436e-01
-3.52851748e-01 5.13152719e-01 -6.08088613e-01 -6.24799430e-01
5.15801609e-01 2.77947374e-02 -6.24365062e-02 1.67669371e-01
-2.68594444e-01 2.80703962e-01 -6.05172575e-01 6.81087375e-01
8.33092690e-01 5.83738135e-03 8.19981620e-02 1.97887674e-01
-6.53319508e-02 3.53285909e-01 -1.11764407e+00 3.07888567e-01
-3.47973377e-01 5.03094256e-01 4.85175908e-01 -1.15522933e+00
1.23913646e+00 1.65048897e-01 -1.99647471e-02 -1.78122759e-01
5.93141496e-01 3.51341516e-01 -1.52483106e-01 -5.35680950e-01
3.09824627e-02 -3.60813439e-02 1.13608316e-01 8.43796492e-01
-2.43796244e-01 4.08495277e-01 1.70116916e-01 4.05465245e-01
1.39289260e+00 -5.66966176e-01 3.98292184e-01 -4.97421563e-01
1.13355637e+00 2.29001209e-01 -1.03369370e-01 9.80193615e-01
-5.23069382e-01 -1.21257417e-01 7.06009507e-01 -6.22490108e-01
-9.03305233e-01 -1.02402842e+00 -2.65135199e-01 1.31546843e+00
-1.47220358e-01 -4.55249511e-02 -6.13315463e-01 -8.98167431e-01
4.76362228e-01 8.50360215e-01 1.41299590e-01 7.84806311e-02
-5.32086194e-01 -7.64815271e-01 3.20990264e-01 -2.40673199e-01
7.65948594e-01 -1.35948598e+00 -6.43579289e-02 6.32776618e-01
4.83940318e-02 -6.86000824e-01 1.32146612e-01 1.91256076e-01
-1.20364046e+00 -9.54009831e-01 -2.79228896e-01 -8.95186663e-01
6.52059674e-01 3.09566557e-01 9.63856041e-01 3.36377293e-01
-4.46428359e-01 6.36320114e-02 -5.83983064e-01 -4.09098327e-01
-1.20908332e+00 6.70069903e-02 6.67306036e-02 2.67437190e-01
1.19466424e+00 -7.98753738e-01 -3.79466385e-01 7.70616055e-01
-7.83623338e-01 -4.34209049e-01 5.94345927e-01 5.44493854e-01
-7.09648728e-01 2.12104067e-01 8.45519841e-01 -1.08783317e+00
1.04580879e+00 -6.14026308e-01 -5.23097038e-01 -2.65448719e-01
-5.67289531e-01 -2.33864754e-01 1.00565970e+00 -1.81476325e-01
-9.31368113e-01 1.08061403e-01 -2.80941159e-01 5.32377064e-01
-8.40765417e-01 -3.05583119e-01 -4.63418737e-02 -3.25666964e-01
1.09600401e+00 3.06833804e-01 4.62895006e-01 -9.30761158e-01
4.38904837e-02 1.88832545e+00 -4.69689816e-02 -1.29378187e-02
7.47191608e-01 5.59767425e-01 -4.93945032e-01 -1.02786994e+00
-4.91261750e-01 -1.17681694e+00 -5.62454462e-01 -3.85981888e-01
4.15247560e-01 -1.51240230e-01 -1.29567993e+00 7.50494421e-01
-1.31764197e+00 5.55993199e-01 3.50640297e-01 5.33197671e-02
1.83902383e-01 7.63224840e-01 -1.07198942e+00 -1.09660792e+00
-7.43441641e-01 -5.90795457e-01 4.81873900e-01 -5.25795529e-03
-7.51782596e-01 -8.13455582e-01 -3.89083475e-01 7.43505597e-01
6.82981491e-01 -3.57160091e-01 1.03613031e+00 -1.43684936e+00
-3.30744028e-01 -9.09417689e-01 -5.81765056e-01 8.27551246e-01
5.53566962e-02 -3.05596441e-01 -8.61331642e-01 6.55214116e-02
3.21551621e-01 9.42688733e-02 7.47673452e-01 -8.82422924e-02
9.56869483e-01 -6.20249212e-01 -2.19041437e-01 -4.01106864e-01
1.09510338e+00 2.39754528e-01 6.50422394e-01 5.81201017e-01
1.27255484e-01 1.02418768e+00 1.18669309e-01 4.66599494e-01
-1.85799733e-01 5.14239110e-02 3.55567515e-01 6.40062332e-01
4.17445689e-01 1.47929296e-01 4.48237002e-01 4.62962478e-01
5.57035387e-01 2.34629065e-01 -7.57256448e-01 1.57631040e-01
-1.89667761e+00 -9.77683723e-01 -8.37150693e-01 1.97620261e+00
7.57809281e-01 3.08422893e-01 3.88499260e-01 5.15820324e-01
1.08964920e+00 -2.78109282e-01 -3.05946730e-02 -9.38418925e-01
3.49246800e-01 5.14168339e-03 6.98743880e-01 5.74002087e-01
-1.18997908e+00 9.68522906e-01 5.31653500e+00 9.72245097e-01
-7.36180127e-01 5.10072485e-02 5.53041518e-01 3.99353713e-01
3.44739854e-01 2.79329658e-01 -1.08146036e+00 9.68113244e-01
1.00892305e+00 6.07490428e-02 5.45350671e-01 1.17141247e+00
5.22710264e-01 -5.33668339e-01 -5.37481308e-01 7.83988118e-01
-1.58720955e-01 -1.01874387e+00 9.39629301e-02 -1.96907029e-01
1.79800555e-01 -1.96938545e-01 -5.75384021e-01 3.60178649e-01
4.90873754e-01 -8.80570352e-01 -9.74494889e-02 3.17277640e-01
-5.66235781e-01 -6.27521813e-01 9.86047029e-01 7.63561249e-01
-3.82245392e-01 -3.48206282e-01 -5.38109422e-01 -6.01967752e-01
1.39711529e-01 1.11543202e+00 -1.22561872e+00 -4.97588962e-02
8.97505462e-01 2.02694833e-01 -8.29591095e-01 1.11303675e+00
1.65984288e-01 9.57453609e-01 -4.77254927e-01 -8.97409201e-01
2.66125619e-01 -3.85780066e-01 7.74635971e-01 1.22077310e+00
-4.05462742e-01 -3.60450327e-01 3.81795540e-02 7.44875789e-01
1.14358760e-01 2.91518420e-01 -2.85226315e-01 2.62285676e-03
9.70687196e-02 1.31640351e+00 -9.37931299e-01 -4.72167075e-01
-1.07822925e-01 8.43599856e-01 -2.65354395e-01 -5.08678518e-03
-1.44822821e-01 -7.47901499e-01 3.84046823e-01 7.17750609e-01
-3.98065627e-01 -1.34053484e-01 -6.54023767e-01 -4.49009776e-01
-4.47841734e-02 -1.00064003e+00 3.82304102e-01 -3.75651360e-01
-1.89222419e+00 2.70538360e-01 -3.49229753e-01 -9.04363096e-01
-1.87923864e-03 -9.78066206e-01 -7.79871047e-01 9.85172272e-01
-6.87800109e-01 -7.27817655e-01 -2.98448026e-01 3.97001177e-01
4.27574068e-01 -7.37848341e-01 6.76053286e-01 4.03374553e-01
-1.28744647e-01 -2.31599584e-01 4.23711985e-01 2.82902122e-01
6.91731691e-01 -1.22770309e+00 4.28768516e-01 1.69409141e-01
-4.56043094e-01 6.41390443e-01 8.28546286e-01 -9.00174797e-01
-1.04601741e+00 -8.23251367e-01 1.43061006e+00 -6.39073193e-01
1.02354872e+00 -4.80773240e-01 -9.93727386e-01 3.14847350e-01
1.30689926e-02 -8.05954754e-01 4.24945533e-01 -2.58536935e-01
-1.18706241e-01 -3.13391805e-01 -1.48895812e+00 4.80801374e-01
2.47181296e-01 -2.79807895e-01 -7.28510559e-01 9.86578882e-01
2.71260291e-01 4.63345528e-01 -4.15855050e-01 -2.40219072e-01
4.16013926e-01 -1.39906323e+00 8.61132145e-01 -9.00778830e-01
5.60973287e-02 -6.68835342e-02 1.53744876e-01 -7.93141782e-01
7.10820705e-02 -6.78607404e-01 1.61660776e-01 1.23444474e+00
1.54197514e-01 -1.12796378e+00 1.04792058e+00 6.17832422e-01
4.87818569e-01 -3.54501531e-02 -6.84619248e-01 -4.77691412e-01
-1.25187058e-02 -4.74631578e-01 2.01029569e-01 6.37538075e-01
5.60147166e-02 4.06001627e-01 -2.29953855e-01 -2.44591415e-01
8.69259477e-01 -1.00819729e-01 8.41570914e-01 -1.72234178e+00
-2.88330186e-02 -4.26654309e-01 -6.75899029e-01 -1.01220095e+00
-3.87642235e-02 -8.36452603e-01 -5.03431380e-01 -1.29406774e+00
-1.67075574e-01 -2.99510270e-01 4.60558593e-01 1.62042409e-01
3.94120634e-01 -2.04088792e-01 -1.18020564e-01 5.15673280e-01
-2.97801793e-01 -1.51096612e-01 7.45038807e-01 -1.31498143e-01
-1.47791132e-01 1.05438411e+00 -4.91255552e-01 1.03866100e+00
1.21698928e+00 -8.18335772e-01 4.01694059e-01 3.33685040e-01
2.22909018e-01 -4.18824494e-01 6.50402427e-01 -6.41257644e-01
5.11152327e-01 6.87438771e-02 3.26504886e-01 -7.64773488e-01
-6.37144670e-02 -1.06607878e+00 -3.56156021e-01 9.52896476e-01
-1.63497850e-01 -2.61851937e-01 -3.35196227e-01 6.57556653e-01
-1.55464604e-01 -9.25897181e-01 1.17804348e+00 -6.71927571e-01
-4.96400923e-01 -3.74472886e-01 -1.23828292e+00 -4.03479874e-01
9.22151804e-01 -2.07933784e-01 -5.16528547e-01 -5.82476079e-01
-7.32153356e-01 3.11936319e-01 1.29141837e-01 5.43706834e-01
2.88450271e-01 -6.84605718e-01 -7.42296994e-01 2.87009865e-01
-1.36867225e-01 -7.50045240e-01 -1.15308464e-01 8.03944945e-01
-1.22140646e+00 5.26529431e-01 -2.83436954e-01 -4.15805191e-01
-1.61988103e+00 2.17511848e-01 1.90278694e-01 1.01540163e-02
-5.06433666e-01 5.45564115e-01 -6.90845728e-01 -1.11078274e+00
6.88241184e-01 3.01231205e-01 -4.84324396e-01 1.52432740e-01
7.72412777e-01 1.14797807e+00 3.37882578e-01 -3.52453113e-01
-3.28280658e-01 -1.59656107e-01 -3.41326296e-01 6.94408119e-02
7.75909603e-01 2.68607643e-02 -1.12372112e+00 1.01484291e-01
1.42343342e+00 -3.25666726e-01 8.65985900e-02 -1.60779968e-01
6.07113540e-01 -6.47196710e-01 -1.87961325e-01 -7.47697055e-01
-2.04931945e-01 7.13988006e-01 5.70741296e-01 1.04142368e+00
4.39721674e-01 2.11951826e-02 8.52557778e-01 8.58304024e-01
1.20507300e-01 -1.44765818e+00 -1.39548331e-01 6.63996100e-01
5.86581826e-01 -1.49570489e+00 -5.22799045e-02 -5.52178741e-01
-3.53395045e-01 1.23706555e+00 2.81337053e-01 -2.64922172e-01
1.05146444e+00 1.60337556e-02 1.03646040e-01 -1.74806770e-02
-5.19112289e-01 -4.81279455e-02 -2.16943488e-01 5.37539065e-01
5.04247129e-01 1.35846183e-01 -7.97462940e-01 3.35282594e-01
-3.08746755e-01 -3.06709319e-01 3.68094176e-01 1.04054296e+00
-1.32146537e+00 -1.17506742e+00 -9.68335152e-01 1.20778048e+00
-6.79080129e-01 -3.08088795e-03 -1.16941369e+00 2.14531228e-01
-3.27228040e-01 1.59902537e+00 1.05704956e-01 -2.58831084e-01
6.35518953e-02 3.67602021e-01 -2.88668483e-01 -4.25231397e-01
-1.44167006e+00 -4.92652088e-01 3.29449922e-01 -1.05840251e-01
1.05991870e-01 -3.73472124e-01 -1.02109194e+00 -9.78421926e-01
-2.24137694e-01 5.33415437e-01 1.02234948e+00 8.57610345e-01
5.28976768e-02 -1.85682312e-01 8.76937628e-01 -3.59670401e-01
-1.07085955e+00 -1.26797020e+00 -7.31009781e-01 5.74903846e-01
-2.37222448e-01 -5.67375831e-02 -9.23694491e-01 -1.32741764e-01]
|
[7.810455799102783, 10.020471572875977]
|
34caf033-d6b2-4436-b0b4-083082c562e2
|
keyphrase-generation-a-multi-aspect-survey
|
1910.05059
| null |
https://arxiv.org/abs/1910.05059v1
|
https://arxiv.org/pdf/1910.05059v1.pdf
|
Keyphrase Generation: A Multi-Aspect Survey
|
Extractive keyphrase generation research has been around since the nineties, but the more advanced abstractive approach based on the encoder-decoder framework and sequence-to-sequence learning has been explored only recently. In fact, more than a dozen of abstractive methods have been proposed in the last three years, producing meaningful keyphrases and achieving state-of-the-art scores. In this survey, we examine various aspects of the extractive keyphrase generation methods and focus mostly on the more recent abstractive methods that are based on neural networks. We pay particular attention to the mechanisms that have driven the perfection of the later. A huge collection of scientific article metadata and the corresponding keyphrases is created and released for the research community. We also present various keyphrase generation and text summarization research patterns and trends of the last two decades.
|
['Ondřej Bojar', 'Erion Çano']
|
2019-10-11
| null | null | null | null |
['keyphrase-generation']
|
['natural-language-processing']
|
[ 6.05510890e-01 2.41770208e-01 -4.65957016e-01 3.14726532e-01
-9.77090478e-01 -4.13117260e-01 9.56329525e-01 6.53461158e-01
-5.07887304e-01 1.20546293e+00 1.07948935e+00 -1.49442092e-01
-2.58676887e-01 -4.86621231e-01 -6.84056938e-01 -5.41750073e-01
6.18878938e-02 1.97341815e-01 -9.98867527e-02 -3.90249610e-01
1.09355307e+00 3.43831301e-01 -1.64995456e+00 4.97170508e-01
1.07024753e+00 6.82349443e-01 2.92895496e-01 1.12959993e+00
-6.91696107e-01 8.11981618e-01 -1.27362740e+00 -5.49298167e-01
-1.61848456e-01 -6.39104664e-01 -8.07901382e-01 -2.80380666e-01
7.84330428e-01 -2.81161278e-01 -6.49827778e-01 9.30934608e-01
7.74993420e-01 -1.40447021e-01 6.90804958e-01 -1.06144285e+00
-1.03002656e+00 1.26630712e+00 -4.27077830e-01 4.35676068e-01
6.60807252e-01 -3.24716896e-01 1.30156898e+00 -7.99908042e-01
8.32248807e-01 9.10870135e-01 2.64452845e-01 4.95017976e-01
-6.93770349e-01 -5.08635998e-01 -5.39060868e-02 4.16351736e-01
-1.04240263e+00 -3.89200956e-01 9.01869297e-01 -2.43624732e-01
1.12301970e+00 3.33804429e-01 8.76255155e-01 1.11962557e+00
7.24702954e-01 1.28860056e+00 7.36340344e-01 -7.28749335e-01
8.15856084e-02 -1.73466861e-01 2.56312847e-01 5.53596735e-01
6.37361825e-01 -3.11439008e-01 -9.82877791e-01 -2.52261281e-01
4.21030283e-01 -2.45137393e-01 -3.15060556e-01 1.49720266e-01
-1.62155676e+00 7.58238912e-01 -2.32061297e-01 3.69828969e-01
-6.12941742e-01 -6.95712492e-02 7.06338227e-01 3.65536690e-01
6.24822855e-01 1.12004268e+00 -4.60620910e-01 -3.40803623e-01
-1.49864149e+00 8.07234108e-01 1.22116327e+00 8.93302798e-01
3.05092156e-01 1.95538685e-01 -6.63708329e-01 6.19797111e-01
-2.02991083e-01 2.80843765e-01 7.02777207e-01 -8.59164059e-01
6.18378341e-01 6.12854838e-01 3.16605978e-02 -9.03134346e-01
-1.84186414e-01 -3.69801968e-01 -1.05949628e+00 -4.08016682e-01
-4.37116623e-02 -4.33696359e-01 -9.19579625e-01 9.90173697e-01
-1.75483406e-01 -2.53373384e-01 1.89109966e-01 6.23651892e-02
1.46861362e+00 1.06819832e+00 -3.21549028e-01 -5.31806469e-01
1.28481162e+00 -9.42934632e-01 -1.23824489e+00 1.72487587e-01
2.33307943e-01 -9.93473291e-01 4.31827605e-01 6.20030940e-01
-1.48518527e+00 -3.14221263e-01 -1.21017635e+00 -1.99495345e-01
-7.46349931e-01 1.96866468e-01 4.12407398e-01 1.96812913e-01
-1.12176406e+00 7.72961497e-01 -3.28499466e-01 -4.31004554e-01
4.51001614e-01 3.22736725e-02 -1.08446270e-01 3.11706245e-01
-1.37985754e+00 9.42947984e-01 7.66407073e-01 -2.76780516e-01
-4.61726964e-01 -1.04607713e+00 -4.81964827e-01 3.01480442e-02
5.69327652e-01 -1.00013447e+00 1.40929103e+00 -3.86870146e-01
-1.67783689e+00 5.24983108e-01 -4.69626859e-02 -6.62016869e-01
2.15278745e-01 -6.73966527e-01 -4.68323380e-01 4.85383689e-01
-2.61146743e-02 5.25174618e-01 7.24212408e-01 -9.19167697e-01
-9.42293644e-01 7.39633217e-02 -1.37956858e-01 2.37409934e-01
-5.15284956e-01 3.37795019e-01 -2.67920554e-01 -1.27238667e+00
-4.36560601e-01 -5.66017866e-01 -3.41432728e-02 -5.26970983e-01
-8.08175027e-01 -6.19816363e-01 5.05666912e-01 -9.29220200e-01
1.80698764e+00 -1.37851632e+00 4.89314198e-01 -4.11165982e-01
2.22053781e-01 4.20176566e-01 -3.25993448e-02 1.34663665e+00
-7.44784027e-02 3.02078754e-01 -2.84609765e-01 2.03539282e-02
8.80518407e-02 -2.81077176e-01 -7.82566369e-01 4.14670706e-02
-4.03170884e-02 1.08399260e+00 -8.96249533e-01 -5.30461371e-01
-1.83177188e-01 8.08121786e-02 -4.09142226e-02 3.15162033e-01
-4.04495418e-01 -1.59164965e-01 -5.63351274e-01 5.39079070e-01
1.88280940e-01 -1.37517005e-01 -2.65008032e-01 -1.82078913e-01
-4.85150725e-01 4.95656967e-01 -7.68289626e-01 1.69844961e+00
-1.97771322e-02 1.01015878e+00 -2.49786690e-01 -9.00021017e-01
6.87177122e-01 6.22044444e-01 7.19268024e-01 -3.20271999e-01
3.36496830e-02 3.28988492e-01 -2.59879023e-01 -4.10609305e-01
1.50056815e+00 2.71233499e-01 -1.58095479e-01 6.52217925e-01
1.12574607e-01 -3.53990078e-01 8.42109561e-01 7.62375116e-01
1.00372589e+00 1.92926422e-01 7.80003250e-01 -1.21283829e-01
5.29195368e-01 1.51650161e-01 3.86310704e-02 1.10199535e+00
2.78095186e-01 6.65219605e-01 4.99745995e-01 -3.60341519e-01
-1.39316285e+00 -6.70238912e-01 8.84892344e-02 7.93090463e-01
-3.24814230e-01 -9.18899119e-01 -8.78675342e-01 -4.26861137e-01
-1.28631279e-01 8.97228599e-01 -5.80389619e-01 -2.17066854e-01
-7.84933209e-01 -6.61908269e-01 9.35670614e-01 3.43560249e-01
4.11870390e-01 -1.38923633e+00 -5.33352315e-01 3.42213333e-01
-2.44039759e-01 -8.78155112e-01 -4.44252610e-01 3.15947570e-02
-8.27004135e-01 -5.32549083e-01 -1.48144031e+00 -6.04320884e-01
4.56421405e-01 1.29013792e-01 1.28415620e+00 -2.47830912e-01
-2.90697902e-01 4.31639016e-01 -7.32872248e-01 -8.52127135e-01
-7.00701952e-01 6.40458524e-01 -4.17849869e-02 -3.77026707e-01
2.31538847e-01 -4.40014422e-01 -2.52413779e-01 -8.27765465e-01
-1.10787928e+00 2.45698795e-01 1.19374907e+00 6.78475380e-01
2.43329585e-01 -2.43159473e-01 8.31864357e-01 -8.26973438e-01
1.34679639e+00 -3.31246555e-01 -2.23098919e-01 3.31126362e-01
-9.24965143e-01 4.55886215e-01 6.66378736e-01 -2.75280982e-01
-1.06264019e+00 -4.68696892e-01 -1.64614826e-01 9.21224952e-02
1.51749328e-02 8.94266903e-01 2.71399915e-01 4.27000046e-01
5.81889868e-01 8.86425555e-01 -2.93790907e-01 -6.49366736e-01
8.59075069e-01 9.37710762e-01 5.47725439e-01 -4.95090634e-01
8.12866211e-01 5.67331687e-02 -8.93251747e-02 -1.17576730e+00
-8.51435006e-01 -2.63018370e-01 -4.41421181e-01 -2.49660432e-01
6.50463879e-01 -7.68831551e-01 -3.57024640e-01 5.84875703e-01
-1.47946155e+00 3.03298861e-01 -5.40687263e-01 2.31506005e-01
-3.16165417e-01 7.52908289e-01 -7.39744425e-01 -4.40821946e-01
-1.21541452e+00 -9.56241727e-01 1.09273243e+00 6.29243910e-01
-5.59393167e-01 -6.99347317e-01 2.25543991e-01 -1.28217656e-02
4.21495974e-01 1.19175896e-01 1.04936004e+00 -7.32232511e-01
-3.52363288e-01 -3.04471433e-01 -1.40461490e-01 1.08421050e-01
6.31022081e-02 1.70738667e-01 -4.46373463e-01 -7.95136020e-02
-3.77293080e-01 -1.74586058e-01 1.41024888e+00 6.07025623e-01
1.07580554e+00 -8.25063050e-01 -3.46950352e-01 1.44715920e-01
1.07566142e+00 1.98025122e-01 6.67104959e-01 4.62316781e-01
7.09031761e-01 5.31196296e-01 2.02833444e-01 4.75342304e-01
5.13057053e-01 3.54800612e-01 -6.85292929e-02 1.44609511e-01
-2.68376172e-01 -4.03253376e-01 4.75461215e-01 1.41059351e+00
-3.35238546e-01 -5.41621745e-01 -5.15938342e-01 5.52767634e-01
-1.77613080e+00 -1.20402312e+00 -9.79284123e-02 1.78216720e+00
1.38263094e+00 1.76587999e-01 2.62470767e-02 2.47573834e-02
5.78376353e-01 7.48651266e-01 -2.62075812e-01 -5.79945207e-01
-3.30323279e-01 3.76309425e-01 5.74232817e-01 2.57582426e-01
-8.76839221e-01 9.44541454e-01 7.23842573e+00 1.00246775e+00
-9.43366766e-01 -5.61433375e-01 2.54995048e-01 -7.55480528e-02
-4.08261508e-01 -5.85009418e-02 -9.53775227e-01 2.83594042e-01
9.55799103e-01 -7.85602868e-01 2.94239409e-02 6.77539945e-01
7.74193332e-02 -2.93904841e-01 -8.33096981e-01 9.60121632e-01
4.51543689e-01 -2.04660106e+00 6.89956009e-01 2.85002254e-02
1.04580617e+00 -5.84361888e-02 -1.37309656e-01 1.29690945e-01
3.37496936e-01 -8.05695832e-01 6.06721580e-01 1.03071618e+00
4.85108525e-01 -8.00605237e-01 6.10853434e-01 2.62169719e-01
-6.62896931e-01 6.16256110e-02 -3.24630231e-01 -8.42549130e-02
1.81728780e-01 9.70102549e-01 -5.81846118e-01 1.01675963e+00
4.61277157e-01 9.29678380e-01 -5.27551949e-01 1.16331255e+00
-4.15765911e-01 7.05532610e-01 7.09485859e-02 -6.88127995e-01
2.21280620e-01 -1.36169776e-01 1.06801772e+00 1.69280219e+00
4.28481132e-01 -4.69389595e-02 4.15047817e-03 5.66075325e-01
-4.61169243e-01 2.77431667e-01 -4.75755334e-01 -7.34754622e-01
4.50044692e-01 1.29002190e+00 -6.28098309e-01 -8.14635873e-01
-1.90006658e-01 9.69332874e-01 -5.33034019e-02 3.42154622e-01
-3.34476620e-01 -1.02826858e+00 3.18716466e-01 -3.02168816e-01
2.96027154e-01 -2.88772047e-01 -1.55888990e-01 -1.26251554e+00
5.61479852e-02 -1.21427417e+00 2.20692337e-01 -7.30513453e-01
-1.20346689e+00 4.68451351e-01 2.77475864e-01 -9.30400789e-01
-4.40822661e-01 -2.20303193e-01 -5.79694152e-01 6.20737374e-01
-1.31650066e+00 -8.10474396e-01 2.09662035e-01 -1.91301107e-01
8.70105624e-01 -5.57169259e-01 8.66651475e-01 -4.86266091e-02
-4.65156049e-01 2.84167409e-01 5.44794917e-01 6.50684610e-02
7.78465986e-01 -1.39890897e+00 6.35938704e-01 8.46480370e-01
1.80625543e-01 8.80088687e-01 1.06474388e+00 -7.51220703e-01
-1.71496749e+00 -6.57056332e-01 1.41430151e+00 -4.66076374e-01
7.12723017e-01 -1.78715870e-01 -7.17435718e-01 2.44341061e-01
1.12382782e+00 -8.20694387e-01 4.06616807e-01 -2.14718685e-01
-5.63540272e-02 -6.16990440e-02 -3.62673163e-01 1.05766451e+00
6.18390441e-01 -3.37165803e-01 -1.21917868e+00 3.12030017e-01
9.42280293e-01 -5.04776657e-01 -7.49204516e-01 2.52195150e-01
7.56696165e-01 -4.88787562e-01 8.49962413e-01 -6.25111461e-01
9.76195931e-01 -1.68039382e-01 3.01692337e-01 -1.50495124e+00
-3.58374059e-01 -1.16964245e+00 -7.11321592e-01 1.17038918e+00
4.46705520e-01 -2.94415683e-01 5.94625771e-01 -1.83851928e-01
-2.80665725e-01 -9.22863543e-01 -5.56502461e-01 -5.16286373e-01
3.20364863e-01 -6.27728030e-02 4.46088672e-01 5.80381036e-01
3.03687990e-01 7.85275221e-01 -6.55519247e-01 -7.14464486e-01
4.84474599e-01 1.65629461e-01 6.96614563e-01 -1.27978778e+00
7.56019130e-02 -8.98999035e-01 6.63048401e-02 -1.24896967e+00
-7.90832285e-03 -9.46335793e-01 -1.32484198e-01 -2.20119405e+00
4.38784212e-01 5.98533571e-01 -5.69974035e-02 3.10571581e-01
-5.19661725e-01 -9.25626680e-02 1.06591068e-01 2.54240453e-01
-5.91656744e-01 5.88423908e-01 1.33005857e+00 -4.10289884e-01
-2.29828686e-01 -1.90733895e-01 -1.18146420e+00 4.06425893e-01
7.62401283e-01 -4.60956633e-01 -1.69972271e-01 -5.58577850e-02
6.99290216e-01 -8.91539678e-02 -6.14193454e-02 -8.28857422e-01
4.52072322e-01 -1.50663406e-01 4.36924547e-01 -1.08770359e+00
-4.55780253e-02 -1.85750995e-03 -1.61616713e-01 4.06911850e-01
-8.69341433e-01 3.38851899e-01 2.36657158e-01 4.60238010e-01
-3.24065059e-01 -3.62956613e-01 2.32696220e-01 -4.36108589e-01
-5.57552338e-01 7.35848546e-02 -8.18004489e-01 2.40289971e-01
4.91877168e-01 -6.80060834e-02 -2.68295139e-01 -5.29949903e-01
-6.48197383e-02 -9.11111981e-02 -9.35469568e-02 6.35813475e-01
6.79157674e-01 -8.99921417e-01 -1.44803786e+00 -4.57597584e-01
1.07954361e-01 -3.55777681e-01 -8.46041590e-02 5.81096590e-01
-4.97719526e-01 1.10349882e+00 -1.36186182e-01 8.39393288e-02
-1.15146053e+00 4.01734412e-01 -3.16674769e-01 -6.13795519e-01
-9.27843034e-01 5.04224241e-01 -2.08058119e-01 1.92874819e-02
2.76829362e-01 -3.55047494e-01 -7.27093577e-01 5.98121524e-01
9.60185170e-01 6.36635125e-01 1.30791068e-01 -3.23790520e-01
6.88352138e-02 3.02516997e-01 -6.41849339e-01 -1.59448072e-01
1.61277986e+00 4.01005559e-02 -3.98772717e-01 4.48814511e-01
1.04148126e+00 1.43654138e-01 -5.25953948e-01 -3.02875638e-01
2.20335200e-01 3.14472526e-01 1.94496572e-01 -1.03011692e+00
-5.74905813e-01 6.55295551e-01 -6.50023818e-02 4.24965113e-01
9.97658730e-01 -1.27206162e-01 1.10351205e+00 6.88698888e-01
-1.16182297e-01 -1.49799776e+00 9.26713571e-02 8.21483314e-01
1.03237426e+00 -6.20679259e-01 6.59098148e-01 1.30653813e-01
-3.44764918e-01 1.51523304e+00 -1.07051998e-01 -2.89804712e-02
2.31482953e-01 1.77615911e-01 -2.97636211e-01 -1.88811660e-01
-7.48275816e-01 4.70794030e-02 5.74533820e-01 2.84140050e-01
7.10064292e-01 -1.82358444e-01 -8.12318146e-01 4.24906582e-01
-5.68203509e-01 2.18394786e-01 9.71234083e-01 1.05355203e+00
-7.78320491e-01 -1.12871218e+00 -3.58868986e-01 8.35337877e-01
-9.48020399e-01 -5.19193053e-01 -8.82454216e-01 6.34551704e-01
-3.66112351e-01 7.96065509e-01 -3.21868479e-01 -1.73883602e-01
2.58358806e-01 2.98340261e-01 5.70267200e-01 -5.08280396e-01
-7.56507337e-01 -8.89992714e-02 2.51355827e-01 -5.76801179e-03
-4.52477008e-01 -6.30939722e-01 -1.01674795e+00 -4.44888622e-01
-1.68837532e-01 5.18691897e-01 6.13424718e-01 9.82954979e-01
5.64585268e-01 8.44224691e-01 3.43183935e-01 -9.51972604e-01
-4.80621845e-01 -1.41147029e+00 -3.70581329e-01 -2.14391574e-01
5.07467210e-01 1.55661013e-02 -7.59404227e-02 3.94163340e-01]
|
[12.46025276184082, 9.30049991607666]
|
2d5b0d61-8e95-4426-b706-33a4ffca85c0
|
scfusion-real-time-incremental-scene
|
2010.13662
| null |
https://arxiv.org/abs/2010.13662v3
|
https://arxiv.org/pdf/2010.13662v3.pdf
|
SCFusion: Real-time Incremental Scene Reconstruction with Semantic Completion
|
Real-time scene reconstruction from depth data inevitably suffers from occlusion, thus leading to incomplete 3D models. Partial reconstructions, in turn, limit the performance of algorithms that leverage them for applications in the context of, e.g., augmented reality, robotic navigation, and 3D mapping. Most methods address this issue by predicting the missing geometry as an offline optimization, thus being incompatible with real-time applications. We propose a framework that ameliorates this issue by performing scene reconstruction and semantic scene completion jointly in an incremental and real-time manner, based on an input sequence of depth maps. Our framework relies on a novel neural architecture designed to process occupancy maps and leverages voxel states to accurately and efficiently fuse semantic completion with the 3D global model. We evaluate the proposed approach quantitatively and qualitatively, demonstrating that our method can obtain accurate 3D semantic scene completion in real-time.
|
['Federico Tombari', 'Nassir Navab', 'Keisuke Tateno', 'Shun-Cheng Wu']
|
2020-10-26
| null | null | null | null |
['3d-semantic-scene-completion']
|
['computer-vision']
|
[ 4.20580477e-01 1.43339485e-01 7.16341883e-02 -3.85906935e-01
-5.43558180e-01 -7.49832988e-01 5.71036160e-01 2.92570829e-01
-3.59731525e-01 4.09088910e-01 1.52947217e-01 -3.30901563e-01
-7.31485263e-02 -9.23358679e-01 -9.42132771e-01 -9.10533592e-02
4.18700963e-01 8.48176181e-01 2.29514405e-01 1.77320316e-02
4.21012580e-01 7.02519596e-01 -1.74077237e+00 3.42836529e-02
9.08615530e-01 8.77061903e-01 6.68667197e-01 5.16207337e-01
-2.85812050e-01 5.48690200e-01 1.70748644e-02 1.60932288e-01
4.42945510e-01 1.05210207e-01 -5.05626559e-01 2.80141324e-01
3.56935859e-01 -7.39484787e-01 -5.87631702e-01 9.22381103e-01
2.00998113e-01 2.23151132e-01 3.83857220e-01 -7.81956196e-01
5.45825437e-02 -1.17467813e-01 -4.09478098e-01 -3.96473825e-01
6.14919841e-01 8.26080218e-02 6.71911538e-01 -9.46933508e-01
6.63893998e-01 1.14856982e+00 6.46943152e-01 2.20762804e-01
-1.29539537e+00 -3.17747980e-01 4.62077200e-01 -8.02392289e-02
-1.20810258e+00 -4.41157222e-01 9.86654043e-01 -4.07482803e-01
1.03249371e+00 1.50220245e-01 8.79653096e-01 8.13835382e-01
4.72167060e-02 6.54246747e-01 9.29521203e-01 -1.26831725e-01
6.74805820e-01 -2.40967155e-01 -1.47670820e-01 5.88780761e-01
2.50198722e-01 1.66946933e-01 -6.26972377e-01 6.08585514e-02
1.10970962e+00 1.84100851e-01 -1.43866614e-01 -1.00264478e+00
-1.20800805e+00 5.15431345e-01 5.75912118e-01 -2.39721626e-01
-4.99599844e-01 3.32427859e-01 4.50078622e-02 -7.16105402e-02
5.02350330e-01 3.92303735e-01 -3.68975222e-01 -1.44926697e-01
-6.81967854e-01 3.61193627e-01 5.23510098e-01 1.00292540e+00
1.06318009e+00 -1.27047881e-01 4.21198845e-01 5.19814372e-01
4.03253615e-01 5.97366333e-01 2.44126096e-03 -1.27738166e+00
4.49940413e-01 7.02074766e-01 4.42008555e-01 -8.69857013e-01
-5.89990795e-01 -3.61615598e-01 -5.57413340e-01 4.11522418e-01
2.74914593e-01 4.56025332e-01 -1.07733464e+00 1.53719556e+00
6.67469740e-01 3.78343642e-01 -1.38763055e-01 1.11690259e+00
4.80718732e-01 3.52546871e-01 -7.31746778e-02 1.06526725e-01
9.18193758e-01 -7.33919442e-01 -3.08351278e-01 -5.87071359e-01
5.02184153e-01 -5.16258299e-01 8.44920456e-01 3.82196486e-01
-9.26542640e-01 -3.53349805e-01 -1.03891563e+00 -4.70894814e-01
-2.17529852e-02 -2.64801711e-01 8.06083918e-01 3.54529262e-01
-1.03746045e+00 5.56113482e-01 -1.23618972e+00 -3.05809379e-01
3.66005719e-01 3.52387369e-01 -5.35344660e-01 -3.98626059e-01
-5.35563290e-01 8.83394063e-01 2.71457970e-01 2.92343140e-01
-8.36321175e-01 -6.78735197e-01 -1.08063483e+00 -1.33718839e-02
4.21834081e-01 -1.03945267e+00 1.25473571e+00 -3.40670377e-01
-1.44486535e+00 7.20876276e-01 -3.54557693e-01 -3.36783201e-01
5.32537937e-01 -4.00755465e-01 2.31306553e-01 2.79213078e-02
9.92237478e-02 7.24792242e-01 4.25544739e-01 -1.57667232e+00
-4.52251852e-01 -9.03100073e-01 4.31740999e-01 6.13328397e-01
1.56658962e-01 -9.39357102e-01 -5.87114513e-01 -1.79399513e-02
1.05563641e+00 -8.37983191e-01 -6.84307992e-01 4.54350859e-01
-3.35808277e-01 5.19137502e-01 5.90950966e-01 -6.22588992e-01
4.52742934e-01 -1.86306572e+00 3.46645653e-01 2.03450456e-01
2.52209157e-01 -2.04913616e-01 -4.20825742e-02 2.23586857e-01
3.88603777e-01 -3.23970944e-01 -4.54318523e-01 -9.31011260e-01
2.30810139e-02 5.19647777e-01 -4.29422617e-01 5.42908549e-01
-4.29908298e-02 8.49916875e-01 -1.05315578e+00 -2.20294029e-01
9.12236392e-01 7.58812129e-01 -8.57640207e-01 1.54931292e-01
-5.53138912e-01 9.76348162e-01 -6.81671679e-01 6.05316103e-01
7.00933754e-01 -9.35195535e-02 2.71801263e-01 -2.15445995e-01
-3.56019318e-01 5.12423337e-01 -1.03461409e+00 2.67229843e+00
-8.65327656e-01 4.04676378e-01 1.54427245e-01 -8.04813087e-01
8.24426651e-01 2.46897973e-02 6.97222173e-01 -9.69894767e-01
1.44653916e-01 3.90479088e-01 -6.38588846e-01 -8.98782462e-02
8.94029915e-01 -9.23015270e-03 -6.25193119e-02 3.49926263e-01
-2.63725579e-01 -7.06457078e-01 -5.33912241e-01 1.10756502e-01
1.17027724e+00 7.78802574e-01 2.34071136e-01 1.05309077e-01
1.63960978e-01 2.65684992e-01 3.92306983e-01 7.40140140e-01
9.19597298e-02 7.32069373e-01 3.31878550e-02 -4.83022630e-01
-1.35402966e+00 -1.27785659e+00 1.39640905e-02 2.79151589e-01
7.09822774e-01 -1.07530445e-01 -4.83431220e-01 -2.94473797e-01
1.71711907e-01 6.87280476e-01 -3.10368836e-01 -2.20511928e-02
-5.41461706e-01 -2.60704696e-01 -1.28514022e-01 3.30769151e-01
2.51450300e-01 -7.37416446e-01 -1.14212811e+00 4.86250669e-01
-2.53112823e-01 -1.43734848e+00 5.12888618e-02 1.87330037e-01
-1.30502605e+00 -1.03443301e+00 -2.22162738e-01 -3.30158293e-01
7.45367944e-01 7.33017802e-01 9.28282559e-01 -4.18523513e-02
-3.12860459e-01 4.22229558e-01 -2.24674225e-01 1.04656518e-02
-2.43083462e-01 2.30887104e-02 -3.41039035e-04 -1.91540495e-01
-1.68219537e-01 -1.02364063e+00 -6.97565377e-01 1.61433175e-01
-8.34233403e-01 7.07553864e-01 3.92444521e-01 5.69914162e-01
9.37658072e-01 -2.02896416e-01 2.32777238e-01 -8.05529833e-01
-5.91454022e-02 -3.61365676e-01 -9.80906010e-01 -2.03113854e-01
-4.25473064e-01 1.56595871e-01 3.82458985e-01 2.96887066e-02
-1.05057251e+00 7.13764250e-01 -3.15665483e-01 -7.34657884e-01
-2.22793758e-01 4.27740961e-01 -2.75352210e-01 -1.45332783e-01
4.26083207e-01 1.70024246e-01 -4.25606817e-02 -6.43935442e-01
4.86842901e-01 3.95679593e-01 7.59698093e-01 -5.13543785e-01
6.49729550e-01 9.80660319e-01 2.73933351e-01 -5.79713702e-01
-9.19463754e-01 -5.45202553e-01 -8.77320707e-01 -2.57476628e-01
5.82317591e-01 -1.15249825e+00 -6.68277919e-01 4.61507291e-01
-1.31806803e+00 -5.57502806e-01 -1.79083928e-01 6.41427040e-01
-9.28013086e-01 4.41697270e-01 -2.98460186e-01 -9.72912908e-01
9.19034109e-02 -1.12134171e+00 1.45503509e+00 -1.89758558e-02
-7.17073679e-02 -6.85665369e-01 -8.87767505e-03 3.88663054e-01
9.86112505e-02 4.80061561e-01 6.60012484e-01 1.61643848e-01
-1.26990592e+00 -1.77805692e-01 -3.27684671e-01 -3.52899969e-01
9.20552462e-02 -5.58140635e-01 -1.25699675e+00 -3.57354470e-02
-5.97554967e-02 -1.64219722e-01 6.92169130e-01 3.17074269e-01
1.05861366e+00 1.33071110e-01 -3.47451627e-01 8.59695435e-01
1.56174397e+00 2.34469946e-04 5.92395902e-01 3.00801486e-01
8.52028310e-01 6.24625325e-01 7.23330021e-01 5.33788383e-01
7.52161801e-01 7.86565363e-01 9.77444589e-01 1.19290188e-01
-1.42704248e-01 -6.77774847e-01 -1.81280315e-01 6.78870916e-01
2.89578974e-01 -1.46978810e-01 -9.57755446e-01 4.92714435e-01
-2.07279754e+00 -3.87792647e-01 -3.05122999e-03 2.42405581e+00
3.23082864e-01 1.48858249e-01 -5.02955556e-01 6.50082901e-02
3.37033868e-01 8.95473212e-02 -9.84490037e-01 2.45553572e-02
6.22078665e-02 1.08570620e-01 6.71344817e-01 8.21693778e-01
-7.94492245e-01 1.04645836e+00 5.55078125e+00 1.78881049e-01
-1.03120685e+00 2.05925614e-01 2.85835296e-01 -1.47628874e-01
-7.61243761e-01 2.89943427e-01 -3.23459059e-01 6.09335192e-02
5.54460704e-01 1.67938024e-01 7.33377337e-01 7.71255851e-01
3.84852767e-01 -4.15042758e-01 -1.16454649e+00 1.22400844e+00
-1.71746314e-01 -1.27557302e+00 -4.30937260e-02 3.29880565e-01
6.76009417e-01 2.87938803e-01 -1.53525367e-01 -9.26082507e-02
3.17558795e-01 -7.26458251e-01 1.16216874e+00 4.52881634e-01
8.04970384e-01 -5.62201023e-01 3.09152961e-01 6.70129597e-01
-1.06011653e+00 -2.43488457e-02 -2.51808345e-01 -4.03581738e-01
6.52030349e-01 8.76586378e-01 -8.86688292e-01 6.53114498e-01
3.32347184e-01 6.69303536e-01 -1.00441217e-01 1.16646242e+00
-1.44000992e-01 -1.29481882e-01 -6.67079210e-01 2.85861939e-01
9.76620764e-02 -1.90859094e-01 5.55717885e-01 3.81064862e-01
4.62511212e-01 1.17066219e-01 2.31634811e-01 1.02566230e+00
1.63209900e-01 -2.37414628e-01 -8.57744992e-01 1.87548116e-01
6.43463135e-01 9.09308255e-01 -1.04336202e+00 -1.22111998e-01
-2.36796215e-01 1.10828030e+00 4.90060627e-01 3.27533334e-01
-6.01924360e-01 1.94694087e-01 6.64427996e-01 2.17278212e-01
9.89278480e-02 -9.32902634e-01 -1.00097811e+00 -1.13835502e+00
3.17853779e-01 -8.56285915e-02 -2.39281133e-01 -9.45087492e-01
-7.50802219e-01 3.69197845e-01 -1.74979225e-01 -1.18572211e+00
-2.45800495e-01 -4.69813347e-01 -2.49282867e-02 6.91267312e-01
-1.70267963e+00 -1.02163899e+00 -7.12592125e-01 2.90223926e-01
4.66429323e-01 4.86957908e-01 8.28288555e-01 2.65995324e-01
1.12266941e-02 -1.10853858e-01 -1.34139553e-01 -4.75898564e-01
1.88588396e-01 -9.83492315e-01 8.93448889e-01 6.49848998e-01
2.24387720e-01 2.85497546e-01 7.11014748e-01 -7.06608295e-01
-1.87225485e+00 -9.94616389e-01 4.47238117e-01 -5.33894837e-01
1.83379918e-01 -6.02865219e-01 -7.36950457e-01 5.24137318e-01
-7.74884999e-01 1.01362139e-01 -3.42230089e-02 1.04263328e-01
-4.20305192e-01 1.05196662e-01 -1.19110584e+00 6.25392854e-01
1.71439898e+00 -7.18330145e-01 -1.59495607e-01 2.61564285e-01
9.04600918e-01 -9.74047720e-01 -5.09370208e-01 4.40724969e-01
7.14261711e-01 -1.21850049e+00 1.11536944e+00 3.33381221e-02
2.98032135e-01 -5.15344501e-01 -4.10493731e-01 -9.72823083e-01
8.46478194e-02 -2.71794379e-01 -3.77368122e-01 3.53431106e-01
6.09569065e-02 -6.95087194e-01 1.09602273e+00 7.00988233e-01
-6.04462802e-01 -4.06561941e-01 -1.22946644e+00 -4.30303305e-01
-4.66230035e-01 -9.65268075e-01 7.08245099e-01 5.61154842e-01
-3.45410436e-01 1.15430444e-01 -2.28058934e-01 5.94119787e-01
7.54680157e-01 4.27678496e-01 9.96272385e-01 -1.32770038e+00
-1.95453897e-01 -1.55196205e-01 -6.66596889e-01 -1.73655307e+00
2.41320327e-01 -6.17828727e-01 5.41705489e-01 -1.83228779e+00
-1.20112240e-01 -9.27485764e-01 7.66920894e-02 3.05288553e-01
1.84210807e-01 2.97738314e-01 1.78354718e-02 2.88436860e-01
-5.05900800e-01 7.99233794e-01 1.11820960e+00 1.31524563e-01
-3.55878472e-01 -2.90699780e-01 -3.61805737e-01 6.35425389e-01
4.54627484e-01 -2.08087444e-01 -5.86829066e-01 -8.33279788e-01
2.94854730e-01 3.76190215e-01 7.35520601e-01 -1.12545180e+00
2.93392062e-01 -2.80975163e-01 2.79098213e-01 -8.91411781e-01
1.02188718e+00 -9.97999370e-01 3.22671264e-01 2.42368370e-01
7.61032999e-02 -2.40239426e-01 1.86458826e-01 8.16322565e-01
1.24881133e-01 3.19245681e-02 3.91194761e-01 -2.38449648e-01
-9.31073666e-01 5.22040784e-01 8.23486075e-02 -4.52801466e-01
7.17590511e-01 -4.57540065e-01 5.26984446e-02 -3.71031642e-01
-6.73828840e-01 1.87773883e-01 1.12242150e+00 3.89053255e-01
8.18971097e-01 -1.09229863e+00 -2.49530613e-01 4.25030053e-01
1.98946744e-01 8.60850275e-01 6.18403137e-01 5.85667849e-01
-8.04316640e-01 5.19584060e-01 1.35462349e-02 -9.77558613e-01
-6.68974221e-01 4.40892071e-01 4.02737141e-01 -8.17138031e-02
-1.05630255e+00 3.92037481e-01 5.59680521e-01 -7.47754216e-01
7.93547034e-02 -4.01340097e-01 3.28858554e-01 -6.30143583e-01
1.37826398e-01 8.84381384e-02 2.32045144e-01 -5.69839239e-01
-1.50442258e-01 6.78121388e-01 2.98214346e-01 -4.89109457e-01
1.22490931e+00 -5.75543284e-01 1.00727491e-01 5.34692168e-01
9.36077058e-01 -3.06672186e-01 -1.74811578e+00 -2.81588554e-01
-1.94705337e-01 -8.38195264e-01 1.79553553e-01 -4.99905318e-01
-6.93381131e-01 7.84015119e-01 3.53424728e-01 -3.74516338e-01
8.07406306e-01 -8.70979428e-02 8.95438612e-01 5.20438910e-01
1.05558503e+00 -8.11650157e-01 -9.71823707e-02 6.05388999e-01
5.75973928e-01 -1.10969496e+00 6.34035766e-02 -7.23285615e-01
-5.88884130e-02 9.18466270e-01 4.59156245e-01 -1.29959196e-01
4.80514437e-01 1.85992301e-01 -1.60707995e-01 -1.73347339e-01
-5.30038774e-01 -6.72701076e-02 -5.42773679e-02 5.58544159e-01
-5.20634390e-02 1.03602290e-01 3.24103117e-01 1.72937512e-01
-2.34784633e-01 1.02233700e-02 4.60472614e-01 1.21572769e+00
-5.27469456e-01 -9.40712392e-01 -3.08075935e-01 1.98439956e-01
2.92075425e-01 1.96937412e-01 -7.47336149e-02 3.97455633e-01
-9.05864909e-02 6.90774381e-01 3.71897936e-01 -3.86128187e-01
3.29123557e-01 -9.48916897e-02 8.06577086e-01 -6.94363952e-01
1.43882066e-01 4.12120782e-02 -3.38836797e-02 -1.01525044e+00
-1.58531040e-01 -5.00066459e-01 -1.49395216e+00 -1.18016899e-01
-8.78750160e-02 -3.15487981e-01 1.23836565e+00 9.95803118e-01
6.11676812e-01 4.27859694e-01 5.42086899e-01 -1.40491068e+00
-2.16568589e-01 -5.71972728e-01 -4.02112693e-01 2.22361013e-01
5.45046151e-01 -9.10654843e-01 -4.53220718e-02 -3.43487501e-01]
|
[8.513659477233887, -2.8795320987701416]
|
97abf50d-700e-48db-8c13-5e86c494b5ea
|
structured-autocorrelation-matrix-estimation
|
2008.12369
| null |
https://arxiv.org/abs/2008.12369v1
|
https://arxiv.org/pdf/2008.12369v1.pdf
|
Structured Autocorrelation Matrix Estimation for Coprime Arrays
|
A coprime array receiver processes a collection of received-signal snapshots to estimate the autocorrelation matrix of a larger (virtual) uniform linear array, known as coarray. By the received-signal model, this matrix has to be (i) Positive-Definite, (ii) Hermitian, (iii) Toeplitz, and (iv) its noise-subspace eigenvalues have to be equal. Existing coarray autocorrelation matrix estimates satisfy a subset of the above conditions. In this work, we propose an optimization framework which offers a novel estimate satisfying all four conditions. Numerical studies illustrate that the proposed estimate outperforms standard counterparts, both in autocorrelation matrix estimation error and Direction-of-Arrival estimation.
|
['Panos P. Markopoulos', 'Dimitris G. Chachlakis']
|
2020-08-27
| null | null | null | null |
['direction-of-arrival-estimation']
|
['audio']
|
[ 3.07629079e-01 -4.14471596e-01 2.61312813e-01 1.72489986e-01
-6.62909210e-01 -1.03899550e+00 1.61745578e-01 -4.19886857e-01
-4.53684032e-02 5.55383682e-01 1.16336152e-01 -5.82864344e-01
-5.71385324e-01 -1.54087469e-01 -2.51301229e-01 -9.98581946e-01
-5.52752495e-01 -7.53921270e-02 -4.17109877e-01 2.01058850e-01
-1.64124817e-02 7.27890193e-01 -7.34396458e-01 -4.71300304e-01
6.75089538e-01 1.27976918e+00 -7.44691864e-02 1.00357437e+00
6.25978649e-01 5.99739552e-01 -4.67797041e-01 -5.66010848e-02
5.23078620e-01 -5.91017008e-01 2.15658963e-01 3.21320742e-01
2.58610900e-02 -1.34894565e-01 -4.41332012e-01 1.21872628e+00
5.13362288e-01 -1.32648021e-01 9.79451358e-01 -1.12999237e+00
-4.37477827e-01 3.67325008e-01 -8.27696204e-01 1.36024863e-01
1.77189127e-01 -3.53975713e-01 6.78823769e-01 -1.41477501e+00
1.08487103e-02 8.51166725e-01 7.67272949e-01 -2.58803666e-01
-9.67623174e-01 -6.99543715e-01 -3.75788569e-01 -2.61907995e-01
-2.09432602e+00 -8.12694609e-01 6.10455573e-01 -3.67248386e-01
4.56983715e-01 7.74233878e-01 5.99918365e-01 6.76284909e-01
4.37950134e-01 2.61545420e-01 1.02078128e+00 -4.44055080e-01
2.08302364e-01 -2.20941484e-01 1.49885476e-01 4.70150739e-01
8.50717306e-01 1.62501380e-01 -9.21898708e-02 -5.33083558e-01
9.96759415e-01 -1.53196171e-01 -4.57613349e-01 -3.30391258e-01
-1.47540236e+00 4.54936475e-01 -6.62368014e-02 5.61609566e-01
-6.33797646e-01 2.01796636e-01 -1.29017442e-01 4.84321207e-01
1.44588307e-01 3.47690761e-01 8.06019530e-02 7.51778111e-02
-7.47217953e-01 -1.71243057e-01 1.27419996e+00 1.23813951e+00
6.92369401e-01 7.62845635e-01 -3.05638555e-02 5.11877477e-01
5.55711687e-01 1.60660505e+00 3.99078205e-02 -5.89833081e-01
5.47551095e-01 -2.43707314e-01 5.86009026e-01 -1.58293319e+00
-5.77007711e-01 -1.33713138e+00 -1.58235407e+00 -3.22421402e-01
3.68313640e-01 -8.02087009e-01 -5.06131589e-01 1.21248055e+00
-4.24780548e-02 6.81025803e-01 3.54094476e-01 1.06784058e+00
1.77732706e-01 7.93160737e-01 -6.69574618e-01 -8.42804968e-01
9.22423244e-01 -4.46101278e-01 -1.00143600e+00 -5.01129568e-01
-4.35581058e-02 -1.28436100e+00 -5.32731824e-02 3.31206799e-01
-1.00972438e+00 -3.06996077e-01 -1.31660700e+00 8.76537204e-01
6.84684157e-01 5.59673846e-01 2.83143789e-01 1.20154178e+00
-1.15941286e+00 -3.07850182e-01 -3.65751356e-01 2.02657744e-01
-6.29349172e-01 1.14278547e-01 -2.41987079e-01 -1.83209926e-02
-7.11226046e-01 3.30066711e-01 -2.22179502e-01 7.13222921e-01
-6.48463905e-01 -3.33555907e-01 -7.50322878e-01 2.86999166e-01
1.92608625e-01 -6.43270016e-01 8.77610028e-01 -9.32570875e-01
-1.20024335e+00 2.10422963e-01 -3.94241363e-01 -9.55132842e-02
2.22338855e-01 -2.19808370e-01 -8.47407401e-01 1.76639006e-01
-7.12048933e-02 -8.34553659e-01 1.35975134e+00 -1.46468353e+00
-3.86700183e-01 -3.32877547e-01 -8.41845751e-01 7.93278441e-02
-1.48299700e-02 1.21186087e-02 -4.14938897e-01 -8.13292623e-01
8.62467349e-01 -1.09755635e+00 -7.57399857e-01 -6.94419026e-01
-6.96883857e-01 5.14671981e-01 3.00763071e-01 -5.79489529e-01
1.75753498e+00 -2.48028517e+00 -2.13399649e-01 1.05972493e+00
2.35746473e-01 3.50419357e-02 -2.37833887e-01 1.12666523e+00
-2.64512271e-01 -2.62162060e-01 8.16521496e-02 4.12881225e-02
-3.23828757e-01 -3.99576247e-01 -4.43346709e-01 1.16850328e+00
-2.11181387e-01 4.53631014e-01 -8.12306643e-01 -4.09888960e-02
1.98459439e-02 2.51464009e-01 -3.77126187e-01 2.38834634e-01
6.97897792e-01 4.78228450e-01 -7.16032624e-01 7.86691010e-01
1.13685703e+00 -1.95944294e-01 4.66920108e-01 -3.30382347e-01
-5.22795975e-01 -6.01013958e-01 -1.94001257e+00 6.79239511e-01
-4.62633967e-01 8.43327582e-01 5.01660466e-01 -6.89965904e-01
1.02270794e+00 3.90799940e-01 6.27433300e-01 -8.27427804e-02
2.57322073e-01 5.98033249e-01 4.66668546e-01 -3.05656314e-01
3.69811594e-01 4.93843518e-02 5.36388718e-02 5.42587459e-01
-3.13491225e-01 2.31904864e-01 2.72826534e-02 2.21419483e-01
1.13808000e+00 -9.54073250e-01 1.00702667e+00 -5.28225362e-01
7.28542209e-01 -5.98382413e-01 5.79483211e-01 1.17713308e+00
8.98100361e-02 5.94053507e-01 2.37831116e-01 3.08701098e-01
-9.45053220e-01 -1.07605481e+00 2.46609058e-02 2.92986423e-01
1.77369267e-01 -7.02472553e-02 -3.82616967e-01 3.50127965e-01
-1.12219632e-01 3.50807041e-01 -6.81817606e-02 3.41692626e-01
-6.35154009e-01 -5.62394261e-01 5.05091488e-01 2.36246228e-01
3.98264617e-01 1.05446920e-01 -2.90288717e-01 3.49600434e-01
-4.58192676e-02 -1.22897673e+00 -7.38310516e-01 9.15508717e-02
-4.48017329e-01 -1.03372741e+00 -9.70197916e-01 -8.51058304e-01
1.02888203e+00 1.06994784e+00 6.98799431e-01 -8.62841308e-02
1.22831881e-01 9.73232567e-01 -4.23672915e-01 -2.61958212e-01
-4.70218575e-03 -5.43652236e-01 2.48223811e-01 8.19644451e-01
-2.55881935e-01 -7.20506072e-01 -6.69068992e-01 6.89907134e-01
-6.17103219e-01 -4.00734812e-01 1.11799002e+00 8.85056973e-01
4.74945664e-01 3.71112347e-01 4.62180167e-01 -3.58404160e-01
8.68936360e-01 -5.45175672e-01 -9.67607439e-01 2.53086895e-01
-4.23358649e-01 -3.30789268e-01 4.97841388e-01 -6.33936167e-01
-1.03167927e+00 2.50982881e-01 1.56351864e-01 -2.56570399e-01
4.38926548e-01 9.84879732e-01 -2.32995898e-01 -3.73628587e-01
5.43833911e-01 4.07421619e-01 -2.32554540e-01 -3.63419913e-02
2.11085141e-01 6.98375344e-01 6.45839572e-01 -7.75594562e-02
1.47743189e+00 4.00841802e-01 4.99750644e-01 -1.41124880e+00
-3.59294623e-01 -1.10759687e+00 -5.03561139e-01 -3.62939030e-01
1.70733437e-01 -1.05335855e+00 -7.65699387e-01 4.48392034e-01
-1.03476167e+00 1.10852428e-01 4.12300557e-01 1.09475291e+00
-1.39985949e-01 5.39879858e-01 -4.46117193e-01 -1.57098329e+00
-3.02644134e-01 -5.77606320e-01 6.43109560e-01 -2.35374607e-02
-1.21013504e-02 -8.32133830e-01 5.72890341e-02 -5.00650465e-01
5.38684726e-01 8.28442723e-02 2.69940168e-01 -4.80773270e-01
-6.26326621e-01 -6.81964874e-01 -4.05041486e-01 3.74415010e-01
8.76123607e-02 -1.63635910e-01 -3.47709119e-01 -5.42328119e-01
4.67635781e-01 5.32541215e-01 1.23199508e-01 7.01873958e-01
4.63847548e-01 -7.06606090e-01 -4.40416813e-01 6.67307377e-01
1.61267185e+00 4.69823927e-01 5.81348360e-01 -1.85407214e-02
3.05510223e-01 5.72941117e-02 6.77187383e-01 1.07178903e+00
-1.35847136e-01 2.75082916e-01 1.51159421e-01 -7.12561607e-03
3.87705714e-01 1.86022073e-01 2.87283957e-01 1.51297307e+00
3.50800343e-03 -5.87328374e-01 -8.80038023e-01 3.55366439e-01
-1.72364342e+00 -1.03663886e+00 -9.39247429e-01 2.28040648e+00
2.23925203e-01 -3.87727648e-01 -3.20398629e-01 8.34658369e-02
5.58062851e-01 1.28397703e-01 -3.02441031e-01 -1.15825478e-02
-3.92100036e-01 -1.66257158e-01 1.07193696e+00 4.90644485e-01
-9.29792225e-01 7.42449015e-02 6.75556421e+00 4.72562432e-01
-9.23079669e-01 -1.62216052e-01 5.96299134e-02 6.14302278e-01
-3.10463458e-01 3.15345041e-02 -4.76478636e-01 1.67074651e-01
7.68323600e-01 -5.69587052e-01 2.87210107e-01 6.65938318e-01
4.05926704e-02 -2.85363525e-01 -8.86985421e-01 1.46803784e+00
3.59983921e-01 -8.23206186e-01 -4.38581884e-01 2.89825499e-01
8.73459756e-01 -3.20656747e-01 4.29979116e-01 3.44971642e-02
-1.13687413e-02 -8.29473376e-01 4.41638798e-01 7.08392501e-01
6.94035590e-01 -7.23140776e-01 9.00058985e-01 3.68971825e-01
-1.32600868e+00 -3.14917624e-01 -3.35673809e-01 -9.64636207e-02
3.81979406e-01 1.24910951e+00 -9.00220335e-01 9.32400942e-01
8.24994594e-02 3.60753119e-01 -1.01078101e-01 1.60345697e+00
-7.30969235e-02 8.38019073e-01 -3.13490152e-01 -2.42600784e-01
3.12261492e-01 -9.50686336e-01 1.03525579e+00 1.36529410e+00
1.18258464e+00 5.95596731e-01 2.36825183e-01 1.96916223e-01
3.57052892e-01 2.28827223e-01 -6.91094041e-01 1.06460251e-01
7.46141315e-01 1.23393726e+00 -5.64326763e-01 -9.75643098e-02
-5.43055117e-01 8.44934404e-01 -6.52308285e-01 1.04230928e+00
-4.51413006e-01 -5.89772761e-01 5.87088645e-01 -1.34918004e-01
2.44984135e-01 -9.18014467e-01 -3.72674584e-01 -8.67886901e-01
3.28765111e-03 -8.50989938e-01 -5.43473251e-02 -6.18122458e-01
-1.13933539e+00 6.22689307e-01 -3.38604003e-01 -1.91568363e+00
-4.35374022e-01 -5.51335990e-01 -5.37595570e-01 9.34694588e-01
-9.36862767e-01 -7.43416309e-01 -1.01131842e-01 5.14512181e-01
-1.39877677e-01 -3.87878776e-01 6.87304914e-01 6.38212919e-01
-5.02819777e-01 4.24969852e-01 7.68235922e-01 1.77823067e-01
4.66044039e-01 -8.93120825e-01 -9.10199508e-02 1.31987333e+00
-2.32916865e-02 1.08418727e+00 1.04366624e+00 -5.68110943e-01
-2.01864648e+00 -8.71239007e-01 5.90367675e-01 -1.32422030e-01
9.09556508e-01 -6.87568411e-02 -3.41549546e-01 9.24050510e-01
2.24211693e-01 -1.22192219e-01 1.11922228e+00 9.96132270e-02
-3.12046558e-02 -1.43430471e-01 -4.81048048e-01 5.89115024e-01
6.35113239e-01 -5.38212880e-02 -3.02491248e-01 3.14832121e-01
-7.76560232e-02 -5.11344194e-01 -6.82288349e-01 5.50514698e-01
8.83665323e-01 -8.21677566e-01 1.04017949e+00 9.08799097e-03
-3.30463707e-01 -5.35959244e-01 -5.32503307e-01 -1.31862497e+00
-8.67108822e-01 -1.08239794e+00 4.56340015e-02 1.01062012e+00
4.16832596e-01 -9.23828781e-01 2.66769111e-01 5.42877503e-02
-2.67925441e-01 -3.86798531e-01 -9.28336143e-01 -1.19439411e+00
-6.46935761e-01 -3.86913657e-01 2.63616979e-01 8.56410861e-01
-6.10196926e-02 6.80422544e-01 -9.75340247e-01 1.11904657e+00
8.02932560e-01 1.39699891e-01 9.34491038e-01 -1.00839496e+00
-4.23907578e-01 -1.92725226e-01 -2.58628726e-01 -1.67777789e+00
-3.50007378e-02 -5.17752349e-01 1.59768403e-01 -1.49860263e+00
-8.13661516e-02 -7.63933539e-01 -9.14024636e-02 -4.41604286e-01
-1.20566646e-02 6.77289441e-02 1.42001482e-02 2.15183705e-01
-3.61781329e-01 2.73988456e-01 9.18710113e-01 2.86484003e-01
-1.30151659e-01 7.30160713e-01 -4.73028064e-01 4.60066170e-01
5.60160875e-01 -2.17751592e-01 -3.63063008e-01 -3.57776940e-01
5.46911538e-01 7.88684666e-01 1.79208945e-02 -1.14502847e+00
3.86798799e-01 -1.98259261e-02 5.40243268e-01 -9.11954641e-01
3.09538305e-01 -1.22177804e+00 6.46043062e-01 5.92899680e-01
2.94419199e-01 6.15554452e-02 -1.22562520e-01 8.55229795e-01
-3.27666193e-01 -3.01371336e-01 3.87311637e-01 5.61532736e-01
-2.33626440e-01 1.81256384e-01 -9.33063090e-01 -2.36108854e-01
8.89734089e-01 -2.00700849e-01 -6.25534058e-02 -8.80360365e-01
-2.91546375e-01 6.40054643e-02 -5.23005910e-02 -2.66521662e-01
6.27105117e-01 -1.42577934e+00 -9.82845187e-01 2.75629401e-01
-7.78700337e-02 -6.79702520e-01 3.60414624e-01 1.26190317e+00
-3.80104899e-01 7.12607980e-01 2.49201909e-01 -6.44988775e-01
-1.27870893e+00 5.22436142e-01 3.19728367e-02 -1.04468562e-01
-5.17726839e-02 5.91876030e-01 2.93249398e-01 -9.49292183e-02
-7.18550757e-02 1.30278701e-02 -6.62992299e-02 -8.28940570e-02
7.43146300e-01 5.66666245e-01 -9.62898210e-02 -8.34664643e-01
-2.19443366e-01 6.50474846e-01 8.12806308e-01 -3.70554239e-01
9.37304378e-01 -5.15712261e-01 -5.60986102e-01 1.47459865e-01
1.31830907e+00 1.14025748e+00 -7.09438205e-01 -3.52117091e-01
-5.67426868e-02 -7.63895810e-01 -2.85349220e-01 -2.30605692e-01
-5.84093034e-01 3.25363964e-01 5.11578262e-01 3.42321038e-01
1.36196637e+00 -3.10563952e-01 2.14243829e-01 6.29534483e-01
3.86429608e-01 -4.38210458e-01 2.44651377e-01 6.16371512e-01
1.06752729e+00 -6.20023370e-01 1.56258941e-01 -8.34564865e-01
-3.17956388e-01 1.16324794e+00 8.08852017e-02 -2.87603319e-01
9.36059058e-01 4.37679768e-01 1.21009335e-01 1.29037872e-01
-3.22599292e-01 -2.13000134e-01 6.50811493e-01 5.62248588e-01
3.97304982e-01 2.68883467e-01 -4.95145172e-01 5.90194046e-01
-1.50060654e-01 -6.06145382e-01 7.51035929e-01 5.65753877e-01
-6.67605281e-01 -7.53613174e-01 -1.11473799e+00 4.86243933e-01
-2.57708102e-01 -2.65073985e-01 8.51364285e-02 5.68493485e-01
-8.06562603e-01 1.40355062e+00 -1.20254487e-01 -1.55211270e-01
5.61742723e-01 -6.42073929e-01 1.01755731e-01 -4.35730100e-01
5.21374941e-02 7.98383534e-01 3.33369881e-01 -7.05919713e-02
-1.79353148e-01 -6.16078138e-01 -7.60056973e-01 -1.86541714e-02
-5.23183286e-01 4.44715470e-01 5.14749944e-01 5.58281779e-01
5.55589311e-02 4.30253118e-01 1.16566610e+00 -5.59526265e-01
-8.20546985e-01 -8.74139369e-01 -1.03135753e+00 -3.56782496e-01
6.43878102e-01 -1.19866215e-01 -4.82127279e-01 -1.64728895e-01]
|
[6.455741882324219, 1.3487093448638916]
|
aac854c6-5b32-4bbc-b1dd-a451657f3e6b
|
disc-diff-disentangled-conditional-diffusion
|
2303.13933
| null |
https://arxiv.org/abs/2303.13933v2
|
https://arxiv.org/pdf/2303.13933v2.pdf
|
DisC-Diff: Disentangled Conditional Diffusion Model for Multi-Contrast MRI Super-Resolution
|
Multi-contrast magnetic resonance imaging (MRI) is the most common management tool used to characterize neurological disorders based on brain tissue contrasts. However, acquiring high-resolution MRI scans is time-consuming and infeasible under specific conditions. Hence, multi-contrast super-resolution methods have been developed to improve the quality of low-resolution contrasts by leveraging complementary information from multi-contrast MRI. Current deep learning-based super-resolution methods have limitations in estimating restoration uncertainty and avoiding mode collapse. Although the diffusion model has emerged as a promising approach for image enhancement, capturing complex interactions between multiple conditions introduced by multi-contrast MRI super-resolution remains a challenge for clinical applications. In this paper, we propose a disentangled conditional diffusion model, DisC-Diff, for multi-contrast brain MRI super-resolution. It utilizes the sampling-based generation and simple objective function of diffusion models to estimate uncertainty in restorations effectively and ensure a stable optimization process. Moreover, DisC-Diff leverages a disentangled multi-stream network to fully exploit complementary information from multi-contrast MRI, improving model interpretation under multiple conditions of multi-contrast inputs. We validated the effectiveness of DisC-Diff on two datasets: the IXI dataset, which contains 578 normal brains, and a clinical dataset with 316 pathological brains. Our experimental results demonstrate that DisC-Diff outperforms other state-of-the-art methods both quantitatively and visually.
|
['Chao Li', 'Xi Chen', 'Lan Jiang', 'Ye Mao']
|
2023-03-24
| null | null | null | null |
['image-enhancement']
|
['computer-vision']
|
[ 3.95725042e-01 -3.28618556e-01 9.96142104e-02 -3.69712830e-01
-1.18354309e+00 -2.10937306e-01 5.28111339e-01 -2.80610919e-01
-3.74697089e-01 7.67066002e-01 4.83063608e-01 2.41800606e-01
-6.71198368e-01 -3.61252785e-01 -1.90127864e-01 -1.01326764e+00
-4.79048282e-01 4.78393763e-01 2.29477763e-01 3.80837880e-02
4.45970744e-02 4.69362617e-01 -1.06373429e+00 3.50791782e-01
1.25332916e+00 7.55220711e-01 7.16595590e-01 3.81137878e-01
3.36873263e-01 6.91200018e-01 -1.19312853e-01 -1.45484852e-02
2.43260592e-01 -4.81089354e-01 -8.50652874e-01 -1.84761032e-01
1.62580937e-01 -6.60597742e-01 -4.93095517e-01 1.31709898e+00
8.68050396e-01 -3.71214785e-02 5.84416330e-01 -5.78696728e-01
-8.02072883e-01 8.20442617e-01 -9.12966013e-01 1.13018084e+00
-4.85388786e-02 3.44350904e-01 5.11574328e-01 -8.32345366e-01
6.53422713e-01 8.97004664e-01 3.76881987e-01 6.04390562e-01
-1.57315123e+00 -7.23570883e-01 -1.22836128e-01 5.11519849e-01
-1.06910682e+00 -4.18683350e-01 6.85466528e-01 -5.25103927e-01
6.25944078e-01 1.31351307e-01 5.65607607e-01 1.09124792e+00
4.80819911e-01 5.49504578e-01 1.90593541e+00 2.08070949e-01
-1.86835006e-02 -5.57715595e-01 -2.94570997e-02 4.40199971e-01
2.27551416e-01 4.32780713e-01 -5.02423704e-01 -2.12847859e-01
1.17022431e+00 -5.61249666e-02 -6.48319066e-01 -1.44366547e-01
-1.40531385e+00 5.69668293e-01 5.06606102e-01 4.75002974e-01
-7.86370695e-01 -2.74722010e-01 2.72508234e-01 -2.08669975e-01
7.94032574e-01 4.45400357e-01 -1.44518644e-01 8.03445186e-03
-1.24929857e+00 9.15137082e-02 1.76913978e-03 1.82013512e-01
9.52538922e-02 1.27990171e-01 -3.40663314e-01 9.11015987e-01
8.21165070e-02 4.18568611e-01 6.19135976e-01 -1.08611381e+00
1.23234510e-01 1.45273775e-01 -2.88020700e-01 -7.63607919e-01
-6.76777422e-01 -8.44490409e-01 -1.29400957e+00 3.68914485e-01
2.58231968e-01 -9.41575840e-02 -9.03654099e-01 1.68053961e+00
2.79360116e-01 3.67929310e-01 -3.21481854e-01 1.48909974e+00
6.79245114e-01 3.92324366e-02 -6.91260472e-02 -5.36708474e-01
1.36458468e+00 -6.17083132e-01 -8.77692521e-01 -1.04005300e-01
-1.17454633e-01 -2.92149603e-01 6.57251894e-01 5.65470874e-01
-1.46420550e+00 -1.38424607e-02 -1.02613270e+00 1.58906505e-01
3.49745750e-01 -3.08914572e-01 5.10397434e-01 2.26781368e-01
-9.54067707e-01 7.56391883e-01 -1.35521209e+00 5.67415714e-01
8.85475218e-01 2.76272565e-01 -4.88661110e-01 -5.41639507e-01
-1.30866051e+00 1.06976902e+00 4.17929105e-02 1.67415008e-01
-1.14507914e+00 -1.44018722e+00 -5.73640764e-01 -1.12844370e-01
2.40307227e-01 -7.19439983e-01 8.34873557e-01 -2.43783921e-01
-1.25508046e+00 6.05471671e-01 -4.80675064e-02 -3.60082120e-01
7.22175479e-01 -3.97802778e-02 -5.61431527e-01 9.57242966e-01
1.97455689e-01 3.44633877e-01 8.59622598e-01 -1.24810505e+00
-7.16912523e-02 -7.98341632e-01 -3.22707176e-01 1.48532271e-01
2.05698997e-01 3.50862056e-01 1.48489624e-01 -6.68949962e-01
4.68383998e-01 -5.36089838e-01 -4.05158013e-01 -1.06724203e-01
-3.14941376e-01 4.95725989e-01 4.39927042e-01 -1.24142718e+00
1.01822889e+00 -1.76944387e+00 4.97739702e-01 2.97018234e-02
1.09306550e+00 1.04223460e-01 6.47247434e-02 -4.43053842e-01
-3.78509998e-01 7.80325830e-02 -6.45704269e-01 -1.31936356e-01
-4.22907472e-01 -1.44919500e-01 1.61021590e-01 8.07238042e-01
2.68671274e-01 8.28741133e-01 -1.18654001e+00 -5.63556373e-01
9.85185727e-02 1.01333559e+00 -4.90588129e-01 1.23654753e-01
3.63363087e-01 1.33242512e+00 -2.82139629e-01 3.68415594e-01
9.32609439e-01 -5.83061993e-01 2.73052752e-01 -6.07075989e-01
-1.19350806e-01 -3.88133302e-02 -8.69430304e-01 1.86868882e+00
-2.51575619e-01 3.16226155e-01 2.98466623e-01 -9.16374326e-01
4.14394647e-01 4.74186808e-01 1.03455937e+00 -1.06249905e+00
-2.03037430e-02 1.92337766e-01 2.64506370e-01 -7.31852770e-01
-3.24805155e-02 -5.30969620e-01 4.87271279e-01 6.39755249e-01
6.39715046e-02 -1.00624777e-01 6.45683035e-02 3.04442823e-01
1.20384705e+00 1.01957377e-02 7.61822537e-02 -4.00344193e-01
2.23398969e-01 -5.07934809e-01 7.05511808e-01 5.63679218e-01
-4.53087926e-01 9.71018493e-01 3.39592040e-01 -1.60384048e-02
-1.06143510e+00 -1.23107505e+00 -5.85317910e-01 3.97896588e-01
-1.24943502e-01 1.35005608e-01 -7.62066364e-01 -2.71662980e-01
-3.42676908e-01 2.62381285e-01 -6.46007121e-01 -2.00423114e-02
-7.01474428e-01 -1.46011305e+00 3.79039675e-01 4.09702212e-01
6.48073733e-01 -5.76574326e-01 -6.27961934e-01 3.14394295e-01
-6.55149102e-01 -1.27621055e+00 -5.02615631e-01 -5.32083102e-02
-1.00366592e+00 -1.05144644e+00 -1.14496744e+00 -2.10711360e-01
5.15484571e-01 4.16472048e-01 9.75227118e-01 -1.17779694e-01
-7.21964002e-01 1.02233320e-01 -4.77703512e-02 2.17665657e-01
-3.19317043e-01 -2.61934727e-01 1.59215346e-01 -5.32316118e-02
-2.26878911e-01 -1.07212877e+00 -1.14334381e+00 2.38315448e-01
-1.07954466e+00 2.63743371e-01 8.20109487e-01 9.73289788e-01
8.40136826e-01 1.34941742e-01 7.29041398e-01 -5.64908862e-01
7.26757228e-01 -8.05380344e-01 -3.92266721e-01 1.77014336e-01
-7.02179134e-01 3.82379502e-01 2.67503470e-01 -5.35824239e-01
-1.44055462e+00 -3.89881819e-01 7.92448409e-03 -4.79526460e-01
-3.14556696e-02 6.60252452e-01 4.06569522e-03 -2.85728276e-01
6.57198608e-01 3.33902955e-01 3.31863880e-01 -5.03203034e-01
2.05563337e-01 2.56056279e-01 7.31366038e-01 -5.11259913e-01
3.89039725e-01 8.52009177e-01 2.67702818e-01 -5.92819870e-01
-8.59808743e-01 -2.73783982e-01 -6.47853136e-01 -4.15857464e-01
9.63761449e-01 -8.17023695e-01 -5.35340667e-01 6.65834963e-01
-7.88777411e-01 -1.03203490e-01 9.93852839e-02 7.83075035e-01
-3.71434838e-01 5.38240671e-01 -1.00416946e+00 -4.55241293e-01
-5.43333113e-01 -1.72776687e+00 6.63392603e-01 1.50913343e-01
1.82629734e-01 -9.14377034e-01 8.42548981e-02 6.65829539e-01
8.17615807e-01 5.40666103e-01 1.06949115e+00 -2.05601141e-01
-5.86604774e-01 3.59278500e-01 -5.56500018e-01 3.52022737e-01
2.27573290e-01 -5.94211996e-01 -8.35266232e-01 -2.61347592e-01
4.10994589e-01 -1.95930451e-01 9.13284302e-01 9.01692033e-01
1.23556471e+00 1.30618541e-02 6.99259937e-02 8.16372812e-01
1.36312735e+00 -2.10360065e-01 5.68786919e-01 2.10331261e-01
5.43383837e-01 5.05281389e-01 4.95743416e-02 4.44731861e-01
4.61109757e-01 4.88920182e-01 4.66333091e-01 -1.93979405e-02
-4.85580057e-01 2.87577063e-01 -2.03323755e-02 1.03195429e+00
-4.44607705e-01 5.49270093e-01 -9.92968142e-01 5.89079082e-01
-1.39061856e+00 -1.18085873e+00 -2.58761078e-01 1.97343767e+00
1.19834208e+00 -5.32232746e-02 5.98151125e-02 -9.56980363e-02
7.76371479e-01 1.94066763e-01 -1.06955576e+00 6.03590965e-01
-4.45719719e-01 2.95282125e-01 2.87732244e-01 5.09075820e-01
-8.03319573e-01 1.70035094e-01 6.39477539e+00 5.41162431e-01
-1.23124683e+00 6.75593615e-01 5.56267619e-01 -4.81946796e-01
-5.24002910e-01 -4.40507174e-01 -2.33968467e-01 4.27013069e-01
7.27238119e-01 -2.01999024e-01 8.12957287e-01 1.24343321e-01
4.87163812e-01 -1.43899411e-01 -7.15985894e-01 9.64710951e-01
2.12612860e-02 -1.30474389e+00 -1.87355712e-01 1.24470457e-01
7.73162782e-01 5.22296548e-01 5.20964146e-01 -3.41749966e-01
2.61736006e-01 -1.16549611e+00 4.20142978e-01 8.26569438e-01
1.06177914e+00 -5.86334825e-01 4.04665828e-01 1.07424304e-01
-7.47590005e-01 -2.14645360e-03 5.77327758e-02 3.70129824e-01
5.36487937e-01 1.00617886e+00 -3.28671634e-01 6.77286625e-01
7.88219094e-01 5.57770312e-01 -3.26045722e-01 1.15067840e+00
-1.42796487e-01 4.45613265e-01 6.73436299e-02 8.38405132e-01
-1.25997504e-02 -2.18104184e-01 9.16746736e-01 9.01443481e-01
1.26576900e-01 5.08968115e-01 -1.49585620e-01 1.35627413e+00
2.72841215e-01 -5.05678713e-01 3.11717279e-02 2.27909520e-01
2.16969535e-01 1.46208680e+00 -5.96418023e-01 -1.51507795e-01
-2.85413384e-01 6.86941922e-01 2.99762279e-01 4.78430092e-01
-7.12660968e-01 1.87800258e-01 6.62607908e-01 2.55148500e-01
-6.44291863e-02 -4.23759282e-01 -3.26204449e-01 -1.44127154e+00
-1.59880202e-02 -9.86781180e-01 2.84205556e-01 -8.36693227e-01
-1.53525054e+00 1.00034392e+00 3.94571275e-01 -8.08568180e-01
-1.45826042e-01 -2.07669050e-01 -1.80465296e-01 1.13229752e+00
-1.94988453e+00 -9.61524308e-01 -3.09837669e-01 6.44489646e-01
2.54239082e-01 1.74177870e-01 5.30925930e-01 4.81900930e-01
-5.51420867e-01 7.23482892e-02 1.60082251e-01 -1.13760561e-01
5.84799647e-01 -1.22206736e+00 -1.08925223e-01 1.04877043e+00
-4.76656139e-01 5.92386782e-01 7.67190099e-01 -7.88196325e-01
-1.14564598e+00 -7.17656016e-01 6.85052052e-02 -1.61207661e-01
8.22352350e-01 1.04946457e-01 -1.33093822e+00 2.25384101e-01
1.41776100e-01 5.32807112e-01 6.78079724e-01 -4.07467246e-01
-2.30627194e-01 1.26158163e-01 -1.55259407e+00 3.70738059e-01
1.07302487e+00 -6.31074369e-01 -5.79931676e-01 1.49658129e-01
6.72699928e-01 -6.37041807e-01 -1.47677696e+00 4.83935088e-01
5.12253582e-01 -9.94233012e-01 1.25763285e+00 -5.33961952e-01
7.39941955e-01 -4.84735407e-02 6.48131743e-02 -1.81299901e+00
-6.73629642e-01 -4.08262759e-01 -3.39167386e-01 6.39410853e-01
4.04179469e-02 -7.08666146e-01 2.43964270e-01 8.61443758e-01
-1.63164780e-01 -7.97361970e-01 -1.08122742e+00 -5.63602149e-01
2.96470731e-01 -2.53359616e-01 6.55749798e-01 1.05284119e+00
-1.27010375e-01 -1.71806201e-01 -2.51142085e-01 3.49787652e-01
1.48478913e+00 -1.32067531e-01 -4.44093704e-01 -1.18234468e+00
-2.84553111e-01 -5.86390078e-01 -5.39160483e-02 -4.86909032e-01
2.42935911e-01 -1.03397810e+00 -2.86192298e-01 -1.51985765e+00
7.54380047e-01 -4.06109273e-01 -5.56967616e-01 1.68825535e-03
-3.25382441e-01 2.64692813e-01 1.12548927e-02 4.12977159e-01
-3.00664395e-01 4.93145734e-01 2.00280976e+00 -1.78645134e-01
1.79412305e-01 -4.31615114e-01 -7.91066587e-01 6.18447721e-01
7.04324663e-01 -4.88623917e-01 -3.43243301e-01 -6.78880095e-01
8.64452943e-02 5.99052370e-01 6.62942767e-01 -7.96730340e-01
2.50486344e-01 -9.97684747e-02 5.01357913e-01 -2.69225955e-01
3.41901004e-01 -5.15512586e-01 1.36546060e-01 3.16067666e-01
-3.62557381e-01 1.73387770e-02 -4.33386415e-02 3.96423161e-01
-3.72109264e-02 6.30069301e-02 1.07249773e+00 -4.33223575e-01
-3.80310804e-01 7.47276306e-01 -2.36538842e-01 3.91977847e-01
5.38598657e-01 1.87627524e-01 -4.82886404e-01 -6.46698400e-02
-1.01763427e+00 7.50524476e-02 3.51329595e-02 2.24944413e-01
1.05462158e+00 -1.16041589e+00 -1.10403466e+00 1.28885075e-01
-3.35708767e-01 3.87726203e-02 1.08936143e+00 1.48381579e+00
-1.38432845e-01 2.28035569e-01 -5.90506613e-01 -7.07531095e-01
-9.59056020e-01 3.80364776e-01 6.50759459e-01 -5.25749505e-01
-1.03350496e+00 5.53065777e-01 1.58718318e-01 -1.06963642e-01
-3.86585653e-01 -2.79527843e-01 -3.72642606e-01 -2.01495782e-01
1.03604209e+00 5.86909890e-01 1.82024240e-01 -8.11038077e-01
-3.69723916e-01 3.77185374e-01 -2.26903319e-01 -4.78983700e-01
1.78334689e+00 -4.03910667e-01 -3.67176741e-01 1.14845999e-01
1.05527163e+00 -3.96762431e-01 -1.60973048e+00 -6.01614594e-01
-2.16860861e-01 -5.06545186e-01 8.78133833e-01 -1.11113250e+00
-1.56302893e+00 8.71719062e-01 7.83026814e-01 -1.72632039e-01
1.24811590e+00 -1.18305728e-01 8.27181399e-01 -4.19946432e-01
6.48710787e-01 -6.53651059e-01 1.69509009e-01 6.76033795e-02
1.13009739e+00 -1.29370821e+00 1.26385331e-01 -2.20006421e-01
-7.68471360e-01 8.84739697e-01 4.73337084e-01 9.89176780e-02
7.78528750e-01 5.75110674e-01 -1.82185888e-01 -5.13177574e-01
-6.28055453e-01 6.67034313e-02 3.95015061e-01 6.61949992e-01
2.61495769e-01 1.08340457e-01 -1.90206498e-01 8.32043111e-01
1.95654646e-01 1.50492638e-01 5.06309927e-01 6.06116831e-01
-9.32111889e-02 -6.61019921e-01 -3.66318613e-01 6.63125992e-01
-8.30634475e-01 -2.64046460e-01 3.40538830e-01 2.99367428e-01
-1.30970672e-01 8.13553691e-01 -2.86503732e-01 9.62563790e-03
-2.71430984e-02 -2.95150578e-01 9.80413854e-01 -3.08684707e-01
-2.95754313e-01 8.62505510e-02 -2.68165171e-01 -7.33776093e-01
-7.13715434e-01 -8.85094225e-01 -1.38492525e+00 -1.56067893e-01
-2.33960032e-01 -1.45507455e-01 6.98495269e-01 1.10989106e+00
4.89948958e-01 9.36803043e-01 4.89274949e-01 -9.96962726e-01
-7.05344796e-01 -9.26415265e-01 -7.39734828e-01 2.86611199e-01
6.01839721e-01 -9.84456003e-01 -3.65948379e-01 -2.28885844e-01]
|
[13.641218185424805, -2.3994524478912354]
|
f14c52bb-884f-4393-88fe-80c0cab8eda6
|
unsupervised-primitive-discovery-for-improved-1
|
1906.03650
| null |
https://arxiv.org/abs/1906.03650v1
|
https://arxiv.org/pdf/1906.03650v1.pdf
|
Unsupervised Primitive Discovery for Improved 3D Generative Modeling
|
3D shape generation is a challenging problem due to the high-dimensional output space and complex part configurations of real-world objects. As a result, existing algorithms experience difficulties in accurate generative modeling of 3D shapes. Here, we propose a novel factorized generative model for 3D shape generation that sequentially transitions from coarse to fine scale shape generation. To this end, we introduce an unsupervised primitive discovery algorithm based on a higher-order conditional random field model. Using the primitive parts for shapes as attributes, a parameterized 3D representation is modeled in the first stage. This representation is further refined in the next stage by adding fine scale details to shape. Our results demonstrate improved representation ability of the generative model and better quality samples of newly generated 3D shapes. Further, our primitive generation approach can accurately parse common objects into a simplified representation.
|
['Nick Barnes', 'Salman H. Khan', 'Munawar Hayat', 'Yulan Guo']
|
2019-06-09
|
unsupervised-primitive-discovery-for-improved
|
http://openaccess.thecvf.com/content_CVPR_2019/html/Khan_Unsupervised_Primitive_Discovery_for_Improved_3D_Generative_Modeling_CVPR_2019_paper.html
|
http://openaccess.thecvf.com/content_CVPR_2019/papers/Khan_Unsupervised_Primitive_Discovery_for_Improved_3D_Generative_Modeling_CVPR_2019_paper.pdf
|
cvpr-2019-6
|
['3d-shape-generation']
|
['computer-vision']
|
[ 2.43191496e-01 1.33053094e-01 2.50321209e-01 -1.91388085e-01
-5.71362972e-01 -7.40886033e-01 7.49775887e-01 -5.90530969e-02
4.37500596e-01 5.04838586e-01 2.22273275e-01 -2.71360646e-03
9.01535302e-02 -1.16718197e+00 -5.97891331e-01 -4.87260014e-01
1.49693623e-01 1.14521801e+00 2.68795341e-01 1.19102187e-02
1.42474487e-01 1.10118210e+00 -1.55507791e+00 1.73629016e-01
9.84233320e-01 6.74153805e-01 2.91828126e-01 7.94175982e-01
-5.93135655e-01 -1.27979875e-01 -4.72424656e-01 -3.29265594e-01
4.61922914e-01 -2.61791736e-01 -5.38304389e-01 5.99445224e-01
1.53857931e-01 -3.38996470e-01 1.92313060e-01 8.15157950e-01
3.16907287e-01 1.71758562e-01 1.18105876e+00 -1.17162681e+00
-6.18540764e-01 4.89072800e-01 -3.16862047e-01 -6.39143527e-01
3.41576219e-01 1.53651029e-01 8.26465786e-01 -1.19430816e+00
6.87553883e-01 1.70737863e+00 3.81799430e-01 5.41418433e-01
-1.66512275e+00 -4.69063222e-01 2.50421524e-01 -5.99447131e-01
-1.44131875e+00 -1.26907319e-01 1.16894019e+00 -6.34438097e-01
8.25021923e-01 1.41257405e-01 1.06474733e+00 6.90189719e-01
3.32720168e-02 9.30778682e-01 8.28206539e-01 -3.28171104e-01
3.85866702e-01 -1.12986743e-01 -2.86623150e-01 7.17061698e-01
4.74237531e-01 -7.61412978e-02 -1.66407496e-01 -6.05127215e-01
1.39757347e+00 1.59566239e-01 3.29346746e-01 -6.85001194e-01
-9.69688773e-01 7.38122821e-01 2.03798115e-01 4.08880003e-02
-5.14864802e-01 1.60949826e-01 -4.43432122e-01 -3.43873352e-01
5.19315422e-01 4.24252838e-01 -3.84061724e-01 -1.11347355e-01
-9.04418945e-01 6.98071718e-01 7.88838804e-01 1.44538426e+00
9.60434318e-01 1.95513561e-01 -2.77486235e-01 8.46367240e-01
6.79274321e-01 7.30988085e-01 -1.27316013e-01 -9.55540180e-01
2.26114355e-02 9.95160401e-01 -1.50482701e-02 -7.03034461e-01
-1.18868230e-02 -4.62962657e-01 -9.23993409e-01 4.17839855e-01
4.46959287e-02 2.25227863e-01 -1.44599354e+00 1.43615901e+00
6.81706727e-01 3.56742367e-02 -2.62492508e-01 4.39741582e-01
6.10510051e-01 6.38187885e-01 -2.25648247e-02 7.38285407e-02
1.12782347e+00 -5.24470091e-01 -2.26412609e-01 2.17335492e-01
-6.52144402e-02 -8.98271382e-01 7.41722226e-01 1.26937613e-01
-1.34329331e+00 -6.84838057e-01 -7.72351742e-01 -3.98481786e-02
-8.76784697e-02 1.13631949e-01 8.30584586e-01 7.00711191e-01
-8.30916882e-01 5.27940631e-01 -9.91737485e-01 1.70481391e-02
8.43453884e-01 2.45353058e-01 -2.45257355e-02 -9.98754725e-02
-4.96543735e-01 3.43016177e-01 2.53575385e-01 -2.44076639e-01
-1.02425098e+00 -9.14166510e-01 -8.35545838e-01 1.82035509e-02
1.23921543e-01 -1.36466384e+00 1.34548795e+00 -1.10455684e-01
-1.72982609e+00 6.82866395e-01 -4.22546744e-01 1.97467938e-01
2.82641411e-01 -5.16210981e-02 8.95157233e-02 -1.22718573e-01
-3.21389735e-02 8.21872592e-01 1.16486084e+00 -1.90810144e+00
-4.57010120e-01 -2.39617273e-01 2.79057445e-03 1.64391518e-01
3.77666473e-01 -4.29648846e-01 -4.67694551e-01 -9.47581530e-01
5.21040082e-01 -9.35841501e-01 -6.57756925e-01 8.06850493e-02
-5.99049687e-01 -3.50597531e-01 6.90689802e-01 -1.68862522e-01
9.75740254e-01 -1.88886011e+00 3.31241876e-01 5.97313046e-01
3.62880766e-01 -5.55118918e-03 -9.48389620e-02 4.27240312e-01
1.60823390e-01 5.09603143e-01 -4.50724095e-01 -6.52913928e-01
2.28314668e-01 3.95960391e-01 -4.80758578e-01 -2.94436008e-01
6.69301450e-01 9.95008767e-01 -8.59707415e-01 -4.63325590e-01
1.68902367e-01 5.18715799e-01 -8.97186935e-01 3.32444727e-01
-6.19019628e-01 5.10220051e-01 -8.82785320e-01 8.90216649e-01
9.58670676e-01 -2.17927709e-01 -2.00595930e-01 -3.56001645e-01
4.37425869e-03 1.13432921e-01 -1.34611642e+00 1.70909250e+00
-2.41587132e-01 -2.76899904e-01 -1.76266119e-01 -3.00818294e-01
1.25331473e+00 1.22440800e-01 5.91809392e-01 9.45339054e-02
2.36221645e-02 2.27191702e-01 -1.62411273e-01 1.78337201e-01
6.76685691e-01 -3.67402345e-01 -1.17252029e-01 8.20417881e-01
1.29108652e-01 -1.17280030e+00 1.36907741e-01 3.03203821e-01
6.52197897e-01 5.57022274e-01 3.15054893e-01 -2.78228521e-02
1.09691225e-01 -1.06816165e-01 5.54666638e-01 5.78799009e-01
4.77054954e-01 9.63859379e-01 1.64254576e-01 -3.42567384e-01
-1.29513419e+00 -1.53533113e+00 -8.24860707e-02 4.17321920e-01
-3.14789787e-02 -6.19233251e-01 -5.91016650e-01 -5.57329476e-01
2.82615811e-01 7.22231984e-01 -4.78133738e-01 -5.72017245e-02
-6.53468907e-01 -6.51724935e-01 1.80895969e-01 4.94156897e-01
-1.67518146e-02 -8.98871481e-01 -4.23267037e-01 3.00978124e-01
1.75529853e-01 -8.41296256e-01 -4.93505150e-01 -9.96968895e-02
-1.37190139e+00 -7.00970709e-01 -7.75052011e-01 -7.07915187e-01
1.21929741e+00 1.08755194e-01 1.26655483e+00 1.59715474e-01
-4.61722970e-01 4.04811144e-01 -3.10341775e-01 -5.33306599e-01
-7.83492744e-01 -7.51943439e-02 -6.03414476e-02 1.46074942e-03
-1.19437128e-01 -9.50046062e-01 -3.76988560e-01 3.06725111e-02
-1.00358963e+00 5.13446569e-01 8.08389843e-01 6.12389624e-01
1.20014846e+00 2.20538870e-01 1.53248623e-01 -7.17731059e-01
5.41559935e-01 -2.50933260e-01 -5.78607321e-01 1.04477249e-01
-3.02784026e-01 5.76715589e-01 4.97911781e-01 -6.55853450e-01
-1.35997486e+00 3.67266148e-01 -1.75227508e-01 -5.46885729e-01
-5.64686894e-01 1.03595540e-01 -4.93610978e-01 2.59066999e-01
3.42165440e-01 3.56522530e-01 -1.28539407e-03 -8.99046421e-01
5.85817754e-01 1.91825643e-01 2.42809132e-01 -1.20021236e+00
1.44143188e+00 4.80577320e-01 3.59494209e-01 -7.39523053e-01
-4.77966845e-01 7.89066032e-02 -1.05312097e+00 -6.55601025e-02
6.47573590e-01 -6.41797721e-01 -3.99978995e-01 5.90595901e-01
-1.26205528e+00 -2.83385634e-01 -6.80361807e-01 2.09172428e-01
-7.76269436e-01 3.54146332e-01 -3.89150620e-01 -1.07561421e+00
-2.58630604e-01 -1.06529796e+00 1.65741944e+00 2.67626435e-01
-2.38791734e-01 -6.74459755e-01 1.96099803e-01 -1.15014508e-01
1.87599406e-01 4.62284595e-01 1.42005861e+00 -1.61549687e-01
-1.15767193e+00 -1.13374814e-01 4.79838289e-02 1.09215491e-02
5.02133608e-01 3.39857042e-01 -6.35734200e-01 -3.23143564e-02
-2.28909403e-01 1.27650527e-02 4.89826113e-01 4.68842089e-01
1.04979205e+00 -3.28834355e-02 -5.36035061e-01 6.72795534e-01
1.19615555e+00 2.59868860e-01 4.91808802e-01 -4.28618670e-01
8.29297483e-01 4.22150373e-01 4.21643704e-01 7.27559745e-01
3.85487497e-01 4.00083542e-01 1.94875851e-01 1.17698237e-01
-4.03499424e-01 -8.96938503e-01 -7.54357129e-02 8.87767613e-01
-3.38605791e-01 -1.00366287e-01 -7.99725294e-01 4.12253529e-01
-1.51923645e+00 -8.43673348e-01 -1.17746770e-01 2.19863749e+00
7.96518803e-01 1.60348907e-01 1.81161180e-01 6.65847138e-02
5.68336666e-01 -2.76175469e-01 -5.20950437e-01 -2.46131569e-02
-5.54851480e-02 4.95641798e-01 -1.35189086e-01 3.87554169e-01
-7.60191858e-01 9.64875698e-01 6.93611383e+00 8.59696150e-01
-6.16431713e-01 -4.60628420e-01 5.19625425e-01 2.16544375e-01
-1.01913285e+00 2.82439172e-01 -1.10808849e+00 2.48849690e-01
9.46358144e-02 -3.26520115e-01 3.02353323e-01 7.78553367e-01
-1.58335343e-02 -2.54659876e-02 -1.03882921e+00 9.78422284e-01
-1.11132421e-01 -1.34213161e+00 7.81928182e-01 3.75970572e-01
9.57138240e-01 -6.18143916e-01 -4.83991764e-02 8.29369202e-02
7.57796407e-01 -1.01178908e+00 8.86874616e-01 8.44832718e-01
9.40165639e-01 -8.72969151e-01 -2.09896956e-02 6.02990866e-01
-1.55821502e+00 3.97326648e-01 -4.90670800e-01 1.58486798e-01
6.53620660e-01 8.92444611e-01 -1.11800408e+00 6.05692446e-01
2.80994833e-01 4.31468636e-01 -4.00633276e-01 1.15585160e+00
-2.27160811e-01 4.58010435e-01 -6.49198234e-01 -2.61823330e-02
-2.52664894e-01 -3.26295406e-01 8.21124673e-01 8.83783460e-01
7.36308038e-01 3.84907484e-01 3.82199436e-01 1.50836682e+00
7.64181763e-02 -1.76529720e-01 -6.82752311e-01 -8.96479040e-02
5.87255716e-01 1.22327316e+00 -1.04140985e+00 -3.25399876e-01
6.18470721e-02 7.82761335e-01 1.98836878e-01 3.43297541e-01
-5.32758713e-01 7.73002580e-03 6.76271558e-01 1.48145810e-01
4.41363186e-01 -7.72343695e-01 -5.15784144e-01 -1.08097768e+00
-2.62783498e-01 -5.37159801e-01 -8.61539841e-02 -7.67820835e-01
-1.45865536e+00 5.37989080e-01 2.75948375e-01 -1.32573247e+00
-4.63276505e-01 -4.24191028e-01 -5.85522354e-01 1.04813838e+00
-1.06672251e+00 -1.50446701e+00 -1.30111545e-01 4.18889284e-01
6.21126413e-01 -9.84487310e-02 1.06026042e+00 -1.74393103e-01
4.86188605e-02 4.21444654e-01 -3.08512777e-01 -8.48477483e-02
1.29631639e-01 -1.31183171e+00 9.04983878e-01 7.78806031e-01
2.30976701e-01 7.41171181e-01 3.99865508e-01 -1.29465413e+00
-1.47508132e+00 -1.07564414e+00 6.38686597e-01 -6.38017356e-01
9.08305049e-02 -6.10079050e-01 -6.81771457e-01 4.12460506e-01
-4.26777780e-01 -2.54292667e-01 7.64985681e-01 -8.55360255e-02
-2.68974125e-01 4.40286458e-01 -1.28421497e+00 7.33426154e-01
1.50804102e+00 -2.83820778e-01 -5.37463665e-01 -1.47344708e-01
7.67193019e-01 -4.11002398e-01 -1.07424188e+00 5.09843469e-01
6.33216500e-01 -5.03675342e-01 1.23460710e+00 -5.28738022e-01
2.42764696e-01 -5.20887852e-01 -2.40593880e-01 -1.28782749e+00
-7.46485472e-01 -8.56675804e-01 -5.11789739e-01 1.24854946e+00
3.43665302e-01 -1.16384760e-01 9.39299464e-01 6.52291298e-01
-2.47721672e-01 -7.83229649e-01 -5.44364512e-01 -7.38255978e-01
5.19496612e-02 -6.26157641e-01 1.12830222e+00 3.94457489e-01
-6.75801516e-01 3.83349925e-01 -7.68291354e-02 3.12751345e-02
9.79757190e-01 9.21450913e-01 1.07459736e+00 -1.71754289e+00
-4.43995029e-01 -4.87931311e-01 -2.52016604e-01 -1.67043304e+00
-1.44280910e-01 -1.09359622e+00 1.11494437e-02 -1.66174495e+00
3.22294384e-01 -9.72088218e-01 3.63907456e-01 2.28316173e-01
-3.00039738e-01 1.64790750e-01 3.12695295e-01 1.47922546e-01
2.97015021e-03 9.49852645e-01 1.93896127e+00 -4.83677387e-02
-4.13810909e-01 3.93283427e-01 -6.79700434e-01 6.43498123e-01
6.61968112e-01 -3.67094874e-01 -4.71260697e-01 -3.21104288e-01
7.05418810e-02 -5.02019823e-02 2.84761250e-01 -6.70525432e-01
-2.01877087e-01 -4.03096259e-01 6.78590178e-01 -1.25161147e+00
6.85049534e-01 -6.91491246e-01 5.93499303e-01 1.81511104e-01
1.23208977e-01 -9.52789113e-02 9.34294164e-02 5.17517567e-01
1.30092353e-01 -1.97722256e-01 6.91002786e-01 -3.42344463e-01
-1.67196006e-01 8.26074302e-01 -2.63215661e-01 -1.60002306e-01
7.44471550e-01 -6.12875640e-01 2.98175812e-01 -2.27952033e-01
-9.54645336e-01 -1.06358960e-01 7.75786340e-01 4.04797554e-01
9.28928554e-01 -1.75447905e+00 -7.28591084e-01 6.37366772e-01
-1.32366136e-01 7.34566212e-01 1.40536517e-01 1.45566240e-01
-2.10646510e-01 6.90531507e-02 -1.62837058e-01 -7.10923612e-01
-9.89263237e-01 3.31663400e-01 -2.74151359e-02 -2.31933326e-01
-6.44450068e-01 7.39074409e-01 4.62751567e-01 -5.92118561e-01
-2.15840697e-01 -6.35940373e-01 1.51245490e-01 -2.29445338e-01
1.60521820e-01 2.66459972e-01 -2.17143923e-01 -7.21387565e-01
-1.35642868e-02 9.33448255e-01 1.98929667e-01 -1.97078586e-01
1.46649206e+00 1.05169453e-01 -3.57866026e-02 3.21563035e-01
8.18948448e-01 3.45945507e-01 -1.38724458e+00 -6.50759190e-02
-3.23459208e-01 -7.06184208e-01 -4.02801782e-01 -6.05916977e-01
-7.79565692e-01 7.18328595e-01 7.20327869e-02 1.43407390e-01
9.33341444e-01 3.39126825e-01 7.98613608e-01 1.59761414e-01
6.69261813e-01 -7.59664834e-01 1.49773344e-01 7.69692898e-01
1.08719814e+00 -6.18891597e-01 -9.79081262e-03 -9.59055781e-01
-2.78703839e-01 1.08892834e+00 4.29530233e-01 -2.54881352e-01
8.69463801e-01 5.04379392e-01 -4.52422082e-01 -1.69867098e-01
-6.55518413e-01 -2.53264099e-01 6.20876312e-01 9.50986207e-01
1.51390418e-01 2.60424763e-01 1.41239613e-01 7.98976719e-01
-4.08373892e-01 -2.06913605e-01 1.78040892e-01 8.72012556e-01
-3.50533336e-01 -1.67063379e+00 -5.28156757e-01 5.36873341e-01
5.52205092e-05 4.67900261e-02 -4.61458355e-01 3.26090902e-01
1.74138442e-01 4.84753758e-01 6.85034841e-02 -3.55206013e-01
4.10874516e-01 1.87564835e-01 8.76739800e-01 -1.00836802e+00
-1.50624022e-01 5.42848885e-01 -2.52907693e-01 -3.30692470e-01
-2.01003045e-01 -8.90539289e-01 -1.33991838e+00 3.08004469e-02
-1.67767957e-01 -3.60244699e-02 5.48751175e-01 5.70400059e-01
7.92722344e-01 3.49591434e-01 6.89679027e-01 -1.48075271e+00
-3.53975564e-01 -6.48672223e-01 -5.04499018e-01 4.89232123e-01
-5.33212312e-02 -9.90795135e-01 -6.66711926e-02 3.77798617e-01]
|
[8.827556610107422, -3.629094123840332]
|
fd528a59-66a2-4cd6-ac13-f1152e1860a9
|
regularization-techniques-for-fine-tuning-in
|
1707.09920
| null |
http://arxiv.org/abs/1707.09920v1
|
http://arxiv.org/pdf/1707.09920v1.pdf
|
Regularization techniques for fine-tuning in neural machine translation
|
We investigate techniques for supervised domain adaptation for neural machine
translation where an existing model trained on a large out-of-domain dataset is
adapted to a small in-domain dataset. In this scenario, overfitting is a major
challenge. We investigate a number of techniques to reduce overfitting and
improve transfer learning, including regularization techniques such as dropout
and L2-regularization towards an out-of-domain prior. In addition, we introduce
tuneout, a novel regularization technique inspired by dropout. We apply these
techniques, alone and in combination, to neural machine translation, obtaining
improvements on IWSLT datasets for English->German and English->Russian. We
also investigate the amounts of in-domain training data needed for domain
adaptation in NMT, and find a logarithmic relationship between the amount of
training data and gain in BLEU score.
|
['Barry Haddow', 'Ulrich Germann', 'Antonio Valerio Miceli Barone', 'Rico Sennrich']
|
2017-07-31
|
regularization-techniques-for-fine-tuning-in-1
|
https://aclanthology.org/D17-1156
|
https://aclanthology.org/D17-1156.pdf
|
emnlp-2017-9
|
['l2-regularization']
|
['methodology']
|
[ 3.05796802e-01 2.77628422e-01 -5.79835892e-01 -6.56790495e-01
-1.29043078e+00 -9.10377264e-01 5.30463755e-01 -1.11962095e-01
-8.24159503e-01 1.25999570e+00 2.61314929e-01 -7.12306559e-01
2.47705534e-01 -5.21180570e-01 -1.15631735e+00 -3.46348226e-01
5.76521635e-01 7.96816051e-01 7.13238940e-02 -4.01811272e-01
-1.53172761e-01 2.47946829e-01 -5.90641677e-01 3.84940416e-01
1.16432822e+00 4.38069075e-01 8.31702948e-02 4.35970098e-01
-2.49846533e-01 4.20299202e-01 -6.42150044e-01 -6.59853220e-01
3.51882786e-01 -9.88708973e-01 -8.23953688e-01 9.21343490e-02
4.87236977e-01 -5.44698127e-02 -2.57978052e-01 8.02361488e-01
6.55698359e-01 9.74774361e-02 8.30975056e-01 -5.85793495e-01
-9.10729408e-01 7.19685555e-01 -4.11870599e-01 2.72156715e-01
-7.60846660e-02 5.78165948e-02 5.25931478e-01 -8.80222499e-01
8.10548604e-01 1.06491292e+00 5.40310204e-01 8.62504959e-01
-1.59618890e+00 -5.62304437e-01 3.37152183e-02 -1.63202047e-01
-7.75222659e-01 -6.02297127e-01 6.05572104e-01 -2.24232301e-01
1.22668099e+00 -1.50253177e-01 9.09467563e-02 1.47003651e+00
8.61993339e-03 7.54540026e-01 1.11646712e+00 -8.02840650e-01
1.47364527e-01 6.27555192e-01 -1.63067535e-01 3.14053625e-01
6.30925074e-02 2.06395656e-01 -5.25493085e-01 -9.03255790e-02
9.16468143e-01 -5.69060683e-01 -1.00418329e-01 -4.07961428e-01
-1.09793568e+00 9.80667591e-01 2.57708997e-01 2.20518932e-01
-1.59803733e-01 -1.22161537e-01 5.47728121e-01 9.31321263e-01
9.68774915e-01 6.31055474e-01 -1.08317542e+00 -2.50427902e-01
-9.80419099e-01 -9.40728188e-02 7.94713318e-01 1.12328076e+00
8.71636748e-01 8.87692869e-02 -1.58843964e-01 1.21310949e+00
-3.18022460e-01 4.60356534e-01 6.78349853e-01 -9.32022929e-01
1.04466867e+00 4.27254885e-01 3.83959450e-02 1.06870435e-01
-1.85571052e-02 -6.14289641e-01 -6.52344763e-01 1.22813851e-01
6.28483534e-01 -6.47131264e-01 -1.03237641e+00 1.99424756e+00
2.61302926e-02 -2.52197474e-01 4.19144303e-01 5.49686491e-01
3.25244099e-01 5.88757217e-01 9.37461667e-03 -4.16202039e-01
8.54153216e-01 -1.19851899e+00 -5.11821747e-01 -5.53733349e-01
1.06888747e+00 -7.44436264e-01 1.44807589e+00 1.90713763e-01
-1.33991063e+00 -4.66976911e-01 -7.06710994e-01 -2.37353191e-01
-4.04950917e-01 3.02570373e-01 1.51704818e-01 5.70100069e-01
-1.03780890e+00 8.80083621e-01 -6.33150160e-01 -6.00311935e-01
3.32818240e-01 6.80410922e-01 -4.39999521e-01 -2.06924796e-01
-1.23690176e+00 1.20826840e+00 4.92046505e-01 -3.79749149e-01
-5.85974693e-01 -6.71728075e-01 -6.62390888e-01 -8.59327689e-02
2.75164545e-02 -7.90892720e-01 1.41051185e+00 -1.48723626e+00
-1.86702394e+00 1.02937698e+00 -1.65337533e-01 -6.25560045e-01
6.57634199e-01 -2.43897200e-01 -2.20689312e-01 -2.68520236e-01
-1.73314512e-01 6.48309767e-01 7.59362578e-01 -8.43834341e-01
-3.19885969e-01 -3.26556802e-01 -1.29786670e-01 4.30114925e-01
-6.19753122e-01 2.49797046e-01 -1.48629159e-01 -6.07011080e-01
-2.79143691e-01 -9.35931683e-01 -2.09565297e-01 -3.54974777e-01
1.02835596e-01 -1.71766341e-01 5.77149272e-01 -9.45451975e-01
1.04159760e+00 -2.03392196e+00 4.31176305e-01 -4.37584408e-02
-3.29801649e-01 5.67750216e-01 -4.75351810e-01 2.32484013e-01
-4.11780328e-02 2.15537488e-01 -4.49273139e-01 -3.72142345e-01
-1.36663705e-01 4.52013016e-01 -1.22778021e-01 1.16523124e-01
6.40674114e-01 9.29630816e-01 -8.23826015e-01 -2.52018929e-01
-1.91145673e-01 3.27517539e-01 -7.21683562e-01 3.20933193e-01
-4.47402328e-01 8.76798391e-01 -1.99347734e-01 2.05295995e-01
5.69977105e-01 5.87777533e-02 1.13120742e-01 3.75599802e-01
2.08351463e-02 8.59068930e-01 -4.81022686e-01 2.00757980e+00
-7.28798926e-01 6.05274200e-01 -9.44989845e-02 -1.07716954e+00
1.14032710e+00 4.66954440e-01 -4.64284383e-02 -9.15417552e-01
9.87392291e-02 7.28052199e-01 -2.03481801e-02 -3.92027646e-01
3.46329868e-01 -3.76302361e-01 -7.20217451e-02 3.41168970e-01
5.43847978e-01 -3.65344398e-02 2.39882022e-01 -1.78470477e-01
1.04342592e+00 5.77893674e-01 1.30940929e-01 -3.01074415e-01
1.97543934e-01 3.31034005e-01 5.86277962e-01 5.02220511e-01
-8.90604034e-02 6.38298869e-01 5.22080839e-01 -2.96703666e-01
-1.47256351e+00 -8.37337792e-01 -4.70720343e-02 1.37746501e+00
-6.37642741e-01 9.65280011e-02 -1.05841208e+00 -1.01387894e+00
-1.57343253e-01 9.29459453e-01 -3.56265694e-01 -4.75696623e-01
-1.03679729e+00 -8.92028511e-01 6.83186114e-01 6.31476045e-01
3.77507359e-01 -9.76638138e-01 1.41096422e-02 3.23513240e-01
-2.94452310e-01 -1.08312654e+00 -5.64467669e-01 9.35942233e-01
-1.50740516e+00 -3.82059574e-01 -1.09444058e+00 -1.16242480e+00
7.31910825e-01 -2.90767610e-01 1.45957065e+00 -3.94070774e-01
5.61182082e-01 -1.54351845e-01 -3.30084890e-01 -3.75903130e-01
-8.48031402e-01 7.51019418e-01 1.34555459e-01 -4.31747347e-01
7.15692997e-01 -5.22401869e-01 -1.82110652e-01 2.54128754e-01
-6.44885659e-01 -2.32757613e-01 9.18655217e-01 1.13101852e+00
3.99968445e-01 -6.01256669e-01 6.37418091e-01 -1.20207965e+00
8.70115280e-01 -4.31296974e-01 -4.94028836e-01 3.86974275e-01
-5.52523434e-01 3.38259101e-01 7.70648122e-01 -9.23144281e-01
-1.16131604e+00 3.88425738e-02 -5.94611689e-02 -4.45837259e-01
-1.38094008e-01 5.51580369e-01 -1.76271930e-01 -2.90086120e-02
1.17795730e+00 1.21817775e-01 -1.97749928e-01 -7.45650649e-01
3.41089427e-01 9.09064054e-01 2.97627181e-01 -7.51838088e-01
7.00033486e-01 -2.79327899e-01 -3.98844421e-01 -6.30050182e-01
-6.83035672e-01 -2.15599597e-01 -9.97388303e-01 4.69487756e-01
5.01383841e-01 -1.09522378e+00 2.22536236e-01 1.99330091e-01
-1.20578980e+00 -8.95832300e-01 -4.97292995e-01 7.38035381e-01
-7.41561711e-01 -1.75374337e-02 -7.94470966e-01 -3.53800237e-01
-3.01013231e-01 -9.87188756e-01 7.49355495e-01 1.47560522e-01
-2.88957030e-01 -1.25349748e+00 3.18482786e-01 3.32354695e-01
5.23889840e-01 -1.64947033e-01 1.16623569e+00 -1.07485580e+00
-2.19296068e-01 -4.69414964e-02 -8.74571130e-02 7.91655242e-01
8.33999068e-02 -5.36628842e-01 -7.25433171e-01 -5.68843842e-01
8.97207018e-03 -5.95138371e-01 7.55263269e-01 3.88954848e-01
5.56381047e-01 -3.88433695e-01 -6.36316538e-02 6.78651333e-01
1.33635664e+00 8.42852369e-02 6.22044206e-01 5.40380001e-01
4.58802074e-01 4.60865915e-01 4.58188623e-01 -5.46767674e-02
-3.93560864e-02 7.37597644e-01 -2.75742918e-01 -3.29010069e-01
-2.77886122e-01 -4.21710223e-01 7.06859946e-01 1.08778608e+00
-1.26523376e-01 -2.61602044e-01 -8.44495356e-01 7.67243147e-01
-1.55427074e+00 -3.83971125e-01 -1.02545172e-02 2.36083984e+00
1.49003923e+00 2.01391429e-01 2.78763056e-01 -3.42607945e-01
6.43166184e-01 -3.73810172e-01 -5.97299814e-01 -9.58210051e-01
-3.24989200e-01 4.89426553e-01 7.36771584e-01 5.34844279e-01
-7.86664963e-01 1.41912842e+00 6.78048086e+00 7.45211601e-01
-1.15414369e+00 4.14138794e-01 6.40930891e-01 -1.02516718e-01
-3.02012116e-01 1.61750421e-01 -8.25378895e-01 3.81076157e-01
1.49204695e+00 5.12447283e-02 6.15129113e-01 6.72197938e-01
1.60294563e-01 2.23280281e-01 -1.29503036e+00 4.43429172e-01
-5.35150208e-02 -9.13718164e-01 1.86105501e-02 1.53581455e-01
1.12132537e+00 2.94665128e-01 2.41873991e-02 6.51061296e-01
4.80319768e-01 -7.23951757e-01 2.33575001e-01 6.49566054e-02
1.12453592e+00 -7.99771190e-01 6.80423975e-01 5.50883353e-01
-3.78099531e-01 1.78773150e-01 -6.85065925e-01 -1.16674444e-02
-3.42580639e-02 5.29669344e-01 -1.03881133e+00 3.29260230e-01
4.48523253e-01 5.14148891e-01 -4.28622544e-01 7.39736855e-01
-4.98469472e-01 9.38959360e-01 -2.75065571e-01 2.66686380e-01
1.68219745e-01 -4.09993023e-01 3.60089511e-01 1.34340727e+00
4.31408316e-01 -2.55112559e-01 -2.29535609e-01 7.73665547e-01
-5.10880709e-01 1.98774874e-01 -7.31405973e-01 -3.85340117e-02
2.23519549e-01 6.37612760e-01 -1.24985985e-01 -3.72752488e-01
-4.94034171e-01 1.31437266e+00 7.31926143e-01 6.81645751e-01
-6.71854496e-01 -3.51124108e-01 5.30853391e-01 9.90829244e-02
3.49033564e-01 -3.57348055e-01 -4.63824391e-01 -1.31822979e+00
2.38831624e-01 -1.16006851e+00 1.36869103e-01 -5.35641849e-01
-1.15441835e+00 7.31898725e-01 -1.08310290e-01 -1.19238806e+00
-4.98564363e-01 -5.67689776e-01 -4.62445140e-01 1.19947124e+00
-1.74653614e+00 -9.73800957e-01 3.17704678e-01 5.48901916e-01
7.03945100e-01 -4.33155477e-01 1.02034080e+00 4.48009342e-01
-4.96302813e-01 9.88952279e-01 7.35757530e-01 2.21875846e-01
1.30731690e+00 -1.13300025e+00 8.32982957e-01 7.22539842e-01
5.84790260e-02 5.60512424e-01 4.79502648e-01 -6.33141398e-01
-8.81322026e-01 -1.14822316e+00 1.44299746e+00 -7.07367182e-01
4.92359757e-01 -5.28818965e-01 -1.09893894e+00 1.02598345e+00
4.00853544e-01 -2.71471590e-01 6.71780586e-01 3.68682176e-01
-3.87279987e-01 7.59776533e-02 -1.13840353e+00 5.56915402e-01
1.09978282e+00 -5.77859581e-01 -7.68757284e-01 3.11543316e-01
9.10117507e-01 -4.84244972e-01 -9.49467242e-01 3.41182649e-01
1.47956505e-01 -4.94461954e-01 8.25831413e-01 -1.07684708e+00
5.29409349e-01 2.51909524e-01 -9.57254786e-03 -1.87665915e+00
-2.39308968e-01 -8.75130713e-01 4.52988818e-02 1.31002629e+00
1.00307488e+00 -5.81292033e-01 9.16687310e-01 4.31859404e-01
-1.50152341e-01 -4.62482214e-01 -9.45402443e-01 -1.10249639e+00
1.05545819e+00 1.94314420e-01 2.24408641e-01 1.06729293e+00
7.19209164e-02 6.86009705e-01 -4.32298571e-01 -3.39536369e-01
3.82537395e-01 -3.51102740e-01 6.87942326e-01 -1.07641757e+00
-3.69752198e-01 -1.35533020e-01 1.42955527e-01 -1.17838788e+00
2.35309377e-01 -1.05884147e+00 -6.07850440e-02 -1.35375166e+00
9.64978039e-02 -2.69718915e-01 -3.39946300e-01 4.89561260e-01
-1.42876118e-01 1.64029285e-01 3.23852003e-02 2.86909163e-01
-1.78458005e-01 3.99855226e-01 1.20367563e+00 -8.90092999e-02
-5.39001763e-01 1.00709297e-01 -5.97686648e-01 4.20514494e-01
1.09889245e+00 -9.61310923e-01 -2.32554063e-01 -1.00654614e+00
3.63648986e-03 1.02490827e-01 -7.71440715e-02 -8.71487975e-01
-1.54640842e-02 -1.50395572e-01 4.39800233e-01 -1.32872865e-01
1.85313195e-01 -6.45367205e-01 -3.67596507e-01 1.69835091e-01
-6.53436422e-01 1.18194424e-01 4.43241447e-01 2.80432492e-01
-2.98116207e-01 -3.76004666e-01 9.43510532e-01 -2.53340155e-01
-3.02064002e-01 -9.80852023e-02 -2.83343881e-01 5.02664566e-01
4.80799645e-01 -1.45417601e-01 -2.25584432e-01 -2.34200984e-01
-8.09673071e-01 2.88417023e-02 5.87026834e-01 4.01710838e-01
1.33788154e-01 -1.33048928e+00 -1.14363420e+00 2.92086720e-01
2.09221523e-02 -1.47539645e-01 -2.42865250e-01 8.26579094e-01
-2.48253450e-01 4.59785283e-01 -3.67719352e-01 -3.53974104e-01
-1.04238999e+00 4.76882041e-01 3.38382393e-01 -6.08437061e-01
-1.76159218e-01 8.88148665e-01 -3.14237662e-02 -1.00393546e+00
2.28774503e-01 -2.29994059e-01 2.26499528e-01 -1.43116921e-01
4.31682952e-02 9.34934467e-02 3.43568504e-01 -2.20233083e-01
-4.25751582e-02 6.02769911e-01 -3.24509650e-01 -2.90806234e-01
1.29590786e+00 -1.12282299e-01 1.66783839e-01 4.16127443e-01
1.38376677e+00 -5.16844988e-02 -1.33756661e+00 -6.82929277e-01
2.81997353e-01 -1.21256135e-01 -2.16458410e-01 -1.14581954e+00
-6.24068379e-01 1.16010153e+00 4.99710858e-01 -4.29145217e-01
1.20453346e+00 -6.80591390e-02 8.49521279e-01 6.11100137e-01
3.22916955e-01 -1.43430924e+00 -3.59140933e-02 9.47497666e-01
6.36431992e-01 -1.38566637e+00 -2.28578717e-01 -1.63108334e-02
-8.01819205e-01 9.53977525e-01 5.45726597e-01 -2.09699050e-01
1.43734053e-01 2.29215249e-01 3.27366471e-01 4.73043084e-01
-7.32587337e-01 -1.47004187e-01 1.64393187e-01 6.95771158e-01
8.33226740e-01 -1.12631105e-01 -4.47476178e-01 4.61650610e-01
-1.43695205e-01 3.39744002e-01 3.63294721e-01 7.94010699e-01
-2.95109063e-01 -1.82986212e+00 -1.98414147e-01 1.69214711e-01
-5.51563084e-01 -3.14815819e-01 -7.82786071e-01 7.07986891e-01
-6.65876940e-02 7.81007648e-01 -1.58525795e-01 -2.61961401e-01
6.25638306e-01 6.35843396e-01 5.52021444e-01 -7.90480733e-01
-9.38941181e-01 2.35204712e-01 3.49339813e-01 -2.40468774e-02
-2.40580067e-01 -5.38122237e-01 -8.57946932e-01 -1.04482688e-01
-3.41757089e-01 2.54148066e-01 7.47365177e-01 9.05881703e-01
4.28371042e-01 4.00650471e-01 4.32395041e-01 -5.22987783e-01
-9.56163228e-01 -1.30701554e+00 -2.09756434e-01 4.46160764e-01
3.62629354e-01 -1.09484546e-01 -4.01919305e-01 1.72297850e-01]
|
[11.654426574707031, 10.284634590148926]
|
7d7ccc9f-df1d-47ac-b22c-3996a04876b9
|
it-s-ai-match-a-two-step-approach-for-schema
|
2203.04366
| null |
https://arxiv.org/abs/2203.04366v1
|
https://arxiv.org/pdf/2203.04366v1.pdf
|
It's AI Match: A Two-Step Approach for Schema Matching Using Embeddings
|
Since data is often stored in different sources, it needs to be integrated to gather a global view that is required in order to create value and derive knowledge from it. A critical step in data integration is schema matching which aims to find semantic correspondences between elements of two schemata. In order to reduce the manual effort involved in schema matching, many solutions for the automatic determination of schema correspondences have already been developed. In this paper, we propose a novel end-to-end approach for schema matching based on neural embeddings. The main idea is to use a two-step approach consisting of a table matching step followed by an attribute matching step. In both steps we use embeddings on different levels either representing the whole table or single attributes. Our results show that our approach is able to determine correspondences in a robust and reliable way and compared to traditional schema matching approaches can find non-trivial correspondences.
|
['Carsten Binnig', 'Andreas Schmidt', 'Michael Truong-Ngoc', 'Benjamin Hättasch']
|
2022-03-08
| null | null | null | null |
['data-integration']
|
['knowledge-base']
|
[ 2.18788788e-01 -3.84599455e-02 1.24399938e-01 -5.28739214e-01
-8.57090533e-01 -7.11568773e-01 6.36340797e-01 1.11962390e+00
-4.93222237e-01 3.49743545e-01 2.09071904e-01 -1.12216529e-02
-3.74700308e-01 -1.09554827e+00 -4.71497744e-01 -2.35556737e-01
4.06547129e-01 9.66199696e-01 3.55551779e-01 -4.25175399e-01
2.09088847e-01 5.48360348e-01 -1.85527074e+00 4.13283885e-01
6.81944668e-01 9.22859013e-01 2.86708057e-01 2.99846023e-01
-8.31764460e-01 3.90258670e-01 -2.46690735e-01 -7.37594783e-01
4.46773052e-01 -5.78889728e-01 -1.00849807e+00 -1.53388306e-01
4.66384083e-01 1.94211811e-01 2.47850358e-01 1.29528463e+00
4.68474865e-01 8.07100683e-02 5.30171096e-01 -1.08827186e+00
-1.12419203e-01 7.28084803e-01 -9.77867618e-02 -6.78522587e-02
6.29384518e-01 -4.48171198e-01 1.21654773e+00 -8.67520392e-01
8.97734046e-01 1.00858724e+00 6.63968980e-01 3.61924320e-01
-1.50111163e+00 -2.38526270e-01 -3.04595679e-01 3.42029691e-01
-1.31081772e+00 -3.80875230e-01 8.20794284e-01 -4.56245810e-01
7.80833066e-01 2.16101959e-01 6.98319018e-01 7.07930744e-01
-2.94127136e-01 3.87470633e-01 1.03350568e+00 -7.54314661e-01
3.04249406e-01 6.60773993e-01 2.39998564e-01 3.90543729e-01
5.72715938e-01 -2.43780345e-01 -2.93961227e-01 -2.29175612e-01
4.05154884e-01 -1.57910526e-01 1.01486750e-01 -8.65803480e-01
-1.13620925e+00 6.84889138e-01 5.02108872e-01 9.08667386e-01
-5.54862201e-01 -1.06171258e-01 5.83872318e-01 4.47998136e-01
1.32069230e-01 7.15509832e-01 -1.48696274e-01 -1.36657700e-01
-8.33950579e-01 4.30539131e-01 1.05140674e+00 9.21258450e-01
1.02915573e+00 -6.21545911e-01 2.74655282e-01 7.26985455e-01
1.80422470e-01 3.63644250e-02 4.24902678e-01 -5.73085964e-01
5.65648556e-01 1.22341621e+00 1.68819100e-01 -1.19305503e+00
-3.96265000e-01 -8.09316989e-03 -5.10989487e-01 2.77618408e-01
4.90958124e-01 3.05609286e-01 -4.53373164e-01 1.59278595e+00
6.82026029e-01 -2.31096134e-01 2.72465408e-01 6.63588166e-01
6.46105826e-01 3.00540358e-01 7.56950080e-02 1.78291336e-01
1.55200636e+00 -4.10608113e-01 -6.88276887e-01 6.87293112e-02
8.55943680e-01 -9.67847407e-01 8.75277042e-01 5.12603335e-02
-1.13535404e+00 -6.78034306e-01 -1.32862222e+00 -3.88968080e-01
-1.12689483e+00 -1.71795130e-01 2.04924524e-01 6.48743689e-01
-7.91283965e-01 8.33186090e-01 -5.18622398e-01 -7.44571924e-01
2.04261482e-01 3.78974408e-01 -8.10070395e-01 -1.19163074e-01
-1.10213089e+00 1.18217540e+00 1.03885579e+00 -5.53064458e-02
-3.57536338e-02 -5.56307614e-01 -9.14832354e-01 1.45470008e-01
6.45774782e-01 -9.17066574e-01 9.15286481e-01 -9.54087436e-01
-8.37923884e-01 1.00477970e+00 -2.06812307e-01 -4.94010448e-01
5.63905239e-01 3.49962004e-02 -4.03807640e-01 -1.50267124e-01
1.64037775e-02 4.40368295e-01 2.59322643e-01 -1.30539954e+00
-8.26495051e-01 -6.81400657e-01 4.40425538e-02 1.83235914e-01
-3.24881256e-01 1.77645713e-01 -4.92390245e-01 -3.46946388e-01
2.60326475e-01 -6.62862897e-01 5.13833500e-02 -9.58587378e-02
-7.46509284e-02 -2.21988931e-01 5.17179549e-01 -7.06346989e-01
1.19825280e+00 -1.82629359e+00 4.14028138e-01 4.04455513e-01
2.05120906e-01 4.22378719e-01 4.78994772e-02 9.41169500e-01
-2.33874977e-01 2.00715084e-02 -5.06375253e-01 -4.69118357e-01
2.66779989e-01 1.83468416e-01 -2.02481672e-02 4.45182025e-02
2.86985248e-01 7.33810425e-01 -8.04466307e-01 -5.60557127e-01
2.80843109e-01 4.85674828e-01 -2.87002444e-01 3.59228760e-01
-3.88087153e-01 -1.33786812e-01 -2.37081006e-01 4.08917636e-01
5.72740257e-01 1.19975187e-01 4.47812110e-01 -4.30685371e-01
-1.16365425e-01 3.97282988e-01 -1.74683058e+00 1.99783909e+00
-5.63185930e-01 4.33577925e-01 -1.59455195e-01 -1.04366982e+00
1.22746408e+00 3.36250961e-01 5.55294394e-01 -7.31439829e-01
3.06245565e-01 5.86434126e-01 -2.26530716e-01 -4.05375868e-01
7.35303283e-01 -2.15636998e-01 -1.78430319e-01 6.26054525e-01
7.93867335e-02 1.16857484e-01 3.49325210e-01 4.56458470e-03
8.90616059e-01 1.50896072e-01 6.43387735e-01 -1.76679641e-01
1.02951097e+00 3.28845948e-01 3.68875921e-01 2.31341094e-01
2.27268919e-01 4.17361706e-01 5.97957075e-01 -7.10576117e-01
-1.31791484e+00 -7.57308424e-01 9.83676165e-02 6.32661879e-01
1.56955048e-01 -5.96724331e-01 -9.39963698e-01 -8.22201490e-01
1.38971642e-01 6.25031471e-01 -7.62213886e-01 -3.80288474e-02
-6.83617175e-01 -1.85453743e-01 2.94482946e-01 5.81061840e-01
3.00642043e-01 -9.31132138e-01 -7.94922411e-01 5.40511668e-01
-9.09603238e-02 -8.77414227e-01 -1.38232514e-01 2.00001881e-01
-8.43926609e-01 -1.16623580e+00 -2.50690341e-01 -5.89245737e-01
5.73067605e-01 1.30064711e-01 1.23298502e+00 -9.81747732e-03
-3.39277476e-01 1.11626312e-01 -3.16892833e-01 -3.23262721e-01
-8.23972225e-01 3.01735222e-01 -2.58849591e-01 1.34318754e-01
7.59421945e-01 -4.76430625e-01 -1.99024543e-01 2.43232042e-01
-1.38561416e+00 -1.49096519e-01 6.86057627e-01 5.66497564e-01
6.93636954e-01 -9.49868634e-02 3.61098439e-01 -1.18007576e+00
7.37997234e-01 -5.30488074e-01 -8.74953866e-01 6.91637039e-01
-8.56550753e-01 5.93881071e-01 6.27065241e-01 -6.63604438e-02
-8.39128375e-01 3.33790243e-01 -3.92819405e-01 -2.29960769e-01
-4.50846225e-01 6.53196633e-01 -5.39638460e-01 1.29033372e-01
5.65726936e-01 5.19376062e-02 2.46913224e-01 -8.27776432e-01
5.51620841e-01 7.40447402e-01 4.27577317e-01 -4.97104824e-01
7.95784950e-01 1.90387577e-01 -2.98698572e-03 -3.92245263e-01
-3.80530953e-01 -6.72495544e-01 -1.27785635e+00 2.61574350e-02
7.64208496e-01 -5.74991524e-01 -3.63965273e-01 -1.13955711e-03
-1.20258558e+00 2.07214996e-01 -4.49518174e-01 8.15523043e-02
-5.77304244e-01 2.42837846e-01 6.78170845e-02 -3.89142543e-01
-3.34378928e-01 -9.90402043e-01 8.16031635e-01 1.67998746e-01
-4.38855559e-01 -1.12618697e+00 4.83437955e-01 2.18784034e-01
4.54052448e-01 2.22553462e-01 1.07889080e+00 -1.21433640e+00
-6.27936780e-01 -4.17449683e-01 -2.84597248e-01 1.07741110e-01
3.72350633e-01 -1.55071989e-01 -7.41336286e-01 -7.25504532e-02
-1.19588017e-01 1.50187360e-02 4.58931148e-01 -3.58811796e-01
4.31361079e-01 -5.45627903e-03 -2.96379417e-01 3.76784593e-01
1.91167116e+00 2.32333958e-01 6.77299559e-01 6.23499334e-01
5.61327517e-01 1.09231913e+00 7.39764273e-01 1.15286283e-01
4.62058902e-01 1.04887283e+00 1.41823336e-01 7.41950870e-02
-1.67565286e-01 -2.99702346e-01 -3.00854333e-02 6.09037578e-01
4.44394648e-01 -1.21344328e-01 -1.00287724e+00 7.56076276e-01
-2.00556803e+00 -9.81981933e-01 -1.79817945e-01 2.55171895e+00
5.91730416e-01 3.29225026e-02 2.91526139e-01 4.35143888e-01
5.57067037e-01 -1.71255052e-01 -1.04442224e-01 -6.68832600e-01
8.75837505e-02 1.38994351e-01 2.65048236e-01 3.96442533e-01
-9.27371442e-01 6.97999656e-01 5.86917973e+00 3.96060377e-01
-8.85754168e-01 2.39926688e-02 -5.68921231e-02 2.72057325e-01
-6.30391657e-01 1.92951113e-01 -7.38123119e-01 3.81151378e-01
9.33504343e-01 -4.04347122e-01 1.83248475e-01 6.48907244e-01
-3.66799772e-01 -1.89586282e-01 -1.28019953e+00 9.47540641e-01
1.07537881e-01 -1.40019011e+00 5.88153414e-02 3.42986733e-02
2.68364042e-01 -3.56255084e-01 -5.25620461e-01 -1.27399579e-01
9.85095128e-02 -7.07817495e-01 5.27958453e-01 7.13333488e-01
4.19495136e-01 -9.37906325e-01 9.79193211e-01 1.76814497e-01
-1.42502475e+00 1.73805207e-01 -3.05285633e-01 3.61813694e-01
1.15157545e-01 4.41655606e-01 -9.63035762e-01 1.10100377e+00
3.54651958e-01 4.37068045e-01 -6.53212130e-01 1.14488459e+00
3.19473222e-02 -2.55108148e-01 -2.79373676e-01 -1.45058572e-01
1.24754980e-01 -4.06593442e-01 4.76413906e-01 1.19879985e+00
3.90780717e-01 -3.42782259e-01 -1.02084249e-01 1.16354179e+00
-2.46480256e-02 4.51380819e-01 -9.30088520e-01 -1.12486064e-01
5.77397883e-01 1.24707937e+00 -7.85972059e-01 -4.63382244e-01
-4.24130291e-01 9.92788494e-01 4.56622869e-01 -2.99932450e-01
-5.36498010e-01 -6.61686778e-01 7.75620103e-01 8.83582756e-02
2.85017669e-01 -2.59796262e-01 -2.42262974e-01 -8.96079421e-01
3.02801013e-01 -8.26572180e-01 6.79229856e-01 -6.22051954e-01
-9.98656511e-01 9.01528835e-01 1.50172010e-01 -1.12576306e+00
-6.12353384e-01 -3.64789695e-01 -3.14527482e-01 1.17920935e+00
-1.25062919e+00 -1.12335610e+00 -4.88787025e-01 4.76045787e-01
2.74521798e-01 5.78788016e-03 1.01842523e+00 5.54726541e-01
-3.65955859e-01 4.60177124e-01 -1.28842937e-03 2.17517793e-01
5.57911158e-01 -1.45400763e+00 5.61114907e-01 9.94778037e-01
6.07245803e-01 7.93262005e-01 8.55530977e-01 -5.21100998e-01
-1.40831566e+00 -6.48142159e-01 1.53600800e+00 -4.78900254e-01
6.06103003e-01 -4.19609576e-01 -1.31990469e+00 4.93320078e-01
3.39886516e-01 -3.10551196e-01 8.26471031e-01 9.35245231e-02
-6.56881809e-01 -4.71091181e-01 -1.21954966e+00 2.97157168e-01
8.36936891e-01 -5.43917656e-01 -9.50537920e-01 -3.32412988e-01
3.66954327e-01 -1.94161013e-01 -1.16593730e+00 1.55460820e-01
5.45690238e-01 -1.08107972e+00 8.62450242e-01 -6.54487073e-01
1.65166497e-01 -4.49300349e-01 -2.31255695e-01 -1.21385229e+00
-2.56089326e-02 -2.98760861e-01 2.45179921e-01 1.59814036e+00
4.65022743e-01 -7.03611970e-01 6.88868344e-01 7.94673085e-01
1.92780092e-01 -3.98468435e-01 -8.67023230e-01 -7.26024270e-01
-1.76089644e-01 -2.44593218e-01 1.06111932e+00 1.07046270e+00
-1.07842796e-01 1.96067214e-01 6.12971969e-02 3.18052620e-02
5.25297344e-01 5.03000438e-01 9.08027470e-01 -1.72532225e+00
5.77393025e-02 -6.70897961e-01 -6.84485614e-01 -3.05444598e-01
-7.81853870e-02 -9.63606060e-01 -2.31769204e-01 -1.78504932e+00
-5.57619035e-02 -7.64331579e-01 -4.45426583e-01 3.00383717e-01
-2.33211517e-02 -1.31391048e-01 2.12937102e-01 1.05425119e-01
-1.06848784e-01 1.45142570e-01 3.66391778e-01 -5.12889922e-02
-1.52588144e-01 -2.66913980e-01 -7.45868564e-01 3.40882212e-01
8.85203838e-01 -8.41263235e-01 -3.89465421e-01 -4.19213921e-01
5.71012437e-01 -9.73198414e-02 1.63415104e-01 -1.26137447e+00
4.99553472e-01 1.10023066e-01 3.69988605e-02 -6.68124914e-01
2.93632895e-01 -1.34907627e+00 6.45172834e-01 2.25839555e-01
-3.69285554e-01 5.41655242e-01 1.83758423e-01 4.08534825e-01
-5.61469018e-01 -7.22159207e-01 6.28263474e-01 -1.72360972e-01
-8.37039888e-01 -5.44233844e-02 5.61904684e-02 -1.88564092e-01
1.13892281e+00 -2.85758406e-01 4.16008979e-02 1.30943105e-01
-7.20558763e-01 -2.59247664e-02 8.55486989e-01 6.22458398e-01
4.40326035e-01 -1.52263296e+00 -4.78289753e-01 1.51853099e-01
5.67941844e-01 -2.54659832e-01 -2.31708348e-01 6.19419038e-01
-6.16214871e-01 3.79310787e-01 -5.11631787e-01 -2.15591133e-01
-1.59273231e+00 8.73449922e-01 2.76748151e-01 -2.86439270e-01
-4.57341194e-01 4.55335557e-01 -6.76949084e-01 -5.07437706e-01
1.68684155e-01 -2.58377403e-01 -3.62351060e-01 6.70298696e-01
5.67624450e-01 3.01054925e-01 5.47253609e-01 -8.27991962e-01
-3.61594111e-01 7.08865702e-01 1.62375197e-02 -2.74709105e-01
1.27545321e+00 -7.03712851e-02 -3.85379106e-01 5.54670811e-01
1.46457326e+00 7.42354840e-02 -4.35114056e-01 -3.96614581e-01
7.11256146e-01 -7.14986563e-01 -1.71368957e-01 -7.23093033e-01
-7.97162592e-01 8.96446824e-01 6.63417876e-01 5.44734895e-01
1.10344529e+00 2.65498236e-02 7.94568419e-01 4.73176986e-01
4.18102235e-01 -1.11812520e+00 -5.10455787e-01 9.88010094e-02
7.66431391e-01 -1.29604602e+00 -9.95418429e-02 -3.41496795e-01
-4.41921443e-01 1.44566417e+00 3.17345411e-01 2.72430211e-01
2.98869431e-01 9.12189931e-02 1.51325613e-01 -3.98337901e-01
-6.57608747e-01 -7.31758058e-01 5.73535144e-01 4.54306006e-01
4.52693731e-01 -2.59596288e-01 -5.74060142e-01 2.14774504e-01
6.30228147e-02 -3.81094627e-02 3.04284930e-01 9.99793589e-01
-4.24183398e-01 -1.98917282e+00 -2.06807002e-01 1.48607463e-01
-2.14789808e-01 2.28206292e-01 -8.74727845e-01 9.49596107e-01
2.79644102e-01 6.60464942e-01 1.51085973e-01 -4.97336537e-01
7.85842717e-01 4.78307903e-01 4.18531090e-01 -4.42214072e-01
-8.86753142e-01 -2.65144736e-01 2.76245326e-01 -7.18222439e-01
-5.58423400e-01 -7.49755919e-01 -1.06848693e+00 -2.63767332e-01
-1.63542181e-01 1.91574425e-01 1.08756030e+00 6.94243968e-01
6.46958947e-01 2.70073444e-01 3.91024262e-01 -2.75878251e-01
-2.82513857e-01 -6.04878068e-01 -1.93400115e-01 9.39068556e-01
1.94631331e-03 -6.12000108e-01 1.18285090e-01 -1.16102248e-01]
|
[9.264275550842285, 8.148992538452148]
|
64e8ced4-f7e2-4fc9-b216-dc8bf6b5cc4a
|
probabilistic-prompt-learning-for-dense
|
2304.00779
| null |
https://arxiv.org/abs/2304.00779v1
|
https://arxiv.org/pdf/2304.00779v1.pdf
|
Probabilistic Prompt Learning for Dense Prediction
|
Recent progress in deterministic prompt learning has become a promising alternative to various downstream vision tasks, enabling models to learn powerful visual representations with the help of pre-trained vision-language models. However, this approach results in limited performance for dense prediction tasks that require handling more complex and diverse objects, since a single and deterministic description cannot sufficiently represent the entire image. In this paper, we present a novel probabilistic prompt learning to fully exploit the vision-language knowledge in dense prediction tasks. First, we introduce learnable class-agnostic attribute prompts to describe universal attributes across the object class. The attributes are combined with class information and visual-context knowledge to define the class-specific textual distribution. Text representations are sampled and used to guide the dense prediction task using the probabilistic pixel-text matching loss, enhancing the stability and generalization capability of the proposed method. Extensive experiments on different dense prediction tasks and ablation studies demonstrate the effectiveness of our proposed method.
|
['Kwanghoon Sohn', 'Jinhyun Jang', 'Jin Kim', 'Somi Jeong', 'Taeyong Song', 'Hyeongjun Kwon']
|
2023-04-03
| null |
http://openaccess.thecvf.com//content/CVPR2023/html/Kwon_Probabilistic_Prompt_Learning_for_Dense_Prediction_CVPR_2023_paper.html
|
http://openaccess.thecvf.com//content/CVPR2023/papers/Kwon_Probabilistic_Prompt_Learning_for_Dense_Prediction_CVPR_2023_paper.pdf
|
cvpr-2023-1
|
['text-matching']
|
['natural-language-processing']
|
[ 4.86294538e-01 -5.16880676e-02 -4.27055359e-01 -7.45533764e-01
-7.82958984e-01 -3.78942817e-01 9.54151452e-01 1.83029220e-01
-3.88496071e-01 5.99388778e-01 4.00650024e-01 1.35418698e-01
2.81945411e-02 -5.76157331e-01 -7.33914018e-01 -9.14812028e-01
4.79147255e-01 6.01113915e-01 2.34120205e-01 3.22993785e-01
7.69674480e-02 3.42702001e-01 -1.66138434e+00 6.28755450e-01
8.44835222e-01 1.39496243e+00 8.82301271e-01 3.76961857e-01
-3.79523396e-01 1.05799353e+00 -1.12881199e-01 -3.21387470e-01
6.63112998e-02 -8.61951709e-02 -4.69968617e-01 2.12253273e-01
7.85780907e-01 -4.46696609e-01 -4.40884292e-01 8.81743848e-01
2.83415139e-01 7.61274844e-02 1.00738811e+00 -1.16358566e+00
-9.13901269e-01 5.51114023e-01 -6.98325753e-01 2.55389940e-02
3.34275104e-02 3.38775635e-01 1.29179275e+00 -1.18116677e+00
4.39988524e-01 1.26366985e+00 5.32364249e-01 6.78539455e-01
-1.32526588e+00 -6.40105784e-01 7.01664865e-01 4.56130415e-01
-1.31129873e+00 -3.45937133e-01 8.62575769e-01 -5.50079107e-01
8.29120159e-01 -1.80327743e-01 7.24305749e-01 1.38069725e+00
4.68718410e-02 1.25454021e+00 1.09927690e+00 -2.87434906e-01
4.52894270e-02 4.07157689e-01 2.04206020e-01 8.47678065e-01
2.13266566e-01 2.91883320e-01 -7.30579853e-01 -7.68117979e-02
4.60584879e-01 4.09303546e-01 -1.33430094e-01 -7.49017775e-01
-1.11461318e+00 9.07684863e-01 7.76200473e-01 -1.87966853e-01
-6.00102007e-01 2.76350111e-01 1.52329892e-01 -3.40907544e-01
4.55084413e-01 -2.55403947e-02 -2.65338153e-01 2.48347670e-01
-1.00876749e+00 1.28469303e-01 4.18529928e-01 1.17821574e+00
9.99012947e-01 2.09482580e-01 -8.54031920e-01 9.89221811e-01
5.55457473e-01 8.45861018e-01 2.18979612e-01 -7.94182181e-01
2.51085728e-01 5.06904244e-01 -7.95912594e-02 -6.66418910e-01
-1.77571937e-01 -6.27658427e-01 -9.44080651e-01 1.44506559e-01
2.56087095e-01 1.49412230e-01 -1.31070054e+00 1.67493856e+00
6.99759871e-02 2.27021113e-01 2.54273042e-02 7.96808958e-01
8.07109535e-01 7.41582632e-01 7.35333860e-01 8.71268287e-03
1.08552945e+00 -1.05621135e+00 -4.38731045e-01 -5.70903540e-01
1.49794504e-01 -6.36339843e-01 1.30883813e+00 7.79756382e-02
-6.96651399e-01 -6.00764096e-01 -7.67280459e-01 -1.86200485e-01
-1.29371539e-01 3.66651267e-01 8.94361377e-01 3.17307115e-01
-9.47681427e-01 5.66035919e-02 -7.01904356e-01 -2.89665192e-01
1.19621694e+00 2.10290700e-01 -6.16398454e-02 -5.36127388e-01
-7.87163138e-01 6.29101694e-01 4.73073840e-01 -3.94453332e-02
-1.55561256e+00 -7.78672934e-01 -8.98041606e-01 1.78480804e-01
1.50350511e-01 -9.62018371e-01 1.24391365e+00 -1.02654123e+00
-1.15778959e+00 1.01668644e+00 -4.01988596e-01 -6.23239577e-01
2.88452923e-01 -3.59392673e-01 1.96206316e-01 1.45988733e-01
1.79727912e-01 1.14913607e+00 1.17879367e+00 -1.46390462e+00
-8.56545746e-01 -3.68416071e-01 -3.56386006e-02 3.34544748e-01
-5.02838016e-01 -3.46767157e-01 -6.08036578e-01 -5.90849936e-01
-5.44851124e-02 -8.02487493e-01 -2.93092072e-01 5.05110145e-01
-5.24685800e-01 -3.21121365e-01 8.03013563e-01 -2.41681173e-01
6.48773611e-01 -2.15959072e+00 -1.23573365e-04 -6.28412664e-02
2.86547154e-01 1.45712778e-01 -2.57515013e-01 4.68035787e-02
4.23731416e-01 -3.88981581e-01 -2.60240644e-01 -6.25079453e-01
3.80363651e-02 3.72623980e-01 -6.84027612e-01 3.68581533e-01
4.66801554e-01 1.05791402e+00 -7.85248756e-01 -7.10360885e-01
4.71191376e-01 6.49282038e-01 -4.39151108e-01 4.51014519e-01
-7.07840383e-01 1.96935728e-01 -7.46688128e-01 8.15661550e-01
3.99249882e-01 -5.06556571e-01 -2.02733979e-01 -3.30385685e-01
3.45754325e-01 -4.96272929e-02 -6.59134746e-01 1.79695964e+00
-5.85235059e-01 5.52418351e-01 -2.25878567e-01 -1.09920573e+00
9.68313813e-01 -1.40738457e-01 1.22970261e-01 -6.84633791e-01
-7.37910196e-02 -1.37735412e-01 -2.45944679e-01 -3.03974599e-01
2.19166026e-01 -2.03839272e-01 9.05032456e-02 1.69608429e-01
2.20218927e-01 -3.09648421e-02 -1.61061019e-01 4.02098805e-01
6.43531621e-01 3.91281277e-01 2.22241551e-01 -6.18617050e-02
4.02014285e-01 1.80259496e-01 4.43985403e-01 1.13831592e+00
-3.37459087e-01 7.11542785e-01 9.03015137e-02 -3.73342812e-01
-9.75222945e-01 -1.32600844e+00 -2.81472087e-01 1.33883560e+00
2.52658814e-01 -1.05540261e-01 -2.70234883e-01 -7.56132424e-01
8.47062394e-02 8.54042470e-01 -7.12223589e-01 -3.54074478e-01
-8.13107193e-02 -6.65167987e-01 1.56568512e-01 8.39856327e-01
5.15399456e-01 -1.15522337e+00 -3.52276206e-01 4.09537517e-02
-1.02004625e-01 -1.31688225e+00 -3.30411464e-01 3.78093243e-01
-7.99566865e-01 -6.61345005e-01 -7.79808342e-01 -9.09589648e-01
7.09270716e-01 2.91939706e-01 1.11244965e+00 -2.66343027e-01
-5.28123856e-01 7.47345150e-01 -1.55805603e-01 -4.64611948e-01
-1.15284592e-01 -5.37296943e-02 -1.35237217e-01 2.44454801e-01
5.36208689e-01 -4.18680906e-01 -6.78258657e-01 -1.37704238e-01
-4.06973183e-01 3.40231121e-01 1.05572605e+00 1.13937426e+00
1.07057798e+00 -4.02273625e-01 3.92379910e-01 -1.03129792e+00
2.11813867e-01 -4.72159892e-01 -4.14904654e-01 3.55462849e-01
-6.68138444e-01 3.13849598e-01 5.48291981e-01 -6.28902912e-01
-1.27472949e+00 6.73831820e-01 7.69236460e-02 -6.73622072e-01
-3.63315523e-01 2.74401337e-01 -2.87723899e-01 9.89680216e-02
5.33288360e-01 6.83479726e-01 -1.69304907e-01 -3.34821552e-01
7.14009285e-01 3.96195114e-01 6.72283769e-01 -6.45937502e-01
7.67672896e-01 5.60901403e-01 -3.75417992e-02 -6.46821856e-01
-1.42819071e+00 -5.35208941e-01 -5.19285560e-01 -4.41625267e-02
9.05709565e-01 -1.30795848e+00 -3.98638606e-01 3.91912788e-01
-1.16354454e+00 -3.58680934e-01 -2.65231520e-01 4.40879017e-01
-7.71378100e-01 8.46044123e-02 -4.13463444e-01 -8.65217686e-01
-4.99107957e-01 -9.86204743e-01 1.31209910e+00 3.29743296e-01
8.88767317e-02 -8.29543352e-01 -2.15926379e-01 2.88701326e-01
3.74104887e-01 -1.81335002e-01 1.10879052e+00 -6.21430755e-01
-9.44059849e-01 -2.31696457e-01 -7.11147249e-01 3.39537561e-01
-8.54701474e-02 -1.09030165e-01 -1.28487122e+00 -4.06203009e-02
-4.12201077e-01 -8.69226336e-01 1.33153927e+00 5.94554543e-01
1.64889622e+00 -1.98011979e-01 -5.66530645e-01 7.87512720e-01
1.52309084e+00 -1.77702934e-01 3.93448919e-01 1.27016559e-01
1.03154957e+00 5.07651150e-01 7.40997612e-01 6.13096595e-01
4.58433449e-01 5.01553237e-01 5.00571787e-01 -1.87593728e-01
-5.56557536e-01 -5.90301991e-01 2.36519158e-01 6.37755170e-02
1.86203703e-01 -1.45683840e-01 -9.23874617e-01 5.72813272e-01
-1.79376352e+00 -9.99873340e-01 1.90841258e-01 1.85357082e+00
9.07556415e-01 1.80088267e-01 -1.51761055e-01 -3.83705020e-01
5.70669889e-01 2.63464481e-01 -7.88717508e-01 1.93190232e-01
-2.47252852e-01 2.07459908e-02 2.72937924e-01 3.28287989e-01
-1.22740924e+00 1.18440187e+00 5.96717024e+00 9.22908425e-01
-1.09504461e+00 -6.97346106e-02 7.11112559e-01 -1.49489552e-01
-5.25937557e-01 -9.06692371e-02 -1.24366379e+00 1.84125185e-01
5.58063984e-01 -8.37503932e-03 8.62637386e-02 1.13304520e+00
6.70485795e-02 -4.80438359e-02 -1.22830057e+00 1.23582900e+00
1.80620581e-01 -1.55756128e+00 6.15777075e-01 -2.01466352e-01
6.79903150e-01 1.70020223e-01 5.36079943e-01 4.22367245e-01
4.00455505e-01 -1.11893034e+00 6.78141117e-01 7.43184566e-01
8.46508920e-01 -5.04111707e-01 3.71075392e-01 2.70187050e-01
-9.74430561e-01 -3.01849723e-01 -7.04152286e-01 1.56811774e-01
-2.05664933e-02 5.51720917e-01 -1.01499128e+00 2.38356902e-03
7.28848100e-01 1.10217702e+00 -6.60815597e-01 1.07489920e+00
-2.34563559e-01 7.31998086e-01 -1.82861388e-01 -2.27457866e-01
4.32382584e-01 5.63893579e-02 3.08117270e-01 1.22877896e+00
1.19071439e-01 2.05980409e-02 4.32663053e-01 1.15997648e+00
-1.10076241e-01 8.52222461e-03 -6.01892352e-01 -1.25494301e-01
4.47905093e-01 1.25766766e+00 -3.29052299e-01 -2.59367973e-01
-7.03027606e-01 9.75601852e-01 6.30332410e-01 6.21988297e-01
-5.68341434e-01 9.59465578e-02 5.94896078e-01 1.00856863e-01
5.70382655e-01 -1.16860911e-01 -5.27564406e-01 -9.03528273e-01
-1.57857925e-01 -4.77123916e-01 4.05584991e-01 -9.36561882e-01
-1.64845538e+00 5.85492790e-01 -1.21037746e-02 -1.03588986e+00
-3.19581628e-01 -7.70425379e-01 -4.65277612e-01 9.77106631e-01
-1.78724313e+00 -1.80870843e+00 -6.31850362e-01 9.30535376e-01
8.21374774e-01 -5.26557982e-01 9.03836370e-01 -4.11870703e-02
-2.82480896e-01 6.36607885e-01 1.27386093e-01 -5.66823147e-02
7.26829827e-01 -1.26806819e+00 -1.17804505e-01 6.61896586e-01
4.77920294e-01 3.85518283e-01 5.55871189e-01 -5.17475724e-01
-1.26284504e+00 -1.39631951e+00 5.12414396e-01 -5.78657925e-01
6.07453763e-01 -6.01345658e-01 -8.94201040e-01 6.14779949e-01
5.95483482e-02 4.88960505e-01 6.28838539e-01 3.20826203e-01
-6.78699970e-01 -4.27052170e-01 -7.33622372e-01 3.87667149e-01
9.86229479e-01 -7.98618674e-01 -4.81272608e-01 4.05564964e-01
6.89645290e-01 -5.11852950e-02 -3.23974013e-01 3.97179574e-01
4.19007689e-01 -6.13919616e-01 1.09414637e+00 -5.21256924e-01
5.83052933e-01 -2.37554163e-01 -5.07854640e-01 -9.36307371e-01
-5.24588585e-01 5.79523528e-03 -2.38564953e-01 1.29375756e+00
5.15309453e-01 -6.57904893e-02 8.62637818e-01 4.79363471e-01
-1.68632716e-01 -8.05998802e-01 -5.88262618e-01 -4.27019984e-01
5.03459871e-02 -5.13897300e-01 2.04823226e-01 5.45000196e-01
-5.76298535e-01 6.20805264e-01 -5.67135930e-01 2.61786997e-01
9.67320859e-01 5.57662666e-01 4.62031454e-01 -1.30786431e+00
-3.90182227e-01 -2.84538478e-01 -4.06640828e-01 -1.35433018e+00
4.85087961e-01 -1.00778079e+00 4.01962399e-01 -1.69706810e+00
8.35172772e-01 -6.95900261e-01 -4.67350721e-01 6.48523211e-01
-4.43011820e-01 2.11891398e-01 1.40147403e-01 3.73660624e-01
-8.32909703e-01 1.03924203e+00 1.20754290e+00 -6.17667615e-01
2.26821885e-01 1.17586426e-01 -7.69727945e-01 6.20658278e-01
4.56405222e-01 -4.13693577e-01 -8.30997527e-01 -4.12453234e-01
-1.29768372e-01 -3.04985583e-01 7.29573905e-01 -8.86670053e-01
2.62878329e-01 -2.94226617e-01 9.62872863e-01 -7.07001150e-01
4.94990259e-01 -8.87860715e-01 -4.75545287e-01 1.50420427e-01
-6.70453727e-01 -6.14989281e-01 1.96102411e-01 1.14076054e+00
-2.08917290e-01 6.23720931e-03 9.16903853e-01 -5.32926731e-02
-1.32019651e+00 8.78600597e-01 -1.06013931e-01 8.29183236e-02
1.02409351e+00 -1.98809534e-01 -2.38603964e-01 -3.52996498e-01
-8.31783414e-01 2.44350776e-01 3.10026228e-01 4.67350066e-01
1.02949333e+00 -1.27283096e+00 -8.08545470e-01 1.47767842e-01
7.04367816e-01 6.86463993e-03 2.83902377e-01 4.69336867e-01
5.54028824e-02 3.26043338e-01 -2.60786295e-01 -1.10340595e+00
-1.00637901e+00 7.98929751e-01 2.52588570e-01 1.03725344e-01
-9.01615500e-01 1.08114612e+00 9.23184454e-01 -6.31986037e-02
5.12106776e-01 -8.79641250e-02 -2.14562431e-01 -2.07453415e-01
6.58805370e-01 -4.71269161e-01 -4.38789010e-01 -4.43630010e-01
-3.38541955e-01 5.11436284e-01 -4.15124655e-01 8.34513754e-02
1.24158764e+00 -3.62064004e-01 3.02334607e-01 5.11190891e-01
9.55083072e-01 -5.05538642e-01 -1.91529870e+00 -7.80091405e-01
-4.90052067e-02 -4.19861883e-01 2.92658776e-01 -1.05910027e+00
-1.03317499e+00 1.33482301e+00 8.30720842e-01 -5.29712021e-01
1.13909245e+00 3.70949626e-01 5.15612960e-01 4.57014054e-01
2.03207329e-01 -7.47850180e-01 3.04915786e-01 4.41436291e-01
8.32314074e-01 -1.57541001e+00 -2.07938254e-01 -4.90695298e-01
-1.06828821e+00 9.21096981e-01 9.95691955e-01 3.92342843e-02
5.53914487e-01 1.39323041e-01 -1.96138751e-02 3.87485069e-03
-9.21772003e-01 -4.91532534e-01 5.70380151e-01 1.14760387e+00
3.15620273e-01 8.86238180e-03 3.12377602e-01 7.53369868e-01
2.65877247e-01 -1.20796464e-01 8.43130499e-02 4.08775926e-01
-6.97810173e-01 -7.94507205e-01 -5.74327260e-03 7.32861876e-01
-8.03762153e-02 -4.65626538e-01 -2.08713427e-01 4.86830533e-01
6.28466206e-03 5.82894146e-01 2.02413127e-01 -1.29476935e-01
6.82491623e-03 3.78070660e-02 6.21436298e-01 -8.10916662e-01
5.70954680e-02 -6.98445663e-02 -1.35323524e-01 -4.77102846e-01
-3.80792767e-01 -6.15092456e-01 -1.14095855e+00 9.91338566e-02
1.38036937e-01 -3.09616804e-01 3.31554860e-01 9.79942620e-01
2.64169246e-01 2.75072157e-01 4.37151879e-01 -7.47541666e-01
-6.64306760e-01 -7.87768185e-01 -5.53600550e-01 4.75924164e-01
4.69426394e-01 -8.81153643e-01 -1.25975922e-01 3.07312340e-01]
|
[10.080221176147461, 1.885063886642456]
|
bf998df3-ee7b-4c67-9c7c-1e28ceea10ae
|
make-a-voice-unified-voice-synthesis-with
|
2305.19269
| null |
https://arxiv.org/abs/2305.19269v1
|
https://arxiv.org/pdf/2305.19269v1.pdf
|
Make-A-Voice: Unified Voice Synthesis With Discrete Representation
|
Various applications of voice synthesis have been developed independently despite the fact that they generate "voice" as output in common. In addition, the majority of voice synthesis models currently rely on annotated audio data, but it is crucial to scale them to self-supervised datasets in order to effectively capture the wide range of acoustic variations present in human voice, including speaker identity, emotion, and prosody. In this work, we propose Make-A-Voice, a unified framework for synthesizing and manipulating voice signals from discrete representations. Make-A-Voice leverages a "coarse-to-fine" approach to model the human voice, which involves three stages: 1) semantic stage: model high-level transformation between linguistic content and self-supervised semantic tokens, 2) acoustic stage: introduce varying control signals as acoustic conditions for semantic-to-acoustic modeling, and 3) generation stage: synthesize high-fidelity waveforms from acoustic tokens. Make-A-Voice offers notable benefits as a unified voice synthesis framework: 1) Data scalability: the major backbone (i.e., acoustic and generation stage) does not require any annotations, and thus the training data could be scaled up. 2) Controllability and conditioning flexibility: we investigate different conditioning mechanisms and effectively handle three voice synthesis applications, including text-to-speech (TTS), voice conversion (VC), and singing voice synthesis (SVS) by re-synthesizing the discrete voice representations with prompt guidance. Experimental results demonstrate that Make-A-Voice exhibits superior audio quality and style similarity compared with competitive baseline models. Audio samples are available at https://Make-A-Voice.github.io
|
['Dong Yu', 'Zhou Zhao', 'Chao Weng', 'Ziyue Jiang', 'Zhenhui Ye', 'Luping Liu', 'Dongchao Yang', 'Yongqi Wang', 'Chunlei Zhang', 'Rongjie Huang']
|
2023-05-30
| null | null | null | null |
['voice-conversion', 'voice-conversion', 'singing-voice-synthesis']
|
['audio', 'speech', 'speech']
|
[ 6.36159629e-02 -4.64902958e-03 -9.39518660e-02 -2.54878402e-01
-1.07053220e+00 -7.91116476e-01 3.27221841e-01 -3.18817556e-01
2.60050625e-01 3.45759243e-01 5.86601019e-01 -2.97029793e-01
3.04585010e-01 -4.80206877e-01 -6.50815189e-01 -5.15680432e-01
3.35493386e-01 1.94249719e-01 -1.01744786e-01 -4.03592259e-01
-3.85557920e-01 4.35051084e-01 -1.77255416e+00 3.67305338e-01
6.89466894e-01 9.66076493e-01 3.14038604e-01 8.99143815e-01
-3.42811376e-01 3.39491487e-01 -6.98406875e-01 2.59191380e-03
1.79647550e-01 -6.96408451e-01 -5.62505305e-01 1.93838716e-01
1.63628086e-01 -1.22186512e-01 -5.64104281e-02 7.90574610e-01
9.45526421e-01 3.45841706e-01 5.26151001e-01 -1.19465411e+00
-6.42244220e-01 7.90046275e-01 2.19324186e-01 -1.92778185e-01
5.17844260e-01 5.14206767e-01 1.29752123e+00 -1.12738132e+00
5.11846840e-01 1.51281488e+00 5.87115407e-01 1.09303248e+00
-1.46283352e+00 -8.08569968e-01 5.80661781e-02 -3.31765383e-01
-1.15967226e+00 -1.17875409e+00 9.29030895e-01 -3.98067772e-01
8.01944137e-01 5.65813661e-01 5.28743863e-01 1.31111777e+00
-2.79367715e-01 8.40411603e-01 8.65968227e-01 -5.63919902e-01
2.99728334e-01 1.86756611e-01 -3.76652986e-01 3.79457325e-01
-6.98249102e-01 3.03007126e-01 -7.34549105e-01 -1.08977504e-01
7.37138510e-01 -4.49145377e-01 -3.59758168e-01 9.96150672e-02
-1.11890852e+00 7.01774895e-01 6.19919598e-02 1.93872586e-01
-3.02083164e-01 1.76355958e-01 4.55743879e-01 4.32816058e-01
2.88968980e-01 5.31648219e-01 -4.76319075e-01 -2.17466712e-01
-1.06414890e+00 4.66148347e-01 9.52014863e-01 1.12325668e+00
4.39586818e-01 9.18193996e-01 -2.42940068e-01 1.39151263e+00
4.66461539e-01 7.90389180e-01 9.02132988e-01 -1.09131253e+00
3.60434443e-01 -2.30817884e-01 -9.23098698e-02 -3.42501998e-01
-2.13958919e-01 -3.59961182e-01 -5.64958572e-01 -2.55457149e-03
1.20994963e-01 -3.20998877e-01 -9.59076047e-01 1.95068610e+00
3.23142141e-01 2.42784664e-01 1.80200592e-01 9.02583957e-01
1.14612770e+00 9.68326628e-01 -6.44851401e-02 -2.09245726e-01
1.21525013e+00 -1.13334882e+00 -1.06205821e+00 -2.10792109e-01
7.98221380e-02 -1.08708763e+00 1.73679888e+00 2.10767269e-01
-1.46306741e+00 -9.62581336e-01 -8.06351840e-01 -3.71819139e-02
-1.90136477e-01 1.57055289e-01 2.78507978e-01 7.33396471e-01
-1.09171391e+00 4.23581630e-01 -5.43317258e-01 8.73701833e-03
-1.18504435e-01 2.10106775e-01 -7.82575682e-02 5.40952802e-01
-1.35308337e+00 2.19872087e-01 -2.62675099e-02 -1.06585659e-01
-1.14415407e+00 -1.06096077e+00 -9.86864090e-01 6.89801872e-02
1.85442612e-01 -7.10114598e-01 1.85459566e+00 -9.09709096e-01
-2.39149523e+00 5.03181458e-01 -4.15151954e-01 -1.31735787e-01
2.33928889e-01 -2.57148802e-01 -8.18582892e-01 -8.74963775e-02
1.32943718e-02 7.26283431e-01 1.16487718e+00 -1.40655398e+00
-4.64598715e-01 2.79447109e-01 -4.84835654e-01 1.91673443e-01
-1.81451946e-01 2.44177803e-01 -2.41821945e-01 -1.08500051e+00
-5.02090901e-02 -8.91047180e-01 -4.83532250e-02 -2.16799691e-01
-6.24132097e-01 -5.00499643e-02 6.55215263e-01 -5.49288392e-01
1.33004785e+00 -2.35872889e+00 1.16485968e-01 8.12784582e-02
-1.86529130e-01 4.13129151e-01 -4.28741634e-01 4.45389301e-01
-1.65064380e-01 2.57463217e-01 -1.05063491e-01 -5.82182586e-01
3.04226041e-01 8.61133114e-02 -9.64939058e-01 -1.63387150e-01
4.43664700e-01 7.74500132e-01 -8.32533240e-01 -3.47577214e-01
2.06090197e-01 4.77802008e-01 -1.07659090e+00 6.19691849e-01
-5.53552151e-01 7.15266705e-01 -2.75178611e-01 7.65343070e-01
7.19759688e-02 3.83274168e-01 -1.03696827e-02 -2.52189159e-01
-1.41280651e-01 8.90767038e-01 -1.43885756e+00 1.59022152e+00
-8.10387015e-01 3.33948821e-01 7.21221447e-01 -3.97602767e-01
1.12257206e+00 8.85410249e-01 3.84600610e-01 -3.13595116e-01
1.10835418e-01 4.93543178e-01 1.96148381e-02 -4.30802613e-01
4.81337190e-01 -5.27627170e-01 -1.10406384e-01 2.59508371e-01
3.54217261e-01 -1.02362096e+00 -1.03142574e-01 -3.06346595e-01
6.26939297e-01 1.64676696e-01 -6.41642185e-03 -9.34789777e-02
4.32197809e-01 -4.00862247e-01 8.31118941e-01 2.81864852e-01
-1.15433030e-01 8.98426354e-01 4.09967043e-02 3.17664146e-01
-8.86690497e-01 -1.44218957e+00 2.03026198e-02 1.30404639e+00
-3.37629795e-01 -4.89722520e-01 -1.04790998e+00 6.17540032e-02
8.99767950e-02 9.99490559e-01 5.36402091e-02 -1.35108829e-01
-5.74881673e-01 -2.33635064e-02 9.30069029e-01 4.77318585e-01
-1.02958038e-01 -1.35218370e+00 1.05034776e-01 5.22779524e-01
-2.08887130e-01 -1.11908638e+00 -1.07663858e+00 2.12296858e-01
-7.29990900e-01 -4.08391535e-01 -4.59565103e-01 -8.01920950e-01
2.48411801e-02 -2.40227953e-02 9.56121504e-01 -2.38886341e-01
-1.30545363e-01 5.13304770e-01 -2.40380436e-01 -6.85853779e-01
-1.04986763e+00 -3.16183977e-02 6.83184266e-01 2.20232621e-01
-3.84826481e-01 -8.33782732e-01 -2.98954338e-01 3.76112312e-01
-8.44574451e-01 -1.83220074e-01 3.09698194e-01 7.61899769e-01
9.09345686e-01 -6.38030767e-02 1.24273574e+00 -6.90504968e-01
9.99471307e-01 -3.70490164e-01 -2.68486679e-01 -7.16431364e-02
-3.19298834e-01 -2.36761808e-01 1.14850056e+00 -7.21321046e-01
-1.05374253e+00 5.81558757e-02 -6.57628417e-01 -7.88272083e-01
-3.41553152e-01 2.72654444e-01 -6.86205328e-01 5.63529313e-01
7.16883838e-01 2.15004668e-01 1.11109212e-01 -7.51128137e-01
8.52169633e-01 1.13075185e+00 8.91355455e-01 -9.11830485e-01
9.65785205e-01 -4.09806520e-02 -4.69712526e-01 -1.21132660e+00
-5.47276497e-01 -3.58731240e-01 -4.55370158e-01 -1.54071487e-02
6.72946155e-01 -1.04014063e+00 -5.36435306e-01 4.08655941e-01
-1.02038860e+00 -7.56660044e-01 -7.76391506e-01 5.82589388e-01
-8.61784160e-01 -6.96746260e-02 -7.21644461e-01 -9.10849214e-01
-4.55334991e-01 -1.35490584e+00 1.14891040e+00 1.07218154e-01
-5.38548112e-01 -9.34450567e-01 -1.04085602e-01 3.47398788e-01
6.54491484e-01 -3.98655087e-02 8.68552685e-01 -5.41001201e-01
-1.86158940e-01 3.17992747e-01 5.44320345e-01 7.36926436e-01
5.50123394e-01 2.58407772e-01 -1.45272696e+00 -7.43019953e-02
-6.36859462e-02 -4.79727179e-01 3.00360113e-01 3.14317524e-01
1.00853121e+00 -4.22264308e-01 2.22060457e-01 7.01863050e-01
5.51216662e-01 1.86492488e-01 2.88168520e-01 -3.91474724e-01
6.60343468e-01 5.28438628e-01 4.40745980e-01 3.60743821e-01
1.73865005e-01 7.10341990e-01 -7.82368239e-03 -1.72022879e-01
-8.68365824e-01 -7.36206651e-01 8.49647999e-01 1.69589031e+00
4.28422600e-01 -1.91229492e-01 -5.24110198e-01 6.41571701e-01
-1.25749493e+00 -7.65611649e-01 1.30270362e-01 1.97560751e+00
1.42388892e+00 3.36663574e-02 3.17701459e-01 3.89745653e-01
7.65834570e-01 2.34629095e-01 -5.72683036e-01 -7.23419666e-01
1.64302588e-01 5.91119170e-01 -1.96144804e-01 7.47956157e-01
-7.11997092e-01 1.29529810e+00 6.17324114e+00 9.69154358e-01
-1.56314850e+00 1.02079205e-01 1.89552352e-01 -3.37134123e-01
-7.35299468e-01 -1.97551921e-01 -9.21114326e-01 3.51176649e-01
1.14624250e+00 -1.22775368e-01 9.16250765e-01 7.99991310e-01
7.04910755e-01 7.64095366e-01 -1.35731888e+00 8.47299755e-01
-2.15287477e-01 -1.16310740e+00 3.27770263e-01 -3.40008318e-01
5.56198657e-01 -1.91565633e-01 2.51756459e-01 5.06889224e-01
1.77192122e-01 -1.08536780e+00 1.39415157e+00 4.54368472e-01
1.43173635e+00 -4.63306248e-01 -1.46794990e-01 2.69828320e-01
-1.45046473e+00 5.42614348e-02 8.47692937e-02 2.70452410e-01
5.12195766e-01 3.86280894e-01 -9.94339049e-01 3.08833718e-01
3.60500425e-01 2.75378168e-01 1.67596534e-01 6.34175420e-01
-3.39748383e-01 1.29547775e+00 -1.85680345e-01 1.61528572e-01
3.39447893e-02 1.52297571e-01 8.94941688e-01 1.31539965e+00
3.96405607e-01 -5.53416833e-03 3.44538510e-01 1.19929957e+00
-2.26241305e-01 2.99144059e-01 -3.18161339e-01 -4.24048990e-01
1.02495372e+00 9.79152262e-01 -1.65817559e-01 -1.57273673e-02
-1.02803625e-01 7.54853725e-01 -1.89598516e-01 4.52660382e-01
-6.25207245e-01 -4.83692735e-01 1.06844687e+00 2.53550649e-01
1.98596314e-01 -1.52154297e-01 -9.21160579e-02 -8.55621696e-01
-1.94612667e-01 -1.33593714e+00 -1.14097886e-01 -8.51899385e-01
-1.26043069e+00 7.62704790e-01 -3.48969907e-01 -1.14284158e+00
-6.81315660e-01 -4.99067456e-01 -7.79405594e-01 1.07936394e+00
-1.42752302e+00 -1.19795537e+00 3.51381861e-02 6.08257174e-01
1.08440590e+00 -4.09123510e-01 1.12857985e+00 3.82233113e-01
-5.07629931e-01 8.37005913e-01 -6.50777370e-02 -5.04386351e-02
9.00568128e-01 -1.22169471e+00 7.11198747e-01 4.72681493e-01
2.17835233e-01 6.05187058e-01 6.56002462e-01 -3.53924781e-01
-1.50830209e+00 -1.24021935e+00 7.70258427e-01 -3.52588117e-01
7.63538003e-01 -6.18183434e-01 -9.09644365e-01 4.79838043e-01
-6.16866201e-02 1.37581319e-01 9.10776794e-01 -5.27332537e-02
-3.58910620e-01 -3.39669645e-01 -9.71244276e-01 7.67807484e-01
9.13161218e-01 -1.03577685e+00 -5.92997670e-01 -2.15245672e-02
1.25614262e+00 -5.55574477e-01 -8.83856237e-01 3.05149496e-01
3.50988030e-01 -5.80188751e-01 9.26599383e-01 -6.49268091e-01
2.25167066e-01 -3.99220347e-01 -4.37000722e-01 -1.69353342e+00
-1.50062829e-01 -1.48016953e+00 -1.28662691e-01 1.84607255e+00
6.42176151e-01 -5.51968455e-01 7.21429214e-02 3.67894858e-01
-6.78215563e-01 -6.02822542e-01 -7.35775292e-01 -8.83854806e-01
2.03344017e-01 -8.35992038e-01 9.98626173e-01 8.41249347e-01
-6.99381754e-02 4.03943539e-01 -2.78331310e-01 2.08565280e-01
3.69725786e-02 -3.59379570e-03 8.65287960e-01 -9.23295379e-01
-6.33969843e-01 -5.32648325e-01 3.02976131e-01 -1.14579666e+00
3.53979588e-01 -1.16504323e+00 4.67086285e-01 -1.19114769e+00
-7.17350543e-01 -7.03891873e-01 -2.83541352e-01 5.41959882e-01
-9.84040201e-02 -6.85189515e-02 4.36068028e-01 2.17325062e-01
8.63001123e-02 8.17452490e-01 1.37899947e+00 1.15645595e-01
-6.81175113e-01 4.17575538e-01 -6.57878876e-01 7.49750555e-01
7.32028544e-01 -2.79427171e-01 -5.96339703e-01 -1.98230460e-01
-3.53758156e-01 5.21547735e-01 1.96890995e-01 -8.89866471e-01
5.62189929e-02 -3.13849598e-01 -2.09042281e-01 -8.68955776e-02
6.46337867e-01 -4.57968205e-01 1.24036491e-01 -7.65494537e-03
-6.01049304e-01 -2.51865506e-01 3.72188389e-01 2.49332264e-01
-3.71033132e-01 -1.90229252e-01 7.42359340e-01 8.56081620e-02
-2.93242693e-01 1.02966845e-01 -5.32074094e-01 4.36014265e-01
3.14408511e-01 1.32814541e-01 -3.33799049e-02 -6.23446226e-01
-8.86966288e-01 -1.04332782e-01 4.34559621e-02 7.66471386e-01
5.39510071e-01 -1.48993218e+00 -7.49915540e-01 5.22615612e-01
-7.45799094e-02 4.88432087e-02 7.17010051e-02 2.92698562e-01
-1.09253265e-02 2.99606413e-01 2.46933743e-01 -4.05055434e-01
-1.00912058e+00 2.78248280e-01 4.25795466e-01 3.03900540e-01
-4.97578442e-01 7.89214432e-01 2.10626289e-01 -9.16442573e-01
4.65556234e-01 -5.39813817e-01 1.77439928e-01 -5.30802063e-04
3.24750423e-01 2.90392160e-01 -5.06276041e-02 -7.94898272e-01
-1.87198102e-01 3.78930569e-01 4.31530893e-01 -6.18861854e-01
1.15360224e+00 -1.27118751e-01 2.52579391e-01 8.79107535e-01
9.63196278e-01 6.69342756e-01 -1.21185720e+00 -2.38061771e-01
-2.21031025e-01 -1.10003941e-01 1.05418563e-01 -8.69732082e-01
-9.40468192e-01 9.49413538e-01 2.04260543e-01 1.36088878e-01
1.02221930e+00 -9.98160243e-02 1.31368339e+00 2.13970572e-01
4.98604551e-02 -1.29299915e+00 1.24149285e-01 6.67248547e-01
1.31821513e+00 -6.87548280e-01 -5.82302153e-01 -4.65776414e-01
-9.39392626e-01 9.37248230e-01 3.64521205e-01 1.27456293e-01
7.71330416e-01 6.48466825e-01 6.26277208e-01 2.61258721e-01
-8.84142339e-01 -2.48281658e-01 6.22553170e-01 5.80544770e-01
6.66781306e-01 2.59991229e-01 2.79446065e-01 1.14111233e+00
-9.12680089e-01 -1.80348501e-01 2.53212661e-01 3.63889664e-01
-3.36374730e-01 -1.29657745e+00 -5.37805080e-01 8.40763226e-02
-4.01253670e-01 -2.27332681e-01 -3.93162608e-01 3.31809402e-01
6.93057477e-02 1.40722311e+00 -1.65322796e-01 -5.07708371e-01
6.84152007e-01 5.20382822e-01 5.33744581e-02 -1.00134599e+00
-8.93532693e-01 6.03790283e-01 3.40935260e-01 -4.65900064e-01
1.10507585e-01 -6.81713700e-01 -1.76291800e+00 7.44878575e-02
-3.69936645e-01 2.00660527e-01 8.60987365e-01 7.08047092e-01
4.82547492e-01 9.97765660e-01 1.09483838e+00 -9.86753583e-01
-9.54305887e-01 -9.10719812e-01 -6.69570446e-01 3.29510480e-01
6.25449598e-01 -4.25511599e-01 -6.36347115e-01 3.79158229e-01]
|
[15.103572845458984, 6.511280536651611]
|
5050f452-967b-4a16-b55c-347e889a8a03
|
mmkgr-multi-hop-multi-modal-knowledge-graph
|
2209.01416
| null |
https://arxiv.org/abs/2209.01416v1
|
https://arxiv.org/pdf/2209.01416v1.pdf
|
MMKGR: Multi-hop Multi-modal Knowledge Graph Reasoning
|
Multi-modal knowledge graphs (MKGs) include not only the relation triplets, but also related multi-modal auxiliary data (i.e., texts and images), which enhance the diversity of knowledge. However, the natural incompleteness has significantly hindered the applications of MKGs. To tackle the problem, existing studies employ the embedding-based reasoning models to infer the missing knowledge after fusing the multi-modal features. However, the reasoning performance of these methods is limited due to the following problems: (1) ineffective fusion of multi-modal auxiliary features; (2) lack of complex reasoning ability as well as inability to conduct the multi-hop reasoning which is able to infer more missing knowledge. To overcome these problems, we propose a novel model entitled MMKGR (Multi-hop Multi-modal Knowledge Graph Reasoning). Specifically, the model contains the following two components: (1) a unified gate-attention network which is designed to generate effective multi-modal complementary features through sufficient attention interaction and noise reduction; (2) a complementary feature-aware reinforcement learning method which is proposed to predict missing elements by performing the multi-hop reasoning process, based on the features obtained in component (1). The experimental results demonstrate that MMKGR outperforms the state-of-the-art approaches in the MKG reasoning task.
|
['Lei Zhao', 'Wei Chen', 'Hongzhi Yin', 'Jianfeng Qu', 'Weiqing Wang', 'Shangfei Zheng']
|
2022-09-03
| null | null | null | null |
['multi-modal-knowledge-graph']
|
['knowledge-base']
|
[-1.15662873e-01 4.49827284e-01 -1.87793270e-01 1.44488756e-02
-8.23702574e-01 -7.04968125e-02 4.84290302e-01 -4.53060567e-02
-2.01817930e-01 7.20461726e-01 4.52345312e-01 -9.29915383e-02
-6.00111067e-01 -1.14693749e+00 -6.18184566e-01 -6.25512242e-01
4.06026185e-01 3.60233724e-01 3.42797309e-01 -6.97760522e-01
1.10898726e-01 2.06136517e-03 -1.75426733e+00 5.65620542e-01
1.23751307e+00 1.07127452e+00 3.22874397e-01 1.50457770e-01
-4.76907313e-01 1.26577508e+00 -2.65736371e-01 -6.64124370e-01
-1.04398809e-01 -4.98697758e-01 -9.33248341e-01 -8.29148814e-02
-1.07023545e-01 -5.79401672e-01 -6.98457479e-01 1.10596085e+00
4.87418860e-01 3.74909639e-01 4.99232829e-01 -1.36317492e+00
-1.31574309e+00 9.30539131e-01 -4.36751157e-01 1.49987534e-01
5.94370246e-01 1.22451626e-01 1.24863434e+00 -8.77400994e-01
3.97261292e-01 1.50648689e+00 4.70120907e-01 3.52067173e-01
-6.64760649e-01 -5.43287218e-01 2.88076758e-01 9.65993941e-01
-1.55061305e+00 -1.66061550e-01 1.15934277e+00 -1.72657654e-01
9.33891475e-01 -5.02149463e-02 6.50075853e-01 1.04438984e+00
-5.37059829e-02 1.02296519e+00 8.97261143e-01 -4.68935490e-01
-1.27578065e-01 1.03387021e-01 1.57370657e-01 9.86083448e-01
-4.93625291e-02 -7.11373612e-02 -5.15963197e-01 4.74583767e-02
6.42745197e-01 7.31135607e-02 -2.86474258e-01 -4.18857997e-03
-1.34062278e+00 9.56943274e-01 7.73350060e-01 4.59243327e-01
-5.68240881e-01 -7.90453404e-02 2.34123051e-01 2.01692358e-01
1.31347179e-01 1.94887012e-01 -5.14511883e-01 1.62658557e-01
-2.74442792e-01 7.78816864e-02 4.87069041e-01 1.06193674e+00
9.59340930e-01 3.43084000e-02 -4.59921867e-01 7.97533751e-01
6.27492845e-01 4.75518703e-01 6.32446587e-01 -7.00723231e-01
8.96185875e-01 1.27869403e+00 -2.02216402e-01 -1.36027551e+00
-5.28219819e-01 -2.02411696e-01 -9.30581391e-01 -3.57752562e-01
1.11897839e-02 -2.34576184e-02 -7.29082823e-01 1.85572624e+00
4.08900708e-01 8.17222297e-02 4.17350501e-01 9.60672736e-01
1.65453148e+00 4.74988878e-01 8.40752572e-02 -6.97744638e-02
1.53414142e+00 -1.23563242e+00 -1.03355443e+00 -1.48625761e-01
3.50622684e-01 -4.07793611e-01 9.69938397e-01 4.00911681e-02
-8.30829561e-01 -7.22738206e-01 -9.69122767e-01 -2.92218208e-01
-9.16809440e-01 1.78589776e-01 7.32384741e-01 1.56136453e-01
-8.70879471e-01 2.07591087e-01 -2.91282117e-01 -1.11669019e-01
4.19633865e-01 2.30256006e-01 -3.97062987e-01 -5.06413162e-01
-1.95306146e+00 1.07565081e+00 1.04236615e+00 4.27676499e-01
-4.78255332e-01 -5.09007931e-01 -1.11404407e+00 1.32019341e-01
9.84685481e-01 -9.31152403e-01 6.05207443e-01 -6.32523417e-01
-1.17009342e+00 3.44871789e-01 6.42971173e-02 1.61861971e-01
3.83205235e-01 -1.53651282e-01 -8.40837419e-01 4.31784183e-01
2.15648860e-01 5.64925611e-01 7.12448061e-01 -1.38085783e+00
-6.95380092e-01 -5.38271785e-01 5.31790197e-01 6.49775922e-01
-4.71844286e-01 -5.90537727e-01 -7.62228966e-01 -5.53897083e-01
9.99087002e-03 -5.26822388e-01 7.41769969e-02 -4.85734880e-01
-6.41395211e-01 -4.63256329e-01 8.82474363e-01 -1.01102793e+00
1.28757751e+00 -2.00798059e+00 4.48309839e-01 7.24273250e-02
3.69320810e-01 3.35914969e-01 -3.29879791e-01 4.82103974e-01
1.02347180e-01 -3.70868482e-02 6.28535450e-02 8.75720605e-02
8.98278579e-02 3.70920062e-01 -8.84290338e-02 -1.83924839e-01
3.38901937e-01 1.41029429e+00 -1.21086323e+00 -8.57793152e-01
4.32483584e-01 5.33136368e-01 -1.32217348e-01 1.20305613e-01
-2.89999664e-01 2.85152942e-01 -8.88027847e-01 9.00075793e-01
5.40868580e-01 -4.56817091e-01 7.04392046e-02 -8.75179827e-01
4.23313141e-01 -1.15874924e-01 -1.19258666e+00 1.71206093e+00
-3.27426791e-01 -6.68897778e-02 -3.71825933e-01 -1.11136127e+00
7.03216732e-01 3.24238330e-01 3.99263114e-01 -7.56748259e-01
2.98856258e-01 -7.79282376e-02 -5.38791828e-02 -9.14149463e-01
4.05570835e-01 -2.56320924e-01 -1.76656768e-02 2.31776744e-01
3.21451575e-01 1.35170281e-01 2.71704435e-01 5.13977170e-01
1.16460931e+00 2.12121189e-01 1.95776671e-01 2.40898043e-01
9.16606367e-01 -8.58165696e-02 5.98255396e-01 4.58980381e-01
-1.86362013e-01 2.18855981e-02 2.77980626e-01 -1.56233191e-01
-5.19843221e-01 -9.27622914e-01 3.18043679e-01 8.56301308e-01
6.52991235e-01 -4.94399518e-01 -3.77310008e-01 -8.20326746e-01
1.59017995e-01 7.39506543e-01 -8.53242636e-01 -7.07351148e-01
-1.17443904e-01 -7.13252127e-01 4.98983800e-01 7.90627122e-01
1.06013012e+00 -1.39342904e+00 -1.12715542e-01 2.20623121e-01
-7.84570694e-01 -1.25892973e+00 -1.57543570e-01 -1.49818227e-01
-4.90056515e-01 -1.44724226e+00 -2.53739744e-01 -6.41265869e-01
6.96255207e-01 3.84335160e-01 9.95081961e-01 4.05305743e-01
-1.13751076e-01 7.02352107e-01 -8.20519149e-01 -1.15645982e-01
-8.56514126e-02 -1.06709100e-01 -2.35393852e-01 1.26803070e-01
6.78862870e-01 -4.45517868e-01 -3.97256166e-01 1.15624428e-01
-1.08188307e+00 2.81011999e-01 1.10615563e+00 1.12152028e+00
5.78232944e-01 7.85411239e-01 9.37683105e-01 -7.39192426e-01
8.15334618e-01 -6.63801908e-01 -2.46505812e-03 7.00032890e-01
-4.91180688e-01 1.29702121e-01 9.07642961e-01 -3.33776236e-01
-1.44763958e+00 -3.63464206e-01 -1.16892993e-01 -5.99294543e-01
7.45828077e-02 9.87224519e-01 -5.12340367e-01 -7.75299743e-02
2.25224048e-01 5.13188899e-01 -7.48069435e-02 -1.29675895e-01
6.38880670e-01 5.28806448e-01 3.63891363e-01 -7.32946515e-01
8.62117529e-01 2.20658585e-01 -5.04829660e-02 -5.23508012e-01
-1.16219580e+00 -1.40875533e-01 -5.14519334e-01 -3.02945703e-01
9.04030859e-01 -1.06135070e+00 -9.39942777e-01 2.91664660e-01
-9.41840291e-01 1.74133331e-02 -1.49412453e-01 5.87839842e-01
-4.36904639e-01 5.81237793e-01 -8.31288993e-01 -8.11174214e-01
-3.34345043e-01 -1.04659057e+00 9.09679890e-01 4.73245889e-01
2.53041655e-01 -9.78768766e-01 -5.22688389e-01 1.08408308e+00
1.65864274e-01 1.64640978e-01 1.29585612e+00 -6.25527442e-01
-8.03904593e-01 4.35301736e-02 -5.16922832e-01 2.86469012e-01
2.15678304e-01 -3.04038852e-01 -8.43938887e-01 -6.52758591e-03
-3.42795104e-01 -7.18834102e-01 8.87689233e-01 4.30830605e-02
1.12928891e+00 -1.94639564e-01 -2.75606096e-01 1.15324341e-01
1.42611182e+00 2.51641929e-01 6.96508348e-01 3.16451460e-01
1.08880615e+00 4.91382658e-01 8.12679887e-01 3.16086501e-01
1.06951225e+00 2.67842948e-01 6.15161836e-01 1.10767983e-01
-1.41425937e-01 -5.37066162e-01 3.23126093e-02 1.16594422e+00
-3.48012656e-01 -5.16089499e-02 -6.20521963e-01 6.70520842e-01
-2.27493620e+00 -1.06704855e+00 2.54168790e-02 1.51999187e+00
7.25238740e-01 -5.23424055e-03 -2.02755436e-01 2.03791991e-01
6.70863628e-01 6.65948763e-02 -8.28135490e-01 2.54525423e-01
-3.73787701e-01 -3.52950841e-01 -1.15783498e-01 2.64024198e-01
-8.51851881e-01 9.49284494e-01 5.00541067e+00 1.10099542e+00
-3.54512334e-01 5.52760474e-02 1.13208860e-01 4.39140290e-01
-7.49940336e-01 -1.00712396e-01 -6.48598969e-01 4.55192208e-01
1.93012834e-01 8.51318836e-02 7.98564672e-01 4.44743901e-01
-4.75361735e-01 -5.51481582e-02 -6.76932693e-01 1.13325500e+00
4.75594342e-01 -1.04440093e+00 4.83660817e-01 -1.33916005e-01
5.64695895e-01 -4.22015131e-01 -2.42195442e-01 9.49698448e-01
4.19426739e-01 -7.65089035e-01 5.07791340e-01 9.34326768e-01
5.97917438e-01 -9.10647690e-01 1.00225878e+00 5.41941345e-01
-1.50018895e+00 -3.94167185e-01 -2.44515419e-01 1.04180612e-01
1.12404570e-01 6.09833479e-01 -2.35020235e-01 1.45689702e+00
6.39465213e-01 8.30292583e-01 -7.02475488e-01 3.99921209e-01
-5.95847607e-01 1.48721799e-01 -1.18517941e-02 3.72507162e-02
2.33963117e-01 -2.10291818e-01 2.03441069e-01 6.39259815e-01
2.86717951e-01 5.14779150e-01 2.66680717e-01 1.01271677e+00
-8.21408927e-02 6.91853240e-02 -3.66527438e-01 -2.20406443e-01
4.85144794e-01 1.28011286e+00 -5.31863630e-01 -4.52320755e-01
-9.15462852e-01 8.35806131e-01 7.22496629e-01 5.39374888e-01
-8.43279421e-01 -5.02921820e-01 1.88777849e-01 -5.55745006e-01
4.47855204e-01 2.96221584e-01 5.28213903e-02 -1.39511931e+00
1.86032832e-01 -6.89809680e-01 8.44829142e-01 -1.24638009e+00
-1.81137145e+00 5.89178562e-01 -2.68368423e-02 -9.21032548e-01
-7.16574416e-02 -4.38037843e-01 -1.94192946e-01 8.39004159e-01
-1.84966123e+00 -1.73319352e+00 -4.95503038e-01 1.15382862e+00
2.61687636e-01 -3.58937591e-01 8.03948581e-01 5.41287363e-01
-5.83933473e-01 4.97449577e-01 -3.51940066e-01 3.97288650e-01
3.02777231e-01 -1.08971560e+00 -4.35714662e-01 5.52551925e-01
-8.56033266e-02 4.81855661e-01 1.99399546e-01 -8.37148249e-01
-1.72964287e+00 -9.93761241e-01 7.14168429e-01 -3.27493876e-01
6.79952681e-01 2.61024952e-01 -1.03096008e+00 9.11736369e-01
-1.78609695e-02 -6.66659772e-02 7.46265888e-01 2.32836679e-01
-3.70401829e-01 -7.87108690e-02 -1.11870682e+00 6.72639191e-01
1.05372083e+00 -5.85070312e-01 -1.09658623e+00 1.20368406e-01
8.87335539e-01 -3.94147485e-01 -1.16075683e+00 5.45204520e-01
3.36921990e-01 -8.72506917e-01 1.06680763e+00 -5.95609426e-01
7.83234000e-01 -5.44211686e-01 -2.78617978e-01 -1.29788852e+00
-6.22585952e-01 5.67758828e-02 -6.13115907e-01 1.47442877e+00
2.68508673e-01 -6.73106253e-01 3.32548976e-01 4.22987491e-01
-1.45695716e-01 -9.53127801e-01 -7.82611191e-01 -5.10028183e-01
-3.32446933e-01 -2.92167872e-01 8.62657726e-01 1.22543180e+00
2.64159113e-01 7.88080513e-01 -4.84580070e-01 2.73047984e-01
4.01131064e-01 4.98673558e-01 3.73636961e-01 -1.18009651e+00
-9.17686746e-02 -2.71020919e-01 -3.60246778e-01 -8.26341271e-01
4.87432510e-01 -8.78743887e-01 -1.69032723e-01 -2.14192915e+00
4.60446298e-01 -1.54704109e-01 -5.03140688e-01 9.00667667e-01
-7.48320580e-01 -1.70697913e-01 2.12621063e-01 -1.05145469e-01
-9.07241344e-01 1.02259588e+00 1.80513096e+00 -2.63196260e-01
1.22489505e-01 -5.22389233e-01 -1.16159916e+00 6.46952391e-01
4.65785414e-01 4.99174781e-02 -8.68960083e-01 -2.65030444e-01
5.14610291e-01 4.31994557e-01 5.96802115e-01 -7.70565152e-01
3.16749722e-01 -2.05017850e-01 6.08623147e-01 -8.39883566e-01
6.25079036e-01 -1.12976372e+00 -1.06000327e-01 1.44820437e-01
-1.03954054e-01 -9.06668529e-02 1.57841697e-01 8.65346789e-01
-5.56359172e-01 -3.49818096e-02 1.81411013e-01 -2.66837507e-01
-1.30882096e+00 3.15394551e-01 5.50987348e-02 2.09697932e-01
9.35636342e-01 -7.52425417e-02 -6.55356646e-01 -3.14143717e-01
-7.32819676e-01 7.00418293e-01 -8.33309144e-02 6.56136811e-01
7.82944441e-01 -1.89616096e+00 -5.83250165e-01 4.11764495e-02
4.14794147e-01 1.45640699e-02 9.18132126e-01 7.98175693e-01
-2.42460575e-02 3.17212999e-01 -2.51344264e-01 -8.80980417e-02
-8.77431154e-01 1.04093671e+00 1.11567684e-01 -4.89661902e-01
-6.18759394e-01 5.49995422e-01 -6.15534633e-02 -6.15866244e-01
-8.06540772e-02 -4.29034233e-02 -6.31190538e-01 2.38807201e-01
4.52450007e-01 3.36422026e-01 -2.27693301e-02 -8.59410942e-01
-4.43622828e-01 5.75507522e-01 -7.03633923e-05 2.38770291e-01
1.14546752e+00 -3.13720852e-01 -1.00274615e-01 4.63658839e-01
8.56889963e-01 -3.43182623e-01 -8.91072750e-01 -6.65445685e-01
-5.80526531e-01 -3.42387795e-01 1.22525357e-01 -1.03827798e+00
-1.34217906e+00 6.54394567e-01 6.89068437e-02 1.58291504e-01
1.37553382e+00 8.31731409e-02 9.84544933e-01 4.59118664e-01
4.63083565e-01 -1.31417835e+00 3.07642192e-01 4.57522839e-01
8.79737973e-01 -1.39497018e+00 3.64541970e-02 -6.25423908e-01
-8.12488735e-01 1.07375991e+00 1.03043318e+00 3.22904885e-01
7.65776217e-01 -2.43435264e-01 -6.09714314e-02 -4.97922659e-01
-6.52192533e-01 -6.18805289e-01 5.26246130e-01 6.00393713e-01
-1.76020451e-02 -2.06060056e-02 -2.94387877e-01 1.14059865e+00
3.92562430e-03 1.62977785e-01 3.54441511e-03 8.86256874e-01
-2.42781237e-01 -6.50390983e-01 -3.88416171e-01 6.91007912e-01
1.80134717e-02 -1.48193225e-01 -3.15670550e-01 9.08983231e-01
5.05991042e-01 1.29341400e+00 -4.75630373e-01 -7.16280639e-01
4.79690433e-01 5.49060069e-02 4.40602064e-01 -2.30430737e-01
-2.82505333e-01 -3.60127866e-01 -4.49146889e-03 -4.64780867e-01
-7.88613737e-01 -1.80167317e-01 -1.35254622e+00 -3.18999559e-01
-5.57292521e-01 -2.43020500e-03 7.55052492e-02 1.49444854e+00
2.62032211e-01 1.15702319e+00 1.54759943e-01 -4.19899493e-01
-3.31952870e-01 -1.01333725e+00 -7.99263358e-01 8.14156592e-01
-5.25214151e-03 -1.15287018e+00 -2.12396652e-01 -1.41497672e-01]
|
[8.934564590454102, 7.759623050689697]
|
591a9cc4-0eda-47ca-87c8-4880a5766d02
|
atlas-based-automated-detection-of-swim
|
1902.06130
| null |
http://arxiv.org/abs/1902.06130v1
|
http://arxiv.org/pdf/1902.06130v1.pdf
|
Atlas-based automated detection of swim bladder in Medaka embryo
|
Fish embryo models are increasingly being used both for the assessment of
chemicals efficacy and potential toxicity. This article proposes a methodology
to automatically detect the swim bladder on 2D images of Medaka fish embryos
seen either in dorsal view or in lateral view. After embryo segmentation and
for each studied orientation, the method builds an atlas of a healthy embryo.
This atlas is then used to define the region of interest and to guide the swim
bladder segmentation with a discrete globally optimal active contour.
Descriptors are subsequently designed from this segmentation. An automated
random forest clas-sifier is built from these descriptors in order to classify
embryos with and without a swim bladder. The proposed method is assessed on a
dataset of 261 images, containing 202 embryos with a swim bladder (where 196
are in dorsal view and 6 are in lateral view) and 59 without (where 43 are in
dorsal view and 16 are in lateral view). We obtain an average precision rate of
95% in the total dataset following 5-fold cross-validation.
|
['Noémie De Crozé', 'Marc Léonard', 'Jean Cousty', 'Hugues Talbot', 'Diane Genest']
|
2019-02-16
| null | null | null | null |
['bladder-segmentation']
|
['medical']
|
[ 2.31793016e-01 2.28204682e-01 3.07006389e-01 -2.26196691e-01
-2.33252257e-01 -1.03026450e+00 4.89331871e-01 6.29443467e-01
-9.65897977e-01 2.90312350e-01 -1.00643754e-01 6.51213527e-02
-1.02402776e-01 -8.42898309e-01 -4.27819490e-01 -9.17792559e-01
-1.04460530e-01 5.71189344e-01 6.02667749e-01 2.72152334e-01
4.01202142e-01 9.55419421e-01 -1.21152842e+00 -1.80130735e-01
4.69751775e-01 5.25793433e-01 3.93552780e-01 8.93772602e-01
3.66679020e-03 1.43137395e-01 -3.84494692e-01 -3.14672649e-01
3.67931277e-01 -6.80307269e-01 -5.74977279e-01 8.53729174e-02
1.81164637e-01 -1.87451527e-01 3.79820079e-01 7.60847747e-01
6.37825489e-01 -2.17757989e-02 1.14922404e+00 -5.79309285e-01
1.31545082e-01 3.88720870e-01 -7.69868195e-01 3.02182585e-01
2.94477791e-01 7.10319204e-04 4.63187963e-01 -7.96875715e-01
8.36836100e-01 7.42054582e-01 6.47094011e-01 5.18621206e-01
-1.10227025e+00 -7.08556235e-01 -3.12626481e-01 -4.47648197e-01
-1.24031329e+00 -2.45197579e-01 2.49407560e-01 -8.77409518e-01
2.72633225e-01 7.97114000e-02 1.00374341e+00 -1.29589308e-02
5.21651983e-01 3.79703701e-01 1.24308181e+00 -4.67903614e-01
5.35310149e-01 -2.86980689e-01 -2.29357053e-02 6.63335741e-01
3.32976520e-01 1.80935189e-01 -3.96402806e-01 -1.16445847e-01
8.34385693e-01 -2.25114480e-01 -2.42987424e-02 -3.05404156e-01
-5.81619799e-01 5.53677738e-01 -1.00780521e-02 4.24733043e-01
-2.61131704e-01 -1.58956736e-01 2.20204666e-01 -1.57166407e-01
1.73720911e-01 3.78000110e-01 -2.39429682e-01 2.56967276e-01
-1.24970663e+00 2.81465143e-01 5.40345311e-01 4.83077228e-01
6.68008566e-01 -1.23618975e-01 3.18021476e-01 6.50422454e-01
6.55683875e-01 5.96729934e-01 4.81392831e-01 -6.30113542e-01
-1.07546322e-01 9.59930718e-01 -1.06220037e-01 -6.33519292e-01
-6.58501804e-01 1.13377042e-01 -3.60647082e-01 6.88828349e-01
6.57334507e-01 -4.04968768e-01 -1.23003542e+00 1.08223557e+00
6.85270667e-01 -8.76832679e-02 2.70995319e-01 5.97908854e-01
1.09097528e+00 6.98503673e-01 1.54370621e-01 -3.24488312e-01
1.48502314e+00 -8.45214650e-02 -4.48454559e-01 5.12781739e-02
5.14061868e-01 -6.10484838e-01 2.03493699e-01 3.30555499e-01
-1.15847921e+00 -1.15412153e-01 -1.16019285e+00 3.31493706e-01
-4.77825582e-01 2.42928401e-01 -7.47333542e-02 5.90443254e-01
-8.51291835e-01 6.86407387e-01 -1.11204958e+00 -4.57886726e-01
1.60194024e-01 5.14421999e-01 -8.94153416e-01 3.54361385e-01
-6.99391782e-01 1.02644789e+00 4.19200212e-01 2.78087795e-01
-1.20625114e+00 -4.44992512e-01 -1.14534092e+00 -2.22102553e-01
-2.03813776e-01 3.84326242e-02 9.40502942e-01 -4.73356724e-01
-1.56835878e+00 1.34742725e+00 2.19404876e-01 -4.31210548e-01
6.93599224e-01 7.51833171e-02 4.13053110e-02 5.14114439e-01
2.09621564e-01 5.01768887e-01 3.37697744e-01 -1.25354564e+00
-8.12861025e-01 -7.54281282e-01 -9.89592671e-02 3.79645973e-01
-1.14730224e-02 3.04767966e-01 -6.59180224e-01 -3.99663866e-01
7.17063487e-01 -9.56947744e-01 -2.88790226e-01 2.21712396e-01
-1.66469634e-01 6.15184084e-02 7.70753443e-01 -8.26125443e-01
9.41977799e-01 -1.81415427e+00 3.94425076e-03 4.40422267e-01
5.25574982e-02 2.99594820e-01 1.47629008e-01 4.10917103e-01
5.00223972e-02 -1.03184044e-01 -4.80233282e-01 -1.58779591e-01
-4.32140917e-01 7.33827800e-02 3.82950604e-01 9.07727659e-01
-5.59515134e-02 1.78435206e-01 -8.83095443e-01 -8.26500356e-01
1.38131455e-01 4.56619620e-01 -1.56976061e-03 3.86363745e-01
3.11600000e-01 4.38682973e-01 -4.70030367e-01 6.31976604e-01
8.88259947e-01 5.29250383e-01 3.25041056e-01 1.13707997e-01
-6.97978854e-01 -5.86837471e-01 -1.11968923e+00 1.28800273e+00
-2.92606205e-01 5.64627171e-01 3.64393383e-01 -5.75615466e-01
1.13321662e+00 6.33400738e-01 4.46789920e-01 -3.89013857e-01
2.59311080e-01 1.43237129e-01 -1.23883709e-01 -5.63362241e-01
6.43838644e-02 -5.93854666e-01 2.00493783e-01 2.21171141e-01
1.99616075e-01 -3.54472041e-01 7.29685009e-01 -4.43195440e-02
6.93854570e-01 4.53689456e-01 2.88630724e-01 -8.10932517e-01
5.66225767e-01 -1.69443920e-01 6.51546896e-01 1.00009404e-01
-3.70396152e-02 7.94262707e-01 7.67660439e-01 -6.43142164e-01
-6.65304005e-01 -8.45130384e-01 -6.17166579e-01 7.17924118e-01
3.62619013e-01 1.11030914e-01 -1.14220083e+00 -5.61935127e-01
-1.41745657e-01 1.27630189e-01 -1.11616540e+00 3.16678822e-01
-2.71960795e-01 -7.03791678e-01 7.25675285e-01 3.11457455e-01
2.88292527e-01 -9.62896943e-01 -1.31384635e+00 1.64947063e-01
4.76201147e-01 -4.59169239e-01 1.11687563e-01 2.61014640e-01
-9.23254907e-01 -1.27572942e+00 -1.24818313e+00 -7.21479058e-01
1.06358635e+00 -3.32213700e-01 7.30645955e-01 9.25193056e-02
-3.52771193e-01 -1.04301572e-01 -3.03989977e-01 -5.20208001e-01
-6.04064286e-01 -5.91947079e-01 -9.15677845e-02 -1.03559390e-01
-4.62187901e-02 -1.89090341e-01 -5.60712755e-01 6.70908213e-01
-1.26797855e+00 -2.41985932e-01 4.46066409e-01 4.79296565e-01
6.96701884e-01 -2.09015995e-01 2.18637526e-01 -9.12703872e-01
-2.42658090e-02 -4.97843117e-01 -1.22267640e+00 2.25560218e-01
-2.30301604e-01 -1.81167036e-01 3.75573546e-01 -1.17676355e-01
-8.92413437e-01 5.65229356e-01 -2.61925399e-01 1.58377796e-01
-5.36634147e-01 5.25006413e-01 5.68412431e-02 -1.34332910e-01
7.47518957e-01 4.32702862e-02 8.33820030e-02 -3.65067333e-01
-3.33215237e-01 4.99276102e-01 3.09692413e-01 -1.18522346e-01
6.38943374e-01 6.59196615e-01 3.16436499e-01 -1.07420874e+00
-2.61762559e-01 -5.90889513e-01 -9.33641374e-01 -6.37302756e-01
1.04556549e+00 -6.41232848e-01 -5.19271433e-01 5.78503311e-01
-6.94116175e-01 -4.21789795e-01 8.58945996e-02 6.18428826e-01
-3.50931436e-01 5.17960966e-01 -6.32867694e-01 -9.31943119e-01
-5.41084230e-01 -1.24219692e+00 8.84428322e-01 7.53815055e-01
-2.09860429e-01 -9.93511796e-01 5.20701289e-01 4.50666919e-02
-5.31951785e-02 6.18617237e-01 5.00884593e-01 -6.49585128e-01
2.78159052e-01 -7.33481467e-01 2.76513577e-01 -6.57026097e-02
-2.04609986e-02 5.53777933e-01 -1.12641811e+00 -2.10918367e-01
-4.95926350e-01 -8.17641839e-02 5.94935834e-01 6.52413249e-01
2.50972927e-01 3.37363243e-01 -5.41515589e-01 3.86930496e-01
1.65450597e+00 9.39542115e-01 4.77323622e-01 1.51207358e-01
4.64751059e-03 9.89303529e-01 8.70838404e-01 4.10103738e-01
-1.51102304e-01 3.82932723e-01 5.11718929e-01 -2.67332464e-01
2.71180123e-01 -2.79389210e-02 3.62586498e-01 -1.20718174e-01
-5.97711325e-01 -2.66670555e-01 -1.18545449e+00 8.30616474e-01
-1.07124710e+00 -5.92378020e-01 -3.03996354e-01 2.30613399e+00
5.62522352e-01 -8.32868591e-02 3.83389235e-01 2.95626581e-01
9.39856470e-01 -4.70894933e-01 -1.16371557e-01 -5.23525119e-01
1.12630039e-01 1.46309599e-01 6.68455422e-01 6.62413538e-01
-1.14727855e+00 6.49976790e-01 6.32018757e+00 4.71960247e-01
-1.24735355e+00 -5.34433246e-01 5.13263106e-01 4.63711590e-01
2.13427559e-01 6.09408282e-02 -1.09123814e+00 4.65729862e-01
5.31738400e-01 -4.32303660e-02 -5.47500670e-01 4.86844152e-01
2.07811520e-01 -7.01462030e-01 -8.64586592e-01 2.75364459e-01
-1.05102375e-01 -1.14811504e+00 -2.37287402e-01 1.09717801e-01
6.03320599e-01 -2.15116456e-01 -4.06153291e-01 -3.67187947e-01
8.23149756e-02 -8.72475863e-01 7.74543047e-01 4.99038756e-01
1.18213689e+00 -6.65640116e-01 1.06934977e+00 5.18028557e-01
-1.31279683e+00 2.44581401e-01 -3.24751586e-01 8.55984539e-02
9.75285023e-02 1.56795591e-01 -1.10343087e+00 3.29033732e-01
6.25978172e-01 3.40018630e-01 -4.03376311e-01 1.46896970e+00
1.04885017e-02 5.26282907e-01 -6.69966102e-01 -1.08562283e-01
1.69308737e-01 -5.59336424e-01 5.12803555e-01 1.39871287e+00
4.74972486e-01 1.96452960e-01 -5.03053546e-01 4.51044202e-01
4.67957348e-01 4.34324116e-01 -7.59300649e-01 -3.67186069e-02
3.59588891e-01 1.41919637e+00 -1.83828127e+00 -1.79439291e-01
-9.01378691e-02 6.27742469e-01 -2.84752965e-01 -2.67555304e-02
-3.01128447e-01 -4.86448526e-01 1.87761575e-01 2.16900274e-01
1.01505853e-01 2.51860023e-01 3.93190719e-02 -4.10442293e-01
-6.32771432e-01 4.18659076e-02 6.15110099e-01 -4.10363853e-01
-5.84706128e-01 5.51208675e-01 3.87327045e-01 -1.43063045e+00
-5.65827489e-02 -6.81075692e-01 -8.98889244e-01 8.81936610e-01
-7.46467233e-01 -9.81561184e-01 -1.79963708e-01 -2.67777443e-01
3.27751517e-01 -4.93886471e-02 9.82941329e-01 6.88140839e-02
-3.92295569e-01 1.75686806e-01 2.13157758e-01 3.86559159e-01
2.75757492e-01 -1.45742691e+00 -4.32163738e-02 8.37618053e-01
-1.54342756e-01 5.11909783e-01 8.95762563e-01 -8.01362395e-01
-8.51786315e-01 -7.32210398e-01 8.06166053e-01 -2.99540401e-01
1.23562358e-01 -7.05911452e-03 -4.77679342e-01 3.47254366e-01
1.60225853e-02 6.67096600e-02 1.08746898e+00 -7.12476194e-01
4.78002578e-01 1.24113366e-01 -1.55291057e+00 3.13807756e-01
1.70452297e-01 2.67580748e-01 -3.76917720e-01 -2.97569390e-02
-5.12843788e-01 -7.99451590e-01 -1.03165770e+00 3.76833498e-01
9.39356685e-01 -1.04248011e+00 5.24956763e-01 -2.58002967e-01
5.15039206e-01 -7.51966774e-01 4.96650815e-01 -1.02822709e+00
8.84001926e-02 -3.25281978e-01 8.68510127e-01 1.13827586e+00
5.74850917e-01 -2.64709532e-01 1.08001196e+00 2.63201296e-01
-2.61577994e-01 -7.35604763e-01 -8.90384197e-01 -4.39189672e-01
3.29079598e-01 1.89823762e-01 -8.87134392e-03 6.82859242e-01
1.26945779e-01 -1.08667493e-01 2.42931977e-01 3.69481742e-01
6.70837998e-01 2.06485257e-01 4.71976459e-01 -1.21128750e+00
1.57593235e-01 -2.10446313e-01 -9.79343295e-01 -4.85239357e-01
-3.81466985e-01 -5.28859437e-01 3.38273317e-01 -1.69782197e+00
2.34490141e-01 -1.66655719e-01 1.10229492e-01 4.03406411e-01
1.69844940e-01 6.60986662e-01 -1.00462906e-01 -2.41094023e-01
-6.74140751e-02 -2.49445304e-01 1.01897418e+00 3.32001626e-01
-1.44665495e-01 1.30224660e-01 -1.63844898e-01 1.20623708e+00
4.51869726e-01 -5.60242772e-01 -1.04357712e-01 -8.03265050e-02
1.19962348e-02 4.60300028e-01 -1.39641672e-01 -9.85267162e-01
1.74355105e-01 -4.92704473e-02 6.67294025e-01 -7.22351670e-01
1.66954651e-01 -8.82994413e-01 5.47745645e-01 1.03360164e+00
-3.13414901e-01 -9.57975462e-02 2.84818858e-01 4.22502667e-01
-1.75626859e-01 -9.07729447e-01 1.32012105e+00 -3.13216686e-01
-4.31172103e-01 9.03826877e-02 -9.12696421e-01 -2.26237953e-01
1.55684757e+00 -6.86067045e-01 3.46633911e-01 1.12960497e-02
-1.10811830e+00 2.41540194e-01 8.55724752e-01 -3.95969301e-01
6.62930191e-01 -6.95463836e-01 -7.19609737e-01 1.45527303e-01
4.16562445e-02 4.22718793e-01 3.80052269e-01 7.62225449e-01
-1.63103330e+00 -5.72970212e-02 -4.63498175e-01 -5.63565373e-01
-1.76847053e+00 -7.57875368e-02 6.34492934e-01 -9.03990194e-02
-4.29947078e-01 1.15008438e+00 3.92760813e-01 -2.04804644e-01
-1.01470046e-01 -3.32611084e-01 -1.05523753e+00 4.90631402e-01
4.76319045e-01 5.39152026e-01 -2.69163754e-02 -1.04904807e+00
-3.24300349e-01 1.26517355e+00 4.55860585e-01 -3.02986652e-01
1.30165792e+00 1.85977846e-01 7.52586499e-03 3.67188722e-01
8.39111745e-01 4.07218575e-01 -1.47002244e+00 3.66788626e-01
3.80756594e-02 -3.73568803e-01 -2.20203772e-01 -5.62949240e-01
-1.04099345e+00 9.56664920e-01 7.54541099e-01 3.62931877e-01
9.94356036e-01 -8.55034515e-02 1.12913512e-01 -2.09595203e-01
1.35834470e-01 -9.16746080e-01 -3.52108210e-01 1.52792558e-01
7.03726053e-01 -8.81390095e-01 1.91231549e-01 -4.57469106e-01
-8.10749650e-01 1.45981598e+00 3.18107575e-01 -3.60013068e-01
5.74069023e-01 7.53262997e-01 4.96754527e-01 -4.06343609e-01
-2.87211090e-01 -1.37750879e-01 1.28929809e-01 6.17598772e-01
6.66192830e-01 -2.98745900e-01 -6.99263990e-01 3.37845922e-01
8.84054080e-02 -1.45506263e-01 6.36679530e-01 1.24809897e+00
-5.23196399e-01 -6.78608477e-01 -3.27242732e-01 2.55534679e-01
-9.87107635e-01 4.82956380e-01 -6.01382256e-01 7.81230748e-01
5.32101870e-01 5.79912007e-01 2.36522391e-01 1.85544919e-02
3.62049639e-01 -1.51761353e-01 3.62583190e-01 -8.24048162e-01
-7.05511391e-01 4.43116575e-01 -1.17474981e-02 3.97473574e-02
-4.99817073e-01 -7.64373779e-01 -1.80819726e+00 4.50572997e-01
-6.41552031e-01 6.41785562e-01 9.15180206e-01 8.41136098e-01
-5.01446068e-01 -6.84882179e-02 5.41952848e-01 -8.45485091e-01
1.96870893e-01 -9.71294761e-01 -1.03791201e+00 7.71190599e-03
1.13089614e-01 -6.84517980e-01 -5.64096987e-01 6.12538695e-01]
|
[14.403270721435547, -2.9927220344543457]
|
d41321b4-6879-4423-94e1-f7de117bf83e
|
data-efficient-end-to-end-information
|
2211.01692
| null |
https://arxiv.org/abs/2211.01692v1
|
https://arxiv.org/pdf/2211.01692v1.pdf
|
Data-efficient End-to-end Information Extraction for Statistical Legal Analysis
|
Legal practitioners often face a vast amount of documents. Lawyers, for instance, search for appropriate precedents favorable to their clients, while the number of legal precedents is ever-growing. Although legal search engines can assist finding individual target documents and narrowing down the number of candidates, retrieved information is often presented as unstructured text and users have to examine each document thoroughly which could lead to information overloading. This also makes their statistical analysis challenging. Here, we present an end-to-end information extraction (IE) system for legal documents. By formulating IE as a generation task, our system can be easily applied to various tasks without domain-specific engineering effort. The experimental results of four IE tasks on Korean precedents shows that our IE system can achieve competent scores (-2.3 on average) compared to the rule-based baseline with as few as 50 training examples per task and higher score (+5.4 on average) with 200 examples. Finally, our statistical analysis on two case categories--drunk driving and fraud--with 35k precedents reveals the resulting structured information from our IE system faithfully reflects the macroscopic features of Korean legal system.
|
['Minjoon Seo', 'Hai Jin Park', 'Hanuhl Lee', 'Saehee Eom', 'Wonseok Hwang']
|
2022-11-03
| null | null | null | null |
['instance-search']
|
['computer-vision']
|
[ 1.44187376e-01 3.97241652e-01 -5.24285316e-01 -4.48062479e-01
-1.41468990e+00 -6.09548569e-01 6.86666369e-01 2.12520868e-01
-6.64609909e-01 9.88254249e-01 3.28948319e-01 -9.05319810e-01
-3.41203749e-01 -6.13220453e-01 -5.65487981e-01 -1.42779961e-01
2.64907420e-01 7.36605227e-01 3.55982363e-01 -3.78639221e-01
5.33525586e-01 2.94418216e-01 -1.05785680e+00 5.73463202e-01
1.56334627e+00 6.19289875e-01 -3.81385386e-02 2.07516566e-01
-1.71847492e-01 1.01531160e+00 -7.45428681e-01 -1.22346365e+00
4.61218208e-01 2.32514814e-02 -8.39933753e-01 -3.55668962e-01
8.72841358e-01 -7.33624578e-01 -6.28956079e-01 9.81174529e-01
3.58497649e-01 -7.52831623e-02 8.15809727e-01 -1.00125909e+00
-8.35882127e-01 8.06702256e-01 -5.81335247e-01 4.98456925e-01
4.04602289e-01 5.17190620e-02 1.28261733e+00 -7.22697020e-01
1.00215685e+00 1.16985178e+00 4.11147356e-01 6.30944788e-01
-1.06523895e+00 -8.18411529e-01 2.30050355e-01 6.48976386e-01
-9.73861098e-01 -5.48816383e-01 5.72423756e-01 -7.10435033e-01
1.07722783e+00 -8.62138495e-02 2.52859443e-01 1.14964676e+00
4.36965674e-01 1.00340259e+00 6.68377042e-01 -2.97377080e-01
7.47959614e-02 1.14978418e-01 6.42055094e-01 2.84160376e-01
9.84967589e-01 -3.75590883e-02 -2.18605399e-01 -3.75944316e-01
2.51725703e-01 -8.47749040e-02 1.71618596e-01 3.90372425e-01
-7.27956533e-01 8.18431973e-01 -1.23343885e-01 2.06375480e-01
-4.43930268e-01 -1.07620217e-01 5.29818237e-01 3.44623744e-01
-1.08757932e-02 5.20045340e-01 -2.96857476e-01 -3.47911775e-01
-1.01868677e+00 1.02769637e+00 8.10674608e-01 1.10518312e+00
4.24067944e-01 -3.81628752e-01 -6.04813099e-01 9.79276478e-01
1.67660877e-01 5.65509140e-01 3.02715778e-01 -9.75916564e-01
1.37244713e+00 7.23777056e-01 2.06172958e-01 -8.33068073e-01
-1.00503460e-01 -1.55579448e-01 -4.99632895e-01 6.14424832e-02
7.63230920e-01 -2.97549337e-01 -1.10144985e+00 1.30814850e+00
-6.28799424e-02 -5.16196907e-01 8.90917107e-02 4.89044845e-01
5.98707736e-01 4.49998617e-01 3.67240041e-01 -1.70958281e-01
1.50884712e+00 -5.94183624e-01 -9.64770079e-01 -6.33212805e-01
4.53656495e-01 -7.98986495e-01 9.62242007e-01 5.33488989e-01
-1.15114200e+00 -8.51965696e-02 -8.37861180e-01 -4.67873544e-01
-2.37468228e-01 2.17410251e-01 6.06557727e-01 5.59408724e-01
-4.32638168e-01 4.97375220e-01 -3.53458196e-01 -1.34605780e-01
7.30749965e-01 -6.96063787e-02 -2.42853522e-01 -2.80678481e-01
-1.53290749e+00 1.13481736e+00 2.38869369e-01 -1.25412166e-01
-3.29175174e-01 -6.92213893e-01 -7.84895957e-01 5.57812862e-02
7.39137948e-01 -3.52788776e-01 1.51011872e+00 2.95134425e-01
-8.20645988e-01 6.76847875e-01 -4.09010887e-01 -5.98257244e-01
8.83345604e-01 -4.96208608e-01 -6.93770647e-01 1.51713148e-01
7.94647574e-01 1.61526650e-01 4.39496875e-01 -8.15545380e-01
-9.70691144e-01 -2.06922650e-01 4.14526984e-02 -9.12863240e-02
1.42088458e-02 4.07580405e-01 -6.16035581e-01 -5.83067119e-01
-5.04622385e-02 -7.19787478e-01 -3.81917536e-01 -4.34136868e-01
-6.25554144e-01 -5.75584590e-01 4.54164326e-01 -9.58353400e-01
1.79522312e+00 -1.77380192e+00 -6.23438179e-01 3.08070809e-01
1.96906269e-01 3.86037767e-01 1.84106499e-01 5.46214283e-01
1.02150239e-01 4.22762781e-01 -1.68513522e-01 1.02559581e-01
2.06212446e-01 2.31971070e-01 -7.20184505e-01 -3.18757147e-02
2.32474223e-01 9.50427592e-01 -9.26811695e-01 -9.05920208e-01
-2.05829576e-01 -1.62697375e-01 -6.23545349e-01 -2.77557522e-01
-1.74314380e-01 -1.39410779e-01 -7.83618271e-01 8.58018041e-01
5.65262854e-01 -1.89542398e-01 1.11057267e-01 1.65304899e-01
2.07301453e-01 8.07583213e-01 -8.80960047e-01 1.28801775e+00
-1.41880885e-01 6.14021242e-01 -2.50473559e-01 -7.61521339e-01
5.60389280e-01 2.97545463e-01 2.00954482e-01 -1.15507770e+00
-2.21766859e-01 3.50491285e-01 2.77217150e-01 -8.96542490e-01
5.87044954e-01 -8.48912075e-02 -3.98326784e-01 5.98988712e-01
-4.08828676e-01 7.10162744e-02 1.16338110e+00 6.91120803e-01
1.28698981e+00 -1.94798529e-01 4.27004844e-01 7.49944225e-02
2.90225893e-01 3.13979596e-01 8.28734219e-01 1.14507914e+00
-1.58208966e-01 2.76162326e-01 9.77946818e-01 -5.21469414e-01
-1.00114393e+00 -8.94612312e-01 -5.40796340e-01 6.74279451e-01
-1.64648667e-01 -5.71637094e-01 -5.60041368e-01 -1.04378903e+00
3.92296702e-01 1.01977181e+00 -4.85211164e-01 3.06792736e-01
-7.86240160e-01 -5.95641792e-01 5.57063818e-01 5.45658529e-01
3.03473502e-01 -1.19633353e+00 -5.28915763e-01 4.67389464e-01
-3.19764018e-01 -1.37973762e+00 -4.94783729e-01 -3.44925374e-01
-6.67352557e-01 -1.32169616e+00 -4.42432314e-01 -3.32534730e-01
6.24120355e-01 -1.74276575e-01 8.93695533e-01 1.22241758e-01
-3.30478340e-01 -2.54891366e-01 -1.38032675e-01 -6.86080575e-01
-5.92117310e-01 1.24779195e-01 -1.47849143e-01 -3.89569342e-01
8.82496297e-01 -3.86438757e-01 -4.79369789e-01 1.06824920e-01
-8.43362391e-01 -1.08097173e-01 9.66788352e-01 6.97017729e-01
9.13745612e-02 -1.78911522e-01 1.00103164e+00 -1.53516698e+00
1.37729454e+00 -4.08822775e-01 -7.15983510e-01 5.12432873e-01
-1.19559884e+00 2.05008268e-01 5.90307713e-01 -1.71299294e-01
-1.27218592e+00 -2.65495837e-01 -5.21210954e-02 2.11695865e-01
-1.61735594e-01 7.63870418e-01 8.45841542e-02 7.67004967e-01
1.03509068e+00 1.02470510e-01 -1.81321353e-01 -4.25477564e-01
3.43562096e-01 9.37028706e-01 6.88685238e-01 -8.15903187e-01
8.71298730e-01 1.99676350e-01 -2.92050630e-01 -2.73108065e-01
-1.01098979e+00 -5.21504939e-01 -4.97051656e-01 -2.54490413e-02
5.06294072e-01 -6.66902244e-01 -6.24584913e-01 -9.17755626e-03
-1.33059394e+00 -1.07453413e-01 -1.72299370e-02 3.33458424e-01
-4.10728790e-02 6.11990571e-01 -8.28752518e-01 -8.80687594e-01
-5.60032845e-01 -1.14613593e+00 8.51405561e-01 6.76549925e-03
-5.65546989e-01 -5.77064455e-01 -9.02089924e-02 9.45101738e-01
1.43283963e-01 5.83864525e-02 1.37022793e+00 -1.00492644e+00
-4.85532552e-01 -7.30521441e-01 -3.74934465e-01 1.57387942e-01
-3.63965072e-02 -1.62987020e-02 -6.41278207e-01 2.61692792e-01
-4.62987006e-01 -2.12418959e-01 9.33003485e-01 1.56507909e-01
7.61466861e-01 -6.33882701e-01 -7.02072203e-01 -1.75938457e-01
9.57443416e-01 7.26827383e-01 7.93911338e-01 4.96556669e-01
3.39080542e-01 6.45263672e-01 7.95376658e-01 3.03830743e-01
4.71290290e-01 5.93942404e-01 -3.07037681e-01 2.72891939e-01
6.25978336e-02 -4.29172963e-01 1.92405984e-01 1.36035591e-01
-1.90967605e-01 -1.22762047e-01 -1.13618588e+00 8.08152318e-01
-2.06582999e+00 -1.42587423e+00 -1.28829807e-01 1.97208762e+00
1.19396842e+00 5.26763201e-01 8.39792565e-02 1.56739309e-01
5.29653668e-01 -2.48559847e-01 -6.97640300e-01 -3.89931202e-01
2.49341592e-01 2.59087533e-02 4.58400935e-01 4.72203135e-01
-9.15815532e-01 1.00937414e+00 6.88682222e+00 9.87545609e-01
-6.98225319e-01 -2.33504474e-01 6.81652963e-01 -2.33769953e-01
-4.23106581e-01 7.95323327e-02 -1.26667380e+00 7.86289096e-01
7.81744063e-01 -8.04219961e-01 6.73659593e-02 1.09442723e+00
2.86733389e-01 -1.78429544e-01 -1.22176909e+00 8.36214662e-01
-2.17094511e-01 -1.47261786e+00 3.16864431e-01 4.66985911e-01
6.49525702e-01 -8.34256113e-02 -2.05328673e-01 5.68224311e-01
6.02718294e-01 -1.00914335e+00 8.15174758e-01 3.21829826e-01
7.00624645e-01 -4.11098838e-01 7.45945275e-01 5.06503105e-01
-6.14500165e-01 -4.13727313e-01 -4.66332674e-01 1.08283050e-02
6.05774820e-01 7.76026130e-01 -1.19325709e+00 4.92207408e-01
5.12682199e-01 4.55796629e-01 -4.67609644e-01 1.04212260e+00
-5.64365923e-01 6.55236185e-01 -2.80015379e-01 -1.66619062e-01
2.14778438e-01 -1.63116291e-01 4.31976020e-01 1.32436216e+00
2.71872222e-01 4.33376193e-01 -6.99775293e-02 7.68354654e-01
-2.46907890e-01 2.82029122e-01 -7.50073373e-01 5.07882722e-02
7.31174231e-01 1.00363529e+00 -3.95700485e-01 -6.32214844e-01
-4.23738599e-01 3.60978723e-01 4.40512151e-01 5.10824382e-01
-6.33304834e-01 -6.86016202e-01 3.67257088e-01 5.88106334e-01
1.39798135e-01 5.98623864e-02 -4.53392863e-01 -1.01570797e+00
4.93429244e-01 -9.75954771e-01 7.05044985e-01 -4.57762957e-01
-1.42978060e+00 5.69614828e-01 3.01548541e-01 -1.23925066e+00
-7.34219313e-01 -5.07857502e-01 -6.35606408e-01 7.51931489e-01
-1.54476810e+00 -6.33937895e-01 2.15940580e-01 2.55028099e-01
5.79832673e-01 -3.88718754e-01 2.88549095e-01 6.90226614e-01
-5.52943647e-01 8.36914659e-01 -2.45882258e-01 4.71796155e-01
9.55267727e-01 -1.22403610e+00 6.57034636e-01 1.00155675e+00
-2.42957491e-02 1.13302183e+00 5.10815859e-01 -1.06428695e+00
-9.01781917e-01 -7.69851685e-01 1.63879991e+00 -7.68739879e-01
7.73079038e-01 -1.45205840e-01 -9.35651064e-01 6.27811670e-01
1.70515358e-01 -5.39198518e-01 6.17758632e-01 3.48911613e-01
-4.53037977e-01 -2.29472294e-01 -1.10868382e+00 8.97982657e-01
1.19974446e+00 -3.83008599e-01 -1.10908473e+00 6.25567436e-01
2.06304461e-01 -4.84042138e-01 -5.93988657e-01 2.12816149e-01
6.82233453e-01 -7.36688435e-01 8.32318604e-01 -8.56640458e-01
7.74137974e-01 -7.16248974e-02 3.56225222e-01 -6.86827779e-01
-3.43148410e-01 -7.61418462e-01 5.83089516e-02 9.70386386e-01
1.19817543e+00 -5.47323644e-01 6.41800880e-01 1.53524184e+00
-5.02855442e-02 -9.46534276e-01 -1.01924026e+00 -8.12400103e-01
2.62554288e-01 -7.30185032e-01 5.29009163e-01 8.20960701e-01
5.90705812e-01 5.93630791e-01 -2.22861230e-01 -1.67561233e-01
6.13619328e-01 1.63507134e-01 7.06200480e-01 -1.31730759e+00
-7.38101453e-02 -6.33222163e-01 3.34030725e-02 -9.87743497e-01
1.72385499e-01 -1.00333738e+00 -2.18371660e-01 -1.96307564e+00
3.81481618e-01 -4.06697065e-01 -5.35395145e-02 9.61225152e-01
-3.56119663e-01 -3.47235054e-01 4.77202386e-02 3.59703481e-01
-4.41243500e-01 -8.46211687e-02 1.21053052e+00 -2.70206034e-01
-2.07207754e-01 5.67174517e-02 -1.32210279e+00 6.13663375e-01
4.75582272e-01 -5.23420393e-01 -4.46959734e-01 -2.81421542e-01
5.27902544e-01 8.16994309e-02 1.40499789e-02 -4.94814783e-01
6.14475191e-01 -4.31319088e-01 1.13976084e-01 -7.29554772e-01
-5.65376244e-02 -5.16802907e-01 -2.77997404e-01 1.73977137e-01
-3.93329799e-01 -1.21163435e-01 2.08536595e-01 4.91839498e-01
-3.13450962e-01 -6.54565215e-01 2.26856649e-01 -6.37774467e-02
-5.08663297e-01 1.54960796e-01 -5.25092959e-01 3.59348029e-01
8.12827110e-01 -3.32216114e-01 -8.26539993e-01 -4.21814114e-01
-5.45749426e-01 6.09990001e-01 -4.52065505e-02 2.64093220e-01
5.45507491e-01 -1.07413793e+00 -1.05449855e+00 -1.31733656e-01
1.56619534e-01 9.38545093e-02 1.31587356e-01 5.65704107e-01
-4.49946344e-01 8.38120461e-01 2.84637839e-01 7.95740262e-03
-1.10087299e+00 3.05757612e-01 -1.85388744e-01 -8.00608337e-01
-8.61832142e-01 4.08202410e-01 -1.17400654e-01 -1.82585430e-03
1.36236221e-01 -4.95295703e-01 -4.24533576e-01 3.23014170e-01
7.56478012e-01 4.21856910e-01 1.65469348e-01 -1.43098071e-01
-3.34888518e-01 3.87946606e-01 -9.85892296e-01 -2.71170259e-01
1.40405989e+00 5.10378420e-01 6.30794093e-02 -3.81339341e-01
6.93825781e-01 3.89678061e-01 -8.56650949e-01 -2.57703274e-01
5.99656403e-01 -5.81914961e-01 -3.54615927e-01 -1.12508571e+00
-6.94385648e-01 5.94298661e-01 -2.15151310e-01 2.76709944e-01
6.69844747e-01 1.57179102e-01 9.65146720e-01 9.79153812e-01
2.70830005e-01 -1.45187068e+00 -3.12796801e-01 5.11307418e-01
1.13414335e+00 -1.15269387e+00 2.73290247e-01 -5.15162170e-01
-9.94780779e-01 9.81186450e-01 4.94545460e-01 1.49420813e-01
3.64827722e-01 3.02496403e-01 2.63027728e-01 -3.55185390e-01
-9.09526467e-01 1.32613018e-01 4.86662656e-01 2.26834461e-01
3.45934838e-01 -2.61785716e-01 -8.37548852e-01 8.89839470e-01
-4.67179000e-01 3.90620008e-02 6.83046103e-01 8.39223504e-01
-3.56206477e-01 -1.47576976e+00 -1.98959559e-01 1.00459790e+00
-9.53835905e-01 -4.46819872e-01 -3.57443124e-01 9.21547055e-01
-1.37902975e-01 1.09610152e+00 -2.21996680e-01 4.76553962e-02
7.15599716e-01 2.85882264e-01 1.71325073e-01 -7.26985395e-01
-4.35997486e-01 4.92768735e-02 6.18948221e-01 -3.73127669e-01
6.22677989e-02 -9.55121279e-01 -1.37963498e+00 -1.11468270e-01
-1.53453067e-01 3.22496265e-01 2.37758726e-01 1.06834948e+00
3.92175913e-01 1.16313346e-01 -5.70176058e-02 -4.68624011e-03
-8.02431047e-01 -1.16264772e+00 -4.27358866e-01 6.73673987e-01
4.22751345e-02 -4.83583927e-01 -1.10508867e-01 2.29531020e-01]
|
[9.971440315246582, 9.19672966003418]
|
36cd6ec9-66d9-4da8-b06b-370f9905dc1c
|
versatile-audio-visual-learning-for-handling
|
2305.07216
| null |
https://arxiv.org/abs/2305.07216v1
|
https://arxiv.org/pdf/2305.07216v1.pdf
|
Versatile Audio-Visual Learning for Handling Single and Multi Modalities in Emotion Regression and Classification Tasks
|
Most current audio-visual emotion recognition models lack the flexibility needed for deployment in practical applications. We envision a multimodal system that works even when only one modality is available and can be implemented interchangeably for either predicting emotional attributes or recognizing categorical emotions. Achieving such flexibility in a multimodal emotion recognition system is difficult due to the inherent challenges in accurately interpreting and integrating varied data sources. It is also a challenge to robustly handle missing or partial information while allowing direct switch between regression and classification tasks. This study proposes a \emph{versatile audio-visual learning} (VAVL) framework for handling unimodal and multimodal systems for emotion regression and emotion classification tasks. We implement an audio-visual framework that can be trained even when audio and visual paired data is not available for part of the training set (i.e., audio only or only video is present). We achieve this effective representation learning with audio-visual shared layers, residual connections over shared layers, and a unimodal reconstruction task. Our experimental results reveal that our architecture significantly outperforms strong baselines on both the CREMA-D and MSP-IMPROV corpora. Notably, VAVL attains a new state-of-the-art performance in the emotional attribute prediction task on the MSP-IMPROV corpus. Code available at: https://github.com/ilucasgoncalves/VAVL
|
['Carlos Busso', 'Berrak Sisman', 'Wei-Cheng Lin', 'Seong-Gyun Leem', 'Lucas Goncalves']
|
2023-05-12
| null | null | null | null |
['video-emotion-recognition', 'multimodal-emotion-recognition', 'emotion-classification', 'emotion-classification', 'speech-emotion-recognition', 'multimodal-emotion-recognition']
|
['computer-vision', 'computer-vision', 'computer-vision', 'natural-language-processing', 'speech', 'speech']
|
[-4.27590460e-02 -3.63227338e-01 -3.40743037e-03 -6.58418834e-01
-1.09327865e+00 -6.79227293e-01 3.78258467e-01 -3.95648405e-02
-3.87743950e-01 3.41647655e-01 9.33185667e-02 -3.11913947e-03
1.41772479e-01 -1.90019816e-01 -5.69898546e-01 -5.61759770e-01
3.97205465e-02 3.98670942e-01 -5.31624675e-01 -2.82567650e-01
-3.10290873e-01 3.86531085e-01 -2.10443521e+00 9.00680065e-01
4.00668263e-01 1.61392450e+00 1.23321302e-02 1.11165893e+00
-1.27497137e-01 9.36863661e-01 -3.51841271e-01 -4.41408396e-01
-3.69923115e-02 -2.87727952e-01 -6.03929043e-01 -1.02481633e-01
3.55473191e-01 -1.85450926e-01 -2.22274005e-01 4.84928608e-01
8.43084931e-01 1.86792761e-01 7.61584699e-01 -1.82511163e+00
-4.50894803e-01 3.31668019e-01 -3.23743254e-01 -1.97098002e-01
6.73304796e-01 9.65976864e-02 1.05281126e+00 -1.28148234e+00
4.22384471e-01 1.17535007e+00 5.26234806e-01 6.62738681e-01
-1.28541100e+00 -6.95451558e-01 2.58155406e-01 5.50571442e-01
-1.39932930e+00 -9.12958384e-01 8.21392059e-01 -4.51618493e-01
1.20147920e+00 6.27161026e-01 6.09508395e-01 1.68297696e+00
-1.18747145e-01 1.10240114e+00 7.11021900e-01 -3.20072293e-01
3.61011624e-02 3.92490536e-01 5.78158945e-02 4.16484416e-01
-7.79236495e-01 -2.24423930e-01 -9.85348463e-01 -1.51992112e-01
1.27055183e-01 -7.50624109e-03 -3.28416795e-01 -2.00015619e-01
-9.61348951e-01 5.47675371e-01 1.27995893e-01 1.79208219e-01
-5.08558750e-01 1.59780428e-01 8.37494016e-01 8.48013222e-01
2.09056392e-01 1.20171472e-01 -5.25861323e-01 -5.58515728e-01
-9.84773397e-01 -1.74732119e-01 6.53995097e-01 7.72927463e-01
5.34814000e-01 4.72986072e-01 2.63871495e-02 1.20473337e+00
2.65272975e-01 5.26944697e-01 3.38061929e-01 -1.11440516e+00
1.91218719e-01 3.60898256e-01 -8.35726485e-02 -7.98991024e-01
-4.96945173e-01 -5.02728857e-02 -8.92822683e-01 4.20815080e-01
1.46416649e-01 -2.11400464e-01 -5.89990020e-01 1.87185752e+00
7.63749033e-02 8.42009392e-03 2.60827005e-01 9.30675566e-01
1.38014638e+00 8.78301978e-01 2.50645041e-01 -1.50462002e-01
1.27352417e+00 -7.44481087e-01 -7.70801008e-01 -2.89408594e-01
3.94047588e-01 -7.86765218e-01 1.33660817e+00 6.76450074e-01
-1.14749062e+00 -4.78185803e-01 -9.92979884e-01 -2.31911972e-01
-4.19367015e-01 4.16086555e-01 6.09005272e-01 4.69533622e-01
-1.06749630e+00 4.39343154e-02 -6.45951331e-01 -1.90159574e-01
2.05738246e-01 3.43295753e-01 -8.66856396e-01 3.12909447e-02
-1.17601609e+00 6.71469390e-01 8.22082534e-02 3.40293497e-01
-1.01444912e+00 -4.89911318e-01 -1.04014313e+00 1.19449571e-01
1.03385255e-01 -4.57634747e-01 1.19238782e+00 -1.67022300e+00
-1.56069040e+00 7.96356082e-01 -2.52109855e-01 3.36806029e-02
3.45675230e-01 -1.47542402e-01 -6.23953760e-01 3.49451631e-01
-4.85554695e-01 8.61817062e-01 1.02064812e+00 -1.29843509e+00
-3.23784649e-01 -3.01480651e-01 -1.98443651e-01 1.83110297e-01
-6.38673544e-01 2.56110758e-01 -5.74283063e-01 -3.25917184e-01
-1.71873078e-01 -8.78858745e-01 3.05609524e-01 1.07440971e-01
-1.87742472e-01 -2.56482549e-02 7.79095292e-01 -6.66962743e-01
1.11067593e+00 -2.56637526e+00 3.96844476e-01 1.99134201e-01
-1.33066267e-01 -2.74433531e-02 -5.42981565e-01 4.87196505e-01
-4.17852134e-01 5.49437851e-03 2.50273030e-02 -8.01300287e-01
2.99367785e-01 1.31637439e-01 -3.98869246e-01 2.43967667e-01
3.34278911e-01 6.99728489e-01 -3.36419314e-01 -4.61740732e-01
2.05560446e-01 1.00085521e+00 -4.25007045e-01 4.40062106e-01
-1.63135286e-02 4.13407534e-01 2.95009930e-02 1.05457747e+00
4.43556637e-01 4.22684429e-03 1.82832196e-01 -3.17550331e-01
5.43952361e-02 8.40299204e-03 -1.37377954e+00 1.81870997e+00
-8.40705216e-01 8.48963261e-01 5.39672256e-01 -7.62069345e-01
9.26255286e-01 8.16002488e-01 5.65965056e-01 -6.56132758e-01
2.43959770e-01 1.65689439e-01 -2.88889349e-01 -6.09846115e-01
4.71573412e-01 -1.40228361e-01 -3.27938080e-01 3.22466254e-01
5.10483265e-01 7.15762749e-02 -1.88551590e-01 1.88590512e-01
8.98264289e-01 8.30028206e-02 -9.77525394e-03 4.02927637e-01
5.05707324e-01 -2.66730726e-01 5.64574957e-01 4.71516073e-01
-1.93661690e-01 7.55945981e-01 5.80941975e-01 -1.65503383e-01
-7.20410645e-01 -1.03297901e+00 -1.54526204e-01 1.83359873e+00
-2.34790772e-01 -6.06887877e-01 -1.66283369e-01 -4.04260755e-01
-1.59688175e-01 6.67965770e-01 -5.73538721e-01 -1.74265146e-01
4.50598560e-02 -2.97991782e-01 7.92993426e-01 5.88859916e-01
5.34189232e-02 -1.16188896e+00 -4.87837106e-01 4.76592071e-02
-6.23262227e-01 -9.93098617e-01 -8.79798830e-02 4.23559874e-01
-2.77358174e-01 -7.90519416e-01 -3.49789560e-01 -6.59814715e-01
6.55824915e-02 -2.27954891e-02 1.16717601e+00 -2.98235983e-01
-2.08365694e-01 1.00100172e+00 -3.56557339e-01 -3.87498170e-01
-2.01457202e-01 -1.47536948e-01 -1.48583993e-01 2.97380626e-01
2.96842486e-01 -6.29222512e-01 -4.05825675e-01 2.56998032e-01
-8.24735820e-01 -6.16171323e-02 2.63034672e-01 9.53671694e-01
6.21472776e-01 -4.11341965e-01 7.93744445e-01 -2.34120101e-01
5.38681865e-01 -6.55937612e-01 -1.51025858e-02 2.93720096e-01
-1.21682875e-01 -3.13210845e-01 4.22229081e-01 -7.48114407e-01
-9.91569996e-01 2.98923135e-01 -6.09266996e-01 -9.54395294e-01
-4.14119750e-01 6.33603632e-01 -4.40002620e-01 1.27903178e-01
4.75715250e-01 1.07830055e-02 8.15173760e-02 -2.74559081e-01
4.46996778e-01 1.09470046e+00 6.03871405e-01 -5.53004563e-01
-1.59769915e-02 3.29666376e-01 -3.44144434e-01 -9.27245021e-01
-2.16793269e-01 -4.67784703e-01 -3.89265180e-01 -6.65372252e-01
7.92127252e-01 -1.37627685e+00 -1.04571164e+00 3.97865504e-01
-1.03287756e+00 -3.59784484e-01 -7.29543567e-02 4.78695869e-01
-7.66267896e-01 9.15785730e-02 -6.32045388e-01 -1.05117881e+00
-3.60013276e-01 -1.18123555e+00 1.20387459e+00 -1.08398497e-01
-5.51940322e-01 -7.66912460e-01 -2.21109495e-01 3.19984883e-01
3.77197921e-01 3.20189774e-01 7.13231266e-01 -6.35828376e-01
2.10388377e-02 -3.31768215e-01 1.67156849e-02 4.28515643e-01
-3.93769532e-01 3.26666743e-01 -1.51419687e+00 -1.64306387e-01
-2.10047022e-01 -1.22199035e+00 9.49853241e-01 1.14799023e-01
1.30312955e+00 -5.64473085e-02 1.77967682e-01 6.89816356e-01
1.05780363e+00 5.47629483e-02 6.12861574e-01 2.39833575e-02
5.33341229e-01 7.61793852e-01 5.97012877e-01 8.08791339e-01
7.28649080e-01 6.74420595e-01 6.38373315e-01 -1.53795615e-01
1.67403474e-01 3.93408686e-02 7.20523417e-01 8.39621246e-01
4.35911752e-02 -4.08950210e-01 -9.08764184e-01 4.98518348e-01
-2.00088787e+00 -1.13965964e+00 -5.22769615e-02 2.06114125e+00
7.70265341e-01 -4.45030212e-01 1.34031966e-01 2.30992317e-01
3.48582178e-01 5.90114854e-02 -5.09252071e-01 -8.26152802e-01
-3.25188279e-01 6.80799931e-02 -4.28802222e-01 4.37986970e-01
-1.18280005e+00 7.18652070e-01 5.56013870e+00 7.01051235e-01
-1.41519928e+00 2.70869672e-01 6.31525218e-01 -7.10023046e-01
-3.57697189e-01 -4.81693745e-01 -3.69897306e-01 1.98448017e-01
1.27246630e+00 3.32636237e-01 4.70032841e-01 7.78072834e-01
-9.89471842e-03 1.18077807e-01 -1.40241051e+00 1.72912097e+00
2.57595688e-01 -8.36766958e-01 -1.34692013e-01 -4.06212479e-01
-1.94823127e-02 9.22576264e-02 2.72243977e-01 6.33189380e-01
-1.61021829e-01 -1.28184640e+00 9.08385456e-01 6.98597729e-01
1.15065444e+00 -8.62679660e-01 6.21689260e-01 1.70053408e-01
-1.24301875e+00 -1.97094470e-01 3.06594931e-02 1.30628809e-01
7.45618641e-02 9.22993422e-02 -3.78075033e-01 4.39603448e-01
1.03656554e+00 6.72521055e-01 -2.49396503e-01 6.02965236e-01
-5.01474068e-02 5.27011931e-01 -2.98600227e-01 2.14317933e-01
-1.35698179e-02 1.38885379e-01 3.61796260e-01 1.53402340e+00
4.15051371e-01 -1.14110969e-01 1.51928380e-01 4.49467510e-01
-1.06654771e-01 3.96704018e-01 -5.66953063e-01 -1.03369504e-01
3.83414239e-01 1.47240746e+00 -1.01314880e-01 -8.80848318e-02
-5.75273633e-01 1.08056796e+00 4.85543877e-01 4.80015486e-01
-9.85894084e-01 -3.41460854e-01 8.69968057e-01 -3.75128537e-01
1.36050910e-01 -8.28396305e-02 -4.34548594e-02 -1.29032624e+00
-1.16789989e-01 -1.14433062e+00 6.36568308e-01 -1.22690153e+00
-1.41111898e+00 8.30188215e-01 -2.87679940e-01 -1.13299596e+00
-5.99510372e-01 -5.99410236e-01 -3.61190259e-01 8.08854759e-01
-1.28706944e+00 -1.42312407e+00 -4.76057202e-01 1.09789860e+00
4.16864693e-01 -2.98484027e-01 1.21792889e+00 5.63665688e-01
-4.44081098e-01 8.63842666e-01 -9.75811779e-02 -1.09644242e-01
1.20489156e+00 -1.00308859e+00 -5.33837974e-01 2.92691231e-01
2.80575275e-01 1.25617325e-01 5.70926964e-01 -6.93901107e-02
-1.71451569e+00 -7.57020593e-01 7.80978560e-01 -3.37638438e-01
5.91527343e-01 -6.91612661e-01 -9.86944973e-01 6.69223487e-01
5.02166331e-01 -5.84158190e-02 1.33316517e+00 3.17114592e-01
-7.67515421e-01 -2.35945389e-01 -9.16293025e-01 4.14126337e-01
5.71915805e-01 -1.01883614e+00 -2.62373716e-01 2.23703794e-02
4.63452131e-01 -8.58955756e-02 -9.94016349e-01 4.13504660e-01
9.19667423e-01 -9.84520137e-01 9.56158638e-01 -6.06869161e-01
5.39608181e-01 -1.00951761e-01 -6.69822514e-01 -1.27900982e+00
2.16859579e-02 -4.56782728e-01 -1.25063419e-01 1.54385865e+00
4.82325703e-01 -3.46704811e-01 1.85413823e-01 7.68970311e-01
-1.29615858e-01 -7.17008054e-01 -1.05249393e+00 -3.48372906e-01
-1.72551215e-01 -1.05041158e+00 3.62185240e-01 1.15241778e+00
2.74517328e-01 4.02370363e-01 -6.55210733e-01 5.59014678e-02
7.87263662e-02 2.11636931e-01 7.82437563e-01 -9.86067533e-01
-3.69741380e-01 -4.48098451e-01 -5.26197076e-01 -5.38170040e-01
6.16943121e-01 -9.73459244e-01 -9.48192850e-02 -1.32075977e+00
8.76814201e-02 -9.73448306e-02 -5.36665678e-01 1.00315619e+00
4.10074145e-01 4.37222481e-01 4.47342962e-01 -3.08433305e-02
-7.40993738e-01 8.84745777e-01 6.34623051e-01 -3.07048202e-01
-2.99779952e-01 -2.21530199e-01 -5.65814257e-01 6.52458310e-01
7.74040937e-01 -1.18505836e-01 -4.40962970e-01 -3.84764791e-01
4.46045965e-01 4.37397093e-01 5.34110069e-01 -8.11865151e-01
2.07826272e-01 1.40525043e-01 5.59864402e-01 -4.67425883e-01
9.88404095e-01 -1.03402972e+00 3.51099700e-01 -3.31069767e-01
-5.04084468e-01 1.57221913e-01 6.07116044e-01 3.25853407e-01
-5.60029626e-01 6.43848479e-02 6.42494142e-01 2.31327415e-01
-8.23177397e-01 9.84648541e-02 -5.63753605e-01 -8.57718363e-02
8.82393658e-01 -1.37446031e-01 -3.56396765e-01 -8.37826550e-01
-1.20943534e+00 3.98058891e-01 3.44957590e-01 6.44553900e-01
9.15947378e-01 -1.53152978e+00 -6.92441761e-01 2.35240638e-01
4.89223778e-01 -6.20573103e-01 7.10699856e-01 9.46034193e-01
2.19351128e-02 -8.46372098e-02 -4.97644335e-01 -6.83931410e-01
-1.71771336e+00 3.37134778e-01 4.47014064e-01 2.00717300e-01
-1.32925808e-01 8.10383081e-01 8.00998881e-02 -6.46857858e-01
6.24653637e-01 1.73940182e-01 -2.46128872e-01 5.73719144e-01
5.61201274e-01 1.66965216e-01 2.03545451e-01 -9.20923054e-01
-4.92232233e-01 2.42858142e-01 1.73054650e-01 -3.53180319e-01
1.46841192e+00 -2.71579057e-01 -9.53585654e-02 1.13615429e+00
1.25782049e+00 -1.62706614e-01 -1.12602139e+00 -7.71963969e-02
-5.22186518e-01 -1.67982280e-01 1.27927646e-01 -1.10814250e+00
-1.02804518e+00 1.35297811e+00 8.43591928e-01 -2.40677036e-02
1.55583310e+00 2.34249141e-02 3.33679706e-01 4.29482043e-01
1.13418080e-01 -1.17370236e+00 1.98710084e-01 5.28223634e-01
1.31056511e+00 -1.45464146e+00 -3.49492252e-01 2.74035987e-02
-1.36147022e+00 1.15192628e+00 5.37068307e-01 3.71985853e-01
6.23335958e-01 4.15233314e-01 4.09909070e-01 -3.95919867e-02
-1.36353528e+00 -1.72316134e-01 4.82022405e-01 4.63698119e-01
7.70145476e-01 3.56639139e-02 3.18798155e-01 1.05912650e+00
4.23689112e-02 -3.41825157e-01 2.43806437e-01 7.75291562e-01
-5.11467904e-02 -9.32030320e-01 -3.69599491e-01 1.67345852e-01
-4.31802988e-01 -9.78157744e-02 -6.07891083e-01 6.95014000e-01
1.70708466e-02 1.20283937e+00 2.21492574e-01 -6.42591059e-01
3.17033738e-01 6.46257341e-01 3.10651392e-01 -1.49051175e-01
-8.74643683e-01 4.58565086e-01 3.11275780e-01 -8.16303611e-01
-3.92970443e-01 -7.28226066e-01 -1.22233963e+00 -1.82584330e-01
9.00506601e-02 -4.29018065e-02 8.01363051e-01 4.41803157e-01
6.37086868e-01 5.87166250e-01 5.90107501e-01 -1.03056192e+00
-1.19658574e-01 -7.43197680e-01 -6.15070701e-01 3.50211442e-01
4.46049362e-01 -5.64765036e-01 -3.83102477e-01 1.08290188e-01]
|
[13.273238182067871, 5.147307395935059]
|
a8e86060-9fca-4be8-8554-9af34e8db166
|
improving-the-generalization-of-meta-learning
|
2107.11056
| null |
https://arxiv.org/abs/2107.11056v1
|
https://arxiv.org/pdf/2107.11056v1.pdf
|
Improving the Generalization of Meta-learning on Unseen Domains via Adversarial Shift
|
Meta-learning provides a promising way for learning to efficiently learn and achieves great success in many applications. However, most meta-learning literature focuses on dealing with tasks from a same domain, making it brittle to generalize to tasks from the other unseen domains. In this work, we address this problem by simulating tasks from the other unseen domains to improve the generalization and robustness of meta-learning method. Specifically, we propose a model-agnostic shift layer to learn how to simulate the domain shift and generate pseudo tasks, and develop a new adversarial learning-to-learn mechanism to train it. Based on the pseudo tasks, the meta-learning model can learn cross-domain meta-knowledge, which can generalize well on unseen domains. We conduct extensive experiments under the domain generalization setting. Experimental results demonstrate that the proposed shift layer is applicable to various meta-learning frameworks. Moreover, our method also leads to state-of-the-art performance on different cross-domain few-shot classification benchmarks and produces good results on cross-domain few-shot regression.
|
['Yao Gao', 'Pinzhuo Tian']
|
2021-07-23
| null | null | null | null |
['cross-domain-few-shot']
|
['computer-vision']
|
[ 3.16161275e-01 -1.73388392e-01 -2.19202071e-01 -3.25843990e-01
-9.79222894e-01 -4.16754156e-01 6.17704451e-01 -2.67623961e-01
-2.44739905e-01 9.16570723e-01 -1.49898008e-01 -3.01620457e-02
3.67821828e-02 -9.14145768e-01 -9.42235470e-01 -6.54384494e-01
3.22394669e-01 3.85602951e-01 3.76879513e-01 -5.55638194e-01
-1.37173638e-01 -8.95250961e-02 -1.32864833e+00 5.46118021e-01
1.04174554e+00 8.16148460e-01 2.04556376e-01 3.16149592e-01
-6.77437857e-02 7.86367476e-01 -8.70416880e-01 -4.95043844e-01
2.86117703e-01 -5.27166307e-01 -6.92496896e-01 2.10231356e-02
2.63402134e-01 -3.51851910e-01 -4.96288627e-01 1.09755850e+00
5.50867438e-01 5.02881050e-01 8.36330712e-01 -1.44005108e+00
-1.21345317e+00 3.54941994e-01 -4.30115402e-01 2.57373720e-01
2.05424681e-01 2.97825396e-01 5.82273960e-01 -7.84136653e-01
4.89268303e-01 1.26384628e+00 5.60802698e-01 1.16332483e+00
-1.10679245e+00 -9.78618681e-01 2.10509002e-01 2.29436919e-01
-1.07466316e+00 -2.06093907e-01 9.45980847e-01 -2.21045256e-01
6.95869803e-01 -3.05064619e-01 5.83595522e-02 1.98796439e+00
1.77087009e-01 8.56064320e-01 1.30688810e+00 -2.64229059e-01
6.20412946e-01 1.09613828e-01 -1.07394449e-01 1.87707916e-01
2.24090829e-01 1.62631020e-01 -2.18115628e-01 -1.51543960e-01
4.94185627e-01 4.32291716e-01 -2.18158126e-01 -5.22464633e-01
-1.04391193e+00 8.85217905e-01 4.42448139e-01 1.58700570e-01
-1.22748919e-01 -2.29370855e-02 7.04562962e-01 6.64108634e-01
8.13405514e-01 5.04126370e-01 -4.10361826e-01 3.30890566e-02
-5.59513330e-01 4.21098202e-01 6.70629799e-01 1.23360729e+00
6.97386265e-01 2.57651836e-01 -1.17032185e-01 1.17701185e+00
-2.89428681e-01 4.47522044e-01 9.73637164e-01 -5.86730599e-01
9.41500306e-01 4.83686864e-01 1.08005710e-01 -5.41063249e-01
-1.14258938e-01 -4.30407345e-01 -1.00626969e+00 1.42079592e-01
1.58929050e-01 -4.80059445e-01 -1.12586069e+00 1.92055309e+00
1.95132613e-01 7.68419385e-01 5.03043830e-01 7.32059300e-01
8.69955599e-01 7.75285542e-01 2.65570849e-01 1.07317708e-01
1.01924384e+00 -1.12683511e+00 -3.42437387e-01 -6.91104174e-01
6.00401402e-01 -2.37975568e-01 1.26382887e+00 1.80159226e-01
-8.71603549e-01 -9.60403681e-01 -1.30220652e+00 1.93230242e-01
-5.88150263e-01 -5.19290090e-01 1.69355705e-01 3.32460761e-01
-4.06179637e-01 7.29667187e-01 -6.22692943e-01 -2.82344788e-01
6.53743207e-01 -7.49336630e-02 -3.23719144e-01 -5.87936819e-01
-1.53653061e+00 1.02216315e+00 9.63271797e-01 -5.94794571e-01
-1.28849936e+00 -9.55754220e-01 -1.08752739e+00 1.28028160e-02
4.16589558e-01 -8.26178610e-01 1.40368152e+00 -1.07250237e+00
-1.33959174e+00 8.57579947e-01 2.42395744e-01 -5.10784149e-01
6.57668769e-01 -1.32497922e-01 -8.44195604e-01 -1.13663189e-01
1.42910033e-01 4.89469141e-01 1.15674067e+00 -1.28670716e+00
-6.69950128e-01 -3.27640861e-01 9.88366157e-02 1.39253765e-01
-6.51234210e-01 -4.89714801e-01 -5.31562977e-02 -9.86488461e-01
-3.94181997e-01 -8.64975691e-01 -1.75628170e-01 -3.84471238e-01
-7.73738772e-02 -1.73207819e-01 7.70393312e-01 -2.40376651e-01
8.44834566e-01 -2.24933171e+00 1.29338369e-01 -3.64266604e-01
5.48488982e-02 6.19680822e-01 -4.71008301e-01 4.24189061e-01
-2.26630598e-01 -9.74196345e-02 -4.12830621e-01 -2.58257359e-01
-1.23594835e-01 4.56357747e-01 -7.31461644e-01 1.70684591e-01
3.36709112e-01 1.01795352e+00 -1.17155635e+00 -1.97254509e-01
1.63532466e-01 2.10383892e-01 -3.32076907e-01 4.75716650e-01
-3.13247472e-01 4.87421334e-01 -5.79044521e-01 5.71625113e-01
8.42969060e-01 -2.67274052e-01 2.15419177e-02 2.59190470e-01
5.98603845e-01 -3.72679718e-02 -8.16801846e-01 2.02946186e+00
-8.26597214e-01 2.46173039e-01 -4.69359785e-01 -1.41670001e+00
1.09458470e+00 2.08333656e-01 9.03826207e-02 -7.69328594e-01
8.72260407e-02 2.44902462e-01 -1.11547440e-01 -4.62203950e-01
2.77371407e-01 -5.86093366e-01 -4.00563538e-01 3.91806185e-01
5.86249650e-01 -1.19312637e-01 -8.63387138e-02 -2.33524233e-01
1.04183233e+00 1.27261803e-01 4.76663291e-01 3.73285380e-03
4.37552899e-01 1.04613237e-01 6.71314001e-01 8.02319109e-01
-3.75535458e-01 5.77947259e-01 1.50494993e-01 -5.15375495e-01
-1.13283682e+00 -1.27304006e+00 4.95120585e-02 1.46123815e+00
4.27949995e-01 5.83766997e-02 -7.60143936e-01 -1.11466789e+00
1.94793776e-01 1.01409960e+00 -8.95931244e-01 -8.60565126e-01
-3.76337439e-01 -7.12232411e-01 4.73147929e-01 8.32510233e-01
7.64745414e-01 -1.07588637e+00 -2.63896525e-01 4.16074961e-01
-9.59276482e-02 -1.18742907e+00 -1.95645764e-01 1.70833752e-01
-9.73801315e-01 -1.02104533e+00 -1.25730217e+00 -1.01230717e+00
4.12775725e-01 5.89503407e-01 1.12596774e+00 -4.49690163e-01
-1.03386961e-01 2.14206114e-01 -5.75871229e-01 -5.10459065e-01
-7.39204764e-01 2.09174886e-01 1.87777683e-01 -8.30284357e-02
6.49650872e-01 -7.73463428e-01 -3.78963947e-01 5.83926022e-01
-1.15007007e+00 -2.70370245e-01 4.48198944e-01 1.25795507e+00
4.76174921e-01 1.92440033e-01 1.23457921e+00 -1.23646045e+00
8.17250669e-01 -1.03298581e+00 -3.23246628e-01 3.81661057e-01
-3.88100982e-01 5.80034927e-02 1.29762387e+00 -9.55787838e-01
-1.25078070e+00 -2.84075558e-01 1.04797117e-01 -8.38307440e-01
-3.79554898e-01 1.67003527e-01 -3.96905869e-01 4.69325185e-02
1.02773046e+00 4.50695276e-01 -2.17738137e-01 -4.89512146e-01
5.38603902e-01 6.89732850e-01 6.43020928e-01 -7.45235324e-01
1.05721819e+00 5.92511714e-01 -2.39719570e-01 -5.61812460e-01
-1.16758227e+00 -4.06552583e-01 -4.90843296e-01 1.96768865e-01
6.19503200e-01 -1.19525623e+00 2.26858500e-02 5.31682789e-01
-9.12658274e-01 -4.87411708e-01 -2.38158882e-01 2.16800675e-01
-8.57871532e-01 1.27690554e-01 -3.03656578e-01 -3.23606700e-01
-3.07811707e-01 -1.02627242e+00 8.09465826e-01 3.16406786e-01
8.94089341e-02 -1.35071647e+00 4.10606205e-01 1.32225260e-01
3.28046590e-01 4.32590753e-01 8.99396598e-01 -1.19317997e+00
4.76507284e-02 -1.22590855e-01 -6.68293089e-02 7.18190491e-01
2.61252433e-01 -7.39910781e-01 -1.15493512e+00 -4.93032932e-01
1.94884375e-01 -8.65435481e-01 1.11096168e+00 -1.09311096e-01
1.29868984e+00 -1.90941319e-01 -3.67834479e-01 7.99015760e-01
1.45062423e+00 1.12284653e-01 5.65030515e-01 5.51769435e-01
5.25842011e-01 3.87227297e-01 8.37603331e-01 3.97449225e-01
1.72698885e-01 4.79163945e-01 4.04646903e-01 1.05557352e-01
-1.01674125e-01 -4.98050034e-01 3.34251612e-01 5.21720529e-01
1.37504056e-01 -1.70597777e-01 -8.39955211e-01 7.19641685e-01
-1.87310195e+00 -1.10903108e+00 5.56613684e-01 2.03791523e+00
8.98965836e-01 3.15589160e-01 1.80470824e-01 -1.74800217e-01
9.28815007e-01 3.56660455e-01 -9.86624360e-01 -3.24566036e-01
-1.32344454e-03 4.40406591e-01 2.41502911e-01 -5.57453968e-02
-1.29387903e+00 1.07872021e+00 5.71342468e+00 1.00570273e+00
-1.07412481e+00 4.35675323e-01 4.24328089e-01 -1.59568561e-03
-9.11418423e-02 -3.34138930e-01 -7.26121545e-01 7.51934886e-01
8.05683911e-01 -5.26717067e-01 3.90427202e-01 1.48422170e+00
-3.92560691e-01 5.03907084e-01 -1.33072901e+00 8.70355904e-01
2.41399601e-01 -1.05800259e+00 2.84318924e-01 -2.25560278e-01
1.22608340e+00 1.43010542e-02 3.44075084e-01 1.15453398e+00
5.47262073e-01 -8.97795796e-01 2.24884987e-01 9.69093665e-02
9.10636365e-01 -9.19015646e-01 6.69529796e-01 6.87982857e-01
-8.92206430e-01 -3.80883962e-01 -1.01234257e+00 7.97742009e-02
-8.86516273e-02 1.63123459e-01 -8.33359659e-01 6.67269826e-01
4.41208869e-01 8.10036778e-01 -4.52211648e-01 8.24710548e-01
-1.81485921e-01 3.66753340e-01 3.69029939e-01 5.08205704e-02
2.86935866e-01 2.04735190e-01 4.56273615e-01 1.16957986e+00
3.49268466e-01 1.16259977e-02 2.22621903e-01 8.52945924e-01
-4.67546135e-01 -2.22202852e-01 -1.01048601e+00 6.98550865e-02
5.76951087e-01 9.24858570e-01 -3.96430343e-02 -4.21779454e-01
-6.69019580e-01 1.21282542e+00 6.12773716e-01 5.81359386e-01
-1.07710135e+00 -6.73163414e-01 8.43147159e-01 -3.33316773e-02
3.28183264e-01 1.07059613e-01 -6.18454367e-02 -1.55494416e+00
5.79583757e-02 -1.10484302e+00 5.74564874e-01 -6.52281463e-01
-1.97446489e+00 5.42780459e-01 1.04061171e-01 -1.76924145e+00
-4.90822405e-01 -7.41349459e-01 -9.01519001e-01 8.93905640e-01
-1.89901364e+00 -1.17614925e+00 -5.63980937e-01 9.16069627e-01
8.89600694e-01 -6.11939490e-01 9.31942761e-01 4.23321947e-02
-3.74313623e-01 8.84268761e-01 4.42193091e-01 3.37615818e-01
1.16290021e+00 -1.21318042e+00 7.59749830e-01 5.98655105e-01
-1.14734359e-01 4.75650132e-01 4.35997605e-01 -5.40216982e-01
-1.23411763e+00 -1.60309100e+00 9.10038054e-02 -4.87182081e-01
7.90353239e-01 -3.81492496e-01 -1.33097482e+00 8.29722464e-01
1.05891868e-01 3.08474809e-01 7.52656579e-01 -4.89888862e-02
-8.08394313e-01 -3.02509576e-01 -1.25962269e+00 5.05778372e-01
1.13569355e+00 -4.76384342e-01 -1.35488272e+00 3.30131114e-01
9.26171303e-01 -4.44485366e-01 -8.16015005e-01 4.40834701e-01
3.01103860e-01 -7.33613312e-01 1.20859778e+00 -1.34100044e+00
9.38933611e-01 1.60912246e-01 -2.17935041e-01 -2.12544918e+00
-4.22366440e-01 -2.72389382e-01 -3.68361652e-01 1.27600169e+00
2.16659129e-01 -7.61171937e-01 6.24079525e-01 3.03878039e-01
-2.14450613e-01 -5.09894073e-01 -8.65297854e-01 -1.52768028e+00
6.42582178e-01 -1.77376226e-01 8.15375388e-01 1.20806336e+00
-7.51920715e-02 3.73277843e-01 -4.99853730e-01 6.49122596e-02
8.30151379e-01 2.63842553e-01 1.02667880e+00 -1.26416087e+00
-4.86189872e-01 -1.27642423e-01 -3.19313407e-01 -8.45995903e-01
7.16748357e-01 -1.02366531e+00 -5.02255596e-02 -1.21990287e+00
2.02632189e-01 -2.62074858e-01 -7.42679060e-01 4.31042790e-01
-6.18859768e-01 -1.09736817e-02 1.79169402e-01 1.72153533e-01
-7.21731007e-01 7.47572720e-01 1.43197954e+00 -3.99595052e-01
-1.17799595e-01 2.42289037e-01 -8.80420804e-01 6.43351495e-01
9.55044210e-01 -6.21474326e-01 -6.93416476e-01 -5.48377454e-01
-3.41492832e-01 -1.48099452e-01 3.82536292e-01 -1.15725493e+00
4.50711399e-02 -4.34044898e-01 6.11734390e-01 -4.54617776e-02
4.34252828e-01 -7.74306595e-01 -3.46968979e-01 3.54282349e-01
-2.97241718e-01 -2.33782470e-01 2.96481252e-01 9.29558396e-01
-3.03503007e-01 -3.90162319e-01 1.12453508e+00 -3.34875852e-01
-1.14581060e+00 4.53071028e-01 2.87022978e-01 8.82959843e-01
1.37901676e+00 4.20857593e-03 -6.39364421e-01 -2.40308702e-01
-6.72576189e-01 3.04307520e-01 5.80129623e-01 8.32428098e-01
6.55965626e-01 -1.60999727e+00 -7.83492982e-01 1.83300912e-01
7.19576240e-01 2.33914591e-02 5.15533328e-01 5.84307015e-02
-9.93534178e-02 5.60866781e-02 -6.13360941e-01 -5.59221625e-01
-6.43562555e-01 1.15965784e+00 3.12421173e-01 -3.01951319e-01
-5.63426375e-01 6.79753602e-01 4.77115870e-01 -5.55696070e-01
1.15552351e-01 -7.42129385e-02 1.24852508e-01 -1.07310824e-01
8.37247670e-01 5.16808987e-01 -5.02992757e-02 -1.58799887e-01
-1.97096542e-01 4.39751595e-01 -3.98085117e-01 5.76475300e-02
1.37865293e+00 2.00994402e-01 5.99483252e-01 5.04537404e-01
1.38511980e+00 -5.32521367e-01 -1.59352362e+00 -5.44291317e-01
-1.73889518e-01 -5.12557268e-01 -4.16319877e-01 -7.58053362e-01
-6.95118129e-01 1.18532717e+00 3.79103273e-01 1.14296049e-01
1.08755434e+00 -1.49337485e-01 1.06327522e+00 5.58255136e-01
5.22055328e-01 -1.21451056e+00 5.06200373e-01 5.64429700e-01
7.39518344e-01 -1.61681890e+00 -2.54839569e-01 1.65546909e-02
-1.00782824e+00 9.99641120e-01 1.07562828e+00 -4.37496901e-01
4.84343976e-01 -1.39511824e-01 -1.41389400e-01 3.84051234e-01
-8.02575409e-01 -3.68413739e-02 1.54595017e-01 1.12400496e+00
-6.29925802e-02 3.65734883e-02 2.16070324e-01 1.05162334e+00
1.94188833e-01 2.43631795e-01 3.99068832e-01 9.57481146e-01
-5.51563025e-01 -1.20008135e+00 -1.50552467e-01 2.30158284e-01
-3.19871932e-01 3.91836986e-02 -2.50314504e-01 9.71831977e-01
5.97768053e-02 8.60231578e-01 -1.56595960e-01 -5.02175748e-01
6.01455688e-01 3.77362221e-01 4.90538180e-01 -9.67814386e-01
-3.72904420e-01 -5.30105710e-01 -2.24938452e-01 -2.83048213e-01
-9.49625447e-02 -1.17272213e-01 -8.68137538e-01 -2.10377499e-01
7.14835078e-02 2.15048026e-02 2.44053885e-01 8.91096950e-01
4.37463343e-01 6.36736214e-01 8.45117271e-01 -6.56466603e-01
-1.25035107e+00 -1.07691443e+00 -6.16997242e-01 9.28383529e-01
4.12135780e-01 -1.06403494e+00 -3.40999901e-01 -2.29788106e-02]
|
[10.168701171875, 3.0557456016540527]
|
f4345d30-76bf-4a23-8352-b79b13a61ac7
|
improving-truthfulness-of-headline-generation
|
2005.00882
| null |
https://arxiv.org/abs/2005.00882v2
|
https://arxiv.org/pdf/2005.00882v2.pdf
|
Improving Truthfulness of Headline Generation
|
Most studies on abstractive summarization report ROUGE scores between system and reference summaries. However, we have a concern about the truthfulness of generated summaries: whether all facts of a generated summary are mentioned in the source text. This paper explores improving the truthfulness in headline generation on two popular datasets. Analyzing headlines generated by the state-of-the-art encoder-decoder model, we show that the model sometimes generates untruthful headlines. We conjecture that one of the reasons lies in untruthful supervision data used for training the model. In order to quantify the truthfulness of article-headline pairs, we consider the textual entailment of whether an article entails its headline. After confirming quite a few untruthful instances in the datasets, this study hypothesizes that removing untruthful instances from the supervision data may remedy the problem of the untruthful behaviors of the model. Building a binary classifier that predicts an entailment relation between an article and its headline, we filter out untruthful instances from the supervision data. Experimental results demonstrate that the headline generation model trained on filtered supervision data shows no clear difference in ROUGE scores but remarkable improvements in automatic and manual evaluations of the generated headlines.
|
['Naoaki Okazaki', 'Sho Takase', 'Kazuki Matsumaru']
|
2020-05-02
|
improving-truthfulness-of-headline-generation-1
|
https://aclanthology.org/2020.acl-main.123
|
https://aclanthology.org/2020.acl-main.123.pdf
|
acl-2020-6
|
['headline-generation']
|
['natural-language-processing']
|
[ 2.4806423e-01 7.4556339e-01 -4.6399388e-01 -3.0406964e-01
-1.1511314e+00 -6.7289686e-01 9.3359065e-01 4.9673161e-01
-7.9367213e-02 1.1639702e+00 1.1889824e+00 -2.5870293e-01
2.1673734e-01 -5.3503025e-01 -1.0434827e+00 -1.4424014e-01
4.3561661e-01 3.1226131e-01 -2.4899581e-02 -4.2944768e-01
7.8714138e-01 -1.2905422e-01 -1.3040156e+00 6.6645604e-01
1.2552435e+00 4.2239606e-01 -9.5873363e-02 8.9687943e-01
8.9545220e-02 1.3378927e+00 -1.3051794e+00 -9.3001330e-01
-1.7806089e-01 -8.3299506e-01 -1.0902485e+00 2.9823014e-01
1.0503175e+00 -4.3442225e-01 -2.9335433e-01 1.1683602e+00
2.0067999e-01 -1.4043291e-01 9.4304508e-01 -1.1519929e+00
-8.4116709e-01 1.1896582e+00 -3.8989842e-01 4.7034121e-01
6.7551923e-01 8.7599553e-02 1.3775244e+00 -5.7410431e-01
7.1394324e-01 9.9748152e-01 4.7831780e-01 4.2130724e-01
-9.5851707e-01 -4.1490346e-01 -1.7465064e-01 3.0616277e-01
-9.6929979e-01 -7.3081434e-01 6.9111657e-01 -5.3244454e-01
1.0099860e+00 5.3448242e-01 3.2877839e-01 1.4538755e+00
6.3816845e-01 8.1722462e-01 7.0637167e-01 -4.7143412e-01
-4.9005736e-02 4.3491754e-01 6.3544703e-01 6.6889894e-01
7.7617812e-01 -2.2934759e-01 -7.4356854e-01 -1.7810836e-01
5.5338241e-02 -5.0775069e-01 -6.7721379e-01 4.7194082e-01
-1.1311084e+00 1.0231423e+00 1.0018116e-01 2.9514897e-01
-4.1566396e-01 -1.5794972e-01 7.0225489e-01 2.2633196e-01
7.1995932e-01 1.0032082e+00 -1.2392623e-01 -2.6569167e-01
-1.5600761e+00 3.2661068e-01 1.1768695e+00 1.1767704e+00
5.9038603e-01 -8.1587985e-02 -7.0691514e-01 4.3435889e-01
-2.1580316e-01 4.3705642e-01 6.9080436e-01 -8.6398101e-01
7.4427140e-01 6.7019188e-01 2.5861427e-01 -1.3184779e+00
-1.5020929e-02 -6.4607966e-01 -7.6142055e-01 -5.0775254e-01
4.6331257e-02 -1.9416386e-01 -4.4065019e-01 1.3085093e+00
-3.0046278e-01 -2.5810933e-01 2.5732878e-01 6.7624074e-01
1.0800593e+00 8.6237442e-01 -3.2912683e-01 -5.7888234e-01
1.1351053e+00 -1.2141159e+00 -1.2238564e+00 -2.4390990e-01
8.7988591e-01 -9.7709185e-01 1.2637715e+00 1.7078409e-01
-1.2744472e+00 -4.9193797e-01 -1.4987335e+00 -3.1325036e-01
-3.8833775e-02 4.6521941e-01 6.9169804e-02 2.2678675e-01
-6.5165532e-01 9.0924871e-01 -2.3090513e-01 -2.7436516e-01
2.2764879e-01 -3.6388141e-01 -2.8952909e-01 6.9668718e-02
-1.4101955e+00 1.2651553e+00 2.7174494e-01 -4.0477085e-01
-8.5024595e-01 -7.1165496e-01 -9.4920254e-01 1.9348596e-01
5.9841454e-01 -5.8117217e-01 1.6148511e+00 -7.3943901e-01
-1.0435222e+00 7.8984571e-01 -4.2324051e-01 -8.1867969e-01
4.9823993e-01 -4.9283785e-01 -5.5402958e-01 1.8086354e-01
5.2339363e-01 1.4330074e-04 7.7083457e-01 -1.1070819e+00
-5.6477916e-01 7.7125281e-02 -1.7890497e-01 1.1245264e-02
-1.4964020e-01 -1.3309848e-01 1.7940980e-01 -6.7860293e-01
-3.8999307e-01 -5.3222883e-01 3.5969463e-01 -8.7631381e-01
-1.2554045e+00 -3.8649988e-01 4.7673145e-01 -9.0387541e-01
1.8530927e+00 -1.7774835e+00 -3.5485426e-01 -3.2086432e-01
2.9728904e-01 1.4517584e-01 2.9463148e-02 8.2430005e-01
2.1588255e-02 5.2667016e-01 -4.0083181e-02 -9.3589529e-02
6.5630183e-02 4.6294756e-02 -9.3868518e-01 3.4019601e-01
2.6878342e-01 8.7065911e-01 -1.1060357e+00 -7.2212863e-01
-3.0412161e-01 -1.3874307e-01 -3.3641803e-01 5.0772595e-01
-3.1114459e-01 -2.9009193e-02 -3.1470892e-01 1.0200301e-01
3.2458955e-01 -3.5943514e-01 -4.4036466e-01 -5.3821129e-01
-1.8048677e-01 1.0377650e+00 -4.2664996e-01 1.1494621e+00
-2.7633548e-01 1.0434620e+00 -6.9897223e-01 -4.5176202e-01
8.9216864e-01 4.8927334e-01 -2.4740377e-01 -5.0614017e-01
1.6365093e-01 2.2539762e-01 -8.2050711e-02 -7.9071045e-01
1.3234210e+00 -3.0965978e-01 -3.3656520e-01 7.2152585e-01
2.3667158e-01 -4.2188302e-01 5.7185912e-01 8.2501328e-01
1.1170969e+00 -3.8048846e-01 5.6033468e-01 -8.7264277e-02
3.3957368e-01 5.5887580e-01 2.4330309e-01 1.0615469e+00
-6.6055834e-02 8.2163030e-01 8.4988815e-01 -6.8694457e-02
-1.0902905e+00 -7.0494413e-01 9.8635897e-02 7.8973627e-01
-9.7346485e-02 -1.0139687e+00 -9.5419186e-01 -8.7366343e-01
-3.4052038e-01 1.7846711e+00 -8.7278670e-01 -4.9018902e-01
-3.0297893e-01 -2.1273272e-01 8.4031153e-01 1.7681745e-01
4.8263431e-01 -6.9503784e-01 -6.8187314e-01 1.6890857e-02
-8.1817544e-01 -1.1198827e+00 -7.3831296e-01 -4.5224786e-02
-6.4146829e-01 -1.0753648e+00 -2.9477102e-01 -3.0842656e-01
4.9238914e-01 2.1414074e-01 1.4539937e+00 9.3538910e-02
3.9126909e-01 -6.2110882e-02 -4.7622117e-01 -5.0646424e-01
-1.2505496e+00 2.1921900e-01 -1.7382661e-01 -3.9682606e-01
4.3624920e-01 1.2729466e-03 -1.9582519e-01 -1.1149864e-01
-8.0883372e-01 1.9571516e-01 3.9980584e-01 7.9800588e-01
1.2632413e-03 2.5616214e-02 6.7116940e-01 -1.1987422e+00
1.2457182e+00 -4.6181935e-01 1.4201889e-02 4.3511957e-01
-7.1443713e-01 3.0645657e-01 7.8617871e-01 -1.5977111e-01
-1.0465034e+00 -6.6920924e-01 -3.9515272e-04 -1.0734157e-01
-4.1843085e-03 6.5414685e-01 1.1822547e-01 8.3276272e-01
1.0678835e+00 2.5114673e-01 -1.5685138e-01 -1.5670449e-01
4.3543595e-01 8.5064632e-01 8.1110322e-01 -2.5610933e-01
6.4032447e-01 1.3513256e-02 -4.9154893e-01 -8.7852263e-01
-1.7342815e+00 -4.2757538e-01 -3.1370145e-01 -2.6518428e-01
5.6335431e-01 -7.5987804e-01 -1.6284826e-01 -2.1954539e-01
-1.6289736e+00 3.1109941e-01 -4.7638661e-01 1.7218815e-01
-5.2563310e-01 4.8068613e-01 -5.2941632e-01 -5.4061472e-01
-6.5256160e-01 -9.1340196e-01 9.2350125e-01 1.5310767e-01
-1.2103207e+00 -9.8046350e-01 3.2839108e-01 5.4192662e-01
1.6776770e-01 2.4120700e-01 1.0471954e+00 -1.2545218e+00
5.0111521e-02 -7.6882523e-01 3.9132126e-02 4.4643921e-01
3.0619323e-01 4.2425787e-01 -9.3773407e-01 1.6189145e-01
3.0649856e-01 -5.3522587e-01 8.1297010e-01 1.6405462e-01
7.0305783e-01 -1.4009037e+00 -4.5659970e-02 -1.6704100e-01
9.9583387e-01 -3.5557103e-01 6.8483764e-01 1.6424739e-01
5.4113299e-01 7.1197528e-01 6.6275227e-01 3.9422518e-01
3.8462922e-01 3.1626815e-01 1.3371065e-01 2.4147181e-01
-1.8761092e-01 -9.4582701e-01 6.0789359e-01 1.1340520e+00
3.7921101e-01 -6.4107203e-01 -4.2112088e-01 7.7962214e-01
-1.6092811e+00 -1.3705715e+00 -5.6818473e-01 2.0086925e+00
1.3077214e+00 4.7296381e-01 -6.9013305e-02 1.9989952e-01
7.0984191e-01 4.8014373e-01 -7.3313765e-02 -8.6428308e-01
-3.0206108e-01 -2.8574863e-01 3.2883525e-01 7.2719824e-01
-7.8327507e-01 7.3564130e-01 6.2019358e+00 7.9976988e-01
-9.5490909e-01 -4.4986546e-02 5.5775750e-01 -6.9005951e-02
-5.1150203e-01 -4.4063635e-02 -9.4237459e-01 6.4667785e-01
1.2442242e+00 -9.2884195e-01 -2.9418409e-01 7.9663658e-01
6.5444219e-01 -3.1027979e-01 -1.4978361e+00 4.1783333e-01
6.9873804e-01 -1.7030621e+00 5.0115085e-01 -1.9835156e-01
8.2581300e-01 -2.9216596e-01 -2.3351054e-01 3.2709974e-01
2.2962549e-01 -9.4447017e-01 9.6245635e-01 4.8531228e-01
4.1338170e-01 -6.1001796e-01 1.1649921e+00 6.8064415e-01
-7.0786647e-02 5.2602714e-01 -4.0842500e-01 -1.7439081e-01
2.0482153e-01 9.6856940e-01 -1.3309801e+00 3.7678581e-01
8.7316826e-02 7.9514408e-01 -8.9708424e-01 7.5597131e-01
-6.9527566e-01 9.4381148e-01 2.1825349e-01 -4.1465813e-01
2.6294386e-01 1.7398992e-01 8.3513397e-01 1.5364358e+00
1.4247851e-01 -1.0554847e-01 -3.1676137e-01 1.1555041e+00
-4.1562283e-01 -4.7561027e-02 -8.7452424e-01 -3.3646810e-01
3.9845228e-01 9.6749157e-01 -1.9972864e-01 -7.2137702e-01
-9.3247116e-02 9.9676824e-01 2.7937806e-01 6.7311630e-02
-8.1341374e-01 -3.7405711e-01 -5.5589147e-02 1.9976249e-01
-3.9816819e-02 3.5876569e-01 -7.5969672e-01 -1.2601600e+00
6.8688884e-02 -1.0344038e+00 2.0483068e-01 -1.1034801e+00
-1.1577528e+00 7.0957601e-01 -4.0057972e-02 -1.1806357e+00
-4.9436679e-01 7.5503953e-02 -1.0379803e+00 6.9441605e-01
-1.2952263e+00 -6.2205225e-01 -1.8615872e-01 7.6580718e-03
9.4198859e-01 1.3018399e-01 6.2592268e-01 -3.1335935e-01
-6.1976254e-01 6.4688814e-01 -3.6565113e-01 6.8175554e-02
9.7962219e-01 -1.3141100e+00 3.6043793e-01 1.1743159e+00
1.7255956e-01 8.3666766e-01 1.5724283e+00 -1.0249381e+00
-9.2327756e-01 -1.1434536e+00 1.6713119e+00 -7.1407151e-01
8.8887841e-01 2.1560951e-01 -1.1078914e+00 6.9962013e-01
9.6227801e-01 -8.7456512e-01 8.3582759e-01 -2.2322945e-02
-4.0737310e-01 3.8591355e-01 -8.3143008e-01 6.8958724e-01
3.0383545e-01 -5.5090034e-01 -1.5209339e+00 6.4025742e-01
8.9404750e-01 -2.8763098e-01 -4.6312857e-01 7.6825209e-02
1.4888808e-01 -8.5913044e-01 1.8408448e-01 -8.0930996e-01
1.6278732e+00 -1.2602016e-01 5.1420890e-02 -1.7170906e+00
-1.6705446e-01 -5.8228910e-01 -3.4377995e-01 1.4762504e+00
7.9769158e-01 -8.3284900e-03 3.8176751e-01 6.0293370e-01
-5.1389432e-01 -3.6724252e-01 -5.6460363e-01 -6.5183455e-01
2.1292001e-01 -1.3944118e-02 2.9891884e-01 8.1902039e-01
6.7768091e-01 1.1885968e+00 -4.9721903e-01 -1.2173003e-01
4.8425880e-01 1.5705726e-01 7.7608299e-01 -6.9247293e-01
-5.8974132e-02 -2.7689674e-01 5.7364933e-02 -9.3240148e-01
2.2406009e-01 -8.4488225e-01 2.8939846e-01 -1.6918318e+00
5.6976187e-01 5.3193879e-01 3.4286585e-01 2.0648247e-01
-3.9558142e-01 -1.5192805e-01 2.7970783e-02 2.9462174e-01
-7.9181558e-01 5.5878305e-01 1.2516358e+00 -2.5244394e-01
5.9882540e-02 -2.8746955e-02 -1.1534439e+00 7.0420444e-01
6.4673436e-01 -5.1521719e-01 -3.2360107e-01 -2.9473877e-01
3.1034011e-01 4.4769056e-02 3.5740659e-01 -7.9956609e-01
3.5463598e-01 -2.4946444e-02 -7.0335515e-02 -9.4594842e-01
-2.0885134e-01 -2.7337241e-01 -2.1494752e-01 2.7818924e-01
-1.0536931e+00 9.8359466e-02 2.9180985e-02 3.5525358e-01
-4.2712876e-01 -7.5444007e-01 4.9497002e-01 -2.2117183e-01
-8.0696493e-02 -4.2104003e-01 -6.4600617e-01 6.4973819e-01
5.2612805e-01 -1.5474162e-01 -1.0396143e+00 -7.4737263e-01
-1.5295435e-01 1.4432218e-02 5.2155161e-01 2.9283822e-01
6.1032170e-01 -8.5710454e-01 -1.1304440e+00 -2.8253868e-01
3.4185791e-01 -3.9152247e-01 -2.5547236e-02 7.6186365e-01
-3.5348701e-01 7.8688592e-01 1.9185450e-02 -1.9437382e-01
-1.0556254e+00 4.8972988e-01 6.9699064e-02 -4.0393054e-01
-5.4269218e-01 5.6853938e-01 -1.0451859e-01 2.7506721e-01
1.9108020e-01 -4.8669377e-01 -2.8562486e-01 2.7437329e-01
8.7767196e-01 6.9905204e-01 3.3535257e-01 -6.8694073e-01
3.1614732e-02 -6.8033852e-02 -5.3690469e-01 1.5611004e-02
1.0303576e+00 -2.1985528e-01 -4.5785341e-02 7.4730897e-01
1.3934435e+00 3.8259065e-01 -8.0026877e-01 1.4294720e-01
1.5476194e-01 -2.2893542e-01 1.4315842e-01 -1.0257961e+00
-4.0385562e-01 8.1253994e-01 -5.5101967e-01 6.1350000e-01
5.9168947e-01 6.9418818e-02 9.1986060e-01 4.6478337e-01
-2.3170359e-02 -1.1292671e+00 1.2048389e-01 7.6068449e-01
1.4369187e+00 -1.2047338e+00 3.6246786e-01 -3.8366541e-01
-1.0402938e+00 1.0115499e+00 4.3469411e-01 -1.6909170e-01
-4.9076423e-02 1.1719735e-01 -5.3792585e-02 -4.2064574e-01
-1.0785611e+00 3.6943945e-01 5.2471679e-01 2.1199054e-01
7.8921777e-01 9.4218194e-02 -5.8631158e-01 8.7082779e-01
-1.0304868e+00 -1.3992435e-01 1.5076981e+00 5.6496239e-01
-6.6190290e-01 -2.3662373e-01 -2.8668588e-01 8.0091715e-01
-6.4519268e-01 -3.5543692e-01 -9.7441673e-01 4.9920326e-01
-4.2568868e-01 1.3664031e+00 -1.2930351e-01 -4.4689596e-01
4.0109891e-01 9.6188463e-02 1.7315656e-01 -8.5800791e-01
-8.7457228e-01 -1.4675876e-01 7.6486075e-01 -3.0239052e-01
-1.9298051e-01 -4.2962009e-01 -1.0861580e+00 -5.8995903e-01
-5.7280195e-01 6.0713518e-01 3.4180421e-01 1.1188025e+00
3.3613324e-01 5.3088564e-01 6.3406932e-01 -2.9697973e-01
-9.5129621e-01 -1.3190725e+00 -3.5150704e-01 6.7654085e-01
6.9200766e-01 -2.4494323e-01 -7.7773589e-01 3.7600768e-01]
|
[12.237347602844238, 9.330570220947266]
|
2f26d466-45ca-4147-814f-4b5002d6a15c
|
evolin-benchmark-evaluation-of-line-detection
|
2303.05162
| null |
https://arxiv.org/abs/2303.05162v1
|
https://arxiv.org/pdf/2303.05162v1.pdf
|
EVOLIN Benchmark: Evaluation of Line Detection and Association
|
Lines are interesting geometrical features commonly seen in indoor and urban environments. There is missing a complete benchmark where one can evaluate lines from a sequential stream of images in all its stages: Line detection, Line Association and Pose error. To do so, we present a complete and exhaustive benchmark for visual lines in a SLAM front-end, both for RGB and RGBD, by providing a plethora of complementary metrics. We have also labelled data from well-known SLAM datasets in order to have all in one poses and accurately annotated lines. In particular, we have evaluated 17 line detection algorithms, 5 line associations methods and the resultant pose error for aligning a pair of frames with several combinations of detector-association. We have packaged all methods and evaluations metrics and made them publicly available on web-page https://prime-slam.github.io/evolin/.
|
['Anastasiia Kornilova', 'Gonzalo Ferrer', 'Kirill Ivanov']
|
2023-03-09
| null | null | null | null |
['line-detection']
|
['computer-vision']
|
[-1.19185023e-01 -2.56527454e-01 2.84381032e-01 -6.23881161e-01
-6.62086248e-01 -7.91953981e-01 7.32868612e-01 3.38178456e-01
-5.50605893e-01 7.57644415e-01 -2.80869007e-01 -2.70252496e-01
-1.50541142e-01 -6.52146637e-01 -7.76333869e-01 -7.85973668e-02
-4.59236622e-01 8.16604197e-01 5.92620909e-01 -3.36070657e-01
1.88099056e-01 8.05950165e-01 -1.35528755e+00 -2.41267636e-01
5.57370961e-01 8.20030749e-01 1.41090841e-03 7.09687769e-01
1.01099811e-01 3.97577286e-01 -1.81373939e-01 -5.44012010e-01
5.26941359e-01 -3.23041290e-01 -4.98122126e-01 2.02445179e-01
9.00831342e-01 -1.31467745e-01 -4.38506991e-01 8.18582118e-01
7.24474132e-01 -3.89673449e-02 5.11055291e-01 -1.56163275e+00
3.97628605e-01 1.44278124e-01 -7.03216910e-01 3.32478099e-02
1.06529427e+00 2.22187787e-01 8.53415191e-01 -1.14462173e+00
1.00321257e+00 8.98726344e-01 1.09767282e+00 -7.35605136e-02
-9.65251088e-01 -3.53906125e-01 -3.73299778e-01 1.31835714e-01
-1.52410197e+00 -4.59723502e-01 5.38504004e-01 -6.27955377e-01
8.60900939e-01 3.72491986e-01 9.94967639e-01 8.20233524e-01
6.26759678e-02 5.41857541e-01 1.23217034e+00 -4.78820831e-01
6.95331767e-02 1.96401421e-02 4.82263742e-03 9.03378367e-01
5.62684357e-01 1.90326288e-01 -8.30118477e-01 2.23257333e-01
7.80718803e-01 -2.39012107e-01 -2.99802512e-01 -1.19313192e+00
-1.48637974e+00 4.67232943e-01 6.56064808e-01 9.86058637e-03
-2.69594807e-02 3.93174559e-01 1.91399321e-01 2.89734840e-01
3.30015272e-02 2.22631127e-01 -4.04432356e-01 -2.39636332e-01
-8.57452869e-01 2.26821736e-01 7.80456781e-01 1.47215164e+00
1.27670741e+00 -5.45960605e-01 3.18976104e-01 3.61034572e-01
3.94473374e-01 8.43783915e-01 -3.93140800e-02 -7.46215224e-01
5.37755311e-01 4.51164633e-01 3.20782274e-01 -9.83799577e-01
-9.81311083e-01 -4.15749580e-01 -5.60627997e-01 2.90861845e-01
5.64617932e-01 -1.00100458e-01 -6.02549791e-01 9.02374625e-01
4.79199290e-01 -1.57793164e-01 -2.22964719e-01 8.93883646e-01
8.17525625e-01 1.75249800e-01 -7.31252432e-01 2.18389690e-01
1.20887399e+00 -1.10872006e+00 -5.05940080e-01 -5.21763742e-01
7.56590188e-01 -1.31094205e+00 8.59299362e-01 4.98908013e-01
-7.13503838e-01 -4.26438719e-01 -1.32923639e+00 -2.96150088e-01
-5.25362730e-01 3.26277584e-01 6.82905257e-01 4.11681890e-01
-9.18269515e-01 5.56239486e-01 -7.57608235e-01 -1.03464031e+00
3.55748758e-02 2.53633618e-01 -6.52900636e-01 -6.33487999e-02
-6.84122860e-01 9.03746367e-01 2.53789783e-01 1.84637591e-01
-3.19794387e-01 -3.65006089e-01 -9.91279542e-01 -6.36062860e-01
3.63328516e-01 -8.65233541e-01 1.16800690e+00 -4.96779293e-01
-1.10907888e+00 1.12944627e+00 -2.42515672e-02 -5.00975609e-01
1.28517425e+00 -6.48170710e-01 -2.40762487e-01 -1.00730598e-01
1.93424687e-01 6.57038510e-01 3.63966286e-01 -1.40651250e+00
-7.71929979e-01 -2.63348371e-01 -2.00436771e-01 3.13814461e-01
5.01621544e-01 -3.13670844e-01 -7.16688871e-01 -8.23744535e-02
6.72146797e-01 -9.87494886e-01 -2.81132162e-01 2.68478602e-01
-7.20948100e-01 3.20399195e-01 6.67415559e-01 -4.16989923e-01
6.94567323e-01 -1.90536106e+00 -1.42385900e-01 4.92083251e-01
-2.44220980e-02 -2.11010352e-01 2.19032452e-01 8.57507288e-01
2.13570356e-01 -2.03274131e-01 -2.06210315e-01 -7.05177784e-01
1.18704118e-01 1.35739237e-01 -5.91977797e-02 1.09950566e+00
-2.56676644e-01 7.43346334e-01 -9.89008248e-01 -6.43304288e-01
8.59609008e-01 4.52609181e-01 -1.55351266e-01 4.65200208e-02
-1.01845879e-02 6.14923060e-01 -3.43245529e-02 7.10506558e-01
8.85726869e-01 1.55096889e-01 -1.12202048e-01 -2.50366062e-01
-4.95581865e-01 1.10829853e-01 -1.57996082e+00 2.15311503e+00
-2.09717602e-01 1.03528392e+00 -4.72865671e-01 1.87233146e-02
1.13640785e+00 -1.10821821e-01 5.64504862e-01 -7.57052600e-01
9.60166231e-02 6.22129560e-01 -2.81284064e-01 -1.09591089e-01
8.31681669e-01 5.80500305e-01 -1.39361367e-01 -7.98099712e-02
1.60637096e-01 -7.66768157e-01 4.05868322e-01 4.17417698e-02
1.15505767e+00 6.33873761e-01 6.94554985e-01 -1.89989582e-01
5.95253229e-01 4.30634886e-01 1.70332551e-01 7.45198786e-01
-1.34726971e-01 8.54295075e-01 1.26642257e-01 -6.35438502e-01
-1.44204378e+00 -1.26987147e+00 -2.12086037e-01 3.89734566e-01
5.17685056e-01 -7.23967016e-01 -2.87986785e-01 -2.83336699e-01
2.18594924e-01 1.80058509e-01 -3.24020863e-01 5.52562594e-01
-3.09271008e-01 -9.76636037e-02 4.51201648e-01 2.32415095e-01
5.37099779e-01 -5.57335734e-01 -9.39719439e-01 -6.02741465e-02
1.24162935e-01 -1.33051753e+00 -2.16244489e-01 4.93059695e-01
-5.18098891e-01 -1.37780058e+00 -3.32695335e-01 -3.96407753e-01
5.60093701e-01 3.80477518e-01 1.22768807e+00 5.69624566e-02
-4.01922494e-01 6.21103585e-01 -4.61920589e-01 -4.79806811e-01
1.14891917e-01 1.21638127e-01 1.25247434e-01 -4.16786224e-01
5.02535738e-02 -2.82999843e-01 -6.46013021e-01 4.57452834e-01
-2.85542309e-01 2.67381519e-01 4.22214866e-01 3.29786003e-01
7.02121258e-01 -4.14664835e-01 -6.43547952e-01 -5.02412200e-01
4.13764939e-02 -7.51914531e-02 -1.05042493e+00 6.83958605e-02
-1.81206927e-01 -2.24589854e-01 5.31500615e-02 5.10246575e-01
-5.13748348e-01 6.96246743e-01 -2.94942707e-01 -9.52912867e-02
-3.76780599e-01 1.20794117e-01 7.48600066e-02 -4.52903360e-01
8.70238483e-01 5.07671805e-03 -3.30157548e-01 -1.69854373e-01
7.13005781e-01 3.31360579e-01 6.22953355e-01 -3.65288019e-01
1.33022738e+00 8.22166920e-01 3.13802779e-01 -1.14557409e+00
-6.29378259e-01 -1.20760787e+00 -1.33260465e+00 -5.29038548e-01
5.72781384e-01 -8.14541221e-01 -4.80754614e-01 5.82752764e-01
-1.10540867e+00 -4.77120280e-01 -2.89412647e-01 5.38770080e-01
-9.59798157e-01 5.35180032e-01 -2.13288739e-01 -7.46649802e-01
-2.73044910e-02 -1.06061113e+00 1.24066794e+00 2.87964910e-01
-1.50146320e-01 -9.13056433e-01 4.91329968e-01 -7.20957965e-02
-5.15022054e-02 6.25761151e-01 -2.15829507e-01 -3.09137762e-01
-7.37723827e-01 -2.29245260e-01 -1.67378306e-01 -2.80646235e-01
-5.79520278e-02 3.20813924e-01 -9.56061661e-01 -2.54028946e-01
-7.32247889e-01 -2.19653919e-01 7.22818613e-01 1.63694266e-02
3.21244121e-01 1.18081890e-01 -5.39171219e-01 9.90519881e-01
1.82743299e+00 -8.00261050e-02 6.29377067e-01 1.08258104e+00
6.36787176e-01 5.00223041e-01 1.04861701e+00 5.93306541e-01
4.48877454e-01 9.55571234e-01 5.27356565e-01 -3.45450312e-01
-8.06494653e-02 -3.20414603e-01 1.57230183e-01 4.78025228e-01
-1.78861201e-01 -2.33086661e-01 -1.40034223e+00 1.74210086e-01
-1.93191051e+00 -7.62708724e-01 -1.05597317e+00 2.33443213e+00
3.02018762e-01 2.49885768e-01 1.04125164e-01 2.29316875e-01
3.04333836e-01 9.01085734e-02 -4.21094187e-02 -7.05769658e-02
-2.88079709e-01 -7.24315569e-02 1.16242480e+00 7.29820371e-01
-1.37194383e+00 1.02358305e+00 5.83923006e+00 3.51518333e-01
-8.90735865e-01 -1.11481905e-01 -7.94824585e-02 3.46323878e-01
-1.90692380e-01 3.56016129e-01 -8.51749718e-01 4.22555096e-02
5.84131896e-01 2.16973126e-01 -6.21811263e-02 8.60570073e-01
1.77853242e-01 -7.36418247e-01 -1.10343933e+00 1.23648608e+00
-2.07394995e-02 -1.19370770e+00 -5.57226121e-01 2.27283370e-02
7.49126852e-01 6.28437996e-01 -3.95853609e-01 -2.35472411e-01
3.08017850e-01 -6.77316487e-01 1.09879923e+00 6.72927082e-01
7.54504800e-01 -4.60638463e-01 7.10548043e-01 4.92175668e-02
-1.42215264e+00 3.68056864e-01 -2.02260688e-01 -1.22321300e-01
3.17847371e-01 6.09735906e-01 -1.19354212e+00 1.13997436e+00
6.70273781e-01 7.87643909e-01 -1.10908353e+00 1.84975505e+00
-4.52747673e-01 -9.51404311e-03 -7.97491074e-01 1.28159732e-01
2.51293510e-01 -5.35570681e-01 3.33026439e-01 1.39412153e+00
5.47832191e-01 -5.57390273e-01 2.21022695e-01 3.21297497e-01
4.10420448e-01 2.13338315e-01 -9.48517025e-01 6.51539803e-01
5.28568387e-01 1.42120159e+00 -1.02994788e+00 8.26430470e-02
-4.60485548e-01 1.14979064e+00 1.31054893e-01 9.08427238e-02
-8.88986826e-01 -2.70767748e-01 5.66750348e-01 2.67811209e-01
-3.89482528e-02 -1.02358830e+00 -1.54960796e-01 -1.02325988e+00
1.07246272e-01 -2.24221379e-01 2.22371981e-01 -1.21373737e+00
-8.18212271e-01 4.66557324e-01 -1.56728879e-01 -1.60895717e+00
-1.54584393e-01 -6.89967513e-01 -1.64192185e-01 3.52533281e-01
-1.53995609e+00 -1.36081135e+00 -1.08241153e+00 6.66243136e-01
4.22879010e-01 3.35479885e-01 4.88423735e-01 3.60814482e-01
-2.35125035e-01 1.65929258e-01 3.05784017e-01 1.93588525e-01
1.01766753e+00 -1.46969450e+00 6.70161664e-01 1.11820829e+00
6.13790929e-01 2.16760546e-01 1.10233963e+00 -5.82336485e-01
-1.42807221e+00 -7.92739630e-01 7.80615687e-01 -7.20147908e-01
7.21477926e-01 -5.76347411e-01 -2.45899200e-01 1.04007983e+00
9.47251469e-02 1.10409163e-01 2.85871267e-01 -2.30504647e-01
-1.15188114e-01 -2.23555356e-01 -7.47137845e-01 3.37226689e-01
1.39488018e+00 -2.47027680e-01 -2.27081388e-01 5.98778069e-01
2.25511119e-01 -1.02201629e+00 -7.77647853e-01 3.41049105e-01
4.82298464e-01 -1.61462975e+00 1.11735630e+00 3.05742770e-01
-1.74408764e-01 -8.23862851e-01 -2.21046820e-01 -1.10226142e+00
7.55288005e-02 -5.75673759e-01 3.95774722e-01 1.27804470e+00
2.85678387e-01 -6.90549731e-01 6.31144643e-01 -2.15807900e-01
-4.22785312e-01 -3.32188219e-01 -9.59572732e-01 -8.70001674e-01
-6.22471452e-01 -7.47008920e-01 4.17093426e-01 6.38718843e-01
-1.95265263e-01 1.14559336e-03 -4.31983054e-01 3.59741628e-01
9.47128713e-01 2.71490552e-02 1.55579829e+00 -1.03052866e+00
1.85319841e-01 -3.89835477e-01 -9.61392045e-01 -9.24267828e-01
-4.48769718e-01 -6.86981678e-01 2.17021748e-01 -1.88961220e+00
-2.66166329e-01 -5.91853261e-01 3.93978298e-01 2.77304411e-01
3.23635787e-01 5.40279090e-01 1.83441713e-01 4.47849452e-01
-8.86669517e-01 2.66596645e-01 7.08620548e-01 1.39073014e-01
1.00480124e-01 -1.67616978e-01 4.20911282e-01 9.71141517e-01
8.78961325e-01 -2.97891736e-01 3.24331373e-02 -2.82943040e-01
5.59789479e-01 -9.01052132e-02 6.55563712e-01 -1.81598675e+00
4.94569093e-01 4.94662561e-02 5.78676879e-01 -1.13471186e+00
5.79355478e-01 -1.02551925e+00 4.68949556e-01 6.77973628e-01
2.88122356e-01 2.32597470e-01 1.06574886e-01 1.50653630e-01
-6.77183867e-02 -8.91089886e-02 5.63701570e-01 -7.38444626e-02
-1.60057950e+00 2.93462306e-01 1.00289531e-01 -2.72695690e-01
1.35232890e+00 -6.44162238e-01 -4.29567039e-01 -3.10125053e-01
-5.16393602e-01 3.32736433e-01 1.19883609e+00 3.27300608e-01
6.52104437e-01 -1.30952716e+00 -5.13791203e-01 2.22618371e-01
7.02289224e-01 1.80578291e-01 -6.54284805e-02 1.11987615e+00
-1.50529289e+00 5.49374938e-01 -5.28073549e-01 -1.07096219e+00
-1.15989006e+00 3.56670231e-01 5.98201632e-01 1.62185609e-01
-5.39775372e-01 6.74856007e-01 -4.19830322e-01 -6.73387468e-01
1.29668131e-01 -3.67798448e-01 2.93053776e-01 1.27504662e-01
-1.67876221e-02 5.55484891e-01 3.38315040e-01 -8.30365837e-01
-7.99625874e-01 1.12311924e+00 7.69125938e-01 -2.84546435e-01
1.14817679e+00 -3.92518848e-01 -3.22978944e-02 6.15856230e-01
9.21060979e-01 5.18257916e-01 -1.31013644e+00 5.64182252e-02
3.29741925e-01 -6.84430599e-01 -4.70268220e-01 -4.42892581e-01
-5.16624033e-01 6.28731489e-01 7.22583890e-01 -5.06796986e-02
7.47514606e-01 3.59770246e-02 2.73120493e-01 4.74795938e-01
1.11479986e+00 -8.52175832e-01 -3.56576368e-02 6.09808505e-01
8.25441480e-01 -1.39041340e+00 4.25504327e-01 -7.11576998e-01
-3.37657243e-01 1.46246409e+00 4.49630320e-01 -2.10993752e-01
2.09431976e-01 5.22941351e-01 1.97205007e-01 -3.34814876e-01
1.18988799e-02 -6.54698133e-01 2.34575450e-01 8.55892956e-01
4.83444452e-01 1.00425482e-02 -2.06765175e-01 -4.62642103e-01
-6.96186781e-01 -1.36126667e-01 6.25936031e-01 1.04904366e+00
-5.47121823e-01 -1.20272887e+00 -6.59833908e-01 1.89382471e-02
3.46847415e-01 8.34087804e-02 -4.99994427e-01 1.33425248e+00
8.98480713e-02 5.92068136e-01 -7.20889354e-03 -4.73570943e-01
6.05202794e-01 -2.51182854e-01 7.64140844e-01 -2.61303008e-01
-2.07180768e-01 -2.12264001e-01 3.74279171e-01 -1.00887012e+00
-4.71574515e-01 -1.04414189e+00 -1.14342058e+00 -3.45212460e-01
-2.77174115e-01 -7.20456615e-02 9.09427345e-01 4.80765074e-01
-3.55223380e-02 1.39842853e-01 2.48861387e-01 -1.11298335e+00
-4.83074039e-02 -6.71130836e-01 -6.35462224e-01 3.46564144e-01
2.74041325e-01 -6.15299225e-01 -1.68887749e-01 -2.25416824e-01]
|
[7.399188995361328, -2.207569122314453]
|
17eb4dcb-9d71-4657-a5da-f5306912943a
|
toward-multilingual-identification-of-online
| null | null |
https://aclanthology.org/W19-6130
|
https://aclanthology.org/W19-6130.pdf
|
Toward Multilingual Identification of Online Registers
|
We consider cross- and multilingual text classification approaches to the identification of online registers (genres), i.e. text varieties with specific situational characteristics. Register is the most important predictor of linguistic variation, and register information could improve the potential of online data for many applications. We introduce the first manually annotated non-English corpus of online registers featuring the full range of linguistic variation found online. The data set consists of 2,237 Finnish documents and follows the register taxonomy developed for the Corpus of Online Registers of English (CORE). Using CORE and the newly introduced corpus, we demonstrate the feasibility of cross-lingual register identification using a simple approach based on convolutional neural networks and multilingual word embeddings. We further find that register identification results can be improved through multilingual training even when a substantial number of annotations is available in the target language.
|
['Sampo Pyysalo', 'Douglas Biber', 'Jesse Egbert', 'Roosa Kyllönen', 'Veronika Laippala']
| null | null | null | null |
ws-nodalida-2019-9
|
['multilingual-word-embeddings', 'multilingual-text-classification']
|
['methodology', 'miscellaneous']
|
[-3.05450559e-01 -2.97516654e-03 -7.85368800e-01 -2.33654991e-01
-1.11795914e+00 -1.01163626e+00 8.48926485e-01 4.67760116e-01
-6.98528290e-01 3.14196199e-01 8.14555228e-01 -5.04818499e-01
-1.14251584e-01 -4.76147771e-01 -4.06796306e-01 -1.10874467e-01
-1.42137870e-01 6.75810158e-01 -1.77585497e-01 -7.70331919e-01
7.94128180e-02 3.32462400e-01 -1.38028216e+00 1.52970389e-01
7.14630723e-01 6.43133461e-01 2.62015104e-01 2.36559004e-01
-4.37601924e-01 3.91941547e-01 -4.70176578e-01 -8.73350739e-01
6.86056614e-02 6.77939728e-02 -8.32033992e-01 -2.10639760e-01
5.65005362e-01 1.09007388e-01 -4.87569332e-01 1.06675327e+00
4.96768385e-01 -3.39079380e-01 4.95156586e-01 -2.95918584e-01
-7.77869642e-01 1.40892017e+00 -3.06522489e-01 4.91271615e-01
5.64103425e-01 -2.54625291e-01 1.50047624e+00 -5.61791420e-01
1.23923922e+00 1.22237551e+00 1.09972095e+00 1.38660967e-01
-1.40062726e+00 -5.34549892e-01 -1.05591439e-01 -1.36730984e-01
-1.65784085e+00 -7.89471447e-01 7.11742401e-01 -7.55288422e-01
1.06167126e+00 5.75040989e-02 4.84156549e-01 9.83008623e-01
1.80420652e-01 4.90116328e-01 9.03524637e-01 -5.91752112e-01
-4.73912269e-01 3.93578112e-01 3.47894430e-01 5.92602670e-01
3.73250753e-01 -6.52663782e-02 -5.98385632e-01 -1.82531655e-01
4.79361773e-01 -5.13133407e-01 7.43880272e-02 3.32442415e-03
-1.35770202e+00 9.85692739e-01 -1.20901406e-01 8.69271219e-01
1.66406572e-01 -3.53185922e-01 9.56522942e-01 7.59715974e-01
9.95376408e-01 5.53848326e-01 -8.80031347e-01 -3.01292300e-01
-7.43154109e-01 -8.41234475e-02 7.21004248e-01 8.34488332e-01
6.97443485e-01 -1.20686732e-01 2.68297374e-01 1.41858089e+00
2.06109196e-01 4.86339211e-01 9.28259552e-01 -3.24504793e-01
8.27054560e-01 8.35189879e-01 -2.65443087e-01 -5.65144360e-01
-7.34481335e-01 -1.19886622e-01 -1.84563726e-01 -4.65311885e-01
6.48038924e-01 -3.30220133e-01 -3.79094779e-01 1.42309308e+00
1.52913630e-01 -4.59636509e-01 2.48472571e-01 5.39904982e-02
1.08927155e+00 3.07019919e-01 -8.71948823e-02 -6.19461760e-02
1.52536654e+00 -2.44868383e-01 -9.28312242e-01 -1.12121895e-01
1.15952706e+00 -1.00984383e+00 9.66147065e-01 -1.44318596e-01
-8.53074551e-01 -5.54810166e-01 -1.05365109e+00 -3.67919087e-01
-8.18378448e-01 2.79107243e-01 8.04046631e-01 1.08523226e+00
-1.00057721e+00 4.03258026e-01 -6.23126090e-01 -3.61553013e-01
1.84228793e-01 5.65696359e-01 -8.02765965e-01 2.08114818e-01
-1.24684393e+00 8.82838309e-01 4.43319529e-01 -1.86323211e-01
1.92916282e-02 -6.36620164e-01 -1.38586867e+00 -5.79628408e-01
-5.59806153e-02 6.32989287e-01 1.14608359e+00 -1.02538729e+00
-1.24418783e+00 1.42741096e+00 -1.22164816e-01 -2.39234343e-01
1.64866120e-01 1.48550078e-01 -8.13814402e-01 -2.31121600e-01
2.70113915e-01 8.98239911e-02 1.60545543e-01 -7.07098305e-01
-7.42359519e-01 -3.78230602e-01 -1.19977206e-01 -2.05518818e-03
-5.65483809e-01 7.18235195e-01 -3.12778592e-01 -6.88750088e-01
-5.11500500e-02 -8.82597446e-01 1.92805097e-01 -7.84708083e-01
-2.70160418e-02 -5.66713929e-01 1.96361721e-01 -1.42524683e+00
1.36366570e+00 -2.34442186e+00 -8.10250938e-02 1.26139149e-01
-3.17373537e-02 -3.45513714e-03 -1.13142833e-01 4.71970320e-01
-5.41869551e-02 4.79833513e-01 4.11936194e-01 -4.36792672e-01
2.33938128e-01 1.27748147e-01 8.28811005e-02 8.15012217e-01
3.81657988e-01 1.01752937e+00 -8.62992465e-01 -3.96534145e-01
-9.38987210e-02 3.06817144e-01 -4.54657406e-01 -3.39624137e-01
1.33175552e-01 3.02300513e-01 -1.02734357e-01 7.95969665e-01
2.57819563e-01 5.37927508e-01 6.35870218e-01 -3.36576849e-02
-5.01551449e-01 8.18128109e-01 -9.14943635e-01 1.54345095e+00
-9.66852784e-01 1.21880221e+00 1.22425025e-02 -6.09391749e-01
1.12546492e+00 3.33647966e-01 4.41769898e-01 -7.44823456e-01
3.19726944e-01 4.45379913e-01 4.38185483e-01 -5.15790880e-01
1.17506146e+00 1.60063684e-01 -8.39219630e-01 3.65621775e-01
4.02852386e-01 2.61541396e-01 3.45326275e-01 -2.83936828e-01
8.69559526e-01 1.27621721e-02 4.75631356e-01 -5.67631721e-01
3.53699982e-01 2.27668121e-01 5.37901223e-01 2.62747258e-01
-2.03437805e-01 1.09969959e-01 5.21076441e-01 -4.81976897e-01
-1.14862490e+00 -8.62629831e-01 -1.07970679e+00 1.51059389e+00
-3.31956387e-01 -5.38955629e-01 -4.13710564e-01 -3.92232865e-01
9.76888388e-02 2.38230512e-01 -7.05389142e-01 3.49372476e-01
-8.38099658e-01 -8.59128833e-01 9.25894797e-01 2.47119948e-01
-3.94910350e-02 -1.01860142e+00 3.62188041e-01 1.92696765e-01
-1.31870344e-01 -1.32479131e+00 -5.69701850e-01 2.29619622e-01
-4.81970131e-01 -1.04945827e+00 -1.92094535e-01 -1.21585631e+00
2.19676301e-01 -3.21598440e-01 1.21822381e+00 -1.29689053e-01
-7.09230527e-02 3.72225165e-01 -6.01169646e-01 -1.57403976e-01
-1.05944049e+00 7.87513494e-01 1.78308889e-01 -1.06989019e-01
9.08774614e-01 5.29445261e-02 3.83974373e-01 4.05112235e-03
-4.49837446e-01 -7.16769099e-01 3.35650235e-01 8.47101808e-01
3.44230533e-01 -1.84947789e-01 7.19234884e-01 -1.12669754e+00
5.78814209e-01 -5.16980886e-01 -5.62166691e-01 7.66105205e-02
-1.83636963e-01 1.94021761e-01 2.98920870e-01 -6.60459697e-01
-8.46016109e-01 1.44639254e-01 -4.54347432e-01 3.79053652e-01
-1.20355792e-01 9.25307631e-01 -2.99927056e-01 -1.62772208e-01
6.18937492e-01 -7.34353140e-02 -9.12964270e-02 -6.11989737e-01
3.51401955e-01 1.34360754e+00 5.16400576e-01 -4.75055158e-01
5.38769364e-01 -2.27649277e-03 -5.88805079e-01 -1.06940961e+00
-3.82387996e-01 -6.58101380e-01 -1.23123598e+00 1.67439284e-03
9.66515303e-01 -1.26151025e+00 -5.62249541e-01 1.71777442e-01
-9.64826465e-01 -4.67339903e-01 -1.99476838e-01 3.87620479e-01
-2.79510856e-01 2.04316854e-01 -8.26505244e-01 -6.78765655e-01
-5.86836925e-03 -1.15641499e+00 1.24777472e+00 -3.09690773e-01
-5.52845657e-01 -1.57103467e+00 3.96002024e-01 2.60394067e-01
1.00739403e-02 2.41950229e-01 9.57875609e-01 -1.06859887e+00
5.97306862e-02 -5.07244825e-01 1.55994773e-01 -1.43554270e-01
2.66121894e-01 -4.50358763e-02 -1.03803575e+00 -1.93794519e-01
-4.63253766e-01 -1.58464625e-01 6.77265704e-01 9.02480558e-02
2.89248943e-01 2.50600521e-02 -2.35895440e-01 5.90019464e-01
1.36216128e+00 7.96632469e-02 4.41256672e-01 7.36415088e-01
8.56100380e-01 7.55249977e-01 2.82844365e-01 1.44122526e-01
7.78814614e-01 8.21340859e-01 -3.85532469e-01 9.86222103e-02
-1.15399193e-02 -2.77103990e-01 8.65520716e-01 1.50159419e+00
-1.21722579e-01 4.27572817e-01 -1.08103693e+00 9.46821272e-01
-1.37257707e+00 -9.33892190e-01 -1.84257135e-01 2.35168505e+00
1.33413219e+00 -3.68773751e-02 3.84243309e-01 1.16825081e-01
8.79020691e-01 1.34872213e-01 -5.26627600e-02 -7.23830819e-01
-4.28389072e-01 6.18971646e-01 1.06196356e+00 5.67384660e-01
-1.38904965e+00 1.30460393e+00 6.99202204e+00 9.56247628e-01
-9.02781665e-01 4.10533816e-01 2.95206606e-01 2.08071932e-01
-1.61555022e-01 -3.13502848e-01 -1.38299298e+00 3.69759232e-01
1.48047221e+00 -1.50869370e-01 4.61518139e-01 5.86931825e-01
-2.36302726e-02 2.69413143e-01 -1.04103768e+00 8.98009896e-01
1.55571222e-01 -1.26010275e+00 -4.57693577e-01 3.60220462e-01
7.80093551e-01 5.55459082e-01 -2.29665497e-03 4.24569279e-01
4.65062171e-01 -8.89270544e-01 8.67712677e-01 -2.37817578e-02
1.38982093e+00 -9.01275456e-01 1.08513641e+00 -2.01736942e-01
-1.38809443e+00 -2.41273586e-02 -4.88852292e-01 1.54137075e-01
-8.11628103e-02 2.66913921e-02 -6.66251957e-01 5.15482247e-01
2.76844770e-01 1.05061126e+00 -1.00715971e+00 5.76436937e-01
1.35684058e-01 5.97846627e-01 -1.50597140e-01 -1.48982033e-01
-6.15782179e-02 -1.78139374e-01 3.39099675e-01 1.44789529e+00
2.33335316e-01 -7.98003912e-01 3.33105683e-01 1.81980953e-01
-2.39286110e-01 7.60943651e-01 -7.57551491e-01 -2.77189076e-01
6.62693202e-01 1.12636375e+00 -4.88814861e-01 -6.02500215e-02
-8.86762738e-01 5.48015177e-01 6.89139545e-01 -6.80698897e-04
-6.21164069e-02 -4.68628556e-01 6.99926555e-01 2.17292026e-01
3.08490783e-01 -5.31465650e-01 -2.00008705e-01 -1.10953474e+00
-1.06115043e-01 -9.95142579e-01 3.60476464e-01 1.23454995e-01
-1.49060106e+00 7.98097014e-01 -2.32270226e-01 -1.17782736e+00
-4.54715818e-01 -1.05703199e+00 -2.35679168e-02 9.42423284e-01
-1.40363348e+00 -1.50556171e+00 2.95920670e-01 3.77667546e-01
4.79495734e-01 -8.10187757e-01 1.00136161e+00 6.20421886e-01
-4.95436728e-01 1.12221205e+00 6.86443985e-01 7.41429448e-01
9.60003138e-01 -1.39611340e+00 5.89043379e-01 3.52706015e-01
5.46930909e-01 5.78041494e-01 2.66793013e-01 -8.09492230e-01
-1.46905529e+00 -1.02653623e+00 1.50835228e+00 -7.62522340e-01
1.64943326e+00 -8.41420889e-01 -5.64861059e-01 1.03212035e+00
2.99127609e-01 -4.72123742e-01 9.62458968e-01 7.74142623e-01
-3.30408633e-01 1.05604284e-01 -8.84908438e-01 6.42508686e-01
1.03983915e+00 -1.25961530e+00 -6.86829209e-01 4.58685935e-01
6.42868102e-01 -4.82090205e-01 -1.55256653e+00 -2.88231194e-01
6.60312593e-01 -3.16928446e-01 6.35011554e-01 -2.06445023e-01
1.46691516e-01 2.01443493e-01 -1.77701503e-01 -1.41280019e+00
-1.93375036e-01 -7.62236476e-01 6.61562085e-01 1.68942177e+00
8.29631388e-01 -8.34549487e-01 2.46160150e-01 2.03580052e-01
-4.04979348e-01 5.47284074e-02 -1.26496530e+00 -8.40827167e-01
5.89628220e-01 -6.32684946e-01 6.82530880e-01 1.41988099e+00
6.18660510e-01 3.68055791e-01 6.06626831e-03 -2.91909389e-02
-2.03386303e-02 -3.99135113e-01 6.64962173e-01 -1.61487985e+00
-1.46391407e-01 -6.96464121e-01 -8.69924426e-01 -4.32576925e-01
9.12310779e-01 -1.59215081e+00 -1.27910018e-01 -8.07788789e-01
-8.23203474e-02 -6.52730405e-01 1.83376223e-01 3.40145946e-01
-3.42170596e-02 4.38457280e-01 -1.53217152e-01 1.38226360e-01
-1.31430358e-01 -3.58783975e-02 6.98188961e-01 -1.40525356e-01
-6.03013933e-01 -2.43041098e-01 -6.49863124e-01 5.00160277e-01
8.08041871e-01 -2.05419943e-01 2.27389097e-01 -3.31248403e-01
4.86597180e-01 -3.20492595e-01 -3.48036170e-01 -4.18287188e-01
-1.65191412e-01 2.56210387e-01 1.70099631e-01 -3.26945275e-01
-3.57175395e-02 -5.40290713e-01 -7.76204839e-02 8.66229758e-02
-3.68714988e-01 4.59295928e-01 5.19903779e-01 2.60382980e-01
-3.19551647e-01 -4.97626901e-01 3.54638666e-01 1.20902948e-01
-6.98879361e-01 4.43321206e-02 -6.68348014e-01 2.03724444e-01
4.28045809e-01 -9.53850001e-02 -5.01800656e-01 1.65791120e-02
-5.61985612e-01 -2.09300756e-01 6.43820226e-01 6.49045706e-01
-3.15987200e-01 -1.50030804e+00 -7.73729205e-01 4.77341056e-01
5.61235487e-01 -7.13858068e-01 -2.97742724e-01 5.24186790e-01
-6.06570184e-01 3.03054988e-01 2.13247519e-02 -2.10110381e-01
-1.22991800e+00 3.52557808e-01 1.86963573e-01 -4.64647412e-01
-5.14527500e-01 2.37751648e-01 -4.69860256e-01 -8.01158488e-01
-1.98445003e-03 -2.67432243e-01 -8.85294974e-01 6.89161062e-01
4.70255017e-01 5.48659079e-02 2.87226081e-01 -1.63336909e+00
-1.95321500e-01 6.51563048e-01 -2.23467052e-01 -3.71975213e-01
1.46911430e+00 -2.01799944e-01 -1.97402850e-01 9.59489167e-01
1.30591094e+00 8.19901288e-01 -4.27711397e-01 -3.27802986e-01
6.37591362e-01 -9.39142182e-02 2.22476512e-01 -2.87589937e-01
-8.57876718e-01 4.29728746e-01 7.08581686e-01 3.50279987e-01
3.96771103e-01 3.40268791e-01 6.02683365e-01 2.34783128e-01
2.93240339e-01 -1.46679723e+00 -5.80186367e-01 9.63758409e-01
5.74557543e-01 -1.34341490e+00 -1.76353246e-01 -2.17385292e-01
-6.12251163e-01 1.11101329e+00 -5.50582670e-02 -9.27469581e-02
9.23905313e-01 5.08825123e-01 4.04884964e-01 -2.77585566e-01
-2.86429793e-01 -7.19645321e-01 4.63874906e-01 8.60704839e-01
1.11781228e+00 3.39519560e-01 -6.88759744e-01 8.11210990e-01
-8.32282245e-01 -6.89557970e-01 5.06702662e-01 4.25125569e-01
-3.95360813e-02 -1.62006378e+00 -3.21967900e-01 6.42394006e-01
-1.06304348e+00 -4.63560998e-01 -5.06081760e-01 1.11913884e+00
2.14763448e-01 9.05357897e-01 6.71149254e-01 -5.92271566e-01
6.75420463e-02 2.12704852e-01 3.47886115e-01 -7.92329669e-01
-1.06889880e+00 2.57877916e-01 8.16817701e-01 4.86899763e-02
-6.37958169e-01 -1.30870593e+00 -7.32959986e-01 -3.06444138e-01
-3.28807712e-01 -1.28170833e-01 7.27116406e-01 1.10067701e+00
-1.00332372e-01 2.22724020e-01 6.45707488e-01 -9.11963463e-01
-5.82896173e-02 -1.26905394e+00 -8.02166760e-01 4.37565655e-01
1.33235082e-01 -5.68868876e-01 -2.86288023e-01 2.79548675e-01]
|
[10.586483001708984, 10.052136421203613]
|
9a2ae520-ac29-48b2-af6c-5b6ea0e681f1
|
blind-video-temporal-consistency-via-deep
|
2010.11838
| null |
https://arxiv.org/abs/2010.11838v1
|
https://arxiv.org/pdf/2010.11838v1.pdf
|
Blind Video Temporal Consistency via Deep Video Prior
|
Applying image processing algorithms independently to each video frame often leads to temporal inconsistency in the resulting video. To address this issue, we present a novel and general approach for blind video temporal consistency. Our method is only trained on a pair of original and processed videos directly instead of a large dataset. Unlike most previous methods that enforce temporal consistency with optical flow, we show that temporal consistency can be achieved by training a convolutional network on a video with the Deep Video Prior. Moreover, a carefully designed iteratively reweighted training strategy is proposed to address the challenging multimodal inconsistency problem. We demonstrate the effectiveness of our approach on 7 computer vision tasks on videos. Extensive quantitative and perceptual experiments show that our approach obtains superior performance than state-of-the-art methods on blind video temporal consistency. Our source codes are publicly available at github.com/ChenyangLEI/deep-video-prior.
|
['Qifeng Chen', 'Yazhou Xing', 'Chenyang Lei']
|
2020-10-22
| null |
http://proceedings.neurips.cc/paper/2020/hash/0c0a7566915f4f24853fc4192689aa7e-Abstract.html
|
http://proceedings.neurips.cc/paper/2020/file/0c0a7566915f4f24853fc4192689aa7e-Paper.pdf
|
neurips-2020-12
|
['video-temporal-consistency']
|
['computer-vision']
|
[-1.53778434e-01 -5.41198134e-01 -2.66530842e-01 -2.29756206e-01
-7.81171560e-01 -5.24712086e-01 4.82796907e-01 -4.96327102e-01
-4.73710716e-01 4.62935835e-01 4.27460819e-01 -1.29964292e-01
-1.19056322e-01 -1.18989013e-02 -9.11948085e-01 -3.48798901e-01
-1.72270402e-01 -2.20064700e-01 1.76312238e-01 2.68489510e-01
4.37039971e-01 -8.70336220e-03 -1.27058637e+00 2.65150219e-01
6.35191560e-01 9.43143368e-01 2.02306017e-01 8.15227151e-01
3.96557480e-01 1.14797962e+00 -1.36261553e-01 -5.06396472e-01
6.65407836e-01 -6.73029482e-01 -9.31439817e-01 4.75080520e-01
1.11034882e+00 -9.97810721e-01 -8.73021483e-01 1.36882591e+00
2.93835431e-01 2.56620109e-01 1.48707390e-01 -1.23353267e+00
-8.69230390e-01 1.56807631e-01 -5.19346833e-01 5.73104322e-01
6.18496120e-01 3.03651303e-01 7.41575420e-01 -9.41916585e-01
5.81486106e-01 1.03247154e+00 8.43463600e-01 5.28366625e-01
-1.06845200e+00 -5.06296337e-01 4.07923520e-01 7.21762180e-01
-1.16270232e+00 -8.45565021e-01 5.05784333e-01 -5.10812759e-01
8.04174185e-01 5.86030483e-02 8.00993860e-01 1.01581323e+00
-1.00929961e-01 6.16154313e-01 8.00636232e-01 -1.48460343e-01
-3.37292217e-02 -5.72448373e-01 -1.60675362e-01 9.04093206e-01
2.24074036e-01 3.27094108e-01 -8.02944064e-01 6.42790422e-02
1.05040026e+00 9.71425045e-03 -8.82062256e-01 -2.50061065e-01
-1.55367446e+00 4.83596206e-01 2.76158839e-01 1.55685872e-01
-2.59144366e-01 5.76147616e-01 3.50371838e-01 6.14673972e-01
3.15660387e-01 7.61528760e-02 -2.10038215e-01 -1.78574309e-01
-1.16084981e+00 2.17774540e-01 4.42829847e-01 9.85421956e-01
5.30974686e-01 -5.20805866e-02 -2.27913409e-01 5.62911749e-01
5.04609525e-01 3.17942649e-01 4.92952824e-01 -1.63418937e+00
3.77650589e-01 -1.21157557e-01 3.43758047e-01 -8.64664793e-01
5.17308228e-02 1.06654480e-01 -6.67340696e-01 2.42234603e-01
6.59180701e-01 -3.12430784e-02 -9.87571239e-01 1.70252335e+00
1.42568350e-01 6.61682069e-01 -1.40267178e-01 1.30632651e+00
6.55456424e-01 3.94143045e-01 -2.19766840e-01 -4.51900810e-01
9.85154927e-01 -1.41160107e+00 -9.53706086e-01 -4.34699655e-02
1.04716137e-01 -9.84732330e-01 6.80176020e-01 3.18943262e-01
-1.48934448e+00 -3.81889313e-01 -9.25005019e-01 -2.08283871e-01
3.05932134e-01 8.31487998e-02 5.84538817e-01 2.77582765e-01
-1.46242821e+00 5.46698868e-01 -1.02859271e+00 -5.52208126e-01
2.94992805e-01 3.16205144e-01 -6.81855083e-01 -4.38048124e-01
-7.69950330e-01 6.29400373e-01 1.38546526e-01 4.64888602e-01
-1.12770927e+00 -5.60706317e-01 -1.03617430e+00 -3.04372817e-01
2.26862475e-01 -1.10692382e+00 1.66633070e+00 -1.39265633e+00
-1.62805581e+00 7.24081814e-01 -6.80371821e-01 -4.10145760e-01
8.45589578e-01 -4.02614474e-01 -2.08389387e-01 6.28801107e-01
1.17472135e-01 6.08803332e-01 1.38886023e+00 -1.13857317e+00
-5.75522125e-01 1.18515365e-01 2.44585425e-01 1.43931434e-01
-3.53401572e-01 2.98448056e-01 -1.26295662e+00 -9.60100234e-01
1.49315760e-01 -1.01589358e+00 -8.87133032e-02 4.12500471e-01
-8.93799439e-02 2.38288790e-01 6.14674330e-01 -1.04427421e+00
1.14322269e+00 -2.11559081e+00 3.89720440e-01 -1.11414731e-01
1.49338394e-01 1.99117348e-01 -2.75397152e-01 5.35672903e-02
-1.49783552e-01 -1.99975044e-01 -3.88364434e-01 -6.41085327e-01
-1.55335113e-01 -2.13383865e-02 -2.29918361e-01 1.01295578e+00
-5.14204465e-02 6.60419762e-01 -1.00905061e+00 -4.48261410e-01
3.35842252e-01 6.16506994e-01 -8.37225020e-01 4.70461369e-01
3.81468013e-02 7.75886595e-01 7.75093585e-02 7.29635060e-01
9.05550241e-01 -3.06186378e-01 2.96089798e-01 -4.75231946e-01
2.49734316e-02 1.89022124e-01 -1.02767146e+00 2.17777967e+00
-2.17857569e-01 8.55992496e-01 1.49138749e-01 -7.99558818e-01
-8.42587929e-03 6.26543343e-01 5.68377435e-01 -7.56837308e-01
1.32643223e-01 1.76019102e-01 -1.49634704e-01 -7.77580202e-01
4.65162009e-01 1.60621386e-02 5.00753582e-01 4.37482059e-01
1.14370577e-01 1.70946211e-01 3.89626116e-01 2.18792453e-01
9.99989450e-01 3.39935720e-01 -4.27788794e-02 1.55338589e-02
6.16702437e-01 -2.70758599e-01 8.23711812e-01 6.38079226e-01
-7.89026618e-01 1.03369129e+00 2.87625611e-01 -5.54438949e-01
-1.15076828e+00 -1.09054959e+00 -5.31596616e-02 6.13579631e-01
5.97705483e-01 -4.67270881e-01 -5.36316693e-01 -5.13796449e-01
-8.93140808e-02 -1.12740308e-01 -5.10850430e-01 2.52571017e-01
-5.87662280e-01 -3.40108603e-01 2.77766109e-01 4.79513794e-01
8.08602512e-01 -6.00204587e-01 -2.23530650e-01 -1.09245874e-01
-6.24062896e-01 -1.57661533e+00 -1.10019839e+00 -7.62362182e-01
-9.16867375e-01 -1.35266864e+00 -9.04451787e-01 -1.23428249e+00
9.45975780e-01 7.20331788e-01 9.07494724e-01 5.35238504e-01
2.89351828e-02 7.45510697e-01 -2.63933331e-01 4.88705158e-01
-1.11228056e-01 -7.31034935e-01 2.77060926e-01 1.64661169e-01
-4.25110431e-03 -3.40166390e-01 -8.93020868e-01 3.29525203e-01
-1.01285899e+00 2.86431294e-02 2.34249100e-01 9.00590122e-01
3.75658125e-01 3.04456688e-02 6.89418986e-02 -1.65859833e-01
2.86595315e-01 -1.84517115e-01 -8.11468899e-01 2.92045116e-01
-3.04256737e-01 -6.70832545e-02 2.62222290e-01 -4.87745225e-01
-1.11376393e+00 1.14019573e-01 1.94229141e-01 -9.78148162e-01
1.45861939e-01 3.96113873e-01 2.20826447e-01 -4.39322233e-01
7.14121759e-02 5.16897179e-02 2.06517056e-01 -1.97279692e-01
3.59891623e-01 1.82328030e-01 1.03741217e+00 -4.09404933e-01
8.55989754e-01 7.33869135e-01 -2.78320044e-01 -4.63377714e-01
-5.55314004e-01 -6.47393942e-01 -5.62116444e-01 -3.79891306e-01
9.03737187e-01 -1.37993157e+00 -7.32552469e-01 9.35264707e-01
-1.37476301e+00 -5.60368478e-01 3.95072162e-01 1.01119268e+00
-6.26163065e-01 9.59820986e-01 -1.02897716e+00 -4.81851339e-01
-3.94123085e-02 -1.20267987e+00 8.89883518e-01 9.06489138e-03
-3.06155030e-02 -1.10520554e+00 1.41730264e-01 4.23458517e-01
2.76406646e-01 -1.98934421e-01 3.02305341e-01 1.62484840e-01
-1.02477098e+00 -9.43323784e-03 -4.38288182e-01 3.69368017e-01
2.72202313e-01 1.25563383e-01 -7.53816664e-01 -6.40877724e-01
1.18259475e-01 -1.96174726e-01 1.13046098e+00 6.98197126e-01
1.07692623e+00 -4.63746220e-01 7.68525004e-02 9.87773597e-01
1.53140748e+00 -1.01479843e-01 7.26050019e-01 5.52424729e-01
7.17831314e-01 3.89769793e-01 4.99685198e-01 3.56454253e-01
4.15564179e-01 7.34606206e-01 4.89822716e-01 1.37899175e-01
-4.31966096e-01 -1.90508701e-02 7.23103523e-01 7.50598311e-01
-3.74896675e-01 -1.74775556e-01 -5.67231357e-01 8.42229247e-01
-2.24082088e+00 -1.36493814e+00 6.12114519e-02 2.36148047e+00
7.57051945e-01 -3.14182997e-01 4.92682867e-02 -3.37400101e-02
1.08642054e+00 2.41397962e-01 -2.60607958e-01 2.23254025e-01
-1.07113495e-01 -3.09890836e-01 4.44025487e-01 1.01624513e+00
-1.34544969e+00 8.48260820e-01 6.69853544e+00 1.73505992e-01
-1.05964851e+00 1.69842258e-01 2.67550260e-01 -5.97087383e-01
-1.85560912e-01 1.48499981e-01 -1.57743052e-01 5.60887933e-01
5.27357578e-01 -1.60039350e-01 7.35079288e-01 1.92738459e-01
4.41874743e-01 -1.55186251e-01 -1.16554213e+00 1.33584547e+00
3.64729762e-01 -1.49046147e+00 -7.88407326e-02 -1.25410974e-01
1.08271062e+00 -2.89611090e-02 2.66296744e-01 -3.43200803e-01
1.08325318e-01 -8.28548491e-01 9.79072452e-01 5.66335559e-01
7.81841099e-01 -3.05766821e-01 6.04372323e-01 -5.09568036e-01
-1.38617182e+00 -1.38989076e-01 -1.60387129e-01 9.10825282e-03
4.35018271e-01 3.69672805e-01 -1.51608795e-01 5.06446779e-01
1.14154184e+00 1.42815912e+00 -5.10022700e-01 1.61061084e+00
-3.25931460e-01 2.33629882e-01 7.69542381e-02 7.36233532e-01
1.24562524e-01 -1.00299001e-01 6.90205395e-01 1.02564752e+00
4.57074821e-01 1.58806483e-03 -3.27706486e-02 6.07002020e-01
-1.79153606e-01 -3.40304822e-01 -6.06230497e-01 -8.20104927e-02
2.26531520e-01 9.36354339e-01 -2.43843079e-01 -3.85362685e-01
-9.37931240e-01 1.53568923e+00 1.75785571e-01 7.74933517e-01
-9.40416873e-01 -2.48264894e-02 8.75788271e-01 -3.10630828e-01
5.62593639e-01 -6.67741895e-01 1.52111843e-01 -1.74871194e+00
4.89559323e-01 -9.59027708e-01 5.37266135e-01 -9.90821183e-01
-1.43314028e+00 5.36087096e-01 -2.77674675e-01 -1.70352256e+00
-2.10678264e-01 -5.67732930e-01 -6.35364056e-01 6.38322592e-01
-1.82065892e+00 -1.03977442e+00 -6.34101331e-01 1.18396115e+00
5.78592122e-01 -1.50920212e-01 2.84684986e-01 5.61726153e-01
-6.25162125e-01 5.44500887e-01 1.58687178e-02 2.82323182e-01
1.24936545e+00 -8.92310262e-01 2.81593174e-01 1.76509023e+00
-8.41872010e-04 6.02992892e-01 6.17239952e-01 -6.20956659e-01
-1.57244790e+00 -1.09576368e+00 8.10510457e-01 -4.15423512e-01
9.94689763e-01 2.15933606e-01 -8.83966982e-01 1.04691052e+00
7.60627270e-01 2.93117523e-01 3.39755684e-01 -3.59718442e-01
-6.29992187e-01 1.81683898e-02 -7.86364734e-01 6.42406940e-01
1.23352647e+00 -9.68069911e-01 -4.87524271e-01 4.02541816e-01
6.38575733e-01 -6.15220189e-01 -6.70146644e-01 4.09538746e-01
6.02598965e-01 -1.25980330e+00 8.76616061e-01 -4.63406056e-01
5.89043617e-01 -6.18225098e-01 -3.40053141e-01 -8.89632642e-01
-2.57521093e-01 -1.00582242e+00 -3.51961434e-01 7.99145579e-01
1.02478184e-01 -4.24873054e-01 5.07196546e-01 7.14205623e-01
-1.80413187e-01 -1.50504038e-01 -9.40207362e-01 -1.03204513e+00
-2.18315810e-01 -4.93614495e-01 9.57159325e-02 9.62449372e-01
8.54866430e-02 -3.81318927e-01 -7.71452904e-01 6.46846116e-01
8.56519222e-01 8.90138000e-02 5.68754256e-01 -5.00173986e-01
-4.99328226e-01 -4.41312730e-01 -5.83467901e-01 -1.18466806e+00
3.55959058e-01 -5.60247719e-01 2.32698113e-01 -1.50191259e+00
3.14491123e-01 2.15514302e-01 -1.63648441e-01 2.93316483e-01
-2.36674383e-01 6.39195561e-01 2.35586107e-01 5.32687664e-01
-8.84994924e-01 4.79948550e-01 1.17641628e+00 -2.21943706e-01
8.55013262e-03 -4.29782897e-01 -3.28625023e-01 5.82917571e-01
3.87374789e-01 -1.20958634e-01 -2.92284429e-01 -1.10973930e+00
3.24464478e-02 1.84129387e-01 8.36758852e-01 -8.86779547e-01
6.48365140e-01 -3.29332538e-02 2.80361027e-01 -4.93170135e-02
2.77122974e-01 -8.02213609e-01 1.67831957e-01 5.15744925e-01
-1.09724991e-01 3.53601068e-01 2.61897773e-01 7.67446697e-01
-5.53287208e-01 -3.33629549e-02 8.62446547e-01 -5.48380688e-02
-1.05924571e+00 7.42876589e-01 -2.63796806e-01 -8.73844996e-02
7.50923991e-01 9.48053524e-02 -6.33222938e-01 -6.79200828e-01
-6.56378090e-01 2.60278255e-01 8.28166366e-01 4.59802538e-01
7.99641848e-01 -1.54181862e+00 -6.37022018e-01 1.84620842e-01
2.40426697e-03 -3.77134323e-01 4.86921132e-01 1.42560673e+00
-8.27672362e-01 2.83662468e-01 -2.65127808e-01 -6.47278249e-01
-1.41079247e+00 7.26894557e-01 6.69436336e-01 3.86869967e-01
-8.67119968e-01 8.91546488e-01 1.38707355e-01 1.88445017e-01
5.43972611e-01 -2.84732968e-01 2.48006463e-01 -3.04858059e-01
7.47558534e-01 2.04443008e-01 -9.02205184e-02 -6.68875635e-01
-4.85064626e-01 7.64878869e-01 3.67828608e-02 -4.37344462e-01
9.99198258e-01 -5.63759506e-01 -4.31100845e-01 -5.72878011e-02
1.32661009e+00 7.06321597e-02 -1.71124506e+00 -2.86014020e-01
-2.58003503e-01 -1.08179915e+00 8.40773582e-02 -3.74182254e-01
-1.38134706e+00 4.42443967e-01 5.67340970e-01 -1.04390129e-01
1.37738287e+00 -9.76806954e-02 7.40342617e-01 3.02195549e-01
1.84142411e-01 -7.90046394e-01 3.36586118e-01 3.99066001e-01
9.79299903e-01 -1.75378072e+00 8.04557353e-02 -3.40862155e-01
-5.14626205e-01 1.13265955e+00 5.31974137e-01 -1.63486570e-01
6.64175868e-01 -6.56777024e-02 3.18756819e-01 1.24331206e-01
-7.24592149e-01 -1.75990716e-01 4.36895519e-01 4.41159129e-01
4.50116158e-01 -5.17285228e-01 -1.49924055e-01 6.04631640e-02
2.23831981e-01 2.36688107e-01 6.99070096e-01 8.80691469e-01
5.65077923e-02 -1.07741451e+00 -3.41974258e-01 -1.29540369e-01
-5.30576825e-01 -3.70330513e-01 -1.19921476e-01 5.28972089e-01
-1.82430357e-01 1.13936162e+00 1.88344434e-01 -2.87084103e-01
-1.41099870e-01 -3.29094291e-01 9.12089527e-01 -6.59645945e-02
-1.73182040e-01 2.85180926e-01 9.58338603e-02 -1.01658165e+00
-1.16724336e+00 -7.56991744e-01 -1.01115167e+00 -5.52261114e-01
2.40513921e-01 -9.25837755e-02 2.25312188e-01 8.94395351e-01
2.30776653e-01 1.52875423e-01 6.07623518e-01 -1.09450531e+00
-3.14622629e-03 -6.49780035e-01 -3.90649050e-01 6.87389553e-01
1.01491725e+00 -4.81965810e-01 -7.32420325e-01 6.02319956e-01]
|
[10.713915824890137, -1.488459587097168]
|
63e8b61a-1b86-4b18-aeb3-ca4fc0d2a887
|
a-slow-shifting-concerned-machine-learning
|
2303.17782
| null |
https://arxiv.org/abs/2303.17782v1
|
https://arxiv.org/pdf/2303.17782v1.pdf
|
A Slow-Shifting Concerned Machine Learning Method for Short-term Traffic Flow Forecasting
|
The ability to predict traffic flow over time for crowded areas during rush hours is increasingly important as it can help authorities make informed decisions for congestion mitigation or scheduling of infrastructure development in an area. However, a crucial challenge in traffic flow forecasting is the slow shifting in temporal peaks between daily and weekly cycles, resulting in the nonstationarity of the traffic flow signal and leading to difficulty in accurate forecasting. To address this challenge, we propose a slow shifting concerned machine learning method for traffic flow forecasting, which includes two parts. First, we take advantage of Empirical Mode Decomposition as the feature engineering to alleviate the nonstationarity of traffic flow data, yielding a series of stationary components. Second, due to the superiority of Long-Short-Term-Memory networks in capturing temporal features, an advanced traffic flow forecasting model is developed by taking the stationary components as inputs. Finally, we apply this method on a benchmark of real-world data and provide a comparison with other existing methods. Our proposed method outperforms the state-of-art results by 14.55% and 62.56% using the metrics of root mean squared error and mean absolute percentage error, respectively.
|
['Chau Yuen', 'Yong Liang Guan', 'Yan Qin', 'Zann Koh']
|
2023-03-31
| null | null | null | null |
['feature-engineering']
|
['methodology']
|
[ 8.19858611e-02 -7.35445082e-01 -3.48651469e-01 -1.97493315e-01
-3.30093980e-01 -5.06726233e-03 2.88855702e-01 -2.86059201e-01
-2.09397212e-01 9.22351480e-01 1.98974282e-01 -6.88028336e-01
-4.26997006e-01 -8.82160604e-01 -1.71460599e-01 -7.57402420e-01
-1.68047681e-01 2.26377100e-01 3.43393028e-01 -4.21158165e-01
1.86778709e-01 5.03448188e-01 -1.65856147e+00 -4.39268537e-02
1.26802921e+00 1.07042539e+00 9.41245109e-02 3.04557055e-01
-4.54437971e-01 8.89189243e-01 -4.34410185e-01 -4.95562032e-02
2.31840059e-01 -1.56695351e-01 -5.50520599e-01 1.11813068e-01
-4.39748168e-02 -2.05065474e-01 -7.14993477e-01 6.74127042e-01
1.81180760e-01 6.35275185e-01 5.48680305e-01 -1.59259057e+00
-9.93216336e-02 4.76651732e-03 -8.28965724e-01 8.63543332e-01
-1.59817964e-01 3.73400629e-01 6.63102329e-01 -5.35932243e-01
-1.28179505e-01 1.22303891e+00 5.07101059e-01 3.15437555e-01
-1.05374265e+00 -7.98114657e-01 4.21319127e-01 5.74747622e-01
-1.33971250e+00 -6.67034090e-01 8.85013163e-01 -7.43290663e-01
8.79820526e-01 2.84923613e-01 4.50348318e-01 5.33011973e-01
3.48693252e-01 6.14266038e-01 7.64718115e-01 -1.31336050e-02
-9.18663070e-02 8.07382632e-03 1.05813347e-01 2.59362280e-01
-3.03699216e-03 1.55061454e-01 -1.39265239e-01 7.10666329e-02
3.38085383e-01 4.05637860e-01 -8.00756663e-02 4.52807575e-01
-9.70477700e-01 5.33171713e-01 3.86834204e-01 2.81922251e-01
-6.66493177e-01 6.89595984e-03 5.35339534e-01 9.28725675e-02
8.20279121e-01 -3.19308192e-01 -3.57718319e-01 -4.18711692e-01
-9.90996242e-01 8.19985196e-02 3.23897690e-01 6.80951059e-01
8.68672609e-01 4.63153630e-01 -3.06397706e-01 8.12227726e-01
2.37577453e-01 9.69740212e-01 4.04181153e-01 -8.47817242e-01
1.05904138e+00 3.82004559e-01 2.34085679e-01 -1.46460760e+00
-6.13154709e-01 -5.47266662e-01 -1.12451434e+00 -3.04912955e-01
3.99919719e-01 -3.90662521e-01 -5.27341664e-01 1.64268994e+00
1.46991268e-01 7.87702978e-01 -1.39050797e-01 6.66428864e-01
3.14469367e-01 1.15326190e+00 2.49259502e-01 -6.18830323e-01
1.04871011e+00 -9.06367838e-01 -8.39432597e-01 -1.35011584e-01
5.38138449e-01 -7.02030420e-01 5.97244203e-01 -6.89600483e-02
-7.11164653e-01 -7.24688172e-01 -3.88279110e-01 5.34130335e-01
-3.02593112e-01 2.59883031e-02 4.78034496e-01 4.89742160e-01
-6.04309499e-01 2.67183036e-01 -6.33672357e-01 -1.37295246e-01
4.56229150e-01 1.09707877e-01 2.35790829e-03 8.46453384e-02
-1.44805598e+00 5.66873550e-01 2.99027473e-01 6.07614696e-01
-2.31861234e-01 -1.07800782e+00 -6.63399994e-01 2.18429744e-01
3.53347927e-01 -4.86263812e-01 9.68792677e-01 -7.12406933e-01
-1.19377303e+00 9.50662345e-02 -8.26024592e-01 -3.31649750e-01
4.69630510e-01 9.22045037e-02 -1.24464309e+00 -2.01863959e-01
2.29257479e-01 2.80120909e-01 8.07864726e-01 -8.11809361e-01
-1.14261615e+00 -1.45612568e-01 -3.12166601e-01 -1.64911553e-01
-3.95554811e-01 -2.88327783e-01 -6.08997792e-02 -6.24333441e-01
-1.67233720e-01 -8.63565683e-01 -1.69190630e-01 -6.48224771e-01
-1.89564466e-01 -4.47161615e-01 9.96660352e-01 -7.74450600e-01
1.93641651e+00 -2.05559301e+00 -6.17983580e-01 3.30078900e-01
9.11020264e-02 6.05037928e-01 -3.80331278e-02 5.86859465e-01
-1.28203064e-01 -8.58326927e-02 -8.38480741e-02 -1.94818974e-01
-1.96841359e-01 2.73973972e-01 -6.85500145e-01 2.87165880e-01
3.72267246e-01 6.88143253e-01 -8.09980512e-01 -2.89602220e-01
5.72505593e-01 3.15652996e-01 -2.15720028e-01 3.14254947e-02
1.57559052e-01 7.49551296e-01 -4.70936388e-01 2.79150903e-01
1.03581131e+00 -1.03502363e-01 -3.19075525e-01 1.36303967e-02
-5.13370156e-01 2.13442355e-01 -1.18001342e+00 9.77601171e-01
-8.72782767e-01 8.97361517e-01 -3.90783012e-01 -1.21958947e+00
9.62675810e-01 3.92470747e-01 8.43330443e-01 -9.96286392e-01
1.45739997e-02 2.62696594e-01 -3.41678672e-02 -8.85932028e-01
3.73760432e-01 -1.98195860e-01 1.86132506e-01 4.38875467e-01
-6.24258757e-01 7.16183901e-01 5.68079650e-01 -1.32517233e-01
8.69653165e-01 -5.12523949e-01 -2.23446012e-01 -2.25259930e-01
1.08843708e+00 -2.67730534e-01 9.76153791e-01 2.32529983e-01
-4.67218876e-01 5.64101152e-02 3.49867225e-01 -8.09569895e-01
-6.67846501e-01 -7.28316307e-01 -1.30960748e-01 8.86267781e-01
-1.48766851e-02 1.28442898e-01 -2.76461005e-01 -4.31600720e-01
1.77926630e-01 7.26057172e-01 -4.89589661e-01 -1.38258874e-01
-8.49083841e-01 -9.43672717e-01 2.23677292e-01 5.66205621e-01
7.45833874e-01 -8.64323318e-01 -2.40088433e-01 5.24725854e-01
-5.24871469e-01 -1.20879555e+00 -6.04991496e-01 -6.67307377e-01
-8.36943686e-01 -9.20823395e-01 -6.30796373e-01 -3.78054678e-01
4.71185505e-01 9.69097376e-01 9.17358518e-01 1.82849839e-01
8.24034140e-02 -1.25224665e-01 3.16502415e-02 -2.37419412e-01
-9.89306197e-02 3.61303478e-01 1.07548311e-01 6.34397388e-01
6.34089589e-01 -8.17041278e-01 -7.63889432e-01 6.45473838e-01
-8.80553782e-01 -1.00341607e-02 4.59874600e-01 6.01311982e-01
1.55279562e-01 5.29521823e-01 1.21995902e+00 -6.20161831e-01
6.11978292e-01 -9.60653126e-01 -6.61740661e-01 1.08300827e-01
-6.91633880e-01 -1.09647676e-01 8.76767278e-01 -1.80493951e-01
-1.32369804e+00 -2.79566854e-01 -1.20195590e-01 -1.90366477e-01
-2.08118573e-01 5.29800355e-01 3.36783752e-02 3.98566008e-01
1.26808405e-01 4.26506639e-01 -1.94943286e-02 -3.89957190e-01
-1.71047196e-01 8.31820905e-01 2.53183722e-01 -2.34198570e-01
1.00239038e+00 5.84524214e-01 3.37182224e-01 -9.85882163e-01
-7.02792823e-01 -7.76914120e-01 -5.43360710e-01 -4.31566983e-01
5.13876796e-01 -9.62255180e-01 -1.08198583e+00 5.84922552e-01
-1.10631657e+00 3.66208740e-02 1.61133081e-01 6.14744902e-01
-3.45992923e-01 3.00788522e-01 -3.72510493e-01 -1.07887948e+00
-9.77840647e-02 -8.83993983e-01 5.73492050e-01 3.13973278e-01
1.30870745e-01 -1.34021151e+00 2.39150211e-01 4.82117474e-01
7.93015480e-01 1.38291910e-01 8.63095343e-01 -1.84340656e-01
-3.87863696e-01 -1.72729477e-01 -3.97565842e-01 3.14392120e-01
4.57645506e-01 9.01312083e-02 -9.11608458e-01 -1.23841412e-01
-1.89666718e-01 6.04041576e-01 9.43719983e-01 5.66088200e-01
1.07413518e+00 -4.04851496e-01 -4.53257740e-01 2.73779362e-01
1.13967705e+00 4.59387273e-01 7.70089626e-01 1.79309919e-01
5.72410882e-01 8.92649293e-01 5.92300177e-01 5.81826389e-01
6.71372235e-01 5.34022510e-01 1.96631074e-01 -1.50067601e-02
1.13669276e-01 -3.08661222e-01 2.38113895e-01 1.04942894e+00
-9.92453545e-02 -2.81813174e-01 -1.01137304e+00 7.49821544e-01
-2.13684583e+00 -1.43664527e+00 -5.58985472e-01 2.23983932e+00
1.42003551e-01 3.04743022e-01 5.18030465e-01 4.21423107e-01
9.27952349e-01 2.90262103e-01 -4.44159746e-01 -1.90043867e-01
1.46273017e-01 -3.10125053e-01 6.22351050e-01 5.57269394e-01
-1.03845668e+00 6.97608232e-01 5.73633242e+00 1.14701819e+00
-1.36779237e+00 -9.85644311e-02 7.94706702e-01 1.09272465e-01
-4.11511362e-02 -1.90636128e-01 -8.05035710e-01 1.13560522e+00
1.23982275e+00 -3.05002689e-01 5.52762270e-01 3.39683950e-01
1.20959342e+00 -8.07053745e-02 -4.44892079e-01 9.66620207e-01
-3.34724724e-01 -1.14564931e+00 -4.79485057e-02 6.25791177e-02
7.94163644e-01 -1.28799960e-01 1.37328118e-01 4.78922695e-01
-3.43650967e-01 -9.59331930e-01 1.03867933e-01 9.62271571e-01
3.65489602e-01 -1.06877851e+00 8.50601017e-01 5.31562746e-01
-1.70344245e+00 -3.56130630e-01 -1.72862634e-01 -5.19127786e-01
6.55194879e-01 9.54026222e-01 -4.49404985e-01 5.46444476e-01
2.83770263e-01 9.96106446e-01 -2.96312004e-01 1.20336568e+00
1.62125558e-01 9.08819616e-01 -2.02572316e-01 3.16026747e-01
1.89297169e-01 -2.67264724e-01 5.77084720e-01 1.02496684e+00
3.58436495e-01 3.75552885e-02 2.41155878e-01 5.86381555e-01
1.34894937e-01 -4.67145927e-02 -4.66019779e-01 3.35750788e-01
5.82558036e-01 1.08909690e+00 -4.39787835e-01 -2.98197538e-01
-5.54865897e-01 2.29285911e-01 -1.29938245e-01 6.84275866e-01
-1.02863038e+00 -5.55540919e-01 8.81427526e-01 3.94757450e-01
1.88341424e-01 -3.88624400e-01 -1.81523860e-01 -1.07518566e+00
1.97489873e-01 -3.33023727e-01 2.44693592e-01 -2.64658332e-01
-1.32187533e+00 5.59990525e-01 4.36874814e-02 -1.59153259e+00
-2.67576158e-01 -2.47865438e-01 -1.09075582e+00 1.08228838e+00
-2.06073189e+00 -8.16718876e-01 -3.06323647e-01 5.17172337e-01
7.90595174e-01 -2.22661749e-01 1.92276940e-01 9.89332855e-01
-1.18324125e+00 2.77108520e-01 2.69958615e-01 1.19420655e-01
4.44330662e-01 -6.35370255e-01 5.00912428e-01 9.37756062e-01
-3.81893843e-01 2.15951383e-01 6.28996611e-01 -4.24246281e-01
-1.00873816e+00 -1.42268836e+00 1.25428975e+00 -1.55071631e-01
7.66995072e-01 1.49652302e-01 -9.59778488e-01 4.09192592e-01
-1.10772066e-01 1.05209276e-01 5.28472722e-01 -1.27415791e-01
1.78189203e-01 -7.30898917e-01 -8.85356486e-01 3.37284386e-01
8.20811331e-01 -2.09644273e-01 -1.84843659e-01 1.09410271e-01
4.26295668e-01 1.48562253e-01 -6.34350359e-01 3.36991072e-01
3.89395207e-01 -7.58480489e-01 6.46780968e-01 -6.53317809e-01
1.39475465e-02 -5.29151142e-01 7.73269534e-02 -1.45475352e+00
-6.26765013e-01 -7.35381007e-01 -1.89705700e-01 1.41324270e+00
4.04481322e-01 -1.05968034e+00 6.05838656e-01 7.32109666e-01
-2.32868851e-03 -4.82219994e-01 -1.11693215e+00 -9.23647165e-01
-1.07565016e-01 -7.33619392e-01 9.72284257e-01 6.69073522e-01
-1.91732094e-01 3.82084936e-01 -6.49394095e-01 9.81612131e-03
4.94325727e-01 2.11884916e-01 7.92764306e-01 -1.41760349e+00
3.47894639e-01 -5.59427083e-01 -3.96061718e-01 -1.39188159e+00
5.04300237e-01 -5.25334895e-01 -1.04886591e-01 -1.18147361e+00
-1.76401883e-01 -4.71664339e-01 -4.70163703e-01 2.73596141e-02
-3.18885177e-01 1.00103347e-02 5.26937321e-02 2.95599848e-01
-5.38486242e-01 9.82116997e-01 1.07810903e+00 -1.01218164e-01
-4.13818896e-01 7.85157800e-01 -4.50560927e-01 5.08955956e-01
1.08717263e+00 -4.29845393e-01 -5.03801823e-01 -4.33864146e-01
6.26157373e-02 1.82322308e-01 2.17689723e-01 -1.17644012e+00
2.89777249e-01 -6.16656363e-01 -6.68178573e-02 -8.58751535e-01
3.23895812e-02 -1.16916037e+00 7.56098181e-02 5.82978606e-01
-1.35532366e-02 3.80767465e-01 3.69052261e-01 8.34174156e-01
-3.57393444e-01 2.90800422e-01 3.95837098e-01 3.98728788e-01
-8.22857618e-01 6.15920544e-01 -9.07092571e-01 1.25405520e-01
1.01930046e+00 -1.36463031e-01 -4.02614117e-01 -4.81110096e-01
-3.57083529e-01 7.48296738e-01 -3.38858932e-01 6.90035999e-01
3.89175683e-01 -1.64888549e+00 -7.58756161e-01 3.87091905e-01
-6.31653070e-02 -3.57828170e-01 9.63414907e-01 1.35438156e+00
-3.24765384e-01 9.48097706e-01 -4.09186408e-02 -6.48537159e-01
-7.79192209e-01 6.09525204e-01 2.68874139e-01 -2.42508039e-01
-2.54583687e-01 2.86483802e-02 3.66510660e-03 -1.26901060e-01
-3.52126658e-02 -3.75199229e-01 -4.24504936e-01 1.63829044e-01
7.91864097e-01 1.20511734e+00 1.09214373e-01 -1.12725914e+00
-4.24830288e-01 6.57496393e-01 2.93864310e-01 3.25547516e-01
1.07380211e+00 -6.38699710e-01 -6.85578287e-02 5.68069398e-01
1.21683431e+00 -1.82653040e-01 -1.24420726e+00 -3.15869868e-01
-5.65877259e-02 -7.43050158e-01 1.28557742e-01 -2.37789109e-01
-1.43437672e+00 1.08900344e+00 6.59881771e-01 6.15461886e-01
1.17148018e+00 -7.11642563e-01 1.50226641e+00 2.67491072e-01
5.17418124e-02 -1.03582549e+00 -5.30291975e-01 6.85360014e-01
3.16827714e-01 -1.56332016e+00 -4.41946685e-01 -4.30892974e-01
-6.02422118e-01 1.13361359e+00 4.41801131e-01 5.49401417e-02
1.10370028e+00 -2.25136265e-01 1.34633780e-01 1.70856744e-01
-9.41736758e-01 -3.34996700e-01 3.61771196e-01 4.10212934e-01
1.99127600e-01 1.64743572e-01 -2.29617506e-01 1.60315812e-01
9.82617363e-02 2.62172282e-01 1.69539720e-01 4.70258236e-01
-5.12958407e-01 -7.62352705e-01 -3.21548373e-01 5.20827770e-01
-2.90687203e-01 7.97355548e-02 4.71764237e-01 4.54186410e-01
1.40354726e-02 1.55675125e+00 3.63800734e-01 -4.82455790e-01
4.21811730e-01 -7.29829445e-02 -3.87586325e-01 -6.63118884e-02
-2.07502887e-01 -1.55408382e-01 -1.10152788e-01 -3.46515566e-01
-5.12217462e-01 -6.15011334e-01 -9.52968419e-01 -1.01871228e+00
-1.64456114e-01 3.18018138e-01 2.48126760e-01 1.34090829e+00
6.17585301e-01 6.12788737e-01 1.36144578e+00 -6.68784082e-01
-1.66790530e-01 -9.23720717e-01 -4.59797621e-01 4.30273682e-01
6.42131209e-01 -8.71714592e-01 -4.79964912e-01 -7.49830231e-02]
|
[6.410472393035889, 2.0829200744628906]
|
e32f9c0d-d240-476a-ba9e-67709e7cddc0
|
delegated-classification
|
2306.11475
| null |
https://arxiv.org/abs/2306.11475v1
|
https://arxiv.org/pdf/2306.11475v1.pdf
|
Delegated Classification
|
When machine learning is outsourced to a rational agent, conflicts of interest might arise and severely impact predictive performance. In this work, we propose a theoretical framework for incentive-aware delegation of machine learning tasks. We model delegation as a principal-agent game, in which accurate learning can be incentivized by the principal using performance-based contracts. Adapting the economic theory of contract design to this setting, we define budget-optimal contracts and prove they take a simple threshold form under reasonable assumptions. In the binary-action case, the optimality of such contracts is shown to be equivalent to the classic Neyman-Pearson lemma, establishing a formal connection between contract design and statistical hypothesis testing. Empirically, we demonstrate that budget-optimal contracts can be constructed using small-scale data, leveraging recent advances in the study of learning curves and scaling laws. Performance and economic outcomes are evaluated using synthetic and real-world classification tasks.
|
['Nir Rosenfeld', 'Inbal Talgam-Cohen', 'Eden Saig']
|
2023-06-20
| null | null | null | null |
['classification-1']
|
['methodology']
|
[ 9.13014635e-02 5.34963012e-01 -7.99265802e-01 -6.96289003e-01
-8.88531029e-01 -7.64042795e-01 1.43295273e-01 -1.08334921e-01
-8.71113181e-01 9.84795392e-01 -3.20271067e-02 -6.65570915e-01
-7.32453346e-01 -4.01077718e-01 -8.42776537e-01 -7.58261859e-01
-3.88187498e-01 6.26486361e-01 -4.69878018e-01 4.40218031e-01
2.66609224e-03 3.67342085e-01 -1.25646448e+00 -4.83603962e-02
7.22854614e-01 9.86746550e-01 -1.18997872e-01 7.70365357e-01
5.97570360e-01 6.64540648e-01 -6.55889511e-01 -7.54501760e-01
9.04462218e-01 -1.58287048e-01 -7.43179262e-01 9.26980600e-02
-5.75453416e-02 -7.31850505e-01 -2.17061818e-01 1.02383149e+00
3.63763541e-01 -2.45344192e-01 5.58517039e-01 -1.83404791e+00
-3.22393894e-01 7.68872738e-01 -4.85683501e-01 5.55057935e-02
-1.67968720e-01 9.33114812e-02 1.62118089e+00 5.43201305e-02
5.44689655e-01 1.08430266e+00 5.28303742e-01 7.01998293e-01
-1.37849593e+00 -5.62348604e-01 -1.75390974e-01 -1.52233571e-01
-6.56721532e-01 -1.47882774e-01 5.11907101e-01 -4.19605196e-01
6.16531551e-01 3.74141753e-01 4.18219596e-01 7.89390385e-01
2.89084494e-01 1.16847026e+00 1.49915731e+00 -3.90152633e-01
4.58552957e-01 2.15469077e-01 -3.91526818e-02 3.94960821e-01
6.72110915e-01 6.90235496e-01 -5.70936501e-01 -6.88173234e-01
6.30652130e-01 7.23534226e-02 -2.35835031e-01 -7.61912942e-01
-7.54962683e-01 1.19986010e+00 9.02148783e-02 -1.93705663e-01
-7.10829973e-01 3.32830817e-01 4.46662456e-01 7.78753281e-01
4.14552122e-01 5.87845087e-01 -8.00106585e-01 -1.39074236e-01
-4.92009133e-01 6.24600828e-01 1.07568920e+00 1.05350208e+00
1.92010447e-01 -4.67604280e-01 -4.14879858e-01 3.15051317e-01
2.76177246e-02 3.93401951e-01 1.91120297e-01 -1.73198617e+00
5.83523095e-01 9.93382037e-02 8.19665611e-01 -1.92998752e-01
-3.03987116e-01 -7.30172813e-01 -3.56114835e-01 3.23708445e-01
6.73784375e-01 -4.59682524e-01 -7.19918832e-02 2.08630109e+00
7.20256045e-02 -2.88101614e-01 2.55405575e-01 9.22894478e-01
-3.45607579e-01 -6.99188411e-02 3.37518081e-02 -6.62119746e-01
1.22121525e+00 -5.68161488e-01 -4.90194470e-01 1.40743405e-01
8.95173252e-01 -2.86768824e-01 9.94694650e-01 3.34806383e-01
-1.00514615e+00 3.04496676e-01 -7.67968595e-01 3.19331259e-01
3.74543548e-01 -2.27209046e-01 1.15161765e+00 9.08670783e-01
-8.15989017e-01 8.43649328e-01 -5.76638103e-01 1.14629082e-01
8.30742180e-01 4.78539258e-01 -1.74876228e-02 1.24027371e-01
-1.05549955e+00 3.47370565e-01 -1.02653511e-01 -2.84283876e-01
-1.09383667e+00 -8.45130980e-01 -2.86492318e-01 2.54120976e-01
7.10362256e-01 -8.79513323e-01 2.18929839e+00 -9.28287506e-01
-1.19947958e+00 9.22260165e-01 3.06021392e-01 -7.86907494e-01
1.05571508e+00 -8.36338028e-02 4.73203033e-01 -1.14889227e-01
2.14908317e-01 1.82539523e-01 4.98623848e-01 -9.76006329e-01
-1.22095084e+00 -5.59141695e-01 4.74716425e-01 3.07397932e-01
-6.00129008e-01 1.76353902e-01 3.73239607e-01 -4.85349685e-01
-4.02758479e-01 -1.04315269e+00 -4.03225780e-01 -1.56975672e-01
3.32395732e-02 -5.91276050e-01 3.56233865e-01 -2.59753048e-01
7.62162626e-01 -1.92040181e+00 -3.63195419e-01 8.77284631e-02
3.21460217e-01 -5.25894642e-01 8.87239501e-02 -2.87142210e-02
2.71746516e-01 2.42904812e-01 -4.30646241e-02 -4.16337281e-01
6.92934394e-01 4.25052583e-01 -3.12615782e-01 7.08056867e-01
-4.63712066e-01 1.00616288e+00 -6.24137282e-01 -2.26550519e-01
-6.77699685e-01 -4.83287960e-01 -8.23181152e-01 3.01759034e-01
-1.69371471e-01 -1.50157899e-01 -8.29917312e-01 4.11874712e-01
4.75568831e-01 -2.80995697e-01 3.93199563e-01 6.82204723e-01
2.75117934e-01 2.65476406e-01 -8.56209636e-01 1.13439941e+00
-2.99415350e-01 4.75871205e-01 7.50808477e-01 -1.23495233e+00
2.42118552e-01 3.41641009e-01 7.65264690e-01 -3.37221503e-01
2.44516358e-01 2.27789432e-01 -5.31460978e-02 -4.16140139e-01
2.26174220e-01 -4.20725495e-01 -3.37016225e-01 1.10723066e+00
-2.67979532e-01 1.27494425e-01 -2.92903930e-01 1.22816496e-01
1.19369328e+00 -2.58194259e-03 2.37653553e-01 -5.25754094e-01
-2.82761693e-01 1.72558665e-01 8.65825653e-01 1.17443657e+00
-3.93882662e-01 -2.45361164e-01 1.05820370e+00 -2.24503130e-01
-1.22188771e+00 -8.00301969e-01 -2.05189794e-01 1.31137645e+00
-3.06322873e-01 1.79790765e-01 -1.02085888e+00 -9.81686652e-01
8.15312505e-01 4.17632312e-01 -6.76862776e-01 4.53655161e-02
9.71682463e-03 -8.86813164e-01 4.66346502e-01 4.49445277e-01
4.17418331e-01 -7.24069297e-01 -1.25962114e+00 -5.49921254e-03
1.43594161e-01 -9.71107364e-01 -8.43683243e-01 2.61454254e-01
-1.04535532e+00 -1.22982883e+00 -7.08900511e-01 -1.58958808e-01
4.88460779e-01 2.23701492e-01 1.03103137e+00 -1.42536610e-01
-1.95118710e-01 6.99590981e-01 5.36131188e-02 -8.38497281e-01
-3.44379067e-01 5.59119880e-02 1.77512228e-01 -1.55641690e-01
5.22810400e-01 -2.47366130e-01 -7.66440630e-01 3.06207687e-01
-8.21160972e-01 -1.54253036e-01 7.37376988e-01 1.11136627e+00
2.99125165e-01 1.74329460e-01 8.79157186e-01 -1.15301490e+00
1.07237768e+00 -4.16584313e-01 -1.33569634e+00 2.79650867e-01
-1.44277894e+00 2.01350793e-01 4.49637443e-01 -4.04886842e-01
-1.06257486e+00 6.67036176e-02 8.80573034e-01 -3.93918455e-01
9.11820978e-02 4.65025544e-01 -3.38527411e-01 8.87341890e-03
6.00906968e-01 -2.27972016e-01 2.41171062e-01 -4.78214294e-01
1.26901090e-01 1.04534662e+00 4.26991016e-01 -1.17847598e+00
4.76664096e-01 4.24071044e-01 3.57109964e-01 -7.37070963e-02
-8.08636665e-01 -2.43010551e-01 -2.34886214e-01 1.35000683e-02
3.79379600e-01 -8.29931438e-01 -1.66171014e+00 1.59844488e-01
-8.54777098e-01 -5.19502759e-01 -6.74951017e-01 7.56026506e-01
-1.17812634e+00 2.69289091e-02 -6.82048619e-01 -1.31400990e+00
-4.48113047e-02 -9.46359754e-01 7.00220704e-01 2.05289572e-01
2.29951382e-01 -8.01011503e-01 -1.26831800e-01 6.98202014e-01
3.90716881e-01 1.06693298e-01 8.15879107e-01 -6.60141408e-01
-7.52722323e-01 -1.53429449e-01 -4.63614352e-02 2.00511709e-01
-3.02281916e-01 -4.76509303e-01 -1.12781775e+00 -4.51471090e-01
3.33332807e-01 -5.39254487e-01 5.81397831e-01 6.66894734e-01
1.04287469e+00 -1.04340565e+00 -3.57691020e-01 5.22252798e-01
1.20251346e+00 1.93624407e-01 -4.74742949e-02 4.14555997e-01
6.48456439e-03 9.69075799e-01 8.36993396e-01 8.95754218e-01
4.15461421e-01 6.18057966e-01 2.26597667e-01 1.43643513e-01
8.52967381e-01 -3.58384728e-01 1.04332238e-01 -1.63971558e-01
-1.64550796e-01 2.58729339e-01 -5.45333803e-01 3.37554961e-01
-2.24059439e+00 -9.43558872e-01 2.10155472e-01 2.55530596e+00
1.10721028e+00 2.47021720e-01 6.54954493e-01 -1.01327091e-01
4.27980393e-01 -4.71748054e-01 -1.04768121e+00 -5.33747613e-01
-2.20113486e-01 -4.54311818e-03 1.33315825e+00 5.31194866e-01
-9.52997208e-01 3.81693661e-01 7.22337294e+00 7.02180564e-01
-4.44930464e-01 6.06721699e-01 1.15475166e+00 -4.54400212e-01
-5.36464036e-01 1.78149283e-01 -3.86986196e-01 2.19863281e-01
8.76589715e-01 -8.87444496e-01 6.98928177e-01 1.25893390e+00
1.97693303e-01 -8.55670273e-02 -1.39794731e+00 7.37537384e-01
-6.15159750e-01 -1.16348839e+00 -8.79578888e-01 8.58609140e-01
8.66853774e-01 -4.80359681e-02 4.06910777e-02 2.63627678e-01
1.03147614e+00 -1.03678119e+00 7.08928585e-01 2.50473559e-01
6.86873317e-01 -1.06703901e+00 8.45629990e-01 5.59624672e-01
-4.21595484e-01 -6.10408425e-01 -3.67128104e-01 -4.33755159e-01
-4.22258466e-01 5.82731783e-01 -8.96804512e-01 3.53573322e-01
4.92818743e-01 8.20167065e-02 7.73947872e-03 7.85563350e-01
-6.29128441e-02 7.56010830e-01 -2.17609137e-01 -1.34166688e-01
-1.37892989e-02 -1.60548210e-01 1.78424016e-01 1.01391959e+00
-1.18517622e-01 3.25764358e-01 -2.52528097e-02 1.02240050e+00
-5.29866755e-01 -5.33903278e-02 -5.98396003e-01 -1.18732706e-01
4.74082172e-01 1.13652277e+00 -3.33459675e-01 -1.07352234e-01
-4.28282350e-01 6.35591388e-01 2.14150950e-01 4.30080771e-01
-5.70812047e-01 -2.26406202e-01 9.23694611e-01 -3.06497663e-01
1.21895850e-01 1.68531477e-01 -6.80692077e-01 -8.86954665e-01
3.92882735e-01 -7.85632968e-01 8.37700903e-01 -1.96341097e-01
-1.41538274e+00 -1.52318224e-01 1.52858660e-01 -7.25469410e-01
-5.96984386e-01 -5.64951360e-01 -2.93601900e-01 8.05826187e-01
-1.45994318e+00 -5.51029861e-01 2.97490627e-01 3.62023920e-01
4.71837670e-02 -2.48885453e-01 6.88386142e-01 -1.74469233e-01
-2.09306180e-01 9.38872457e-01 4.20921236e-01 -1.36738256e-01
2.41352841e-01 -1.55222166e+00 -5.99267632e-02 4.53485668e-01
-3.21146809e-02 1.90244630e-01 7.90720642e-01 -3.55811954e-01
-1.60572600e+00 -6.70090199e-01 6.08657241e-01 -2.85060048e-01
7.36921012e-01 -3.84072930e-01 -4.48520839e-01 5.93847573e-01
-6.47675544e-02 4.29661237e-02 7.12610245e-01 3.44142169e-01
-3.70132416e-01 -3.95343482e-01 -1.41337216e+00 2.10400850e-01
1.19412267e+00 -5.50806522e-01 -2.74209648e-01 6.15320444e-01
7.43017673e-01 -1.89473733e-01 -1.02114999e+00 2.85410851e-01
7.80536234e-01 -7.90867925e-01 4.06316340e-01 -1.26740575e+00
2.66600668e-01 4.36521351e-01 -1.70595214e-01 -1.08103037e+00
-2.46359874e-02 -1.34866607e+00 9.26288441e-02 8.06360185e-01
4.56737369e-01 -8.51021171e-01 1.19113946e+00 1.25771260e+00
4.16411698e-01 -7.50156522e-01 -1.30399346e+00 -1.25483203e+00
5.44358194e-01 -3.18630040e-01 7.28556037e-01 9.60391104e-01
3.96982044e-01 -5.11035137e-02 -2.71907657e-01 -1.64402798e-01
1.29317987e+00 3.91896427e-01 7.21508682e-01 -1.28978348e+00
-9.76074457e-01 -7.59400368e-01 -6.67791292e-02 -8.95167053e-01
5.06257236e-01 -8.24469805e-01 -2.24247556e-02 -8.97199154e-01
7.77604699e-01 -8.28525424e-01 -3.49175692e-01 5.93523383e-01
2.04230640e-02 -4.92944717e-01 3.91429365e-01 5.49468398e-01
-4.33288902e-01 3.39692622e-01 1.01517439e+00 3.86883542e-02
1.42813772e-01 6.90219820e-01 -1.06654370e+00 3.04205507e-01
8.51601005e-01 -7.64708579e-01 -4.43495184e-01 -3.28300856e-02
2.62008101e-01 5.10839343e-01 3.99253219e-01 -4.13802639e-02
2.47582093e-01 -7.82434285e-01 -1.31874785e-01 1.98930234e-01
-1.90701693e-01 -8.73146236e-01 -5.36507294e-02 9.07911718e-01
-9.57272768e-01 4.66164947e-02 -4.99849707e-01 6.55254781e-01
3.23502928e-01 -3.90003949e-01 4.40082014e-01 7.64368623e-02
1.90683901e-01 3.13398808e-01 -4.76766285e-03 2.93899000e-01
1.27471137e+00 4.02673692e-01 -3.48019451e-01 -7.50170052e-01
-2.05581054e-01 4.93430316e-01 4.47608829e-01 4.87517193e-02
2.30617374e-01 -1.13153136e+00 -9.01191533e-01 -9.04221535e-02
3.64111736e-02 -1.73060253e-01 -1.35436639e-01 7.04853594e-01
2.03281850e-01 7.16475546e-01 1.01514012e-01 -3.99031341e-01
-1.20943213e+00 5.89978814e-01 4.98155445e-01 -3.81713390e-01
-1.57231584e-01 5.22801220e-01 1.93279311e-01 -1.95822909e-01
5.43520868e-01 -2.26078600e-01 4.85345513e-01 -2.97022462e-01
1.82769001e-01 3.05201292e-01 -1.72450468e-01 1.84643909e-01
-9.59858745e-02 -2.28131264e-01 -2.63975002e-03 -7.62243032e-01
1.33302450e+00 1.10604696e-01 7.47338086e-02 -3.63023318e-02
1.03667378e+00 -2.00469196e-01 -1.78240740e+00 -5.22137880e-01
5.49815416e-01 -6.98347747e-01 -3.06397229e-02 -6.02525711e-01
-1.02262330e+00 4.86078888e-01 3.33308429e-01 4.79683131e-01
1.02836442e+00 1.55608267e-01 1.72108427e-01 5.28836668e-01
7.58825123e-01 -1.43348801e+00 -3.09419304e-01 -5.90950623e-02
5.99429727e-01 -1.16961539e+00 -1.39626220e-01 9.21064764e-02
-7.66585886e-01 6.27986729e-01 2.88749874e-01 1.25454485e-01
7.31809556e-01 5.02233386e-01 -6.85125217e-02 -2.09679580e-04
-1.22393858e+00 1.59866229e-01 -7.97101259e-02 5.22816598e-01
1.58042789e-01 9.26161647e-01 -7.05120802e-01 1.22764444e+00
-3.68311435e-01 2.81246722e-01 5.76914012e-01 1.05685747e+00
-2.68688858e-01 -1.18801570e+00 -3.27086210e-01 6.85328007e-01
-9.02956545e-01 3.26829433e-01 -2.93224573e-01 6.56345367e-01
-3.02474678e-01 8.77794743e-01 1.80748254e-01 -7.40016773e-02
1.34149179e-01 -4.85328212e-02 5.17269671e-01 -3.04679841e-01
-2.80766070e-01 1.93004534e-01 -9.51326340e-02 -5.02006650e-01
-4.64927047e-01 -9.75446403e-01 -9.78086829e-01 -5.20990372e-01
-1.90394565e-01 4.95408565e-01 7.80466497e-01 8.82149577e-01
2.86085308e-01 7.65196010e-02 1.19615066e+00 -2.69346982e-02
-1.99020600e+00 -7.40033507e-01 -1.00531173e+00 4.30719018e-01
2.93201625e-01 -3.92180651e-01 -6.79721177e-01 -2.05992386e-01]
|
[4.534945011138916, 3.285856246948242]
|
b516c676-a160-4f13-b492-031d31a0a9d9
|
a-self-supervised-contrastive-learning-method
|
2306.14437
| null |
https://arxiv.org/abs/2306.14437v1
|
https://arxiv.org/pdf/2306.14437v1.pdf
|
A Self-supervised Contrastive Learning Method for Grasp Outcomes Prediction
|
In this paper, we investigate the effectiveness of contrastive learning methods for predicting grasp outcomes in an unsupervised manner. By utilizing a publicly available dataset, we demonstrate that contrastive learning methods perform well on the task of grasp outcomes prediction. Specifically, the dynamic-dictionary-based method with the momentum updating technique achieves a satisfactory accuracy of 81.83% using data from one single tactile sensor, outperforming other unsupervised methods. Our results reveal the potential of contrastive learning methods for applications in the field of robot grasping and highlight the importance of accurate grasp prediction for achieving stable grasps.
|
['Xinyu Wu', 'Zhengkun Yi', 'Yupo Zhang', 'Ke Mai', 'Yuanzhe Su', 'YiWen Liu', 'Binhua Huang', 'Chengliang Liu']
|
2023-06-26
| null | null | null | null |
['contrastive-learning', 'contrastive-learning']
|
['computer-vision', 'methodology']
|
[ 6.88352808e-02 -3.22517276e-01 -4.84041631e-01 -2.44028717e-01
-5.77764273e-01 -9.21852961e-02 2.07929779e-02 2.59492159e-01
-1.99671164e-01 4.08370972e-01 -6.64170608e-02 2.15024710e-01
-6.24479353e-01 -6.35758400e-01 -7.76791692e-01 -9.92330551e-01
-7.64810801e-01 6.11616075e-01 -4.31233682e-02 -1.44039094e-01
7.14339375e-01 5.96759379e-01 -1.60847199e+00 2.74875343e-01
8.06534886e-01 1.36925149e+00 7.99476922e-01 2.94408917e-01
2.50230670e-01 6.23500884e-01 -2.61819065e-01 -1.36256441e-01
4.67378318e-01 4.05694306e-01 -6.33388460e-01 -1.61261007e-01
1.50204942e-01 -7.52285779e-01 -4.29411083e-01 7.77624249e-01
5.61928213e-01 6.18095323e-02 1.01365852e+00 -1.02204120e+00
-6.88341439e-01 6.39576852e-01 -2.51513839e-01 -7.89108574e-02
4.04775560e-01 5.99775910e-02 1.12696385e+00 -1.31820834e+00
4.38042045e-01 1.08233142e+00 6.27960980e-01 3.92257720e-01
-1.06485415e+00 -3.60671073e-01 -1.51431918e-01 4.32044834e-01
-8.63539279e-01 -2.24177301e-01 1.05229461e+00 -5.22646129e-01
9.07138586e-01 -9.04413685e-03 4.81577218e-01 1.25103104e+00
6.53980076e-01 1.30802333e+00 1.20737064e+00 -4.48736846e-01
3.81248832e-01 -6.83412492e-01 8.93476829e-02 8.50996256e-01
2.42268816e-01 5.59735954e-01 -9.23876107e-01 -1.00522541e-01
9.57029641e-01 2.07052678e-01 1.89308330e-01 -7.94286489e-01
-1.31418502e+00 7.26019979e-01 5.67087412e-01 1.51063994e-01
-1.01151133e+00 1.80927441e-01 4.36926872e-01 2.94876039e-01
3.53341192e-01 6.77820146e-01 -6.72708035e-01 -2.00879216e-01
-5.43348789e-01 2.85762757e-01 1.07185173e+00 1.15424645e+00
4.47824508e-01 7.47211501e-02 -2.22233608e-01 8.33401918e-01
3.03122312e-01 7.38153815e-01 2.77372807e-01 -1.21258759e+00
2.77656347e-01 4.37443197e-01 4.25015956e-01 -1.11536050e+00
-3.90583783e-01 8.70482922e-02 -7.32575536e-01 2.73422837e-01
2.63277620e-01 1.77180052e-01 -1.02130699e+00 1.10513127e+00
-4.94824573e-02 -4.22805846e-01 9.29877609e-02 8.84570539e-01
7.23309517e-01 2.85610586e-01 1.85984716e-01 -2.73685038e-01
3.75544041e-01 -1.04329455e+00 -7.59262085e-01 6.25326261e-02
9.90593899e-03 -8.37743402e-01 1.08901310e+00 8.29278171e-01
-1.11031330e+00 -4.65719163e-01 -8.06123912e-01 3.25329036e-01
-9.94192138e-02 5.72292089e-01 1.22416914e+00 -1.08749717e-01
-7.44250059e-01 1.21973336e+00 -1.20713067e+00 -3.46418768e-01
4.12965238e-01 5.62301874e-01 -1.49507299e-01 -1.95773348e-01
-5.17695427e-01 1.15808415e+00 5.03211200e-01 2.71569669e-01
-1.21455097e+00 -3.09526473e-01 -6.01888180e-01 -7.99318105e-02
1.58662587e-01 1.89682424e-01 1.04562962e+00 -1.13665402e-01
-1.68844211e+00 5.72358429e-01 3.78723480e-02 -2.55940050e-01
3.57409805e-01 -8.08817565e-01 3.63563269e-01 3.51708174e-01
4.10118625e-02 6.04217708e-01 1.13539374e+00 -1.56383336e+00
-3.31297219e-01 -3.41985524e-01 -1.47489354e-01 -2.29104608e-02
-6.34656310e-01 -4.70596284e-01 -1.00290529e-01 -6.70679152e-01
6.34441435e-01 -8.34420085e-01 -1.35135710e-01 5.58357239e-02
-2.84933835e-01 -5.97622991e-01 7.15623140e-01 -7.73931563e-01
6.47995472e-01 -1.89104211e+00 6.19663477e-01 1.99540511e-01
9.20315273e-03 2.67177001e-02 -2.19301373e-01 7.82483637e-01
4.94934112e-01 -6.16601646e-01 -1.31052673e-01 -4.35396284e-02
1.37143329e-01 3.93490463e-01 -4.99926180e-01 1.40456557e-01
2.00423092e-01 1.06414413e+00 -1.22154951e+00 -4.01407152e-01
4.41254497e-01 -2.02218384e-01 -4.43406284e-01 8.90099168e-01
-1.19222283e-01 2.96989053e-01 -6.64513350e-01 1.48601508e+00
5.54708540e-01 -7.54055157e-02 3.13571244e-01 -6.64588630e-01
-2.81581879e-01 -1.76348966e-02 -6.93210244e-01 1.75914240e+00
-1.93878308e-01 2.05745190e-01 1.24181740e-01 -1.32037795e+00
1.39178181e+00 5.43200262e-02 1.10153878e+00 -4.46082681e-01
-4.84262705e-02 5.24615169e-01 -5.14373854e-02 -9.58510160e-01
2.35584363e-01 7.84087256e-02 1.98647484e-01 1.73524722e-01
5.29456079e-01 -4.07133967e-01 5.53765334e-02 -2.32038185e-01
1.08842897e+00 2.15176672e-01 -1.15669645e-01 -6.84364140e-01
-1.97783604e-01 1.60043925e-01 1.86612248e-01 9.72525239e-01
-1.01027250e-01 3.59419137e-01 -1.39903143e-01 -5.03585279e-01
-1.06845796e+00 -1.41226113e+00 -1.65081397e-01 1.17701674e+00
4.96087849e-01 -1.03575312e-01 -1.81666836e-01 -2.65998185e-01
8.68005216e-01 2.65482724e-01 -5.36285639e-01 -2.10207686e-01
-4.84427154e-01 -5.47914505e-01 4.00072187e-02 1.09642923e+00
3.35416585e-01 -1.46443808e+00 -8.33548903e-01 4.93144900e-01
-3.66219543e-02 -8.77630115e-01 2.97949612e-01 6.96148336e-01
-1.27265704e+00 -1.28750718e+00 -7.30968416e-01 -1.19936287e+00
5.71278632e-01 1.94708183e-01 6.83969617e-01 7.45152235e-02
-5.94098270e-01 1.00750899e+00 -6.56552672e-01 -5.28446853e-01
-1.72838375e-01 3.51822153e-02 4.66567308e-01 -7.17351913e-01
1.96912289e-01 -5.33352792e-01 -4.73305136e-01 1.96819514e-01
-3.73767287e-01 -1.69815615e-01 8.83419871e-01 1.12521732e+00
5.02774060e-01 -2.38864213e-01 9.74385083e-01 9.72760748e-03
9.00354385e-01 -4.20773625e-01 -2.54337579e-01 4.83932048e-01
-7.31650829e-01 -7.26298392e-02 2.68341988e-01 -5.93125463e-01
-8.82367134e-01 1.24348521e-01 9.16012153e-02 -5.28873622e-01
9.08723846e-02 7.28060901e-01 3.97388160e-01 -4.25678611e-01
3.03817660e-01 3.77830833e-01 5.12665093e-01 -6.55282319e-01
-8.98847282e-02 5.34391284e-01 4.55959737e-01 -1.13059103e+00
5.09983599e-01 2.09815919e-01 3.16649079e-02 -6.50314212e-01
-5.86759448e-01 -5.00550985e-01 -9.43890154e-01 -5.55305779e-01
4.68349218e-01 -5.82534850e-01 -1.12760127e+00 9.86931324e-01
-9.81160462e-01 -6.76399887e-01 -1.31135955e-01 6.02523267e-01
-1.03041255e+00 4.37089741e-01 -1.07644403e+00 -9.64368880e-01
-6.63965225e-01 -9.77974236e-01 1.12592304e+00 -7.22956508e-02
1.35475531e-01 -5.98036528e-01 -3.26591253e-01 -2.81529278e-02
4.87563699e-01 2.01790795e-01 9.27749097e-01 -4.98061925e-01
-4.46943223e-01 -3.20379347e-01 -4.00128700e-02 5.31006575e-01
7.76864886e-01 -7.02864826e-02 -4.52505648e-01 -7.62134671e-01
5.86469807e-02 -1.03784418e+00 8.51403773e-01 3.69074881e-01
1.29749131e+00 -2.11817190e-01 -3.71293366e-01 -9.41856876e-02
1.32661819e+00 3.33347082e-01 4.43667412e-01 2.48263434e-01
5.70895791e-01 6.25074208e-01 1.12863433e+00 7.55059898e-01
1.18052833e-01 2.95071632e-01 8.02706361e-01 3.63083333e-01
2.43031099e-01 -1.76236600e-01 2.20116545e-02 1.24946153e+00
-6.65610313e-01 1.62509412e-01 -1.04083312e+00 7.08002150e-01
-2.05213618e+00 -6.15034163e-01 4.14346784e-01 1.97186661e+00
8.11891437e-01 -3.69595326e-02 -2.00325176e-01 7.11242855e-02
4.01571155e-01 -1.17999896e-01 -7.98003674e-01 -1.87362179e-01
1.83546707e-01 4.33351636e-01 3.34491581e-01 1.18111581e-01
-1.27428758e+00 1.10940087e+00 7.88127899e+00 5.43814659e-01
-9.43689287e-01 -3.20269167e-01 -5.48448898e-02 3.33845407e-01
8.99640247e-02 -4.38730210e-01 -3.43773007e-01 4.68982995e-01
1.94733609e-02 4.12656777e-02 6.55364871e-01 1.19445252e+00
-8.90317112e-02 -2.48845220e-01 -1.41480660e+00 8.38809431e-01
-1.42895594e-01 -1.14154053e+00 6.33313805e-02 -3.09979051e-01
5.14685392e-01 1.53039813e-01 -3.05609917e-03 1.95435748e-01
2.08733454e-01 -8.92787814e-01 6.23647451e-01 7.45139122e-01
4.07451063e-01 -3.61308992e-01 6.11573339e-01 2.35752985e-01
-9.99158382e-01 -5.67401588e-01 -7.17520177e-01 -2.56304860e-01
1.08266078e-01 4.78449374e-01 -6.97639167e-01 5.06685376e-01
1.19657612e+00 1.06784773e+00 5.09503931e-02 1.03566015e+00
-1.67042106e-01 4.77633297e-01 -2.69555390e-01 -3.36658001e-01
1.34397730e-01 -6.51589036e-02 6.43734753e-01 1.05522990e+00
8.88731554e-02 2.82204807e-01 4.85215276e-01 6.91389143e-01
2.34365121e-01 -1.02362610e-01 -6.66584849e-01 -2.31823191e-01
5.62373519e-01 7.35579550e-01 -3.61810982e-01 3.81138660e-02
-7.11473152e-02 6.77871764e-01 6.02135301e-01 8.48609805e-02
-1.43658832e-01 -1.76956519e-01 2.39330590e-01 -4.95539278e-01
4.82938886e-01 -8.78299296e-01 -3.87913227e-01 -1.04905021e+00
3.99798870e-01 -6.83344841e-01 -1.70671284e-01 -6.36196256e-01
-1.92140293e+00 5.18719107e-02 1.56254262e-01 -1.22011995e+00
-5.43915331e-02 -1.24389434e+00 -4.82818872e-01 2.42272764e-01
-1.40339923e+00 -9.58847523e-01 -5.32187521e-01 3.39982331e-01
7.35238492e-01 -4.80138928e-01 1.16295791e+00 -1.50273770e-01
-6.45458326e-02 2.34223321e-01 4.67079610e-01 -8.81890878e-02
6.04043365e-01 -1.14636981e+00 -8.05939585e-02 -1.23977456e-02
-1.41483843e-01 6.15871012e-01 7.09895730e-01 -7.84009874e-01
-2.29765177e+00 -5.05480468e-01 1.63675651e-01 -1.12526186e-01
8.00589561e-01 7.63361529e-02 -8.61873269e-01 4.18571174e-01
-5.56660220e-02 -3.05526197e-01 2.68196672e-01 1.42830387e-01
9.44655463e-02 -1.09723315e-01 -1.20015872e+00 2.59259671e-01
1.05037212e+00 -2.98138738e-01 -9.88673866e-01 5.79039633e-01
2.09658027e-01 -4.99420285e-01 -1.41917741e+00 1.42875159e+00
1.04009712e+00 -5.03254533e-01 1.14236844e+00 -7.57426083e-01
9.90359604e-01 5.76698601e-01 -5.75155020e-01 -1.01927257e+00
-3.66436988e-01 -9.87392664e-02 -6.57686651e-01 7.57331431e-01
-1.23857938e-01 -3.41368914e-01 8.07219625e-01 5.07232726e-01
-1.33467436e-01 -1.27171767e+00 -6.83034420e-01 -1.08652377e+00
1.91873744e-01 -6.41154125e-02 -2.64390185e-02 6.57825172e-01
3.51725340e-01 -4.71046716e-01 -4.03185725e-01 -8.20303261e-02
1.11232269e+00 5.62004387e-01 3.70819926e-01 -1.36415243e+00
7.72245601e-02 -6.36913404e-02 -2.82384574e-01 -1.18100286e+00
4.59842712e-01 -7.44596601e-01 5.69908619e-01 -1.67557466e+00
4.68120277e-01 -9.42039967e-01 -5.79855263e-01 6.66480362e-01
-6.99529871e-02 -1.22631952e-01 2.58062899e-01 7.62368500e-01
-4.59331840e-01 9.45534348e-01 1.39164376e+00 -5.43241084e-01
1.90547314e-02 -2.52567288e-02 1.31376043e-01 6.45988345e-01
9.01728332e-01 -1.76379964e-01 -8.10810104e-02 -7.08702564e-01
-4.26915169e-01 -1.18491715e-02 5.14078677e-01 -8.33836377e-01
2.89649516e-01 -2.71740198e-01 3.43814611e-01 -6.12337291e-01
3.83668631e-01 -7.57609189e-01 -6.18885994e-01 8.46263766e-01
-3.95557046e-01 -1.15692995e-01 3.03728729e-01 7.73108721e-01
-1.45081878e-01 -4.49557483e-01 1.99069947e-01 -9.93761718e-02
-1.04814696e+00 2.14894801e-01 -4.02787417e-01 -4.28201169e-01
8.15523148e-01 -6.66907430e-02 7.93029461e-03 -5.63636683e-02
-9.16222990e-01 2.17428058e-01 2.21901491e-01 6.40615821e-01
1.14244545e+00 -1.47233284e+00 -4.23296541e-01 2.33979020e-02
4.02023718e-02 -9.76732150e-02 -2.06175774e-01 5.77996612e-01
-4.37736571e-01 2.83324927e-01 -8.02756965e-01 -8.62190723e-01
-7.61436701e-01 3.28565478e-01 -2.19064787e-01 9.09511000e-02
-6.00346029e-01 8.20053160e-01 -4.90635455e-01 -4.92722511e-01
7.00805068e-01 -1.55880854e-01 -3.36639695e-02 -2.54146010e-01
-1.93798363e-01 7.55406976e-01 -2.77864630e-03 1.03675045e-01
-4.21798259e-01 3.99659157e-01 -1.33155853e-01 3.88915569e-01
1.93730128e+00 5.05820513e-01 -4.71426785e-01 5.55917203e-01
9.52509165e-01 -5.64991057e-01 -1.59083390e+00 -7.89655671e-02
3.93621951e-01 -3.81251872e-01 -2.51946807e-01 -9.86179829e-01
-9.60136712e-01 5.09317994e-01 7.61643052e-01 1.34042799e-02
9.01438773e-01 1.94753855e-01 6.90350831e-01 1.15915489e+00
1.31570125e+00 -1.37301004e+00 7.97829628e-01 7.22149372e-01
1.34280813e+00 -1.83372879e+00 1.01350285e-01 -5.47079861e-01
-4.53015685e-01 1.44084549e+00 7.81100273e-01 -6.88226283e-01
8.15977216e-01 4.30277765e-01 -1.45491019e-01 -3.51250023e-01
-4.72440600e-01 1.09627292e-01 2.71664500e-01 5.55729926e-01
2.00792298e-01 3.08962762e-01 -4.82701629e-01 4.14102465e-01
1.29628763e-01 1.54035076e-01 -1.62140913e-02 1.68556225e+00
-5.84448576e-01 -1.01714551e+00 1.16346210e-01 8.03419590e-01
-4.48382273e-02 2.20926739e-02 -2.72760153e-01 7.27505803e-01
-5.21360695e-01 9.18139100e-01 -2.05553591e-01 -6.21177554e-01
3.31445158e-01 -1.07227437e-01 1.24722755e+00 -6.14272892e-01
-2.43510649e-01 -2.55689859e-01 -3.22626889e-01 -7.83133209e-01
-7.33946204e-01 -7.43572652e-01 -1.12838042e+00 7.31028840e-02
-5.44940710e-01 -6.16953336e-02 7.15981960e-01 9.86631453e-01
2.24465355e-01 1.14175893e-01 8.99804950e-01 -1.56103802e+00
-1.36424780e+00 -1.18617916e+00 -7.56450891e-01 3.33481550e-01
-8.59883428e-03 -1.42457294e+00 -2.07492888e-01 -1.18543310e-02]
|
[5.797428131103516, -0.8270968198776245]
|
c241ca00-089b-4534-b63b-9c8288b068bb
|
stsc-snn-spatio-temporal-synaptic-connection
|
2210.05241
| null |
https://arxiv.org/abs/2210.05241v1
|
https://arxiv.org/pdf/2210.05241v1.pdf
|
STSC-SNN: Spatio-Temporal Synaptic Connection with Temporal Convolution and Attention for Spiking Neural Networks
|
Spiking Neural Networks (SNNs), as one of the algorithmic models in neuromorphic computing, have gained a great deal of research attention owing to temporal information processing capability, low power consumption, and high biological plausibility. The potential to efficiently extract spatio-temporal features makes it suitable for processing the event streams. However, existing synaptic structures in SNNs are almost full-connections or spatial 2D convolution, neither of which can extract temporal dependencies adequately. In this work, we take inspiration from biological synapses and propose a spatio-temporal synaptic connection SNN (STSC-SNN) model, to enhance the spatio-temporal receptive fields of synaptic connections, thereby establishing temporal dependencies across layers. Concretely, we incorporate temporal convolution and attention mechanisms to implement synaptic filtering and gating functions. We show that endowing synaptic models with temporal dependencies can improve the performance of SNNs on classification tasks. In addition, we investigate the impact of performance vias varied spatial-temporal receptive fields and reevaluate the temporal modules in SNNs. Our approach is tested on neuromorphic datasets, including DVS128 Gesture (gesture recognition), N-MNIST, CIFAR10-DVS (image classification), and SHD (speech digit recognition). The results show that the proposed model outperforms the state-of-the-art accuracy on nearly all datasets.
|
['Erping Li', 'Aili Wang', 'Gaoang Wang', 'Da Li', 'Zheming Gu', 'Chengting Yu']
|
2022-10-11
| null | null | null | null |
['gesture-recognition']
|
['computer-vision']
|
[ 4.52574253e-01 -6.58650756e-01 2.05149636e-01 -1.94182619e-01
2.36583441e-01 -3.97898585e-01 8.74078989e-01 -3.03520083e-01
-9.51854229e-01 7.74458468e-01 -8.81128162e-02 -2.40051270e-01
-4.35059220e-01 -6.94181681e-01 -7.34793544e-01 -8.80987883e-01
-3.57391000e-01 -3.15530181e-01 1.08597851e+00 -1.33458212e-01
4.65591073e-01 7.14701831e-01 -1.68870664e+00 7.37115443e-01
4.96092618e-01 1.28805614e+00 4.58343178e-01 5.05323052e-01
-3.15817654e-01 4.51583087e-01 -4.00752217e-01 -1.68136969e-01
3.70562784e-02 -4.47906047e-01 -2.67528147e-01 -6.25700176e-01
-1.93205163e-01 1.22355141e-01 -7.34173357e-01 7.08128333e-01
5.61508358e-01 -3.90946828e-02 5.49426734e-01 -1.15340710e+00
-5.93001783e-01 5.35700321e-01 -1.85698703e-01 9.70071733e-01
-2.01219127e-01 4.68872696e-01 5.63314259e-01 -6.96275175e-01
5.89044929e-01 1.03102863e+00 8.32584679e-01 8.83586824e-01
-1.17269635e+00 -7.89741158e-01 1.22115687e-01 3.67641658e-01
-9.24016595e-01 -3.12161744e-01 2.91305810e-01 -2.61423469e-01
1.60006571e+00 -1.52541220e-01 1.14257324e+00 1.52818108e+00
6.82103693e-01 6.52043581e-01 1.20327842e+00 -6.12009829e-03
5.60986102e-01 -7.65616477e-01 3.31793249e-01 1.19225286e-01
5.05794538e-03 2.79729664e-01 -1.27956223e+00 1.65198132e-01
1.23544979e+00 4.54013050e-01 -4.81870808e-02 3.53149593e-01
-1.20312572e+00 1.25561088e-01 7.49745131e-01 4.97853696e-01
-5.11133373e-01 7.13630378e-01 3.51636738e-01 3.04615557e-01
-1.71594650e-01 1.40296835e-02 -3.67390215e-01 -3.71058166e-01
-8.12641978e-01 7.31987581e-02 2.21249193e-01 5.22030294e-01
1.48499772e-01 1.46573797e-01 -3.72466624e-01 6.58888936e-01
2.45898187e-01 4.30353284e-01 9.21512246e-01 -7.07975984e-01
2.06040040e-01 6.96356297e-01 -3.80842984e-01 -2.97397643e-01
-4.78606164e-01 -9.13534239e-02 -9.78165865e-01 4.14300144e-01
5.64082682e-01 3.07344466e-01 -1.24886775e+00 1.79503882e+00
-3.40234876e-01 5.11695147e-01 9.34772640e-02 8.76422703e-01
5.64094603e-01 3.42403829e-01 3.86613131e-01 -7.01930188e-03
1.39386928e+00 -3.37275922e-01 -6.35458350e-01 -1.87173113e-01
-6.39517382e-02 -3.32992733e-01 6.66872382e-01 2.72452682e-01
-1.12141097e+00 -5.48552155e-01 -1.15046728e+00 3.78170311e-02
-6.04065955e-01 -1.83929160e-01 8.69301140e-01 3.98653597e-01
-1.16463518e+00 1.14741266e+00 -1.44098151e+00 -5.76409996e-01
8.21541011e-01 8.49659622e-01 -1.78368852e-01 3.01603764e-01
-1.19238627e+00 6.86093271e-01 2.50643283e-01 9.94005203e-02
-8.77562344e-01 -4.78180707e-01 -2.25002334e-01 6.16771281e-02
-5.36404848e-01 -6.84065402e-01 9.91340637e-01 -5.80416977e-01
-1.37802684e+00 4.99282956e-01 -3.74902695e-01 -9.00152922e-01
5.84889911e-02 3.38096589e-01 -3.89152825e-01 2.44947776e-01
-4.28484350e-01 1.04177296e+00 5.20282686e-01 -3.31119061e-01
-4.31087613e-01 -7.09399223e-01 -1.78437307e-01 -1.56513453e-01
-6.03926241e-01 8.00473168e-02 -2.29641542e-01 -7.46699154e-01
3.59923393e-01 -7.63145804e-01 -6.52380362e-02 3.54165882e-01
3.71626094e-02 -3.29042941e-01 7.48683929e-01 -2.87735045e-01
1.07499218e+00 -2.38506007e+00 6.51410595e-02 -2.16462892e-02
-3.66818309e-02 4.13038701e-01 -1.46345034e-01 2.45971560e-01
4.39764783e-02 -3.52078751e-02 -3.77774805e-01 -1.76130503e-01
-2.04020455e-01 5.60459018e-01 -4.16598141e-01 1.17482975e-01
6.43156052e-01 1.26942861e+00 -5.44332445e-01 -9.86670256e-02
3.40670273e-02 7.53004491e-01 -2.74308711e-01 -3.72129947e-01
-2.15066284e-01 4.85792011e-01 -2.50345856e-01 6.10038102e-01
4.50564891e-01 -2.17258379e-01 5.13153970e-02 -4.35902737e-02
-4.64919358e-01 6.28760517e-01 -6.88639998e-01 1.88043368e+00
-1.38905412e-02 7.58933961e-01 -5.00416577e-01 -9.70892966e-01
1.01572430e+00 3.84556592e-01 4.62992847e-01 -1.46737409e+00
2.33702302e-01 4.65074837e-01 5.49564898e-01 -2.94117689e-01
-1.14441395e-01 -9.53780040e-02 2.39792123e-01 3.44674736e-01
4.25902873e-01 5.13111413e-01 2.05434322e-01 -1.86957847e-02
1.59698594e+00 1.28873885e-01 -3.19089532e-01 -1.90406516e-01
4.79896599e-03 -3.52481186e-01 4.96029258e-01 6.01412714e-01
-3.71318042e-01 5.67805231e-01 3.06428403e-01 -4.23439234e-01
-9.04975951e-01 -1.41429937e+00 -3.54860246e-01 8.91536057e-01
1.51288509e-01 -2.26757210e-02 -6.20532095e-01 -6.11647591e-02
5.43253087e-02 1.33922234e-01 -8.55930209e-01 -2.68453956e-01
-6.94675863e-01 -1.04651773e+00 1.18669665e+00 1.03790355e+00
6.90869212e-01 -1.57261014e+00 -1.35587347e+00 5.23669720e-01
2.60324001e-01 -1.06341684e+00 -7.41525963e-02 1.03989851e+00
-1.34857845e+00 -8.55936468e-01 -5.97529829e-01 -8.10218394e-01
2.40670502e-01 8.40962399e-03 5.25435984e-01 -3.04714501e-01
-6.39650285e-01 2.44200019e-05 -7.10596591e-02 -4.40623522e-01
3.80404681e-01 -1.23986520e-01 3.21392529e-02 -1.27306610e-01
6.56448781e-01 -1.35252976e+00 -7.68067777e-01 3.83602411e-01
-1.14865017e+00 8.13660920e-02 6.65523946e-01 7.45221913e-01
7.58039057e-01 -3.59015286e-01 7.68423975e-01 -3.45660895e-01
3.99906188e-01 -2.78178841e-01 -3.00718099e-01 3.19326408e-02
-3.57776970e-01 3.79616737e-01 6.33199930e-01 -8.73574078e-01
-8.86370957e-01 -1.14626601e-01 -1.00242488e-01 -3.40246826e-01
-7.57179260e-02 2.69266427e-01 2.53115594e-01 -1.16963349e-01
6.90079272e-01 6.33389652e-01 -1.34053737e-01 -3.76951873e-01
-3.46947163e-01 2.26130694e-01 6.09285593e-01 -4.08158004e-01
1.18486762e-01 8.79998147e-01 1.11121766e-01 -6.08622015e-01
2.97159329e-02 -2.79230654e-01 -6.41735077e-01 -9.03142989e-02
8.95737410e-01 -5.31901419e-01 -8.95933926e-01 9.42378581e-01
-1.41050541e+00 -5.33648014e-01 -2.10290313e-01 5.73938191e-01
-4.91710961e-01 -1.75147921e-01 -1.04835284e+00 -8.81038666e-01
-2.99622715e-01 -8.10865462e-01 7.33624160e-01 7.16282904e-01
-4.10161614e-02 -5.80055058e-01 -3.14819068e-02 -4.06674564e-01
7.41569340e-01 -3.25227827e-02 9.36300576e-01 -3.80839378e-01
-6.49848223e-01 2.94541091e-01 -3.59678298e-01 -8.58365148e-02
-1.54868007e-01 -4.14016768e-02 -1.25608516e+00 1.26087904e-01
-2.00502440e-01 -2.03678787e-01 1.45896780e+00 5.83367825e-01
1.24402106e+00 2.19733000e-01 -3.96008641e-01 6.68278873e-01
1.34083617e+00 5.89195013e-01 1.09941363e+00 1.56528458e-01
-2.39970651e-03 4.21202242e-01 -1.93454564e-01 5.24407506e-01
-2.06787791e-02 4.19235498e-01 6.29381537e-01 3.54799151e-01
-4.40275759e-01 -9.20352340e-02 4.36427534e-01 7.44879961e-01
-5.86841643e-01 -1.19123660e-01 -7.44698882e-01 6.62948310e-01
-2.01472330e+00 -1.05711305e+00 -1.90182760e-01 1.97784352e+00
8.24707329e-01 3.47231597e-01 -9.82242674e-02 3.49745333e-01
5.81492305e-01 -1.22196518e-01 -9.50577974e-01 -3.97009611e-01
-7.44658947e-01 8.55324447e-01 3.31018299e-01 -3.68276477e-01
-6.83208406e-01 7.06554949e-01 5.72638083e+00 5.55070877e-01
-1.33166659e+00 9.85475034e-02 3.44220579e-01 -4.46070045e-01
-8.09206963e-02 -3.75725031e-01 -9.17813778e-01 6.82427049e-01
1.25364125e+00 2.61705667e-01 7.43684471e-01 3.64536904e-02
2.08514258e-01 8.02650210e-03 -1.09016871e+00 1.05960226e+00
-5.16529322e-01 -1.59327626e+00 -2.68278308e-02 3.01774684e-02
5.61564624e-01 4.31767344e-01 2.91091591e-01 2.02462777e-01
3.30360048e-02 -1.26226902e+00 7.13119686e-01 7.42379069e-01
7.20253885e-01 -4.20805424e-01 5.40212214e-01 2.53575921e-01
-1.34207416e+00 -2.50768483e-01 -4.28523242e-01 -2.74298877e-01
5.50361956e-03 3.29979897e-01 -1.57432005e-01 -1.28930770e-02
1.28603852e+00 8.70959699e-01 -4.51560229e-01 1.45472729e+00
-3.98942502e-03 5.08149743e-01 -5.72573006e-01 -5.43654919e-01
2.30279818e-01 2.37017035e-01 2.83155628e-02 1.35028529e+00
4.47618186e-01 4.61007535e-01 -6.84039414e-01 1.02556705e+00
-2.06245840e-01 -5.21061599e-01 -4.10117865e-01 -3.18590939e-01
4.68160152e-01 8.60213518e-01 -1.11465847e+00 1.90128982e-02
-2.54720926e-01 1.04036176e+00 2.90150434e-01 4.68118072e-01
-7.72879779e-01 -4.35785532e-01 8.18840265e-01 1.62487645e-02
6.03030264e-01 -4.17052418e-01 -7.35724509e-01 -7.75984108e-01
1.61058664e-01 -1.88383982e-01 1.28362522e-01 -6.79415286e-01
-1.15699196e+00 6.58811748e-01 -4.08690155e-01 -9.89726186e-01
6.42236918e-02 -1.15803552e+00 -7.81029880e-01 7.69488335e-01
-1.55140615e+00 -7.30943441e-01 -1.21962771e-01 8.89149964e-01
3.76487255e-01 1.50109366e-01 7.76643634e-01 4.80021954e-01
-3.01555753e-01 3.19692492e-01 -6.87488243e-02 1.31836191e-01
3.93447161e-01 -8.81338239e-01 6.59936011e-01 8.28516901e-01
1.76513359e-01 9.35293972e-01 1.49951279e-01 -4.91858512e-01
-1.44336390e+00 -1.08038545e+00 9.24060643e-01 2.94195430e-04
7.07859874e-01 -4.77324098e-01 -9.92055714e-01 2.04413384e-01
9.46251899e-02 3.73844236e-01 5.47870457e-01 -4.21736866e-01
-6.34469926e-01 -4.27427813e-02 -1.02594042e+00 6.07278705e-01
1.65889537e+00 -5.37285566e-01 -6.79609239e-01 -2.24846840e-01
4.79595602e-01 1.38315991e-01 -5.74118853e-01 4.71859723e-01
9.97426569e-01 -1.13168538e+00 8.98173153e-01 -6.16845489e-01
3.57278556e-01 -3.44859540e-01 -2.55034596e-01 -8.33655477e-01
-2.17414349e-01 -3.02842557e-01 -1.72723904e-01 8.56360257e-01
3.70206594e-01 -7.28154778e-01 7.90411890e-01 3.94152850e-01
-2.77795911e-01 -9.22725797e-01 -1.37677252e+00 -1.01462245e+00
-5.29213250e-02 -6.47460580e-01 3.00184131e-01 3.14373195e-01
1.44404203e-01 4.47197817e-03 2.14043915e-01 -3.54072452e-02
3.57585877e-01 -1.41111210e-01 -1.69097811e-01 -1.23682868e+00
-4.30157959e-01 -8.09654951e-01 -7.00147152e-01 -1.00318980e+00
-2.44525954e-01 -8.38715494e-01 4.12809029e-02 -1.47410476e+00
1.39724225e-01 -3.23073357e-01 -9.73377287e-01 8.18414569e-01
2.80576587e-01 5.86324275e-01 1.61384735e-02 2.74806529e-01
-5.69961488e-01 4.65541691e-01 1.00627100e+00 -2.41291821e-02
-9.09320861e-02 -7.66332597e-02 -1.11524619e-01 4.88489181e-01
8.33199441e-01 -4.81328279e-01 -2.44144261e-01 -6.86733007e-01
-4.04667063e-03 -1.63048729e-01 7.26527095e-01 -1.42817807e+00
8.70986938e-01 1.81462541e-01 8.28007996e-01 -4.06784832e-01
5.76879561e-01 -4.98709530e-01 9.05243456e-02 8.82271230e-01
-3.96860421e-01 1.53736070e-01 5.04200757e-01 7.05336213e-01
-1.79967880e-01 6.42491691e-03 7.13326275e-01 -5.08924872e-02
-8.74317229e-01 4.42643046e-01 -8.37071598e-01 -2.51261890e-01
7.07439959e-01 -6.40541613e-01 -6.00160480e-01 4.98516053e-01
-6.54240489e-01 -1.95960075e-01 5.26650921e-02 3.57828707e-01
9.91689265e-01 -1.25365567e+00 -2.07089946e-01 4.27140743e-01
-1.86263472e-02 -4.21813726e-01 2.62391746e-01 9.60489810e-01
-1.57926992e-01 5.98775446e-01 -9.31429148e-01 -7.24470317e-01
-8.93366575e-01 3.25861990e-01 2.50696480e-01 -4.64428999e-02
-4.57148224e-01 1.15000498e+00 2.02797931e-02 -1.97622459e-03
4.17171925e-01 -7.80302584e-01 -2.79080749e-01 -1.81175351e-01
5.44441879e-01 1.54136643e-01 2.23090336e-01 3.67358066e-02
-6.04656637e-01 4.38484937e-01 1.66107088e-01 -4.07904595e-01
1.64571214e+00 3.15113157e-01 -1.12174705e-01 6.08023882e-01
8.56300414e-01 -8.19070756e-01 -1.65527809e+00 -1.45047471e-01
2.10878789e-01 1.31041005e-01 -2.84577519e-01 -9.36168671e-01
-1.07089591e+00 1.46507406e+00 7.89084196e-01 -2.90249512e-02
1.33619463e+00 -1.88124374e-01 9.40545261e-01 4.55082715e-01
6.75441504e-01 -8.32492173e-01 3.77418935e-01 7.39122927e-01
5.32796979e-01 -6.14111364e-01 -7.25933909e-01 -3.32966968e-02
-2.16718584e-01 1.46858931e+00 6.08501077e-01 -4.50944334e-01
7.11281061e-01 7.06150770e-01 -4.31296915e-01 -1.08640946e-01
-1.14836848e+00 -2.27672502e-01 2.53216997e-02 5.70135772e-01
3.86211336e-01 -1.59901261e-01 -3.26488435e-01 1.03085208e+00
2.19322234e-01 3.38938475e-01 -4.83582243e-02 1.00736630e+00
-4.53397185e-01 -1.08801067e+00 8.26889351e-02 6.53961718e-01
-4.83402789e-01 -4.07358259e-01 -2.34801635e-01 4.66566473e-01
3.33813936e-01 5.93972683e-01 3.16004544e-01 -5.75936377e-01
2.83098668e-01 2.57422179e-01 8.74477923e-01 -2.65513897e-01
-9.85303521e-01 -2.23893970e-02 -4.37572241e-01 -5.56227922e-01
-4.24554557e-01 -7.25178063e-01 -1.97684574e+00 -6.88547716e-02
6.16937838e-02 -3.20463747e-01 9.68411744e-01 1.02828586e+00
6.19953990e-01 9.05853629e-01 3.79736559e-03 -8.40093136e-01
-2.67873913e-01 -8.33219111e-01 -3.76663208e-01 3.37267280e-01
5.17408699e-02 -6.31202757e-01 -5.29110841e-02 2.19535738e-01]
|
[8.227150917053223, 2.4232289791107178]
|
fd2b267f-26de-41ef-8de7-399d2f4fb0d4
|
towards-deep-and-representation-learning-for
|
1809.06473
| null |
http://arxiv.org/abs/1809.06473v1
|
http://arxiv.org/pdf/1809.06473v1.pdf
|
Towards Deep and Representation Learning for Talent Search at LinkedIn
|
Talent search and recommendation systems at LinkedIn strive to match the
potential candidates to the hiring needs of a recruiter or a hiring manager
expressed in terms of a search query or a job posting. Recent work in this
domain has mainly focused on linear models, which do not take complex
relationships between features into account, as well as ensemble tree models,
which introduce non-linearity but are still insufficient for exploring all the
potential feature interactions, and strictly separate feature generation from
modeling. In this paper, we present the results of our application of deep and
representation learning models on LinkedIn Recruiter. Our key contributions
include: (i) Learning semantic representations of sparse entities within the
talent search domain, such as recruiter ids, candidate ids, and skill entity
ids, for which we utilize neural network models that take advantage of LinkedIn
Economic Graph, and (ii) Deep models for learning recruiter engagement and
candidate response in talent search applications. We also explore learning to
rank approaches applied to deep models, and show the benefits for the talent
search use case. Finally, we present offline and online evaluation results for
LinkedIn talent search and recommendation systems, and discuss potential
challenges along the path to a fully deep model architecture. The challenges
and approaches discussed generalize to any multi-faceted search engine.
|
['Cagri Ozcaglar', 'Rohan Ramanath', 'Krishnaram Kenthapadi', 'Xianren Wu', 'Qi Guo', 'Hakan Inan', 'Bo Hu', 'Sahin Cem Geyik', 'Gungor Polatkan']
|
2018-09-17
| null | null | null | null |
['learning-semantic-representations']
|
['methodology']
|
[-4.30608302e-01 1.42300174e-01 -7.26234257e-01 -3.74178320e-01
-3.55055898e-01 -4.97817308e-01 7.67797828e-01 1.89421952e-01
-3.47135395e-01 4.00413305e-01 6.11187696e-01 -3.89273643e-01
-1.11500657e+00 -1.14805484e+00 -4.57337826e-01 2.36677721e-01
5.83421104e-02 1.42414320e+00 -2.98908651e-01 -7.06579745e-01
3.40628326e-01 5.73366523e-01 -1.50798106e+00 1.29942074e-01
8.39500427e-01 9.05595899e-01 8.95341742e-04 4.51619208e-01
-1.47251785e-01 8.22862446e-01 -4.35102254e-01 -7.28486836e-01
3.67270172e-01 1.18290029e-01 -1.27508664e+00 -4.29877490e-01
1.03204763e+00 -7.43965209e-02 -9.71617043e-01 5.11947572e-01
6.32259488e-01 5.28756142e-01 6.14457190e-01 -9.09006536e-01
-1.12765408e+00 7.39709258e-01 -1.20887868e-01 6.11227155e-01
1.70666158e-01 -1.44971147e-01 1.66931868e+00 -9.84298110e-01
6.62264585e-01 1.33781135e+00 7.82569230e-01 2.23994002e-01
-8.95001292e-01 -6.85837924e-01 3.04172486e-01 1.86627433e-01
-9.53932106e-01 -2.21514612e-01 3.85755986e-01 -5.50157905e-01
9.67278063e-01 2.37121210e-01 9.14105713e-01 1.06429601e+00
2.23869473e-01 7.94763207e-01 1.11195433e+00 -1.43022388e-01
-3.62219244e-01 -3.77760381e-02 8.74175608e-01 1.09577954e+00
4.01620060e-01 4.57316130e-01 -8.90451133e-01 -4.66415346e-01
1.02685571e+00 6.07748330e-01 3.65338236e-01 -1.90000743e-01
-9.06223953e-01 1.11433911e+00 7.93376863e-01 3.98061126e-01
-5.42890906e-01 4.26046491e-01 -1.11297593e-01 5.65433443e-01
5.57081401e-01 1.31775570e+00 -3.15886647e-01 1.64850801e-01
-1.22246695e+00 4.05116767e-01 9.38540339e-01 9.34021711e-01
1.11598134e+00 3.16903032e-02 -9.06744242e-01 8.24865341e-01
2.37825289e-02 2.54032105e-01 4.64048594e-01 -8.23921800e-01
1.02960281e-01 8.92318606e-01 -1.91687241e-01 -1.11600900e+00
-6.51676595e-01 -1.42932606e+00 -2.17123762e-01 -5.06960034e-01
2.07766984e-02 -1.47741735e-01 -1.12276661e+00 1.33577359e+00
-3.52122039e-01 2.17101395e-01 -3.64386350e-01 9.77816343e-01
1.39677334e+00 9.37930793e-02 2.67184108e-01 3.53800416e-01
1.28417826e+00 -1.12885463e+00 -4.97771770e-01 -6.89805329e-01
4.01693135e-01 -6.39817774e-01 8.75192106e-01 -1.14392273e-01
-1.43679953e+00 -7.05194592e-01 -4.34649318e-01 -3.33647519e-01
-6.51795924e-01 1.84075087e-01 1.36109221e+00 4.64064181e-01
-1.21489620e+00 7.62650311e-01 -4.36363399e-01 -5.07064342e-01
4.80616391e-01 8.19787323e-01 1.17946364e-01 -2.72349834e-01
-1.67545629e+00 1.03921950e+00 -1.38996407e-01 -8.09378102e-02
-9.82130229e-01 -7.69167244e-01 -4.67734158e-01 7.39944220e-01
4.34816271e-01 -1.15080285e+00 1.10022461e+00 -4.26490873e-01
-7.61793435e-01 8.70346844e-01 3.21062505e-02 -3.46961915e-01
8.16271082e-02 -3.60818863e-01 -4.88709986e-01 -4.38827187e-01
2.59521097e-01 3.57860416e-01 1.13407314e-01 -8.79558682e-01
-5.50888717e-01 -5.26685238e-01 4.42954391e-01 5.34028113e-01
-6.66936994e-01 -9.33728274e-03 -8.00378561e-01 -4.26295727e-01
-9.33373198e-02 -9.56962824e-01 -4.85718697e-01 -9.24731553e-01
-1.79381415e-01 -6.90676332e-01 1.07529849e-01 -3.97600144e-01
1.48952758e+00 -1.31862414e+00 2.33013019e-01 6.80812299e-01
5.46790302e-01 1.18742719e-01 -3.89384598e-01 5.40426672e-01
2.48376653e-01 2.51121402e-01 9.33167279e-01 -1.66422978e-01
1.05374761e-01 -5.81980534e-02 -1.92572564e-01 -2.61174981e-02
-1.51253402e-01 1.57377279e+00 -1.10477901e+00 -3.60346586e-01
-4.03980732e-01 -5.35091683e-02 -5.77757418e-01 6.20399415e-02
9.81379896e-02 -1.65994644e-01 -9.35481012e-01 1.04252005e+00
-1.59624517e-02 -5.75204313e-01 1.21294923e-01 -3.31280008e-02
1.23597123e-01 6.79704607e-01 -6.20083034e-01 1.44888830e+00
-7.09301829e-01 4.71523941e-01 2.58901119e-01 -8.95748496e-01
9.71488297e-01 -1.49572209e-01 6.20057702e-01 -9.17647302e-01
-1.98362932e-01 3.38738680e-01 -5.49370050e-02 -1.29305080e-01
1.21246517e+00 5.16768545e-02 -3.09022248e-01 7.01603293e-01
3.51698548e-01 2.28705496e-01 1.50797993e-01 6.06006920e-01
1.40412128e+00 -2.73611993e-01 -3.78973544e-01 -5.14510274e-01
-1.27692014e-01 2.27558553e-01 2.70885080e-01 1.36557150e+00
2.66171753e-01 1.10511974e-01 2.48255014e-01 -6.22614145e-01
-6.21690929e-01 -6.43552959e-01 1.03719771e-01 1.83939910e+00
1.01210229e-01 -6.05399787e-01 4.03316598e-03 -7.36738741e-01
5.34601331e-01 3.34527969e-01 -6.24940157e-01 -4.26519424e-01
-5.54123938e-01 -5.21548331e-01 3.17271471e-01 5.70528686e-01
1.83567122e-01 -9.98504877e-01 4.37225811e-02 7.45204836e-02
-1.36525065e-01 -5.59067667e-01 -5.69078863e-01 2.92490840e-01
-8.81845117e-01 -1.04820740e+00 -5.43056548e-01 -8.57423842e-01
3.02120179e-01 3.49957168e-01 1.77344429e+00 5.32319665e-01
-2.84440279e-01 1.10220444e+00 8.61184075e-02 -2.97041863e-01
2.86391824e-01 7.89306760e-01 2.92171717e-01 -5.04778326e-01
6.74628198e-01 -2.45245934e-01 -9.02874172e-01 4.04725701e-01
-3.87110591e-01 -3.43447447e-01 9.16206479e-01 8.91384959e-01
4.11658138e-01 -9.93705988e-02 7.62420774e-01 -1.46299851e+00
1.32444036e+00 -9.68113482e-01 -1.06907077e-01 5.80861688e-01
-1.40252995e+00 -9.79377106e-02 1.74327657e-01 -2.92028487e-01
-7.05207050e-01 -3.70209664e-01 2.18506038e-01 -5.54854989e-01
5.16894102e-01 1.05211878e+00 5.38567781e-01 -1.88508227e-01
9.51720119e-01 -9.83908996e-02 -3.35638970e-01 -6.12658441e-01
5.44206321e-01 3.66378874e-01 2.87388980e-01 -7.46415675e-01
8.44798923e-01 2.67035067e-01 6.00232668e-02 -2.66011864e-01
-1.15272033e+00 -8.02848577e-01 -3.68316054e-01 -1.23971611e-01
3.19598407e-01 -1.13644481e+00 -8.11143816e-01 -2.09753409e-01
-6.48137152e-01 -7.86479563e-02 -4.53424245e-01 4.35781866e-01
-2.36448124e-01 -1.93178222e-01 -9.36884701e-01 -5.19033551e-01
-5.26476204e-01 -8.71254742e-01 9.13160264e-01 4.65290993e-01
-1.56831518e-01 -1.18421209e+00 1.21966880e-02 9.19773698e-01
9.13414776e-01 -4.34755653e-01 9.68899846e-01 -1.18729973e+00
-9.52315271e-01 -5.17210245e-01 -2.78144985e-01 -4.05207813e-01
-1.72198504e-01 -8.58604968e-01 -6.13003433e-01 -5.66206992e-01
-7.77419984e-01 -6.31422222e-01 1.39258099e+00 4.42246884e-01
1.04469311e+00 -3.45471412e-01 -8.48174334e-01 5.37286997e-01
8.59865487e-01 -3.57893586e-01 9.99865830e-02 2.74638653e-01
8.97812843e-01 7.37765670e-01 7.44681895e-01 2.24917680e-01
5.29228985e-01 7.80322552e-01 3.13348681e-01 -3.84742618e-01
-3.06517184e-01 -6.87919080e-01 -5.54038808e-02 3.41768533e-01
-6.01552129e-01 -1.84026763e-01 -8.21177006e-01 6.87477291e-01
-1.91440761e+00 -1.13162410e+00 -5.94495656e-03 1.92765403e+00
6.01274312e-01 -1.94492146e-01 2.00583547e-01 -7.46941030e-01
5.10269403e-01 1.60686702e-01 -6.84510529e-01 -2.74038374e-01
1.93693116e-02 7.57184505e-01 7.71620810e-01 5.44027865e-01
-9.23394084e-01 1.37562943e+00 6.69442511e+00 8.79586875e-01
-5.89097142e-01 1.38097554e-01 7.16303289e-01 -3.09285969e-01
-7.56727159e-01 2.24371672e-01 -1.17690718e+00 -5.57214096e-02
8.90249372e-01 -4.75742519e-01 6.89878285e-01 9.38784003e-01
-2.86148310e-01 6.26140594e-01 -1.28154528e+00 6.98763192e-01
1.42625943e-01 -1.64355314e+00 6.40965551e-02 6.08937800e-01
1.05887616e+00 1.50834814e-01 4.93138492e-01 1.09318161e+00
9.96736884e-01 -1.63375700e+00 1.46449968e-01 9.73805606e-01
5.63126385e-01 -6.53863668e-01 6.74916387e-01 2.15709940e-01
-1.08098912e+00 -7.11989343e-01 -5.30763268e-01 -1.69339523e-01
-3.40949267e-01 5.21443665e-01 -1.09667194e+00 5.93264580e-01
6.77961767e-01 8.06644619e-01 -7.32646048e-01 8.79641652e-01
1.76390648e-01 7.04179943e-01 4.20450754e-02 -1.90226957e-01
3.04258198e-01 -4.83013652e-02 2.18350798e-01 1.17909944e+00
5.08111119e-01 -3.57130207e-02 4.52321380e-01 1.12958121e+00
-6.50311708e-01 1.82081565e-01 -5.95642328e-01 -3.73017788e-01
5.51739454e-01 1.51007259e+00 -2.34570742e-01 -1.55601680e-01
-3.51895958e-01 6.31985068e-01 6.20373666e-01 6.05142176e-01
-1.80468574e-01 1.03142690e-02 6.21805549e-01 6.89826429e-01
-1.66486949e-01 -8.86785835e-02 -1.43671498e-01 -9.03054893e-01
-5.18715441e-01 -9.47627604e-01 7.50775158e-01 -3.87401670e-01
-1.70853376e+00 4.31013107e-01 -1.57134727e-01 -6.23918593e-01
-2.57876664e-01 -4.82889742e-01 -6.30205154e-01 1.31905079e+00
-1.57244444e+00 -1.40618050e+00 -4.25221771e-01 4.77704018e-01
4.04947609e-01 -7.96131313e-01 7.08982229e-01 4.39563781e-01
-2.86369383e-01 5.78133821e-01 -1.59761161e-01 3.33732553e-02
7.10238099e-01 -1.10869980e+00 5.11463463e-01 1.90903947e-01
5.01117706e-01 1.29881525e+00 8.15729126e-02 -9.48620558e-01
-1.60214305e+00 -9.42598641e-01 1.35697901e+00 -9.86166120e-01
6.26789927e-01 -1.21701837e-01 -3.43151152e-01 8.72483909e-01
9.79602933e-02 -1.35445073e-01 7.96314359e-01 1.36952698e+00
4.68347073e-02 6.84809610e-02 -5.11534929e-01 2.52386808e-01
1.57318497e+00 -6.19362831e-01 -4.51412588e-01 7.71802664e-01
4.29573387e-01 -3.19605112e-01 -1.03282082e+00 3.51180077e-01
4.33302730e-01 -7.38176167e-01 1.51263571e+00 -1.27259195e+00
5.12712955e-01 4.14228022e-01 2.99355060e-01 -1.34467268e+00
-1.27840066e+00 -4.10918206e-01 -3.73825878e-01 8.90134275e-01
5.07130563e-01 -2.45685667e-01 1.20906901e+00 8.79312694e-01
-3.15781772e-01 -1.14890492e+00 -4.63984638e-01 -4.52315897e-01
1.38667494e-01 3.79817933e-01 7.47609317e-01 1.23839068e+00
-2.27462336e-01 7.97510624e-01 -5.21649957e-01 -1.81060821e-01
2.28548899e-01 6.38752162e-01 5.07707417e-01 -2.06056905e+00
-5.60354114e-01 -8.32666814e-01 7.18976855e-02 -1.26556695e+00
4.13570851e-01 -1.50158679e+00 -6.01927042e-01 -2.13663769e+00
5.47735095e-01 -1.01836085e+00 -7.37725258e-01 6.04204535e-01
-4.14608032e-01 7.33779296e-02 1.55769633e-02 4.97939944e-01
-7.87578464e-01 3.81232530e-01 1.36577380e+00 -3.89575928e-01
-6.60966486e-02 5.86806238e-01 -1.41866648e+00 3.36422056e-01
2.87276685e-01 -2.06132919e-01 -5.39333344e-01 -7.12125361e-01
8.34691942e-01 1.44374728e-01 3.27385396e-01 -5.42035520e-01
5.67281842e-01 -3.22520673e-01 6.11286461e-01 -2.76321024e-01
5.70253551e-01 -5.73329985e-01 -1.62162259e-01 2.20720872e-01
-7.26159155e-01 2.07514629e-01 -1.14124201e-01 7.45378196e-01
-7.20469654e-02 -1.85488597e-01 3.24712545e-02 -4.73368376e-01
-6.35551035e-01 8.12920272e-01 4.83983010e-02 -2.39201635e-02
2.82060683e-01 3.81308906e-02 -8.05924058e-01 -7.17051446e-01
-8.97997975e-01 8.35327387e-01 1.91104114e-01 6.38574600e-01
5.25256395e-01 -1.42065525e+00 -8.28619063e-01 -1.24631792e-01
4.12379839e-02 -4.86935616e-01 1.38988150e-02 5.28090060e-01
1.10312412e-02 8.39150429e-01 -8.79000649e-02 1.79123104e-01
-1.03698170e+00 3.73746157e-01 2.95907676e-01 -1.01467359e+00
1.47105128e-01 1.19477046e+00 6.41855374e-02 -7.73899257e-01
1.89982548e-01 3.66889566e-01 -7.06312537e-01 1.76553190e-01
-2.00390473e-01 6.27056956e-01 1.99936569e-01 -4.62015033e-01
-2.61173099e-01 4.60182995e-01 -3.36365998e-01 2.82918602e-01
1.35856855e+00 2.32314974e-01 2.22694613e-02 -1.12617441e-01
6.26683235e-01 2.20467821e-01 -3.08690518e-01 -6.00701928e-01
2.06683561e-01 -4.13845032e-01 3.79978091e-01 -1.07216799e+00
-1.19498944e+00 5.98766446e-01 4.94176090e-01 2.14278638e-01
5.68581462e-01 4.41365391e-01 7.39617705e-01 7.50105917e-01
2.35495657e-01 -1.21206570e+00 4.84407425e-01 7.84096539e-01
6.27021253e-01 -1.12793422e+00 3.27708542e-01 -6.79794103e-02
-5.22971272e-01 8.25247526e-01 9.60842490e-01 -4.18031290e-02
5.55056930e-01 -4.93362933e-01 -1.67197064e-01 -1.01408994e+00
-9.35798466e-01 -5.62713265e-01 9.27335560e-01 2.20125556e-01
9.85734999e-01 1.52199358e-01 -4.86466438e-01 7.89134324e-01
-5.31536460e-01 -1.38984248e-01 -2.40023471e-02 5.09032607e-01
-6.08755648e-01 -1.36811519e+00 -2.91375928e-02 1.29260409e+00
-4.52363104e-01 -5.37581444e-01 -9.03967500e-01 3.08791369e-01
1.35041684e-01 9.84552801e-01 6.30134419e-02 -6.72293723e-01
5.40030122e-01 5.53465113e-02 2.56078929e-01 -9.81505752e-01
-1.23111820e+00 -2.53488839e-01 5.34522355e-01 -6.05738938e-01
5.15021943e-02 -4.06740397e-01 -8.37700069e-01 -5.15073061e-01
-4.62668478e-01 6.27525866e-01 4.17151153e-01 5.41720390e-01
8.13173950e-01 6.98187292e-01 3.70750487e-01 -4.74556595e-01
-6.22169435e-01 -1.20306599e+00 -6.69941306e-01 3.81257653e-01
-1.35423854e-01 -7.21464455e-01 6.80652335e-02 -7.05465317e-01]
|
[10.432619094848633, 6.2814621925354]
|
4197cf5f-48e7-4622-bf1e-3e193b179808
|
lectures-on-jacques-herbrand-as-a-logician
|
0902.4682
| null |
http://arxiv.org/abs/0902.4682v5
|
http://arxiv.org/pdf/0902.4682v5.pdf
|
Lectures on Jacques Herbrand as a Logician
|
We give some lectures on the work on formal logic of Jacques Herbrand, and
sketch his life and his influence on automated theorem proving. The intended
audience ranges from students interested in logic over historians to logicians.
Besides the well-known correction of Herbrand's False Lemma by Goedel and
Dreben, we also present the hardly known unpublished correction of Heijenoort
and its consequences on Herbrand's Modus Ponens Elimination. Besides Herbrand's
Fundamental Theorem and its relation to the Loewenheim-Skolem-Theorem, we
carefully investigate Herbrand's notion of intuitionism in connection with his
notion of falsehood in an infinite domain. We sketch Herbrand's two proofs of
the consistency of arithmetic and his notion of a recursive function, and last
but not least, present the correct original text of his unification algorithm
with a new translation.
|
['Christoph Benzmueller', 'Claus-Peter Wirth', 'Serge Autexier', 'Joerg Siekmann']
|
2009-02-26
| null | null | null | null |
['formal-logic']
|
['reasoning']
|
[ 1.91487521e-02 8.51093650e-01 -2.69493014e-01 -2.67417759e-01
-1.04489617e-01 -8.66888046e-01 5.15742123e-01 3.10087889e-01
-3.76939476e-01 1.09965098e+00 -3.62449855e-01 -1.27243412e+00
-3.80759448e-01 -1.01277125e+00 -6.56342983e-01 -3.19413424e-01
-2.62508214e-01 4.88976240e-01 4.88546550e-01 -6.83647931e-01
5.89642167e-01 3.63703936e-01 -1.36643875e+00 1.05511874e-01
7.82247961e-01 8.49007130e-01 -1.78205609e-01 9.79698002e-01
-1.37050480e-01 1.27540493e+00 -3.22551847e-01 -9.52255607e-01
1.17041424e-01 -5.40570617e-01 -1.41228604e+00 -2.34771401e-01
1.54916555e-01 -2.93403149e-01 -3.53632838e-01 1.37146306e+00
-2.48280779e-01 -3.00957769e-01 4.12429154e-01 -1.73164105e+00
-3.94840330e-01 1.08851969e+00 -2.30751306e-01 -1.72232874e-02
3.82755667e-01 -3.99194919e-02 1.30946469e+00 -3.47050011e-01
4.49626416e-01 1.22472203e+00 7.68225491e-01 6.51005924e-01
-1.20101154e+00 -1.77576005e-01 -2.63563186e-01 6.27985597e-01
-1.39284217e+00 -3.78632993e-01 4.75657135e-01 -2.11017445e-01
7.61518121e-01 8.97055149e-01 7.52807140e-01 -2.39493791e-02
5.55573285e-01 7.90552080e-01 1.01539004e+00 -1.02636814e+00
1.05777472e-01 2.68143773e-01 3.07968080e-01 1.02395320e+00
9.03247178e-01 2.32848581e-02 2.38692947e-02 -3.94480526e-01
8.06158304e-01 -3.88541102e-01 -2.48719260e-01 -2.55978346e-01
-1.21022391e+00 1.10946965e+00 -1.34893298e-01 6.46939874e-01
1.96284920e-01 4.59829897e-01 5.40736437e-01 6.04131997e-01
-3.44704598e-01 2.69169837e-01 -6.51527703e-01 2.35803023e-01
-6.20260417e-01 6.56421661e-01 1.31145442e+00 9.84223962e-01
5.29128194e-01 -2.08872691e-01 4.74770665e-01 -3.53652626e-01
3.89756441e-01 3.83003414e-01 4.11150791e-02 -1.41341472e+00
1.08303964e-01 3.95109057e-01 3.53120267e-01 -9.73867059e-01
-2.44688153e-01 -2.78904110e-01 -5.42244315e-01 3.10720891e-01
6.36357427e-01 2.58967895e-02 1.23583548e-01 1.75961530e+00
1.15645491e-02 -1.13862999e-01 4.25398529e-01 5.59064925e-01
6.06601954e-01 3.78869355e-01 -1.32228434e-01 -8.12888622e-01
1.11129177e+00 -4.68703300e-01 -7.63982773e-01 3.01579386e-01
9.73284245e-01 -4.66279984e-01 4.59653199e-01 6.73253179e-01
-1.60535467e+00 1.54815257e-01 -1.01609826e+00 -3.22141737e-01
-1.38779014e-01 -4.17459458e-01 1.61928308e+00 7.45529532e-01
-9.62477982e-01 5.47924161e-01 -4.26338941e-01 -1.07065409e-01
8.01285431e-02 7.16744304e-01 -1.94516778e-01 2.63036072e-01
-1.50125492e+00 8.91208827e-01 5.63320518e-01 9.58917215e-02
-4.10415411e-01 -4.79163200e-01 -8.46196353e-01 -3.11800372e-02
8.46759558e-01 -9.78221714e-01 1.63593364e+00 -1.02096510e+00
-1.24840832e+00 1.27624846e+00 -3.03076953e-01 -9.99692738e-01
7.82232523e-01 5.85474133e-01 -5.45473695e-01 6.50468618e-02
1.46985874e-01 -2.21773475e-01 -1.25829771e-01 -9.23349380e-01
-6.56774104e-01 -4.61674690e-01 9.65648234e-01 -2.03081414e-01
5.51874101e-01 -3.99696641e-02 4.58241552e-01 -1.56977162e-01
5.31131387e-01 -4.14171040e-01 -1.50930598e-01 -3.20409238e-01
-2.38382220e-01 -5.17696202e-01 1.92484006e-01 -1.35903537e-01
1.38985646e+00 -1.82348204e+00 -1.35953411e-01 4.46949184e-01
4.93021846e-01 -2.24535931e-02 4.24187452e-01 5.22292733e-01
-4.87132490e-01 4.12102908e-01 -1.98714375e-01 7.84062982e-01
6.33666515e-01 2.99599200e-01 -3.86758238e-01 8.19096088e-01
-9.07398909e-02 1.04317701e+00 -1.02433014e+00 -7.82716930e-01
2.15385601e-01 -2.25064546e-01 -7.63849199e-01 -4.03130144e-01
-5.65893292e-01 -3.52277994e-01 -4.23597902e-01 3.82481366e-01
7.94385493e-01 -3.61966163e-01 4.45504636e-01 -1.75001353e-01
-3.61034900e-01 6.39000952e-01 -1.40444946e+00 1.41101360e+00
-1.45591825e-01 3.10114056e-01 3.07708174e-01 -9.93131876e-01
4.41185653e-01 5.30171216e-01 1.87761426e-01 -3.94449890e-01
4.34511423e-01 4.55829710e-01 2.34596744e-01 -3.51639241e-01
4.91834432e-01 -1.10102725e+00 -2.23140568e-01 5.34256577e-01
-5.80645025e-01 -6.21992230e-01 5.65072775e-01 6.03221059e-01
9.10622656e-01 1.24176562e-01 7.68385708e-01 -7.81753898e-01
1.27065814e+00 3.96276504e-01 5.28388917e-01 6.32193923e-01
-2.82654352e-02 -1.00330643e-01 9.27553117e-01 -6.28132463e-01
-8.33550334e-01 -1.05151236e+00 -4.85308081e-01 6.40930116e-01
2.21105248e-01 -7.68891871e-01 -5.02415836e-01 -4.62220907e-01
7.85817131e-02 1.30194545e+00 -4.62021112e-01 -1.07464995e-02
-5.11199832e-01 -2.81245291e-01 9.05193508e-01 1.00699179e-01
6.00836277e-01 -4.75792110e-01 -7.00296462e-01 3.31892930e-02
-3.00495148e-01 -6.28311396e-01 4.10529852e-01 2.77736753e-01
-1.05405343e+00 -1.54392040e+00 1.29451662e-01 -5.94633102e-01
4.59703028e-01 -2.59054184e-01 1.22782850e+00 6.74413025e-01
4.96145003e-02 8.03451091e-02 2.33339638e-01 -4.93755102e-01
-1.10804188e+00 -4.40959692e-01 -1.55629620e-01 -8.91884327e-01
5.77058613e-01 -4.57346350e-01 -5.37381731e-02 6.33164048e-02
-8.38772297e-01 -2.59272940e-02 3.24156694e-03 6.67976081e-01
5.02845831e-02 6.85024202e-01 1.37979910e-01 -1.09397137e+00
1.60796955e-01 -2.02903003e-02 -1.15191233e+00 5.39862990e-01
-5.65558910e-01 2.39510864e-01 7.74162531e-01 2.76547343e-01
-8.61927450e-01 -4.90333319e-01 1.03087984e-01 4.68136907e-01
2.56646365e-01 4.44233656e-01 -3.40885341e-01 -5.61592951e-02
6.04967117e-01 -2.05937438e-02 -2.00697705e-01 1.30752668e-01
4.39087749e-01 5.35013258e-01 1.01342297e+00 -9.99800980e-01
8.72645676e-01 5.36420047e-01 8.65388393e-01 -4.57571208e-01
-7.97126710e-01 -1.22062797e-02 -5.14210820e-01 2.46396154e-01
1.99416950e-01 -5.90958536e-01 -1.55598545e+00 -5.52056469e-02
-1.51420546e+00 -2.05835536e-01 -5.29615045e-01 3.33888173e-01
-1.10982275e+00 7.92274654e-01 -6.30733907e-01 -1.36351728e+00
2.58949008e-02 -7.31015146e-01 4.61274683e-01 -2.34511957e-01
-3.90255183e-01 -1.43333864e+00 4.24906053e-02 1.68580160e-01
-1.02213539e-01 2.18052268e-01 1.40683019e+00 -5.02800822e-01
-4.59086657e-01 -3.07664275e-01 -3.19644511e-01 1.53787568e-01
-1.77055135e-01 -2.18723826e-02 -7.35859394e-01 2.81533182e-01
4.36355770e-01 1.57815933e-01 3.14215928e-01 1.26572937e-01
6.49752796e-01 -8.52078557e-01 -3.53018120e-02 1.89283624e-01
1.96030962e+00 -1.86163157e-01 1.14320087e+00 7.46693134e-01
-2.14760154e-02 8.59624222e-02 3.55042666e-01 3.07999074e-01
4.17473823e-01 2.20226333e-01 3.12366277e-01 4.45936531e-01
3.72824281e-01 -1.45319970e-02 1.83582455e-01 4.11976278e-01
-6.52981102e-01 3.80482346e-01 -5.74363351e-01 2.58305520e-01
-1.88346767e+00 -1.53881633e+00 -7.54647315e-01 2.33218288e+00
1.39289951e+00 3.61807734e-01 1.74539357e-01 7.45299816e-01
9.00599778e-01 -3.44780356e-01 1.83605403e-01 -9.54827487e-01
-3.30649465e-01 4.13894147e-01 8.34971905e-01 1.19030035e+00
-5.31652570e-01 1.00219738e+00 7.08624411e+00 5.69558084e-01
-5.75759649e-01 9.97642428e-02 -1.80243254e-02 3.23615104e-01
-9.16937470e-01 8.13227534e-01 -4.51239467e-01 2.06798036e-02
8.64975452e-01 -8.43818486e-01 6.11839116e-01 7.33370900e-01
-1.64186835e-01 -6.84978545e-01 -1.38147640e+00 6.08137310e-01
-9.25732926e-02 -1.22718251e+00 -3.66423517e-01 -1.56866595e-01
5.88456273e-01 -6.93977058e-01 -4.84564483e-01 7.99700096e-02
4.92008924e-01 -1.05415332e+00 9.38260198e-01 1.95479974e-01
3.80918533e-01 -9.36620653e-01 8.82738888e-01 4.24532682e-01
-7.22058654e-01 -1.01050129e-02 -4.32599634e-01 -9.44708288e-01
-1.40181109e-01 7.74535298e-01 -5.35056710e-01 6.66788518e-01
-3.98269445e-01 1.90148070e-01 -5.67818619e-02 8.55690658e-01
-4.85508591e-01 2.61492580e-01 -4.81464207e-01 -1.18633136e-01
1.82683215e-01 -3.34142029e-01 2.96114057e-01 9.41607416e-01
-1.04726292e-01 6.49848104e-01 -5.22148609e-01 1.29180276e+00
3.69182676e-01 -1.36687502e-01 -2.58317113e-01 4.78912555e-02
3.26264165e-02 7.19962239e-01 -6.33930624e-01 -8.16989064e-01
-2.63959020e-01 3.83548170e-01 -4.08984482e-01 5.16155809e-02
-9.49422240e-01 -6.27246439e-01 1.68433070e-01 4.01854485e-01
-2.93766148e-02 -1.24369852e-01 -7.09051788e-01 -1.10406101e+00
6.68993741e-02 -8.01182628e-01 4.54351604e-01 -3.93749654e-01
-7.14562356e-01 9.52773262e-03 1.46920353e-01 -4.63843137e-01
-3.68359774e-01 -7.92508006e-01 -6.79778159e-01 1.08769333e+00
-1.51875377e+00 -1.03955066e+00 6.76014721e-01 5.49511850e-01
-1.95221737e-01 2.95520574e-01 9.18935359e-01 -1.33401126e-01
1.22977279e-01 3.28267604e-01 -3.10813338e-01 -6.83484506e-03
3.99631739e-01 -1.36513722e+00 7.49102756e-02 8.85097504e-01
-4.11602944e-01 1.04176366e+00 1.18834984e+00 -5.08208394e-01
-1.94749153e+00 -4.28911030e-01 1.70560110e+00 -3.06427717e-01
1.19887388e+00 3.35319042e-01 -3.77840608e-01 1.28420377e+00
1.17229506e-01 -3.31569493e-01 7.86135867e-02 1.64674111e-02
-5.46064019e-01 -2.79348284e-01 -1.35822940e+00 4.97116208e-01
8.59543920e-01 -4.41874921e-01 -1.12687230e+00 6.06938839e-01
3.94059479e-01 -1.71338245e-01 -6.43792927e-01 4.40872699e-01
9.33671474e-01 -1.52329922e+00 5.87880731e-01 -4.80294764e-01
5.29307723e-01 -5.57723820e-01 -3.06293875e-01 -2.84893513e-01
-4.93004352e-01 -9.62917805e-01 2.51603901e-01 6.50145411e-01
3.19488078e-01 -1.00009203e+00 3.58848959e-01 7.37019718e-01
-7.80684082e-03 -4.15057749e-01 -9.11189854e-01 -7.19767213e-01
5.98048270e-01 -6.35119617e-01 5.26257336e-01 1.02964175e+00
1.20905280e+00 4.05601323e-01 1.07100822e-01 -6.70584291e-02
6.50356710e-01 2.01025635e-01 7.53854156e-01 -1.47706699e+00
-6.07386410e-01 -7.93062031e-01 -7.77762055e-01 -6.69679046e-01
3.65564913e-01 -1.27917147e+00 -1.50776431e-01 -1.48710096e+00
1.41453311e-01 -2.48693421e-01 1.25808954e-01 3.66075188e-01
5.61041474e-01 1.74474433e-01 8.51821378e-02 -9.73498076e-02
-8.25937927e-01 -2.11812988e-01 1.26591313e+00 -1.02597043e-01
1.48322538e-01 5.97035624e-02 -1.09729123e+00 1.21771991e+00
6.66860938e-01 -2.22884297e-01 -3.12119931e-01 -7.42862672e-02
1.21108115e+00 2.54290402e-01 6.54040694e-01 -8.90621364e-01
2.67505348e-01 -5.94140828e-01 -1.19414136e-01 -3.41090083e-01
-3.49262774e-01 -8.84160340e-01 3.68971944e-01 9.12913799e-01
-2.23303095e-01 1.17875941e-01 2.10202739e-01 -1.25182584e-01
1.73230797e-01 -8.98773849e-01 6.88619494e-01 -5.32832444e-01
-6.79651439e-01 -3.15332502e-01 -3.78803998e-01 1.02402687e-01
9.66480732e-01 2.44260117e-01 -5.04386902e-01 -1.57692924e-01
-7.90383518e-01 8.66101757e-02 8.90089095e-01 -6.71659946e-01
3.28570575e-01 -9.76630807e-01 -5.25378942e-01 -2.82996967e-02
-3.16996247e-01 -1.28942773e-01 1.56626292e-02 1.33582723e+00
-9.51878071e-01 6.45565271e-01 6.97245002e-02 -1.20197646e-01
-1.28857994e+00 6.44299746e-01 3.58155042e-01 -2.52309740e-01
-5.72819650e-01 3.39175463e-01 2.59669840e-01 5.54821789e-02
-5.93816005e-02 -4.07560110e-01 4.54436690e-01 -7.36140311e-01
7.64547110e-01 6.50620639e-01 -1.67506680e-01 -3.49805951e-01
-7.45432973e-01 4.47267443e-01 2.28948399e-01 -1.87850147e-01
1.13741696e+00 -4.05772090e-01 -1.17949986e+00 8.22041690e-01
8.39686453e-01 4.01510030e-01 8.73626843e-02 1.46040559e-01
-1.16514184e-01 -2.87869498e-02 -3.14421624e-01 -8.35544825e-01
-1.88838840e-01 6.86052203e-01 -2.24236399e-01 6.72323585e-01
8.27809215e-01 1.94456384e-01 6.39267862e-01 9.09468770e-01
6.66991413e-01 -9.60669637e-01 -1.04554856e+00 5.08789122e-01
7.55213380e-01 -8.76712143e-01 6.93735838e-01 -6.03199363e-01
-2.75761425e-01 1.45608187e+00 -1.51948750e-01 -3.70099992e-02
2.95902193e-01 4.38724428e-01 -1.61361903e-01 -1.43555969e-01
-8.44578147e-01 -1.27479360e-01 -2.41317540e-01 2.25115687e-01
6.88691139e-01 2.54267454e-01 -1.02133572e+00 6.45142794e-01
-7.67958045e-01 7.50517726e-01 1.05569828e+00 9.25746739e-01
-6.39358819e-01 -1.24642444e+00 -7.10672736e-01 -6.72529126e-03
-6.71330094e-01 -2.62264758e-01 -2.10399956e-01 1.26017296e+00
4.30701286e-01 8.46767366e-01 -3.83201778e-01 5.73924556e-02
-1.77616075e-01 1.68786585e-01 1.39533031e+00 -3.95798177e-01
-2.39473358e-01 1.29672840e-01 3.03762525e-01 -1.26145080e-01
-7.86079645e-01 -3.74043971e-01 -1.61479855e+00 -1.46554136e+00
-5.76566219e-01 1.01271009e+00 4.05088931e-01 1.26529193e+00
-7.90158868e-01 1.66345027e-03 3.04006875e-01 1.30444029e-02
-8.38013828e-01 -4.53504384e-01 -1.15576482e+00 -2.44470313e-01
3.93466502e-01 5.78527153e-02 -8.12292099e-01 2.01016963e-01]
|
[8.749448776245117, 6.795022964477539]
|
b4d10bb1-d24e-4366-afea-66f6c4c349b9
|
a-marker-based-neural-network-system-for
|
2212.12800
| null |
https://arxiv.org/abs/2212.12800v1
|
https://arxiv.org/pdf/2212.12800v1.pdf
|
A Marker-based Neural Network System for Extracting Social Determinants of Health
|
Objective. The impact of social determinants of health (SDoH) on patients' healthcare quality and the disparity is well-known. Many SDoH items are not coded in structured forms in electronic health records. These items are often captured in free-text clinical notes, but there are limited methods for automatically extracting them. We explore a multi-stage pipeline involving named entity recognition (NER), relation classification (RC), and text classification methods to extract SDoH information from clinical notes automatically. Materials and Methods. The study uses the N2C2 Shared Task data, which was collected from two sources of clinical notes: MIMIC-III and University of Washington Harborview Medical Centers. It contains 4480 social history sections with full annotation for twelve SDoHs. In order to handle the issue of overlapping entities, we developed a novel marker-based NER model. We used it in a multi-stage pipeline to extract SDoH information from clinical notes. Results. Our marker-based system outperformed the state-of-the-art span-based models at handling overlapping entities based on the overall Micro-F1 score performance. It also achieved state-of-the-art performance compared to the shared task methods. Conclusion. The major finding of this study is that the multi-stage pipeline effectively extracts SDoH information from clinical notes. This approach can potentially improve the understanding and tracking of SDoHs in clinical settings. However, error propagation may be an issue, and further research is needed to improve the extraction of entities with complex semantic meanings and low-resource entities using external knowledge.
|
['Anthony Rios', 'Xingmeng Zhao']
|
2022-12-24
| null | null | null | null |
['relation-classification']
|
['natural-language-processing']
|
[ 1.39526084e-01 4.32634205e-01 -4.27132130e-01 -1.62959695e-01
-1.25958443e+00 -3.11114490e-01 -1.87555458e-02 1.09123468e+00
-6.62351489e-01 6.85438275e-01 8.65776420e-01 -5.72363555e-01
-5.66561580e-01 -7.61659980e-01 -3.57548743e-01 -2.08663434e-01
-1.28927141e-01 5.44823647e-01 -4.13876176e-02 1.83373421e-01
1.34822026e-01 2.93905318e-01 -6.57529235e-01 6.19934201e-01
1.22711587e+00 5.78276038e-01 -1.94521400e-03 7.07031906e-01
-2.57961899e-01 1.09693861e+00 -4.90842730e-01 -7.06254125e-01
-1.32784456e-01 -4.69235986e-01 -1.03889871e+00 -3.99975955e-01
-1.14490790e-02 2.53149569e-02 8.31386894e-02 7.37148285e-01
1.07218099e+00 -2.96085179e-01 6.05157733e-01 -6.16576791e-01
-6.42058671e-01 6.92779183e-01 5.47931809e-03 3.09850305e-01
5.39528728e-01 -5.18792961e-03 1.07973933e+00 -7.90611684e-01
1.30471313e+00 5.97525418e-01 1.48892522e+00 4.18937534e-01
-1.04695749e+00 -7.36961961e-01 -4.63189125e-01 -1.42174125e-01
-1.43300581e+00 -4.53897148e-01 -1.24558352e-01 -7.41513371e-01
1.36811757e+00 2.08320782e-01 7.90530026e-01 6.50282204e-01
4.04675335e-01 3.29973876e-01 8.70622814e-01 -3.02293241e-01
1.27094314e-01 1.58711269e-01 3.79563808e-01 8.45316410e-01
6.50281489e-01 -4.12056923e-01 -4.15159523e-01 -1.01888132e+00
3.81306648e-01 3.27468485e-01 -1.93901449e-01 2.97562271e-01
-1.19088173e+00 6.40928149e-01 2.48691559e-01 5.27822852e-01
-7.10034490e-01 -5.29590607e-01 4.39216554e-01 -1.17611513e-01
5.15601933e-01 7.91318715e-01 -7.79308498e-01 -1.33107722e-01
-8.94395649e-01 4.13185284e-02 1.22068942e+00 8.38424027e-01
5.05853407e-02 -7.85291553e-01 -4.12859976e-01 8.43317151e-01
3.34114045e-01 4.50640827e-01 3.86573344e-01 -6.39469266e-01
6.60373926e-01 1.16133833e+00 -5.00026066e-03 -1.07182062e+00
-1.03729022e+00 -3.91658694e-01 -6.76229775e-01 -6.39253974e-01
3.19595069e-01 -5.80587626e-01 -8.17446351e-01 1.35098267e+00
4.94345963e-01 1.05807640e-01 4.47794497e-01 4.25876707e-01
1.31772888e+00 -1.99466590e-02 7.51189232e-01 -1.66604012e-01
1.91926038e+00 -5.10710239e-01 -1.22144008e+00 8.95179063e-02
1.27088451e+00 -8.43873620e-01 4.02368158e-01 -1.46285519e-01
-8.56940269e-01 1.99714616e-01 -5.46930671e-01 -2.05341075e-02
-5.70079684e-01 5.65913133e-02 4.83396679e-01 6.39800489e-01
-7.28750408e-01 5.22737145e-01 -9.35596824e-01 -6.72888577e-01
7.55618453e-01 3.41504902e-01 -5.56013703e-01 7.09026009e-02
-1.35844958e+00 9.78152573e-01 1.81166023e-01 -1.64811760e-01
-1.41209915e-01 -1.37301016e+00 -1.05498099e+00 9.63935480e-02
3.05273622e-01 -1.27855742e+00 9.34952736e-01 -2.37956554e-01
-7.32623935e-01 9.51334000e-01 -3.56863022e-01 -2.11857885e-01
1.06119081e-01 -1.55282125e-01 -7.56304324e-01 3.56014013e-01
4.11187738e-01 1.62605777e-01 -1.69350952e-01 -5.88092685e-01
-6.55209064e-01 -5.19537091e-01 -5.96251070e-01 7.85136744e-02
-2.81716049e-01 5.18634915e-01 -4.83970623e-03 -7.26793945e-01
-2.78930187e-01 -9.41322803e-01 -4.70050991e-01 -1.75047949e-01
-5.01227856e-01 -2.55605042e-01 -4.59983647e-02 -1.18327665e+00
1.93231773e+00 -1.95078659e+00 -4.48565573e-01 1.05388522e-01
6.82362318e-01 4.42755401e-01 3.65541756e-01 6.41242981e-01
-7.52467364e-02 8.71370494e-01 -2.48445913e-01 -1.96987897e-01
-3.75552326e-01 -2.37055346e-01 3.40475410e-01 2.36641258e-01
4.23812568e-01 1.13727486e+00 -9.75499332e-01 -8.09276760e-01
-4.15636152e-01 6.61392510e-01 -7.29485631e-01 2.23147780e-01
2.71060556e-01 2.08603889e-01 -5.67156315e-01 8.05616736e-01
3.22587132e-01 -9.33732152e-01 5.26367009e-01 -1.78271696e-01
-3.91328260e-02 7.38329887e-01 -1.06455290e+00 1.44215608e+00
8.40311274e-02 7.73425475e-02 -7.40592331e-02 -2.13440567e-01
4.41765606e-01 6.76521540e-01 9.15626884e-01 -4.08714116e-02
-4.08520997e-02 3.79600763e-01 7.76083022e-02 -8.93835366e-01
2.30474919e-01 -2.97884524e-01 -1.13488995e-01 2.14160562e-01
-2.67535776e-01 4.73717690e-01 -1.14326060e-01 2.94303715e-01
1.68904495e+00 -3.63486350e-01 1.05995059e+00 -1.58814639e-01
2.26906568e-01 4.94770914e-01 1.12394214e+00 6.30279422e-01
-2.21085861e-01 4.45568383e-01 4.19044435e-01 -9.01155695e-02
-5.46496570e-01 -5.41244268e-01 -7.45555043e-01 5.08325577e-01
-5.23395419e-01 -9.73827541e-01 -5.08715749e-01 -9.10810590e-01
1.44592196e-01 4.33039486e-01 -7.49872863e-01 1.88882440e-01
-3.63524437e-01 -1.02995706e+00 1.07692742e+00 5.90477526e-01
1.69170260e-01 -1.00156856e+00 -5.86334884e-01 5.40622592e-01
-4.25451577e-01 -1.18667543e+00 -7.23306894e-01 -2.04965547e-02
-6.64611757e-01 -1.62142563e+00 -8.19078445e-01 -6.51833534e-01
6.57469869e-01 -2.96577871e-01 1.11162543e+00 9.29222107e-02
-4.29999530e-01 3.91766369e-01 -3.14601600e-01 -8.23690295e-01
-4.93566662e-01 2.88686126e-01 -2.18704611e-01 -4.74209219e-01
9.65787232e-01 4.29183170e-02 -8.38515937e-01 6.18544109e-02
-8.68051171e-01 -1.73635125e-01 8.15564930e-01 7.35180080e-01
6.54173076e-01 -2.28049561e-01 8.72119427e-01 -1.54972708e+00
6.09108567e-01 -9.46977019e-01 7.78533444e-02 2.57260233e-01
-1.22641778e+00 -1.47923589e-01 3.34281266e-01 -1.28954694e-01
-8.97413075e-01 -1.17306709e-01 -1.97989598e-01 3.46209258e-01
-1.89850703e-01 9.33649838e-01 4.47183363e-02 2.33776480e-01
5.76781392e-01 -4.54456270e-01 4.53437194e-02 -5.67296684e-01
-3.49263936e-01 1.22385561e+00 -2.06790119e-01 -2.86808796e-02
2.03670412e-01 1.81237087e-01 -3.51545326e-02 -5.95020413e-01
-1.07025051e+00 -9.89199638e-01 -2.76288509e-01 4.38237995e-01
1.38826311e+00 -1.16076815e+00 -6.74857259e-01 3.36638749e-01
-9.07745183e-01 -2.19301581e-01 -2.31549054e-01 7.86331773e-01
2.14667752e-01 1.52211860e-01 -1.02709579e+00 -4.54286575e-01
-8.09522033e-01 -8.64711642e-01 1.00612128e+00 1.52038053e-01
-6.82211220e-01 -1.16142142e+00 4.59352314e-01 6.00910783e-01
2.89674461e-01 4.03366148e-01 1.12761462e+00 -1.27425992e+00
-6.75096065e-02 -1.00851811e-01 -3.36716473e-01 -2.34723896e-01
4.39129263e-01 -2.27648303e-01 -3.54881287e-01 7.09672943e-02
-2.46310264e-01 2.24148750e-01 6.18961573e-01 2.77143747e-01
6.07596755e-01 -6.54528916e-01 -7.22898483e-01 5.77338874e-01
1.31797528e+00 3.49079907e-01 5.07316351e-01 4.35434103e-01
7.12780595e-01 5.63465238e-01 3.97843778e-01 4.93446589e-01
1.03860199e+00 4.20251191e-01 -2.75957227e-01 -3.65224063e-01
-5.47485240e-02 -1.05457745e-01 -3.89982201e-02 8.27596784e-01
-8.23818818e-02 -5.30841276e-02 -1.55719626e+00 7.94464946e-01
-1.59436047e+00 -7.18721330e-01 -4.05998290e-01 1.81900561e+00
1.18963695e+00 -1.06351897e-01 1.09038968e-02 -2.35007092e-01
6.95632756e-01 -4.21435297e-01 -5.28723180e-01 -2.79828645e-02
-4.08008210e-02 3.93776089e-01 6.95215344e-01 9.81776863e-02
-8.19693387e-01 4.36959267e-01 6.05948400e+00 1.75422475e-01
-6.44014955e-01 3.27214897e-01 4.10105139e-01 1.50552884e-01
-2.02024177e-01 -1.38913378e-01 -1.29387248e+00 4.57603157e-01
1.21505165e+00 -7.76485279e-02 -2.30672792e-01 4.78141129e-01
2.88941145e-01 3.91541198e-02 -8.79793167e-01 5.90004265e-01
8.63896981e-02 -1.60016561e+00 -2.21058697e-01 4.63092029e-01
6.51558399e-01 1.82727069e-01 -3.79504979e-01 1.78977102e-01
3.52396667e-01 -8.94391835e-01 -4.18326264e-04 8.79599571e-01
9.95657921e-01 -2.46988147e-01 1.30047822e+00 -2.13234127e-03
-1.17377114e+00 -5.93438968e-02 -1.17289480e-02 5.08930266e-01
2.84240186e-01 8.72206092e-01 -1.47758389e+00 8.37588310e-01
7.24335968e-01 8.56261849e-01 -6.24366701e-01 1.20696735e+00
3.75214741e-02 8.57373834e-01 -6.01574853e-02 8.09773207e-02
-1.20806411e-01 2.51021951e-01 3.78851563e-01 1.66070712e+00
3.97302687e-01 7.02383757e-01 1.37073724e-02 5.45182705e-01
-4.15375084e-01 6.60539806e-01 -5.27237296e-01 -4.49688047e-01
5.83557665e-01 1.29056644e+00 -6.69427931e-01 -5.34926593e-01
-5.40426016e-01 4.36542034e-01 2.90777057e-01 4.00896296e-02
-5.54036140e-01 -5.91403008e-01 4.77885157e-01 5.24679840e-01
1.87267199e-01 3.60894918e-01 -3.74871105e-01 -1.10629523e+00
-4.10497665e-01 -9.84082222e-01 1.12331080e+00 -3.22002709e-01
-1.50706923e+00 4.93697941e-01 -4.33298528e-01 -1.07951760e+00
-1.90419406e-01 -2.81606048e-01 3.12454645e-02 7.99003720e-01
-1.52145123e+00 -1.04562068e+00 -2.31361657e-01 4.88518298e-01
-2.96537690e-02 -3.57172601e-02 1.32225156e+00 6.47560298e-01
-8.02772164e-01 7.25960255e-01 -8.52773152e-03 6.46098793e-01
1.24110794e+00 -1.29605103e+00 4.32296935e-03 2.71169096e-01
-5.41342080e-01 1.11406338e+00 4.48333956e-02 -1.35151386e+00
-1.03625202e+00 -1.30310440e+00 1.79796445e+00 -7.91162670e-01
5.56546390e-01 1.23874590e-01 -1.08683169e+00 7.70415902e-01
6.80519640e-02 -1.22154534e-01 1.80748487e+00 3.19341183e-01
-1.89428359e-01 2.73067296e-01 -1.45535016e+00 3.05949837e-01
1.16060257e+00 -5.64901769e-01 -9.29914296e-01 2.32818514e-01
7.13715374e-01 -3.64169598e-01 -1.88590050e+00 5.08850634e-01
5.38972199e-01 -3.05217683e-01 8.40575218e-01 -9.60635781e-01
4.75016415e-01 -1.16827786e-01 1.63375825e-01 -1.10677552e+00
-4.87198263e-01 -5.31193554e-01 -8.74089170e-03 1.25870693e+00
1.00551462e+00 -9.59049106e-01 4.81139600e-01 1.16649055e+00
-8.97600427e-02 -8.92427266e-01 -6.01744175e-01 -2.25628421e-01
-2.13990837e-01 -4.77138022e-03 4.93057430e-01 1.51086080e+00
4.53538597e-01 5.00101268e-01 3.09200883e-01 3.46248686e-01
2.94387758e-01 -1.66606575e-01 2.70037472e-01 -1.34348750e+00
-2.09588960e-01 -1.92903668e-01 -2.48528212e-01 -1.34316608e-01
-5.67458160e-02 -1.11825919e+00 -3.72335166e-01 -2.24196100e+00
7.34679937e-01 -8.23363662e-01 -4.89808083e-01 8.56135488e-01
-7.01830447e-01 -1.92014247e-01 -1.44315526e-01 4.66661632e-01
-6.56564176e-01 -4.70569544e-02 9.72418308e-01 1.66878507e-01
-5.18648505e-01 -1.59259230e-01 -1.13009715e+00 6.05619907e-01
6.11820102e-01 -1.16281116e+00 1.26513228e-01 -1.71403661e-01
4.68334109e-01 2.76531756e-01 4.03419919e-02 -5.96679866e-01
5.27906775e-01 -2.77898218e-02 3.12132120e-01 -4.29496676e-01
-2.42841393e-01 -7.37880349e-01 4.96833503e-01 7.21308470e-01
-4.15029913e-01 2.52753764e-01 1.84604719e-01 5.56564331e-01
-1.74050611e-02 -1.73896402e-01 2.87955076e-01 -2.74361581e-01
7.29125589e-02 -6.39013574e-03 -6.10372901e-01 4.32356358e-01
8.84141326e-01 2.11750083e-02 -6.56133294e-01 8.35402459e-02
-7.52028465e-01 1.18234560e-01 2.57321537e-01 -5.48758693e-02
3.13937604e-01 -9.41956401e-01 -9.04099941e-01 -2.67003298e-01
2.88790375e-01 -1.49809318e-02 3.15941989e-01 1.43413162e+00
-4.26103354e-01 5.64568639e-01 2.17194632e-01 -1.30689919e-01
-1.32937968e+00 5.18383980e-01 3.11276466e-01 -9.43698585e-01
-6.82276547e-01 4.06474441e-01 -1.85481042e-01 -4.83445168e-01
-1.72228776e-02 -3.75653863e-01 -6.03674650e-01 3.90336871e-01
7.41069496e-01 4.99881446e-01 2.68771052e-01 -5.84422112e-01
-7.94780731e-01 3.11034441e-01 -1.82983786e-01 2.18926042e-01
1.63943410e+00 -3.35052982e-03 -2.46525854e-01 2.02644229e-01
1.10700917e+00 5.79370916e-01 -2.07002595e-01 -2.09375665e-01
4.75668490e-01 9.73689705e-02 -1.87447071e-01 -1.08290386e+00
-5.57102203e-01 1.97848886e-01 4.98575449e-01 7.35652447e-02
8.97990227e-01 3.20689082e-02 9.55814958e-01 2.15283066e-01
-4.60727699e-02 -8.73350918e-01 -4.00123596e-01 4.37466085e-01
2.70586014e-01 -1.18237722e+00 8.89146142e-03 -6.11455262e-01
-7.32193828e-01 7.46269345e-01 2.06699044e-01 5.30082226e-01
9.17572856e-01 3.78939092e-01 1.54180989e-01 -6.78996086e-01
-8.11113715e-01 -1.31246328e-01 5.69808781e-01 2.57874429e-01
9.01000798e-01 1.19533077e-01 -6.51944160e-01 1.25879884e+00
9.29392725e-02 4.02530015e-01 4.48420346e-01 1.02353263e+00
1.62726361e-02 -1.27405667e+00 -1.22934148e-01 1.02824771e+00
-1.45892489e+00 -7.00997710e-01 -4.74382311e-01 5.97663343e-01
3.12541068e-01 1.04257572e+00 -3.64041805e-01 -1.32579774e-01
6.38337970e-01 3.13905060e-01 -2.55827487e-01 -1.00135446e+00
-1.40152109e+00 -1.27362102e-01 7.25300312e-01 -4.78466332e-01
-7.39161074e-01 -8.01647663e-01 -1.64510036e+00 4.08801287e-02
-4.43928808e-01 4.30164248e-01 4.42325920e-01 7.75109708e-01
1.29097521e+00 8.18247437e-01 5.65761812e-02 3.49419445e-01
-1.95959479e-01 -8.72713566e-01 -2.14499995e-01 4.54895586e-01
2.19177052e-01 -2.70923436e-01 -1.79031327e-01 1.12032734e-01]
|
[8.418601989746094, 8.675188064575195]
|
082599c8-8a5a-43eb-a4a2-00ff76188874
|
multitask-learning-for-instrument-activation
|
2008.00616
| null |
https://arxiv.org/abs/2008.00616v1
|
https://arxiv.org/pdf/2008.00616v1.pdf
|
Multitask learning for instrument activation aware music source separation
|
Music source separation is a core task in music information retrieval which has seen a dramatic improvement in the past years. Nevertheless, most of the existing systems focus exclusively on the problem of source separation itself and ignore the utilization of other~---possibly related---~MIR tasks which could lead to additional quality gains. In this work, we propose a novel multitask structure to investigate using instrument activation information to improve source separation performance. Furthermore, we investigate our system on six independent instruments, a more realistic scenario than the three instruments included in the widely-used MUSDB dataset, by leveraging a combination of the MedleyDB and Mixing Secrets datasets. The results show that our proposed multitask model outperforms the baseline Open-Unmix model on the mixture of Mixing Secrets and MedleyDB dataset while maintaining comparable performance on the MUSDB dataset.
|
['Yun-Ning Hung', 'Alexander Lerch']
|
2020-08-03
| null | null | null | null |
['music-source-separation']
|
['music']
|
[ 2.25103125e-01 -6.25273943e-01 -2.10289970e-01 8.31693318e-03
-1.60656750e+00 -7.01158226e-01 4.28097576e-01 -1.52649790e-01
-4.25331235e-01 5.73608637e-01 4.99056697e-01 1.86551455e-02
-4.91594702e-01 -4.57258709e-02 -4.86107945e-01 -8.57121289e-01
5.40047586e-02 3.88539582e-01 -7.60674626e-02 -2.05754489e-01
1.05348706e-01 -8.36880282e-02 -1.63983262e+00 4.38357234e-01
8.55750918e-01 7.12869704e-01 2.46759936e-01 6.56665444e-01
-1.55208096e-01 6.54944539e-01 -8.99651825e-01 -2.05603629e-01
4.62448597e-01 -6.53934479e-01 -5.26959658e-01 -3.98185551e-01
6.15799010e-01 8.53169709e-02 -1.41446888e-01 1.06628799e+00
1.01924539e+00 1.16057314e-01 5.03320098e-01 -1.22960603e+00
-2.64566362e-01 1.20057213e+00 -9.19870794e-01 3.28191310e-01
4.12875921e-01 -5.79276741e-01 1.25388777e+00 -6.68785572e-01
1.68275848e-01 1.16898263e+00 7.46163011e-01 1.48812428e-01
-1.44212937e+00 -1.27373874e+00 -1.43150389e-01 3.45506072e-01
-1.42458665e+00 -9.45658505e-01 1.13384140e+00 -2.13951603e-01
6.27293825e-01 3.43129903e-01 1.23498887e-01 1.31050122e+00
-2.58867592e-01 1.08652866e+00 1.26580477e+00 -5.92242301e-01
-1.08394191e-01 1.03419408e-01 2.14225039e-01 1.50928587e-01
1.72574192e-01 -1.23028599e-01 -1.12660134e+00 -4.70141649e-01
5.94843805e-01 -4.18484211e-01 -5.40626109e-01 -4.04581130e-02
-1.69693983e+00 5.15790701e-01 7.13544935e-02 4.98372078e-01
-1.58080667e-01 1.47210434e-01 4.30983096e-01 6.71888471e-01
4.42168832e-01 4.96376961e-01 -4.76646036e-01 -3.37574035e-01
-1.47724843e+00 3.81406218e-01 9.47973728e-01 9.50124860e-01
2.92770088e-01 3.26796114e-01 -1.96479395e-01 1.21635306e+00
1.61950350e-01 7.54546046e-01 6.83252037e-01 -8.29119086e-01
6.95269823e-01 1.12013295e-01 1.28649473e-01 -7.07010329e-01
-4.63657349e-01 -8.99403453e-01 -7.40739226e-01 -2.41096355e-02
5.32353282e-01 -1.47188127e-01 -6.27381861e-01 1.99470985e+00
1.12381734e-01 6.30853176e-01 1.71120331e-01 9.01411295e-01
9.82528090e-01 4.70997363e-01 -4.46409851e-01 -3.85535270e-01
1.27836144e+00 -1.10215962e+00 -1.01547992e+00 1.25044270e-03
1.96479142e-01 -1.38503444e+00 9.44053710e-01 8.89491856e-01
-1.11934841e+00 -8.72146606e-01 -1.25346398e+00 9.96871386e-03
4.98756720e-03 3.96994352e-01 7.16627777e-01 8.11890781e-01
-7.00529933e-01 4.09543306e-01 -4.53815937e-01 -1.08640090e-01
1.06006987e-01 3.77379745e-01 -9.11661685e-02 6.58567250e-02
-1.12158740e+00 3.83249581e-01 2.75328964e-01 -1.66586787e-01
-7.72658050e-01 -5.34128010e-01 -2.81391978e-01 -7.31655490e-03
6.43902004e-01 -6.99865997e-01 1.29138076e+00 -9.83856261e-01
-1.51755202e+00 5.49752176e-01 -2.62251914e-01 -2.79850066e-01
2.60410786e-01 -6.22629344e-01 -8.05617154e-01 -1.88657507e-01
1.90992244e-02 2.03198671e-01 1.13160956e+00 -1.25878251e+00
-5.31891704e-01 -3.93938005e-01 -1.50182113e-01 3.74535501e-01
-3.45745593e-01 3.78152072e-01 -6.60053670e-01 -1.30016565e+00
1.66311994e-01 -1.27365983e+00 1.22931994e-01 -7.89036572e-01
-6.72054589e-01 9.43407491e-02 4.54381078e-01 -5.46935022e-01
1.47466469e+00 -2.40964532e+00 5.09688079e-01 1.58099979e-01
-2.37039179e-01 2.56003708e-01 -4.10798579e-01 5.07474661e-01
-3.19441736e-01 -3.66738081e-01 -6.36331439e-02 -6.15534127e-01
-8.01728070e-02 -1.27625316e-01 -6.12468064e-01 2.70781338e-01
-4.91709709e-01 6.15194082e-01 -5.94370782e-01 -3.99612546e-01
-2.49576375e-01 2.59495169e-01 -3.76785278e-01 -1.63707864e-02
-5.60042113e-02 5.79057932e-01 -2.81104147e-01 7.14231253e-01
5.97948432e-01 -8.26377869e-02 3.39605957e-02 -6.84544444e-02
7.97646791e-02 4.60796297e-01 -1.73535299e+00 2.57482338e+00
-3.02828074e-01 5.42963624e-01 4.63128865e-01 -6.93039000e-01
6.69324338e-01 6.88922703e-01 8.13085139e-01 -2.61118978e-01
9.56890881e-02 4.08400834e-01 4.63010132e-01 -1.39083862e-01
6.32965505e-01 -2.04291254e-01 -3.07913929e-01 6.73158765e-01
1.28199011e-01 -4.61098254e-02 6.09760657e-02 2.08004639e-01
9.02012050e-01 3.37415785e-01 5.22172973e-02 -2.44531363e-01
4.51005220e-01 -6.64426461e-02 5.60540020e-01 9.15941179e-01
9.75217856e-03 7.67456293e-01 1.19174525e-01 4.26903874e-01
-4.40396965e-01 -9.95629966e-01 -1.59056261e-01 1.43668091e+00
5.74619370e-03 -6.78284705e-01 -5.96962094e-01 -4.21369612e-01
7.38990530e-02 5.63065886e-01 -2.81450599e-01 1.48646742e-01
-3.99218291e-01 -1.18946588e+00 1.16763496e+00 2.53732532e-01
4.52545434e-01 -8.09498668e-01 -1.20997563e-01 3.50282162e-01
-7.90346622e-01 -9.42929327e-01 -5.11165559e-01 4.75928158e-01
-7.86735117e-01 -8.58990610e-01 -7.92641699e-01 -5.83481908e-01
-3.72680485e-01 6.27576470e-01 1.04343128e+00 -5.73972344e-01
-1.32224217e-01 2.84446508e-01 -4.09721911e-01 -9.73294973e-01
-2.81935722e-01 3.52626234e-01 2.45486438e-01 2.99208879e-01
3.54167819e-01 -7.77864158e-01 -3.67495835e-01 2.53203660e-01
-7.90383756e-01 -2.12771557e-02 7.34105170e-01 6.77915096e-01
4.01471138e-01 3.26561630e-01 1.07487988e+00 -8.48946989e-01
8.14498961e-01 -4.84950125e-01 -1.42036065e-01 1.58948228e-01
-7.81865478e-01 9.25377458e-02 3.50680381e-01 -7.68799603e-01
-1.14995193e+00 -2.05015287e-01 -7.39986226e-02 -5.16273975e-01
-1.09724641e-01 4.63262290e-01 -4.31499422e-01 8.54222924e-02
6.39957964e-01 1.39276579e-01 -2.90362418e-01 -1.04669535e+00
2.91621178e-01 9.14495409e-01 7.34973609e-01 -7.30582535e-01
8.06847990e-01 3.78183007e-01 -3.64565141e-02 -6.36622965e-01
-8.04559231e-01 -1.02355230e+00 -4.35720712e-01 1.99940920e-01
5.26280165e-01 -1.39356542e+00 -3.58481795e-01 5.89559019e-01
-8.44332457e-01 7.61090312e-03 -2.19715893e-01 7.26064622e-01
-4.49697644e-01 2.86995918e-01 -6.55423284e-01 -7.08820939e-01
-3.50868434e-01 -1.13204265e+00 1.14749730e+00 -1.31610855e-01
-3.20700318e-01 -5.62479317e-01 4.76547688e-01 4.86065447e-01
3.65964055e-01 -7.63898715e-02 7.43932128e-01 -9.50342059e-01
-4.19136792e-01 -7.51146227e-02 1.36041306e-02 3.02007735e-01
2.83004135e-01 -4.71798420e-01 -1.43217301e+00 -2.58791804e-01
2.21581653e-01 -4.74460661e-01 1.11557972e+00 4.42944050e-01
7.89096713e-01 1.97287053e-01 -1.99784517e-01 6.29570961e-01
1.15797913e+00 1.18090048e-01 3.55161995e-01 2.20917627e-01
7.42518067e-01 4.55240816e-01 6.03919864e-01 3.71154606e-01
2.54021704e-01 9.69633400e-01 1.74159016e-02 -1.34499937e-01
-4.31526154e-01 -5.54937124e-02 5.43892324e-01 1.31577134e+00
-2.07298696e-01 -2.35052168e-01 -5.00833452e-01 4.17118013e-01
-1.91506732e+00 -1.07205081e+00 -3.91400993e-01 2.32855272e+00
9.36345994e-01 -2.34267130e-01 3.66605878e-01 5.57901442e-01
4.54526305e-01 4.51008886e-01 -4.03212607e-01 8.05840939e-02
-4.37053055e-01 3.20367157e-01 2.76988924e-01 1.83144182e-01
-1.22767770e+00 6.11708462e-01 6.58180475e+00 1.15712011e+00
-9.27756429e-01 3.76992166e-01 -1.50565267e-01 -6.26667440e-01
-1.92853272e-01 -1.10804759e-01 -7.31053352e-01 3.98528069e-01
9.68975961e-01 -5.11265174e-02 6.01760268e-01 2.96881318e-01
9.15976912e-02 -1.12809658e-01 -1.15565896e+00 1.32640684e+00
5.18406987e-01 -6.42343283e-01 -1.05794698e-01 7.80788660e-02
7.34442115e-01 1.10991880e-01 1.05176784e-01 4.01433021e-01
1.26049101e-01 -8.07969093e-01 8.21474612e-01 3.46042752e-01
6.74192667e-01 -7.38793194e-01 4.90814745e-01 5.76614916e-01
-1.12125409e+00 -9.06985477e-02 -6.04818650e-02 1.91476330e-01
5.64414561e-02 5.02535462e-01 -4.70211983e-01 1.19774866e+00
7.69525945e-01 6.14561915e-01 -5.32077491e-01 1.10458076e+00
4.64203767e-02 9.30177987e-01 -3.06632638e-01 5.31490505e-01
-1.60088986e-01 -1.96403772e-01 9.01396275e-01 1.31393611e+00
4.25847232e-01 -2.71071494e-01 1.77826256e-01 3.86994213e-01
-1.13185175e-01 3.03465515e-01 -3.81573886e-01 8.44947901e-03
1.67299837e-01 1.24457145e+00 -4.41050887e-01 -2.68095255e-01
-5.12126386e-01 9.89903152e-01 -8.87318552e-02 4.03716773e-01
-8.97491097e-01 -3.11320394e-01 6.23008013e-01 -2.91368425e-01
3.09667915e-01 1.66249350e-02 -2.25637630e-01 -1.42990506e+00
-1.67702630e-01 -1.50604880e+00 5.20178378e-01 -6.26047730e-01
-1.20297790e+00 5.44763267e-01 1.15219250e-01 -1.41581476e+00
-2.68366963e-01 -1.36372715e-01 -4.06196505e-01 9.84853387e-01
-1.53907824e+00 -1.02279544e+00 3.71319726e-02 1.05397093e+00
5.85028350e-01 -3.96391988e-01 1.01298499e+00 7.10592091e-01
-5.60923338e-01 7.16536164e-01 5.83790064e-01 -1.77970752e-01
1.57186878e+00 -1.31881762e+00 -1.17468843e-02 8.67045820e-01
1.14033926e+00 7.56448746e-01 7.24804282e-01 -3.86429191e-01
-1.38105464e+00 -6.98789358e-01 6.08117282e-01 -5.51576674e-01
5.24702489e-01 -4.59107995e-01 -9.63434100e-01 5.29632568e-01
4.36259806e-01 -6.08978570e-01 1.26962125e+00 7.18731821e-01
-5.93044102e-01 -3.32692683e-01 -6.68936014e-01 2.41020069e-01
1.04676664e+00 -6.62497878e-01 -7.89743602e-01 9.05052386e-03
4.47640866e-01 -2.63110757e-01 -7.43336916e-01 5.43811858e-01
8.23566258e-01 -8.84436131e-01 1.24719417e+00 -3.68490487e-01
3.65149230e-02 -4.65759993e-01 -3.20770562e-01 -1.52024281e+00
-3.50918442e-01 -7.69908667e-01 -6.65284395e-02 1.66261780e+00
4.03829753e-01 -3.50316346e-01 4.90275562e-01 -8.38797018e-02
-8.78963340e-03 5.34732677e-02 -8.90202165e-01 -8.66667807e-01
-1.17010616e-01 -7.68452883e-01 4.82306242e-01 1.04368317e+00
-5.03884144e-02 6.39101565e-01 -8.42220664e-01 4.79846150e-02
7.73724496e-01 6.85216546e-01 1.08058178e+00 -1.46136117e+00
-1.05651045e+00 -3.97674114e-01 2.97448337e-01 -1.04663479e+00
1.46397665e-01 -1.19997525e+00 8.70902538e-02 -1.23430288e+00
4.06735539e-01 -5.07363021e-01 -9.30978954e-01 2.73420006e-01
-4.61030632e-01 4.66364503e-01 3.80189091e-01 6.74852133e-01
-6.16856217e-01 3.43858600e-01 9.73195851e-01 -8.57661441e-02
-4.47133273e-01 3.39435786e-01 -1.01298964e+00 8.10522556e-01
5.99683642e-01 -6.81616366e-01 -6.97801292e-01 -4.87083226e-01
1.28056809e-01 2.69730240e-01 -2.12198161e-02 -1.15748763e+00
3.14963400e-01 3.36468428e-01 2.67312825e-01 -5.84812284e-01
6.59438550e-01 -8.02721739e-01 4.87195939e-01 -4.11808603e-02
-3.50052327e-01 -9.91127267e-02 3.64272326e-01 8.08724999e-01
-4.66008365e-01 -2.06233919e-01 2.54654735e-01 1.54467717e-01
-2.63165742e-01 -1.63465068e-01 -1.54142037e-01 6.57154843e-02
5.05681634e-01 1.96518213e-01 -1.26631916e-01 -4.74571198e-01
-4.86068159e-01 -1.85666233e-01 1.90311745e-01 7.74462819e-01
2.05131173e-01 -1.29058552e+00 -9.50958014e-01 1.59553692e-01
2.78927654e-01 -3.07315618e-01 1.10807806e-01 9.18505728e-01
2.78800756e-01 3.96886140e-01 7.15955719e-02 -3.69598001e-01
-1.58965886e+00 5.39784431e-01 -1.15921289e-01 -4.88293767e-01
-3.45685720e-01 8.38493049e-01 1.94558993e-01 -2.78033495e-01
5.53953707e-01 -1.63907632e-01 -2.87051469e-01 4.31006759e-01
4.74991560e-01 5.93922019e-01 6.03873394e-02 -7.20810294e-01
-3.02800953e-01 5.14335036e-01 5.72828278e-02 -5.90935767e-01
1.10995257e+00 -9.66836587e-02 -5.18921427e-02 1.09377289e+00
7.99272478e-01 8.91372263e-01 -6.22502506e-01 -6.71698153e-01
2.38606796e-01 -4.52454060e-01 1.08461373e-01 -1.05466485e+00
-7.97534406e-01 7.71753669e-01 5.08934855e-01 4.33297195e-02
1.36398125e+00 -7.27157071e-02 7.32453763e-01 3.34295034e-01
5.47349632e-01 -9.27731395e-01 -1.55698629e-02 1.53330043e-01
8.73423398e-01 -1.22741520e+00 1.78168230e-02 -2.32520655e-01
-6.03045285e-01 6.94161475e-01 6.06768392e-02 1.40829295e-01
5.88766575e-01 4.86266166e-01 3.36000979e-01 -7.64267705e-03
-4.91471648e-01 -4.58470881e-01 6.31914616e-01 2.69754261e-01
5.49729466e-01 -1.59764476e-02 -5.59218191e-02 9.61587369e-01
-2.34042808e-01 1.54536581e-02 2.98579663e-01 7.03049421e-01
-1.89629257e-01 -1.60623217e+00 -8.64378452e-01 3.70336533e-01
-1.06495261e+00 -3.95514518e-01 -4.79055226e-01 5.71857989e-01
1.81582838e-01 1.43384171e+00 -5.05994856e-01 -3.25276673e-01
2.39315912e-01 3.65049273e-01 6.70056403e-01 -5.55545568e-01
-8.93199563e-01 1.02841270e+00 2.63838470e-01 -2.95267522e-01
-9.80171502e-01 -8.43576312e-01 -8.38885546e-01 -9.06201378e-02
-6.42554164e-01 3.23986083e-01 7.03949571e-01 6.07218444e-01
3.68000239e-01 8.20883512e-01 4.07466173e-01 -7.81797349e-01
-6.41533971e-01 -1.35494184e+00 -9.31172788e-01 3.61513972e-01
3.63823950e-01 -5.72987080e-01 -3.28215986e-01 -2.78782658e-02]
|
[15.627286911010742, 5.3729567527771]
|
d1af3cd1-a982-420b-ad8f-85e5596cd64f
|
multi-modal-sarcasm-detection-and-humor
|
2105.09984
| null |
https://arxiv.org/abs/2105.09984v2
|
https://arxiv.org/pdf/2105.09984v2.pdf
|
Multi-modal Sarcasm Detection and Humor Classification in Code-mixed Conversations
|
Sarcasm detection and humor classification are inherently subtle problems, primarily due to their dependence on the contextual and non-verbal information. Furthermore, existing studies in these two topics are usually constrained in non-English languages such as Hindi, due to the unavailability of qualitative annotated datasets. In this work, we make two major contributions considering the above limitations: (1) we develop a Hindi-English code-mixed dataset, MaSaC, for the multi-modal sarcasm detection and humor classification in conversational dialog, which to our knowledge is the first dataset of its kind; (2) we propose MSH-COMICS, a novel attention-rich neural architecture for the utterance classification. We learn efficient utterance representation utilizing a hierarchical attention mechanism that attends to a small portion of the input sentence at a time. Further, we incorporate dialog-level contextual attention mechanism to leverage the dialog history for the multi-modal classification. We perform extensive experiments for both the tasks by varying multi-modal inputs and various submodules of MSH-COMICS. We also conduct comparative analysis against existing approaches. We observe that MSH-COMICS attains superior performance over the existing models by > 1 F1-score point for the sarcasm detection and 10 F1-score points in humor classification. We diagnose our model and perform thorough analysis of the results to understand the superiority and pitfalls.
|
['Tanmoy Chakraborty', 'Md Shad Akhtar', 'Shivani Kumar', 'Manjot Bedi']
|
2021-05-20
| null | null | null | null |
['multi-modal-classification']
|
['miscellaneous']
|
[-3.81550401e-01 9.10100043e-02 -2.64877826e-01 -3.80596638e-01
-6.43882036e-01 -4.11011755e-01 6.55079782e-01 6.59679994e-02
-1.87843487e-01 5.39021671e-01 9.21427667e-01 -1.68178871e-01
3.22222054e-01 -3.23524833e-01 -9.19644311e-02 -4.63152617e-01
3.15693855e-01 4.06934142e-01 -3.03768204e-03 -7.83731401e-01
4.97696728e-01 -8.59504193e-02 -1.05139577e+00 5.69341660e-01
7.79979467e-01 8.13231170e-01 6.47037774e-02 7.13553488e-01
4.26711254e-02 1.65365279e+00 -5.89955270e-01 -4.63891745e-01
-2.36236036e-01 -8.59937072e-01 -1.16626978e+00 6.85736686e-02
-7.54866749e-02 -6.32339120e-01 -5.83993018e-01 8.11685503e-01
5.19073129e-01 3.25147599e-01 6.61835909e-01 -1.23163497e+00
-8.58064234e-01 1.01156461e+00 -4.69426930e-01 5.74721843e-02
4.49017674e-01 2.88248688e-01 1.34836423e+00 -1.08221388e+00
2.89454788e-01 1.38924789e+00 4.75418717e-01 7.58041978e-01
-5.93152702e-01 -3.99453998e-01 -2.19349831e-01 5.03670394e-01
-9.30691481e-01 -4.25894558e-01 1.24018633e+00 -6.06879771e-01
8.79154503e-01 3.18213075e-01 2.59562194e-01 1.24856555e+00
1.34687359e-02 1.22801054e+00 1.08586156e+00 -2.27694049e-01
7.25595132e-02 3.33473235e-01 6.63307846e-01 5.29600620e-01
-7.93424904e-01 -2.84620970e-01 -6.85135782e-01 -1.56095251e-01
2.89112628e-01 4.23309766e-02 -1.79627791e-01 2.41235986e-01
-1.02776027e+00 1.37819576e+00 5.84882975e-01 3.68939817e-01
-1.15124270e-01 -1.42957225e-01 8.32375884e-01 3.69906187e-01
1.92984670e-01 5.97653508e-01 -2.74364650e-02 -5.07351577e-01
-6.85984194e-01 3.43058228e-01 8.73565495e-01 8.97070289e-01
2.75021225e-01 1.27867460e-01 -4.12888110e-01 1.31824839e+00
1.83362469e-01 2.52579331e-01 8.75937343e-01 -8.57823610e-01
5.34614384e-01 8.29289734e-01 -1.13277607e-01 -1.10426724e+00
-6.91792727e-01 -1.14018671e-01 -8.27684939e-01 -1.95295438e-01
1.19612209e-01 -1.13208808e-01 -2.09718823e-01 1.75917721e+00
-2.94958837e-02 -5.17900765e-01 3.65034312e-01 1.16223526e+00
1.52348959e+00 7.20285833e-01 -6.10266030e-02 -3.64205182e-01
1.42525899e+00 -1.57493782e+00 -8.88547301e-01 -2.07349017e-01
6.12684309e-01 -7.88632691e-01 1.65468442e+00 1.55229688e-01
-1.10570848e+00 -4.83939022e-01 -1.09645402e+00 -5.13765633e-01
-8.09830651e-02 2.03614086e-01 5.21447957e-01 2.44950399e-01
-3.65007937e-01 1.80559695e-01 -4.23044026e-01 -2.83677965e-01
-2.84401374e-03 4.50612903e-02 1.08772382e-01 3.91164482e-01
-1.47981691e+00 9.62303638e-01 1.97806120e-01 -4.90623750e-02
-9.99401450e-01 -3.44096154e-01 -9.20235753e-01 1.38492286e-01
2.75790602e-01 -1.43271729e-01 1.65708911e+00 -9.78707850e-01
-1.74405789e+00 8.08497012e-01 1.19234221e-02 -4.35381621e-01
2.66700059e-01 -1.88032106e-01 -2.34222040e-01 1.08743608e-02
-5.57326451e-02 3.90668631e-01 4.21264529e-01 -1.07700622e+00
-9.08907801e-02 -1.70504212e-01 4.86265063e-01 4.60194856e-01
-5.99793971e-01 3.26266646e-01 -4.70413119e-02 -5.43296278e-01
-1.45494580e-01 -1.03863287e+00 1.34402052e-01 -6.67131305e-01
-6.03452861e-01 -4.40139383e-01 9.59891081e-01 -6.23339355e-01
1.61095393e+00 -2.12225246e+00 3.74995738e-01 -6.68367267e-01
3.72118980e-01 1.12505034e-01 1.39538199e-01 8.21978509e-01
2.87090242e-01 -1.22841597e-01 -1.97904155e-01 -5.81185460e-01
1.55701086e-01 7.13832304e-02 -4.46320951e-01 1.92626342e-01
3.91677879e-02 1.10123289e+00 -7.01327026e-01 -6.72806382e-01
2.36697122e-01 -1.23659745e-02 -6.27573788e-01 7.30629086e-01
-1.06263258e-01 4.04778838e-01 -3.56154293e-01 6.83678269e-01
2.51709700e-01 -3.34144354e-01 1.32698745e-01 -1.11376680e-01
-6.07369840e-02 5.14813781e-01 -3.61585557e-01 1.53008926e+00
-6.33785188e-01 7.12074518e-01 2.54455265e-02 -9.05955970e-01
9.63889539e-01 5.25049269e-01 2.97612965e-01 -4.98089671e-01
5.96172452e-01 1.24164596e-01 3.29371691e-01 -6.94424093e-01
7.42534578e-01 -5.39389849e-01 -6.28007650e-01 4.92106646e-01
1.46396637e-01 -2.00891241e-01 -8.97472948e-02 5.46895444e-01
8.69772375e-01 -5.13164043e-01 7.02607155e-01 -3.48782271e-01
9.19035792e-01 -4.54719886e-02 4.52954322e-01 4.87992018e-01
-6.54291332e-01 6.07274055e-01 7.61494696e-01 -2.38118023e-01
-9.12183821e-01 -5.83796442e-01 -9.07324404e-02 1.62248707e+00
1.17255166e-01 -4.74208176e-01 -6.42558396e-01 -6.33415103e-01
-2.68055141e-01 8.31025898e-01 -7.00992942e-01 -1.22061290e-01
-5.09407520e-01 -6.85212135e-01 6.46203101e-01 6.98954165e-01
7.10576952e-01 -1.45361376e+00 -7.97158241e-01 1.50079075e-02
-4.98156726e-01 -1.18247199e+00 -6.15094900e-01 1.93551317e-01
-4.84636903e-01 -9.33386087e-01 -3.83014023e-01 -8.47119033e-01
1.92147214e-02 3.35016757e-01 1.05408180e+00 3.09512198e-01
1.86944664e-01 1.20523527e-01 -6.73539042e-01 -1.01139002e-01
-7.10187614e-01 1.40196905e-01 -1.36862412e-01 -2.77736425e-01
6.04446650e-01 -4.20756310e-01 -3.27947974e-01 3.29981595e-01
-5.10837138e-01 3.54698896e-01 2.70282984e-01 1.32609153e+00
-2.42912114e-01 -5.03686130e-01 1.04635632e+00 -1.03555977e+00
1.05462325e+00 -8.53970528e-01 1.57786295e-01 -1.45009607e-01
-3.50664854e-01 -2.82211661e-01 9.53136265e-01 -3.92826796e-01
-1.08950436e+00 -2.48205930e-01 -2.34297752e-01 -1.68857574e-01
4.49740775e-02 8.28549922e-01 -6.59890398e-02 4.37236756e-01
7.36965239e-01 8.89995322e-02 -1.91490706e-02 -3.15685272e-01
4.12251949e-01 1.38023412e+00 7.75801003e-01 -6.00014329e-01
3.84992748e-01 -5.41982520e-03 -5.01466870e-01 -9.29817855e-01
-1.23696542e+00 -6.54337168e-01 -6.29827917e-01 -3.00695211e-01
9.82158482e-01 -9.48080361e-01 -1.12492537e+00 4.47749764e-01
-1.23875153e+00 -3.21591645e-01 2.60002971e-01 2.45297372e-01
-6.80600703e-01 6.36271894e-01 -1.37655592e+00 -1.18864238e+00
-7.13789344e-01 -1.13122261e+00 6.66610658e-01 2.46222734e-01
-5.24734855e-01 -8.69322538e-01 1.31119087e-01 1.05913877e+00
2.61272639e-01 5.68760000e-02 1.13371801e+00 -1.04289496e+00
8.64147842e-02 5.27127236e-02 -2.25090981e-01 3.25067222e-01
-1.52627707e-01 -3.12303662e-01 -1.11088955e+00 -2.70487875e-01
4.24385041e-01 -1.33288825e+00 6.34034932e-01 -1.69070810e-02
9.38388288e-01 -5.84232688e-01 3.75441432e-01 -1.87801309e-02
8.35747480e-01 2.54231930e-01 4.72900063e-01 2.16698244e-01
6.60849929e-01 7.53399730e-01 7.22988904e-01 8.29335809e-01
6.67359173e-01 7.83806860e-01 4.61498082e-01 8.99406075e-02
-1.58754904e-02 -3.00368249e-01 5.63636899e-01 1.61141944e+00
2.63725907e-01 -1.33833468e-01 -7.83842862e-01 6.87027514e-01
-2.13479257e+00 -9.90223289e-01 -3.64153266e-01 1.62020075e+00
1.07766449e+00 -7.21184984e-02 5.88615835e-01 5.73349325e-03
4.25238311e-01 4.46199626e-01 -4.96073008e-01 -6.17099464e-01
-1.89294070e-01 -3.68085533e-01 -3.34407479e-01 7.58247733e-01
-1.03042173e+00 1.04696178e+00 5.74651480e+00 5.22255421e-01
-9.76319969e-01 3.25265437e-01 4.95123982e-01 -1.52902558e-01
-1.21897042e-01 -1.07368365e-01 -4.78560805e-01 6.30392015e-01
9.13914919e-01 -1.50364101e-01 4.60105538e-01 1.13457716e+00
2.46258646e-01 1.49231225e-01 -1.06197345e+00 9.65745032e-01
5.12125134e-01 -9.42478299e-01 -3.43195260e-01 -2.80698508e-01
6.75676644e-01 -1.01003200e-01 -5.91966994e-02 8.20143223e-01
1.55476213e-01 -1.20923150e+00 6.81414127e-01 4.24402095e-02
2.84516245e-01 -7.23478496e-01 1.04188907e+00 6.79378629e-01
-7.60679603e-01 -3.46467823e-01 -3.76320720e-01 -4.10558820e-01
2.58141667e-01 -6.84766620e-02 -8.60634387e-01 1.64565161e-01
3.77261758e-01 7.87869990e-01 -4.02197331e-01 2.62482077e-01
-1.80720374e-01 9.74509120e-01 1.99226320e-01 -4.72520828e-01
4.86690551e-01 8.97859316e-03 4.12854582e-01 1.47660530e+00
-1.62049845e-01 3.99388701e-01 4.68138635e-01 1.00919569e+00
-1.00816466e-01 4.84146476e-01 -3.27011377e-01 -1.27606809e-01
5.12959123e-01 1.43223095e+00 -1.62276551e-01 -2.34864235e-01
-6.19838357e-01 9.17027056e-01 5.13988733e-01 3.33722383e-02
-9.55285251e-01 -2.99921602e-01 2.44615614e-01 -1.77048162e-01
-2.44416296e-01 -1.27308205e-01 -5.68120003e-01 -1.20646274e+00
-3.34770292e-01 -1.11704159e+00 5.66079676e-01 -6.98573947e-01
-1.50882983e+00 7.17539787e-01 -6.70181140e-02 -9.25576270e-01
-4.35390979e-01 -5.19204855e-01 -9.08250988e-01 6.25201821e-01
-1.15864336e+00 -1.57776701e+00 -5.69877028e-01 5.69805086e-01
1.15524805e+00 -5.35159171e-01 7.36471057e-01 1.21877611e-01
-6.44739449e-01 7.06298113e-01 -1.62639946e-01 3.34198952e-01
7.29108632e-01 -1.29411352e+00 -1.17759362e-01 4.38654929e-01
-2.52211422e-01 7.11641014e-01 8.65207076e-01 -3.87913257e-01
-1.13165987e+00 -4.29965496e-01 8.86291385e-01 -5.40348411e-01
1.02809095e+00 -4.67366725e-01 -1.10733175e+00 7.21245766e-01
7.22524941e-01 -7.21446276e-01 9.08858657e-01 4.75589454e-01
-4.10548568e-01 4.57832575e-01 -8.74587238e-01 5.80837786e-01
5.16124249e-01 -7.25362539e-01 -8.90293956e-01 3.29011232e-01
8.83910418e-01 -3.45623404e-01 -8.68664026e-01 3.00821483e-01
4.29412335e-01 -1.14100850e+00 5.44581592e-01 -9.17249322e-01
1.36935627e+00 3.46405432e-02 -5.14549792e-01 -1.02778244e+00
-2.85501987e-01 -4.85760957e-01 -2.92223632e-01 1.39678705e+00
2.34269813e-01 -3.68753448e-02 6.97690129e-01 3.80839854e-01
-4.84485418e-01 -9.10113692e-01 -7.68721819e-01 -2.48979092e-01
5.95104337e-01 -1.51972294e-01 3.04460406e-01 1.12957990e+00
8.89655292e-01 1.24131870e+00 -1.19520962e+00 -4.06162143e-01
9.71897244e-02 5.17323196e-01 9.09478724e-01 -8.57303739e-01
-5.55817306e-01 -5.10408700e-01 -6.43857643e-02 -1.38816082e+00
5.14010012e-01 -8.88438880e-01 2.99583316e-01 -9.72462535e-01
9.17385638e-01 -3.45112220e-03 -2.82349512e-02 4.56102967e-01
-2.71207392e-01 4.04453687e-02 4.71698403e-01 4.83627468e-01
-6.63483083e-01 1.06452358e+00 1.10133040e+00 -1.07870504e-01
-2.99867809e-01 -1.81785241e-01 -8.09113622e-01 7.41849661e-01
7.99368203e-01 -7.83652067e-02 -6.04960263e-01 -1.62789710e-02
-1.08277321e-01 6.26568198e-01 2.83954322e-01 -6.65105760e-01
1.45229161e-01 -5.39353132e-01 -3.01289558e-01 -7.77591705e-01
6.69064283e-01 -2.76471078e-01 -5.74881196e-01 2.70225167e-01
-8.48048270e-01 -4.71591949e-02 -6.50934130e-02 3.08667839e-01
-3.66202682e-01 -5.88508129e-01 1.09748125e+00 -1.43117234e-01
-5.92915893e-01 -3.24825525e-01 -6.12885237e-01 3.28951895e-01
7.30551481e-01 3.08147430e-01 -7.14212060e-01 -1.02145231e+00
-2.56273389e-01 3.39261234e-01 3.12713683e-01 5.82924128e-01
5.08737266e-01 -1.40990508e+00 -7.60895669e-01 -2.22218230e-01
4.05141264e-01 -4.99525279e-01 5.06241083e-01 1.07344103e+00
-2.90906847e-01 4.77744550e-01 -3.46497446e-01 -4.04831797e-01
-1.41765034e+00 6.98198318e-01 2.24185959e-01 -2.55907238e-01
-4.26971376e-01 5.18609703e-01 5.13330042e-01 -7.50871480e-01
1.99971780e-01 1.48824707e-01 -4.38926041e-01 -7.06401840e-03
4.95099753e-01 3.73012543e-01 -3.85039121e-01 -9.99499381e-01
-2.56909281e-01 3.52514237e-02 -2.29433760e-01 -1.53827995e-01
1.06209075e+00 -3.35224569e-01 -1.39213324e-01 9.39167738e-01
1.19430256e+00 -2.20960416e-02 -8.03877532e-01 -2.06042424e-01
-1.70579657e-01 -2.18272030e-01 -1.39924243e-01 -9.44194913e-01
-6.64591312e-01 1.21074605e+00 -4.19497676e-02 2.48953417e-01
8.86233389e-01 1.79038331e-01 1.10761476e+00 4.98525530e-01
-8.68526399e-02 -1.28342640e+00 8.01758945e-01 1.02452385e+00
1.27971375e+00 -1.47719800e+00 -1.07094504e-01 -1.30210340e-01
-1.54688537e+00 1.11683214e+00 1.10821176e+00 5.19202277e-02
2.91159242e-01 -6.22992106e-02 3.95549089e-01 -4.52742428e-01
-9.78506565e-01 1.31714672e-01 1.43697560e-01 6.08915538e-02
1.07553256e+00 8.41444731e-02 -6.46727562e-01 1.26571405e+00
-5.73382735e-01 -4.39860791e-01 8.26749504e-01 5.69773316e-01
-6.88906252e-01 -6.55371487e-01 -6.91037998e-02 8.20609033e-02
-2.87995726e-01 -9.70380902e-02 -9.84232426e-01 6.53886437e-01
-3.97844255e-01 1.30916214e+00 -4.20035064e-01 -7.35070884e-01
9.92191955e-02 1.70274571e-01 -1.40009731e-01 -5.45983076e-01
-1.04234886e+00 3.73580307e-02 2.55830526e-01 -1.25435442e-01
-3.72074127e-01 -4.93150353e-01 -1.44702685e+00 -6.28274441e-01
-2.40857571e-01 2.54567772e-01 1.39412880e-01 9.77173090e-01
1.50915338e-02 3.43362689e-01 9.70608354e-01 -5.96061289e-01
-1.08372140e+00 -1.53692973e+00 -5.69996238e-01 8.18916798e-01
1.86488256e-01 -7.37249792e-01 -4.28770602e-01 -2.04079852e-01]
|
[12.999309539794922, 5.609080791473389]
|
a944999c-48ad-4ddc-a8c7-cbad2d2b5690
|
a-large-scale-dataset-and-benchmark-for
|
1701.05766
| null |
http://arxiv.org/abs/1701.05766v2
|
http://arxiv.org/pdf/1701.05766v2.pdf
|
A Large-scale Dataset and Benchmark for Similar Trademark Retrieval
|
Trademark retrieval (TR) has become an important yet challenging problem due
to an ever increasing trend in trademark applications and infringement
incidents. There have been many promising attempts for the TR problem, which,
however, fell impracticable since they were evaluated with limited and mostly
trivial datasets. In this paper, we provide a large-scale dataset with
benchmark queries with which different TR approaches can be evaluated
systematically. Moreover, we provide a baseline on this benchmark using the
widely-used methods applied to TR in the literature. Furthermore, we identify
and correct two important issues in TR approaches that were not addressed
before: reversal of contrast, and presence of irrelevant text in trademarks
severely affect the TR methods. Lastly, we applied deep learning, namely,
several popular Convolutional Neural Network models, to the TR problem. To the
best of the authors, this is the first attempt to do so.
|
['Cemal Aker', 'Sinan Kalkan', 'Osman Tursun']
|
2017-01-20
| null | null | null | null |
['trademark-retrieval']
|
['computer-vision']
|
[ 2.81483740e-01 -3.86521816e-01 -1.00492172e-01 -6.05457947e-02
-1.02762830e+00 -8.48090053e-01 7.95331180e-01 -1.90839782e-01
-2.54893005e-01 6.72942877e-01 -1.49799228e-01 -4.89560962e-01
-6.05721891e-01 -5.90434134e-01 -7.36551881e-01 -3.14068109e-01
2.60885715e-01 4.08015460e-01 2.58474141e-01 -2.66908169e-01
8.21303546e-01 7.87222862e-01 -1.62199664e+00 2.80979782e-01
6.83329642e-01 1.39604580e+00 -3.23542804e-01 -5.41844070e-02
-3.71570557e-01 8.01802993e-01 -8.44141304e-01 -6.56179547e-01
6.98807240e-01 -1.10526964e-01 -7.54107594e-01 -4.33786720e-01
9.38973725e-01 -3.06856513e-01 -4.68694150e-01 1.00307250e+00
3.44448835e-01 3.86648215e-02 7.69242942e-01 -1.18396819e+00
-1.10504913e+00 4.37351197e-01 -7.10812330e-01 3.43340993e-01
1.52670532e-01 -3.60595211e-02 1.36093986e+00 -7.73116887e-01
8.47039521e-01 1.26840246e+00 5.68426073e-01 1.80981249e-01
-8.47640574e-01 -9.68785942e-01 -8.88643190e-02 3.60505670e-01
-1.08669686e+00 -2.54618645e-01 7.46540785e-01 -3.24912190e-01
7.66696930e-01 2.29254201e-01 1.25748158e-01 1.42608774e+00
3.22720408e-01 8.41816783e-01 1.35195291e+00 -2.54942745e-01
-3.22836116e-02 4.61255759e-02 4.07807708e-01 1.21275224e-01
4.32219565e-01 4.35943425e-01 -2.58448720e-01 -2.61288792e-01
6.25753224e-01 6.09037019e-02 -4.72275168e-02 -3.50068480e-01
-8.81875038e-01 7.02110231e-01 3.25568259e-01 6.99785948e-01
-1.74853563e-01 2.34172001e-01 1.77537501e-01 7.59092391e-01
4.88798708e-01 8.80887687e-01 -5.47822654e-01 1.85744837e-02
-8.85612190e-01 4.08338100e-01 1.03722394e+00 7.98766494e-01
4.61711109e-01 -3.35621059e-01 -2.07391933e-01 6.74430907e-01
1.09317929e-01 3.19377393e-01 3.50241780e-01 -8.04901958e-01
3.04443896e-01 3.79241109e-01 2.15848222e-01 -1.11712921e+00
-4.96470556e-02 -7.71559894e-01 -3.47084284e-01 4.89353202e-02
5.05012274e-01 2.66143620e-01 -8.69978845e-01 1.14203191e+00
-2.22536102e-01 1.64485231e-01 -1.23112366e-01 6.57011688e-01
6.99924946e-01 3.04755509e-01 -2.07085282e-01 -1.41709119e-01
1.21195185e+00 -1.17065728e+00 -9.65727806e-01 3.44503187e-02
3.20015639e-01 -1.26828969e+00 8.90864730e-01 6.44802868e-01
-7.39789009e-01 -3.83523792e-01 -1.09434998e+00 7.70828053e-02
-1.00500631e+00 -1.22136541e-01 7.37990916e-01 8.11354339e-01
-9.04588223e-01 8.22237790e-01 -9.64922011e-02 -3.65322441e-01
6.28066599e-01 1.95121467e-01 -5.31694517e-02 -1.73729107e-01
-1.46325111e+00 1.05699432e+00 1.84021518e-01 2.67652810e-01
-6.89513087e-01 -5.55118680e-01 -1.87245235e-01 5.81932999e-02
1.01647592e+00 -1.69810131e-01 1.38494825e+00 -6.84741259e-01
-1.17284632e+00 7.99395978e-01 2.93106347e-01 -5.04099429e-01
7.74553180e-01 -7.19507217e-01 -9.08596158e-01 -5.70960455e-02
1.00909173e-01 2.08115682e-01 8.11966181e-01 -1.18816519e+00
-5.51672757e-01 -4.50484604e-01 1.14356935e-01 -2.72218049e-01
-2.60143399e-01 1.98729709e-01 -4.42928255e-01 -8.18600118e-01
-1.25070989e-01 -8.48534346e-01 5.93427159e-02 -3.49711120e-01
-1.84658155e-01 -3.67831409e-01 7.16192245e-01 -4.31625247e-01
1.35021579e+00 -2.16599703e+00 -4.68555182e-01 1.97746441e-01
1.21131308e-01 4.98218775e-01 -2.72271842e-01 7.26820469e-01
-1.84512604e-03 6.33581698e-01 -5.85014969e-02 1.06952548e-01
3.39685380e-01 -3.38681880e-03 -8.93433213e-01 2.55695045e-01
9.25902054e-02 1.25096118e+00 -7.76873767e-01 -2.15524286e-01
-1.36901960e-01 2.11252838e-01 3.41109559e-02 2.44288519e-02
-4.99485344e-01 2.01714113e-01 -7.49638319e-01 8.69421422e-01
7.70697296e-01 -2.40438595e-01 -1.62693068e-01 -1.67261183e-01
-3.12746257e-01 2.49781087e-01 -9.75042164e-01 1.22975671e+00
-1.72278330e-01 6.46004796e-01 -3.96357507e-01 -7.67550111e-01
8.63198996e-01 4.30139266e-02 5.71252167e-01 -1.24326861e+00
1.02403395e-01 7.84862876e-01 -2.37949774e-01 -5.03639460e-01
7.75621176e-01 4.61403057e-02 1.21203482e-01 6.68933094e-01
-2.98448116e-01 3.42380464e-01 2.21314356e-01 -9.47490633e-02
1.34946382e+00 1.29662350e-01 -1.30577954e-02 2.91571748e-02
4.19917434e-01 9.29983854e-02 1.60886481e-01 1.22321343e+00
-1.90159470e-01 7.22884893e-01 4.47456062e-01 -5.17555118e-01
-6.72975600e-01 -8.05288732e-01 -3.03283632e-01 9.60714281e-01
2.29736745e-01 -6.60895929e-02 -3.91220421e-01 -9.46172059e-01
2.24997923e-01 5.36719620e-01 -6.04768932e-01 1.06832765e-01
-7.14148521e-01 -5.80659211e-01 7.23733366e-01 3.27982157e-01
6.06809556e-01 -1.31592739e+00 -2.78116524e-01 3.13304514e-01
1.70020506e-01 -1.02552986e+00 -4.89134818e-01 4.66165952e-02
-6.46209538e-01 -1.47238004e+00 -9.06860590e-01 -5.99937499e-01
1.13958821e-01 5.29671669e-01 1.01077306e+00 -7.27509905e-04
-2.90735692e-01 2.42959693e-01 -5.10326624e-01 -4.74150270e-01
-8.30117837e-02 3.84648532e-01 -4.29864705e-01 -2.98533915e-03
7.43555725e-01 -1.05370410e-01 -4.21582639e-01 5.15154779e-01
-1.32460177e+00 -9.48349774e-01 1.12149930e+00 6.74509346e-01
3.31708908e-01 1.01382747e-01 8.75830948e-01 -1.39008188e+00
9.53846812e-01 -4.35410261e-01 -9.50217664e-01 5.21833420e-01
-9.80976820e-01 1.29812449e-01 3.65892649e-01 -3.81274402e-01
-9.68694925e-01 -5.06028473e-01 -1.48932204e-01 -5.05893946e-01
-1.23799510e-01 7.35785484e-01 9.27756913e-03 -3.14884067e-01
5.73533416e-01 9.86390859e-02 -2.41641194e-01 -9.73918915e-01
4.32215273e-01 7.10231066e-01 1.50084764e-01 -4.45001334e-01
9.32940066e-01 3.75818253e-01 4.47073095e-02 -3.73846352e-01
-1.23843372e+00 -6.36851728e-01 -4.50937808e-01 2.07384229e-01
3.59330714e-01 -2.91047454e-01 -6.86438978e-01 3.38101089e-01
-1.14682710e+00 3.02584618e-01 -6.89985156e-02 1.10037923e-01
-2.76151150e-01 5.92666447e-01 -4.70722347e-01 -6.92082882e-01
-3.36428612e-01 -1.08229363e+00 8.53451312e-01 -9.94467083e-03
1.53465986e-01 -7.45393872e-01 1.57220721e-01 5.95055699e-01
9.34467614e-01 2.46583104e-01 9.16546226e-01 -1.23432791e+00
-9.24435794e-01 -2.88502812e-01 -4.73966807e-01 2.52408117e-01
1.84368491e-01 -6.21876158e-02 -1.15615571e+00 -3.47998679e-01
3.47329769e-04 -1.51241049e-01 1.27811432e+00 8.75979438e-02
1.13044846e+00 -1.64000615e-01 -4.59357709e-01 2.66164273e-01
1.31184340e+00 5.77819645e-01 7.99601078e-01 8.82980943e-01
4.41112459e-01 7.47857034e-01 9.78530288e-01 8.26349854e-02
-1.22055873e-01 6.86236084e-01 4.16928202e-01 9.65385512e-02
-2.03418527e-02 -9.15653259e-02 2.11768582e-01 2.72041440e-01
1.49925902e-01 -5.66605270e-01 -7.64795542e-01 3.15108567e-01
-1.91591358e+00 -8.37432206e-01 -2.74181992e-01 2.23370767e+00
3.75940681e-01 1.76184073e-01 -7.17846826e-02 -6.43685460e-02
6.80336893e-01 3.77493888e-01 -5.01854837e-01 -1.07760333e-01
-3.26363027e-01 4.47602510e-01 5.58392525e-01 3.62849720e-02
-1.29183388e+00 1.10637498e+00 6.66077518e+00 9.81464028e-01
-1.06220305e+00 -8.82589743e-02 3.39446217e-01 1.56083256e-01
-2.74400741e-01 6.25370219e-02 -8.67344737e-01 3.19320947e-01
9.10717726e-01 -2.47568250e-01 3.34775597e-01 6.34418368e-01
-3.51218849e-01 2.94115424e-01 -1.33289182e+00 8.55656803e-01
9.30918455e-02 -1.11822915e+00 1.37663022e-01 3.63625616e-01
6.07145250e-01 4.98639699e-03 6.87202215e-01 4.87055153e-01
1.29910648e-01 -1.19085979e+00 3.74743789e-01 4.62722093e-01
3.80705237e-01 -7.19366252e-01 1.05926120e+00 -1.82989135e-01
-6.94518805e-01 -1.94792688e-01 -4.12952155e-01 4.13477451e-01
-1.83699846e-01 5.68018854e-01 -4.83772755e-01 9.08260942e-01
6.75156057e-01 8.37340295e-01 -7.24221945e-01 1.28293300e+00
-1.95622563e-01 4.21666235e-01 7.11322995e-03 -2.80309506e-02
5.38966537e-01 -4.75522615e-02 5.11384428e-01 9.15239394e-01
4.47112292e-01 -3.22922021e-01 -1.16638742e-01 1.04309773e+00
-3.22426379e-01 2.70830601e-01 -9.00555372e-01 -4.40272957e-01
3.20431620e-01 1.19101298e+00 -4.25371021e-01 1.08020632e-02
-7.04504430e-01 6.06695116e-01 4.35189344e-02 7.06430018e-01
-6.33319855e-01 -4.22826767e-01 3.55849326e-01 3.91929299e-01
5.11183023e-01 4.92640771e-02 -8.08207244e-02 -1.07831419e+00
2.94268072e-01 -9.98170793e-01 7.10750461e-01 -5.67465782e-01
-1.74547398e+00 5.47350407e-01 -7.98445567e-02 -1.38900256e+00
-7.06974790e-02 -7.35997736e-01 -9.05470699e-02 6.42434835e-01
-2.09463620e+00 -1.00757945e+00 9.16895568e-02 2.44360954e-01
2.25755677e-01 -4.61423159e-01 5.05998969e-01 6.17419064e-01
-4.13286716e-01 7.33646274e-01 6.63398027e-01 -5.72050437e-02
1.17220008e+00 -1.02304447e+00 4.81006056e-01 5.04574060e-01
1.88277856e-01 9.76612568e-01 4.24567997e-01 -7.77924597e-01
-1.41844773e+00 -9.93185461e-01 9.06009138e-01 -6.38080716e-01
1.23696887e+00 -1.18594490e-01 -1.09368193e+00 6.31916940e-01
2.60762632e-01 -3.37270081e-01 6.00660741e-01 1.51476786e-01
-6.33209229e-01 -2.68773407e-01 -1.11847425e+00 5.46083808e-01
8.99652958e-01 -3.78419816e-01 -8.36154521e-01 3.72081578e-01
5.86685896e-01 -8.96552503e-02 -9.68455613e-01 3.04072440e-01
7.73714781e-01 -6.72730267e-01 1.04065180e+00 -5.04515350e-01
4.68013942e-01 -1.62783399e-01 -1.05243996e-01 -9.77023125e-01
-1.76173657e-01 -5.56012750e-01 -6.18812330e-02 1.37416112e+00
2.79717535e-01 -9.82288480e-01 7.03007162e-01 5.36286175e-01
1.08157203e-01 -5.90138078e-01 -8.38559806e-01 -1.21793807e+00
2.65129656e-01 -1.25845641e-01 7.60427296e-01 1.05388212e+00
-4.72242147e-01 1.88446850e-01 -4.66718316e-01 -1.87475294e-01
7.00369835e-01 4.87454653e-01 5.29864788e-01 -1.59277940e+00
-1.50799900e-01 -6.30670846e-01 9.12875310e-02 -1.06614745e+00
1.70992091e-01 -7.96350479e-01 -2.63291091e-01 -1.30044234e+00
8.93291980e-02 -2.95025557e-01 -9.42064881e-01 4.46331799e-01
3.83126587e-02 1.50813788e-01 9.96363536e-02 6.95223272e-01
-6.96919084e-01 2.40474597e-01 1.32826972e+00 -4.43643630e-01
8.32009539e-02 1.17364982e-02 -9.47009683e-01 2.35926256e-01
7.59156942e-01 -4.89113182e-01 -9.99538898e-02 -4.06133145e-01
3.82360607e-01 -8.77776220e-02 2.40906686e-01 -5.53420663e-01
3.79216641e-01 -9.03133899e-02 1.86206803e-01 -9.72880900e-01
8.83509144e-02 -1.02317667e+00 2.80498385e-01 2.74376243e-01
-6.49408102e-01 1.45717382e-01 9.95369405e-02 7.00471342e-01
-4.98979568e-01 -4.35388714e-01 4.74793911e-01 -3.41350585e-01
-8.02423716e-01 2.91504920e-01 -1.29610285e-01 -2.33190991e-02
8.46365213e-01 -2.40741789e-01 -7.34837830e-01 -2.56366283e-01
-2.81663626e-01 2.50102520e-01 3.26867849e-01 8.86046171e-01
5.02231598e-01 -1.19694018e+00 -5.65925658e-01 -1.59457088e-01
3.39444786e-01 -5.06567121e-01 -3.45911980e-02 5.67834377e-01
-3.82684231e-01 9.17641819e-01 -1.77376255e-01 -7.14588612e-02
-9.76338267e-01 1.01188266e+00 1.86615154e-01 -5.89500189e-01
-4.25695509e-01 2.81317741e-01 1.97125137e-01 -2.19689086e-01
4.38685209e-01 -7.96596333e-02 -5.47707319e-01 3.73743504e-01
5.79850912e-01 5.28153598e-01 3.79440486e-01 -4.16089445e-01
-2.79632151e-01 7.12108195e-01 -8.52703393e-01 -1.57723501e-02
1.52642584e+00 1.16873682e-01 -1.85075372e-01 2.57337928e-01
1.28082776e+00 -9.27450061e-02 -5.89802921e-01 -3.81927162e-01
6.43279731e-01 -7.60976970e-01 5.73478937e-02 -1.08316433e+00
-1.18851113e+00 8.30014706e-01 6.68435931e-01 4.67285991e-01
8.94514382e-01 -2.53979921e-01 9.78109300e-01 7.10146904e-01
4.51678127e-01 -1.17449450e+00 2.02032685e-01 6.26285672e-01
9.23927963e-01 -1.22734988e+00 -2.55472008e-02 -3.38716269e-01
-1.79431975e-01 1.14552724e+00 3.39223802e-01 4.50364910e-02
4.61278886e-01 -1.11852527e-01 1.90040424e-01 -5.01821220e-01
-6.27077281e-01 -1.78756982e-01 3.33256572e-01 3.35020512e-01
5.51529586e-01 -4.25978184e-01 -7.64545083e-01 2.61562973e-01
1.55944750e-01 2.27969632e-01 4.75301206e-01 1.02369535e+00
-3.28434229e-01 -1.43680465e+00 -4.02097493e-01 6.68203175e-01
-1.05383670e+00 -2.09303573e-01 -9.04164195e-01 1.08736134e+00
-3.40041667e-01 9.74935949e-01 -4.58027512e-01 -2.48168945e-01
6.22379124e-01 -1.58375353e-01 3.24942201e-01 -2.77409494e-01
-1.01158273e+00 3.19988370e-01 2.42824294e-02 -4.87572432e-01
-3.24320763e-01 -5.06247580e-01 -6.22305155e-01 -2.15568170e-01
-3.78155023e-01 5.81680536e-02 5.17634451e-01 8.58002543e-01
3.27156723e-01 3.57071936e-01 5.53996682e-01 -1.51034370e-01
-9.38361585e-01 -8.50098670e-01 -7.23011851e-01 5.41295111e-01
1.96784303e-01 -7.69044340e-01 -5.39943933e-01 -4.67056692e-01]
|
[9.857081413269043, 8.277076721191406]
|
bd822c1c-4152-4a00-a7fd-764283d38677
|
boosting-facial-expression-recognition-by-a
|
2205.14361
| null |
https://arxiv.org/abs/2205.14361v1
|
https://arxiv.org/pdf/2205.14361v1.pdf
|
Boosting Facial Expression Recognition by A Semi-Supervised Progressive Teacher
|
In this paper, we aim to improve the performance of in-the-wild Facial Expression Recognition (FER) by exploiting semi-supervised learning. Large-scale labeled data and deep learning methods have greatly improved the performance of image recognition. However, the performance of FER is still not ideal due to the lack of training data and incorrect annotations (e.g., label noises). Among existing in-the-wild FER datasets, reliable ones contain insufficient data to train robust deep models while large-scale ones are annotated in lower quality. To address this problem, we propose a semi-supervised learning algorithm named Progressive Teacher (PT) to utilize reliable FER datasets as well as large-scale unlabeled expression images for effective training. On the one hand, PT introduces semi-supervised learning method to relieve the shortage of data in FER. On the other hand, it selects useful labeled training samples automatically and progressively to alleviate label noise. PT uses selected clean labeled data for computing the supervised classification loss and unlabeled data for unsupervised consistency loss. Experiments on widely-used databases RAF-DB and FERPlus validate the effectiveness of our method, which achieves state-of-the-art performance with accuracy of 89.57% on RAF-DB. Additionally, when the synthetic noise rate reaches even 30%, the performance of our PT algorithm only degrades by 4.37%.
|
['Weihong Deng', 'Jing Jiang']
|
2022-05-28
| null | null | null | null |
['facial-expression-recognition']
|
['computer-vision']
|
[ 1.22296326e-01 -9.79674160e-02 -3.35595161e-01 -8.01026702e-01
-1.08935392e+00 5.76696694e-02 -1.21591337e-01 -3.27290505e-01
-4.86837715e-01 1.06596434e+00 -1.67939305e-01 4.13345546e-01
2.23897770e-01 -5.10959864e-01 -5.67019403e-01 -9.35802996e-01
2.69641101e-01 2.86115229e-01 -7.73896724e-02 -2.22469062e-01
-3.29957157e-01 2.50607461e-01 -1.73017871e+00 3.43905538e-01
1.09233701e+00 1.69227123e+00 3.09471264e-02 -1.49574563e-01
-2.56133020e-01 1.09363306e+00 -5.93045533e-01 -6.07956409e-01
1.06152914e-01 -3.79702717e-01 -5.93565106e-01 4.19997156e-01
4.15692478e-01 -4.44600999e-01 -1.15815751e-01 1.34912348e+00
6.25375450e-01 -1.65405557e-01 3.21982026e-01 -1.36902773e+00
-3.75857145e-01 2.77663171e-01 -9.16360617e-01 -7.78104365e-02
-1.63537070e-01 -2.12471765e-02 6.77093446e-01 -1.24369299e+00
5.98797798e-01 1.05616450e+00 6.61733270e-01 8.17915618e-01
-1.00598454e+00 -1.26783931e+00 3.22688222e-02 1.17696010e-01
-1.72533369e+00 -7.45804906e-01 9.80073929e-01 -2.31358200e-01
1.08360946e-01 9.72030237e-02 3.98627788e-01 1.04813969e+00
-3.80969524e-01 9.72830057e-01 1.25226986e+00 -2.73731291e-01
1.81471422e-01 2.04055011e-01 1.12789154e-01 9.68537271e-01
-1.40902065e-02 -6.37975931e-02 -6.82538629e-01 -1.04283124e-01
4.34504122e-01 -2.74413209e-02 -2.27093920e-01 -4.25127633e-02
-4.90467191e-01 6.09283864e-01 2.95334637e-01 2.67239928e-01
-3.45353842e-01 -2.68372416e-01 6.33577406e-01 2.36419201e-01
8.16909492e-01 -4.74392548e-02 -5.35676241e-01 -9.52850580e-02
-1.09361732e+00 -1.84581652e-01 3.10795069e-01 8.41663182e-01
9.96836603e-01 4.43161696e-01 2.39600614e-03 1.37141573e+00
1.94906637e-01 6.17008269e-01 4.27782446e-01 -9.33851957e-01
4.12786335e-01 5.90459406e-01 -1.29175447e-02 -9.83245969e-01
-6.26909360e-02 -5.40824711e-01 -1.14187920e+00 -2.09659878e-02
4.37613100e-01 -1.48488626e-01 -1.00910008e+00 1.90857375e+00
2.59579182e-01 2.49737993e-01 1.80435181e-02 1.05152392e+00
9.29966569e-01 4.38553780e-01 3.36417168e-01 -7.05423117e-01
1.05792284e+00 -1.03155029e+00 -9.54976022e-01 8.69065598e-02
6.81398392e-01 -6.47571146e-01 1.02585387e+00 6.26512885e-01
-8.50721657e-01 -6.70016289e-01 -8.82771432e-01 3.24407011e-01
5.29509820e-02 6.67000055e-01 6.20244801e-01 4.72213924e-01
-7.22924709e-01 1.76234260e-01 -6.57675326e-01 7.51635209e-02
1.12175179e+00 3.45225632e-01 -6.64264500e-01 -3.13437015e-01
-1.19975412e+00 3.18718642e-01 1.50400519e-01 3.38934094e-01
-1.07678485e+00 -6.19062960e-01 -7.40154982e-01 -1.77072778e-01
4.93028551e-01 5.40679991e-02 1.14056194e+00 -1.83409560e+00
-1.53596866e+00 1.05469596e+00 -1.51942536e-01 -1.37622327e-01
5.32460332e-01 -1.70286372e-01 -7.30387390e-01 2.66526669e-01
4.13843617e-02 6.34171426e-01 9.13583219e-01 -1.38456643e+00
-5.89667559e-01 -7.14999676e-01 -3.40376616e-01 -1.09683655e-01
-6.84325874e-01 1.31700307e-01 -5.59216976e-01 -5.46880305e-01
-5.67327160e-03 -8.21740627e-01 4.93238010e-02 2.54549235e-01
-1.73648611e-01 -2.29929492e-01 1.00020802e+00 -6.56500816e-01
1.06423950e+00 -2.34786010e+00 -3.85078013e-01 9.58095789e-02
1.59592807e-01 6.19524121e-01 -2.87522972e-01 -3.08024406e-01
-1.84890226e-01 -1.98682882e-02 -2.13743299e-01 -5.29966772e-01
-4.30835396e-01 3.81633341e-01 -1.25972599e-01 4.40339148e-01
4.32327777e-01 5.93930602e-01 -8.79504383e-01 -8.16265404e-01
-1.34096682e-01 6.19602859e-01 -3.83940279e-01 4.35265720e-01
-4.40770499e-02 6.16340697e-01 -6.70474827e-01 1.07637084e+00
8.68775487e-01 -2.41545156e-01 1.83411296e-02 -4.98398215e-01
3.84039700e-01 -2.68541098e-01 -1.11602783e+00 1.64290869e+00
-3.46865624e-01 3.32359552e-01 3.03482503e-01 -1.29775655e+00
1.18930662e+00 3.09807837e-01 7.27433205e-01 -8.58238637e-01
2.92126507e-01 5.03545582e-01 -4.14661467e-01 -6.20711803e-01
2.10099265e-01 -2.78115273e-01 2.54476696e-01 5.25897406e-02
2.90043890e-01 4.81102079e-01 8.26424211e-02 -5.43199070e-02
8.11675370e-01 1.68284699e-01 -2.56363004e-02 1.37891597e-03
6.22526228e-01 -1.28098831e-01 1.27708626e+00 1.05655901e-01
-5.07653058e-01 4.32060450e-01 2.95459300e-01 -4.27723706e-01
-6.86647058e-01 -7.12238789e-01 -4.23069596e-01 1.05948257e+00
1.06746510e-01 -2.30646059e-01 -8.66677642e-01 -1.02061522e+00
-2.28957653e-01 1.58607200e-01 -6.49011433e-01 -3.06352168e-01
-2.82802790e-01 -9.71790195e-01 6.92627788e-01 4.17891115e-01
9.83912468e-01 -1.01558733e+00 -9.50350389e-02 -4.81494367e-02
-4.26842660e-01 -1.13737750e+00 -2.03017056e-01 -5.78571260e-02
-7.14861751e-01 -1.13868093e+00 -7.02749789e-01 -7.10368216e-01
1.05897069e+00 1.21939085e-01 1.02488875e+00 1.61408290e-01
-2.08005354e-01 -2.04067126e-01 -4.08010393e-01 -2.72414893e-01
-2.84437746e-01 -3.28379393e-01 4.67519537e-02 5.26362002e-01
4.75272536e-01 -3.70567799e-01 -4.56462055e-01 6.78813577e-01
-9.09078896e-01 -1.89394146e-01 5.23339629e-01 1.41813123e+00
1.02912748e+00 2.16248453e-01 8.41861010e-01 -8.64075899e-01
1.48493350e-01 -4.07807380e-01 -4.76724535e-01 2.76890635e-01
-5.50149322e-01 -2.14858040e-01 7.30105221e-01 -4.91204113e-01
-1.34127891e+00 8.16102698e-02 -2.14120835e-01 -8.40322196e-01
-9.98098776e-02 4.45785254e-01 -5.88496745e-01 -1.68214783e-01
4.53165621e-01 7.97596872e-02 8.05154592e-02 -5.09351492e-01
-9.49763358e-02 9.91013169e-01 5.30354798e-01 -7.06770778e-01
4.44981515e-01 5.06486714e-01 -5.10212369e-02 -5.65862775e-01
-1.22901559e+00 -2.27947533e-01 -3.14403445e-01 -2.81091571e-01
4.63059574e-01 -1.25156355e+00 -4.12521720e-01 7.58550525e-01
-6.95108771e-01 -2.44386092e-01 -3.46492261e-01 3.23643327e-01
-2.36141860e-01 1.69249713e-01 -6.34182930e-01 -9.47414875e-01
-5.17182350e-01 -1.23890316e+00 1.36918068e+00 2.71298051e-01
3.14048171e-01 -4.88783449e-01 -2.81126589e-01 7.77866542e-01
2.51431197e-01 2.45004252e-01 2.74791151e-01 -5.69370091e-01
-1.59200579e-01 -2.17181265e-01 -3.95458370e-01 9.35792685e-01
2.15999782e-01 5.89287095e-02 -1.23408175e+00 -2.13741124e-01
-6.54599145e-02 -1.16381598e+00 8.98266137e-01 -9.34436619e-02
1.73726308e+00 -2.24697605e-01 -6.57803705e-03 6.37077689e-01
1.31605375e+00 4.95795161e-02 6.66413248e-01 6.55730367e-02
6.04484320e-01 6.82202041e-01 1.16201866e+00 6.20755911e-01
2.16593936e-01 6.89018905e-01 2.43110180e-01 -2.60612279e-01
-1.93627104e-01 -2.99537301e-01 2.81913608e-01 8.52497160e-01
-5.53682931e-02 7.48549476e-02 -6.12471223e-01 2.45693907e-01
-1.64874864e+00 -8.02511513e-01 1.47124022e-01 2.11337638e+00
1.31810462e+00 -5.56783266e-02 -3.61707807e-02 2.02381149e-01
6.71834469e-01 -6.24633506e-02 -6.59526169e-01 2.93088347e-01
-4.71127599e-01 2.62391329e-01 3.08346450e-01 9.77452472e-03
-1.13228703e+00 9.95368421e-01 4.55686235e+00 1.42164457e+00
-1.41747487e+00 3.64710808e-01 1.25833452e+00 3.77614191e-03
6.46041706e-02 -4.00244206e-01 -8.39713395e-01 6.44542754e-01
6.67568147e-01 3.87245983e-01 1.38192773e-01 1.15995026e+00
3.10906917e-01 1.29917907e-02 -5.50018311e-01 1.37972462e+00
3.00753295e-01 -8.43779206e-01 -7.64971226e-02 -2.08862931e-01
7.70385027e-01 -1.51577696e-01 -2.15826905e-04 3.82448792e-01
1.00776002e-01 -8.79158020e-01 5.33738375e-01 4.69520867e-01
1.24224341e+00 -9.26685154e-01 1.02396798e+00 1.68637663e-01
-9.40861702e-01 3.39389108e-02 -5.33078849e-01 3.45370263e-01
-6.92250580e-02 9.52817380e-01 -2.95628965e-01 4.86714542e-01
8.28203440e-01 8.79046917e-01 -5.46383977e-01 6.88472629e-01
-2.82588929e-01 9.67014253e-01 -4.08592373e-01 3.97755057e-01
2.49331206e-04 -2.54864395e-01 2.19261423e-02 7.97463834e-01
1.85978755e-01 7.61972070e-02 3.30704808e-01 5.31359971e-01
-5.98945975e-01 3.55686188e-01 -1.60443708e-01 -2.53436528e-02
4.66045201e-01 1.35564387e+00 -2.10984379e-01 -3.02540213e-01
-2.45720103e-01 8.48701298e-01 5.63750029e-01 4.10559297e-01
-8.07468593e-01 -1.20836698e-01 5.10714531e-01 5.25737852e-02
1.70271024e-02 2.47131377e-01 8.54164660e-02 -1.28464794e+00
1.71475768e-01 -1.07444191e+00 4.52092111e-01 -7.75326610e-01
-1.45558345e+00 8.96358848e-01 -4.81528878e-01 -1.26436353e+00
5.07840849e-02 -2.88303822e-01 -5.98020181e-02 4.49455410e-01
-1.76939082e+00 -1.30064321e+00 -6.34831846e-01 8.23397219e-01
4.36872065e-01 -4.34238821e-01 8.75086606e-01 8.41540158e-01
-9.93995667e-01 1.13255990e+00 1.31359607e-01 4.01548028e-01
1.10707402e+00 -7.12743700e-01 -6.31850123e-01 5.83614171e-01
1.04791291e-01 2.30307221e-01 2.91278064e-01 -4.02352363e-01
-1.23060215e+00 -1.31817627e+00 4.74756867e-01 2.03476965e-01
4.77823257e-01 -1.51367888e-01 -9.46733057e-01 2.75682747e-01
-3.05171132e-01 8.13494027e-01 8.70773613e-01 -1.41235456e-01
-5.53166926e-01 -7.76911974e-01 -1.45235598e+00 3.01293343e-01
1.00216305e+00 -4.52333927e-01 -1.72125036e-03 3.95575851e-01
3.61619115e-01 -2.97764003e-01 -9.71677363e-01 7.52329290e-01
4.60811675e-01 -8.48369896e-01 5.12754858e-01 -5.82475364e-01
4.25762445e-01 -2.56561041e-01 -3.15890938e-01 -1.10930705e+00
1.95037469e-01 -4.33266282e-01 7.53923282e-02 1.57136381e+00
1.08825766e-01 -4.62194204e-01 9.92076993e-01 5.90507329e-01
2.47496143e-02 -1.00573790e+00 -9.62646723e-01 -6.56400263e-01
-3.67892802e-01 -3.93653840e-01 4.50052381e-01 1.19075036e+00
-3.39891195e-01 2.26055995e-01 -6.33315563e-01 -1.53815478e-01
7.40008414e-01 7.88182616e-02 7.22182453e-01 -1.07069027e+00
1.20926872e-02 -3.00403982e-02 -4.18524414e-01 -8.43214571e-01
5.86422145e-01 -5.95327556e-01 1.22047536e-01 -8.86669576e-01
3.91455472e-01 -8.24391067e-01 -4.87668902e-01 9.32887495e-01
-2.77409673e-01 6.65918529e-01 -7.21864542e-03 3.99270296e-01
-1.01371181e+00 9.94358599e-01 1.26759541e+00 -2.45554253e-01
5.93574308e-02 -2.04321727e-01 -6.15030766e-01 8.17849457e-01
8.39467943e-01 -5.45237184e-01 -3.69196206e-01 -3.32532525e-01
-2.14839429e-01 -1.34983763e-01 1.62795737e-01 -7.70617008e-01
-1.32029578e-01 -1.05276637e-01 4.31661487e-01 -1.72263443e-01
3.05010617e-01 -8.85246515e-01 -6.99876472e-02 1.87142882e-02
-2.83611387e-01 -5.02059758e-01 7.79759511e-02 5.07242620e-01
-7.04395175e-01 5.40515371e-02 1.14284229e+00 1.40914902e-01
-8.13697338e-01 6.90235794e-01 2.19950125e-01 2.58971095e-01
7.96607137e-01 -5.61018009e-04 -2.10107893e-01 -3.74141425e-01
-5.73314071e-01 2.43101254e-01 3.13654453e-01 2.64087260e-01
6.21660292e-01 -1.46275437e+00 -7.12953210e-01 2.72016883e-01
4.05685633e-01 9.01862383e-02 3.99882436e-01 8.67098927e-01
-1.67125627e-01 -1.44922867e-01 -2.48138815e-01 -5.27992845e-01
-1.45099723e+00 2.39728838e-01 4.63479996e-01 -2.31235385e-01
-8.46990794e-02 9.76316154e-01 3.77273336e-02 -3.25215548e-01
3.65557551e-01 2.85989434e-01 -1.73972085e-01 6.09455816e-02
7.92588890e-01 3.70748229e-02 1.19361348e-01 -1.02666259e+00
-2.64896393e-01 4.50953543e-01 -6.13906309e-02 3.61885667e-01
1.39683640e+00 -7.85785168e-02 -2.41016209e-01 2.69785643e-01
1.47535384e+00 -1.30368486e-01 -1.29601872e+00 -4.03541058e-01
-2.95622528e-01 -6.70158684e-01 2.62186915e-01 -8.11220825e-01
-1.75367045e+00 9.10824537e-01 9.21999037e-01 -5.75637341e-01
1.63767755e+00 -2.90735632e-01 9.73896921e-01 3.25379848e-01
6.12740338e-01 -1.30761373e+00 2.73919731e-01 9.48814675e-02
6.18741989e-01 -1.55293703e+00 -1.53816149e-01 -5.20497203e-01
-8.44272017e-01 8.65431130e-01 9.45466399e-01 2.37123162e-01
6.82842433e-01 3.01720589e-01 3.20985109e-01 -7.75947869e-02
-5.48376024e-01 -1.68044984e-01 7.46154487e-02 4.07300353e-01
2.82096446e-01 -1.29317269e-01 -2.67808020e-01 1.01786900e+00
1.43804401e-01 5.58351994e-01 -2.77031381e-02 7.39792466e-01
-2.53635883e-01 -9.94827867e-01 -1.88662991e-01 5.15314400e-01
-6.86855018e-01 1.12402067e-01 -1.01901859e-01 5.42877138e-01
3.91094625e-01 1.02204645e+00 -1.58061966e-01 -5.27085006e-01
2.43509635e-01 4.39051501e-02 2.56339878e-01 -2.11423814e-01
-1.72930360e-01 2.13437632e-01 2.42438361e-01 -6.83988392e-01
-8.15662324e-01 -3.40809256e-01 -1.22650480e+00 -4.66453396e-02
-5.15861928e-01 3.28192681e-01 4.83889937e-01 9.06238616e-01
4.30649191e-01 2.25697756e-01 9.33797717e-01 -2.86845922e-01
-6.37351096e-01 -9.69912350e-01 -9.22755003e-01 7.49915183e-01
5.45045175e-02 -9.11036253e-01 -2.72142678e-01 2.46650368e-01]
|
[13.603584289550781, 1.6525810956954956]
|
e72ce040-2c8c-4e78-b0d3-35239f5acb5e
|
a-perceptual-quality-metric-for-video-frame
|
2210.01879
| null |
https://arxiv.org/abs/2210.01879v1
|
https://arxiv.org/pdf/2210.01879v1.pdf
|
A Perceptual Quality Metric for Video Frame Interpolation
|
Research on video frame interpolation has made significant progress in recent years. However, existing methods mostly use off-the-shelf metrics to measure the quality of interpolation results with the exception of a few methods that employ user studies, which is time-consuming. As video frame interpolation results often exhibit unique artifacts, existing quality metrics sometimes are not consistent with human perception when measuring the interpolation results. Some recent deep learning-based perceptual quality metrics are shown more consistent with human judgments, but their performance on videos is compromised since they do not consider temporal information. In this paper, we present a dedicated perceptual quality metric for measuring video frame interpolation results. Our method learns perceptual features directly from videos instead of individual frames. It compares pyramid features extracted from video frames and employs Swin Transformer blocks-based spatio-temporal modules to extract spatio-temporal information. To train our metric, we collected a new video frame interpolation quality assessment dataset. Our experiments show that our dedicated quality metric outperforms state-of-the-art methods when measuring video frame interpolation results. Our code and model are made publicly available at \url{https://github.com/hqqxyy/VFIPS}.
|
['Feng Liu', 'Abhijay Ghildyal', 'Qiqi Hou']
|
2022-10-04
| null | null | null | null |
['video-frame-interpolation']
|
['computer-vision']
|
[-2.46949315e-01 -7.70179987e-01 -2.95210570e-01 -4.27213520e-01
-9.31092322e-01 -2.98888743e-01 3.15913588e-01 7.05953836e-02
-3.11107725e-01 5.92716515e-01 3.09941918e-01 -1.61729142e-01
1.44384116e-01 -7.39380538e-01 -8.76394093e-01 -4.07802671e-01
-2.21190378e-01 -4.55291808e-01 4.44979340e-01 -1.46666497e-01
3.89165789e-01 1.75932005e-01 -1.58488882e+00 6.83399081e-01
8.30138564e-01 1.31293738e+00 1.57950804e-01 8.95202398e-01
2.05782503e-01 1.06367111e+00 -6.20420456e-01 -3.72099906e-01
3.51378679e-01 -5.11546850e-01 -6.92405701e-01 -1.68359563e-01
5.61756194e-01 -8.53347301e-01 -6.62686646e-01 9.73250747e-01
4.80861962e-01 -2.25216877e-02 3.72806191e-01 -1.54789853e+00
-1.10931110e+00 3.98041815e-01 -3.83467168e-01 4.78774518e-01
8.57705653e-01 4.39723611e-01 9.78438973e-01 -1.06133211e+00
3.60089153e-01 1.15140438e+00 9.02558029e-01 3.13301593e-01
-9.67607915e-01 -6.88055098e-01 -2.18926057e-01 8.86630833e-01
-1.36911118e+00 -5.78306198e-01 7.31391609e-01 -6.08789265e-01
7.37042964e-01 2.89888859e-01 6.93961620e-01 9.80533481e-01
1.82535037e-01 8.30567539e-01 9.50586677e-01 -2.51657307e-01
2.60904044e-01 -3.59466523e-01 -2.53005058e-01 4.48567808e-01
-2.57045627e-01 2.88191140e-01 -6.43038511e-01 1.42246664e-01
1.08214176e+00 2.01473013e-02 -4.03666586e-01 -1.58907115e-01
-1.48801017e+00 5.15525162e-01 5.17692506e-01 3.67296606e-01
-2.70108879e-01 2.83989787e-01 5.23921907e-01 5.32402039e-01
3.82032931e-01 -1.50146969e-02 -3.15108359e-01 -5.79725862e-01
-1.31092501e+00 4.68269467e-01 4.70633358e-01 1.02362287e+00
8.57690811e-01 9.10748169e-02 -5.36668837e-01 7.97358453e-01
1.67759225e-01 3.11508745e-01 3.16459209e-01 -1.59874439e+00
2.94752568e-01 2.00881928e-01 3.57070535e-01 -1.18689227e+00
-7.51500800e-02 1.64498344e-01 -7.67869830e-01 4.28339362e-01
5.77669144e-01 1.46332517e-01 -4.89960253e-01 1.43924081e+00
-6.47876337e-02 4.50047463e-01 -4.00435865e-01 1.23434711e+00
1.00225604e+00 7.15347052e-01 1.44573271e-01 -1.03528239e-01
1.00464320e+00 -1.01120961e+00 -7.55658746e-01 4.90255594e-01
3.22763056e-01 -1.04817474e+00 1.42240262e+00 5.41784227e-01
-1.51181054e+00 -1.12910330e+00 -1.16587532e+00 -3.18438172e-01
-5.41240275e-02 -1.44183235e-02 3.55829686e-01 5.56376338e-01
-1.43876994e+00 1.02299178e+00 -7.47098327e-01 -6.10171519e-02
4.90409493e-01 1.01630084e-01 -2.06652284e-01 -5.42581268e-02
-1.35324121e+00 6.59684300e-01 1.70416459e-02 -4.23292071e-02
-9.98917103e-01 -8.27656507e-01 -8.75810683e-01 -1.21423811e-01
1.00509919e-01 -6.03511930e-01 1.45454454e+00 -1.27012396e+00
-1.47681963e+00 4.54291046e-01 -3.92770827e-01 -3.30537498e-01
7.52951860e-01 -3.83095890e-01 -6.43686116e-01 3.15976292e-01
6.61686733e-02 8.06923747e-01 9.88814414e-01 -1.20573783e+00
-8.07548583e-01 1.13556355e-01 3.87190104e-01 -1.63743988e-01
-2.47025728e-01 2.18829438e-01 -5.74150145e-01 -9.31086004e-01
-3.03416044e-01 -4.98211831e-01 1.03233121e-01 4.87488955e-01
-5.21536125e-03 -2.23614708e-01 8.39489520e-01 -8.16224277e-01
1.60248625e+00 -2.14218378e+00 -1.27823755e-01 -2.75156200e-01
1.72724441e-01 2.47533277e-01 -1.86429828e-01 3.69486451e-01
1.29359961e-02 2.19493523e-01 -3.54237147e-02 -2.92454690e-01
4.54034731e-02 -1.36117548e-01 -1.04148373e-01 3.49276036e-01
2.20087767e-01 6.69784427e-01 -1.07480884e+00 -6.87755227e-01
5.48266888e-01 7.93896973e-01 -7.63104737e-01 3.16937864e-01
-9.88825038e-03 6.47185147e-01 -1.74520183e-02 8.10932696e-01
8.32274616e-01 -2.12302223e-01 -2.50369102e-01 -6.63858593e-01
-2.41997272e-01 1.78013876e-01 -1.09322405e+00 2.07364130e+00
-4.72984880e-01 1.00214493e+00 -4.22318846e-01 -6.66846335e-01
5.26781142e-01 5.19966662e-01 6.75225437e-01 -8.33871901e-01
-1.41990641e-02 1.76484630e-01 -1.35279939e-01 -7.73033023e-01
6.84934497e-01 3.62318784e-01 3.59034210e-01 4.18248698e-02
6.23748079e-02 4.68060076e-02 4.45224226e-01 6.61549121e-02
1.12837410e+00 5.02565086e-01 1.98736012e-01 -9.58221704e-02
6.45651877e-01 -3.81893188e-01 6.94238544e-01 5.68778932e-01
-6.29692674e-01 1.11122131e+00 3.07726145e-01 -6.94901228e-01
-1.43844283e+00 -1.44743538e+00 -1.70753270e-01 1.02494824e+00
3.41784120e-01 -7.25973964e-01 -9.21381354e-01 -3.35726738e-01
-1.65806428e-01 2.94610053e-01 -5.62487602e-01 1.93291217e-01
-5.99802792e-01 -8.92313495e-02 4.73563254e-01 6.04156673e-01
7.86844075e-01 -1.02081311e+00 -6.51384115e-01 2.01147065e-01
-4.37480450e-01 -1.18619514e+00 -7.02027678e-01 -7.70518720e-01
-7.61334836e-01 -9.48755860e-01 -9.99370456e-01 -6.80573404e-01
1.96194068e-01 3.17276061e-01 1.45255244e+00 4.81771439e-01
-9.20718387e-02 1.54545307e-01 -5.50663710e-01 3.04530114e-02
-3.89901072e-01 -2.54057944e-01 1.72789453e-03 -1.28306136e-01
4.25170064e-01 -6.51986241e-01 -1.17648578e+00 6.36413634e-01
-9.40792680e-01 1.59132078e-01 2.11188495e-01 5.53516984e-01
5.22577465e-01 -2.16152444e-01 4.62431163e-01 -1.26522213e-01
5.42835891e-01 -3.73317301e-01 -3.83698851e-01 -1.14734145e-02
-1.85447931e-01 -2.98643082e-01 7.81981647e-01 -3.48356098e-01
-7.36749113e-01 -3.90433878e-01 -3.21333706e-01 -7.61155963e-01
-2.31946275e-01 3.21768880e-01 -5.37938066e-02 -1.04572168e-02
6.89653754e-01 -1.05879828e-02 -2.82430291e-01 -3.30834359e-01
5.71657196e-02 6.22788608e-01 6.05402946e-01 -5.00186563e-01
5.74344933e-01 4.18557316e-01 -2.21825004e-01 -6.31259799e-01
-4.89914119e-01 -1.98967859e-01 -5.71736991e-01 -6.48812711e-01
8.03311467e-01 -1.10779762e+00 -8.45110536e-01 5.00435591e-01
-1.17231953e+00 -4.80971605e-01 -5.62856756e-02 5.68947613e-01
-7.81788230e-01 6.87400639e-01 -9.06065404e-01 -4.91097778e-01
-1.19854562e-01 -1.51007104e+00 1.06884539e+00 1.34035781e-01
-1.82638094e-01 -8.01480889e-01 -5.99143188e-03 2.15454161e-01
6.59488559e-01 2.96705693e-01 4.07868087e-01 3.34689796e-01
-6.91845536e-01 6.87295794e-02 -5.03780544e-01 5.25563300e-01
3.18571448e-01 4.50155318e-01 -9.88781631e-01 -2.18248188e-01
-2.24749118e-01 -1.37173399e-01 7.27110863e-01 6.45542681e-01
1.62735617e+00 -1.72280714e-01 2.14368314e-01 7.79949844e-01
1.40073252e+00 1.26507878e-01 1.14037740e+00 4.05640930e-01
6.34169400e-01 1.16205685e-01 6.30554795e-01 6.67148650e-01
5.75447023e-01 8.78242135e-01 3.33261758e-01 -4.02529910e-03
-2.88061529e-01 -1.63333282e-01 4.98653829e-01 7.58126915e-01
-6.71681464e-01 -1.76007152e-01 -7.79907525e-01 3.72363806e-01
-1.81785357e+00 -1.31331706e+00 -6.39342740e-02 2.13298225e+00
9.81523514e-01 1.51875064e-01 4.05281514e-01 5.41556656e-01
5.69021106e-01 1.34024367e-01 -3.78210127e-01 -3.24359208e-01
-1.54452780e-02 8.10482204e-02 2.97019273e-01 3.99862826e-01
-1.29963696e+00 5.85813761e-01 6.29820633e+00 7.21149385e-01
-1.07424557e+00 2.10873768e-01 6.99701130e-01 -1.09113067e-01
-2.50437856e-01 -7.23783374e-02 -1.34787664e-01 7.97477543e-01
8.99256706e-01 -1.22508973e-01 4.66856062e-01 6.49583459e-01
7.25503385e-01 -1.29943475e-01 -1.37307119e+00 1.47856104e+00
-1.69896558e-01 -1.53069007e+00 -1.77005492e-02 -2.91536480e-01
6.66754603e-01 -1.67795092e-01 2.82745689e-01 1.39874965e-01
-1.34422183e-01 -1.09353590e+00 1.06328130e+00 8.01054537e-01
9.61734474e-01 -5.77731729e-01 6.08510256e-01 -1.83405325e-01
-1.44903982e+00 7.27104992e-02 -4.39655006e-01 -3.43651086e-01
1.96809754e-01 6.88068807e-01 -1.27868161e-01 4.55220282e-01
1.23437369e+00 1.06994522e+00 -7.10477233e-01 1.43290365e+00
3.12789828e-02 5.55850387e-01 1.80701703e-01 3.62237632e-01
-3.47468704e-02 -2.31696269e-03 2.87815422e-01 1.35195518e+00
6.28574014e-01 5.43612465e-02 6.25207424e-02 8.22534919e-01
-1.17077515e-01 1.03532232e-01 -3.63438576e-01 2.83522606e-01
4.28983152e-01 9.44197953e-01 -2.72056937e-01 -4.69941825e-01
-7.91096926e-01 9.56975102e-01 -7.31943175e-02 4.03432071e-01
-1.35242271e+00 -3.38295996e-01 1.01598716e+00 1.86335802e-01
1.64460674e-01 -4.89300877e-01 -3.00971240e-01 -1.30491579e+00
2.55501509e-01 -9.11887348e-01 1.45114645e-01 -9.45488572e-01
-1.25386655e+00 6.17297649e-01 -4.92905360e-03 -1.92587256e+00
-2.16445744e-01 -3.37594539e-01 -4.99087483e-01 7.18656421e-01
-1.38781083e+00 -8.24569821e-01 -7.01056600e-01 8.64098191e-01
8.83120358e-01 -3.52745503e-02 6.09345019e-01 7.91949987e-01
-3.21790785e-01 8.19899797e-01 -7.30279908e-02 4.00688171e-01
9.24897373e-01 -9.80135500e-01 5.50311983e-01 9.54787850e-01
-9.66727510e-02 4.45383549e-01 8.50581825e-01 -3.12516361e-01
-1.43764293e+00 -8.97114098e-01 6.39830232e-01 -3.68548214e-01
5.30155838e-01 -9.32032894e-03 -1.02386069e+00 2.81580478e-01
5.13126433e-01 3.53654742e-01 6.08571649e-01 -3.68087769e-01
-4.10693526e-01 -2.77355045e-01 -1.12343919e+00 6.83780968e-01
1.14272702e+00 -5.62208951e-01 -2.39124298e-01 -1.79298464e-02
6.85334206e-01 -3.59747499e-01 -1.26279938e+00 4.71093893e-01
8.38433027e-01 -1.60910916e+00 1.12196362e+00 -1.86230734e-01
8.80861044e-01 -5.95554888e-01 -3.79088521e-01 -1.03764164e+00
-4.96382952e-01 -4.89415765e-01 -2.57613391e-01 1.02417374e+00
1.76343903e-01 -4.52034883e-02 5.78866541e-01 5.11664212e-01
-8.13196450e-02 -6.25501871e-01 -8.11559916e-01 -8.47559929e-01
-6.28341362e-02 -6.48738205e-01 8.02277148e-01 8.81907940e-01
-6.05712458e-02 -2.47742355e-01 -6.13766015e-01 -5.84761463e-02
6.65177584e-01 -8.38266686e-02 7.55971074e-01 -7.90281653e-01
-2.35632211e-01 -5.29718578e-01 -7.53717840e-01 -9.00616407e-01
-1.28717437e-01 -4.46563333e-01 -1.00894064e-01 -1.46419108e+00
2.90105045e-02 -1.83087215e-01 -4.45360512e-01 1.62738472e-01
-1.59808621e-01 6.16924465e-01 4.89347607e-01 2.51352191e-01
-7.34644651e-01 4.52764988e-01 1.33855760e+00 -2.04494134e-01
-1.00684524e-01 -2.76330620e-01 -2.18147367e-01 7.28236496e-01
9.05221820e-01 -2.64368299e-02 -3.15402746e-01 -8.95712197e-01
1.42225683e-01 2.13844642e-01 6.11352623e-01 -1.64436173e+00
7.87668005e-02 -1.93522170e-01 7.60239661e-01 -4.67787474e-01
2.87252545e-01 -6.95310473e-01 3.40728730e-01 2.56346613e-01
-3.37959170e-01 4.43762362e-01 5.16950451e-02 2.44093761e-01
-5.08089066e-01 1.20100200e-01 9.25467491e-01 -1.37405274e-02
-1.06144416e+00 5.68686545e-01 -2.36997038e-01 -1.36943355e-01
7.00433135e-01 -3.56498778e-01 -1.37876332e-01 -6.35451913e-01
-5.76696634e-01 -1.36597767e-01 8.60729337e-01 4.86824065e-01
9.69000041e-01 -1.74855089e+00 -9.07074869e-01 6.71441704e-02
9.08914860e-03 -4.41198200e-01 4.06239480e-01 8.02130163e-01
-8.63881469e-01 3.50976400e-02 -6.03597939e-01 -8.86314452e-01
-1.09605682e+00 7.39167213e-01 1.73274830e-01 2.27936819e-01
-4.63695198e-01 5.67243814e-01 3.47003303e-02 2.29601368e-01
2.54928023e-01 -6.99340940e-01 -1.76443961e-02 -1.14463560e-01
9.19701278e-01 6.24897540e-01 -2.15719640e-02 -6.55414343e-01
-2.45562568e-01 5.81403196e-01 3.31404716e-01 -4.69597019e-02
1.02215600e+00 -2.91859031e-01 1.24283962e-01 4.50987488e-01
1.42382789e+00 -2.50978172e-01 -1.55028963e+00 -1.50247052e-01
-2.16481030e-01 -1.17298138e+00 -1.21099442e-01 -4.55396533e-01
-1.31234634e+00 8.76497030e-01 8.90914559e-01 3.18985105e-01
1.55184960e+00 -4.29921627e-01 1.02335215e+00 -1.10712700e-01
6.08480096e-01 -1.05911052e+00 1.15378156e-01 4.08385694e-01
1.03560162e+00 -1.51061463e+00 -3.36971357e-02 -3.10335308e-01
-4.83192116e-01 1.27121127e+00 4.85183716e-01 -3.31897974e-01
8.04103792e-01 1.79077178e-01 1.17067963e-01 4.69877213e-01
-6.63362920e-01 -5.51244570e-03 3.89011800e-01 6.29296243e-01
9.18187618e-01 1.07288901e-02 -3.60048056e-01 2.65956491e-01
-2.38909841e-01 5.76369107e-01 5.04084706e-01 7.41576016e-01
-2.79100806e-01 -1.16553867e+00 -3.31853628e-01 2.50225395e-01
-6.45296454e-01 -1.11866817e-01 2.67446011e-01 5.57772458e-01
1.75893798e-01 1.20451927e+00 3.95819806e-02 -6.86370909e-01
3.43519717e-01 -4.03713524e-01 5.97669601e-01 6.61248043e-02
-4.70941782e-01 -1.00716263e-01 -1.21987805e-01 -1.10576642e+00
-8.50054085e-01 -6.01345003e-01 -1.00916016e+00 -7.65211344e-01
3.48290682e-01 -1.66818634e-01 4.12025720e-01 6.17909670e-01
2.79821664e-01 4.72365528e-01 6.58670783e-01 -1.36710060e+00
-8.28326214e-03 -8.68853867e-01 -1.79780751e-01 7.63476133e-01
6.03261769e-01 -6.64002478e-01 -1.01710804e-01 5.74164391e-01]
|
[11.385066986083984, -1.6466529369354248]
|
99e10e4e-807d-44f5-ad47-3648a2ff11a0
|
improving-keyphrase-extraction-with-data
|
2209.04951
| null |
https://arxiv.org/abs/2209.04951v1
|
https://arxiv.org/pdf/2209.04951v1.pdf
|
Improving Keyphrase Extraction with Data Augmentation and Information Filtering
|
Keyphrase extraction is one of the essential tasks for document understanding in NLP. While the majority of the prior works are dedicated to the formal setting, e.g., books, news or web-blogs, informal texts such as video transcripts are less explored. To address this limitation, in this work we present a novel corpus and method for keyphrase extraction from the transcripts of the videos streamed on the Behance platform. More specifically, in this work, a novel data augmentation is proposed to enrich the model with the background knowledge about the keyphrase extraction task from other domains. Extensive experiments on the proposed dataset dataset show the effectiveness of the introduced method.
|
['Thien Huu Nguyen', 'Franck Dernoncourt', 'Nicole Meister', 'Amir Pouran Ben Veyseh']
|
2022-09-11
| null | null | null | null |
['keyphrase-extraction']
|
['natural-language-processing']
|
[ 3.06920052e-01 5.54155111e-02 -5.73316038e-01 6.38496280e-02
-6.58562541e-01 -7.79657006e-01 1.01444042e+00 5.01133800e-01
-4.79367852e-01 9.66394007e-01 7.27115154e-01 -1.02400362e-01
-5.95788099e-02 -4.73349601e-01 -7.23622382e-01 -6.29454732e-01
1.21881954e-01 -2.50460207e-01 1.78678289e-01 -4.09625918e-02
4.58293438e-01 2.08289102e-01 -1.47490323e+00 3.89333218e-01
4.99939173e-01 1.08090699e+00 1.83566093e-01 6.12740934e-01
-4.52134043e-01 1.25344741e+00 -6.05929315e-01 -3.93939227e-01
1.55226737e-01 -3.56431127e-01 -8.72803867e-01 2.95823812e-01
5.53361356e-01 -5.94204247e-01 -8.06624651e-01 1.08575571e+00
2.37280667e-01 -1.21362796e-02 3.84702742e-01 -1.35521054e+00
-1.64562836e-01 1.11835670e+00 -5.62907875e-01 4.15737003e-01
4.77509499e-01 -4.90822703e-01 1.16222060e+00 -1.02562916e+00
8.77156675e-01 9.22874987e-01 7.93140158e-02 2.13387832e-01
-6.70546949e-01 -6.69531047e-01 3.97211313e-01 3.97379428e-01
-1.25626814e+00 -4.11145896e-01 1.06561756e+00 -2.92185813e-01
4.36738282e-01 1.71269938e-01 7.84920096e-01 1.26088810e+00
-1.40787974e-01 1.49034011e+00 1.32697761e+00 -5.94291806e-01
-1.17959946e-01 2.79266149e-01 3.22403252e-01 4.76384878e-01
3.22434962e-01 -4.43644136e-01 -1.05593252e+00 -5.02482988e-02
6.10679090e-01 -1.04266936e-02 -6.61090016e-01 -3.79561856e-02
-1.33666503e+00 5.69810390e-01 -2.74011761e-01 5.50800025e-01
-5.40668070e-01 -1.35207012e-01 6.65809155e-01 2.65219837e-01
7.96561718e-01 3.63829643e-01 -4.18249398e-01 -5.12504220e-01
-1.08309066e+00 5.26786566e-01 1.02652895e+00 1.06870413e+00
6.03393197e-01 -2.56798208e-01 -6.36414662e-02 5.06515741e-01
1.87287200e-03 8.06790441e-02 3.59729141e-01 -4.23474967e-01
1.00583529e+00 6.65173948e-01 2.80688822e-01 -1.44461787e+00
-2.18617648e-01 -2.86785156e-01 -7.55538881e-01 -7.45712221e-01
2.12481081e-01 -3.24516147e-01 -6.61468029e-01 1.23821461e+00
4.16044116e-01 4.63717073e-01 1.17532477e-01 6.98318779e-01
1.17743301e+00 9.90723073e-01 -3.77565585e-02 -6.23302758e-01
1.46671283e+00 -1.02209091e+00 -1.31082857e+00 -5.32222539e-02
2.76798129e-01 -8.98682714e-01 6.88920140e-01 5.79608679e-01
-7.25840449e-01 -3.35519284e-01 -9.95270967e-01 -7.91388899e-02
-4.12552804e-01 3.60259533e-01 6.41048491e-01 1.78779259e-01
-3.49641442e-01 1.09320894e-01 -6.42524421e-01 -5.82347035e-01
4.68411297e-01 -1.60455361e-01 -4.07723963e-01 -2.10924655e-01
-1.28068030e+00 3.32306325e-01 8.62234712e-01 -6.96421713e-02
-7.31395841e-01 -6.15698218e-01 -5.72119296e-01 -1.15280367e-01
1.20030892e+00 -2.17805922e-01 1.37600434e+00 -7.87530363e-01
-1.39054787e+00 6.38807416e-01 -1.42058842e-02 -7.05819547e-01
3.69635403e-01 -8.20296347e-01 -3.04352403e-01 6.67663455e-01
1.06015755e-02 3.10042202e-01 1.05197656e+00 -9.75032508e-01
-9.89070833e-01 -2.71557182e-01 5.21991968e-01 1.84552521e-01
-1.05056953e+00 3.11504900e-01 -8.84769022e-01 -1.17556894e+00
-2.19146907e-01 -8.17427516e-01 7.70984963e-02 -3.93251628e-01
-5.98048270e-01 -2.66929626e-01 1.09297669e+00 -9.11989152e-01
1.80330157e+00 -2.17877316e+00 2.48579904e-01 7.56515637e-02
2.83371866e-01 1.14681363e-01 2.10279241e-01 9.79236305e-01
1.78915501e-01 1.79298416e-01 4.20205183e-02 -1.43864993e-02
-1.11651652e-01 6.35535941e-02 -7.99450278e-01 3.98419350e-01
-4.04956983e-03 6.12241209e-01 -8.59554648e-01 -7.53184497e-01
-6.76404908e-02 3.11277360e-01 -3.36650461e-01 2.69486696e-01
-3.94314766e-01 3.34788918e-01 -8.78127515e-01 6.30937517e-01
3.75982583e-01 -2.81156570e-01 1.67036820e-02 -4.34900701e-01
-3.16924512e-01 4.06280041e-01 -1.38434291e+00 1.74199343e+00
-1.81284487e-01 7.57733643e-01 2.15902194e-01 -9.75201905e-01
4.52308178e-01 6.88200235e-01 7.65391111e-01 -1.42145514e-01
3.00978750e-01 5.22701368e-02 -3.48363370e-01 -5.16927361e-01
8.94886553e-01 1.58337623e-01 2.86001619e-02 5.00751376e-01
5.76211289e-02 -1.91073284e-01 7.11944938e-01 5.54968476e-01
1.00506353e+00 1.49885461e-01 7.73147941e-01 8.15809444e-02
6.35766387e-01 1.84832513e-01 2.04672605e-01 6.26055598e-01
3.84592749e-02 4.13571358e-01 4.65722322e-01 -1.49976805e-01
-8.15892398e-01 -1.97605148e-01 2.39749141e-02 8.52249622e-01
8.92068446e-03 -1.25383532e+00 -8.58251452e-01 -9.24247622e-01
-2.26908401e-01 2.29792818e-01 -3.83902371e-01 2.27398664e-01
-5.62494874e-01 -5.59397280e-01 5.02134323e-01 3.73024493e-01
7.72982836e-01 -9.18386340e-01 -4.40815389e-01 2.49798685e-01
-5.56157589e-01 -1.94901288e+00 -4.61108357e-01 -1.18291907e-01
-6.70836210e-01 -1.15160596e+00 -6.56260371e-01 -7.07766891e-01
5.77057838e-01 3.87288064e-01 9.05555725e-01 -7.42089562e-03
-1.84817016e-02 7.72674561e-01 -9.62345064e-01 -6.37185633e-01
-1.62393034e-01 2.12948814e-01 2.87688822e-02 4.60780084e-01
3.90349925e-01 -4.08966184e-01 -3.71856034e-01 2.01298832e-03
-1.16839719e+00 3.45913529e-01 7.72669852e-01 6.37216389e-01
5.34355879e-01 5.84398389e-01 2.40580171e-01 -1.03801131e+00
8.53702843e-01 -5.18244326e-01 -4.67988998e-01 -1.11584798e-01
-3.81249905e-01 -8.21931735e-02 7.36102521e-01 -7.68088579e-01
-1.12357008e+00 4.56981501e-03 2.55859256e-01 -1.82827398e-01
-2.60892451e-01 1.00214720e+00 -1.69342220e-01 1.38024718e-01
1.51666060e-01 5.79933703e-01 -6.32735908e-01 -6.89808726e-01
4.08824235e-01 8.08363676e-01 5.29795468e-01 -7.40372121e-01
9.75919604e-01 4.76682633e-01 -9.89109874e-02 -1.33902037e+00
-1.16744137e+00 -7.67409563e-01 -5.31558335e-01 -3.34412187e-01
6.36792064e-01 -1.06764770e+00 -4.43970114e-01 3.64918411e-01
-1.14956939e+00 3.09513777e-01 -1.94757298e-01 7.12631226e-01
-3.10281575e-01 5.54096162e-01 -4.44040030e-01 -7.90999413e-01
-4.35287535e-01 -7.54608154e-01 9.64257658e-01 1.82529449e-01
-1.49646118e-01 -7.15588450e-01 1.02145284e-01 4.87651676e-01
-1.14858046e-01 2.26463079e-01 5.47421336e-01 -9.44337666e-01
-6.46591127e-01 -5.21638989e-01 -2.47341633e-01 4.08760339e-01
3.65677029e-01 -7.49538690e-02 -7.93515384e-01 -9.24682319e-02
-1.81777906e-02 -5.45664489e-01 9.84523714e-01 -3.24675292e-02
1.21726298e+00 -8.29672337e-01 -2.50575900e-01 2.03192845e-01
1.18460202e+00 1.26389161e-01 4.54933971e-01 5.17452121e-01
7.91654646e-01 6.06089056e-01 9.81921792e-01 8.00528705e-01
3.05096924e-01 5.12106657e-01 1.92389160e-01 2.63888627e-01
1.12756595e-01 -5.45514166e-01 4.79040861e-01 1.15367901e+00
-2.98930168e-01 -4.83261108e-01 -8.20450246e-01 6.85060441e-01
-1.81997204e+00 -8.92686188e-01 9.41977128e-02 1.52105176e+00
1.06789899e+00 2.26104259e-01 4.96951230e-02 5.27951002e-01
5.15344739e-01 6.94579065e-01 1.40267089e-01 2.24642932e-01
-1.35139003e-01 9.81835648e-02 3.35940093e-01 -6.06672652e-03
-1.43395996e+00 1.03827965e+00 4.82547188e+00 9.95873690e-01
-9.60933089e-01 -2.20482126e-02 1.31276205e-01 1.25864819e-01
9.99202877e-02 1.49888024e-01 -1.03689539e+00 2.79379249e-01
4.98209625e-01 -1.96445256e-01 1.24620937e-01 7.28460908e-01
2.39244357e-01 -2.45182574e-01 -1.06273162e+00 1.19515777e+00
3.01750779e-01 -1.29534745e+00 2.24164620e-01 -5.51675446e-02
7.40299284e-01 -4.38394636e-01 -1.52696505e-01 -3.30759436e-02
-2.25448817e-01 -5.36893070e-01 8.10322642e-01 3.06232393e-01
5.24174690e-01 -7.11210847e-01 7.70594180e-01 5.44562221e-01
-1.16672635e+00 1.77709479e-02 -3.68507355e-02 -1.22099891e-01
1.76934794e-01 7.76391983e-01 -9.67378020e-01 7.18657911e-01
6.09215975e-01 1.02877450e+00 -5.14025867e-01 8.36862028e-01
-4.60250735e-01 1.10239077e+00 -3.82932514e-01 -2.72522837e-01
3.73717040e-01 -5.66425845e-02 7.98563957e-01 1.53548098e+00
6.08335547e-02 4.66677159e-01 3.14706951e-01 3.04530412e-01
-4.97501791e-01 5.45339584e-01 -7.28837013e-01 -1.03960288e+00
2.38824651e-01 1.49714220e+00 -8.12961936e-01 -4.98432308e-01
-7.80932784e-01 7.55422771e-01 -1.27424464e-01 4.95895058e-01
-4.74496871e-01 -4.25070643e-01 2.79265344e-01 2.21927166e-01
3.68445516e-01 -4.41107213e-01 3.43724668e-01 -1.55084360e+00
3.72785389e-01 -1.15635860e+00 4.18744713e-01 -3.99460196e-01
-9.89925981e-01 3.18947583e-01 4.50584292e-01 -1.25848663e+00
-1.28095135e-01 -4.77524340e-01 -2.35023901e-01 2.61908144e-01
-1.68192625e+00 -1.18874264e+00 -3.59892309e-01 5.90237558e-01
9.63529348e-01 -2.77181059e-01 4.30409968e-01 4.31265980e-01
-4.39812511e-01 7.33682588e-02 -4.30982672e-02 4.68413353e-01
8.25605154e-01 -1.03046596e+00 1.98762611e-01 1.06845224e+00
6.60116136e-01 8.53966653e-01 9.27892745e-01 -7.89104044e-01
-1.90762842e+00 -7.56090105e-01 8.09361875e-01 -1.57420635e-01
1.11351204e+00 -4.73186344e-01 -7.66107917e-01 7.01561511e-01
5.80855906e-01 -3.27477641e-02 6.03457808e-01 -2.90666729e-01
-2.41875872e-01 -9.96952429e-02 -5.16697526e-01 6.92505181e-01
7.36711323e-01 -7.75085688e-01 -8.10285270e-01 3.70433956e-01
7.36558974e-01 -5.54494321e-01 -7.10287094e-01 3.95377308e-01
5.49605548e-01 -3.02168965e-01 7.52673626e-01 -6.36996508e-01
6.05682254e-01 -2.77948290e-01 -1.47672355e-01 -1.12121415e+00
5.01268923e-01 -9.84569252e-01 -7.85800099e-01 1.67316985e+00
1.48655279e-02 -3.90288793e-02 7.86919117e-01 3.30614485e-02
2.29068920e-01 -5.73167980e-01 -6.47034347e-01 -4.23134834e-01
-5.67541420e-01 -7.87296057e-01 2.25957796e-01 1.04772925e+00
1.35196075e-01 6.77593827e-01 -8.34952354e-01 2.77798213e-02
4.34171081e-01 1.19064882e-01 1.11712837e+00 -1.09324086e+00
-1.75807610e-01 -4.60077077e-02 -1.90823004e-01 -1.47261810e+00
2.09169999e-01 -5.99061191e-01 -1.79010019e-01 -1.48466563e+00
4.06515598e-01 2.23215356e-01 -2.44990457e-02 3.17811728e-01
-1.90447628e-01 -8.95607099e-02 2.34393105e-01 1.95263505e-01
-8.16815078e-01 5.27109265e-01 1.30281818e+00 -3.72660905e-01
-1.60135090e-01 -8.49206522e-02 -7.58772731e-01 9.34322298e-01
6.19506180e-01 -6.52317405e-01 -5.22178531e-01 -2.90153399e-02
6.34455502e-01 -2.86830068e-02 1.51181206e-01 -6.83076441e-01
3.80440384e-01 -3.71832401e-01 -1.49528742e-01 -8.43977988e-01
1.40969813e-01 -1.07398307e+00 -3.41525286e-01 1.20117161e-02
-3.27911675e-01 -1.12677060e-01 3.43204558e-01 8.01241636e-01
-8.01865637e-01 -1.99570343e-01 5.22266775e-02 -1.53474778e-01
-6.73431456e-01 4.15720999e-01 -4.52273577e-01 3.49854439e-01
6.97167277e-01 -9.67616739e-04 -1.99064091e-01 -6.32171452e-01
-3.04065198e-01 1.45322189e-01 -3.60350544e-03 4.89588022e-01
8.01635742e-01 -1.12467194e+00 -7.43276894e-01 -2.82534182e-01
3.11809659e-01 5.12139238e-02 -3.22350673e-02 9.24352527e-01
-2.74358779e-01 7.59991467e-01 9.70851853e-02 -9.26484242e-02
-1.42616427e+00 5.22256017e-01 -4.14858311e-01 -3.83680493e-01
-7.52262056e-01 4.95292872e-01 2.75772214e-01 2.31258988e-01
4.30401921e-01 -5.49769104e-01 -8.38780999e-01 6.42478168e-01
8.51742089e-01 2.51797885e-01 -7.95445666e-02 -6.80266738e-01
-8.18365663e-02 4.53662068e-01 -4.70112771e-01 -2.00774983e-01
1.48064280e+00 -3.22744638e-01 -2.77103275e-01 5.57922244e-01
1.02649355e+00 4.09411788e-01 -8.10326457e-01 -5.52747488e-01
2.54013419e-01 -4.57107157e-01 1.93628252e-01 -3.17434222e-01
-7.47813344e-01 6.53796434e-01 -3.27481180e-02 2.98656553e-01
1.24609470e+00 -4.58227359e-02 1.06034231e+00 8.10171783e-01
3.16265762e-01 -1.31306422e+00 2.44866386e-01 5.36185622e-01
9.50343609e-01 -1.17701554e+00 4.20427144e-01 -4.82327431e-01
-3.88345212e-01 1.32308376e+00 3.22962552e-01 3.31687450e-01
7.48607934e-01 2.82235801e-01 -2.69564629e-01 -3.14087719e-01
-7.49732375e-01 -1.68284014e-01 4.47941273e-01 1.20045662e-01
4.34247017e-01 -3.93601775e-01 -7.39302397e-01 9.79651988e-01
-6.72485679e-02 2.68942147e-01 7.56617904e-01 1.37218010e+00
-3.19242328e-01 -9.32048082e-01 -3.38530362e-01 3.66364121e-01
-1.25701010e+00 -2.35814855e-01 -8.15459669e-01 8.62618446e-01
-5.91092296e-02 9.06161010e-01 -3.94982278e-01 -1.40522078e-01
2.16146439e-01 2.36902907e-02 3.76413703e-01 -6.41622543e-01
-5.21805227e-01 2.65917748e-01 4.14871424e-01 -3.74916762e-01
-1.09233332e+00 -5.62391281e-01 -1.12700379e+00 7.18634129e-02
-3.40984344e-01 4.70855325e-01 5.30097663e-01 1.16256773e+00
-4.23151739e-02 4.20569748e-01 4.07047361e-01 -5.83243668e-01
-9.49554071e-02 -1.06592619e+00 -5.74826598e-01 3.29739541e-01
5.78892112e-01 -4.81487423e-01 -2.87117422e-01 5.35529017e-01]
|
[12.268649101257324, 8.889451026916504]
|
f96327ac-9714-46d5-af68-cac3bec748c0
|
sdc-uda-volumetric-unsupervised-domain-1
|
2305.11012
| null |
https://arxiv.org/abs/2305.11012v1
|
https://arxiv.org/pdf/2305.11012v1.pdf
|
SDC-UDA: Volumetric Unsupervised Domain Adaptation Framework for Slice-Direction Continuous Cross-Modality Medical Image Segmentation
|
Recent advances in deep learning-based medical image segmentation studies achieve nearly human-level performance in fully supervised manner. However, acquiring pixel-level expert annotations is extremely expensive and laborious in medical imaging fields. Unsupervised domain adaptation (UDA) can alleviate this problem, which makes it possible to use annotated data in one imaging modality to train a network that can successfully perform segmentation on target imaging modality with no labels. In this work, we propose SDC-UDA, a simple yet effective volumetric UDA framework for slice-direction continuous cross-modality medical image segmentation which combines intra- and inter-slice self-attentive image translation, uncertainty-constrained pseudo-label refinement, and volumetric self-training. Our method is distinguished from previous methods on UDA for medical image segmentation in that it can obtain continuous segmentation in the slice direction, thereby ensuring higher accuracy and potential in clinical practice. We validate SDC-UDA with multiple publicly available cross-modality medical image segmentation datasets and achieve state-of-the-art segmentation performance, not to mention the superior slice-direction continuity of prediction compared to previous studies.
|
['Dosik Hwang', 'Taejoon Eo', 'Yohan Jun', 'Sewon Kim', 'Hyeongyu Kim', 'Hyungseob Shin']
|
2023-05-18
|
sdc-uda-volumetric-unsupervised-domain
|
http://openaccess.thecvf.com//content/CVPR2023/html/Shin_SDC-UDA_Volumetric_Unsupervised_Domain_Adaptation_Framework_for_Slice-Direction_Continuous_Cross-Modality_CVPR_2023_paper.html
|
http://openaccess.thecvf.com//content/CVPR2023/papers/Shin_SDC-UDA_Volumetric_Unsupervised_Domain_Adaptation_Framework_for_Slice-Direction_Continuous_Cross-Modality_CVPR_2023_paper.pdf
|
cvpr-2023-1
|
['unsupervised-domain-adaptation', 'pseudo-label']
|
['methodology', 'miscellaneous']
|
[ 6.23228252e-01 4.81586277e-01 -4.75662082e-01 -6.53192282e-01
-1.26239312e+00 -4.79611158e-01 7.45050609e-02 2.63339162e-01
-6.03048801e-01 6.99251235e-01 -1.09118201e-01 -5.35717726e-01
3.43321227e-02 -6.43676817e-01 -5.91819704e-01 -7.15253055e-01
8.13499093e-02 9.29697216e-01 5.78156590e-01 3.09421599e-01
-2.88460851e-01 4.43782240e-01 -6.53823674e-01 3.61641347e-01
1.18084621e+00 9.64928508e-01 4.52449262e-01 5.69598556e-01
-2.21425682e-01 4.81568396e-01 -6.91740662e-02 -1.97922349e-01
1.91474333e-01 -6.03536546e-01 -1.34648597e+00 3.93231422e-01
3.73106480e-01 -3.32589269e-01 1.70565784e-01 9.36235607e-01
5.90711415e-01 -2.59851128e-01 8.72480571e-01 -7.64770508e-01
-5.24852931e-01 6.38435423e-01 -6.24702871e-01 2.38130942e-01
-2.29002401e-01 2.01330781e-01 5.05816281e-01 -3.16838413e-01
8.69676828e-01 4.76939827e-01 9.97658491e-01 7.23916590e-01
-1.50508428e+00 -3.83380324e-01 -2.84170777e-01 -2.72178203e-01
-1.23240256e+00 1.47973627e-01 6.51859105e-01 -7.81892896e-01
6.18261814e-01 2.61739731e-01 6.55107558e-01 6.58635676e-01
2.94270366e-01 9.51318026e-01 1.35802281e+00 -3.04500580e-01
3.07314038e-01 -1.18098602e-01 -9.42034796e-02 8.04315507e-01
-9.00096297e-02 -6.50457442e-02 1.80205688e-01 1.32385477e-01
1.16247368e+00 -1.66736886e-01 -2.78692752e-01 -5.68600476e-01
-1.51745319e+00 7.16651857e-01 8.65928650e-01 5.14955997e-01
-5.97882032e-01 -1.44855291e-01 5.90481877e-01 -1.87507704e-01
5.51065326e-01 5.40281236e-01 -5.38991749e-01 2.26384088e-01
-1.45437133e+00 -2.75694311e-01 2.83094883e-01 7.27469504e-01
2.87741005e-01 -7.27681518e-02 -4.22683120e-01 9.26732481e-01
2.22093537e-01 3.48693907e-01 7.73278296e-01 -1.08069193e+00
1.36391586e-02 5.09418428e-01 -2.84513950e-01 -3.62354517e-01
-8.73975217e-01 -5.13175964e-01 -1.20893693e+00 1.91899568e-01
5.86488724e-01 -1.00911133e-01 -1.48727620e+00 1.44016039e+00
6.02189302e-01 9.98167619e-02 -1.64620042e-01 1.14630663e+00
9.15295243e-01 1.36688709e-01 4.31725115e-01 -4.44887340e-01
1.25727606e+00 -1.05634570e+00 -5.43299675e-01 -2.69247051e-02
8.32057297e-01 -5.55655658e-01 1.15152359e+00 1.65808767e-01
-1.08401680e+00 -3.65882456e-01 -9.25067723e-01 2.90917419e-02
6.96757138e-02 -9.92488191e-02 7.83883035e-01 7.26437569e-01
-1.10317981e+00 4.28970546e-01 -1.18144369e+00 -1.00076966e-01
1.06988358e+00 5.41397154e-01 -3.84888560e-01 -6.70296475e-02
-9.91246164e-01 8.17313075e-01 6.23943627e-01 -2.34911934e-01
-7.28256345e-01 -1.30632758e+00 -8.40482175e-01 -5.94958365e-01
4.89121437e-01 -7.25496948e-01 1.39053988e+00 -8.99304152e-01
-1.41746700e+00 1.39591134e+00 5.23707606e-02 -5.79405308e-01
8.23627055e-01 3.79146487e-01 -2.07208648e-01 4.80954796e-01
5.15444875e-01 1.22237694e+00 5.14704704e-01 -1.37007916e+00
-3.44613671e-01 -4.80378151e-01 -4.97292042e-01 9.22221616e-02
3.09427589e-01 -3.88711780e-01 -6.08982980e-01 -7.09344089e-01
4.36979920e-01 -9.06846225e-01 -5.64930916e-01 2.75285482e-01
-5.39838076e-01 4.84978855e-02 5.98443985e-01 -8.22713375e-01
8.80556881e-01 -1.80966020e+00 4.13566791e-02 2.32240424e-01
4.30850476e-01 2.62483388e-01 2.02282295e-01 -6.44131482e-01
-5.71741723e-02 2.37294018e-01 -9.99597311e-01 -2.45686382e-01
-4.46728796e-01 3.30122888e-01 4.89768237e-01 4.94064003e-01
1.10695243e-01 1.28418660e+00 -9.47662532e-01 -1.17378008e+00
4.77773219e-01 2.73603886e-01 -6.46541655e-01 1.47271812e-01
-3.26614082e-01 1.48692203e+00 -4.43462074e-01 7.85372198e-01
6.97245181e-01 -5.93416035e-01 2.58264422e-01 -2.57396996e-01
2.86787245e-02 -2.43801281e-01 -5.63967407e-01 2.09094357e+00
-5.48707902e-01 2.78359473e-01 6.75728917e-02 -1.06420302e+00
4.58865464e-01 4.58814800e-01 1.16082633e+00 -9.37946022e-01
3.04980636e-01 4.43212807e-01 4.80236970e-02 -6.69430435e-01
-2.51020212e-02 -6.55283689e-01 1.19552948e-01 3.02767158e-01
2.10897118e-01 -6.10540271e-01 5.70664965e-02 -1.48321362e-02
6.41179860e-01 9.67714116e-02 1.22757867e-01 -4.39697236e-01
5.50209880e-01 2.78820187e-01 5.09843230e-01 6.65351391e-01
-5.65635800e-01 1.08897865e+00 2.38209277e-01 -2.19038367e-01
-1.22613358e+00 -1.12168741e+00 -6.84074461e-01 7.49007225e-01
9.18193758e-02 2.61281759e-01 -1.11564755e+00 -1.04886711e+00
-2.92430639e-01 5.59379876e-01 -7.92082071e-01 1.76837146e-01
-4.56732124e-01 -8.60465288e-01 5.08294642e-01 6.80890203e-01
5.72306216e-01 -1.09631813e+00 -5.91862679e-01 4.32248026e-01
-5.10776937e-01 -1.33541811e+00 -5.56724429e-01 2.25968003e-01
-1.21613598e+00 -1.14681542e+00 -1.37521327e+00 -1.00591540e+00
8.16212475e-01 -2.65528262e-01 1.32803810e+00 -9.11656171e-02
-5.42558908e-01 1.82281107e-01 -7.87686035e-02 -1.86702192e-01
-7.13250220e-01 1.50413662e-01 -4.62188452e-01 -3.36546659e-01
2.06963476e-02 -3.30448002e-01 -8.75747025e-01 3.58402014e-01
-1.07201385e+00 4.72427517e-01 6.22429252e-01 9.97483015e-01
1.25137568e+00 -2.02654898e-01 4.55934346e-01 -1.35974658e+00
2.35093206e-01 -2.73338020e-01 -4.39245641e-01 4.23329473e-01
-7.78744161e-01 -1.72279328e-01 1.93501100e-01 -1.60204738e-01
-1.16192043e+00 4.19659108e-01 -4.90465105e-01 -1.51286364e-01
-4.29692179e-01 5.30153811e-01 3.56725991e-01 -1.87311321e-01
7.21523881e-01 3.56300950e-01 2.62307197e-01 -1.32033169e-01
3.27384472e-01 4.13574189e-01 8.53004098e-01 -4.97161508e-01
1.67857096e-01 6.45593166e-01 9.58466977e-02 -4.66572255e-01
-9.83195007e-01 -5.80879271e-01 -1.27239954e+00 -4.36556250e-01
1.51991451e+00 -6.12890840e-01 -3.87328386e-01 5.55530906e-01
-8.05775166e-01 -6.96327150e-01 -4.40846473e-01 4.31289345e-01
-6.90402567e-01 3.79625648e-01 -7.92665303e-01 -1.64431781e-01
-5.57594299e-01 -1.83236706e+00 1.31089211e+00 2.58477300e-01
-1.52323514e-01 -1.37188423e+00 -2.76798848e-02 6.46235764e-01
4.21707153e-01 6.24522269e-01 9.47086394e-01 -4.93228823e-01
-3.62064600e-01 6.78899884e-02 -3.41264635e-01 5.07190347e-01
2.65331954e-01 -4.36042458e-01 -6.06327176e-01 -5.07508069e-02
-9.98492464e-02 -4.31872725e-01 7.07958817e-01 1.14929032e+00
1.52709866e+00 3.80112112e-01 -2.77942896e-01 7.73639798e-01
1.32509840e+00 1.29792750e-01 4.89050061e-01 3.40184689e-01
8.32686782e-01 2.10434794e-01 5.26720822e-01 1.06933832e-01
3.32511306e-01 6.38722658e-01 3.05714130e-01 -8.93580258e-01
-4.75227147e-01 -1.03394436e-02 -6.42158031e-01 6.65759861e-01
4.67449874e-02 2.00851575e-01 -1.21595025e+00 7.04251587e-01
-1.42526591e+00 -2.06346586e-01 -4.44578737e-01 1.83192050e+00
1.32753074e+00 7.54412487e-02 1.12794600e-01 -1.43292010e-01
6.04198754e-01 -3.25407863e-01 -7.59477556e-01 4.59638946e-02
7.90526494e-02 3.48800242e-01 7.59789348e-01 5.24803996e-01
-1.38706648e+00 8.72087717e-01 6.59462309e+00 9.37277615e-01
-1.39920175e+00 7.14519083e-01 1.22151661e+00 2.43642911e-01
-1.17836744e-01 -5.41784585e-01 -2.19963819e-01 3.62532705e-01
5.88139534e-01 4.74527061e-01 -3.14526260e-02 7.95632839e-01
1.29055798e-01 -4.46142077e-01 -9.86613333e-01 7.73977041e-01
-1.48816809e-01 -1.55270505e+00 -2.61847615e-01 -1.42320544e-01
1.17289722e+00 1.10608310e-01 5.93154691e-02 1.96497858e-01
6.04860969e-02 -9.98488903e-01 3.62182081e-01 2.77457327e-01
1.31762111e+00 -4.56653118e-01 8.92230511e-01 1.93297505e-01
-7.00455010e-01 5.89214742e-01 -1.97090991e-02 7.36377835e-01
4.39775139e-01 7.21456885e-01 -1.21996605e+00 6.16898656e-01
6.57612324e-01 5.16625583e-01 -4.58947182e-01 1.07573783e+00
-3.15028578e-02 7.22200513e-01 -1.61086202e-01 6.26193941e-01
4.42621469e-01 -1.60540372e-01 1.85010388e-01 1.19277310e+00
1.49446668e-03 1.99008867e-01 4.03068602e-01 8.80239546e-01
-1.45719960e-01 1.93544805e-01 -1.25771076e-05 3.96586806e-01
-1.25752110e-02 1.13619351e+00 -1.32415271e+00 -5.76313913e-01
-2.03415185e-01 1.11545885e+00 -5.16502261e-02 1.06360771e-01
-1.00355601e+00 3.38656276e-01 -1.74751073e-01 3.09251219e-01
1.03309937e-01 9.60465427e-03 -8.86324346e-01 -7.51760304e-01
-3.43915075e-01 -5.33018291e-01 5.33273280e-01 -5.65408170e-01
-1.26329434e+00 6.70580566e-01 1.24741336e-02 -1.12401891e+00
-6.47504553e-02 -4.55267310e-01 -1.49139807e-01 7.36532986e-01
-1.60157692e+00 -1.35818911e+00 -3.23068053e-01 6.14162326e-01
4.56136584e-01 5.80557734e-02 7.75063276e-01 4.67129946e-01
-2.80730277e-01 5.61499357e-01 1.76180959e-01 2.76195318e-01
7.10314035e-01 -1.42920816e+00 -7.71014858e-03 5.00793397e-01
-2.67463893e-01 1.57911688e-01 3.00516039e-01 -8.28542233e-01
-7.09796965e-01 -1.26212275e+00 2.90000081e-01 -3.37749034e-01
5.18911958e-01 1.24893226e-01 -9.65616703e-01 6.25307620e-01
5.67038730e-02 5.26729107e-01 8.23814988e-01 -1.55790985e-01
9.89667028e-02 1.70812786e-01 -1.72277248e+00 3.12897950e-01
6.70064688e-01 -2.31856957e-01 -4.65844959e-01 6.23824775e-01
7.49402285e-01 -1.01137459e+00 -1.40720141e+00 7.15744019e-01
2.21574664e-01 -8.84941995e-01 1.06326532e+00 -3.07847410e-01
6.17700219e-01 -1.43826813e-01 2.94572145e-01 -1.02396560e+00
-1.49524316e-01 -1.69324409e-02 2.82847285e-01 8.77154469e-01
4.65918064e-01 -4.68711704e-01 9.68131185e-01 7.94421732e-01
-5.85613549e-01 -8.62375021e-01 -1.13697600e+00 -4.98861611e-01
5.38671315e-01 -5.18154740e-01 2.64585167e-01 1.13009095e+00
-2.86473870e-01 -2.26982474e-01 3.26022841e-02 6.22802787e-02
8.46337557e-01 6.94719404e-02 7.60492384e-02 -1.02969134e+00
-2.71366358e-01 -6.38715804e-01 -1.13349192e-01 -7.53771365e-01
5.10732122e-02 -1.33533049e+00 2.90646642e-01 -1.74615312e+00
4.17240471e-01 -7.61712790e-01 -3.53454053e-01 5.27082205e-01
-1.37491316e-01 8.27430248e-01 -1.45433038e-01 4.02435631e-01
-7.39119530e-01 1.92184314e-01 2.13171053e+00 -3.78428668e-01
-1.20039918e-01 7.98041001e-02 -3.47812414e-01 6.33086085e-01
7.17811644e-01 -4.97239858e-01 -3.70092779e-01 -3.89748037e-01
-5.86353540e-01 4.12736863e-01 3.50813478e-01 -8.81444454e-01
1.98147483e-02 -4.21641916e-02 6.14294648e-01 -6.27853096e-01
-1.79667637e-01 -8.29807997e-01 -7.12521821e-02 6.94427907e-01
-4.55231756e-01 -5.79375148e-01 2.75524884e-01 2.38925025e-01
-2.01718703e-01 -2.08628565e-01 1.24352884e+00 -4.81029809e-01
-7.16525435e-01 4.93951052e-01 -3.59565973e-01 2.87902772e-01
1.15616894e+00 -2.52340674e-01 2.43304491e-01 3.26588675e-02
-1.24560320e+00 2.34471008e-01 2.38135219e-01 1.19747492e-02
3.66050571e-01 -1.12489986e+00 -7.20309079e-01 -5.56825027e-02
5.21660261e-02 5.60121953e-01 8.29087138e-01 1.43287694e+00
-8.06578517e-01 5.95437944e-01 -2.39613727e-01 -1.32636535e+00
-9.37431335e-01 4.01186317e-01 6.31900728e-01 -5.73539793e-01
-7.67704070e-01 1.02803588e+00 2.36652032e-01 -7.95054257e-01
-1.43968254e-01 -4.99184966e-01 1.11786118e-02 -3.71294200e-01
1.64085254e-01 -2.08738610e-01 3.13242495e-01 -6.69489026e-01
-4.04643923e-01 5.78444600e-01 -1.62998587e-01 5.42846322e-02
1.03533363e+00 -1.09842390e-01 -4.18060943e-02 2.68131673e-01
1.16630733e+00 -4.86885160e-01 -1.46373701e+00 -4.06731457e-01
-1.15170978e-01 -1.34708032e-01 4.73000675e-01 -1.38470638e+00
-1.41620004e+00 7.69405782e-01 1.04864156e+00 -1.59249529e-01
1.20995200e+00 2.86346495e-01 9.91281331e-01 -4.28389907e-01
4.40161198e-01 -9.62340415e-01 -1.44936755e-01 9.04088169e-02
6.27322078e-01 -1.81900203e+00 -2.36136224e-02 -6.94107533e-01
-9.54174697e-01 8.60483646e-01 5.88560522e-01 1.68222308e-01
5.91181397e-01 2.98056006e-01 3.70216429e-01 -1.97470590e-01
1.08499378e-01 -1.46451458e-01 6.33769333e-01 8.40069413e-01
7.36860752e-01 4.72979724e-01 -3.41858357e-01 3.42904985e-01
1.61932319e-01 2.63615221e-01 7.31991976e-02 8.20645630e-01
-8.71282741e-02 -1.03932083e+00 -2.81469524e-01 4.44250941e-01
-6.73205733e-01 -1.09015353e-01 1.33277655e-01 9.34401453e-01
2.74457991e-01 4.75913942e-01 -4.67316881e-02 1.62714809e-01
1.15668356e-01 -1.05880804e-01 5.99537492e-01 -6.35754287e-01
-5.20629644e-01 4.08556849e-01 -3.82038653e-01 -3.90346438e-01
-6.13438547e-01 -6.62716806e-01 -1.84556592e+00 1.73250958e-01
-1.68827120e-02 -1.44401953e-01 7.59362936e-01 1.14546847e+00
7.68689066e-02 1.00825322e+00 3.26579630e-01 -7.77077377e-01
-1.18024396e-02 -8.58331084e-01 -4.75110859e-01 6.23724222e-01
1.88135371e-01 -5.33022463e-01 2.25267738e-01 3.77493382e-01]
|
[14.613288879394531, -2.2020652294158936]
|
52ed3228-9558-4cbf-a02a-04335c4715b4
|
doc-deep-occlusion-estimation-from-a-single
|
1511.06457
| null |
http://arxiv.org/abs/1511.06457v4
|
http://arxiv.org/pdf/1511.06457v4.pdf
|
DOC: Deep OCclusion Estimation From a Single Image
|
Recovering the occlusion relationships between objects is a fundamental human
visual ability which yields important information about the 3D world. In this
paper we propose a deep network architecture, called DOC, which acts on a
single image, detects object boundaries and estimates the border ownership
(i.e. which side of the boundary is foreground and which is background). We
represent occlusion relations by a binary edge map, to indicate the object
boundary, and an occlusion orientation variable which is tangential to the
boundary and whose direction specifies border ownership by a left-hand rule. We
train two related deep convolutional neural networks, called DOC, which exploit
local and non-local image cues to estimate this representation and hence
recover occlusion relations. In order to train and test DOC we construct a
large-scale instance occlusion boundary dataset using PASCAL VOC images, which
we call the PASCAL instance occlusion dataset (PIOD). This contains 10,000
images and hence is two orders of magnitude larger than existing occlusion
datasets for outdoor images. We test two variants of DOC on PIOD and on the
BSDS occlusion dataset and show they outperform state-of-the-art methods.
Finally, we perform numerous experiments investigating multiple settings of DOC
and transfer between BSDS and PIOD, which provides more insights for further
study of occlusion estimation.
|
['Peng Wang', 'Alan Yuille']
|
2015-11-20
| null | null | null | null |
['occlusion-estimation']
|
['computer-vision']
|
[ 9.94426459e-02 3.99392880e-02 -3.05769920e-01 -5.08682370e-01
-2.35256284e-01 -5.48602998e-01 4.69289392e-01 -1.18938848e-01
-1.37404919e-01 5.26430428e-01 9.61082354e-02 -3.24317515e-01
5.26489504e-02 -9.42584991e-01 -1.23568165e+00 -4.33636427e-01
-4.43501584e-02 7.37438977e-01 4.62679416e-01 2.60945529e-01
-4.16354463e-02 8.51311028e-01 -1.63311911e+00 5.19534171e-01
6.07938111e-01 1.44694829e+00 -3.69771980e-02 7.66809881e-01
3.80729623e-02 7.08511710e-01 -7.61149108e-01 -2.87004888e-01
5.71476996e-01 1.40881091e-01 -8.02338362e-01 3.92926276e-01
1.24555910e+00 -5.85975409e-01 -5.75650275e-01 6.04379714e-01
2.61920571e-01 -1.94879230e-02 6.99062109e-01 -1.36225986e+00
-6.67101324e-01 -3.05143949e-02 -6.85349226e-01 4.56230164e-01
1.68910131e-01 1.61990091e-01 9.07324374e-01 -6.63512468e-01
1.14740527e+00 1.47185993e+00 7.10413277e-01 -4.25273366e-02
-1.33097124e+00 -3.72224182e-01 4.07121092e-01 1.79942995e-01
-1.10653269e+00 -3.25006992e-01 7.22819388e-01 -6.41434252e-01
9.50504959e-01 3.21865946e-01 5.68092585e-01 1.01934159e+00
2.88814366e-01 1.08091128e+00 1.01342535e+00 -2.97224969e-01
3.31086926e-02 -3.58681493e-02 2.91269630e-01 5.20305932e-01
3.62473726e-01 1.03404246e-01 -3.52783144e-01 2.39807755e-01
7.52537310e-01 -1.51574418e-01 -3.08597147e-01 -1.03905046e+00
-1.00626063e+00 6.07005000e-01 9.04399812e-01 -1.96256995e-01
-1.60651937e-01 3.14801455e-01 4.02341992e-01 1.40400410e-01
4.25190270e-01 1.90129876e-01 -4.48991120e-01 2.99012929e-01
-4.05154020e-01 4.95492667e-01 1.00621080e+00 1.06153452e+00
9.23314035e-01 -3.76829654e-01 -2.84584880e-01 6.81669533e-01
2.58216739e-01 3.56709033e-01 -2.54149407e-01 -1.32512867e+00
6.36196733e-01 5.90573370e-01 3.55240822e-01 -1.03809619e+00
-4.42028821e-01 -4.90178108e-01 -5.01375258e-01 4.08832818e-01
6.51022255e-01 2.02982157e-01 -1.18622804e+00 1.68697846e+00
6.86617613e-01 6.91856220e-02 -1.85739264e-01 1.04779243e+00
1.16662741e+00 3.78003329e-01 -2.63732765e-02 4.21755105e-01
1.62606573e+00 -1.09520710e+00 -4.83115166e-01 -4.88402337e-01
4.68036801e-01 -8.00286770e-01 7.46722519e-01 2.94815779e-01
-9.16633964e-01 -7.60539353e-01 -1.08205497e+00 -6.67173207e-01
-6.27610683e-01 3.77468556e-01 8.62251043e-01 3.97637695e-01
-8.33019078e-01 3.31766516e-01 -8.23547602e-01 -2.81382084e-01
8.85866165e-01 2.89712191e-01 -4.41635489e-01 -3.10903013e-01
-9.71947730e-01 7.86968172e-01 2.72278726e-01 3.01217347e-01
-1.07865024e+00 -7.42368817e-01 -1.29996049e+00 4.62892503e-02
3.86593103e-01 -6.58946157e-01 8.61337364e-01 -7.43231297e-01
-8.78985941e-01 1.37050509e+00 -2.19671309e-01 -5.78900754e-01
8.47913980e-01 -4.16704088e-01 -1.70547813e-01 1.91775486e-01
3.22582543e-01 1.22392654e+00 7.39683867e-01 -1.53232121e+00
-7.11998880e-01 -3.01968902e-01 5.35801828e-01 2.92140897e-03
2.70381331e-01 -1.19270444e-01 -8.60935867e-01 -4.38471824e-01
2.82062829e-01 -8.34528267e-01 2.24952996e-01 5.13933837e-01
-6.94182754e-01 -3.80214512e-01 1.24220788e+00 -6.90906942e-01
8.34851146e-01 -2.05932236e+00 4.48104553e-02 3.08127119e-03
4.88940567e-01 3.00147533e-01 -2.15820581e-01 1.67245105e-01
-1.35396048e-01 -7.06015080e-02 -1.80230618e-01 -4.87149268e-01
3.27306598e-01 5.08904338e-01 -5.52665591e-01 6.73098385e-01
4.58220273e-01 1.02188909e+00 -5.84669888e-01 -3.76933962e-01
5.43389082e-01 6.43788874e-01 -5.91032505e-01 3.49575251e-01
-4.21964616e-01 3.63807678e-01 -2.29627788e-01 7.64398336e-01
1.16431153e+00 -1.82586998e-01 -2.42009424e-02 -5.35334408e-01
-1.84534162e-01 4.09279585e-01 -1.19839585e+00 1.54480267e+00
-4.59795952e-01 1.29021037e+00 -4.26044501e-02 -9.88842309e-01
8.95489872e-01 -3.09642758e-02 2.50468165e-01 -6.64040208e-01
3.09932649e-01 -8.41726735e-02 -2.10709050e-01 -6.97221935e-01
2.82193065e-01 7.14208066e-01 9.83600318e-02 2.60420412e-01
-2.10388497e-01 -1.04364052e-01 5.43150425e-01 1.36362106e-01
1.13567388e+00 5.78230321e-01 -4.57393825e-02 -3.92907828e-01
4.36921000e-01 -2.10051790e-01 7.60938108e-01 9.01285052e-01
-5.32454669e-01 7.58794785e-01 1.06775999e+00 -1.07799637e+00
-9.40408349e-01 -1.32859623e+00 -6.62751198e-01 7.53147006e-01
5.71095049e-01 -1.28380001e-01 -6.52133822e-01 -6.58925712e-01
6.19047761e-01 3.76017362e-01 -9.19251680e-01 2.44330943e-01
-9.59847271e-01 -2.55231202e-01 5.05607389e-02 8.15725267e-01
8.85712326e-01 -1.24721146e+00 -6.40812814e-01 -6.41054064e-02
-2.80005336e-01 -1.59282911e+00 -2.91744292e-01 2.33373061e-01
-3.82302135e-01 -1.41318595e+00 -3.19051385e-01 -9.47951913e-01
6.31078541e-01 3.17936569e-01 1.61885047e+00 -5.99492937e-02
-5.24578750e-01 1.76371500e-01 1.67225614e-01 -4.65024889e-01
1.21026775e-02 1.84902370e-01 -3.17913055e-01 -4.64122258e-02
5.75054586e-01 -5.35248041e-01 -7.30978608e-01 5.87470710e-01
-8.73840094e-01 1.48491904e-01 4.03114945e-01 5.25596619e-01
8.01686466e-01 -1.37389332e-01 7.27584511e-02 -9.12024736e-01
-4.76447120e-02 -3.18581432e-01 -9.87325609e-01 2.29294568e-01
3.51180404e-01 -7.69840032e-02 2.66956210e-01 -1.53645441e-01
-9.67321217e-01 1.36201248e-01 7.10350648e-02 -5.20750761e-01
-4.47154015e-01 -1.00691967e-01 -4.85484183e-01 -7.49600008e-02
4.53985810e-01 -5.13994396e-01 -5.68782985e-01 -4.84222531e-01
4.54106152e-01 5.74040115e-01 8.84739876e-01 -6.99330986e-01
5.93080103e-01 9.54731226e-01 -4.39205989e-02 -6.94582045e-01
-1.40478945e+00 -4.51042414e-01 -1.06223285e+00 -1.48546383e-01
1.18634140e+00 -9.92118955e-01 -9.26845729e-01 4.80251729e-01
-1.44354856e+00 -6.37285531e-01 -2.34541520e-01 2.45652512e-01
-5.73006332e-01 -1.37767181e-01 -6.75900996e-01 -4.03694302e-01
1.75207436e-01 -1.34872460e+00 1.60366189e+00 1.44788072e-01
1.87580157e-02 -8.96267474e-01 -1.55422643e-01 5.46940684e-01
1.37425676e-01 8.16396713e-01 7.42254853e-01 8.06451663e-02
-1.24760342e+00 -1.81814164e-01 -8.73991609e-01 2.10980073e-01
-9.59145948e-02 9.10412744e-02 -1.22136927e+00 -2.37681717e-01
-1.72265247e-01 -4.09512222e-01 1.18251133e+00 6.32468164e-01
1.44987249e+00 1.71267167e-01 -7.21313715e-01 9.57912862e-01
1.25012016e+00 -7.97607005e-02 8.55723321e-01 4.80526060e-01
8.75320315e-01 6.12414181e-01 5.00408351e-01 1.37870640e-01
5.73210835e-01 8.84456515e-01 6.99774265e-01 -4.40667242e-01
-6.74410403e-01 -8.94969106e-02 -1.57826126e-01 3.32106054e-01
2.41107434e-01 -5.65052688e-01 -8.35267544e-01 5.74217737e-01
-1.72237730e+00 -4.34161603e-01 -1.65605783e-01 1.92527390e+00
6.28273249e-01 3.98381859e-01 -2.84778297e-01 -3.04100942e-02
7.01910615e-01 4.17495400e-01 -6.63346827e-01 -4.27515984e-01
-2.21749619e-01 5.05389832e-02 4.89146888e-01 6.42497480e-01
-1.78864658e+00 1.00605750e+00 5.84679317e+00 4.78934556e-01
-9.48224485e-01 -1.38074988e-02 8.50433350e-01 1.59916401e-01
2.30095275e-02 -1.38247445e-01 -1.04958832e+00 2.01055050e-01
2.71450698e-01 6.26249313e-01 2.26990536e-01 9.24118638e-01
-1.72769025e-01 -3.84075761e-01 -1.36366105e+00 7.70261705e-01
6.78127557e-02 -1.12483752e+00 -2.38804311e-01 2.33993113e-01
9.77212548e-01 1.94614410e-01 -6.67093694e-02 1.30268112e-01
5.22340000e-01 -8.27656031e-01 8.10133994e-01 3.32236707e-01
6.95200384e-01 -3.20388615e-01 7.09097862e-01 -5.89177199e-02
-1.46279573e+00 6.96646795e-02 -5.89971781e-01 -7.88388923e-02
9.91286710e-02 6.20348394e-01 -4.62809473e-01 2.07776010e-01
1.04910898e+00 9.33919251e-01 -5.66002250e-01 1.11014235e+00
-7.04139054e-01 1.50398746e-01 -4.56356972e-01 5.57095766e-01
1.91517875e-01 -9.53226388e-02 5.24169624e-01 1.03016675e+00
-3.19506556e-01 -1.91697627e-01 2.60507107e-01 1.12107623e+00
-4.37960982e-01 -4.98574138e-01 -5.64901829e-01 5.18687844e-01
4.13616955e-01 1.21293056e+00 -8.52930307e-01 -5.63588321e-01
-3.20923269e-01 7.74303019e-01 7.32968748e-01 6.84079111e-01
-1.15308559e+00 -4.26497370e-01 9.50508475e-01 1.10787787e-01
7.92822003e-01 -2.69022733e-01 -2.21089423e-01 -1.03932011e+00
3.14997524e-01 -6.55367613e-01 2.33080462e-01 -9.30731654e-01
-1.10911465e+00 5.00557065e-01 4.13343543e-03 -1.03051662e+00
2.80237198e-01 -9.34329450e-01 -5.11144876e-01 7.04484105e-01
-1.76802063e+00 -1.10875762e+00 -9.11409557e-01 9.82535705e-02
4.59016532e-01 4.35448229e-01 4.33370143e-01 4.76244003e-01
-8.31862509e-01 3.64553958e-01 -6.41717464e-02 3.65769535e-01
6.09347045e-01 -1.22562432e+00 7.84188092e-01 7.56495059e-01
1.75801739e-01 2.44881541e-01 4.54297960e-01 -5.37000775e-01
-1.16057217e+00 -1.35610652e+00 9.62949514e-01 -8.31262231e-01
3.22858453e-01 -7.89448202e-01 -8.09369564e-01 1.04487288e+00
6.12612872e-04 5.72298884e-01 8.43240842e-02 -9.24079716e-02
-4.73533630e-01 -2.68142700e-01 -9.87744391e-01 4.41953570e-01
1.52032161e+00 -4.27219391e-01 -5.79883397e-01 6.45266950e-01
8.78661454e-01 -9.42148507e-01 -8.11772883e-01 5.66263378e-01
6.62336349e-01 -1.34116077e+00 1.29286766e+00 -4.56729054e-01
6.04713619e-01 -4.13179219e-01 -2.40233168e-01 -7.76256740e-01
-1.20683730e-01 -1.42331943e-01 -2.31118932e-01 1.12124240e+00
-1.26880810e-01 -7.24560559e-01 8.97056460e-01 6.03207648e-01
-3.83487344e-01 -9.45141315e-01 -9.05083895e-01 -7.32891142e-01
-1.93767086e-01 -4.82653052e-01 7.92419791e-01 6.91059411e-01
-8.79366696e-01 -6.32550940e-02 -1.28598101e-02 2.49568880e-01
5.74970782e-01 4.94585842e-01 1.21386528e+00 -1.15580297e+00
3.13586555e-02 -2.56171227e-01 -6.94568098e-01 -1.59143722e+00
3.69856983e-01 -6.03910089e-01 3.77105479e-03 -1.54660082e+00
-1.94895476e-01 -7.59830475e-01 -1.57042280e-01 3.90623301e-01
2.40167137e-02 8.02776754e-01 -6.75144196e-02 1.29696444e-01
-7.46178210e-01 4.35699880e-01 1.33558345e+00 -4.35924262e-01
-1.04702204e-01 -1.48020521e-01 -1.97025299e-01 9.09120739e-01
5.01028538e-01 -4.11709487e-01 -1.14507064e-01 -9.04960990e-01
-9.70116109e-02 -2.30780005e-01 5.87117612e-01 -1.01337755e+00
-2.08127305e-01 1.66470379e-01 7.96473444e-01 -1.02929032e+00
5.29778361e-01 -7.87660122e-01 -1.81168318e-02 3.23184967e-01
-3.11799377e-01 -1.97424337e-01 2.93800503e-01 4.54913735e-01
-9.39794183e-02 5.75695485e-02 5.81080973e-01 -2.31413282e-02
-8.10673118e-01 5.03441870e-01 1.61760718e-01 2.32764184e-01
1.09802198e+00 -2.36697897e-01 -6.56265259e-01 8.98721104e-04
-4.48283762e-01 4.20050532e-01 5.10159433e-01 6.66950166e-01
3.09526235e-01 -1.44665778e+00 -4.56543654e-01 5.47763884e-01
1.47236332e-01 3.18730682e-01 1.50260152e-02 6.75901353e-01
-1.09891129e+00 5.92359066e-01 -1.90729961e-01 -1.03855288e+00
-1.01840913e+00 6.08833790e-01 5.26968777e-01 -1.33900911e-01
-9.74683762e-01 9.80125427e-01 8.44952583e-01 -3.74167472e-01
4.77627337e-01 -6.38568223e-01 -1.60108268e-01 -9.14192051e-02
3.42862457e-01 1.64269045e-01 1.71852186e-02 -7.38057077e-01
-4.20116007e-01 8.25231254e-01 1.17943492e-02 3.86015385e-01
9.97053742e-01 -1.03901334e-01 -2.89770782e-01 1.52988642e-01
1.45341825e+00 -2.02499807e-01 -1.80364764e+00 -3.59673023e-01
-1.90669373e-01 -9.89969432e-01 1.07448660e-02 -5.73138654e-01
-1.34301949e+00 1.00603497e+00 6.20724142e-01 1.85270429e-01
8.96126807e-01 1.18306540e-01 7.47168064e-01 2.98828065e-01
1.29867643e-01 -9.02327657e-01 -9.13492888e-02 3.40512276e-01
1.02685285e+00 -1.51328313e+00 2.32098415e-01 -8.43770862e-01
-1.44765407e-01 9.26594794e-01 8.74234617e-01 -3.73444319e-01
5.84675014e-01 2.08698660e-01 1.85538426e-01 -3.94561052e-01
-5.15226841e-01 -3.58390868e-01 5.79568028e-01 6.67471468e-01
1.80049717e-01 1.31615922e-01 6.61875233e-02 8.11705831e-03
-2.74676323e-01 -1.99825078e-01 8.05982575e-02 7.99530089e-01
-2.96318650e-01 -8.62907648e-01 -4.14321125e-01 3.61580431e-01
-1.54859394e-01 2.73472518e-01 -3.31978977e-01 1.07829320e+00
6.76960647e-01 7.55685806e-01 6.34118676e-01 8.72155577e-02
3.33738416e-01 -4.56952840e-01 4.56828088e-01 -2.20771655e-01
-1.94269314e-01 -2.14172423e-01 2.53540903e-01 -8.86298180e-01
-4.35487121e-01 -5.54166317e-01 -9.03323233e-01 -1.95881829e-01
-3.78703624e-01 -2.53993213e-01 6.04999542e-01 7.85818994e-01
3.86271089e-01 8.32525074e-01 3.20301324e-01 -1.26513374e+00
1.31365433e-01 -7.98572958e-01 -2.31388003e-01 7.21507192e-01
3.38207096e-01 -1.09365606e+00 -3.35114568e-01 7.93507844e-02]
|
[9.411972999572754, 0.22151699662208557]
|
e4ac1d0f-3f0d-4f20-936c-a29aaf152e52
|
beyond-prompting-making-pre-trained-language
|
2210.16637
| null |
https://arxiv.org/abs/2210.16637v2
|
https://arxiv.org/pdf/2210.16637v2.pdf
|
Beyond Prompting: Making Pre-trained Language Models Better Zero-shot Learners by Clustering Representations
|
Recent work has demonstrated that pre-trained language models (PLMs) are zero-shot learners. However, most existing zero-shot methods involve heavy human engineering or complicated self-training pipelines, hindering their application to new situations. In this work, we show that zero-shot text classification can be improved simply by clustering texts in the embedding spaces of PLMs. Specifically, we fit the unlabeled texts with a Bayesian Gaussian Mixture Model after initializing cluster positions and shapes using class names. Despite its simplicity, this approach achieves superior or comparable performance on both topic and sentiment classification datasets and outperforms prior works significantly on unbalanced datasets. We further explore the applicability of our clustering approach by evaluating it on 14 datasets with more diverse topics, text lengths, and numbers of classes. Our approach achieves an average of 20% absolute improvement over prompt-based zero-shot learning. Finally, we compare different PLM embedding spaces and find that texts are well-clustered by topics even if the PLM is not explicitly pre-trained to generate meaningful sentence embeddings. This work indicates that PLM embeddings can categorize texts without task-specific fine-tuning, thus providing a new way to analyze and utilize their knowledge and zero-shot learning ability.
|
['Mrinmaya Sachan', 'Roger Wattenhofer', 'Zhao Meng', 'Ping Nie', 'Yu Fei']
|
2022-10-29
| null | null | null | null |
['sentence-embeddings', 'sentence-embeddings']
|
['methodology', 'natural-language-processing']
|
[-3.51619460e-02 3.15141119e-02 -3.69888186e-01 -4.32672679e-01
-9.01207328e-01 -3.89713079e-01 9.38532770e-01 4.85280305e-01
-5.58716297e-01 2.75085241e-01 3.86328369e-01 -3.05681467e-01
3.97897214e-02 -8.19194674e-01 -2.93538660e-01 -5.77135623e-01
3.29984576e-01 6.37361467e-01 2.84025103e-01 -2.56802946e-01
4.94353831e-01 -1.29916996e-01 -1.74690521e+00 1.72101021e-01
9.02316570e-01 4.76198494e-01 1.59371644e-01 7.23624766e-01
-7.70851731e-01 3.16901594e-01 -7.54708648e-01 -4.40314472e-01
-2.73041464e-02 -2.32848182e-01 -4.96851385e-01 2.50162005e-01
2.97836125e-01 -8.62238780e-02 -1.19160704e-01 7.93210089e-01
5.27624846e-01 6.65080309e-01 1.16931057e+00 -1.22756708e+00
-8.43941867e-01 8.54948580e-01 -6.26097083e-01 5.06676501e-03
4.27099429e-02 -1.33489547e-02 1.16116858e+00 -1.19351518e+00
4.31690335e-01 1.34034359e+00 7.08643675e-01 5.89485228e-01
-1.28893185e+00 -5.92687070e-01 1.59714326e-01 2.04246268e-01
-1.26103675e+00 -2.79235154e-01 5.73698521e-01 -7.43865252e-01
1.09657633e+00 -1.92900136e-01 1.98498547e-01 1.30281985e+00
4.66904268e-02 8.10658991e-01 5.11599302e-01 -8.84977996e-01
6.17019832e-01 5.87671816e-01 7.99490750e-01 2.65372038e-01
3.83156329e-01 -6.10404372e-01 -4.46813911e-01 -2.85081863e-01
4.19852026e-02 3.75978649e-01 -3.72909717e-02 -6.84515297e-01
-1.02045751e+00 1.42528737e+00 -1.02791160e-01 5.70640326e-01
1.06630642e-02 -1.72514230e-01 5.92345595e-01 -1.95285231e-02
7.96379387e-01 6.92154169e-01 -4.63783026e-01 -2.73069292e-01
-1.14869225e+00 -1.00633033e-01 8.30551028e-01 9.15216684e-01
8.52563858e-01 -2.94682961e-02 -3.38406056e-01 1.14898753e+00
2.50043154e-01 9.76048559e-02 1.14794052e+00 -5.43494821e-01
2.95902818e-01 5.87144315e-01 1.61086377e-02 -7.79115379e-01
-2.79651284e-01 -7.05384761e-02 -4.92886722e-01 -4.17500697e-02
3.31550270e-01 -3.16192985e-01 -9.97449815e-01 1.42132080e+00
1.67901024e-01 3.46465647e-01 1.78105414e-01 3.73484045e-01
5.86977124e-01 9.35109973e-01 3.49995673e-01 -9.70377326e-02
1.48701477e+00 -1.10859799e+00 -8.69427145e-01 -1.93133384e-01
9.56352592e-01 -7.02899694e-01 1.60256290e+00 3.75821233e-01
-5.93507469e-01 -5.10952055e-01 -1.22965336e+00 -8.02557245e-02
-8.07564259e-01 -1.29521489e-01 4.35687214e-01 1.02743530e+00
-6.29977286e-01 6.96015179e-01 -7.72289932e-01 -7.32399702e-01
4.76737201e-01 3.64168361e-02 -5.15414737e-02 -1.46498680e-01
-1.23558950e+00 7.95892596e-01 4.72140282e-01 -8.54234397e-01
-5.37918568e-01 -9.60832298e-01 -1.03946829e+00 3.51594746e-01
2.64361531e-01 -3.33409578e-01 1.22884250e+00 -2.70621151e-01
-1.44675541e+00 6.31258667e-01 -2.71354973e-01 -3.64588797e-01
7.69090187e-03 -1.65079862e-01 -2.80554384e-01 1.12527370e-01
1.22324936e-01 7.47003853e-01 1.00979161e+00 -1.07251918e+00
-5.58775604e-01 -1.95317879e-01 -2.14502484e-01 1.01359308e-01
-1.30657983e+00 -7.23374560e-02 -3.75843316e-01 -6.80632889e-01
-1.23288341e-01 -6.62784040e-01 -2.24208251e-01 -1.47299036e-01
-3.30731767e-04 -7.50310898e-01 9.91343856e-01 -1.29360721e-01
1.41580665e+00 -2.28879428e+00 -1.90128729e-01 -1.22793123e-01
1.99581683e-01 3.88503283e-01 -1.32215247e-01 5.17244637e-01
1.37880936e-01 4.51506555e-01 -3.16099562e-02 -6.71029925e-01
4.60351944e-01 2.78996527e-01 -3.41390967e-01 2.41744339e-01
4.60621752e-02 7.75563419e-01 -1.05512559e+00 -7.15885997e-01
6.10163748e-01 5.44525385e-01 -6.47138774e-01 4.01257500e-02
-2.10269973e-01 -2.41343349e-01 -5.59652150e-02 2.43502989e-01
3.83878559e-01 -4.52840954e-01 1.77502930e-01 3.52338068e-02
1.94588929e-01 2.05771893e-01 -1.18756604e+00 1.64484680e+00
-6.34677172e-01 8.91516507e-01 -5.01553357e-01 -1.22560966e+00
1.02044153e+00 4.12982523e-01 4.00353849e-01 -1.55237958e-01
3.79508615e-01 -1.58022419e-01 -2.02697754e-01 -4.20921952e-01
7.72457361e-01 -4.05228972e-01 -2.49014691e-01 9.46584523e-01
6.20135784e-01 -1.85966149e-01 5.19803822e-01 4.35605913e-01
7.88522899e-01 -3.62269402e-01 3.59316587e-01 -3.02182823e-01
6.46226034e-02 -9.34792310e-02 1.30181193e-01 8.90220642e-01
-2.87996590e-01 6.37461424e-01 3.81381035e-01 -1.76597521e-01
-1.10649109e+00 -1.03442287e+00 -3.98848742e-01 1.82857203e+00
-2.80391648e-02 -7.68910110e-01 -5.90764582e-01 -7.13348031e-01
-1.40441284e-02 1.37164056e+00 -8.59300792e-01 -2.80367702e-01
1.53283747e-02 -9.70895112e-01 2.70739555e-01 5.55682898e-01
-3.16942036e-01 -7.58983552e-01 -4.02043670e-01 2.89577901e-01
-7.58092552e-02 -8.04159820e-01 -5.26996970e-01 4.99082386e-01
-7.66420424e-01 -8.15299809e-01 -8.32894623e-01 -7.92060018e-01
6.24074876e-01 6.53466344e-01 9.08920407e-01 -3.11514437e-01
-3.52583677e-01 5.25499761e-01 -7.41997540e-01 -5.93176484e-01
-2.74111032e-01 2.74045646e-01 2.35973731e-01 -1.92450464e-01
1.14285576e+00 -3.73425841e-01 -2.51067221e-01 2.29798123e-01
-9.03162658e-01 -3.15043658e-01 1.84108764e-01 9.70614254e-01
2.70039253e-02 1.71830758e-01 6.18070662e-01 -9.87223983e-01
9.26479638e-01 -6.92119002e-01 -4.05200273e-02 2.90047735e-01
-8.32559168e-01 1.90264612e-01 5.06987870e-01 -8.45189273e-01
-9.95315790e-01 -2.77873635e-01 1.24857411e-01 -4.07447189e-01
-2.54797250e-01 2.31807724e-01 2.11479720e-02 4.95383173e-01
9.33822513e-01 -4.82350178e-02 -2.52575614e-02 -4.07466084e-01
8.04716289e-01 9.75777090e-01 4.13186476e-02 -5.33118844e-01
6.47873938e-01 4.00730282e-01 -6.84187293e-01 -1.12018609e+00
-1.04057610e+00 -9.64744449e-01 -8.77102375e-01 9.08318460e-02
8.92045915e-01 -8.40364099e-01 -1.70583144e-01 -1.91678852e-02
-9.09400463e-01 -2.54434347e-01 -4.58067566e-01 6.41070247e-01
-2.80267775e-01 5.48411906e-01 -6.29300416e-01 -8.31541717e-01
-2.93623716e-01 -9.47468519e-01 1.01364279e+00 8.61843303e-02
-7.15289831e-01 -1.33936381e+00 1.89805329e-01 1.50685593e-01
3.73789579e-01 -3.86885911e-01 1.01760733e+00 -1.22335911e+00
1.50743276e-01 -2.99457848e-01 -9.47077386e-03 2.16952875e-01
2.55507171e-01 1.59681022e-01 -1.09795940e+00 -2.73258150e-01
-4.71170731e-02 -4.48853701e-01 9.85313654e-01 4.21782702e-01
9.58970129e-01 -1.43772166e-04 -4.29076344e-01 2.47563303e-01
1.22027314e+00 -9.93553642e-03 3.87597233e-01 4.19386089e-01
5.47098815e-01 7.85026908e-01 6.13539636e-01 7.80589223e-01
2.23995134e-01 3.41903448e-01 -3.44461761e-02 2.31765851e-01
1.13181256e-01 -2.16882303e-01 4.23453301e-01 1.01032829e+00
5.15612662e-01 -3.69947314e-01 -8.76920819e-01 6.88709855e-01
-1.74083173e+00 -9.69054580e-01 2.75470447e-02 1.97062552e+00
9.40919280e-01 2.62619466e-01 -3.14169340e-02 2.96985954e-01
7.97606826e-01 1.46301553e-01 -3.65374774e-01 -5.11778951e-01
2.47168139e-01 4.34463739e-01 2.33765230e-01 4.64050710e-01
-1.24354839e+00 1.00015903e+00 6.35410786e+00 1.13733327e+00
-8.50671232e-01 3.03796023e-01 3.17554772e-01 -3.15503150e-01
-3.88116300e-01 -7.56407678e-02 -1.13856757e+00 4.97010142e-01
1.22784913e+00 -6.20726407e-01 -1.65981784e-01 1.12151802e+00
9.59370509e-02 9.58653018e-02 -9.81313765e-01 8.79150748e-01
5.46630979e-01 -1.22638106e+00 1.33365303e-01 -4.40331884e-02
8.82730901e-01 -1.56652495e-01 4.29489417e-03 1.02047563e+00
6.53705955e-01 -8.25668454e-01 2.36391544e-01 1.51747450e-01
5.14337480e-01 -8.26478541e-01 7.73329675e-01 4.35863525e-01
-8.89593840e-01 -4.77698147e-02 -8.64269793e-01 -6.99472427e-02
1.00976005e-01 5.56510806e-01 -1.01697123e+00 1.60589531e-01
5.50390899e-01 6.29488647e-01 -7.14921534e-01 7.18477547e-01
-5.08813933e-02 8.40752780e-01 -1.33332998e-01 -4.13057327e-01
2.65474260e-01 -3.75262089e-02 1.31738201e-01 1.44706821e+00
5.09851575e-01 -7.25894198e-02 2.93205023e-01 5.83435476e-01
4.37488519e-02 4.05378968e-01 -5.66876709e-01 -2.60044396e-01
5.55738449e-01 1.41889441e+00 -9.28645849e-01 -8.12177479e-01
-6.26458585e-01 7.27520823e-01 1.98966727e-01 2.22234234e-01
-7.03509629e-01 -7.67439246e-01 6.86938405e-01 -1.69536881e-02
5.03316164e-01 -1.97521493e-01 -4.28787529e-01 -1.41457105e+00
-5.84666729e-01 -4.65419263e-01 4.10123438e-01 -5.16069651e-01
-1.71501410e+00 2.85765618e-01 5.01962565e-02 -1.24189639e+00
-3.68146896e-01 -7.41792500e-01 -8.90122056e-01 5.13035715e-01
-1.09199643e+00 -8.32038760e-01 -1.42144501e-01 2.80042648e-01
1.01786304e+00 -3.24100852e-01 1.04032159e+00 1.15706496e-01
-6.52430773e-01 7.69788206e-01 5.39649844e-01 7.02748522e-02
1.18750083e+00 -1.36305571e+00 3.39847863e-01 5.53639472e-01
4.07183319e-01 8.17724109e-01 8.61113250e-01 -3.53648424e-01
-1.05502200e+00 -1.09829342e+00 9.15002406e-01 -7.10833907e-01
1.01892889e+00 -6.90797210e-01 -1.23930955e+00 5.17887414e-01
5.06411195e-01 -2.95441687e-01 1.43334138e+00 5.94137311e-01
-4.70318556e-01 3.07504088e-01 -8.06958616e-01 6.10692322e-01
3.89867604e-01 -3.72060716e-01 -1.18032002e+00 4.70243394e-01
9.58604395e-01 3.37192595e-01 -8.25376809e-01 1.49639174e-02
2.23446533e-01 -6.38954043e-01 7.73788571e-01 -6.31208956e-01
4.45705771e-01 -3.30624655e-02 -2.25471959e-01 -1.48007202e+00
-3.48313183e-01 -2.02864841e-01 -2.14797452e-01 1.45597720e+00
3.69849980e-01 -3.30183357e-01 7.76556194e-01 5.98069608e-01
-7.93698058e-02 -5.04526913e-01 -5.37324488e-01 -8.33709478e-01
4.21664566e-01 -6.99102521e-01 2.17067629e-01 1.32899332e+00
6.43875003e-01 7.18807936e-01 -1.99950889e-01 -2.70012587e-01
5.42547584e-01 -1.52325839e-01 9.09451902e-01 -1.46941400e+00
-1.95077643e-01 -6.35739625e-01 -2.09874257e-01 -9.29222941e-01
3.77748907e-01 -9.88091886e-01 2.10368201e-01 -1.59832859e+00
3.27491224e-01 -1.45827815e-01 -3.37947249e-01 3.54981333e-01
-4.43666607e-01 1.95560992e-01 1.26165137e-01 1.35985419e-01
-8.81194174e-01 7.51100838e-01 6.40995026e-01 -2.90142149e-01
-2.70609081e-01 -2.64162898e-01 -8.30304563e-01 7.98008263e-01
8.26266766e-01 -4.75248367e-01 -5.89923501e-01 -1.10092402e-01
-1.61514610e-01 -6.02354944e-01 -1.89532742e-01 -8.73922706e-01
3.24704736e-01 -8.66602454e-03 3.27059388e-01 -7.22936809e-01
2.92807966e-01 -4.49665189e-01 -6.59705281e-01 1.80509418e-01
-4.73672360e-01 -4.10190225e-01 1.95414722e-01 6.53533041e-01
-8.18039402e-02 -8.43617082e-01 7.30226934e-01 -6.76688552e-02
-8.52237582e-01 9.07804742e-02 -7.99045086e-01 2.42994994e-01
1.15684581e+00 -1.98194429e-01 -1.74498096e-01 -2.96430141e-01
-6.86590850e-01 2.88709521e-01 5.26461899e-01 6.19966507e-01
3.61058950e-01 -1.14227152e+00 -4.06148344e-01 2.15469703e-01
5.24821520e-01 -2.42962942e-01 2.76361793e-01 4.91802782e-01
-6.49756119e-02 4.10230964e-01 1.47300765e-01 -7.29357302e-01
-9.99435365e-01 9.86527085e-01 -2.74330288e-01 -2.05800980e-01
-6.67802930e-01 6.16980493e-01 1.61883503e-01 -6.08701050e-01
4.51804519e-01 -1.08172752e-01 -4.26665306e-01 6.52928829e-01
7.75914073e-01 3.45229924e-01 -9.76914093e-02 -1.77534655e-01
-6.34030178e-02 5.05497098e-01 -3.74383867e-01 -2.73918211e-01
1.33705306e+00 -1.82088599e-01 3.28203321e-01 1.03544867e+00
1.38735521e+00 -1.41173854e-01 -1.13294697e+00 -2.03855366e-01
2.95207292e-01 -3.29874218e-01 -5.16606160e-02 -1.72253400e-01
-4.55630392e-01 1.47109389e+00 3.36171508e-01 3.08891028e-01
5.16137242e-01 1.77817717e-01 7.40247071e-01 5.60132205e-01
4.88785915e-02 -1.37673414e+00 5.45964122e-01 7.15577781e-01
1.68620780e-01 -1.49134779e+00 -3.12630907e-02 -2.96234079e-02
-8.39031935e-01 1.17555833e+00 5.54460108e-01 -1.26684636e-01
9.55542326e-01 1.22209199e-01 1.68608114e-01 6.26448169e-03
-9.70277846e-01 -1.97164953e-01 2.47026995e-01 6.19179249e-01
7.07781434e-01 -4.66163419e-02 -3.60919654e-01 8.29783499e-01
-2.94871122e-01 -3.57220858e-01 6.04457676e-01 9.52784657e-01
-1.01582873e+00 -1.16734791e+00 -2.79083043e-01 6.44903421e-01
-1.94049358e-01 -1.95191994e-01 -8.46209824e-02 7.08491802e-01
-1.32624611e-01 1.13397753e+00 4.33454186e-01 -2.57584631e-01
1.30701354e-02 7.37891376e-01 1.76736206e-01 -1.25350511e+00
-1.33675620e-01 1.56327069e-01 -3.26234877e-01 7.54264966e-02
-1.10697530e-01 -6.10899031e-01 -1.19721937e+00 -1.82185709e-01
-5.54845095e-01 3.43272418e-01 6.44828737e-01 9.50919032e-01
3.30385149e-01 5.96486866e-01 5.02481401e-01 -9.16041315e-01
-7.77655780e-01 -1.40092492e+00 -5.53942502e-01 7.16434002e-01
-1.83975905e-01 -9.57088351e-01 -7.65568078e-01 1.51441306e-01]
|
[10.517535209655762, 7.212742805480957]
|
8e9523b3-bf8b-4183-9282-84a19476278a
|
the-best-of-both-modes-separately-leveraging
|
1907.13236
| null |
https://arxiv.org/abs/1907.13236v2
|
https://arxiv.org/pdf/1907.13236v2.pdf
|
The Best of Both Modes: Separately Leveraging RGB and Depth for Unseen Object Instance Segmentation
|
In order to function in unstructured environments, robots need the ability to recognize unseen novel objects. We take a step in this direction by tackling the problem of segmenting unseen object instances in tabletop environments. However, the type of large-scale real-world dataset required for this task typically does not exist for most robotic settings, which motivates the use of synthetic data. We propose a novel method that separately leverages synthetic RGB and synthetic depth for unseen object instance segmentation. Our method is comprised of two stages where the first stage operates only on depth to produce rough initial masks, and the second stage refines these masks with RGB. Surprisingly, our framework is able to learn from synthetic RGB-D data where the RGB is non-photorealistic. To train our method, we introduce a large-scale synthetic dataset of random objects on tabletops. We show that our method, trained on this dataset, can produce sharp and accurate masks, outperforming state-of-the-art methods on unseen object instance segmentation. We also show that our method can segment unseen objects for robot grasping. Code, models and video can be found at https://rse-lab.cs.washington.edu/projects/unseen-object-instance-segmentation/.
|
['Yu Xiang', 'Dieter Fox', 'Christopher Xie', 'Arsalan Mousavian']
|
2019-07-30
| null | null | null | null |
['unseen-object-instance-segmentation']
|
['computer-vision']
|
[ 7.27473497e-01 3.56983840e-01 3.37163210e-01 -3.61792296e-01
-7.30930924e-01 -9.90340114e-01 2.74525076e-01 -1.80148080e-01
-2.36481264e-01 4.10286635e-01 -4.65948075e-01 -8.17985013e-02
4.32410575e-02 -5.75440288e-01 -1.21493304e+00 -5.22817373e-01
1.79902747e-01 8.28799665e-01 5.30179203e-01 -2.57848471e-01
3.02362353e-01 4.99871284e-01 -1.70517135e+00 4.91644353e-01
6.98660433e-01 1.02844417e+00 6.23120844e-01 8.62172067e-01
1.89207166e-01 3.12017679e-01 -4.23543692e-01 -2.27077216e-01
9.27029729e-01 -7.92801529e-02 -9.28873539e-01 5.11143863e-01
4.67684239e-01 -7.63760269e-01 -2.41759583e-01 1.01584792e+00
1.77160740e-01 8.21722001e-02 5.44686794e-01 -1.34402990e+00
-6.28749907e-01 4.58265960e-01 -3.55022639e-01 -1.98157012e-01
3.93457919e-01 4.89500433e-01 6.34961307e-01 -9.29505944e-01
7.35955715e-01 1.24991548e+00 4.48078811e-01 7.98266470e-01
-1.18043375e+00 -3.48916709e-01 2.45719180e-01 -2.95018554e-01
-1.08230817e+00 -3.32743883e-01 6.71817482e-01 -5.38497210e-01
6.65007293e-01 4.86266389e-02 6.98300660e-01 1.33951104e+00
-2.71221191e-01 1.17243671e+00 9.33324456e-01 -3.72778684e-01
3.47937852e-01 -6.70147408e-03 1.95041616e-02 4.92716908e-01
4.25960898e-01 3.23985107e-02 -1.60046861e-01 1.94205999e-01
1.13635755e+00 3.58715802e-01 -3.54845315e-01 -8.65737557e-01
-1.56189549e+00 3.81578863e-01 4.52959329e-01 5.84746562e-02
-2.54748613e-01 3.37856323e-01 -1.42473117e-01 1.65389583e-01
1.09850459e-01 6.77764237e-01 -6.62524402e-01 -2.61025488e-01
-6.16059721e-01 4.68143880e-01 9.19462025e-01 1.58216691e+00
6.95050418e-01 -4.83137637e-01 2.77268708e-01 5.31341791e-01
4.55286391e-02 3.47130418e-01 2.57948786e-01 -1.47369015e+00
5.44012725e-01 6.29918218e-01 6.84917331e-01 -5.02761364e-01
-3.74067336e-01 1.86847717e-01 -1.46829993e-01 4.10781622e-01
7.99167991e-01 -1.26925290e-01 -1.56883752e+00 1.24758232e+00
5.43611586e-01 -7.37929046e-02 1.02566652e-01 1.10423994e+00
4.89654839e-01 3.34251612e-01 -4.93267566e-01 3.45229805e-01
1.04216504e+00 -1.20990753e+00 -3.14968646e-01 -5.29343367e-01
4.49614942e-01 -5.43747067e-01 1.28748918e+00 7.81331360e-01
-1.33373141e+00 -4.63996828e-01 -1.13794267e+00 -3.68679553e-01
-4.92283076e-01 9.63683333e-03 7.37710118e-01 4.25530225e-01
-7.94665396e-01 7.23063707e-01 -1.30112219e+00 -4.94306773e-01
6.65496051e-01 6.43800318e-01 -3.18105042e-01 -5.67934155e-01
-4.98022765e-01 6.27747893e-01 7.42484152e-01 3.66843253e-01
-1.04850769e+00 -3.91168028e-01 -8.11374068e-01 -3.09936434e-01
7.65199006e-01 -5.19508421e-01 1.69389629e+00 -8.86759400e-01
-1.43860674e+00 7.22522736e-01 1.62059650e-01 -5.54622635e-02
8.17720532e-01 -4.58899856e-01 3.40020537e-01 4.29016322e-01
-1.34201109e-01 1.09103692e+00 9.20484543e-01 -1.82158935e+00
-6.99064195e-01 -6.13793314e-01 5.05751610e-01 1.63755581e-01
-8.98781940e-02 -5.41920125e-01 -7.05098689e-01 -3.91287118e-01
4.74069744e-01 -1.18192446e+00 -2.84448922e-01 2.37384081e-01
-5.33414066e-01 6.55516386e-02 9.95464087e-01 -3.00695628e-01
3.11898679e-01 -2.17917371e+00 2.52696157e-01 -1.54532297e-02
1.77888677e-01 6.44915923e-02 -1.26509771e-01 3.00940126e-01
3.38194996e-01 1.83601379e-02 -5.22801578e-01 -4.72294807e-01
3.10476989e-01 4.53414321e-01 -3.36931348e-01 3.10566336e-01
4.19102073e-01 1.10440922e+00 -1.19867086e+00 -3.16337049e-01
3.76691341e-01 2.93631434e-01 -6.52648807e-01 2.22436890e-01
-7.46086538e-01 7.53138900e-01 -6.87910438e-01 1.02862203e+00
7.59540319e-01 -3.44771057e-01 2.65711714e-02 -2.81252209e-02
1.09429240e-01 -1.67393431e-01 -1.23419333e+00 2.10635686e+00
-9.54113901e-02 3.74899328e-01 2.41583660e-01 -8.57367218e-01
6.83339715e-01 -7.36835301e-02 4.98602360e-01 -2.37962827e-01
2.49769747e-01 5.58620334e-01 -1.05200060e-01 -9.05484140e-01
6.14037514e-01 1.09163195e-01 -6.76133260e-02 2.47858852e-01
-1.01653554e-01 -8.59541774e-01 2.71478057e-01 7.26145059e-02
1.32896674e+00 6.84001982e-01 -2.85031885e-01 8.91398638e-02
-2.25214317e-01 4.54744339e-01 2.46293515e-01 8.57229471e-01
-2.32106954e-01 1.01739359e+00 3.06647956e-01 -3.98153067e-01
-1.28848052e+00 -1.14089215e+00 -1.53536439e-01 8.30488682e-01
7.79642105e-01 1.58159792e-01 -9.92466450e-01 -7.63740957e-01
2.20031917e-01 4.95898753e-01 -6.38666391e-01 1.65684685e-01
-6.39335930e-01 -3.35751474e-01 2.13171393e-01 7.04641819e-01
4.21372235e-01 -1.31229222e+00 -9.91774917e-01 1.85070887e-01
1.21000148e-01 -1.49264205e+00 -1.30813003e-01 3.94560128e-01
-1.01820064e+00 -1.36906695e+00 -7.47078478e-01 -9.21315789e-01
9.49084103e-01 4.08695459e-01 9.35479462e-01 4.43352908e-02
-4.49059397e-01 7.94686615e-01 -6.73192143e-01 -5.32583833e-01
-2.91280478e-01 2.25485384e-01 -1.21659383e-01 -3.22063476e-01
2.42966518e-01 -5.00317097e-01 -8.15363765e-01 4.70743626e-01
-1.27423573e+00 1.14718579e-01 7.76742399e-01 5.04985452e-01
6.25447035e-01 -2.14645475e-01 1.80107251e-01 -6.89971745e-01
2.19929934e-01 -3.56544882e-01 -6.68422997e-01 2.32876584e-01
1.82535592e-02 2.82274210e-03 4.22833055e-01 -7.57593632e-01
-7.33505130e-01 6.83397651e-01 1.72823146e-01 -6.23702109e-01
-4.05454844e-01 -1.43957332e-01 -1.36264950e-01 -1.11640275e-01
6.72788978e-01 -1.15025677e-01 -1.36568636e-01 -5.19466221e-01
3.68004620e-01 8.50169003e-01 7.32127964e-01 -8.32813680e-01
8.26049089e-01 7.94115663e-01 -2.53774732e-01 -7.38176763e-01
-8.66134465e-01 -4.00582582e-01 -1.00189793e+00 3.04311588e-02
9.25619245e-01 -6.71148181e-01 -6.41665757e-01 5.24179757e-01
-1.13160610e+00 -1.13068795e+00 -4.93322015e-01 3.95411342e-01
-9.75400507e-01 1.58808589e-01 -5.58805645e-01 -8.22554052e-01
8.31943974e-02 -1.32870734e+00 1.61780083e+00 1.65439680e-01
1.01798408e-01 -3.65715444e-01 -4.04960811e-01 5.16723514e-01
7.04738274e-02 6.30483210e-01 3.92645597e-01 -6.10042393e-01
-1.16520905e+00 -1.81680053e-01 -2.34917447e-01 3.81945550e-01
1.71997517e-01 -1.00885138e-01 -8.90565574e-01 -3.37112129e-01
-5.47747733e-03 -8.00158679e-01 6.99222088e-01 2.62531843e-02
1.34037817e+00 -5.78277968e-02 -4.22120899e-01 6.23801172e-01
1.15621626e+00 7.90050775e-02 5.30526400e-01 3.54147077e-01
9.26690996e-01 5.59992909e-01 9.30606604e-01 2.51529425e-01
3.56955856e-01 5.37676156e-01 8.00486803e-01 6.11392073e-02
1.90948427e-01 -1.25735462e-01 7.12670833e-02 2.28163481e-01
-1.59225501e-02 -5.02007186e-01 -1.01697612e+00 7.40418971e-01
-1.99875665e+00 -5.08881450e-01 -6.65750280e-02 1.94696081e+00
6.05534077e-01 3.68072242e-01 3.95885576e-03 1.49251044e-01
5.85803449e-01 -3.73806566e-01 -9.97998893e-01 -9.34761390e-02
2.19762936e-01 1.86780944e-01 5.92980266e-01 3.66709203e-01
-1.04575944e+00 1.11434305e+00 5.17602444e+00 2.03963667e-01
-8.40606451e-01 -1.12959243e-01 2.48411402e-01 -1.98012590e-01
-2.29594305e-01 -1.48760483e-01 -6.14222765e-01 2.88101882e-01
4.22198474e-01 4.85454053e-01 7.13060200e-01 1.03502834e+00
-1.38985708e-01 -2.53596544e-01 -1.51499414e+00 9.36817884e-01
-3.71536575e-02 -8.50769758e-01 -1.44873872e-01 9.50322151e-02
9.14745629e-01 2.24946290e-01 -2.69701667e-02 1.99276701e-01
4.31977063e-01 -9.48916733e-01 9.90109384e-01 3.33556145e-01
4.56669539e-01 -2.03445122e-01 4.32313770e-01 5.29742897e-01
-8.35270405e-01 -3.00136715e-01 -3.61381412e-01 6.08221740e-02
1.25335321e-01 3.18796068e-01 -1.03233469e+00 2.70724148e-01
8.94568205e-01 6.11764312e-01 -4.06780392e-01 1.04497433e+00
-2.70159423e-01 3.96069512e-02 -6.68749213e-01 8.85681435e-02
2.79079407e-01 -7.42897741e-04 3.50045025e-01 7.62582064e-01
3.31281781e-01 3.89427871e-01 3.10405523e-01 9.57459509e-01
-6.73909113e-02 -4.83650446e-01 -5.90372980e-01 -2.22206086e-01
3.30957711e-01 1.15127289e+00 -1.07674611e+00 -2.13565350e-01
-1.13645658e-01 1.42459273e+00 2.68490553e-01 4.37608182e-01
-6.74668312e-01 -4.43694890e-01 3.63813043e-01 7.49646798e-02
7.47090340e-01 -6.82151794e-01 -3.27683151e-01 -1.08275950e+00
4.36698437e-01 -8.20265532e-01 -4.87904064e-02 -1.01333988e+00
-1.19658768e+00 4.97468859e-01 7.77769238e-02 -1.10618615e+00
-2.75049359e-01 -1.18868828e+00 -1.12966709e-01 3.50600392e-01
-1.39227831e+00 -1.24663484e+00 -9.58150208e-01 5.03893673e-01
9.48457778e-01 5.31968892e-01 5.77948332e-01 -4.82270233e-02
-2.75921196e-01 1.22046456e-01 3.59822772e-02 -1.00655193e-02
5.11921644e-01 -1.43754101e+00 6.38575137e-01 5.84567010e-01
-1.01462193e-01 5.26474833e-01 6.79260731e-01 -6.55587435e-01
-1.67299843e+00 -9.02047276e-01 -1.10261388e-01 -1.04741669e+00
5.47521293e-01 -8.61377418e-01 -7.21104145e-01 1.08883905e+00
-2.41940498e-01 4.21862006e-01 4.48754430e-02 -4.92495239e-01
-6.23262562e-02 2.77765304e-01 -1.32575047e+00 5.69157183e-01
1.43161488e+00 -5.85562848e-02 -7.05016732e-01 5.16308963e-01
1.03631198e+00 -9.62560415e-01 -8.95887017e-01 6.44115508e-01
9.26762104e-01 -8.66281033e-01 1.03991067e+00 -4.96944040e-01
5.91722310e-01 -3.84216845e-01 -4.05385971e-01 -9.99688208e-01
2.41640285e-01 -7.93792248e-01 -3.63838881e-01 7.57254004e-01
5.58701098e-01 -4.93306518e-01 1.08963168e+00 1.07543159e+00
-4.66315687e-01 -7.77913988e-01 -4.31150943e-01 -9.78256226e-01
-4.66503650e-02 -4.85127449e-01 5.40598512e-01 6.70094967e-01
-2.41516575e-01 -2.13798642e-01 1.66064903e-01 3.26105356e-01
6.18502200e-01 2.47509643e-01 1.14843214e+00 -1.14435601e+00
-3.79672617e-01 -4.46338654e-02 -2.82612264e-01 -1.36882818e+00
-1.93570629e-01 -4.60458994e-01 4.96851206e-01 -1.78187215e+00
-5.76226823e-02 -7.21547365e-01 1.24604389e-01 5.35245240e-01
-5.16300723e-02 5.24009585e-01 3.10582161e-01 3.18051696e-01
-6.41113997e-01 3.12751561e-01 1.62358487e+00 -1.35741860e-01
-5.36991119e-01 -5.45879416e-02 -4.13788646e-01 8.74669909e-01
9.55936134e-01 -4.50694591e-01 -3.69854093e-01 -7.81360626e-01
-2.94861421e-02 -2.87260205e-01 5.82875013e-01 -1.10436118e+00
2.62561571e-02 -2.46667534e-01 5.38447857e-01 -4.94562685e-01
7.37716436e-01 -9.50410903e-01 -1.49771363e-01 4.04518038e-01
-9.97931063e-02 -1.79929689e-01 2.05575123e-01 5.88780880e-01
2.94823140e-01 -4.92223620e-01 4.75130200e-01 -5.59088647e-01
-6.80147767e-01 2.58465528e-01 1.45714477e-01 1.31953418e-01
1.21647537e+00 -6.53048933e-01 -3.33077282e-01 -3.00364755e-02
-7.19472051e-01 3.44890296e-01 1.12772739e+00 6.41190827e-01
6.72745705e-01 -8.23792338e-01 -4.04889643e-01 1.62239596e-01
2.59254485e-01 9.43658650e-01 -3.85715961e-02 6.39271975e-01
-8.15032005e-01 3.30595411e-02 -6.79758489e-02 -8.42198253e-01
-6.75918400e-01 7.54248738e-01 1.27451688e-01 1.79735929e-01
-7.25361347e-01 9.37518418e-01 2.84672499e-01 -5.24980724e-01
3.62856269e-01 -8.62308860e-01 5.83642960e-01 -4.27274108e-01
2.69562066e-01 3.13651592e-01 5.04694246e-02 -1.34355798e-01
-1.39980987e-01 5.24168253e-01 -5.94484992e-02 -2.64161468e-01
1.53508174e+00 -5.23571149e-02 2.06448823e-01 5.31354010e-01
1.09637201e+00 -2.74787366e-01 -1.90079701e+00 1.99320436e-01
-5.73378839e-02 -5.81091046e-01 -3.37727100e-01 -7.63582408e-01
-8.30423594e-01 8.30047011e-01 4.18964297e-01 2.70965308e-01
9.64029193e-01 2.37619758e-01 1.05786693e+00 8.67721260e-01
8.19252551e-01 -1.07675564e+00 4.66251224e-01 3.66693556e-01
9.27751660e-01 -1.48982716e+00 -2.97992259e-01 -6.54618025e-01
-4.61701602e-01 1.23439908e+00 7.67986953e-01 -3.44349861e-01
1.74545303e-01 3.51027399e-01 -2.81632300e-02 -4.46233422e-01
-2.27428138e-01 -2.58689642e-01 4.05116268e-02 6.68341279e-01
-2.78884023e-01 -8.94972160e-02 4.07039315e-01 3.67916971e-01
-3.11546654e-01 1.45670757e-01 7.10811913e-01 1.61317360e+00
-5.11259258e-01 -1.04475653e+00 -4.56023693e-01 3.34560722e-01
-2.43823454e-01 1.98984861e-01 -6.39831603e-01 1.04423416e+00
1.95830047e-01 9.10706401e-01 -7.88295525e-04 -3.63516927e-01
4.42740768e-01 -9.96866003e-02 1.04047644e+00 -9.10264254e-01
-2.61000246e-01 -1.75889477e-01 -1.95388600e-01 -7.95276701e-01
-3.72063011e-01 -6.56255960e-01 -1.44471002e+00 2.03638092e-01
-2.94904709e-01 -3.13522428e-01 9.98867810e-01 8.26194048e-01
2.35243529e-01 3.26866269e-01 3.71454567e-01 -1.76127470e+00
-3.75552803e-01 -8.42409194e-01 -3.68286788e-01 5.93279719e-01
5.83769381e-01 -8.03185821e-01 -5.01650274e-01 3.01533878e-01]
|
[6.147257328033447, -1.0453237295150757]
|
d27d6a29-2ee4-486e-9cd8-193162cc843a
|
east-an-efficient-and-accurate-scene-text
|
1704.03155
| null |
http://arxiv.org/abs/1704.03155v2
|
http://arxiv.org/pdf/1704.03155v2.pdf
|
EAST: An Efficient and Accurate Scene Text Detector
|
Previous approaches for scene text detection have already achieved promising
performances across various benchmarks. However, they usually fall short when
dealing with challenging scenarios, even when equipped with deep neural network
models, because the overall performance is determined by the interplay of
multiple stages and components in the pipelines. In this work, we propose a
simple yet powerful pipeline that yields fast and accurate text detection in
natural scenes. The pipeline directly predicts words or text lines of arbitrary
orientations and quadrilateral shapes in full images, eliminating unnecessary
intermediate steps (e.g., candidate aggregation and word partitioning), with a
single neural network. The simplicity of our pipeline allows concentrating
efforts on designing loss functions and neural network architecture.
Experiments on standard datasets including ICDAR 2015, COCO-Text and MSRA-TD500
demonstrate that the proposed algorithm significantly outperforms
state-of-the-art methods in terms of both accuracy and efficiency. On the ICDAR
2015 dataset, the proposed algorithm achieves an F-score of 0.7820 at 13.2fps
at 720p resolution.
|
['Xinyu Zhou', 'He Wen', 'Yuzhi Wang', 'Weiran He', 'Shuchang Zhou', 'Jiajun Liang', 'Cong Yao']
|
2017-04-11
|
east-an-efficient-and-accurate-scene-text-1
|
http://openaccess.thecvf.com/content_cvpr_2017/html/Zhou_EAST_An_Efficient_CVPR_2017_paper.html
|
http://openaccess.thecvf.com/content_cvpr_2017/papers/Zhou_EAST_An_Efficient_CVPR_2017_paper.pdf
|
cvpr-2017-7
|
['curved-text-detection']
|
['computer-vision']
|
[ 9.30481926e-02 -5.87606072e-01 1.55117452e-01 -1.36054829e-01
-7.93763876e-01 -5.03189921e-01 5.66173673e-01 3.32970083e-01
-5.81965864e-01 1.81950495e-01 -4.66679186e-02 -2.82600611e-01
3.26668382e-01 -8.08690608e-01 -6.73586965e-01 -4.30141211e-01
3.29310030e-01 5.14855564e-01 6.44307256e-01 1.45837948e-01
4.66875464e-01 4.34193850e-01 -1.63897765e+00 4.98147100e-01
8.00490618e-01 1.08614969e+00 2.65213341e-01 9.46477175e-01
-3.75252545e-01 9.21055198e-01 -6.44302309e-01 -7.09133983e-01
1.84620604e-01 2.82548890e-02 -6.03331685e-01 1.25451088e-01
7.75516570e-01 -6.36027157e-01 -5.80456614e-01 8.10598791e-01
7.02110469e-01 -1.86557338e-01 4.99355048e-01 -7.63678014e-01
-3.69737864e-01 6.38596296e-01 -9.30438280e-01 5.55795394e-02
1.59009427e-01 6.97637722e-02 1.12681913e+00 -1.21669388e+00
3.29052448e-01 1.16102505e+00 8.22404861e-01 2.69237664e-02
-1.09307015e+00 -5.05665421e-01 1.51462123e-01 1.89912543e-01
-1.62358475e+00 -4.49937403e-01 3.55389476e-01 -4.53186184e-01
1.14403224e+00 1.24058485e-01 2.79308826e-01 8.79733145e-01
1.14277296e-01 1.22367465e+00 6.72642946e-01 -4.03321892e-01
5.06210551e-02 -9.79458839e-02 2.81943142e-01 7.49177217e-01
4.86996919e-01 -5.83080113e-01 -7.76414275e-01 4.78480980e-02
4.98634458e-01 -1.00668721e-01 -3.86778340e-02 -5.00234589e-02
-1.14809203e+00 6.42893314e-01 1.08996205e-01 1.90020487e-01
-1.83907777e-01 1.74810350e-01 6.55659556e-01 -1.33453250e-01
3.53385955e-01 5.35347983e-02 -3.77353579e-01 -1.30685300e-01
-1.28860736e+00 2.84404516e-01 6.80299342e-01 1.03117073e+00
3.19679797e-01 1.10989241e-02 -4.62578654e-01 1.12043536e+00
9.46582928e-02 7.28145897e-01 1.83881104e-01 -3.70400071e-01
7.90846646e-01 6.19944692e-01 2.79152915e-02 -1.11474741e+00
-5.21942794e-01 -4.80667263e-01 -9.21461582e-01 -1.25288695e-01
4.33825523e-01 -1.46445096e-01 -1.07772446e+00 8.41110885e-01
4.61593010e-02 -1.48037672e-01 -2.79013962e-01 7.53624022e-01
9.38733935e-01 8.21648419e-01 -1.72109723e-01 1.52474061e-01
1.48971796e+00 -1.34262025e+00 -6.22381866e-01 -3.98027778e-01
5.69879234e-01 -1.18295562e+00 1.27282655e+00 7.46466696e-01
-1.15223706e+00 -5.41604936e-01 -1.00662065e+00 -4.33272183e-01
-3.50949794e-01 7.83313751e-01 3.39845896e-01 5.87674201e-01
-9.90310371e-01 2.91059852e-01 -7.34051943e-01 -5.60967267e-01
6.51946902e-01 3.35009128e-01 1.28086030e-01 -2.30883230e-02
-6.83802366e-01 3.55249107e-01 3.83709192e-01 5.17282970e-02
-5.64053714e-01 -4.27463293e-01 -4.77864593e-01 4.14341778e-01
5.08917987e-01 -4.66940343e-01 1.33306897e+00 -6.61403596e-01
-1.48624444e+00 7.89484501e-01 -4.22586352e-02 -8.01218152e-01
9.46120620e-01 -7.02605188e-01 -1.36386707e-01 3.53760689e-01
-1.19959377e-01 9.60948050e-01 9.30569530e-01 -9.10583436e-01
-1.03064024e+00 -1.89330071e-01 -1.71158746e-01 1.86662957e-01
-6.38178885e-01 3.68321508e-01 -1.13782036e+00 -7.20184267e-01
-1.00357197e-01 -7.37411916e-01 5.22315167e-02 2.61444151e-01
-7.81205595e-01 -1.59855708e-01 1.03141820e+00 -5.96077740e-01
1.28087544e+00 -1.97460985e+00 -4.27519619e-01 -1.75199836e-01
1.00398138e-01 5.64072609e-01 6.80300146e-02 3.85266006e-01
3.23366046e-01 2.01672822e-01 3.16552036e-02 -6.38274312e-01
1.37748942e-01 -3.10796738e-01 -4.24469292e-01 4.37375784e-01
7.68314749e-02 8.52571845e-01 -3.47740054e-01 -8.01546633e-01
6.21238053e-01 6.14893496e-01 -4.43258047e-01 -3.45305800e-02
-3.18867117e-01 -2.46518761e-01 -2.91536778e-01 7.26922989e-01
8.33792388e-01 -5.42676628e-01 1.49928126e-03 -1.92473933e-01
-2.52786994e-01 3.11788410e-01 -1.15301132e+00 1.46703303e+00
-1.65308207e-01 1.10859394e+00 -3.22652347e-02 -7.93037832e-01
8.73560250e-01 5.60610509e-03 2.74853230e-01 -8.08576763e-01
3.93080086e-01 1.43544953e-02 -3.57363164e-01 -3.25919032e-01
9.97963607e-01 6.68686688e-01 7.28051066e-02 2.40484446e-01
-1.73837170e-01 9.68568847e-02 4.98356134e-01 8.21237713e-02
1.10316086e+00 2.46812664e-02 6.38826638e-02 -1.13753967e-01
5.05454600e-01 1.10297851e-01 6.45942315e-02 9.90909398e-01
-7.22615048e-02 8.94469917e-01 6.54897392e-01 -6.77414358e-01
-1.17618787e+00 -8.89832914e-01 -2.81448662e-01 1.22079766e+00
1.45412445e-01 -6.57482326e-01 -7.64210880e-01 -3.97227108e-01
-2.79110819e-01 5.44287860e-01 -4.02532697e-01 3.84745985e-01
-7.21070528e-01 -9.41497922e-01 8.50146234e-01 6.05403006e-01
8.04355145e-01 -8.93211186e-01 -7.40752697e-01 2.10352078e-01
-1.03297174e-01 -1.81401551e+00 -3.74648571e-01 6.34126067e-02
-6.71232700e-01 -8.31879497e-01 -7.42777944e-01 -7.73298085e-01
4.47572559e-01 5.72446823e-01 1.21137404e+00 -5.14503792e-02
-5.64487457e-01 -7.01697543e-03 -3.76154810e-01 -2.28761956e-01
-5.30861691e-02 2.53486603e-01 -5.80568492e-01 -8.24452639e-02
3.46620232e-01 -2.46030595e-02 -7.41632164e-01 2.93527097e-01
-9.17858243e-01 4.52741981e-01 7.27977335e-01 7.54195094e-01
6.17107689e-01 2.07271501e-01 8.94256011e-02 -7.13386118e-01
4.16433603e-01 7.91351721e-02 -9.54765916e-01 2.39968464e-01
-4.53794211e-01 -2.90113002e-01 8.36180925e-01 -2.34874666e-01
-1.05819428e+00 3.67677152e-01 -2.91828662e-01 -2.19969288e-01
-2.86841244e-01 2.53572226e-01 1.02388911e-01 1.98173672e-01
5.44402122e-01 4.48621988e-01 -6.25360608e-01 -5.27696967e-01
2.50253499e-01 8.57853651e-01 6.03044271e-01 -4.09591675e-01
5.52418768e-01 7.11018682e-01 -2.09948078e-01 -1.10764098e+00
-7.58042574e-01 -4.45165724e-01 -5.95014572e-01 2.02878937e-02
8.33563030e-01 -1.08362329e+00 -7.30353832e-01 9.77682352e-01
-1.22765398e+00 -2.46492624e-01 1.82782203e-01 1.23565391e-01
-2.68504292e-01 5.38224995e-01 -8.79711449e-01 -6.59152865e-01
-8.19012523e-01 -1.07685065e+00 1.49159515e+00 1.66295364e-01
2.81232968e-02 -5.97278535e-01 -3.68047804e-01 4.15653795e-01
2.25358382e-01 -6.68313280e-02 7.39068925e-01 -6.84237599e-01
-6.66375935e-01 -4.38289613e-01 -7.59420574e-01 1.68980300e-01
-2.28447706e-01 5.00946283e-01 -1.06468642e+00 -9.15552899e-02
-6.01744771e-01 -3.90602767e-01 1.06716859e+00 5.20048857e-01
1.58777189e+00 -1.44724235e-01 -1.74097925e-01 5.45895576e-01
1.54152012e+00 6.01650402e-03 6.32182360e-01 4.32345122e-01
8.11363697e-01 3.96389842e-01 5.49967825e-01 7.29631901e-01
3.83507103e-01 5.51267266e-01 4.18139935e-01 -2.59219706e-01
-2.01265663e-01 -6.65307045e-02 1.96682408e-01 5.18022060e-01
4.62501526e-01 -7.43954539e-01 -1.26766169e+00 5.46489596e-01
-1.85467052e+00 -6.89587951e-01 -6.34035707e-01 2.07432652e+00
5.16667426e-01 4.93183941e-01 7.10200742e-02 2.62796253e-01
8.74919832e-01 3.06245893e-01 -4.36583459e-01 -3.16529989e-01
-4.11942959e-01 1.37695208e-01 5.11674345e-01 1.91016197e-01
-1.51363099e+00 1.17981243e+00 6.31779623e+00 1.06183279e+00
-1.25616217e+00 -2.05697104e-01 9.07248795e-01 -2.72463471e-01
3.98634434e-01 -4.62867498e-01 -1.10439432e+00 3.91131312e-01
6.87474787e-01 1.57366619e-01 2.10440993e-01 7.65902519e-01
2.36791298e-01 -2.31833085e-01 -7.94084191e-01 1.10666931e+00
1.62033439e-01 -1.41408968e+00 7.34746754e-02 -2.15197176e-01
7.22566426e-01 3.36608171e-01 1.76483870e-01 1.38733476e-01
1.39810905e-01 -1.16538596e+00 9.08566356e-01 3.01049501e-02
7.41786063e-01 -8.04866910e-01 8.55186999e-01 3.08488518e-01
-1.13443947e+00 4.77261618e-02 -4.59606916e-01 1.82250589e-01
-1.11309282e-01 7.55126238e-01 -7.98912108e-01 2.98586518e-01
1.02529502e+00 6.38051152e-01 -7.64703035e-01 1.07057774e+00
-1.39457941e-01 6.40181422e-01 -5.14420629e-01 -2.19820693e-01
3.15798759e-01 1.29194722e-01 6.06296137e-02 1.69147182e+00
2.59437233e-01 -3.89414966e-01 6.95948526e-02 5.92279732e-01
-3.91797334e-01 5.04801691e-01 -2.33081937e-01 -1.00440040e-01
4.09853220e-01 1.43577230e+00 -1.23637688e+00 -4.16053653e-01
-6.66558802e-01 9.79226291e-01 1.79614127e-01 1.15346596e-01
-1.10416210e+00 -5.77878594e-01 2.00316072e-01 8.87163728e-03
8.99296939e-01 -3.02943915e-01 -7.28188336e-01 -1.09392965e+00
3.38741750e-01 -9.21818972e-01 2.61339635e-01 -7.94826984e-01
-9.13080037e-01 6.05667174e-01 -4.71295267e-01 -1.05716097e+00
1.88389093e-01 -8.12898099e-01 -5.54911017e-01 4.77789789e-01
-1.47231042e+00 -1.24082613e+00 -5.21224320e-01 4.50305730e-01
1.03630817e+00 9.19754952e-02 4.81305629e-01 4.67579216e-01
-9.55699801e-01 7.06787944e-01 4.50025231e-01 3.52765858e-01
8.56099308e-01 -1.05409276e+00 8.15387845e-01 1.06275034e+00
2.46642143e-01 1.23271100e-01 6.28379047e-01 -3.88442606e-01
-1.44535148e+00 -1.15790462e+00 6.90075099e-01 -1.21780857e-01
7.02356517e-01 -7.21626759e-01 -8.98151577e-01 2.78639406e-01
3.62224460e-01 1.02669664e-01 2.60259092e-01 -3.76435667e-02
-3.40975374e-01 -3.23412269e-02 -7.92335093e-01 7.85258353e-01
7.81265557e-01 -2.53482878e-01 -1.34656474e-01 5.09434819e-01
5.56341529e-01 -5.64554214e-01 -5.48063695e-01 2.35968426e-01
7.18109250e-01 -1.27392364e+00 9.00912881e-01 -9.95968580e-02
6.46114707e-01 -2.30073974e-01 -7.37267360e-02 -4.46701825e-01
-3.04806139e-02 -4.92546141e-01 -2.06683084e-01 1.23923576e+00
3.22610050e-01 -2.34032065e-01 9.38188732e-01 6.17007576e-02
-6.13841489e-02 -8.80269527e-01 -6.33975625e-01 -5.08494139e-01
4.71633486e-02 -5.99607348e-01 3.63993496e-01 5.38809955e-01
-4.74694848e-01 4.37777728e-01 -4.70872909e-01 1.30776316e-01
5.30290961e-01 1.41188994e-01 1.03634667e+00 -1.07044041e+00
-1.53454989e-01 -7.16935277e-01 -1.58635810e-01 -1.40788138e+00
-6.98526800e-02 -3.21628600e-01 1.86948285e-01 -1.66729891e+00
3.13327402e-01 -1.28508285e-01 1.22605450e-01 4.50021654e-01
-2.83197433e-01 3.64380240e-01 2.15171486e-01 2.02004805e-01
-9.68903244e-01 3.99942398e-01 9.03151989e-01 -2.59905607e-01
-5.24664372e-02 -2.29942665e-01 -4.69892412e-01 1.07413685e+00
9.30420220e-01 -3.09374601e-01 -9.62544084e-02 -8.31393957e-01
3.12248826e-01 -3.44180405e-01 2.67446697e-01 -1.13746440e+00
4.96974736e-01 2.14322224e-01 6.37728095e-01 -1.44117939e+00
2.35268459e-01 -6.98657990e-01 -2.91989326e-01 4.29180175e-01
-3.34394723e-01 -2.76521896e-04 4.18632746e-01 3.22929204e-01
-1.53009936e-01 -1.98311895e-01 9.13866043e-01 3.40708137e-01
-5.76278448e-01 1.25584990e-01 -5.15342832e-01 8.56065899e-02
8.44785750e-01 -2.44005054e-01 -7.33779430e-01 -2.04053491e-01
-1.30232036e-01 2.37731904e-01 5.26288092e-01 4.82514352e-01
4.92127478e-01 -7.45270252e-01 -8.36652398e-01 8.65666848e-03
8.25664923e-02 2.64748633e-01 2.18790457e-01 7.66145825e-01
-1.05386114e+00 7.08507419e-01 5.71100488e-02 -9.39808190e-01
-1.42618096e+00 3.50289941e-01 2.60011941e-01 -5.38199782e-01
-8.51378381e-01 7.51671791e-01 3.56675774e-01 -8.38264152e-02
5.77224374e-01 -4.08640891e-01 -4.89250496e-02 6.13976866e-02
6.55727506e-01 4.74758297e-01 3.47316831e-01 -4.51442152e-01
-2.88910985e-01 6.83183551e-01 -4.18783784e-01 2.25384206e-01
1.12511289e+00 -9.75634456e-02 7.01652393e-02 1.24181628e-01
9.91138339e-01 8.64927545e-02 -1.24424291e+00 -3.16326439e-01
1.88687053e-02 -4.00112450e-01 2.72124559e-01 -7.85037696e-01
-9.39803243e-01 1.13894415e+00 5.66960931e-01 2.63268292e-01
1.25066626e+00 -2.05492347e-01 9.45328474e-01 6.95141852e-01
-6.15835376e-02 -1.28368020e+00 2.49571130e-01 7.99237609e-01
5.85954070e-01 -1.18047678e+00 2.47551486e-01 -5.38843393e-01
-3.65765721e-01 1.35658598e+00 6.34260058e-01 -7.31971040e-02
2.92013973e-01 6.27510011e-01 -2.13844720e-02 2.01349221e-02
-8.21272314e-01 -1.09228842e-01 2.58581966e-01 7.09072277e-02
5.73618114e-01 -1.85032576e-01 -1.49513528e-01 2.22047240e-01
-9.69859883e-02 -2.50714123e-01 4.92083549e-01 8.24101210e-01
-6.32116258e-01 -6.96516752e-01 -5.01528680e-01 6.14449441e-01
-1.00057089e+00 -4.16945755e-01 -5.02947867e-01 8.71833205e-01
-2.13221774e-01 8.91676188e-01 3.68195653e-01 -1.95033669e-01
2.88997859e-01 -2.22653776e-01 1.81345224e-01 -3.75464261e-01
-6.07889295e-01 4.57536221e-01 1.27938122e-01 -3.90765309e-01
-2.06731126e-01 -5.65648973e-01 -1.26546586e+00 -6.42136812e-01
-4.90663618e-01 -4.01635677e-01 7.86029637e-01 6.21087730e-01
5.52003503e-01 5.75607240e-01 4.55415457e-01 -6.76207781e-01
-2.66381383e-01 -8.81808162e-01 -3.08314979e-01 -9.84302070e-03
9.08995941e-02 -1.81875378e-01 -7.19094276e-02 3.00058812e-01]
|
[12.015491485595703, 2.296715021133423]
|
ed5b5a71-eb7f-41a5-95f6-8e7ddeb5dd96
|
text-based-localization-of-moments-in-a-video
|
2008.08716
| null |
https://arxiv.org/abs/2008.08716v2
|
https://arxiv.org/pdf/2008.08716v2.pdf
|
Text-based Localization of Moments in a Video Corpus
|
Prior works on text-based video moment localization focus on temporally grounding the textual query in an untrimmed video. These works assume that the relevant video is already known and attempt to localize the moment on that relevant video only. Different from such works, we relax this assumption and address the task of localizing moments in a corpus of videos for a given sentence query. This task poses a unique challenge as the system is required to perform: (i) retrieval of the relevant video where only a segment of the video corresponds with the queried sentence, and (ii) temporal localization of moment in the relevant video based on sentence query. Towards overcoming this challenge, we propose Hierarchical Moment Alignment Network (HMAN) which learns an effective joint embedding space for moments and sentences. In addition to learning subtle differences between intra-video moments, HMAN focuses on distinguishing inter-video global semantic concepts based on sentence queries. Qualitative and quantitative results on three benchmark text-based video moment retrieval datasets - Charades-STA, DiDeMo, and ActivityNet Captions - demonstrate that our method achieves promising performance on the proposed task of temporal localization of moments in a corpus of videos.
|
['Amit K. Roy-Chowdhury', 'Sudipta Paul', 'Niluthpol Chowdhury Mithun']
|
2020-08-20
| null | null | null | null |
['moment-retrieval']
|
['computer-vision']
|
[ 3.68241742e-02 -4.57156271e-01 -4.92643595e-01 -2.89174497e-01
-1.15976620e+00 -5.80089688e-01 7.22963631e-01 2.77239084e-01
-6.41769290e-01 2.25413054e-01 5.48397779e-01 2.60796666e-01
-8.29839036e-02 -2.65635550e-01 -8.07612538e-01 -7.40150094e-01
-3.92528087e-01 -5.06489305e-03 3.11790258e-01 1.02371134e-01
4.89660263e-01 2.90579587e-01 -1.21551776e+00 6.09393418e-01
2.34750342e-02 1.23901367e+00 3.08317602e-01 8.69542181e-01
1.02835298e-01 1.35409892e+00 -6.03762388e-01 -7.52390921e-02
1.03804916e-01 -6.32354856e-01 -9.13722932e-01 2.06157744e-01
7.65629411e-01 -5.86077213e-01 -8.50768149e-01 1.03298628e+00
3.81714702e-01 5.21132886e-01 5.34215331e-01 -1.39652419e+00
-5.22622466e-01 5.99022269e-01 -4.60672408e-01 8.66328001e-01
8.34529400e-01 -1.21121295e-01 1.08793342e+00 -8.07808399e-01
1.18914020e+00 8.90891731e-01 3.43093425e-01 2.57311821e-01
-7.49809206e-01 -2.02218443e-01 2.33762175e-01 7.87982285e-01
-1.71035266e+00 -4.91439164e-01 1.17131734e+00 -5.03057182e-01
9.34486628e-01 1.12711459e-01 5.99230051e-01 1.32142282e+00
2.92431056e-01 8.70100617e-01 3.04414690e-01 -1.64066613e-01
3.26409489e-01 -2.68598050e-01 -2.54521549e-01 6.17049217e-01
-6.45600259e-01 -4.86655772e-01 -1.07956183e+00 4.42613959e-02
5.53472638e-01 1.64898232e-01 -3.60242367e-01 -5.07049501e-01
-1.75501788e+00 7.38502264e-01 1.68433428e-01 6.53558731e-01
-5.98194242e-01 1.85801163e-01 8.54646981e-01 4.02841210e-01
4.98500675e-01 1.93671659e-01 -2.72338778e-01 -4.50812042e-01
-1.42932940e+00 7.62046874e-02 8.53679895e-01 8.84943068e-01
7.31579125e-01 -3.26619446e-01 -2.92339057e-01 4.26352829e-01
-6.49009645e-02 3.88805360e-01 5.09563327e-01 -1.12521231e+00
6.53600097e-01 1.85971215e-01 1.28399178e-01 -1.52451539e+00
2.64790673e-02 1.80826951e-02 -2.91709840e-01 -6.58895016e-01
3.41220349e-01 1.18095659e-01 -6.34292781e-01 1.78597462e+00
3.59414876e-01 6.08263731e-01 -2.34332681e-02 1.08194196e+00
6.29734516e-01 9.40719664e-01 7.59593248e-02 -4.31468308e-01
1.23202944e+00 -8.58084679e-01 -6.93233132e-01 -2.38774866e-02
5.88562131e-01 -7.31622934e-01 8.72979343e-01 6.68120831e-02
-9.68395591e-01 -4.08865005e-01 -8.71342361e-01 -3.14720720e-01
-3.75375360e-01 5.41905733e-03 1.88746259e-01 -9.40252393e-02
-1.14102674e+00 3.26372892e-01 -8.35299611e-01 -7.31039345e-01
-2.44880076e-02 1.34090990e-01 -6.20103240e-01 -2.27297828e-01
-1.22053039e+00 4.30184036e-01 5.07580757e-01 1.07912876e-01
-1.13429105e+00 -5.47430933e-01 -1.12076235e+00 5.81552088e-02
5.95655262e-01 -4.90901738e-01 1.11089826e+00 -1.09716249e+00
-1.07805145e+00 8.02311599e-01 -3.89475852e-01 -7.15039074e-01
3.03433746e-01 -3.45158815e-01 -3.90049189e-01 1.21557915e+00
3.65543783e-01 8.40055525e-01 1.00522041e+00 -7.06106722e-01
-7.36178458e-01 8.90198536e-03 5.11890113e-01 3.50158662e-01
-5.12558520e-01 3.32130790e-01 -9.37429607e-01 -6.75242603e-01
1.12804845e-01 -6.86978877e-01 2.34823987e-01 -2.12661549e-02
-1.36116743e-01 -2.93722302e-01 1.39541042e+00 -8.81686866e-01
1.36259890e+00 -2.17037988e+00 2.18717992e-01 -1.31392345e-01
4.97688912e-02 -3.41125757e-01 -3.51957291e-01 8.28893125e-01
1.78253260e-02 -1.37164831e-01 3.08109552e-01 -2.53420025e-01
-7.98449889e-02 1.65614128e-01 -6.23273492e-01 7.88517356e-01
-1.07973188e-01 9.29038823e-01 -1.20777476e+00 -1.11779976e+00
2.89751410e-01 6.80936098e-01 -5.40462017e-01 2.18493000e-01
-1.13603942e-01 4.02143061e-01 -4.10160512e-01 7.25544751e-01
3.26687917e-02 -3.60022336e-01 -1.01361880e-02 -6.66278183e-01
-5.07288985e-02 5.98745840e-03 -8.10172856e-01 2.36409616e+00
-2.18783632e-01 1.03418732e+00 -2.57123977e-01 -1.15257061e+00
2.50948787e-01 5.97517729e-01 1.28041625e+00 -7.61004686e-01
-5.80244027e-02 -2.11103946e-01 -5.00240922e-01 -9.57862258e-01
6.28470182e-01 3.89862537e-01 -2.48287156e-01 4.72919196e-01
3.77561957e-01 2.47045442e-01 5.66421032e-01 5.20940602e-01
1.27250457e+00 2.23288745e-01 6.40851110e-02 5.27532622e-02
5.33178627e-01 -2.43218929e-01 3.71807873e-01 7.53356218e-01
-5.56319594e-01 6.20739043e-01 7.08021462e-01 -3.93617094e-01
-1.06458306e+00 -9.91773367e-01 2.76406825e-01 1.29018772e+00
3.76037270e-01 -7.54840732e-01 -6.88230574e-01 -8.11479390e-01
-4.07245576e-01 2.14375138e-01 -7.16291964e-01 1.97897125e-02
-7.42689908e-01 -7.94730186e-02 2.21439436e-01 2.83253849e-01
5.39497197e-01 -8.61374378e-01 -6.49340808e-01 6.11531101e-02
-8.73898685e-01 -1.70290351e+00 -1.11347032e+00 -3.58762175e-01
-5.87910175e-01 -1.10417140e+00 -7.18590617e-01 -9.56900656e-01
5.80755651e-01 5.11274815e-01 9.80288267e-01 -1.79800868e-01
-1.60293937e-01 1.09162152e+00 -6.16809964e-01 4.18837070e-01
4.93315272e-02 -1.01404622e-01 8.26669019e-03 3.16598177e-01
3.45739573e-01 -5.06536365e-01 -9.45732594e-01 3.41413975e-01
-1.08101952e+00 -2.13571921e-01 2.55282193e-01 5.67353070e-01
7.15830564e-01 -3.59675065e-02 1.71857148e-01 -1.24654517e-01
3.10720474e-01 -7.64894307e-01 -1.95400372e-01 3.72755378e-01
1.54928043e-02 -2.44729847e-01 4.91497517e-01 -6.23754025e-01
-5.79304814e-01 2.19665036e-01 2.28807762e-01 -9.44783449e-01
1.78022422e-02 6.71211898e-01 1.18141435e-01 5.35798781e-02
2.84031957e-01 5.91514885e-01 -3.79961967e-01 -7.98474327e-02
2.56797373e-01 3.18541676e-01 8.28919053e-01 -4.52327639e-01
5.93781054e-01 7.95109093e-01 1.49016939e-02 -1.08291709e+00
-8.89183104e-01 -1.08868384e+00 -9.52827215e-01 -6.05641484e-01
1.19485593e+00 -1.20087862e+00 -7.73106635e-01 5.64596197e-03
-1.15000999e+00 -1.57142028e-01 -6.71382770e-02 6.34021401e-01
-8.83588016e-01 9.23243225e-01 -6.77935004e-01 -3.65497172e-01
-1.26079425e-01 -9.83494043e-01 1.38907540e+00 -1.21922299e-01
-3.25310707e-01 -1.05148864e+00 1.29299179e-01 3.09370011e-01
3.09859142e-02 3.53142798e-01 4.70671296e-01 -6.73430085e-01
-7.25831389e-01 -4.53241318e-01 -4.36370298e-02 8.15122128e-02
1.77350864e-01 5.40319979e-02 -6.58713877e-01 -4.67708260e-01
2.32037619e-01 -2.69983083e-01 6.48567855e-01 5.17464042e-01
1.02268064e+00 -5.35676360e-01 -2.39148781e-01 6.66552484e-01
1.23475552e+00 1.91226617e-01 4.06576157e-01 2.92821854e-01
6.19837344e-01 3.97403210e-01 8.55424881e-01 5.18160582e-01
3.76817822e-01 7.67400086e-01 4.04540092e-01 3.21851939e-01
1.45928994e-01 -3.71450186e-01 6.71525300e-01 9.49741840e-01
1.46322861e-01 -3.15005630e-01 -6.28846765e-01 8.00609112e-01
-2.12695098e+00 -1.55494702e+00 6.03226304e-01 1.99050152e+00
5.62018394e-01 -2.06185117e-01 1.78220034e-01 -2.44770572e-01
6.61930323e-01 7.25375354e-01 -4.88066316e-01 1.88290268e-01
-1.35265201e-01 -4.91610050e-01 1.52018026e-01 2.95790583e-01
-1.50196743e+00 7.84734964e-01 5.67109537e+00 6.84164464e-01
-1.39598882e+00 2.88273752e-01 4.83780146e-01 -6.19014680e-01
1.59126133e-01 3.03948782e-02 -3.93731564e-01 6.90457225e-01
1.09710205e+00 -1.86506778e-01 3.53438258e-01 8.55102658e-01
5.25509238e-01 -3.30022544e-01 -1.40400755e+00 1.24988806e+00
6.18206024e-01 -1.40907681e+00 1.73273593e-01 -2.59515345e-01
8.22911441e-01 -2.21721781e-03 6.88383058e-02 3.24828804e-01
-6.70474529e-01 -5.63128531e-01 9.55652475e-01 7.28895724e-01
5.71305990e-01 -7.14615583e-01 5.86136162e-01 1.35965034e-01
-1.48644340e+00 1.49421766e-02 -2.34904096e-01 2.67287403e-01
2.72301644e-01 1.99493185e-01 -7.76171207e-01 4.69528347e-01
9.05750215e-01 1.11967301e+00 -6.24080300e-01 9.34642851e-01
-7.16467351e-02 3.83559763e-01 -1.22320600e-01 2.78520614e-01
4.32237923e-01 -1.95701756e-02 5.87038577e-01 1.39454222e+00
3.82955939e-01 -5.37306890e-02 4.19605345e-01 4.97748762e-01
-1.42927706e-01 2.20314682e-01 -7.12024212e-01 -3.50097597e-01
3.31452847e-01 1.09744811e+00 -1.01277435e+00 -3.23333353e-01
-6.23139739e-01 1.50564766e+00 9.32595953e-02 6.74115360e-01
-1.08852446e+00 -4.87031907e-01 5.11413813e-01 -5.17369695e-02
4.35557842e-01 -4.91205364e-01 6.91736341e-01 -1.42849493e+00
2.94607848e-01 -5.45473695e-01 5.55859268e-01 -1.00986469e+00
-9.30738509e-01 3.34446758e-01 1.38098434e-01 -1.47396863e+00
-6.39446616e-01 -1.56490803e-01 -4.84429419e-01 3.88708949e-01
-1.24611020e+00 -1.24341547e+00 -2.68538028e-01 1.01779306e+00
8.66303444e-01 1.46176413e-01 2.94812322e-01 6.02456629e-01
-3.15520138e-01 3.70405674e-01 5.30177727e-02 4.22302276e-01
1.03258669e+00 -9.50058460e-01 -2.58656517e-02 1.09511828e+00
4.72542375e-01 6.31983936e-01 7.75796354e-01 -6.05447590e-01
-1.74167562e+00 -1.09901559e+00 1.00463891e+00 -5.53871036e-01
1.08612466e+00 -4.02331829e-01 -6.12696946e-01 7.89360642e-01
3.29548717e-01 1.54484779e-01 4.00683463e-01 -4.67960238e-01
-2.78931171e-01 -1.48747742e-01 -6.72033966e-01 5.71718216e-01
8.83857667e-01 -1.40481806e+00 -4.46605533e-01 7.67576158e-01
8.35154533e-01 -4.67380643e-01 -9.31645691e-01 -3.27580832e-02
4.83748317e-01 -8.14706504e-01 1.22090316e+00 -5.06166697e-01
6.40589356e-01 -3.97144467e-01 -3.78873527e-01 -7.19355106e-01
2.57343262e-01 -8.85996699e-01 -3.85717511e-01 1.13082254e+00
-1.58397213e-01 1.19520858e-01 6.80526137e-01 3.85271400e-01
-1.02313943e-01 -5.93976974e-01 -1.16564679e+00 -7.09034920e-01
-5.89843631e-01 -5.55909336e-01 1.86511293e-01 1.06358266e+00
2.02084482e-01 9.78453085e-02 -6.42376065e-01 4.03179854e-01
4.23772901e-01 1.98139131e-01 6.83274448e-01 -5.30241787e-01
-8.45636874e-02 -2.01926783e-01 -7.58452952e-01 -1.30131209e+00
4.65365887e-01 -7.25559056e-01 1.68171853e-01 -1.33780205e+00
5.57149291e-01 6.20506585e-01 -5.35995662e-01 1.17501780e-01
2.76232779e-01 2.63855219e-01 1.24616370e-01 4.59621459e-01
-1.50433898e+00 3.50511074e-01 9.45600688e-01 -2.91779935e-01
-3.69044393e-02 -4.97942150e-01 6.69232011e-03 4.75408733e-01
3.07522774e-01 -4.86752629e-01 -6.32360339e-01 -3.73980254e-01
3.35835993e-01 5.09216428e-01 6.76785231e-01 -1.03330016e+00
7.50210822e-01 -2.38852844e-01 2.87881166e-01 -8.58937025e-01
5.67952275e-01 -8.64462912e-01 -3.44328955e-02 1.80491254e-01
-4.89582956e-01 2.60785520e-01 -8.91991332e-02 8.27530086e-01
-6.23293638e-01 -6.82059452e-02 3.67676854e-01 5.44308824e-03
-1.20437920e+00 5.21944463e-01 -2.68188238e-01 6.44300058e-02
1.14272845e+00 -3.50980252e-01 -3.63490969e-01 -8.17553997e-01
-8.21214676e-01 2.12929517e-01 5.24658501e-01 5.44219255e-01
7.80070603e-01 -1.29436421e+00 -2.60093391e-01 -2.34501794e-01
2.03745380e-01 -5.65412521e-01 5.46579957e-01 1.15650225e+00
-4.34395403e-01 6.98522806e-01 -2.10612789e-02 -9.02120411e-01
-1.22185528e+00 9.09526706e-01 1.50003999e-01 4.46005948e-02
-6.45691097e-01 6.27121925e-01 3.12150478e-01 3.37069452e-01
3.85003537e-01 -3.28582048e-01 -1.25210747e-01 3.57301176e-01
5.00455320e-01 1.19209550e-01 -2.51350880e-01 -1.17110884e+00
-3.48567069e-01 6.73690498e-01 -7.62816742e-02 -1.21503443e-01
1.21712387e+00 -5.65742791e-01 -1.58330277e-01 5.98071814e-01
1.86440217e+00 -8.48479122e-02 -1.41448486e+00 -3.78556281e-01
1.05069928e-01 -4.52406913e-01 -2.45846286e-02 -2.69538552e-01
-9.09224212e-01 7.44417250e-01 4.27119344e-01 -6.79744482e-02
1.23271215e+00 1.83681190e-01 1.02865064e+00 6.41550422e-01
2.57099420e-01 -1.16297376e+00 5.63817084e-01 5.60607791e-01
7.74447739e-01 -1.23804080e+00 4.15336192e-02 2.33850390e-01
-4.37719464e-01 1.21686339e+00 2.55404294e-01 1.83944460e-02
6.07555091e-01 -3.65631044e-01 -1.27782449e-01 -3.42750877e-01
-7.71718621e-01 -2.92876810e-02 6.16683841e-01 1.62711188e-01
1.14680208e-01 -5.12859523e-01 -8.01327005e-02 1.60710335e-01
1.10767335e-01 -1.40414000e-01 4.11949128e-01 1.16691482e+00
-2.96153188e-01 -4.81873661e-01 -1.12456903e-01 4.48148362e-02
-6.97263777e-01 -5.58196940e-02 -3.27289999e-01 8.66414428e-01
-2.54296333e-01 8.50784779e-01 2.80047357e-01 -2.74902344e-01
-5.91025539e-02 1.43261686e-01 3.44462901e-01 -4.40305114e-01
-1.94685951e-01 4.28312391e-01 -2.79175729e-01 -1.03559315e+00
-7.98449337e-01 -7.50291049e-01 -1.05066776e+00 -5.68440668e-02
2.42349416e-01 4.29759175e-01 5.69604516e-01 9.49780405e-01
3.66896361e-01 1.72305599e-01 6.85457349e-01 -1.07323778e+00
-1.08458839e-01 -6.93243861e-01 -5.62146127e-01 7.67007947e-01
6.89995587e-01 -3.96700025e-01 -5.64481199e-01 6.09079540e-01]
|
[10.09813117980957, 0.7659195065498352]
|
734f229d-ba00-4876-96c1-45d7cc44a2d1
|
assessor360-multi-sequence-network-for-blind
|
2305.10983
| null |
https://arxiv.org/abs/2305.10983v2
|
https://arxiv.org/pdf/2305.10983v2.pdf
|
Assessor360: Multi-sequence Network for Blind Omnidirectional Image Quality Assessment
|
Blind Omnidirectional Image Quality Assessment (BOIQA) aims to objectively assess the human perceptual quality of omnidirectional images (ODIs) without relying on pristine-quality image information. It is becoming more significant with the increasing advancement of virtual reality (VR) technology. However, the quality assessment of ODIs is severely hampered by the fact that the existing BOIQA pipeline lacks the modeling of the observer's browsing process. To tackle this issue, we propose a novel multi-sequence network for BOIQA called Assessor360, which is derived from the realistic multi-assessor ODI quality assessment procedure. Specifically, we propose a generalized Recursive Probability Sampling (RPS) method for the BOIQA task, combining content and detailed information to generate multiple pseudo viewport sequences from a given starting point. Additionally, we design a Multi-scale Feature Aggregation (MFA) module with Distortion-aware Block (DAB) to fuse distorted and semantic features of each viewport. We also devise TMM to learn the viewport transition in the temporal domain. Extensive experimental results demonstrate that Assessor360 outperforms state-of-the-art methods on multiple OIQA datasets.
|
['Yujiu Yang', 'Yinqiang Zheng', 'Jing Xiao', 'Mingdeng Cao', 'Haoming Cai', 'Shuwei Shi', 'Tianhe Wu']
|
2023-05-18
| null | null | null | null |
['image-quality-assessment']
|
['computer-vision']
|
[-4.70007062e-02 -5.72016120e-01 4.33073580e-01 -5.19986629e-01
-9.45185959e-01 -5.53926945e-01 6.05859935e-01 -3.65533084e-01
-2.66244024e-01 3.47874045e-01 5.78649640e-01 -2.50615567e-01
-4.34081912e-01 -7.13044345e-01 -5.34811795e-01 -4.62176174e-01
2.46762514e-01 9.07440633e-02 1.10526614e-01 -2.84874618e-01
2.59720474e-01 3.65440339e-01 -1.81562436e+00 3.68742317e-01
1.15305662e+00 1.01564300e+00 4.41435575e-01 8.13029408e-01
2.69891739e-01 8.99593771e-01 -4.40699607e-01 -5.82993984e-01
4.31532145e-01 -2.59931743e-01 -6.70157552e-01 9.53995585e-02
8.01502585e-01 -9.10672724e-01 -4.71927047e-01 1.20404410e+00
9.10059690e-01 2.63208479e-01 3.75831068e-01 -1.19983542e+00
-5.50848842e-01 -8.96360055e-02 -2.39329740e-01 4.46555793e-01
6.98526919e-01 4.65907246e-01 9.18933272e-01 -1.13313770e+00
6.46806419e-01 1.29488897e+00 2.75707275e-01 3.28087837e-01
-8.08743834e-01 -4.82739717e-01 9.72343534e-02 7.77682245e-01
-1.13954508e+00 -4.78533834e-01 5.99800587e-01 -3.30740452e-01
5.95088243e-01 3.20760757e-01 8.58771443e-01 1.09654546e+00
1.21890865e-01 9.41977978e-01 1.38607812e+00 1.28842458e-01
3.04475904e-01 -1.15300156e-01 -3.62301916e-01 7.32602596e-01
-1.51020065e-01 4.95827705e-01 -6.70088649e-01 2.52035111e-01
7.79427350e-01 -1.01428367e-01 -4.60117966e-01 -9.04590487e-01
-1.39776862e+00 3.48596096e-01 3.82998347e-01 -1.24702089e-01
-5.39738119e-01 -2.42626041e-01 2.57803172e-01 3.84076566e-01
2.18942419e-01 2.47004554e-01 -3.53658527e-01 -3.33517998e-01
-7.10944057e-01 3.03592324e-01 3.17260742e-01 8.38211715e-01
4.72162575e-01 8.51752535e-02 -3.25644732e-01 9.31005836e-01
3.37837517e-01 6.74392283e-01 2.69999474e-01 -1.37314892e+00
5.64950824e-01 3.40832382e-01 5.15294135e-01 -8.93913209e-01
-3.29066306e-01 -6.29146099e-01 -7.08549201e-01 5.54756284e-01
2.77432412e-01 2.40759984e-01 -9.18549776e-01 1.36762869e+00
4.11918312e-01 -1.23144279e-03 3.77376676e-02 1.46639550e+00
1.13779330e+00 5.88713586e-01 -2.25676641e-01 3.99839953e-02
1.30313802e+00 -1.03505063e+00 -5.78177333e-01 7.19746351e-02
1.25079080e-01 -6.31107271e-01 1.32246971e+00 9.58407223e-01
-1.26013041e+00 -8.76488030e-01 -1.18843651e+00 -8.02698880e-02
-7.28084594e-02 1.80490538e-01 4.91080582e-01 7.33339429e-01
-1.26668227e+00 2.56057501e-01 -4.10019606e-01 -5.21312505e-02
1.75390363e-01 3.18917334e-02 -6.19107306e-01 -7.31239557e-01
-1.10427356e+00 8.50426078e-01 2.32350603e-02 2.91177392e-01
-1.57504690e+00 -7.25863636e-01 -9.35222089e-01 -1.16144620e-01
4.97663558e-01 -1.11031616e+00 1.10867524e+00 -6.52326167e-01
-1.74860871e+00 6.59820795e-01 -1.42446175e-01 -1.62984669e-01
7.34760702e-01 -3.24038684e-01 -7.02609897e-01 4.25429612e-01
6.47525564e-02 5.07878006e-01 9.03225780e-01 -1.61348712e+00
-8.93588006e-01 -4.73337829e-01 5.84323227e-01 8.63772213e-01
7.68456534e-02 2.40242537e-02 -7.28725433e-01 -3.71530503e-01
2.57065624e-01 -6.25741482e-01 -5.15318662e-02 1.62861094e-01
-3.17166597e-01 7.43205175e-02 2.90905625e-01 -1.11811590e+00
1.01068652e+00 -2.01642632e+00 3.00696880e-01 4.15856624e-03
3.64737421e-01 3.05066884e-01 -3.90190154e-01 1.44045264e-01
9.29004326e-02 -4.42600965e-01 -1.16519496e-01 -5.00908852e-01
-1.49574310e-01 6.91367371e-04 1.12621471e-01 4.83228475e-01
-1.34153888e-01 6.55660450e-01 -1.23011363e+00 -3.62816691e-01
7.81047523e-01 4.37646657e-01 -7.84618318e-01 5.91629922e-01
2.24764496e-01 8.51904392e-01 -1.32405274e-02 6.68016136e-01
1.08089280e+00 -1.69422895e-01 -8.56413022e-02 -5.43460727e-01
-1.86495423e-01 3.70521128e-01 -1.49247837e+00 2.20936537e+00
-8.81751537e-01 2.88271815e-01 -1.67487375e-02 -5.37977159e-01
7.28157401e-01 3.18832815e-01 4.82960224e-01 -1.23790169e+00
-6.93486556e-02 1.82868466e-01 -8.89674425e-02 -6.55597627e-01
6.97288752e-01 1.22574262e-01 1.63342267e-01 3.68968695e-02
2.41515949e-01 -2.32469097e-01 6.04806393e-02 2.13541940e-01
9.88329113e-01 2.74666458e-01 2.70438105e-01 1.22461230e-01
8.14350784e-01 -4.16452169e-01 6.69075489e-01 6.25362813e-01
-5.67087710e-01 1.03598952e+00 2.19781194e-02 -4.02386159e-01
-1.18132472e+00 -1.68762457e+00 9.46571305e-03 6.13074064e-01
5.33131599e-01 -2.80584633e-01 -5.25590718e-01 -7.19515920e-01
-3.56234312e-01 6.49841547e-01 -3.49451035e-01 6.20232746e-02
-3.60824227e-01 -4.94738609e-01 -5.03463186e-02 3.14022005e-01
9.21381772e-01 -8.06021094e-01 -5.09349823e-01 1.18688434e-01
-8.62110972e-01 -1.26917362e+00 -5.21978915e-01 -4.78532284e-01
-5.16279459e-01 -9.07282233e-01 -9.72023964e-01 -3.18762481e-01
2.78241128e-01 8.28718185e-01 1.20070994e+00 -3.73279482e-01
-2.11723614e-04 5.93328476e-01 -3.70132685e-01 8.57127905e-02
-2.78221160e-01 -3.34392935e-01 2.74525322e-02 2.37033188e-01
1.49860187e-02 -6.50136590e-01 -1.22697222e+00 7.02277780e-01
-8.73721421e-01 4.37126309e-01 5.89172721e-01 6.58709407e-01
5.04371583e-01 1.05427265e-01 2.60171473e-01 -5.93847372e-02
3.86915207e-01 -3.85972917e-01 -5.13681471e-01 8.85770842e-02
-5.50010145e-01 -1.62747398e-01 5.76950014e-01 -3.44861820e-02
-1.49043119e+00 -3.91251326e-01 -4.38529491e-01 -3.82160842e-01
-1.97718889e-01 2.12557718e-01 -6.81649685e-01 4.66841608e-02
4.61943686e-01 4.75492418e-01 -1.81719467e-01 -4.23615187e-01
5.13234377e-01 6.99194312e-01 7.32048750e-01 -1.36142910e-01
7.77623296e-01 8.15162599e-01 -8.67199674e-02 -5.84839642e-01
-8.76406312e-01 -8.01397145e-01 -2.67794758e-01 -6.37296319e-01
8.73505294e-01 -1.26231539e+00 -9.06829953e-01 6.60830021e-01
-1.03013718e+00 -1.10797863e-03 -1.90431625e-01 7.29646564e-01
-7.20635653e-01 5.21954536e-01 -3.68845969e-01 -5.75485885e-01
-1.22593194e-01 -1.53068841e+00 1.09453773e+00 1.76505819e-01
4.18903142e-01 -5.06164193e-01 8.45849663e-02 8.39306235e-01
2.20005855e-01 -1.49930805e-01 5.36320388e-01 8.84495825e-02
-1.04260898e+00 2.33518437e-01 -6.39172494e-01 8.37438047e-01
1.33373812e-01 -5.21126628e-01 -9.62573528e-01 -3.76559258e-01
-6.53508212e-03 -3.01586121e-01 5.25282979e-01 4.94021922e-01
1.11749637e+00 -9.18661579e-02 3.76229823e-01 9.37427938e-01
1.52917659e+00 2.33983025e-01 8.59150112e-01 5.23676157e-01
9.31938589e-01 6.62173867e-01 8.56512785e-01 5.67632675e-01
1.05013311e+00 1.02852654e+00 8.72721970e-01 -8.91899168e-02
-3.52465183e-01 -3.89251739e-01 3.98792446e-01 9.28229332e-01
-3.60580504e-01 -2.58464336e-01 -6.66535556e-01 7.10236967e-01
-1.39888299e+00 -9.06008363e-01 -5.07036448e-02 2.25014758e+00
4.10552353e-01 -1.82032630e-01 6.95304573e-02 2.05431730e-01
4.23432350e-01 2.94960320e-01 -5.71851194e-01 -1.01205818e-01
-2.07649499e-01 -2.46055722e-01 3.96710783e-01 5.15121222e-01
-8.25053811e-01 5.38097501e-01 5.22119284e+00 7.14655340e-01
-6.83438897e-01 1.69822603e-01 1.62595212e-01 -9.35890079e-02
-5.80290735e-01 -1.41907737e-01 -5.17130435e-01 4.47810888e-01
7.11144984e-01 1.43438280e-01 8.84371340e-01 6.26575291e-01
5.34221113e-01 -2.05890700e-01 -8.46412838e-01 1.39538288e+00
2.14467347e-01 -8.82046342e-01 2.94126958e-01 2.08079472e-01
7.87035286e-01 1.34222314e-01 3.75320196e-01 1.30857900e-01
4.08556849e-01 -8.04698884e-01 9.66475725e-01 8.13442051e-01
8.92943501e-01 -7.73320138e-01 6.27041936e-01 1.21628121e-01
-1.19841540e+00 -3.19846690e-01 -3.38864028e-01 3.93859744e-01
5.58558524e-01 4.93587643e-01 -5.63746929e-01 1.08856809e+00
1.03649986e+00 7.29010344e-01 -7.56230414e-01 1.41042972e+00
-2.59898037e-01 1.87159494e-01 9.95110944e-02 5.38543642e-01
1.65231347e-01 -2.76198298e-01 9.69416261e-01 5.93135536e-01
5.62920511e-01 -5.18350638e-02 -2.65123338e-01 5.04327774e-01
2.27449477e-01 -2.03190316e-02 -5.27104676e-01 5.89259207e-01
2.95496672e-01 1.09604073e+00 -5.04882224e-02 -2.92765141e-01
-4.99586463e-01 1.41148949e+00 -6.33894354e-02 4.17155296e-01
-7.17519760e-01 -7.30555207e-02 8.66882145e-01 -9.84330773e-02
2.92850435e-01 -3.12327385e-01 6.58862814e-02 -1.47216630e+00
2.91543961e-01 -1.26335514e+00 2.74718553e-01 -1.49239969e+00
-9.76368368e-01 6.42992973e-01 -1.06047951e-01 -1.80379522e+00
-1.72334537e-01 -3.72528255e-01 -1.26839027e-01 9.35882211e-01
-1.99303317e+00 -1.18692136e+00 -9.34934258e-01 8.83361578e-01
7.36066639e-01 -7.03347400e-02 4.30926561e-01 6.17446601e-01
-6.83832094e-02 4.49766546e-01 1.27467951e-02 -2.14022920e-01
7.15989888e-01 -1.30860388e+00 6.82206690e-01 1.00521004e+00
-6.47651479e-02 4.01241183e-01 7.49676049e-01 -2.84253269e-01
-1.35349488e+00 -9.58723605e-01 5.06377816e-01 -5.69563806e-01
2.70864010e-01 -1.16682835e-01 -7.08716094e-01 1.41786143e-01
-3.24018523e-02 2.11025968e-01 3.15175027e-01 -1.22066684e-01
-2.68694490e-01 -3.72432530e-01 -1.18515384e+00 6.68042958e-01
1.56191921e+00 -7.10373998e-01 -3.52604777e-01 -2.04448670e-01
6.63846076e-01 -3.90875727e-01 -9.22891438e-01 4.47523326e-01
8.17453325e-01 -1.68891954e+00 1.34550202e+00 -4.83369716e-02
5.22345364e-01 -7.23789573e-01 -4.76507187e-01 -1.69079518e+00
-3.34863126e-01 -2.68776894e-01 -9.32306796e-02 1.01841056e+00
-1.73158005e-01 -4.28797841e-01 4.59320635e-01 -1.85013451e-02
-2.96074003e-01 -2.77913302e-01 -9.97622430e-01 -8.55636179e-01
-4.75051939e-01 -5.16281128e-01 8.45705926e-01 5.37775218e-01
-6.24670804e-01 1.36456102e-01 -7.36395597e-01 5.28808951e-01
1.00859439e+00 -1.49630219e-01 8.99884880e-01 -1.02560043e+00
-5.15644968e-01 -1.62050538e-02 -5.82284451e-01 -1.34259534e+00
-4.11935002e-01 -6.21698618e-01 5.41078448e-02 -1.95576346e+00
9.05273333e-02 -2.88296491e-01 -3.99812758e-01 -4.82925296e-01
-2.45179057e-01 2.42600933e-01 1.57301620e-01 2.18862131e-01
-8.06500852e-01 9.86506701e-01 1.79245245e+00 -6.78755343e-02
-3.53403986e-02 -1.23735685e-02 -4.72774833e-01 5.66624522e-01
4.27231789e-01 -9.65349525e-02 -7.68326759e-01 -6.06220901e-01
5.74066758e-01 5.66950858e-01 7.53713906e-01 -1.31721532e+00
-1.09127155e-02 -4.28677397e-03 2.76545852e-01 -8.90266836e-01
4.67723638e-01 -7.99053788e-01 6.70336559e-02 6.58684298e-02
1.22035025e-02 2.46977374e-01 -2.40445852e-01 6.82692170e-01
-3.51025313e-01 1.04422636e-01 8.64665329e-01 -1.14285864e-01
-1.12550271e+00 5.25422335e-01 -2.16876328e-01 -6.86651617e-02
5.27899623e-01 -2.47280583e-01 -2.56587088e-01 -5.12312651e-01
-4.58823144e-01 1.34088367e-01 5.98477066e-01 6.13638878e-01
1.06472540e+00 -1.41140938e+00 -7.25889623e-01 2.31617272e-01
6.60798013e-01 -3.67245488e-02 9.75661993e-01 6.60502076e-01
-4.60692614e-01 2.61143476e-01 -5.12276113e-01 -6.76337659e-01
-1.15608156e+00 5.34498096e-01 4.67151999e-01 -3.59157950e-01
-7.33032584e-01 6.59871101e-01 4.65156645e-01 -6.75705552e-01
2.03524940e-02 -1.67349502e-01 -5.83118737e-01 -2.11017102e-01
7.68576920e-01 7.20770061e-01 2.32601121e-01 -7.08113968e-01
-1.92627802e-01 4.22764570e-01 3.25787365e-02 -5.89742303e-01
1.14100003e+00 -8.53497624e-01 2.34037355e-01 2.43405476e-01
1.07856798e+00 -1.81413349e-02 -1.68041241e+00 -2.62929559e-01
-6.21219993e-01 -8.99754584e-01 2.91389823e-01 -1.13519120e+00
-9.13793266e-01 8.59201491e-01 1.14481795e+00 -2.15030864e-01
1.33488429e+00 -3.85042131e-01 8.49520981e-01 1.20510019e-01
8.42495263e-01 -9.71371412e-01 2.99739063e-01 2.03332290e-01
1.01837122e+00 -1.44071341e+00 -2.25647464e-01 -1.85680345e-01
-8.09081912e-01 6.25606537e-01 6.20960891e-01 2.12523907e-01
3.78704399e-01 -3.62145901e-01 2.80079871e-01 -1.82992235e-01
-5.02290010e-01 -2.69287348e-01 5.59502184e-01 9.95564938e-01
-1.95858434e-01 -1.92591529e-02 3.71795371e-02 2.81274617e-01
-2.46492207e-01 8.57358798e-02 6.54088795e-01 6.03370965e-01
-1.24848276e-01 -7.19454706e-01 -2.46707603e-01 1.93689451e-01
-2.36033395e-01 -2.58260369e-01 4.96455550e-01 2.51718372e-01
1.97594106e-01 1.21427774e+00 -2.43373215e-01 -6.71526909e-01
5.86324334e-01 -5.94368875e-01 5.62604606e-01 -2.34546199e-01
-2.54763037e-01 -2.65768528e-01 2.36100823e-01 -1.08148170e+00
-6.03729486e-01 -7.23454237e-01 -4.62613314e-01 -3.02112043e-01
1.40678048e-01 -1.65780962e-01 1.03159857e+00 7.70754874e-01
2.36824572e-01 7.71919608e-01 8.47027957e-01 -8.13584805e-01
-4.34112459e-01 -7.80451119e-01 -6.50786638e-01 4.02764469e-01
6.05131269e-01 -8.03210855e-01 -4.34571236e-01 -1.33594185e-01]
|
[11.810975074768066, -1.9013748168945312]
|
519e9d7e-7451-4b4b-bf5b-ab1ca6195422
|
test-positive-at-w-nut-2020-shared-task-3
|
2009.14262
| null |
https://arxiv.org/abs/2009.14262v1
|
https://arxiv.org/pdf/2009.14262v1.pdf
|
TEST_POSITIVE at W-NUT 2020 Shared Task-3: Joint Event Multi-task Learning for Slot Filling in Noisy Text
|
The competition of extracting COVID-19 events from Twitter is to develop systems that can automatically extract related events from tweets. The built system should identify different pre-defined slots for each event, in order to answer important questions (e.g., Who is tested positive? What is the age of the person? Where is he/she?). To tackle these challenges, we propose the Joint Event Multi-task Learning (JOELIN) model. Through a unified global learning framework, we make use of all the training data across different events to learn and fine-tune the language model. Moreover, we implement a type-aware post-processing procedure using named entity recognition (NER) to further filter the predictions. JOELIN outperforms the BERT baseline by 17.2% in micro F1.
|
['Chieh-Yang Huang', 'Yaqi Hou', 'Yang Shi', 'Jiaqi Wang', 'Enyan Dai', 'Chacha Chen']
|
2020-09-29
| null | null | null | null |
['extracting-covid-19-events-from-twitter']
|
['natural-language-processing']
|
[-3.44452858e-02 -9.51137319e-02 -1.79236457e-01 -6.07772529e-01
-1.10899103e+00 -4.70800608e-01 6.77391827e-01 6.64580941e-01
-9.74343598e-01 8.18702579e-01 3.80033851e-01 -3.57404910e-02
1.86628729e-01 -1.10697877e+00 -6.83704078e-01 -3.31660479e-01
2.50340819e-01 6.43059790e-01 3.60129744e-01 -2.11138785e-01
-1.16669357e-01 2.97140807e-01 -1.41852129e+00 4.96034116e-01
5.67529023e-01 7.69570231e-01 2.34019496e-02 5.78523636e-01
-3.88125986e-01 8.60442340e-01 -7.40836561e-01 -6.26777351e-01
-3.09000194e-01 -1.59328401e-01 -9.17002916e-01 -3.09828401e-01
4.46820743e-02 -5.91858663e-03 1.75574988e-01 7.38139510e-01
3.74736279e-01 2.17005745e-01 7.36320496e-01 -1.08569527e+00
1.04807608e-01 9.15732324e-01 -4.71521497e-01 5.03835797e-01
1.98505893e-01 -3.45636606e-01 1.07326984e+00 -8.63429666e-01
6.28273845e-01 1.09309721e+00 7.28386283e-01 3.04661661e-01
-1.00718939e+00 -9.18434501e-01 1.88729346e-01 1.21716745e-01
-1.53019357e+00 -3.58154804e-01 4.23709810e-01 -5.28603315e-01
7.84506500e-01 4.08421218e-01 7.92886037e-03 1.42004395e+00
2.05431268e-01 6.39742136e-01 8.04960251e-01 -3.63750070e-01
4.03121673e-02 4.16021079e-01 3.59056830e-01 1.70871288e-01
1.46711916e-01 -4.38286215e-01 -5.31687260e-01 -3.40027511e-01
1.09339699e-01 4.52118963e-02 3.56126368e-01 4.75394934e-01
-9.43019509e-01 8.64394307e-01 -3.82266603e-02 4.68949497e-01
-5.33295214e-01 -1.59183413e-01 6.50935650e-01 -4.89254408e-02
6.73900068e-01 3.91447961e-01 -9.55902278e-01 -7.35760704e-02
-9.23372030e-01 4.64100033e-01 1.20549667e+00 6.37277663e-01
1.12069559e+00 -4.07697797e-01 -5.70507169e-01 7.76490867e-01
3.66048872e-01 4.46835279e-01 3.28186661e-01 -3.63303006e-01
6.74137831e-01 6.54156089e-01 2.23355711e-01 -7.81556964e-01
-8.37697744e-01 -4.28842038e-01 -6.31914616e-01 -6.06012821e-01
4.03527021e-01 -8.83134842e-01 -5.78132153e-01 1.70277536e+00
6.34478033e-01 5.67517042e-01 -6.97622970e-02 3.66883725e-01
1.05576539e+00 8.54999304e-01 9.13172245e-01 -1.79741472e-01
1.81100917e+00 -5.14846265e-01 -7.12107718e-01 -6.06048465e-01
7.47571051e-01 -8.11099589e-01 5.84625781e-01 -1.31827652e-01
-6.87603354e-01 -6.07821882e-01 -6.11533165e-01 2.84264125e-02
-7.25409091e-01 2.80478299e-01 3.57836396e-01 6.11552656e-01
-4.00432646e-01 5.10614067e-02 -6.97037399e-01 -4.25548136e-01
-1.70723107e-02 2.40238711e-01 -9.80373248e-02 4.33496118e-01
-1.80547357e+00 8.66759419e-01 7.43309915e-01 -1.85629934e-01
-5.53367376e-01 -9.15700018e-01 -9.25783217e-01 1.14635609e-01
3.72366488e-01 -6.15380824e-01 1.32566941e+00 -7.02313721e-01
-1.11307895e+00 1.13636398e+00 -5.04522264e-01 -6.62911952e-01
2.88817763e-01 -3.80428255e-01 -9.02252555e-01 -1.08519539e-01
4.85297978e-01 2.81825483e-01 5.76676667e-01 -7.87475109e-01
-1.36112714e+00 -1.95083201e-01 -2.01633766e-01 -6.94241896e-02
-3.21014851e-01 7.22306907e-01 -6.17225587e-01 -5.33822000e-01
-7.37996697e-01 -8.77428234e-01 -3.87087137e-01 -9.06509578e-01
-3.89278919e-01 -6.37094676e-01 6.82565451e-01 -8.71832967e-01
1.67855227e+00 -2.11973691e+00 -4.87257957e-01 2.67002434e-01
8.36816505e-02 3.62664849e-01 9.43134427e-02 4.49097902e-01
1.83890998e-01 9.16231051e-02 3.65134984e-01 -5.09291530e-01
-1.63353533e-02 -1.58474618e-03 -2.20147684e-01 1.00763299e-01
4.39078331e-01 7.64807224e-01 -7.03432739e-01 -7.18734205e-01
-1.54448412e-02 5.16648591e-01 -3.89914095e-01 3.04217666e-01
-2.48983130e-01 4.25809145e-01 -6.33653522e-01 1.78264990e-01
5.74647486e-01 -2.57491320e-01 1.50533617e-01 -1.19572416e-01
-3.12848598e-01 6.03629053e-01 -1.42507374e+00 1.00762141e+00
-6.47943676e-01 3.34387332e-01 -5.16322441e-02 -8.79507363e-01
8.93990040e-01 4.47475195e-01 4.74942386e-01 -4.81017649e-01
2.62254447e-01 1.34813458e-01 -4.35553432e-01 -6.27491653e-01
5.97208738e-01 -8.21497440e-02 -6.71488047e-01 3.10302824e-01
1.31202772e-01 5.93271732e-01 4.12342131e-01 -8.02992955e-02
9.60639656e-01 -1.17111921e-01 5.35190642e-01 -8.33890773e-03
9.49469268e-01 -1.15217149e-01 8.99603307e-01 8.42383385e-01
-1.49846733e-01 2.81546146e-01 3.18323642e-01 -5.82905650e-01
-7.18457043e-01 -8.42281282e-01 -1.67106345e-01 1.72082937e+00
-1.53958306e-01 -5.82109809e-01 -6.21245801e-01 -7.72743583e-01
-2.02315435e-01 8.30003917e-01 -5.97279787e-01 1.54419839e-01
-7.65936553e-01 -1.12478185e+00 7.67469645e-01 3.16923171e-01
3.90828639e-01 -1.08712447e+00 -3.00983399e-01 6.00040257e-01
-6.47258461e-01 -1.44946253e+00 -3.20606083e-01 1.72856688e-01
-1.52961925e-01 -8.27902734e-01 -3.03513229e-01 -7.24961817e-01
5.82944490e-02 -4.56528664e-01 1.34219480e+00 -2.84300327e-01
4.22433317e-02 1.43355411e-02 -4.59123731e-01 -7.33502507e-01
-3.94362748e-01 6.84193969e-01 -2.67538786e-01 4.28213418e-01
1.03185654e+00 -3.26049656e-01 -2.62039661e-01 2.69347459e-01
-6.74559593e-01 -8.54045600e-02 5.16154706e-01 3.61023813e-01
6.04869783e-01 1.16951698e-02 8.15377772e-01 -1.26835346e+00
4.34982240e-01 -9.13961649e-01 -3.46216261e-01 4.90200669e-01
-1.93304956e-01 4.05799076e-02 4.73620623e-01 -5.05998790e-01
-1.26985967e+00 1.04769155e-01 -6.30320907e-01 3.48396450e-01
-5.03964305e-01 6.32300854e-01 -2.31992468e-01 5.77579319e-01
5.84975898e-01 1.08361403e-02 -7.64883816e-01 -4.95993167e-01
5.91207147e-02 8.37986052e-01 3.83586407e-01 -3.73688579e-01
7.26348639e-01 2.86204457e-01 -3.38995278e-01 -8.87408972e-01
-1.55318987e+00 -9.52331722e-01 -5.78970909e-01 -3.63078117e-02
1.28775549e+00 -1.33672023e+00 -6.10958159e-01 3.54703367e-01
-1.30922723e+00 -2.51170784e-01 8.91891196e-02 5.57579637e-01
-5.08938320e-02 -2.57190317e-01 -6.26926780e-01 -8.40971947e-01
-6.18935168e-01 -7.35705495e-01 1.13067663e+00 4.70615953e-01
-4.98296052e-01 -1.25992095e+00 2.47787789e-01 3.78783524e-01
2.77564138e-01 2.82465994e-01 4.98590171e-01 -1.40123320e+00
-7.88478032e-02 -1.76923111e-01 -2.37747237e-01 -9.66824591e-03
-1.40565947e-01 -5.09099625e-02 -8.55711222e-01 1.41325682e-01
-2.15983421e-01 -2.16513693e-01 8.19429398e-01 3.20252955e-01
1.04500306e+00 -2.89833128e-01 -6.93207741e-01 4.21656936e-01
1.13956892e+00 -2.17480913e-01 4.53288317e-01 5.92721939e-01
6.27892077e-01 5.68186700e-01 6.13585174e-01 7.54244328e-01
1.05670822e+00 6.78997099e-01 -1.48476407e-01 -6.05820417e-02
1.77890152e-01 -4.51943994e-01 4.56972957e-01 4.34514642e-01
1.64266095e-01 -2.61382282e-01 -1.00058579e+00 5.89110017e-01
-1.77057862e+00 -1.07859826e+00 -5.23413479e-01 1.86353815e+00
1.11111939e+00 2.61083543e-01 1.95036978e-01 -1.73294246e-01
8.70180845e-01 2.06699163e-01 -6.10314533e-02 -3.01373482e-01
-4.87327501e-02 4.25586373e-01 6.51107430e-01 3.96546125e-01
-1.71056712e+00 1.05010474e+00 5.49146318e+00 9.90665197e-01
-1.12242198e+00 3.71912271e-01 6.23297215e-01 3.16753566e-01
-8.44359472e-02 5.55095845e-04 -1.85581768e+00 5.34828603e-01
1.57351518e+00 -2.87823528e-01 -1.76035553e-01 5.44899106e-01
4.11999553e-01 5.25279045e-02 -6.97399259e-01 7.79182971e-01
1.32959977e-01 -1.19836688e+00 -1.71283275e-01 -1.81624547e-01
6.20492160e-01 2.77414382e-01 -4.00304735e-01 8.70369494e-01
4.71347183e-01 -5.60002506e-01 6.68785512e-01 7.46803761e-01
4.26740199e-01 -8.38641524e-01 7.99923480e-01 6.71766758e-01
-1.45059180e+00 6.19633235e-02 -9.27987974e-03 -7.16558471e-02
3.24616671e-01 9.05696332e-01 -1.15571976e+00 5.85990131e-01
7.34664559e-01 4.65959191e-01 -7.45719016e-01 1.05998600e+00
-4.59469676e-01 9.53500271e-01 -4.86752212e-01 -1.48562789e-01
1.29158884e-01 2.81347781e-01 4.13449943e-01 1.78607571e+00
3.26345891e-01 -1.38022929e-01 4.37986583e-01 5.89062810e-01
-3.48935694e-01 4.18765515e-01 -8.23865607e-02 1.42586172e-01
5.63786447e-01 1.56622124e+00 -5.84989488e-01 -4.59219366e-01
-4.94348705e-01 7.50091434e-01 3.94536763e-01 2.55224049e-01
-1.00537574e+00 -5.46460211e-01 5.90176165e-01 2.02225298e-01
4.56408173e-01 1.06872275e-01 -5.81028126e-02 -1.15153611e+00
-2.83849448e-01 -4.38004375e-01 8.72792006e-01 -2.83003330e-01
-1.31670892e+00 5.81013203e-01 -7.27892518e-02 -8.46301675e-01
-5.01164079e-01 -3.02873373e-01 -8.56682003e-01 8.41122270e-01
-1.66566825e+00 -1.30141616e+00 -2.74638180e-03 4.73640382e-01
3.21810216e-01 6.70767799e-02 8.73190701e-01 7.67755032e-01
-7.18598783e-01 6.89905286e-01 -2.20009342e-01 6.25020802e-01
1.12277114e+00 -1.11324453e+00 2.85545200e-01 6.59531116e-01
1.12371348e-01 2.94302493e-01 7.86442876e-01 -5.87712646e-01
-7.75793314e-01 -1.67343688e+00 1.94651365e+00 -6.66972399e-01
8.24589372e-01 -5.43088377e-01 -8.52465987e-01 8.34513187e-01
1.62965693e-02 -1.52122334e-01 1.14923584e+00 6.26263201e-01
-3.24061066e-01 -1.71933815e-01 -8.82967353e-01 1.99088529e-01
4.24615353e-01 -3.96158993e-01 -6.49737537e-01 3.94978285e-01
7.94691026e-01 -3.01874131e-01 -8.74262691e-01 3.31219465e-01
2.19635904e-01 -5.65210700e-01 8.64837646e-01 -6.32372499e-01
-3.64672556e-03 -1.44962102e-01 1.72687873e-01 -1.07894516e+00
-2.70629406e-01 -4.55459207e-01 1.35327980e-01 1.93663228e+00
9.43835199e-01 -5.93128979e-01 4.00417656e-01 6.10618174e-01
2.91810930e-01 -2.51878440e-01 -6.90438747e-01 -3.34644258e-01
-9.39424336e-02 -7.62514114e-01 6.07395709e-01 9.05745924e-01
-2.17315748e-01 7.89312363e-01 -5.05046964e-01 6.00892544e-01
2.50351757e-01 1.80347264e-02 8.09108794e-01 -1.53184545e+00
-1.04776606e-01 -9.97346938e-02 6.38234778e-04 -5.33068120e-01
5.26947737e-01 -7.75031507e-01 -9.85181332e-02 -1.34926617e+00
2.48143584e-01 -4.05408978e-01 -3.48209292e-01 6.04144871e-01
-6.32208467e-01 1.32283390e-01 4.13715653e-02 -2.60341614e-02
-1.10560310e+00 1.25639156e-01 4.74750131e-01 1.84041157e-01
-3.96670908e-01 4.82696950e-01 -6.34405792e-01 5.89029968e-01
9.08639431e-01 -7.92954266e-01 3.04543108e-01 -2.36926898e-01
7.17121422e-01 -8.86726603e-02 1.29146159e-01 -8.35897386e-01
3.44261527e-01 -2.16185138e-01 2.86470950e-01 -6.93036973e-01
-7.74676865e-03 -6.06743038e-01 1.14749201e-01 1.39490381e-01
-3.99429321e-01 -6.64312690e-02 1.94203183e-01 4.15390402e-01
-4.26089019e-01 -2.61595756e-01 5.08321643e-01 9.36984569e-02
-5.98717391e-01 3.92625362e-01 -5.79615295e-01 4.03753072e-01
1.07283628e+00 5.01371026e-01 -2.02659816e-01 -9.94417369e-02
-7.44008780e-01 5.41219354e-01 -1.79012239e-01 5.04552245e-01
-8.50364417e-02 -1.12277436e+00 -1.11289954e+00 6.18017390e-02
2.90605694e-01 -2.05718294e-01 3.10741097e-01 7.90102065e-01
-1.22266114e-01 2.96855301e-01 1.60218298e-01 -1.51112571e-01
-1.18181777e+00 2.12996349e-01 1.77068606e-01 -9.71721709e-01
-5.34764268e-02 8.62999439e-01 -5.69613241e-02 -6.40037358e-01
7.40904659e-02 -1.38173729e-01 -7.96029925e-01 6.85135245e-01
8.53752911e-01 1.95643604e-01 2.16616273e-01 -7.78564334e-01
-5.55948496e-01 7.78451748e-03 -9.05645639e-02 -6.25540912e-02
1.43786097e+00 -1.12209544e-01 -1.09164953e-01 5.99309921e-01
1.25560737e+00 3.43502909e-01 -8.41377079e-01 -4.48977143e-01
5.11631370e-01 1.77464724e-01 -3.40293087e-02 -6.39707685e-01
-6.32347882e-01 5.21400809e-01 1.80586502e-01 4.06710029e-01
8.75397265e-01 2.31942207e-01 9.19914126e-01 1.67187050e-01
-1.88015327e-02 -1.23820424e+00 -4.58344221e-01 1.09045875e+00
4.03176010e-01 -1.26196265e+00 -2.71091133e-01 -2.63047785e-01
-7.06595778e-01 9.55257297e-01 4.88099426e-01 1.14071220e-01
9.98541653e-01 3.07728797e-01 5.98489381e-02 -1.11434288e-01
-8.19360435e-01 -6.46230817e-01 4.94967103e-01 1.97755650e-01
5.39885640e-01 1.85578644e-01 -4.59975809e-01 1.15309882e+00
-3.68250817e-01 -1.18806437e-01 4.51727696e-02 5.48390567e-01
-5.27318120e-01 -1.30815148e+00 -4.42523897e-01 5.09865761e-01
-1.00139427e+00 9.32115167e-02 -1.73998877e-01 4.48236823e-01
4.88932908e-01 1.13102400e+00 1.31044522e-01 -3.43584180e-01
5.00298083e-01 3.94596517e-01 -2.54184872e-01 -8.06913614e-01
-1.04247415e+00 3.08989361e-02 4.86117423e-01 -1.33265138e-01
-6.14537060e-01 -9.22107100e-01 -1.29712820e+00 2.68712249e-02
-1.23053432e-01 3.78417462e-01 4.44446504e-01 1.34878445e+00
3.75697941e-01 4.36085373e-01 6.25826240e-01 -4.41745430e-01
-1.38153121e-01 -1.01804018e+00 -2.31160015e-01 4.32844520e-01
2.49119610e-01 -3.38273197e-01 -2.96175838e-01 1.85807839e-01]
|
[9.54062271118164, 9.478062629699707]
|
b08d851d-fcab-4f33-8317-466e07d15434
|
professor-forcing-a-new-algorithm-for
|
1610.09038
| null |
http://arxiv.org/abs/1610.09038v1
|
http://arxiv.org/pdf/1610.09038v1.pdf
|
Professor Forcing: A New Algorithm for Training Recurrent Networks
|
The Teacher Forcing algorithm trains recurrent networks by supplying observed
sequence values as inputs during training and using the network's own
one-step-ahead predictions to do multi-step sampling. We introduce the
Professor Forcing algorithm, which uses adversarial domain adaptation to
encourage the dynamics of the recurrent network to be the same when training
the network and when sampling from the network over multiple time steps. We
apply Professor Forcing to language modeling, vocal synthesis on raw waveforms,
handwriting generation, and image generation. Empirically we find that
Professor Forcing acts as a regularizer, improving test likelihood on character
level Penn Treebank and sequential MNIST. We also find that the model
qualitatively improves samples, especially when sampling for a large number of
time steps. This is supported by human evaluation of sample quality. Trade-offs
between Professor Forcing and Scheduled Sampling are discussed. We produce
T-SNEs showing that Professor Forcing successfully makes the dynamics of the
network during training and sampling more similar.
|
['Alex Lamb', 'Saizheng Zhang', 'Anirudh Goyal', 'Yoshua Bengio', 'Ying Zhang', 'Aaron Courville']
|
2016-10-27
|
professor-forcing-a-new-algorithm-for-1
|
http://papers.nips.cc/paper/6099-professor-forcing-a-new-algorithm-for-training-recurrent-networks
|
http://papers.nips.cc/paper/6099-professor-forcing-a-new-algorithm-for-training-recurrent-networks.pdf
|
neurips-2016-12
|
['handwriting-generation']
|
['computer-vision']
|
[ 6.03775203e-01 7.21830487e-01 -3.99819911e-01 -3.80239516e-01
-9.71972764e-01 -9.86369193e-01 8.38680804e-01 -5.63655078e-01
-2.90592045e-01 9.91757870e-01 3.99182826e-01 -4.86729026e-01
3.08287501e-01 -7.72413969e-01 -9.43234026e-01 -4.24193621e-01
-4.79867831e-02 5.82899809e-01 1.33540086e-03 -1.60368681e-01
8.05729032e-02 5.48938632e-01 -1.03547454e+00 4.70678180e-01
5.87174356e-01 2.66520381e-01 2.17956543e-01 1.17187393e+00
2.12565958e-02 1.06099153e+00 -1.08189476e+00 -3.17347586e-01
2.49970078e-01 -1.00615573e+00 -7.59120166e-01 1.23462625e-01
3.18657964e-01 -4.73669708e-01 -4.43032175e-01 1.02100551e+00
4.56237257e-01 3.36183608e-01 7.42530763e-01 -8.23615313e-01
-8.06294858e-01 1.31223142e+00 8.22764705e-04 2.94918805e-01
2.29397789e-01 5.62368393e-01 1.00712097e+00 -5.60437024e-01
8.21348369e-01 1.42837286e+00 5.99108458e-01 9.65542614e-01
-1.37695849e+00 -8.32983851e-01 1.57360211e-01 -4.35470670e-01
-8.97496641e-01 -5.84240913e-01 7.96344221e-01 -2.89754719e-01
8.83469403e-01 8.52536559e-02 5.57087600e-01 1.48666918e+00
2.14246809e-01 8.93234432e-01 7.88989127e-01 -6.51767135e-01
1.98886439e-01 2.00655237e-01 -2.13854253e-01 5.91994286e-01
-3.84526491e-01 5.10383725e-01 -4.09587532e-01 -2.97177821e-01
1.19211245e+00 -4.89855826e-01 -1.96911290e-01 3.45570505e-01
-1.18132603e+00 8.52401435e-01 -1.32995902e-03 1.51708990e-01
-4.01938617e-01 5.18357396e-01 3.30423385e-01 7.05797017e-01
4.41983134e-01 8.95942450e-01 -4.70041245e-01 -1.76148653e-01
-1.05296195e+00 4.44035709e-01 7.19016790e-01 9.37804401e-01
3.34033608e-01 7.84664035e-01 -3.14799488e-01 9.13556516e-01
1.14085317e-01 5.92713356e-01 8.28334451e-01 -1.45720696e+00
2.46955708e-01 -2.37098649e-01 9.72838253e-02 -3.20341587e-01
8.37579295e-02 -4.92711872e-01 -4.76806253e-01 3.71578813e-01
6.63516700e-01 -5.56607425e-01 -1.12572825e+00 2.20824480e+00
-1.84930980e-01 4.88039941e-01 1.79425508e-01 3.91124338e-01
2.65095145e-01 9.56935585e-01 -1.81669682e-01 -3.73042703e-01
7.03167796e-01 -8.25529456e-01 -7.51878083e-01 -2.12368205e-01
5.27748048e-01 -6.33623421e-01 1.26226091e+00 5.59947014e-01
-1.60753560e+00 -7.52690732e-01 -9.32731688e-01 3.95384252e-01
1.69630706e-01 1.30835518e-01 2.33215630e-01 5.47906399e-01
-1.15524292e+00 1.11044109e+00 -8.72994125e-01 1.70436546e-01
2.10035369e-01 2.75970101e-01 1.07793711e-01 2.53364444e-01
-1.30010414e+00 8.77498984e-01 1.28860429e-01 -2.86904931e-01
-1.43953371e+00 -5.86387336e-01 -7.33189821e-01 -8.23134407e-02
-1.59781635e-01 -5.24557948e-01 1.97116148e+00 -1.26335716e+00
-2.10793138e+00 4.88393456e-01 -3.59593093e-01 -9.15435791e-01
5.39892852e-01 -2.00685516e-01 -4.73208666e-01 2.47564875e-02
-1.17302060e-01 8.60680282e-01 1.14357591e+00 -9.91006255e-01
-2.34302938e-01 3.02501589e-01 -6.28143176e-02 8.21250081e-02
-1.20201223e-02 -1.76535815e-01 2.62938924e-02 -9.01265860e-01
-3.65309834e-01 -9.04286325e-01 -3.39276701e-01 -3.91457260e-01
-6.06401086e-01 -3.52122605e-01 5.67604840e-01 -5.00488698e-01
1.08366108e+00 -1.99204087e+00 -1.74622554e-02 3.26621830e-01
-2.08899200e-01 2.21509084e-01 -6.60730422e-01 2.25242019e-01
-4.34292108e-01 3.20116162e-01 -7.69235939e-02 -2.89948523e-01
-9.93741229e-02 3.02357167e-01 -8.43740880e-01 2.93145865e-01
4.40385044e-01 1.00140858e+00 -9.02330220e-01 -9.76332426e-02
-1.41267955e-01 1.70212269e-01 -7.80989170e-01 4.82191116e-01
-7.87326396e-01 3.52582961e-01 -9.26299021e-02 1.76781610e-01
3.15214484e-03 -1.53753504e-01 1.46273613e-01 4.83050913e-01
1.28117114e-01 8.76732945e-01 -7.82633901e-01 1.50220764e+00
-5.07475615e-01 8.51005256e-01 -2.23430380e-01 -6.38076842e-01
1.03140903e+00 6.92496836e-01 -3.85354936e-01 -4.04220700e-01
5.31216897e-02 2.69505292e-01 3.71633470e-01 -2.76157260e-01
3.10471207e-01 -5.31358838e-01 1.99801028e-01 1.05162644e+00
5.41914403e-02 -6.14784718e-01 9.31353047e-02 2.28376165e-01
1.06020081e+00 1.65699333e-01 -1.15050972e-01 -8.90256464e-02
1.35449558e-01 -5.73758893e-02 4.08207268e-01 1.01579475e+00
1.48217708e-01 5.92877865e-01 4.80411500e-01 -2.47106552e-01
-1.66174209e+00 -1.18564355e+00 2.18367547e-01 1.16081762e+00
-7.26541638e-01 1.28602386e-01 -8.71392190e-01 -5.38464546e-01
-1.83489010e-01 1.29689741e+00 -6.98385835e-01 -2.25944012e-01
-1.01019228e+00 -1.22200266e-01 9.88098860e-01 5.71941555e-01
-6.88885003e-02 -1.65305555e+00 -2.64374167e-01 4.97097731e-01
1.63811997e-01 -7.81297743e-01 -8.20756733e-01 3.33938062e-01
-1.04987228e+00 -5.14033973e-01 -8.80545378e-01 -8.75474095e-01
6.27769053e-01 -4.04863179e-01 1.25045586e+00 8.51291865e-02
1.88741922e-01 -2.16015149e-02 3.49707529e-02 -4.31566626e-01
-1.34350550e+00 2.47687981e-01 3.01793754e-01 -5.46465278e-01
-2.14099959e-01 -8.38085353e-01 -5.40671758e-02 1.27300426e-01
-8.79412472e-01 -3.48308496e-02 4.61130053e-01 1.08453524e+00
2.20431894e-01 -2.13542745e-01 9.57579017e-01 -1.11886120e+00
1.06025505e+00 -2.99330890e-01 -5.00921249e-01 2.89341118e-02
-3.94091725e-01 4.01890785e-01 1.01419055e+00 -1.07690382e+00
-9.70445395e-01 -2.99266607e-01 -2.76263952e-01 -6.73540592e-01
1.13139106e-02 1.60748243e-01 2.67974019e-01 4.66394991e-01
1.08947837e+00 4.65270013e-01 2.43364900e-01 -1.50939062e-01
4.95396495e-01 3.85307312e-01 7.76827633e-01 -7.94181108e-01
9.50950861e-01 4.27406356e-02 -5.71050942e-01 -5.85319579e-01
-8.02386284e-01 3.55183631e-01 -2.99347550e-01 -1.60721183e-01
4.59055960e-01 -7.95638144e-01 -4.46109116e-01 3.19845051e-01
-1.32838893e+00 -1.20939732e+00 -9.02840555e-01 3.88432533e-01
-8.31314564e-01 -8.95007253e-02 -8.58790934e-01 -1.03425813e+00
-3.25307339e-01 -9.23391581e-01 5.60003519e-01 1.35721475e-01
-7.37725616e-01 -1.08930361e+00 3.10519010e-01 -3.34423184e-01
3.31134588e-01 4.03799675e-02 9.86642122e-01 -8.41442645e-01
-4.75543708e-01 2.83883855e-04 4.40106034e-01 6.81055307e-01
2.94467639e-02 2.70802438e-01 -1.03835022e+00 -1.61641538e-01
-2.30950937e-02 -5.11139810e-01 5.19766092e-01 5.66286683e-01
9.14328575e-01 -5.64990461e-01 7.10634813e-02 3.95690382e-01
8.18598390e-01 4.76165384e-01 6.80329025e-01 5.57179600e-02
2.39231303e-01 5.11364937e-01 3.81251216e-01 2.39602074e-01
-3.13244104e-01 1.39040172e-01 1.15087181e-01 1.30641311e-01
-1.72619134e-01 -6.33039474e-01 8.68821442e-01 1.01114869e+00
2.93511748e-01 -3.20423096e-01 -6.77988172e-01 8.21994603e-01
-1.36337066e+00 -1.35200250e+00 4.38270181e-01 1.90757072e+00
1.46297264e+00 7.55280197e-01 2.18884930e-01 9.62317139e-02
7.01489627e-01 1.63015470e-01 -7.89189935e-01 -8.13193977e-01
-4.16358225e-02 5.98523736e-01 3.70538384e-01 9.94843483e-01
-3.87404531e-01 1.02503490e+00 8.09064770e+00 8.33769858e-01
-1.13610792e+00 -5.23406863e-02 8.95638704e-01 -3.63244712e-01
-7.97138751e-01 -2.78965622e-01 -8.14096451e-01 4.17016327e-01
1.52110088e+00 -3.72941554e-01 8.20293009e-01 7.69501209e-01
3.92631650e-01 3.38278234e-01 -1.35919285e+00 2.70448744e-01
-2.74408489e-01 -1.77483130e+00 6.46402761e-02 -1.43426046e-01
9.65113521e-01 2.21574172e-01 2.62761027e-01 4.61209416e-01
1.14271820e+00 -1.33832788e+00 8.72479916e-01 3.66933256e-01
1.00644588e+00 -9.55709815e-01 2.05864072e-01 7.64942706e-01
-6.20720327e-01 2.54568271e-02 -3.20445925e-01 -1.37050211e-01
1.70062035e-01 4.23329860e-01 -1.54211950e+00 -3.34396631e-01
-9.54339728e-02 2.86891878e-01 3.77235487e-02 3.57884765e-01
-5.07405221e-01 1.16941130e+00 -2.85748065e-01 -1.02344848e-01
3.54008436e-01 -5.92930876e-02 5.91939330e-01 1.38714504e+00
2.08208680e-01 -2.69275103e-02 3.41613069e-02 1.09967375e+00
-3.47859949e-01 -4.41572279e-01 -9.58071291e-01 -3.00737709e-01
8.84564102e-01 6.71684265e-01 -3.27180773e-01 -5.85809350e-01
2.80387491e-01 8.18807721e-01 2.10646436e-01 5.18447340e-01
-7.47867882e-01 -3.97512853e-01 6.43765092e-01 1.61793977e-01
2.11971089e-01 -2.48511389e-01 -3.69691819e-01 -7.61489570e-01
-4.72628087e-01 -1.15739489e+00 -2.39352956e-01 -9.99220908e-01
-9.77567554e-01 6.74153507e-01 -1.13385424e-01 -8.97972763e-01
-1.23512220e+00 -2.45343059e-01 -9.37240362e-01 1.21559155e+00
-1.19611192e+00 -4.82097119e-01 6.09220862e-01 3.18010986e-01
1.03090084e+00 -5.16323745e-01 9.30859149e-01 -2.76596963e-01
-3.33006680e-01 8.89717638e-01 -5.51213548e-02 3.93396437e-01
4.44335967e-01 -1.25928521e+00 1.06899989e+00 8.77635002e-01
3.17462564e-01 8.56537163e-01 1.02270937e+00 -6.28204525e-01
-8.26329947e-01 -1.10422802e+00 7.08199620e-01 -4.09080625e-01
7.56458700e-01 -2.91633964e-01 -9.66146708e-01 9.62319314e-01
4.10157204e-01 -2.22501859e-01 4.06274617e-01 -1.90310836e-01
-3.66392612e-01 3.64195764e-01 -1.17296135e+00 8.06732416e-01
8.19152832e-01 -7.94981599e-01 -7.58455038e-01 3.59669000e-01
1.11682808e+00 -5.36290586e-01 -6.12060487e-01 -1.00746881e-02
5.63165784e-01 -6.26931131e-01 7.24485576e-01 -8.46985340e-01
7.74838269e-01 1.36960074e-01 6.91823848e-03 -1.74618876e+00
-1.99234828e-01 -1.20267808e+00 -1.11100353e-01 1.05358744e+00
7.64714360e-01 -6.18441880e-01 1.07147503e+00 2.39166394e-01
-1.60463229e-01 -5.92788696e-01 -6.02056265e-01 -8.67166400e-01
5.19420564e-01 -4.57717121e-01 4.18968737e-01 8.81075859e-01
-8.90376270e-02 3.56607646e-01 -3.42295378e-01 2.29084617e-04
3.62009764e-01 -2.60733217e-01 6.37186110e-01 -7.53971279e-01
-8.73143315e-01 -3.72044683e-01 4.12310809e-01 -1.31438220e+00
3.63168389e-01 -7.91284442e-01 3.20230812e-01 -8.27449501e-01
-3.38768154e-01 -2.99618006e-01 9.48143825e-02 3.64999056e-01
-2.59472337e-02 5.49201071e-02 2.65081674e-01 1.37337863e-01
1.21870190e-01 2.52536327e-01 1.52024198e+00 5.79463840e-02
-3.06283832e-01 1.99150845e-01 -6.24559224e-01 7.60995924e-01
1.00254309e+00 -6.75085783e-01 -7.43526161e-01 -3.43921304e-01
1.10765114e-01 4.30076122e-01 -2.38075722e-02 -8.91447723e-01
1.20316781e-02 -2.98249990e-01 4.79466319e-01 -3.74540597e-01
3.42065960e-01 -1.65957600e-01 5.62865846e-02 6.74345613e-01
-1.13778448e+00 2.63457417e-01 2.78532445e-01 3.44596088e-01
9.63709354e-02 -5.57201684e-01 1.06964374e+00 -3.77970278e-01
3.93851250e-02 -1.67958345e-02 -7.26585865e-01 3.98557276e-01
4.89325285e-01 -4.89788949e-02 -2.45779883e-02 -8.61179650e-01
-9.72835422e-01 7.13731200e-02 3.11611295e-01 9.41869617e-02
5.67406595e-01 -1.29044998e+00 -8.64627898e-01 4.02741462e-01
-6.94916964e-01 3.87287438e-02 -2.74540216e-01 5.46211265e-02
-1.52825370e-01 1.81445330e-01 5.80571368e-02 -4.36239570e-01
-1.06717956e+00 -1.84319876e-02 4.66282398e-01 -5.34112573e-01
-4.07422870e-01 1.19378984e+00 -7.11988956e-02 -6.06120825e-01
3.88109118e-01 -4.44836318e-01 1.18714757e-01 -1.13588125e-01
6.12668216e-01 -3.56299691e-02 -3.14813346e-01 2.02884600e-01
3.28682840e-01 1.18507899e-01 -3.62044215e-01 -8.17289829e-01
1.33442354e+00 2.99342722e-01 3.17708611e-01 7.83222854e-01
1.01092827e+00 1.35163441e-01 -1.59201896e+00 -3.64979237e-01
-2.11950868e-01 4.13024351e-02 -3.79587084e-01 -8.31260979e-01
-7.56906807e-01 8.97204995e-01 1.34303644e-01 2.71668106e-01
6.91439629e-01 -2.87957996e-01 8.07008743e-01 5.76638520e-01
1.40943334e-01 -9.41292524e-01 5.12577236e-01 8.81454229e-01
8.52460802e-01 -5.34285963e-01 -2.36045733e-01 1.00597441e-01
-6.50549948e-01 1.07907891e+00 3.82895559e-01 -7.73509204e-01
2.63867140e-01 6.76173389e-01 1.87613219e-01 2.54159689e-01
-1.26275623e+00 4.36164320e-01 -4.23562042e-02 5.35697639e-01
6.32485628e-01 -9.59634557e-02 2.23530918e-01 1.36822864e-01
-7.55620241e-01 1.47378430e-01 7.92858481e-01 6.42810047e-01
-7.32621431e-01 -1.26432908e+00 -4.11063284e-01 5.09270012e-01
-6.44197226e-01 -2.74668604e-01 -2.24827006e-01 7.40367293e-01
-2.35619187e-01 4.96561736e-01 2.19065458e-01 -2.97187895e-01
-9.57563240e-03 6.33503973e-01 4.98939127e-01 -9.72369611e-01
-8.31185162e-01 1.59972668e-01 3.39684159e-01 -1.94695204e-01
1.19839637e-02 -7.87716329e-01 -1.45142984e+00 -1.89896449e-01
-3.37870479e-01 2.75507212e-01 5.43453574e-01 9.38329637e-01
-1.06214359e-02 7.83910990e-01 9.03546691e-01 -8.88491452e-01
-1.17334950e+00 -1.27246892e+00 -3.80327284e-01 1.22303464e-01
6.18383765e-01 5.72642386e-02 -4.87711221e-01 3.66080105e-01]
|
[11.879307746887207, 9.293447494506836]
|
8fb338cb-9fa5-4023-9b9f-dcb7bf7868fe
|
a-technique-to-jointly-estimate-depth-and
|
2305.19780
| null |
https://arxiv.org/abs/2305.19780v1
|
https://arxiv.org/pdf/2305.19780v1.pdf
|
A technique to jointly estimate depth and depth uncertainty for unmanned aerial vehicles
|
When used by autonomous vehicles for trajectory planning or obstacle avoidance, depth estimation methods need to be reliable. Therefore, estimating the quality of the depth outputs is critical. In this paper, we show how M4Depth, a state-of-the-art depth estimation method designed for unmanned aerial vehicle (UAV) applications, can be enhanced to perform joint depth and uncertainty estimation. For that, we present a solution to convert the uncertainty estimates related to parallax generated by M4Depth into uncertainty estimates related to depth, and show that it outperforms the standard probabilistic approach. Our experiments on various public datasets demonstrate that our method performs consistently, even in zero-shot transfer. Besides, our method offers a compelling value when compared to existing multi-view depth estimation methods as it performs similarly on a multi-view depth estimation benchmark despite being 2.5 times faster and causal, as opposed to other methods. The code of our method is publicly available at https://github.com/michael-fonder/M4DepthU .
|
['Marc Van Droogenbroeck', 'Michaël Fonder']
|
2023-05-31
| null | null | null | null |
['depth-aleatoric-uncertainty-estimation', 'autonomous-vehicles', 'monocular-depth-estimation', 'trajectory-planning']
|
['computer-vision', 'computer-vision', 'computer-vision', 'robots']
|
[-2.43675053e-01 1.24590941e-01 -1.59981549e-01 -3.19571495e-01
-1.14013910e+00 -1.01218688e+00 7.64269412e-01 -1.52474850e-01
-1.43815011e-01 7.67426610e-01 1.29369125e-01 -1.82865500e-01
4.77820113e-02 -1.01418233e+00 -6.73567891e-01 -6.63780332e-01
-1.45386025e-01 5.41980207e-01 6.91550732e-01 7.49617815e-02
2.98750520e-01 3.64698946e-01 -1.68831968e+00 -4.51850072e-02
6.57294571e-01 1.30768275e+00 6.34959415e-02 8.51339579e-01
4.70782727e-01 6.46435440e-01 -3.43765289e-01 -1.38535261e-01
4.59728330e-01 9.66736004e-02 -8.63501847e-01 1.46605866e-02
6.56568348e-01 -1.07038367e+00 -3.74318838e-01 8.79402459e-01
3.16940010e-01 3.14571522e-02 7.87024200e-01 -1.49560738e+00
6.94962516e-02 1.72081321e-01 -5.64177096e-01 -1.20483404e-02
4.74406362e-01 1.88733071e-01 8.68335843e-01 -8.55329633e-01
6.08293891e-01 1.10367739e+00 5.68968058e-01 3.55662465e-01
-7.40893424e-01 -4.66048628e-01 1.80522442e-01 7.38884732e-02
-1.44370210e+00 -3.26558650e-01 1.39338240e-01 -6.18795812e-01
9.32999671e-01 -1.71757713e-01 5.87422073e-01 1.05842209e+00
6.63579524e-01 8.82708311e-01 7.51221836e-01 -1.33880030e-03
3.84011924e-01 -1.87456414e-01 -3.95668626e-01 9.01115477e-01
4.16183829e-01 6.76616609e-01 -5.01935959e-01 -2.15594098e-01
4.96633619e-01 -2.96891183e-01 -3.63416225e-01 -7.23872960e-01
-1.35908639e+00 7.39863992e-01 4.22798395e-01 -5.25159061e-01
-2.57908076e-01 6.60119593e-01 1.18678883e-01 -1.69425145e-01
7.07952082e-01 1.19922020e-01 -5.39473653e-01 -4.53840315e-01
-8.15640867e-01 4.80522454e-01 6.93446338e-01 1.43786132e+00
7.69662976e-01 -1.56971484e-01 -2.66939718e-02 1.10897958e-01
5.58565021e-01 7.45449305e-01 -7.71915466e-02 -1.45460284e+00
4.33889478e-01 3.15551728e-01 4.85848576e-01 -6.18970752e-01
-2.88262576e-01 1.71637625e-01 -4.98619258e-01 7.74755657e-01
4.14503068e-01 -4.45692509e-01 -7.82671690e-01 1.42234921e+00
4.91972446e-01 2.35996157e-01 3.74972999e-01 8.11914265e-01
7.66396940e-01 7.12160826e-01 -3.72594982e-01 2.76005447e-01
1.05899644e+00 -1.04700470e+00 -3.37354571e-01 -5.56068182e-01
6.42203152e-01 -6.78197324e-01 4.29093182e-01 6.03249550e-01
-8.01523387e-01 -2.00356603e-01 -1.17087913e+00 6.58319443e-02
-1.91483513e-01 2.78431207e-01 5.50568640e-01 3.88058037e-01
-9.66934979e-01 7.24109054e-01 -1.01545763e+00 -3.36916864e-01
3.22869211e-01 1.58787742e-02 -3.82713079e-01 -2.91763604e-01
-9.77075994e-01 8.59143913e-01 4.84995127e-01 -2.20810235e-01
-1.56466746e+00 -5.86091459e-01 -1.13104856e+00 -4.69219476e-01
5.09008825e-01 -8.24615896e-01 1.61041629e+00 -1.04394920e-01
-1.54912543e+00 4.61356461e-01 -9.37676579e-02 -5.31806350e-01
6.63517773e-01 -5.71715415e-01 3.32332760e-01 2.67210960e-01
2.56595165e-01 1.17397082e+00 7.36495197e-01 -1.26244318e+00
-1.11977696e+00 -4.36154574e-01 5.09190798e-01 3.25411916e-01
3.17321450e-01 -5.38813710e-01 -5.53693771e-01 -1.25827461e-01
2.50628918e-01 -1.15689707e+00 -2.83064216e-01 3.93307477e-01
-3.01070511e-01 5.24355881e-02 8.49999845e-01 -2.35357568e-01
7.91099131e-01 -1.73435462e+00 4.22560275e-01 -1.79640979e-01
2.73780912e-01 -2.76401192e-01 2.53383964e-01 5.45035541e-01
5.89587629e-01 -1.69032142e-02 -5.58501065e-01 -4.74368632e-01
-5.06090187e-02 1.85923334e-02 -2.94813156e-01 7.95581937e-01
-2.07954794e-02 6.10184491e-01 -1.16453946e+00 -2.98495382e-01
6.99464321e-01 4.40116644e-01 -4.02215093e-01 1.90935716e-01
-4.05934334e-01 5.33713281e-01 -4.17476088e-01 9.71888602e-01
8.87258887e-01 4.00213487e-02 -1.08866431e-01 -3.42657454e-02
-3.40236902e-01 -5.25583215e-02 -9.15402830e-01 1.95790696e+00
-4.98440057e-01 8.24484289e-01 -5.43586202e-02 -7.71598443e-02
7.21830189e-01 1.12295412e-01 4.54680741e-01 6.39771447e-02
3.61129284e-01 5.08322828e-02 -5.95198631e-01 -7.49056339e-02
8.32057416e-01 2.12965712e-01 -1.92478299e-01 1.67220533e-01
2.67623295e-03 -1.09250414e+00 -1.20305521e-02 3.10102880e-01
1.17968416e+00 5.74432492e-01 4.79208291e-01 -5.02781235e-02
1.75675690e-01 2.55468518e-01 4.67826009e-01 3.89058799e-01
-4.71745223e-01 7.52537966e-01 4.93191510e-01 -2.75990367e-01
-8.18648100e-01 -1.15974402e+00 -2.72749096e-01 3.06319326e-01
5.27498186e-01 -7.07511485e-01 -7.29052782e-01 -8.52242172e-01
1.85756505e-01 7.12797284e-01 -5.90013921e-01 7.65102729e-02
2.49486670e-01 -4.09856975e-01 6.43342912e-01 6.49434865e-01
6.88228965e-01 -2.67794251e-01 -1.13602984e+00 5.21341674e-02
-2.91751623e-01 -1.48699677e+00 -3.19934748e-02 -3.55654657e-02
-8.44506919e-01 -1.20715821e+00 -6.92766786e-01 4.39909548e-02
3.15387785e-01 6.19078517e-01 1.12810075e+00 -2.44853318e-01
-2.52838731e-01 8.48763645e-01 -6.13477707e-01 -7.20314682e-01
-4.14414555e-02 -4.70299236e-02 1.80814356e-01 -6.49011075e-01
1.76816255e-01 -3.77941161e-01 -7.81025052e-01 6.05174899e-01
-7.52950609e-01 -1.02081567e-01 2.71692634e-01 4.91665035e-01
7.93452382e-01 1.54071555e-01 -4.36448529e-02 -4.99385774e-01
1.11361340e-01 -7.08383679e-01 -1.37191856e+00 -2.04926670e-01
-7.26752818e-01 -3.96173559e-02 2.54321061e-02 1.18183017e-01
-8.27055335e-01 3.54061514e-01 -4.49054874e-02 -5.10936201e-01
-2.74851650e-01 2.10880369e-01 -4.49171364e-02 -3.49232763e-01
7.06483364e-01 -1.54688224e-01 -9.26585868e-02 1.98775098e-01
4.99879241e-01 4.68148828e-01 2.58003712e-01 -3.76871318e-01
8.16302240e-01 1.02054358e+00 3.75358552e-01 -7.22580016e-01
-8.77382100e-01 -5.49350500e-01 -8.31537664e-01 -4.80441391e-01
8.29793155e-01 -1.25737870e+00 -5.17692447e-01 5.02407253e-01
-1.28875673e+00 -5.96624017e-01 -1.17260078e-03 6.99020743e-01
-9.26708341e-01 5.54866731e-01 -1.03012078e-01 -9.72036600e-01
-5.57937436e-02 -1.38288772e+00 1.57852650e+00 6.38917461e-03
2.75479071e-02 -1.00607634e+00 2.67080098e-01 1.94220096e-01
1.68332502e-01 5.45690358e-01 2.36777917e-01 2.23867893e-01
-1.15009558e+00 -2.40912840e-01 -3.28021377e-01 2.47306913e-01
3.86829115e-02 2.83093691e-01 -1.21592546e+00 -3.91074598e-01
-4.04519051e-01 -3.83861601e-01 1.05802917e+00 7.91522324e-01
9.83709216e-01 -5.12964539e-02 -4.00354594e-01 7.78469205e-01
1.52772152e+00 -8.97384435e-02 6.74906671e-01 5.91597438e-01
5.21519601e-01 6.72973156e-01 1.42823207e+00 1.00946379e+00
7.78773189e-01 6.42078698e-01 1.29307926e+00 4.58155662e-01
1.44596562e-01 -2.36909032e-01 4.63800639e-01 -3.48340167e-04
-1.42004743e-01 -5.83564281e-01 -1.01068234e+00 5.16687155e-01
-1.98835576e+00 -7.62708902e-01 -1.70333654e-01 2.39497685e+00
6.78396598e-02 6.23449944e-02 2.55027623e-03 -2.24892870e-01
1.27428889e-01 3.94231588e-01 -6.44347727e-01 3.09069101e-02
1.84914902e-01 -4.01958346e-01 1.11723888e+00 7.63029575e-01
-1.24894893e+00 1.20033610e+00 6.67749214e+00 5.74509799e-01
-5.07996202e-01 -5.23166955e-02 1.70913324e-01 -1.55227482e-01
-3.86317730e-01 2.01027006e-01 -1.11662328e+00 1.02533706e-01
8.98944676e-01 -9.41437110e-02 2.84363359e-01 1.17845452e+00
5.11147594e-03 -9.21340585e-01 -1.04852986e+00 1.06540656e+00
-3.42765674e-02 -1.16956210e+00 -1.89358354e-01 2.75970101e-01
8.79256964e-01 4.36031252e-01 1.29881576e-01 1.41090468e-01
4.81696248e-01 -8.48178208e-01 8.36728096e-01 2.93679118e-01
7.59137392e-01 -1.07909417e+00 9.97820020e-01 4.42561328e-01
-1.49401581e+00 8.64770710e-02 -7.50537038e-01 -2.47186609e-02
3.52924973e-01 8.13677132e-01 -9.24048781e-01 8.80571842e-01
9.53673899e-01 8.92103016e-01 -3.22109103e-01 1.10378253e+00
-5.72288990e-01 1.07525721e-01 -2.85223961e-01 2.59042442e-01
4.53613937e-01 -2.70795114e-02 8.28591704e-01 7.54715025e-01
7.69774079e-01 1.99272126e-01 6.15159683e-02 7.00208187e-01
1.96947604e-01 -3.95377398e-01 -1.42060220e+00 1.13942660e-01
5.35922468e-01 1.28142858e+00 -5.25167823e-01 -1.44151643e-01
-3.14009547e-01 1.03040648e+00 2.88378373e-02 2.19243020e-03
-1.14053619e+00 -2.87596017e-01 1.02961862e+00 -4.22625709e-03
2.40840971e-01 -5.74745595e-01 -2.02211980e-02 -9.48930621e-01
-3.22970748e-02 -3.45491946e-01 1.24987282e-01 -9.70139682e-01
-8.19127262e-01 8.44078720e-01 4.36710209e-01 -1.91514766e+00
-6.36760890e-01 -8.71470392e-01 -2.71088094e-01 5.92800915e-01
-1.67786729e+00 -1.09798670e+00 -9.66180623e-01 1.27798066e-01
7.67120481e-01 -3.51809501e-03 7.23238826e-01 -1.52298525e-01
-1.07122451e-01 7.05319177e-03 -3.77053060e-02 -4.04363126e-01
7.65133858e-01 -1.27620137e+00 7.87035942e-01 1.06375921e+00
-1.02717198e-01 -1.35098204e-01 7.95556962e-01 -7.61448562e-01
-1.49112988e+00 -1.35707402e+00 1.79561421e-01 -1.08775735e+00
6.02207899e-01 1.60689652e-01 -2.67864883e-01 8.43432903e-01
1.00749560e-01 2.99429428e-02 2.91435331e-01 -2.32035264e-01
-9.65439379e-02 2.50617445e-01 -1.21544421e+00 3.53231370e-01
9.45535839e-01 -2.93854833e-01 -2.52927691e-01 2.41518512e-01
8.07822883e-01 -1.01497126e+00 -9.11685288e-01 8.74727190e-01
7.40380883e-01 -1.50792456e+00 9.86881733e-01 1.99668422e-01
6.15423799e-01 -6.03227615e-01 -6.44988000e-01 -1.53154874e+00
-5.24418578e-02 -3.13585222e-01 -3.25762391e-01 8.33750844e-01
3.05305153e-01 -5.97596586e-01 7.70756483e-01 3.94345552e-01
-3.91129762e-01 -8.15420091e-01 -9.83570158e-01 -1.07528830e+00
-6.62275925e-02 -6.66201472e-01 5.27569056e-01 2.49180287e-01
-3.87862951e-01 -5.47472425e-02 -3.97788256e-01 8.25572670e-01
9.98558819e-01 -1.43405825e-01 9.97992337e-01 -1.21987653e+00
7.33598098e-02 -1.47851124e-01 -6.15997553e-01 -1.31770301e+00
2.59034455e-01 -3.85522008e-01 4.43394542e-01 -1.96136212e+00
-1.81295916e-01 -3.28474909e-01 4.56425995e-01 2.16668576e-01
2.73719579e-01 6.10891320e-02 -1.76328942e-01 4.16920036e-02
-5.20419061e-01 8.84822190e-01 1.01442325e+00 -1.09117866e-01
-5.34231542e-03 2.80112118e-01 -3.06139439e-01 1.12619889e+00
8.55313301e-01 -3.95443708e-01 -6.59532011e-01 -6.35838807e-01
1.83046818e-01 2.06306025e-01 5.11390984e-01 -1.44469595e+00
2.24621028e-01 -2.38975167e-01 3.26429486e-01 -1.21471095e+00
8.34625840e-01 -8.66427124e-01 -1.42386332e-01 2.02595398e-01
4.12521780e-01 1.14438303e-01 5.53962409e-01 9.37702239e-01
-2.68734306e-01 -3.21236134e-01 6.70886457e-01 -2.35532373e-01
-1.22597909e+00 7.52557039e-01 -5.59381425e-01 4.38505597e-02
1.36850595e+00 -1.72749385e-01 -5.44866741e-01 -6.43072188e-01
-1.25881836e-01 3.86014998e-01 8.74656975e-01 2.74279505e-01
1.01442444e+00 -1.24133801e+00 -6.19175136e-01 -1.64069489e-01
7.02339232e-01 5.38874865e-01 2.87426591e-01 6.98660672e-01
-7.85816371e-01 6.44371688e-01 -3.24166678e-02 -9.17456925e-01
-1.05450952e+00 2.52267331e-01 1.51197374e-01 2.00150058e-01
-5.81061959e-01 9.85232592e-01 2.81160563e-01 -5.86195409e-01
2.62318283e-01 -5.99121153e-01 8.71567130e-02 -5.69283068e-02
6.28940046e-01 6.74749255e-01 -5.39075918e-02 -4.63370621e-01
-5.31716764e-01 9.16117787e-01 4.29341674e-01 -5.09219646e-01
1.00248408e+00 -4.52144355e-01 7.18152374e-02 4.12611276e-01
8.07661057e-01 -3.17231156e-02 -1.66768682e+00 2.02760354e-01
-3.47594708e-01 -7.77330697e-01 5.45372248e-01 -4.65291828e-01
-7.58679867e-01 8.32908511e-01 4.87142146e-01 -1.32236645e-01
8.92512798e-01 -5.80907911e-02 6.26188993e-01 5.20653605e-01
1.17306161e+00 -7.30845869e-01 -8.90537864e-04 8.34534705e-01
8.97210181e-01 -1.68932319e+00 2.46565074e-01 -5.35069585e-01
-8.69164407e-01 1.17229652e+00 7.51101255e-01 -8.72793328e-03
7.31236994e-01 4.34707344e-01 -4.31166813e-02 -1.71925470e-01
-9.54072773e-01 -4.52569455e-01 2.37021476e-01 8.01924944e-01
-2.85787620e-02 7.94813111e-02 4.76142883e-01 8.12404081e-02
-1.23757429e-01 -6.43281415e-02 8.15720737e-01 9.89912927e-01
-6.80379808e-01 -9.18489814e-01 -3.02430600e-01 7.10242391e-02
2.65409760e-02 -1.11551046e-01 -3.12733531e-01 8.85691226e-01
-1.76426813e-01 1.00277197e+00 -9.36295092e-02 -7.30991423e-01
1.88691974e-01 -4.28331971e-01 6.61879301e-01 -6.46108687e-01
2.24362329e-01 -3.51781338e-01 2.61684179e-01 -1.12977648e+00
-3.70731622e-01 -5.74563324e-01 -1.12916231e+00 -4.60536391e-01
-3.11185688e-01 -1.62916467e-01 8.55767071e-01 6.19897604e-01
3.64564478e-01 3.57495219e-01 5.79212189e-01 -1.34801865e+00
-4.02742833e-01 -8.11983764e-01 -5.21007836e-01 -6.34929657e-01
4.58689362e-01 -1.23946309e+00 -7.71529675e-01 -3.45016241e-01]
|
[7.947906017303467, -2.2115612030029297]
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.